WO2004005550A2

WO2004005550A2 - Method of screening for antidiabetic agents

Info

Publication number: WO2004005550A2
Application number: PCT/JP2003/008523
Authority: WO
Inventors: Takao Fujimura; Hiroyuki Sakuma; Chiaki Kimura; Tomoya Oe; Ichiro Aramori
Original assignee: Fujisawa Pharmaceutical Co., Ltd.
Priority date: 2002-07-04
Filing date: 2003-07-04
Publication date: 2004-01-15
Also published as: WO2004005550A3; AUPS337802A0

Abstract

A method of screening for an antidiabetic agent that lowers blood glucose level without inducing obesity and/or insulin resistance, is provided. The method comprises measuring the expression level of a nuclear peroxisome proliferator-activated receptor g (PPARg)-induced adipogenic gene in mature adipocytes in the presence of a test compound or a PPARg full agonist and selecting a compound that results in an expression level significantly lower than that measured using the PPARg full agonist. The method may further comprise the same steps except for using differentiating adipocytes to select a compound that results in an expression level comparable to or higher than that measured using the PPARg full agonist. A kit used for the above screening method, a compound obtainable by the screening method, a pharmaceutical composition containing such a compound, and a method lowering the blood glucose level without causing obesity and/or insulin resistance using the compound, are also provided.

Description

METHOD OF SCREENING FOR ANTIDIABETIC AGENTS

Technical Field

This invention relates to a method of screening for an antidiabetic agent that lowers blood glucose level without inducing obesity and/or insulin resistance.

Background Art

The nuclear peroxisome proliferator-activated receptor γ (PPARγ) is a I igand-dependent transcriptional factor that is important for the regulation of adipogenesis and glucose ho eostasis. When a specific 1 igand binds to PPARγ, PPARγ undergoes conformational changes, which promotes recruitment of transcriptional coactivators or corepressors, and regulates transactivation of the genes critical for adipocyte differentiation and glucose metabol ism. Recent studies showed that PPARγ plays dual roles in the regulation of insulin sensitivity; one role is adipocyte differentiation that contributes to insul in sens i tizat ion (Gerhold, D.L. et al., Endocrinology 143(6) :2106-2118, 2002) and the other is adipocyte hypertrophy that leads to obesity-induced insulin resistance (Kubota, N. et al., Molecular Cell 4:597-609, 1999).

Type 2 (non-insul in dependent) diabetes mel I itus (NIDDM) is characterized by decreased insul in sensi tivi ty in peripheral tissues. Thiazol idinedione (TZD) compounds such as rosigl i tazone and pioglitazone are antidiabetic compounds that improve insul in sensi tivi ty and are used widely in the treatment of type 2 diabetes. These compounds are known to have activities as full agonists for PPARγ. Upon binding to the protein, each of these compounds activates PPARγ to express a lot of PPARγ-induced genes involved in glucose/I ipid metabolism, and is expected to exert therapeutic effect of an antidiabetic agent. In general, TZD compounds show excellent glucose-lowering activities through the activation of PPARγ. However, excess stimulation of PPARγ can often cause adipocyte hypertrophy that leads to obesity and insulin resistance. NIDDM patients often suffer from obesity, and progress of obesity in these patients must be avoided. Thus, development of an antidiabetic agent that lowers blood glucose level without inducing obesity and/or insulin resistance has been desired.

Several methods are known for screening for PPARγ agonists/antagonists that can be used for treating diabetes and obesity. W099/10532 discloses the use of a reporter gene assay for identifying PPARγ agonists/antagonists by detecting changes in I igand-dependent interaction between PPARγ and coactivators in the presence of a test compound. W001/30343 discloses a method of selecting a compound that could be used as an anti-obesity agent by measuring the PPARγ antagonism or partial agonis of candidate compounds compared to a PPARγ full agonist. The selection can be made by known methods including the homogeneous time-resolved fluorescence (HTRF) assay detecting PPARγ-CBP complex formation, GAL4 chimeric receptor transcriptional assay, and mouse 3T3-L1 pre-adipocyte differentiation assay. However, these methods have not so far provided a method of screening for an effective antidiabetic agent that can lower blood glucose level wi thout inducing obesity and/or insulin resistance.

Disclosure of the Invention

An objective of this invention is to provide a method of screening for a candidate compound for an antidiabetic agent that lowers blood glucose level without inducing obesity and/or insulin resistance. Another objective of this invention is to provide a kit for the above screening method.

A further objective of this invention is to provide a compound obtainable by the above screening method, or its pharmaceutically acceptable salt.

A further objective of this invention is to provide a pharmaceutical composition comprising a compound obtainable by the above screening method, or its pharmaceutically acceptable salt, and a pharmaceutically acceptable carrier.

A still further objective of this invention is to provide a method for lowering the blood glucose level without causing obesi ty and/or insul in resistance by administering the above compound, or its pharmaceutically acceptable salt, to a subject in need thereof.

The present inventors found that a selective PPARγ modulator, which behaves as a PPARγ ful I agonist in differentiating adipocytes and behaves as a PPARγ partial agonist in mature adipocytes, can be a potential candidate for an antidiabetic agent that lowers the blood glucose level without causing obesity and/or insul in resistance.

The above objectives can be achieved by the following inventions:

A method of screening for a candidate compound for an antidiabetic agent that lowers blood glucose level without inducing obesi ty and/or insul in resistance, wherein said method comprises the steps of: (a) providing a compound that is a PPARγ agonist;

(b) contacting mature adipocytes with the compound provided by step (a), and measuring the expression level for a PPARγ-induced adipogenic gene in the mature adipocytes;

(c) contacting mature adipocytes wi th the PPARγ ful I agonist and measuring the expression level for a PPARγ-induced adipogenic gene in the mature adipocytes; (d) comparing the expression level measured in step (b) with that measured in step (c) ; and

(e) selecting a compound whose contact results in the significantly lower expression level in step (b) in comparison with that in step (c) ; A kit for screening for a compound that lowers blood glucose level without inducing obesity and/or insulin resistance, the kit comprising

(a) adipocytes;

(b) a culture medium for the adipocytes;

(c) a PPARγ ful I agonist; and (d) an ol igonucleotide primer set for the PPARγ-induced gene;

A compound obtainable by the screening method described above, which lowers blood glucose level without inducing obesity and/or insulin resistance;

A pharmaceutical composition comprising the compound described above, or i ts pharmaceutical ly acceptable salt, and a pharmaceutical ly acceptable carrier; and

A method for lowering the blood glucose level wi thout causing obesity and/or insulin resistance by administering the compound described above, or its pharmaceutically acceptable salt, to a subject in need thereof.

The term "a PPARγ agonist" used herein means a PPARγ ligand that binds to PPARγ and activates it to facilitate the expression of PPARγ-induced genes, and includes both "a PPARγ full agonist" and "a PPARγ partial agonist". The term "a PPARγ full agonist" as used herein means a PPARγ ligand that has an activity to ful ly induce the expression of PPARγ-induced adipogenic genes in mature adipocytes. The term "a PPARγ partial agonist" as used herein means a PPARγ ligand that has an activity to partially induce the expression of PPARγ-induced adipogenic genes in mature adipocytes. There has been reported a lot of synthetic compounds, in addition to endogenous (naturally-occurring) ligands, which can act as a PPARγ full agonist. The endogenous ligand compounds acting as a PPARγ full agonist includes 9S-H0DE, 13-HODE, eicosapentanoic acid, and 15-deoxy-del ta12, 14- prostaglandin J2. 9S-H0DE and 15-deoxy-del ta12, 14-prostaglandin J2 are preferably used. The synthetic compounds acting as a PPARγ full agonist include, for example, rosigl i tazone and piogl i tazone.

Screening methods In one aspect, this invention relates to methods of screening for candidate compounds for antidiabetic agents that lower blood glucose level without inducing obesi ty and/or insul in resistance. A screening method of this invention comprises the steps of:

(a) providing a compound that is a PPARγ agonist; (b) contacting mature adipocytes with the compound provided by step (a) and measuring the expression level for a PPARγ-induced adipogenic gene in the mature adipocytes;

(e) selecting a compound whose contact results in the significantly lower expression level in step (b) in comparison with that in step (c).

Using this method, a compound that functions as a PPARγ partial agonist in mature adipocytes can be screened. Each of the steps (a) through (e) is explained in detai I below.

Step (a)

A test compound that has an activity as a PPARγ agonist is provided. Test compounds may derive from ei ther synthetic compounds or natural ly-occurring I igand compounds. The synthetic compounds include those produced by both a chemical process and a biological process.

Whether a compound has an activi ty as a PPARγ agonist or not can be determined according to the following criteria: A. binding to PPARγ;

B. protease protection;

C. recruitment of a coactivator;

D. dissociation of a corepressor; or

E. inhibition of the binding of an endogenous ligand or a synthetic PPAR r ful I agonist

The criteria A through E above can be assayed as set forth below.

A. Binding to PPARγ

Receptor-binding assay can be used to detect the binding between PPARγ and a test compound in the presence of a PPARγ ful I agonist. As one of typical examples, the assay can be performed as follows. A PPARγ I igand-binding domain expression pi asmid is prepared and used to express the I igand-binding domain in an appropriate host. Any known vector-host system such as E. col i expression system can be used for the expression of PPARγ I igand-binding domain. For example, tritium can be used to ra iolabel a PPARγ ful I agonist such as rosi l i tazone or piogl i tazone. A test compound is contacted with the I igand-binding domain in the presence of the radiolabeled PPARγ full agonist. If the test compound could bind to the PPARγ I igand-binding domain even in the presence of the competing radiolabeled PPARγ ful I agonist, the compound is regarded as a PPARγ agonist. Reporter gene assay can also be used as described in Berger, J. et al., Journal of Biological Chemistry 274 : 671 8-6725, 1 999.

B. Protease protect i on

Protease protection assay can be used and performed as follows. PPARγ is radiolabeled, for example, wi th [⁵S]methionine, and digested wi th a protease such as trypsin in the presence of both a PPARγ ful I agonist and a test compound. The digestion product is separated by, for example, SDS-PAGE, followed by autoradiography to visual ize radiolabeled digestion products, and the digestion prof i les are compared. If the conforma i onaI change of PPARγ induced by the test compound was different from that induced by a known agonist, the compound can be regarded as a different type of PPARγ agonist.

C. Recruitment of a coactivator

Recruitment of a coactivator can be assayed by either an In vitro GST (glutathione S-transferase) pull-down assay or a mammalian two-hybrid assay as described below.

In vitro GST (glutathione S-transferase) pull-down assay can be used and performed as fol lows.

A coactivator is radiolabeled with, for example, [³⁵S]methionine, and incubated with GST-human PPARγ fusion protein and glutathione sepharose beads in the presence of a test compound. The beads are recovered by centrifugation and washed to remove the unbound coactivator. The beads are separated by SDS-PAGE, followed by autoradiography to visualize the radiolabeled coactivator. The compound having an activity to promote coactivator recruitment can be regarded as a PPARγ agonist.

Coactivator recrui tment can also be examined by mammal ian two-hybrid assay as fol lows.

A VP16- human PPARγ fusion protein expression vector and a GAL4-coactivator fusion protein expression vector are generated. These vectors and a luciferase reporter piasmid containing GAL4 binding si tes are used to cotransfeet mammal ian cells. A test compound is contacted with the transfectants, and luciferase activities in the cells are then determined. Other known two-hybrid systems can also be used.

Alternatively, known methods can be used, including GAL4 chimeric receptor transcriptional assay (Berger, J. et al., Journal of Biological Chemistry, 274, 6718-1725, 1999) and the PPAR-CBP HTRF assay (Zhou, G. et al., Molecular Endocrinology, 12:1594-1604, 1998).

Known coactivators that can be used in this assay includes PGC-2 (Casti I lo,

G. et al., EMBO J. 18(13): 3676-87 (1999)), PGC-1 (Puigserver, P. et al., Cell 92:829-839 (1998), steroid receptor coactivator-1 (SRC-1) (Zhu, Y. et al., Gene Expr. 6:185-195 (1996), thyroid hormone receptor-associated protein (TRAP) 220 (DRIP205) /PBP (Yuan, C.X. et al., Proc. Nat I. Acad. Sci. USA, 95:7939-7944 (1998) and Mol. Cell Biol. 20(21) :8008-17 (2000)), p300/CREB binding protein (CBP) (Gelman, L. et al., J. Biol. Chem.274(12) :7681-8 (1999)), and receptor interacting protein (RIP)-140 (Treuter, E. et al., Mol. Endocrinol. 12:864-881 (1998)).

A screening test compound provided by the step (a) described above may be preferably derived from compounds that result in a significantly lower binding activity between PPARr and a non-selective coactivator as compared with a PPAR

T full agonist, and the compounds can be selected by the method comprising the steps of:

(a-1) contacting PPARr and a non-selective coactivator with a compound, and measuring the binding activity of PPARr to the non-selective coactivator;

(a-2) contacting PPARr and a non-selective coactivator with a PPARr ful I agonist, and measuring the binding activity of PPARr to the non-selective coactivator;

(a-3) comparing the binding activity measured in (a-1) with that measured in (a-2) ; and

(a-4) selecting a compound whose contact results in a significantly lower binding activity in (a-1) in comparison with that in (a-2).

Here, non-selective coactivators mean coactivators that interact with a wide range of nuclear receptors in add i t ion to PPARr . Non-selective coactivators that can be used in this assay preferably include CREB binding protein (CBP) and steroid receptor coactivator-1 (SRC-1). The term "significantly lower" means that the binding activity of PPAR r to a test compound is at most 70%, more preferably at most 50%, of that to a PPARγ full agonist.

Specifically, this method can be performed using either in vitro GST pull-down assay or mammalian two-hybrid assay. Furthermore, a screening test compound provided by the step (a) described above may be preferably derived from compounds that resul t in a substantial ly same binding activity between PPARr and a PPAR r -selective coactivator, and the compounds can be selected by the method comprising the steps of:

(a-5) contacting PPARr and a PPARr -selective coactivator wi th a compound and measuring the binding activity of PPARr to the selective coactivator;

(a-6) contacting PPARr and a PPARr-selective coactivator with a PPAR r full agonist and measuring the binding activity of PPARr to the selective coactivator;

(a-7) comparing the binding activity measured in (a-5) with that measured in (a-6) ; and (a-8) selecting a compound whose contact results in a substantially same binding activity in (a-5) in comparison with that in (a-6).

PPAR r -selective coactivators used herein mean coactivators that selectively interact wi th PPARr- PPARr -selective coactivators that can be used in this assay preferably include peroxisome proliferator-activated receptor r coactivator-1 (PGC-1) and TRAP220.

Here, the term "substantial ly same" means that the binding activity of PPAR r to the selective coactivator to a test compound ranges from 50% to 150%, more preferably from 70% to 130%, even more preferably from 90% to 110%, of that to a PPARγ ful I agonist.

Specifically, this method can be performed using either in vitro GST pull-down assay or mammalian two-hybrid assay.

A compound selected by these assays has activities to induce recruitment of PPARγ-selective coactivators to PPARγ effectively but induce recruitment of non-selective coactivators to PPARγ weakly, as compared wi th a PPARγ ful I agonist. Such a compound can be a selective PPARγ modulator, which behaves as a ful I agonist in differentiating adipocytes and behaves as a partial agonist in mature adipocytes.

D. Dissociation of a corepressor

The coactivator recruitment can be measured using the method such as mammalian two-hybrid assay, and can be performed as described above except for using a corepressor in place of a coactivator.

Corepressors to be used include nuclear corepressor (NCoR) and silencing mediator of retinoid and thyroid receptors (SMRT).

E. Inhibition of the binding of an endogenous ligand or a synthetic PPARr full agonist

Inhibition of the binding of an endogenous ligand to PPAR γ by an test compound can correlate with the reduced activity of binding between PPARr and a non-selective coactivator, and that can be assayed by the method comprising the steps of:

(a-9) contacting PPARr and a non-selective coactivator with either an endogenous ligand for PPARr or a synthetic PPARr full agonist in the presence of a test compound, and measuring the binding activity of PPAR r to the non-selective coactivator;

(a-10) contacting PPARr and a non-selective coactivator with either an endogenous ligand for PPARr or a synthetic PPARr full agonist in the absence of a test compound, and measuring the binding activity of PPAR r to the non-selective coactivator; (a-11) comparing the binding activity measured in (a-9) with that measured in (a-10) ; and

(a-12) selecting a compound whose presence results in the reduced binding activity in (a-9) in comparison with that in (a-10). The endogenous ligands that can be used in this assay include (S)-hydroxyoctadecad i eno i c acid (9S-H0DE), 13 (S)-ydroxyoctadecad i enoi c acid (13S-H0DE), eicosapentaenoic acid, and 15-deoxy-del ta 12, 14-prostagl and i n J2.

Specifically, this method can be performed using either in vitro GST pull-down assay or mammalian two-hybrid assay. Mammalian two-hybrid assay can be carried out in a simi lar manner as described above for coactivator recrui tment, except that an endogenous ligand for PPARr or a synthetic PPARr full agonist is added as well as a test compound.

Step (b) A compound provided in step (a) is contacted with mature adipocytes. The contact can be performed by adding the compound or the agonist to a culture medium containing adipocytes.

For this purpose, any human adipocytes or mouse 3T3-L1 cells can be used as adipocytes. For instance, human adipocyte precursor eel Is (preadipocytes) can be isolated from subcutaneous adipose tissue. Liposuction should be undergone from healthy donors of the age between 20 and 60 years. The preadipocytes are treated with collagenase, and isolated by centrifugal force.

Mature human adipocytes can be prepared as follows. Preadipocytes are inoculated in a preadipocyte medium (Dul Ibeccos' modified Ea le's essential medium (DMEM)/Ham™s F-10 medium (1:1, v/v) containing 15 mM HEPES (pH 7.4), 10% fetal bovine serum (FBS), 100 U/mL penicillin, 100 μg/mL streptomycin and 0.25 μg/mL aphotericin B), and cultured at 37°C as growing precursor cells.

Mouse 3T3-L1 preadipocytes are available from JCRB Cell Bank, and 3T3-L1 mature adipocytes can be prepared and contacted with a test compound provided by Step (a) as fol lows.

Typical ly, 3T3-L1 eel Is (about 3x10⁴ eel Is/cm²) are incubated wi th 3 mL of a culture medium (DMEM containing 10% FBS, 100 U/mL penicillin and 0.1 mg/mL streptomycin) in a 6-wel I plate at 37°C with a 5% C0₂ atmosphere on Day 1. After 2-day incubation, the medium is replaced with 2.5 mL of a differentiation-inducing medium (DMEM containing 10% FBS, 100 U/mL penici 11 in, 0.1 mg/mL streptomycin, 0.25 μM DEX, 0.5 mM IBMX and 10 μg/mL insul in). After three-day incubation, the medium is replaced with 2.5 mL of a maintai nig medium (DMEM containing 10% FBS, 100 U/mL penicillin, 0.1 mg/mL streptomycin, and 10 μg/mL insulin). After further 2-day incubation, the medium is replaced with 2.5 mL of the culture medium. The medium is replaced with 2.5 mL of the flesh culture medium on Day 10 and Day 13. On Day 14, 2.5 μL of the 1000 x concentrations of a test compound solution is added to the medium in the culture wells.

In general, the compound provided in step (a) can be added into the culture medium for the adipocytes, which have been cultured for about 12 to 21 days as described above. The culture medium containing the adipocytes and the compound is incubated further at 37°C for 48 hrs.

Preferably, the final concentration of a test compound in the medium should range from about 1x10^~9 M to about 1x10^~5 M, and the final concentration of dimethyl sulfoxide (DMSO), solvent of the compound, in the medium should be adjusted to 0.1%. As for the non-treated control eel Is, final 0.1% of DMSO without any test compound should be added to the medium.

The expression level for a PPARγ-induced adipogenic gene is measured. Examples of the PPARγ-induced adipogenic gene include adipocyte fatty acid binding protein (aP2) gene, fatty acid binding protein (FABP) gene, and adiponectin gene. The gene expression level can be measured by quanti tative real-time RT-PCR analysis among others. Specif icaI ly, a supernatant of the eel I cul ture is removed by centrifugation and test compound-treated eel Is are lysed by adding TRIZOL (GIBCO BRL). Cell lysate can be stored at -80°C until total RNA will be isolated. A reverse transcription (RT) product is synthesized from total RNAs isolated from the cell lysate using a reverse transeriptase system such as Taq Man Reverse Transcription Reagent (Applied Biosystems). The RT product is amplified and detected using a real-time PCR system (e.g. ABI PRISM 7700).

Step (c) Mature adipocytes are contacted with a PPARγ full agonist, and the expression level of a PPARγ-induced adipogenic gene in the cells is measured in a simi lar manner as described in Step (b), except that a PPARγ ful I agonist is used instead of a compound provided in step (a).

Step (d)

The expression level measured in step (b) is compared with that measured in step (c). Each of the expression levels is calculated as a percentage to the control (the expression level for the control eel Is to which no compound had been added).

Step (e)

A compound whose contact results in a significantly lower expression level in step (b) in comparison with that in step (c) is selected. Here, the term "significantly lower" means that the expression level for a test compound is at most 70%, more preferably at most 50%, of that for a PPARγ full agonist. The compound selected by the above method can be further screened by a method comprising the steps of:

(f) contacting differentiating adipocytes with the compound selected in step (e) and evaluating the differentiation stage of the adipocytes;

(g) contacting differentiating adipocytes wi th the PPARγ ful I agonist and evaluating the differentiation stage of the adipocytes;

(h) comparing the differentiation stage evaluated in step (f) with that evaluated in step (g) ; and (i) selecting a compound whose contact resul ts in a substantial ly same stage of adipocyte differentiation in step (f) in comparison with that in step (g).

Step (f)

Step (f) can be carried out in the same manner as Step (b) except for using differentiating adipocytes. Differentiating adipocytes includes human differentiating adipocytes or mouse 3T3-L1 differentiating adipocytes.

Human preadipocytes are isolated as described in Step (b), and human differentiating adipocytes can be prepared from the preadipocytes cultured with medium supplemented with adipogenic and lipogenic hormones according to the simi lar procedure described by Halvorsen, Y. etal. (Metabol ism 50:407-413, 2001) or Harp, J. et al. (Biochemical and Biophysical Research Communications 281 : 907-912, 2001). Typical I y, the enti re volume of the culture medium for preadipocytes is exchanged with a differentiation medium (DMEM/Ham™s F-10 medium (1:1, v/v) containing 15 mM HEPES (pH 7.4), 3% (v/v) FBS, 33 μM biotin, 17 μM pantothenate, 100 nM human insulin, 1 μM dexamethasone (DEX), 100 U/mL penicillin, 100 μg/mL streptomycin, 0.25 μg/mL a photericin B, 0.20 mM 3-i sobuty 1-1-methy I xanth i ne (IBMX), andlOμMPPAPγ agonist). After 3-day incubation at 37°C and 5% C0₂, cells are fed with an adipocyte medium (the differentiation medium without the PPAPγ agonist). When feeding, only half of the medium in a we 11 should be replaced and replenished wi th an equal volume fresh medium. After about four feedings (12 days), rounded eel Is with I ipid droplets wi 11 appear and be sui table for adipocyte assays such as insulin responses and lipolysis. Three weeks after the initiation of differentiation, cells should appear rounded with large lipid droplets apparent in the eel Is. 3T3-L1 differentiating adipocytes can be prepared by incubating 3T3-L1 preadipocytes in a culture medium (DMEM containing 10% FBS, 100 uni ts/mL penici 11 in, and 0.1 mg/mL streptomycin) at 37°C for two days, replacing the medium with a di fferentiat ion-i nducing medium (DMEM containing 10% FBS, 100 uni ts/mL penici 11 in, 0.1 mg/mL streptomycin, 0.25 μM DEX, 0.5 mM IBMX, and 10 μg/mL insul in), incubating the cells for three days, replacing the medium with the culture medium, and incubating the cells for two days in a similar manner as described in step (b) for preparing mature adipocytes.

The adipocytes cultured for about 2 to 5 days can be used to contact with a test compound. The differentiation stage of the adipocytes can be evaluated in terms of the expression level of a PPARγ-induced adipogenic gene or accumulation of triglyceride in the adipocytes. The former can be evaluated by measuring the expression level as described in Step (c). The latter can be evaluated by measuring the content of triglyceride, for example, as follows. The culture medium for the adipocytes is removed and intracel lular triglyceride is extracted by the add i t ion of 6 L of isopropanol. Typical ly, the amount of the extracted triglyceride is quantitated by the acetylacetone method using a ki t commercial ly avai lable from Wako Pure Chemicals Industries, Ltd. (Osaka, Japan). The concentration of intracel lular protein is determined using, for example, DC Protein Assay (BIO RAD), after solubi I izing it with 0.1 mol/L of NaOH (0.8 mL/well). The intracel lular triglyceride content is calculated as per mg protein.

Step (g) Differentiating adipocytes are contacted wi th a PPARγ ful I agonist, and the differentiation stage of the adipocytes is evaluated in a similar manner as described in Step (f), except that a PPARγ ful I agonist is used instead of a compound provided in step (e).

Step (h)

The differentiation stage evaluated in step (f) is compared with that evaluated in step (g). When the differentiation stages were evaluated in terms of triglyceride accumulation, the intracel lular triglyceride contents are compared between the test compound-treated group and the PPARγ ful I agonist-treated group. When the differentiation stages were evaluated in terms of the expression level of PPARγ-induced adipogenic gene such as aP2 and FABP, the expression levels in cells are compared between the test compound-treated group and the PPARγ full agonist-treated group.

Step (i)

A compound whose contact results in a substantially same adipocyte differentiation in step (f) in comparison with that in step (g) is selected. Here, the term "substantial ly same" means that the intracel lular triglyceride contents for a test compound ranges from 70% to 130%, more preferably from 90% to 110%, of that for a PPARγ full agonist, when the differentiation stage of the cells is evaluated in terms of triglyceride contents. Otherwise, the term "substantial I y same" means that the expression level of PPARγ-induced adipogenic gene in the presence of a test compound ranges from 70% to 130%, more preferably from 90% to 110%, of that in the presence of a PPARγ full agonist, when the differentiation stages were evaluated in terms of the expression level of PPARγ-induced adipogenic gene such as aP2 and FABP.

After Step (e) or (i) of the above-described methods, the following steps may further be carried out:

(j) contacting mature adipocytes with the compound selected in step (e) or step (i) and measuring the expression level of a gene, of which down-regulation is involved in the development of insulin resistance;

(k) contacting mature adipocytes wi th the PPARγ ful I agonist and measuring the expression level of a gene, of which down-regulation is involved in the development of insulin resistance; (I) comparing the expression level measured in step (j) with that measured in (k) ; and

(m) selecting a compound that results in a significantly higher expression level in (j) in comparison with that in (k).

Step (j)

The mature adipocytes can be cultured and contacted with the test compound selected in step (e) or step (i) in a similar manner as Step (b). To examine the long term effects on the adipocytes of the test compound, the eel Is can be cul tured and contacted with the compound by the following method among others. Seven volumes of A solution (acid-soluble type I collagen solution (pH 3, Nitta Gelatin)), 2 volumes of B solution (5-fold concentrated DMEM (without NaHC0₃)) and 1 volume of C solution (reconstruction buffer (2.2% NaHC0₃, 200 mM HEPES, 0.05 N NaOH)) are thoroughly mixed and kept on ice, and then mixed with the dissociated 3T3-L1 cells (2x10⁵ cells/mL). Typically, each 800 μL of the mixture containing 3T3-L1 cells are dispensed into each well of a type I col lagen-coated 12- we 11 plate, incubated at 37°C for 30 mi n, and, after gelation, overlaid with 3 mL of the culture medium. The cells are incubated at 37°C with a 5% C0₂ atmosphere. After 2-day incubation, the medium in each wel I is replaced with 3 mL of the differentiation-inducing medium. After a further 3-day incubation, the medium is replaced with 3 mL of the maintainig medium. On Day 8, the medium in each well is replaced with 3 mL of the culture medium and 3.8 μL of the 1000 x concentrations of the test compound solution is added to each wel Is. Each test compound is added to the final concentration of IX. Preferably, the final concentration of the test compound in the medium should range from about 1x10^"9M to about 1x10^~5 M, and the final concentration of DMSO, solvent of the compound, in the medium should be adjusted to 0.1%. As for the non-treated control eel Is, final 0.1% of DMSO without any test compound should be added to the medium. On Day 10, 13, 15, 17, 20 and 22 the medium in each well is replaced to 3 mL of the flesh cul ture medium and 3 μL of the 1000 x concentrations of the test compound solution is added to the replaced medium. On Day 24, the medium is removed and the adipocyte-embedded gel is scraped off and minced in the test tube. The culture medium is added up to 1.5 L in the test tube. The gel is digested by addition of 30 μL of Collagenase S-1 (10 mg/mL, Nitta Geratin) and incubated at 37°C for 30 in. The expression level of a gene, of which down-regulation is involved in the development of insul in resistance, can be measured in a simi lar manner as Step (b). The gene, of which down-regulation is involved in the development of insul in resistance, includes leptin and phosphodiesterase 3B (PDE3B) gene. Down-regulation of such genes in mature adipocytes is thought to be involved in the development of obesity-induced insulin resistance in fat cells.

Step (k)

Mature adipocytes are contacted with a PPARγ full agonist, and the expression level of a gene, of which down-regulation is involved in the development of insulin resistance, is measured in a similar manner as described in Step (j), except that a PPARγ full agonist is used instead of the compound selected in step

(e) or step (i).

Step (I) The expression level measured in step (j) is compared with that measured in step (k). Each of the expression levels is calculated as a percentage to the control (the expression level for the control eel Is to which no compound had been added).

Step (m)

A compound whose contact resu I ts in a significantly hi her expression level in step (j) in comparison with that in step (k) is selected. Here, the term "significantly higher" means that the expression level for a test compound is at least 150%, more preferably at least 200%, of that for a PPARγ full agonist.

Kits

In another aspect, the present invention provides kits for screening for compounds that lower blood glucose level wi thout inducing obesity and/or insul in resistance, which can be used for performing the above-described methods. A kit of the present invention comprises, (a) adipocytes ;

(b) a culture medium for the adipocytes;

(c) a PPARγ full agonist; and

(d) an ol igonucleotide primer set for the PPARγ-induced gene. Adipocytes (a) are at least one member selected from the group consisting of (1) mature adipocytes; (2) differentiating adipocytes; and (3) preadipocytes. The cells may be either human cells or mouse 3T3-L1 cells. The mature adipocytes can be prepared from differentiating eel Is as described above. An example of the culture medium is DMEM supplemented with 10% FBS, 100 units/mL penicillin, and 0.1 mg/mL streptomycin. The PPARγ full agonist is selected from the group consisting of 9S-H0DE, 13S-H0DE, eicosapentaenoic acid, 15-deoxy-del ta 12, 14-prostagl andi n J2, rosigl i tazone, and piogl i tazone. The PPARγ-induced gene includes adipocyte fatty acid binding protein (aP2) gene and fatty acid binding protein (FABP) gene.

Compounds obtainable by the screening methods

In a further aspect, the present invention provides compounds obtainable by the above-described screening methods. Such compounds are candidates for antidiabetic agents that lower blood glucose level wi thout inducing obesity and/or insul in resistance.

Examples of the compounds are benzimidazole derivatives described in W097/24334, W099/00373, and W000/39099. A preferable example of the benzimidazole derivatives is FK614.

Pharmaceutical compositions

In a still further aspect, the present invention provides pharmaceutical compositions comprising a compound obtainable by the above-described screening methods, or its pharmaceutically acceptable salt, and a pharmaceutically acceptable carrier. When a screened compound is used as an antidiabetic agent, it may be formulated into a pharmaceutical composition of a solid form such as tablets, granules, powders, and capsules or of a I iquid form such as solutions, suspensions, syrups, emulsions, and lemonades, together with a pharmaceutically acceptable carrier, such as an organic or inorganic sol id or I iquid vehicle sui table for oral administration, parenteral administration, or external application.

If desired, the pharmaceutical compositions may further contain a pharmaceutical aid, a stabilizer, a wetting agent, and also any ordinary addi tive, for example, lactose, citric acid, tartaric acid, stearicacid, magnesium stearate, terra alba, sucrose, corn starch, talc, gelatin, agar, pectin, peanut oil, olive oil, cacao butter, ethylene glycol, etc. A pharmaceutical ly acceptable sal t of a compound of the present invention is a non-toxic, ordinary pharmaceutically acceptable salt, such as a salt with a base as well as an acid addition salt. Examples thereof include a salt with an inorganic base, such as a salt with an alkali metal (e.g., sodium, potassium); a sal t wi th an alkal ine earth metal (e.g. , calcium, magnesium) ; an ammonium sal t; a salt wi th organic amine (e.g. , tr i ethyl ami ne, pyridine, pi col ine, ethanolamine, triethanola ine, d i eye lohexy I ami ne, N, N'-dibenzylethylenediamine) ; asaltwithan inorganic acid (e.g., hydrochloric acid, hydrobromic acid, sulfuric acid, phosphoric acid) ; a sal t wi th an organic carboxy I ic acid (e.g. , formic acid, acetic acid, trif luoroacet ic acid, maleic acid, tartaric acid); a salt with a sulfonic acid (e.g., ethanesulfonic acid, benzenesulfonic acid, p-to I uenesu I fon i c acid); a salt with a basic or acidic amino acid (e.g., arginine, aspartic acid, glutamic acid).

The compounds or their pharmaceutically acceptable salts are contained in the pharmaceutical compositions in an amount effective to lower the blood glucose level without causing obesity and/or insulin resistance. The amount of the compounds or their pharmaceutically acceptable salts may vary depending on the age and the condition of patients, the type and the state of the disease, and the type of the compounds used.

Methods of lowering the blood glucose level

In a still further aspect, the present invention provides methods for lowering the blood lucose level wi thout causing obesi ty and/or insul in resistance by administering a compound or i ts pharmaceutical l acceptable salt of the present invention to a subject in need thereof.

A subject in need of the treatment of the present invention is, for example, a patient suffering from diabetes, particularly, NIDDM.

In general, for oral administration, the dose of a compound may be from 1 to 100 mg/kg; and for intramuscular injection or intravenous injection, it may be from 0.1 to 10 mg/kg. Such a unit dose may be applied to a subject once to four times a day.

Any patents, published patent applications, and publications cited herein are incorporated by reference.

Brief Description of the Drawings

Fig. 1 is a photograph showing SDS-PAGE patterns of trypsin-digested products of a rosigl i tazone-PPARγ complex and an FK614-PPARγ complex.

Fig. 2 shows effects of FK614, rosigl i tazone, and piogl itazone on recruitment of CBP to PPARγ. The error bar shows mean ± S.E. (n=5). Fig. 3 shows effects of FK614, rosigl itazone, and piogl itazone on recruitment of SRC-1 to PPARγ. The error bar shows mean ± S.E. (n=5).

Fig. 4 shows effects of FK614, rosigl itazone, and piogl itazone on recruitment of PGC-1 to PPARγ. The error bar shows mean ± S.E. (n=5). Fig. 5 shows effects of FK614, rosigl i tazone, and piogl itazone on release of NCoR from hPPARγ2. The error bar shows mean ± S.E. (n=3). $$ (FK614), #* (rosigl itazone), and U (piogl itazone) means significant difference (P<0.01).

Fig. 6 shows effects of FK614, rosigl itazone, and piogl itazone on release of SMRT from hPPARγ2. The error bar shows mean ± S.E. (n=3). $ (FK614), * (rosigl i tazone), and & (piogl i tazone) mean significant difference (P<0.05) .

$$ (FK614), *# (rosigl i tazone), and && (piogl itazone) mean significant difference

(P<0.01) .

Fig. 7 shows effects of FK614, rosigl itazone, and piogl itazone on recruitment of CBP to hPPARγ2. The error bar shows mean ± S.E. (n=3). $$ (FK614), *# (rosigl i tazone), and && (piogl i tazone) mean significant difference (P<0.01).

Fig. 8 shows effects of FK614, rosigl itazone, and piogl itazone on recruitment of SRC-1 to hPPARγ2. The error bar shows mean ±S.E. (n=3). $$ (FK614),

** (rosi l i tazone), and && (piogl i tazone) mean significant difference (P<0.01) .

Fig. 9 shows effects of FK614, rosigl itazone, and piogl itazone on recruitment of TRAP220 to hPPARr2.

Fig. 10 shows effects of FK614, rosigl i tazone, and pioglitazone on recruitment of CBP to hPPARr 2. The effects were measured by mammal ian two-hybri assay using each compound at 1 X 10^"5M (Mean, n=3).

Fig. 11 shows effects of FK614, rosigl i tazone, and pioglitazone on recruitment of SRC-1 to hPPARr . The effects were at 1 X 10^"5M measured by mammalian two-hybrid assay using each compound (Mean, n=3).

Fig. 12 shows effects of FK614, rosigl i tazone, and pioglitazone on recruitment of TRAP220 to hPPARr2. The effects were measured by mammalian two-hybrid assay using each compound at 1 X 10^"5 (Mean, n=2). Fig. 13 shows effect of FK614 on recruitment of CBP to PPARγ induced by 9(S)-H0DE. The error bar shows mean ± S.E. (n=3).

Fi . 14 shows effects of FK614, rosigl i tazone and piogl itazone on aP2 gene expression in mature 3T3-L1 adipocytes. The error bar shows mean ± S.E. (n=3).

Fig. 15 shows effects of FK614, rosigl i tazone, and piogl itazone on aP2 gene expression in differentiating mouse 3T3-L1 adipocytes. The error bar shows mean

± S.E. (n=3).

Fig. 16 shows effects of FK614, rosigl itazone, and pioglitazone on the triglyceride accumulation in differentiating mouse 3T3-L1 cells. The error bar shows mean ± S.E. (n=3). * and ** mean P<0.05 and P<0.01, respectively. Fig. 17 shows long-term effect of FK614, rosigl i tazone and pioglitazone on PDE3B gene expression in mature 3T3-L1 adipocytes. The error bar shows mean ± S.E. (n=3). C means non-treated control group.

Best Mode for Carrying out the Invention

The present invention will be described in detail using Examples below, but it is not to be construed as being limited thereto.

Example 1: Identification of PPARγ agonists 1-1. Protease Protection

In order to determine whether FK614 can directly induce conformational changes in PPARγ or not, PPARγ was digested with a protease in the presence or absence of FK614 or rosigl itazone.

Test compounds were dissolved in dimethyl sulfoxide (DMSO) at the concentration of 1.0 x 10^"M. The dissolved compounds were di luted 100-folds wi th Dulbecco's phosphate buffered saline at 10 x concentration of the final, and then added into the transcription/translation reaction mixtures. (The final concentrations of the test compounds and DMSO were 1.0 x 10^"5M and 0.1%, respectively.) According to the protocol by the manufacturer (Promega Corporation, WI, USA), the pCDMδ-hPPARγl plasmids were used to synthesize [³⁵S]met ionine-labeled full-length human PPARγl polypeptides in a coupled transcription/translation system. Each of the equally aliquoted transcription/translation reaction mixtures was mixed with diluted solution for test compounds. The mixture was incubated for 30 min at room temperature, and was then added distilled water or increasing concentrations' solutions of trypsin. The protease digestions were allowed to proceed for 20 min at 25°C, and then terminated by the addition of SDS sample loa ing buffer and boi led for 1 min at 100°C. The resultants were resolved by SDS-PAGE, fol lowed by autoradiography using BAS2000 (Fuji Photo Fi Im Co. , Ltd., Tokyo, Japan) to visualize radiolabeled digestion products. Experiments were repeated three times.

Autoradiograms were analyzed by using the image analyzing software ATTO Densito graph 4.0 (ATTO Co., Ltd., Tokyo, Japan), and the molecular mass of the protected fragments was measured. The values for the molecular mass were obtained from three independent experiments, and the average among them was calculated.

Fig. 1 shows SDS-PAGE patterns of the trypsin-digestion products of a rosigl i tazone- PPARγ complex and an FK614- PPARγ complex. There appeared a major band of 22-kDa and a minor band of 32-kDa, together with other bands for partial digests, that would derive from protease-resistant fragments for the rosigl i tazone-PPARγ complex. Though trypsin-digestion of the FK614-PPARγ complex also yielded simi lar protected fragments, distinct digestion pattern was observed. The 32-kDa fragment was detected more clearly in the FK614-treated samples than in the rosi l itazone-treated samples, showing that complex formation with FK614 generated stronger protection for the fragments in comparison wi th that with rosigl i tazone. In addition, a 25-kDa fragment, that was not present in the rosigl itazone-treated samples, was faintly detected. Thus the results indicate that each of these two compounds is a PPARγ agonist that can interact with PPARγ directly and induce distinct PPARγ conformation changes.

1-2. Recruitment of coactivators measured by in vitro GST pull-down assay

Complex formation between PPARγ and each of the coactivators, CBP, SRC-1, and PGC-1, was induced by adding each of FK614 and two thiazol idinedione compounds, rosi l itazone and piogl i tazone, and the interaction between them was exa ined by in vitro GST pull-down assay. Each compound of FK614, pioglitazone, and rosigl itazone was dissolved in DMSO and was diluted with the binding buffer containing 8mM Tris- HCI (pH 7.6), 120mM KCI, 1mM di thiothreitol (DTT), 8% glycerol, 0.5% 3-[(3-Col ami dopropy l)-d imethy Iammonio]-1-propane-su I fonate (CHAPS), 1mg/mL bovine serum albumin (BSA), 10μg/mL aprotinin, 10μg/mL leupeptin, and lOμg/mL pepstatin. Each of the final concentrations of FK614, pioglitazone, and rosigl i tazone was 1.0 x 10^"5M, and the final concentration of DMSO in the solutions was 1%.

According to the protocol by the manufacturer (Pro ega Corporation, WI, USA), [³⁵S]methionine-labeled coactivators were synthesized in a coupled transcription/translation system using the expression pi asmids encoding the human full-length CBP, SRC-1, and PGC-1. The transcription/translation reaction mixtures were mixed with GST-hPPARγ2 fusion protein and glutathione sepharose in the presence of test compounds, and then incubated overnight at 4°C with a gentle rotation. Then, the sepharose beads were pel leted by centri fugation, washed four times to remove unbound coactivators, and were suspended in SDS sample loading buffer. The resultant materials were boiled to denature the bound proteins, and separated by SDS-PAGE, fol lowed by autoradiography using BAS2000 (Fuji Photo Film Co., Ltd., Tokyo, Japan) to visualize radiolabeled coactivators. To ensure the generation of the coactivators in a coupled transcription/translation system, a small quantity of the transcription/translation reactions was also resolved by SDS-PAGE, and analyzed by BAS2000. Experiments were repeated five times.

Recruited [³⁵S]methionine-labeled coactivators were quantified by measuring the radioactivity for the corresponding bands using BAS2000. The extent of I igand-dependent coactivator recruitment was calculated in accordance wi th the fol lowing formula. [fold increase]={[the radioactivity of recruited coactivators in the presence of test compounds]-[back ground radioactivi ty]}/{[the radioactivity of recruited coactivators in the absence of test compounds]-[back ground radioactivi ty]} Figs. 2-4 show effects of each of FK614, pioglitazone, and rosigl itazone on recruitment of the coactivators. All the three compounds were found to promote recruitment of the coactivators to PPARγ. However, the magnitudes of recruitment of individual coactivators were varied wi th regard to each I igand. The magni tude of CBP recruitment induced by either of FK614, pioglitazone, and rosigl itazone was measured to be, respectively, 1.91-fold, 4.20-fold, and 5.09-fold above the basal level. Likewise, the magnitudes of SRC-1 recruitment induced by either of FK614, pioglitazone, and rosigl i tazone were, respectively, 3.31-fold, 19.80-fold, and 26.63-fold above the basal level.

Ei ther amount of CBP recrui tment and SRC-1 recruitment by FK614 was smal ler than that by either of the other two compounds. In contrast, the magnitude of PGC-1 recruitment by ei ther of FK614, piogl itazone, and rosigl itazone was measured to be, respectively, 2.11-fold, 2.17-fold, and 2.37-fold above the basal level, showing that each of these three compounds similarly induced the PGC-1 binding to PPARγ.

1-3. Dissociation of corepressor and recruitment of coactivators measured by mammalian two-hybrid assay

Mammalian two-hybrid assay was used to examine the abilities of FK614, rosigl i tazone and pioglitazone to release the corepressors, NCoR and SMRT, from hPPARγ2 or to recruit CBP and SRC-1 to hPPARγ2.

CV-1 eel Is were obtained from RIKEN Cel I Bank. Dulbeccos' modified Eagle's essential medium (DMEM), fetal bovine serum (FBS), and penici 11 in (10000 unit/mL) - streptomycin (10000 μg/mL) liquid were purchased from Nikken Bio Medical Laboratory, MOREGAT and GIBCO BRL, respectively. FBS charcoal treated (C-FBS) was obtained from HyClone. DMEM supplemented with 10% FBS, 100 uni ts/mL penicillin, and 0.1 mg/mL streptomycin was used as the culture medium. DMEM supplemented with 10% C-FBS was used as the assay medium.

A VP16 and human PPARγ2 (hPPARγ2) fusion protein expression vector pACT-hPPARγ2 was generated by inserting the cDNA fragment encoding the ful l-length hPPARγ2 into the expression vector pACT (Promega) which contains a VP16 activation domain-coding region upstream of a multiple cloning region. The expression vector pBIND (Promega) contains GAL4 DNA-binding domain upstream of a multiple cloning region and expresses the renilla reniformis luciferase whose activity can be monitored for normalization of the transfection efficiency. The GAL4-cofactor fusion protein expression vectors of pBIND-SRC-1, pBIND-CBP, pBIND-TRAP220, pBIND-NCoR and pBIND-SMRT were constructed by inserting into the pBIND vectors the cDNAs encoding, respectively, the region of amino acids 592-782 for human SRC-1 (Onate, S.A. et al., Science 270:1354-1357, 1995), the region of amino acids 1-115 for human CBP (Genbank Accession No. U85962), the region of amino acids 500-738 for human TRAP220 (Genbank Accession No. AF055994), the region of amino acids 2225-2286 for human NCoR (Genbank Accession No. AF044209) and the region of amino acids 1281-1344 for human SMRT (Chen, J.D. et al., Nature 377:454-457, 1995).

Plasmid pG5Luc (Promega), which contains five GAL4 binding sites upstream of a minimal TATA box, which in turn is upstream of the firefly luciferase gene, was used as reporter plasmid.

Each test compound of FK614, rosigl itazone, and piogl i tazone was dissolved in DMSO at the concentration of 1.0 x 10^"z M, and was then serial ly di luted 10-folds with DMSO to 1000 x concentration of the final. 1000 x concentrates were further diluted with the assay medium to 2 x concentration of the final. CV-1 cells (1.2 x 10^δ cells) were pre-cu I tured with the culture medium in 100-mm tissue culture dishes at 37°C with a 5% C0₂ atmosphere for one day. The medium was replaced with 4.8 mL of the Opti-MEM (GIBCO BRL). Transfection mixes contained 2.4 μg of pACT-hPPARγ2, 1.2 μg of GAL4-cofactor fusion protein expression plasmid and 4.8 μg of pG5Luc. For mock group, pBIND was added to transfection mix instead of the GAL4-cofactor fusion protein expression plasmid. Transfection was performed with LipofectAMINE PLUS (GIBCO BRL) for 3 hrs at 37°C. The transfected cells were harvested from the 100-mm tissue culture dishes and suspended in the assay medium. The suspended cells (4.0 x 10⁵ cells/mL) were distributed 50 μL each into each well in a 96-well white plate with clear bottom (Corning Costar).

After 60 min, 50 μL of 2 x concentrates of the test compound solution was added to the medium in the culture wells. Thus, each test compound was added at 1 x concentration of the final with 0.1% of DMSO. Final 0.1% of DMSO was contained in the medium for the non-treated control cells without test compound. Then, the test compound-treated eel Is were incubated for 24 hrs. Luciferase act ivi ties were determined using the Dual-Luciferase Reporter Assay System (Promega) according to the manufacture's protocol. Briefly, the medium was removed and cells were lysed by addi tion of 20 μL of 1 x lysis buffer (suppl ied by Dual-Luciferase Reporter Assay System). Then, the dual luciferase activi ties were measured by Mul ti label counter (WALLAC). The blank value was measured from the wel Is without eel Is, to eliminate all of the background of luminescence in the instrument, wells and substrates for f i ref ly and reni I la luciferase. Each experiment was performed in duplicates for each test compound.

The firefly and renilla luciferase activities were calculated by subtracting the mean of blank values from actually measured values [A]. For normalization of the transfection efficiency, the firefly luciferase activities were divided by the renilla luciferase activities and defined as the normalized firefly activities [B], Transfection-dependent interaction of cofactors with hPPARγ2 was determined as the normalized specific firefly activities, by subtracting the mean of the normal ized firefly activities for mock group from the mean of the normalized firefly activities for the cells transfected with GAL4-cofacor fusion protein expression plasmid [C]. Additionally, the ligand- promoted interaction between cofactors and hPPARγ2 was calculated as percentage to that of the non-treated control group [D]. The formulas for each calculation are shown as fol lows.

A: [the firefly or renilla luciferase activity] = [actually measured value of firefly or renilla luciferase activity] - [mean of blank values]

B: [the normalized firefly luciferase activity] = [the firefly luciferase activity] / [the renilla luciferase activity]

C: [the normalized specific firefly luciferase activity] = [the mean of the normalized firefly luciferase activity for GAL4-cofactor] - [the mean of the normalized firefly luciferase activity for mock group]

D: [% of non-treated control] = [the normalized specific firefly luciferase activity for the concentration of each compound] / [the normalized specific firefly luciferase activity for the non-treated control] x 100.

All experiments were carried out three times in duplicates. Data are expressed as mean ± S.E. Statistical significance between non-treated control group and each of the drug-treated groups was assessed by analysis of variance, based on randomized block design, followed by Dunnett's multiple comparisons.

Statistical significance was assumed at P < 0.05.

Table 1 shows association of corepressors with hPPARγ2 in the absence of I igand.

Table 1

GAL4-NCoR GAL4-SMRT GAL4 (mock)

Normal ized firefly 3.59x10^"2 ± 1.99

9.49 ± 1.12 10.2 ± 0.958 luciferase xlO^"3 activi ty

Data is expressed as mean ± S.E. (n=3). Normalized firefly luciferase activity indicates the interaction of GAL4-corepressor fusion protein or GAL4 with hPPARγ2.

The results in Table 1 show that the corepressor NCoR interacted consti tutively with hPPARγ2 in the absence of the ligands. Fig. 5 shows effect of each I igand on the release of NCoR from hPPARγ2. Each of FK614, rosigl i tazone, and piogl i tazone dismissed NCoR from hPPARγ2 in a concentration-dependent manner. The significant dissociation of NCoR-hPPARγ2 complex was initiated at the concentrations of 1 x 10^"6, 1 x 10^"7, and 1 x 10^"6 M for FK614, rosigl i tazone, and pioglitazone, respectively. The hPPARγ2 bindings to NCoR were 7.5%, 5.7%, and 17.1% of the control level at the highest concentration for FK614, rosigl itazone, and piogl i tazone, respectively. Table 1 also shows that the corepressor SMRT was associated with unliganded hPPARγ2, and Fig. 6 shows the effects of each ligand on the release of SMRT from hPPARγ2. The results shown in Fig. 6 indicate a concentration-dependent release of SMRT from hPPARγ2 in response to the addition of each of the three ligands. Each of FK614 and rosigl i tazone significantly induced the dissociation of SMRT-hPPARγ2 complex at 1 x 10^"9 M and higher concentrations. Pioglitazone significantly induced the dissociation of SMRT-hPPARγ2 complex at 1 x 10^"9 M, 1 x 10^"r M, and higher concentrations. The hPPARγ2 bindings to SMRT were 4.6%, 3.0%, and 12.1% of the control level at the highest concentration for FK614, rosigl i tazone, and piogl itazone, respectively. Accordingly, these results indicate that the FK614 promotes the release of corepressors which interact with uni iganded PPARγ, as efficiently as the other TZD compounds.

Fig. 7 shows effects of FK614, rosigl i tazone, and pioglitazone on the recruitment of CBP to hPPARγ2. These results indicate concentration-dependent increases in the association of CBP wi th hPPARγ2 in response to FK614, rosigl itazone, and pioglitazone. The significant induction of CBP recruitment by FK614, rosi l itazone, and piogl itazone was ini tiated at the concentrations of 1 x 10^~6, 1 x 10^~6, and 1 x 10^~6 M, respectively. The induction of CBP recruitment by FK614, rosigl itazone, and piogl i tazone at the highest concentration was 480.2%, 1538.4%, and 1083.1% of the basal level, respectively.

Fig. 8 shows effects of FK614, rosigl itazone, and pioglitazone on the recruitment of SRC-1 to hPPARγ2. A concentration-dependent association of SRC-1 with hPPARγ2 was observed in response to the addition of each of the three I igands. Each of FK614, rosigl i tazone, and pioglitazone significantly induced SRC-1 recrui tment to hPPARγ2 at the concentrations of 1 x 10^~5 M. The induction of SRC-1 recruitment by 1 x 10^~5 M of FK614, 1 x 10^~5 M of rosigl i tazone, and 1 x 10^~5 M of piogl itazone were, respectively, 1289.5%, 8039.2%, and 3418.3% of the basal level.

Fig. 9 shows effects of FK614, pioglitazone, and rosigl itazone on the recruitment of TRAP220 to hPPARγ2. These results indicate concentration-dependent increases in the association of TRAP220 with hPPARγ2 in response to FK614, pioglitazone, and rosigl itazone. The initiation concentrations for inducing TRAP220 recrui tment by either of FK614, pioglitazone, and rosigl itazone were, respectively, 1 x 10^"6 M, 1 x 10^"δ M, and 1 x 10^"7 M. The amount of TRAP220 recruitment induced by either of FK614, pioglitazone, and rosiglitazone at 1 x 10^~5 M was, respectively, 390.1%, 540.9%, and 668.5% above the basal level. All the three compounds promoted recruitment of each of the coactivators, CBP, SRC-1, and TRAP220, to hPPARγ2. However, the magnitudes of recruitment of individual coactivators were varied with respect to each compound. The magnitude of CBP recrui tment induced by FK614, piogl itazone, and rosigl i tazone at the highest concentration (1 x 10^"5 M) was 480.2%, 1083.1%, and 1538.1% of basal level, respectively (Fig. 10). Likewise, the magnitude of SRC-1 recruitment induced by each compound at the highest concentration (1 x 10^~5 M) was 1289.5%, 3418.3%, and 8039.2% of basal level, respectively (Fig. 11). The amount of CBP and SRC-1 recruitment by FK614 is much lower than the other compounds. In contrast, FK614, pioglitazone, and rosiglitazone induced TRAP220 binding to hPPARγ2 with 390.1%, 540.9%, and 668.5% of basal level within the similar range of magnitude at the highest concentration (1 x 10^~5 M) (Fig. 12). FK614, pioglitazone, and rosiglitazone result in different PPARγ-coactivator interaction.

FK614 was much less efficacious than each of the other TZDs as for the recruitment of CBP and SRC-1, whereas the levels of TRAP220 recruitment measured using coactivator recruitment analysis were with similar magnitude of response for all three ligands. Thus, these results suggest that different PPARγ ligands can have different properties in PPARγ activation in different tissues, cells or the stages of differentiation depending on the types or the content of coactivators in the target eel Is. TRAP220 is known to act as a PPARγ-selective coactivator and to be required for PPARγ-stimulated adipogenesis (Ge, K. , et al., Nature 417:563-567, 2002).

1-4. Inhibition of endogenous I igand-binding

Effect of FK614 on 9(S)-H0DE-promoted coactivator recrui tment was examined by in vi tro GST pull-down assay. The assay was performed in a similar manner as described above in 1-2 except for adding 9(S)-H0DE in the reaction mixture and using only CBP as a coactivator. 9(S)-H0DE (solution in 95% ethanol) was purchased from Larodan Fine Chemicals. To change the solvent, ethanol was evaporated under a gentle stream of nitrogen gas, and remaining 9(S)-H0DE was dissolved in DMSO. Fig. 13 shows effects of FK614 on the recruitment of CBP to PPARγ-induced by 9(S)-H0DE. 9(S)-H0DE promoted the recruitment of CBP to PPARγ. The induction of CBP recruitment by 9(S)-H0DE (1.0 x 10^"4M) was 4.77 fold above the basal level, and about 55% of that observed with rosiglitazone (8.72 fold above the basal at the concentration of 1.0 x 10^"5M). FK614 also induced the coactivator recruitment. The magnitude of the coactivator recruitment induced by FK614 (1.0 x 10^"5M) was 2.14 fold above the basal, and was about 45% of that achieved by 9(S)-H0DE. As 1.0 x 10^"5M FK614 shows its maximal response in this assay, these results indicate that 9(S)-H0DE is more efficacious than FK614 in this assay.

When examined in the presence of 1.0 x 10^"4 M 9(S)-H0DE, FK614 was found to cause the inhibition of the coactivator recruitment in a concentration-dependent manner, with the maximum inhibition level being same as observed when stimulated by FK614 alone.

These results demonstrate that the biological effects of 9(S)-H0DE, an efficacious endogenous agonist for PPARγ, can be modulated negatively by FK614.

Example 2: aP2 Gene expression in mature adipocytes

Activation of PPARγ by each of FK614, rosigl i tazone, and piogl itazone was analyzed by measuring the gene expression levels of adipocyte fatty acid binding protein (aP2) in mature 3T3-L1 adipocytes. FK614, rosiglitazone, and pioglitazone were dissolved in DMSO at the concentration of 1.0 x 10^"Z . Dissolved test compounds were then serial ly di luted 10-folds with DMSO to 1000 x concentration of the final.

For the preparation of test compound-treated mature adipocytes, 3T3-L1 eel Is (2x10⁴ eel Is/mL; obtained from the JCRB Cel I Bank) were incubated with 3 mL of the culture medium, which was prepared by supplementing DMEM with 10% FBS, 100 units/mL penicillin, and 0.1 mg/mL streptomycin, in a 6—we11 plate at 37°C with a 5 %C0₂ atmosphere on Day 1. After a two-day incubation, the medium was replaced with 2.5 mL of a differentiation-inducing medium. The differentiation-inducing medium is prepared from the culture medium by adding dexamethasone (DEX), 3-isobuty 1-1-methylxanthi ne (IBMX), and insulin (SIGMA ) to the final concentrations of 0.25 μM, 0.5mM, and 10 μg/mL, respectively. After a three-day incubation, the medium was replaced with 2.5 mL of the maintaining medium, which was prepared by adding insulin to the culture medium at a final concentration of 10 μg/mL. After further two-day incubation, the medium was replaced with 2.5 mL of the culture medium. The medium was replaced with 2.5 mL of the flesh culture medium on Day 10 and Day 13. On Day 14, 2.5 μL of the 1000 x concentrations of test compound solution was added to the medium in the culture wells. Thus, each test compound was added at 1 x concentration of the final with 0.1% of DMSO. Final . 0.1% of DMSO was contained in the medium for the non-treated control eel Is without test compound. Then, the test compound-treated cells were incubated for 48 hrs. The medium was removed and the treated cells were lysed by addition of 1 mL of TRIZOL. Then, the lysed samples were stored at -80°C unti I total RNA was isolated to serve as experimental samples.

The aP2 gene expression level was measured by quantitative real-time RT-PCR analysis as fol lows.

Total RNA was isolated from the test compound-treated adipocytes using TRIZOL. Total RNA was treated wi th DNase RT Grade (Nippon gene) to remove residual genomic DNA. After determining the concentration of total RNA solution by measuring absorption at 260 nm, 100 μL of reverse transcription (RT) product was synthesized from 1 μg of total RNA using Taq Man Reverse Transcription Reagent (Applied Biosystems) and Random Hexamer (Applied Biosystems) as primers. One-fortieth (2.5 μL) of each RT product was ampI if ied by ABI PRISM 7700 Sequence Detection System (Applied Biosystems) using SYBR Green PCR Master Mix (Applied Biosystems). After initial denaturation at 50°C for 2 min and 95°C for 10 min, reactions were cycled 40 times using the following parameters: 95°C for 15 sec, primer annealing and extension at 60°C for 1 min. Following each cycle, the fluorescence of the double-stranded products were detected. As an internal control, the expression of the house keeping gene cyclophilin B (CPB) was also evaluated. The following ol igonucleotide primers specific for mouse aP2 gene (GenBank accession no. K02109) and mouse CPB gene (GenBank accession no. M60456) were used: aP2, ACTGGGCGTGGAATTCGAT (sense, SEQ ID NO: 1) and CGCCATCTAGGGTTATGATGCT (ant i sense, SEQ ID NO: 2); CPB, CAGGAGAGAAAGGATTTGGCTACA (sense, SEQ ID NO: 3) and TCCACCCTGGATCATGAAGTC (antisense, SEQ ID NO: 4). RT-PCR products of 71 bp (aP2) and 74 bp (CPB) were confirmed by a 2.5% (w/v) agarose gel electrophoresis. Reverse Transcription products from the mature adipocyte sample was amplified using each set of primers, dNTPs and AmpliTaq Gold DNA Poly erase (Applied Biosystems) with the attached buffer instead of SYBR Green PCR Master Mix, and then those PCR products were used as PCR standards. Three-fold serial dilutions of these PCR products were amplified in parallel with the experimental samples as described above. Based on the amplification curves of the standards, a standard curve was generated for each PCR product using the Sequence Detection System software (Applied Biosystems). The amplification curves of the experimental samples were plotted against these standard curves to generate an estimated arbitrary unit of gene-specific mRNA. To account for differences in RT efficiency among the experimental samples, aP2 and CPB genes were ampl if ied from the same experimental RT product and the data were normal ized and expressed as a ratio, which was termed the relative expression of aP2, of aP2 arbitrary uni t/CPB arbitrary unit. Furthermore, PCR was carried out in paraI lei reactions in which the RT product was omitted as no template control. Data are expressed as mean ± S.E. An increase in the expression level of aP2 gene in the test compound-treated groups was calculated as percentage to that in the control group.

Fig. 14 shows the effects of FK614, rosiglitazone, and piogl i tazone on aP2 gene expression in mature 3T3-L1 adipocytes. The concentration-response relationship for mature adipocytes indicated that FK614, rosiglitazone, and piogl i tazone ini t iated to induce aP2 gene expression at 1 x 10^-7 M, 1 x 10^"8 M, and 1 x 10^"r, respectively. The maximal induction of aP2 gene expression by FK614, rosiglitazone, and pioglitazone in mature cells were about 290%, 650%, and 440% of the basal level, respectively. This shows that whi le being as potent as TZDs, FK614 is less efficacious than the other TZD compounds in mature adipocytes. The results indicate that FK614 does not excessively activate PPARγ in mature adipocytes, suggesting that it eliminates excessive hypertrophy in mature adipocytes.

Example 3: aP2 Gene expression in differentiating adipocytes

Effects of FK614, rosiglitazone, and piogl i tazone, on PPARγ activation were measured by using aP2 gene expression in differentiating 3T3-L1 adipocytes. The experiment was performed in a similar manner as described in Example 2 except for using differentiating adipocytes in place of mature adipocytes. Test compound-treated differentiating adipocytes were prepared as follows.

3T3-L1 eel Is (2x10⁴ eel Is/mL) were incubated wi th 3 L of the culture medium in a 6 well plate at 37°C with a 5% C0₂ atmosphere. After a two-day incubation, the medium was replaced with 2.5 mL of a differentiation-inducing medium. After a further two-day incubation, the medium was replaced with 2.5 L of the culture medium. Next day, 2.5 μL of the 1000 x concentrations of test compound solution was added to the medium in the culture wel Is. Thus, each test compound was added to 1 x concentration of the final with 0.1% of DMSO. Final 0.1% of DMSO was contained in the medium for the non-treated control cells without test compound. Then, the test compound-treated cells were incubated for 48 hrs. The medium was removed, and treated cells were lysed by addition of 1 mL of TRIZOL (GIBCO BRL). Then, the lysed samples were stored at -80 °C until total RNA was isolated.

Fig. 15 shows effects of FK614, rosiglitazone, and piogl i tazone on aP2 gene expression in differentiating 3T3-L1 adipocytes. Analysis of aP2 gene expression in differentiating adipocytes revealed that aP2 gene expression level was increased in response to each of FK614, rosiglitazone, and pioglitazone in a concentration-dependent manner. It was observed that FK614, rosiglitazone, and piogl i tazone ini tiated to induce aP2 gene expression at 1 x 10^~7 M, 1 x 10^~8 M, and 1 x 10^~7 M, respectively. The maximal levels of aP2 gene expression induced by FK614, rosigl itazone, and piogl i tazone were about 220%, 230%, and 200% of the basal level, respectively. Accordingly, the results indicate that FK614 is a potent ligand that activates PPARγ as efficacious as the other TZD compounds in differentiating adipocytes.

The results of Examples 2 and 3 indicate that FK614, rosiglitazone, and pioglitazone exert stage-specific different effects on the induction of the aP2 gene expression in differentiating and mature 3T3-L1 adipocytes. They also show that FK614 behaves as a full agonist for PPARγ in differentiating adipocytes and as a partial agonist for PPARγ in mature adipocytes.

Example 4: Measurement of triglyceride accumulation The ability of each of FK614, rosiglitazone, and pioglitazone to promote adipocyte differentiation was measured in terms of the accumulation of triglyceride in differentiating mouse 3T3-L1 cells. All materials used in this experiment were same as those used in Example 2.

Each test compound of FK614, rosigl i tazone, and piogl i tazone, was dissolved in DMSO at the concentration of 1.0 x 10^~2M, and 1000 x concentration of the final was prepared by 10-folds serial dilution with DMSO. The 1000 x concentrates were further diluted with the culture medium to 1 x concentration of the final, and then added into each well (the final concentration of DMSO was 0.1%).

3T3-L1 eel Is (2x10⁴ eel Is/L) were incubated with 3 mL of the culture medium in type I collagen-coated 6-well plate at 37°C with a 5% C0₂ atmosphere. After a two-day incubation, the medium was replaced with the differentiation-inducing medium and incubated at 37°C for two days. The medium was then replaced with the culture medium containing 1 x concentration of the test compound. After three days, the medium was replaced with a fresh culture medium containing 1 x concentration of the test compound. Two days later, the medium was removed and the intracel lular triglycerides were extracted by the addition of 6 mL of isopropanol. Then the extracted solution was di luted 1 :3 using isopropanol. The extracted triglyceride content was assayed by the acetylacetone method using a kit from Wako Pure Chemical Industries, Ltd. The concentration of cellular protein was determined using DC Protein Assay (BIO RAD), after solubi I izing with 0.1 mol/L of NaOH (0.8 mL/well). Cellular triglyceride content was calculated as per mg protein.

Data are expressed as mean ± S.E. Statistical significance between the non-treated control group and drug treated groups was assessed by analysis of variance, based on randomized block design, followed by Dunnett's multiple comparisons. Statistical significance was assumed at P < 0.05.

Fig. 16 shows effects of FK614, rosiglitazone, and pioglitazone on the triglyceride accumulation in differentiating mouse 3T3-L1 cells. Concentration-dependent increases in triglyceride content were observed in response to FK614, rosiglitazone, and pioglitazone. Significant increases in triglyceride content were initiated at the concentrations of 1 x 10^~7 M, 1 x 10^~8 M, and 1 x 10^"7 M for FK614, rosiglitazone, and piogl i tazone, respectively. It was observed that FK614, rosiglitazone, and pioglitazone initiated to increase triglyceride content in the cells significantly at the concentration of 1 x 10^-7 M, 1 x 10^"8 M, and 1 x 10^~7 M, respectively. Simi lar maximal levels of triglyceride content, about 2.5-fold of the basal levels, were observed for all the three compounds. Accordingly, the results indicate that the effect of each compound to promote adipocyte differentiation can be assessed by measuring triglyceride accumulation in adipocytes.

Example 5: PDE3B gene expression

TNFα has been reported to induce the down-regulation of PDE3B gene expression (Rahn Landstro , T. et al., Biochem. J. 346:337-343, 2000). In this experiment, this function of TNFα was confirmed (data not shown) and long-term effects of FK614, rosiglitazone, and pioglitazone on PDE3B gene expression in mature adipocytes were examined. Al I materials used in this experiment were same as those used in Example 2 unless otherwise specified.

Three solutions for three-dimensional col lagen gel cul ture were prepared: A solution, acid-soluble type I col lagen solution (pH3, Ni tta Gelatin) ; B solution, 5-fold concentrated DMEM (without NaHC0₃) ; and C solution, reconstruction buffer (2.2% NaHC0₃, 200 mM HEPES, 0.05 N NaOH).

Seven volumes A solution, 2 volumes B solution and 1 volume C solution were sufficiently mixed and kept in ice. The mixture was then mixed with the dissociated 3T3-L1 cells (2x10⁵ cells/mL). Eight hundred microlitters of the mixture containing 3T3-L1 cells were placed in a type I collagen-coated 12—we11 plate and were incubated at 37°C for 30 min. After gelation, the wells were overlaid wi th 3 mL of the cul ture medium. The eel Is were incubated at 37°C with a 5% C0_Z atmosphere. After a two-day incubation, the medium was replaced with 3 mL of the di fferenti at ion-inducing medium. After a further three-day incubation, the medium was replaced with 3 mL of the maintainig medium. On Day 8, the medium was replaced with 3 mL of the cul ture medium and 3.8 μL of the 1000 x concentrations of the test compound solution. Thus, each test compound was added at 1 x concentration of the final with 0.1% of DMSO. Final 0.1% of DMSO was contained in the medium for the non-treated control eel Is without test compound. The medium was replaced with 3 mL of the flesh culture medium, and 3 μL of the 1000 x concentrations of the test compound solution was repeatedly added to the replaced medium on Day 10, 13, 15, 17, 20 and 22. On Day 24, the medium was removed and the adipocyte-embedded gel was scraped off and minced in the test tube. The culture medium was added up to 1.5 mL in the test tube. The gel was digested by addition of 30 μL of Collagenase S-1 (10 mg/mL, Nitta Geratin) and incubated at 37°C for 30 min. After centrifugation, supernatant was removed and then the adipocytes were lysed by addition of 1 mL of TRIZOL (GIBCO BRL). The lysate was stored at -80°C until total RNA was isolated.

The PDE3B gene expression level was measured by quantitative real-time RT-PCR analysis in a simi lar manner as described in Example 2 except that 45 cycles of PCR was performed using the fol lowing ol i onucleotide primers specific for mouse PDE3B gene(GenBank accession no. AJ132271) to obtain RT-PCR products of 80 bp in length.

CCAGGTGTGCATCAAATTAGCA (sense, SEQ ID NO: 5) CAATGCCTTCTGTCCATCTCAA (ant i sense, SEQ ID NO: 6)

Fig. 17 shows long-term effects of FK614, rosiglitazone, and piogl i tazone on PDE3B gene expression in mature 3T3-L1 adipocytes. Concentration-dependent decreases in PDE3B gene expression were observed in response to rosigl i tazone and pioglitazone. The maximal percentage of the decrease of PDE3B gene expression level from the control was 42.0 % for rosiglitazone and 28.3 % for pioglitazone. A marked contrast was that down-regulation of PDE3B gene expression was not observed in response to FK614. Thus, the long-term treatment of mature adipocytes with rosiglitazone and pioglitazone induced the down-regulation of PDE3B mRNA, whereas FK614 had no effect on PDE3B gene expression in ful ly differentiated mature adipocytes.

Industrial Appl icabi I i ty

Thiazol idinedione (TZD) compounds such as rosiglitazone and pioglitazone are known to be PPARγ full agonists, and are expected to be antidiabetic agents that improve insulin sensitivity, as used widely in the treatment of NIDDM. However, their long-term administration to NIDDM patients tends to cause obesity that often leads to insulin resistance. Thus, an antidiabetic agent that lowers blood glucose level without inducing obesity and/or insulin resistance has been strongly desired from the cl inical aspect. The screening methods of the present invention can provide candidate compounds for antidiabetic agents, in particular, useful for treating NIDDM. Furthermore, the screening methods of the present invention can provide the compounds that can lower blood glucose level without inducing obesity and/or insulin resistance. It thus has less side effects compared to the TZD compounds without reducing its blood glucose-lowering effect.

Claims

1. A method of screening for a candidate compound for an antidiabetic agent that lowers blood glucose level wi thout inducing obesity and/or insul in resistance, wherein said method comprises the steps of:

(a) providing a compound that is a PPARγ agonist;

(b) contacting mature adipocytes with the compound provided by step (a) and measuring the expression level for a PPARγ-induced adipogenic gene in the mature adipocytes; (c) contacting mature adipocytes with the PPARγ ful I agonist and measuring the expression level for a PPARγ-induced adipogenic gene in the mature adipocytes;

(d) comparing the expression level measured in step (b) with that measured in step (c) ; and

(e) selecting a compound whose contact results in a significantly lower expression level in step (b) in comparison with that in step (c).

2. The method of claim 1, further comprising the steps of:

(f) contacting differentiating adipocytes with the compound selected in step (e) and evaluating the differentiation stage of the adipocytes; (g) contacting differentiating adipocytes with the PPARγ full agonist and evaluating the differentiation stage of the adipocytes;

(h) comparing the differentiation stage evaluated in step (f) with that evaluated in step (g) ; and

(i) select ing a compound whose contact results in a substantial ly same stage of adipocyte differentiation in step (f) in comparison with that in step (g).

3. The method of claim 2, wherein the differentiation stage is determined by measuring the expression level of the PPARγ-induced adipogenic gene or the accumulation of triglycerides in adipocytes.

4. The method of claim 1 or 3, wherein the PPARγ-induced adipogenic gene is either adipocyte fatty acid binding protein (aP2) gene or fatty acid binding protein (FABP) gene.

5. The method of any one of claims 1 to 4, wherein the adipocytes are either human adipocytes or mouse adipocytes.

6. The method of any one of claims 1 to 5, wherein the PPARγ full agonist is at least one member selected from the group consisting of 9(S)-hydroxyoctadecadienoic acid (9S-H0DE), 13 (S)-hydroxyoctadecad i enoi c acid (13S-H0DE), eicosapentaenoic acid, 15-deoxy-del ta 12, 14-prostagl and i n J2, rosiglitazone, and pioglitazone.

7. The method of any one of claims 1 to 6, wherein step (a), the PPARr agonist is selected by the steps of:

(a-1) contacting PPARr and a non-selective coactivator with a compound and measuring the binding activity of PPARr to the non-selective coactivator;

(a-2) contacting PPARr and a non-selective coactivator with a PPARr full agonist and measuring the binding activity of PPARr to the non-selective coactivator;

8. The method of claim 1, wherein step (a), the PPARr agonist is selected by the steps of:

(a-5) contacting PPARr and a PPARr-selective coactivator wi th a compound and measuring the binding activity of PPARr to the selective coactivator;

(a-7) comparing the binding activity measured in (a-5) with that measured in (a-6) ; and

(a-8) selecting a compound whose contact results in a substantially same binding activity in (a-5) in comparison with that in (a-6).

9. The method of claim 1, wherein step (a), the PPARr agonist is selected by the steps of:

(a-9) contacting PPARr and a non-selective coactivator with either an endogenous ligand for PPARr or a synthetic PPARr full agonist in the presence of a test compound and measuring the binding activity of PPARr to the non-selective coactivator; (a-10) contacting PPARr and a non-selective coactivator with either an endogenous ligand for PPARr or a synthetic PPARr full agonist in the absence of a test compound and measuring the binding activity of PPARr to the non-selective coactivator;

(a-11) comparing the binding activi ty measured in (a-9) with that measured in (a-10) ; and (a-12) selecting a compound whose presence results in the reduced binding activity in (a-9) in comparison with that in (a-10).

10. The method of claim 7 or 9, wherein said non-selective coactivator is either CREB binding protein (CBP) or steroid receptor coactivator-1 (SRC-1).

11. The method of claim 8, wherein said PPARr-selective coactivator is either peroxisome prol iferator-activated receptor r coactivator-1 (PGC-1) or TRAP220.

12. The method of claim 9, wherein said endogenous ligand for PPARr is 9(S)-hydroxyoctadecadienoic acid (9S-H0DE), 13 (S)-hydroxyoctadecad ienoi c acid (13S-H0DE), eicosapentaenoic acid, or 15-deoxy-del ta 12, 14-prostagland i n J2

13. The method of any one of claims 7, 8, and 9, wherein said binding activity between PPARr and said coactivator is measured using ei ther in vitro QSl pul l-down assay or mammalian two-hybrid assay.

14. The method of any one of claims 1-13, further comprising the steps of: (j) contacting mature adipocytes with the compound selected in step (e) or step (i) and measuring the expression level of a gene, of which down-regulation is involved in the development of insulin resistance;

(k) contacting mature adipocytes wi th the PPARγ full agonist and measuring the expression level of a gene, of which down-regulation is involved in the development of insulin resistance; (I) comparing the expression level measured in step (j) with that measured in (k) ; and

15. The method of claim 14, wherein the gene, of which down-regulation is involved in the development of insulin resistance, is leptin or phosphodiesterase 3B (PDE3B) gene.

16. A kit for screening for a compound that lowers blood glucose level without inducing obesity and/or insulin resistance, the kit comprising

(a) adipocytes;

(b) a culture medium for the adipocytes;

(c) a PPARγ full agonist; and

(d) an ol igonucleotide primer set for the PPARγ-induced gene.

17. The kit of claim 16, wherein the adipocytes (a) are at least one member selected from the group consisting of (1) mature adipocytes; (2) differentiating adipocytes; and (3) preadipocytes.

18. The kit of claim 16, wherein the adipocytes are 3T3-L1 cells.

19. A compound obtainable by the method of any one of claims 1-15_% which lowers blood glucose level without inducing obesity and/or insulin resistance.

20. A pharmaceutical composition comprising the compound of claim 19, or its pharmaceutically acceptable salt, and a pharmaceutically acceptable carrier.

21. A method for lowering the blood glucose level wi thout causing obesi ty and/or insulin resistance by administering the compound of claim 19, or its pharmaceutically acceptable salt, to a subject in need thereof.

22. Use of the compound of claim 19 for the manufacture of a medicament for lowering the blood glucose level without causing obesity and/or insulin resistance.