It is now well established that the parameters of GR-mediated gene induction (Amax, EC50, and PAA) can be modulated by changing the concentrations of involved factors. Most of these studies have been conducted by looking at the activities of saturating and subsaturating concentrations of agonists, occasionally with saturating concentrations of antisteroids, in transiently transfected tissue culture cells. Our recent studies in human peripheral mononuclear cells (PBMCs) have confirmed that changes in factor concentration affects the induction parameters of endogenous, as well as exogenous, GR-regulated genes. These results provide strong support for our hypothesis that the modulation of GR induction parameters is a relevant feature of human physiology. Yet another mechanism for altering GR induction parameters appears to be by modifying the activity of GR-bound cofactors. As we first reported in 2008, mutations in the ligand binding domain (LBD) of glucocorticoid receptors (GRs) marginally affect the binding affinity of the glucocorticoids dexamethasone (Dex) and deacylcortivazol (DAC) while dramatically altering their Amax and EC50 in a steroid-dependent manner. At the same time, the PAA changed. These results were proposed to derive, in part, from altered protein-protein interactions of GR with the coactivator TIF2 despite normal TIF2 binding affinities. We now find that increasing concentrations of Ubc9 and of GR, to probe unidentified cofactors, do not reverse the effects of GR LBD mutations with the GREtkLUC reporter in both CV-1 and U2OS cells. Thus changes in simple equilibrium binding affinities of factors to mutant receptors cannot account for the modified transcriptional properties. Similar effects are seen with Ubc9 for three endogenous genes in U2OS cells. This behavior is most dramatic with Ubc9 and the isolated GR LBD fused to the GAL4 DNA binding domain, despite the normal binding of Ubc9 to mutant GRs. In all cases, the nuclear translocation of Dex- and DAC-bound wild type and mutant receptors is the same. These results extend the earlier results with TIF2 and support the hypothesis that small changes in the GR LBD can alter the activities of bound cofactor without modifying cofactor binding. We propose that this separation of binding and transactivation parameters occurs for a wide variety of GR-associated cofactors. In collaboration with Carson Chow, we sought to provide a firm theoretical underpinning of the above general observations. At the outset, it was well known that ligand-mediated gene induction by steroid receptors appears to precisely follow a first-order Hill dose-response curve (FHDC), which has been explained by receptor-ligand binding being the rate-limiting step. However, this view predicts the same EC50 and Amax for all genes induced by a given receptor-steroid complex, which is challenged by the findings that various cofactors/reagents can alter both parameters. Furthermore, it is very difficult for a multi-step mechanism to yield a FHDC. Thus, we faced two major questions: how can a FHDC arise from a reaction with multiple steps and how do cofactors modify both potency and maximal activity? We derived a general mathematical theory of gene induction for which the most general form of the equilibrium and mass-conservation equations yield a FHDC for plots of the final product vs. initial ligand concentration. This theory requires that individual reactions leading to the final product dissociate from downstream reactions, implying that complexes are weakly bound or exist only transiently. When applied to glucocorticoid receptor-mediated gene induction, this theory predicts eight previously unidentified classes of mechanisms for the effects of cofactors/reagents on gene induction potency and maximal activity, matches the unusual experimental data for the effects of Ubc9 with different concentrations of receptor, and requires that FHDC plots cannot arise from the DNA binding of pre-formed receptor dimers, which was experimentally supported. An important finding of these studies was the demonstration that quite accurate modeling of a complex reaction pathway can be achieved without having first to characterize all of the contributing steps. Furthermore, our model indicates that new graphical methods of data analysis can be used to determine not only how the factor modulates these induction parameters but also where the factor acts, relative to a reference point called the concentration limiting step (CLS). The CLS is the steady state analog of the rate limiting step of enzyme kinetics. Such information is not currently available by any other method. We therefore believe that we have formulated a new general model for steroid hormone action that correctly accounts for the observed changes in Amax and EC50 and is capable of yielding previously unobtainable information. It should also be noted that this theoretical model, and the associated graphical analyses of the data, are applicable to any inducible system that affords a FHDC plot. As a result of the above studies, we have gained new molecular information both about the determinants of glucocorticoid steroid activity and about the modulation of the dose-response curve of agonists. These modulatory factors permit a continuum of responses and constitute new therapeutic targets for differential control of gene expression by steroid hormones during development, differentiation, homeostasis, and endocrine therapies. These combined findings contribute to our long-term goal of defining the action of steroid hormones at a molecular level and of understanding their role in human physiology.
Showing the most recent 10 out of 16 publications
