Peroxisome proliferator-activated receptor (PPAR) γ is a ligand-activated transcription factor and a member of the nuclear hormone receptor superfamily that is thought to be the master regulator of fat storage; however, the relationship between PPARγ and insulin sensitivity is highly controversial. We show here that supraphysiological activation of PPARγ by PPARγ agonist thiazolidinediones (TZD) markedly increases triglyceride (TG) content of white adipose tissue (WAT), thereby decreasing TG content of liver and muscle, leading to amelioration of insulin resistance at the expense of obesity. Moderate reduction of PPARγ activity by heterozygous PPARγ deficiency decreases TG content of WAT, skeletal muscle, and liver due to increased leptin expression and increase in fatty acid combustion and decrease in lipogenesis, thereby ameliorating high fat diet-induced obesity and insulin resistance. Moreover, although heterozygous PPARγ deficiency and TZD have opposite effects on total WAT mass, heterozygous PPARγ deficiency decreases lipogenesis in WAT, whereas TZD stimulate adipocyte differentiation and apoptosis, thereby both preventing adipocyte hypertrophy, which is associated with alleviation of insulin resistance presumably due to decreases in free fatty acids, and tumor necrosis factor α, and up-regulation of adiponectin, at least in part. We conclude that, although by different mechanisms, both heterozygous PPARγ deficiency and PPARγ agonist improve insulin resistance, which is associated with decreased TG content of muscle/liver and prevention of adipocyte hypertrophy. Peroxisome proliferator-activated receptor (PPAR) γ is a ligand-activated transcription factor and a member of the nuclear hormone receptor superfamily that is thought to be the master regulator of fat storage; however, the relationship between PPARγ and insulin sensitivity is highly controversial. We show here that supraphysiological activation of PPARγ by PPARγ agonist thiazolidinediones (TZD) markedly increases triglyceride (TG) content of white adipose tissue (WAT), thereby decreasing TG content of liver and muscle, leading to amelioration of insulin resistance at the expense of obesity. Moderate reduction of PPARγ activity by heterozygous PPARγ deficiency decreases TG content of WAT, skeletal muscle, and liver due to increased leptin expression and increase in fatty acid combustion and decrease in lipogenesis, thereby ameliorating high fat diet-induced obesity and insulin resistance. Moreover, although heterozygous PPARγ deficiency and TZD have opposite effects on total WAT mass, heterozygous PPARγ deficiency decreases lipogenesis in WAT, whereas TZD stimulate adipocyte differentiation and apoptosis, thereby both preventing adipocyte hypertrophy, which is associated with alleviation of insulin resistance presumably due to decreases in free fatty acids, and tumor necrosis factor α, and up-regulation of adiponectin, at least in part. We conclude that, although by different mechanisms, both heterozygous PPARγ deficiency and PPARγ agonist improve insulin resistance, which is associated with decreased TG content of muscle/liver and prevention of adipocyte hypertrophy. peroxisome proliferator-activated receptor white adipose tissue triglyceride thiazolidinediones retinoid X receptor free fatty acids tumor necrosis factor high carbohydrate sterol regulatory element-binding protein insulin receptor substrate high fat phosphatidylinositol brown adipose tissue Peroxisome proliferator-activated receptor (PPAR)1 γ is a ligand-activated transcription factor and a member of the nuclear hormone receptor superfamily that functions as a heterodimer with a retinoid X receptor (RXR) (1Kersten S. Desvergne B. Wahli W. Nature. 2000; 405: 421-424Crossref PubMed Scopus (1701) Google Scholar, 2Spiegelman B.M. Flier J.S. Cell. 1996; 87: 377-389Abstract Full Text Full Text PDF PubMed Scopus (1175) Google Scholar, 3Gonzalez F.J. Biochimie (Paris). 1997; 79: 139-144Crossref PubMed Scopus (97) Google Scholar, 4Auwerx J. Diabetologia. 1999; 42: 1033-1049Crossref PubMed Scopus (587) Google Scholar, 5Saltiel A.R. Olefsky J.M. Diabetes. 1996; 45: 1661-1669Crossref PubMed Scopus (0) Google Scholar). Agonist-induced activation of PPARγ/RXR is known to increase insulin sensitivity (6Lehmann J.M. Moore L.B. Smith-Oliver T.A. Wilkison W.O. Willson T.M. Kliewer S.A. J. Biol. Chem. 1995; 270: 12953-12956Abstract Full Text Full Text PDF PubMed Scopus (3492) Google Scholar, 7Mukherjee R. Davies P.J. Crombie D.L. Bischoff E.D. Cesario R.M. Jow L. Hamann L.G. Boehm M.F. Mondon C.E. Nadzan A.M. Paterniti Jr., J.R. Heyman R.A. Nature. 1997; 386: 407-410Crossref PubMed Scopus (578) Google Scholar). Thiazolidinediones (TZD), which have the ability to directly bind and activate PPARγ (6Lehmann J.M. Moore L.B. Smith-Oliver T.A. Wilkison W.O. Willson T.M. Kliewer S.A. J. Biol. Chem. 1995; 270: 12953-12956Abstract Full Text Full Text PDF PubMed Scopus (3492) Google Scholar) and to stimulate adipocyte differentiation (2Spiegelman B.M. Flier J.S. Cell. 1996; 87: 377-389Abstract Full Text Full Text PDF PubMed Scopus (1175) Google Scholar, 8Okuno A. Tamemoto H. Tobe K. Ueki K. Mori Y. Iwamoto K. Umesono K. Akanuma Y. Fujiwara T. Horikoshi H. Yazaki Y. Kadowaki T. J. Clin. Invest. 1998; 101: 1354-1361Crossref PubMed Scopus (937) Google Scholar), are used clinically to reduce insulin resistance and hyperglycemia in type 2 diabetes (1Kersten S. Desvergne B. Wahli W. Nature. 2000; 405: 421-424Crossref PubMed Scopus (1701) Google Scholar, 2Spiegelman B.M. Flier J.S. Cell. 1996; 87: 377-389Abstract Full Text Full Text PDF PubMed Scopus (1175) Google Scholar, 4Auwerx J. Diabetologia. 1999; 42: 1033-1049Crossref PubMed Scopus (587) Google Scholar, 5Saltiel A.R. Olefsky J.M. Diabetes. 1996; 45: 1661-1669Crossref PubMed Scopus (0) Google Scholar). We and others (9Kubota N. Terauchi Y. Miki H. Tamemoto H. Yamauchi T. Komeda K. Satoh S. Nakano R. Ishii C. Sugiyama T. Eto K. Tsubamoto Y. Okuno A. Murakami K. Sekihara H. Hasegawa G. Naito M. Toyoshima Y. Tanaka S. Shiota K. Kitamura T. Fujita T. Ezaki O. Aizawa S. Kadowaki T. Mol. Cell. 1999; 4: 597-609Abstract Full Text Full Text PDF PubMed Scopus (1243) Google Scholar, 10Miles P.D. Barak Y. He W. Evans R.M. Olefsky J.M. J. Clin. Invest. 2000; 105: 287-292Crossref PubMed Scopus (377) Google Scholar) have reported that heterozygous PPARγ-deficient mice are protected from high fat (HF) diet- or aging-induced adipocyte hypertrophy, obesity, and insulin resistance. Consistent with this, the Pro-12 → Ala polymorphism in human PPARγ2, which moderately reduces the transcriptional activity of PPARγ, has been shown to confer resistance to type 2 diabetes (11Deeb S.S. Fajas L. Nemoto M. Pihlajamaki J. Mykkanen L. Kuusisto J. Laakso M. Fujimoto W. Auwerx J. Nat. Genet. 1998; 3: 284-287Crossref Scopus (1225) Google Scholar, 12Hara K. Okada T. Tobe K. Yasuda K. Mori Y. Kadowaki H. Hagura R. Akanuma Y. Kimura S. Ito C. Kadowaki T. Biochem. Biophys. Res. Commun. 2000; 271: 212-216Crossref PubMed Scopus (281) Google Scholar, 13Altshuler D. Hirschhorn J.N. Klannemark M. Lindgren C.M. Vohl M.C. Nemesh J. Lane C.R. Schaffner S.F. Bolk S. Brewer C. Tuomi T. Gaudet D. Hudson T.J. Daly M. Groop L. Lander E.S. Nat. Genet. 2000; 26: 76-80Crossref PubMed Scopus (124) Google Scholar). This apparent paradox raises the following important unresolved issue (14Kahn R.C. Chen L. Cohen S.E. J. Clin. Invest. 2000; 106: 1305-1307Crossref PubMed Scopus (151) Google Scholar) which we addressed experimentally in this study. We attempted to explain how insulin resistance could be improved by two opposite PPARγ activity states, supraphysiological activation of PPARγ and moderate reduction. We did so by using heterozygous PPARγ-deficient mice and a pharmacological activator of PPARγ in wild-type mice. We show here that supraphysiological activation of PPARγ by TZD markedly increases triglyceride (TG) content of white adipose tissue (WAT), thereby decreasing TG content of liver and muscle, leading to amelioration of insulin resistance at the expense of obesity. Moderate reduction of PPARγ activity by heterozygous PPARγ deficiency decreases TG content of WAT, skeletal muscle, and liver due to increased leptin expression and increase in fatty acid combustion and decrease in lipogenesis, thereby ameliorating HF diet-induced obesity and insulin resistance. Moreover, although heterozygous PPARγ deficiency and TZD have opposite effects on total WAT mass, heterozygous PPARγ deficiency decreases lipogenesis in WAT, whereas TZD stimulate adipocyte differentiation and apoptosis, thereby both preventing adipocyte hypertrophy. This results in a decrease in molecules causing insulin resistance such as free fatty acids (FFA) (15Shulman G.I. J. Clin. Invest. 2000; 106: 171-176Crossref PubMed Scopus (2220) Google Scholar) and tumor necrosis factor (TNF) α (16Hotamisligil G.S. J. Intern. Med. 1999; 245: 621-625Crossref PubMed Scopus (707) Google Scholar) and up-regulation of insulin-sensitizing hormone adiponectin (17Yamauchi T. Kamon J. Waki H. Terauchi Y. Kubota N. Hara K. Mori Y. Ide T. Murakami K. Tsuboyama-Kasaoka N. Ezaki O. Akanuma Y. Gavrilova O. Vinson C. Reitman M.L. Kagechika H. Shudo K. Yoda M. Nakano Y. Tobe K. Nagai R. Kimura S. Tomita M. Froguel P. Kadowaki T. Nat. Med. 2001; 7: 941-946Crossref PubMed Scopus (4165) Google Scholar), thereby leading to alleviation of insulin resistance. We conclude that, although by different mechanisms, both heterozygous PPARγ deficiency and PPARγ agonist improve insulin resistance, which is associated with decreased TG content of muscle/liver and prevention of adipocyte hypertrophy. Rosiglitazone was synthesized as described elsewhere (6Lehmann J.M. Moore L.B. Smith-Oliver T.A. Wilkison W.O. Willson T.M. Kliewer S.A. J. Biol. Chem. 1995; 270: 12953-12956Abstract Full Text Full Text PDF PubMed Scopus (3492) Google Scholar). Wy-14,643 was purchased from Biomol (Plymouth Meeting, PA). All other materials were from the sources given in Refs. 8Okuno A. Tamemoto H. Tobe K. Ueki K. Mori Y. Iwamoto K. Umesono K. Akanuma Y. Fujiwara T. Horikoshi H. Yazaki Y. Kadowaki T. J. Clin. Invest. 1998; 101: 1354-1361Crossref PubMed Scopus (937) Google Scholar and9Kubota N. Terauchi Y. Miki H. Tamemoto H. Yamauchi T. Komeda K. Satoh S. Nakano R. Ishii C. Sugiyama T. Eto K. Tsubamoto Y. Okuno A. Murakami K. Sekihara H. Hasegawa G. Naito M. Toyoshima Y. Tanaka S. Shiota K. Kitamura T. Fujita T. Ezaki O. Aizawa S. Kadowaki T. Mol. Cell. 1999; 4: 597-609Abstract Full Text Full Text PDF PubMed Scopus (1243) Google Scholar. Heterozygous PPARγ-deficient mice have been described (9Kubota N. Terauchi Y. Miki H. Tamemoto H. Yamauchi T. Komeda K. Satoh S. Nakano R. Ishii C. Sugiyama T. Eto K. Tsubamoto Y. Okuno A. Murakami K. Sekihara H. Hasegawa G. Naito M. Toyoshima Y. Tanaka S. Shiota K. Kitamura T. Fujita T. Ezaki O. Aizawa S. Kadowaki T. Mol. Cell. 1999; 4: 597-609Abstract Full Text Full Text PDF PubMed Scopus (1243) Google Scholar). All other animals were purchased from Nippon CREA Co., Ltd. Six-week-old mice were fed powdered chow according to methods described previously (9Kubota N. Terauchi Y. Miki H. Tamemoto H. Yamauchi T. Komeda K. Satoh S. Nakano R. Ishii C. Sugiyama T. Eto K. Tsubamoto Y. Okuno A. Murakami K. Sekihara H. Hasegawa G. Naito M. Toyoshima Y. Tanaka S. Shiota K. Kitamura T. Fujita T. Ezaki O. Aizawa S. Kadowaki T. Mol. Cell. 1999; 4: 597-609Abstract Full Text Full Text PDF PubMed Scopus (1243) Google Scholar). Drugs were given as food admixtures (8Okuno A. Tamemoto H. Tobe K. Ueki K. Mori Y. Iwamoto K. Umesono K. Akanuma Y. Fujiwara T. Horikoshi H. Yazaki Y. Kadowaki T. J. Clin. Invest. 1998; 101: 1354-1361Crossref PubMed Scopus (937) Google Scholar, 9Kubota N. Terauchi Y. Miki H. Tamemoto H. Yamauchi T. Komeda K. Satoh S. Nakano R. Ishii C. Sugiyama T. Eto K. Tsubamoto Y. Okuno A. Murakami K. Sekihara H. Hasegawa G. Naito M. Toyoshima Y. Tanaka S. Shiota K. Kitamura T. Fujita T. Ezaki O. Aizawa S. Kadowaki T. Mol. Cell. 1999; 4: 597-609Abstract Full Text Full Text PDF PubMed Scopus (1243) Google Scholar), and there was no toxicity observed including liver damage. The area of glucose and insulin curves was calculated by multiplying the cumulative mean height of the glucose values (1 mg ml−1 = 1 cm) and insulin values (1 ng ml−1 = 1 cm), respectively, by time (60 min = 1 cm) as described in Ref. 7Mukherjee R. Davies P.J. Crombie D.L. Bischoff E.D. Cesario R.M. Jow L. Hamann L.G. Boehm M.F. Mondon C.E. Nadzan A.M. Paterniti Jr., J.R. Heyman R.A. Nature. 1997; 386: 407-410Crossref PubMed Scopus (578) Google Scholar. The results are expressed as the percentage of the value of each controls. The insulin resistance index (7Mukherjee R. Davies P.J. Crombie D.L. Bischoff E.D. Cesario R.M. Jow L. Hamann L.G. Boehm M.F. Mondon C.E. Nadzan A.M. Paterniti Jr., J.R. Heyman R.A. Nature. 1997; 386: 407-410Crossref PubMed Scopus (578) Google Scholar) was calculated from the product of the areas of glucose and insulin × 10−2 in glucose tolerance test (9Kubota N. Terauchi Y. Miki H. Tamemoto H. Yamauchi T. Komeda K. Satoh S. Nakano R. Ishii C. Sugiyama T. Eto K. Tsubamoto Y. Okuno A. Murakami K. Sekihara H. Hasegawa G. Naito M. Toyoshima Y. Tanaka S. Shiota K. Kitamura T. Fujita T. Ezaki O. Aizawa S. Kadowaki T. Mol. Cell. 1999; 4: 597-609Abstract Full Text Full Text PDF PubMed Scopus (1243) Google Scholar). The results are expressed as the ratio of the value of each wild-type controls on the high carbohydrate (HC) diet (9Kubota N. Terauchi Y. Miki H. Tamemoto H. Yamauchi T. Komeda K. Satoh S. Nakano R. Ishii C. Sugiyama T. Eto K. Tsubamoto Y. Okuno A. Murakami K. Sekihara H. Hasegawa G. Naito M. Toyoshima Y. Tanaka S. Shiota K. Kitamura T. Fujita T. Ezaki O. Aizawa S. Kadowaki T. Mol. Cell. 1999; 4: 597-609Abstract Full Text Full Text PDF PubMed Scopus (1243) Google Scholar). Leptin was assayed with the enzyme-linked immunosorbent assay-based Quantikine M mouse leptin immunoassay kit (R & D Systems) according to the manufacturer's instructions. For leptin sensitivity (9Kubota N. Terauchi Y. Miki H. Tamemoto H. Yamauchi T. Komeda K. Satoh S. Nakano R. Ishii C. Sugiyama T. Eto K. Tsubamoto Y. Okuno A. Murakami K. Sekihara H. Hasegawa G. Naito M. Toyoshima Y. Tanaka S. Shiota K. Kitamura T. Fujita T. Ezaki O. Aizawa S. Kadowaki T. Mol. Cell. 1999; 4: 597-609Abstract Full Text Full Text PDF PubMed Scopus (1243) Google Scholar), leptin (PeproTech) was administered to mice as a daily intraperitoneal injection of 10 µg/g body weight/day. Isotonic sodium chloride solution was administered to the controls. Food intake and body weight were measured to assess the effects of leptin administration. Adipose tissue was removed from each animal, fixed in 10% formaldehyde/phosphate-buffered saline, and maintained at 4 °C until used. Fixed specimens were dehydrated, embedded in tissue-freezing medium (Tissue-Tek OCT compound; Miles), and frozen in dry ice and acetone. WAT was cut into 10-µm sections, and the sections were mounted on silanized slides. The adipose tissue was stained with hematoxylin and eosin. Mature white adipocytes were identified by their characteristic multilocular appearance. Total adipocyte areas were traced manually and analyzed with Win ROOF software (Mitani Co., Ltd., Chiba, Japan). White adipocyte areas were measured in 400 or more cells per mouse in each group according to the methods described previously (8Okuno A. Tamemoto H. Tobe K. Ueki K. Mori Y. Iwamoto K. Umesono K. Akanuma Y. Fujiwara T. Horikoshi H. Yazaki Y. Kadowaki T. J. Clin. Invest. 1998; 101: 1354-1361Crossref PubMed Scopus (937) Google Scholar, 9Kubota N. Terauchi Y. Miki H. Tamemoto H. Yamauchi T. Komeda K. Satoh S. Nakano R. Ishii C. Sugiyama T. Eto K. Tsubamoto Y. Okuno A. Murakami K. Sekihara H. Hasegawa G. Naito M. Toyoshima Y. Tanaka S. Shiota K. Kitamura T. Fujita T. Ezaki O. Aizawa S. Kadowaki T. Mol. Cell. 1999; 4: 597-609Abstract Full Text Full Text PDF PubMed Scopus (1243) Google Scholar). Sections of adipose tissues from mice treated for 14 days were stained by the terminal deoxynucleotidyltransferase-mediated dUTP nick end-labeling technique with a kit (In Situ Cell Death Detection Kit, AP; Roche Molecular Biochemicals) to detect apoptotic nuclei as described (8Okuno A. Tamemoto H. Tobe K. Ueki K. Mori Y. Iwamoto K. Umesono K. Akanuma Y. Fujiwara T. Horikoshi H. Yazaki Y. Kadowaki T. J. Clin. Invest. 1998; 101: 1354-1361Crossref PubMed Scopus (937) Google Scholar), with slight modifications. The numbers of all nuclei and apoptosis positive-stained nuclei were counted to calculate the ratio to the number of apoptotic nuclei to total number of nuclei. Total RNA was prepared from tissues with TRIzol (Life Technologies, Inc.) according to the manufacturer's instructions. Total RNA from 5 to 10 mice in each group was pooled, and aliquots were subjected to Northern blot analysis with the probes for rat acyl-CoA oxidase (Dr. T. Hashimoto), mouse CD36, UCP2, and adiponectin cDNA or RNase protection assay to measure mRNAs of TNFα performed using a standard protocol (8Okuno A. Tamemoto H. Tobe K. Ueki K. Mori Y. Iwamoto K. Umesono K. Akanuma Y. Fujiwara T. Horikoshi H. Yazaki Y. Kadowaki T. J. Clin. Invest. 1998; 101: 1354-1361Crossref PubMed Scopus (937) Google Scholar, 9Kubota N. Terauchi Y. Miki H. Tamemoto H. Yamauchi T. Komeda K. Satoh S. Nakano R. Ishii C. Sugiyama T. Eto K. Tsubamoto Y. Okuno A. Murakami K. Sekihara H. Hasegawa G. Naito M. Toyoshima Y. Tanaka S. Shiota K. Kitamura T. Fujita T. Ezaki O. Aizawa S. Kadowaki T. Mol. Cell. 1999; 4: 597-609Abstract Full Text Full Text PDF PubMed Scopus (1243) Google Scholar, 18Murakami K. Tobe K. Ide T. Mochizuki T. Ohashi M. Akanuma Y. Yazaki Y. Kadowaki T. Diabetes. 1998; 47: 1841-1847Crossref PubMed Scopus (244) Google Scholar, 19Motojima K. Passilly P. Peters J.M. Gonzalez F.J. Latruffe N. J. Biol. Chem. 1998; 273: 16710-16714Abstract Full Text Full Text PDF PubMed Scopus (456) Google Scholar). The radioactivity in each band was quantified, and the fold change in each mRNA was calculated after correction for loading differences by measuring the amount of 28 S rRNA. Very low levels (<10%) of adipocyte P2 mRNA were detected in muscle as compared with those in WAT. By contrast, CD36, SCD1, acyl-CoA oxidase, and UCP2 mRNAs were detected in muscle at levels comparable to those in WAT. These findings suggest that the results for muscle tissue essentially represent the results for the muscle cells, although the muscle was contaminated by a small amount (<10%) of inter-myocyte fat (20Vidal-Puig A. Jimenez-Linan M. Lowell B.B. Hamann A. Hu E. Spiegelman B. Flier J.S. Moller D.E. J. Clin. Invest. 1996; 97: 2553-2561Crossref PubMed Scopus (598) Google Scholar). The procedures used for PI3-kinase assay, immunoprecipitation, and immunoblotting have been described previously (21Yamauchi T. Tobe K. Tamemoto H. Ueki K. Kaburagi Y. Yamamoto-Honda R. Takahashi Y. Yoshizawa F. Aizawa S. Akanuma Y. Sonenberg N. Yazaki Y. Kadowaki T. Mol. Cell. Biol. 1996; 16: 3074-3084Crossref PubMed Scopus (251) Google Scholar). Representative data from one of three independent experiments are shown. The measurements of [14C]CO2 production from [1-14C]palmitic acid and lipogenesis from [1-14C]acetate were performed using liver, muscle, and WAT slices, as described (18Murakami K. Tobe K. Ide T. Mochizuki T. Ohashi M. Akanuma Y. Yazaki Y. Kadowaki T. Diabetes. 1998; 47: 1841-1847Crossref PubMed Scopus (244) Google Scholar, 22Mannaerts G.P. Debeer L.J. Thomas J. De Schepper P.J. J. Biol. Chem. 1979; 254: 4585-4595Abstract Full Text PDF PubMed Google Scholar). Liver and muscle homogenates were extracted, and their TG content was determined as described previously (18Murakami K. Tobe K. Ide T. Mochizuki T. Ohashi M. Akanuma Y. Yazaki Y. Kadowaki T. Diabetes. 1998; 47: 1841-1847Crossref PubMed Scopus (244) Google Scholar), with some modifications. To explain how insulin resistance could be improved by two opposite PPARγ activity states, supraphysiological activation of PPARγ and moderate reduction, we studied the phenotypes of untreated or PPARγ agonist-treated wild-type mice and untreated heterozygous PPARγ-deficient mice. We assessed PPARγ activity in vivoby measuring expression levels of lipoprotein lipase (23Schoonjans K. Peinado-Onsurbe J. Lefebvre A.M. Heyman R.A. Briggs M. Deeb S. Staels B. Auwerx J. EMBO J. 1996; 15: 5336-5348Crossref PubMed Scopus (1037) Google Scholar), fatty-acid translocase (FAT)/CD36 (24Tontonoz P. Nagy L. Alvarez J.G. Thomazy V.A. Evans R.M. Cell. 1998; 93: 241-252Abstract Full Text Full Text PDF PubMed Scopus (1633) Google Scholar), and adipocyte fatty acid-binding protein/adipocyte P2 (25Tontonoz P. Hu E. Spiegelman B.M. Cell. 1994; 79: 1147-1156Abstract Full Text PDF PubMed Scopus (3175) Google Scholar) (Fig.1A), whose promoters contain peroxisome proliferator response element, in WAT, where PPARγ is expressed most predominantly in vivo. As expected, rosiglitazone-treated wild-type mice exhibited a significant increase in PPARγ activity as compared with untreated wild-type mice (Fig.1A, lanes 1 and 2), whereas untreated heterozygous PPARγ-deficient mice showed a moderate decrease in PPARγ activity (Fig. 1A, lanes 1 and3). Untreated wild-type mice on the HF diet gained significantly more body weight than the mice on the HC diet (data not shown). Administration of rosiglitazone to wild-type mice increased significantly more body weight than vehicle on the HF diet (Fig. 1B, lanes 1 and 2). In contrast, heterozygous PPARγ deficiency reduced the increase in body weight on the HF diet (Fig. 1B, lanes 2 and 3). Treatment of wild-type mice with rosiglitazone significantly increased WAT mass (Fig. 1C,lanes 1 and 2), whereas untreated heterozygous PPARγ-deficient mice were protected from HF diet-induced increase in WAT mass (Fig. 1C, lanes 2 and 3). These data suggested that PPARγ determines the adiposity in proportion to its activity. Treatment of wild-type mice with rosiglitazone improved hyperglycemia (Fig. 1D, lanes 1) and hyperinsulinemia (Fig.1E, lane 1) on the HF diet as compared with untreated wild-type mice (Fig. 1, D and E, lane 2). Untreated heterozygous PPARγ-deficient mice were also protected from HF diet-induced hyperglycemia (Fig. 1D, lanes 2 and 3) and hyperinsulinemia (Fig. 1E, lanes 2 and 3). These findings indicate that TZD improve insulin sensitivity at the expense of obesity, whereas moderate reduction of PPARγ activity has potential as anti-obesity and anti-diabetic drugs. The rectal temperature was lower in rosiglitazone-treated wild-type mice than that in untreated wild-type mice (Fig. 2A, lanes 1 and 2); on the contrary, it was significantly higher in untreated heterozygous PPARγ-deficient mice (Fig. 2A, lanes 2 and 3). The serum leptin (26Friedman J.M. Nature. 2000; 404: 632-634Crossref PubMed Scopus (634) Google Scholar) levels were slightly but not significantly lower in rosiglitazone-treated wild-type mice than those of untreated wild-type mice (Fig. 2B,lanes 1 and 2), whereas they were significantly higher in untreated heterozygous PPARγ-deficient mice (Fig.2B, lanes 2 and 3). Thus the serum leptin levels were parallel to the rectal temperature. It was also noted that serum leptin levels were negatively correlated with PPARγ activity, suggesting that the serum leptin levels were parallel to the repression of leptin gene transcription by PPARγ/RXR (27Hollenberg A.N. Susulic V.S. Madura J.P. Zhang B. Moller D.E. Tontonoz P. Sarraf P. Spiegelman B.M. Lowell B.B. J. Biol. Chem. 1997; 272: 5283-5290Abstract Full Text Full Text PDF PubMed Scopus (233) Google Scholar). Moreover, leptin sensitivity as assessed by reductions in food intake and body weight change in response to exogenously administered leptin was significantly increased in heterozygous PPARγ-deficient mice as compared with wild-type mice on the HF diet (Fig. 2C andD, lanes 3–6). The degree of change in body weight induced by leptin treatment differed significantly (p < 0.01) between wild-type (−0.67 ± 0.09%) and heterozygous PPARγ-deficient mice (−1.38 ± 0.11%). These data raised the possibility that increased leptin effects may contribute to the effects of the heterozygous PPARγ deficiency. In WAT from untreated heterozygous PPARγ-deficient mice, expressions of lipoprotein lipase and CD36 were reduced, which may contribute to decreased TG content. In addition, expressions of lipogenic enzymes such as sterol regulatory element-binding protein (SREBP) 1c and SCD (stearoyl-CoA desaturase) 1 were reduced, and lipid synthesis was indeed significantly decreased in WAT from heterozygous PPARγ-deficient mice as compared with that in wild-type mice on the HF diet (Fig. 2F). Expression of β3-adrenergic receptor (Fig. 2E, lanes 2 and 3) was increased presumably due to their increased leptin effects (Fig. 2, B–D) (28Soukas A. Cohen P. Socci N.D. Friedman J.M. Genes Dev. 2000; 14: 963-980PubMed Google Scholar) and decreased PPARγ/RXR effects (29Bakopanos E. Silva J.E. Diabetes. 2000; 49: 2108-2115Crossref PubMed Scopus (35) Google Scholar), and fatty acid oxidation was increased (data not shown). Decreased lipid synthesis and increased fatty acid oxidation as well as presumably decreased fatty acid influx in heterozygous PPARγ-deficient mice may in concert prevent adipocyte hypertrophy (Fig. 3A), and therefore obesity (Fig. 1, B and C, lanes 2 and 3), on the HF diet. Interestingly, supraphysiological activation of PPARγ significantly reduced the average size of adipocytes under the HF diet (Fig.3A, lane 1 and 2) as a result of a marked increase in the number of newly differentiated small adipocytes and significant decrease in the number of large adipocytes with a concomitant induction of apoptosis of adipocyte (Fig. 3B, lanes 1 and 2, and Fig. 3C) (8Okuno A. Tamemoto H. Tobe K. Ueki K. Mori Y. Iwamoto K. Umesono K. Akanuma Y. Fujiwara T. Horikoshi H. Yazaki Y. Kadowaki T. J. Clin. Invest. 1998; 101: 1354-1361Crossref PubMed Scopus (937) Google Scholar). On the other hand, heterozygous PPARγ deficiency appeared to prevent HF diet-induced adipocyte hypertrophy without a significant change in the total number of adipocytes (Fig. 3B, lanes 2 and3 and Fig. 3C). These data suggest that the heterozygous PPARγ deficiency decreases lipogenesis in WAT, whereas TZD stimulate adipocyte differentiation and apoptosis, thereby both preventing adipocyte hypertrophy. We next attempted to experimentally clarify the relationships between adipocyte hypertrophy and insulin resistance. To this end, we induced adipocyte hypertrophy by high fat feeding, leptin receptor deficiency, or agouti overexpression. The size of the adipose cells and the insulin resistance were increased in mice on a HF diet compared with those in mice on a HC diet (Fig. 3D). The size of the adipose cells and the insulin resistance of db/db mice were also increased compared with their wild-type controls on both the HC and HF diet (Fig. 3D). We obtained essentially similar results by using KKAy mice and their wild-type controls (KK) (Fig. 3E). These findings support a close correlation between adipocyte hypertrophy and insulin resistance, even though a cause and effect relationship is again unproven. In this context, protection from adipocyte hypertrophy due to decreased lipid synthesis in WAT from heterozygous PPARγ-deficient mice (Fig. 2F) may cause an increase in insulin sensitivity (Fig. 3F). We tried to clarify the molecular link between adipocyte hypertrophy and insulin resistance. We examined the levels of expression of molecules secreted by WAT that regulate insulin sensitivity under the following four different conditions: HC feeding, HF feeding, HF feeding with heterozygous PPARγ deficiency, and HF feeding with PPARγ agonist. The HF diet significantly increased adipocyte size and at the same time increased the molecules causing insulin resistance, such as FFA and TNFα (Fig. 3,G and H), and decreased the molecules causing insulin sensitivity, such as adiponectin (Fig. 3I), in mice that exhibited insulin resistance as compared with mice on the HC diet (Fig. 1, D and E, lanes 1 and2). (Replenishment of adiponectin in KKAy mice on the HF diet partially reverses insulin resistance even at the doses that do not significantly change adipocyte size (17Yamauchi T. Kamon J. Waki H. Terauchi Y. Kubota N. Hara K. Mori Y. Ide T. Murakami K. Tsuboyama-Kasaoka N. Ezaki O. Akanuma Y. Gavrilova O. Vinson C. Reitman M.L. Kagechika H. Shudo K. Yoda M. Nakano Y. Tobe K. Nagai R. Kimura S. Tomita M. Froguel P. Kadowaki T. Nat. Med. 2001; 7: 941-946Crossref PubMed Scopus (4165) Google Scholar).) In addition, treatment of wild-type mice with the PPARγ agonist rosiglitazone or heterozygous PPARγ deficiency, both of which resulted in protection against HF diet-induced adipocyte hypertrophy, significantly decreased FFA and TNFα (Fig. 3, G and H) and increased adiponectin (Fig. 3I) and at the same time ameliorated insulin resistance (Fig. 1, D and E) on the HF diet. However, treatment of wild-type mice with a PPARγ agonist increased adipose tissue mass (Fig. 1C, lane 1) and body weight (Fig. 1B, lane 1), whereas heterozygous PPARγ deficiency significantly decreased them (Fig. 1,B and C, lane 3). These findings raised the possibility that levels of expression of molecules regulating insulin sensitivity may be more closely related to adipocyte size than PPARγ activity, adipose tissue mass, or body weightin vivo. Large adipocytes are known to be resistant to anti-lipolytic action of insulin, thereby releasing a large amount of FFA (30Askew E.W. Huston R.L. Plopper C.G. Hecker A.L. J. Clin. Invest. 1975; 56: 521-529Crossref PubMed Scopus (70) Google Scholar); however, the mechanisms underlying the correlation between larger adipocytes and up-regulation of TNFα and/or down-regulation of adiponectin remain to be elucidated. Interestingly, both supraphysiological activation of PPARγ and moderate reduction of PPARγ activity significantly reduced tissue TG content in muscle and liver (Fig.4A and Fig.5A), suggesting that insulin resistance has an excellent correlation with tissue TG content in muscle and liver (Fig. 1, D and E, Fig.4A, and Fig. 5A) (15Shulman G.I. J. Clin. Invest. 2000; 106: 171-176Crossref PubMed Scopus (2220) Google Scholar).Figure 5TZDindirectly decreases molecules involved inFFAinflux into the liver, whereas heterozygous PPARγ deficiency combusts fatty acid and decreases lipogenesis, thereby both decreasing tissueTGcontent in liver. Tissue triglyceride (A), amounts of the mRNAs of fatty-acid translocase (FAT)/CD36, SREBP 1, stearoyl Co-A desaturase (SCD) 1, acyl-CoA oxidase (ACO), and uncoupling protein (UCP) 2 (B), lipid synthesis from [14C]acetate in the liver (C), fatty acid (FA) oxidation (D), glucokinase, phosphoenolpyruvate carboxykinase (PEPCK) and glucose-6-phosphatase (G6Pase) (E), in liver of wild-type (WT) and heterozygous PPARγ-deficient mice (+/−) untreated (−) or treated with rosiglitazone (Rosi) for 4 weeks while on the HF diet. Rosiglitazone was given as a 0.01% food admixture. Fatty acid oxidation was assessed by the measurements of [14C]CO2 production from [14C]palmitic acid (D). Each bar represents the mean ± S.E. (n = 5–10). *, p< 0.05; **, p < 0.01; compared with untreated wild-type mice.View Large Image Figure ViewerDownload Hi-res image Download (PPT) In wild-type mice treated with rosiglitazone, the decreased tissue TG content in muscle/liver, where PPARγ was less abundantly expressed as compared with what was in WAT, was presumably via reduced expression of molecules involved in FFA influx into muscle/liver (Fig. 4Eand Fig. 5B, lanes 1 and 2). On the other hand, heterozygous PPARγ deficiency reduced expression of lipogenic enzymes such as SCD1 (Fig. 4E and Fig. 5B, lanes 2 and 3) and SREBP 1 (Fig.5B, lanes 2 and 3), and indeed significantly reduced lipogenesis (Fig. 5C), presumably due to increased leptin effects (Fig. 2, B–D) (28Soukas A. Cohen P. Socci N.D. Friedman J.M. Genes Dev. 2000; 14: 963-980PubMed Google Scholar), may reduce tissue TG content in muscle/liver (Fig. 4A and Fig.5A, lanes 2 and 3). Moreover, in muscle/liver from heterozygous PPARγ-deficient mice, increased expression of enzymes involved in β-oxidation such as acyl-CoA oxidase and that of molecules involved in energy dissipation such as UCP2 (Fig. 4E and Fig.5B, lanes 2 and 3) were observed. Fatty acid oxidation was indeed significantly increased in muscle/liver from heterozygous PPARγ-deficient mice as compared with that in wild-type mice on the HF diet (Fig. 4F and Fig.5D). These alterations may be an additional mechanism for reduced TG content in muscle/liver of heterozygous PPARγ-deficient mice. Since these effects were recapitulated by treatment of wild-type mice with Wy-14,643, a PPARα agonist as reported (18Murakami K. Tobe K. Ide T. Mochizuki T. Ohashi M. Akanuma Y. Yazaki Y. Kadowaki T. Diabetes. 1998; 47: 1841-1847Crossref PubMed Scopus (244) Google Scholar, 19Motojima K. Passilly P. Peters J.M. Gonzalez F.J. Latruffe N. J. Biol. Chem. 1998; 273: 16710-16714Abstract Full Text Full Text PDF PubMed Scopus (456) Google Scholar, 31Kelly L.J. Vicario P.P. Thompson G.M. Candelore M.R. Doebber T.W. Ventre J. Wu M.S. Meurer R. Forrest M.J. Conner M.W. Cascier I.M.A. Moller D.E. Endocrinology. 1998; 139: 4920-4927Crossref PubMed Scopus (258) Google Scholar) (data not shown), PPARα pathways appeared to be activated in the liver of heterozygous PPARγ-deficient mice. In the BAT, where PPARα was relatively abundantly expressed compared to that in WAT, significant increases in the expression of molecules involved in fatty acid combustion presumably via activation of PPARα pathways (18Murakami K. Tobe K. Ide T. Mochizuki T. Ohashi M. Akanuma Y. Yazaki Y. Kadowaki T. Diabetes. 1998; 47: 1841-1847Crossref PubMed Scopus (244) Google Scholar, 19Motojima K. Passilly P. Peters J.M. Gonzalez F.J. Latruffe N. J. Biol. Chem. 1998; 273: 16710-16714Abstract Full Text Full Text PDF PubMed Scopus (456) Google Scholar, 31Kelly L.J. Vicario P.P. Thompson G.M. Candelore M.R. Doebber T.W. Ventre J. Wu M.S. Meurer R. Forrest M.J. Conner M.W. Cascier I.M.A. Moller D.E. Endocrinology. 1998; 139: 4920-4927Crossref PubMed Scopus (258) Google Scholar) and β3-adrenergic receptor (Fig. 6), due to increased leptin effects (28Soukas A. Cohen P. Socci N.D. Friedman J.M. Genes Dev. 2000; 14: 963-980PubMed Google Scholar) and decreased PPARγ effects (29Bakopanos E. Silva J.E. Diabetes. 2000; 49: 2108-2115Crossref PubMed Scopus (35) Google Scholar), were observed. These alterations in concert may provide the mechanism that increased energy expenditure by heterozygous PPARγ deficiency (Fig.2A). Increased tissue TG content has been reported to interfere with insulin-stimulated phosphatidylinositol (PI) 3-kinase activation and subsequent GLUT4 translocation and glucose uptake (15Shulman G.I. J. Clin. Invest. 2000; 106: 171-176Crossref PubMed Scopus (2220) Google Scholar). Next, we tried to experimentally clarify the relationships between tissue TG content and insulin resistance. To do so, we increased tissue TG content by high fat feeding, leptin receptor deficiency, or agouti overexpression. The tissue TG content of skeletal muscle and insulin resistance were increased in mice on the HF diet compared with those in mice on the HC diet (Fig. 4B). The tissue TG content of skeletal muscle and insulin resistance of db/db mice were also increased compared with their wild-type controls on both the HC and HF diets (Fig.4B). We obtained essentially similar results by using KKAy mice and their wild-type controls (KK) (Fig. 4C). These findings raise the possibility that increases in tissue TG content are associated with insulin resistance. Conversely, decreased tissue TG content due to decreased lipid synthesis and increased fatty acid oxidation in muscle/liver from heterozygous PPARγ-deficient mice (Fig. 4F and Fig. 5, C and D) may cause an increase in insulin sensitivity (Fig. 4D). Shulman and co-workers (32Abel E.D. Peroni O. Kim J.K. Kim Y.B. Boss O. Hadro E. Minnemann T. Shulman G.I. Kahn B.B. Nature. 2001; 409: 729-733Crossref PubMed Scopus (983) Google Scholar) proposed a cause and effect relationship between the accumulation of intracellular fatty acid-derived metabolites and insulin resistance. However, there are instances in which tissue TG content actually does not change in another scenario that also causes insulin resistance, i.e. adipose-selective targeting of the GLUT4 gene (32Abel E.D. Peroni O. Kim J.K. Kim Y.B. Boss O. Hadro E. Minnemann T. Shulman G.I. Kahn B.B. Nature. 2001; 409: 729-733Crossref PubMed Scopus (983) Google Scholar). Thus, interpretation should be done with caution, and decreased tissue TG content in muscle/liver is one possible mechanism for the results of increased insulin sensitivity in heterozygous PPARγ-deficient mice. Consistent with this possibility, decreased TG content in muscle of heterozygous PPARγ-deficient mice indeed improved insulin signal transduction in muscle, as demonstrated by increases in insulin-induced tyrosine phosphorylation of insulin receptor, insulin receptor substrate (IRS)-1 and IRS-2, and insulin-stimulated PI3-kinase activity in phosphotyrosine, IRS-1 and IRS-2 immunoprecipitates, and insulin-stimulated Akt activity in skeletal muscle (Fig.4G). The reduction of TG content in liver of heterozygous PPARγ-deficient mice was associated with increased expression of glucokinase and decreased expression of enzymes involved in gluconeogenesis such as phosphoenolpyruvate carboxykinase and glucose-6-phosphatase (Fig. 5E), indicating increased insulin actions also in liver. We attempted to explain how insulin resistance could be improved by the following two opposite PPARγ activity states: a potent activation of PPARγ and its moderate reduction. We did so by using heterozygous PPARγ-deficient mice and a pharmacological activator of PPARγ in wild-type mice. On the basis of experimental results obtained in this study, we propose the following hypothesis on the mechanisms for the regulation of insulin sensitivity by PPARγ (Fig.7). As shown in the Fig. 7, panel 2, on the HF diet, “normal” amounts of PPARγ activity seen in wild-type mice increase TG content in WAT, skeletal muscle, and liver due to a combination of increased fatty acid influx into WAT, skeletal muscle, and liver and HF diet-induced leptin resistance, leading to insulin resistance and obesity. Moreover, hypertrophic adipocytes may increase the secretion of molecules causing insulin resistance, such as FFA (15Shulman G.I. J. Clin. Invest. 2000; 106: 171-176Crossref PubMed Scopus (2220) Google Scholar) and TNFα (16Hotamisligil G.S. J. Intern. Med. 1999; 245: 621-625Crossref PubMed Scopus (707) Google Scholar), and decrease that of an insulin-sensitizing hormone, such as adiponectin (17Yamauchi T. Kamon J. Waki H. Terauchi Y. Kubota N. Hara K. Mori Y. Ide T. Murakami K. Tsuboyama-Kasaoka N. Ezaki O. Akanuma Y. Gavrilova O. Vinson C. Reitman M.L. Kagechika H. Shudo K. Yoda M. Nakano Y. Tobe K. Nagai R. Kimura S. Tomita M. Froguel P. Kadowaki T. Nat. Med. 2001; 7: 941-946Crossref PubMed Scopus (4165) Google Scholar). As shown in Fig. 7, panel 1, supraphysiological activation of PPARγ way beyond that by TZD stimulates adipogenesis, which promotes a flux of FFA from liver and muscle into WAT, leading to a decrease in TG content in liver and muscle and improvement of insulin sensitivity at the expense of increased WAT mass, i.e.obesity. Moreover, TZD induce adipocyte differentiation and apoptosis, thereby increasing the number of small adipocytes, which finally lead to alleviation of insulin resistance presumably via a decrease in molecules causing insulin resistance, such as FFA and TNFα, and up-regulation of insulin-sensitizing hormone adiponectin, at least in part. By contrast, as shown in the Fig. 7, panel 3, moderate reduction of PPARγ activity observed in untreated heterozygous PPARγ-deficient mice decreases TG content in WAT, skeletal muscle, and liver. This effect is due to a combination of increased leptin expression by antagonism of PPARγ-mediated suppression of the gene, thereby reducing expression of lipogenic enzymes, and consequent activation of PPARα pathway in liver, BAT, and skeletal muscle, leading to an increase in expression of UCP2 and enzymes involved in β-oxidation. These observations fit well with the recently demonstrated effects of PPARα agonists on insulin resistance (33Guerre-Millo M. Gervois P. Raspe E. Madsen L. Poulain P. Derudas B. Herbert J.M. Winegar D.A. Willson T.M. Fruchart J.C. J. Biol. Chem. 2000; 275: 16638-16642Abstract Full Text Full Text PDF PubMed Scopus (551) Google Scholar) and decreased fatty acid combustion in PPARα-deficient mice (34Kersten S. Seydoux J. Peters J.M. Gonzalez F.J. Desvergne B. Wahli W. J. Clin. Invest. 1999; 103: 1489-1498Crossref PubMed Scopus (1390) Google Scholar). Moreover, direct antagonism of PPARγ to reduce lipogenesis in WAT prevents adipocyte hypertrophy under the HF diet, thereby reducing the molecules causing insulin resistance, such as FFA and TNFα, and up-regulating the insulin-sensitizing hormone adiponectin, at least in part. These alterations lead to prevention against HF diet-induced obesity and insulin resistance. The data showing that moderate reduction of PPARγ activity resulted in increased insulin sensitivity were further confirmed by the observation that treatment of heterozygous PPARγ-deficient mice with a low dose of TZD caused the re-emergence of insulin resistance (9Kubota N. Terauchi Y. Miki H. Tamemoto H. Yamauchi T. Komeda K. Satoh S. Nakano R. Ishii C. Sugiyama T. Eto K. Tsubamoto Y. Okuno A. Murakami K. Sekihara H. Hasegawa G. Naito M. Toyoshima Y. Tanaka S. Shiota K. Kitamura T. Fujita T. Ezaki O. Aizawa S. Kadowaki T. Mol. Cell. 1999; 4: 597-609Abstract Full Text Full Text PDF PubMed Scopus (1243) Google Scholar). This study has thus revealed the mechanisms whereby both PPARγ agonist and heterozygous PPARγ deficiency have a similar effect on insulin sensitivity. However, it should also be noted that PPARγ agonist and heterozygous PPARγ deficiency have an opposite effect on adiposity and energy expenditure which appear to be more directly regulated by PPARγ activity. Although both heterozygous PPARγ deficiency and PPARγ agonist finally improve insulin resistance via decreased TG content in muscle/liver and prevention of adipocyte hypertrophy, there are some important differences between them. First, although both reduced TG content in muscle/liver, heterozygous PPARγ deficiency did so via activation of fatty acid combustion and energy dissipation, whereas TZD did so via potent stimulation of adipogenesis, thereby increasing fatty acid flux from muscle/liver into WAT. Second, both prevented HF diet-induced adipocyte hypertrophy, and TZD markedly increased the number of newly differentiated small adipocytes, whereas heterozygous PPARγ deficiency appeared not to change the total number of adipocytes. Taken together, all of these differences are consistent with the notion that activation of PPARγ plays a role in energy storage and adiposity, and reduction of PPARγ causes energy dissipation and prevention of adiposity. In conclusion, although by different mechanisms, both heterozygous PPARγ deficiency and PPARγ agonist improve insulin resistance via decreased TG content in muscle/liver and prevention of adipocyte hypertrophy (Fig. 7). We thank S. Uchida, K. Kirii, S. Sakata, and T. Nagano for excellent technical assistance.