点击显示 收起
【关键词】 Differential
Despite their proven antidiabetic efficacy, widespread use of peroxisome proliferator-activated receptor (PPAR) agonists has been limited by adverse cardiovascular effects. To overcome this shortcoming, selective PPAR modulators (SPPARMs) have been identified that have antidiabetic efficacy comparable with full agonists with improved tolerability in preclinical species. The results of structural studies support the proposition that SPPARMs interact with PPAR differently from full agonists, thereby providing a physical basis for their novel activities. Herein, we describe a novel PPAR ligand, SPPARM2. This compound was a partial agonist in a cell-based transcriptional activity assay, with diminished adipogenic activity and an attenuated gene signature in cultured human adipocytes. X-ray cocrystallography studies demonstrated that, unlike rosiglitazone, SPPARM2 did not interact with the Tyr473 residue located within helix 12 of the ligand binding domain (LBD). Instead, SPPARM2 was found to bind to and activate human PPAR in which the Tyr473 residue had been mutated to alanine (hPPARY473A), with potencies similar to those observed with the wild-type receptor (hPPARWT). In additional studies, we found that the intrinsic binding and functional potencies of structurally distinct SPPARMs were not diminished by the Y473A mutation, whereas those of various thiazolidinedione (TZD) and non-TZD PPAR full agonists were reduced in a correlative manner. These results directly demonstrate the important role of Tyr473 in mediating the interaction of full agonists but not SPPARMs with the PPAR LBD, thereby providing a precise molecular determinant for their differing pharmacologies.
The peroxisome proliferator-activated receptors (PPARs) , , and compose a nuclear receptor subfamily that modulates the transcription of a large compendium of genes encoding proteins that regulate lipid metabolism, cell differentiation, and signal transduction in a ligand-dependent manner (Berger and Moller, 2002). PPAR has been shown to be a master regulator of adipogenesis and nutrient metabolism in adipocytes where it is highly expressed. Thiazolidinedione (TZD) PPAR full agonists have demonstrated clinical efficacy for the treatment of type 2 diabetes mellitus (T2DM) patients (Berger et al., 2005). However, the use of these insulin-sensitizing agents has been restricted because of their association with several adverse effects, including plasma volume expansion and edema that can cause or exacerbate congestive heart failure, as well as weight gain, which seems to result from both increased adiposity and plasma volume. Thus, there exists a critical need to develop PPAR-targeted insulin-sensitizing agents that display robust antidiabetic efficacy with improved tolerability.
To meet this challenge, we and others have identified and characterized promising selective PPAR modulators (SPPARMs). SPPARMs are PPAR ligands that serve as partial agonists of the receptor in cell-based transcriptional activity and adipogenesis assays (Berger et al., 2005). They have also been shown to generate attenuated gene signatures in rodent adipocytes in vitro and adipose tissue in vivo. The superior therapeutic window of several SPPARMs has been established in preclinical species. For example, nTZDpa demonstrated robust insulin-sensitizing activity in obese C57/BL6 mice with attenuated adverse effects on weight gain, adiposity, and cardiac hypertrophy relative to a potent PPAR full agonist (Berger et al., 2003). Compounds 24 (Acton et al., 2005) and 5 (Dropinski et al., 2005) are structurally unique SPPARMs that demonstrated antihyperglycemic efficacy comparable with rosiglitazone in diabetic db/db mice with reduced weight gain at similar plasma exposures. The SPPARM FK614 has been shown to have significant antidiabetic activity in mice with diminished hemodilution and cardiac hypertrophy in rats (Minoura et al., 2004). Remarkably, the angiotensin receptor blocker telmisartan exhibited SPPARM activity, including robust amelioration in insulin resistance in diet-induced obese mice without increasing weight gain (Schupp et al., 2005). A novel SPPARM, PA-082, has been described that, compared with PPAR full agonists, has a diminished ability to induce lipid accumulation during in vitro adipogenesis of preadipocytes but maintains the ability to ameliorate cytokine-induced insulin resistance in cultured adipocytes to a similar extent (Burgermeister et al., 2006).
Previously published data seem to indicate that the unique properties of the SPPARMs may be due to their distinct physical interaction with the receptor, resulting in diminished conformational stability of the receptor compared with full agonists. Indeed, X-ray cocrystallographic studies of the PPAR ligand binding domain (LBD) complexed with the full agonist rosiglitazone indicated that the negatively charged nitrogen of the TZD ring of rosiglitazone was within hydrogen-bonding distance of the tyrosine (Y) 473 side-chain hydroxyl group in helix 12 of the human PPAR LBD (Nolte et al., 1998). In contrast, X-ray cocrystallography and molecular modeling studies performed with the PPAR LBD and SPPARMs (Oberfield et al., 1999; Burgermeister et al., 2006) indicate that the carboxylic acid moiety of such ligands was not within hydrogen-bonding distance of Tyr473. Consonant with their different binding modes, biochemical and NMR studies have demonstrated that SPPARMs induce a unique and less stable receptor conformation than PPAR full agonists (Berger et al., 2003). Taken together, these studies suggest that Tyr473 is a critical site of interaction between the PPAR LBD and full agonists but not SPPARMs. Thus, Tyr473 is implicated as having an important role in the stabilization of the LBD of the receptor by the former but not the latter class of ligands. Furthermore, because Tyr473 lies within the helix 12 transcriptional activation function-2 domain that composes a section of the transcriptional coactivator binding pocket of LBD (Nolte et al., 1998; Sheu et al., 2005), the inability of SPPARMs to bind to and thereby directly stabilize this region may serve as the physical basis for the differential receptor-coactivator interactions, attenuated transcriptional activity, and resulting improved tolerability observed in preclinical species with these ligands (Kintscher et al., 2000; Berger et al., 2003; Minoura et al., 2004; Acton et al., 2005; Dropinski et al., 2005; Schupp et al., 2005).
Here, we describe a novel, potent PPAR ligand, SPPARM2. We found that in comparison with rosiglitazone, SPPARM2 served as a partial agonist of human PPAR in a cell-based transcriptional activity assay, had a diminished ability to induce adipogenesis and triglyceride accumulation in primary cultures of human preadipocytes, and generated an attenuated gene signature in human adipocytes. X-ray cocrystallography studies localized the bound ligand at a considerable distance from the Tyr473 residue within the LBD of the receptor. It is noteworthy that neither the binding nor the transcriptional activity of SPPARM2 was significantly affected when it was assayed with a mutant PPAR in which the Tyr473 residue was changed to an alanine (hPPARY473A). In additional studies, we found that the Tyr473A mutation did not reduce the binding or transcriptional activities of a structurally diverse cohort of SPPARMs. In contrast, the intrinsic and functional potencies of various TZD and non-TZD PPAR full agonists were significantly reduced in a correlative manner.
Site-Directed Mutagenesis and Protein Purification. Generation of the chimeric receptor hPPARY473A/GAL4 was performed using the previously described mammalian expression plasmid containing the wild-type (WT) human PPAR LBD, pcDNA3-hPPAR/GAL4 (Elbrecht et al., 1999), oligonucleotides GCTCCTGCAGGAGATCGCCAAGGACTTGTACTAG and CTAGTACAAGTCCTTGGCGATCTCCTGCAGGAGC, and the QuikChange site-directed mutagenesis kit (Stratagene, La Jolla, CA) to produce the vector pcDNA3-hPPARY473A/GAL4. GST fusion protein constructs containing the hPPARWT or hPPARY473A LBD were generated by transferring the nucleotide fragments encoding amino acids 176 to 477 from the above-mentioned pcDNA constructs into a similarly cleaved pGEX-5X-2 vector (GE Healthcare, Chalfont St. Giles, UK). The fusion proteins were then expressed in Escherichia coli bacteria and purified on a glutathione-Sepharose column as described previously (Berger et al., 2003).
SPA Binding Assays. A previously described competition binding assay using GST-hPPAR and the radioligand [3H2]nTZD3 (Berger et al., 2003) was used to identify a series of structurally distinct PPAR ligands. To compare the potencies of PPAR ligands on the wild-type and Y473A mutant PPAR, novel competition binding assays using [3H]SPPARM1 were developed and performed using modified versions of previously described protocols (Berger et al., 2003). To characterize the binding activity of radiolabeled SPPARM1, saturation binding assays were performed using recombinant GST-hPPARWT LBD or GST-hPPARY473A LBD, anti-GST antibody (GE Healthcare), and radioligand [3H]SPPARM1 at 0 to 30 nM ± 50 µM nonradioactive SPPARM1 combined in the same assay buffer and assayed under similar conditions as described previously (Berger et al., 2003). Specific binding was determined from the difference between the total and nonspecific binding. The Kd value was derived from nonlinear regression analysis using the one-site binding equation in Prism (GraphPad Software Inc., San Diego, CA). Subsequently, competition binding assays were performed using recombinant GST-hPPARWT LBD or GST-hPPARY473A LBD, anti-GST antibody (GE Healthcare), and 2.5 nM [3H]SPPARM1 in the same buffers and under the same conditions as the aforementioned assays.
Cell Culture and Transient Transactivation Assay. COS-1 cells were cultured and maintained as described previously (Berger et al., 1999). Cells were transfected with the pcDNA3-hPPARWT/GAL4 or pcDNA3-hPPARY473A/GAL4 expression vector, pUAS(5X)-tk-luc reporter vector, and pCMV-lacZ (as an internal control for transactivation efficiency) using Lipofectamine (Invitrogen, Carlsbad, CA). After a 48-h exposure to compounds, cell lysates were produced, and luciferase and β-galactosidase activity in cell extracts was determined. Inflection points of normalized luciferase activity were calculated by a four-parameter logistic equation in Prism (GraphPad Software Inc.).
Adipogenic Differentiation of Primary Human Preadipocytes. Primary human preadipocytes derived from subcutaneous adipose depots (Zen-Bio, Inc., Research Triangle Park, NC) were differentiated by adding 10 µM rosiglitazone or SPPARM2 in a final concentration of 0.1% dimethyl sulfoxide in differentiation media (Zen-Bio, Inc.) for 5 days. The media were then replaced with adipocyte media (Zen-Bio, Inc.) for 11 days. On day 14, cells were washed two times with 1x phosphate-buffered saline (Invitrogen) and fixed with 10% Formalin for 10 min. The cells were then stained with Oil Red O (Sigma-Aldrich, St. Louis, MO) for 1 h at room temperature, and digital images were captured using a phase contrast inverted Diaphot 300 microscope (Nikon, Tokyo, Japan) at a magnification of 10x.
Quantitation of Triglyceride in Differentiated Primary Human Preadipocytes. The intracellular triglyceride content within human preadipocytes differentiated for 14 days as described above with 0 to 10 µM rosiglitazone, SPPARM2, or nTZDpa was assayed with Adipored reagent using the protocol provided by the manufacturer (Lonza Bioscience, Basel, Switzerland).
Assessment of Gene Expression in Differentiated Primary Human Preadipocytes. RNA from primary human adipocytes differentiated for 14 days as described above with 10 µM rosiglitazone, SPPARM2, or nTZDpa was isolated using mRNA Catcher plates according to protocols from the manufacturer (Sequitur, Natick, MA). After reverse transcription of RNA using random hexamers, quantitative PCR of each target cDNA was performed with Taqman PCR reagent kits in the ABI Prism 7700 sequence detection system according to the protocols provided by the manufacturer (Applied Biosystems, Foster City, CA). The levels of mRNA were normalized to the amount of glyceraldehyde-3-phosphate dehydrogenase mRNA detected in each sample.
X-Ray Cocrystallography Studies of PPAR Ligands and the PPAR LBD. Purified PPARLBD (residues Gln203 to Tyr477) (Berger et al., 2003) at 10 to 15 mg/ml was mixed with SPPARM2 at a 1.1:1 M ratio of compound/protein on ice and allowed to stand at 4°C overnight. The solution was clarified, if necessary, by centrifugation before use. Crystals were grown by vapor diffusion at room temperature in 2-µl "sitting" drops that contained equal volumes of the protein complex solution and a reservoir solution consisting of 100 mM Tris-HCl, pH 8.0, 0.65 to 0.90 M sodium citrate, and 1 mM tris(2-carboxyethyl) phosphine hydrochloride against 0.5 to 1.0 ml of reservoir solution in a Cryschem MVD-24 crystallization tray (Charles Supper, Natick, MA). Suitable crystals would appear in 1 to 3 weeks. Single crystals were transferred to 10 to 20 µl of 100 mM Tris-HCl, pH 8.0, 1.44 M Na3 citrate, and 1 mM tris(2-carboxyethyl) phosphine hydrochloride for 5 to 10 min at room temperature, and they were vitrified by plunging the nylon-loop-captured crystals directly into liquid nitrogen. Crystals were shipped overnight in a dry shipper at liquid nitrogen temperature to the Advanced Photon Source at Argonne National Laboratories (Argonne, IL). The dry shipper was refilled with liquid nitrogen, and single crystals were cryogenically transferred to a magnetic mount under a cold (-170°C) nitrogen gas stream generated by a CryoStream model 600 (Oxford Cryosystems, Oxford, UK) for the duration of the diffraction experiment. X-ray diffraction data were collected at the ID-17 beamline of the Industrial Macromolecular Crystallography Association Collaborative Access Team as a series of 540 0.5° rotation images using an ADSC Quantum 210 2 x 2 charge-coupled device area detector (Area Detection Systems Corporation, Poway, CA) mounted on a Crystal Logic goniometer (Crystal Logic, Los Angeles, CA). Exposure time was set to keep low-resolution reflection overloads at a minimum without significantly affecting the quality of the high-resolution data (1-3 s/exposure). Data were measured from a single crystal exhibiting C2 symmetry to a resolution of 2.0 Å. Measured intensities were reduced and scaled with DENZO and SCALEPACK from the HKL suite (Otwinowski and Minor, 1997) that yielded an average redundancy on measurements of approximately 5 and an Rmerge of 0.042 (Table 1). The structure was solved by molecular replacement. SPPARM2 was incorporated between rounds of model building using CHAIN (Sack, 1988) or XTALVIEW (McRee, 1999), and refinement was accomplished using CNX, version 2002 (Accelrys, San Diego, CA) (Brünger et al., 1987, 1998). A summary of the crystallographic data are found in Table 1. The quality of diffraction data is generally determined by the final resolution achieved (2.35 Å), the number of unique reflections collected (29,681, or 99.0% of all data to 2.35 Å) compared with the number of nonhydrogen atoms in the model (3973, or 7.5 reflections/atom), and how well the individual diffraction measurements combine with each other (Rsym <10% overall). The quality of the atomic model is best judged visually by how well it fits the electron density derived from the experimental data. However, resolution (<2.5 Å) will indicate the expected level of detail to be seen, Rcryst (<25)/Rfree (<30) will provide estimates of how well diffraction data calculated from the model will compare with the experimentally collected data. Plotting the two main chain angles ( and ) for each residue on a Ramachandran plot (>90% in most favored regions and 0% in disallowed regions for nonglycine residues) and determining the root mean square deviations of bond lengths, bond angles, etc., provides additional support for the accuracy of the atomic model. The coordinates may be found in the Protein Data Bank under ID code 2P4Y.
TABLE 1 Crystallographic data summary for the cocrystal structure of SPPARM2 with human PPARLBD
Identification of Novel SPPARMs. To identify novel SPPARMs, a series of indole-containing small molecules were chemically synthesized (Acton et al., 2005; Dropinski et al., 2005; Liu et al., 2005) and characterized by two in vitro pharmacological assays. First, we determined which of the compounds were PPAR ligands by performing SPA-based competition binding assays as described previously (Berger et al., 2003). Subsequently, we carried out PPAR cell-based transcriptional activity assays as described previously (Berger et al., 1999) to assess the functional activities of the ligands. Through these efforts, we were able to identify a series of novel indole-containing SPPARMs that demonstrate <40% of the maximal PPAR transcriptional activity of rosiglitazone (Table 2). An example of such a compound is SPPARM2, which was able to completely displace the radioligand [3H2]nTZD3 from PPAR, with greater potency than rosiglitazone in competition binding assays (Fig. 1A). SPPARM2 also activated the receptor with greater potency than rosiglitazone but only to a maximal level that was 20% of that attained with the full agonist (Fig. 1B). The same binding and transcriptional activity assays were used to identify and characterize a cohort of structurally diverse PPAR full agonists that possess 100 ± 30% of the maximal PPAR transcriptional activity of rosiglitazone (Table 2).
TABLE 2 Properties of PPAR ligands
Fig. 1. SPPARM2 is a potent PPAR ligand and partial agonist. A, binding of SPPARM2 and rosiglitazone to PPAR. Competition binding curves were generated by incubation of 5 nM [3H2]nTZD3 with GST-hPPAR. The displacement of radioligand after incubation in the presence of the indicated concentrations of SPPARM2 or rosiglitazone for 16 h is plotted. Similar results were obtained in at least two independent experiments performed in duplicate. B, activation of hPPAR by SPPARM2 or rosiglitazone. COS-1 cells were transiently cotransfected with pSG5-hPPAR/GAL4, pUAS(5X)-tk-luciferase and pCMV-lacz, and then they were incubated with the indicated concentrations of rosiglitazone, SPPARM2, or rosiglitazone for 48 h. Cell lysates were produced, and luciferase and β-galactosidase activity in cell extracts was determined. The means of triplicate samples from a representative experiment are plotted. Similar results were obtained in at least two independent experiments.
SPPARMs Induced Diminished Adipogenic Differentiation of Human Preadipocytes Compared with PPAR Full Agonists. The essential role of PPAR in adipocyte differentiation and the robust effects of PPAR full agonists on adipogenesis are well established (Berger et al., 2005). We have demonstrated that the SPPARM nTZDpa had a diminished ability to induce adipogenic differentiation of mouse 3T3-L1 preadipocytes in comparison with PPAR full agonists such as rosiglitazone (Berger et al., 2003); similar studies with SPPARM2 provided comparable results (data not shown). To extend this observation by assessing the human adipogenic potential of SPPARMs, we first examined the effects of SPPARM2, nTZDpa, and rosiglitazone on lipid accumulation in human subcutaneous preadipocytes. At day 14 of the differentiation regimen, the lipid content within the cells was first assessed by microscopic examination of cells stained with Oil Red O and then by a quantitative fluorescent lipid accumulation assay using Adipored reagent as described under Materials and Methods. Micrographs presented in Fig. 2A demonstrate that incubation of the preadipocytes with SPPARM2 resulted in the generation of fewer lipid-containing adipocytes than rosiglitazone. Consistent with these results, data from the Adipored assay indicated that although both the SPPARMs and rosiglitazone promoted lipid accumulation in a dose-responsive manner, maximum lipid accretion by cells incubated with the SPPARMs was greatly diminished compared with those treated with the full agonist (Fig. 2B).
Fig. 2. SPPARMs display diminished adipogenic action on human preadipocytes in comparison with a PPAR full agonist. A, differentiation of preadipocytes into lipid-filled adipocytes. Primary human preadipocytes derived were differentiated by adding 10 µM rosiglitazone or SPPARM2 for 5 days. The differentiation media were then replaced with adipocyte media for 11 days. On day 14, cells were washed, fixed, and stained with Oil Red O, and then phase contrast micrographs were taken at 10x magnification. B, quantitation of triglyceride in differentiated primary human preadipocytes. The triglyceride content of human preadipocytes differentiated for 14 days as described above with 0 to 10 µM rosiglitazone, SPPARM2, or nTZDpa was assayed with Adipored reagent as described under Materials and Methods. The means of triplicate samples from a representative experiment are plotted. Similar results were obtained in at least two independent experiments. C, determination of adipocyte marker gene expression in differentiated primary human adipocytes. RNA from primary human adipocytes differentiated for 14 days as described above with 10 µM rosiglitazone, SPPARM2, or nTZDpa was isolated and reverse transcribed. Quantitative PCR of FABP4, PEPCK, and ADFP cDNA was performed, and the level of each mRNA was normalized to the amount of glyceraldehyde-3-phosphate dehydrogenase mRNA detected in each sample. The figure plots the mean ± S.E.M. of triplicate samples of a representative experiment. Similar results were obtained in at least two independent experiments.
The adipogenic activity of PPAR activators are apparently mediated by their ability to alter the expression of gene constellations that regulate various elements of the adipocytic phenotype. A number of genes that are regulated by PPAR, including the direct, PPAR-response element-containing targets FABP4, PEPCK, and ADFP, have been shown to serve as markers of adipocyte differentiation (Hunt et al., 1986; Jiang and Serrero, 1992; Tontonoz et al., 1995). Therefore, we isolated RNA from cells that had undergone 14 days of adipogenic differentiation, and we compared the regulatory effects of SPPARMs and rosiglitazone on the expression of these three genes. As demonstrated in Fig. 2C, SPPARM2 and nTZDpa displayed a diminished ability to induce FABP4, PEPCK, and ADFP mRNA expression compared with rosiglitazone. Taken together, our data indicate that SPPARMs are less effective than a prototypical PPAR full agonist in promoting adipogenesis of human preadipocytes.
SPPARM2 Did Not Interact with the PPAR LBD in the Vicinity of Tyr473. The hPPAR-LBD/SPPARM2 complex crystallized with a monoclinic habit that contains two monomers in the asymmetric unit (a side-by-side view of the two monomers is depicted in Fig. 3). Monomer A assumed an "activated" conformation with the displaced helix 12 of a symmetry-related monomer occupying the coactivator peptide binding site, stabilizing the position of the local helix 12 so that the side chain of residue Tyr473 was pointing into the ligand binding pocket. It should be noted that the binding of the coactivator peptide analog (sequence H466-P-L-L-Q-E-I472) binds to monomer A in the same orientation seen in the ternary complex of hPPAR-LBD/rosiglitazone/SRC-1 peptide (Nolte et al., 1998), but the "charge clamp" residues Lys301 and Glu471 do not make direct hydrogen bonds to the backbone atoms of the bound helix. Monomer B, in contrast, assumed an "unactivated" conformation with the connected helix 12 displaced so that Tyr473 is unable to enter the ligand binding pocket. Furthermore, there was very little strong and continuous electron density for the connecting loop between helices 11 and 12 to allow for the unambiguous placement of residues 457 to 465. Other regions of considerable flexibility common to both monomers include the elbow between helix 2 and β strand 1 (residues 238-244) and the insert region between β strand 1 and helix 3 (residues 255-278). Model coordinates are omitted for these regions. In spite of this, SPPARM2 bound to both monomers in the same configuration, with the carboxylic acid hydrogen bonded to the main-chain nitrogen atom of Ser342 (Fig. 4). In monomer A only, there was additional electron density between the bound SPPARM2 and the residues that typically interact with the acid functionalities of full agonists His323, His449, and Tyr473. The shape of this density was consistent with tris(hydroxymethyl)-aminomethane, the buffer molecule that is present at nearly 100 mM, and has been modeled as such. However, this density may instead indicate the presence of a low (<5%) occupancy conformation of the bound SPPARM2 that may have bound as a full agonist with the compound's carboxylic acid interacting directly with the phenolic hydroxyl of Tyr473.
Fig. 3. The hPPARLBD crystallizes as an asymmetric unit containing activated and unactivated forms of the receptor monomer. The two monomers of the asymmetric unit are depicted side by side to highlight structural differences between the activated form on the left and the unactivated on the right. SPPARM2 and tris(hydroxymethyl)-aminomethane are depicted with orange and yellow carbons, respectively. Regions of missing density are depicted as light yellow strands. The crystals were generated and analyzed as described under Materials and Methods.
Fig. 4. SPPARM2 interacts with different residues within the hPPARLBD than rosiglitazone. View of SPPARM2 (orange) and "tris" (yellow) as cocrystallized with hPPAR and residues that fall within 5 Å of either. Rosiglitazone (purple) as cocrystallized with hPPAR (Protein Data Bank under ID code 2PRG) references the binding orientation for the typical agonist.
Effect of the Y473A Mutation on the Functional Activity of PPAR Ligands. We and others have provided X-ray cocrystallography and molecular modeling data supporting the proposition that PPAR full agonists interact directly with the Tyr473 residue in helix 12 of the LBD of the receptor, whereas SPPARMs do not (Oberfield et al., 1999; Burgermeister et al., 2006). To directly determine the functional importance of this differential receptor-ligand interaction, we used recombinant mutagenesis techniques to change the Tyr473 residue of PPAR to an alanine. The activities of SPPARM2 and rosiglitazone were then examined in transcriptional activity assays performed as described previously (Berger et al., 1999) in COS-1 cells transiently transfected with a GAL4-responsive reporter construct and a chimeric receptor composed of the GAL4 DNA-binding domain and either the hPPARWT or hPPARY473A LBD. As seen in Fig. 5A, the activity of rosiglitazone was nearly completely abrogated with hPPARY473A compared with hPPARWT. In contrast, only a negligible alteration in SPPARM2 potency was seen between the mutant and wild-type receptors. Furthermore, when the activities of numerous structurally diverse PPAR full agonist and SPPARM collections were determined in similar transactivation assays with hPPARWT and hPPARY473A, it was clearly demonstrated that there were marked reductions in activation maxima and potency for full agonists but not with SPPARMs (Table 2; Fig. 5B). To directly probe the significance of the carboxylic acid on SPPARM activity, the alkynyl-substituted SPPARM4, which is completely devoid of the acidic moieties present in on all other ligands, was synthesized. SPPARM4 exhibited comparable binding and functional activity in hPPARWT and hPPARY473A assays. In cocrystallization studies with hPPAR-LBD, SPPARM4 was found in the same position and orientation observed for SPPARM2 (our unpublished data). Rather than interact with the Ser342 NH like the carboxylate, the alkynyl substituent extended directly toward bulk solvent adjacent to the C of Gly284. The 15-fold decrease in potency observed when comparing the carboxylate SPPARM1 with its alkynyl analog SPPARM4 was commensurate with what might be expected for loss of a neutral-neutral hydrogen bond rather than for a charge-reinforced hydrogen bond, perhaps as a consequence of exposure to bulk solvent (Davis and Teague, 1999). These data further support the novel binding model (Schupp et al., 2005; Burgermeister et al., 2006) discussed for SPPARM2 and the lack of direct interactions with Tyr473.
Fig. 5. Transcriptional activities of PPAR full agonists but not SPPARMs are diminished with PPARY473A versus PPARWT. A, activation of PPARWT and PPARY473A by rosiglitazone or SPPARM2. COS-1 cells were transiently cotransfected with pcDNA3-hPPARWT/GAL4 or pcDNA3-hPPARY473A/GAL4 and both pUAS(5X)-tk-luciferase and pCMV-lacz, and then they were incubated with the indicated concentrations of rosiglitazone or SPPARM2 for 48 h. Cell lysates were produced, luciferase, and β-galactosidase activities in cell extracts were determined as. The figure plots the mean ± S.E.M. of triplicate samples of a representative experiment. Similar results were obtained in at least two independent experiments. B, ratio of potencies of SPPARMs and PPAR full agonists in the hPPARWT and hPPARY473A transcriptional activity assays. Assays were performed as described above, and EC50 values were determined by four-parameter fit analysis. The figure plots the EC50 hPPARY473A/EC50 hPPARWT ratios for the PPAR ligands.
Development of Novel hPPARWT and hPPAR-Y473A and SPA Binding Assays. Having identified a negative effect of the Y473A mutation on the functional activities of PPAR full agonists, it was necessary to confirm whether these changes were due to a reduction in the intrinsic binding affinities of these ligands or whether defective functional activation occurred despite normal binding potency. Thus, we developed ligand binding assays that could be used to compare the potencies of full agonists and SPPARMs with hPPARWT and hPPARY473A. We could not use our previously reported radioligand [3H]nTZD3 (Berger et al., 2003) for such studies, because it is a full agonist that demonstrated greatly diminished potency with hPPARY473A in transcriptional activity assays similar to those described above (data not shown). Thus, SPPARM1 was selected as a candidate radioligand with which to develop parallel binding assays, because its functional activity was unaffected by the Y473A mutation (Table 1).
Using [3H]SPPARM1 as the radioligand at concentrations from 0 to 30 nM in the presence or absence of 50 µM excess cold SPPARM1, saturation binding assays were performed with recombinant hPPARWT LBD or hPPARY473A LBD, as described under Materials and Methods. Nonlinear regression analyses of data from these studies indicated that the Y473A mutation did not have a significant effect on SPPARM1 binding, because its Kd value was determined to be 1.83 and 1.60 nM for hPPARWT (Fig. 6A) and hPPARY473A (Fig. 6B), respectively. Therefore, [3H]SPPARM1 proved to be a suitable radioligand for subsequent competition binding assays with which to compare the intrinsic potencies of PPAR full agonists and SPPARMs.
Fig. 6. Radiolabeled SPPARM1 is a potent ligand of hPPARWT and hPPARY473A. A, nonlinear regression analysis of data from hPPARWT saturating binding studies performed using GST-hPPARWT LBD, anti-GST antibody, and radioligand [3H]SPPARM1 at 0 to 30 nM ± 50 mM cold SPPARM1. B, nonlinear regression analysis of data from hPPARY473A saturating binding studies performed using GST-hPPARY473A LBD, anti-GST antibody, and radioligand [3H]SPPARM1 at 0 to 30 nM ± 50 µM cold SPPARM1. Similar results were obtained in two independent experiments performed in duplicate.
Binding Potencies of PPAR Full Agonists but Not SPPARMs Were Diminished with the hPPARY473A Mutant. In initial experiments, we compared the binding of rosiglitazone and SPPARM2 to hPPARWT and hPPAR-Y473A in SPA assays using [3H]SPPARM1. The affinity of the full agonist but not the SPPARM was greatly reduced with the mutant receptor in comparison with the wild-type receptor (Fig. 7A). Subsequently, the binding potencies of the same diverse collection of SPPARMs and PPAR full agonists studied in the aforementioned cell-based transcriptional activity assays were determined in the competition binding assays. Data from these studies revealed that PPAR full agonists consistently displayed a marked reduction in binding affinity to hPPARY473A in comparison with hPPARWT (Fig. 7B). In contrast, the binding characteristics of the SPPARMs were not notably affected by the Y473A mutation. Such results largely paralleled those obtained previously in the transcriptional activity assays described above. In fact, the shifts in potencies of the tested ligands in the two assays were highly correlated: r2 = 0.86, p < 0.0001 (Fig. 8).
Fig. 7. Binding potencies of PPAR full agonists but not SPPARMs are diminished with hPPARY473A versus hPPARWT. A, binding of SPPARM2 and rosiglitazone to PPARWT and hPPARY473A. Competition curves generated by incubation of 2.5 nM [3H]SPPARM1 with GST-hPPARWT LBD or GST-hPPARY473A LBD. The displacement of radioligand after incubation in the presence of the indicated concentration of SPPARM2 or rosiglitazone is plotted. Similar results were obtained in at least two independent experiments performed in duplicate. B, ratio of potencies of SPPARMs and PPAR full agonists in the hPPARWT and hPPARY473A transcriptional activity assays. Assays were performed as described above, and Ki values were determined by four-parameter fit analysis. The figure plots the Ki hPPARY473A/Ki hPPARWT ratios for the PPAR ligands.
Fig. 8. Diminished functional potencies of PPAR full agonists with hPPARY473A versus hPPARWT correlate with their diminished binding potencies with hPPARY473A versus hPPARWT. The EC50 hPPAR-Y473A/EC50 hPPARWT and the Ki hPPARY473A/Ki hPPARWT ratios of the PPAR ligands presented in Table 2 were plotted against each other on log10 scale x- and y-axes.
PPAR agonists, including the currently marketed TZDs pioglitazone and rosiglitazone, have proven efficacious in the treatment of T2DM patients. However, despite the significant antidiabetic activity of such compounds, they have been used in a limited manner because of several untoward effects associated with their use. These include plasma volume expansion and edema, which can lead to or exacerbate congestive heart failure, as well as weight gain resulting from increased adiposity and plasma volume. To extend PPAR-targeted diabetes therapy to a broader patient population, considerable efforts have recently focused on the identification of novel efficacious PPAR ligands with improved tolerability. Such endeavors have been motivated by the success of previous drug researchers in identifying selective modulators of steroid nuclear receptors that demonstrated cell-, tissue-, and gene-specific activities (Miller, 2002). For example, selective estrogen receptor modulators were found to retain beneficial effects in the treatment of osteoporosis while inducing diminished adverse effects. Thus, we and others have proceeded to generate and characterize SPPARMs that might serve as an improved second generation of insulin-sensitizing agents for the amelioration of T2DM and other metabolic disorders.
Here, we present examples of novel indole-containing SPPARMs that are structurally distinct from such ligands described previously by ourselves and others (Berger et al., 2003; Minoura et al., 2004; Acton et al., 2005; Dropinski et al., 2005; Schupp et al., 2005; Burgermeister et al., 2006). These SPPARMs were identified by their ability to bind the PPAR with high affinity and to serve as potent partial agonists of the receptor in cell-based transcriptional activity assays. Additional studies performed on an exemplary member of this cohort, SPPARM2, revealed that in comparison with PPAR full agonists, this ligand (like nTZDpa) possessed reduced adipogenic activity and an attenuated regulatory effect on the PPAR target genes FABP4, PEPCK, and ADFP in human adipocytes. These findings extend and concur with those obtained previously with SPPARMs in mouse 3T3-L1 cells and adipose tissue (Oberfield et al., 1999; Berger et al., 2003; Schupp et al., 2005).
X-ray studies of the PPAR LBD cocrystallized with the PPAR full agonist rosiglitazone support the proposition that the negatively charged nitrogen of the TZD head group forms a hydrogen bond with the helix 12 Tyr473 side-chain hydroxyl head group (Nolte et al., 1998). Additional hydrogen bond interactions are thought to exist between the ligand and the His323, His449, and Lys367 residues. These interactions govern the conformation of the TZD head group and the amino acids of the ligand binding domain. In subsequent X-ray crystallography studies, the carboxylic acid moiety of the tyrosine-based PPAR full agonist GI262570 and an -aryloxyphenylacetic-derived PPAR/ agonist were also shown to reside within hydrogen binding distance of the Tyr473 moiety of the PPAR LBD (Gampe et al., 2000; Shi et al., 2005). In contrast, crystal structures of the PPAR LBD with the structurally distinct SPPARMs GW0072 (Oberfield et al., 1999), PA-082 (Burgermeister et al., 2006), and SPPARM2 (data above) revealed that these ligands were bound in the entry-end of the ligand binding pocket at a distance from helix 12 that preclude direct contact with Tyr473.
As depicted in Fig. 4 for monomer A, the phenolic O of Tyr473 was more than 6 Å removed from its closest approach to SPPARM2 at the trifluoromethyl tail. Nor is Tyr473 properly positioned to hydrogen bond with tris(hydroxymethyl)-aminomethane, which instead apparently hydrogen bonded to the Gln286, His323, and His449 side chains. As a result of the repositioning of helix 12, Tyr473 was situated outside of the binding cavity in monomer B. Other than this difference, the rest of the residues proximate to the partial agonist were the same in either monomer and similarly situated to create the hydrophobic interior to the binding cavity. A unique consequence of the meta-substitution pattern is that the lactate was unable to achieve a conformation that would allow it, like its ortho-substituted analog Full1, to act as a potent full agonist presumably via hydrogen bonding to Tyr473 (Acton et al., 2005). Rather, this structural constraint forced the lactate to hydrogen bond with the main-chain NH of Ser342. This interaction was similar to that observed crystallographically for a 2-benzoylaminobenzoic acid analog complexed with hPPAR (Ostberg et al., 2004), and it was quite distinct from that seen between agonists such as rosiglitazone and Tyr473 as in Fig. 4.
On the basis of the above-mentioned observations, we first assessed the role of Tyr473 in determining the functional activities of structurally diverse PPAR full agonists and SPPARMs in cell-based transcriptional activity assays using wild-type hPPAR and a mutant receptor in which the tyrosine residue 473 in helix 12 was mutated to an alanine. We found that the clinically available PPAR full-agonist rosiglitazone underwent a large diminution in activity with hPPARY473A compared with hPPARWT. Moreover, we also provide evidence indicating that Tyr473 was also required for the activity of a variety of structurally distinct TZD and non-TZD full agonists. In striking contrast to these results, the Tyr473A mutation did not influence the functional activity of SPPARM2, nTZDpa, or the remainder of the structurally diverse SPPARMs that were tested.
To clarify whether the observed decreases in functional activity of PPAR full agonists were due to decreases in their intrinsic binding potencies or reduced ability of such ligands to induce a productive receptor conformation after binding in the LBD, we developed a novel SPA-based competition binding assay for hPPARWT and hPPARY473A using [3H]SPPARM1 as the radioligand. We found that [3H]-SPPARM1 was well suited for this purpose, because, in concordance with cell-based transactivation studies showing that its functional activity was not significantly affected by the Y473A mutation, nonlinear regression analyses of ligand binding studies revealed that the Y473A mutation had a negligible influence on the affinity of [3H]SPPARM1 for PPAR.
When the binding activities of the PPAR full agonist rosiglitazone and SPPARM2 were examined in radioligand displacement assays using [3H]SPPARM1, we found that the potency of the full agonist was decreased more than 10-fold with hPPARY473A in comparison with hPPARWT, whereas the affinity of SPPARM2 was undiminished by the Y473A mutation. In subsequent binding assays performed with our full array of PPAR full agonists and SPPARMs, all of the former ligands underwent notable decreases in potency with the mutant receptor that correlated well with the similar large reductions in potencies observed with PPARY473A in the transcriptional activity studies. Thus, Tyr473 was clearly required for both receptor binding and subsequent activation by PPAR full agonists; in contrast, the SPPARMs demonstrated similar potencies in binding assays performed with hPPARWT and hPPARY473A. Such results support the proposition that the SPPARMs retained their functional activities with hPPARY473A because their intrinsic binding activities were largely unchanged with the mutant receptor.
Taken together, our results provide robust biochemical support for the X-ray crystallography data from our studies as well as those of others indicating that PPAR full agonists but not SPPARMs interact directly with the Tyr473 residue located in the activation function-2 helix of the PPAR LBD (Oberfield et al., 1999; Burgermeister et al., 2006). Although Tyr473 is not a component of the charge clamp that contacts backbone atoms of LXXLL helices within transcriptional coactivators, it is in proximity to Glu471, one of the two major residues implicated in this PPAR-coactivator interaction (Nolte et al., 1998; Sheu et al., 2005). We have previously shown by NMR spectroscopy studies that PPAR full agonists induce a more stable conformation of the PPAR LBD, including the helix 12, than SPPARMs (Berger et al., 2003). Thus, the differing abilities of PPAR full agonists and SPPARMs to interact with Tyr473 and directly stabilize the aforementioned charge clamp seem to result, at least in part, in the differential receptor-coactivator interactions that have been observed with these two classes of ligands, thereby serving as a physical basis for their distinct gene transcription and resulting pharmacological effects. In support of this proposition, it has been demonstrated that compared with PPAR full agonists such as rosiglitazone or pioglitazone, the structurally distinct SPPARMs telmisartan, irbesartan, PA-082, and FK-614 induce diminished interaction of PPAR with transcriptional coactivators, including transcriptional intermediary factor 2, steroid receptor coactivator 1, and CREB-binding protein, that was associated with attenuated adipocyte gene regulation by the latter ligands (Schupp et al., 2005; Burgermeister et al., 2006; Fujimura et al., 2006). In addition, when two of these SPPARMs, telmisartan and FK614, were studied in rodent models of T2DM, they displayed comparable insulinsensitizing activity to PPAR full agonists with reduced mechanism-based adverse effects, including hemodilution, cardiomegaly, and body weight gain (Minoura et al., 2004; Schupp et al., 2005).
In summary, we have presented data demonstrating that a novel selective PPAR modulator, SPPARM2, is a potent PPAR ligand that displays diminished maximal activity compared with the PPAR full agonist rosiglitazone in cell-based assays of transcriptional and adipogenic activity. In additional studies using the wild-type and mutant receptors, hPPARWT and hPPARY473A, respectively, to characterize the binding and functional activities of structurally distinct PPAR full agonists and SPPARMs, we demonstrated that the LBD helix 12 residue Tyr473 plays a critical role in the interaction of the former but not the latter ligands with the receptor. Such results provide a physical basis for the novel biological activities of SPPARMs. Finally, it is worth noting that based on our results, hPPARY473A might prove useful as the basis for a high-throughput assay that would be biased toward the identification of additional, structurally unique SPPARMs. Thus, by advancing our understanding of PPAR molecular pharmacology and structural biology, we hold the potential to further enrich a population of ligands that should prove useful in further probing PPAR biology and that may provide novel PPAR-targeted therapeutic agents of superior efficacy and tolerability than are currently available for the treatment of T2DM patients.
Acknowledgements
We thank the Merck Research Laboratories medicinal chemists for providing the Merck ligands used in this study. We are grateful for the technical assistance of Neelam Sharma.
ABBREVIATIONS: PPAR, peroxisome proliferator-activated receptor; TZD, thiazolidinedione; SPPARM, selective peroxisome proliferator-activated receptor modulator; T2DM, type 2 diabetes mellitus; LBD, ligand binding domain; WT, wild type; GST, glutathione transferase; PCR, polymerase chain reaction; nTZDpa, nonthiazolidinedione-selective peroxisome proliferator-activated receptor modulator; FABP4, fatty acid-binding protein 4; PEPCK, phosphoenolpyruvate carboxykinase; ADFP, adipose differentiation-related protein; FK-614, 3-(2,4-dichlorobenzyl)-2-methyl-N-(pentylsulfonyl)-3H-benzimidazole-5-carboxamide; GI262570, (2S)-2-[[2-(benzoyl)phenyl]amino]-3-[4-[2-(5-methyl-2-phenyl-1,3-oxazol-4-yl)ethoxy]phenyl]propanoic acid; GW0072, 4-[4-[(2S,5S)-5-[2-(bis(phenylmethyl)amino)-2-oxoethyl]-2-heptyl-4-oxo-1,3-thiazolidin-3-yl]butyl]benzoic acid; PA-082, 1-(3,4-dimethoxy-benzyl)-6,7-dimethoxy-4-[4-(2-methoxy-phenyl)-piperidin-1-ylmethyl]-isoquinoline.
1 Current affiliation: Vitae Pharmaceuticals, Inc., Fort Washington, Pennsylvania.
2 Current affiliation: Pharmasset, Inc., Princeton, New Jersey.
3 Current affiliation: Lilly Research Laboratories, Indianapolis, Indiana.
【参考文献】
Acton JJ 3rd, Black RM, Jones AB, Moller DE, Colwell L, Doebber TW, Macnaul KL, Berger J, and Wood HB (2005) Benzoyl 2-methyl indoles as selective PPARgamma modulators. Bioorg Med Chem Lett 15: 357-362.
Berger J, Leibowitz MD, Doebber TW, Elbrecht A, Zhang B, Zhou G, Biswas C, Cullinan CA, Hayes NS, Li Y, et al. (1999) Novel PPAR and PPAR ligands produce distinct biological effects. J Biol Chem 274: 6718-6725.[Abstract/Free Full Text]
Berger J and Moller DE (2002) The mechanism of action of PPARs. Annu Rev Med 53: 409-435.
Berger JP, Akiyama TE, and Meinke PT (2005) PPARs: therapeutic targets for metabolic disease. Trends Pharmacol Sci 26: 244-251.
Berger JP, Petro AE, Macnaul KL, Kelly LJ, Zhang BB, Richards K, Elbrecht A, Johnson BA, Zhou G, Doebber TW, et al. (2003) Distinct properties and advantages of a novel peroxisome proliferator-activated receptor selective modulator. Mol Endocrinol 17: 662-676.[Abstract/Free Full Text]
Brünger AT, Adams PD, Clore GM, DeLano WL, Gros P, Grosse-Kunstleve RW, Jiang JS, Kuszewski J, Nilges M, Pannu NS, et al. (1998) Crystallography & NMR system: a new software suite for macromolecular structure determination. Acta Crystallogr D Biol Crystallogr 54: 905-921.
Brünger AT, Kuriyan J, and Karplus M (1987) Crystallographic R-factor refinement by molecular dynamics. Science 235: 458-460.[Abstract/Free Full Text]
Burgermeister E, Schnoebelen A, Flament A, Benz J, Stihle M, Gsell B, Rufer A, Ruf A, Kuhn B, Marki HP, et al. (2006) A novel partial agonist of peroxisome proliferator-activated receptor-gamma (PPARgamma) recruits PPARgamma-coactivator-1alpha, prevents triglyceride accumulation, and potentiates insulin signaling in vitro. Mol Endocrinol 20: 809-830.[Abstract/Free Full Text]
Davis AM and Teague SJ (1999) Hydrogen bonding, hydrophobic interactions and failure of the rigid receptor hypothesis. Angew Chem Int Ed Engl 30: 736-749.
Dropinski JF, Akiyama T, Einstein M, Habulihaz B, Doebber T, Berger JP, Meinke PT, and Shi GQ (2005) Synthesis and biological activities of novel aryl indole-2-carboxylic acid analogs as PPARgamma partial agonists. Bioorg Med Chem Lett 15: 5035-5038.
Elbrecht A, Chen Y, Adams A, Berger J, Griffin P, Klatt T, Zhang B, Menke J, Zhou G, Smith RG, et al. (1999) L-764406 is a partial agonist of human peroxisome proliferator-activated receptor . The role of Cys313 in ligand binding. J Biol Chem 274: 7913-7922.[Abstract/Free Full Text]
Fujimura T, Kimura C, Oe T, Takata Y, Sakuma H, Aramori I, and Mutoh S (2006) A selective peroxisome proliferator-activated receptor gamma modulator with distinct fat cell regulation properties. J Pharmacol Exp Ther 318: 863-871.[Abstract/Free Full Text]
Gampe RT Jr, Montana VG, Lambert MH, Miller AB, Bledsoe RK, Milburn MV, Kliewer SA, Willson TM, and Xu HE (2000) Asymmetry in the PPARgamma/RXRalpha crystal structure reveals the molecular basis of heterodimerization among nuclear receptors. Mol Cell 5: 545-555.
Hunt CR, Ro JHS, Dobson DE, Min HY, and Spiegelman B (1986) Adipocyte P2 gene developmental expression and homology of 5'-flanking sequences among fat cell-specific genes. Proc Natl Acad Sci U S A 83: 3786-3790.[Abstract/Free Full Text]
Jiang HP and Serrero G (1992) Isolation and characterization of a full-length cDNA coding for an adipose differentiation-related protein. Proc Natl Acad Sci U S A 89: 7856-7860.[Abstract/Free Full Text]
Kintscher U, Goetze S, Wakino S, Kim S, Nagpal S, Chandraratna RA, Graf K, Fleck E, Hsueh WA, and Law RE (2000) Peroxisome proliferator-activated receptor and retinoid X receptor ligands inhibit monocyte chemotactic protein-1-directed migration of monocytes. Eur J Pharmacol 401: 259-270.
Liu K, Black RM, Acton JJ 3rd, Mosley R, Debenham S, Abola R, Yang M, Tschirret-Guth R, Colwell L, Liu C, et al. (2005) Selective PPARgamma modulators with improved pharmacological profiles. Bioorg Med Chem Lett 15: 2437-2440.
McRee DE (1999) XtalView/Xfit-A versatile program for manipulating atomic coordinates and electron density. J Struct Biol 125: 156-165.
Miller CP (2002) SERMs: evolutionary chemistry, revolutionary biology. Curr Pharm Des 8: 2089-2111.
Minoura H, Takeshita S, Ita M, Hirosumi J, Mabuchi M, Kawamura I, Nakajima S, Nakayama O, Kayakiri H, Oku T, et al. (2004) Pharmacological characteristics of a novel nonthiazolidinedione insulin sensitizer, FK614. Eur J Pharmacol 494: 273-281.
Nolte RT, Wisely GB, Westin S, Cobb JE, Lambert MH, Kurokawa R, Rosenfeld MG, Willson TM, Glass CK, and Milburn MV (1998) Ligand binding and co-activator assembly of the peroxisome proliferator-activated receptor-gamma. Nature 395: 137-143.
Oberfield JL, Collins JL, Holmes CP, Goreham DM, Cooper JP, Cobb JE, Lenhard JM, Hull-Ryde EA, Mohr CP, Blanchard SG, et al. (1999) A peroxisome proliferator-activated receptor gamma ligand inhibits adipocyte differentiation. Proc Natl Acad Sci U S A 96: 6102-6106.[Abstract/Free Full Text]
Ostberg T, Svensson S, Selen G, Uppenberg J, Thor M, Sundbom M, Sydow-Backman M, Gustavsson AL, and Jendeberg L (2004) A new class of peroxisome proliferator-activated receptor agonists with a novel binding epitope shows antidiabetic effects. J Biol Chem 279: 41124-41130.[Abstract/Free Full Text]
Otwinowski Z and Minor W (1997) Processing of X-ray diffraction data collected in oscillation mode. Methods Enzymol 276: 307-326.
Sack JS (1988) CHAIN—a crystallographic modeling program. J Mol Graph 6: 224-235.
Schupp M, Clemenz M, Gineste R, Witt H, Janke J, Helleboid S, Hennuyer N, Ruiz P, Unger T, Staels B, et al. (2005) Molecular characterization of new selective peroxisome proliferator-activated receptor gamma modulators with angiotensin receptor blocking activity. Diabetes 54: 3442-3452.[Abstract/Free Full Text]
Sheu SH, Kaya T, Waxman DJ, and Vajda S (2005) Exploring the binding site structure of the PPAR gamma ligand-binding domain by computational solvent mapping. Biochemistry 44: 1193-1209.
Shi GQ, Dropinski JF, McKeever BM, Xu S, Becker JW, Berger JP, MacNaul KL, Elbrecht A, Zhou G, Doebber TW, et al. (2005) Design and synthesis of alphaaryloxyphenylacetic acid derivatives: a novel class of PPARalpha/gamma dual agonists with potent antihyperglycemic and lipid modulating activity. J Med Chem 48: 4457-4468.
Tontonoz P, Hu E, Devine J, Geale EG, and Spiegelman BM (1995) PPAR2 regulates adipose expression of the phosphoenolpyruvate carboxykinase gene. Mol Cell Biol 15: 351-357.
作者单位：Departments of Metabolic Disorders (M.E., T.E.A., G.A.C., C.F.W., D.E.M., J.P.B.) and Medicinal Chemistry (B.M., R.T.M., J.W.B., P.T.M., H.B.W.), Merck Research Laboratories, Rahway, New Jersey