PPAR complex coactivator interaction
This is part 3 of a three-part series on insulin sensitizers for treatment of type 2 diabetes.
In part 1 of the series (posted August 23, 2010), we focused on safety issues with the two marketed thiazolidinedione (TZD) peroxisome proliferator-activated receptor gamma (PPARγ) agonists–rosiglitazone and pioglitazone (Takeda’s Actos). Both of these insulin sensitizing, antidiabetic agents induce weight gain, and carry an increased risk of edema and heart failure. In addition, rosiglitazone carries an increased risk of myocardial infarction. Because of the latter risk, some critics would like to see it removed from the market. However, on July 15, 2010, by a close vote the FDA’s Endocrinologic and Metabolic Drugs Advisory Committee voted to leave rosiglitazone on the market, with some new restrictions.
In part 2 of the series (posted on August 29, 2010, we discussed a breakthrough discovery by Bruce Spiegelman (Dana-Farber Cancer Institute and Harvard Medical School, Boston MA) and his colleagues, published in the 22 July issue of Nature. It was the Spiegelman group that originally identified PPARγ as the master regulator of adipocyte biology and differentiation, which eventual led to the development of the TZD drugs.
In the new research, the Spiegelman group found evidence that the insulin sensitizing and antidiabetic effects of PPARγ agonists may not be due to the agonistic effects of these compounds on PPARγ, but to their ability to inhibit phosphorylation (at Ser 273) of PPARγ by the enzyme cyclin-dependent kinase 5 (CDK5). A weak PPARγ agonist, the benzoyl 2-methyl indole (non-TZD) MRL24, inhibits CDK5 phosphorylation of PPARγ as well as rosiglitazone, and also has very good antidiabetic activity. MRL24 had been discovered by Merck in 2005.
CDK5 phosphorylation of PPARγ does not change the ability of PPARγ to upregulate transcription of genes involved in adipocyte differentiation. However, it inhibits the ability of PPARγ to upregulate transcription of a set of genes, including the gene for the adipokine adiponectin, that induce insulin sensitivity and resistance to obesity. Although both rosiglitazone and MRL24 inhibit CDK5 phosphorylation of PPARγ, treatment with the strong agonist rosiglitazone results in upregulation of both the adipogenic and the pro-insulin resistance sets of genes, while treatment with MRL24 results only in upregulation of the pro-insulin resistance set.
The Spiegelman group’s research indicates that the difference between the action of rosiglitazone and MRL24 is due to the different conformational changes induced in the PPARγ protein molecule by binding of the two compounds. The researchers hypothesized that these two different conformational changes may change the way in which PPARγ interacts with coregulator proteins. Nuclear receptors work together with coregulators (i.e., coactivators and corepressors) to regulate specific sets of genes. Different ligands (natural or synthetic) that modulate nuclear receptor interactions with its coregulators can give different results in terms of which genes are unregulated or downregulated.
Researchers hypothesize that it is the upregulation of the adipogenic gene set that is responsible for the adverse effects of strong agonists of PPARγ–weight gain, edema, and the risk of heart failure. In contrast, the upregulation of adiponectin and the other members of the pro-insulin resistance gene set is thought to be responsible for the desirable, antidiabetic effect of PPARγ agonists. If researchers could develop synthetic PPARγ ligands that would induce PPARγ upregulation of the pro-insulin resistance gene set but not the adiopgenic gene set, these compounds might constitute improved, second-generation insulin sensitizers that would have the desirable effects of the TZDs with fewer adverse effects.
In the 22 July 2010 issue of Nature, there are two essays that discuss using the new breakthrough results of the Spiegelman group to discover and develop such improved insulin sensitizers. These are a News and Views article by metabolic disease researchers Riekelt Houtkooper and Johan Auwerx (Ecole Polytechnique Federale de Lausanne [EPFL] in Switzerland), and a Nature News article by Boston-based journalist Heidi Ledford, Ph.D. of Nature.
In the Houtkooper and Auwerx article, the authors advocate changing screening strategies for PPARγ-modulating drugs, to look for those that inhibit CDK5 phosporylation of PPARγ rather than those that are strong PPARγ agonists. Potential PPARγ-modulating drugs would also induce conformational changes in the PPARγ protein that support its recruitment of coregulators that have favorable effects on metabolism (e.g., induce insulin sensitivity and/or protect against obesity). Specifically, these authors suggest that researchers determine whether CDK5-mediated PPARγ phosphorylation enhances recruitment of coactivators such as TIF2/SRC-2 and RIP140 (which unfavorably affect metabolism), and inhibits recruitment of coactivators such as SRC-1 and PGC-1, which have a favorable effect on metabolism. If this is true, drugs that inhibit CDK5 phosporylation of PPARγ should shift the balance toward recruitment of cofactors that favorably affect metabolism (i.e., promote insulin sensitivity and resistance to obesity).
Houtkooper and Auwerx also advocate biochemical and genetic research on the multiple roles of CDK5 and its upstream regulators in metabolic pathways and disorders, as well as research to identify a PPARγ phosphatase that reverses the effects of CDK5 on PPARγ. These authors do not advocate screening for CDK5 inhibitors. (Several already exist, such as roscovitine, which also inhibits several CDKs that are involved in the cell cycle, and is being developed as the anticancer drug Seliciclib by Cyclacel). CDK5 inhibitors would interfere with CDK5′s other functions, including in the central nervous system. Instead, drug discovery researchers should focus on discovering compounds that specifically change the conformation of PPARγ so that CDK5 phosphorylation of that molecule is inhibited.
In Heidi Ledford’s News article, she also advocates that pharmaceutical companies screen for inhibition of PPARγ phosphorylation, rather than strong activation of PPARγ. She also reviews the rocky history of the “glitazone” class of drugs (i.e., TZDs and related compounds), many of which have exhibited various safety problems that have kept them from the market or removed them from it. (You can see our take on this history in two 2006 articles on our website [here and here). Some commentators therefore believe that companies are unlikely to perform any new screens to identify PPARγ modulators, given the recent conflicts over Avandia and the checkered history of the glitazones. Other commentators counter that because of the large and growing type 2 diabetes market ($20.2 billion in 2008 and a projected $37.9 billion in 2018), companies will be tempted to try some new drug discovery efforts based on the new Spiegelman research. However, others counter that the new Spiegelman results are only preliminary. In any case, if drug developers would start the new screens now, it would take around 10-15 years for any new drugs to appear on the market.
But some companies may not need to start from scratch, by screening for novel compounds that inhibit CDK5 phosphorylation of PPARγ. These companies already have potential second generation insulin sensitizers in clinical development. In our published 2006 article "Safety Issues Hamper Dual PPAR Agonists Is Partial Antagonism the Solution?", we discussed safety issues both with PPARγ agonists, and with two PPARγ/PPARα dual agonists that were in late-stage development but were discontinued in 2006 because of adverse cardiovascular events. (PPARα is another PPAR nuclear receptor that is involved in control of lipid metabolism, especially of serum triglycerides and high density lipoprotein [HDL].) We proposed that development of what we called “partial agonists” of PPARs might be a solution to these safety problems. “Partial agonists” of PPARγ and of other PPARs are also called “selective PPAR modulators”. In the case of PPARγ, selective modulators are compounds that are less active in activating pathways that result in adipogenesis than strong agonists such as rosiglitzone and pioglitazone, while still upregulating pathways involved in insulin resistance.
In our 2006 article, we especially focused on a compound then called metaglidasen, which was being developed by Metabolex (Hayward, CA) and Johnson & Johnson (JNJ). This non-TZD drug is now called MBX-102/JNJ39659100. Metabolex and JNJ had tested MBX-102 for treatment of type 2 diabetes in eight Phase I and four Phase II clinical studies. However, according to the Metabolex website, the companies have repurposed the drug to treat gout, after discovering that MBX-102 is an effective uricosuric agent (i.e., it increases the excretion of uric acid in the urine).
MBX-102 is a single optical isomer of halofenate, a compound that had been studied in the 1970′s by Merck as a lipid-lowering agent. Halofenate was serendipitously found to be an insulin sensitizer; Metabolex produced a form of the drug that contained only the active optical isomer, and developed it as MBX-102.
in a 2009 paper, researchers showed that MBX-102 had insulin sensitizing and serum glucose-lowering effects in diabetic rat models (but with much lower potency than rosiglitazone), without the weight gain seen with TZDs. In vitro, MBX-102 did not drive human and mouse adipocyte differentiation, unlike TZDs. Moreover, MBX-102 had a greatly reduced ability to recruit PPARγ coactivators. MBX-102′s ability to recruit coactivators that favorably affect metabolism (SRC-1 and PGC-1) was significantly greater than its ability to recruit coactivators that unfavorably affect metabolism (such as TIF2/SRC2). MBX-102 also potently mediated transrepression of proinflammatory genes in vitro and in vivo. Earlier preclinical studies, discussed in our 2006 article, indicated that MBX-102 produced less cardiac hypertrophy in animals than rosglitazone, and also preserved the function of pancreatic beta cells.
Metabolex had also been developing a second selective PPARγ modulator, MBX-2044, a more potent follow-on compound to MBX-102, which reached Phase II of development. However, due to the company’s limited resources, it cannot conduct Phase III clinical trials without partners, especially the large Phase III trials required for diabetes. Therefore, Metabolex has shelved MBX-2044 for now, and has repurposed MBX-102 for gout, in partnership with JNJ. In a June 23 news release, Metabolex characterized itself as “a research-based company…not a commercial company” However, Metabolex is likely to seek to retain co-promotion rights in any commercialization agreement with pharmaceutical companies.
Meanwhile, another biotech company, InteKrin Therapeutics (Los Altos, CA) is developing another selective non-TZD PPARγ modulator, INT131. InteKrin had licensed the drug (formerly known as AMG131 and originally discovered by Tularik) from Amgen in January 2007. Unlike MBX-102, INT131 was purposefully designed to be a selective PPARγ modulator. INT131 has been in Phase II clinical trials for treatment of type 2 diabetes. in a 2009 paper, Amgen researchers showed that INT131 exhibited a different pattern of PPARγ coregulator recruitment from TZDs. in adipocytes, INT131 only minimally activated genes involved in adipogenesis, and exhibited a greater degrees of activation of genes involved in mediating insulin sensitivity. In a diabetic rat model, INT131 had a similar effect on glucose metabolism to rosiglitazone, with similar potency. But unlike rosigliatazone, INT131 did not induce weight gain, cardiac hypertrophy, or edema, in both rodents and nonhuman primates. In Phase IIa clinical trials (as published by InteKrin in 2010, INT131 showed potent glucose lowering (even at the low, 1 mg dose), without weight gain and fluid retention. High-dose (10 mg) INT131 provided apparently greater benefit to glucose metabolism than maximal-dose TZDs, without weight gain and fluid retention. InteKrin has tested INT131 in Phase IIb clinical trials, in which the drug gave comparable glycemic control to pioglitazone, but without edema and with minimal weight gain. The company has completed an End-of-Phase II meeting with the FDA, and is moving INT131 into Phase III clinical trials.
In order to “close the loop” in comparing studies of such compounds as MBX-102, MBX-2044, and INT131 with the new Spiegelman studies of the mechanism of insulin sensitization by PPARγ modulators, researchers would need to test these compounds for inhibition of phosphorylation of PPARγ by CDK5. Such studies have not yet been done.
While we were completing preparation of this three-part series of articles for our blog, a two-page “News and Analysis” article on selective PPAR (including PPARγ and the two other human PPARs) modulators was published in this month’s (September 2010) issue of Nature Reviews Drug Discovery (NRD). As with our blog posts, this article discussed (in a briefer format) the July 2005 FDA Advisory Committee recommendations on Avandia, the recently published Spiegelman article on the effects of insulin sensitizers on PPARγ phosphorylation by CDK5, and the prospects for selective PPARγ modulators as therapeutics for type 2 diabetes.
The NRD article not only discusses selective PPARγ modulators developed by InteKrin and Metabolex (with the latter company’s PPARγ modulators only being listed in a table), but also selective agonists of other PPARs, including single compounds that address multiple PPARs. Among these compounds is Metabolex’ selective PPARδ agonist MBX-8025 (formerly RWJ-800025, in-licensed from Janssen Pharmaceutica). PPARδ regulates genes involved in multiple aspects of the metabolic syndrome, including lipid storage and transport, fatty acid oxidation, and insulin sensitivity. MBX-8025 is being developed for treatment of mixed dyslipidemia and metabolic syndrome. In late 2008, MBX-8025 completed a Phase II proof-of-concept clinical trial, in which the compound was found to substantially reduce serum triglycerides, increase HDL, lower low-density lipoprotein (LDL), and improve insulin sensitivity in obese patients with metabolic syndrome. The effects of MBX-8025 are complementary to the LDL-lowering effects of statins such as atorvastatin (Pfizer’s Lipitor).
Another compound discussed in the NRD article is Roche’s rationally designed balanced dual PPARα/PPARγ activator aleglitazar, which is in Phase III development for treatment of patients with type 2 diabetes who have experienced a recent acute coronary syndrome. This compound is designed to combine the improvements in insulin sensitivity associated with activation of PPARγ with the amelioration of dyslipidemia associated with activation of PPARα. Roche believes that aleglitazar will avoid the adverse cardiovascular effects seen with earlier dual PPARα/PPARγ activators, or “glitazars”, as discussed in our 2006 article. A 2009 clinical trial showed that aleglitazar had a positive effect on lipid and glucose metabolism, with no induction of edema or congestive heart failure, and less weight gain than for pioglitazone, over a 16-week treatment period. These short-term results are encouraging, but must be confirmed by the results of ongoing Phase III trials.
Meanwhile, other companies, including Merck, the discoverer of MRL24, have been continuing to develop selective PPARγ modulators for treatment of type 2 diabetes.
In conclusion, the development of selective PPARγ agonists for type 2 diabetes, as we postulated in our 2006 article, is a promising approach to overcoming the issues with current PPARγ agonists, especially rosiglitazone/Avandia. However, it will be important to overcome the cloud that hangs over PPAR modulators, as the result of safety issues with such drug classes as TZDs and glitazars. Encouraging data from ongoing trials of second-generation, selective modulators will hopefully overcome these doubts, enabling companies to develop them and regulators to review them without prejudgment.
It will also be important to “close the loop” with the recent Spiegelman studies, by looking at the effects of these agents on PPARγ phosphorylation by CDK5. Moreover, screening of compounds for their effects on PPARγ phosphorylation may lead to the development of even better agents, especially if agents now in the clinic fail or give less than optimal results in late-stage trials. However, we hope that agents now in development (especially InteKrin’s INT131 and/or the two Metabolex agents if that company can find the resources from partners to complete development for diabetes) will prove to be safe and effective in clinical studies. In the longer term, it will be important to confirm that any new insulin sensitizers work to preserve beta-cell function (which would prevent progression of type 2 diabetes), and to determine if they lower the incidence of cardiovascular complications of diabetes. These are major unmet needs in treatment of type 2 diabetes.