Diet-induced insulin resistance in mice lacking adiponectin/ACRP30

NATURE MEDICINE • VOLUME 8 • NUMBER 7 • JULY 2002 731

ARTICLES

Adipose tissue produces and secretes several bioactive sub-stances1–5 known as adipocytokines6. Dysregulated productionof adipocytokines such as tumor necrosis factor-α (TNF-α)7,8

leptin9,10 and plasminogen-activator inhibitor type 1 (PAI-1)6 isassociated with the pathophysiology of obesity-related insulinresistance and thrombosis.

Adiponectin/ACRP30 is an adipocytokine we identified byscreening adipose-specific genes in the human cDNA project11.The mouse homolog of adiponectin has been cloned as ACRP30(ref. 12) and AdipoQ (ref. 13). The protein contains a signal se-quence, and a collagen-repeat domain at the N terminus and aglobular domain at the C terminus14. Adiponectin/ACRP30mRNA is exclusively expressed in adipose tissue in humans11,monkeys15 and mice12,13. Adiponectin/ACRP30 protein is abun-dantly present in plasma (range: 5–30 µg/ml)16, but adiposemRNA and plasma protein levels decrease in obesity and type 2diabetes17,18. Hyperinsulinemic and euglycemic clamp studies inhumans and monkeys show that plasma adiponectin/ACRP30levels correlate significantly with insulin sensitivity in thewhole body15,19. Moreover, adiponectin/ACRP30 treatment im-proves diabetes in mice20–22. These data suggest thatadiponectin/ACRP30 is an insulin-sensitizing hormone, and

that reduced production of adiponectin/ACRP30 might relate tothe pathophysiology of insulin resistance.

Here we used adiponectin/ACRP30-knockout (KO) mice toshow that mice lacking adiponectin/ACRP30 exhibited diet-in-duced insulin resistance. Based on these results, we suggest thatthe two adipocytokines adiponectin/ACRP30 and TNF-α sup-press each other’s production locally in adipose tissue and su-press each other’s function remotely in muscle.

Gene disruption and phenotype on normal dietTo disrupt the adiponectin gene we replaced exon 2, which con-tains the translation initiation sites, with the neomycin-resis-tance gene (neoR) (Fig. 1a). We confirmed disruption of thetarget allele by Southern blotting (Fig. 1b), using SpeI-digestedgenomic DNA (Fig. 1a). We also confirmed the lack ofadiponectin mRNA in fat and adiponectin protein in plasma ofKO mice (Fig. 1c and d).

During development up to 16 weeks of age, there was no dif-ference in growth rate and food intake between wild-type (WT)and KO mice. Body weight, epididymal fat, brown fat, liver, gas-trocnemius muscle and heart were similar in 12- and 16-week-old WT and KO mice. Levels of 12 h-fasted plasma glucose,

Diet-induced insulin resistance in mice lackingadiponectin/ACRP30

NORIKAZU MAEDA1, IICHIRO SHIMOMURA1,*, KEN KISHIDA1, HITOSHI NISHIZAWA1,MORIHIRO MATSUDA1, HIROYUKI NAGARETANI1, NAOKI FURUYAMA1, HIDEHIKO KONDO1,

MASAHIKO TAKAHASHI1, YUKIO ARITA1, RYUTARO KOMURO1, NORIYUKI OUCHI1, SHINJI KIHARA1,YOSHIHIRO TOCHINO1, KEIICHI OKUTOMI2, MASATO HORIE2, SATOSHI TAKEDA2,

TOSHIFUMI AOYAMA3, TOHRU FUNAHASHI1 & YUJI MATSUZAWA1

1Department of Internal Medicine and Molecular Science, Graduate School of Medicine, Osaka University, Osaka, Japan

2Otsuka GEN Research Institute, Otsuka Pharmaceutical Co. Ltd., Tokushima, Japan3Department of Aging Biochemistry, School of Medicine, Shinshu University, Nagano, Japan

*Current address: Department of Organismal Biosystems, Graduate School of Frontier Bioscience, Department of Medicine and Pathophysiology, Graduate School of Medicine, Osaka University, Osaka, Japan

K.K. and H.N. contributed equally to this study.Correspondence should be addressed to I.S.; email: [email protected] or [email protected]

Published online: 17 June 2002, doi:10.1038/nm724

Here we investigated the biological functions of adiponectin/ACRP30, a fat-derived hormone,by disrupting the gene that encodes it in mice. Adiponectin/ACRP30-knockout (KO) miceshowed delayed clearance of free fatty acid in plasma, low levels of fatty-acid transport protein 1(FATP-1) mRNA in muscle, high levels of tumor necrosis factor-α (TNF-α) mRNA in adipose tissueand high plasma TNF-α concentrations. The KO mice exhibited severe diet-induced insulin resis-tance with reduced insulin-receptor substrate 1 (IRS-1)-associated phosphatidylinositol 3 kinase(PI3-kinase) activity in muscle. Viral mediated adiponectin/ACRP30 expression in KO mice re-versed the reduction of FATP-1 mRNA, the increase of adipose TNF-α mRNA and the diet-in-duced insulin resistance. In cultured myocytes, TNF-α decreased FATP-1 mRNA, IRS-1-associatedPI3-kinase activity and glucose uptake, whereas adiponectin increased these parameters. Our re-sults indicate that adiponectin/ACRP30 deficiency and high TNF-α levels in KO mice reducedmuscle FATP-1 mRNA and IRS-1-mediated insulin signaling, resulting in severe diet-induced in-sulin resistance.

©20

02 N

atu

re P

ub

lish

ing

Gro

up

h

ttp

://m

edic

ine.

nat

ure

.co

m

732 NATURE MEDICINE • VOLUME 8 • NUMBER 7 • JULY 2002

ARTICLES

insulin and free fatty acid (FFA), and performance in glucose-,and insulin-tolerance tests on normal chow were also similar be-tween the two groups. These metabolic parameters also did notdiffer in KO mice compared with WT littermate mice at 30 weeksof age (data not shown).

Molecular physiology of KO miceWe examined the expression of mRNAs encoding proteins forfatty-acid and glucose metabolism (Fig. 2a–d). Fatty-acid trans-port protein 1 (FATP-1) mRNA was significantly lower in skeletalmuscle of KO mice (Fig. 2b), but was unchanged in white adiposetissue (WAT) and liver (Fig. 2a and c) compared with WT mice.The mRNAs of other FATP isoforms23 (FATP-2 through FATP-5)were similar in KO and WT mice in each tissue (data not shown).Levels of CD36 and lipoprotein lipase (LPL) mRNA in muscleand WAT were also similar. Expression of acyl-CoA oxidase(ACO), glucose transporter 4 (GLUT-4) and uncoupling protein 2(UCP-2) in muscle of KO mice were similar to those of WT mice(Fig. 2b). The mRNA levels of hepatic gluconeogenic enzymes,including glucose-6-phosphatase (G6P) and phosphoenolpyru-vate carboxykinase (PEPCK), were similar in the two groups (Fig.2c). Adipogenic marker genes, including peroxisome prolifera-tor-activated receptor-γ (PPAR-γ) and leptin, were normally ex-pressed in WAT (Fig. 2a) and brown adipose tissue (BAT) (datanot shown) of KO mice. mRNA expression of insulin receptorsubstrate 1 (IRS-1), IRS-2, GLUT1 and GLUT4 were also similar inWAT of the two groups (data not shown). Adipose mRNA levelsand plasma concentrations of TNF-α were higher in KO micecompared with WT mice, but were less than half of those indb/db mice (Fig. 2d). The levels TNF-α mRNA in other tissues(liver, skeletal muscle, heart and BAT) did not change in KOmice (data not shown).

Fruebis et al. demonstrated adiponectin treatment acceleratedFFA clearance in plasma22. A fat-loading test revealed delayedFFA clearance in KO mice (Fig. 2e).

Adiponectin adenovirus increases FATP-1 and decreases TNF-αTo determine whether altered regulations of muscle FATP-1mRNA and adipose TNF-α mRNA in KO mice are reversed by ex-ogenous adiponectin, mice were infected with adenovirus pro-ducing full-length adiponectin. Infection resulted in high levelsof plasma adiponectin(Fig. 2f). The decreased FATP-1 mRNA ob-served in muscle of KO mice infected with β-galactosidase (β-gal)

adenovirus contrasts with normal levels in the adiponectin ade-novirus–infected KO mice (Fig. 2f). High levels of adipose TNF-αmRNA and plasma TNF-α in β-gal adenovirus–infected KO micewere suppressed by viral-mediated production of adiponectin inplasma (Fig. 2g). The changes in FATP-1 and TNF-α levels oc-curred in KO, but not WT mice.

Severe diet-induced insulin resistance in KO miceAdiponectin KO mice showed disturbed FFA catabolism and el-evated plasma TNF-α concentrations in the basal state. In lightof this, we challenged adiponectin KO mice with a high-fat/high-sucrose (HF/HS) diet. At two weeks after challenge,plasma glucose and insulin concentrations were significantlyhigher in KO mice than in WT mice (Fig. 3a). The high levels ofhomeostasis model–assessment of insulin resistance (HOMA-IR) indicated that KO mice developed severe insulin resistance.HF/HS diet also increased plasma FFA concentrations in KOmice. After six weeks on HF/HS diet there was still no signifi-cant difference in body weight and adiposity between WT andKO mice. Mice fed with HF/HS diet for two weeks were sub-jected to glucose- and insulin-tolerance tests (Fig. 3b and c).The insulin-resistance index calculated from glucose-tolerancetest was higher in KO mice (Fig. 3b). Insulin-mediated suppres-sion of plasma glucose deteriorated in KO mice (Fig. 3c).Decreases in muscle FATP-1 mRNA and increases in adiposeTNF-α mRNA were observed in WT mice on the HF/HS diet, butnot as much as in KO mice (Fig. 3d). Insulin stimulation of IRS-1-associated phosphatidylinositol 3-kinase (PI3-kinase) on theHF/HS diet was lower in muscle of KO mice than in WT mice(Fig. 3e). IRS-2-associated PI3-kinase activity in muscle and IRS-1- or IRS-2-associated PI3-kinase activity in liver were not sig-nificantly altered in KO mice fed either normal chow or HF/HSdiet (data not shown). The triglyceride contents in liver andmuscle were similar in the two groups under each diet (datanot shown). The fatty-acid oxidation activities in gastrocne-mius muscle were not altered in the KO mice under either diet(normal chow, 9.7 ± 0.3 (WT) and 9.4 ± 0.5 (KO) pmol/min/mgprotein; HF/HS diet 13.0 ± 0.5 (WT) and 12.5 ± 0.6 (KO)pmol/min/mg protein). We also performed the same diet studyin heterozygote mice. There was no significant difference inmetabolic parameters and in glucose- and insulin-tolerancetests between heterozygote and WT mice under each diet (datanot shown).

a b c d

Fig. 1 Targeted disruption of the mouse adiponectin gene. a, Schematicrepresentation of the gene targeting strategy. Partial restriction map of themouse adiponectin locus for WT allele (top). Targeting construct of mouseadiponectin was generated by replacing the regions of exon 2 containingthe translation initiation site ATG and parts of intron 1 and 2 with the neoR

cassette (middle). Transcriptional direction of the neoR gene is indicated bythe arrow. Expected disrupted allele was obtained by homologous recom-bination (bottom). S, E, Sn, A and V indicate SpeI, EcoRI, SnaBI, AatII and

VspI restriction enzyme sites, respectively. b, SpeI-digested genomicSouthern-blot analysis. The 17-kb band corresponds to the WT allele (+/+)and the 7-kb band to the disrupted allele (–/–), using the probe denoted ina. c, Northern blotting of WAT. Adiponectin mRNA signal was detected as asingle band at ∼ 1.4 kb in WT mice. d, Immunoblotting of plasmaadiponectin. Immunoblot analysis using a rabbit polyclonal antibodyagainst mouse adiponectin as the primary antibody. The specific signal formouse adiponectin was detected at around 30 kD in WT mice.

©20

02 N

atu

re P

ub

lish

ing

Gro

up

h

ttp

://m

edic

ine.

nat

ure

.co

m

Fig. 2 Characterization of KOmice in basal state. a–c, Northern-blot analysis of mRNAs related tofatty-acid and glucose metabolism.Total RNAs of pooled samples (n =4) from WAT (a), skeletal muscle (b)and liver (c) subjected to elec-trophoresis and blot hybridizationwith cDNA. d, Upper, quantificationof adipose TNF-α mRNA levels.Aliquots of total RNA from epididy-mal fat tissue were subjected toRNAse protection assay (inset). TNF-α mRNA levels in KO (�) mice anddb/db (�) mice were 4.3- and 9.9-fold higher than those in WT (�) mice, respectively. Lower, plasma TNF-αconcentrations. TNF-α plasma levels were 15.9 ± 3.5, 36.2 ± 3.1, and 82.6± 9.8 pg/ml (mean ± s.e.m.) in WT, KO and db/db mice (n = 3–5), respec-tively. *, P < 0.01, Fisher’s PLSD test. e, Fat-loading test. 2 groups of 14-wk-old male mice were injected with Intrafat into the jugular vein after 12-hfast. �, WT, n = 5; �, KO, n = 4. 3 min after injection, FFA levels increasedfrom 1.00 ± 0.13 to 1.96 ± 0.14 mM in WT mice, and from 0.99 ± 0.09 to1.82 ± 0.14 mM in KO mice (mean ± s.e.m.). Clearance of plasma FFA atlater time points was normalized to FFA levels at 3 min after injection, con-


ARTICLES

Adiponectin adenovirus ameliorates insulin resistanceWe examined the effect of the adenoviral production ofadiponectin on the diet-induced insulin resistance in WT andKO mice. Adiponectin overproduction markedly decreasedplasma glucose, insulin and FFA levels in KO mice treated withHF/HS diet. Decreases in these parameters after adiponectinoverproduction were observed even in the HF/HS-treated WTmice (Fig. 4a). Adiponectin overproduction improved the moreimpaired insulin sensitivity of the HF/HS-treated KO (Fig. 4b).Plasma glucose at 30 min after insulin injection in WT mice in-fected with adiponectin adenovirus also declined, although itwas not statistically significant (P = 0.07).

Adiponectin and TNF-α exhibit opposite effects in myocytesThe globular, as opposed to other forms of adiponectin, hasbeen suggested to be active on muscle 20,22. Using an antibody di-rected against the globular region of the adiponectin protein, wedetected a trace amount of globular adiponectin (less than 1% offull-length adiponectin) in human and mouse plasma (data notshown). We investigated the competitive effects of globularadiponectin and TNF-α on FATP-1 mRNA, IRS-1-associated PI3-kinase activity and 2-deoxy-[3H]glucose (2-DG) uptake in C2C12myocytes. TNF-α reduced FATP-1 mRNA, whereas globularadiponectin induced this mRNA (Fig. 5a). TNF-α-induced reduc-tion of FATP-1 mRNA was abrogated by adiponectin. TNF-α re-duced insulin stimulation of IRS-1-associated PI3-kinase activity

as described24 (Fig. 5b). Adiponectin enhanced insulin-stimu-lated activation of IRS-1-associated PI3-kinase in the absence ofTNF-α (Fig. 5b). Adiponectin also increased TNF-α-mediated sup-pression of IRS-1-associated PI3-kinase activity (Fig. 5b). Insulinstimulated 2-DG uptake by approximately 1.4-fold (Fig. 5c), andTNF-α significantly decreased 2-DG uptake (Fig. 5c) as de-scribed24 (Fig. 5c). Adiponectin enhanced basal level and insulin-stimulated glucose uptake (Fig. 5c), and also increasedTNF-α-mediated suppression of glucose uptake, both in the ab-sence and presence of insulin (Fig. 5c).

DiscussionInsulin-clamp studies performed on humans and monkeys sug-gest that adiponectin is an insulin-sensitizing hormone15,19.Adiponectin injection also improves diabetes in mice20–22. Theseeffects are thought to be mediated by adiponectin’s enhance-ment of FFA oxidation in muscle20,22 and clearance of plasma FFA(ref. 22), and/or by its suppression of glucose release from livercells21,25. Here we demonstrate that adiponectin increased muscleFATP-1 mRNA levels, which may partially account for the aug-mented effect of adiponectin on FFA clearance. We did not de-tect any change in gluconeogenic mRNAs or insulin signaling inliver of KO mice. Further insulin-clamp studies should be con-ducted to investigate whether the severe diet-induced insulin re-sistance in KO mice is solely explained by the effects ofadiponectin in muscle.

a b c d

g

sidered as 100%. Plasma FFA clearance was significantly less in KO mice at30 and 60 min after injection. ∗ , P < 0.05; Student’s t test. f and g, Effect ofadenovirus-mediated production of adiponectin on FATP-1 and TNF-αmRNAs and plasma TNF-α. 10-wk-old male WT and KO mice (n = 4) wereinjected with adenoviruses producing ACRP30 (A) or β-gal (β), respectively.f, Levels of FATP-1 mRNA. Production of adiponectin in plasma was vali-dated by immunoblotting. g, Quantification of adipose mRNAs (upper) andplasma levels (lower) of TNF-α. RNAse protection assay (upper); plasmaTNF-α concentration (lower). *, P < 0.01, ANOVA with Fisher’s PLSD test.

e f

©20

02 N

atu

re P

ub

lish

ing

Gro

up

h

ttp

://m

edic

ine.

nat

ure

.co

m


ARTICLES

TNF-α is thought to be a causative factor of insulin resistancein obesity. Adipose TNF-α mRNA levels are elevated in obese dia-betic mice and humans7,8, and antagonism of TNF-α with neu-tralizing antibodies improves diabetes in mice26. Here,adiponectin KO mice had elevated levels of TNF-α in adipose tis-sue and plasma. Supplementation of plasma adiponectin usingadenovirus decreased TNF-α in KO mice. These results suggestthat adiponectin suppresses TNF-α mRNA expression. This cor-roborates our previous finding that adiponectin suppresseslipopolysaccharide (LPS)-induced TNF-α mRNA expression inmacrophages27. However, in 3T3-L1 adipocytes, TNF-α sup-pressed the transcription of adiponectin gene28, which might ex-plain the lower adiponectin mRNA levels in obesity associatedadipose tissue, where TNF-α production was increased. Thus, it ispossible that adiponectin and TNF-α mutually inhibit each oth-er’s production in adipose tissue.

In addition to these reciprocal suppressive effects in adiposetissue, adiponectin and TNF-α inhibited each other’s function in

muscle. Adiponectin inhibitsTNF-α-mediated activation ofnuclear factor-κB (NF-κB) sig-naling in vascular endothelialcells29,30. Here we reveal oppo-site effects of adiponectin andTNF-α on the insulin stimula-tion of IRS-1-associated PI3-ki-nase, glucose uptake andFATP-1 mRNA expression inmyocytes.

In humans with type 2 dia-betes, FATP-1 mRNA levels inmuscle are low31. Low adipo-nectin and high TNF-α inplasma—effects seen in thispathological condition—mightreduce FATP-1 mRNA in muscle.Fruebis et al.22 show that globu-lar adiponectin increased FFAoxidation and clearance in micefed a high-fat diet, which wasaccompanied by improved glu-cose metabolism. Increased FFAuptake by FATP-1 and enhancedFFA oxidation may account forthe adiponectin-mediated clear-ance of plasma FFA, althoughno change of FA oxidation wasdetectable in the muscle of KOmice.

The mechanism underlyingfor the opposite physiological

effects of adiponectin and TNF-α might relate to the structuralresemblance of adiponectin and TNF-α (ref. 14). Adiponectinand TNF-α might bind to each other’s receptors, due to their

a b

c

d

e

Fig. 3 Severe insulin resistance in adiponectin KO mice under HF/HS diet. 10-wk-old male WT and KO mice werefed either normal chow or HF/HS diet for 2 wk. a, Plasma levels of glucose and insulin, calculated HOMA-IR valuesand plasma FFA levels. b, Insulin-resistance indices in glucose-tolerance test, (n = 4; index normalized in WT miceunder normal chow diet as 100%). c, Glucose curves under insulin-tolerance test (n = 7; plasma glucose levels werenormalized to those at 0 min (100%). �, WT mice fed normal chow diet; �, KO mice fed normal chow diet; �, WTmice fed HF/HS diet; �, KO mice fed HF/HS diet. d, Expression of muscle FATP-1 and adipose TNF-α mRNA.Northern-blot analysis (FATP-1) and RNAse protection assay (TNF-α) were as in Fig. 2. e, Insulin-stimulated IRS-1-as-sociated PI3-kinase activity in skeletal muscles. PI3-kinase activities were measured using protein lysate pooled from4 mice in each group. 32P-labeled 3-phosphatidylinositides were quantified as in Methods. Data represent 3 indepen-dent experiments. For all panels, bars denote mean ± s.e.m.; ∗ , P < 0.05; ∗∗ , P < 0.01 compared with WT mice underHF/HS diet.

a

bFig. 4 Adenovirus-mediated production of adiponectin ameliorates diet-induced insulin resistance in mice. 12-wk-old male mice (n = 4–6) were in-jected with Ad-ACRP30 (A) or Ad-β-gal (β) and fed HF/HS diet for 2 wk as inFig. 2. a, Plasma levels of glucose, insulin and FFA. b, Glucose curves underinsulin tolerance test. �, WT mice infected with Ad-β-gal; �, WT mice in-fected with Ad-ACRP30; �, KO mice infected with Ad-β-gal; �, KO mice in-fected with Ad-ACRP30. Plasma glucose levels were normalized to those at0 min in each group (100%), n = 4. Bars denote mean ± s.e.m. ∗ , P < 0.05;∗∗ , P < 0.01 comparing Ad-ACRP30 infected and the Ad-β-gal-infectedcounterpart.

©20

02 N

atu

re P

ub

lish

ing

Gro

up

h

ttp

://m

edic

ine.

nat

ure

.co

m


ARTICLES

highly similar structures. Alternatively, these two moleculesmay share a signal-transducing pathway with ultimately oppo-site efects, as was seen in NF-κB signaling in vascular endothelialcells29.

We and others have suggested that hypoadiponectinemiacaused by over-nutrition is linked to metabolic disorders such asdiabetes and atherosclerosis16–20,28,32. Only when nutrition wasoverloaded did KO mice develop severe insulin resistance.Without over-nutrition, loss of adiponectin and the associatedmild elevation of plasma TNF-α (less than half of TNF-α levels indb/db mice) may not be sufficient to induce overt metabolic dis-orders, at least in this strain of mice. Over-nutrition in monkeysinitiated hypoadiponectinemia before the development of in-sulin resistance15. Feeding mice a high-fat diet induces TNF-αmRNA in adipose tissue before the development of insulin resis-tance33. Therefore, as long as calorie overloading persists, hy-poadiponectinemia and increased TNF-α may drive andaggravate insulin resistance.

Recent genome-wide scan studies identified a potential sus-ceptibility for locus of type 2 diabetes and the metabolic syn-

drome at chromosome 3q27, where the adiponectin gene is located34,35. Moreover, genetic missense mutations of theadiponectin gene result in reduction of plasma adiponectin andare strongly associated with metabolic syndrome including dia-betes and atherosclerosis36. SNPs in the gene encodingadiponectin, accompany hypoadiponectinemia, and are linkedto insulin resistance37. In sum, hypoadiponectinemia caused bygenetic KO, overfeeding, or naturally occuring genetic muta-tions may produce diet-induced insulin resistance and impairFFA metabolism, both of which are common features of themetabolic syndrome in human.

Fig. 6 presents a model derived from the current and previousstudies28. Adiponectin and TNF-α induce local, reciprocal sup-pression in adipose tissue. Adiponectin and TNF-α also exhibitopposite effects in skeletal muscle. Adiponectin enhances in-sulin sensitivity by activating IRS-1-associated PI3-kinase andglucose uptake, accelerates FFA clearance by enhancing FATP-1mRNA expression and probably fatty-acid oxidation in muscle,and then ameliorates TNF-α-mediated suppression of these para-meters. Decreased adiponectin and increased TNF-α in obesityshould aggravate insulin resistance by decreasing insulin sensi-tivity and fatty-acid clearance in skeletal muscle.

TNF-α is also involved in the initiation and development ofatherosclerosis38. In light of the anti-atherogenic activity ofadiponectin27,29,30, the adiponectin KO mouse with increasedplasma TNF-α may be more susceptible to atherosclerosis.

In conclusion, pharmacological interventions aimed at in-creasing plasma adiponectin levels relative to those of plasmaTNF-α should be useful in the treatment of diabetes and athero-sclerosis.

MethodsConstruction of targeting vectors, embryonic stem-cell culture andgeneration of KO mice. The targeting vector for adiponectin KO micewas constructed using a positive-selection cassette derived from the vectorpPolIIsneobpA, containing the neoR gene. The 1.4-kb VspI-EcoRI region, lo-cated in intron 2 of the mouse adiponectin gene, was inserted into theVspI site of pPolIIsneobpA before the neoR cassette. The 7-kb SnaBI-AatII re-

a b c

Fig. 5 Adiponectin increased FATP1 mRNA, activated IRS-1-associ-ated PI3-kinase and enhanced glucose uptake in C2C12 myocytes. a, Northern blotting of FATP-1 mRNA levels in C2C12 myocytes.mRNA levels were measured as in Fig. 2. b, Effects of adiponectinand/or TNF-α on IRS-1-associated PI3-kinase activity. Visualized 32P-la-

beled cells were quantified as in Fig. 2d. c, Effects of adiponectinand/or TNF-α on glucose uptake in the presence or absence of insulin.For all panels, bars denote mean ± s.e.m. of 4 experiments; *, P < 0.05;**, P < 0.01 using ANOVA with Fisher’s PLSD test.

Fig. 6 Adipo-muscular axis; reciprocal effects of adiponectin and TNF-α. A model illustrating the reciprocal interactions of adiponectin andTNF-α on their local production in adipose tissue and function remote inskeletal muscle.

©20

02 N

atu

re P

ub

lish

ing

Gro

up

h

ttp

://m

edic

ine.

nat

ure

.co

m


ARTICLES

gion located in intron 1 was inserted into SnaBI site, in the 3′ site of theneoR cassette. The targeting construct was linearized with NotI and intro-duced into mouse AB2.2-prime ES cells (Lexicon Genetics, Woodlands,Texas) by electroporation (270 V, 500 mF BTX ECM600). Embryonic stem-cell clones resistant to G418/gancyclovir were isolated and 16 positiveclones were obtained. Homologous recombination was verified bySouthern blotting. Chimeric mice were bred to C57BL/6J mice and 3chimeric males gave offspring that carried the disrupted mouseadiponectin allele through the germline. We analyzed mice backcrossed toC57BL/6J for 5 generations.

Immunoblotting and mRNA analysis. Immunoblotting for plasmaadiponectin and mRNA analysis of northern blotting and RNAse protectionassay were performed as described28,39. Abundance of mRNAs was deter-mined using FAST SCAN Scanning Imager (Molecular Dynamics;Buckinghamshire, UK).

Glucose- and insulin-tolerance tests, and fat-loading test. Insulin toler-ance tests using 0.75 mU/g body weight of human regular insulin and glu-cose tolerance tests using 2 mg/g body weight were conducted asdescribed20,39. HOMA-IR values and insulin-resistance index were calculatedas described20. For the fat-loading test, mice fasted for 12 h were injectedwith 30 µl ‘Intrafat Injection 20%’ (Takeda Chemicals, Osaka, Japan)through jugular vein22.

Preparation and administration of adenovirus. Adenovirus producingthe full-length mouse adiponectin was prepared by using AdenovirusExpression Vector Kit (Takara, Kyoto, Japan). 1 × 108 plaque-forming units(p.f.u.) of adenovirus β-galactosidase or ACRP30 adenovirus were injectedinto the jugular vein. On the day 7 after virus injection, mice were killed foranalysis.

Diet study. Mice were divided at random into 2 groups. 1 group was fedHF/HS diet (AIN93G) and the other was fed normal chow (CRF-1) (OrientalYeast, Suita, Japan).

IRS-associated PI3-kinase activity. In overnight-fasted mice, insulin (1 U/gbody weight) was injected through inferior vena cava, and the gastrocne-mius muscle and liver were excised at 3 and 4 min after injection40. IRS-1- orIRS-2-associated PI3-kinase activities were measured as described41.

Fatty acid β-oxidation activity. Fatty-acid β-oxidation activity was mea-sured as described42,43.

Preparation of recombinant protein. Glutathione S-transferase (GST) fu-sion vector for globular domain of ACRP30 (gACRP30) was constructedwith pGEX-6P-1 (Amersham). GST-gACRP30 protein was produced in strainBL21 of Escherichia coli and purified using glutathione sepharose 4B(Amersham). GST was cleaved from GST-gACRP30 protein by PreScissionProtease (Amersham).

Cell culture. Differentiation of C2C12 myocytes was performed as de-scribed24. At day 5 after differentiation induction, 10 µg/ml of recombi-nant gACRP30 and/or 10 ng/ml of mouse TNF-α (Pepro Tech EC;London, UK) were added and incubated for 24 h. For PI3-kinase activityassay, cells were treated with gACRP30 and/or TNF-α for 24 h and stimu-lated acutely with or without insulin (10 nM) for 5 min before cell collec-tion.

Deoxyglucose uptake. 2-DG uptake assay was performed using C2C12myocytes as described24. Cells were treated with 10 µg/ml gACRP30 for 24 h and 10 ng/ml TNF-α for 1 h and then stimulated with or without in-sulin (100 nM) for 30 min, before cell collection.

Statistical analysis and ethical considerations. Results were expressed asmean ± s.e.m. Differences between groups were examined for statisticalsignificance using Student’s t test or analysis of variance (ANOVA) withFisher’s PLSD test. The experimental protocol was approved by the EthicsReview Committee for Animal Experimentation of Osaka University Schoolof Medicine.

AcknowledgmentsWe thank S. Tanaka and Y. Matsukawa for technical assistance. This studywas supported in part by the ‘Research for the Future’ program from the JapanSociety for the Promotion of Science (JSPS-RFTF97L00801) and grants-in-aidfrom the Ministry of Education, Science, Sports and Culture of Japan(12307022, 12557090 and 12671084).

Competing interests statementThe authors declare that they have no competing financial interests.

RECEIVED 11 JANUARY; ACCEPTED 17 MAY 2002

1. Spiegelman, B.M. & Flier, J.S. Obesity and the regulation of energy balance. Cell104, 531–543 (2001).

2. Friedman, J.M. Obesity in the new millennium. Nature 404, 632–634 (2000).3. Hotamisligil, G.S. & Spiegelman, B.M. Tumor necrosis factor α: A key compo-

nent of the obesity-diabetes link. Diabetes 43, 1271–1278 (1994).4. Saltiel, A.R. & Kahn, C.R. Insulin signaling and the regulation of glucose and

lipid metabolism. Nature 414, 799–806 (2001).5. Berg, A.H., Combs, T.P. & Scherer, P.E. ACRP30/adiponectin: An adipokine reg-

ulating glucose and lipid metabolism. Trends Endocrinol. Metab. 13, 84–89(2002).

6. Shimomura, I. et al. Enhanced expression of PAI-1 in visceral fat: Possible con-tributor to vascular disease in obesity. Nature Med. 2, 800–803 (1996).

7. Hotamisligil, G.S., Shargill, N.S. & Spiegelman, B.M. Adipose expression oftumor necrosis factor-α: Direct role in obesity-linked insulin resistance. Science259, 87–91 (1993).

8. Hotamisligil, G.S., Arner, P., Caro, J.F., Atkinson, R.L. & Spiegelman, B.M.Increased adipose tissue expression of tumor necrosis factor-α in human obesityand insulin resistance. J. Clin. Invest. 95, 2409–2415 (1995).

9. Friedman, J.M. & Halaas, J.L. Leptin and the regulation of body weight in mam-mals. Nature. 395, 763–770 (1998).

10. Kahn, B.B. & Flier, J.S. Obesity and insulin resistance. J. Clin. Invest. 106,473–481 (2000).

11. Maeda, K. et al. cDNA cloning and expression of a novel adipose specific colla-gen-like factor, apM1 (Adipose most abundant gene transcript 1). Biochem.Biophys. Res. Commun. 221, 286–289 (1996).

12. Scherer, E.P., Williams, S., Fogliano, M., Baldini, G. & Lodish, H.F. A novel serumprotein similar to C1q, produced exclusively in adipocytes. J. Biol. Chem. 270,26746–26749 (1995).

13. Hu, E., Liang, P. & Spiegelman, B.M. AdipoQ is a novel adipose-specific genedysregulated in obesity. J. Biol. Chem. 271, 10697–10703 (1996).

14. Shapiro, L. & Scherer, P. E. The crystal structure of a complement-1q family pro-tein suggests an evolutionary link to tumor necrosis factor Curr. Biol. 8, 335–338(1998).

15. Hotta, K. et al. Circulating concentrations of the adipocyte protein, adiponectin,are decreased in parallel with reduced insulin sensitivity during the progressionto type-2 diabetes in rhesus monkeys. Diabetes 50, 1126–1133 (2001).

16. Arita, Y. et al. Paradoxical decrease of an adipose-specific protein, adiponectin,in obesity. Biochem. Biophys. Res. Commun. 257, 79–83 (1999).

17. Halleux, C.M. et al. Secretion of adiponectin and regulation of apM1 gene ex-pression in human visceral adipose tissue. Biochem. Biophys. Res. Commun. 288,1102–1107 (2001).

18. Hotta, K. et al. Plasma concentrations of a novel, adipose-specific protein,adiponectin, in type 2 diabetic patients. Arterioscler. Thromb. Vasc. Biol. 20,1595–1599 (2000).

19. Weyer, C. et al. Hypoadiponectinemia in obesity and type 2 diabetes. J. Clin.Endocrinol. Metab. 86, 1930–1935 (2001).

20. Yamauchi, T. et al. The fat-derived hormone adiponectin reverses insulin resis-tance associated with both lipoatrophy and obesity. Nature Med. 7, 941–946(2001).

21. Berg, A.H., Combs, T.P., Du, X., Brownlee, M. & Scherer, P.E. The adipocyte-se-creted protein Acrp30 enhances hepatic insulin action. Nature Med. 7, 947–953(2001).

22. Fruebis, J. et al. Proteolytic cleavage product of 30-kDa adipocyte complement-related protein increases fatty acid oxidation in muscle and causes weight loss inmice. Proc. Natl. Acad. Sci. USA 98, 2005–2010 (2001).

23. Hirsch, D., Stahl, A. & Lodish, H.F. A family of fatty acid transporters conservedfrom mycobacterium to man. Proc. Natl. Acad. Sci. U S A 95, 8625–8629 (1998).

24. del Aguila, L.F., Claffery, K.P. & Kirwan, J.P. TNF-α impairs insulin signaling andinsulin stimulation of glucose uptake in C2C12 muscle cells. Am. J. Physiol. 276,E849–E855 (1999).

25. Combs, T.P. et al. Endogenous glucose production is inhibited by the adipose-derived protein Acrp30. J. Clin. Invest. 108, 1875–1881 (2001).

26. Hotamisligil, G.S., Budavari, A., Murray, D. & Spiegelman, B.M. Reduced tyro-sine kinase activity of the insulin receptor in obesity-diabetes. Central role oftumor necrosis factor-α. J. Clin. Invest. 94, 1543–1549 (1994).

27. Ouchi, N. et al. Adipocyte-derived plasma protein, adiponectin, suppresses lipidaccumulation and class A scavenger receptor expression in human monocyte-derived macrophages. Circulation 103, 1057–1063 (2001).

28. Maeda, N. et al. PPARγ ligands increase expression and plasma concentration of

©20

02 N

atu

re P

ub

lish

ing

Gro

up

h

ttp

://m

edic

ine.

nat

ure

.co

m


ARTICLES

adiponectin, an adipose-derived protein. Diabetes 50, 2094–2099 (2001).29. Ouchi, N. et al. Adiponectin, an adipocyte-derived plasma protein, inhibits en-

dothelial NF-κB signaling through a cAMP-dependent pathway. Circulation 102,1296–1301 (2000).

30. Ouchi, N. et al. Novel modulator for endothelial adhesion molecules: adipocyte-derived plasma protein, adiponectin. Circulation 100, 2473–2476 (1999).

31. Binnert, C. et al. Fatty acid transport protein-1 mRNA expression in skeletal mus-cle and in adipose tissue in humans. Am. J. Physiol. 279, E1072–1079 (2000).

32. Yamauchi, T. et al. Increased insulin sensitivity despite lipodystrophy in Crebbpheterozygous mice. Nature Genet. 30, 221–226 (2002).

33. Hofmann, C. et al. Altered gene expression for tumor necrosis factor-α and itsreceptors during drug and dietary modulation of insulin resistance.Endocrinology 134, 264–270 (1994).

34. Vionnet, N. et al. Genomewide search for type 2 diabetes-susceptibility genes inFrench whites: Evidence for a novel susceptibility locus for early-onset diabeteson chromosome 3q27-qter and independent replication of a type 2 diabeteslocus on chromosome 1q21-q24. Am. J. Hum. Genet. 67, 1470–1480 (2000).

35. Kissebah, A.H. et al. Quantitative trait loci on chromosomes 3 and 17 influencephenotypes of the metabolic syndrome. Proc. Natl. Acad. Sci. USA 97,14478–14483 (2000).

36. Kondo, H. et al. Association of adiponectin mutation with type 2 diabetes melli-

tus: A candidate gene for the insulin resistance syndrome. Diabetes in press.37. Hara, K. et al. Genetic variation in the gene encoding adiponectin is associated

with an increased risk of type 2 diabetes in the Japanese population. Diabetes51, 536–540 (2002).

38. Ross, R. Atherosclerosis: an inflammatory disease. N. Engl. J. Med. 340, 115–126(1999).

39. Shimomura, I. et al. Insulin resistance and diabetes mellitus in transgenic miceexpressing nuclear SREBP-1c in adipose tissue: model for congenital generalizedlipodystrophy. Genes Dev. 12, 3182–3194 (1998).

40. Shimomura, I. et al. Decreased IRS-2 and increased SREBP-1c lead to mixed in-sulin resistance and sensitivity in livers of lipodystrophic and ob/ob mice. Mol.Cell 6, 77–86 (2000).

41. Yamauchi, T. et al. Insulin signaling and insulin actions in the muscles and liversof insulin-resistant, insulin receptor substrate 1-deficient mice. Mol. Cell. Biol.16, 3074–3084 (1996).

42. Aoyama, T. et al. Altered constitutive expression of fatty acid-metabolizing en-zymes in mice lacking the peroxisome proliferator-activated receptor α (PPARα).J. Biol. Chem. 273, 5678–5684 (1998).

43. Watanabe, K. et al. Constitutive regulation of cardiac fatty acid metabolismthrough peroxisome proliferator-activated receptor α associated with age-de-pendent cardiac toxicity. J. Biol. Chem. 275, 22293–22299 (2000).

©20

02 N

atu

re P

ub

lish

ing

Gro

up

h

ttp

://m

edic

ine.

nat

ure

.co

m

Diet-induced insulin resistance in mice lacking adiponectin/ACRP30

Documents

Transcript of Diet-induced insulin resistance in mice lacking adiponectin/ACRP30