Peroxisome Proliferator–Activated Receptor (PPAR)α Activation Increases Adiponectin Receptors and Reduces Obesity-Related Inflammation in Adipose Tissue: Comparison of Activation of PPARα, PPARγ, and Their Combination

Rosiglitazone was synthesized as described elsewhere (18). Wy-14,643 was purchased from Biomol Research Laboratories (Plymouth Meeting, PA). All materials were obtained from sources described previously (8,9,19).

We purchased 6-week-old male KKAy mice and age-matched KK mice from Nippon CLEA (Shizuoka, Japan). KKAy mice are an obese diabetic model in which the Ay mutation is introduced onto a KK strain background. Therefore, we used KK mice as nondiabetic controls. Mice were housed singly and maintained on a 12-h light/dark cycle. The high-fat diet consisted of 32% (wt/wt) fat as described previously (20). KKAy mice were given the high-fat diet, and KK mice were given normal chow. High-fat feeding was begun 1 week before rosiglitazone or Wy-14,643 administration. Rosiglitazone or Wy-14,643 was given as a 0.01 and 0.05% food admixture, respectively. These doses were chosen because they have been shown to be effective therapeutic doses in diabetic mice (21–24). The same amount of food was given to the pair-fed group of KKAy mice as that fed to the Wy-14,643–treated group of KKAy mice. The animal care procedures and methods were approved by the animal care committee of the University of Tokyo.

The glucose tolerance and insulin tolerance tests were carried out according to previously described methods (9). Plasma glucose, serum free fatty acid, and triglyceride levels were determined by a glucose test, nonesterified fatty acid-C test, and triglyceride L-type (Wako Pure Chemical Industries, Osaka, Japan), respectively. Plasma insulin was measured by an insulin immunoassay (Shibayagi, Gunma, Japan). Plasma leptin and adiponectin levels were determined by a Quintikine M kit (R&D Systems, Minneapolis, MN) and mouse adiponectin immunoassay kit (Otsuka Pharmaceutical, Tokushima, Japan), respectively (20). Detection of multimer species of adiponectin was conducted as described previously (25). In brief, 0.7 μl of serum was subjected to 2–15% SDS-PAGE under nonreducing and non–heat-denaturing conditions. Adiponectin was detected using anti–globular domain antiserum obtained by immunizing rabbits with mouse recombinant adiponectin globular domain produced in Escherichia coli (25).

Epididymal adipose tissue was removed from each animal, fixed in 10% formaldehyde/PBS, and maintained at 4°C until use. Fixed specimens were dehydrated, embedded in tissue-freezing medium (Tissue-Tek OCT compound; Miles, Elkhart, IN), 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 (7) or anti-mouse F4/80 antibody (Serotec, Raleigh, NC) (13).

Total RNA was prepared from cells or tissues with TRIzol (Invitrogen, Carlsbad, CA) according to the manufacturer’s instructions. A real-time PCR method was used to quantify the AdipoRs mRNAs (19). The primer sets and the probes for mAdipoR1 and -R2 were as follows: the forward primer for mAdipoR1 was acgttggagagtcatcccgtat, the reverse primer ctctgtgtggatgcggaagat, and the probe cctgctacatggccacagaccacct; the forward primer for mAdipoR2 was tcccaggaagatgaagggtttat, the reverse primer ttccattcgttcgatagcatga, and the probe atgtccccgctcctacaggccc. For quantification of the other genes, we used a set of predesigned primers and a probe for each gene (Assay on Demand; Applied Biosystems, Foster City, CA). The relative amount of each transcript was normalized to the amount of β-actin transcript in the same cDNA (19).

Isolation of adipocytes and stromal-vascular cells from adipose tissues were performed as described previously, with slight modification (13). Epididymal adipose tissues were isolated from mice and were minced into fine pieces. Minced tissues were incubated in Dulbecco’s modified Eagle’s medium supplemented with 5 mg/ml collagenase (Sigma, St. Louis, MO) and 2.5% BSA (Sigma) at 37°C for 40–45 min. Digested samples were passed through a sterile 250-μm nylon mesh and centrifuged. The pellet cells and the floating cells were washed twice with PBS and collected as the stromal-vascular cells and adipocytes, respectively. In the experiments of direct action of PPAR agonists on primary adipocytes and stromal-vascular cells, 1 × 10⁵ isolated adipocytes were seeded onto 24-well plates in Dulbecco’s modified high-glucose Eagle’s medium (DMEMH) supplemented with 20% charcoal-treated FCS, and 1 × 10⁵ isolated stromal-vascular cells were seeded onto 96-well plates in DMEMH supplemented with 20% charcoal-treated FCS and mouse granulocyte colony-stimulating factor (R&D Systems). Then, the cells were incubated with 3 μmol/l Wy-14,643 or 0.3 μmol/l rosiglitazone for 24 h and then harvested to isolate total RNA.

Mouse 3T3-L1 cells were grown in DMEMH supplemented with 10% FCS. Induction of adipogenic differentiation was carried out according to a method described previously (8). By day 8, 3T3-L1 adipocytes were incubated with 30 μmol/l Wy-14,643 or 0.3 μmol/l rosiglitazone in DMEMH supplemented with 1% BSA for 18 h, and then 1 ng/ml TNF-α was added. The cells were harvested 6 h after TNF-α addition to isolate total RNA. Peritoneal macrophages were isolated as previously described (26). In brief, macrophages were isolated 4 days after intraperitoneal injection of 3 ml thioglycolate medium (Sigma) to male C57BL/6j mice. Then, 1,000,000 cells were plated onto 24-well plates in RPMI-1640/10% FBS (vol/vol) and incubated for 4 h. Then, the macrophages were incubated with 30 μmol/l Wy-14,643, 100 μmol/l fenofibrate, or 0.3 μmol/l rosiglitazone in the presence of 0.1 μg/ml lipopolysaccharide for 24 h and then harvested to isolate total RNA. The concentrations of the compounds in the in vitro cell culture experiments were chosen according to previous studies, and they were comparable to those to which PPARα in animal tissues were exposed in previous in vivo studies (8,10–12,27,28).

Data are the means ± SE. Student’s t test was used for statistical comparison. P < 0.05 was considered statistically significant.

To examine the antidiabetic effects of a dual activation of PPARα and -γ in comparison with a single activation of each alone, we treated obese diabetic KKAy mice with a PPARγ agonist, rosiglitazone, a PPARα agonist, Wy-14,643, or a combination of the two for 8 weeks. We then examined the glucose and lipid metabolism and insulin sensitivity of these mice. Because it has been previously reported that PPARα agonists reduce food intake in rodents (28), pair-fed control mice were given daily the same amount of food as that consumed by Wy-14,643–treated mice (Fig. 1F). As shown in Fig. 1, Wy-14,643 treatment significantly ameliorated hyperglycemia (Fig. 1A), hyperinsulinemia (Fig. 1B), and hyperlipidemia (Fig. 1C and D) compared with ad libitum–fed vehicle-treated KKAy mice. Because these effects of Wy-14,643 were significant even in comparison with pair-fed mice (Fig. 1A–D), it is likely that Wy-14,643 actually exerted its antidiabetic effects via a mechanism other than decreased food intake. Blood glucose (Fig. 1A) and plasma insulin levels (Fig. 1B) in the rosiglitazone-treated mice were significantly lower than those in the ad libitum–fed vehicle-treated mice. However, the serum lipid levels of the rosiglitazone-treated mice were significantly higher than those of the Wy-14,643–treated mice (Fig. 1C and D). Interestingly, a combination of rosiglitazone and Wy-14,643 completely normalized the hyperglycemia (Fig. 1A), hyperinsulinemia (Fig. 1B), and hyperlipidemia (Fig. 1C and D) observed in KKAy mice, and, in particular, the blood glucose (Fig. 1A) and lipid levels (Fig. 1C and D) in the combined rosiglitazone- and Wy-14,643–treated mice were significantly lower than those in the rosiglitazone-treated mice. Furthermore, serum lipid levels in the combined rosiglitazone- and Wy-14,643–treated mice were significantly lower than those in the Wy-14,643–treated mice (Fig. 1C and D). Although body weight was slightly lower in the Wy-14,643–treated mice (P = 0.093) than in the pair-fed vehicle-treated mice, the difference was not statistically significant (Fig. 1E).

We next examined the effects of PPARα and -γ activation on the improvement of insulin resistance in more detail, using a glucose tolerance test and insulin tolerance test. Blood glucose levels during both tests in Wy-14,643–treated mice were significantly lower than those in pair-fed vehicle-treated mice (Fig. 2A and C), suggesting that Wy-14,643 treatment ameliorated insulin resistance. Again, the combination of rosiglitazone and Wy-14,643 ameliorated insulin resistance more effectively than rosiglitazone or Wy-14,643 alone (Fig. 2B and C).

It has been reported that PPARα agonists prevent high-fat diet–induced obesity (10). Thus, we measured epididymal and subcutaneous WAT and intrascapular brown adipose tissue (BAT) weights. Wy-14,643 decreased both WAT and BAT weights compared with pair-fed mice (Fig. 3A–C), suggesting that PPARα activation prevents obesity. The combination of rosiglitazone and Wy-14,643 decreased only epididymal WAT weights compared with pair-fed mice (Fig. 3A). In contrast, rosiglitazone increased subcutaneous WAT and BAT weights compared with vehicle (Fig. 3B and C).

We have previously shown that PPARγ activation by its agonist or a moderate reduction of PPARγ activity prevents adipocyte hypertrophy, which results in an improvement in insulin resistance (7). We next attempted to clarify whether adipocyte hypertrophy and increased fat pad weight are suppressed by PPARα activation. The size of the adipose cells in epididymal WAT was increased in KKAy mice (Fig. 4A, upper left) compared with the wild-type control KK mice (Fig. 4A, lower right). Very interestingly, the size of the adipose cells in epididymal WAT from Wy-14,643–treated mice (Fig. 4A, upper right) was dramatically decreased compared with that from pair-fed mice (Fig. 4A, lower middle) and was comparable to that in the wild-type control KK mice (Fig. 4A, lower right). Although rosiglitazone treatment (Fig. 4A, upper middle) increased the ratio of small adipocytes in epididymal WAT compared with vehicle treatment (Fig. 4A, upper left), the changes were moderate compared with Wy-14,643 treatment (Fig. 4A, upper right). The size of the adipose cells in epididymal WAT from mice treated with a combination of rosiglitazone and Wy-14,643 (Fig. 4A, lower left) was also smaller than that from pair-fed mice (Fig. 4A, lower middle). The size of the adipose cells in subcutaneous WAT from Wy-14,643–treated mice was also decreased compared with that from pair-fed mice (Fig. 4B). There were small nucleated cells and macrophage-specific antigen F4/80–expressing cells in the interstitial spaces between adipocytes in WAT of vehicle-treated KKAy mice (Fig. 4C). In contrast, there were almost no such cells in the WAT of Wy-14,643–treated mice (Fig. 4C), suggesting that macrophage infiltration to WAT may be suppressed by Wy-14,643 treatment, whereas the effect of rosiglitazone treatment seemed to be faint. We obtained similar results with BAT (Fig. 4D), except that the size of the adipocytes in BAT from rosiglitazone-treated mice (Fig. 4D, upper middle) was larger than that in vehicle-treated mice (Fig. 4D, upper left).

We next studied whether the levels of expression of molecules secreted from WAT that regulate insulin sensitivity were changed by PPARα activation. As reported previously (9), serum adiponectin levels in rosiglitazone-treated mice were higher by fourfold than those in vehicle-treated mice (Fig. 3D). In contrast, serum adiponectin levels in Wy-14,643–treated mice were slightly lower than those in pair-fed mice, suggesting that the improvement in insulin resistance by Wy-14,643 was not caused by increased gene expression or secretion of adiponectin (Fig. 3D). The combination of rosiglitazone and Wy-14,643 increased serum adiponectin levels approximately threefold above the vehicle (Fig. 3D). Serum leptin levels in KKAy mice were increased by overexpression of agouti compared with those in wild-type KK mice (Fig. 3E). Wy-14,643 treatment decreased the serum leptin levels to levels comparable to wild-type KK mice (Fig. 3E). Although rosiglitazone treatment slightly (P = 0.057) increased serum leptin levels, combined rosiglitazone and Wy-14,643 treatment decreased them compared with vehicle (Fig. 3E).

PPARα is abundantly expressed in liver and BAT in rodents, and PPARα activation induces fatty acid combustion (29,30). Thus, we examined the gene expression of molecules involved in fatty acid oxidation and lipolysis, using quantitative PCR analyses. As shown in Fig. 5B and C, the expression of uncoupling protein (UCP)-1 and β3-adrenergic receptor in BAT were decreased in vehicle-treated KKAy mice compared with wild-type KK mice, consistent with the adipocyte hypertrophy observed in KKAy mice (Fig. 5C). Wy-14,643 treatment increased the expression of UCP2 in liver (Fig. 5A) and the expression of UCP1 and β3-adrenergic receptor in BAT (Fig. 5B and C). In contrast, rosiglitazone treatment decreased the expression of UCP1 and β3-adrenergic receptor in BAT compared with vehicle (Fig. 5B and C). Combined treatment with rosiglitazone and Wy-14,643 increased the expression of UCP2 (Fig. 5A), but it did not affect the expression of UCP1 and β3-adrenergic receptor in BAT compared with vehicle (Fig. 5B and C). Taken together, these findings suggest that the induction of molecules involved in fatty acid combustion in liver and BAT is, at least in part, responsible for the prevention of adipocyte hypertrophy by PPARα activation.

Recently, it was proposed that obesity-related insulin resistance may be a chronic inflammatory disease initiated in adipose tissue (13,14). Although it has been reported that rosiglitazone suppressed the expression of inflammation genes in WAT of obese diabetic ob/ob mice (14), the effects of PPARα agonists on inflammation in WAT have not yet been studied in detail. We studied whether Wy-14,643 suppressed the expression of inflammatory genes in WAT. As shown in Fig. 5D–F, the expression of TNF-α, monocyte chemotactic protein (MCP)-1, and macrophage antigen-1 in WAT of vehicle-treated KKAy mice were significantly increased compared with wild-type KK mice, suggesting that the accumulation of macrophages and inflammatory responses were induced in WAT of KKAy mice. Interestingly, Wy-14,643 treatment suppressed the increased expression of TNF-α, MCP-1, and macrophage antigen-1 in WAT compared with vehicle (Fig. 5D–F), which was consistent with the suppression of macrophage infiltration into WAT by Wy-14,643 treatment (Fig. 4C). Rosiglitazone treatment partially suppressed the expression of MCP-1 in WAT (Fig. 5F), but it did not affect the expression of TNF-α and macrophage antigen-1 compared with vehicle (Fig. 5D and E). Combined rosiglitazone and Wy-14,643 treatment, as well as Wy-14,643 treatment, suppressed the expression of these genes compared with vehicle (Fig. 5D–F).

It has been reported that MCP-1 decreased the expression of genes involved in adipocyte function and attenuated insulin sensitivity in cultured adipocytes (31). Thus, we examined the effects of PPARα activation on the expression of lipoprotein lipase and β3-adrenergic receptor in WAT. The expression of lipoprotein lipase and β3-adrenergic receptor were decreased in WAT of vehicle-treated KKAy mice compared with wild-type KK mice (Fig. 5G and H). Wy-14,643 treatment increased the expression of lipoprotein lipase and β3-adrenergic receptor in WAT, whereas rosiglitazone treatment showed no effect (Fig. 5G and H). Taken together, these data suggest that Wy-14,643 suppressed inflammation in WAT and normalized gene expression, which were dysregulated by obesity-associated adipocyte hypertrophy and inflammation.

We previously reported that AdipoR1 and -2 are downregulated in WAT, BAT, and skeletal muscles in obese diabetic ob/ob mice, which are correlated with decreased adiponectin sensitivity (32). As shown in Fig. 6A and B, the expression of AdipoR1 and -2 was decreased in WAT of vehicle-treated KKAy mice compared with wild-type KK mice. Wy-14,643 treatment almost completely reversed the decrease in AdipoR1 and -2 expression in KKAy mice (Fig. 6A and B). In contrast, the expression levels of AdipoR1 and -2 were not affected by rosiglitazone treatment, whereas combined rosiglitazone and Wy-14,643 treatment partially reversed the decrease in these genes (Fig. 6A and B). We also obtained similar results in BAT (data not shown).

Adiponectin is known to form characteristic multimers (33,34). It was recently reported that the increase in the ratio of the high molecular weight (HMW) form to total adiponectin is correlated with an improvement in insulin sensitivity by TZD treatment (35). Here, we studied whether a PPARα agonist affects the forms of serum adiponectin. As shown in Fig. 6C, the ratio of HMW to total adiponectin was decreased in vehicle-treated KKAy mice compared with wild-type KK mice. The restriction of food intake by pair-fed mice partially restored the decrease in the ratio of HMW to total adiponectin in KKAy mice (Fig. 6C). Wy-14,643 treatment did not affect the ratio of HMW to total adiponectin, whereas rosiglitazone treatment dramatically increased total adiponectin and the ratio of HMW to total adiponectin (Fig. 6C). Treatment with a combination of rosiglitazone and Wy-14,643 showed results similar to rosiglitazone treatment (Fig. 6C).

There is a possibility that long-term treatment with Wy-14,643 indirectly increased the gene expression in the WAT of KKAy mice via an improvement in insulin resistance or some other mechanism. Therefore, we examined the effects of short-term treatment with Wy-14,643 on the expression of AdipoR1, AdipoR2, MCP-1, and β3-adrenergic receptor in WAT of KKAy mice. As shown in Fig. 7A–D, treatment with Wy-14,643 for 3 days increased the expression of AdipoR1, AdipoR2, and β3-adrenergic receptor and decreased the expression of MCP-1 in WAT of KKAy mice. We obtained similar results with the combined treatment of rosiglitazone and Wy-14,643 (Fig. 7A–D).

WAT of obese diabetic mice is thought to consist of adipocytes and stromal vascular cells, such as infiltrated macrophages (13,14). We next tried to determine in which cell type the observed effects occur by analyzing the adipocyte and stromal-vascular cell subfractions of adipose tissue separately. As shown in Fig. 7F and G, treatment with Wy-14,643 for 3 days increased the expression of AdipoR2 and decreased the expression of MCP-1 in both adipocytes and stromal-vascular cells, whereas rosiglitazone treatment showed similar effects only in adipocytes. The expression pattern of leptin and CD68, markers of adipocytes and stromal-vascular cells, respectively, confirmed that the fractionation was performed properly (Fig. 7H and I).

We next examined the expression of PPARα in the adipose tissue fractions, cultured 3T3-L1 adipocytes, and peritoneal macrophages to determine whether the effects of Wy-14,643 on adipocytes and macrophages were mediated by PPARα. PPARα expression was observed in both the adipose fraction and 3T3-L1 adipocytes as well as in skeletal muscle of KKAy mice (Fig. 8A), whereas PPARα expression levels in stromal-vascular cells and peritoneal macrophages were much lower than those in skeletal muscle of KKAy mice (Fig. 8A). We further examined the direct action of Wy-14,643 on gene expression in cultured adipocytes or macrophages. A 24-h treatment with 30 μmol/l Wy-14,643 significantly increased AdipoR2 expression in 3T3-L1 adipocytes (Fig. 8C), whereas AdipoR1 expression was not significantly increased (Fig. 8B). Wy-14,643 treatment suppressed the expression of MCP-1, which was increased by 1 ng/ml TNF-α stimulation in 3T3-L1 adipocytes (Fig. 8D). Fenofibrate treatment (100 μmol/l) also increased AdipoR2 expression by 25% compared with vehicle in 3T3-L1 adipocytes (data not shown). Wy-14,643 treatment as well as fenofibrate treatment suppressed the increase in MCP-1 expression caused by 0.1 μg/ml lipopolysaccharide treatment and slightly but significantly increased AdipoR2 expression in peritoneal macrophages (Fig. 8F and G). Rosiglitazone treatment increased AdipoR2 expression only in 3T3-L1 adipocytes (Fig. 8C) and suppressed MCP-1 expression in both adipocytes and macrophages (Fig. 8D and G), whereas the effect of Wy-14,643 treatment on macrophages was more potent than that of rosiglitazone treatment (Fig. 8G). Furthermore, we examined the direct action of the lower concentration of Wy-14,643 on gene expression in primary cultured adipocytes or stromal-vascular cells. A 24-h treatment with 3 μmol/l Wy-14,643 significantly increased AdipoR2 expression in adipocytes and stromal-vascular cells (Fig. 8I), whereas AdipoR1 expression was not significantly increased (Fig. 8H). These data suggest that Wy-14,643 directly increased AdipoR2 expression in adipocytes and suppressed MCP-1 expression in both adipocytes and macrophages.

Recent studies have revealed that adipose tissue plays a key role in regulating whole-body glucose metabolism (13,14,36). In this study, we attempted to clarify the mechanisms by which activation of PPARα and/or PPARγ ameliorates insulin resistance, focusing on adipose tissue. We demonstrate here that activation of PPARα prevented adipocyte hypertrophy, at least in part, through mechanisms other than decreased food intake. As shown in Fig. 9, one possible and conventional mechanism for this effect is an increase in systemic energy expenditure by the induction of molecules involved in fatty acid combustion, such as UCPs in liver and BAT (Fig. 5A–C) (27,37,38). We propose that there is also another pathway in which the activation of PPARα in WAT normalizes adipocyte function associated with an increase in β3-adrenergic receptor, which is expressed in WAT and BAT and has been shown to play important roles in lipolysis and thermogenesis (39,40). WAT has not been considered as a main target tissue of PPARα agonist so far (29,30). PPARα mRNA was, however, actually detected by real-time PCR analysis in the adipocyte subfraction of WAT and 3T3-L1 adipocytes (although the expression levels were much lower in adipocytes compared with liver), as well as in skeletal muscle (Fig. 8A). In addition, it was recently reported that PPARα agonists, such as Wy-14,643, directly enhance lipolysis in isolated adipocytes (41). Thus, it is possible that high concentrations of a PPARα agonist can directly activate PPARα in adipocytes. It remains to be determined which tissue mainly contributes to the improvement in insulin resistance by PPARα activation. To clarify this point, it is necessary to investigate the effects of PPARα agonists in PPARα tissue–specific knockout mice.

It was recently reported that a PPARα/γ agonist elevated pyruvate dehydrogenase kinase isozyme 4 (PDK4) mRNA levels in liver compared with vehicle or a PPARγ agonist (42). In the present study, we checked PDK4 mRNA levels in liver and epididymal adipose tissue and found that PDK4 mRNA levels in epididymal adipose tissue of Wy-14,643–treated KKAy mice were significantly higher than those of vehicle-treated KKAy mice (data not shown), indicating that PPARα was actually activated in adipose tissue by Wy-14,643 treatment. However, there was no significant change in the PDK4 mRNA level in liver between vehicle- and Wy-14,643–treated KKAy mice (data not shown). Precise reasons that explain differences between the results of their study and those of our current study are not known at present. However, several factors, such as animal species (rats and mice), treatment periods (1 and 8 weeks), administration route (oral gavage and mixed chow), and sampling time after the final dose (6 and 24 h) are different between the two studies, which may cause differences in gene expression of PDK4 in liver.

Here we have shown that activation of PPARα suppressed obesity-induced increases in inflammatory cytokines such as TNF-α and MCP-1 in WAT. To the best of our knowledge, this is the first report indicating that the activation of PPARα regulates inflammation in WAT, whereas it has been reported that TZDs also suppressed the increased expression of inflammatory genes in WAT of ob/ob mice (14). In the current study, the efficacy of the PPARα agonist appeared to be more potent than that of the PPARγ agonist with respect to suppression of the increased expression of inflammatory molecules in vivo. Although PPARα agonists have been shown to inhibit a nuclear factor-κB pathway stimulated by proinflammatory substances in smooth muscle cells (43,44), the distinct mechanism by which PPARα agonists suppress the expression of inflammatory cytokines in WAT remains to be clarified. However, in this study, a PPARα agonist directly suppressed the increase in MCP-1 expression by proinflammatory substances in cultured adipocytes or macrophages in vitro. Therefore, it is plausible that PPARα can mediate the reduction of proinflammatory substance–induced increase in MCP-1 expression in both adipocytes and infiltrated macrophages in WAT, leading to a reduction of macrophage accumulation and a reversal of adipocyte dysfunction.

Finally, we found that activation of PPARα increased AdipoR1 and -2 expression in both WAT and BAT in vivo. In contrast, activation of PPARα increased only AdipoR2 in cultured cells. This result is consistent with previous studies that found PPARα agonists increased only AdipoR2 expression in cultured macrophages (45). Furthermore, the expression of AdipoR1 and -2 in the liver from Wy-14,643–treated KKAy mice were not significantly changed compared with those from vehicle-treated mice (data not shown). Therefore, cofactors that are expressed in adipocytes and macrophages and bind with PPARα may be involved in the regulation of AdipoR2 expression. Activation of PPARγ also increased the expression of AdipoR2 in adipocytes in vitro and in vivo, but it did not affect the expression of AdipoR2 in either WAT or BAT in vivo because an increase of AdipoR2 expression in stromal-vascular cells in adipose tissues was not observed after PPARγ activation. The increase of AdipoR2 expression in adipocytes may be caused by the secondary effect on adipocyte differentiation by PPARγ activation because AdipoR2 expression was dramatically increased during adipocyte differentiation (T.Y., T.K., unpublished data). It was recently reported that adiponectin inhibits lipopolysaccharide-induced inflammatory responses in adipocytes (46). Therefore, the increase of AdipoR expression by PPARα activation in adipocytes may enhance adiponectin’s anti-inflammatory action in WAT. To date, however, there are no data indicating to what extent increases in AdipoR1 and -2 expression contribute to the improvement in insulin resistance by PPARα activation. Further studies using AdipoR knockout mice will be required to clarify this point. We also showed that activation of PPARγ or food restriction increased the ratio of HMW to total adiponectin and that activation of PPARα did not affect the ratio. This result indicates that an improvement in adipocyte hypertrophy or a reduction in body weight was sufficient to increase the ratio of HMW to total adiponectin. Further investigations will be important to clarify how PPARγ agonists increase HMW adiponectin because HMW adiponectin may be the active form of this protein with respect to its glucose-lowering effect (35) and in the activation of AMP kinase (T.Y., T.K., unpublished data).

In conclusion, as shown in Fig. 9, we have proposed novel mechanisms by which the activation of PPARα and -γ can improve obesity-induced insulin resistance. First, activation of PPARα suppresses inflammation and adipose hypertrophy in WAT. Second, dual activation of PPARα and -γ enhances the action of adiponectin by increasing AdipoR expression and the ratio of HMW to total adiponectin.

This work was supported by the Program for Promotion of Fundamental Studies in Health Sciences of the Organization for Pharmaceutical Safety and Research of Japan; a grant from the Human Science Foundation (to T.K.); a grant-in-aid for the development of innovative technology from the Ministry of Education, Culture, Sports, Science and Technology of Japan (to T.K.); Grant-in Aid for Creative Scientific Research 10NP0201 from the Japan Society for the Promotion of Science (to T.K.); and health science research grants (research on human genome and gene therapy) from the Ministry of Health, Labor and Welfare of Japan (to T.K.).

We are grateful to A. Itoh and A. Okano for their excellent technical assistance.