Modulation of Glucocorticoid Receptor Activities by Coactivators and Corepressors

S. Stoney Simons, Jr.

Steroid Hormones Section, NIDDK/LMCB

National Institutes of Health, Bethesda, MD 20892, U.S.A.

A major response of steroid receptors to steroid hormones is the induction of gene transcription. Two other physiologically relevant, but less studied, properties are the EC₅₀ of receptor-agonist complexes and the partial agonist activity of receptor-antagonist complexes. Agonists are ligands that mimic the activity of endogenous steroid hormones. The dose-response curve of agonists gives the amount of gene induction by any concentration of steroid, with half of the maximal induction occurring at a value called the EC₅₀. The lower the EC₅₀ of a given gene, the greater is the level of induction that is achieved with the circulating concentration of steroid in a cell or animal (Fig. [1]). Antisteroids, or antagonists, block the action of agonist steroids. However, virtually all antisteroids display some partial agonist activity with selected genes. Recently, it has been appreciated that if one can selectively eliminate the trans-activation of a target gene, while reducing the expression of most of the other genes that are regulated by a given receptor (Fig. [1]), then the number of undesirable side effects that usually result from the indiscriminate repression of all responsive genes will be greatly diminished.

Fig. 1 Variations in dose-response curves and partial agonist activity among different genes induced by the same receptor steroid complex

Contrary to earlier expectations, accumulating experiments indicate that the value of neither the EC₅₀ nor the partial agonist activity is constant for a given receptor-steroid complex. Our recent experiments with glucocorticoid (GR), progesterone (PR), and mineralocorticoid (MR) receptors in transiently transfected cells indicate that changing the concentrations of a variety of factors causes the position of the dose-response curve with agonists, and the amount of partial agonist activity of antagonists, to display a continuum of values over a broad range. Two sets of factors that have been extensively examined with GRs are the p160 coactivators (SRC-1, TIF2, and AIB1) and the corepressors (SMRT and NCoR). These studies have led to the two hypotheses (1) that coactivators and corepressors each bind to both GR-agonist and -antagonist complexes and (2) that the final biological response is influenced by the intracellular ratio of coactivators to corepressors (Fig. [2]).

Fig. 2 Equilibrium model for modulation of GR-agonist and -antagonist activities with varying concentrations of coactivators and corepressors.

In an effort to obtain biochemical support for hypothesis #1 of our model, we examined the ability of coactivators and corepressors to interact with GR complexes in mammalian two-hybrid and cell-free pulldown assays. Both agonist and antagonist complexes associate with the coactivator TIF2 in a manner that depends upon the presence of an intact receptor interacting domain (RID) in TIF2. Conversely, both agonist and antagonist complexes of GR also combine with the corepressors NCoR and SMRT. The ability of GR-agonist complexes to interact with corepressors was confirmed by the observation that mutating the RIDs of NCoR eliminated the mammalian two-hybrid interactions with both GR-agonist and antagonist complexes. Furthermore, the association of GR-agonist complexes was also seen with the wild type GR instead of the VP16/GR chimera that was used in the other two-hybrid experiments. Therefore, we conclude that GR-agonist complexes do bind to corepressors and that GR-antagonist complexes bind to coactivators.

In support of hypothesis #2 of our model, we found that overexpressed TIF2 and corepressors each competitively inhibit the association of GR with corepressors or coactivators in the mammalian two-hybrid assay. Furthermore, the position of the GR dose-response curve and the partial agonist activity of GR-antagonist complexes could be modulated simply by overexpressing thyroid receptor beta ± thyroid hormone, thereby altering the ratio of endogenous coactivators to corepressors. In summary, we believe that our data support a model in which the dose-response curve and partial agonist activity of GR complexes can be modulated in an equilibrium manner by the competitive binding of coactivators and corepressors to both GR-agonist and antagonist complexes.