Day 1 | Day 2 | Day 3

Day 2 of the 2009 Jensen Symposium on Nuclear Receptors featured talks from Charles Sawyers, Henry Bryant (in lieu of William Chin), Jan-Ake Gustafsson, John Cidlowski, Robert Roeder, Gordon Hager, Lee Kraus and Keith Yamamoto.

Charles Sawyers (Memorial Sloan Kettering) presented a review of work on a pair of compounds which represent a novel class of antiandrogen compounds. The compounds, RD162 and MDV3100, emerged from a structure-activity relationship (SAR)-based modification of a non-steroidal Roussel-Uclef compound which bound tightly to AR relative to Casodex (bicalutamide). They selectively bind androgen receptor (AR) with 5-fold greater affinity than bicalutamide, do not display AR agonism, and are more potent antagonists than bicalutamide. To address the mechanism of blocking, Sawyers showed that, unlike bicalutamide, RD162 and MDV3100 failed to recruit AR to the PSA enhancer, a phenomenon he attributed at least in part to less efficient nuclear translocation of the receptor in the presence of these compounds. Moving on to preclinical in vivo studies, he demonstrated that RD162 inhibited AR transcriptional function in a mouse containing a probasin promoter-luciferase reporter, and that the compound was superior to Casodex in a castrate-resistant LNCaP xenograft model. Sawyers next reported on a Phase I dose escalation multicenter trial of MDV3100 in castrate-resistant prostate cancer, highlighting a remarkable rate of reduction in serum PSA levels in these patients.

While MDV3100 has since moved onto a Phase III trial, Sawyers pointed out two risks at this point, namely the disconnect between PSA response and survival, and the well-known heterogeneity of prostate cancer. Sawyers concluded with some work focused on defining a biomarker of MDV3100 sensitivity, honing in on AR amplification. He noted that very few prostate cancer models are available with AR amplification – LNCaP cells, for example, while the workhorse of the field, are not AR amplified, and VCAP cells, while they are AR amplified, are difficult to grow. Using siRNA, Sawyers showed that cells with natural AR amplification are more sensitive to AR knockdown than single copy prostate cancer cell lines. Extending this observation to animals, he showed that even modest amounts of AR shRNA are sufficient to reduce proliferation of cancer cell lines in vivo.

Estrogen receptor-β (ERβ) was the focus of the next presentation by Henry Bryant (Eli Lilly Corp) who summarized recent work at the company on benzopyran-based ERβ-selective agonists. Starting on a mechanistic level, he showed that these ERβ-selective agonists exhibit greater potency than 17β-estradiol (E2) in recruitment of coactivators by the receptor. Interestingly, two enantiomers of one of the ERβ selective agonists, LYE2 and LYE1, have different nuclear magnetic resonance (NMR) patterns on ERβ. He then reviewed X-ray crystallography work which showed differential conformations of a benzopyran ERβ-selective agonist depending on whether it was bound to ERα and ERβ and which gave clues as to how the addition of groups to the cyclopentane ring could induce steric hindrance and decrease affinity for ERα.

The hydrogen-deuterium exchange (HDX) assay, a measure of receptor dynamics initially mentioned in Wednesday’s final talk from Geoff Greene, made a second appearance in Bryant’s next set of data slides. He showed that HDX heat map profiles of ERβ bound to different ligands could be correlated to the in vitro tissue selectivity for the LY benzopyran compounds, as measured by their behavior in Ishikawa uterine cell-based assays. Bryant moved on to discuss data on the potential of ERβ selective agonists in senescent diseases such as benign prostatic hyperplasia (BPH). He showed that a ERβ selective agonist reduced prostate weight in rodents, and that it was effective in inhibiting growth in an LNCaP tumor xenograft model. Bryant concluded with a set of slides showing that a ERβ selective agonist reduces feeding and weight in diet-induced obesity rats, and that anorectic biomarkers such as CGRP were induced in an ERβ-specific manner in peripheral fat.

Continuing in the ERβ vein, Jan Ake-Gustafsson’s (University of Houston) talk opened with a review of its ability to inhibit cellular proliferation in models of breast cancer, in T47D xenografts, and colon cancer, in SW480 colon cancer cells. Moving towards a cell biology basis for its pro-differentiative effects, Gustafsson proceeded to show that ERβ causes upregulation of cellular adhesion proteins such as ITGA1, and leads to enhanced focal contacts, improved adherence and decreased migration in ERβ-expressing cells. He then presented data which implicated ERβ as a modulator of TGFβ signaling in the hypothalamic pituitary gonadal axis, highlighting the appearance of ERα-overexpressing ovarian tumors in ERβ null females over 1.5 years. Citing the regulation of many components of the TGF-β/SMAD signaling pathway by TGF-β, Gustafsson suggested that unbalanced TGF-β/SMAD signaling and an absence of inhibin signaling might contribute to the development of these tumors in the ER-β null mouse. The second part of his talk revolved around two diseases, BPH and lymphangioleiomyomatosis (LAM), in which ERβ might afford a potentially novel therapeutic leverage point. Gustafsson presented data that suggested that TGF-β signaling was involved in epithelial to mesenchymal transition in BPH. Interestingly, BPH samples contain a splice variant of ERβ, ERβCX, which he suggested might heterodimerize with, and retard signaling through, ERβ. In the case of LAM, he again implicated epithelial to mesenchymal transition involving overactivation of TGFβ signaling and a reduction of ERβ signaling.

While acknowledging the contribution of coactivators in determining differential patterns of gene regulation by nuclear receptors, John Cidlowski (NIEHS) presented the “one gene, many proteins” case for an important role for alternative translational initiation of the glucocorticoid receptor (GR) in giving rise to functional and tissue-specific diversity for this receptor and its ligand. He showed mechanistic data demonstrating that N-terminal start codons are indispensable for the expression of GR N-terminal isoforms, and followed this up with Western blots of native rodent tissues that conveyed the striking tissue-specific distribution of these different GR isoforms. He then used immunohistochemistry to ask where are these N-terminal isoforms of GRα in the cell? The native receptor, and three isoforms, A, B and C, are diffuse in the absence of ligand and nuclear-translocated in the presence of ligand, whereas the shortest isoform, D, is nuclear in both the presence and absence of ligand. Microarray analysis demonstrated that these GR isoforms regulate distinct sets of genes, and this was reflected in the distinct biology of the isoforms in the context of glucocorticoid induction of apoptotic cell death in U2OS cells. For example, cells expressing the C form of GR were more sensitive to dexamethasone-induced cell death, whereas cells expressing the D form were resistant to cell death. To divine the mechanistic basis for this, Cidlowski showed high selective regulation of the pro-apoptotic protein granzyme A by the C isoform, although he cautioned that the reason for this was not clear. He then showed data that suggested that the basis for resistance to apoptosis of cells expressing the GR-D isoform could be a low affinity of this isoform for NFκB, suggesting that glucocorticoid inhibition of anti-apoptotic signaling may not be as effective for the D isoform. Cidlowski concluded his talk with some data which suggested sexual dimorphism in glucocorticoid signaling in the liver. Beginning with a review of pilot microarray experiments identifying differential glucocorticoid regulation of gene expression between males and females, he extended this to an investigation of the survival effects of glucocorticoids between males and females in an adrenalectomized rodent sepsis model. Strikingly, there was a clear difference between the sexes in the ability of glucocorticoids to reverse sepsis in LPS challenged rats – all females died, whereas many of the males survived.

Robert Roeder (Rockefeller) presented two recent stories from his laboratory, the first concerning the role of MED1/TRAP220 in adipogenesis and the second exploring the functional interplay of MED1/TRAP220, PGC-1α and nuclear receptors on the promoter of the gene encoding uncoupling protein-1 (UCP-1). He first reviewed data from a model system designed to investigate the role of MED1 in adipogenesis of MEFS expressing the PPARγ2 isoform, showing that while MED1 was required for adipogenesis in vitro, the two centrally-located LXXLL motifs of MED1/TRAP220 were not, suggesting that there may be redundancy in coregulator usage in this process, or that MED1 was recruited through intermediary factors. In order to investigate the physiological role of the LXXLL motifs in vivo, his laboratory created a knock in mouse in which the tandem LXXLL motifs were mutated to LXXAA, analysis of which is preliminary and ongoing. Moving on to the second theme of his talk, he reviewed data showing that PGC-1α enhanced TRAP/Mediator dependent transcription by PPARγ from a DNA template. He emphasized a model invoking distinct functions of a dynamic PGC-1α in chromatin remodeling and transcriptional events. Focusing on the UCP-1 promoter, he showed that UCP-1 induction in differentiated adipocytes depended on the C-terminus of MED-1, and that this involved the recruitment of other coregulators. Roeder’s lab had previously demonstrated synergy between MED1 and PGC-1, and he showed that while an LXXLL-containing MED1 fragment inhibited PGC-1 recruitment to the TRα-RXRα enhancer complex, this was enhanced by full-length MED1. He then demonstrated that the LXXLL domains of PGC-1 were not required for MED-1 dependent recruitment to the TRα-RXRα enhancer complex. Roeder stressed the need to demonstrate that coregulators function in their native states, namely large protein complexes, and he showed data demonstrating that in addition to isolated MED1, intact Mediator complex was also able to enhance PGC-1α recruitment in this system.

Gordon Hager (NIH) sketched a dynamic model of GR function in the context of chromatin. He first noted that while the chromatin immunoprecipitation (ChIP)-based view of regulatory site occupancy by nuclear receptors conveyed a static relationship receptor between the receptor and its binding site, photobleaching experiments painted a rather different picture, emphasizing that nuclear receptors are in fact highly mobile on their cognate DNA elements. In the second part of his talk, he linked this receptor mobility to chromatin remodeling, postulating that nucleosomes are being constantly remodeled as receptors arrive at and leave their response elements. This led to a third story in which he summarized data from a project exploring the global interaction of GR with chromatin, using the DNAse hypersensitivity assay as a tool to discern receptor binding. He showed that, similar to what had been shown for other receptors, a high percentage of binding sites for GR were at significant distances from the promoter and, interestingly, that the majority of GR binding events occur at pre-existing DNAse hypersensitive sites, indicating that GR is not a pioneer factor, and often arrives pursuant to remodeling directed by other factors. Hager emphasized that their data was consistent with chromatin remodeling systems being a major determinant of the cell specificity of GR regulation of target genes. In the final part of his talk, Hager highlighted the importance of pulsatile release of corticosterone in directing GR regulation of target genes. He showed that while GR bound to its physiological ligand was able to cycle off experimental arrays, the synthetic ligand dexamethasone did not permit this due to its extremely high affinity for the receptor. He concluded by invoking three separate time scales for endogenous GR mediated signaling: an hour-based scale (hormone release from its tissue of synthesis), a minute-based scale (the chaperone protein cycle), and a second-based scale (the chromatin-binding cycle). Finally he cautioned that adverse side effects in patients treated with synthetic steroid therapeutics might result from inappropriate transcriptional programs arising as a result of the loss of the natural pulsatile rhythm of endogenous ligands.

Lee Kraus (Cornell University) described efforts in his laboratory to define direct E2 targets on a global scale by separating transcription from steady-state mRNA dynamics. He introduced a new technique, global run-on sequencing (GROseq) that combines nuclear run-on assays with deep sequencing to quantify transcriptionally-engaged polymerase density on a genome-wide scale. Looking at very early time points (<1h) after E2 induction, he showed that the immediate and direct E2-regulated transcriptome includes many genes, including coding and non-coding transcripts, including a significant number of down-regulated genes. The gene sets defined by GRO-seq are distinct from those defined by expression microarrays and RNA PolII-ChIP-sequencing, an observation he attributed, at least in part, to sensitivity issues, or to non-transcriptional endpoints and secondary transcriptional effects. Kraus went on to demonstrate that the majority of genes were regulated at the early (<1h) timepoints, and that these genes were enriched in ERα binding sites within 10kb of the promoter; in contrast to the genes induced later (~3h) and genes in the downregulated category. He surmised that for genes in the latter category, regulation by other transcription factors was important, potentially secondary to rapid, non-genomic signaling by E2. He highlighted the ability of GRO-seq to give an estimate of the rate of processivity of RNA PolII, which he estimated at an average of 1.9kb/minute across the entire dataset. Finally he showed that E2-regulated genes were enriched for paused RNA PolII, although this was, somewhat counterintuitively, less evident in the genes regulated at the 10 min and 40 min time points.

The final talk of the day came from Keith Yamamoto (UCSF). He set the stage by highlighting the ability of small differences between ligands for different steroid receptors to give rise to large differences in biology, and the ability of different ligands for the same receptor to elicit distinct effects from that receptor. He reviewed studies in which a panel of C11-substituted arylpyrazole derivatives were tested for their biological effects in different biological endpoint assays. While dexamethasone scored positive in all the assays, the C11-substituted ligands gave mixed results. Of particular interest though, Yamamoto observed, was the specific information that the ligand gave to the receptor that determined whether or not it would be active in a specific assay, and what point along the receptor activation pathway it would reach. For example, some ligands did not permit DNA binding, whereas some gave instructions to the receptor to bind DNA, but curtailed its activity beyond this point. In summary, he stressed that ligand chemistry is a profound determinant of receptor selectivity, and that agonism and antagonism were simplistic terms that belied the importance of the context of the gene and the cellular context. The second part of Yamamoto’s talk reiterated one of his long-held postulates, namely that GRE sequences are important allosteric effectors of GR structure. Interestingly, he showed that an amino acid sequence in the region of the DNA binding domain was akin to a “lever arm” which he suggested read the specific DNA sequence to which the receptor was bound and communicated a structural change to the rest of the receptor to effect a change its activity. He concluded by emphasizing the fact that combinations of signals – including ligand, DNA and other factors - are amplified and integrated by the GR to specify complex transcriptional networks.

Day 1 | Day 2 | Day 3