Mechanism of Amelioration of Insulin Resistance by β₃-Adrenoceptor Agonist AJ-9677 in the KK-A^y/Ta Diabetic Obese Mouse Model

Chemicals. AJ-9677 ([3-[(2R)-[[2R)-(3-chlorophenyl)-2-hydroxyethyl]amino] propyl]-1H-indol-7-yloxy]acetic acid), a specific β₃-adrenoceptor agonist (β₁:β₂:β₃ = 1:2:210), was synthesized at Dainippon Pharmaceutical (Osaka, Japan).

Animals and drug treatment. Male 9-week-old KK-A^y/Ta and 11-week-old C57BL/6J mice were purchased from Clea Japan (Tokyo, Japan) and were allowed to adjust to the facilities and handling for 1 week before all experimental protocols. Under standardized conditions, the mice were given dry food (CE-2; Oriental Yeast, Tokyo) and water ad libitum. Mice were divided into control and treatment groups so that the mean body weight of the two groups was comparable 1 day before the starting day (day 1). The treatment group was given AJ-9677 by gavage at 0.01 mg/kg or 0.1 mg^-1·kg^-1·day^-1 in the morning for 14 days. The control group was given only 0.5% tragacanth solution. At 24 h after the last administration of AJ-9677, blood samples were collected from the cut ends of the tails and centrifuged to separate plasma. Blood samples from 12-week-old C57BL/6J mice were also collected by the same method. For some mice, the rectal temperature was measured after AJ-9677 administration. Some mice were killed by bleeding under anesthesia 20 h after the final administration of AJ-9677. Epididymal white adipose tissues, inguinal white adipose tissues, interscapular brown adipose tissues, and the gastrocnemius were removed and divided into segments for further studies. The same tissues from 12-week-old C57BL/6J mice were also removed. Some KK-A^y/Ta mice were administered AJ-9677 (0.1 mg/kg) after 24-h fasting to observe the acute effect on lipolytic activity in vivo. Blood samples were collected 1 h after the administration.

Oral glucose tolerance test and determination of plasma glucose, insulin, FFAs, and triglycerides. The oral glucose tolerance test (OGTT) was performed on day 15 after 24 h fasting. The mice received a 20% glucose solution (2 g/kg). Blood samples were collected just before and 0.5, 1, 2, and 3 h after glucose loading. Plasma glucose and insulin levels were determined with a Glucose C-II Test from Wako Pure Chemical Industries (Osaka, Japan) and an insulin enzyme-linked immunosorbent assay (ELISA) from Shibayagi, (Gumma, Japan). FFAs were determined with a nonesterified fatty acid (NEFA) C-Test from Wako Pure Chemical Industries, and Triglycerides were detected with a Determiner triglyceride-S555 from Kyowa Medex Co. (Tokyo, Japan).

Lipolytic activity in white adipocytes. White adipocytes were isolated from epididymal white adipose tissues of KK-A^y/Ta mice by collagenase type II (Sigma Chemical, St. Louis, MO) digestion according to the method described by Rodbell (25). Adipocytes were successively filtered through 1,000-, 500-, and 200-μm meshes and washed in Krebs-Ringer phosphate solution (pH 7.4) containing bovine serum albumin (4 g/100 ml) and glucose (1 mg/ml). Isolated adipocytes were suspended in Krebs-Ringer phosphate solution (pH 7.4) containing bovine serum albumin (1 g/100 ml), glucose (1 mg/ml), phenylisopropyl adenosine (100 nmol/l), and adenosine deaminase (1 U/ml). Adipocytes (∼2.5 × 10⁵ cells/ml) were incubated with AJ-9677 or (-)-isoproterenol for 30 min at 37°C. The reaction was terminated with 1N H₂SO₄ and neutralized with 1N NaOH. FFAs were determined with an NEFA C-Test (Wako Pure Chemical Industries). A quantity of the isolated adipocytes was mixed in chloroformmethanol (3:2), shaken vigorously, and centrifuged at 10,000g. The precipitate was dissolved in 1N NaOH, and protein was determined with a bicinchoninic acid protein assay reagent (Pierce, IL).

Expression of TNF-α and leptin. Measurements of TNF-α and leptin proteins secreted in isolated white adipose tissues (400-1,000 mg) were performed as described previously (26). Adipose tissues were minced into small pieces and incubated in a Krebs-Ringer phosphate solution (pH 7.4) containing endotoxin-free bovine serum albumin (4 g/100 ml) and glucose (1 mg/ml) at 37°C with rotation (100 rpm). After a 2-h incubation, the incubation mixture was centrifuged and the supernatant collected. Concentrations of TNF-α and leptin proteins were determined with a mouse TNF-α ELISA system (Amersham Pharmacia Biotech U.K., Buckinghamshire, U.K.) and a mouse leptin ELISA kit (Morinaga Bioscience Institute, Yokohama, Japan). Total DNA was also isolated from the adipose tissues as described previously (27).

Histological analysis. Small pieces of epididymal and inguinal white adipose tissues were removed and rinsed with saline. The tissues were fixed with 10% formalin and embedded in paraffin. Tissue sections were cut at a thickness of 2.5 μm and stained with hematoxylin and eosin. To examine the size of the white adipocytes, the number of adipocytes was counted in five limited areas (7 × 10^-3 mm²) of each stained specimen. The mean value of the five areas was designated as an index of the cell size (i.e., a larger number means smaller size). In this analysis, the multilocular adipocytes were excluded. For immunohistochemistry, the paraffin-embedded sections were stained with anti-cytochrome oxidase antibody (Molecular Probes, Eugene, OR). The signals were detected by a Vectastain avitin-biotin-peroxidase complex system with a diaminobenzidine substrate kit (Vector Laboratories, Burlingame, CA). For electron microscopic observations, the tissues were fixed with 2.5% glutaraldehyde and embedded in Epon 812.

Triglycerides and DNA in adipose tissues. The method for determining the triglyceride and DNA contents in the epididymal and inguinal white adipose tissues was slightly modified from the method described by Okuno et al. (23). Briefly, 50 mg adipose tissue was homogenized in 2 ml of a solution containing 150 mmol/l sodium chloride, 0.1% Triton X-100, and 10 mmol/l Tris, pH 8.0, at 50°C using a polytron homogenizer (NS-310E; Micro Tech Nichion, Chiba, Japan) for 1 min. The triglyceride content of this homogenized solution was determined by a Determiner triglyceride-S555 (Kyowa Medex). For DNA determination, the homogenized solution was mixed with SDS, proteinase K, and EDTA to final concentrations of 0.1%, 100 μg/ml, and 10 mmol/l, respectively. After a 12-16 h incubation at 37°C, the DNA was extracted by the phenol-chloroform extraction method. The DNA pellets were redissolved in a solution containing 20 μg/ml ribonuclease A, 1 mmol/l EDTA, and 10 mmol/l Tris at pH 8.0. After 20 min of incubation at 37°C, the DNA was re-extracted by the phenol-chloroform extraction method. The DNA pellets were finally dissolved with TE buffer (10 mmol/l Tris and 1 mmol/l EDTA, pH 8.0). DNA content was calculated from the absorbance at 260 nm, with the optical density at 260 nm of 50 μg/ml DNA solution taken as equal to 1.0 by a Gene Quant RNA/DNA Calculator (Amersham Pharmacia Biotech, U.K.).

Expression of mRNAs by reverse transcription-competitive polymerase chain reaction. Extraction of total RNA from tissues was carried out by homogenizing 50 mg frozen tissue in 1 ml RNAzol B (Tel-Test, Friendswood, TX). RNA pellets were obtained from the homogenate by repeated extraction with chloroform and alcohol precipitation. The mRNA of TNF-α, leptin, GLUT4, and three different subtypes of UCP were quantitatively determined by the reverse transcription (RT)-competitive polymerase chain reaction (PCR) method described by Auboeuf and Vidal (28); however, the internal standard was prepared according to the method of Celi et al. (29). Table 1 lists both the sequences of primers used for internal standards and RT-competitive PCR and the sizes of each target fragment and competitor fragment generated by RT-competitive PCR. For construction of single-strand cDNA, 1 μg of total RNA was used with the SuperScript preamplification system (Life Technologies, Gaithersburg, MD) using oligo (dT)_12-18 as a primer. The internal standards (competitors) were prepared by PCR using template cDNA constructed with total RNA from epididymal white or interscapular brown adipose tissues and a specific set of oligonucleotides as primers (i.e., forward-short [S] and reverse) (Table 1). RT-competitive PCR was performed with a constant amount of target cDNA and four different amounts of the corresponding internal standard. The PCR conditions were 20-30 cycles of denaturation at 94°C for 15 s, annealing at 55°C (or 62°C for TNF-α) for 30 s, and extension at 72°C for 30 s. The number of cycles was selected to obtain a measurable amount of products depending on the content of the target. The PCR products were subjected to agarose gel electrophoresis and ethidium bromide staining. After photographs were taken, the density of the DNA band was analyzed with the Kodak Digital Science Electrophoresis Documentation and Analysis System 120 (Eastman Kodak Company, Rochester, NY). The initial amount of the target was calculated from the plot of the density ratio of the competitor band to the target band versus the initial amount of the competitor added in the PCR reaction.

Western blot analysis of GLUT4. Crude and plasma membranes were prepared from frozen epididymal white adipose tissues using differential ultracentrifugation as described by Simpson et al. (36) and Kelada et al. (37). Proteins in crude and plasma membrane fractions were separated by SDS-PAGE on 12.5% gels using the system described by Laemmli (38). The separated proteins were transferred to polyvinylidene difluoride (PVDF) membranes at 10 V for 30 min by the semidry blotting system (Bio-Rad Laboratories, Hercules, CA). The PVDF membranes were blocked with Block Ace (Dainippon Pharmaceutical) and incubated for 1 h with polyclonal anti-GLUT4 antibody (Transformation Research, Framingham, MA). The membranes were subsequently incubated with anti-rabbit IgG antibody conjugated with horseradish peroxidase (Santa Cruz Biotechnology, Santa Cruz, CA) for 1 h. The immunolabeled signals were detected with autoradiography film using the enhanced chemiluminescence system (Amersham Pharmacia Biotech).

Statistical analysis. Values are shown as the means ± SE. The statistical significance of the values was analyzed by the Dunnett's two-tailed test when more than three groups were compared or the two-tailed Student's t test or the Welch t test when two groups were compared.

Effects of AJ-9677 on plasma glucose, insulin, FFA, and triglyceride levels. Control KK-A^y/Ta diabetic obese mice showed marked hyperglycemia, hyperinsulinemia, hypertriglyceridemia, and high FFA levels compared with normal C57BL/6J mice (Table 2). AJ-9677 stimulated lipolysis in vitro dose-dependently with higher potency than isoproterenol (Fig. 1). Acute administration of AJ-9677 also stimulated lipolysis, and plasma FFA levels were increased (0.77 ± 0.05 to 1.63 ± 0.05 mEq/l, P < 0.001, n = 8). However, in mice treated with AJ-9677 (0.1 mg/kg) for 14 days, hyperglycemia and high FFA levels were completely normalized and hyperinsulinemia and hypertriglyceridemia were markedly improved (Table 2). The OGTT indicated that the diabetic obese mice treated with AJ-9677 showed significantly lower glucose levels after glucose loading and significantly lower levels of plasma insulin than the untreated control mice (Fig. 2). These data clearly indicate that the AJ-9677 antidiabetic effect is mediated by the amelioration of insulin resistance in the diabetic obese mouse model. Although AJ-9677 treatment did not alter the total food intake, age-dependent weight gain was reduced in the diabetic obese mice (Table 2). These results suggest that AJ-9677 may inhibit weight gain by increasing energy expenditure. Consistent with this idea, rectal temperature was elevated by ∼1°C (36.8 ± 0.2 to 37.7 ± 0.1°C, P < 0.01, n = 8) 1.5 h after the administration on day 1 of 0.1 mg/kg of AJ-9677.

Effects of AJ-9677 on white adipose tissues. The mice treated with AJ-9677 showed an ∼50% reduction in the weight of epididymal white adipose tissues compared with the control mice (Table 2). Moreover, triglyceride levels in the epididymal white adipose tissues normalized to DNA levels (TG/DNA ratios) were also reduced by ∼50%, to almost the same level as the control levels (Table 2), indicating that the triglyceride content per cell was reduced to normal. TG/DNA ratios in inguinal white adipose tissues were also reduced by ∼50% by AJ-9677 treatment (Table 2).

Histological analysis of white adipose tissues. Treatment with AJ-9677 clearly reduced the size of adipocytes in the epididymal white adipose tissues (Fig. 3A and B. The number of white adipocytes in the limited area was larger in the AJ-9677-treated group than that in the control group. This effect was exerted in a dose-dependent manner (i.e., a 120% increase at a dosage of 0.01 mg/kg and a 160% increase at a dosage of 0.1 mg/kg) (Table 3). Moreover, the AJ-9677 treatment caused the formation of some multilocular adipocytes resembling brown adipocytes. Immunostaining of cytochrome oxidase in the AJ-9677-treated group was stronger than that in the control group (Fig. 3C and D. These results were also found in the inguinal white adipose tissues (data not shown). These observations suggest that AJ-9677 treatment caused the exaltation of mitochondrial function and the conversion of white adipocytes into brown adipocytes. Alternatively, β₃-adrenoceptor agonist treatment may cause differentiation of preadipocytes of brown adipose tissue lineage in white adipose tissues. AJ-9677 treatment caused a dramatic increase in the number of mitochondria in the cytoplasm of inguinal white adipocytes (Fig. 3E and F).

Effect of AJ-9677 treatment on the overexpression of TNF-α and leptin mRNA and protein. The TNF-α mRNA levels were markedly higher in control KK-A^y/Ta mice than in C57BL/6J mice. Treatment with AJ-9677 decreased the TNF-α mRNA level to 35% of the control level (Fig. 4A). This reduction of TNF-α mRNA was associated with a parallel reduction in the amount of TNF-α protein secreted from the adipocytes (Fig. 4C). Very similar data were obtained for leptin mRNA and protein secretion (Fig. 4B and D). AJ-9677 also reduced the mRNA expression levels and secretion of both TNF-α and leptin in the inguinal white adipose tissues (data not shown).

Effect of AJ-9677 treatment on expression of UCP-1, -2, and -3. AJ-9677 treatment caused the mRNA expression of UCP-1 in brown adipose tissues to be increased threefold (Fig. 5A). UCP-2 mRNA levels were increased in the brown adipose tissues of control KK-A^y/Ta mice, and AJ-9677 treatment further upregulated the expression of UCP-2 mRNA. UCP-3 mRNA levels were not changed in the brown adipose tissues of the control KK-A^y/Ta mice, and AJ-9677 treatment did not affect the expression of UCP-3 mRNA. UCP-1 was hardly expressed in the epididymal and inguinal white adipose tissues of both control KK-A^y/Ta and C57BL/6J mice. However, administration of AJ-9677 caused a 20- to 80-fold increase in UCP-1 expression levels in the epididymal (Fig. 5B) and inguinal (Fig. 5C) white adipose tissues. UCP-2 expression levels were increased three- to sixfold in both epididymal and inguinal white adipose tissues of the control KK-A^y/Ta mice, as seen in brown adipose tissues (Fig. 5B and C). However, the expression of UCP-2 mRNA was reduced and left unchanged by AJ-9677 in the epididymal and inguinal white adipose tissues, respectively, unlike the effects seen in brown adipose tissues (Fig. 5B and C). The expression of UCP-3 was not altered in white and brown adipose tissues of the control KK-A^y/Ta mice. Although AJ-9677 treatment decreased UCP-3 expression levels in epididymal white adipose tissues (Fig. 5B), it did not affect UCP-3 expression levels in inguinal white adipose tissues (Fig. 5C). The expression levels of UCP-1, -2, and -3 mRNA in the gastrocnemius were very low compared with the adipose tissues, and they were not upregulated by AJ-9677 treatment (Fig. 5D).

Effect of AJ-9677 treatment on GLUT4 expression. GLUT4 mRNA expression in epididymal white adipose tissues was similar in control KK-A^y/Ta diabetic obese and C57BL/6J mice. AJ-9677 treatment increased GLUT4 mRNA expression approximately fourfold. AJ-9677 treatment also caused a two-to threefold increase in GLUT4 mRNA levels in inguinal white adipose tissues, brown adipose tissues, and the gastrocnemius (Table 4). AJ-9677 treatment also increased GLUT4 protein levels in the crude and plasma membrane of epididymal white adipose tissues four- and sixfold, respectively (Fig. 6).

Administration of AJ-9677 for 14 days reduced plasma glucose, insulin, FFA, and triglyceride levels in the diabetic obese mouse model to almost the normal levels seen in C57BL/6J mice. AJ-9677 stimulated lipolysis in adipocytes dose-dependently in vitro. Although the plasma FFA level increased after the first AJ-9677 administration, this effect was not observed after chronic treatment, presumably because plasma FFAs generated by lipolysis may be consumed by thermogenesis. Although we used the dosage of 0.1 mg^-1·kg^-1·day^-1, AJ-9677 can reduce these plasma parameters at lower dosages (data not shown). The effective dosage was much lower for AJ-9677 than for other β₃-adrenoceptor agonists reported previously (14,39). White adipose tissues in humans show lower expression levels of β₃-adrenoceptor than in rodents. However, because AJ-9677 is an equally potent agonist against both human and rat β₃-adrenoceptors in the Chinese hamster ovary cell expression system, it may show the same kind of effects, to some extent, in humans as it does in rodents.

Mitochondrial UCPs, which cause respiration without ATP synthesis, are believed to be involved in the expenditure of excess energy. At least three UCP isoforms (UCP-1, -2, and -3) have been cloned. UCP-1 is primarily expressed in brown adipose tissues, and it is slightly expressed in white adipose tissues. In the present study, administration of AJ-9677 increased the expression of UCP-1 threefold in brown adipose tissues and 20- to 80-fold in epididymal and inguinal white adipose tissues. The expression of UCP-2 was increased by 2.5-fold in brown adipose tissues, but it was decreased in epididymal white adipose tissues and unaltered in inguinal white adipose tissues. The expression of UCP-3 was not changed in brown adipose tissues and inguinal white adipose tissues, but it was decreased in epididymal white adipose tissues. Previous reports showed that the β₃-adrenoceptor agonists BRL 35135 and CL 316,243 (40,41,42) increased the expression of UCP-1 in both brown and white adipose tissues. Our data are consistent with those results. The effect of AJ-9677 on the expression of UCP-2 in brown adipose tissues is different from that of BRL 35135 (40), but consistent with CL 316,243 (41). The effects on UCP-3 expression are inconsistent. Thus, the effects of β₃-adrenoceptor agonists on UCP-2 and UCP-3 expression may depend on the types of β₃-adrenoceptor agonists, the duration of the treatment, and the time of death. Although CL 316,243 increased UCP-1 mRNA in the gastrocnemius (7), AJ-9677 did not increase UCP-1, 2, or 3 mRNA in the tissues. These differences may also depend on the types of β₃-adrenoceptor agonists. The role of UCPs in the muscles may not be very important for the antidiabetic effects of AJ-9677.

The major novel finding of this study is that the β₃-adrenoceptor agonist AJ-9677 reduced the size of white adipocytes, an effect associated with the reduction of mRNA expression and protein secretion of TNF-α and the reduction of plasma FFA levels. The reduction in size of the white adipocytes was confirmed by histological analysis and the triglyceride/DNA ratio. This reduction may be caused by increased lipolysis and energy expenditure mediated by the increased expression of both UCP-1 and UCP-2 after AJ-9677 treatment. We cannot identify the source of UCP-1 overexpressed in white adipose tissues; it is possible that the increased UCP-1 was expressed in the multilocular adipocytes found in the white adipose tissues during AJ-9677 treatment. However, because the absolute expression level of UCP-1 in white adipose tissues was much lower than in brown adipose tissues, white adipose tissues may contribute less to energy expenditure than brown adipose tissues. Thiazolidinediones cause the differentiation of pre-adipocytes into adipocytes (through the peroxisome proliferator—activated receptor γ [PPAR-γ]) to generate small white adipocytes and, concomitantly, cause the apoptosis of large white adipocytes (23). This reduction of the mean size of adipocytes by thiazolidinediones is associated with the normalization of both increased levels of expression of TNF-α and increased production of FFAs in diabetic obese animal models. Because the increase in the proportion of small adipocytes by thiazolidinediones appears to contribute to the amelioration of insulin resistance via decreased TNF-α expression and decreased FFA production, the reduction in the size of adipocytes by AJ-9677 may contribute to the amelioration of insulin resistance via a similar reduction of TNF-α expression and FFA production. Moreover, the reduction in the size of adipocytes by AJ-9677 may also be associated with decreased levels of leptin expression and secretion in white adipose tissues, which may well reflect the abrogation of the need for compensation for insulin resistance.

TNF-α can cause downregulation of GLUT4 expression (43,44), so decreased TNF-α expression probably contributes to increased GLUT4 expression after AJ-9677 treatment. Moreover, GLUT4 expression on the plasma membrane was preferentially increased compared with the crude membrane, suggesting that AJ-9677 treatment stimulated GLUT4 translocation to the plasma membrane in white adipose tissues. Whether this effect is a direct action of AJ-9677 or secondary to its amelioration of high plasma FFA levels is unclear at present. In either case, the change in the localization of GLUT4 to the plasma membrane, in addition to the increased GLUT4 expression levels, may further contribute to the amelioration of insulin resistance and diabetes.

We would like to propose a novel hypothesis for the relationship between the generation of small adipocytes and the amelioration of insulin resistance, as depicted in Fig. 7. The present study clearly shows that the β₃-adrenoceptor agonist AJ-9677 converted large adipocytes into small adipocytes through increased lipolysis and the induction of UCPs (e.g., UCP-1) that are associated with the reduction of TNF-α and FFA levels and thus the amelioration of insulin resistance. We previously showed that thiazolidinediones promote adipocyte differentiation to generate small adipocytes, which is also associated with reduction of TNF-α and FFA levels, leading to the amelioration of insulin resistance (23). Therefore, insulin sensitivity is induced in vivo whether small adipocytes are generated de novo by differentiation via stimulation of PPAR-γ or by conversion from the large adipocytes via β₃-adrenoceptors. Thus, we would like to propose that the generation of small adipocytes is the key event in the amelioration of insulin resistance.

Hypertrophic obesity resulting from adipocyte hypertrophy, which develops with a high-fat diet and sedentary lifestyle, is closely linked to major health issues (e.g., diabetes, hypertension, hyperlipidemia, and cardiovascular diseases) in Western countries and in Japan (45). Insulin resistance, which is usually associated with hypertrophic obesity, is believed to be the major mechanism causing these diseases. Thus, the treatment of hypertrophic adipocytes is one of the most important issues in medical science. Thiazolidinediones may not be the ideal agents because the number of adipocytes tends to increase and promote obesity, even though insulin resistance is ameliorated. This study clearly demonstrated that β₃-adrenoceptor agonists (e.g., AJ-9677) appear to convert hypertrophic adipocytes into small adipocytes, thereby ameliorating insulin resistance and obesity simultaneously.