Oncogene (2006) 25, 6868–6886 & 2006 Nature Publishing Group All rights reserved 0950-9232/06 $30.00 www.nature.com/onc REVIEW Cross-talk between nuclear receptors and nuclear factor jB

K De Bosscher, W Vanden Berghe and G Haegeman

Laboratory for Eukaryotic Expression and (LEGEST), Department of Molecular Biology, Ghent University, Gent, Belgium

A variety of studies have shown that some activated deficiency states. Cortisone, or at that time called nuclear receptors (NRs), especially the glucorticoid Compound E, had been isolated and identified by , the receptor and peroxisome proli- Kendall and Reichstein as early as 1935 (Glyn, 1998). ferator-activated receptor, can inhibit the activity of the As jaundice leads to retention of bile acids in the body factor nuclear factor jB (NF-jB), which and because of a structural relationship between bile plays a key role in the control of involved in acids and ‘cortin substances’ (isolated from the adrenal inflammation, cell proliferation and apoptosis. This review cortex), the combined efforts of several groups even- describes the molecular mechanisms of cross-talk between tually led to the exciting discovery that cortisone was NRs and NF-jB and the biological relevance of this cross- possibly the miracle drug that could relieve the talk. The importance and mechanistic aspects of selective debilitating symptoms of chronic rheumatoid arthritis. NR modulation are discussed. Also included are future However, the spectacular effects of cortisone on research prospects, which will lead to a new era in the field rheumathoid arthritis appeared to be only temporary of NR research with the aim of specifically inhibiting and reversible. Moreover, the finding that long-term NF-jB-driven gene expression for anti-inflammatory, corticosteroid treatment of chronic ailments is accom- anti-tumor and immune-modulatory purposes. panied by the appearance of a range of negative side Oncogene (2006) 25, 6868–6886. doi:10.1038/sj.onc.1209935 effects further overshadowed the initial successes of cortisone. Among these, side effects are increased blood Keywords: NF-kappaB; ; ; sugar levels leading to diabetes, cataracts, excessive ; ; PPAR weight gain, redistributed fat (leading to puffy cheeks, swelling of the abdomen and a hunchback), osteoporo- sis, muscle wasting, psychosis and neurosis. The reason for the occurrence of multiple side effects is not surprising now that we understand many of the Introduction physiological roles of these in the body. Glucocorticoids, the secretion of which is subject to a Approximately 60 years ago, the observation that an strict hypothalamus–pituitary–adrenal-axis feedback attack of jaundice led to excellent remission in a patient control, exert a variety of biological actions, including with severe rheumatoid arthritis sparked the belief that the regulation of glucose and metabolism, blood the early pathology of this inflammatory disease was pressure, electrolyte balance, and modulation of the reversible. As remissions of rheumatoid arthritis had immune, cardiovascular and central nervous systems also been observed during pregnancy, Philip Hench (Reichardt et al., 2000; Sapolsky et al., 2000; de Kloet initiated attempts to identify the ‘anti-rheumatic et al., 2005). Long-term and/or high-dose administra- substance X’, as well as the female , tion of synthetic glucocorticoids or corticosteroids is which might be responsible for these remissions therefore certain to have unwanted effects on a variety (Glyn, 1998). of physiological processes in the organism. Based on the knowledge that we have today, these Despite continuous efforts of the pharmaceutical phenomena no doubt exemplify the negative cross-talk industry to design anti-inflammatory compounds that of glucocorticoid and sex steroid hormones with nuclear can circumvent side effects while maintaining effec- factor kB (NF-kB), a dimeric (TF) tiveness, no other drug has approached cortisone’s recognized as a key regulator of pro-inflammatory gene therapeutic benefits. Thus, corticosteroids remain the expression (see Gilmore, 2006; Hoffmann et al., 2006). mainstay for the treatment of asthma (Barnes, 2006), Corticosteroids, or steroidal hormones, were initially and are also used in combination with other therapeu- only used as a replacement therapy in adrenocortical tics to treat a wide range of chronic autoimmune and inflammatory diseases, including rheumatoid arthritis, Correspondence: Dr K De Bosscher, Laboratory for Eukaryotic Gene atopy, multiple sclerosis, systemic lupus erythematosis, Expression and Signal Transduction, Department of Molecular Biology, Ghent University, KL Ledeganckstraat 35, B-9000 Gent, inflammatory bowel disease, etc. Similarly, due to their Belgium. potent immunosuppressive effects, glucocorticoids are E-mail: [email protected] used to prevent organ rejection after transplantation Cross-talk between nuclear receptors and NF-jB K De Bosscher et al 6869 and, because of their immune cell apoptosis-inducing involved in dimerization and contacting DNA (through abilities, they are also used in the treatment of certain 8 residues that form two zinc ion-containing types of B- and T-cell lymphoma (Jehn and Osborne, tetrahedric structures); a less conserved D region 1997). Because glucocorticoids efficiently block leuko- containing a nuclear localization signal, sometimes cyte migration towards sites of inflammation (Cato and overlapping with the C domain; and a C-terminal Wade, 1996), by inhibiting relevant NF-kB-driven ligand-binding domain (LBD or E domain), which is cytokines and chemokines, the general symptoms of moderately conserved at the level of DNA sequence but inflammation are quickly dampened. has a well conserved secondary alpha-helical structure. Since the discovery of NF-kB, major advances have Included in the E domain is the AF-2 transactivation been made in understanding the molecular mechanisms subdomain, a strong dimerization interface, a second underlying normal and pathologic inflammatory nuclear localization signal and often a transrepression responses, and recent progress has also led to a more function. detailed understanding of fundamental mechanisms of Surprisingly, 48 genes in man (as compared to more gene transcription. For example, much attention now than 270 genes in Drosophila) are enough to code for the focuses on learning how to read and interpret the phylogenetically related, abundant transcriptional regu- ‘histone code’, inferring an additional level of dynamic lator class of NRs. In man, the glucocorticoid receptor gene regulation. Chromatin organization plays a major (GR) and the mineralocorticoid receptor (MR) are role in controlling the dynamic nature of NF-kB most closely related, and are grouped together with the recruitment to target genes and represents an integra- receptor (PR), the (AR) tion point for mediating diverse interactions between and the estrogen receptor (ER) (Robinson-Rechavi TFs (Burkhart et al., 2005; Natoli et al., 2005; Natoli et al., 2003). Together with the peroxisome prolifera- and De Santa, 2006; Vanden Berghe et al., 2006b). This tor-activated receptor a (PPARa) receptor (belonging to information is paramount for researchers interested in a different group), these NRs have been shown to mechanisms of hormonal repression of NF-kB-driven interact and interfere with the transcriptional activity of gene expression, as additional key regulators of this NF-kB, representing an important regulatory link process are still emerging. Novel insights into hormo- between the endocrine and immune systems. nal repression of NF-kB may provide a solution for Regulation of transcription by NRs can be effected patients who have become resistant to classical by three main mechanisms: ligand-independent gene hormone-based therapies. This review describes the repression, ligand-dependent transactivation and ligand- cross-talk between nuclear receptors (NRs), many of dependent transrepression (Ogawa et al., 2005). In the them are receptors for steroid hormones, and NF-kB, classical model of receptor function, and the biological and therapeutic relevance of this unliganded receptors such as GR, MR, PR and ER are cross-talk. kept in the cytoplasm by (such as Hsp70, and immunophilins) (Pratt and Toft, 1997) and only travel into the nucleus after a ligand- Signaling by hormone receptors induced conformational change, which exposes a nuclear localization signal (Figure 1). However, recent Due to their lipophilic nature, steroid hormones (e.g., data indicate that, in fact, a continuous dynamic NR glucocorticoids, , progesterone, mineralocorti- shuttling between the nuclear and cytoplasmic compart- coids, androgens, vitamin D3, ecdysone, oxysterols and ments takes place, and, as hypothesized for ER, this bile acids), retinoic acids, fatty acids and prostaglandins may allow the receptor to exert both genomic (nuclear) diffuse freely across the plasma membrane without and non-genomic (cytoplasmic) activities (Lo¨ sel and the need for a membrane receptor. Once inside the Wehling, 2003; Leclercq et al., 2006). For GR too, the cell, these compounds exert most of their effects by subcellular localization of both liganded and unliganded binding to specific intracellular receptors, termed receptors is at any given time the net result of a dynamic NRs, which belong to a superfamily of ligand-inducible equilibrium determined by the rates of nuclear import TFs. Ligand-activated NRs, mostly as homo- or and export (Savory et al., 1999), although the import/ heterodimers, can directly regulate their individual export mechanisms are not precisely known (Liu and gene programs by contacting promoters/enhancers at DeFranco, 1999). sequence-specific DNA response elements; however, Other NRs, such as RAR/RXR and PPAR constitu- these NR–ligand complexes can also influence other tively reside in the nucleus, bound to DNA in the signaling pathways (Resche-Rigon and Gronemeyer, unliganded state, and are kept inactive through interac- 1998), such as those driven by NF-kB, cAMP-respon- tion with a nuclear complex, consisting of sive-element binding (CREB), activator protein nuclear corepressor (NcoR) or SMRT (silencing me- (AP-1) or signal transducers and activators of transcrip- diator of retinoic-acid and thyroid hormone receptors) tion (STATs). and histone deacetylases (HDACs), which possess Nuclear receptors share a common structural organi- histone deacetylase activity. Indeed, it is generally zation: an N-terminal variable domain or A/B domain, assumed that chromatin condensation and gene repres- with at least one constitutive transactivation subdo- sion are linked to the activity of histone deacetylases, main (AF-1) and several transactivation subdomains; whereas chromatin relaxation and the promotion a central DNA-binding domain (DBD or C domain), of gene expression involve the activity of histone

Oncogene Cross-talk between nuclear receptors and NF-jB K De Bosscher et al 6870

Figure 1 Scheme depicting activation pathways and nuclear cross-talk between NF-kB and GR. TNF, a pro-inflammatory cytokine, binds to the membrane-bound TNF-receptor, which initiates recruitment of various adaptor proteins and activation of a complex cytoplasmic kinase cascade. Representative and important proteins that make up the TNF-R-associated protein complex and IKK complex are depicted in the red and green dotted circles, respectively. Activation of the IKK complex leads to phosphorylation and ubiquitination of IkBa, the cytoplasmic inhibitor of NF-kB. Subsequent degradation of IkBa by the proteasome releases NF-kB (p50/RelA heterodimer) and allows its translocation to the nucleus. RelA is phosphorylated by cytoplasmic kinases but also additionally by the nuclear kinase MSK-1 (Vermeulen et al., 2003). NF-kB is the exclusive transcription factor for induction of the IL-6 gene in response to TNF and is the final switch that activates a largely pre-existing ‘enhanceosome’ multiprotein complex at the level of the IL-6 promoter. The enhanceosome also displays phosphorylation (i.e., MSK- and/or IKKa-dependent) and acetylation (i.e., CBP/HAT-dependent) activity, facilitating IL-6 gene expression (Vanden Berghe et al., 1999a; Vermeulen et al., 2003; Yamamoto and Gaynor, 2004). GR is a transcription factor that predominantly resides in the cytoplasm in an inactive state. Ligand binding induces a conformational change in GR, allowing the release of its cytoplasmic chaperones and enabling GR’s nuclear translocation. At the IL-6 enhanceosome, the activated GR interferes with the transcriptional activity of NF-kB through distinct, but non-exclusive mechanisms. Three such mechanisms are illustrated: direct physical interaction between RelA and GR (Ray and Prefontaine, 1994; De Bosscher et al., 1997), recruitment of corepressor molecules (HDAC2) (Ito et al., 2000), and blocking of a critical phosphorylation on RNA polymerase II (Nissen and Yamamoto, 2000). GR has also been found to interact with NF-kB family members in the cytoplasm (represented by the complex in the gray dotted line). The biological relevance of this cytoplasmic interaction remains to be established, but may reflect a dynamic nucleocytoplasmic shuttling process, acting as a biological sensor. Note that this figure represents a simplified static model of the IL-6 enhanceosome. It is clear that other factors may be present and that not all factors shown may be present on the promoter simultaneously.

acetyltransferases (HATs) (Davie and Spencer, 1999). substitution of the corepressor by a com- With the unliganded DNA-bound NRs, ligand binding plex, harboring HAT activity. This hormone-dependent induces a conformational change in the AF-2 domain, corepressor/coactivator switch is also known as the which is the a-helical region in the receptor LBD. This ‘ exchange’ model (Torchia et al., 1998; Glass conformational change promotes dissociation of the and Rosenfeld, 2000). Coactivator molecules, such as repressor complex (a process called ‘derepression’), CBP, p300 or SRC-1 (steroid receptor coactivator-1) which involves ubiquitination and proteasomal mediate the transcriptional activity of many TFs, degradation of NcoR (Perissi et al., 2004), followed by including the NRs, by means of their HAT functions,

Oncogene Cross-talk between nuclear receptors and NF-jB K De Bosscher et al 6871 ultimately leading to a more relaxed chromatin structure Cross-talk between NRs and NF-jB and, as such, promoting gene activation. A wide range of NR-interacting cofactor molecules, Cross-talk is defined as the interaction between different which are predominantly nuclear proteins that form TFs, resulting in either a cooperative enhancement or a bridge between TFs and the basal transcription ma- inhibition of gene expression, with an important chinery (BTM), have been identified, including the p160 consequence of cross-talk being that the number of family members (e.g., SRC-1, TIF2/GRIP), PPARg possible gene-specific responses is greatly increased in a coactivator-1 (PGC-1), NcoR, SMRT, transducin-b- cell type-dependent manner. In addition to the direct like 1, etc. (for extensive reviews, see McKenna and effects of NRs on target genes by binding to hormone O’Malley, 2002; Perissi et al., 2004; Perissi and response DNA elements, an important second mechan- Rosenfeld, 2005; Spiegelman and Heinrich, 2004). ism of action of activated NRs, including GR, ER, PR Coactivator molecules interact with liganded NRs via and PPAR, is their interference with other signaling conserved LXXLL (X, any ) motifs, which pathways (e.g., inhibition of gene activation by NF-kB make initial contacts with several helices (notably helix or AP-1). This NR-TF cross-talk is thought to occur 12/AF2) in the LBD of the NR (McKenna and primarily in the nucleus, although cytosolic mechanisms O’Malley, 2002). Subsequently, the HAT-containing have also been reported. complex dissociates and a second coactivator complex NF-kB is the collective name given to a family of TRIP/DRIP/ARC is assembled to contact the basal ubiquitous TFs, activated in response to inflammatory transcriptional machinery, finally resulting in target stimuli and environmental stressors, and the activation gene activation. Upon ERa activation, cyclic and of NF-kB is required for the increased expression of dynamic recruitment of a plethora of protein complexes, many crucial inflammatory and immune response genes including chromatin-remodeling factors, histone (see Pahl, 1999; Gilmore, 2006; www.nf-kb.org). The (de)acetyl- and methyl-transferases, the NF-kB TFs are related to one another through a complex (involved in direct recruitment of the general conserved N-terminal DNA-binding/dimerization do- TF and RNA polymerase II to the promoter), protea- main, called the (Gilmore, 2006). somal subunits, heat shock proteins (Hsp’s), general TFs The prototypical NF-kB heterodimer consists of RelA and RNA polymerase II, has been demonstrated at (p65) and p50, but NF-kB complexes can also include ER-targeted promoters (Glass and Rosenfeld, 2000; homodimers and heterodimers of the p52, c-Rel and Malik and Roeder, 2000; Lewis and Reinberg, 2003; RelB subunits. In most cells, NF-kB complexes are in an Me´ tivier et al., 2003, 2006). Although the functionality inactive, cytoplasmic state by virtue of being bound to of HAT (chromatin-relaxing and thus gene-activating) an inhibitor kB(IkB) inhibitor protein. Upon stimula- and HDAC (chromatin-condensing and thus gene- tion of the cell with an appropriate signal, NF-kBis repressing) complexes in gene transcription is well activated by inactivation of IkB, which occurs via IkB established and generally accepted, it remains puzzling kinase (IKK)-mediated phosphorylation and protea- how chromatin relaxation is initiated. As chromatin some-mediated degradation of IkB. This then allows the severely restricts TF access to promoters and determines nuclear translocation and accumulation of NF-kB, TF-binding kinetics, one can question whether TF leading to gene induction once NF-kB binds to target recruitment of HATs is the driving force for chromatin gene DNA (see Hayden and Ghosh, 2004; Gilmore, relaxation, as binding of the TF-cofactor complex is 2006; Scheidereit, 2006) (Figure 1). determined by a ‘pre-relaxed’ chromatin condition (also Because of its clinical relevance in inflammation, a considered as the chicken and egg problem) (Natoli highly explored research topic is the functional inter- et al., 2005). Of special interest, - ference between GR and NF-kB (see Table 1) (Barnes, dependent large-scale chromatin decondensation has 2001; De Bosscher et al., 2003b). However, there is little been observed, which does not require transcriptional consensus about the precise molecular mechanism(s) activation, ligand-induced coactivator interactions or used by glucocorticoids (GCs) to inhibit NF-kB. In histone hyperacetylation (Nye et al., 2002). Alternative truth, GCs probably utilize different mechanisms to to the cyclic cofactor recruitment model for ER, inhibit NF-kB, depending on the ‘environment’, for occupancy of the AR coactivator complex on PSA example, the particular target gene (that is, not all gene-regulatory regions was found to increase only NF-kB-driven gene promoters have the same TF- gradually, with a peak at 16 h after the binding of composition to drive gene expression), the androgen to the AR, and better fits a model of cell type-dependent presence of certain cofactors, the integrated DNA looping and pol II tracking, compatible surrounding histone code (i.e., the particular set of with long-range interactions between enhancer and chromatin modifications), etc. (Natoli and De Santa, promoter regions (Chakalova et al., 2005; Wang et al., 2006; Vanden Berghe et al., 2006b). Gene-repressing 2005; West and Fraser, 2005). Overall, hormone NRs should, therefore, be considered more as context- signaling requires an intimate and dynamic interplay dependent, multitargeting effectors, rather than media- of cofactor complexes with their chromatin tors of repression via a single pathway (De Bosscher environment. Below, it is discussed in detail how cross- et al., 2003a). talk of hormone signaling with the inflammatory stress To provide a glimpse of the possibilities for NR-TF factor NF-kB can specifically modulate target gene cross-talk, we discuss herein the different proposed expression. mechanisms for this cross-talk, particularly as

Oncogene Cross-talk between nuclear receptors and NF-jB K De Bosscher et al 6872 Table 1 NF-kB-driven pro-inflammatory genes that are negatively disrupted DNA binding were still capable of me- regulated by glucocorticoids diating transrepression of NF-kB (Heck et al., 1994; Cytokines Caldenhoven et al., 1995; Helmberg et al., 1995). IL-1, IL-2, IL-3, IL-4, IL-5, IL-6, IL-9, IL-11, IL-12, IL-13, IL-18, Whether the first or second Zn finger in the DBD of TNFa, GM-CSF, SCF GR is most important for transrepression of NF-kBis unresolved (Lide´ n et al., 1997), and may depend on the Chemokines IL-8, CINC/gro, RANTES, MIP-1a, MCP1-4, eotaxin nature of the investigated responsive elements (promoter target gene and chromatin- or histone modification- dependent) in different cell lines (cofactor context- iNOS dependent) (see below). Similarly, the NF-kB-repressing cPLA2 COX-2 capacity of the ER was also fully dependent on a complete ER DBD, as mutations within or overlapping Adhesion molecules with the DBD of ER abolished ER’s ability to repress ICAM, VCAM, E-selectin the NF-kB-dependent interleukin (IL)-6 promoter (Ray and Prefontaine, 1994). Recently, with the aid of kinetic hormone Endothelin-1 imaging studies (FRAP), the real-time DNA-binding kinetics of GR and NF-kB have been measured in vivo, Abbreviations: COX-2, cyclooxygenase-2; GM-CSF, granulocyte– revealing a quick turnover of DNA binding by both macrophage colony-stimulating factor; IL, interleukin; iNOS, induci- factors (Stavreva et al., 2004; Bosisio et al., 2006). It is ble nitric oxide synthase; MCP, monocyte chemoattractant protein; believed that this rapid turnover on DNA allows the cell NF-kB, nuclear factor kB; RANTES, regulated on activation, normal T-cell expressed and secreted; TNFa, tumor necrosis factor a. to continuously sample changes in nucleoplasmic con- centrations of GR and/or NF-kB in response to external stimuli. It would be exciting if one could determine exemplified by the prototypic GR and NF-kB cross-talk. whether mutual physical interactions of GR and NF-kB Following the discussion of the detailed mechanistic affect real-time binding kinetics of each factor specifi- aspects, we also highlight the physiological relevance cally in transrepressing conditions. and cross-talk mechanisms of other NRs that influence Gene repression by activated NRs appears in most NF-kB activity. Proposed mechanisms can be divided cases to be mediated through interaction with DNA- generally into two categories: nuclear models and bound transcription factors without the need for the NR cytoplasmic models. However, as mentioned above, to contact DNA by itself. A decade ago, it was believed the dynamic nature of protein shuttling events make it that GRa could also mediate gene repression through likely that such strict boundaries do not exist in a real contacting the DNA on so-called negative glucocorti- cell setting. coid response element (nGRE) elements. Subsequently, it was demonstrated, even for the ‘prototypical’ nGRE- containing genes, pro-opiomelanocortin and GnRH, NF-jB and GR: mechanistic aspects that transcriptional repression by GR instead occurred through interference with the activity of NGFI-B/Nur77 Nuclear models and Oct-1 TFs, respectively (Phillips et al., 1997; Physical interaction between NF-kB and GR. Activated Chandran and DeFranco, 1999). Direct protein-protein NRs block the activity of many other TFs, of which the interaction between NGFI-B and GR takes place most commonly studied are NF-kB and AP-1. The primarily via their respective DBDs, although for GR cross-talk is in many cases ‘reciprocal’, meaning that DNA binding and homodimerization are not required. activated NF-kB can also interfere with the transac- Thus, a mechanism quite similar to that proposed for tivation properties of NRs directly bound to DNA transrepression between GR and AP-1 or NF-kBisat (Caldenhoven et al., 1995). Therefore, it was most play (Martens et al., 2005). logical to assume that a direct physical interaction would form the basis of these mutual cross-talks, Interference with the BTM. For glucocorticoid- wherein the interacting TFs blocked each other’s induced repression of the osteocalcin gene, direct DNA-binding and/or transactivation functions. For binding of GR to DNA was found to occlude assembly GRa and NF-kB, a physical association in vivo was of a functional basal transcriptional complex, as the demonstrated in the A549 lung epithelial cell line GR-responsive DNA-binding element partially overlaps (Adcock et al., 1997). A series of deletion and point with the TATA box region (Meyer et al., 1997). In a mutants of GRa pointed to an indispensable role for the simplified cellular NF-kB activation system, circumvent- DBD of GR for interaction with both the Rel homology ing the need for a cytoplasmic signaling cascade through domain and the C-terminal transactivation domains of the use of an exclusively nuclear fusion protein of the RelA (Heck et al., 1994; Caldenhoven et al., 1995; GAL4 DBD with RelA, which could constitutively drive Helmberg et al., 1995; Ray et al., 1995; Lide´ n et al., expression from a GAL4-dependent luciferase con- 1997; Wissink et al., 1997; Moras and Gronemeyer, struct, it was clear that GR-mediated repression was a 1998). Although the GR DBD was important for the nuclear event (De Bosscher et al., 1997). The specificity interaction with RelA, DNA binding per se was not of glucocorticoid repression was further found to required, as point mutants in GR that specifically depend on the TATA box context and thus to involve

Oncogene Cross-talk between nuclear receptors and NF-jB K De Bosscher et al 6873 the BTM (De Bosscher et al., 2000b). More evidence for of how cofactors could only target a certain gene in a an entirely nuclear glucocorticoid repression mechanism designated compartment in the nucleus while leaving the is supported by an unchanged in vivo footprinting same factors associated with different genes in other pattern of NF-kB, bound to its response element in the compartments intact could be answered by adopting the GC-repressed ICAM promoter (Lide´ n et al., 2000). In concept of territorial subdivision (Francastel et al., 2000; addition, chromatin immunoprecipitation (ChIP) ana- Hager et al., 2000; Lemon and Tjian, 2000; Stenoien lysis of NF-kB RelA bound to the IL-8 promoter et al., 2000; Stewart and Crabtree, 2000). Indeed, revealed that under conditions of gene repression, RelA specific nuclear matrix-targeting signals have been found remains bound to its response element (Nissen and in different domains of GR (DeFranco and Guerrero, Yamamoto, 2000). 2000). Experimentally, however, it has been demon- Comparison of a GC-repressed NF-kB-driven IL-8 strated that activation of GR does not affect the promoter to an NF-kB-driven GC-unresponsive IkBa interaction between RelA and CBP and that a nuclear promoter revealed that recruitment of the polymerase II GAL4-RelA point mutant, defective in CBP recruit- Ser2 C-terminal domain phosphorylating kinase pTEFb ment, could still be functionally repressed by GR (De is a prerequisite for GC-mediated repression (Luecke Bosscher et al., 2000b), arguing against a general CBP- and Yamamoto, 2005). Whether this repression mechan- competition model for inhibition or RelA by GR. ism is exclusive for GR, or is also applicable to other Instead, CBP may act as an integrator, facilitating the NRs, is not clear yet. It also remains to be addressed physical interaction between GRa and NF-kB (McKay how pTEFb/RNApol II regulation is affected at and Cidlowski, 2000). Although GR interactions with GR-unresponsive NF-kB-driven reporter constructs specific coactivators are critical for direct GR target with a viral TATA box, or in cells of GC-resistant gene transactivation, they appear to be dispensable for patients (De Bosscher et al., 2000b). How GC can at least certain aspects of GR-mediated transrepression accomplish specific repression of the general transcrip- of NF-kB (Wu et al., 2004). tion machinery is remarkable. However, new evidence Histone acetylation plays a critical role in the demonstrates the existence of differentially phosphory- regulation of inflammatory genes, and increasing lated RNA pol II forms with varying processivity and evidence suggests that this activity is targeted by activity, depending on the C-terminal domain phos- corticosteroids, while mediating their anti-inflammatory phorylation pattern (CTD code), which offers new effects. The activated GR is able to recruit HDAC2 to opportunities for fine-tuning the basal transcription activated inflammatory gene complexes, thus switching complex (Buratowski, 2003; Eissenberg and Shilatifard, off NF-kB-driven gene expression (Ito et al., 2000). 2006; Yamada et al., 2006). Insights into the mechanisms of action of steroid hormones have also come from studies comparing Cofactor-based models. As with ligand-dependent acti- agonists with antagonists. Because the GR antagonist vation of NRs, transrepression by NRs also involves the RU486 does not recruit HDAC2 to the RelA/HAT- recruitment of complexes of various cofactor molecules, containing complex, it was concluded that GCs require a process that happens in a cell/tissue- and promoter- HDAC2 recruitment for their full anti-inflammatory specific manner, although the precise molecular basis for effect (Ito et al., 2001). Along the same lines, an increase this effect remains largely unknown. The early view that in inflammation and corticosteroid resistance, observed cofactors are exclusively nuclear and provide a static in chronic obstructive pulmonary disease and in physical link between the TFs and the transcription asthmatic patients who smoke, is believed to be due to machinery appeared too simplistic. Indeed, besides using a reduction in HDAC2 activity and expression (Barnes, their enzymatic activities (histone acetylation, methyla- 2006). The development of specific HDAC2-activating tion, etc.) to modify the chromatin and to regulate drugs may, therefore, represent a novel approach for transcription, transcriptional cofactors are themselves the treatment of GC-resistant chronic inflammatory subject to regulation by signal-mediated post-transla- diseases. tional modifications, such as phosphorylation, acetyla- Using the bioluminescence resonance energy transfer tion, sumoylation, etc. (Xu et al., 2004; Wu et al., assay, Garside et al. (2004) found that RU486 efficiently 2005b). Counterintuitively, cofactors can even be recruited the corepressor NCoR to GR, unlike de- detected in the cytoplasm, suggesting a broader func- xamethasone (a strong synthetic GR agonist), which tionality (Kao et al., 2001; Miska et al., 2001). recruited the coactivator SRC1. With respect to As both NRs and NF-kB can interact with CBP/p300 transrepression, RU486 was found to impair the (HAT-containing coactivators) to activate their target interaction between GR and RelA, leading to the genes, it was proposed that competition between RelA reduced recruitment of GR to the NF-kB-responsive and GR for limiting amounts of CBP/p300 could region of the IL-8 promoter. Taken together, the account for transrepression of NF-kB-dependent genes mechanism whereby activated GR targets NF-kB-driven by GR (Sheppard et al., 1998). In theory, this model gene expression may involve recruitment of a fully immediately struggles with a lack of specificity, given functional corepressor complex that minimally contains that a large number of other TFs, which are not HDAC2 and NcoR. NcoR and SMRT have been found susceptible to GR-mediated gene repression, also utilize to interact not only with NRs, but also with NF-kB, CBP/p300 for enhancing their transactivation properties suggesting a more general role in mediating repression (De Bosscher et al., 2000b, 2001). However, the question (Lee et al., 2000). This hypothesis is further supported

Oncogene Cross-talk between nuclear receptors and NF-jB K De Bosscher et al 6874 by the finding that, in NcoR-deficient macrophages, promoter-specific transrepression. Considering the cell- de-repression of NF-kB and AP-1-dependent gene specific repertoire of cofactors (qualitative and quanti- networks involved in inflammation, cell migration and tative differences), the exact mechanism deployed by a collagen catabolism can occur (Ogawa et al., 2004). given NR to mediate gene repression will inevitably also For SMRT, it was shown that de-repression involves be highly cell context-specific. Moreover, as different IKKa-dependent phosphorylation, which stimulates NRs repress overlapping but also distinct sets of NF-kB a proteasome-dependent exchange of corepressor for target genes, another level of specificity is reached, coactivator complexes, serving as a prerequisite for opening up the way for combinatorial therapy strategies NF-kB-mediated transcription and survival. It was in combating inflammatory diseases (Ogawa et al., further found that the SMRT-dependent de-repression 2005). It remains puzzling how apparently identical coincides with p300-dependent acetylation of RelA liganded NRs are able to recruit either coactivator or itself (Hoberg et al., 2006). The question thus remains corepressor complexes depending on a transactivation whether NRs can block NF-kB-dependent transcription or transrepression context. In this respect, cofactor by deacetylase-mediated reversal of RelA acetylation studies on liganded monomeric or dimeric GR con- and consequently promote recruitment of the initial formations may be very informative. These results, (recycled) hormone receptor repressor complex or even taken together with the recently discovered signal- a newly composed repressor complex, or alternatively, specific regulation of the cofactors themselves (Wu whether the NRs prevent the assembly of a transcrip- et al., 2005b), suggest that it is unlikely that NR- tion-competent activator complex. mediated transrepression could result from a simple Recently, it was also shown that the ability of certain competition mechanism between activated NF-kB and NRs to inhibit inflammatory responses can be highly activated NRs for a limiting amount of the same signal-specific. For example, the same gene, Cxcl10, that coactivator(s). A plethora of arguments rather point to is activated by NF-kB following stimulation of Toll-like cofactor exchange models as a more likely explanation receptor-4 (TLR-4) with the pro-inflammatory com- (Rosenfeld et al., 2006; Glass and Ogawa, 2006). pound LPS and is sensitive to GR-mediated repression, becomes resistant to GR-mediated repression when Cytoplasmic models activated by the TLR-3 ligand polyinosinic-polycytidilic Cytoplasmic sequestration of NF-kB. More than a acid (poly I:C). The molecular basis for this selectivity decade ago, the discovery that glucocorticoids could involves signal-dependent usage of NF-kB RelA as an induce upregulation of IkBa sparked the belief that GCs obligate TLR-4-specific coactivator of IRF3 transcrip- could effect gene repression by IkBa-dependent seques- tion factor-driven Cxcl10 gene expression (Ogawa et al., tration of NF-kB in the cytoplasm (Auphan et al., 1995; 2005; Reily et al., 2006). Alternatively, the molecular Scheinman et al., 1995). Other NRs have also been basis determining responsivity to GC repression for the shown to upregulate IkBa protein levels; these NRs Cxcl9 gene as compared to the unresponsive IKBA include PR, AR, PPAR (reviewed in De Bosscher et al., (NFKBIA) gene involves the use of IRF3 as a specific 2003b) and retinoid-related orphan receptor alpha 1 coactivator of DNA-bound NF-kB (Ogawa et al., 2005; (Delerive et al., 2001). It remains puzzling, however, that Reily et al., 2006). The role of IRF3 is not exclusive, no classical GRE has been found in the promoter for however, as another factor, positive transcription IkBa and that the GC-induced expression of IkBa is elongation factor b (pTEFb), has been found to be a strictly cell type-dependent. Thus, upregulation of IkBa distinguishing cofactor determining responsivity to GCs cannot be the sole mechanism through which NRs for tumor necrosis factor (TNF)-mediated induction of inhibit NF-kB-driven gene expression, particularly gene expression of IL-8 as opposed to IkBa (Luecke and because transrepression also proceeds in the absence of Yamamoto, 2005), two NF-kB-controlled promoters novel protein synthesis and without removing NF-kB whose NF-kB-binding sites differ by only one base pair. from its response elements in vivo (De Bosscher et al., As a single nucleotide difference between two functional 1997, 2003b; Lide´ n et al., 2000; Nissen and Yamamoto, NF-kB-binding sites is sufficient to dictate alternative 2000). coactivator recruitment (Leung et al., 2004), it is not Contrasting with classical models, in cytosolic rat surprising that different NF-kB-driven genes can display liver extracts, the NF-kB p50/RelA and IkBa complexes a different susceptibility to GR-mediated repression. have been shown to interact with GR. Surprisingly, Along the same lines, nuclear localization of a recently these interactions can occur even in the absence of identified signaling component, SIMPL, is required for glucocorticoids or of an inflammatory signal (Wide´ n TNFa-dependent but not IL-1-dependent NF-kB acti- et al., 2003) (Figure 1). Furthermore, the NF-kB-IkB vation, raising the question of whether and how NRs complex had already been found to shuttle between the also target this TNFa-specific coactivator (Kwon et al., cytoplasm and the nucleus in the absence of signal 2004). (Johnson et al., 1999; Carlotti et al., 2000; Ghosh and In conclusion, NRs cannot be regarded as general Karin, 2002). It was further found that cytoplasmic blocking agents of NF-kB-driven gene expression, sequestration of RelA by IkBa leads to the relocaliza- because only a subset of NF-kB-driven genes is subject tion of nuclear SMRT/NCoR to the to NR-mediated gene repression. NRs can prevent the cytoplasm. Both RelA and IkBa are able to directly interaction of NF-kB with some required coactivators, bind SMRT, and RelA acetylation negatively modulates but not others, providing the basis for signal- and this interaction with SMRT (Espinosa et al., 2003).

Oncogene Cross-talk between nuclear receptors and NF-jB K De Bosscher et al 6875 One piece of evidence that NRs can affect NF-kB (Doucas et al., 2000). However, the presence of shuttling was provided through the use of fluorescent ‘multiple’ nuclear localization signals in GR together fusion proteins. Activated GR was found to alter the with the recent observation that both unliganded and timing of activated NF-kB RelA nuclear occurrence by liganded GRs continuously shuttle between the cyto- increasing the nuclear export rate of RelA (Nelson, plasm and nucleus (Savory et al., 1999) might indicate 1995). However, a series of data argue that no exclusive that the D589–697 GR mutant can still enter the relationship between NF-kB relocalization from the nucleus, where it can mediate effects on NF-kB-driven nucleus to cytoplasm, reduced NF-kB/DNA binding gene activation. Moreover, substitution of RelA and elevation in expression levels of IkBa during GC Ser-276 with Cys in the context of the exclusively repression appears to exist (De Bosscher et al., 2003b). nuclear GAL4-p65 fusion protein, capable of activating a GAL4-dependent reporter gene, demonstrated that Interference with other signal transduction pathways. transactivation by the RelA S276C mutant could still Glucocorticoids have also been shown to interfere with be efficiently repressed by activated GR, ruling out cytoplasmic signaling pathways that impinge on the an exclusive dependence of PKA signaling for the activation of NF-kB. In certain cell types, treatment transrepression mechanism (De Bosscher et al., 2000a). with GCs leads to upregulation of mitogen-activated Although NF-kB and GR are widely recognized as protein kinase phosphatase 1 (MKP-1), which de- mutually antagonistic regulators of adaptive immunity phosphorylates and thus prevents activation of ERK and inflammation, synergistic activities have been and p38 (Kassel et al., 2001; Lasa et al., 2002) and, as reported too, for example, when NF-kB- and GR- such, may inhibit downstream mitogen-actived protein responsive elements were juxtaposed within a single kinase (MAPK)-dependent NF-kB transactivation promoter (Hofmann and Schmitz, 2002). Along the (Vanden Berghe et al., 1998). In line with this model, same line, GC and NF-kB were found to cooperatively impaired MKP-1 activity exacerbates inflammation regulate gene expression of Toll-like receptor-2, a (Salojin et al., 2006). As p38 MAPK is also implicated component of innate immunity, both in a direct manner in the stabilization of pro-inflammatory gene mRNAs and perhaps also indirectly, as GC-mediated negative (e.g., of COX-2 (cyclooxygenase-2), TNFa, IL-6 and cross-talk with p38 MAPK was also implicated in the IL-8), MKP-1 induction by GCs consequently results in synergistically enhanced TLR-2 expression (Shuto mRNA destabilization of the transcripts from these et al., 2002; Hermoso et al., 2004; Smoak and genes (Lasa et al., 2002). However, the fact that in Cidlowski, 2004). presence of cycloheximide, which blocks new protein Other signaling pathways also impact on NR synthesis (of MKP-1), GC-mediated gene repression of functionality. For instance, the GR protein can be IL-6 still occurs proves that MKP-1 induction is not phosphorylated by the MAPK signaling pathway, a universal transrepression mechanism (De Bosscher and this phosphorylation inhibits GR-mediated tran- et al., 1997). In addition, because neither TNF- scriptional activation (Rogatsky et al., 1998). How activated ERK nor p38 activity is affected by GCs in phosphorylation of GR might affect its subcellular L929sA fibroblasts, whereas GR efficiently represses localization, rate of nuclear import/export or turnover IL-6 gene transcription in these cells, MKP-1 induction requires further investigation. As glucocorticoid-resis- is unlikely to be a major determinant for explaining tant asthma patients do benefit from a combination of anti-inflammatory effects of GCs (De Bosscher et al., MAPK inhibitor and GC therapy, this suggests a role 2001). for phosphorylation in GR functionality with respect Another signaling pathway affecting NF-kB/GR to NF-kB transrepression mechanisms (Irusen et al., cross-talk, in both directions, is the protein kinase A 2002). (PKA) pathway. PKA is able to phosphorylate RelA at In the case of ER, allosteric control of the receptor Ser-276 within the Rel homology domain, allowing was demonstrated by the ERK-activated p90 ribosomal complex formation with the coactivator molecule CBP S6 kinase 2 (RSK2), which phosphorylates Ser-167 (Zhong et al., 1998). Ser-276 phosphorylation was on ER. Interestingly, antiestrogens can block RSK2- further found to be essential for RelA-mediated mediated activation of ER (Clark et al., 2001). repression of GR-dependent transcription (the ‘recipro- Similarly, as the RSK2 kinase has also been linked to cal’ transrepression mechanism) (Doucas et al., 2000). NF-kB signaling, inhibition of MAPK/RSK2 and GC The catalytic subunit of PKA (PKAc) can associate therapy may reveal beneficial therapeutic effects (Panta with GR in vivo, and can potentiate GR-dependent et al., 2004). Taken together, the observed differential transcription. Surprisingly, an exclusively cytoplasmic results on MAPK signaling components, even with the mutant of RelA (with a deletion of the nuclear same NR and depending on the cellular context, localization signal) could still transrepress a GR- emphasize the diversity of regulatory possibilities of activated and GRE-driven mouse mammary tumor NRs when modulating the signaling pathways leading virus reporter gene. Similarly, a GRa mutant (D589– to activation of NF-kB. 697) with predominantly cytoplasmic localization could Furthermore, because of the above-mentioned (see also inhibit RelA-driven gene expression. Taken to- section: cofactor-based models) signal-specific and gether with the above-mentioned observation, these promoter-specific cofactor surroundings associated with results argue for a role for PKAc in mediating mutual either NR- or NF-kB-mediated gene activation, it is cross-talk between GR and NF-kB in the cytoplasm more likely that the reciprocal mechanisms of cross-talk

Oncogene Cross-talk between nuclear receptors and NF-jB K De Bosscher et al 6876 between NRs and NF-kB do not necessarily target the estrogens can block DNA binding of NF-kBtokB site same cellular components. promoters (Stein and Yang, 1995; Galien and Garcia, 1997; An et al., 1999). However, a recent study by Cvoro et al. (2006) investigated the mechanism of ER-mediated Cross-talk between NF-jB and the ER repression of the endogenous TNFa gene, containing a composite AP-1/NF-kB element, by means of ChIP Estrogens play an important role not only in reproduc- analysis and re-ChIP analysis. The latter technique tion, but also in the skeletal, cardiovascular, immune allows identification of protein complexes that are and central nervous system (Jordan, 2002, 2004; simultaneously bound at the promoter in vivo. Not only McDonnell and Norris, 2002; Pearce et al., 2003). Their did they find that treatment does not alter biological effects are produced through binding to two recruitment of c-Jun and NF-kB to the TNF-induced distinct receptors, ERa and ERb, which can activate or TNFa promoter, but they also discovered that the pro- repress gene transcription (Nilsson et al., 2001). The inflammatory stimulus on its own leads to recruitment classical (genomic) mechanism of ER action involves of unliganded ERa to the promoter at later time points, estrogen binding to its cognate receptors, after which the where ERa associates with NF-kB and c-Jun and receptors dimerize and bind to specific response functions as a coactivator. The addition of estradiol elements, known as estrogen response elements (EREs) leads to loss of detection of ERa on the TNFa promoter, located in the promoters of target genes. Alternatively, and the concomitant recruitment of one member of the ligand-bound ERs can regulate gene expression without p160 coactivator family member, GRIP, here function- directly binding to DNA through protein–protein ing as a corepressor (Cvoro et al., 2006). The notion interactions with other DNA-binding TFs in the that cofactors, previously identified either as solely nucleus. In addition, membrane-associated ERs mediate coactivators or as corepressors, can switch functionality the non-genomic actions of estrogens, which can lead depending on the promoter context, probably by both to altered functions of proteins in the cytoplasm adopting different conformations, has been described and to effects on the regulation of gene expression before (Rogatsky et al., 2002). Further studies are (Bjo¨ rnstro¨ m and Sjo¨ berg, 2005; Gururaj et al., 2006; needed to address whether this mechanism can be Manavathi and Kumar, 2006). The functional interac- generalized and extended to other cell types and other tion, or cross-talk, between ER and NF-kB has been NF-kB target genes in vivo, or whether it is a peculiarity suggested to play a key role in estrogen’s ability to of the composite NF-kB/AP-1 element in the TNFa prevent age-related conditions and tumorigenesis in vivo promoter. As a matter of fact, a series of earlier studies (Dijsselbloem et al., 2004). Pro- or anti-inflammatory using EMSA analysis, in different cell types, points to roles of estrogens on NF-kB-dependent gene expression the loss of NF-kB binding as the explanation for ER- have been shown to strongly depend on cell type mediated transrepression, although again ER is likely to (signaling circuitries, cofactor dynamics, repertoire of affect NF-kB by a variety of molecular mechanisms ER isoforms and chromatin promoter context) (Cerillo (reviewed in Kalaitzidis and Gilmore, 2005). With et al., 1998; Maret et al., 1999; Wissink et al., 2001; regard to the specific role of cofactors in mediating Cheung et al., 2003; Cutolo et al., 2003; Bhat-Nakshatri cell- and ligand-specific ER responses (Jordan, 2002; et al., 2004; Kalaitzidis and Gilmore, 2005; Vanden Katzenellenbogen and Katzenellenbogen, 2002; Shang Berghe et al., 2006a) or ligand metabolism (Cutolo et al., and Brown, 2002), the development of pharmaceutical 2004). Of special note, loss of ER function has been agents that target ER/cofactor interactions as therapeu- associated with constitutive NF-kB activity and hyper- tics for estrogen-associated pathologies is currently a active MAPK, because of constitutive secretion of subject of intensive research (Heldring et al., 2004; Hall cytokine and growth factors, which ultimately culmi- and McDonnell, 2005; Kong et al., 2005). nates in aggressive, metastatic, hormone-resistant Much effort has also focused on the development of cancers (Ali and Coombes, 2002; Pratt et al., 2003; Lu novel SERMs (selective estrogen receptor modulators), et al., 2004). because the transcriptional activation of proliferation- One of the first reports describing a functional promoting ER target genes is mainly associated with antagonism between ERa and NF-kB looked at regula- breast cancer (Foster et al., 2001a, b; Wu et al., 2005c), tion of the IL-6 cytokine gene (Ray et al., 1994). Because whereas transrepression of NF-kB activity by ER is of the role of IL-6 in osteoclastogenesis (the formation mainly coupled to ER’s beneficial anti-inflammatory of cells that function in the breakdown and resorption of effect and its protective role on bone metabolism. bone tissue), an increased production of IL-6, as a Tamoxifen is an anti-estrogen that is widely used in consequence of the estrogen-depleted menopausal state, the clinic as an effective treatment for ER-positive contributes to the increased resporption of bone which breast cancer, but with clear drawbacks: besides the fact leads to osteoporosis (Ray et al., 1994). Recently, it has that certain tumors are resistant to tamoxifen, treatment become clear that an overall increase in cytokine with tamoxifen also has been associated with an production, including TNFa, also increases the inci- increased incidence of endometrial tumors (Zhou dence of osteoporosis and other inflammatory condi- et al., 2005a, b; Wu et al., 2005c). cDNA microarray tions such as cardiovascular diseases and Alzheimer’s analysis indicates that estrogens and SERMs exert disease (Pacifici, 1996; Ammann et al., 1997; Cacquevel tissue-specific effects through ERa and ERb, resulting et al., 2004). From in vitro studies it was found that in unique profiles of induced target genes (Kian Tee

Oncogene Cross-talk between nuclear receptors and NF-jB K De Bosscher et al 6877 et al., 2004). The existence of ERa point mutants that Interestingly, non-genomic mechanisms of cross-talk favor NF-kB-based transrepression over DNA-bound between ER and NF-kB have also been described with ER target gene transactivation suggests that subtle important biological implications for human health. changes in ER conformation can provoke distinct and Indeed, through a non-genomic effect of ER, the differential biological outcomes. Because most results activation of the NF-kB-dependent gene product suggest that inhibition of the NF-kB pathway is a valid COX-2 leads to production of the atheroprotective strategy to combat breast cancer in many individuals, prostacyclin, PGI2 (Egan et al., 2004), suggesting an continued research now focuses on the development of explanation for why chronic use of selective COX-2 SERMs that are not only anti-proliferative but also anti- inhibitors (used in cases of chronic inflammatory inflammatory and possibly pro-apoptotic. Many cur- diseases) may potentially undermine the cardiovascular rently used SERMS do not have the capacity to protective effects of estrogen. efficiently block NF-kB (Biswas et al., 2005); however, raloxifene has recently been shown to remove RelA from DNA in an ERa-dependent manner in multiple Cross-talk between NF-jB and the myeloma cells (Olivier et al., 2006). In addition, the recently developed non-steroidal compound WAY- Activation of the progesterone receptor can lead to 169916 can selectively block NF-kB signaling through inhibition of NF-kB-driven gene expression (Kalkhoven binding to either ERa or ERb, but fails to induce et al., 1996). This cross-talk has a biological significance classical ER-mediated target gene effects, suggesting in keeping the uterus quiescent before the onset of labor that such compounds could be valuable for the and parturition. At term, fetal lung surfactant protein treatment of ER-positive breast cancers where inhibition secretion activates fetal amniotic fluid macrophages, of NF-kB can induce apoptosis (Chadwick et al., 2005). which in turn activate NF-kB in the uterine wall Estrogen and NF-kB activation can both increase cell (Condon et al., 2003; Mendelson and Condon, 2005). proliferation by inducing expression of the cell cycle Activated NF-kB leads to the expression of proteins that regulatory protein Cyclin D1; moreover, NF-kB can are believed to facilitate uterine contractility, such as enhance cell survival by inducing the expression of COX-2, but NF-kB also antagonizes PR-activated several anti-apoptotic genes (see Dutta et al., 2006). The target genes that modulate uterine contractility by advantage of NF-kB-selective ER modulators may lie in directly interacting with PR, reducing its DNA binding their ability to block the proliferative and pro-survival and transcriptional activity (called ‘reciprocal’ repres- effects of enhanced NF-kB signaling, which currently is sion). Recently, it was found that an NF-kB-induced an inevitable side effect and a possible cancer-promoting change in the expression of a uterine PR isoform event, when estrogen signaling is completely inhibited by also forms the basis of a decrease in PR DNA-binding current ER-positive breast cancer treatments with capacity in the laboring myometrium (Condon et al., aromatase inhibitors (which block estrogen synthesis) 2006). (Biswas et al., 2005). The importance of ER-mediated inhibition of NF-kB in breast cancer is further supported by the observation that active DNA-bound Cross-talk between NF-jB and the AR NF-kB is detected in the majority of ER-negative breast tumors, whereas active NF-kB is absent in their ER- Reciprocal negative cross-talk has also been observed positive counterparts (Biswas et al., 2004). Thus, NF-kB between NF-kB and the AR. Palvimo et al. (1996) found inhibitors may be valuable for the treatment of ER- that activated AR does not cause increased levels of negative breast cancer and, perhaps even in the treat- IkBa, which could sequester and inactivate RelA. ment of hormone-resistant ER-positive breast cancers. Instead, they found that a physical association, albeit In contrast to the negative cross-talk between ER and weak, between AR and RelA could explain their mutual NF-kB, positive synergism between these transcription repression. In a later study, transcriptional interference factors has also been reported. In particular, synergistic between AR and AP-1 or NF-kB was suggested to be activation by ER and NF-kB has been shown for the mediated, at least in part, through competition for gene encoding the serotonin 5HT1A receptor (Wissink intracellular CBP (Aarnisalo et al., 1998). However, et al., 2001). The molecular basis for this synergy needs similar to the reservations for the role of coactivator to be further explored, and this positive interaction squelching in GR-NF-kB cross-talk (described above), it between ER and NF-kB may be rare and highly seems unlikely that coactivator competition between AR promoter-specific. In another experimental setting, and NF-kB would provide sufficient biological specifi- related to gender-specific regulation of longevity genes, city. At the prostate-specific antigen (PSA) promoter, it an estradiol-mediated increase in MAPK activation was was found that cross-modulation between AR and RelA found to activate the NF-kB signaling pathway, thus may occur by a novel mechanism involving binding to a driving the expression of antioxidant enzymes such as common cis-DNA element (Cinar et al., 2005). Mn-superoxide dismutase and glutathione peroxidase; Androgens have been reported to inhibit IL-6 this mechanism provides an elegant explanation for production by bone marrow-derived stromal cells the antioxidant properties of estrogen and its effects (Bellido et al., 1995). Suppression of osteoblast IL-6 on longevity-related genes (Borras et al., 2005; Vina production by androgens may cause, at least in part, the et al., 2005). anti-resorptive effects of androgens on bone (Hofbauer

Oncogene Cross-talk between nuclear receptors and NF-jB K De Bosscher et al 6878 and Khosla, 1999). NF-kB has been implicated in the macrophage foam cell formation in the peritoneal cavity repression of the AR gene promoter (Supakar et al., (Li et al., 2004). 1995; Nakajima et al., 1996). Paradoxically, in Sertoli PPARa-mediated repression of NF-kB signaling cells, NF-kB elements in the AR promoter have been involves a direct protein-protein interaction between identified as being responsible for increased AR expres- PPARa and RelA (Delerive et al., 1999). In addition, sion, representing an important (cell type-specific) activated PPARa can cause a small upregulation of regulatory mechanism required to maintain efficient IkBa protein levels. The transcriptional cooperation spermatogenesis (Delfino et al., 2003). Much like the between ligand-activated PPARa and NF-kB in upre- reciprocal cross-talk between ER and NF-kB in breast gulating IkBa involves recruitment of DRIP/TRAP cancer, NF-kB activation blocks the proliferation of complexes onto the Sp1 site flanking the kB site in the androgen-dependent prostate cancer cells, whereas the IkBa promoter (Delerive et al., 2002). Whether this proliferation of cancer cells devoid of AR is not affected mechanism can be generalized for other NRs is not by NF-kB activation (Nakajima et al., 1996). Converse- known. Nevertheless, as discussed above for GR, the ly, constitutive NF-kB activation has been demonstrated functional contribution of IkBa upregulation to NR- in AR-negative prostate cancer cell lines (Suh et al., mediated NF-kB transrepression is believed to be 2002). As relatively few NF-kB target genes have been marginal. identified in cDNA expression microarray analyses of Besides PPARa, PPARg has also been shown to have prostate cancer cells, further studies will be required to anti-inflammatory and NF-kB-inhibitory effects in other determine whether NF-kB plays a role in human experimental settings or target cells. For instance, prostate cancer development and/or progression, and activation of PPARg in macrophages and foam cells to characterize the potential of NF-kB as a therapeutic inhibits the expression of activated pro-inflammatory target in prostate cancer (Suh and Rabson, 2004). NF-kB target genes such as inducible nitric oxide Finally, with respect to ageing processes, an increase in synthase (iNOS), matrix metalloproteinase-9 (MMP9) NF-kB activity in the liver is believed to be related to the and scavenger receptor A. PPARg may also affect the androgen desensitization that occurs during liver cell recruitment of monocytes to atherosclerotic lesions senescence (Supakar et al., 1995). (Neve et al., 2000) and can suppress inflammation in intestinal epithelial cells (Eun et al., 2006). A novel anti-inflammatory mechanism in the gut involves Cross-talk between NF-jB and PPAR/ PPARg-dependent increased nuclear export of tran- scriptionally active NF-kB (Kelly et al., 2004). Taken PPARs (including subtypes PPARa, b and g) and LXR together, it is clear that the PPAR class of NRs can also (Liver X receptor) are heterodimeric NRs that form a be defined as potential targets for the treatment of complex with RXR and belong to the class of adopted chronic inflammatory diseases, including atherosclero- orphan receptors. Although the endogenous ligands that sis, inflammatory bowel disease and rheumatoid arthri- activate these receptors in vivo are poorly characterized, tis, based on the ability of PPARs to inhibit NF-kB all PPAR members can be activated by various fatty activity. acids (Forman et al., 1996; Kliewer and Willson, 1998). Novel mechanisms of action of clinically used Besides their role as transcriptional regulators of fatty drugs are also being discovered. Statins, which are acid metabolism, PPARs also function as regulators of competitive inhibitors of 3-hydroxy-3-methylglutaryl- inflammatory responses. Thus, PPAR agonists could coenzyme A reductase and are used as cholesterol- potentially be used as a second line of defense and lowering agents, have additional anti-inflammatory protection in combinatorial treatment with glucocorti- effects on NF-kB-driven genes by restoring PPARa coids as such agonists have shown therapeutic effective- transrepression activity upon inhibition of PKCa ness in chronic inflammatory disease models, such as (Paumelle et al., 2006). In this respect, statins are atherosclerosis or obesity-induced insulin-resistance beneficial for patients suffering from the metabolic (Ogawa et al., 2005). Notably, some PPAR ligands syndrome, which encompasses inflammatory, diabetic, have been shown to reduce NF-kB activity in a receptor- and cardiovascular disease states (Sowers, 2003; independent manner, for example, the PPARg-agonist Haffner et al., 2006). LXRs, activated by oxysterols 15dPGJ2 can directly inhibit IKK and the DNA- (oxygenated cholesterol derivatives) function not binding activity of NF-kB (Rossi et al., 2000; Straus only as sensors of cellular cholesterol homeostasis, et al., 2000). but also play a role in adaptive immune responses PPARa ligands, but not PPARg ligands, inhibit IL-1- and have been characterized as regulators of the induced production of IL-6, prostaglandin and COX-2 macrophage inflammatory pathways. LXR agonists in human aortic smooth-muscle cells, which play an tone down the inflammatory response to bacterial important role in plaque formation and post-angio- pathogens by inhibiting the production of a range of plasty re-stenosis. In the vascular wall, PPARa inhibits NF-kB-dependent cytokines, including IL-1b, IL-6, the inflammatory response of aortic smooth-muscle iNOS, MMP9, COX-2. In contrast, LXR knockout cells and decreases the concentration of plasma acute- mice display an exaggerated response to LPS, which is a phase proteins, having a positive outcome on the potent activator of NF-kB. The precise molecular process of atherosclerosis and re-stenosis (Staels et al., mechanism of LXR-mediated inflammatory gene re- 1998; Delerive et al., 1999). PPARa also prevents pression is not yet known, but clearly involves inhibition

Oncogene Cross-talk between nuclear receptors and NF-jB K De Bosscher et al 6879

Figure 2 The molecular rationale behind the separation of GR functionalities. In the absence of ligand, GR is predominantly cytoplasmic and is associated with heat shock and chaperone proteins. Binding of glucocorticoid (black filled circle) or the non- steroidal GR modulator Compound A (CpdA, yellow filled circle) induces the nuclear translocation of GR. Glucocorticoid-activated DNA-bound dimeric GR leads to the expression of metabolic gene products (indicated by the plus sign), which contribute to the occurrence of side effects associated with long-term glucocorticoid usage. A minor pathway leads to the expression of anti- inflammatory proteins, including lipocortin-1, secretory leukoprotease inhibitor, CC protein 10, IL1-receptor antagonist and in some cases IkBa. The anti-inflammatory effects of glucocorticoids are mainly mediated through interference of GR (presumably via GR in its monomeric form) with the transactivation potential of NF-kB, blocking NF-kB-driven gene expression (indicated by the minus sign). CpdA-activated GR does not induce GRE-driven gene expression and only causes transrepression of NF-kB-driven genes. This dissociated property of CpdA makes it not only an interesting fundamental research tool to explore distinct GR-mediated biological effects but also opens avenues for anti-inflammatory applications in the clinic, because a better side effect profile is expected. of NF-kB (Castrillo et al., 2003; Joseph and Tontonoz, and therefore incapable of binding DNA indicate that 2003; Joseph et al., 2003). the NF-kB transrepression pathway is still functional. Apparently, different NR family members recruit From these experiments, it can be concluded that the different cofactors to cross-talk with NF-kB. For exam- anti-inflammatory action of glucocorticoids mainly ple, the repression of inflammatory genes by PPARg arises from the ‘negative’ cross-talk of GR with agonists requires NcoR, whereas estrogen-dependent NF-kB/AP-1 and can be separated from GR’s DNA inhibition of NF-kB requires GRIP, a cofactor origin- binding-dependent transactivation pathway (Reichardt ally identified as a coactivator for GRE-driven gene et al., 1998, 2000, 2001). Third, the finding that the expression. These surprising differences in NR-depen- interactions of GR with specific coactivators appear to dent cofactor functionalities again demonstrate that be dispensable, to some extent, for GR-mediated specific NR-mediated repression of NF-kB-driven genes transrepression of NF-kB supports a model wherein occurs by distinct mechanisms (Pascual et al., 2005; GR-mediated repression is not intrinsically linked to Cvoro et al., 2006). activation and can be separated mechanistically (Wu et al., 2004). Figure 2 illustrates the two main pathways that GR uses to modulate gene expression and the Separating NR functionalities? rationale behind the strategy of attempting to develop molecules that can separate NR functionalities. Three important findings favor the existence of a molecular basis for the separation of the beneficial effects of corticosteroids or other steroids (i.e., transre- Selective NR modulators pression or inhibition of NF-kB and AP-1-driven cytokine gene expression) from their ‘detrimental’ side Since the discovery of cortisone, much effort has been effects (i.e., transactivation or activation of steroid- devoted to modifying the steroid backbone to achieve an responsive genes regulating metabolic processes). First, improved therapeutic index. However, the overall the identification and characterization of in vitro results with respect to side effect profiles have been selective agonists/antagonists of NRs suggest that only rather disappointing (Markham and Bryson, 1995). a subset of the NR functions, induced by the natural The field was re-energized when pharmacologists and ligand, can selectively be elicited (Resche-Rigon and chemists realized the need to separate GR-mediated Gronemeyer, 1998; Vanden Berghe et al., 1999b). transcriptional activation from transcriptional repres- Second, studies with knock-in mice that express mutant sion. Still, reduction of GR transactivation properties by GR receptors defective in their dimerization functions promising ‘dissociating’ ligands seemed to be highly cell

Oncogene Cross-talk between nuclear receptors and NF-jB K De Bosscher et al 6880 type-dependent and could not be sustained in in vivo crystallization studies (Necela and Cidlowski, 2003), situations (Vayssie` re et al., 1997; Vanden Berghe et al., how changes in the topology of the ligand-binding 1999b; Belvisi et al., 2001a, b; Miner, 2002). pocket induced by these novel ligand structures affect Attention now focuses on the development of non- the specific NR’s activity spectrum. steroidal steroid receptor modulators. One recent example is CpdA (Compound A), a plant-derived non- steroidal compound that interacts with GR and elicits Novel prospects and research areas the desired dissociated properties: CpdA potently inhibits NF-kB-driven gene expression in a GR-depen- It is now clear that in order to unravel cross-talk dent manner, but fails to activate GRE-dependent gene mechanisms between NF-kB and NRs, we cannot limit expression, consequently lacking unwanted diabetogenic ourselves to the study of only these two proteins. effects in vivo (De Bosscher et al., 2005). It needs to be Instead, we need to look beyond the first level of simple further investigated whether this type of compound interaction and to expand our research area to other affects other GR target genes involved in osteoporosis, layers of regulatory mechanisms, imposed by chroma- muscle wasting and psychosis/neurosis. Nevertheless, tin-embedded histone or transcription- and/or cofactor- such dissociating ligands may become successful lead modifying proteins. Undoubtedly, the dissection of compounds for drug development. Conceivably, the molecular pathways used by SnuRM (Selective Nuclear molecular rationale for the dissociated effect probably Receptor Modulators) will be of paramount importance resides in a differential cofactor requirement that may be to allow the design of increasingly selective ligands, dictated by the different conformations, which different which can regulate the activity of a receptor in a unique ligands can impose upon the GR. tissue- and/or cell-specific manner (e.g., only where the Today, various examples of structurally unrelated GR inflammation takes place). Future developments for the modulators exist, which seem to support this differential clinic could include the use of peptide aptamers (protein cofactor usage hypothesis. First, ursodeoxycholic acid recognition agents that block protein–protein interac- (UDCA), which is defined as a despite carrying tions), which could interfere with NR activity in a a steroidal structure and is used as an immunomodu- specific manner by mimicking NR/cofactor interactions latory agent in the treatment of various liver diseases, is (Gronemeyer et al., 2004). capable of functionally modulating GR (although its specificity of binding to GR is a debated issue) and of Post-translational modifications of transcription factors suppressing NF-kB-driven gene expression (Miura et al., and/or cofactors 2001; Bellentani, 2005). UDCA attenuates the interac- Apart from well-known ‘classical’ (de)phosphorylation tion between GR and the p160 family protein TIF2/ processes that can switch on/off or modulate the activity GRIP1, indicating that UDCA-stimulated GR cannot of many proteins, it has become clear that TF activity activate AF-2, a mechanism reminiscent of RU486 itself can also be fine-tuned by a variety of other post- antagonist-bound GR (Roux et al., 1999; Miura et al., translational protein modifications, such as acetylation, 2001; Schulz et al., 2002). Second, AL-438, derived by methylation and sumoylation (see also Perkins, 2006). modifying a synthetic progestin scaffold, was found to For example, acetylation of ERa by the p300 acetylase be completely competent for transcriptional repression in a ligand- and SRC-dependent manner modulates the of NF-kB-driven genes, but only a partial agonist for gene-regulatory activity of this NR (Kim et al., 2006). GR-mediate transcriptional activation, and thus capable Recently, it was also found that glucocorticoid binding of separating at least some GR activities in a gene- induces acetylation of the GR protein itself and that specific manner. The mechanism of action of AL-438 subsequent HDAC2-mediated deacetylation of GR is involves differential coactivator recruitments of GRIP1/ necessary to enable its NF-kB-inhibitory activity (Ito TIF2 versus PGC-1; the latter seems preferentially used et al., 2006). in steroid-mediated glucose upregulation (Coghlan It will be interesting to further investigate how ligand- et al., 2003). Third, a series of related arylpyrazole dependent NR activation leads to a specific pattern of compounds can elicit different expression patterns on 17 factor modifications linked to transactivation versus endogenous GR target genes in different cell types (i.e., transrepression. In a recent study, ligand-dependent lung, preadipocyte and preosteoblast), demonstrating sumoylation of the PPARg LBD was found to recruit that subtle differences in ligand structure can have NCoR/HDAC3 complexes to inflammatory gene pro- profound effects on transcriptional regulatory activities moters and to prevent corepressor clearance of the of NRs (Wang et al., 2006). promoter by proteasome components, which maintain Structurally different steroidal ER ligands can also the promoter in a transrepressed state. Of special note, induce dramatically different gene expression profiles in sumoylation rather than ligand binding is the critical a single cell type (Shang and Brown, 2002; Jordan, 2004; event that determines whether PPARg attracts coacti- Kian Tee et al., 2004), further supporting the idea that vator or corepressor complexes (Pascual et al., 2005). even with the same set of cofactors present in the cell, a Recently, not only NRs but also their coregulators different biological outcome can still be produced, have been shown to be targets of post-translational depending on the inducing ligand. In order to predict modifications (e.g., phosphorylation, acetylation, methy- and achieve a specific biological outcome with new NR lation, etc.), upon activation of diverse cellular signaling ligands, an important goal will now be to determine, by pathways (Rosenfeld et al., 2006; Baek and Rosenfeld,

Oncogene Cross-talk between nuclear receptors and NF-jB K De Bosscher et al 6881 2004). It will be a challenge to unravel the combinatorial in NF-kB transrepression remains to be proven (Xu TF/cofactor code required to achieve specific gene et al., 2004). Finally, cell- and hormone-specific responses (Xu et al., 2003; Wu et al., 2005b). responses were also found to depend on the promoter DNA methylation status. As an example, glucocorti- Histone methylation, and coids were found to regulate DNA demethylation within DNA methylation a key enhancer of the rat liver-specific tyrosine Histone methylation has long been considered to be aminotransferase (Tat) gene. As a stronger subsequent exclusively linked to gene repression. However, a recent glucocorticoid response is observed, demethylation study investigating histone modifications involved in appears to provide memory of the first stimulation. (TR)-mediated transcrip- During development, this demethylation occurs before tional regulation revealed that repression by unliganded birth, at a stage when the Tat gene is not yet inducible, TR is associated with a substantial increase in methyla- and it could thus prepare the enhancer for subsequent tion of histone H3 Lys-9 (H3-K9), but a decrease in stimulation by hypoglycemia at birth. In this respect, methylation of H3 Lys-4 (H3-K4) and of H3 Arg-17 demethylation appears to contribute to the fine-tuning (H3-R17). Upon transcriptional activation by liganded of the enhancer and to the molecular memory of a TR, a substantial decrease in both H3-K4 and regulatory event during development (Thomassin et al., H3-K9 methylation and a robust increase in H3-R17 2001). In another context, it was shown that the partial methylation were observed (Li et al., 2002). The estrogenic activity of tamoxifen in the uterus, which methyltransferase CARM1 (coactivator-associated argi- increases the incidence of endometrial cancer, is nine methyltransferase 1) has been found to enhance dependent on hypomethylation of the PAX2 gene (Wu transcriptional activation by liganded NRs through et al., 2005a). Finally, the GR promoter has also been interaction with the coactivators p160 and CBP and shown to be susceptible to regulation by DNA through methylation of histone H3 at Arg-17 (H3-R17). methylation. Specifically, it is reported that certain CARM1 has recently also been linked to NF-kB- maternal behaviors (increased pup licking and grooming mediated transcription through H3-R17 methylation. and arched-back nursing) by rat mothers can alter the In addition, CARM1 physically interacts with NF-kB offspring’s epigenome, including the DNA methylation and is recruited to NF-kB-regulated genes (Miao et al., status of the GR gene promoter in the hippocampus 2006). In contrast, the histone demethylase lysine- (Weaver et al., 2004; Meaney and Szyf, 2005; Szyf et al., specific histone demethylase 1 (LSD1) has been found 2005). However, further study is required to determine to promote AR-dependent transcription. Whether the whether NF-kB transrepession by hormones is also variety of (de)methylases (CARM1, PRMT2, G9A, susceptible to dynamic regulation by DNA methylation LSD1, Ezh2) are preferred targets involved in hor- (Kirillov et al., 1996; Bird, 2003; Murayama et al., 2006; mone-dependent NF-kB transrepression, has not yet Vanden Berghe et al., 2006b). been fully addressed (Varambally et al., 2002; Xu et al., 2003, 2004; Metzger et al., 2005; Lee et al., 2006). Post-transcriptional effects Today, histone methylation biology has reached addi- Up to this point, this review has described only tional complexity, since the discovery of mono-, di- or transcriptional effects of cross-talk between NR and trimethylbranches on lysine residus and the identifica- NF-kB. However, activated GR can also affect NF-kB- tion of various (de)methylating enzymes, suggesting that dependent gene expression by post-transcriptional me- the original histone methylation – repression paradigm chanisms. A classical post-transcriptional effect of GR is is too simplistic (Dutnall, 2003; Peters et al., 2003; destabilization of mRNA from NF-kB-driven genes Schotta et al., 2004; Lee et al., 2006; Shilatifard, 2006). (i.e., iNOS, TNFa, GM-CSF, COX-2) (Tobler et al., Recent evidence has shown that RelA-mediated 1992; Delany et al., 1995; Chaudhary and Avioli, 1996; repression of an integrated GRE-driven MMTV pro- Lasa et al., 2001). Recently, a novel mechanism of post- moter construct requires the brahma-related gene 1 transcriptional regulation by NR cofactors has been (Brg1) chromatin-remodeling activity (Burkhart et al., unveiled. Because of their structural relationship to 2005; Chen et al., 2006). Reciprocally, differential proteins involved in splicing, further exploration NF-kB target gene expression was recently found to rely demonstrated that some of the NR transcriptional on different (antagonistic) functions of the chromatin- coregulators are also part of the spliceosome. Thus, a remodeling complexes Brg1, Brm, Mi-2b (nucleosome subset of NR coregulators acts as dual-function remodeling and histone deacetylase (NuRD)) (Ramirez- ‘coupling’ proteins between transcription and splicing, Carrozzi et al., 2006; Vanden Berghe et al., 2006b). providing the molecular basis for a quality control Of special interest, expression of the Mi-2/NuRD checkpoint. Hormonal stimuli thus control not only the complex component metastasis associated protein quantity but also the ‘nature’ (exon content) of their (MTA) is hormonally regulated and suggests that the target gene products (Auboeuf et al., 2005). In 1999, stoechiometry of chromatin-remodeling complexes is O’Malley and co-workers described the isolation and under hormone control too (Fujita et al., 2003, 2004; functional characterization of SRA, a bona fide Bowen et al., 2004). Furthermore, interesting associa- transcriptional coactivator selective for the AF1 trans- tions have been reported of hormone receptors with a activation domain of steroid receptors and present in a nucleosomal methylation complex, which includes steady-state coregulator complex with the AF2 coacti- Brg1 as well as CARM1, but its potential involvement vator SRC-1, but very much differing from other known

Oncogene Cross-talk between nuclear receptors and NF-jB K De Bosscher et al 6882 protein coregulators, as SRA is an RNA transcript GR modulators and selective PPAR modulators). (Lanz et al., 1999, 2002, 2003). Similarly, there has been This strategy may prove successful in achieving a wave of newly discovered mRNA expression regula- specific receptor regulation and a decreased side effect tors, termed microRNAs, which act at multiple levels as profile. For chronic inflammatory disorders, patient- regulators of gene expression in eukaryotes, and which specific tailor-made symptomatic relief may evolve to are processed from larger transcripts by sequential either multifocal therapeutics or to alternating thera- action of nuclear and cytoplasmic ribonuclease III-like peutics, which might prevent the occurrence of therapy endonucleases (Lund et al., 2004). Consequently, it is resistance. likely that more of these RNA regulators of mRNA expression will undoubtedly soon emerge in the field of NF-kB and NR research. Abbreviations

Concluding remarks AP-1, activator protein-1; AR, androgen receptor; Brg1, brahma-related gene 1; BTM, basal transcription machinery; CARM, coactivator-associated arginine methyltransferase; Today, various studies focusing on the benefits, side COX, cyclooxygenase; DBD, DNA-binding domain; DRIP, effects and risks of hormone therapies have generated receptor-interacting protein; ER, estrogen receptor; debate and confusion among clinicians, researchers, and FRAP, fluorescence recovery after photobleaching; GILZ, the lay public. Herein, we have reviewed how NR glucocorticoid-induced ; GR, glucocorticoid modulators elicit complex tissue- and gene-specific receptor; GRIP, glucocorticoid receptor-interacting protein; responses and what their impact is on NF-kB-driven (n)GRE, (negative) glucocorticoid response element; HAT, gene expression. Major advances in our knowledge of histone acetyltransferase; HDAC, histone deacetylase; Hsp, molecular steroid hormone endocrinology have signifi- ; IkB, inhibitor kB; IKK, IkB kinase; IL, cantly affected modern hormone therapy protocols and interleukin; iNOS, inducible nitric oxide synthase; LBD, ligand-binding domain; LSD1, lysine-specific histone demethy- pharmaceutical development. Seventy years after the lase 1; LXR, liver X receptor; MAPK, mitogen-actived protein discovery of ‘Compound E’ (i.e., the well-known kinase; MKP-1, mitogen-activated protein kinase phospha- steroidal compound cortisone), ‘Compound A’ (CpdA), tase 1; MMP, matrix metalloproteinase; MR, mineralocorti- a plant-derived non-steroidal GR-modulator that speci- coid receptor; MTA, metastasis-associated protein; NcoR, fically inhibits NF-kB in the absence of ‘metabolic’ nuclear corepressor; NF-kB, nuclear factor kB; NR, nuclear GRE-driven side effects, may start a new chapter in GC receptor; PGC-1, PPARg coactivator-1; PPAR, peroxisome pharmacology, which is directed at improved medical proliferator-activated receptor; PK, protein kinase; PLA, anti-inflammatory applications. Because of recent tech- phospholipase A; PR, progesterone receptor; PSA, prostate- nological advances that aid in making traditional specific antigen; p-TEFb, transcription elongation factor; pol, natural product-based libraries more compatible with polymerase; RAR/RXR, retinoid acid receptor/; SEGRM, selective GR modulators; SERM, selective high-throughput screening, there is a renewed interest ER modulators; SMRT, silencing mediator of retinoic-acid in the chemical diversity of natural products in the and thyroid hormone receptors; SRC-1, steroid receptor field of NR drug discovery (Koehn and Carter, 2005). coactivator-1; TBL1, transducin-b-like 1; TF, transcription The outcome of this research may reveal novel factor; TIF, transcription intermediary factor; TLR, toll-like unexpected concepts in hormone pharmacology. More receptor; TNF, tumor necrosis factor; TR, thyroid hormone research will reveal whether CpdA and variants thereof receptor; TRAP, thyroid receptor-activated protein. or other NR-modulating (natural) compounds will emerge as effective anti-inflammatory drugs and Acknowledgements whether they will succeed in knocking cortisone from its pedestal. We thank the University of Gent, the IAP and Marato´ A fertile area for future therapeutic exploration programs, and the FWO-Vlaanderen for financial support. involves the use of combinations of several compounds KDB and WVB are postdoctoral fellows with the FWO- that act on different NR pathways (e.g., selective Vlaanderen.

References

Aarnisalo P, Palvimo JJ, Ja¨ nne OA. (1998). Proc Natl Acad Sci Auphan N, DiDonato JA, Rosette C, Helmberg A, Karin M. USA 95: 2122–2127. (1995). Science 270: 286–290. Adcock IM, Newton R, Barnes PJ. (1997). Biochem Soc Trans Baek SH, Rosenfeld MG. (2004). Biochem Biophys Res 25: 154s. Commun 319: 707–714. Ali S, Coombes C. (2002). Nat Rev Mol Cell Biol 2: 101–115. Barnes PJ. (2001). Allergy 56: 928–936. Ammann P, Rizzoli R, Bonjour JP, Bourrin S, Meyer JM, Barnes PJ. (2006). Eur J Pharmacol 533: 2–14. Vassalli P et al. (1997). J Clin Invest 99: 1699–1703. Bellentani S. (2005). Eur J Gastroenterol Hepatol 17: 137–140. An J, Ribeiro RC, Webb P, Gustafsson JA, Kushner PJ, Baxter JD Bellido T, Jilka RL, Boyce BF, Girasole G, Broxmeyer H, et al. (1999). Proc Natl Acad Sci USA 96: 15161–15166. Dalrymple SA et al. (1995). J Clin Invest 95: 2886–2895. Auboeuf D, Dowhan DH, Dutertre M, Martin N, Berget SM, Belvisi MG, Brown TJ, Wicks S, Foster ML. (2001a). Pulm O’Malley BW. (2005). Mol Cell Biol 25: 5307–5316. Pharmacol Ther 14: 221–227.

Oncogene Cross-talk between nuclear receptors and NF-jB K De Bosscher et al 6883 Belvisi MG, Wicks SL, Battram CH, Bottoms SE, Redford JE, De Bosscher K, Vanden Berghe W, Beck IM, Van Molle W, Woodman P et al. (2001b). J Immunol 166: 1975–1982. Hennuyer N, Hapgood J et al. (2005). Proc Natl Acad Sci Bhat-Nakshatri P, Campbell RA, Patel NM, Newton TR, USA 102: 15827–15832. King AJ, Marshall MS et al. (2004). Br J Cancer 90: De Bosscher K, Vanden Berghe W, Haegeman G. (2000a). 853–859. J Neuroimmunol 109: 16–22. Bird A. (2003). Nat Immunol 4: 208–209. De Bosscher K, Vanden Berghe W, Haegeman G. (2001). Mol Biswas DK, Shi Q, Baily S, Strickland I, Ghosh S, Pardee AB Endocrinol 15: 219–227. et al. (2004). Proc Natl Acad Sci USA 101: 10137–11042. De Bosscher K, Vanden Berghe W, Haegeman G. (2003a) In: Biswas DK, Singh S, Shi Q, Pardee AB, Iglehart JD. (2005). Beyaert R (ed). Nuclear Factor-kappaB: Regulation and Role Sci STKE 2005: pe27. in Diseases. Kluwer Academic Publishers: The Netherlands, Bjo¨ rnstro¨ m L, Sjo¨ berg M. (2005). Mol Endocrinol 19: 833–842. pp 221–239. Borras C, Gambini J, Gomez-Cabrera MC, Sastre J, De Bosscher K, Vanden Berghe W, Haegeman G. (2003b). Pallardo FV, Mann GE et al. (2005). Aging Cell 4: 113–118. Endocr Rev 24: 488–522. Bosisio D, Marazzi I, Agresti A, Shimizu N, Bianchi M, De Bosscher K, Vanden Berghe W, Vermeulen L, Plaisance S, Natoli G. (2006). EMBO J 25: 798–810. Boone E, Haegeman G. (2000b). Proc Natl Acad Sci USA Bowen NJ, Fujita N, Kajita M, Wade PA. (2004). Biochim 97: 3919–3924. Biophys Acta 1677: 52–57. de Kloet E, Joels M, Holsboer F. (2005). Nat Rev Neurosci 6: Buratowski S. (2003). Nat Struct Biol 10: 679–680. 463–475. Burkhart BA, Hebbar PB, Trotter KW, Archer TK. (2005). DeFranco DB, Guerrero J. (2000). Crit Rev Eukaryot Gene J Biol Chem 280: 6349–6358. Exp 10: 39–44. Cacquevel M, Lebeurrier N, Cheenne S, Vivien D. (2004). Curr Delany AM, Gabbitas BY, Canalis E. (1995). J Cell Biochem Drug Targets 5: 529–534. 57: 488–494. Caldenhoven E, Lide´ n J, Wissink S, Van de Stolpe A, Delerive P, De Bosscher K, Besnard S, Vanden Berghe W, Raaijmakers J, Koenderman L et al. (1995). Mol Endocrinol Peters JM, Gonzalez FJ et al. (1999). J Biol Chem 274: 9: 401–412. 32048–32054. Carlotti F, Dower SK, Qwarnstrom EE. (2000). J Biol Chem Delerive P, De Bosscher K, Vanden Berghe W, Fruchart JC, 275: 41028–41034. Haegeman G, Staels B. (2002). Mol Endocrinol 16: Castrillo A, Joseph SB, Marathe C, Mangelsdorf DJ, 1029–1039. Tontonoz P. (2003). J Biol Chem 278: 10443–10449. Delerive P, Monte D, Dubois G, Trottein F, Fruchart-Najib J, Cato AC, Wade E. (1996). Bioessays 18: 371–378. Mariani J et al. (2001). EMBO Rep 2: 42–48. Cerillo G, Rees A, Manchanda N, Reilly C, Brogan I, White A Delfino FJ, Boustead JN, Fix C, Walker WH. (2003). Mol Cell et al. (1998). J Steroid Biochem Mol Biol 67: 79–88. Endocrinol 201: 1–12. Chadwick CC, Chippari S, Matelan E, Borges-Marcucci L, Dijsselbloem N, Vanden Berghe W, De Naeyer A, Haegeman G. Eckert AM, Keith Jr JC et al. (2005). Proc Natl Acad Sci (2004). Biochem Pharmacol 68: 1171–1185. USA 102: 2543–2548. Doucas V, Shi Y, Miyamoto S, West A, Verma I, Evans RM. Chakalova L, Debrand E, Mitchell JA, Osborne CS, Fraser P. (2000). Proc Natl Acad Sci 97: 11893–11898. (2005). Nat Rev Genet 6: 669–677. Dutnall RN. (2003). Mol Cell 12: 3–4. Chandran UR, DeFranco DB. (1999). Behav Brain Res 105: Dutta J, Fan Y, Gupta N, Fan G, Ge´ linas C. (2006). Oncogene 29–36. 25: 6800–6816. Chaudhary LR, Avioli LV. (1996). J Biol Chem 271: Egan KM, Lawson JA, Fries S, Koller B, Rader DJ, 16591–16596. Smyth EM et al. (2004). Science 306: 1954–1957. Chen J, Kinyamu HK, Archer TK. (2006). Mol Endocrinol 20: Eissenberg JC, Shilatifard A. (2006). Curr Opin Genet Dev 16: 1–13. 184–190. Cheung E, Schwabish MA, Kraus WL. (2003). EMBO J 22: Espinosa L, Ingles-Esteve J, Robert-Moreno A, Bigas A. 600–611. (2003). Mol Biol Cell 14: 491–502. Cinar B, De Benedetti A, Freeman MR. (2005). Cancer Res 65: Eun CS, Han DS, Lee SH, Paik CH, Chung YW, Lee J et al. 2547–2553. (2006). Dig Dis Sci 51: 693–697. Clark DE, Poteet-Smith CE, Smith JA, Lannigan DA. (2001). Forman BM, Chen J, Evans RM. (1996). Ann NY Acad Sci EMBO J 20: 3484–3494. 804: 266–275. Coghlan MJ, Jacobson PB, Lane B, Nakane M, Lin CW, Foster JS, Henley DC, Ahamed S, Wimalasena J. (2001a). Elmore SW et al. (2003). Mol Endocrinol 17: 117–127. Trends Endocrinol Metab 12: 320–327. Condon JC, Jeyasuria P, Faust JM, Wilson JW, Mendelson CR. Foster JS, Henley DC, Bukovsky A, Seth P, Wimalasena J. (2003). Proc Natl Acad Sci USA 100: 9518–9523. (2001b). Mol Cell Biol 21: 794–810. Condon JC, Hardy DB, Kovaric K, Mendelson CR. (2006). Francastel C, Schubeler D, Martin DI, Groudine M. (2000). Mol Endocrinol 20: 764–775. Nat Rev Mol Cell Biol 1: 137–143. Cutolo M, Capellino S, Montagna P, Villaggio B, Sulli A, Fujita N, Jaye DL, Kajita M, Geigerman C, Moreno CS, Seriolo B et al. (2003). Clin Exp Rheumatol 21: 687–690. Wade PA. (2003). Cell 113: 207–219. Cutolo M, Villaggio B, Seriolo B, Montagna P, Capellino S, Fujita N, Kajita M, Taysavang P, Wade PA. (2004). Mol Straub RH et al. (2004). Autoimmun Rev 3: 193–198. Endocrinol 18: 2937–2949. Cvoro A, Tzagarakis-Foster C, Tatomer D, Paruthiyil S, Galien R, Garcia T. (1997). Nucleic Acids Res 25: 2424–2429. Fox MS, Leitman DC. (2006). Mol Cell 21: 555–564. Garside H, Stevens A, Farrow S, Normand C, Houle B, Davie JR, Spencer VA. (1999). J Cell Biochem Suppl 32–33: Berry A et al. (2004). J Biol Chem 279: 50050–50059. 141–148. Ghosh S, Karin M. (2002). Cell 109(Suppl): S81–S96. De Bosscher K, Schmitz ML, Vanden Berghe W, Plaisance S, Gilmore TD. (2006). Oncogene 25: 6680–6684. Fiers W, Haegeman G. (1997). Proc Natl Acad Sci USA 94: Glass CK, Rosenfeld MG. (2000). Genes Dev 14: 121–141. 13504–13509. Glass CK, Ogawa S. (2006). Nat Rev Immunol 6: 44–55.

Oncogene Cross-talk between nuclear receptors and NF-jB K De Bosscher et al 6884 Glyn J. (1998). J Royal Soc Med 91: 513–517. Kwon HJ, Breese EH, Vig-Varga E, Luo Y, Lee Y, Goebl MG Gronemeyer H, Gustafsson JA, Laudet V. (2004). Nat Rev et al. (2004). Mol Cell Biol 24: 9317–9326. Drug Discov 3: 950–964. Lanz RB, Chua SS, Barron N, Soder BM, DeMayo F, Gururaj AE, Rayala SK, Vadlamudi RK, Kumar R. (2006). O’Malley BW. (2003). Mol Cell Biol 23: 7163–7176. Clin Cancer Res 12: 1001s–11007s. Lanz RB, McKenna NJ, Onate SA, Albrecht U, Wong J, Haffner SM, Ruilope L, Dahlof B, Abadie E, Kupfer S, Tsai SY et al. (1999). Cell 97: 17–27. Zannad F. (2006). J Cardiovasc Pharmacol 47: 469–475. Lanz RB, Razani B, Goldberg AD, O’Malley BW. (2002). Hager GL, Fletcher TM, Xiao N, Baumann CT, Muller WG, Proc Natl Acad Sci USA 99: 16081–16086. McNally JG. (2000). Biochem Soc Trans 28: 405–410. Lasa M, Abraham SM, Boucheron C, Saklatvala J, Clark AR. Hall JM, McDonnell DP. (2005). Mol Interv 5: 343–357. (2002). Mol Cell Biol 22: 7802–7811. Hayden MS, Ghosh S. (2004). Genes Dev 18: 2195–2224. Lasa M, Brook M, Saklatvala J, Clark RA. (2001). Mol Cell Heck S, Kullmann M, Gast A, Ponta H, Rahmsdorf HJ, Biol 21: 771–780. Herrlich P et al. (1994). EMBO J 13: 4087–4095. Leclercq G, Lacroix M, Laios I, Laurent G. (2006). Curr Heldring N, Nilsson M, Buehrer B, Treuter E, Gustafsson JA. Cancer Drug Targets 6: 39–64. (2004). Mol Cell Biol 24: 3445–3459. Lee DY, Northrop JP, Kuo MH, Stallcup MR. (2006). J Biol Helmberg A, Auphan N, Caelles C, Karin M. (1995). EMBO J Chem 281: 8476–8485. 14: 452–460. Lee SK, Kim JH, Lee YC, Cheong J, Lee JW. (2000). J Biol Hermoso MA, Matsuguchi T, Smoak K, Cidlowski JA. (2004). Chem 275: 12470–12474. Mol Cell Biol 24: 4743–4756. Lemon B, Tjian T. (2000). Genes Dev 14: 2551–2569. Hoberg JE, Popko AE, Ramsey CS, Mayo MW. (2006). Mol Leung TH, Hoffmann A, Baltimore D. (2004). Cell 118: Cell Biol 26: 457–471. 453–464. Hofbauer LC, Khosla S. (1999). Eur J Endocrinol 140: 271–286. Lewis BA, Reinberg D. (2003). J Cell Sci 116: 3667–3675. Hoffmann A, Natoli G, Ghosh G. (2006). Oncogene 25: Li AC, Binder CJ, Gutierrez A, Brown KK, Plotkin CR, 6758–6780. Pattison JW et al. (2004). J Clin Invest 114: 1564–1576. Hofmann TG, Schmitz ML. (2002). Biol Chem 383: 1947–1951. Li J, Lin Q, Yoon HG, Huang ZQ, Strahl BD, Allis CD et al. Irusen E, Matthews JG, Takahashi A, Barnes PJ, Chung KF, (2002). Mol Cell Biol 22: 5688–5697. Adcock IM. (2002). J Allergy Clin Immunol 109: 649–657. Lide´ n J, Delaunay F, Rafter I, Gustafsson J, Okret S. (1997). Ito K, Barnes PJ, Adcock IM. (2000). Mol Cell Biol 20: J Biol Chem 272: 21467–21472. 6891–6903. Lide´ n J, Rafter I, Truss M, Gustafsson JA, Okret S. (2000). Ito K, Jazrawi E, Cosio B, Barnes PJ, Adcock IM. (2001). Biochem Biophys Res Commun 273: 1008–1014. J Biol Chem 276: 30208–30215. Liu J, DeFranco DB. (1999). Mol Endocrinol 13: 355–365. Ito K, Yamamura S, Essilfie-Quaye S, Cosio B, Ito M, Lo¨ sel R, Wehling M. (2003). Nat Rev Mol Cell Biol 4: 46–56. Barnes PJ et al. (2006). J Exp Med 203: 7–13. Lu T, Sathe SS, Swiatkowski SM, Hampole CV, Stark GR. Jehn BM, Osborne BA. (1997). Crit Rev Eukaryot Gene Exp 7: (2004). Oncogene 23: 2138–2145. 179–193. Luecke HF, Yamamoto KR. (2005). Genes Dev 19: 1116–1127. Johnson C, Van Antwerp D, Hope TJ. (1999). EMBO J 18: Lund E, Guttinger S, Calado A, Dahlberg JE, Kutay U. 6682–6693. (2004). Science 303: 95–98. Jordan VC. (2002). Cancer Cell 1: 215–217. Malik S, Roeder RG. (2000). Trends Biochem Sci 25: 277–283. Jordan VC. (2004). Cancer Cell 5: 207–213. Manavathi B, Kumar R. (2006). J Cell Physiol 207: 594–604. Joseph SB, Castrillo A, Laffitte BA, Mangelsdorf DJ, Maret A, Clamens S, Delrieu I, Elhage R, Arnal JF, Bayard F. Tontonoz P. (2003). Nat Med 9: 213–219. (1999). Endocrinology 140: 2876–2882. Joseph SB, Tontonoz P. (2003). Curr Opin Pharmacol 3: Markham A, Bryson HM. (1995). Drugs 50: 317–333. 192–197. Martens C, Bilodeau S, Maira M, Gauthier Y, Drouin J. Kalaitzidis D, Gilmore TD. (2005). Trends Endocrinol Metab (2005). Mol Endocrinol 19: 885–897. 16: 46–52. McDonnell DP, Norris JD. (2002). Science 296: 1642–1644. Kalkhoven E, Wissink S, van der Saag PT, van der Burg B. McKay LI, Cidlowski JA. (2000). Mol Endocrinol 14: (1996). J Biol Chem 271: 6217–6224. 1222–1234. Kao HY, Verdel A, Tsai CC, Simon C, Juguilon H, Khochbin S. McKenna NJ, O’Malley BW. (2002). Endocrinology 143: (2001). J Biol Chem 276: 47496–47507. 2461–2465. Kassel O, Sancono A, Kra¨ tzschmar J, Kreft B, Stassen M, Meaney MJ, Szyf M. (2005). Dialogues Clin Neurosci 7: Cato CB. (2001). EMBO J 20: 7108–7116. 103–123. Katzenellenbogen BS, Katzenellenbogen JA. (2002). Science Mendelson CR, Condon JC. (2005). J Steroid Biochem Mol 295: 2380–2381. Biol 93: 113–119. Kelly D, Campbell JI, King TP, Grant G, Jansson EA, Me´ tivier R, Penot G, Hubner MR, Reid G, Brand H, Kos M Coutts AG et al. (2004). Nat Immunol 5: 104–112. et al. (2003). Cell 115: 751–763. Kian Tee M, Rogatsky I, Tzagarakis-Foster C, Cvoro A, An J, Me´ tivier R, Reid G, Gannon F. (2006). EMBO Rep 7: Christy RJ et al. (2004). Mol Biol Cell 15: 1262–1272. 161–167. Kim MY, Woo EM, Chong YT, Homenko DR, Kraus WL. Metzger E, Wissmann M, Yin N, Muller JM, Schneider R, (2006). Mol Endocrinol 20: 1479–1493. Peters AH et al. (2005). Nature 437: 436–439. Kirillov A, Kistler B, Mostoslavsky R, Cedar H, Wirth T, Meyer T, Carlstedt Duke J, Starr DB. (1997). J Biol Chem 272: Bergman Y. (1996). Nat Genet 13: 435–441. 30709–30714. Kliewer SA, Willson TM. (1998). Curr Opin Genet Dev 8: Miao F, Li S, Chavez V, Lanting L, Natarajan R. (2006). 576–581. Mol Endocrinol 20: 1562–1573. Koehn FE, Carter GT. (2005). Nat Rev Drug Discov 4: 206–220. Miner JN. (2002). Biochem Pharmacol 64: 355–361. Kong EH, Heldring N, Gustafsson JA, Treuter E, Hubbard RE, Miska EA, Langley E, Wolf D, Karlsson C, Pines J, Pike AC. (2005). Proc Natl Acad Sci USA 102: 3593–3598. Kouzarides T. (2001). Nucleic Acids Res 29: 3439–3447.

Oncogene Cross-talk between nuclear receptors and NF-jB K De Bosscher et al 6885 Miura T, Ouchida R, Yoshikawa N, Okamoto K, Makino Y, Reichardt HM, Tuckermann JP, Gottlicher M, Vujic M, Nakamura T et al. (2001). J Biol Chem 276: 47371–47378. Weih F, Angel P et al. (2001). EMBO J 20: 7168–7173. Moras D, Gronemeyer H. (1998). Curr Opin Cell Biol 10: Reily MM, Pantoja C, Hu X, Chinenov Y, Rogatsky I. (2006). 384–391. EMBO J 25: 108–117. Murayama A, Sakura K, Nakama M, Yasuzawa-Tanaka K, Resche-Rigon M, Gronemeyer H. (1998). Curr Opin Chem Biol Fujita E, Tateishi Y et al. (2006). EMBO J 25: 1081–1092. 2: 501–507. Nakajima Y, DelliPizzi AM, Mallouh C, Ferreri NR. (1996). Robinson-Rechavi M, Escriva Garcia H, Laudet V. (2003). Prostate 29: 296–302. J Cell Sci 116: 585–586. Natoli G, Saccani S, Bosisio D, Marazzi I. (2005). Nat Rogatsky I, Logan SK, Garabedian MJ. (1998). Proc Natl Immunol 6: 439–445. Acad Sci USA 95: 2050–2055. Natoli G, De Santa F. (2006). Cell Death Differ 13: 693–696. Rogatsky I, Luecke HF, Leitman DC, Yamamoto KR. (2002). Necela BM, Cidlowski JA. (2003). Trends Pharmacol Sci 24: Proc Natl Acad Sci USA 99: 16701–16706. 58–61. Rosenfeld MG, Lunyak VV, Glass CK. (2006). Genes Dev 20: Nelson HC. (1995). CurrOpin Genet Dev 5: 180–189. 1405–1428. Neve BP, Fruchart JC, Staels B. (2000). Biochem Pharmacol Rossi A, Kapahi P, Natoli G, Takahashi T, Chen Y, Karin M 60: 1245–1250. et al. (2000). Nature 403: 103–108. Nilsson S, Makela S, Treuter E, Tujague M, Thomsen J, Roux S, Terouanne B, Couette B, Rafestin-Oblin ME, Andersson G et al. (2001). Physiol Rev 81: 1535–1565. Nicolas JC. (1999). J Biol Chem 274: 10059–10065. Nissen RM, Yamamoto KR. (2000). Genes Dev 14: 2314–2329. Salojin KV, Owusu IB, Millerchip KA, Potter M, Platt KA, Nye AC, Rajendran RR, Stenoien DL, Mancini MA, Oravecz T. (2006). J Immunol 176: 1899–1907. Katzenellenbogen BS, Belmont AS. (2002). Mol Cell Biol Sapolsky RM, Romero LM, Munck AU. (2000). Endocr Rev 22: 3437–3449. 21: 55–89. Ogawa S, Lozach J, Jepsen K, Sawka-Verhelle D, Perissi V, Savory JG, Hsu B, Laquian IR, Giffin W, Reich T, Hache RJ Sasik R et al. (2004). Proc Natl Acad Sci USA 101: et al. (1999). Mol Cell Biol 19: 1025–1037. 14461–14466. Scheinman R, Cogswell PC, Lofquist A, Baldwin AS. (1995). Ogawa S, Lozach J, Benner C, Pascual G, Tangirala RK, Science 270: 283–286. Westin S et al. (2005). Cell 122: 707–721. Scheidereit C. (2006). Oncogene 25: 6685–6705. Olivier S, Close P, Castermans E, de Leval L, Tabruyn S, Schotta G, Lachner M, Peters AH, Jenuwein T. (2004). Chariot A et al. (2006). Mol Pharmacol 69: 1615–1623. Novartis Found Symp 259: 22–37. Pacifici R. (1996). J Bone Miner Res 11: 1043–1051. Schulz M, Eggert M, Baniahmad A, Dostert A, Heinzel T, Pahl HL. (1999). Oncogene 18: 6853–6866. Renkawitz R. (2002). J Biol Chem 277: 26238–26243. Palvimo JJ, Reinikainen P, Ikonen T, Kallio PJ, Moilanen A, Shang Y, Brown M. (2002). Science 295: 2465–2468. Ja¨ nne OA. (1996). J Biol Chem 271: 24151–24156. Sheppard K-A, Phelps KM, Williams AJ, Thanos D, Glass CK, Panta GR, Kaur S, Cavin LG, Cortes ML, Mercurio F, Rosenfeld MG et al. (1998). JBiolChem273: 29291–29294. Lothstein L et al. (2004). Mol Cell Biol 24: 1823–1835. Shilatifard A. (2006). Annu Rev Biochem 75: 243–269. Pascual G, Fong AL, Ogawa S, Gamliel A, Li AC, Perissi V Shuto T, Imasato A, Jono H, Sakai A, Xu H, Watanabe T et al. (2005). Nature 437: 759–763. et al. (2002). J Biol Chem 277: 17263–17270. Paumelle R, Blanquart C, Briand O, Barbier O, Duhem C, Smoak KA, Cidlowski JA. (2004). Mech Ageing Dev 125: Woerly G et al. (2006). Circ Res 98: 361–369. 697–706. Pearce ST, Liu H, Jordan VC. (2003). J Biol Chem 278: Sowers JR. (2003). Am J Cardiol 91: 14B–22B. 7630–7638. Spiegelman BM, Heinrich R. (2004). Cell 119: 157–167. Perissi V, Aggarwal A, Glass CK, Rose DW, Rosenfeld MG. Staels B, Koenig W, Habib A, Merval R, Lebret M, Torra IP (2004). Cell 116: 511–526. et al. (1998). Nature 393: 790–793. Perissi V, Rosenfeld MG. (2005). Nat Rev Mol Cell Biol 6: Stavreva DA, Muller WG, Hager GL, Smith CL, McNally JG. 542–554. (2004). Mol Cell Biol 24: 2682–2697. Perkins ND. (2006). Oncogene 25: 6717–6730. Stein B, Yang MX. (1995). Mol Cell Biol 15: 4971–4979. Peters AH, Kubicek S, Mechtler K, O’Sullivan RJ, Stenoien DL, Mancini MG, Patel K, Allegretto EA, Smith CL, Derijck AA, Perez-Burgos L et al. (2003). Mol Cell 12: Mancini MA. (2000). Mol Endocrinol 15: 518–534. 1577–1589. Stewart S, Crabtree GR. (2000). Nature 408: 46–47. Phillips A, Maira M, Mullick A, Chamberland M, Lesage S, Straus DS, Pascual G, Li M, Welch JS, Ricote M, Hsiang CH Hugo P et al. (1997). Mol Cell Biol 17: 5952–5959. et al. (2000). Proc Natl Acad Sci USA 97: 4844–4849. Pratt MA, Bishop TE, White D, Yasvinski G, Menard M, Suh J, Payvandi F, Edelstein LC, Amenta PS, Zong WX, Niu MY et al. (2003). Mol Cell Biol 23: 6887–6900. Ge´ linas C et al. (2002). Prostate 52: 183–200. Pratt WB, Toft DO. (1997). Endocr Rev 18: 306–360. Suh J, Rabson AB. (2004). J Cell Biochem 91: 100–117. Ramirez-Carrozzi V, Nazarian A, Sarah C, Gore L, Sridharan R, Supakar PC, Jung MH, Song CS, Chatterjee B, Roy AK. Imbalzano A et al. (2006). Genes Dev 20: 282–296. (1995). J Biol Chem 270: 837–842. Ray A, Prefontaine KE. (1994). Proc Natl Acad Sci USA 91: Szyf M, Weaver IC, Champagne FA, Diorio J, Meaney MJ. 752–756. (2005). Front Neuroendocrinol 26: 139–162. Ray A, Prefontaine KE, Ray P. (1994). J Biol Chem 269: Thomassin H, Flavin M, Espinas ML, Grange T. (2001). 12940–12946. EMBO J 20: 1974–1983. Ray A, Zhang DH, Siegel MD, Ray P. (1995). Ann NY Acad Tobler A, Meier R, Seitz M, Dewald B, Baggiolini M, Sci 762: 79–87. Fey MF. (1992). Blood 79: 45–51. Reichardt HM, Kaestner KH, Tuckermann J, Kretz O, Torchia J, Glass C, Rosenfeld MG. (1998). Curr Opin Cell Biol Wessely O, Bock R et al. (1998). Cell 93: 531–541. 10: 373–383. Reichardt HM, Tronche F, Berger S, Kellendonk C, Schu¨ tz G. Vanden Berghe W, De Bosscher K, Boone E, Plaisance S, (2000). Adv Pharmacol 47: 1–21. Haegeman G. (1999a). J Biol Chem 274: 32091–32098.

Oncogene Cross-talk between nuclear receptors and NF-jB K De Bosscher et al 6886 Vanden Berghe W, Francesconi E, De Bosscher K, Re` sche- Wide´ n C, Gustafsson JA, Wikstrom AC. (2003). Biochem J Rigon M, Haegeman G. (1999b). Mol Pharmacol 56: 373: 211–220. 797–806. Wissink S, van Heerde EC, Schmitz ML, Kalkhoven E, Vanden Berghe W, Dijsselbloem N, Vermeulen L, Ndlovu M, van der Burg B, Baeuerle PA et al. (1997). J Biol Chem 272: Boone E, Haegeman G. (2006a). Cancer Res 66: 22278–22284. 4852–4862. Wissink S, van der Burg B, Katzenellenbogen BS, van der Saag Vanden Berghe W, Ndlovu M, Hoya-Arias R, Dijsselbloem N, PT. (2001). Mol Endocrinol 15: 543–552. Gerlo S, Haegeman G. (2006b). Biochem Pharmacol. Wu H, Chen Y, Liang J, Shi B, Wu G, Zhang Y et al. (2005a). doi:10.1016/j.bq.2006.07.012. Nature 438: 981–987. Vanden Berghe W, Plaisance S, Boone E, De Bosscher K, Wu J, Li Y, Dietz J, Lala DS. (2004). Mol Endocrinol 18: 53–62. Schmitz ML, Fiers W et al. (1998). J Biol Chem 273: Wu RC, Smith CL, O’Malley BW. (2005b). Endocr Rev 26: 3285–3290. 393–399. Varambally S, Dhanasekaran SM, Zhou M, Barrette TR, Wu YL, Yang X, Ren Z, McDonnell DP, Norris JD, Kumar-Sinha C, Sanda MG et al. (2002). Nature 419: Willson TM et al. (2005c). Mol Cell 18: 413–424. 624–629. Xu W, Cho H, Evans RM. (2003). Methods Enzymol 364: Vayssie` re BM, Dupont S, Choquart A, Petit F, Garcia T, 205–223. Marchandeau C et al. (1997). Mol Endocrinol 11: 1245–1255. Xu W, Cho H, Kadam S, Banayo EM, Anderson S, Yates III Vermeulen L, De Wilde G, Van Damme P, Vanden Berghe W, JR et al. (2004). Genes Dev 18: 144–156. Haegeman G. (2003). EMBO J 22: 1313–1324. Yamada T, Yamaguchi Y, Inukai N, Okamoto S, Mura T, Vina J, Borras C, Gambini J, Sastre J, Pallardo FV. (2005). Handa H. (2006). Mol Cell 21: 227–237. FEBS Lett 579: 2541–2545. Yamamoto Y, Gaynor RB. (2004). Trends Biochem Sci 29: Wang Q, Carroll JS, Brown M. (2005). Mol Cell 19: 631–642. 72–79. Wang JC, Shah N, Pantoja C, Meijsing SH, Ho JD, Zhong H, Voll RE, Ghosh S. (1998). Mol Cell 1: 661–671. Scanlan TS et al. (2006). Genes Dev 20: 689–699. Zhou Y, Eppenberger-Castori S, Eppenberger U, Benz CC. Weaver IC, Cervoni N, Champagne FA, D’Alessio AC, (2005a). Endocr Relat Cancer 12: S37–S46. Sharma S, Seckl JR et al. (2004). Nat Neurosci 7: Zhou Y, Eppenberger-Castori S, Marx C, Yau C, Scott GK, 847–854. Eppenberger U et al. (2005b). Int J Biochem Cell Biol 37: West AG, Fraser P. (2005). Hum Mol Genet 14: R101–R111. 1130–1144.

Oncogene