US20030135873A1

US20030135873A1 - Method for activating peroxisome proliferator activated receptor-gamma

Info

Publication number: US20030135873A1
Application number: US10/354,015
Authority: US
Inventors: Barry Forman
Original assignee: Individual
Current assignee: Individual
Priority date: 1999-06-18
Filing date: 2003-01-30
Publication date: 2003-07-17
Also published as: AU2001268460A1; CA2407788A1; WO2001098536A2; US20010032337A1; EP1294761A2; WO2001098536A3; JP2004500894A; US6548739B2

Abstract

The invention relates to a method of identifying nuclear receptor controlled genes in specific tissues. In particular, the method also provides a method of activating PPARγ nuclear receptor controlled target genes in vivo in a tissue-specific manner.

Description

CROSS REFERENCE TO RELATED APPLICATIONS

This application is a continuation of co-pending application Ser. No. 09/778,032, filed Feb. 7, 2001, which is a continuation-in-part of co-pending prior application Ser. No. 09/596,083, filed Jun. 16, 2000, which claims priority from provisional application Ser. No. 60/139,718, filed Jun. 18, 1999.[0001]

BACKGROUND OF THE INVENTION

1. Technical Field

This application relates to a method of activating PPARγ nuclear receptor controlled target genes in vivo. Activation of PPARγ in certain tissues is sufficient to prevent the development of diabetes.

2. Background of the Invention

The thiazolidinedione class of antidiabetic drugs represent one of the few treatments of diabetes that alleviate insulin resistance, hyperglycemia and hyperlipidemia in patients with NIDDM. Thiazolidinediones are ligands for peroxisome proliferator activated receptor-γ (PPARγ), a member of the nuclear receptor superfamily. This molecular linkage implies that thiazolidinediones achieve their insulin resistance effects by regulating PPARγ target genes. However, the precise pathway connecting PPARγ activation to insulin sensitization remains a mystery. In particular, the target tissue for PPARγ action is unknown. Therefore, it is unclear which PPARγ target genes contribute to the normalization of insulin response.

Molecular cloning studies have demonstrated that nuclear receptors for steroid, retinoid, vitamin D and thyroid hormones comprise a superfamily of regulatory proteins that are structurally and functionally related. Nuclear receptors contain a central DNA binding domain that binds to cis-acting elements (response elements) in the promoters of their target genes. Once bound to a response element, nuclear receptors activate transcription of specific genes through their conserved C-terminal ligand binding domains which bind hormones with high affinity and specificity. The ligand binding domain is a complex entity containing several embedded subdomains. These include a C-terminal transactivation function and a dimerization interface. Binding of the specific ligands to the nuclear receptor results in a conformation change that promotes interactions between the transactivation domain and several transcriptional co-activator complexes. These complexes destabilize chromatin and activate transcription. Through this mechanism, nuclear receptors directly regulate transcription in response to their specific ligands.

An important advance in the characterization of this superfamily of regulatory proteins has been the discovery of a growing number of gene products which possess the structural features of hormone receptors but which lack known ligands. These are known as orphan receptors, which like the classical members of the nuclear receptor superfamily, possess DNA and ligand binding domains. They are believed to be receptors for yet to be identified signaling molecules.

The peroxisome proliferator activated receptors (PPARs) represent a subfamily of structurally related nuclear receptors. Three subtypes have been described: PPARα, γ, and δ. The DNA recognition sequences for all PPAR subtypes are composed of a directly repeating core-site separated by 1 nucleotide. A common recognition sequence is XXXAGGTCAXAGGTCA (SEQ ID NO:1), however, the core-site (AGGTCA; SEQ ID NO:2) is variable and may change by one or more nucleotide. To bind DNA, PPARs must heterodimerize with the 9-cis retinoic acid receptor (RXR).

The α subtype of PPAR is expressed at high levels in liver and was originally identified as a molecule that mediates the transcriptional effects of drugs that induce peroxisome proliferation in rodents. In addition, PPARα binds to and regulates transcription of a variety of genes involved in fatty acid degradation (β- and ω-oxidation). Mice lacking functional PPARα exhibit decreased β-oxidation capacity and are incapable of increasing this capacity in response to PPARα ligands). Further, these mice inappropriately accumulate lipid in response to pharmacologic stimuli and develop late-onset obesity. Taken together, these data indicate that PPARαacts as both a sensor and an effector in a feedback loop that induces lipid catabolism in response to fatty acid signals.

In contrast to PPARα, the γ subtype of PPAR plays a critical role in the opposing process of fatty acid storage. PPARγ is expressed at high levels in adipocytes where it has been shown to be critical for adipogenesis. Indeed, forced expression of PPARγ in fibroblasts initiates a transcriptional cascade that leads to the expression of adipocyte-specific genes and ultimately to triglyceride accumulation. This implies that signals which modulate PPARγ activity may serve a primary role in regulatory energy homestasis.

PPARδ is ubiquitously expressed and binds several polyunsatured fatty acids as well as carbaprostacyclin, a synthetic analog of PGI ₂. PPARδ has been suggested to contribute to the control of embryo implantation and the inhibitory effects of non-steroidal anti-inflammatory drugs on colon cancer.

That PPARs regulate lipid homeostasis implies that putative PPAR ligands represented endogenous regulators of lipid homeostasis. One ligand for PPARγ has been identified 15-deoxy-Δ ^12,14-prostaglandin J₂(15d-J₂). The thiazolidinedione class of anti-diabetic agents mimic 15d-J₂, acting as potent ligands. Moreover, the potency of thiazolidinediones as antidiabetic agents correlates with their in vitro affinity for PPARγ. Forman et al., Cell 83:803-812 (1995); Wilson et al., J. Med. Chem. 39:665-668 (1996). These data suggest that PPARγ mediates the antidiabetic activity of these compounds.

Several other studies have shown that thiazolidinediones simultaneously promote insulin sensitization and increases in adipose cell number/mass in rodent models of NIDDM. Similarly, a genetic analysis suggested a link between obesity and a lower degree of insulin resistance in humans harboring an activating mutation in the N-terminus PPARγ. Ristow et al., N. Engl. J. Med. 339:953-959 (1998). That activation of PPARγ can induce adiopogenesis in cell culture as well as promote insulin sensitization in vivo appears paradoxical given the epidemiological studies that link weight gain and obesity to NIDDM. However, like the pharmacologic data in rodents, this genetic data suggests that PPARγ activation dissociates adipogenesis from insulin resistance.

Thiazolidinediones reverse insulin resistance in skeletal muscle, adipose tissue and hepatocytes. Komers and Vrana, Physiol. Res. 47(4):215-225 (1998). An increase insulin responsiveness is accompanied by a normalization of a wide range of metabolic abnormalities associated with NIDDM, including elevated levels of glucose, insulin, triglycerides, free fatty acids and LDL-cholesterol. Thiazolidinediones do not promote insulin secretion nor do they act as hypoglycemic agents in non-diabetic animals, implying that PPARγ regulates genes that reverse a critical step in the development of insulin resistance.

Several interesting hypotheses have been proposed to explain what causes insulin resistance and how PPARγ activation reverses this process. Insulin resistance may result from increase in circulating levels of free fatty acids. McGarry, Science 258:766-770 (1992). If this is the case, PPARγ activation would be predicted to reverse insulin resistance by promoting an increase in fatty acid storage in adipocyes. However, this does not account for the observation that free fatty acids are not elevated in all diabetic models and that a lowering of fatty acids using other treatments is not sufficient to promote insulin sensitization. Alternatively, Spiegelman and colleagues have suggested that insulin resistance results from an increased production of TNFα in the adipose tissue of diabetics. Uysal et al., Nature 389:610-614 (1997). Under this theory, PPARγ ligands act by blocking the TNFα-mediated inhibition of insulin signaling, however this is not consistent with all models of NIDDM. How PPARγ normalizes insulin resistance thus remains unclear.

PPARγ is expressed at high levels in both brown (BAT) and white adipose tissue (WAT). In vivo administration of PPARγ ligands have been shown to increase the size of both fat depots. In principle, therefore, both of these tissues could be involved in regulating insulin responsiveness. Transgenic mice with decreased levels of both BAT and WAT may be made by expressing the diphtheria toxin in these tissues using the adipocyte specific aP2 promoter. Burant et al., J. Clin. Invest. 100:2900-2908 (1997). By 8-10 months of age these mice apparently lack subcutaneous or intra-abdominal triglyceride-containing adipose tissue. The loss of triglyceride containing cells was accompanied by a progressive increase in insulin resistance and the development of diabetes. Despite the apparent loss of adipose tissue, administration of thiazolidinediones to these mice still resulted in insulin sensitization. These findings indicate that the antidiabetic action of thiazolidinedione occurs independently of thiazolidinedione-induced increases in adipocyte triglyceride content, and perhaps independent of adipose tissue. Burant et al., J. Clin. Invest. 100:2900-2908 (1997). However, this may depend on how adipocyte is defined. It is known that PPARγ is induced early in the course of adipogenesis and that PPARγ expression is required for subsequent activation of the aP2 promoter in adipocytes. This transcriptional cascade is followed by massive triglyceride accumulation. The strategy employed by Graves and colleagues for generation of “fat-free” mice depends on expression of a toxic transgene under control of the fat-specific aP2 promoter. However, since the expression of the toxic transgene in fat requires the presence of PPARγ, these mice should possess adipocyte-precursors which express PPARγ in the atrophic remnants of adipose tissue. Thus, it may be more precise to state that thiazolidinedione action is independent of mature adipose tissue. Previous studies have not been designed to address the issue of whether the antidiabetic effects of thiazolidinediones are mediated by cells of the adipocyte lineage.

PPARγ may also be expressed in skeletal muscle and in the liver but its level of expression is at least 10-fold lower in these tissues than in fat. The analysis of PPARγ expression in skeletal muscle has been complicated by the presence of fat cells that interdigitate among the myocytes. Since PPARγ is expressed at high levels in fat, it is possible that PPARγ transcripts seen on northern blots are derived from the contaminating fat cells. Immunohistochemical assays with PPARγ-specific antibodies have shown that PPARγ is expressed in myocyte nuclei at low levels. Despite the ability of thiazolidinediones to promote glucose uptake sensitization of skeletal muscle in vivo, these compounds had no detectable effect on glucose uptake after a five-hour exposure. Since the antidiabetic effects of thiazolidinediones are observed only after 1-2 weeks of treatment, a longer duration of exposure may be required to elicit the antidiabetic response, however it is difficult to maintain phenotypic myocytes in culture for this length of time. For similar reasons, it is not clear whether the liver is a direct target for the antidiabetic effects of thiazolidinediones.

“Knockout” mice lacking the PPARγ gene have an embryonic lethal phenotype. Thus, these mice are not useful in studying the effects of PPARγ in the adult animal. In principle, it might be possible to bypass the embryonic lethal phenotype by generating tissue-specific knockouts of PPARγ. In practice, this approach has been complicated by the need to express sufficient quantities of the cre-recombinase in the target tissue. Assuming these technical difficulties can be overcome, the resulting mice would be useful in an analysis of the physiological consequences resulting from the loss of PPARγ function. In any case, these mice would not be useful to study the consequences of PPARγ activation. A method of studying what the effects would be in individual tissues upon activation of PPARγ by a drug, or the like would be enormously useful.

SUMMARY OF THE INVENTION

Accordingly, this invention provides a method of identifying nuclear receptor controlled genes in a specific tissue of an animal, which comprises providing an expression vector containing a constitutively active nuclear receptor gene which is fused at the N-terminus to the transcriptional activation domain of the Herpes viral VP16 protein, and is linked to a-promoter element which drives tissue-specific expression, transferring the constitutively active nuclear receptor gene to the animal, expressing the constitutively active nuclear receptor gene in the animal, determining the level of expression of candidate target genes of the nuclear receptor in the tissue, and identifying genes which exhibit altered expression.

In another embodiment, the invention provides a method of modulating the in vivo expression of nuclear receptor controlled genes in a specific tissue of an animal, which comprises providing an expression vector containing a constitutively active nuclear receptor gene which is fused at the N-terminus to the transcriptional activation domain of the Herpes viral VP16 protein, and is linked to a promoter element which drives tissue-specific expression, transferring the constitutively active nuclear receptor gene to the animal, and expressing the constitutively active nuclear receptor gene in vivo in the animal.

In yet another embodiment, the invention provides a transgenic non-human animal harboring a constitutively active nuclear receptor gene which is fused at the N-terminus to the transcriptional activation domain of the Herpes viral VP16 protein, and is linked to a promoter element which drives tissue-specific expression. Preferably, the nuclear receptor gene is PPARγ and the promoter element drives brown adipose tissue-specific expression. A preferred promoter is the promoter from uncoupling protein 1.

In yet a further embodiment, the invention provides a method of preventing diabetes in a mammal having brown adipose tissue which comprises activity PPARγ expression specifically in brown adipose tissue.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a histogram depicting reporter-activity of PPARγ and VP-PPARγ. [0023]
FIGS. [0024] 2A-2D show which animals contain the indicated exogenous VP-PPARγ gene by RT-PCR using two primer pairs with confirmation by Southern blot.
FIG. 3 presents RT-PCR analysis (3A; primer pair A, 740 bp; 3B: primer pair B, 679 bp) indicating that the VP-PPARγ1 transgene is specifically expressed in brown fat. [0025]
FIG. 4 presents data on metabolic parameters in aged female wild type (wt) mice and transgenic (Tg) mice expressing VP-PPARγ in brown fat. [0026]
FIG. 4A: fasting glucose; [0027]
FIG. 4B: insulin levels; [0028]
FIG. 4C: leptin levels; [0029]
FIG. 4D: serum triglycerides (mg/dL); [0030]
FIG. 4E: serum cholesterol (mg/dL); [0031]
FIG. 4F: food intake (g); [0032]
FIG. 4G: body weight (g); [0033]
FIG. 4H: daytime rectal temperature (°C.). [0034]
FIG. 5 shows brown adipose tissue from wild type and transgenic mice and demonstrates decreased steatosis in brown fat expressing the UCP-VPP-PPARγ transgene.[0035]

DETAILED DESCRIPTION OF THE INVENTION

Transgenic mice have been developed that express a constitutively active form of PPARγ in either brown fat, white fat, skeletal muscle or liver. Each mouse line will be evaluated for parameters of insulin responsiveness under normal and diabetogenic conditions. This will allow determination of whether activation of PPARγ target genes in any one (or combination) of tissue(s) is sufficient to relieve insulin resistance in vivo. As insulin resistance is a defining featured of NIDDM, a better understanding of this phenomenon should aid in the development of more effective therapeutic strategies. [0036]
A constitutively active form of PPARγ was generated by fusing the transcriptional activation domain of the Herpes viral VP16 protein to the N-terminus of PPARγ (VP-PPARγ). N-terminus fusions were created as this does not alter the DNA binding or dimerization activity of receptors. The VP16 transactivation domain was chosen because it is well characterized and is known to function independent of nuclear receptor ligands. Recent studies indicate that VP16 recruits a transcription activation complex that is indistinguishable from that utilized by PPARγ and other nuclear receptors (TRAP/SMCC). The altered receptor activated transcription via PPARγ response elements in the absence of ligand, yet has similar DNA binding and transactivation specificity as the wild-type receptor. As shown in FIG. 1, the VP-PPARγ chimera activates PPARγ response elements in the absence of ligand but has no effect on non-PPARγ response elements including a closely related DR-1 type response element (CRBPII) specific for RXR homodimers. [0037]
To further confirm the biological activity of VP-PPARγ, this receptor was tested for induction of adipocyte differentiation in the absence of PPARγ ligands using the adipogenic system established by Tontonoz and Spiegelman. Tontonoz and Spiegelman, [0038] Cell 79:1147-1156 (1994). The NIH 3T3 cells in this system lack PPARγ and thus cannot differentiate into adipocytes. However, when infected with PPARγ-expressing retroviruses, these cells will undergo adipocyte differentiation in the presence of PPARγ ligands. VP-PPARγ was cloned into a replication defective retroviral expression vector (pBABE). Expression of the wild-type receptor in NIH-3T3 cells resulted in cells that underwent nearly 100% differentiation in response to PPARγ ligands. In contrast, the VP-PPARγ expressing retroviruses led to an equally efficient adipocyte conversion in the absence of PPARγ ligand (data not shown). These results are consistent with the transfection studies and confirm that VP-PPARγ acts as a constitutively active form of VP-PPARγ in a biologically relevant system.
Transgenic expression vectors that would specifically direct expression of this chimera to the BAT, WAT, skeletal muscle and liver of transgenic mice were then constructed. The transgenic expression vectors contained VP-PPARγ linked to the following promoter elements which are known to drive tissue-specific expression: creatine kinase in skeletal muscle (Moller et al. [0039] Endocrinol. 137:2397-2405 (1996)); major urinary protein in liver (Held et al. EMBO J. 8:183-191 (1989)); aP2 in total adipose tissue (Ross et al., Genes Dev. 7:1318-1324 (1993)); and uncoupling protein 1 in brown adipose tissue (Lowell et al., Nature 336:740-742 (1993)). The tissue specificity of these promoters have been well documented and they are transcribed very late in fetal development or within a few weeks after birth. Thus, this method allows expression of the chimeric receptor in a tissue and temporal-specific manner that avoids the potential problem of developmental defects resulting from transgenic expression.
The transgenes described above were injected into zygotes derived from C57BL/KsJ mice to create several lines of transgenic mice. These mice are known to be genetically susceptible to the development of NIDDM. After birth, the transgenic mice were screened for integration of the transgene using two different sets of PCR primers. Positive mice were confirmed by Southern blot analysis using a VP16 probe that specifically recognizes the transgene. For each promoter construct, several founders were identified that have incorporated an apparently intact transgene (FIG. 2). [0040]
The founder mice were screened for tissue specific expression of PPARγ-specific probes. Levels of expressed chimeric protein will be determined by western blot analysis using a monoclonal antibody (12CA5) that specifically recognizes a 9 amino acid epitope tag engineered into the original VP-PPARγ chimera. Several lines containing intact transgenes were further analyzed for tissue specific expression of the transgene by RT-PCT analysis using two independent primer pairs that span the 5′ and 3′ end of the transgene. As indicated in FIG. 3, several lines expressed the transgene in brown fat, but not in other tissues critical for glucose homeostasis, including white fat, skeletal muscles and liver. [0041]
To confirm the functionality of the transgene, levels of expression of several known PPARγ target genes (e.g., UCP, aP2) were measured. Elevated levels of expression of these target genes would be expected in transgenic animals that express a functional VP-PPARγ. Different lines of mice expressing VP-PPARγ were selected for each promoter construct. These mice have been analyzed by RT-PCR and the results indicate that the transgenes show the expected pattern of expression. For example, the mice containing the VP-PPARγ transgene expressed the transgene in brown fat but not in white fat, liver, skeletal muscle or other tissues that were examined. Colonies of the mice were expanded for the analyses described below. [0042]
A cursory phenotypic analysis of the mice suggests that the transgenes are appropriately expressed. Specifically, the UCP-VP-PPARγ mice would be expected to be expressed uniquely in brown fat and to tend to an overproduction of uncoupling [0043] protein 1. Tai et al., J. Biol. Chem. 271:29909-29914 (1996). As UCP-1 activity is associated with the burning of fat, these mice might be expected to have decreased white adipose stores.
To examine the effect of VP-PPARγ expression in brown fat on the development of diabetes and insulin resistance, the metabolic effects of transgene expression in old mice (1-1.5 years) that spontaneously develop diabetes was examined. As shown in FIG. 4, wild-type female mice had elevated fasting glucose (FIG. 4A) and insulin levels (FIG. 4B) whereas their transgenic counterparts had circulating glucose values in the normal range. These data indicate that the VP-PPARγ expression in brown fat prevents the development of insulin resistance and diabetes. Transgenic mice also displayed lower leptin levels (FIG. 4C) but showed no differences in the levels of serum triglycerides (FIG. 4D) and cholesterol (FIG. 4E). Examination of body weight (FIG. 4F) and food intake (FIG. 4G) indicated that food intake is higher in the transgenic mice than in wild-type controls. However, despite the increased food intake, these mice did not show an elevated body weight, suggesting that their metabolism is more efficient. Transgenic mice had elevated body temperatures, suggesting that the transgene promotes increased energy consumption via thermogenesis (FIG. 4H). [0044]
Histologic analysis of brown fat from wild-type and transgenic mice indicated that transgenic mice have smaller fat cells with decreased accumulation of triglyceride (FIG. 5). Brown fat was collected, sectioned and stained with Oil-Red-O, a triglyceride specific dye. The appearance of the transgenic brown fat was remarkably similar to that of brown fat from younger mice. These histological features are consistent with improved function of brown adipose tissue (increased thermogenesis) in the transgenic mice. [0045]
The data presented above indicate that activation of PPARγ in brown fat (or UCP1-expressing cells) is sufficient to prevent the development of diabetes. Therefore, tissue-specific drugs such as TZD analogs which specifically activate PPARγ in brown fat can act as specific antidiabetic agents, free of the other systematic effects of PPARγ activation. Moreover, the invention disclosed here provides a rational justification as well as critical reagents for a screen to identify this putative regulator of insulin responsiveness. Gene therapies, such as that described above, in which PPARγ is specifically activated in brown adipose tissue, may be used to prevent or treat diabetes. [0046]
The data disclosed above show that the antidiabetic effects of TZDs/PPARγ can be dissociated from the deleterious effects of enhanced white adipose mass which are known to result from TZDS. Moreover, dissociation of the insulin sensitizing effects of PPARγ from its lipid lowering effects implies that elevated triglyceride levels are not an underlying cause of NIDDM. The data presented here also indicate that PPARγ can promote insulin sensitization independent of a direct effect on resistin or leptin genes. However, PPARγ activation in brown fat can produce an indirect compensatory effect on white adipose tissue, since leptin levels were lower in UCP1-VP-PPARγ mice compared to wild type. [0047]
1 2 1 16 DNA mammalian misc_feature (1)..(1) n = any nucleotide base 1 nnnaggtcan aggtca 16 2 6 DNA mammalian 2 aggtca 6

Claims

1. A transgenic non-human animal, the genome of which comprises a gene that encodes a constitutively active nuclear receptor which is fused at the N-terminus to the transcriptional activation domain of the Herpes viral VP16 protein, wherein said gene encoding said nuclear receptor is linked to a promoter element which drives tissue-specific expression.

2. A transgenic animal of claim 1 wherein said nuclear receptor gene is PPARγ.

3. A transgenic animal of claim 1 wherein said promoter element drives expression specifically in uncoupling protein 1-expressing cells.

4. A transgenic animal of claim 1 wherein said promoter element is the promoter for uncoupling protein 1.