(1)
European Molecular Biology Laboratory (EMBL), Meyerhofstraße 1, 69117 Heidelberg, Germany
Email: Steven.Johnsen@embl.de
Email: Frank.Gannon@embo.org
Abstract
In recent years, there has been a growing realization that a static two-dimensional model of gene activation by transcription factors is inadequate. Based on the work from a number of groups (Kang et al. 2002; Liu and Bagchi 2004; Metivier et al. 2003; Park et al. 2005; Reid et al. 2003; Shang et al. 2000; Sharma and Fondell 2002; Vaisanen et al. 2005), it is becoming clear that transcriptional regulation by nuclear receptors is a dynamic and cyclical process (Metivier et al. 2006). There are significant consequences that arise from this shift in understanding, from nuclear receptors as ligand activated factors that bind to a response element to activate expression of a target gene to a process where the receptor repeatedly binds in order to achieve transcription. New insights that arise from viewing the activation process as cyclical and the consequences of this for developing new strategies that modulate the activity of the estrogen receptor are outlined in this chapter.
As we and others have described, the binding of the estrogen receptor-α (ERα) to an estrogen response element within the promoter of a target gene is the first of many steps that ultimately lead to the engagement and functional activation of RNA polymerase II (PolII) (Liu and Bagchi 2004; Metivier et al. 2003; Park et al. 2005; Reid et al. 2003; Shang et al. 2000). The first cellular response is the binding of SWI/SWF complex that alters the chromatin context. This is followed by modifications to the histone code by histone methyl transferase (HMT) and histone acetyl transferase (HAT)-mediated post-translational modifications to local histones at the promoter region. These changes are multiple and complex and occur in a defined sequence on a given promoter. The outcome of these alterations is the generation of a new chromatin landscape that is permissive to the recruitment of the transcription complex, mediators, elongation factors, and ultimately PolII. This rendition is not much different than the previous description of the mechanism of action of the ERα except that the cyclical process requires that the complexes assembled during the first phase need to be disassembled in order to allow for subsequent activation cycles. Consistent with this model, ERα is a target of the ubiquitin-proteasome pathway (Nawaz et al. 1999; Preisler-Mashek et al. 2002; Reid et al. 2003; Stenoien et al. 2001). Following functional attainment, events previously regarded as transcriptionally repressive then take place to reset the promoter, thereby allowing the commencement of the next cycle. The histone code is changed to a nonpermissive state through the deacetylation of histones by the action of histone deacetylases (HDACs), and the methylation of histones is also reversed. Chromatin condensation is then carried out by the chromatin remodeling complexes prior to ERα initiating the entire process again.
ERα-initiated cycles consist of an initial association phase followed by a subsequent clearance phase. The degradative action of the proteasome and the removal of the post-translational tags associated with an active histone environment lead to this conceptualization of the transcription process. The insight that ERα achieved transcriptional activation in a cyclical manner implies that interfering with any step in the process, including classical repressors of estrogen action, would likely block the assembly of the activation complex and subsequent recruitment of PolII and hence block the induction of gene expression. In many cases of breast cancer, where ERα is expressed and remains central to oncogenic proliferation, analogs of estrogen are used to block the proliferative function of ERα. It follows that an understanding of the consequences of other inhibitors that might act in an indirect way could have significant consequences. For example, inhibitors or activators of some of the steps in ERα-mediated transcriptional activation might lead to alterations in the activity of ERα and could give rise to new therapeutic reagents for the treatment of estrogen-dependent endocrine disorders such as ERα-positive breast cancer and postmenopausal osteoporosis.
1 Inhibiting the Proteasome
The ubiquitin-proteasome system is now well acknowledged as a major regulator of normal cellular physiology. Initially viewed as an end point for the degradation of proteins, it is now recognized to be a significant contributor to the maintenance of the status quo of a well-functioning cell and plays a particularly important role in the regulation of gene transcription (Muratani and Tansey 2003). The estrogen receptor itself is a direct target for ubiquitination (Nawaz et al. 1999; Nirmala and Thampan 1995). It was shown by the O'Malley (Lonard et al. 2000) and Mancini (Stenoien et al. 2001) groups that the widely used proteasome inhibitor, MG132, interfered with the ability of ERα to activate target gene transcription. This inhibition occurs despite an increase in the levels of ERα (through decreased degradation by the proteasome). We confirmed this observation and showed that the same outcome is achieved with another proteasome inhibitor lactacystin (Reid et al. 2003). Critically, and in keeping with the cyclical action of estrogen receptor mediated transactivation, we demonstrated that inhibiting polymerase activity prevents degradation of ERα, while conversely, inhibition of proteasome activity blocks mRNA synthesis from responsive genes.
In order to define the role of the proteasome in ERα-dependent transcriptional activation more clearly, we utilized chromatin immunoprecipitation analysis (ChIP) of the ERα target gene pS2 (Trefoil Factor 1) in the ERα-positive breast cancer cell line MCF7 as a model system. Induction of pS2, under the experimental conditions used, is solely dependent on ERα. We showed that proteins involved in the ubiquitin-proteasome pathway associate with the promoter of pS2 in ChIP assays. Furthermore, using transcriptionally synchronized cells, we showed that components of the ubiquitin-proteasome pathway, including certain proteasome components, were cyclically recruited to the pS2 promoter in an ERα-dependent fashion. It follows that these ERα-recruited ubiquitin-proteasome pathway proteins likely play an important role in disassembling the ERα transcriptional activation complex. This would allow for rapid and precise regulation of transcriptional activation by priming the promoter for subsequent rounds of transcriptional activation when sufficient ligand is available or by blocking additional activation when ligand is no longer present. Consistent with the vital role of the proteasome in ERα-dependent transcription, the proteasome inhibitor MG132 dramatically alters the profile of the cyclic binding of ERα to the pS2 promoter. Although ERα bound to the target promoter for a longer period of time in the presence of MG132, PolII was not recruited (Reid et al. 2003). The data obtained from these kinetic ChIP experiments are in keeping with the data from the inhibition of transactivation of a target gene in a cell and further strengthen the role of cyclic binding of ERα in the activation of gene expression.
The role of the proteasome in regulating ERα-dependent transcription may also have clinical relevance. The dipeptidyl boronic acid proteasome inhibitor PS-341 was recently accepted for therapeutic use under the name Valcade (Bortezomib) for the treatment of multiple myeloma. Interestingly, this drug also shows promise in both in vitro and xenograft experiments for use in the treatment of breast cancer (Teicher et al. 1999). From the above data, it would appear that proteasome inhibitors may potentially be included in the drugs that are used in conjunction with front-line treatments by antiestrogens for the treatment of ERα-positive cancers.
2 Inhibiting Specific Ubiquitin Ligases
Prior to degradation by the proteasome, target proteins are tagged by a series of ubiquitin moieties that are added by the action of a cascade of enzymes that culminate in the polyubiquitination of the substrate protein (Ciechanover 2005). The final component of this system, which also provides specificity, is the E3 ubiquitin ligase. It is estimated that the human genome contains more than 400 ubiquitin ligases. We showed that ERα was ubiquitinated and that both MDM2 and E6AP, members of two different classes of ubiquitin ligases (RING finger and HECT domain, respectively), were in fact present in complexes with ERα on the pS2 promoter (Reid et al. 2003). In a kinetic ChIP experiment, we showed that these ubiquitin ligases were recruited in synchrony and subsequent to ERα, but prior to the arrival of the proteasome components. This is consistent with the hypothesis that the degradation of ERα is an important component of the transcription cycle of ERα. In the presence of proteasome inhibitor, the E3 ligases, but not the proteasome component Rpt6 (or PolII), were still recruited to the pS2 promoter together with ERα. One would predict, therefore, that inhibitors of specific ER-associated E3 ligases should also interfere with the ability of ERα to transactivate a target gene. One such candidate is the Nutlin family of compounds developed by Hoffman-La Roche. The Nutlins are specific inhibitors of the interaction between MDM2 and p53 (Vassilev et al. 2004) and show great potential for therapeutic use in the treatment of many types of cancer. The impact of these compounds on ERα action will be very instructive and could potentially point to a further novel approach to the treatment of ERα-positive cancers.
3 HDAC Inhibition
As indicated above, an integral part of the transcription cycle for estrogen-regulated genes is the removal of acetylation tags from histones. The histone deacetylases (HDACs) are a family of enzymes responsible for this step and have been shown to be recruited to the pS2 promoter during transcriptional activation at the time when the binding phase of ERα has been completed (Metivier et al. 2003). Although deacetylation is frequently associated with the shut-down of a chromatin locus, it was critical to determine whether the inhibition of the HDAC activity indeed promoted gene expression by interfering with the shutting down of the chromatin locus or inhibited it by blocking the integrated steps in the cyclical binding of the transcription factor ERα. A medically relevant alternative to Trichostatin A (TSA), the reagent most commonly used to block HDAC activity in the laboratory, is valproic acid (VPA). Interestingly, VPA is widely used clinically for the treatment of convulsions, epilepsy, and related disorders and is currently in phase II trials as a potential cancer therapy (Chavez-Blanco et al. 2005). The reported side effects of VPA treatment include many that one might anticipate from an interruption of the normal physiological function of ERα such as reproductive abnormalities (Duncan 2001; Isojarvi et al. 1993; O'Donovan et al. 2002), endocrine disorders (Rattya et al. 2001), and decreased bone mass (Sato et al. 2001). Consistent with these results, we demonstrated that when cells were transfected with an estrogen-responsive reporter construct (ERE-TK-luciferase) and treated with VPA (or TSA), the transactivation was greatly diminished (Reid et al. 2005). The consequences at the cellular level were also profound, with the compounds showing a cytotoxicity profile that matched their inhibition of reporter gene activation. Using expression microarray assays, it was further shown that there was a 90% overlap between the effects of VPA and tamoxifen (the estrogen-receptor ligand used extensively in the treatment of breast cancer) on a family of estrogen-responsive genes. Taken together it is clear that the inhibition of histone deacetylase activity did not stimulate transcription of ERαactivated genes, but rather inhibited the expression of these genes, thus adding further support to the proposed essential role for the integration of all steps in the cyclical induction of transcription by ERα.
Interestingly, the biological effects of VPA also appear to be linked to the ubiquitin-proteasome pathway. For example, while VPA treatment decreased ERα mRNA expression, it also decreased ERαprotein levels in a proteasome-dependent manner (Reid et al. 2005). Furthermore, VPA treatment has been shown to increase targeted protein degradation by specifically increasing the expression of the ubiquitin-conjugating enzyme UBC8, thereby promoting the ubiquitination and subsequent proteasomal degradation of substrate proteins by the ubiquitin ligase RNF12 (Kramer et al. 2003). We are currently investigating whether RNF12 also plays a role in the regulation of ERα-dependent transcription and whether this may account for the proteasome-dependent function of VPA in mediating increased ERα degradation.
Arising from these studies, yet another family of therapeutic agents can be viewed as having potential in the treatment of breast cancer. As indicated above, VPA is widely used at high concentrations for other disorders and is beginning to be investigated for its therapeutic efficacy in cancer. A study is being prepared to follow the epidemiology of women that received VPA for a different dysfunction with respect to the incidence of breast cancer. In light of our results, one would predict that women who have undergone VPA treatment for other disorders would have a lower risk of breast cancer. However, our lack of understanding regarding some of the early stages of the onset of breast cancer could give rise to a different outcome. It would perhaps be surprising if there were no effect on ERα-mediated cancers and therefore the outcome of this study will be of great interest.
4 Inhibition of DNA Methylation
The mechanism of action of VPA on estrogen-regulated transcription was examined in detail by ChIP experiments using the pS2 promoter as an indicator of ERα binding and activity. These experiments suggested a role for the involvement of the maintenance DNA methyl transferase-1 (DNMT1) and methyl CpG binding protein-2 (MeCP2) in the activation of gene transcription. In the absence of the HDAC inhibitor, association of DNMT1 and MeCP2 was readily detected by ChIP analysis, whereas treatment with VPA increased binding of MeCP2, but decreased the presence of DNMT1 on the promoter (Reid et al. 2005). Furthermore, assays using the methylation-sensitive restriction enzyme HpaII as an indicator of the methylation status of CpG dinucleotides in the vicinity of the promoter demonstrated that a change in the methylation status occurred at specific CpGs near the transcriptional start site (Reid et al. 2005). Moreover, re-expression of ERα in an ERα-negative cell line MDA-MB-231 reversed the repressed state of the pS2 promoter (Metivier et al. 2004; Reid et al. 2005). All of these data suggest that there are variations in the methylation status of the actively transcribed promoter under study in MCF7 cells, and that this may be linked in some way with ER cycling and HDAC recruitment to the locus. It follows that interference with the methylation status of a promoter should also block the activation of the gene.
5 Inhibiting RNA Polymerase II
The obvious and inevitable consequence of inhibiting PolII would be the blocking of transcription. But should that have an influence on the ERα in the transactivation cycle? It could have been that there was no effect, but if ERα and PolII are components of the same transcription environment, we reasoned that there could be some feedback on the activator (ERα). In support of this, ERα protein levels decrease to approximately 10% of that in cells grown in estrogen-free medium upon the addition of estrogen. The engagement of ERα in activating transcription therefore usually results in a general destabilization of ERα. Interestingly, upon the addition of tamoxifen, ERα levels increase. It is known that tamoxifen recruits co-repressors in breast cancer cells and suggests that the stabilization may arise from a conformational change of ERα in the presence of tamoxifen or because of some steric hindrance to the degradative process by the presence of the co-repressors. When we tested the stability of a naturally occurring truncated ERα isoform (ER46) (Flouriot et al. 2000), we found that this version of ERα is stable even in the presence of E2. All of these data suggest that the regulation of ERα stability is at the center of complex interactions that appear to be related to its engagement in gene activation. For these reasons, it was not surprising that the use of inhibitors of PolII resulted in the stabilization of ERα (Reid et al. 2003). The image is therefore reinforced that ERα and PolII are the start and end points of the cycle of gene activation and that, like cogs in a wheel, any blockage in this cycle inevitably leads to effects on the other components of the system, even if their activities are required at different time points in the cycle.
6 Nontargeted Inhibition
Since the ERα transcription activation cycle was described, approximately 50 proteins with a variety of functions were shown to cycle on the pS2 target promoter in synchrony with the ERα. The total range of participants in the process, however, could be much greater that that. The inhibition studies outlined above had a specific target. We reasoned that new components of the transcription activation cycle could be identified by a screen of diverse chemical compounds. Therefore we tested a library of 55,000 compounds with drug potential for their effects on the MCF7 (ERα-positive) cell line transiently transfected with an estrogen response element reporter construct. Most compounds identified in the screen had no effect. However, approximately 1% inhibited the induction of ERα activity while 1% increased the activity. Focusing on the inhibitors, we showed that five different chemical classes were involved with multiple hits coming from closely related compounds. The inhibition of ERα action was achieved with approximately the same concentration of the compounds required to inhibit proliferation. The analysis to date suggests that none of these compounds have targets that have been analyzed for their role in ERα-mediated transcription. However, it should be noted that the effects on ERα could be indirect and that further studies on the kinetics of the inhibitors will be required. If the ERα-modulated process is not a target, but is coincidentally blocked due to destruction or misregulation of some other key steps in the physiology of the cell, it could be equally instructive, as the goal in cancer treatment is generally to destroy the tumor cells.
7 Conclusions
The primary goal of most medically relevant research is to understand the basic biology of a system so that this knowledge can then be used in a practical manner to develop new therapies. For those working on ERα, longer-term targets include breast cancer and osteoporosis. By focusing on the actions of ERα when it acts as a transcription factor, it has become clear that ERα binds to the promoter, recruits, in a sequential and programmed manner, a series of proteins that change the promoter region profoundly and prepare the local context for the engagement of PolII, thus leading to transcriptional activation. The binding of ERα and the assembly of the activation complex has been shown to be cyclic and all evidence obtained to date is concordant with a very tight interlinkage between all steps of the cycle and between the subsequent cycles. The productive process eventually stops when the level of estrogen in the system diminishes, as happens under normal physiological conditions such as during the estrous cycle and at the onset of menopause. The new insights on how this transcription factor works are of general interest, but the focus of this report is the consequence of this knowledge for treatment of cancer. We have shown that when steps in the cycle have been blocked by well-known inhibitors, the whole process was interrupted. As many of these inhibitors are used in clinical settings for diseases other than breast cancer, it follows that new mechanisms of modulating ERα activity have been identified. In addition to inhibitors of the proteasome, specific E3 ligases and HDACs, the above-described chemical screen demonstrates that ERα-mediated transcription is a potentially fruitful source of new and improved therapeutic agents. Obviously, the steps between an inhibitor and a drug are fraught with difficulties and many such approaches lead to disappointment. Nonetheless, the fact that basic research repeatedly gives rise to such hopes validates the general view that investment in the generation of knowledge is the correct approach to identifying new therapeutic compounds.
References
Chavez-Blanco A, Segura-Pacheco B, Perez-Cardenas E et al. (2005) Histone acetylation and histone deacetylase activity of magnesium valproate in tumor and peripheral blood of patients with cervical cancer. A phase I study. Mol Cancer 4:22PubMedCrossRef
Ciechanover A (2005) Proteolysis: from the lysosome to ubiquitin and the proteasome. Nat Rev Mol Cell Biol 6:79–87PubMedCrossRef
Duncan S (2001) Polycystic ovarian syndrome in women with epilepsy: a review. Epilepsia 42(Suppl 3):60–65PubMedCrossRef
Flouriot G, Brand H, Denger S et al. (2000) Identification of a new isoform of the human estrogen receptor-alpha (hER-alpha) that is encoded by distinct transcripts and that is able to repress hER-alpha activation function 1. EMBO J 19:4688–4700PubMedCrossRef
Isojarvi JI, Laatikainen TJ, Pakarinen AJ et al. (1993) Polycystic ovaries and hyperandrogenism in women taking valproate for epilepsy. N Engl J Med 329:1383–1388PubMedCrossRef
Kang Z, Pirskanen A, Janne OA et al. (2002) Involvement of proteasome in the dynamic assembly of the androgen receptor transcription complex. J Biol Chem 277:48366–48371PubMedCrossRef
Kramer OH, Zhu P, Ostendorff HP et al. (2003) The histone deacetylase inhibitor valproic acid selectively induces proteasomal degradation of HDAC2. EMBO J 22:3411--3420PubMedCrossRef
Liu XF, Bagchi MK (2004) Recruitment of distinct chromatin-modifying complexes by tamoxifen-complexed estrogen receptor at natural target gene promoters in vivo. J Biol Chem 279:15050–15058PubMedCrossRef
Lonard DM, Nawaz Z, Smith CL et al. (2000) The 26S proteasome is required for estrogen receptor-alpha and coactivator turnover and for efficient estrogen receptor-alpha transactivation. Mol Cell 5:939–948PubMedCrossRef
Metivier R, Penot G, Hubner MR et al. (2003) Estrogen receptor-alpha directs ordered, cyclical, and combinatorial recruitment of cofactors on a natural target promoter. Cell 115:751–763PubMedCrossRef
Metivier R, Penot G, Carmouche RP et al. (2004) Transcriptional complexes engaged by apo-estrogen receptor-alpha isoforms have divergent outcomes. EMBO J 23:3653–3666PubMedCrossRef
Metivier R, Reid G, Gannon F (2006) Transcription in four dimensions: nuclear receptor-directed initiation of gene expression. EMBO Rep 7:161–167PubMedCrossRef
Muratani M, Tansey WP (2003) How the ubiquitin-proteasome system controls transcription. Nat Rev Mol Cell Biol 4:192–201PubMedCrossRef
Nawaz Z, Lonard DM, Dennis AP et al. (1999) Proteasome-dependent degradation of the human estrogen receptor. Proc Natl Acad Sci USA 96:1858–1862PubMedCrossRef
Nirmala PB, Thampan RV (1995) Ubiquitination of the rat uterine estrogen receptor: dependence on estradiol. Biochem Biophys Res Commun 213:24–31PubMedCrossRef
O'Donovan C, Kusumakar V, Graves GR et al. (2002) Menstrual abnormalities and polycystic ovary syndrome in women taking valproate for bipolar mood disorder. J Clin Psychiatry 63:322–330PubMedCrossRef
Park KJ, Krishnan V, O'Malley BW et al. (2005) Formation of an IKKalpha-dependent transcription complex is required for estrogen receptor-mediated gene activation. Mol Cell 18:71–82PubMedCrossRef
Preisler-Mashek MT, Solodin N, Stark BL et al. (2002) Ligand-specific regulation of proteasome-mediated proteolysis of estrogen receptor-alpha. Am J Physiol Endocrinol Metab 282:E891–E898PubMed
Rattya J, Pakarinen AJ, Knip M et al. (2001) Early hormonal changes during valproate or carbamazepine treatment: a 3-month study. Neurology 57:440–444PubMedCrossRef
Reid G, Hubner MR, Metivier R et al. (2003) Cyclic, proteasome-mediated turnover of unliganded and liganded ERalpha on responsive promoters is an integral feature of estrogen signaling. Mol Cell 11:695–707PubMedCrossRef
Reid G, Metivier R, Lin CY et al. (2005) Multiple mechanisms induce transcriptional silencing of a subset of genes, including oestrogen receptor alpha, in response to deacetylase inhibition by valproic acid and trichostatin A. Oncogene 24:4894–4907PubMedCrossRef
Sato Y, Kondo I, Ishida S et al. (2001) Decreased bone mass and increased bone turnover with valproate therapy in adults with epilepsy. Neurology 57:445–449PubMedCrossRef
Shang Y, Hu X, DiRenzo J et al. (2000) Cofactor dynamics and sufficiency in estrogen receptor-regulated transcription. Cell 103:843–852PubMedCrossRef
Sharma D, Fondell JD (2002) Ordered recruitment of histone acetyltransferases and the TRAP/Mediator complex to thyroid hormone-responsive promoters in vivo. Proc Natl Acad Sci USA 99:7934–7939PubMedCrossRef
Stenoien DL, Patel K, Mancini MG et al. (2001) FRAP reveals that mobility of oestrogen receptor-alpha is ligand- and proteasome-dependent. Nat Cell Biol 3:15–23PubMedCrossRef
Teicher BA, Ara G, Herbst R et al. (1999) The proteasome inhibitor PS-341 in cancer therapy. Clin Cancer Res 5:2638–2645PubMed
Vaisanen S, Dunlop TW, Sinkkonen L et al. (2005) Spatio-temporal activation of chromatin on the human CYP24 gene promoter in the presence of 1alpha,25-Dihydroxyvitamin D3. J Mol Biol 350:65–77PubMedCrossRef
Vassilev LT, Vu BT, Graves B et al. (2004) In vivo activation of the p53 pathway by small-molecule antagonists of MDM2. Science 303:844–848PubMedCrossRef