Kun Lian,1 Chaosheng Du,1 Yi Liu,1 Di Zhu,1 Wenjun Yan,1 Haifeng Zhang,2 Zhibo Hong,1 Peilin Liu,1,3

Lijian Zhang,1 Haifeng Pei,1 Jinglong Zhang,1 Chao Gao,1 Chao Xin,1 Hexiang Cheng,1 Lize Xiong,4 andLing Tao1

Impaired Adiponectin SignalingContributes to DisturbedCatabolism of Branched-ChainAmino Acids in Diabetic MiceDiabetes 2015;64:49–59 | DOI: 10.2337/db14-0312

The branched-chain amino acids (BCAA) accumulatedin type 2 diabetes are independent contributors toinsulin resistance. The activity of branched-chain a-ketoacid dehydrogenase (BCKD) complex, rate-limiting en-zyme in BCAA catabolism, is reduced in diabetic states,which contributes to elevated BCAA concentrations.However, the mechanisms underlying decreased BCKDactivity remain poorly understood. Here, we demon-strate that mitochondrial phosphatase 2C (PP2Cm),a newly identified BCKD phosphatase that increasesBCKD activity, was significantly downregulated in ob/oband type 2 diabetic mice. Interestingly, in adiponectin(APN) knockout (APN2/2) mice fed with a high-fat diet(HD), PP2Cm expression and BCKD activity were signif-icantly decreased, whereas BCKD kinase (BDK), whichinhibits BCKD activity, was markedly increased. Concur-rently, plasma BCAA and branched-chain a-keto acids(BCKA) were significantly elevated. APN treatment mark-edly reverted PP2Cm, BDK, BCKD activity, and BCAA andBCKA levels in HD-fed APN2/2 and diabetic animals.Additionally, increased BCKD activity caused by APNadministration was partially but significantly inhibited inPP2Cm knockout mice. Finally, APN-mediated upregula-tion of PP2Cm expression and BCKD activity were abol-ished when AMPK was inhibited. Collectively, we haveprovided the first direct evidence that APN is a novel reg-ulator of PP2Cm and systematic BCAA levels, suggestingthat targeting APN may be a pharmacological approachto ameliorating BCAA catabolism in the diabetic state.

The branched-chain amino acids (BCAA) are essentialamino acids such as leucine, isoleucine, and valine; theirhomeostasis is determined largely by catabolic activities ina number of organs including liver, muscle and adiposetissue (1–3). The first step of BCAA catabolism generatesa set of corresponding branched-chain a-keto acids(BCKA), which are irreversibly decarboxylated by thebranched-chain a-keto acid dehydrogenase (BCKD) com-plex (4). As with most nutrients, maintaining of the phys-iological level of BCAA is critical for cell metabolism andsurvival. However, many researchers have described in-creased BCAA and BCKA levels in diabetes and obesity(3,5–8). Furthermore, BCAA and their catabolites arestrongly associated with insulin resistance (9–11), andelevated BCAA contributes to the development of insulinresistance (10,12). Mechanistically, elevated BCAA levelsactivate mTOR/p70S6 kinase, resulting in an increased Iinsulin receptor substrate-1 phosphorylation, therebyinhibiting phosphatidylinositol 3-kinase. This inhibitionof phosphatidylinositol 3-kinase in turn leads to impairedinsulin signaling (13,14). It is also reported that BCAA areindependent predictors of insulin resistance, diabetes,and cardiovascular events (15–17). Therefore, it is neces-sary to determine the mechanisms of abnormal BCAAcatabolism in order to better understand their associationwith metabolic-related pathogenesis.

The BCKD complex is the rate-limiting enzyme inBCAA catabolism (4,12); regulation of BCKD activity is

1Department of Cardiology, Xijing Hospital, Fourth Military Medical University,Xi’an, China2Experiment Teaching Center, Fourth Military Medical University, Xi’an, China3Department of Cardiology, 306th Hospital of PLA, Beijing, China4Department of Anesthesiology, Xijing Hospital, Fourth Military Medical University,Xi’an, China

Corresponding author: Ling Tao, lingtao2006@gmail.com.

Received 22 February 2014 and accepted 9 July 2014.

This article contains Supplementary Data online at http://diabetes.diabetesjournals.org/lookup/suppl/doi:10.2337/db14-0312/-/DC1.

K.L., C.D., and Y.L. contributed equally to this work.

© 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, January 2015 49

METABOLISM

therefore important for maintaining the homeostasis ofsystemic BCAA and BCKA. The complex consists of threecatalytic components: a heterotetrameric (a2b2) branched-chain a-keto acid decarboxylase (E1), a homo-24 mericdihydrolipoyltransacylase (E2), and a homodimeric dihy-drolipoamide dehydrogenase (E3). The activity of BCKDcomplex is controlled by the reversible phosphorylationof its E1a subunit (Ser293) by specific BCKD kinase (BDK)and phosphatase (BDP), respectively (4,18). Phosphoryla-tion catalyzed by BDK inhibits the enzymatic activity of theBCKD complex, whereas it becomes activated when theSer293 residue is dephosphorylated by BDP. A series ofreports demonstrated that activation of BCKD was reducedin liver and adipose tissue, resulting in increased plasmaBCAA and BCKA concentrations in diabetic and obese ani-mals (6–8). Moreover, alterations of metabolism can influ-ence BCKD activity partly through changes in BDK (3),suggesting that BDP might be suppressed. The mitochon-drial phosphatase 2C (PP2Cm) is the only identified BDP(18,19), which specifically binds the BCKD complex andinduces dephosphorylation of BCKD at Ser293 in the pres-ence of BCKD substrates (18). Additionally, PP2Cm defi-ciency impairs BCAA catabolism, leading to elevated plasmaBCAA and BCKA concentrations (18). However, the rela-tionship between PP2Cm and reduced diabetic BCKD ac-tivity has not yet been investigated.

Adiponectin (APN) is an adipocytokine predominantlysynthesized in and secreted from adipose tissue. APN helpsregulating glucose and lipid metabolism (20,21) and hasvascular/cardioprotective effects (22,23). More recently,Liu et al. (24) reported that APN corrects altered muscleBCAA metabolism induced by a high-fat diet (HD). In ad-dition, several observations have revealed that obesity andtype 2 diabetes are associated with decreased plasma APNlevels (20,25,26). However, the correlation between de-creased APN levels and reduced BCKD activity has neverbeen investigated. More importantly, the underlying mo-lecular mechanisms by which APN mediates disturbedBCAA catabolism in diabetes are completely unknown.

In the current study, we used both in vitro and in vivoexperiments to identify hypoadiponectinemia as a contrib-uting factor to reduced BCKD activity in diabetes. Indiabetic mice, BCKD activity was reduced, and BCAA andBCKA levels were significantly elevated. APN treatmenteffectively reversed these pathological alterations, whichwere completely abolished by inhibition of AMPK andpartially but significantly attenuated by knockout ofPP2Cm expression. These findings lead us to concludethat impaired APN signaling is an important part of theunderlying mechanism for disturbed BCAA catabolism intype 2 diabetes.

RESEARCH DESIGN AND METHODS

Animal Care and Drug TreatmentAll experiments were performed in adherence to theNational Institutes of Health Guidelines on the Use ofLaboratory Animals and were approved by the Fourth

Military Medical University Committee on Animal Care.Male ob/ob and wild-type (WT) C57BL/6 control micewere purchased from the Department of Pathology, theFourth Military Medical University. Male APN knockout(APN2/2) mice (22) and WT C57BL/6 mice on the samebackground have previously been described. The whole-body PP2Cm knockout (PP2Cm2/2) mice (18,27) werea gift from Yibin Wang of University of California, LosAngeles (Los Angeles, CA).

Mice were rendered type 2 diabetic by the followingprocedures. Four-week-old WT C57BL/6 mice were fedwith an HD (60% of kcal from fat; Research Diets, NewBrunswick, NJ) for 6 weeks and injected with a low doseof streptozotocin (STZ) twice (28) (25 mg/g body wt i.p.;STZ in 0.05 mol/L sodium citrate, pH 4.5, once daily;Sigma, St. Louis, MO). Blood glucose and body weightwere measured daily, and a diabetic condition was con-firmed at 4 weeks after STZ injection by a nonfastingblood glucose level of $200 mg/dL. Fasting blood insulinwas measured, and intraperitoneal glucose tolerance test(IPGTT) and insulin tolerance test (ITT) were carried outin each successful model group.

In some experiments, 7-week-old APN2/2 and WT micewere randomized to receive a normal chow diet (ND) (12%of kcal from fat; control) or HD (45% of kcal from lard;Research Diets) for 4 weeks. Additionally, animals receiveddifferent treatment as follows: 1) ob/ob, HD-fed APN2/2 andPP2Cm2/2 mice received vehicle (PBS) or APN (5 mg/g bodywt; PeproTech, Rocky Hill, NJ) and 2) type 2 diabetic micereceived vehicle, APN (5 mg/g body wt), AMPK activatorAICAR (150 mg/g body wt; Sigma, St. Louis, MO), or APNin conjunction with the AMPK inhibitor compound C (CC)(20 mg/g body wt; Sigma) once daily for an additional3 days. After 12 h of fasting, animals were killed for collec-tion of tissues (liver, epididymal fat pad, and gastrocnemiusmuscle) and 0.5 mL aortic blood at 10–12 weeks of age.

Cell Culture and BCKA ChallengeAML12 mouse hepatocytes (American Type Culture Col-lection, Manassas, VA) were cultured in DMEM medium(Invitrogen, Carlsbad, CA) containing 10% fetal bovineplasma (HyClone, Waltham, MA). Experiments were car-ried out at three or four cell passages. The hepatocyteswere transfected by small interfering RNA (siRNA) andincubated in DMEM medium (Invitrogen) with an addi-tional mixture of all of the BCKA (Sigma) at 2.5 mmol/L.Each set of cells was randomized to receive treatment with10 mg/mL APN or control vehicle. After 1 h of treatment,cells and supernatants were collected for Western blottingand BCKA analysis.

RNA InterferencesiRNA constructs against AMPK a1 or AMPK a2 mRNAwere designed and purchased from Gene Pharma (Shang-hai, China). The siRNA sequences are as follows:AMPKa1, sense 59-GCCGACCCAAUGAUAUCAUTT-39, an-tisense 59-AUGAUAUCAUUGGGUCGGCTT-39; AMPKa2,sense 59-GGACAGGGAAGCCUUAAAUTT-39, antisense

50 Hypoadiponectinemia and Diabetic BCAA Catabolism Diabetes Volume 64, January 2015

59-AUUUAAGGCUUCCCUGUCCTT-39; and scrambledsiRNA, sense 59-UUCUCCGAACGUGUCACGUTT-39, anti-sense 59-ACGUGACACGUUCGGAGAATT-39.

Mouse hepatocytes were transfected with siRNA by usingLipofectamine2000 (Invitrogen) according to the manufac-turer’s instructions. Efficiency of gene knockdown was con-firmed using Western blotting 48 h after siRNA transfection.

BCAA and BCKA AnalysisPlasma and supernatant of cultured cells were collected andsubsequently stored at 280°C. Determination of BCAAconcentrations was performed in triplicate using a commer-cially available BCAA detection kit (Biovision, Milpitas, CA)per the manufacturer’s instructions. BCKA concentrationswere determined by HPLC as described by Loï et al. (29).

BCKD Enzyme Activity AssaysTissue extraction and assessment of BCKD activity wereperformed as previously described (30). BCKD complexwas concentrated from whole tissue extracts using 9%polyethylene glycol. BCKD activity was determined spec-trophotometrically by measuring the rate of NADH pro-duction resulting from the conversion of a-keto-isovalerateto isobutyryl-CoA. A unit of enzyme activity was defined as1 mmol NADH formed per minute at 30°C.

Western Blot AnalysisProteins were separated on SDS-PAGE gels, transferred topolyvinylidene fluoride membranes (Millipore, Billerica,MA), and incubated overnight at 4°C with antibodiesdirected against AMPKa (1:1,000; CST, Danvers, MA),AMPKa1 (1:1,000; CST), AMPKa2 (1:1,000; CST),phospho-AMPKa (Thr172, 1:1,000; CST), PP2Cm (1:1,000;a gift from Yibin Wang, University of California, LosAngeles), BDK (1:2,000; Abcam, Cambridge, MA), phospho-BCKD E1a (1:1,000; Bethyl Laboratories, Montgomery, TX),BCKD E1a (1:500; Santa Cruz Biotechnology, Dallas, TX),and GAPDH (1:5,000; Zhong Shan Golden Bridge Biotech-nology, Beijing, China). After washing to remove excess pri-mary antibody, blots were incubated for 1 h with horseradishperoxidase–conjugated secondary antibody. Binding wasdetected via enhanced chemiluminescence (Millipore).Films were scanned with ChemiDocXRS (Bio-Rad Labora-tory, Hercules, CA). Densitometry was performed usingLaboratory Image software.

Statistical AnalysisAll data are presented as means 6 SEM of n independentexperiments. All data (except densitometry) was subjectedto ANOVA followed by a Bonferroni correction for a posthoc t test. Densitometry was analyzed using the Kruskal-Wallis test, followed by a Dunn post hoc test. Probabilitiesof #0.05 were considered statistically significant.

RESULTS

PP2Cm Is Downregulated in ob/ob and Type 2 DiabeticMiceConsistent with previous reports, plasma BCAA and BCKAlevels were significantly increased in both ob/ob and type2 diabetic mice compared with WT mice (Fig. 1A and B).

Liver is the primary metabolic clearing house of BCKA;the reduction of BCKD activity causes increased circulat-ing BCAA and BCKA concentrations (3). In addition, im-munoreactivity of the BCKD pSer293 antibody is directlycorrelated with BCKD activity (18); we therefore exam-ined the activity of BCKD and phosphorylation of BCKDat Ser293 in liver. We found that BCKD activity was sig-nificantly reduced and phosphorylation of BCKD was sig-nificantly increased in diabetic animals compared withWT mice (Fig. 1C–E and H). Interestingly, the newly iden-tified BCKD phosphatase, PP2Cm, was markedly de-creased in diabetic mice (Fig. 1F and H), indicating thatdownregulation of PP2Cm may be involved in reducedBCKD activity in diabetes.

Considerable evidence indicates that besides liver,adipose tissue, and skeletal muscle also play an importantrole in modulating circulating BCAA homeostasis (1,2);thus, the following experiments were performed. Asshown in Fig. 1, BCKD activity was markedly downregu-lated in diabetic adipose tissue and skeletal muscle (Fig.1C–E and H). Additionally, PP2Cm protein levels weresignificantly reduced in diabetic adipose tissue (Fig. 1Fand H). To our surprise, changes of PP2Cm expressionin skeletal muscle were not observed in diabetic animals(Fig. 1F and H), indicating that skeletal muscle may havea relatively low level of PP2Cm-dependent BCKD activa-tion. Since the activity of BCKD complex is increased byPP2Cm and inhibited by BDK, we than analyzed expres-sion of BDK. Our experimental results demonstrated thatBDK expression was significantly increased in diabetictissues, including liver and adipose tissue as well as skel-etal muscle (Fig. 1G and H). These results indicate thatdecreased PP2Cm and increased BDK may both contributeto reduced BCKD activity in diabetic liver and adiposetissue, whereas increased BDK may be the primary causeof decreased BCKD activity in diabetic skeletal muscle.

APN Deficiency Contributes to Decreased BCKDActivityIt is known that APN reverts altered BCAA metabolism inmuscle (24). In order to determine whether APN contrib-utes to BCKD activity, 7-week-old WT and APN2/2 micewere fed with either ND or 45% HD for a further 4 weeks,resulting in four subgroups: ND WT, HD WT, ND APN2/2,and HD APN2/2. As shown in Fig. 2, APN2/2 micerevealed the significant increase of plasma BCAA, BCKA,and hepatic BCKD phosphorylation levels but reduction ofBCKD activity in response to HD. No response was ob-served in WT mice with the same genetic background (Fig.2A–E and Supplementary Fig. 2). More importantly, he-patic PP2Cm was markedly decreased in HD-fed APN2/2

mice (Fig. 2F and Supplementary Fig. 2). We then soughtto determine whether treatment with APN could signifi-cantly reverse altered BCKD activity. HD-fed APN2/2

mice were given recombinant APN at 5 mg/g body wtonce daily for an additional 3 days after the initial4-week HD-feeding period. As we expected, treatment

diabetes.diabetesjournals.org Lian and Associates 51

with APN completely corrected plasma BCAA and BCKA,hepatic BCKD activity, BCKD phosphorylation, and PP2Cmprotein levels (Fig. 2L–Q and T). However, plasma glucoselevels, glucose tolerance, insulin tolerance, and cholesterolconcentrations did not show a difference between APN andvehicle-treated mice (Fig. 2H–K). Collectively, these find-ings demonstrate that APN deficiency is responsible fordownregulated BCKD activity and PP2Cm expression.

Since BCKD activity is also significantly reduced inadipose tissue and skeletal muscle, we thus determinedthe effect of APN knockout upon BCKD activity in thesetissues. Compared with HD-fed WT mice, HD-fed APN2/2

mice revealed greater reduction of BCKD activity in adi-pose tissue and skeletal muscle (Fig. 2C–E and Supple-mentary Fig. 2), which were completely corrected byAPN treatment (Fig. 2N–P and T ). Interestingly, therewas no change of PP2Cm level in skeletal muscle butnot in adipose tissue from HD-fed APN2/2 mice (Fig.2F and Supplementary Fig. 2). APN treatment markedlyincreased the downregulated PP2Cm expression in adi-pose tissue (Fig. 2Q and T ). In addition, BDK expressionwas significantly increased in liver, skeletal muscle, andadipose tissue from HD-fed APN2/2 mice (Fig. 2G andSupplementary Fig. 2), which was reverted by APN treat-ment (Fig. 2R and T). Collectively, these findings demon-strate for the first time that APN deficiency causesdownregulated BCKD activity and PP2Cm expressionand upregulated BDK levels.

PP2Cm Deficiency Partially Inhibits APN-ActivatedBCKD

Data have shown that plasma APN concentration issignificantly reduced in diabetic patients and animals(20,26). Consequently, we sought to determine whetherexogenous APN administration could change BCKD activ-ity in diabetic mice. Diabetic animals were injected withvehicle or APN (5 mg/g body wt, daily) for 3 days. Asillustrated in Fig. 3, plasma glucose levels, glucose toler-ance, and insulin tolerance did not show a difference be-tween APN and vehicle-treated mice (Fig. 3A–C). As weexpected, there was significant decrease in BCAA and BCKAlevels but increase in BCKD activity in APN-treated micecompared with those only treated with vehicle (Fig. 3D–H). In addition, APN treatment significantly increasedPP2Cm protein expression in diabetic liver and adiposetissue but not in skeletal muscle (Fig. 3I). Injection withAPN markedly reduced BDK levels in diabetic animals(Fig. 3J). Taken together, these results indicate thatAPN can increase downregulated BCKD activity and ame-liorate disturbed BCAA catabolism during diabetes.

To further determine whether PP2Cm is required byAPN to mediate its effects on BCKD activity, we treatedPP2Cm2/2 mice with vehicle or APN. Interestingly, thepositive effects of APN on BCKD were reduced but notcompletely lost in PP2Cm2/2 mice. Specifically, plasmaBCAA and BCKA concentrations were significantly higher,but BCKD activity was markedly lower in PP2Cm2/2 mice

Figure 1—PP2Cm is downregulated in ob/ob and type 2 diabetic mice. A and B: Plasma BCAA and BCKA concentrations in WT, ob/ob,and type 2 diabetic mice. C and D: Analysis of BCKD activity. E: Quantification of pSer293 BCKD E1a. PP2Cm (F ) and BDK (G) protein levelsin liver, adipose tissue, and skeletal muscle from WT, ob/ob, and type 2 diabetic mice. H: Representative immunoblots of pSer293 BCKDE1a, BCKD E1a, PP2Cm, BDK, and GAPDH in WT, ob/ob, and type 2 diabetic mice. All results are presented as mean6 SEM. *Significantdifference between ob/ob or type 2 diabetic group versus WT group. *P< 0.05, **P< 0.01. n = 6–8. AT, adipose tissue; AU, arbitrary unit; L,liver; SM, skeletal muscle.

52 Hypoadiponectinemia and Diabetic BCAA Catabolism Diabetes Volume 64, January 2015

treated with APN compared with those in APN-treateddiabetic and WT control groups (Fig. 3D–H and Supple-mentary Fig. 3A–C). These data suggested that other fac-tors may also contribute the stimulatory effect of APNupon BCKD. Indeed, APN treatment markedly attenuatedthe BDK expression in PP2Cm2/2 mice compared withvehicle-treated ones (Fig. 3K). Taken together, these obser-vations support the hypothesis that APN can regulatePP2Cm (upregulating) and BDK (downregulating) in theopposite ways, thus increasing BCKD activity and amelio-rating disturbed BCAA catabolism in diabetic disease.

AMPK Is Necessary for APN-Mediated BCKDActivationConsiderable evidence exists that AMPK acts as an integratorof nutritional and hormonal signals that monitor systemicand cellular energy status (31). We then investigated whether

AMPK contributes to decreased BCKD activity in diabetes.Consistent with previous reports, phosphorylated AMPKwas markedly downregulated in type 2 diabetic mice(Fig. 4A). We treated these mice with AICAR (21)(a cell-permeable activator of AMPK [150 mg/g body wtdaily]) for 3 days. As shown in Fig. 4, AICAR treatmentresulted in a significant decrease in BCAA and BCKA con-centrations while augmenting BCKD activity and PP2Cmin diabetic mice liver (Fig. 4B–G). To determine whetheradipose tissue and skeletal muscle contribute to systemBCAA homeostasis in response to AICAR treatment, thefollowing experiments were conducted. Treatment withAICAR significantly increased the activity of BCKD in di-abetic adipose tissue and skeletal muscle (Fig. 4D and E).Moreover, AICAR injection significantly upregulatedPP2Cm protein levels in adipose tissue but did not influ-ence PP2Cm expression in skeletal muscle from diabetic

Figure 2—APN deficiency contributes to decreased BCKD activity. Plasma BCAA and BCKA (A and B) and BCKD (C and D) activity andquantification of pSer293 BCKD E1a (E), PP2Cm (F ), and BDK (G) protein levels in liver, adipose tissue, and skeletal muscle from WT andAPN2/2 mice fed with ND or HD. H: Analysis of blood glucose concentrations. I: Glucose tolerance tested by IPGTT. Insulin tolerancetested by ITT (J), blood cholesterol levels (K), and plasma BCAA and BCKA concentrations (L and M) in HD-fed APN2/2 mice after vehicleor APN treatment. Detection of BCKD activity (N and O) and quantitative results of pSer293 BCKD E1a (P), PP2Cm (Q), and BDK proteinlevel (R) testing in liver, adipose tissue, and skeletal muscle from HD-fed APN2/2 mice injected with APN. S: Representative Western blotsof pSer293 BCKD E1a, BCKD E1a, PP2Cm, BDK, and GAPDH in ND or HD-fed WT and APN2/2 mice. T: Representative immunoblots ofpSer293 BCKD E1a, BCKD E1a, PP2Cm, BDK, and GAPDH in HD-fed APN2/2 mice injected with APN or vehicle. All results are presentedas mean 6 SEM. *Significant difference between APN2/2 and WT mice fed with HD. %Significant difference between APN and vehicletreatment. * and %P < 0.05, ** and %%P < 0.01. n = 6–8. AT, adipose tissue; AU, arbitrary unit; L, liver; SM, skeletal muscle.

diabetes.diabetesjournals.org Lian and Associates 53

animals (Fig. 4G). However, AICAR caused a significantreduction of BDK in diabetic skeletal muscle, as well asin liver and adipose tissue (Fig. 4H). Altogether, theseresults suggest that AMPK may be an endogenous regu-lator of BCKD activity and PP2Cm and BDK expression.

Since AMPK activation has been recognized as a mecha-nism of action for APN (20,32), we therefore examinedwhether AMPK contributes to APN-activated BCKD in dia-betic mice. To do so, type 2 diabetic mice were treated in threegroups: vehicle, APN (5 mg/g body wt daily), and APN plusCC (a potent AMPK inhibitor [30 min before APN injection;

20 mg/g body wt daily]) for 3 days. As per our expectations,compound C (CC) completely blocked the effect of APN ondiabetic BCAA, BCKA, and BCKD activity and PP2Cm andBDK expression (Fig. 4I–O). These findings indicate thatAMPK is necessary for APN-activated BCKD in vivo.

The remote actions of AMPK could potentially havesecondary effects on BCKD. Moreover, liver has anextremely high BCKD activity compared with other tissues(3,33). Thus, in vitro experiments were performed in mousehepatocytes. BCKA-challenged mouse hepatocytes were incu-bated with vehicle, APN (10 mg/mL), APN (10 mg/mL) + CC

Figure 3—PP2Cm deficiency partially inhibits APN-activated BCKD. Analysis of blood glucose concentrations (A), glucose tolerance testedby IPGTT (B), and insulin tolerance tested by ITT in type 2 diabetic animals administered with APN or vehicle (C). BCAA and BCKAconcentrations in plasma (D and E) and BCKD activity (F and G) in liver, adipose tissue, and skeletal muscle from ob/ob, type 2 diabetic,and PP2Cm2/2 mice, each injected with APN or vehicle. H: Western blot for hepatic BCKD E1a phosphorylation in ob/ob, type 2 diabetic,and PP2Cm2/2 mice, each administered with APN. PP2Cm (I) and BDK (J) protein levels were detected by Western blot in liver, adiposetissue, and skeletal muscle from ob/ob and type 2 diabetic mice injected with APN. K: BDK levels were assessed by Western blot in liver,adipose tissue, and skeletal muscle from PP2Cm2/2 mice treated with APN. All results are presented as mean 6 SEM. %Significantdifference between APN- and vehicle-injected group. &Significant difference between APN treated ob/ob and PP2Cm2/2 group. % and&P < 0.05, %% and &&P < 0.01. n = 5–8. A, APN; AT, adipose tissue; AU, arbitrary unit; L, liver; SM, skeletal muscle; V, vehicle.

54 Hypoadiponectinemia and Diabetic BCAA Catabolism Diabetes Volume 64, January 2015

(added 30 min before APN treatment [20 pmol/mL]),or AICAR (2 pmol/mL) for 1 h. As shown in SupplementaryFig. 4, AICAR resulted in significantly reduced BCKA andphosphorylated BCKD levels and increased PP2Cm proteinexpression; in contrast, coincubation with APN and CC hadno effect on BCKA catabolism (Supplementary Fig. 4E–H).Taken together, these data directly demonstrate that AMPKcontributes to APN-activated BCKD and BCAA catabolismin mouse hepatocytes.

AMPKa2 Contributes to Total AMPKa Activity in APN-Stimulated BCKD ActivationAMPK exists as a heterotrimeric complex consisting of acatalytic a subunit and two regulatory b and g subunits.Phosphorylation of Thr172 in AMPKa is associated with

activation of both the a1 and a2 subunits of AMPK (34).Thus, we investigated the relative contributions ofAMPKa2 and AMPKa1 to total AMPKa activity in BCKDof mouse hepatocytes challenged with BCKA. Mouse hepato-cytes were transfected with AMPKa2- or AMPKa1-specificsiRNA, and AMPKa2 or AMPKa1 levels were examined byWestern blot. As depicted in Fig. 5A, siRNA transfectionreduced the levels of AMPKa2 and AMPKa1 protein by81% and 83%, respectively. Importantly, PP2Cm expressionand BCKD activity were reduced, and BCKA accumulated inthe AMPKa2 siRNA knockdown group (Fig. 5A–D).

We further examined whether AMPKa2 is involved inAPN-induced BCKD activation and PP2Cm expression.Mouse hepatocytes transfected with either AMPKa2 orAMPKa1 siRNA were challenged with BCKA and then

Figure 4—AMPK is necessary for APN-mediated BCKD activation. A: Expression of phospho-AMPKa at Thr172 and AMPKa levels in liver,adipose tissue, and skeletal muscle from type 2 diabetic mice. Left: Phospho-AMPKa, AMPKa, and GAPDH protein levels were assessedby Western blot analysis. Right: Quantification of Western blot data. BCAA and BCKA concentrations in plasma (B and C) and BCKDactivity (D and E) in liver, adipose tissue, and skeletal muscle from type 2 diabetic mice treated with AICAR. F: Immunoblotting of hepaticBCKD E1a phosphorylation levels in type 2 diabetic mice treated with AICAR. PP2Cm (G) and BDK (H) protein levels were measured byWestern blot in liver, adipose tissue, and skeletal muscle from type 2 diabetic mice treated with AICAR or vehicle. Analysis of BCAA andBCKA in plasma (I and J) and BCKD activity (K and L) in liver, adipose tissue, and skeletal muscle from type 2 diabetic animals treated withAPN or APN plus CC. M: Hepatic BCKD E1a phosphorylation levels in type 2 diabetic mice injected with APN or APN plus CC. PP2Cm (N)and BDK (O) expression was detected in type 2 diabetic animals treated with APN or APN plus CC. All results are presented as mean 6SEM. *Significant difference between type 2 diabetic and normal groups. %Significant difference between AICAR and vehicle treatment.&Significant difference between APN and APN plus CC treatment. % and &P < 0.05; **, %%, and &&P < 0.01. n = 5–8. A, APN; AT,adipose tissue; AU, arbitrary unit; CC, compound C; L, liver; SM, skeletal muscle; V, vehicle.

diabetes.diabetesjournals.org Lian and Associates 55

treated with vehicle or APN (10 mg/mL) for 1 h. As shownin Fig. 5, AMPKa2 siRNA completely blocked the effectsof APN on PP2Cm expression, BCKD activity, and BCKAlevels. This effect was not seen with scrambled siRNA orAMPKa1 siRNA transfection (Fig. 5E, F, and G). Overall,these results directly demonstrate that AMPKa2 deletion/inactivation modulates BCKD activity and that the PP2Cmregulatory function of APN is dependent on AMPKa2in vitro.

DISCUSSION

Here, we have made several important observations. First,we validated that PP2Cm was downregulated in diabeticmice, which may, at least in part, contribute to reducedBCKD activity potentially caused by an APN deficiency.Second, APN treatment reverted downregulated PP2Cmexpression and BCKD activity, as well as elevated plasmaBCAA and BCKA levels in diabetic mice. However, theincreased BCKD activity mediated by APN treatment was

partially abolished in PP2Cm2/2 mice. Third, the downreg-ulated BDK level was involved in APN-activated BCKDcomplex, especially in skeletal muscle. Last, AMPK in di-abetic mice was significantly impaired, which results in re-duced BCKD activity and accumulation of BCAA and BCKA.These effects could be reversed by APN injection. Impor-tantly, APN upregulated PP2Cm expression in hepatocytes,which depends on AMPKa2. Collectively, our studies haveestablished a novel mechanism that impaired APN signalingcontributes to diabetic BCAA catabolism.

BCKD complex is the most important regulatoryenzyme in BCAA catabolism; the activity of BCKD isregulated by a phosphorylation-dephosphorylation cycleand responsive to alterations in various metabolic con-ditions (35). Studies have reported that decreased BCKDactivity causes downregulation of BCAA catabolism in di-abetes (6,7,36). It also has demonstrated that BDK is re-sponsible for phosphorylation and inactivation of BCKDcomplex and considered a primary regulator of BCKD

Figure 5—AMPKa2 contributes to total AMPKa activity in APN-stimulated BCKD activation. A: Representative immunoblots of hepatocytelysate for pSer293 BCKD E1a, BCKD E1a, AMPKa1, AMPKa2, PP2Cm, and GAPDH after transfection with scramble, AMPKa1, or AMPKa2siRNA. The relative levels of PP2Cm (B) and pSer293 BCKD E1a (C) were quantified, and BCKA concentrations (D) were analyzed inhepatocytes transfected with scramble, AMPKa1, or AMPKa2 siRNA. PP2Cm protein (E), BCKD E1a phosphorylation (F ), and BCKA levels(G) in scramble, AMPKa1, or AMPKa2 siRNA transfected hepatocytes treated with or without APN for 1 h. All results are presented asmean 6 SEM. **Significant difference between scramble siRNA and AMPKa2 siRNA group. %Significant difference between APN- andvehicle-treated group. %P < 0.05, ** and %%P < 0.01. n = 6–12 wells. AU, arbitrary unit.

56 Hypoadiponectinemia and Diabetic BCAA Catabolism Diabetes Volume 64, January 2015

activation (6). Diabetic and obese animals showed in-creased BDK expression (6,8). Moreover, alteration ofmetabolic status can influence BCKD activity partly viachanges in BDK (3), indicating that alteration of BDPmay also be important. But no information about BDP isavailable in diabetes. Identification of a specific BDP provedelusive for many years; PP2Cm was identified as the onlyendogenous BDP until recently (18,19,27). In the currentstudy, we demonstrated for the first time that PP2Cm wassignificantly decreased in liver and adipose tissue from di-abetic animals, which may at least partially contribute toreduction of BCKD activity and elevated plasma BCKA lev-els. Altogether, these findings provide a new explanation forthe disturbed regulation of BCAA catabolism in diabetes.

Hypoadiponectinemia is commonly seen in diabetes(20,26), and APN is a critical regulator in lipid and glucosemetabolism (20). Moreover, Liu et al. (24) recently dem-onstrated that HD-fed APN2/2 mice have significantlyincreased levels of muscle BCAA, which can be correctedwith APN supplementation for 2 weeks. However, the roleof APN in systemic BCAA catabolism remains unclear. In ourstudy, APN2/2 mice fed with 45% HD for 4 weeks showedmild insulin resistance (Supplementary Fig. 2C and E) butnormal blood glucose levels (Supplementary Fig. 2A).

In addition, we demonstrated that HD-fed APN2/2

mice had significant reduced BCKD activity and PP2Cmlevels and increased BDK expression, along with highplasma BCAA and BCKA concentrations. Treatment withAPN for 3 days significantly reversed these trends, butinsulin resistance remained (Fig. 2H–J). More impor-tantly, APN treatment completely reverted the reductionof BCKD activity and PP2Cm expression and the elevationof BDK level in diabetic models. These results suggest thathypoadiponectinemia contributes to downregulated BCKDactivity and PP2Cm expression and upregulated BDK level,as well as abnormal BCAA catabolism. This supports thenotion that APN has a wide range of impacts on metab-olism (24).

Additionally, evidence by Newgard et al. (10) indicatesthat changes in BCAA levels contribute to the develop-ment of insulin resistance. An interesting point that war-rants further investigation is that the positive effect ofAPN on BCAA may subsequently contribute to the insu-lin-sensitizing effect of APN. Lastly, BCAA levels changedacutely upon short-time APN administration, whereasplasma glucose, glucose tolerance, insulin tolerance, andplasma cholesterol levels were unchanged (Figs. 2H–K and3A–C). Although the precise mechanism causing the rapidBCAA regulatory response after short-period APN adminis-tration remains unclear, this result further supports thatBCAAs may be useful biomarkers for monitoring the earlyresponse to therapeutic interventions for metabolic disease.

In the current study, we observed that APN deficiencyin the APN2/2 mice did not result in elevated BCAA levelsunder basal physiological condition. However, in review-ing numerous previous publications using this model, wefound that our observation is very consistent with

previous publications, showing that the APN2/2 mousedoes not have phenotypic changes under physiologicalcondition. However, when pathologically challenged,such as by exposure to HD (37) or myocardial ischemia(22), these animals show much severer tissue injury thanWT controls. These results indicate that although othermolecules present in APN2/2 animals are sufficient tocompensate the effect of APN under normal physiologicalconditions, APN plays an essential role in counteractingpathological stress, such as HD. Additionally, APN partlyupregulated BCKD activity and markedly downregulatedBDK expression in PP2Cm-deficient mice, suggesting thatBDK was involved in APN-activated BCKD. It has beenshown in literature that thyroid hormone and sex hor-mones regulate expression of BDK (38); our present dataalso revealed that BDK was the downstream molecular ofAPN and contributed to APN-mediated BCAA catabolism.Thus, APN may signal both BDK and PP2Cm in oppositeways to increase BCKD activity and improve BCAA catab-olism in diabetes.

AMPK has been considered a master switch inregulating glucose and lipid metabolism (39) and playsan essential role in the actions of APN (40). Nevertheless,AMPK can also modulate transcription of specific genesinvolved in energy metabolism (31). In this study, weobserved that BCKD activity was elevated by activationof AMPK both in vitro and in vivo by pharmacological orgenetic methods. Moreover, when the AMPK inhibitor CCor siRNA was applied, the effect of APN on BCKD wasvirtually abolished. These results might provide us witha better understanding of the role of AMPK in the regu-lation of metabolism.

The possibility remains that the remote role of APN-AMPK signaling in glucose and lipid metabolism couldshow secondary effects on BCKD activity in vivo; we thusperformed in vitro experiments. Because BCKD capacitymostly resides in liver (3,33), mouse hepatocytes wereused in present study. Here, mouse hepatocytes chal-lenged with BCKA and incubated with APN or AICARhelped to address the fact that APN-AMPK directly upreg-ulated BCKD activity (Supplementary Fig. 4A–H). Inter-estingly, effect of APN upon BCKA occurred prior to thesignificant upregulation of PP2Cm levels, suggesting thatAPN may improve BCKA catabolism via a signaling systemin addition to upregulation of PP2Cm. Although thedetailed molecular mechanisms cannot be addressed inthe current study, it is possible that APN may enhanceBCKD/PP2Cm interaction, thus increasing BCKD activity–independent PP2Cm expression. This interesting possibilitywill be directly investigated in our future study. To date, itremains unclear how PP2Cm expression is regulated. Pre-vious studies (27,38) and our unpublished data haverevealed that PP2Cm expression can be regulated by theavailability of nutrients as well as stress (i.e., ischemia andheart failure). Here, we demonstrated that APN was thefirst identified endogenous molecule that can upregulatePP2Cm level.

diabetes.diabetesjournals.org Lian and Associates 57

In summary, this study provides evidence that reduc-tions of APN signaling could underlie the decreased BCKDactivity observed in type 2 diabetes. Our study alsoreveals direct evidence that APN can modulate expressionof hepatic PP2Cm via an AMPKa2-dependent pathway.These new insights provide a better understanding of theunderlying regulatory mechanisms involved in diabeticBCKD activity and identify potential therapeutic targetsto mitigate BCAA catabolism in metabolic diseases.

Acknowledgments. The authors thank Dr. Xinliang Ma, Thomas JeffersonUniversity, for help with revising the manuscript. The authors also thank Dr. YibinWang and Dr. Haipeng Sun, University of California, Los Angeles, for the generoussupply of the antibody to PP2Cm and PP2Cm2/2 mice.Funding. This work was supported by the Program for National Science Fundfor Distinguished Young Scholars of China (grant no. 81225001), National KeyBasic Research Program of China (973 Program, 2013CB531204), New CenturyExcellent Talents in University (grant no. NCET-11-0870), National Science Fundsof China (grant nos. 81070676 and 81170186), Program for Changjiang Scholarsand Innovative Research Team in University (no. PCSIRT1053), and Major Scienceand Technology Project of China “Significant New Drug Development” (grant no.2012ZX09J12108-06B).Duality of Interest. No potential conflicts of interests relevant to thisarticle were reported.Author Contributions. K.L. designed methods and experiments, carriedout the laboratory experiments, and wrote the paper. C.D. performed RT-PCRanalysis and analyzed data. Y.L. analyzed data and interpreted the results. D.Z.contributed to the data of Fig. 5. W.Y. reviewed and edited the manuscript. H.Z. andZ.H. researched data and contributed to discussion. P.L., L.Z., and H.P. performeddiabetic animal model and collected the samples. J.Z., C.G., and C.X. performedBCAA and BCKA analysis. H.C. and L.X. contributed to discussion. L.T. defined theresearch theme and revised the manuscript critically. L.T. is the guarantor of thiswork and, as such, had full access to all the data in the study and takes re-sponsibility for the integrity of the data and the accuracy of the data analysis.

References1. Herman MA, She P, Peroni OD, Lynch CJ, Kahn BB. Adipose tissue

branched chain amino acid (BCAA) metabolism modulates circulating BCAA

levels. J Biol Chem 2010;285:11348–113562. Tso SC, Qi X, Gui WJ, et al. Structure-based design and mechanisms of

allosteric inhibitors for mitochondrial branched-chain a-ketoacid dehydrogenase

kinase. Proc Natl Acad Sci USA 2013;110:9728–97333. Adams SH. Emerging perspectives on essential amino acid metabolism in

obesity and the insulin-resistant state. Adv Nutr 2011;2:445–4564. Harris RA, Joshi M, Jeoung NH. Mechanisms responsible for regulation of

branched-chain amino acid catabolism. Biochem Biophys Res Commun 2004;

313:391–3965. Xu F, Tavintharan S, Sum CF, Woon K, Lim SC, Ong CN. Metabolic signature

shift in type 2 diabetes mellitus revealed by mass spectrometry-based metabolomics.

J Clin Endocrinol Metab 2013;98:E1060–E10656. Doisaki M, Katano Y, Nakano I, et al. Regulation of hepatic branched-chain

alpha-keto acid dehydrogenase kinase in a rat model for type 2 diabetes mellitus

at different stages of the disease. Biochem Biophys Res Commun 2010;393:

303–3077. Bajotto G, Murakami T, Nagasaki M, Sato Y, Shimomura Y. Decreased

enzyme activity and contents of hepatic branched-chain alpha-keto acid de-

hydrogenase complex subunits in a rat model for type 2 diabetes mellitus.

Metabolism 2009;58:1489–14958. She P, Van Horn C, Reid T, Hutson SM, Cooney RN, Lynch CJ. Obesity-

related elevations in plasma leucine are associated with alterations in enzymes

involved in branched-chain amino acid metabolism. Am J Physiol EndocrinolMetab 2007;293:E1552–E15639. Tai ES, Tan ML, Stevens RD, et al. Insulin resistance is associated with

a metabolic profile of altered protein metabolism in Chinese and Asian-Indianmen. Diabetologia 2010;53:757–76710. Newgard CB, An J, Bain JR, et al. A branched-chain amino acid-relatedmetabolic signature that differentiates obese and lean humans and contributes to

insulin resistance. Cell Metab 2009;9:311–32611. Shah SH, Crosslin DR, Haynes CS, et al. Branched-chain amino acid levelsare associated with improvement in insulin resistance with weight loss. Dia-

betologia 2012;55:321–33012. Adeva MM, Calviño J, Souto G, Donapetry C. Insulin resistance and the me-

tabolism of branched-chain amino acids in humans. Amino Acids 2012;43:171–18113. Tremblay F, Marette A. Amino acid and insulin signaling via the mTOR/p70S6 kinase pathway. A negative feedback mechanism leading to insulin resistance

in skeletal muscle cells. J Biol Chem 2001;276:38052–3806014. Tremblay F, Jacques H, Marette A. Modulation of insulin action by dietary

proteins and amino acids: role of the mammalian target of rapamycin nutrientsensing pathway. Curr Opin Clin Nutr Metab Care 2005;8:457–46215. Würtz P, Soininen P, Kangas AJ, et al. Branched-chain and aromatic aminoacids are predictors of insulin resistance in young adults. Diabetes Care 2013;36:648–65516. Wang TJ, Larson MG, Vasan RS, et al. Metabolite profiles and the risk ofdeveloping diabetes. Nat Med 2011;17:448–45317. Shah SH, Sun JL, Stevens RD, et al. Baseline metabolomic profiles predictcardiovascular events in patients at risk for coronary artery disease. Am Heart J

2012;163:844–850, e118. Lu G, Sun H, She P, et al. Protein phosphatase 2Cm is a critical regulator ofbranched-chain amino acid catabolism in mice and cultured cells. J Clin Invest

2009;119:1678–168719. Joshi M, Jeoung NH, Popov KM, Harris RA. Identification of a novel PP2C-type

mitochondrial phosphatase. Biochem Biophys Res Commun 2007;356:38–4420. Turer AT, Scherer PE. Adiponectin: mechanistic insights and clinical im-plications. Diabetologia 2012;55:2319–232621. Yamauchi T, Kamon J, Minokoshi Y, et al. Adiponectin stimulates glucoseutilization and fatty-acid oxidation by activating AMP-activated protein kinase. Nat

Med 2002;8:1288–129522. Tao L, Gao E, Jiao X, et al. Adiponectin cardioprotection after myocardialischemia/reperfusion involves the reduction of oxidative/nitrative stress. Circu-

lation 2007;115:1408–141623. Cao Y, Tao L, Yuan Y, et al. Endothelial dysfunction in adiponectin deficiency

and its mechanisms involved. J Mol Cell Cardiol 2009;46:413–41924. Liu Y, Turdi S, Park T, et al. Adiponectin corrects high-fat diet-induceddisturbances in muscle metabolomic profile and whole-body glucose homeo-

stasis. Diabetes 2013;62:743–75225. Li S, Shin HJ, Ding EL, van Dam RM. Adiponectin levels and risk of type 2

diabetes: a systematic review and meta-analysis. JAMA 2009;302:179–18826. Weyer C, Funahashi T, Tanaka S, et al. Hypoadiponectinemia in obesity and

type 2 diabetes: close association with insulin resistance and hyperinsulinemia.J Clin Endocrinol Metab 2001;86:1930–193527. Zhou M, Lu G, Gao C, Wang Y, Sun H. Tissue-specific and nutrient regu-

lation of the branched-chain a-keto acid dehydrogenase phosphatase, proteinphosphatase 2Cm (PP2Cm). J Biol Chem 2012;287:23397–2340628. Zhang M, Lv XY, Li J, Xu ZG, Chen L. The characterization of high-fat dietand multiple low-dose streptozotocin induced type 2 diabetes rat model. Exp

Diabetes Res 2008;2008:70404529. Loï C, Nakib S, Neveux N, Arnaud-Battandier F, Cynober L. Ornithine alpha-ketoglutarate metabolism in the healthy rat in the postabsorptive state. Metab-

olism 2005;54:1108–111430. Nakai N, Kobayashi R, Popov KM, Harris RA, Shimomura Y. Determination of

branched-chain alpha-keto acid dehydrogenase activity state and branched-

58 Hypoadiponectinemia and Diabetic BCAA Catabolism Diabetes Volume 64, January 2015

chain alpha-keto acid dehydrogenase kinase activity and protein in mammaliantissues. Methods Enzymol 2000;324:48–6231. Viollet B, Lantier L, Devin-Leclerc J, et al. Targeting the AMPK pathway for thetreatment of Type 2 diabetes. Front Biosci (Landmark Ed) 2009;14:3380–340032. Wu X, Motoshima H, Mahadev K, Stalker TJ, Scalia R, Goldstein BJ. In-volvement of AMP-activated protein kinase in glucose uptake stimulated by theglobular domain of adiponectin in primary rat adipocytes. Diabetes 2003;52:1355–136333. Hutson SM, Sweatt AJ, Lanoue KF. Branched-chain [corrected] amino acidmetabolism: implications for establishing safe intakes. J Nutr 2005;135(Suppl.):1557S–1564S34. Stein SC, Woods A, Jones NA, Davison MD, Carling D. The regulation ofAMP-activated protein kinase by phosphorylation. Biochem J 2000;345:437–44335. Lu J, Xie G, Jia W, Jia W. Insulin resistance and the metabolism ofbranched-chain amino acids. Front Med 2013;7:53–59

36. Kuzuya T, Katano Y, Nakano I, et al. Regulation of branched-chain aminoacid catabolism in rat models for spontaneous type 2 diabetes mellitus. BiochemBiophys Res Commun 2008;373:94–9837. Yamauchi T, Kadowaki T. Adiponectin receptor as a key player in healthylongevity and obesity-related diseases. Cell Metab 2013;17:185–19638. Lu G, Ren S, Korge P, et al. A novel mitochondrial matrix serine/threonineprotein phosphatase regulates the mitochondria permeability transition poreand is essential for cellular survival and development. Genes Dev 2007;21:784–79639. Kahn BB, Alquier T, Carling D, Hardie DG. AMP-activated protein kinase:ancient energy gauge provides clues to modern understanding of metabolism.Cell Metab 2005;1:15–2540. Lau WB, Tao L, Wang Y, Li R, Ma XL. Systemic adiponectin malfunctionas a risk factor for cardiovascular disease. Antioxid Redox Signal 2011;15:1863–1873

diabetes.diabetesjournals.org Lian and Associates 59