Download - Title Running Title Authors - Diabetes · 2014. 7. 28. · 3 38 Abstract 39 The branched-chain amino acids (BCAA) accumulated in type 2 diabetes mellitus 40 are independent contributors

1

Title 1

Impaired Adiponectin Signaling Contributes to Disturbed Catabolism of 2

Branched-Chain Amino Acids in Diabetic Mouse 3

Running Title 4

Hypoadiponectinemia and Diabetic BCAA Catabolism 5

Authors 6

Kun Lian1, Chaosheng Du

1, Yi Liu

1, Di Zhu

1, Wenjun Yan

1, Haifeng Zhang

2, Zhibo 7

Hong1, Peilin Liu

1,4, Lijian Zhang

1, Haifeng Pei

1, Jinglong Zhang

1, Chao Gao

1, Chao 8

Xin1, Hexiang Cheng

1, Lize Xiong

3 and Ling Tao

1* 9

Affiliations 10

1Department of Cardiology, Xijing Hospital, The Fourth Military Medical University, 11

15 Changlexi Road, Xi’an, 710032, China; 12

2Experiment Teaching Center, The Fourth Military Medical University, 169 Changlexi 13

Road, Xi’an, 710032, China; 14

3Department of Anesthesiology, Xijing Hospital, The Fourth Military Medical Uni-15

versity, 15 Changlexi Road, Xi’an, 710032, China; 16

4Department of Cardiology, The 306th Hospital of PLA, 9 Anxiangbeili Street, Bei-17

jing, 100101, China. 18

Contact Information 19

of 50 Diabetes

Diabetes Publish Ahead of Print, published online July 28, 2014

2

Ling Tao, MD, PhD 20

Department of Cardiology 21

Xijing Hospital 22

The Fourth Military Medical University 23

15 Changlexi Road, 24

Xi’an, 710032, China; 25

E-mail: [email protected]; 26

Tel: 86-29-84771024, 86-29-84775183; 27

Fax: 86-29-84771024. 28

Additional Footnotes 29

Kun Lian, Chaosheng Du and Yi Liu contributed equally to this work. 30

Word Count 31

4411 32

Number of Figures 33

5 34

Number of Supplemental Figures 35

4 36

37

of 50Diabetes

*

3

Abstract 38

The branched-chain amino acids (BCAA) accumulated in type 2 diabetes mellitus 39

are independent contributors to insulin resistance. The activity of branched-chain 40

α-keto acid dehydrogenase (BCKD) complex, rate-limiting enzyme in BCAA catabo-41

lism, is reduced in diabetic states, which contributes to elevated BCAA concentrations. 42

However, the mechanisms underlying decreased BCKD activity remain poorly under-43

stood. Here we demonstrate that mitochondrial phosphatase 2C (PP2Cm), a newly 44

identified BCKD phosphatase which increases BCKD activity, was significantly 45

down-regulated in ob/ob and type 2 diabetic mice. Interestingly, in adiponectin 46

knockout (APN-/-

) mice fed with high-fat-diet (HD), PP2Cm expression and BCKD 47

activity were significantly decreased, whereas BCKD kinase (BDK) which inhibits 48

BCKD activity was markedly increased. Concurrently, plasma BCAA and 49

branched-chain α-keto acids (BCKA) were significantly elevated. APN treatment 50

markedly reverted PP2Cm, BDK, BCKD activity, BCAA and BCKA levels in HD-fed 51

APN-/-

and diabetic animals. Additionally, increased BCKD activity caused by APN 52

administration was partially but significantly inhibited in PP2Cm knockout mice. Fi-53

nally, APN-mediated up-regulation of PP2Cm expression and BCKD activity were 54

abolished when adenosine monophosphate-activated protein kinase (AMPK) was in-55

hibited. Collectively, we have provided the first direct evidence that APN is a novel 56

regulator of PP2Cm and systematic BCAA levels, suggesting that targeting APN may 57

be a pharmacological approach to ameliorating BCAA catabolism in the diabetic state. 58

Key Words: BCAA; BCKD; Diabetes; APN; AMPK; PP2Cm; BDK 59

of 50 Diabetes

4

Introduction 60

The branched-chain amino acids (BCAA) are essential amino acids such as leucine, 61

isoleucine, and valine; their homeostasis is determined largely by catabolic activities 62

in a number of organs including liver, muscle and adipose tissue (1-3). The first step 63

of BCAA catabolism generates a set of corresponding branched-chain α-keto acids 64

(BCKA), which are irreversibly decarboxylated by the branched-chain α-keto acid 65

dehydrogenase (BCKD) complex (4). As with most nutrients, maintaining of physio-66

logical level of BCAA is critical for cell metabolism and survival. However, many 67

researchers have been described that increased BCAA and BCKA levels in diabetes 68

and obesity (3; 5-8). Furthermore, BCAA and their catabolites are strongly associated 69

with insulin resistance (9-11); and elevated BCAA contributes to the development of 70

insulin resistance (10; 12). Mechanistically, elevated BCAA levels activate 71

mTOR/p70S6 kinase, resulting in an increased IRS-1 phosphorylation thereby inhib-72

iting PI3 kinase. This inhibition of PI3K in turn leads to impaired insulin signaling (13; 73

14). It is also reported that BCAA are independent predictors of insulin resistance, di-74

abetes and cardiovascular events (15-17). Therefore, it is necessary to determine the 75

mechanisms of abnormal BCAA catabolism in order to better understand their associ-76

ation with metabolic-related pathogenesis. 77

The BCKD complex is the rate-limiting enzyme in BCAA catabolism (4; 12), reg-78

ulation of BCKD activity is therefore important for maintaining the homeostasis of 79

systemic BCAA and BCKA. The complex consists of three catalytic components: a 80

heterotetrameric (α2β2) branched-chain α-keto acid decarboxylase (E1), a homo-24 81

of 50Diabetes

5

meric dihydrolipoyltransacylase (E2), and a homodimeric dihydrolipoamide dehy-82

drogenase (E3). The activity of BCKD complex is controlled by the reversible phos-83

phorylation of its E1a subunit（Ser293） by specific BCKD kinase (BDK) and 84

phosphatase (BDP), respectively (4; 18). Phosphorylation catalyzed by BDK inhibits 85

the enzymatic activity of the BCKD complex, whereas it becomes activated when the 86

Ser293 residue is dephosphorylated by BDP. A series of reports have demonstrated 87

that activation of BCKD was reduced in liver and adipose tissue, resulting in in-88

creased plasma BCAA and BCKA concentrations in diabetic and obese animals (6-8). 89

Moreover, alterations of metabolism can influence BCKD activity partly through 90

changes in BDK (3), suggesting that BDP might be suppressed. The mitochondrial 91

phosphatase 2C (PP2Cm) is the only identified BDP (18; 19), which specifically binds 92

the BCKD complex and induces dephosphorylation of BCKD at Ser293 in the pres-93

ence of BCKD substrates (18). Additionally, PP2Cm deficiency impairs BCAA catab-94

olism, leading to elevated plasma BCAA and BCKA concentrations (18). However, 95

the relationship between PP2Cm and reduced diabetic BCKD activity has not yet been 96

investigated. 97

Adiponectin (APN) is an adipocytokine predominantly synthesized in and secreted 98

from adipose tissue. APN helps regulating glucose and lipid metabolism (20; 21) and 99

has vascular/cardioprotective effects (22; 23). More recently, Liu and colleagues (24) 100

have reported that APN corrects altered muscle BCAA metabolism induced by a high 101

fat diet (HD). In addition, several observations have revealed that obesity and type 2 102

diabetes are associated with decreased plasma APN levels (20; 25; 26). However, the 103

of 50 Diabetes

6

correlation between decreased APN levels and reduced BCKD activity has never been 104

investigated. More importantly, the underlying molecular mechanisms by which APN 105

mediates disturbed BCAA catabolism in diabetes are completely unknown. 106

In the present study, we employed both in vitro and in vivo experiments to identify 107

hypoadiponectinemia as a contributing factor to reduced BCKD activity in diabetes. 108

In diabetic mice, BCKD activity was reduced, and BCAA and BCKA levels were sig-109

nificantly elevated. APN treatment effectively reversed these pathological alterations, 110

which were completely abolished by inhibition of adenosine monophos-111

phate-activated protein kinase (P-AMPK) and partially but significantly attenuated by 112

knockout of PP2Cm expression. These findings lead us to conclude that impaired 113

APN signaling is an important part of the underlying mechanism for disturbed BCAA 114

catabolism in type 2 diabetes mellitus. 115

116

of 50Diabetes

7

Research Design and Methods 117

Animal Care and Drug Treatment. All experiments were performed in adherence to 118

the National Institutes of Health Guidelines on the Use of Laboratory Animals, and 119

were approved by the Fourth Military Medical University Committee on Animal Care. 120

Male ob/ob and wild type (WT) C57BL/6 control mice were purchased from the De-121

partment of Pathology, The Fourth Military Medical University. Male APN knockout 122

(APN-/-

) mice (22) and WT C57BL/6 mice on the same background have been previ-123

ously described. The whole-body PP2Cm knockout (PP2Cm-/-

) mice (18; 27) were 124

gifted by professor Yibin Wang of UCLA (Los Angeles, CA). 125

Mice were rendered type 2 diabetic by the following procedures. Four-week old 126

WT C57BL/6 mice were fed with a high fat diet (HD, 60% kcal% fat, Research Diets, 127

New Brunswick, NJ) for 6 weeks and intraperitoneally injected with a low-dose of 128

streptozotocin twice (28) [25 mg/kg STZ in 0.05 M sodium citrate, pH 4.5, once daily, 129

Sigma, St. Louis, MO]. Blood glucose and body weight were measured daily, and a 130

diabetic condition was confirmed at 4 weeks after STZ injection by a non-fasting 131

blood glucose level of ≥ 200 mg/dl. Fasting blood insulin was measured, intraperito-132

neal glucose tolerance test (IPGTT) and insulin tolerance test (ITT) were carried out 133

in each successful model group. 134

In some experiments, 7-week old APN-/-

and WT mice were randomized to receive 135

a normal chow diet (ND, 12% kcal from fat, control) or HD (45% kcal from lard, Re-136

search Diets, New Brunswick, NJ) for 4 weeks. Additionally, animals received differ-137

of 50 Diabetes

8

ent treatment as following: ① ob/ob, HD-fed APN-/-

and PP2Cm-/-

mice received 138

vehicle (phosphate buffered saline, PBS) or APN (5 µg/g body wt., PeproTech, Rocky 139

Hill, NJ); ② type 2 diabetic mice received vehicle, APN (5 µg/g body wt.), AMPK 140

activator AICAR (150 µg/g body wt., Sigma, St. Louis, MO) or APN in conjunction 141

with the AMPK inhibitor compound C (20 µg/g body wt., Sigma, St. Louis, MO) once 142

daily for an additional 3 days. After 12 hours of fasting, animals were euthanized for 143

collection of tissues （liver, epididymal fat pad and gastrocnemius muscle） and 0.5 144

ml aortic blood at 10 -12 weeks of age. 145

Cell Culture and BCKA Challenge. AML12 mouse hepatocytes (ATCC, Manassas, 146

VA) were cultured in DMEM medium (Invitrogen, Carlsbad, CA) containing 10% 147

fetal bovine plasma (HyClone Waltham, MA). Experiments were carried out at 3 or 4 148

cell passages. The hepatocytes were transfected by siRNA and incubated in DMEM 149

medium (Invitrogen, Carlsbad, CA) with an additional mixture of all of the BCKA 150

(Sigma, St. Louis, MO) at 2.5 mM. Each set of cells was randomized to receive 151

treatment with 10 µg/ml APN or vehicle control. After 1 hour of treatment, cells and 152

supernatants were collected for Western blotting and BCKA analysis. 153

RNA Interference. siRNA constructs against AMPK α1 or AMPK α2 mRNA were 154

designed and purchased from Gene Pharma (Shanghai, China). The siRNA sequences 155

are as follows: 156

AMPKα1: sense 5'-GCCGACCCAAUGAUAUCAUTT3', anti-sense 157

5'-AUGAUAUCAUUGGGUCGGCTT-3'; 158

of 50Diabetes

9

AMPKα2: sense 5'-GGACAGGGAAGCCUUAAAUTT-3', anti-sense 159

5'-AUUUAAGGCUUCCCUGUCCTT-3'; 160

Scrambled siRNA: sense 5'-UUCUCCGAACGUGUCACGUTT-3', anti-sense 161

5'-ACGUGACACGUUCGGAGAATT-3'. 162

Mouse hepatocytes were transfected with siRNA by using LipofectamineTM

2000 163

(Invitrogen, Carlsbad, CA) according to the manufacturer’s instructions. Efficiency of 164

gene knockdown was confirmed using Western blotting 48 hours after siRNA trans-165

fection. 166

BCAA and BCKA Analysis. Plasma and supernatant of cultured cells were collected 167

and subsequently stored at -80 °C. Determination of BCAA concentrations was per-168

formed in triplicate using a commercially available BCAA detection kit (Biovision, 169

Milpitas, CA) per the manufacturer's instructions. BCKA concentrations were deter-170

mined by HPLC as described by Loi et al.(29). 171

BCKD enzyme activity assays. Tissue extraction and assessment of BCKD activity 172

were performed as described previously (30). BCKD complex was concentrated from 173

whole tissue extracts using 9% polyethylene glycol. BCKD activity was determined 174

spectrophotometrically by measuring the rate of NADH production resulting from the 175

conversion of α-keto-isovalerate to isobu-tyryl-CoA. A unit of enzyme activity was 176

defined as 1 µmol of NADH formed per minute at 30°C. 177

Western Blot Analysis. Proteins were separated on SDS-PAGE gels, transferred to 178

PVDF (Polyvinylidene difluoride) membranes (Millipore, Billerica, MA), and incu-179

of 50 Diabetes

10

bated overnight at 4℃ with antibodies directed against AMPKα (1:1,000,CST, Dan-180

vers, MA), AMPKα1 (1:1,000, CST, Danvers, MA), AMPKα2 (1:1,000, CST, Dan-181

vers, MA), phospho-AMPKα (Thr172, 1:1,000, CST, Danvers, MA), PP2Cm (1:1,000, 182

a gift by professor Yibin Wang of UCLA, Los Angeles, CA), BDK (1:2,000, Abcam, 183

Cambridge, MA), phospho-BCKD E1α (1:1,000, Bethyl Laboratories, Montgomery, 184

TX), BCKD E1α (1:500, Santa Cruz Biotechnology, Dallas, TX) and GAPDH 185

(1:5,000, Zhong Shan Golden Bridge Biotechnology, Beijing). After washing to re-186

move excess primary antibody, blots were incubated for 1 hour with horseradish pe-187

roxidase (HRP) conjugated secondary antibody. Binding was detected via enhanced 188

chemiluminescence (Millipore, Billerica, MA). Films were scanned with Chemi-189

DocXRS (Bio-Rad Laboratory, Hercules, CA). Densitometry was performed using 190

Lab Image software. 191

Statistical Analysis. All values in the text and figures are presented as mean ± SEM 192

of n independent experiments. All data (except densitometry) was subjected to 193

ANOVA followed by a Bonferroni correction for a post hoc t test. Densitometry was 194

analyzed using the Kruskal-Wallis test, followed by a Dunn post hoc test. Probabili-195

ties of 0.05 or less were considered statistically significant. 196

197

of 50Diabetes

11

Results 198

PP2Cm Is Down-regulated in ob/ob and Type 2 Diabetic Mice 199

Consistent with previous reports, plasma BCAA and BCKA levels were signifi-200

cantly increased in both ob/ob and type 2 diabetic mice when compared to WT mice 201

(Figure 1A and 1B). Liver is the primary metabolic clearing house of BCKA; the re-202

duction of BCKD activity causes increased circulating BCAA and BCKA concentra-203

tions (3). In addition, immunoreactivity of the BCKD pSer293 antibody is directly 204

correlated with BCKD activity (18); we therefore examined the activity of BCKD and 205

phosphorylation of BCKD at Ser293 in liver. We found that BCKD activity was sig-206

nificantly reduced and phosphorylation of BCKD was significantly increased in dia-207

betic animals as compared with WT mice (Figure1C, 1D, 1E and 1H). Interestingly, 208

the newly identified BCKD phosphatase, PP2Cm, was markedly decreased in diabetic 209

mice (Figure 1F and 1H), indicating that down-regulation of PP2Cm may be involved 210

in reduced BCKD activity in diabetes mellitus. 211

Considerable evidence indicates that besides liver, adipose tissue and skeletal mus-212

cle also play an important role in modulating circulating BCAA homeostasis (1; 2), 213

following experiments were thus performed. As shown in figure 1, BCKD activity 214

was markedly down-regulated in diabetic adipose tissue and skeletal muscle (Fig-215

ure1C, 1D, 1E and 1H). Additionally, PP2Cm protein levels were significantly re-216

duced in diabetic adipose tissue (Figure 1F and 1H). To our surprise, changes of 217

PP2Cm expression in skeletal muscle were not observed in diabetic animals (Figure 218

of 50 Diabetes

12

1F and 1H), indicating that skeletal muscle may have relatively low level of 219

PP2Cm-dependent BCKD activation. Since the activity of BCKD complex are in-220

creased by PP2Cm and inhibited by BDK, we than analyzed expression of BDK. Our 221

experimental results demonstrated that BDK expression was significantly increased in 222

diabetic tissues including liver, adipose tissue as well as skeletal muscle (Figure 1G 223

and 1H). These results indicate that decreased PP2Cm and increased BDK may both 224

contribute to reduced BCKD activity in diabetic liver and adipose tissue, whereas in-225

creased BDK may be the primary cause of decreased BCKD activity in diabetic skel-226

etal muscle. 227

Adiponectin Deficiency Contributes to Decreased BCKD Activity 228

It is known that APN reverts altered BCAA metabolism in muscle (24). In order to 229

determine whether APN contributes to BCKD activity, 7-week old WT and APN-/-

230

mice were fed with either ND or 45% HD for further 4 weeks, resulting in four sub-231

groups: ND WT, HD WT, ND APN-/-

and HD APN-/-

. As shown in figure 2, APN-/-

232

mice revealed the significant increase of plasma BCAA, BCKA and hepatic BCKD 233

phosphorylation levels, but reduction of BCKD activity in response to HD. No re-234

sponse was observed in WT mice with the same genetic background (Figure 2A, 2B, 235

2C, 2D, 2E and 2S). More importantly, hepatic PP2Cm was markedly decreased in 236

HD-fed APN-/-

mice (Figure 2F and 2S). We then sought to determine whether treat-237

ment with APN could significantly reverse altered BCKD activity. HD-fed APN-/-

238

mice were given recombinant APN at 5µg/g body wt. once daily for an additional 3 239

of 50Diabetes

13

days after the initial 4-week HD feeding period. As we expected, treatment with APN 240

completely corrected plasma BCAA and BCKA, hepatic BCKD activity, BCKD 241

phosphorylation and PP2Cm protein levels (Figure 2L, 2M, 2N, 2O, 2P, 2Q and 2T). 242

However, plasma glucose levels, glucose tolerance, insulin tolerance and cholesterol 243

concentrations did not show difference between APN and vehicle-treated mice (Figure 244

2H, 2I, 2J and 2K). Collectively, these findings demonstrate that APN deficiency is 245

responsible for down-regulated BCKD activity and PP2Cm expression. 246

Since BCKD activity is also significantly reduced in adipose tissue and skeletal 247

muscle, we thus determined the effect of APN knockout upon BCKD activity in these 248

tissues. Compared with HD-fed WT mice, HD-fed APN-/-

mice revealed greater re-249

duction of BCKD activity in adipose tissue and skeletal muscle (Figure 2C, 2D, 2E 250

and 2S), which were completely corrected by APN treatment (Figure 2N, 2O, 2P and 251

2T). Interestingly, there was no change of PP2Cm level in skeletal muscle but not in 252

adipose tissue from HD-fed APN-/-

mice (Figure 2F and 2S). APN treatment markedly 253

increased the down-regulated PP2Cm expression in adipose tissue (Figure 2Q and 2T). 254

In addition, BDK expression was significantly increased in liver, skeletal muscle and 255

adipose tissue from HD-fed APN-/-

mice (Figure 2G and 2S); and that were reverted 256

by APN treatment (Figure 2R and 2T). Collectively, these findings demonstrate for 257

the first time that APN deficiency causes down-regulated BCKD activity and PP2Cm 258

expression, and up-regulated BDK levels. 259

PP2Cm Deficiency Partially Inhibits Adiponectin-activated BCKD 260

of 50 Diabetes

14

Data have shown that plasma APN concentration is significantly reduced in diabet-261

ic patients and animals (20; 26). Consequently, we sought to determine whether ex-262

ogenous APN administration could change BCKD activity in diabetic mice. Diabetic 263

animals were injected with vehicle or APN (5 µg/g body wt., daily) for 3 days. As il-264

lustrated in figure 3, plasma glucose levels, glucose tolerance, insulin tolerance did 265

not show difference between APN and vehicle treated mice (Figure 3A, 3B and 3C). 266

To our expectation, there were significant decrease in BCAA and BCKA levels, but 267

increase in BCKD activity in APN-treated mice when compared to those only treated 268

with vehicle (Figure 3D, 3E, 3F, 3G and 3H). In addition, APN treatment significantly 269

increased PP2Cm protein expression in diabetic liver and adipose tissue, but not in 270

skeletal muscle (Figure 3I). Injection with APN markedly reduced BDK levels in dia-271

betic animals (Figure 3J). Taken together, these results indicate that APN can increase 272

down-regulated BCKD activity and ameliorate disturbed BCAA catabolism during 273

diabetes. 274

To further determine whether PP2Cm is required by APN to mediate its effects on 275

BCKD activity, we treated PP2Cm-/-

mice with vehicle or APN. Interestingly, the pos-276

itive effects of APN on BCKD were reduced but not completely lost in PP2Cm-/-

mice. 277

Specifically, plasma BCAA and BCKA concentrations were significantly higher, but 278

BCKD activity was markedly lower in PP2Cm-/-

mice treated with APN when com-279

pared with those in APN-treated diabetic and WT control groups (Figure3D, 3E, 3F, 280

3G, 3H, S3A, S3B and S3C). These data suggested that other factors may also con-281

tribute the stimulatory effect of APN upon BCKD. Indeed, APN treatment markedly 282

of 50Diabetes

15

attenuated the BDK expression in PP2Cm-/-

mice compared with vehicle treated-ones 283

(Figure 3K). Taken together, these observations support the hypothesis that APN can 284

regulate PP2Cm （up-regulating）and BDK (down-regulating) in the opposite ways, 285

thus increase BCKD activity and ameliorate disturbed BCAA catabolism in diabetic 286

disease. 287

AMPK Is Necessary for Adiponectin - mediated BCKD Activation 288

Considerable evidence exists that AMPK acts as an integrator of nutritional and 289

hormonal signals that monitor systemic and cellular energy status (31). We then in-290

vestigated whether AMPK contributes to decreased BCKD activity in diabetes. Con-291

sistent with previous reports, phosphorylated AMPK was markedly down-regulated in 292

type 2 diabetic mice (Figure 4A). We treated these mice with AICAR (21) (a 293

cell-permeable activator of AMPK, 150 µg/g body wt., daily) for 3 days. As shown in 294

figure 4, AICAR treatment resulted in a significant decrease in BCAA and BCKA 295

concentrations, while augmenting BCKD activity and PP2Cm in diabetic mice liver 296

(Figure 4B, 4C, 4D, 4E, 4F and 4G). To determine whether adipose tissue and skeletal 297

muscle contribute to system BCAA homeostasis in response to AICAR treatment, the 298

following experiments were conducted. Treatment with AICAR significantly in-299

creased the activity of BCKD in diabetic adipose tissue and skeletal muscle (Figure 300

4D and 4E). Moreover, AICAR injection significantly up-regulated PP2Cm protein 301

levels in adipose tissue, but did not influent PP2Cm expression in skeletal muscle 302

from diabetic animals (Figure 4G). However, AICAR caused a significant reduction 303

of 50 Diabetes

16

of BDK in diabetic skeletal muscle, as well as in liver and adipose tissue (Figure 4H). 304

Altogether, these results suggest that AMPK may be an endogenous regulator of 305

BCKD activity, PP2Cm and BDK expression. 306

Since AMPK activation has been recognized as a mechanism of action for APN (20; 307

32), we therefore examined whether AMPK contributes to APN-activated BCKD in 308

diabetic mice. To do so, type 2 diabetic mice were treated in three groups: vehicle, 309

APN（5 µg/g body wt., daily）, and APN plus compound C (a potent AMPK inhibitor, 310

30 mins before APN injection, 20 µg/g body wt., daily) for 3 days. As per our expec-311

tations, compound C (CC) completely blocked the effect of APN on diabetic BCAA 312

and BCKA, BCKD activity, PP2Cm and BDK expression (Figure 4I, 4J, 4K, 4L, 4M, 313

4N and 4O). These findings indicate that AMPK is necessary for APN-activated 314

BCKD in vivo. 315

The remote actions of AMPK could potentially have secondary effects on BCKD. 316

Moreover, liver has an extremely high BCKD activity compared with other tissues (3; 317

33). Thus, in vitro experiments were performed in mouse hepatocytes. BCKA chal-318

lenged mouse hepatocytes were incubated with vehicle, APN (10 µg/ml), APN (10 319

µg/ml) + CC (added 30 mins before APN treatment, 20 pmol/ml), or AICAR (2 320

pmol/ml) for 1 hour. As shown in figure S4, AICAR resulted in significantly reduced 321

BCKA and phosphorylated BCKD levels, and increased PP2Cm protein expression; 322

in contrast, coincubation with APN and CC had no effect on BCKA catabolism (Fig-323

ure S4E, S4F, S4G and S4H). Taken together, these data directly demonstrate that 324

of 50Diabetes

17

AMPK contributes to APN-activated BCKD and BCAA catabolism in mouse hepato-325

cytes. 326

AMPKα2 Contributes to Total AMPKα Activity in Adiponectin-stimulated 327

BCKD Activation 328

AMPK exists as a heterotrimeric complex consisting of a catalytic α subunit, and 2 329

regulatory β and γ subunits. Phosphorylation of Thr 172 in AMPKα is associated with 330

activation of both the α1 and α2 subunits of AMPK (34). Thus, we investigated the 331

relative contributions of AMPKα2 and AMPKα1 to total AMPKα activity in BCKD of 332

mouse hepatocytes challenged with BCKA. Mouse hepatocytes were transfected with 333

AMPKα2 or AMPKα1-specific siRNA, and AMPKα2 or AMPKα1 levels were ex-334

amined by Western blot. As depicted in figure 5A, siRNA transfection reduced the 335

levels of AMPKα2 and AMPKα1 protein by 81% and 83%, respectively. Importantly, 336

PP2Cm expression and BCKD activity were reduced, and BCKA accumulated in the 337

AMPKα2 siRNA knockdown group (Figure 5A, 5B, 5C and 5D). 338

We further examined whether AMPKα2 is involved in APN-induced BCKD activa-339

tion and PP2Cm expression. Mouse hepatocytes transfected with either AMPKα2 or 340

AMPKα1 siRNA were challenged with BCKA and then treated with vehicle or APN341

（10 µg/ml）for 1 hour. As shown in figure 5, AMPKα2 siRNA completely blocked 342

the effects of APN on PP2Cm expression, BCKD activity and BCKA levels. This ef-343

fect was not seen with scrambled siRNA or AMPKα1 siRNA transfection (Figure 5E, 344

5F and 5G). Overall, these results directly demonstrate that AMPKα2 dele-345

of 50 Diabetes

18

tion/inactivation modulates BCKD activity, and that the PP2Cm regulatory function 346

of APN is dependent on AMPKα2 in vitro. 347

348

of 50Diabetes

19

Discussion 349

Here we have made several important observations. First, we validated that PP2Cm 350

was down-regulated in diabetic mice, which may, at least in part, contribute to re-351

duced BCKD activity potentially caused by an APN deficiency. Second, APN treat-352

ment reverted down-regulated PP2Cm expression and BCKD activity, as well as ele-353

vated plasma BCAA and BCKA levels in diabetic mice. However, the increased 354

BCKD activity mediated by APN treatment was partially abolished in PP2Cm-/-

mice. 355

Third, the down-regulated BDK level was involved in APN-activated BCKD complex, 356

especially in skeletal muscle. Last, AMPK in diabetic mice was significantly impaired 357

which results in reduced BCKD activity and accumulation of BCAA and BCKA. 358

These effects could be reversed by APN injection. Importantly, APN up-regulated 359

PP2Cm expression in hepatocytes, which depends on AMPKα2. Collectively, our 360

studies have established a novel mechanism which impaired adiponectin signaling 361

contributes to diabetic BCAA catabolism. 362

BCKD complex is the most important regulatory enzyme in BCAA catabolism; the 363

activity of BCKD is regulated by a phosphorylation-dephosphorylation cycle and re-364

sponsive to alterations in various metabolic conditions (35). Studies have reported that 365

decreased BCKD activity causes down-regulation of BCAA catabolism in diabetes (6; 366

7; 36). It also has demonstrated that BDK is responsible for phosphorylation and inac-367

tivation of BCKD complex and considered a primary regulator of BCKD activation 368

(6). Diabetic and obese animals showed increased BDK expression (6; 8). Moreover, 369

alteration of metabolic status can influence BCKD activity partly via changes in BDK 370

of 50 Diabetes

20

(3), indicating that alteration of BDP may also be important. But no information about 371

BDP is available in diabetes mellitus. Identification of a specific BDP proved elusive 372

for many years; PP2Cm was identified as the only endogenous BDP until recently (18; 373

19; 27). In the present study, we demonstrated for the first time that PP2Cm was 374

significantly decreased in liver and adipose tissue from diabetic animals, which may 375

at least partially contribute to reduction of BCKD activity and elevated plasma BCKA 376

levels. Altogether, these findings provide a new explanation for the disturbed regula-377

tion of BCAA catabolism in diabetes. 378

Hypoadiponectinemia is commonly seen in diabetes mellitus (20; 26), and APN is a 379

critical regulator in lipid and glucose metabolism (20). Moreover, Liu and colleagues 380

(24) recently demonstrated that HD-fed APN-/-

mice have significantly increased lev-381

els of muscle BCAA, which can be corrected with APN supplementation for 2 weeks. 382

However, the role of APN in systemic BCAA catabolism remains unclear. In our study, 383

APN-/-

mice fed with 45% HD for 4 weeks showed mild insulin resistance (Figure 384

S2C and S2E), but normal blood glucose levels (Figure S2A). In addition, we demon-385

strated that HD-fed APN-/-

mice had significant reduced BCKD activity and PP2Cm 386

levels, increased BDK expression, along with high plasma BCAA and BCKA concen-387

trations. Treatment with APN for 3 days significantly reversed these trends but insulin 388

resistance remained (Figure 2H-J). More importantly, APN treatment completely re-389

verted the reduction of BCKD activity and PP2Cm expression, and the elevation of 390

BDK level in diabetic models. These results suggest that hypoadiponectinemia con-391

tributes to down-regulated BCKD activity and PP2Cm expression, up-regulated BDK 392

of 50Diabetes

21

level, as well as abnormal BCAA catabolism. This supports the notion that APN has a 393

wide range of impacts on metabolism (24). Additionally, evidence by Newgard et al. 394

(10) indicates that changes in BCAA levels contribute to the development of insulin 395

resistance. An interesting point that warrants further investigation is that the positive 396

effect of APN on BCAA may subsequently contribute to the insulin sensitizing effect 397

of APN. Lastly, BCAA levels changed acutely upon short-time APN administration, 398

whereas plasma glucose, glucose tolerance, insulin tolerance and plasma cholesterol 399

levels were unchanged (Figure 2H-2K and 3A-3C). Although the precise mechanism 400

causing the rapid BCAA regulatory response after short period APN administration 401

remains unclear, this result further support that BCAA may be useful biomarkers for 402

monitoring the early response to therapeutic interventions for metabolic disease. 403

In the present study, we observed that APN deficiency in the APN-/- mice did not 404

result in elevated BCAA levels under basal physiological condition. However, in re-405

viewing numerous previous publications utilizing this model, our observation is very 406

consistent with previous publication, showing that APN-/-

mouse does not have phe-407

notypic changes under physiological condition. However, when pathologically chal-408

lenged, such as exposure to HD (37) or myocardial ischemia (22), these animals show 409

much severer tissue injury than WT controls. These results indicate that although oth-410

er molecules present in APN-/- animals are sufficient to compensate the effect of APN 411

under physiological condition, APN plays essential role in counteracting the patho-412

logical stress, such as HD. Additionally, the observation that APN-stimulated BCKD 413

activation was partially retained and APN markedly down-regulated BDK expression 414

of 50 Diabetes

22

in PP2Cm-deficient mice, suggesting that BDK was involved in APN-activated 415

BCKD. It has been shown in literature that thyroid hormone and sex hormones regu-416

late expression of BDK (40) and our present data also revealed that BDK was the 417

down-stream molecular which attributes to APN-mediated BCAA catabolism. Thus, 418

APN may signal both BDK and PP2Cm in opposite ways to increase BCKD activity 419

and improve BCAA catabolism in diabetes mellitus. 420

AMPK has been considered a master switch in regulating glucose and lipid metab-421

olism (38), and plays an essential role in the actions of APN (39). Nevertheless, 422

AMPK can also modulate transcription of specific genes involved in energy metabo-423

lism (31). In this study, we observed that BCKD activity was elevated by activation of 424

AMPK both in vitro and in vivo by pharmacological or genetic methods. Moreover, 425

when the AMPK inhibitor CC or siRNA was applied, the effect of APN on BCKD 426

was virtually abolished. These results might provide us with a better understanding of 427

the role of AMPK in the regulation of metabolism. 428

The possibility remains that the remote role of APN-AMPK signaling in glucose 429

and lipid metabolism could show secondary effects on BCKD activity in vivo, we thus 430

performed in vitro experiments. Because BCKD capacity mostly resides in liver (3; 431

33), mouse hepatocytes were used in present study. Here, mouse hepatocytes chal-432

lenged with BCKA were incubated with APN or AICAR helped to address the fact 433

that APN-AMPK directly up-regulated BCKD activity (Figure S4A-S4H). Interest-434

ingly, effect of APN upon BCKA occurred prior to the significant up-regulation of 435

PP2Cm levels, suggesting that APN may improve BCKA catabolism via signaling 436

of 50Diabetes

23

system in addition to up-regulation of PP2Cm. Although the detailed molecular 437

mechanisms cannot be addressed in the current study, it is possible that APN may en-438

hance BCKD/PP2Cm interaction, thus increase BCKD activity independent PP2Cm 439

expression. This interesting possibility will be directly investigated in our future study. 440

To date, it remains unclear how PP2Cm expression is regulated. Previous studies (27; 441

40) and our unpublished data have revealed that PP2Cm expression can be regulated 442

by the availability of nutrients as well as stress (i.e. ischemia and heart failure). Here 443

we demonstrated that APN was the first identified endogenous molecule that can 444

up-regulate PP2Cm level. 445

In summary, this study provides evidence that reductions of APN signaling could 446

underlie the decreased BCKD activity observed in type 2 diabetes mellitus. Our study 447

also reveals direct evidence that APN can modulate expression of hepatic PP2Cm via 448

an AMPKα2 dependent pathway. These new insights provide a better understanding 449

of the underlying regulatory mechanisms involved in diabetic BCKD activity, and 450

identify potential therapeutic targets to mitigate BCAA catabolism in metabolic dis-451

eases. 452

453

of 50 Diabetes

24

Acknowledgments 454

This work was supported by Program for National Science Fund for Distinguished 455

Young Scholars of China(Grant No.81225001), National Key Basic Research Pro-456

gram of China (973 Program, 2013CB531204), New Century Excellent Talents in 457

University (Grant No.NCET-11-0870), National Science Funds of China(Grants No. 458

81070676 and 81170186), Program for Changjiang Scholars and Innovative Research 459

Team in University (No. PCSIRT1053) and Major Science and Technology Project of 460

China “Significant New Drug Development” (Grant No. 2012ZX09J12108-06B). 461

No potential conflicts of interests relevant to this article were reported. 462

K. L. designed methods and experiments, carried out the laboratory experiments, 463

and wrote the paper. C. S. D. performed RT-PCR analysis and analyzed data. Y. L. 464

analyzed the data and interpreted the results. D. Z. contributed to data of figure 5. W. J. 465

Y. reviewed and edited the manuscript. H. F. Z. and Z. B. H. researched data and con-466

tributed to discussion. J. L. Z., P. L. L. and H. F. P. performed animal mode and col-467

lected the samples. J. L. Z., C.G and C. X. performed BCAA and BCKA analysis. 468

H.X.C and L.Z.X contributed to discussion. L. T. defined the research theme and re-469

vised the manuscript critically. L. T. is the guarantor of this work and, as such, had 470

full access to all the data in the study and takes responsibility for the integrity of the 471

data and the accuracy of the data analysis. 472

The authors thank Dr. Xinliang Ma, Thomas Jefferson University, for help with re-473

vising the manuscript. The authors also thank Dr. Yibin Wang and Dr. Haipeng Sun, 474

of 50Diabetes

25

University of California, Los Angeles, for the generous supply of the antibody to 475

PP2Cm and PP2Cm-/-

mice. 476

477

of 50 Diabetes

26

References: 478

1. Herman MA, She P, Peroni OD, Lynch CJ, Kahn BB: Adipose tissue branched 479

chain amino acid (BCAA) metabolism modulates circulating BCAA levels. J Biol 480

Chem 285:11348-11356, 2010 481

2. Tso SC, Qi X, Gui WJ, Chuang JL, Morlock LK, Wallace AL, Ahmed K, Laxman 482

S, Campeau PM, Lee BH, Hutson SM, Tu BP, Williams NS, Tambar UK, Wynn RM, 483

Chuang DT: Structure-based design and mechanisms of allosteric inhibitors for mito-484

chondrial branched-chain alpha-ketoacid dehydrogenase kinase. Proc Natl Acad Sci 485

U S A 110:9728-9733, 2013 486

3. Adams SH: Emerging perspectives on essential amino acid metabolism in obesity 487

and the insulin-resistant state. Adv Nutr 2:445-456, 2011 488

4. Harris RA, Joshi M, Jeoung NH: Mechanisms responsible for regulation of 489

branched-chain amino acid catabolism. Biochem Biophys Res Commun 313:391-396, 490

2004 491

5. Xu F, Tavintharan S, Sum CF, Woon K, Lim SC, Ong CN: Metabolic signature 492

shift in type 2 diabetes mellitus revealed by mass spectrometry-based metabolomics. J 493

Clin Endocrinol Metab 98:E1060-E1065, 2013 494

6. Doisaki M, Katano Y, Nakano I, Hirooka Y, Itoh A, Ishigami M, Hayashi K, Goto 495

H, Fujita Y, Kadota Y, Kitaura Y, Bajotto G, Kazama S, Tamura T, Tamura N, Feng 496

GG, Ishikawa N, Shimomura Y: Regulation of hepatic branched-chain alpha-keto acid 497

dehydrogenase kinase in a rat model for type 2 diabetes mellitus at different stages of 498

the disease. Biochem Biophys Res Commun 393:303-307, 2010 499

7. Bajotto G, Murakami T, Nagasaki M, Sato Y, Shimomura Y: Decreased enzyme 500

activity and contents of hepatic branched-chain alpha-keto acid dehydrogenase 501

complex subunits in a rat model for type 2 diabetes mellitus. Metabolism 502

58:1489-1495, 2009 503

8. She P, Van Horn C, Reid T, Hutson SM, Cooney RN, Lynch CJ: Obesity-related 504

elevations in plasma leucine are associated with alterations in enzymes involved in 505

branched-chain amino acid metabolism. Am J Physiol Endocrinol Metab 506

of 50Diabetes

27

293:E1552-E1563, 2007 507

9. Tai ES, Tan ML, Stevens RD, Low YL, Muehlbauer MJ, Goh DL, Ilkayeva OR, 508

Wenner BR, Bain JR, Lee JJ, Lim SC, Khoo CM, Shah SH, Newgard CB: Insulin re-509

sistance is associated with a metabolic profile of altered protein metabolism in Chi-510

nese and Asian-Indian men. Diabetologia 53:757-767, 2010 511

10. Newgard CB, An J, Bain JR, Muehlbauer MJ, Stevens RD, Lien LF, Haqq AM, 512

Shah SH, Arlotto M, Slentz CA, Rochon J, Gallup D, Ilkayeva O, Wenner BR, Yancy 513

WJ, Eisenson H, Musante G, Surwit RS, Millington DS, Butler MD, Svetkey LP: A 514

branched-chain amino acid-related metabolic signature that differentiates obese and 515

lean humans and contributes to insulin resistance. Cell Metab 9:311-326, 2009 516

11. Shah SH, Crosslin DR, Haynes CS, Nelson S, Turer CB, Stevens RD, Muehlbauer 517

MJ, Wenner BR, Bain JR, Laferrere B, Gorroochurn P, Teixeira J, Brantley PJ, Ste-518

vens VJ, Hollis JF, Appel LJ, Lien LF, Batch B, Newgard CB, Svetkey LP: 519

Branched-chain amino acid levels are associated with improvement in insulin re-520

sistance with weight loss. Diabetologia 55:321-330, 2012 521

12. Adeva MM, Calvino J, Souto G, Donapetry C: Insulin resistance and the metabo-522

lism of branched-chain amino acids in humans. Amino Acids 43:171-181, 2012 523

13. Tremblay F, Marette A: Amino acid and insulin signaling via the mTOR/p70 S6 524

kinase pathway. A negative feedback mechanism leading to insulin resistance in skel-525

etal muscle cells. J Biol Chem 276:38052-38060, 2001 526

14. Tremblay F, Jacques H, Marette A: Modulation of insulin action by dietary pro-527

teins and amino acids: role of the mammalian target of rapamycin nutrient sensing 528

pathway. Curr Opin Clin Nutr Metab Care 8:457-462, 2005 529

15. Wurtz P, Soininen P, Kangas AJ, Ronnemaa T, Lehtimaki T, Kahonen M, Viikari 530

JS, Raitakari OT, Ala-Korpela M: Branched-chain and aromatic amino acids are pre-531

dictors of insulin resistance in young adults. Diabetes Care 36:648-655, 2013 532

16. Wang TJ, Larson MG, Vasan RS, Cheng S, Rhee EP, McCabe E, Lewis GD, Fox 533

CS, Jacques PF, Fernandez C, O'Donnell CJ, Carr SA, Mootha VK, Florez JC, Souza 534

A, Melander O, Clish CB, Gerszten RE: Metabolite profiles and the risk of develop-535

of 50 Diabetes

28

ing diabetes. Nat Med 17:448-453, 2011 536

17. Shah SH, Sun JL, Stevens RD, Bain JR, Muehlbauer MJ, Pieper KS, Haynes C, 537

Hauser ER, Kraus WE, Granger CB, Newgard CB, Califf RM, Newby LK: Baseline 538

metabolomic profiles predict cardiovascular events in patients at risk for coronary ar-539

tery disease. Am Heart J 163:844-850, 2012 540

18. Lu G, Sun H, She P, Youn JY, Warburton S, Ping P, Vondriska TM, Cai H, 541

Lynch CJ, Wang Y: Protein phosphatase 2Cm is a critical regulator of branched-chain 542

amino acid catabolism in mice and cultured cells. J Clin Invest 119:1678-1687, 2009 543

19. Joshi M, Jeoung NH, Popov KM, Harris RA: Identification of a novel PP2C-type 544

mitochondrial phosphatase. Biochem Biophys Res Commun 356:38-44, 2007 545

20. Turer AT, Scherer PE: Adiponectin: mechanistic insights and clinical implications. 546

Diabetologia 55:2319-2326, 2012 547

21. Yamauchi T, Kamon J, Minokoshi Y, Ito Y, Waki H, Uchida S, Yamashita S, 548

Noda M, Kita S, Ueki K, Eto K, Akanuma Y, Froguel P, Foufelle F, Ferre P, Carling 549

D, Kimura S, Nagai R, Kahn BB, Kadowaki T: Adiponectin stimulates glucose utili-550

zation and fatty-acid oxidation by activating AMP-activated protein kinase. Nat Med 551

8:1288-1295, 2002 552

22. Tao L, Gao E, Jiao X, Yuan Y, Li S, Christopher TA, Lopez BL, Koch W, Chan L, 553

Goldstein BJ, Ma XL: Adiponectin cardioprotection after myocardial ische-554

mia/reperfusion involves the reduction of oxidative/nitrative stress. Circulation 555

115:1408-1416, 2007 556

23. Cao Y, Tao L, Yuan Y, Jiao X, Lau WB, Wang Y, Christopher T, Lopez B, Chan 557

L, Goldstein B, Ma XL: Endothelial dysfunction in adiponectin deficiency and its 558

mechanisms involved. J Mol Cell Cardiol 46:413-419, 2009 559

24. Liu Y, Turdi S, Park T, Morris NJ, Deshaies Y, Xu A, Sweeney G: Adiponectin 560

corrects high-fat diet-induced disturbances in muscle metabolomic profile and 561

whole-body glucose homeostasis. Diabetes 62:743-752, 2013 562

25. Li S, Shin HJ, Ding EL, van Dam RM: Adiponectin levels and risk of type 2 dia-563

betes: a systematic review and meta-analysis. JAMA 302:179-188, 2009 564

of 50Diabetes

29

26. Weyer C, Funahashi T, Tanaka S, Hotta K, Matsuzawa Y, Pratley RE, Tataranni 565

PA: Hypoadiponectinemia in obesity and type 2 diabetes: close association with insu-566

lin resistance and hyperinsulinemia. J Clin Endocrinol Metab 86:1930-1935, 2001 567

27. Zhou M, Lu G, Gao C, Wang Y, Sun H: Tissue-specific and Nutrient Regulation 568

of the Branched-chain alpha-Keto Acid Dehydrogenase Phosphatase, Protein Phos-569

phatase 2Cm (PP2Cm). J Biol Chem 287:23397-23406, 2012 570

28. Zhang M, Lv XY, Li J, Xu ZG, Chen L: The characterization of high-fat diet and 571

multiple low-dose streptozotocin induced type 2 diabetes rat model. Exp Diabetes Res 572

2008:704045, 2008 573

29. Loi C, Nakib S, Neveux N, Arnaud-Battandier F, Cynober L: Ornithine al-574

pha-ketoglutarate metabolism in the healthy rat in the postabsorptive state. Metabo-575

lism 54:1108-1114, 2005 576

30. Nakai N, Kobayashi R, Popov KM, Harris RA, Shimomura Y: Determination of 577

branched-chain alpha-keto acid dehydrogenase activity state and branched-chain 578

alpha-keto acid dehydrogenase kinase activity and protein in mammalian tissues. 579

Methods Enzymol 324:48-62, 2000 580

31. Viollet B, Lantier L, Devin-Leclerc J, Hebrard S, Amouyal C, Mounier R, Foretz 581

M, Andreelli F: Targeting the AMPK pathway for the treatment of Type 2 diabetes. 582

Front Biosci 14:3380-3400, 2009 583

32. Wu X, Motoshima H, Mahadev K, Stalker TJ, Scalia R, Goldstein BJ: Involve-584

ment of AMP-activated protein kinase in glucose uptake stimulated by the globular 585

domain of adiponectin in primary rat adipocytes. Diabetes 52:1355-1363, 2003 586

33. Hutson SM, Sweatt AJ, Lanoue KF: Branched-chain [corrected] amino acid me-587

tabolism: implications for establishing safe intakes. J Nutr 135:1557S-1564S, 2005 588

34. Stein SC, Woods A, Jones NA, Davison MD, Carling D: The regulation of 589

AMP-activated protein kinase by phosphorylation. Biochem J 345 Pt 3:437-443, 2000 590

35. Lu J, Xie G, Jia W, Jia W: Insulin resistance and the metabolism of 591

branched-chain amino acids. Front Med 7:53-59, 2013 592

36. Kuzuya T, Katano Y, Nakano I, Hirooka Y, Itoh A, Ishigami M, Hayashi K, 593

of 50 Diabetes

30

Honda T, Goto H, Fujita Y, Shikano R, Muramatsu Y, Bajotto G, Tamura T, Tamura 594

N, Shimomura Y: Regulation of branched-chain amino acid catabolism in rat models 595

for spontaneous type 2 diabetes mellitus. Biochem Biophys Res Commun 373:94-98, 596

2008 597

37. Yamauchi T, Kadowaki T: Adiponectin receptor as a key player in healthy lon-598

gevity and obesity-related diseases. Cell Metab 17:185-196, 2013 599

38. Kahn BB, Alquier T, Carling D, Hardie DG: AMP-activated protein kinase: an-600

cient energy gauge provides clues to modern understanding of metabolism. Cell 601

Metab 1:15-25, 2005 602

39. Lau WB, Tao L, Wang Y, Li R, Ma XL: Systemic adiponectin malfunction as a 603

risk factor for cardiovascular disease. Antioxid Redox Signal 15:1863-1873, 2011 604

40. Lu G, Ren S, Korge P, Choi J, Dong Y, Weiss J, Koehler C, Chen JN, Wang Y: A 605

novel mitochondrial matrix serine/threonine protein phosphatase regulates the mito-606

chondria permeability transition pore and is essential for cellular survival and devel-607

opment. Genes Dev 21:784-796, 2007 608

609

of 50Diabetes

31

Figure Legends 610

Figure 1 PP2Cm Is Down-regulated in ob/ob and Type 2 Diabetic Mice 611

(A and B) Plasma BCAA and BCKA concentrations in WT, ob/ob and type 2 diabetic 612

mice. (C and D) Analysis of BCKD activity, (E) quantification of pSer293 BCKD E1α, 613

(F) PP2Cm and (G) BDK protein levels in liver, adipose tissue and skeletal muscle 614

from WT, ob/ob and type 2 diabetic mice. (H) Representative immunoblots of 615

pSer293 BCKD E1α, BCKD E1α, PP2Cm, BDK and GAPDH in WT, ob/ob and type 616

2 diabetic mice. 617

All results are presented as mean ± SEM. *Significant difference between ob/ob or 618

type 2 diabetic group versus WT group. * P < 0.05, ** P < 0.01. n = 6 - 8. L, liver; AT, 619

adipose tissue; SM, skeletal muscle. 620

621

of 50 Diabetes

32

Figure 2 Adiponectin Deficiency Contributes to Decreased BCKD Activity 622

(A and B) Plasma BCAA and BCKA, (C and D) BCKD activity, (E) quantification of 623

pSer293 BCKD E1α, (F) PP2Cm and (G) BDK protein levels in liver, adipose tissue 624

and skeletal muscle from WT and APN-/-

mice fed with ND or HD. (H) Analysis of 625

blood glucose concentrations, (I) glucose tolerance tested by IPGTT, (J) insulin toler-626

ance tested by ITT, (K) blood cholesterol levels, (L and M) plasma BCAA and BCKA 627

concentrations in HD-fed APN-/-

mice after vehicle or APN treatment. (N and O) De-628

tection of BCKD activity, (P) quantitative results of pSer293 BCKD E1α, (Q) PP2Cm 629

and (R) BDK protein levels in liver, adipose tissue and skeletal muscle from HD-fed 630

APN-/-

mice injected with APN. (S) Representative Western blots of pSer293 BCKD 631

E1α, BCKD E1α, PP2Cm, BDK and GAPDH in ND or HD-fed WT and APN-/-

mice. 632

(T) Representative immunoblots of pSer293 BCKD E1α, BCKD E1α, PP2Cm, BDK 633

and GAPDH in HD-fed APN-/-

mice injected with APN or vehicle. 634

All results are presented as mean ± SEM. *Significant difference between APN-/-

and 635

WT mice fed with HD. %

Significant difference between APN and vehicle treatment. *%

636

P < 0.05, **%%

P < 0.01. n= 6 - 8. L, liver; AT, adipose tissue; SM, skeletal muscle. 637

638

of 50Diabetes

33

Figure 3 PP2Cm Deficiency Partially Inhibits Adiponectin-activated BCKD 639

(A) Analysis of blood glucose concentrations, (B) glucose tolerance tested by IPGTT 640

and (C) insulin tolerance tested by ITT in type 2 diabetic animals administered with 641

APN or vehicle. (D and E) BCAA and BCKA concentrations in plasma, (F and G) 642

BCKD activity in liver, adipose tissue and skeletal muscle from ob/ob, type 2 diabetic 643

and PP2Cm-/- mice each injected with APN or vehicle. (H) Western blot for hepatic 644

BCKD E1α phosphorylation in ob/ob, type 2 diabetic and PP2Cm-/- mice each admin-645

istered with APN. (I) PP2Cm and (J) BDK protein levels were detected by western 646

blot in liver, adipose tissue and skeletal muscle from ob/ob and type 2 diabetes mice 647

injected with APN. (K) BDK levels were assessed by Western blot in liver, adipose 648

tissue and skeletal muscle from PP2Cm-/- mice treated with APN. 649

All results are presented as mean ± SEM. %

Significant difference between APN and 650

vehicle injected group. &

Significant difference between APN treated ob/ob and 651

PP2Cm-/-

group. %&

P < 0.05, %%&&

P < 0.01. n = 5 - 8. A, APN; AT, adipose tissue; L, 652

liver; SM, skeletal muscle; V, vehicle. 653

654

of 50 Diabetes

34

Figure 4 AMPK Is Necessary for Adiponectin - mediated BCKD Activation 655

(A) Expression of phospho-AMPKα at Thr172 and AMPKα levels in liver, adipose 656

tissue and skeletal muscle from type 2 diabetes mice; left, phospho-AMPKα, AMPKα 657

and GAPDH protein levels were assessed by Western blot analysis; right, 658

quantification of Western blot data. (B and C) BCAA and BCKA concentrations in 659

plasma, (D and E) BCKD activity in liver, adipose tissue and skeletal muscle from 660

type 2 diabetes mice treated with AICAR. (F) Immunoblotting of hepatic BCKD E1α 661

phosphorylation levels in type 2 diabetic mice treated with AICAR. (G) PP2Cm and 662

(H) BDK protein levels were measured by Western blot in liver, adipose tissue and 663

skeletal muscle from type 2 diabetic mice treated with AICAR or vehicle. (I and J) 664

Analysis of BCAA and BCKA in plasma, (K and L) BCKD activity in liver, adipose 665

tissue and skeletal muscle from type 2 diabetic animals treated with APN or APN plus 666

CC. (M) hepatic BCKD E1α phosphorylation levels in type 2 diabetic mice injected 667

with APN or APN plus CC. (N) PP2Cm and (O) BDK expression were detected in 668

type 2 diabetic animals treated with APN or APN plus CC. 669


Significant difference between AICAR and 670

vehicle treatment. &

Significant difference between APN and APN plus CC treat-671

ment. %&

P < 0.05, %%&&

P < 0.01. n = 5 - 8. A, APN; AT, adipose tissue; CC, com-672

pound C; L, liver; SM, skeletal muscle; V, vehicle. 673

674

of 50Diabetes

35

Figure 5 AMPKα2 Contributes to Total AMPKα Activity in Adiponec-675

tin-stimulated BCKD Activation 676

(A) Representative immunoblots of hepatocytes lysate for pSer293 BCKD E1α, 677

BCKD E1α, AMPKα1, AMPKα2, PP2Cm and GAPDH after transfection with scram-678

ble, AMPKα1 or AMPKα2 siRNA. (B) The relative levels of PP2Cm, (C) pSer293 679

BCKD E1α were quantified and (D) BCKA concentrations were analyzed in hepato-680

cytes transfected with scramble, AMPKα1 or AMPKα2 siRNA. (E) PP2Cm protein, 681

(F) BCKD E1α phosphorylation and (G) BCKA levels in scramble, AMPKα1 or 682

AMPKα2 siRNA transfected hepatocytes treated with or without APN for 1 hour. 683

All results are presented as mean ± SEM. *Significant difference between scramble 684

siRNA and AMPKα2 siRNA group. %

Significant difference between APN and vehicle 685

treated group. % P < 0.05, **

%%P < 0.01. n = 6 - 12 wells. 686

687

688

of 50 Diabetes

243x138mm (300 x 300 DPI)

of 50Diabetes

279x207mm (300 x 300 DPI)

of 50 Diabetes

236x211mm (300 x 300 DPI)

of 50Diabetes

273x192mm (300 x 300 DPI)

of 50 Diabetes

220x155mm (300 x 300 DPI)

of 50Diabetes

Supplemental Figure Legends

Figure S1 Adiponectin Is Reduced in Type 2 Diabetic Mice

（Related to Figure 1, Figure 3 and Figure 4 ）

(A) Body weight, (B) blood glucose concentrations, (C) glucose tolerance tested by

IPGTT, (D) insulin tolerance tested by ITT and (E) fasting blood insulin levels in

normal controls and type 2 diabetic mice. (F) APN concentrations in plasma and (G)

APN mRNA levels in adipose tissue of type 2 diabetic mice and their controls.

All results are presented as mean ± SEM. *Significant difference between type 2 dia-

betic group and normal group. * P < 0.05, **P < 0.01. n = 5 - 8.

of 50 Diabetes

Figure S2 HD-fed Adiponectin Knockout Reveals Insulin Resistance

（Related to Figure 2）

(A) Blood glucose levels, (B and C) glucose tolerance tested by IPGTT, (D and E)

insulin tolerance tested by ITT and (F) blood cholesterol levels in WT and APN-/-

mice fed with normal diet (ND) or high fat diet (45% HD).

All results are presented as mean ± SEM. *Significant difference between HD and ND

fed group. * P < 0.05, **P < 0.01. n = 6 - 8.

of 50Diabetes

Figure S3 PP2Cm Deficiency Partially Inhibits Adiponectin-activated BCKD

（Related to Figure 3）

(A and B) Plasma BCAA and BCKA concentrations, (C) hepatic BCKD E1α phos-

phorylation levels in WT and PP2Cm-/-

mice treated with vehicle or APN for 3 days.


Significant difference between APN and

vehicle injected group. &

Significant difference between APN treated WT and

PP2Cm-/-

mice. %& P < 0.05,

%%&& P < 0.01. n = 5 - 8.

of 50 Diabetes

Figure S4 APN up-regulates PP2Cm expression and improves BCAA catabolism

which depend on AMPK in vitro （Related to Figure 2, 3 and 4）

(A) BCKA levels, (B) BCKD phosphorylation (E1α at Ser293), (C) PP2Cm protein

and (D) PP2Cm mRNA levels in hepatocytes challenged with BCKA and treated with

APN. Measurements by western blot were made at 30 mins (blot lane 1-4, lane 2 and

4 represent APN treated group); 1 hour (blot lane 5-8, lane 6 and 8 represent APN

treated group); 3 hours (blot lane 9-12, lane 10 and 12 represent APN treated group);

24 hours (blot lane 13-16, lane 14 and 16 represent APN treated group). (E) Bio-

chemical analysis of BCKA concentrations, (F) BCKD phosphorylation, (G) PP2Cm

protein expression and (H) PP2Cm mRNA levels in BCKA challenged hepatocytes

which treated with vehicle, APN, APN + CC or AICAR. Measurements were made at

1 hour.


Significant difference between APN and

vehicle treated group. &

Significant difference between APN and APN plus CC treated

group. *

Significant difference between AICAR and vehicle treated group. %&*P <

0.05, %%&&**

P < 0.01. n = 6 - 12 wells. CC, compound C.

of 50Diabetes

Supplemental Experimental Procedures:

Cholesterol Content Assays.

Plasma cholesterol concentrations were performed in triplicate by the Department of

Clinical Laboratory, Xijing Hospital, The Fourth Military Medical University.

Intraperitoneal Glucose Tolerance Test (IPGTT).

After a 16-hour fast, alert mice were challenged with a glucose load of 1.5g/kg, ad-

ministered via intraperitoneal injection. Tail blood was taken 0, 15, 60, and 120

minutes after the glucose load, and blood glucose levels were determined with a

OneTouch II glucose meter (Lifescan, Milpitas, CA).

Insulin Tolerance Test (ITT)

After a 16-hour fast, insulin (0.5 IU/kg, Sigma, St. Louis, MO) was administered by

intraperitoneal injection. Tail blood samples were collected at 0, 15, 30, 60, 90 and

120 minutes for the measurement of plasma glucose. Blood glucose levels were de-

tected by a OneTouch II glucose meter (Lifescan, Milpitas, CA).

Detection of Plasma Insulin Levels.

Fasting blood insulin concentrations were detected with a mouse insulin ELISA kit

(EMD Millipore, Billerica, MA) in accordance with the manufacturer’s instructions.

Determination of Plasma Total Adiponectin Concentrations.

Endogenous plasma adiponectin levels were determined with a mouse adiponectin

ELISA kit (R&D Systems, Minneapolis, MN) in accordance with the manufacturer’s

instructions.

of 50 Diabetes

Cell Culture and BCKA Challenge.

AML12 mouse hepatocytes (ATCC, Manassas, VA) were cultured in DMEM medium

(Invitrogen, Carlsbad, CA) containing 10% fetal bovine plasma (HyClone Waltham,

MA). Experiments were carried out at 3 or 4 cell passages. After 5 hours of plas-

ma-starvation (plasma-free growth medium incubation), hepatocytes were washed

twice with PBS and incubated in DMEM medium (Invitrogen, Carlsbad, CA) which

contained an additional mixture of all of the BCKA (Sigma, St. Louis, MO) at 2.5mM.

Cell cohorts were randomized to receive one of the following treatments: vehicle

(PBS), APN (10 µg/ml), APN (10 µg/ml) + AMPK inhibitor compound C (added 30

mins before APN treatment, 20 pmol/ml), AMPK activator AICAR (2 pmol/ml). After

30 mins, 1 hour, 3 hours, or 24 hours of treatment, cells and supernatants were col-

lected for Western blotting, real-time RT-PCR and BCKA analysis.

RNA Preparation and Real-Time RT-PCR Analysis. Total RNA from mouse adi-

pocytes, liver, and cultured hepatocytes were prepared with TRIzol (Invitrogen,

Carlsbad, CA) according to the manufacturer’s instructions. Total RNA was reverse

transcribed into first-strand cDNA using the SuperScript First-Strand Synthesis Kit

(TaKaRa, Otsu, Shiga). cDNA transcripts were quantified by the Step-One Plus

RT-PCR System (Bio-Rad, Hercules, CA) using SYBR Green (TaKaRa, Otsu, Shiga).

β-actin ((TaKaRa, Otsu, Shiga) served as an endogenous control. Each reaction was

performed in triplicate and values were averaged to calculate relative expression lev-

els. Primers sequences are available upon request.

of 50Diabetes

204x120mm (300 x 300 DPI)

of 50 Diabetes

162x128mm (300 x 300 DPI)

of 50Diabetes

166x80mm (300 x 300 DPI)

of 50 Diabetes

243x208mm (300 x 300 DPI)

of 50Diabetes