Smad3 Deficiency in Mice Protects Against Insulin Resistance and Obesity Induced by a High-Fat Diet.

2.9 MB·English

by  Tan, Chek Kun

ORIGINALARTICLE
fi
Smad3 De ciency in Mice Protects Against Insulin
Resistance and Obesity Induced by a High-Fat Diet
Chek Kun Tan,1 Nicolas Leuenberger,2 Ming Jie Tan,1 Yew Wai Yan,1 Yinghui Chen,1
Ravi Kambadur,1 Walter Wahli,2 and Nguan Soon Tan1
OBJECTIVE—Obesity and associated pathologies are major Obesityisprimarilycharacterizedbyincreasedfatmass
globalhealthproblems.Transforminggrowthfactor-b/Smad3sig- or white adipose tissue (WAT). WAT consists of adipo-
naling has been implicated in various metabolic processes, in- cytesspecializedinthestorageoffat(2).Inadditiontoits
cluding adipogenesis, insulin expression, and pancreatic b-cell primary function as an energy reservoir, WAT is an endo-
function. However, the systemic effects of Smad3 deficiency on crine organ that secretes adipocytokines (e.g., leptin and
adiposityandinsulinresistanceinvivoremainelusive.Thisstudy resistin) that have been shown to regulate glucose and
investigated the effects of Smad3 deficiency on whole-body glu- lipid metabolism (3). In obesity, adipose secretion of adi-
cose and lipid homeostasis and its contribution to the develop-
pocytokines is disturbed. The mechanisms that underlie
mentofobesityandtype2diabetes.
obesity-associated pathologies, such as insulin resistance,
RESEARCHDESIGNANDMETHODS—Wecomparedvarious are likely to involve communication among different or-
metabolic profiles of Smad3-knockout and wild-type mice. We gans, such as insulin-responsive skeletal muscle and WAT
also determined the mechanism by which Smad3 deficiency (4,5). It has been proposed that adipose lipid storage
affects the expression of genes involved in adipogenesis and functions to prevent peripheral lipotoxicity (5). The exces-
metabolism. Mice were then challenged with a high-fat diet to
sivelipidaccumulationinskeletalmuscleandliverleadsto
study the impact of Smad3 deficiency on the development of
insulin resistance resulting from the adverse effects of
obesityandinsulinresistance.
chroniclipotoxicityonthesetissues(5,6).Indeed,freefatty
RESULTS—Smad3-knockout mice exhibited diminished adi- acids (FFAs) can inhibit insulin activation of insulin re-
posity with improved glucose tolerance and insulin sensitivity. ceptor substrate-1–associated phosphatidylinositol-3-kinase
Chromatin immunoprecipitation assay revealed that Smad3 de-
activity in skeletal muscle (7).
ficiency increased CCAAT/enhancer-binding protein b-C/EBP Studieshaveshownthatthethreeperoxisomeproliferator–
homologous protein 10 interaction and exerted a differential
activatedreceptorisotypes(PPARa,b/d,andg)playcentral
regulation on proliferator-activated receptor b/d and prolifera-
roles in this process (8–10). PPARa activation decreases
tor-activated receptorgexpression in adipocytes.Focused gene
expression profiling revealed an altered expression of genes dyslipidemia and regulates obesity in rodents by both in-
involved in adipogenesis, lipid accumulation, and fatty acid creasing hepatic FFA oxidation and decreasing levels of
b-oxidation,indicativeofalteredadiposephysiology.Despitere- circulating triglycerides responsible for adipocyte hyper-
ducedphysicalactivitywithnomodificationinfoodintake,these trophy and hyperplasia (11,12). The transcriptional upre-
mutant mice were resistant to obesity and insulin resistance in- gulation of PPARg during adipogenesis is well studied.
ducedbyahigh-fat diet.
Adipogenic hormones, such as glucocorticoids, cyclic
CONCLUSIONS—Smad3 is a multifaceted regulator in adipose AMP, and insulin, induce a transient increase in the ex-
physiology and the pathogenesis of obesity and type 2 diabetes, pression of the transcription factors CCAAT/enhancer-
suggestingthatSmad3maybeapotentialtargetforthetreatment binding protein (C/EBP) b and d early in adipocyte
of obesity and its associated disorders. Diabetes 60:464–476, differentiation. Together they induce PPARg expression
2011 in preadipocytes, subsequently triggering full-blown adi-
pocytedifferentiation(13).PPARb/dplaysimportantfunc-
tions in adipose tissue metabolism, weight control, and
O
regulation of insulin sensitivity (14). PPARb/d protects
besity is a global medical issue by virtue of its
against weight gain, hypertriglyceridemia, and insulin re-
association with an array of metabolic abnor-
sistance in mice fed a high-fat diet (HFD) and in animals
malities, including insulin resistance, hyperten-
that are genetically predisposed to obesity (15). Thus,
sion, and hyperlipidemia, collectively termed
“metabolic syndrome” (1). Thus, the need is urgent for available information suggests that obesity and other fac-
ets of metabolic syndrome involve deregulation of signal-
elucidation of the molecular events underlying the devel-
opment of metabolic syndrome and for identification of ing pathways mediated by PPAR.
Transforming growth factor (TGF)-b1 signals through a
novel targets for disease prevention and therapy.
complexoftwomembrane-boundreceptorserine/threonine
kinases thatrecruitandphosphorylateSmad2andSmad3.
Fromthe1SchoolofBiologicalSciences,NanyangTechnologicalUniversity, Oncephosphorylated,Smad2andSmad3oligomerizewith
Singapore;and2CenterforIntegrativeGenomics,NationalResearchCenter
Smad4 and translocate to the nucleus to participate in
FrontiersinGenetics,UniversityofLausanne,Lausanne,Switzerland.
Correspondingauthor:NguanSoonTan,[email protected]. transcriptional regulation (16). TGF-b1 has been reported
Received10June2010andaccepted25October2010. to inhibit adipogenesis, although these findings were de-
DOI:10.2337/db10-0801
rived from in vitro preadipocyte models (17,18). Smad3
This article contains Supplementary Data online at http://diabetes.
diabetesjournals.org/lookup/suppl/doi:10.2337/db10-0801/-/DC1. was shown to bind C/EBP and repress its transactivation
(cid:1)2011bytheAmericanDiabetesAssociation.Readersmayusethisarticleas potential, thus abolishing the expression of PPARg2 (17).
longastheworkisproperlycited,theuseiseducationalandnotforprofit,
However,theinvivoeffectofSmad3onadiposityremains
and the work is not altered. See http://creativecommons.org/licenses/by
-nc-nd/3.0/fordetails. unclear.ElevatedexpressionandplasmalevelsofTGF-b1
464 DIABETES,VOL.60,FEBRUARY2011 diabetes.diabetesjournals.org
C.K.TANANDASSOCIATES
have been reported in WAT from obese mice and in di- from various fat pad depots was reduced by ;63% in KO
abetic patients, respectively (19,20). TGFb1/Smad3 was mice (Fig. 1C–E). Total body composition analysis by
also shown to regulate insulin gene expression and pan- quantitativenuclearmagneticresonancefurtherconfirmed
creatic b-cell function (21). Conceivably, this will have an that fat content in KO mice was reduced by ;63% (WT =
impact on energy homeostasis. However, the effects of 11.3661.51%vs.KO=4.1860.29%;P,0.05),whereasas
Smad3 on communications among insulin-responsive or- expectedfromthisobservation,theproportionofthelean
gans toward the development of diet-induced obesity and mass was consistently higher in KO mice (WT = 72.29 6
type 2 diabetes have not been examined in vivo. 1.24% vs.KO =80.01 6 0.49%; P, 0.01)(Fig.1F).Smad3-
WeshowthatSmad3-knockout(KO)micehavereduced KO epididymal WAT contained smaller adipocytes than
adiposityassociatedwithimpairedlipidaccumulationand their WT counterparts (WT = 2,206.03 6 153.08 mm2 vs.
adipogenesis. We revealed that this distinct phenotype KO = 567.49 6 34.11 mm2; P , 0.001) (Fig. 1G and H).
might be associated with altered expression of PPARg2 Fluorescence-activatedcellsortinganalysiscomparingKO
and b/d in the adipose tissue. Despite their reduced and WT WAT using Nile Red staining revealed a smaller
physical activity arising from muscle atrophy, these KO andreducednumberofadipocytesinKOmice(Fig.1Iand
mice are resistantto HFD-induced obesity. The deficiency J). No significant difference in cell proliferation in WAT
of Smad3 also confers improved glucose tolerance and and liver histology was observed for both genotypes
insulin sensitivity. (Supplementary Fig. 1D and E).
Smad3-KO mice show increased peripheral insulin
sensitivity. To uncover the systemic manifestations of
RESEARCHDESIGNANDMETHODS reduced adiposity, we measured plasma lipid and glucose
Animals.Smad3heterozygousmiceonaC57BL/6Jbackgroundwereinter- parameters in WT and KO mice. Although the plasma tri-
crossedtoproduceSmad3-KOandwild-type(WT)offspring(22).Malemiceof glyceride level was elevated in KO mice, the FFA and
thesameagewereusedinallexperiments.
glycerol levels were reduced, suggesting decreased lipol-
Hyperinsulinemic-euglycemic clamp studies. Clamp studies were per-
ysis (Fig. 2A–C). Plasma cholesterol levels remained un-
formedaspreviouslydescribed(23).
Invitroinsulin-inducedglucoseuptakeassay.Afterfasting(15h),tissues changed (Supplementary Fig. 1F). KO mice displayed
wereharvested,followedbyincubationinaKrebs-Ringerbicarbonatebuffer. hyperinsulinemia and were hypoglycemic regardless of
ThesolutionwasthenreplacedwithaKrebs-Ringerbicarbonatebuffersolution feeding status (Fig. 2D and E). KO mice exhibited signifi-
containing[3H]2-deoxyglucose(2.25mCi/mL)and[14C]mannitol(0.3mCi/mL) cantly faster reduction in blood glucose concentration af-
inthepresenceorabsenceofinsulin(12–14nmol/L).Thetissueswerequickly ter glucose administration (Fig. 2F), and they were more
frozenuntiluse.Thetissueswerelysedwith1mol/LNaOHandcentrifuged.
sensitive to insulin, as indicated by the prolongation of
Thesupernatantwasneutralizedwith1mol/LHCl,followedbymeasuringthe
radioactivity for [3H]2-deoxyglucose and [14C] mannitol by a scintillation blood glucose–lowering effects of insulin in these animals
counter.Netuptakeof[3H]2-deoxyglucoseuptakerateswerecorrectedusing (Fig. 2G).
[14C]mannitol. These findings suggested that Smad3 deficiency im-
Real-time quantitative PCR. Quantitative PCR (qPCR) was performed on proved glucose tolerance and insulin responsiveness. To
threepairedsetsofWTandKOWAT.Thesequencesoftheprimerpairsare
understandthemechanismsofimprovedinsulinsensitivity
summarized in Supplementary Table 1. qPCR was performed as previously
inKOmice,theproteinexpressionlevelsofkeymediators
described(24).
Fattyaciduptake.WTandKOadipocyteswereincubatedinmediacontaining of insulin-signaling cascades were examined. Immunoblot
1mCi/mLof[14C]palmitatecomplexedtoBSAata6.6:1molarratio.Theup- analysis of WAT and skeletal muscle revealed increased
takeandaccumulationof[14C]palmitateweremeasuredusingascintillation GLUT4, insulin receptor substrate-1, phosphorylated
proximityassay(25). FoxO1, GSK-3b, Akt, and ERK expression in Smad3-KO
Incorporationof[14C]palmitateintotriglycerides.WTandKOadipocytes mice when compared with WT mice (Fig. 2H). Next, we
were loaded with [14C]palmitate as described above.Lipids were extracted,
including a recovery standardin each sample (0.02 mCi of[3H]cholesterol), measured in vivo insulin-stimulated glucose uptake by
and analyzed by thin-layer chromatography using hexane:ethyl ether:acetic using hyperinsulinemic-euglycemic clamp studies. During
acid(80:20:1)aspreviouslydescribed(26). constant hyperinsulinemia, a higher glucose infusion rate
Fattyacidoxidationinadipocytes.Substrateoxidationwasmonitoredby wasrequiredtomaintainnormalglucoselevelsinKOmice
incubating cells with [14C]palmitate, with subsequent capture of liberated compared with WT mice (Fig. 2I and J). Todetermine the
14CO aspreviouslydescribed(25).
2 insulin-stimulated glucose uptake of skeletal muscle and
Lipolysis in adipocytes. Lipolysis was determined from freshly isolated
WAT, we assessed 2-deoxy-D-[1,2-3H]glucose (2-DG) up-
adipocytesbymeasurementofglycerolreleasedintothemediumaspreviously
described(27). take during clamp studies. Whole-body glucose utilization
Lipidextractionfromperipheralorgans.Lipidwasextractedfromtissues (peripheral insulin sensitivity) and percentage of insulin-
byaFolchextractionaspreviouslydescribed(28,29). mediated suppression of hepatic glucose production (he-
Statistical analysis. Statistical analysis was performed using two-tailed paticinsulinsensitivity)weresignificantlyincreasedinthe
Mann–Whitneytests.Valueswereexpressedasmean6SEM.P,0.05was KO mice (Fig. 2K and L). We also found a significant in-
considered statistically significant. Detailed description of various methods
crease of 2-DG uptake into the soleus, gastrocnemius
canbeobtainedfromthecorrespondingauthor.
muscles,andWATofKOmice(Fig.2M).Datafrominvitro
2-DG uptake of isolated soleus muscle and WAT were
RESULTS congruentwiththeinvivofindingsthatKOmicearemore
Smad3-KO mice exhibit reduced adiposity, insulin-sensitive than WT (Fig. 2N and O).
hypoglycemia, and hyperinsulinemia. Eight-week-old Smad3-KO mice display decreased physical activity.
Smad3-KO mice were characterized by reduced weight Next,weexploredtheimpactofSmad3deletiononwhole-
andshortbodylengthwhencomparedwithWTmice(Fig. body metabolic activity in the absence of a dietary chal-
1A and B).Toidentifythecause oftheseanomalies inKO lenge. To this end, mice were fed a standard chow diet,
mice, we measured the weight of several organs, and various metabolic parameters were monitored us-
expressed as percent body weight. The weights of the ing metabolic cages. We observed no significant differ-
liver,muscle,andbrownadiposetissueweresimilarinWT ences in food and water consumption (Fig. 3A and B), O
2
and KO mice (Supplementary Fig. 1A–C). WAT weight consumption (Fig. 3C), CO production (Fig. 3D), and
2
diabetes.diabetesjournals.org DIABETES,VOL.60,FEBRUARY2011 465
SMAD32/2MICEARERESISTANTTOOBESITY
FIG.1.ReducedadiposityinSmad3-KOmice.A–E:Meanbodyweight(A),overallmouth-to-anusbodylength(B),andrelativeweightofsub-
cutaneous(C),epididymal(D),andvisceral(E)whiteadiposefatpadsof8-week-oldWTandSmad3-KOmice.Valuesrepresentpercentageof
bodyweight;n=8/group.F:BodyfatcontentandleanmasscompositionanalysisofWTandKOmice;n=8/group.G:Representativehematoxylin–
eosin-stainedparaffin-embeddedsectionofWTandKOepididymalWAT.Scalebars,100mm.H:Meancross-sectionalareaofWTandKOadi-
pocytes(n=2,000/group).I:MeanadipocytenumberinWTandKOepididymalWAT.J:Flowcytometryanalysisofadipocytes(NileRedpositive)
andstromal/vascularcells(Hoechstpositive)inWTandKOepididymalWAT.WATwassubjectedtocollagenasedigestion.Theadipocytelayerwas
gentlyrecoveredandstainedfor5minwith10mLofNileRedstainingsolution(MolecularProbes,Eugene,OR).CellspositiveforNileRedwere
countedusingLSRIIFlowCytometerSystem(BectonDickinson,FranklinLakes,NJ).Dataarerepresentedasmean6SEM.*P,0.05,**P,0.01,
***P,0.001.(Ahigh-qualitycolorrepresentationofthisfigureisavailableintheonlineissue.)
metabolic rate (heat production) (Fig. 3E) between WT reduced adiposity in KO mice did not stem from in-
and KO mice (Supplementary Fig. 2A–D). Both genotypes creased energy expenditure arising from physical ac-
consumed carbohydrates as the main energysourcegiven tivity or increased hepatic b-oxidation.
that their respiratory exchange ratio (RER), which equals Smad3-KO mice show increased b-oxidation and
VO2/VCO2,was.0.92(Fig.3F).TheRERindicateswhether impaired lipolysis in WAT. Two factors contribute to
carbohydrates(RER=1.0)orlipids(RER=0.7)arebeing the expansion of WAT mass: increased size of existing
used to produce energy. adipocytesbecauseoffat accumulationandtheformation
KOmiceexhibitedadecreaseof;74%inbothhorizontal ofnewadipocytesthroughadipogenesis.Fatmassreflects
and rearing movements (Fig. 3G and H, Supplementary theratiobetweenlipogenesisandlipolysis/b-oxidation.To
Fig. 2E and F). Macroscopic comparison between WT and examine the role of Smad3 in adipocyte physiology, we
KO showed a ;10% reduction in muscle fiber size (Fig. 3I performed focused qPCR array on WT and KO WAT. The
and J). Real-time qPCR analysis comparing WT and KO expression of several antiadipogenic factors, such as pre-
musclerevealednodifferenceintheexpressionofgenes adipocyte factor 1 (Pref-1), cyclin D1, and C/EBP homol-
involved in b-oxidation or lipolysis (Fig. 3K). ogous protein 10 (CHOP-10), was elevated by twofold or
ThelivermaycontributetothereducedadiposityinKO more in KO adipocytes, whereas there was a reduced
mice via increased fatty acid (FA) b-oxidation. qPCR PPARg2 expression, suggesting that Smad3 deficiency
analysis revealed no difference in mRNA expression level inhibits adipogenesis (Fig. 4A). PPARg2-regulated genes,
of some genes involved in b-oxidation between WT and such as PEPCK and fatty acid synthase (FAS), were also
KOliver(SupplementaryFig.3A).Invitrototalb-oxidation downregulated(Fig.4A).TheexpressionoftheSREBP-1c,
assay also showed no significant difference in the oxi- which stimulates many genes involved in FA metabolism
dation of [14C]palmitate in KO and WT hepatocytes andpotentiatesthetranscriptionalactivityofPPARg2,was
(Supplementary Fig. 3B). These results indicatethatthe decreased in KO mice. Furthermore, the expression of
466 DIABETES,VOL.60,FEBRUARY2011 diabetes.diabetesjournals.org
C.K.TANANDASSOCIATES
FIG. 2. Smad3-KO mice develop hypoglycemia and insulin hypersensitivity. A–E: Mean plasma triglyceride (A), FFA (B), glycerol (C), blood
glucose(D),andbloodinsulin(E)inWTandKOmice.Bloodglucoseandinsulinlevelsweremeasuredat6-hfastedandfedstates.Bloodglucose
levelsweremeasuredusinganAccu-ChekAdvantageglucometerwithglucosemeasurementstrips(RocheDiagnostics,Indianapolis,IN).Serum
insulinwasmeasuredbyELISA(Millipore,Billerica,MA).Serumtotalcholesterol,triglyceride,glycerol,andFFAconcentrationsweredetermined
by quantitative enzymatic assays using L-type CHO-H (Wako Pure Chemical Industries, Osaka, Japan), serum triglyceride determination kit
(Sigma-Aldrich,St.Louis,MO),freeglyceroldeterminationkit(Sigma-Aldrich),andserum/plasmaFAandglycerolkit(Zen-Bio,ResearchTriangle
Park, NC), respectively. F: Changes in blood glucose concentration at the indicated time points after oral glucose tolerance test (OGTT).
G:ChangesinbloodglucoseconcentrationattheindicatedtimepointsafterIST.FortheOGTT,miceweregavagefedwitha2mgglucose/gbodywt
glucoseload.FortheIST,micefastedfor6hwereintraperitoneallyinjectedwith0.75mUinsulin/gbodywtusinganinsulinsyringe.OGTTandIST
wereperformedonmicefastedfor6h.Dataarerepresentedasmean6SEM,n=8/group.H:ImmunoblotanalysisofindicatedproteinsinWTand
diabetes.diabetesjournals.org DIABETES,VOL.60,FEBRUARY2011 467
SMAD32/2MICEARERESISTANTTOOBESITY
FIG.3.Smad3-KOmiceexhibitreducedphysicalactivity.A–D:Themeantotalfood(A),water(B),O (C)consumptions,andCO emission(D)
2 2
duringa24-hlightanddarkcycleinWTandKOmice.ValuesinCandDweremeasuredbyindirectcalorimetryandexpressedasaverageVO2and
VCO2perkgbodywtperhduringa24-hmonitoringsessionofalight-darkcycle.E:HeatproductionofWTandKOmicewascalculatedfromtheO2
productionandtheRERandexpressedasaveragekcal/h/kgbodywt.F:FuelconsumptionwasdeterminedfromtheratioofCO emittedtothe
2
amountofO consumed.FueltypepreferenceforcarbohydratehasanRERof1.0versusfatof0.7.GandH:Physicalactivitywasassessedbythe
2
numberofhorizontalandverticalbeambreaks,whichrepresentthetotalhorizontal(G)andrearing(H)movementsduringthe24-hmonitoring
period.I:RepresentativephotographshowingthehindlegmusclesofWTandKOmice.Smalldivisionontherulerscale=1mm.J:Representative
hematoxylin–eosin-stainedhistologicsectionoftheWTandKOquadricepsmuscles.Scalebars,100mm.K:RelativefoldchangeinmRNAlevelof
theindicatedgenesinWTandKOmusclewasanalyzedbyreal-timeqPCR.ValueswerenormalizedtoribosomalproteinL27.Resultsarerep-
resentedasfoldinductioncomparedwithWT.PrimersequencesaresummarizedinSupplementaryTable1.Valuesrepresentmean6SEM,n=
8/group.*P,0.05,**P,0.01,***P,0.001.(Ahigh-qualitycolorrepresentationofthisfigureisavailableintheonlineissue.)
diacylglycerol acyltransferase-1 and 2 was reduced by enzymesofthisprocess,weredecreasedbyapproximately
approximately two- to threefold in KO WAT. twofold in the KO WAT. Notably and in contrast with the
In contrast with lipogenesis, mobilization of fat stores reducedexpressionofPPARg2,theexpressionofPPARb/
occurspredominantlythroughthehydrolysisoftriglycerides d in the KO WAT was significantly increased. The mRNA
intoglycerolandFAs.Theexpressionofadiposetriglyceride levels of PPARb/d-regulated genes, such as uncoupling
lipase and hormone-sensitive lipase, the rate-limiting protein(UCP)2,UCP3,andacyl-CoAoxidase1(14),were
KOskeletalmuscleandWAT.b-Tubulinwasusedasloadingandtransfercontrol.Representativeblotsareshown.IandJ:Hyperinsulinemic-
euglycemicclampstudiesofSmad3-KOandWTmice.Glucoseinfusionraterequired tomaintainedeuglycemiainKOandWTmice(I).Blood
glucoselevelsduringtheclampstudy(J).K:Whole-bodyglucoseuptake(peripheralinsulinsensitivity).L:insulin-mediatedsuppressionofhe-
paticglucoseproductionrates(hepaticinsulinsensitivity).Numberrepresentspercentageofsuppressioncomparedwithbasalhepaticglucose
output.M:2-DGuptakebysoleus,gastrocnemiusmuscles,andWATofKOandWTmiceduringclampstudies.Dataarerepresentedasmean6
SEM,n=6/group.NandO:Invitroinsulin-stimulatedglucoseuptakeinskeletalmusclesandWAT.2-DGuptakeintosoleusmuscle(N)andWAT
(O)wasmeasuredfor30min.ThetissuesfromWTandKOwereincubatedintheabsence[insulin(2)]orpresenceofinsulin[insulin(+),14nmol/
L],followedbymeasurementof2-DGuptake.Netuptakeofglucosewasdeterminedbysubtractingtheamountof[14C]mannitolfromthatof2-DG.
Datarepresentmean6SEM,n=10/group.*P,0.05,**P,0.01,***P,0.001vs.WTcontrols.
468 DIABETES,VOL.60,FEBRUARY2011 diabetes.diabetesjournals.org
C.K.TANANDASSOCIATES
FIG.4.Smad3-KOadiposetissueshaveincreasedFAuptakeandb-oxidation.A:RelativefoldchangeinmRNAleveloftheindicatedgenesinWT
andKOadiposetissueasdeterminedbyqPCR.ValuesarenormalizedtoribosomalproteinL27.NormalizedvaluesfromWTmicewerearbitrarily
assignedavalueof1.Onlygeneswithtwofoldorgreaterchangeinexpressionarepresented.Valuesrepresentmean6SEM,n=3/group.Primer
sequencesaresummarizedinSupplementaryTable1.B:ImmunoblotanalysisofindicatedproteinsinWTandKOadiposetissues.b-Tubulinwas
usedasloadingandtransfercontrol.Representativeblotsfromtwomiceforeachgenotypeareshown.C–F:Therateof[14C]palmitateuptake(C),
[14C]palmitate incorporationinto triglycerides (D), lipolysis (E), and [14C]palmitate oxidation (F) at the indicated time points of adipocytes
freshlyisolatedfromtheepididymalfatpadsofWTandKOmice.Dataareexpressedasmean6SEM,n=3/group.*P,0.05,**P,0.01,***P,
0.001vs.WTcontrols.
diabetes.diabetesjournals.org DIABETES,VOL.60,FEBRUARY2011 469
SMAD32/2MICEARERESISTANTTOOBESITY
elevated in the KO WAT, suggesting an increase FA b- with anti-HDAC1 (Fig. 5B). In KO WAT, only weak ampli-
oxidation and thermogenesis (Fig. 4A). The protein levels ficationresultedwithanti-C/EBPb(Fig.5B).Theincreased
of selected genes were similarly altered (Fig. 4B). C/EBPb-CHOP-10 interactions in the KO when compared
To strengthen the observations, we determined the ef- with WT adipocytes (Fig. 5D) also diminished PPARg2
fect of Smad3 deficiency on triglyceride metabolism. We promoter activity (Fig. 5B). These observations show that
measured [14C]palmitate uptake and its conversion into Smad3 deficiency alters adipose physiology, causing
triglyceridesinWTandKOadipocytes.Weobserveda1.7- a change associated with diminished PPARg2 and in-
fold increase in [14C]palmitate uptake in KO adipocytes creased PPARb/d expression (Fig. 5E).
compared with WT (102.5 6 9.4 nmol/105 cells vs. 61.0 6 Smad3-KOmiceareresistanttoHFD-inducedobesity
8.8 nmol/105 cells, respectively; Fig. 4C). Notably, the in- andinsulinresistance.AchronicHFDtreatmentinduces
creasedFAuptakewasnotaccompaniedbyaconcomitant obesity and obesity-associated insulin resistance in mice
increased conversion into triglycerides, which were re- (32). To examine the effect of Smad3 deficiency on HFD-
duced in KO adipocytes when compared with WT (0.26 6 induced obesity and insulin resistance, we placed KO and
0.11 nmol/105 cells vs. 0.49 6 0.17 nmol/105 cells, re- WTmiceonalow-fatdiet(LFD)orHFDfor18weeks.On
spectively;Fig.4D).Wealsoassayedthelevelsofglycerol an HFD, WT mice became obese and were significantly
released into the culture medium as an indicator of lipol- heavierthanthoseonanLFD(Fig.6AandB).Incontrast,
ysis. Smad3 deficiency led to a twofold decrease in glyc- nosignificantdifferenceinbodyweightgainwasobserved
erol release (WT = 2.9 6 0.74 nmol/105 cells vs. KO = in KO mice on either diet, indicative of their resistance to
1.4560.28nmol/105cells;Fig.4E).Wealsofoundthatthe HFD-induced obesity (Fig. 6A and B). There was no dif-
oxidation of [14C]palmitate in KO adipocytes was twofold ference in the mean daily food intake per mouse, regard-
higher than that of WT (0.61 6 0.09 nmol/105 cells vs. less of diet, throughout the study (Supplementary Fig. 4A
0.33 6 0.08 nmol/105 cells, respectively; Fig. 4F). The in- and B).
cubation of adipocytes with insulin reduced palmitate As expected, HFD induced an increase in adipose mass
oxidation in both WT and KO adipocytes; however, the inallfatdepotsinWTmice(Fig.6CandD).Thisincrease
oxidation level in KO adipocytes remained 1.8-fold higher of 33% in WAT was associated with hypertrophy of adi-
than in WT (0.44 6 0.09 nmol/105 cells vs. 0.24 6 0.05 pocytes (Fig. 6E and F) when compared with LFD con-
nmol/105 cells, respectively; Fig. 4F). Altogether, these trols. No significant difference in adipocyte size was
observations show that Smad3 deficiency impairs adipo- observedbetween KO mice on either diet (Fig. 6E and F).
genesis and in parallel moderates fat accumulation by Histologic analysis of the liver from WT mice on an HFD
directing increased FA uptake to b-oxidation. showed macrovesicular hepatic steatosis as evidenced by
Repression of PPARb/d by C/EBPb is relieved on a fatty liver populated with abundant large vacuolar lipid
upregulation of CHOP-10 resulting from Smad3 droplets(Fig.6G).Wealsoobservedsmalllipiddropletsin
deficiency. C/EBPb represses the expression of PPARb/d the liver of WT mice on an LFD. Notably, the liver of KO
in keratinocytes (30). To gain insight into the regulation of mice had few or no lipid droplets and did not develop
PPARb/d in adipocytes, we performed in vivo coimmuno- hepatic steatosis, regardless of treatment (Fig. 6G). HFD
precipitation and chromatin immunoprecipitation (ChIP) induced an increase in triglyceride level in liver, skeletal
usinganti-C/EBPbantibodiesinWTandKOmice(Fig.5A). muscle, and pancreas of WT mice when compared with
ChIP and re-ChIP assays showed that in the WT WAT, LFD controls. No significant difference in tissue tri-
C/EBPb was associated with a specific C/EBP binding site glyceride level was observed between KO mice on either
at 494–485 bp upstream from the PPARb/d promoter. The diet (Fig. 6H–J, Supplementary Fig. 4C and D).
data further showed that the association results in re- KO mice exhibited a higher plasma triglyceride level
cruitment to this element of a transcriptional repressor comparedwithWTmiceonanLFD(Fig.7A).Asexpected,
complex containing histone deacetylase 1 (HDAC1) (Fig. the plasma triglyceride level was increased in WT and KO
5B), indicating a C/EBPb-mediated repression of PPARb/d. mice on an HFD, with the former reaching a level com-
Notably,theassociationofC/EBPb-HDAC1complexeswith parabletothatoftheKOmiceunderanLFD(Fig.7A).The
this binding site was reduced in KO WAT (Fig. 5B). This levelsofglycerolandFFAweresignificantlyelevatedinWT
association was further confirmed by qPCR normalized to and KO mice on an HFD (Fig. 7B and C), but remained
chromatin before immunoprecipitation (i.e., input) (Fig. lowerinKOmiceoneitherdiet.Similarly,regardlessofthe
5C). No amplification was observed using preimmune IgG dietandfeedingstatus,KOmiceconsistentlyshowedlower
or a control DNA sequence (21179 to 2894). blood glucose levels (Fig. 7D). No significant difference
Smad3 deficiency resulted in elevated expression of in plasma cholesterol level was detected (Supplementary
CHOP-10 (Fig. 4B). We examined possible involvement of Fig.4E).Inagreementwiththeseresults,KOmiceexhibited
CHOP-10 in the ability of C/EBPb to regulate PPARb/d higherinsulinlevelsunderanLFD,butincontrastwithWT
transcription. In vivo co-immunoprecipitation with anti-C/ mice, HFD treatment did not significantly increase these
EBPb followed by immunoblot with anti-CHOP-10 (Fig. levels (Fig. 7E).
5A) revealed that the interactions between C/EBPb and Glucose tolerance and insulin sensitivity tests (ISTs)
CHOP-10 are markedly enhanced in KO WAT (Fig. 5D). werecarriedouttodeterminetheeffectsofanHFDonthe
Thus, increased C/EBPb-CHOP-10 interactions removed WTandKOmice.ThebloodglucoseofWTmiceonanHFD
the inhibitory effect of C/EBPb on PPARb/d promoter, only started to decrease at 45 min after glucose adminis-
resulting in increased PPARb/d expression in KO mice tration and failed to return to the basal level even at the
(Fig. 5E). 120-min time point, indicative of impaired glucose toler-
C/EBPb is closely involved in the regulation of PPARg ance in these diet-induced obese mice. By contrast, KO
in adipocytes (31). As expected, ChIP assay involving the mice under an HFD retained the rapid glucose clearance
PPARg2 promoter showed that C/EBPb was bound to the abilityaspreviouslyobservedunderthestandarddiet(Fig.
C/EBP binding site. In contrast with the PPARb/d pro- 7F).In addition, HFD-induced obese WT mice also clearly
moter results, no amplification was detected after re-ChIP displayed obesity-associated insulin resistance (Fig. 7G),
470 DIABETES,VOL.60,FEBRUARY2011 diabetes.diabetesjournals.org
C.K.TANANDASSOCIATES
FIG.5.IncreasedC/EBPb-CHOP-10interactioninSmad3-KOadipocytes.A:Schematicoutlineofprocedureforinvivocoimmunoprecipitation,
ChIP,andre-ChIP.B:DNA-proteincomplexesfromWTorKOadipocyteswereimmunoprecipitated(ChIP)withanti-C/EBPborpreimmuneIgG.
EnrichmentoftheDNAfragmentcontainingtheC/EBPbindingsiteontheendogenousPPARb/d,PPARg2promoter,oracontrolsequenceup-
stream was evaluated by PCR. A second immunoprecipitation step (re-ChIP) was also performed with an anti-HDAC1-specific antibody (H).
I=inputchromatinbeforeimmunoprecipitation,2=negativecontrol(withoutanti-HDAC1antibody),M=DNAmarker,IB=immunoblot.C:Real-
timeqPCRwasperformedonimmunoprecipitatesofC/EBPbantibodiesandnormalizedtoinput.Dataareexpressedasmean6SEM,n=5/group.
***P,0.001.D:Westernblotanalyseswerealsoperformedonproteinsafterimmunoprecipitationwithanti-C/EBPbantibody.Proteinswere
separated by SDS-PAGE, blotted, and probed with anti-CHOP-10 primary antibody as indicated. Bands were detected by chemiluminescence.
Western blot control with b-tubulin antibody was performed using cell lysate before immunoprecipitation. E: Schematic diagram showing the
repressionofPPARb/dbyC/EBPbisrelievedonupregulationofCHOP-10resultingfromdepletionofSmad3.
diabetes.diabetesjournals.org DIABETES,VOL.60,FEBRUARY2011 471
SMAD32/2MICEARERESISTANTTOOBESITY
FIG.6.Smad3-KOmiceareresistanttoHFD-inducedobesityandhepaticsteatosis.AandB:GraphshowsmeanbodyweightofWTandKOmiceon
HFDorLFDover18weeks(A)andattheendofthe18-weekdietregimen(B).C:Representativeimagesofsubcutaneous,epididymal,andvisceral
adiposetissues,andtheliverofWTandKOmicefromthetwodietregimens.Smalldivisionontherulerscale=1mm.D:Meanweightsoftheliver
and the indicated adipose tissues of WT and KO mice after 18 weeks on HFD or LFD. Values are expressed as percentage of body weight.
E:Representativehematoxylin–eosin-stainedhistologicsectionoftheWTandKOEATafter18weeksonHFDorLFDtreatment.Scalebars,100mm.
F:Meancross-sectionalareaofadipocytes(n=400)fromtheWTandKOmiceontheindicateddietregimen.G:Representativehistologicsection(oil
redOandmethylenebluestaining)ofthelivertissues.Scalebars,100mm.Whitearrowsindicateoildroplets.H–J:Quantificationoflipidaccu-
mulationinperipheralorgans.Totaltriglycerideextractedfromfresh-frozenliver(H),skeletalmuscle(I),andpancreas(J)ofWTandKOmicefed
onHFDorLFDdietwasexpressedasamountinmilimoleper1gofprotein.Datarepresentmean6SEM,n=10/group.##P,0.01vs.WTLFD
control;*P,0.05,**P,0.01,***P,0.001.EAT,epididymaladiposetissue;SAT,subcutaneousadiposetissue;VAT,visceraladiposetissue.(Ahigh-
qualitydigitalrepresentationofthisfigureisavailableintheonlineissue.)
which was absent in KO mice. In vivo insulin-stimulated skeletal muscle of HFD-fed KO mice (Fig. 7M). Altogether,
glucose uptake by using hyperinsulinemic-euglycemic these observations indicated that KO mice are resistant to
clampstudies(Fig.7H–J)andinvitrotissue-specificinsulin- HFD-induced obesity and numerous obesity-linked pathol-
stimulated 2-DG uptake (Fig. 7K and L) indicated that KO ogies.
miceweremoretoleranttoHFD-inducedperipheralinsulin
resistance compared with WT mice. In concordance, the DISCUSSION
protein expression levels of indicated key mediators of Obesity increases risk for type 2 diabetes, hypertension,
insulin-signaling cascades were elevated in the WAT and and cardiovascular diseases. Accordingly, much research
472 DIABETES,VOL.60,FEBRUARY2011 diabetes.diabetesjournals.org
C.K.TANANDASSOCIATES
FIG.7.Smad3deficiencyamelioratesHFD-inducedinsulinresistanceandglucoseintolerance.A–E:Themeanplasmatriglyceride(A),FFA(B),
glycerol(C),bloodglucose(D),andbloodinsulin(E)inWTandKOmiceonHFDorLFD.Bloodglucoseandinsulinlevelsweremeasuredat6-h
fasted and fed states. F: Changes in blood glucose concentration at the indicated time points after OGTT. G: Changes in blood glucose
diabetes.diabetesjournals.org DIABETES,VOL.60,FEBRUARY2011 473