About 1 out of 6 prescription drugs produce therapeutic effects by binding to a family of transcription factors
called nuclear receptors. Such nuclear receptor drugs often provide the best treatment option for many
diseases; however, they also cause serious adverse effects. For example, agonists known as
thiazolidinediones (TZDs) activate the nuclear receptor peroxisome proliferator-activated receptor ¿ (PPAR¿)
and are arguably the best treatment for type II diabetes; however, they cause weight gain and weak bones.
Some PPAR¿ partial agonists produce fewer adverse effects but the same beneficial effects in mice
compared to TZDs. While this new class of agonists is promising, they have not reached the clinic. A lack of
understanding of how they produce different effects from TZDs impedes their development into drugs to treat
type II diabetes and other diseases.
We propose that such partial agonists are “biased agonists”. Like TZDs, a biased agonist would bind to
and activate PPAR¿; however, they would produce different effects by activating the receptor differently from
TZDs. How biased agonists could activate the receptor differently is not known.
We know that agonists produce effects by recruiting other proteins, known as coactivators, to PPAR¿.
The best-supported mechanism of biased agonism in nuclear receptors is that they induce what we term
“coactivator bias”. Coactivator bias refers to the ability of an agonist to bias interaction of PPAR¿ toward certain
coactivators or away from others relative to TZDs. It is well-documented that some agonists induce coactivator
bias; however, such bias has never been well-quantified and the mechanism underlying bias is unknown. This
lack of mechanistic understanding limits enthusiasm for and the ability to carry out further development of
biased PPAR¿ agonists.
This proposal will quantify and compare coactivator bias for a panel of agonists and measure the acute
effects of those same agonists on cells. This will help determine how coactivator bias affects PPAR¿ signaling
pathways. Comparison of bias with the published physiologic effects of these agonists may correlate bias with
physiologic effects, including the desired and undesired effects of type II diabetes drugs.
This proposal will also test structural mechanisms of coactivator bias. Our preliminary data show that
there are two distinct structural classes of coactivators and suggest a clear mechanism by which biased
agonists favor binding of one class. Completion of the aims of this proposal will define, in atomic detail,
mechanisms of biased agonism in PPAR¿. Such knowledge is critical to further development of drugs that
produce less adverse effects, but maintain the powerful and unique beneficial effects of TZDs. Because
PPAR¿ is structurally similar to other nuclear receptors, the knowledge gained in this proposal will impact
biased drug development for other nuclear receptors.