Estrogen plays multiple roles in normal physiology and a pathologic one in breast cancer. The completion of the sequencing of the human genome has allowed the near-complete identification of the expressed regions of protein-coding genes, however, little is known concerning the organization of their cis-regulatory elements. In recently published studies we have completed chromosome-scale mapping of estrogen receptor (ER) association with the entire non-repetitive sequence of human chromosomes 21 and 22 by combining chromatin immunoprecipitation (ChIP) with high-density tiled microarrays (ChIP-chip). In these studies we find that ER binds selectively to a limited number of sites on these two chromosomes. Interestingly, the majority of these are at significant distances (>10kb) from the transcription start sites of estrogen-regulated genes. The unbiased sequence interrogation of these genuine ER binding sites suggests that ER binding often involves both estrogen response elements (ERE) and the presence of forkhead factor binding in close proximity. Furthermore, knockdown of the expression of the forkhead factor FoxA1 decreases both the association of ER with these sites and estrogen-induced gene expression. These studies, though limited to only two chromosomes and identifying only 57 binding sites, confirm the feasibility of defining the complete set of the predicted ~4000 ER binding sites across the entire human genome. We hypothesize that once all of the ER binding sites are identified that there will be distinct classes of ER targets based on the transcription factor motifs present within or adjacent to the ER binding regions. In addition, the extension of these studies to include the entire genome, should allow the study of genes both induced and repressed by estrogen. These hypotheses will be addressed in Aim 1. Our findings suggest the hypothesis that FoxA1 is necessary for defining a significant subset of ER targets in breast cancer cells. In addition we hypothesize that the different classes of ER targets defined in Aim 1 will represent different functional classes of target genes.
Aim 2 will test these hypotheses. Finally, the finding that most ER binding sites are distant from the mRNA start sites of the nearest genes raises the more general problem of how to assign putative cis-regulatory elements to specific target genes.
Aim 3 will address this problem.
Showing the most recent 10 out of 23 publications
