Heart- and muscle-derived signaling system dependent · PDF fileHeart- and muscle-derived signaling system dependent on MED13 and Wingless controls obesity in Drosophila Ji-Hoon Lee,

Heart- and muscle-derived signaling system dependenton MED13 and Wingless controls obesity in DrosophilaJi-Hoon Lee, Rhonda Bassel-Duby, and Eric N. Olson1

Department of Molecular Biology, University of Texas Southwestern Medical Center, Dallas, TX 75390-9148

Contributed by Eric N. Olson, May 23, 2014 (sent for review March 27, 2014)

Obesity develops in response to an imbalance of energy homeo-stasis and whole-body metabolism. Muscle plays a central role inthe control of energy homeostasis through consumption of energyand signaling to adipose tissue. We reported previously that MED13,a subunit of the Mediator complex, acts in the heart to control obe-sity in mice. To further explore the generality and mechanistic basisof this observation, we investigated the potential influence ofMED13 expression in heart and muscle on the susceptibility ofDrosophila to obesity. Here, we show that heart/muscle-specificknockdown of MED13 or MED12, another Mediator subunit,increases susceptibility to obesity in adult flies. To identify possiblemuscle-secreted obesity regulators, we performed an RNAi-basedgenetic screen of 150 genes that encode secreted proteins andfound that Wingless inhibition also caused obesity. Consistentwith these findings, muscle-specific inhibition of Armadillo, thedownstream transcriptional effector of the Wingless pathway,also evoked an obese phenotype in flies. Epistasis experimentsfurther demonstrated that Wingless functions downstream ofMED13 within a muscle-regulatory pathway. Together, thesefindings reveal an intertissue signaling system in which Wing-less acts as an effector of MED13 in heart and muscle and sug-gest that Wingless-mediated cross-talk between striated muscleand adipose tissue controls obesity in Drosophila. This signalingsystem appears to represent an ancestral mechanism for thecontrol of systemic energy homeostasis.

Skuld | kohtalo | myokines | metabolic syndrome

Obesity is a systemic disorder caused by an energy imbalancein which energy input exceeds energy utilization, resulting

in accumulation of excess body fat. Muscle plays a central role insystemic energy homeostasis by consuming nutrients and sig-naling in an endocrine manner to other tissues (1–3). Thus, therehas been intense interest in identifying secreted factors frommuscle that modulate the function of adipose tissue.Previously, we reported that cardiac deletion of MED13,

a subunit of the Mediator complex, increases susceptibility toobesity in mice whereas cardiac overexpression of MED13 con-fers a lean phenotype (4), revealing an unforeseen function ofthe heart as a systemic regulator of energy homeostasis. TheMediator is a conserved multisubunit complex that mediates theinteraction between RNA polymerase II and transcription fac-tors and therefore governs transcription in all eukaryotes (5).MED13 and MED12 are among four subunits of the auxiliarykinase module, which confers additional regulatory functionsto the Mediator complex (6). Expression profiling of yeastmutants or gene-depleted Drosophila cell lines revealed a closecorrelation between the gene-expression programs controlled byMED13 and MED12, suggesting their concerted actions in generegulation (7, 8).Drosophila provides a powerful model system for the genetic

analysis of obesity (9–11). The processes that regulate energyhomeostasis, such as energy storage and mobilization of fat inadipose tissue of the fat body, as well as the genetic pathwaysgoverning such functions, are conserved in flies (12). Perturba-tions of such pathways in flies have been shown to generatephenotypes relevant to human diseases, including obesity (13–15).

Furthermore, Drosophila serves as a model system to under-stand systemic effects of interorgan cross-talk via secretedmolecules (3, 16).Here, we show that Drosophila muscle modulates obesity

through the function of MED13 in the context of the Wingless(Wg) signaling pathway. In Drosophila, MED13 and MED12are encoded by skuld (skd) and kohtalo (kto), respectively (17).Muscle-specific knockdown of skd or kto increases fat body massand triglyceride accumulation in adult flies. We describe a geneticscreen for muscle-secreted obesity regulators, which revealed Wgas a muscle-derived suppressor of obesity. Similarly, inhibition ofArmadillo (Arm), the Drosophila β-catenin ortholog and tran-scriptional effector of Wg signaling, suppresses obesity. We alsoidentify functional relationships between skd and the Wg pathwayin which a skd-null mutation dominantly enhances the musclephenotype resulting from arm knockdown, and wg acts down-stream of skd in the regulation of fat accumulation. Our findingsindicate that Wg acts as an effector of MED13 function in muscleto suppress obesity in Drosophila.

ResultsLoss of MED13 Function in Muscle Increases Susceptibility to Obesityin Drosophila. Based on our observation that cardiac deletion ofMED13 confers an obese phenotype and that cardiac-specificoverexpression of MED13 prevents obesity in mice (4), we ex-amined whether muscle expression of MED13, encoded by skd,regulates fat accumulation in Drosophila. We performed RNAi-mediated knockdown experiments using the UAS/Gal4 systemand expressed UAS-RNAi targeting skd mRNA with the Mef2-Gal4 driver, which directs the expression of UAS constructs insomatic, cardiac, and visceral muscle tissues (18–21). By 3 wkof age, we observed increased abdominal fat bodies in adultMef2>skd RNAi flies (Fig. 1A). Lipid droplets in the fat bodycells were also enlarged as seen by Nile Red stain (Fig. 1B).Consistently, total triglyceride amounts were significantly increased

Significance

Obesity is a major health epidemic and develops as a result ofimbalanced energy homeostasis. Previously, we reported thatcardiac expression of MED13, a subunit of the Mediator com-plex, controlled systemic energy homeostasis in mice such thatincreased or decreased expression of MED13 caused leannessor obesity, respectively. Here, we report that MED13 also actswithin muscle of Drosophila to control obesity. The secretedpeptide Wingless acts as a downstream effector of MED13 tomediate cross-talk with adipose tissue and suppress obesity.Our work reveals a conserved signaling system in muscle inwhich MED13 and Wingless act as key controllers of obesity.

Author contributions: J.-H.L., R.B.-D., and E.N.O. designed research; J.-H.L. performed re-search; J.-H.L., R.B.-D., and E.N.O. analyzed data; and J.-H.L., R.B.-D., and E.N.O. wrotethe paper.

The authors declare no conflict of interest.1To whom correspondence should be addressed. E-mail: [email protected].

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1409427111/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1409427111 PNAS | July 1, 2014 | vol. 111 | no. 26 | 9491–9496

CELL

BIOLO

GY

in those flies (Fig. 1C). Using a different fly muscle-specificdriver, Mhc-Gal4 (22), to target skd mRNA, we observed thatMhc>skd RNAi flies also displayed increased total triglycerides(Fig. 1D).To test whether perturbation of MED13 function specifically

in the fly heart modulates total body triglycerides, skd was knockeddown using the heart-specific Tin-Gal4 (23, 24). Tin>skd RNAiflies also showed a significant increase in total triglycerides(Fig. 1E). We also tested whether muscle-specific knockdownof kto, which encodes MED12, affected fat accumulation.Indeed, flies with RNAi knockdown of kto in muscle or heartusing Mef2-Gal4, Mhc-Gal4, or Tin-Gal4 displayed increasedtotal body triglyceride levels similar to the effect of skdknockdown (Fig. 1 C–E).Resistance to starvation is a characteristic of obese flies (25).

We tested whether skd or kto knockdown in muscle conferredresistance to starvation. Three-week-old flies with Mef2-Gal4driven knockdown of skd or kto survived substantially longer

under starvation conditions of 1% agar (Fig. 1F). However,their overall lifespans under normal conditions did not change(Fig. 1G).To confirm that skd or kto knockdown using the RNAi lines

indeed inhibited their expression, we tested the effects in the flyeye, where skd or kto loss-of-function has been shown to preventphotoreceptor differentiation (26), resulting in small eye phe-notypes in adults (27). We expressed the RNAi lines using theeye-specific eyeless-Gal4 and GMR-Gal4 and observed small eyephenotypes from both skd and kto knockdown flies, similar to thedocumented loss-of-function mutant eye phenotype (Fig. S1),confirming effective RNAi knockdown.We tested whether the obese phenotype associated with skd or

kto knockdown in muscle was age-dependent. Flies with muscle-specific knockdown of skd or kto showed no obvious changesin triglycerides 2 wk after eclosion, but a substantial increasewas observed by 3 wk and 4 wk of age. By 5 wk of age, fataccumulation reached a maximum level irrespective of geno-type (Fig. 2A).We next tested whether flies with muscle-specific skd or kto

knockdown were more susceptible to triglyceride accumulationon a high-fat diet, which contains coconut oil as the source ofsaturated fat (28). Newly eclosed flies with Mef2-Gal4–drivenknockdown of skd or kto were fed normal food for 12 d and thenmoved to either fresh normal food or high-fat food containing30% coconut oil. Three days later, the flies fed high-fat foodincreased triglycerides substantially compared with the controlflies under the same growth condition whereas those fed normalfood did not (Fig. 2B).

An RNAi Screen Identifies Muscle-Secreted Proteins ControllingObesity. We hypothesized that the obese phenotype resultingfrom muscle-specific knockdown of skd or kto was mediatedby extracellular factors secreted by muscle cells. To test thishypothesis, we performed an RNAi-based genetic screen usingMef2-Gal4 to identify muscle-secreted proteins that, when in-hibited, caused an obese phenotype, similar to what was ob-served with skd or kto knockdown in muscle. We analyzed 182RNAi lines targeting 150 genes that encode secreted proteins.

Fig. 1. MED13 expression in muscle regulates obesity. (A) Abdominal fatbodies of 3-wk-old adult females expressing either skd RNAi or gfp (control)with Mef2-Gal4. (B) Confocal images of the abdominal fat body fromMef2>skd RNAi or Mef2>gfp (control) flies stained with Nile Red (red) andPhalloidin (green). (Scale bar: 50 μm.) (C–E) Relative triglyceride amounts ofadult females with muscle-specific knockdown of skd, kto, or luciferase(control) using Mef2-Gal4 (C), Mhc-Gal4 (D), or Tin-Gal4 (E). Error bar, SEM;*P < 0.05; **P < 0.01; ***P < 0.001. (F) Survival curves showing resistance ofadult females with indicated genotypes under starvation conditions. Flieswere maintained under normal conditions for 3 wk and then moved to 1%agar. Mef2>skd RNAi (median survival, 108 h, n = 120); Mef2>kto RNAi(median survival, 96 h, n = 80); Control, Oregon R (median survival, 72 h, n =72). (G) Survival curves showing overall lifespan of females with indicatedgenotypes. Mef2>skd RNAi (median survival, 81 d, n = 177); Mef2>kto RNAi(median survival, 81 d, n = 149); Control, Oregon R (median survival, 81 d,n = 109).

Fig. 2. Flies with MED13 or MED12 knockdown display increased suscepti-bility to obesity. (A) Age-dependent changes in total triglyceride levels.Mef2-Gal4 was used to drive knockdown of skd or kto. Control 1, Mef2-Gal4alone; control 2, Mef2>gfp. (B) Effects of high-fat diet on total triglycerideamounts in flies with the indicated genotypes. Twelve-day-old femalesgrown in normal food were transferred to normal (Left) or high-fat (Right)food and maintained for 3 d. With normal food, Mef2>skd RNAi andMef2>kto RNAi increased triglycerides by 22% and 23% on average, re-spectively, but the differences were statistically insignificant (P > 0.05). Withhigh-fat food, Mef2>skd RNAi and Mef2>kto RNAi caused an increase oftriglycerides in flies to 57% and 33% on average, respectively. Controlharbors Mef2>gfp. Error bar, SEM; NS, not significant; *P < 0.05; **P < 0.01.

9492 | www.pnas.org/cgi/doi/10.1073/pnas.1409427111 Lee et al.

Relative total triglyceride levels of these flies at 4 wk of age weremeasured in comparison with the average of six controls (TableS1). Because skd knockdown using either Mef2-Gal4 or Mhc-Gal4 caused the fat accumulation phenotype (Fig. 1 C and D),we performed the knockdown screen again using the same set ofRNAi lines expressed with Mhc-Gal4 (Table S2). We identifiedsix RNAi lines that increased total triglyceride levels greater than60% as a result of knockdown with both Mef2-Gal4 and Mhc-Gal4 (Table 1).

Wg Signaling in Muscle Suppresses Obesity. We identified wg fromthe genetic screen described above (Table 1) and confirmed theeffect of wg using two independent RNAi lines. Flies expressingeither wg RNAi with Mef2-Gal4 showed increased abdominal fatbody mass (Fig. 3A). The lipid droplets therein were enlarged(Fig. 3B), and total triglyceride amounts of the flies were alsoincreased (Fig. 3C). To confirm the effect of wg knockdown inmuscle on total triglycerides and to test the heart-specific effectof wg knockdown, we used Mhc-Gal4 and Tin-Gal4, respectively,to express wg RNAi and found that knockdown of wg with eitherof the drivers resulted in a substantial increase of body tri-glyceride content (Fig. 3 D and E).Consistent with the results from wg knockdown experiments,

wg overexpression in muscle decreased fat accumulation. Twoindependent UAS-wg cDNA lines were tested with Mef2-Gal4.These flies showed a decrease of abdominal fat body mass andlipid droplet size (Fig. 3 A and B). Total triglyceride levels inthose flies were also decreased significantly (Fig. 3C). Reducedfat accumulation was also observed in flies overexpressing wgwith Mhc-Gal4 or Tin-Gal4 (Fig. 3 D and E).To investigate whether autonomous Wg signaling activity in

muscle has a role in regulating obesity, we tested the effect on fataccumulation of arm, encoding the transcriptional effector of Wgsignaling, in muscle. The expression of arm RNAi using Mef2-Gal4 resulted in increased abdominal fat bodies containing en-larged lipid droplets (Fig. 3 A and B). Total triglyceride levelswere also increased by muscle-specific knockdown of arm (Fig.3C). In addition, using Mhc-Gal4 and Tin-Gal4 to knock downarm, we observed an increase in total triglyceride levels (Fig. 3 Dand E). Conversely, muscle-specific overexpression of arm eitheras a wild-type (armS2) or a constitutively active (armS10) formdecreased abdominal fat body mass, lipid droplets in the fatbody, and total triglyceride amounts (Fig. 3).

Functional Interaction Between skd and Arm in Muscle. To test thefunctional interaction of the Wg pathway and MED13 in muscle,we performed genetic-interaction experiments between armand skd. Flies with arm knockdown using Mef2-Gal4 are viablewithout morphological defects. skdT606 is a null allele that ishomozygous lethal (24), but skdT606 heterozygotes are viablewithout morphological defects. However, arm knockdown usingMef2-Gal4 in a skdT606 heterozygous (skd−/+) backgroundcaused complete lethality, which was fully penetrant. To better

understand the nature of this lethality, the somatic musclestructure of the embryos was examined. Embryos expressingluciferase RNAi in muscle in a skd+/+ or skd−/+ backgroundmaintained intact muscle patterns (Fig. 4 A and B). A portionof embryos expressing arm RNAi in muscle in a skd+/+ back-ground had intact muscle patterns (Fig. 4C) whereas otherembryos with the same genotype displayed patterning defectsin the somatic musculature as exemplified by defects in lateraltransverse (LT) and dorsal muscles (Fig. 4C′). However, allobserved embryos expressing arm RNAi using Mef2-Gal4 inthe skd−/+ background had patterning defects in their somaticmusculature (Fig. 4 D and D′).

The Epistatic Relationship of Wg and MED13 in Muscle for ObesityControl. Our finding that overexpression of wg in muscle causeda lean phenotype (Fig. 3) whereas skd knockdown in musclecaused an obese phenotype (Fig. 1), as well as the functionalinteraction between skd and arm (Fig. 4), raised the possibilitythat wg and skd act within a linear pathway in muscle to regulateobesity. To address this possibility, we tested the epistatic re-lationship between wg and skd. Flies expressing either skd RNAior wg cDNA or both with Mef2-Gal4 were grown for 4–4.5 wk,and their total triglyceride levels were compared. Total tri-glyceride levels of flies with both skd knockdown and wg over-expression were significantly decreased compared with those offlies with skd knockdown, but indistinguishable from those offlies with wg overexpression alone (Fig. 5A). We also performedthe same experiments using Mhc-Gal4 and Tin-Gal4 and obtainedconsistent results (Fig. 5 B and C). These data strongly supportthe conclusion that muscle-secreted Wg acts as a downstreameffector of skd function in muscle to suppress fat deposition inthe fat body.Finally, we tested the effect of Wg activation in the fat body.

When wg was overexpressed using the fat body-specific Dcg-Gal4(29, 30), it caused lethality at the pupal stage and severe re-duction of fat body mass in the abdominal region of third instarlarvae (Fig. 5D). These findings support a model in whichmuscle-secreted Wg acts on the fat body to inhibit obesity (Fig. 6and Discussion).

DiscussionOur results reveal a role of muscle in systemic regulation ofobesity via the function of MED13 in Drosophila. We per-formed a genetic screen and identified muscle-secreted obesity-regulating factors, including Wg, and demonstrated that Wgsignaling in muscle is necessary and sufficient to suppressobesity. Furthermore, we showed that a skd-null mutation domi-nantly enhances the arm phenotype in muscle and that wg isepistatic to skd, suggesting that Wg is a downstream effector ofMED13 in muscle.Our results reveal that MED13 in Drosophila muscle functions

to suppress obesity based on several criteria, such as histology,measurement of whole-body triglycerides, tolerance to starvation

Table 1. Six RNAi lines that increase triglyceride amounts >60% from Mef2-Gal4 and Mhc-Gal4screens

Gene name Annotation symbol BDSC no. Mef2 > RNAi* Mhc > RNAi*

Diptericin B CG10794 28975 1.84 2.10Angiotensin converting enzyme CG8827 36749 1.76 1.93SIfamide CG33527 29428 1.74 2.26Wingless CG4889 33902 1.69 1.72Insulin-like peptide 4 CG6736 33682 1.65 2.13Unpaired 3 CG33542 32859 1.63 1.63

BDSC, Bloomington Drosophila Stock Center.*Indicated values are relative triglyceride amounts (triglyceride/protein) compared with controls.

Lee et al. PNAS | July 1, 2014 | vol. 111 | no. 26 | 9493

CELL

BIOLO

GY

stress, and susceptibility to high-fat diet. Similarly, muscle-specificknockdown of MED12 also increased fat accumulation, suggestingthat MED12 and MED13 function similarly in the control of fatdeposition in Drosophila. Our finding that MED12 and MED13modulate energy homeostasis adds to a growing number ofexamples in which components of the kinase module of theMediator complex influence metabolic signaling on an organ-ismal level. For example, the other two components of the kinasemodule, Cyclin-dependent kinase 8 and Cyclin C, have also beenreported as negative regulators of fat accumulation in flies andmice (31). Our finding that the activity of MED13 in cardiac

muscle regulates fat accumulation in Drosophila is consistentwith our earlier observation with mice (4) and suggests that thefunction of cardiac MED13 in systemic regulation of fat storagerepresents an ancestral mechanism conserved in metazoans.Although it seems most likely that the effect of MED13 onobesity is mediated by overall changes in metabolism, it is alsoconceivable that changes in feeding behavior contribute to theobesity phenotypes we observed.Knockdown of MED12 and MED13 using drivers that are

active specifically in the heart using Tin-Gal4 or generally inall muscles using Mef2-Gal4 or Mhc-Gal4 commonly evoked

Fig. 3. The Wg signal in muscle regulates obesity.(A) Abdominal fat bodies of adult females expressingluciferase (luc) RNAi (control), wg cDNA, wg RNAi,armS10 cDNA (constitutively active), armS2 cDNA (wild-type), or arm RNAi with Mef2-Gal4. (B) Confocalimages of adult abdominal fat bodies stained withNile Red (red) and Phalloidin (green). Genotypes areas indicated above (A). (Scale bar: 20 μm.) (C–E)Effects of muscle-specific knockdown or over-expression of Wg or Arm on relative triglycerideamounts in adult females using Mef2-Gal4 (C ),Mhc-Gal4 (D), or Tin-Gal4 (E ). Control was lucif-erase RNAi. Error bar, SEM; *P < 0.05; **P < 0.01;***P < 0.001.

Fig. 4. Genetic interaction between MED13 andArm in muscle. Embryos at St. 16 were immuno-stained with anti-Mhc antibody and shown laterallywith the orientation of dorsal up and anterior right.skd− is skdT606. Embryos in A, B, and C display nor-mal embryonic musculature whereas embryos in C′,D, and D′ have defects in their musculature, some ofwhich are indicated with red asterisks for dorsalmuscle defects and dotted boxes highlighting nor-mal (A, B, and C) or abnormal (C′, D, and D′) pat-terns of LT muscles 1–4. LT, lateral transverse; SBM,segment border muscle.

9494 | www.pnas.org/cgi/doi/10.1073/pnas.1409427111 Lee et al.

comparable obesity phenotypes. Thus, we conclude that MED12and MED13 can control metabolic signaling from the heart,consistent with our prior conclusions regarding the functionsof MED13 in the mouse heart. However, these Gal4 drivers donot enable us to reach conclusions regarding the specific roleof somatic or visceral muscle in this signaling process becauseMhc-Gal4 and Mef2-Gal4 are active in diverse muscle-cell types.Given that MED12 and MED13 are ubiquitously expressed, itis possible that they also act in nonmuscle tissues to regulatemetabolic homeostasis.We hypothesized that muscle-secreted factors mediate the

function of MED13 in Drosophilamuscle to suppress systemic fataccumulation. To identify such factors, we screened for muscle-secreted obesity-regulating proteins using two different muscledrivers, Mef2-Gal4 and Mhc-Gal4. We identified six genes thatincreased fat accumulation of flies in both screens by >60%,including the genes encoding (i) an antimicrobial peptide, Dip-tericin B (32); (ii) a Drosophila homolog of Angiotensin con-verting enzyme (33); (iii) a G protein-coupled receptor ligandSIFamide (34); (iv) one of seven Drosophila Insulin/IGF homo-logs, Insulin-like peptide 4 (35); (v) a JAK/STAT signaling li-gand, Unpaired 3 (36); and (vi) Wg. Interestingly, it has beenshown recently that MED13 and MED12 are required for theexpression of Diptericin B in response to Immune Deficiency(IMD) pathway activation (8), suggestive of additional regula-tory functions of MED13 and the genes identified from ourscreens beyond obesity control.We demonstrated that Wg and its autonomous signaling ac-

tivity, controlled by Arm, in muscle are necessary and sufficientfor systemic regulation of obesity in vivo. Previously, the corre-lation between obesity and the expression of genes involved inthe Wnt signaling pathway in heart has been raised from tran-scriptome analyses using heart biopsies from obese patients (37).Similarly, correlations between obesity and differential expres-sion of genes for Wnt signaling, as well as genes for insulinsensitivity and myogenic capacity, were also found in skeletal-muscle samples from obese rats (38). These findings suggest thatWg signaling activity in muscle serves as an intrinsic rheostat forobesity control.Muscle-specific arm knockdown caused partial-patterning defects

in the embryonic musculature, and a skd-null allele dominantlyenhanced this phenotype to complete lethality. Given the centralrole of Arm in Wg target gene expression, our findings are con-sistent with the established function of wg in the development ofmesoderm and the embryonic musculature (39). Our findings re-veal a close functional connection between MED13 and Arm,suggestive of the role of MED13 in Wg target gene expression. In

fact, in the developing Drosophila eye and wing, MED13 andMED12 are essential for Wg target gene expression, and theMED13/MED12 complex physically interacts with Pygopus,a component of the Wg transcriptional complex (40). Further-more, MED12 hypomorphic mutant mice are embryonic lethalwith impaired expression of Wnt targets (41). Therefore, ourgenetic interaction data along with these previous reportssuggest that MED13 is a general component of the canonicalWg/Wnt pathway.Our epistasis experiments indicate that muscle-secreted Wg

functions downstream of MED13 in muscle to suppress obesity.Because both wg and arm in muscle are crucial for obesity reg-ulation, one function of muscle-secreted Wg might be to acton muscle. Accordingly, the nonautonomous function of Wg tosuppress obesity may occur through autonomous Wg signal ac-tivity in muscle. However, if MED13 functions at the level oftranscriptional control of Wg target genes and the sole functionof muscle-secreted Wg ligand is to activate the Wg signal “in”muscle, Wg should be upstream of MED13, which is contrary toour epistasis studies. Based on our data, it stands to reason thatmuscle-secreted Wg should also act directly on a tissue otherthan muscle for its nonautonomous effect. If so, which tissuemay be the target? Ectopic expression of Wg using a fat body-specific Dcg-Gal4 decreased larval abdominal fat body mass,which demonstrates the role of Wg signaling in the fat body forfat-mass regulation. Similarly, in mammals, autonomous activationof the Wnt pathway in adipose tissue decreases fat mass. Wnt sig-naling blocks mammalian adipogenesis in vitro (42), and, in mice,activation of the canonical Wnt pathway in adipocytes by ectopicexpression of Wnt10b, a Wnt ligand, inhibits obesity (43, 44). Fur-thermore, autonomous activation of the Wnt pathway in adiposeprogenitors with constitutively active β-catenin expression decreasesfat mass (45). Therefore, the reduced fat mass in Dcg > wg larvaeindicates that autonomous Wg signaling activity in the fat bodyserves as a regulator of fat mass. Considered together with our datashowing that muscle-secreted Wg contributes to nonautonomousregulation of adiposity in vivo, we conclude that muscle serves asa source of Wg to regulate adiposity by modulating Wg signalingactivity in fat body. However, we cannot rule out the possibility thatthe systemic effect of Wg from muscle is mediated through an al-ternative tissue, such as nervous system (46).Wg acts on short- and long-range targets. Wg is highly hy-

drophobic and has been shown to diffuse through the extracel-lular space and act on long-range targets by associating withsolubilizing molecules such as lipoprotein particles and SecretedWg-interacting molecule (47, 48). Furthermore, Wnt-1 has beenidentified in serum, and decreased Wnt-1 levels in serum cor-relate with premature myocardial infarction and metabolic syn-drome (49), suggesting that Wg may act on remote organs as anendocrine factor. Therefore, we propose a model in whichmuscle-secreted Wg is a downstream effector of MED13 andacts both to activate the signal in muscle and to act on the fatbody ultimately to achieve systemic inhibition of obesity.

Fig. 5. Wg functions downstream of MED13 in muscle to regulate obesity.(A–C) Epistatic relationship between wg and skd shown with relative tri-glyceride amounts of 4- to 4.5-wk-old adult females with the expression ofwg cDNA or skd RNAi or both in muscle using Mef2-Gal4 (A), Mhc-Gal4 (B),or Tin-Gal4 (C). Error bar, SEM; NS, not significant; **P < 0.01; ***P < 0.001.(D) Images of third instar larvae with Wg overexpression using fat body-specific Dcg-Gal4. Control was Dcg-Gal4 alone. Arrows indicate abdominalregion where fat bodies are severely reduced.

Fig. 6. A model of muscle-derived signaling via MED13 and Wg for obesitycontrol. Expression of MED13 or Wg in muscle suppresses fat accumulationin the fat body, and Wg acts downstream of MED13. Muscle-secreted Wgactivates the Wg signal in both muscle and fat body.

Lee et al. PNAS | July 1, 2014 | vol. 111 | no. 26 | 9495

CELL

BIOLO

GY

Materials and MethodsFly Stocks and Genetics. Fly stocks used are described in SI Materials andMethods. Flies were maintained in vials with Cornmeal-Molasses-Yeast me-dium except for experiments using high-fat food, which was 30% (wt/vol)coconut oil (MP Biomedicals) added to normal food, as described (28). De-tailed procedures of fly genetics are available in SI Materials and Methods.

Triglyceride Assay. Triglyceride assay was performed as described with mod-ifications (30) and is available in SI Materials and Methods.

Immunohistochemistry. Antibodies used were anti-Mhc (a gift of Bruce Paterson,National Cancer Institute, Bethesda) and anti-GFP (Torrey Pines Laboratories).Embryos were immunostained as described (50).

Nile Red Staining. To visualize adult abdominal fat bodies, a vertical incisionwas made along the ventral abdomen of female flies. Other internal organswere removed and fixed in 4% paraformaldehyde for 15 min. Images weretaken using a Zeiss AxioCam. Fixed abdomens were stained with Nile Red(Fluka) at a concentration of 10 μg/mL and Alexa 633 Phalloidin (Life

Technology) overnight, washed four times with PBS for 15 min each, andmounted in VECTASHIELD mounting medium (Vector Laboratories).

Statistics. Prism 6 (GraphPad Software) was used for statistical analyses andgraphical presentations.

ACKNOWLEDGMENTS. We thank Jonathan Graff [University of Texas South-western Medical Center (UT Southwestern)], Michael Buszczak (UT South-western), Bum-Kyu Lee (UT Austin), Aaron Johnson (University of ColoradoDenver), Jin Seo (Rogers State University), and members of the E.N.O.laboratory for helpful discussion; Dylan Tennison and Evelyn Tennison fortechnical support; Jose Cabrera for graphics; Jessica Treisman (New YorkUniversity), Bruce Paterson (National Cancer Institute), Jonathan Graff (Uni-versity of Texas Southwestern Medical Center), and Janice Fischer (Universityof Texas at Austin) for providing reagents; and the Transgenic RNAi Projectat Harvard Medical School [National Institutes of Health (NIH)/National In-stitute of General Medical Sciences Grant R01-GM084947] for providingtransgenic RNAi fly stocks. This work was supported by NIH Grants HL-077439, HL-111665, HL-093039, DK-099653, and U01-HL-100401, grants fromthe Cancer Prevention and Research Institute of Texas, and Robert A. WelchFoundation Grant 1-0025 (to E.N.O.).

1. Pedersen BK, Febbraio MA (2012) Muscles, exercise and obesity: Skeletal muscle asa secretory organ. Nat Rev Endocrinol 8(8):457–465.

2. Bassel-Duby R, Olson EN (2006) Signaling pathways in skeletal muscle remodeling.Annu Rev Biochem 75:19–37.

3. Owusu-Ansah E, Song W, Perrimon N (2013) Muscle mitohormesis promotes longevityvia systemic repression of insulin signaling. Cell 155(3):699–712.

4. Grueter CE, et al. (2012) A cardiac microRNA governs systemic energy homeostasis byregulation of MED13. Cell 149(3):671–683.

5. Boube M, Joulia L, Cribbs DL, Bourbon HM (2002) Evidence for a mediator of RNApolymerase II transcriptional regulation conserved from yeast to man. Cell 110(2):143–151.

6. Taatjes DJ (2010) The human Mediator complex: A versatile, genome-wide regulatorof transcription. Trends Biochem Sci 35(6):315–322.

7. van de Peppel J, et al. (2005) Mediator expression profiling epistasis reveals a signaltransduction pathway with antagonistic submodules and highly specific downstreamtargets. Mol Cell 19(4):511–522.

8. Kuuluvainen E, et al. (2014) Cyclin-dependent kinase 8 module expression profilingreveals requirement of Mediator subunits 12 and 13 for transcription of Serpent-dependent innate immunity genes in Drosophila. J Biol Chem 289(23):16252–16261.

9. Pospisilik JA, et al. (2010) Drosophila genome-wide obesity screen reveals hedgehogas a determinant of brown versus white adipose cell fate. Cell 140(1):148–160.

10. Baumbach J, et al. (2014) A Drosophila in vivo screen identifies store-operated calciumentry as a key regulator of adiposity. Cell Metab 19(2):331–343.

11. Reis T, Van Gilst MR, Hariharan IK (2010) A buoyancy-based screen of Drosophilalarvae for fat-storage mutants reveals a role for Sir2 in coupling fat storage to nu-trient availability. PLoS Genet 6(11):e1001206.

12. Baker KD, Thummel CS (2007) Diabetic larvae and obese flies-emerging studies ofmetabolism in Drosophila. Cell Metab 6(4):257–266.

13. Trinh I, Boulianne GL (2013) Modeling obesity and its associated disorders in Dro-sophila. Physiology (Bethesda) 28(2):117–124.

14. Oldham S (2011) Obesity and nutrient sensing TOR pathway in flies and vertebrates:Functional conservation of genetic mechanisms. Trends Endocrinol Metab 22(2):45–52.

15. Diop SB, Bodmer R (2012) Drosophila as a model to study the genetic mechanisms ofobesity-associated heart dysfunction. J Cell Mol Med 16(5):966–971.

16. Rajan A, Perrimon N (2012) Drosophila cytokine unpaired 2 regulates physiologicalhomeostasis by remotely controlling insulin secretion. Cell 151(1):123–137.

17. Janody F, Martirosyan Z, Benlali A, Treisman JE (2003) Two subunits of the Dro-sophila mediator complex act together to control cell affinity. Development 130(16):3691–3701.

18. Ranganayakulu G, Schulz RA, Olson EN (1996) Wingless signaling induces nautilusexpression in the ventral mesoderm of the Drosophila embryo. Dev Biol 176(1):143–148.

19. Brand AH, Perrimon N (1993) Targeted gene expression as a means of altering cellfates and generating dominant phenotypes. Development 118(2):401–415.

20. Ni JQ, et al. (2008) Vector and parameters for targeted transgenic RNA interference inDrosophila melanogaster. Nat Methods 5(1):49–51.

21. Dietzl G, et al. (2007) A genome-wide transgenic RNAi library for conditional geneinactivation in Drosophila. Nature 448(7150):151–156.

22. Schuster CM, Davis GW, Fetter RD, Goodman CS (1996) Genetic dissection of structuraland functional components of synaptic plasticity. I. Fasciclin II controls synaptic sta-bilization and growth. Neuron 17(4):641–654.

23. Johnson AN, Mokalled MH, Haden TN, Olson EN (2011) JAK/Stat signaling regulatesheart precursor diversification in Drosophila. Development 138(21):4627–4638.

24. Lo PC, Frasch M (2001) A role for the COUP-TF-related gene seven-up in the di-versification of cardioblast identities in the dorsal vessel of Drosophila. Mech Dev104(1-2):49–60.

25. Liu J, et al. (2012) Synphilin-1 alters metabolic homeostasis in a novel Drosophilaobesity model. Int J Obes (Lond) 36(12):1529–1536.

26. Treisman J (2001) Drosophila homologues of the transcriptional coactivation complexsubunits TRAP240 and TRAP230 are required for identical processes in eye-antennaldisc development. Development 128(4):603–615.

27. Loncle N, et al. (2007) Distinct roles for Mediator Cdk8 module subunits in Drosophiladevelopment. EMBO J 26(4):1045–1054.

28. Birse RT, et al. (2010) High-fat-diet-induced obesity and heart dysfunction are regu-lated by the TOR pathway in Drosophila. Cell Metab 12(5):533–544.

29. Hoshizaki DK, et al. (1994) Embryonic fat-cell lineage in Drosophila melanogaster.Development 120(9):2489–2499.

30. Suh JM, et al. (2006) Hedgehog signaling plays a conserved role in inhibiting fatformation. Cell Metab 3(1):25–34.

31. Zhao X, et al. (2012) Regulation of lipogenesis by cyclin-dependent kinase 8-mediatedcontrol of SREBP-1. J Clin Invest 122(7):2417–2427.

32. Lee JH, et al. (2001) Diptericin-like protein: An immune response gene regulated bythe anti-bacterial gene induction pathway in Drosophila. Gene 271(2):233–238.

33. Tatei K, Cai H, Ip YT, Levine M (1995) Race: A Drosophila homologue of the angio-tensin converting enzyme. Mech Dev 51(2-3):157–168.

34. Jørgensen LM, Hauser F, Cazzamali G, Williamson M, Grimmelikhuijzen CJ (2006)Molecular identification of the first SIFamide receptor. Biochem Biophys Res Commun340(2):696–701.

35. Garofalo RS (2002) Genetic analysis of insulin signaling in Drosophila. Trends Endo-crinol Metab 13(4):156–162.

36. Wright VM, Vogt KL, Smythe E, Zeidler MP (2011) Differential activities of the Dro-sophila JAK/STAT pathway ligands Upd, Upd2 and Upd3. Cell Signal 23(5):920–927.

37. Philip-Couderc P, et al. (2004) Uncomplicated human obesity is associated witha specific cardiac transcriptome: Involvement of the Wnt pathway. FASEB J 18(13):1539–1540.

38. Zhou D, Strakovsky RS, Zhang X, Pan YX (2012) The skeletal muscle Wnt pathway maymodulate insulin resistance and muscle development in a diet-induced obese ratmodel. Obesity (Silver Spring) 20(8):1577–1584.

39. Baylies MK, Martinez Arias A, Bate M (1995) wingless is required for the formation ofa subset of muscle founder cells during Drosophila embryogenesis. Development121(11):3829–3837.

40. Carrera I, Janody F, Leeds N, Duveau F, Treisman JE (2008) Pygopus activates Winglesstarget gene transcription through the mediator complex subunits Med12 and Med13.Proc Natl Acad Sci USA 105(18):6644–6649.

41. Rocha PP, Scholze M, Bleiss W, Schrewe H (2010) Med12 is essential for early mousedevelopment and for canonical Wnt and Wnt/PCP signaling. Development 137(16):2723–2731.

42. Ross SE, et al. (2000) Inhibition of adipogenesis by Wnt signaling. Science 289(5481):950–953.

43. Wright WS, et al. (2007) Wnt10b inhibits obesity in ob/ob and agouti mice. Diabetes56(2):295–303.

44. Longo KA, et al. (2004) Wnt10b inhibits development of white and brown adiposetissues. J Biol Chem 279(34):35503–35509.

45. Zeve D, et al. (2012) Wnt signaling activation in adipose progenitors promotes insulin-independent muscle glucose uptake. Cell Metab 15(4):492–504.

46. Benzler J, et al. (2013) Hypothalamic WNT signalling is impaired during obesity andreinstated by leptin treatment in male mice. Endocrinology 154(12):4737–4745.

47. Panáková D, Sprong H, Marois E, Thiele C, Eaton S (2005) Lipoprotein particles arerequired for Hedgehog and Wingless signalling. Nature 435(7038):58–65.

48. Mulligan KA, et al. (2012) Secreted Wingless-interacting molecule (Swim) promoteslong-range signaling by maintaining Wingless solubility. Proc Natl Acad Sci USA109(2):370–377.

49. Goliasch G, et al. (2012) Premature myocardial infarction is associated with low serumlevels of Wnt-1. Atherosclerosis 222(1):251–256.

50. Johnson AN, Mokalled MH, Valera JM, Poss KD, Olson EN (2013) Post-transcriptionalregulation of myotube elongation and myogenesis by Hoi Polloi. Development140(17):3645–3656.

9496 | www.pnas.org/cgi/doi/10.1073/pnas.1409427111 Lee et al.