Identifying a Mechanism for Crosstalk Between the Estrogen and Glucocorticoid Receptors
Nuclear receptors must reorganize local nucleosome structures to recognize and bind to their response elements in chromatin. For a given receptor, a small minority of potential binding elements are available for unimpeded binding (a, c, d, f). Most binding sites, however, are blocked for direct receptor interaction (b, e). At a subset of elements, the action of one receptor can dramatically modulate access of an alternate factor; (g) ER acts to render a site previously closed to GR (b), so that GR can now bind. Similarly, at site (h) GR modulates chromatin to allow binding at a site normally closed to ER (e). Access for a given factor can also be reduced by modification of the local chromatin domain, (I, j).
Estrogen has long been known to play important roles in the development and progression of breast cancer. Its receptor (ER), a member of the steroid receptor family, binds to estrogen response elements (EREs) in DNA and regulates gene transcription. More recently, another steroid receptor family member, the glucocorticoid receptor (GR), has been implicated in breast cancer progression, and ER/GR status is an important predictor of breast cancer outcome.
Treating breast cancer cells with corticosteroids and estrogen cause distinct effects from treatment with either type of hormone alone, suggesting that there may be a complex interaction between the receptors in physiological systems where multiple hormones are present. However, most studies of steroid receptor-regulated transcription have been performed using a single hormone. To understand mechanistically how ER and GR might modulate each other’s functions, Gordon Hager, Ph.D., and colleagues in CCR’s Laboratory of Receptor Biology and Gene Expression, investigated the receptors’ activities in the presence of both dexamethasone (Dex) and estradiol (E2).
Recent studies have shown that access to genomic binding elements in eukaryotic cells is strongly restricted by chromatin structures, and that access is controlled in a highly cell specific manner. Hager and colleagues first examined receptor binding across the genome following 30-minute treatments of Dex only, E2 only, or Dex+E2. For ER and GR, they observed a subset of binding sites that overlapped between treatment with E2 and Dex+E2 or Dex and Dex+E2, respectively. At the same time, some ER bound sites were specific to Dex+E2 while other sites were specific for E2. The same was true for GR in the presence of Dex+E2 and Dex. These results indicated that a global redistribution of receptor binding occurs when both receptors are activated.
To examine the overall changes in ER and GR binding, the researchers combined the data on each receptor and identified nine major binding clusters. Most of the binding sites within these clusters were located between genes or within introns, suggesting they are valid regulatory regions. Two clusters, 5 and 7, represented GR and ER binding, respectively, when activated by their corresponding hormone. In contrast, cluster 2 for ER and cluster 6 for GR represented binding modules that depended on activation of the other receptor, suggesting these sites are regulated by a mechanism known as assisted loading. Changes in receptor binding following Dex+E2 co-treatment within these assisted-loading regions correlated with changes in gene expression. This crosstalk between the receptors agrees with previous studies in human cancer cells.
Because DNA regions need to be accessible to allow receptor binding, the researchers hypothesized that co-treatment might lead to changes in chromatin structure and examined changes in DNA accessibility by identifying regions sensitive to DNaseI digestion. They found that ER binding promoted by GR in cluster 2 occurred at sites of increased sensitivity to DNaseI after co-treatment. Sites in cluster 7 specific for ER binding in the presence of E2 alone, however, showed no change in accessibility. Similar results were found for GR binding sites with co-treatment in cluster 6 and with Dex only treatment in cluster 5, suggesting that activation of ER or GR can cause changes in chromatin structure at specific sites to facilitate the binding of the other receptor.
The investigators then examined the DNA sequences in these assisted-loading regions. Surprisingly, in cluster 2 they found no enrichment for EREs and, instead, identified a high number of glucocorticoid response elements and binding sites for the transcription factor AP-1, a heterodimer of the Fos and Jun proteins. These results suggested that direct ER binding to DNA was not required at assisted-loading sites. To test this idea, the researchers generated an ER DNA binding domain mutant that no longer bound EREs but could still interact with AP-1. When expressed in cells lacking wild type ER, the ER mutant bound to cluster 2 sites with Dex+E2 co-treatment but showed reduced association with E2-specific sites, demonstrating that ERE binding is not necessary for ER assisted loading.
Using data from a previous study, the scientists found that, with Dex treatment, AP-1 binding increased in cluster 2 but not in other clusters where ER binding does not depend on GR. To test whether AP-1 is a factor that can recruit ER to assisted-loading sites, the researchers made a Fos mutant that dimerizes with Jun but does not bind DNA. With co-treatment, expression of the Fos mutant reduced ER binding to cluster 2 but had no effect on GR binding.
Together these data support a model where co-treatment activates GR, which binds specific DNA regions and promotes chromatin remodeling. Other factors like AP-1 then recruit activated ER, leading to changes in gene expression. Likewise, ER can alter the chromatin landscape to facilitate GR binding. These findings implicate epigenetic programming of the genome as an important mechanism in ER and GR crosstalk. The results are of particular importance because of the widespread use of corticosteroids to suppress inflammation in the treatment of many kinds of cancer.Summary Posted: 07/2013
Miranda TB, Voss TC, Sung MH, Baek S, John S, Hawkins M, Grøntved L, Schiltz RL, Hager GL. Reprogramming the Chromatin Landscape: ER and GR Interplay at the Genome Level. Cancer Research. June 26, 2013 PubMed Link