In part 1 of this three-part series (posted August 23, 2010), we discussed the recent action of the Endocrinologic and Metabolic Drugs Advisory Committee regarding rosiglitazone (GlaxoSmithKline’s Avandia). The advisory committee, by a close vote, recommended to the FDA that it leave the drug on the market, with new restrictions (e.g., closer supervision and new label warnings).
Avandia and pioglitazone (Takeda’s Actos) are the only marketed members of the thiazolidinedione (TZD) class of peroxisome proliferator-activated receptor gamma (PPARγ) agonists. PPARγ is a nuclear receptor that controls glucose metabolism and adipocyte differentiation. In treatment of type 2 diabetes, TZD modulation of PPARγ results in decreased insulin resistance, thus enabling tissues such as muscle and fat to utilize insulin more efficiently for the uptake of glucose. Agents that work by decreasing insulin resistance are known as “insulin sensitizers”.
As discussed in our August 23 article, clinical evidence indicates that both Avandia and Actos induce weight gain in type 2 diabetics (who are usually obese to begin with), and carry an increased risk of edema and heart failure. Avandia also carries a significantly increased risk of myocardial infarction (MI). Critics of Avandia who want the drug removed from the market cite the increased risk of MI, and the availability of a safer TZD, Actos.
Despite the major safety issues with TZDs, there is both animal model and human evidence that these agents may work to preserve and/or enhance beta-cell function, and thus to help prevent progression of type 2 diabetes. Moreover, insulin resistance is a major factor in the pathogenesis of the disease. We therefore asked whether it might be possible to discover and develop better, safer insulin sensitizers that would have the desirable properties of the TZDs with fewer adverse effects. In this article, which is part 2 of the series, we discuss a recent breakthrough in the biochemistry of PPARγ that may enable companies to develop better insulin sensitizers. In part 3 of this series, we shall look at how companies might develop such compounds.
It was Bruce Spiegelman (Dana-Farber Cancer Institute and Harvard Medical School, Boston MA) and his colleagues who identified PPARγ as the master regulator of adipocyte biology and differentiation back in 1994. This eventually led to the discovery and development of TZDs such as Avandia and Actos, which are synthetic compounds that are strong agonists of PPARγ. These compounds act as potent insulin sensitizers, and are thus used in the treatment of type 2 diabetes. However, the mechanism of their insulin sensitizing activity is not clear. Administration of TZDs result in decreased expression in adipose tissue of insulin-resistance inducing hormones such as tumor necrosis factor-alpha (TNF-α), interleukin-1 (IL-1) and resistin. They also induce increased expression of adiponectin, an adipose-derived hormone or adipokine (a member of a class of cytokines that are secreted by adipocytes) that acts as a natural insulin sensitizer.
In a July 2010 research article published in Nature, the Spiegelman group noted that the action of insulin-sensitizing PPARγ agonists does not make biological sense. Obese people and people with insulin resistance (including type 2 diabetics) have no deficiency in PPARγ or PPARγ activity. Therefore, why do synthetic activators of PPARγ have such dramatic insulin sensitizing and antidiabetic activity? Since PPARγ controls adipocyte differentiation, it makes good biological sense that PPARγ agonists can induce weight gain, however. But why do these inducers of weight gain also act as insulin sensitizers, while excessive weight is generally associated with increased insulin resistance? Finally, although the strong PPARγ agonists rosiglitazone and pioglitazone are insulin sensitizing and antidiabetic, some selective PPARγ modulators with poor agonist activity, such as the (non-TZD) benzoyl 2-methyl indole MRL24 (discovered by Merck Research Laboratories) have very good antidiabetic activity.
The Spiegelman group found that the enzyme cyclin-dependent kinase 5 (CDK5) phosphorylates PPARγ at serine 273 (Ser 273). There are no other CDK5 phosphorylation sites on PPARγ. Unlike other members of the CDK family, CDK5 is not a cell-cycle kinase that is regulated by a cyclin, but instead is regulated by p35/25, which are targets of numerous cytokines and other proinflammatory signals. Specifically, cytoplasmic p35, possibly in response to proinflammatory signals, is cleaved to form p25. p25 can then enter the nucleus, where it associates with and activates CDK5. The activated CDK5 can then phosphorylate PPARγ at Ser 273.
Treatment of adipocytes with proinflammatory cytokines or free fatty acids results in enhanced formation of p25, and enhanced phosphorylation of PPARγ at Ser 273. The same results occur in vivo, in mice that are fed a high-fat diet over a prolonged period of time, and thus become obese. It is well known that both free fatty acids and proinflammatory cytokines are elevated in obesity.
The researchers also found that phosphorylation of PPARγ at Ser 273 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 certain other genes, including adiponectin. The insulin-sensitizing synthetic compounds rosiglitazone (a strong PPARγ agonist) and MRL24 (a weak PPARγ agonist) both inhibited phosphorylation of PPARγ at Ser 273 in adipocytes. However, treatment of adipocytes with these two compounds gave different results in terms of their effects on PPARγ-regulated genes. Treatment of fat cells with rosiglitazone resulted in upregulation of adiponectin and other genes that are downregulated by Ser 273 phosphorylation. But rosiglitazone treatment also resulted in upregualtion of PPARγ-regulated genes involved in adipogenesis. In contrast, treatment of adipocytes with MRL24 did not upregulate the genes involved in adipogenesis. But it did upregulate the gene set (including adiponectin) that was downregulated by phosphorylation of PPARγ at Ser 273.
Mass spectrometry studies indicated that both rosiglitazone and MRL24 changed the conformation of PPARγ in such a way as to make this protein less favorable for phosphorylation by CDK5. However, rosiglitazone and MRL24 binding results in different conformational changes. The researchers hypothesized that these conformational changes may change the way in which PPARγ interacts with coregulator proteins. Nuclear receptors work together with coregulators 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.
The researchers then studied the action of roslglitazone and of MRL24 on phosphorylation of PPARγ at Ser 273 and on modulation of PPARγ-regulated genes in vivo. In mouse models, these two compounds inhibited PPARγ Ser 273 phosphorylation in adipose tissue, and caused similar changes in PPARγ-regulated gene expression as they do in adipose cells in vitro. Moreover, human diabetes patients treated with rosiglitazone usually exhibited decreased phosphorylation of PPARγ at Ser 273 in biopsied subcutaneous fat. However, in some cases, no such decreased phosphorylation was seen. Improvements in insulin sensitivity in these patients correlated with decreased phosphorylation of PPARγ at Ser 273.
On the basis of these results, the researchers concluded that the insulin sensitizing and antidiabetic effects of PPARγ agonists may not be due to the agonistic effects of these compounds on PPARγ, but on their ability to inhibit CDK5 phosphorylation of PPARγ.
In a News and Views article in the same issue of Nature, Riekelt Houtkooper and Johan Auwerx (Ecole Polytechnique Federale de Lausanne [EPFL] in Switzerland) postulate that the new findings of the Spiegelman group may be used to develop PPARγ modulating drugs that lack full agonist activity, but still inhibit CDK5 phosphorylation of PPARγ. Such compounds (of which MRL24 may be a starting point) would not upregulate adipogenic genes, but would upregulate insulin sensitizing genes such as adiponectin. These authors postulate that these novel compounds may thus have strong insulin sensitizing and antidiabetic effects, but would lack such adverse effects as weight gain, edema, and the risk of heart failure.
We shall discuss strategies for developing improved insulin sensitizers in greater depth in Part 3 of this series.