The Interactome of the Glucocorticoid Receptor and Its Influence on the Actions of Glucocorticoids in Combatting Inflammatory and Infectious Diseases

SUMMARY

Glucocorticoids (GCs) have been widely used for decades as a first-line treatment for inflammatory and autoimmune diseases. However, their use is often hampered by the onset of adverse effects or resistance. GCs mediate their effects via binding to glucocorticoid receptor (GR), a transcription factor belonging to the family of nuclear receptors. An important aspect of GR's actions, including its anti-inflammatory capacity, involves its interactions with various proteins, such as transcription factors, cofactors, and modifying enzymes, which codetermine receptor functionality. In this review, we provide a state-of-the-art overview of the protein-protein interactions (PPIs) of GR that positively or negatively affect its anti-inflammatory properties, along with mechanistic insights, if known. Emphasis is placed on the interactions that affect its anti-inflammatory effects in the presence of inflammatory and microbial diseases.

INTRODUCTION

Glucocorticoids (GCs) are steroid hormones that have important effects on many physiological functions due to their diverse roles in inflammation, growth, metabolism, and development. Owing to their antiproliferative and anti-inflammatory properties, GCs have been used for decades to treat malignancies (1) and various inflammatory and autoimmune diseases (2). Synthetic GCs commonly used in the clinic include dexamethasone (Dex), betamethasone, cortisone, hydrocortisone, methylprednisolone, and prednisolone (3). Endogenous GCs, produced by the adrenal cortex, are strictly regulated by the HPA (hypothalamic-pituitary-adrenal) axis, and their lipophilicity enables them to easily diffuse through the cell membrane. Intracellularly, GCs mediate their effects through the glucocorticoid receptor (GR), which is a transcription factor belonging to the superfamily of nuclear hormone receptors. In the absence of GCs, GR is sequestered in the cytoplasm in a complex with chaperone proteins, which maintain its inactive state and organize the structural conformation of the receptor to accept the ligand when it is present in the cytoplasm. Upon ligand binding, GR becomes active and shuttles dynamically between the cytoplasm and nucleus, where it exerts its actions as a transcription factor, regulating many GR-dependent genes both positively and negatively. Transcriptional induction by GR is triggered mainly by the binding of GR homodimers to promoter regions containing the palindromic glucocorticoid response elements (GREs), a mechanism known as GR-dependent transactivation (TA). On the other hand, GR-dependent transrepression (TR) is mediated by the interaction of GR with DNA-bound transcription factors such as NF-κB and activator protein-1 (AP-1), resulting in the repression of the respective inflammatory signaling cascades (4). For a long time, it was believed that transrepression is mediated by the monomeric GR, but recent data derived from advanced microscopy assays suggest that dimerized GR may also interact with the p65 subunit of NF-κB (5). Interestingly, the interference of GR with the NF-κB and AP-1 pathways can lead to mutual inhibition because both transcription factors have been shown to negatively regulate GR activity. Apart from the classic TR mechanism, it was shown recently that transcriptional repression can also be mediated by direct GR binding onto so-called negative GREs (nGREs) (6) or onto half-site GREs (7, 8), by the cooperative binding of GR with other transcription factors in composite GREs (9), by the binding of GR to competitive GREs, or, finally, by GR-mediated sequestration of transcription factors, preventing their binding to their respective response elements (10, 11).

In addition to the vital functions of GR in the nucleus, GR has been reported to be involved in rapid nongenomic effects that occur within minutes after its activation and significantly contribute to the receptor's anti-inflammatory actions, as discussed below (12).

Throughout its intracellular journey, GR creates physical contacts with coregulator proteins that mediate diverse functions. The protein-protein interactions (PPIs) of GR in different cellular compartments are crucial for the resolution of the physiological processes in which GR is involved. Generally, proteins that bind to GR are transcription factors, enzymes, chromatin modulators, coactivators, corepressors, and cochaperone proteins (13, 14). Although we focus in this review on GR-dependent mechanisms, for the sake of completeness, it must be noted that GR-independent anti-inflammatory actions of GCs have also been postulated. Steroids are highly lipophilic and tend to accumulate in lipid membranes. Hence, they can alter membrane fluidity and might influence the function of embedded proteins such as ion channels or receptor proteins (15). In immune cells, the interaction of GCs with plasma membranes leads to rapidly reduced calcium and sodium cycling across the membranes, which is thought to contribute to immunosuppression and the reduction of inflammation. Additionally, GCs contribute to the suppression of ATP production, essential cytokine synthesis, migration, phagocytosis, and antigen processing and presentation by increasing mitochondrial proton leakage (12).

In this review, we provide a state-of-the-art overview of the PPIs in which GR is involved and that affect its activity in a positive or negative way. Emphasis is placed on GR interactions that affect GC effects at the molecular level on an organism's defense against inflammation and microbial infections. Pathogenic infections with bacteria and viruses trigger both the innate and adaptive arms of the immune system and elicit secretion of inflammatory mediators. Upon infection, HPA axis activation culminates in adrenal GC secretion, the role of which is to desensitize inflammation and to avoid an overshooting immune response, which may be detrimental for cells and tissues and eventually the infected organism (16). When microbes enter the host, different signaling cascades, such as the Toll-like receptor (TLR) signaling cascade, are activated to induce proinflammatory responses. GR interferes with numerous components of these cascades via PPIs that can result in mutual regulation. Many pathological changes associated with infections can be reproduced in vivo by treatment of experimental animal models or by stimulation of in vitro cell systems with cytokines such as interleukin-2 (IL-2), IL-4, or tumor necrosis factor (TNF), which activate proinflammatory responses even in the absence of microorganisms. Due to their potent anti-inflammatory properties, GCs have been used in combinatorial treatments with antibiotics to cure or to ameliorate disease conditions of patients with extensive infectious diseases. In clinical trials, patients with bacterial meningitis, tuberculous meningitis, tuberculous pericarditis, severe typhoid fever, tetanus, or pneumocystis pneumonia with moderate to severe hypoxia, combinatorial treatment with GCs improved patients' survival. For patients with bacterial arthritis, GCs were also beneficial and reduced long-term disability. Moreover, in a number of other infectious conditions, including infectious mononucleosis, pneumococcal pneumonia, and pulmonary tuberculosis, GC administration moderates the disease's symptoms (17).

We discuss the interactome of GR in the cytoplasm, the plasma membrane, and the nucleus as well as the mechanisms by which these interactions modulate GR activity. We also pose questions about the GR interactome that require further investigation, the answers to which could optimize the efficacy of GCs as a treatment against inflammatory and infectious disease states.

THE GLUCOCORTICOID RECEPTOR PROTEIN

The human GR (hGR) protein is organized in three major domains with distinct functions. The N-terminal domain (NTD) of GR consists of the first 421 amino acids of the protein and contains the constitutively active ligand-independent activation function 1 (AF-1), which is essential for maximal transcriptional activation of the receptor (18). AF-1 acts as a docking site for interactions with coregulators and the basal transcription machinery. The DNA binding domain (DBD) of the GR (residues 422 to 486) consists of two highly conserved zinc fingers. This domain mediates the binding of the receptor to the major and minor grooves of the DNA as well as PPIs. Importantly, the DBD contains the amino acids that drive GR homodimerization. The DBD is separated from the C-terminally located ligand binding domain (LBD) by the hinge region (residues 487 to 526), which confers flexibility to GR dimers and facilitates GR DNA binding (19). The LBD consists of about 12 helices that fold overall into a globular structure to form a central pocket, the ligand binding site (20). The LBD (residues 527 to 777) comprises the ligand-dependent activation function 2 (AF-2), which assists in the interaction with cochaperone proteins, coregulators, and other transcription factors via their LXXLL motifs (21). In both the DBD and LBD, nuclear localization signals (NLSs) mediate GR nuclear translocation, while the DBD acts as the nuclear export signal (NES) (22) (Fig. 1).

FIG 1.

Structure of the hGRα protein and amino acid residues subject to posttranslational modifications. (Top) GR domains that mediate protein-protein interactions (PPIs), dimerization, and interaction with Hsps or that contain the nuclear localization signals (NLS) and the activation function (AF) domains. (Bottom) Structure of the human GRα protein with different domains (black numbers represent amino acids) and residues that undergo posttranslational modifications, including phosphorylation (black), sumoylation (green), ubiquitinylation (blue), and acetylation (orange). S, serine; T, threonine; K, lysine; NTD, N-terminal domain; DBD, DNA binding domain; HR, hinge region; LBD, ligand binding domain.

Specific residues throughout the GR protein have been described as targets of modifying enzymes to mediate posttranslational modifications of the receptor, influencing its actions. These modifications, including phosphorylation, acetylation, and ubiquitination, are dependent on specific residues (Fig. 1), and their importance for GR's actions are discussed below.

GR Isoforms Resulting from Alternative mRNA Splicing

The human GR gene is located on chromosome 5q31-32 and is composed of nine exons. Alternative splicing near the end of the primary transcript generates two receptor isoforms, GRα and GRβ, which differ at the ends of their C termini (23). The classic GRα protein is produced by joining the end of exon 8 to exon 9α (Fig. 2). The splice variant GRβ results from an alternative acceptor site so that exon 8 is joined to the more downstream exon 9β. The resulting proteins are identical through amino acid 727, but thereafter, GRα contains an additional 50 amino acids, and GRβ contains an additional nonhomologous 15 amino acids (24). The GRα-specific sequences encode helices 11 and 12 of the LBD, a region crucial not only for GC binding but also for coregulator recruitment by AF-2. On the other hand, GRβ's unique sequence gives this isoform several distinct properties (23, 24). Although GRβ does not bind classic GCs, it can bind the GRα antagonist RU486. As such, GRβ resides constitutively in the nucleus and is transcriptionally active (25). Genome-wide analyses have proven that hGRβ can directly induce or repress many genes that are controlled by GRα. The ability of GRβ to constitutively induce histone deacetylation may account for the repression of certain genes. Moreover, antagonism between the two isoforms has been described, since when coexpressed with GRα, GRβ acts as a dominant negative inhibitor of GRα, inhibiting its activity (26–28). The ability of GRβ to inhibit the transcriptional activity of GRα indicates that alterations in the expression levels of this splice variant modulate cellular sensitivity to GCs. GRβ is widely expressed but is generally found at much lower levels than GRα. It has been shown that exposure of cells to proinflammatory cytokines and other immune activators can selectively increase the expression of GRβ, leading to glucocorticoid resistance (29). Conversely, agents that increase GRα expression at the expense of GRβ sensitize cells to GCs (30). Some patients with glucocorticoid-resistant forms of asthma (31), rheumatoid arthritis (32), ulcerative colitis (33), systemic lupus erythematosus (34), acute lymphoblastic leukemia, or chronic lymphocytic leukemia (35) present with elevated levels of GRβ. Recent reports describe the discovery of GRβ in several animal species other than humans. GRβ isoforms in zebrafish (zGRβ) (36) and mouse (mGRβ) (37) arise from distinct mechanisms of alternative splicing, but they are strikingly similar in structure and function to human GRβ.

FIG 2.

GR isoforms resulting from alternative splicing. (A) Genomic structure of the hGR gene. (B) Alternative hGR mRNA splicing gives rise to different GR isoforms. Black triangles depict deletions. The gray rectangle with the letter I represents an intron retained in this particular isoform. (C) Structure of hGRα protein. The colors of the exons in panels A and B correspond to the respective GR domains that they encode in panel C.

Several other GR isoforms are generated by alternative splicing and affect glucocorticoid signaling. GRγ contains an insertion of a single arginine residue between the two zinc fingers of the DBD and originates from the use of an alternative splice donor site in the intron separating exons 3 and 4 (38). This widely expressed isoform binds GCs and DNA with a capacity similar to that of GRα but with a distinct gene regulatory ability. GRγ expression is associated with GC resistance in small-cell lung carcinoma cells (39), corticotroph adenomas, and childhood acute lymphoblastic leukemia (35, 40). Recently, it was shown that this arginine insertion causes conformational changes to the GR molecule, influencing the GR activity toward its target genes by affecting the selection of binding sites and downstream events (41). GR-A and GR-P lack a hormone binding ability and were initially discovered in glucocorticoid-resistant multiple-myeloma cells (42). GR-A, produced by alternative splicing linking the end of exon 4 to the beginning of exon 8, is missing the N-terminal half of the LBD encoded by exons 5 to 7. GR-P is missing exons 8 and 9, which encode the C-terminal half of the LBD, because no splicing occurs at the boundary between exons 7 and 8 (43, 44). Not much more is known about GR-A, but GR-P has been detected in many tissues and appears to be the predominant receptor variant in several glucocorticoid-insensitive cancer cells (42–44). hGRΔ313–338 is expressed in several tissues, such as lung, subcutaneous adipose tissue, liver, skin, heart muscle, and hippocampus. As the deleted region contains putative phosphorylation sites, it was suggested that this variant may have a differential GC-induced transactivation capacity (45). hGR-S1 has an early termination site because the last intron that contains a stop codon is preserved. Hence, this splice variant generates a truncated protein of 745 amino acids that has less transactivation ability than that of classical hGRα, probably due to insufficient ligand binding (46). Little is also known about the recently identified isoforms hGR-NS1 and hGR-DL1. However, the activity of hGR-NS1 is at least twice that of hGRα, whereas the activity of GR-DL1 is only 10% of the activity of GRα (47). No link between these three isoforms and inflammatory disorders or GC resistance has been reported (Fig. 2).

PROTEIN-PROTEIN INTERACTIONS OF GR IN THE CYTOPLASM

The GR Chaperone Complex

Before ligand binding, GR is sequestered in the cytoplasm in a complex with chaperone proteins that keep it in its inactive form and in a conformation that enables recognition of the ligand by the LBD (48).

Among the first accessory proteins found to interact with GR in the cytoplasm are Hsp90 and Hsp70. These proteins act in an ATP-dependent manner through cleavage of ATP to ADP and P_i. Hsp70 is the first cochaperone that recognizes and binds newly synthesized GR, and this interaction is enhanced by Hsp40, which binds ATP-bound Hsp70 (49). Hsp40 serves as the platform for the binding of Hsp90 dimers to the Hsp70-GR complex and for the assembly of these three cochaperone proteins into the so-called foldosome (50). The foldosome directs the assembly of the complex of GR and Hsp90 and refolds the receptor to the steroid binding state by acting as a self-sufficient protein folding machine (51). The formation of the foldosome is aided by the tetratricopeptide repeat (TPR)-containing protein Hop (Hsp70-Hsp90-organizing protein), which can simultaneously bind Hsp90 and Hsp70. Another molecule crucial for GR maturation and ligand association is immunophilin p23. This cochaperone binds dynamically to the ATP-dependent conformation of Hsp90 to stabilize the GR-Hsp90 complex in its steroid binding form (52). Furthermore, it was shown that GR in histone deacetylase 6 (HDAC6)-deficient cells is defective in ligand binding, nuclear translocation, and gene activation because the absence of HDAC6 leads to hyperacetylation of Hsp90 in the cytosol, resulting in the assembly of a dynamic yet unstable GR-Hsp90 heterocomplex and reduced binding of p23 (53). Thus, the interaction of HDAC6 with the GR-Hsp90 complex is crucial for the correct formation of the GR heterocomplex in the cytoplasm (54) (Fig. 3).

FIG 3.

Assembly of the GR chaperone complex. Binding of Hsp40 and Hsp70 to GR is the first step for ligand-receptive folding and conformation of GR. Conversion of Hsp70 to an ATP-dependent conformation permits binding of Hop. Subsequent binding of Hsp90 completes the assembly of the foldosome. Binding of HDAC6 to Hsp90 is necessary in order to deacetylate Hsp90 so that it forms a stable complex with GR. Conversion to ATP-bound Hsp90 leads to the release of Hop, Hsp40, and Hsp70 and the recruitment of p23 and immunophilins (FKB51), contributing to the maturation of the receptor to a high-affinity conformation. Binding to GCs is followed by an immunophilin switch (FKBP52).

Other proteins escorting GR in the cytoplasm are BAG family molecular chaperone regulator 1 (BAG-1) and Hsp70-interacting protein known as Hip (or ST13). BAG-1 reduces the steroid binding activity of GR, but cotransfection of Hip in COS cells reverted this negative effect, thereby indirectly assisting GR assembly (55). Later, it was shown that the introduction of human Hip into Saccharomyces cerevisiae enhances the hormone-dependent activation of a reporter gene by GR. Hip promotes the functional maturation of GR without increasing steady-state levels of the GR protein (56). CHIP (or STUB1) is a TPR-containing protein with negative effects on GR's steroid binding and transactivation potential. CHIP interacts, via a set of tetratricopeptide repeat motifs, with Hsp70 and Hsp90 and induces the ubiquitination of GR and degradation by the proteasome (57). CHIP might not have a substrate binding site but ubiquitinates unfolded proteins that are bound to Hsp70. Thus, it is the heterodimer CHIP-Hsp70 that acts as a ubiquitin ligase (58).

In the presence of ATP, Hop is released from the GR complex, permitting the immunophilins FKBP52, FKBP51, cyclophilin-40 (Cyp-40) (59), and phosphatase PP5 to anchor to the GR complex. FKBP51 (encoded by the Fkbp5 gene) binds unliganded GR and is thought to reduce GR activity. Upon ligand binding, FKBP51 is replaced by FKBP52 (encoded by the Fkbp4 gene), which activates the GR complex (60). FKBP51 has also been clinically associated with GC resistance in autoimmune diseases. Specifically, in sputum samples from chronic obstructive pulmonary disease (COPD) patients treated with formoterol-budesonide or formoterol-budesonide-theophylline for 4 weeks, the expression level of FKBP51 was higher in the second treatment group (57%), indicating that competition between immunophilins for binding the GR-Hsp90 complex might alter the response to GCs (61). Moreover, in a study focused on the identification of markers of epithelial cell dysfunction in asthma, it was found that GC treatment markedly upregulated the expression of FKBP51, which was associated with a poor response to corticosteroids (62). Although FKBP51 inhibits GR actions, it is induced by activated GR (63), pointing to an autoregulatory mechanism through a negative-feedback loop that GR engages to restrict its own actions.

Other GR Interactors in the Cytoplasm

Apart from the classical GR complex that is essential for GR protein maturation, many cytoplasmic proteins can interact with GR and alter its activity. The 14-3-3 family of proteins is a highly conserved family in all eukaryotic cells. These proteins preferentially bind phosphorylated serine/threonine residues of partner molecules and influence many signal transduction events by altering the subcellular localization of these partner molecules and/or protecting them from proteolysis. They recognize phosphorylated residues in the amino acid sequence in the “high-affinity 14-3-3 binding motif” RSXpSXP. 14-3-3 proteins contain one classic NES in their ninth helix, which helps to localize 14-3-3/partner protein complexes in the cytoplasm. GR has been shown to interact with 14-3-3σ, 14-3-3η, and 14-3-3ζ/δ. T. Kino et al. demonstrated that 14-3-3σ interacts through its NES-containing C terminus with the LBD of GRα but does not bind GRβ. In human colon carcinoma-derived (HCT116) cells, 14-3-3σ functions as a negative regulator in the glucocorticoid signaling pathway by shifting the subcellular localization/circulation of the receptor toward the cytoplasm through its NES. Deletion of the NES domain abolishes 14-3-3-mediated GR cytoplasmic retention (64). Conversely, overexpression of 14-3-3η enhances the transcriptional activity of GR as well as its protein levels in both the cytoplasm and nucleus in response to GCs. The interaction of the GR LBD with the C terminus of 14-3-3η is enhanced by glucocorticoid agonists but inhibited by the antagonist RU486 (65). Of note, 14-3-3η inhibited the ligand-induced downregulation of GR. Although ligand-activated degradation of GR occurs via the ubiquitin-proteasomal degradation pathway, proteasomal inhibition did not induce synergy with the 14-3-3-dependent increase in the GR level in response to GCs, and inhibition of translation did not block elevation of GR levels by 14-3-3η, indicating that 14-3-3η induces the stabilization of GR (66). 14-3-3ζ/δ interacts with GR in a mechanism that cells use to integrate cellular stress-mediated signaling pathways. Hormone-independent human GR phosphorylation at serine 134 (S134) is induced by different stress-activating stimuli in a manner that depends on the p38 mitogen-activated protein kinase (MAPK), and this phosphorylation is required for the interaction between GR and 14-3-3ζ/δ. Under oxidative stress (H₂O₂ treatment), S134 is hyperphosphorylated, resulting in reduced corecruitment of 14-3-3ζ/δ with GR to Igfbp1 and Lad1 promoters, but it does not influence its recruitment to the GC-induced leucine zipper (Gilz) promoter. These observations reveal a mechanism of gene regulation control before ligand binding (67).

THE GR INTERACTOME IN THE NUCLEUS

As a transcription factor, GR influences the expression levels of a number of genes involved in inflammation. In the nucleus, GR binds to promoter regions of GR-dependent genes (mainly transactivation) and tethers to DNA-bound transcription factors, altering their signaling pathways (mainly transrepression). Conversely, these interactions can affect GR activity. In this section, we focus on the interactions of the GR with transcription factors involved in inflammatory processes and the cross talk between them as well as on the interactions with cofactors that enable GR to perform its functions.

GR CROSS TALK WITH THE TLR SIGNALING PATHWAY

Pathogens of quite different biochemical compositions and with entirely different life cycles, including viruses, bacteria, fungi, and protozoa, are recognized by different receptors after crossing the host's primary defense borders. This recognition activates the innate and adaptive immune responses (68). The conserved components of pathogens that are recognized by the hosts' receptors are commonly termed pathogen-associated molecular patterns (PAMPs) (69). The best-studied PAMP receptors that initiate immune responses belong to the family of Toll-like receptors (TLRs). TLRs are transmembrane proteins expressed on the surface of antigen-presenting cells (APCs), including macrophages, dendritic cells (DCs), and B lymphocytes (70). Different TLRs recognize different PAMPs, and it has been shown that Gram-negative bacteria are recognized by TLR4 via the lipid A portion of lipopolysaccharide (LPS) (71), whereas TLR2 detects lipoteichoic acid, lipoproteins, and peptidoglycan of Gram-positive bacteria (72). TLR4 also recognizes the fusion protein of respiratory syncytial virus (RSV) and cellular Hsps. However, flagellin, the major constituent of the motility apparatus of flagellated bacteria, is recognized by TLR5 (73, 74). TLR3 recognizes double-stranded RNA (dsRNA) produced during viral replication (75), whereas TLR7 and TLR8 are activated by single-stranded RNA (ssRNA) (76, 77). TLR9 detects unmethylated CpG DNA in the genomes of viruses and bacteria (78). Engagement of a TLR by a specific PAMP triggers downstream signaling pathways, ultimately leading to the initiation of an antimicrobial proinflammatory response involving the production of proinflammatory cytokines, chemokines, cell adhesion molecules, and immunoreceptors, which together orchestrate the early host response to infection. The TLR-mediated response depends on three signaling pathways: (i) NF-κB and AP-1, (ii) MAPKs, and (iii) interferon (IFN) regulatory factors (IRFs) (69, 79) (Fig. 4). NF-κB and MAPKs are crucial for the induction of proinflammatory responses, while IRFs are essential for the stimulation of IFN production. Insights into the mechanism of GC-mediated interference of GR with microbial infections and TLR signaling showed that Dex inhibits TLR signaling induced by Neisseria meningitides and Streptococcus pneumoniae by targeting the NF-κB pathway at several levels, including inhibition of IκB kinase (IKK)-mediated IκB phosphorylation and NF-κB DNA binding (80). As discussed below, to mediate the anti-inflammatory actions of GCs, GR has been reported to interfere with the three above-mentioned key points of TLR signaling via its various PPIs.

FIG 4.

GR cross talk with the TLR signaling pathway. GR interferes with the three main mediators of TLR signaling, MAPKs, NF-κB, and IRF3. GR decreases the activity of NF-κB and AP-1 by interacting with p38 (1) and JNK (2), respectively. Additionally, in the nucleus, GR interacts with AP-1 via tethering, thereby suppressing the transcription of its target genes (3). GR positively regulates the expression of IκBα (4), which keeps NF-κB in its inactive form in the cytoplasm. The mechanisms employed by GR to suppress NF-κB via PPIs include the interaction of GR with PKAc (5), reducing its availability to phosphorylate NF-κB, which is crucial for its activity; direct interaction of GR with NF-κB in the cytoplasm, leading to sequestration (6); or interaction in the nucleus via sequestration or tethering (7), which is assisted by the deacetylation of GR by HDAC2. Interaction of GR with the p65 subunit of NF-κB or with GRIP1 reduces their availability for an interaction with IRF3. Consequently, the transcription of inflammatory genes that depends on the cooperative binding of both NF-κB and IRF3 or of GRIP1 and IRF3 in their promoters is decreased (8). Ub, ubiquitin.

Extensive research has focused on the interference of GR with the signaling pathways of the transcription factors NF-κB and AP-1, which are important regulators of immunity and inflammation. Many genes involved in immune and inflammatory responses in the cell, particularly proinflammatory cytokines and cell adhesion molecules, are counteracted by GCs, whereas they are induced by NF-κB under inflammatory conditions. In addition, NF-κB has been shown to decrease GR transactivation, leading to the conclusion that NF-κB and GR are physiological antagonists. For this reason, the interaction of NF-κB with GR in immunity and inflammation is a mainstay area of research.

Interaction of GR with NF-κB

NF-κB is a potent proinflammatory transcription factor activated by many extracellular signals, including viruses (e.g., herpes simplex virus and adenovirus) and viral proteins, bacterial products (e.g., LPS), inflammatory cytokines (TNF, IL-1, and IL-2), and various DNA-damaging agents and oxidative stressors (81). In response to these inflammatory stimuli, different NF-κB-responsive genes are activated, including cytokines, chemokines, cell adhesion molecules, immunoreceptors, and growth factors. NF-κB consists of subunits from the Rel family proteins (p65 [RelA], RelB, c-Rel, p50/p105, and p100/p52) and functions as a dimeric DNA binding protein. However, the best-studied mechanism of NF-κB signaling is the one mediated by the prototypical p65/p50 heterodimer, which becomes active mainly through the actions of IKKs but also through its direct phosphorylation by the protein kinase A (PKA) catalytic subunit (PKAc) (82, 83). Here we discuss the latter kinase because of its interaction with GR.

Many different mechanisms have been described for the cross talk between the GR and NF-κB signaling pathways, including cross-coupling of GR with PKAc, GR-mediated reduced activity of IκB kinase, interaction with the p65 subunit, IκBα upregulation and cytoplasmic sequestration of p65, GR-mediated HDAC2 recruitment, interference with the phosphorylation status of RNA polymerase II (Pol II), and competition for binding of cofactors (Fig. 4).

In 1994, Ray and Prefontaine provided the first piece of evidence that the interaction between GR and NF-κB can act as a mechanism of mutual inhibition of these two signaling pathways and that this interaction is mediated by the Rel homology domain of p65 and the DBD of GR (84). The initial hypothesis was that these two factors impaired each other's DNA binding functions and therefore blocked the ability to activate transcription. In support of this notion, it was shown that Dex suppresses TNF- and IL-1β-mediated activation of NF-κB at the level of DNA binding, as demonstrated by electrophoretic mobility shift assays (EMSAs), as well as at the transcriptional level. Interestingly, in COS cells treated with Dex for 72 h, cotransfection of GR caused a significant reduction of transfected p65 but not of transfected p50 or c-Rel subunits (85).

The Dex-mediated downregulation of IL-6 in fibroblasts and epithelial cells (L929sA) was shown to be due to the repression of NF-κB at the transcriptional level by the GR, without any change in the IκBα levels, confirming a nuclear mechanism (86). Furthermore, inhibition of cytokine production by steroids via inhibition of NF-κB was tested in fibroblast-like rheumatoid synoviocytes (FLSs). EMSAs demonstrated that identical amounts of NF-κB were present in the nuclei of FLSs stimulated with TNF with or without pretreatment with Dex, indicating that GR does not interfere with the DNA binding properties of NF-κB (87). The predominant model for GR-mediated NF-κB suppression is that GR interacts with DNA-bound NF-κB, preferentially with its p65 subunit, leading to its transcriptional inhibition.

The use of coactivators by NF-κB is promoter and signal specific. Distinct sequences of κB elements determine whether IRF3 is used as a necessary coactivator of NF-κB-dependent transcription. Conversely, p65 has been shown to act as a coactivator of IRF3 in response to TLR4 activation. The physical interaction of IRF3 with p65 is impaired upon the binding of activated GR to p65. In this way, GR binding to p65 diminishes the availability of p65 as a substrate for coactivators such as IRF3, repressing the subset of NF-κB target genes that require IRF3 as a coactivator as well as IRF3 target genes that require p65 as a coactivator when activated by TLR4 (Fig. 4). In contrast, IRF3-dependent genes become GC resistant in response to TLR3 activation due to the absence of p65 (88).

A similar mechanism accounts for the promoter-specific repression of TNF-inducible genes. TNF induces the transcription of Nfkbia and IL-8 in A543 cells, but only IL-8 is sensitive to GR-mediated transrepression. This is related to the requirement for P-Tefb as a mandatory coactivator in IL-8 induction. P-Tefb is important for the phosphorylation of the C-terminal domain of Pol II. GR is tethered to both IL-8 and Nfkbia promoters in a ligand-dependent manner, but due to its interaction with p65, it blocks the binding of P-Tefb to NF-κB, repressing TNF-induced activation (89).

While the importance of the DBD of GR for the p65-GR interaction was described first, a series of GR deletion mutants tested for their ability to repress p65 proved that more than one domain is important. Deletion of either of the two zinc fingers of the DBD, as well as deletion of the LBD or large portions of the transactivation domain of GR, also abolished the repressive effect of GR on NF-κB. Interestingly, the reciprocal repression of GR by p65 was less selective, since all transcriptionally active GR mutants were repressed by p65 (90). Independent data have demonstrated that two critical amino acids (Arg488 and Lys490) in the C-terminal zinc finger of rat GR (rGR) are indispensable for GC-mediated NF-κB inhibition (91).

During GR-dependent NF-κB inhibition, GR can cooperate with the transcription factor p53. It has been shown in vivo that p53 is essential for the inhibition of NF-κB by GR and that Dex is not protective against LPS in p53 knockout (KO) mice, showing significant inhibition of GR transactivation. The proposed mechanism for this important role of p53 in the GR pathway is that p53 targets GR to NF-κB, inhibiting the binding of NF-κB to DNA. Additionally, p53 might directly target GR to the promoter region of its target genes or help to recruit the transcriptional activation machinery to the promoter regions of GR target genes or otherwise assist in the formation of the proper complexes around GR in order to positively regulate genes (92). In support of this proposed role for p53, patients with rheumatoid arthritis who did not respond to GC treatment showed reduced p53 expression levels in blood mononuclear cells (93, 94). On the other hand, it was shown that the formation of the GR/p53 complex in neuroblastoma cells drives the inhibition of p53-dependent functions, including transactivation, cell cycle arrest, and apoptosis due to the cytoplasmic sequestration of both transcription factors. GR antagonists could revert this effect, indicating that molecules that effectively disrupt the GR/p53 complex might be useful in the treatment of diseases in which p53 is sequestered in the cytoplasm by GR (95). Moreover, endogenous p53 and GR form a ligand-dependent trimeric complex with Hdm2 in the cytoplasm. Disruption of the p53-Hdm2 interaction prevents Dex-induced ubiquitination of GR and p53. These results provide a mechanistic explanation for GR and p53 acting as opposing forces in the decision of cell death or survival (96).

The possibility that the pool of available cofactors such as glucocorticoid receptor-interacting protein 1 (GRIP1) or steroid receptor coactivator 1 (SRC-1) constitutes a target of antagonism between p65 and GR has been investigated, and the findings are elaborated below in the section on GR interactions with classical cofactors.

Beside the above-mentioned nuclear modulation of NF-κB activity, GR can limit the actions of NF-κB via its rapid nongenomic effects. More specifically, it was shown that GR physically associates with p85α/phosphatidylinositol 3-kinase (PI3K) in the skin of GR-overexpressing mice, resulting in decreased Akt and IκB kinase activity, which is essential for NF-κB phosphorylation and subsequent activation. Therefore, GR/PI3K-Akt cross talk is a major mechanism of the antitumor effect of GCs in the skin as well as a nongenomic mechanism to oppose the actions of NF-κB (97).

A few studies supported the notion that the GR-mediated inhibition of NF-κB depends on cytoplasmic mechanisms. In cytoplasmic extracts of rat liver cells and H4-II-E-C3 hepatoma cells, p65, p50, and the inhibitory IκBα complex interact with GR even in the absence of GCs or inflammatory stimuli, and GCs can decrease TNF-induced p65 nuclear entry. These results show that the inhibitory effects of GCs on NF-κB are in the cytoplasmic compartment (98). Additionally, when HeLa cells were cultured in the presence of Dex, the level of nuclear p65 protein decreased after TNF induction, leading to the conclusion that the loss of p65 protein could be due to destabilization or cytoplasmic sequestration (85).

Another cytoplasmic mechanism for mutual inhibition of the GR and NF-κB pathways is based on interactions with PKAc. The interaction of GR with PKA positively regulates the actions of both proteins. In F9 embryonal carcinoma cells, which lack endogenous functional cyclic AMP (cAMP) response element binding protein (CREB), PKA can still control gene transcription through GR by upregulating its hormone-dependent transactivation (99). Through its catalytic subunit, PKA interacts with the DBD of GR, leading to an enhancement of the binding of GR to its cognate DNA response elements. In vitro, PKA phosphorylates rat liver GR independently of the presence of Hsp90 and the transformation state of the receptor (100). Additionally, the in vitro and in vivo associations of GR with PKA are indispensable for the GR-mediated inhibition of NF-κB transactivation. Apart from its positive effect on GR transactivation, PKAc is also crucial for the maximal transactivation capacity of NF-κB by mediating its cytoplasmic phosphorylation at S276. These findings localize the cross-coupling between NF-κB and GR in the cytoplasm and implicate PKA as the molecular interface for this mutual inhibition (101). Interestingly, the same study reported that phosphorylation of NF-κB at S276 is indispensable for NF-κB-mediated GR repression. Despite the repression of GR transactivation, p65 fails to block homologous GR downregulation, indicating that in this case, NF-κB does not hamper GR-DNA binding (102). Although most of the research on NF-κB/GR antagonism has focused on the role of GR in dampening NF-κB actions, the antagonism of GR function by NF-κB is physiologically equally important and should not be disregarded.

The negative effect of NF-κB on GR function may account for the development of GC resistance in many patients with inflammatory or autoimmune diseases, such as steroid-resistant asthma. It will also be important to determine whether these individuals have any abnormality in NF-κB regulation that leads to its excessive activation. A chronic high level of NF-κB activity might lead not only to chronic inflammation but also to GC resistance by blocking the GR signaling pathway.

Interaction of GR with AP-1

GR represses AP-1 via mechanisms resembling those employed for NF-κB, including direct PPIs, which also result in reciprocal transcriptional antagonism (Fig. 4). AP-1 is a transcriptional regulator involved in various aspects of proliferation and differentiation and is also involved in upregulating the expression of many cytokine genes, such as IL-1 and IL-2, which in turn can be repressed by GCs. AP-1 is a member of the basic region leucine zipper (bZIP) family of DNA binding proteins and is composed of a Jun family member (c-Jun, v-Jun, Jun-B, or Jun-D), which can dimerize with another Jun protein or with a Fos protein (c-Fos, Fos-B, Fra-1, or Fra-2). Transcriptional activation by AP-1 depends on its subunit composition. While the c-Fos/c-Jun heterocomplex is ubiquitous in cells, inflammatory cytokines have been shown to promote this dimerization via c-Jun phosphorylation by c-Jun N-terminal kinase (JNK).

GR binds c-Fos/c-Jun dimers via a sequence unique to c-Fos and strongly inhibits the DNA binding and transactivation capacities of this AP-1 dimer. Additionally, GR appears to inhibit c-Jun's DNA binding capacity and transactivation of c-Fos/c-Jun complexes 10-fold more than for Jun homodimers (103).

As mentioned above, GR can repress the transcriptional activity of NF-κB and AP-1 by tethering. Kassel et al. proposed that this tethering is mediated by a nuclear isoform of the focal adhesion LIM domain protein Trip6. As shown by chromatin immunoprecipitation (ChIP), Trip6 is recruited to the promoters of target genes together with AP-1 or NF-κB as a coactivator. In the presence of GCs, GR joins the Trip6 complex. Reducing the level of Trip6 by RNA interference or abolishing its interaction with GR eliminates the transrepression capacity of GR (104).

Regarding AP-1-mediated inhibition of GR, a possible mechanism proposed for the case of GC-resistant asthma is that AP-1, by its DNA binding to its respective response elements to activate genes encoding inflammatory cytokines, prevents DNA binding of GR by direct sequestration (105). Additionally, the c-Jun component of AP-1 can bind to GR and thereby inhibit its actions (106).

The interaction between GR and AP-1 was originally reported to result in mutual antagonism (106, 107). However, surprisingly, AP-1 has emerged as a key partner in GR-regulated transcription of endogenous target genes. AP-1 regulates basal chromatin structure and accessibility, and so it enhances GR binding to specific sites in the genome (108). This suggests that the physiological interaction between GR and AP-1 is more extensive and more complex than previously realized, with current evidences supporting either enhancement or suppression of GR activity (109).

It has been proposed that GCs also negatively regulate AP-1 through a nongenomic mechanism involving inhibition of JNK and subsequent phosphorylation of c-Jun on S63/73, thereby repressing AP-1 activity. It has also been shown that the interaction of GR with JNK contributes to this mechanism. More specifically, GCs induce the disassembly of JNK from mitogen-activated protein kinase kinase 7 (MKK7), which is essential for JNK phosphorylation/activation by promoting its association with GR through the LBD of the receptor (Fig. 4). The GR-inactive JNK complex translocates to the nucleus, loading AP-1 with inactive JNK while masking AP-1 from activated JNK generated by subsequent MAPK activation (110). The activity of JNK and the subsequent activation of AP-1 in relation to GC responsiveness have been implicated in the pathophysiology of asthma. Researchers speculate that the activation of AP-1 is increased due to excessive activation of the JNK pathway, which has been observed in steroid-resistant asthma patients (111).

Interaction of GR with MAPKs

MAPKs are highly conserved among species and control many physiological processes, including gene regulation, cell survival and apoptosis, proliferation, differentiation, and inflammatory responses. There are three main classes of MAPKs in mammals: extracellular signal-regulated kinases (ERKs), JNK, and p38. MAPKs are serine/threonine protein kinases regulated by a series of phosphorylating events, and in turn, they phosphorylate other target proteins. Inflammatory stimuli activate the Ras/Raf signaling pathway, which then activates MAPKs. Activation of MAPKs leads to the production of inflammatory cytokines such as TNF, IL-1, and IL-6. AP-1 is one of the best-characterized targets of MAPKs because JNK was first characterized as the kinase responsible for c-Jun phosphorylation/activation. NF-κB is also a target of MAPKs (p38, ERK1/2, and JNK). GR has been reported to interfere with the signaling pathways of MAPKs, and one mechanism is by physically interacting with them, which leads to mutual suppression of the respective pathways.

JNK interacts with rGR and determines its phosphorylation and sumoylation status, which modify its transcriptional activity. GR is phosphorylated at S246 by the JNK protein family in a rapid and transient manner. The levels of S246 phosphorylation of endogenous GR increased significantly in cells treated with UV radiation, which activates JNK. S246 GR phosphorylation by JNK enables subsequent GR sumoylation at lysines 297 and 313. Sumoylation is mediated by the small ubiquitin-like modifier SUMO2, and in COS-7 cells, these modifications result in decreased GR transcriptional activity, as demonstrated in GR-dependent reporter activity assays and endogenously by Igfbp and Gilz levels (112).

p38 MAPK can also contribute to the phosphorylation of hGR at S211 in lymphoid cells. However, it was shown that treatment with the p38 inhibitor SB203580 diminishes GR phosphorylation at this serine only slightly. On the other hand, phosphorylation of the hGR by p38 at an unidentified residue was associated with reduced GR ligand binding affinity and reduced GC-mediated granulocyte-macrophage colony-stimulating factor (GM-CSF) repression. Induced p38 phosphorylation led to diminished GR DNA binding and transactivation. However, independent data proved that MAPK-dependent GR phosphorylation at S211 may positively regulate GC-induced apoptosis of lymphoid cells (113). p38 and JNK also phosphorylate S226 of hGR and suppress its transcriptional activity by enhancing the nuclear export of the receptor (114).

GCs can exert their anti-inflammatory actions through the interaction of GR with MSK1 (mitogen- and stress-activated protein kinase). This group of kinases includes serine/threonine kinases that are activated downstream of ERK and p38 to regulate gene transcription. NF-κB and CREB are targets of MSK1. MSKs also regulate chromatin conformation by regulating the phosphorylation of histones. GC anti-inflammatory actions are related to the ratio of cytoplasmic MSK1 to nuclear MSK1. Dex prompts the export of MSK1 from the nucleus, whereas RU486 blocks MSK1 nuclear translocation. The shift in the ratio of nuclear MSK1 to cytoplasmic MSK1 could inhibit the activity of NF-κB, thereby transrepressing NK-κB-dependent gene expression as well as inducing a closed chromatin environment at specific inflammatory gene promoters, resulting in reduced inflammatory manifestations (115).

Furthermore, the rapid nongenomic effects of GR interfere with the MAPK transduction cascade. In the cytosolic compartment of hepatocytes from adrenalectomized rats, GR was identified in the same complex with 14-3-3 and Raf-1, and the interaction was enhanced by ligand treatment (116). The authors of that study speculated that the interaction of GR with 14-3-3 and Raf-1, which is an upstream regulator of MAPK, is a mechanism by which GCs interfere with MAPK signaling. In addition, GR directly binds to ERK in the cytoplasm. During a stressful event, activated GR interacts with ERK, leading to its increased phosphorylation, which in turn rapidly leads to the activation of MSK1. This event consequently promotes the phosphorylation/acetylation of histone H3 and the induction of c-Fos and early growth response protein (Egr-1). This nongenomic effect of GCs, specifically in dentate gyrus granule neurons, mediates an increase in ERK signaling and is involved in behavioral adaptation (117).

GR CROSS TALK WITH NUCLEAR RECEPTORS

GRα Homodimerization

Although the major anti-inflammatory effects of GCs were for a long time attributed to their transrepression properties, which inhibit NF-κB and AP-1 signaling pathways, mounting evidence from independent studies confirmed that considerable GC protection against inflammation is mediated by GR homodimerization. GR dimers bind to the respective GR response elements, which are located in the promoter regions of many anti-inflammatory genes, such as Gilz and MAPK phosphatase 1 (Mkp1) (118). The importance of GR dimerization in counteracting inflammation was shown by in vivo experiments on so-called GRdim mice, in which GR dimerization is largely defective due to the replacement of an alanine by a threonine (A548T) in the second zinc finger in the DBD of the GR (119).

A study using GRdim mice assessing antigen-induced arthritis (AIA) as a model of human rheumatoid arthritis demonstrated the impairment of GC-induced suppression of T helper 1 (Th1) and Th17 cell-derived proinflammatory cytokines. It was shown that IL-17-producing T cells are the most important targets for effective GC therapy and that GR homodimerization is indispensable for this process (120). Similarly, GR dimerization has been determined as an important mechanism for protection against septic shock caused by cecal ligation and puncture or by LPS (121). Furthermore, when the contact hypersensitivity model was assessed in GRdim mice, the persistent infiltration of macrophages and neutrophils was explained by impaired repression of inflammatory cytokines and chemokines such as IL-1β, monocyte chemoattractant protein 1, macrophage inflammatory protein 2, and IFN-γ-inducible protein 10. These findings indicate that GR homodimerization is required for GCs to exert their anti-inflammatory effects on their main targets, which are macrophages and neutrophils (122). Finally, previously reported data from our research group demonstrated the importance of GR dimerization for protection against TNF-induced inflammation. Administration of TNF in GRdim, Mkp1^−/−, and Jnk2^−/− mice revealed that JNK2 is a crucial mediator of TNF-induced inflammation and that GR dimerization inhibits JNK2 through Mkp1 and protects from TNF-induced apoptosis and lethal inflammation (123).

At this point, it is essential to discuss a few representative genes with well-documented anti-inflammatory actions that are activated by the binding of GR homodimers to their promoter regions, through so-called GR transactivation. Mkp1 (encoded by the Dusp1 gene) catalyzes dephosphorylation and mediates subsequent inactivation of MAPKs (including ERKs, JNK, and p38 MAPK), which play pivotal roles in the activation of proinflammatory transcription factors such as NF-κB and AP-1 (124–126). Moreover, Gilz (encoded by the Tsc22d3 gene) is another representative of GRE-containing genes with anti-inflammatory actions that bind Ras and Raf-1 and result in negative regulation of the MAPK signaling pathway by GR. Gilz has also been shown to bind the p65 subunit of NF-κB and both subunits of AP-1, c-Jun, and c-Fos (127, 128).

Finally, annexin-1 (lipocortin-1, encoded by the Anxa1 gene) is another GR-inducible gene involved in anti-inflammatory processes. Annexin-1 can bind the p65 subunit and inhibit NF-κB from binding to DNA (129). Additionally, it has been shown that Anxa1^−/− mice are more susceptible to AIA, bleomycin-induced lung fibrosis, and dextran sodium sulfate-induced colitis (130–132).

Heterodimerization of GRα with GRβ

In contrast to the positive effects of GRα homodimerization on GR activity, heterodimerization of GRβ with GRα exerts negative effects on the immunosuppressive actions of GR.

Human GRβ is located primarily in the nucleus of transfected cells independently of hormone administration. In the absence of hGRα, hGRβ exerts no transcriptional activity on a glucocorticoid-responsive enhancer. However, when both isoforms are expressed in the same cell, hGRβ inhibits the hormone-induced, hGRα-mediated stimulation of gene expression. Thus, hGRβ might function as a dominant negative inhibitor of hGRα activity (133). For instance, bronchiolitis patients infected with RSV showed an increase in GRβ expression correlated with disease severity (134). In addition, GRβ and GRγ potently exerted dominant negative effects on GRα-mediated transrepression through NF-κB and inflammation-related genes (135). Thus, the negative effects of GRβ and GRγ on GRα have been reported to be responsible for the GC resistance frequently observed in steroid-resistant nephrotic syndrome, rheumatoid arthritis, hematologic tumors, and other chronic inflammatory diseases under GC therapy. GRβ also interferes with GRα-mediated apoptosis in T cells. Strong constitutive expression of GRβ by human neutrophils may provide a mechanism by which these cells escape GC-induced cell death. Moreover, upregulation of GRβ by proinflammatory cytokines such as IL-8 further enhances the survival of neutrophils in the presence of GCs during inflammation (136).

GR Cross Talk with Steroidal and Nonsteroidal Nuclear Receptor Signaling Pathways

GR may interfere with the signaling pathways of more nuclear receptors (NRs), often resulting in mutual regulation. In this section, we discuss briefly the cross talk of GR with other NRs and its functional effect on the receptors' actions.

Peroxisome proliferator-activated receptor γ (PPARγ) and GR cooperate in diminishing chemokine expression under inflammatory conditions. More specifically, prostaglandin 15d-PGJ2 induces the interaction of PPARγ with GR in TNF-treated human airway smooth muscle (ASM) cells, and their interaction was further enhanced by fluticasone (a synthetic GR agonist) and salmeterol (β2 agonist) stimulation. Combinatorial treatment with prostaglandin 15d-PGJ2, fluticasone, and salmeterol inhibits TNF-induced histone H4 acetylation at the eotaxin promoter and NF-κB p65 binding to the eotaxin promoter while inducing PPARγ and GR associations on the eotaxin promoter, as analyzed by ChIP. Hence, it has been suggested that inhibition of chemokine expression by PPAR agonists, GCs, and β2 agonists, and particularly the expression of eotaxin, may be dependent on the interaction between GR and PPARγ proteins (137). In line with this, pioglitazone administration also mediates the association of PPARγ with GR, supporting the healing of chronic gastric ulcer in rats, as indicated by the suppressed TNF and IL-1β gene expression in gastric mucosa (138). Finally, PPARα also interacts with GR, and they cooperatively act as immunosuppressors. Simultaneous activation of PPARα and GR dose-dependently enhances transrepression of NF-κB-driven gene expression and additively represses cytokine production. However, PPARα agonists inhibit the expression of classical GRE-driven genes in a PPARα-dependent manner by interfering with the recruitment of GR and RNA polymerase II, as shown in vitro for the Gilz promoter (139).

Investigation of the cross talk between GR and liver X receptor (LXR) showed both synergistic and opposing effects. The synergy between GR and LXR apparently involves a role for PPARγ as well. Combinatorial treatment with agonists for GR, PPARγ, and LXRs resulted in additive or synergistic inhibition of a subset of LPS-induced TLR4 target genes both in cultured macrophages and in vivo (88). However, with regard to the impact on GRE-driven transcription, it was reported that LXR binding to GR resulted in the inhibition of the activity of a GR-dependent reporter in HCT116 cells, although the transrepression properties of GR remained unaffected (140).

Furthermore, retinoic acid receptors (RAR) interact with GR and enhance the GC-induced death of T cell lines and mouse thymocytes both in vitro and in vivo. This mechanism involves the binding of the liganded RARα/retinoid X receptor (RXR) with the liganded GR, resulting in enhanced transcriptional activity of the GR without DNA binding of retinoic receptors being a prerequisite (141). Regarding the functional effect of the interaction of GR with progesterone receptor (PR), the evidence is more limited. PR and GR were shown to interact in vitro (142), and in the context of inflammation, reports claim that progesterone acts via GR to repress IL-1β-driven COX-2 activation (143). The androgen receptor (AR) and the GR interact at the transcriptional level, forming heterodimers at a common DNA site both in vitro and in vivo, and this interaction leads to mutual inhibition of transcriptional activity (144). Additionally, the physical interaction of GR with farnesoid X receptor (FXR), involved in metabolic pathways, has been documented (145, 146). FXR was recently reported to exert its anti-inflammatory actions by antagonizing NF-κB signaling (147); nevertheless, it is unknown whether its interaction with GR is involved in this process. Regarding the interaction of the estrogen receptor (ER) with the GR, Bolt et al. found that the two receptors can interact, while the recruitment of MED14, SRC-2, and SRC-3 in the same complex is required for GR binding. The outcome of this interaction can be either cooperation or antagonism on E2-regulated genes in MCF-7 cells, underlying a gene-specific mechanism (148). GR and mineralocorticoid receptor (MR) share considerable structural and functional homology, and their physical interaction can lead to transcriptional suppression of GC-responsive genes. Binding of MR to GR counteracts GC-induced apoptosis by inhibiting the proteolytic process of pro-caspase-3, -8, and -9 in a B-lymphoma cell line (149). In the context of neuronal inflammation, these two NRs can differentially regulate NF-κB activation and neuroinflammatory parameters. MR promotes a neuroinflammatory response mediated by NF-κB activation, which can be blocked by GR activation, albeit it is not specified in this case whether GR suppresses MR function by the formation of heterodimers (150). The action of the orphan receptor Nur77 in mediating corticotropin-releasing hormone (CRH)-dependent induction of Pomc transcription is antagonized by GR at two levels. First, GCs partly blunt the CRH induction of Nur77 mRNA, and second, through PPI, GR antagonizes Nur77-dependent transcription on the Nur77 response element of the Pomc gene (151) via a tethering mechanism (152). Antagonism between GR and Nur77 has been observed in both endocrine and lymphoid cells (153). Finally DAX-1 (dosage-sensitive sex reversal-adrenal hypoplasia congenital critical region on the X chromosome, gene 1) is an orphan nuclear receptor that acts as a corepressor of the transactivation of GR, as shown by the decreased levels of Gilz and Fkbp5 upon DAX-1 overexpression. The mechanism that DAX-1 utilizes to regulate GR activity involves competition of coactivators such as GRIP1 for binding to the receptor (154).

GR INTERFERENCE WITH THE STAT SIGNALING PATHWAY

Cytokines produced in response to an inflammatory trigger such as infection coordinate the activities of many immune cell types; this culminates in innate and adaptive immune responses. Most cytokines act through the Janus kinase/signal transducer and activator of transcription (Jak/STAT) signaling pathway, which leads to the activation of STAT proteins. Inflammation is one of the earliest biological functions associated with STAT proteins, from the antiviral functions of STAT1 to the polarized Th cell responses that require STAT4 and STAT6. There are seven mammalian STATs: STAT1, STAT2, STAT3, STAT4, STAT5a, STAT5b, and STAT6. STATs are important players in the regulation of innate and acquired immunity. More specifically, as STAT1 mediates the effects of IFN-γ, it can be considered a major transcription factor driving macrophages and Th1 cell responses. STAT3 is activated by IL-6 and by IL-10, a potent anti-inflammatory and immunosuppressive cytokine. Consequently, deletion of STAT3 in myeloid cells results in hyperactivation of macrophages, increased inflammation, and inflammatory bowel disease. In line with an anti-infectious role for STAT3, when activated by type I IFNs, STAT3 has been shown to mediate antiviral responses (155, 156). STAT4 is activated by IL-12, IL-23, and IFN-α and promotes Th1 cell differentiation. In contrast, STAT6 is activated by IL-4 and IL-13 and plays a key role in Th2 cell responses. Finally, STAT5a and -b are important in mediating proliferative responses and are activated by IL-2, IL-4, and IL-15. GR has been shown to interfere with the STAT signaling pathway under inflammatory conditions by interacting with STAT3 and STAT5, thereby influencing immune responses (Fig. 5). There is also evidence for transcriptional cooperation between GR and STAT6, but no PPI between them has been described yet. Apart from direct PPIs with the STAT family, GR also interacts with suppressor of cytokine signaling 1 (SOCS1), which modulates the early phase of cross talk between GR and cytokine signaling. The interaction is dependent on the SH2 domain of SOCS1 and the LBD of GR. Furthermore, GC stimulation was found to increase the mRNA levels of SOCS1, and SOCS1 binding to GR did not require ligand binding of the receptor. However, the interaction was abolished after long-term GC stimulation, suggesting a functional role for this PPI in the early phase of GC action. This interaction negatively influences the transcription of two GR-regulated genes, Fkbp5 and Dusp1, as the GC-dependent expression of these genes was inhibited by IFN-γ, which is a SOCS1 inducer, while their levels were enhanced in SOCS1-deficient murine embryonic fibroblasts (157).

FIG 5.

Cross talk of GR with the STAT signaling pathway. Binding of GR to DNA-bound STAT3 suppresses the transcriptional activity of STAT3. However, STAT3 acts as a coactivator for DNA-bound GR, enabling the recruitment of additional coactivators. Simultaneous binding of GR and STAT3 on composite elements leads to synergism. The synergism between STAT5 and GR enhances the expression of TLR2. STAT5 binding to DNA-bound GR may lead to either transcriptional repression or activation of GR target genes. GR enhances the transcriptional activity of STAT5 on the β-casein promoter.

Interaction of GR with STAT3

STAT3 stimulates the transcriptional activity of AR, GR, and ER in a hormone-dependent manner. Reporter assays in CV-1 cells demonstrated that STAT3 acts synergistically with other coactivators, such as SRC-1, p300/CREB binding protein (CBP)-associated factor (pCAF), CBP, and GRIP1, to increase the transcriptional activity of these hormone receptors (158).

In the late 1990s, it was shown that IL-6 has a synergistic effect on glucocorticoid signaling. The data demonstrated that IL-6-activated STAT3 associates with ligand-bound GR to form a transactivating/signaling complex, which can function through either an IL-6 response element or a GRE. These findings reveal a new level of interaction between these two crucial signaling cascades and indicate that activated STAT3 can also act as a transcriptional coactivator without directly associating with its DNA binding motif (159). The interaction of GR with STAT3 has also been described as important in the regulation of haptoglobin, one of the acute-phase plasma proteins that provide protective functions during tissue injury and inflammation by binding the hemoglobin released from erythrocytes and thereby inhibiting its oxidant activity. This mechanism for the increased transcriptional activity of the haptoglobin gene involves the binding of GR to the GRE in the promoter of haptoglobin, in the same complex with STAT3, CEBPβ, and pCAF which cooperatively lead to increased chromatin accessibility via histone acetylation (160).

In a recent genome-wide prevalence study of tethering between GR and STAT3, Langlais et al. observed extensive cooperative genomic recruitment of GR and STAT3. They found that GR tethering to DNA-bound STAT3 represses transcription, whereas STAT3 tethering to GR results in synergy. Furthermore, different patterns of GR and STAT3 corecruitment to regulatory modules, including neighboring and composite binding sites, result in transcriptional synergy. These results provide a genome-wide appraisal of transcriptional regulation by tethering and a molecular basis for the integration of signals mediated by GR and STATs in health and disease (161) (Fig. 5).

Interaction of GR with STAT5

GR acts as a transcriptional coactivator for STAT5 and enhances STAT5-dependent transcription. More specifically, it was shown that prolactin and GCs, through their respective downstream effectors, STAT5 and GR, act synergistically to activate the transcriptional activity of beta-casein in the absence of a GRE (162). However, the formation of the STAT5/GR complex diminishes the response of a GRE-containing promoter upon IL-2 induction (163).

The synergistic effect between STAT5 and GR was further assisted with the binding of coactivators, i.e., CCAAT/enhancer binding proteins (C/EBPs) (164). Wyszomierski et al. further elucidated the synergistic mechanism of GR/STAT5-mediated upregulation of β-casein gene transcription and demonstrated that GR enhances STAT5 DNA binding by modulating the rate of STAT5 dephosphorylation (165). STAT5 has been found to be essential for GR to exert its actions in growth processes. The interaction in this case is mediated through the N-terminal tetramerization domain of STAT5b (166). Owing to the potent anti-inflammatory actions of GCs, quite intriguing was the finding that Dex can upregulate the expression levels of TLR2 upon Haemophilus influenzae infection, which is an important human bacterial pathogen causing otitis media and chronic obstructive pulmonary diseases (167). Although the possible mechanism in that study was the negative cross talk of GR with the p38 MAPK pathway, independent data that surfaced later attributed the Dex-dependent increased TLR2 expression to the synergism of GR with STAT5. More specifically, it was shown that TLR2 is cooperatively induced by TNF and Dex through GR and STAT5 recruitment at the TLR2 promoter in vivo, whereas Dex alone recruits GR but does not induce TLR2 expression (168) (Fig. 5).

The binding of GR to STAT5 creates a regulatory pathway for zinc accumulation in rat AR42J pancreatic acinar cells by modulating the expression of zinc transporters. Dex administration leads to an increase in the expression of ZnT2, which participates in zinc transport into pancreatic zymogen granules, and this mechanism requires the interaction of GR with STAT5 (169). In the presence of adequate zinc levels, NF-κB activity and the subsequent generation of cytokines, including IL-1β, IL-2, IL-6, and TNF, are partially downregulated, and the proinflammatory response is controlled (170, 171). In the view of the high global prevalence rates of zinc deficiency and chronic diseases, the interplay of zinc and inflammation in relation to GR and STAT5 warrants further examination.

CLASSICAL COFACTORS DETERMINE THE ACTIONS OF GR

By the 1990s, it became evident that NRs require additional factors to modulate gene expression. The discovery of NR coregulators, coactivators, and corepressors permitted a quantum leap in the elucidation of NR mechanisms of gene regulation. The interchange of corepressor and coactivator complexes with opposite chromatin-modifying enzyme activity is a possible mechanism for the ligand-induced switch of NRs between transcriptional repression and activation. Furthermore, both the type of ligand (agonist or antagonist) and the response element influence the three-dimensional structure of the LBD and thereby determine the binding of coactivators or corepressors (172–174).

GR Interaction with Classical Coactivators

Although there are numerous coactivators, extensive research has been performed mostly on the p160 family, commonly referred to as SRC-1, SRC-2, and SRC-3. A nomenclature group has codified the names of these coactivators NCoA1 to -3.

NCoA1 (or SRC-1) was first discovered by Onate et al. as a coactivator for PR (175). Next, GRIP1 (or NCoA2) was reported (176). The third p160 protein, reported by several research groups, is most often referred to as SRC-3 (NCoA3, AIB1, or RAC3) (177). Although less extensively studied, ASC-2 (NCoA6) has also been shown to be a bona fide coactivator of GR in CV-1 cells, with possible implications for the development of breast cancer (178).

Each p160 protein contains two highly conserved regions. In the center of the protein is a cluster of three leucine-X-X-leucine-leucine (LXXLL) motifs, in which X denotes any amino acid (179, 180). These are also referred to as NR boxes. The motifs are the sites responsible for the interaction with NRs, since mutations of these sites abrogate NR activation functions (181). p160s also contain a domain that binds to histone acetylases, e.g., CBP, which remodel chromatin by acetylating specific lysines in histones (182), e.g., p300 and pCAF (183).

The interaction of ligand-bound GR in the nucleus with coactivator molecules such as CBP, pCAF, or SRCs, which confer histone acetyltransferase (HAT) activity, causes the acetylation of lysines on histone H4 or H3, leading to the activation of genes encoding anti-inflammatory proteins, including secretory leukoprotease inhibitor (Slpi), Mkp1/Dusp1, IκB-α, and Gilz (111, 184).

One of the best-studied GR-interacting cofactors is GRIP1 (NCoA2 or SRC-2). This cofactor has potent anti-inflammatory properties, as documented by in vitro and in vivo studies. Accordingly, it has been shown in vivo in GRIP1 KO bone marrow-derived macrophages that the Dex-mediated repression of IL-1α, IL-1β, TNF, and CCL4 was significantly attenuated under LPS-induced inflammatory conditions. Besides, GRIP1 is protective against systemic inflammatory responses in vivo, since the mortality rate for GRIP1 KO mice injected with a sublethal dose of LPS is significantly higher. A survey of GRIP1 expression in autoimmune diseases using the NextBio database revealed that GRIP1 was found to be downregulated in synovial fibroblasts from patients with rheumatoid arthritis. In addition, GRIP1 mRNA levels were also reduced in peripheral monocytes of patients with lupus and in lupus models (185). This protein has a dual action upon binding to GR: it can act as both an activator and a repressor, in a gene-dependent context, ultimately serving to enhance the anti-inflammatory properties of GR. Acting as a GR coactivator, GRIP1 has been shown to enhance the transcription of GR-dependent genes crucial for inflammation, such as Gilz and Mkp1. Moreover, combinatorial treatment of ASM cells with TNF/IFN-γ strongly suppressed steroidal fluticasone propionate-dependent gene expression in a reporter system. This suppression, however, was no longer observed upon GRIP-1 overexpression (186).

One possible mechanism by which cofactors might mediate GR-p65 antagonism is by acting as targets for competition. If the amount of common transcriptional cofactors required by both GR and NF-κB for full transactivation is limited, it could result in the repression of one transcription factor's basal or induced activity when the other transcription factor is activated and depletes all the available cofactors. For the induction of downstream genes, both GR and NF-κB use the coactivators CBP and SRC-1. If competition for limiting amounts of CBP or SRC-1 accounts for the inhibitory effect of GR, then increased levels of the coactivators should restore or rescue p65-dependent gene expression. Indeed, the inhibitory effect of GR on p65-dependent gene expression was completely abolished by cotransfection of vectors expressing CBP. The inhibitory effect of p65 on GR-dependent gene expression was eliminated in a dose-dependent manner by the expression of CBP. Likewise, the suppression of GR-dependent gene expression by p65 was significantly decreased by SRC-1. Collectively, these functional studies claimed that CBP and SRC-1 are limiting for both p65- and GR-dependent transactivation and that the coactivators can rescue the repressive interaction between p65 and GR (187). However, to answer the question of cofactor competition binding, De Bosscher et al. showed that GC repression of p65-mediated gene expression is not relieved by overexpression of the coactivator CBP, SRC-1, or p300 and that sustaining repression by overexpressing GR does not result from downmodulation of p65 or CBP levels in the cell. Similar results were obtained for the reciprocal transrepression of a GR element-driven reporter gene by p65 (188).

Accordingly, GR, via its interaction with GRIP1, mediates transrepression on AP-1 actions. In U2OS osteosarcoma cells, collagenase 3, an AP-1-dependent gene, is activated by phorbol esters, and this activation is blocked by GR agonist treatment. Intriguingly, activation of GR causes its tethering to AP-1 and subsequent recruitment of GRIP1, which in this context acts as a corepressor. This corepressor function is dependent on an intrinsic repression domain, which is not shared by other p160/SRC family proteins (189). The phosphorylation of GRIP1 in response to corticoids is a prerequisite for the GR-GRIP1 interaction and potentiates GR-mediated activation of transcription (190). Besides, phosphorylation at S211 of the hGR is required to enable the recruitment of GRIP1 and for the receptor to exert its maximal transcriptional response, as evident via the activity of the synthetic mouse mammary tumor virus (MMTV) reporter and the endogenous Gilz mRNA levels in COS-1 cells (191). In conjunction with GRIP1, with the role of corepressor, GR tethers to DNA-bound AP-1 and NF-κB and represses the transcription of their target proinflammatory cytokine genes. However, these target genes fall into distinct classes depending on the step of the transcription cycle that is rate limiting for their activation: some are controlled through Pol II recruitment and initiation, whereas others undergo signal-induced release of paused elongation complexes into productive RNA synthesis. It has been reported that at the initiation-controlled inflammatory genes in primary macrophages, GR inhibited LPS-induced Pol II occupancy. In contrast, at the elongation-controlled genes, GR did not affect Pol II recruitment or transcription initiation but promoted, in a GRIP1-dependent manner, the accumulation of the pause-inducing negative elongation factor (192). Furthermore, GR interferes with the TLR3/4 signaling cascade through GRIP1 to reduce inflammation. IRF3 is a downstream effector of the TLR3/4 pathway and interacts with GRIP1 to exert its proinflammatory actions. GR and IRF3 competed for GRIP1 binding and in this way antagonized IRF3-mediated transcription, identifying the GRIP1-IRF3 interaction as a novel indirect target for GC-mediated immunosuppression (Fig. 4) (193).

One possible mechanism by which the inflammatory milieu can decrease the anti-inflammatory actions of GR is by changing the binding of cofactors in the GR complex. Accordingly, TNF was described to exert its inhibitory action by inducing the recruitment of negative cofactors to GR, altering its transactivation potential. From a yeast two-hybrid screen, FLASH (FLICE-associated huge protein or CASP8-associated protein 2) was identified to bind GRIP1 at a region between the second and third LXXLL motifs. FLASH is a mediator of TNF and Fas ligand signals, and in functional assays, it predominantly suppressed both GR transactivation and GRIP1 enhancement of the GC signal and inhibited the physical interaction between GR and the GRIP1 in HCT116 cells. These findings indicate that FLASH participates in a TNF-induced blockade of GR transactivation, upstream and independently of NF-κB (194). On the other hand, FLASH was identified by another research group as an interaction partner of GR, through its AF-1 domain. Notably, FLASH was found to increase the activity of the GR in reporter assays in a fusion of embryonic mouse hippocampal and human neuroblastoma cells (HN9.10), suggesting a tissue- and stimulation-specific action of FLASH on GR-mediated transcription activation (195).

The LBD of the GR interacts with both the N terminus and the C terminus of CBP. It is possible that CBP exerts its effects by contacting both the GR LDB and the AF-1 domain simultaneously through different CBP regions (196). However, CBP can also function conditionally as a negative regulator of GR activities. CBP suppressed the responsiveness of the MMTV promoter to Dex in a dose-dependent fashion in HeLa and A204 cells. The inhibitory effect of CBP is localized to its N-terminal domain and is independent of the histone acetyltransferase and coactivator binding domains but dependent upon its GR binding domain. P300 completely antagonized the suppressive effects of CBP in a dose-dependent fashion, probably by competing with this protein at the level of the transcription complex. These findings suggest that CBP and p300 may function additively or antagonistically to each other depending on their relative concentrations and type of target tissue to influence the sensitivity of tissues to GCs (197).

CEBPA (C/EBPα) is important for GC-dependent antiproliferative effects. Upon treatment, GC-sensitive rat BDS1 hepatoma cells induce the expression of CEBPA rapidly, which subsequently colocalize in a complex with GR at the promoters of the p21waf1/cip1 cyclin-dependent kinase inhibitor gene and the alpha-1 acid glycoprotein (AGP) acute-phase response gene, promoting their expression and resulting in G₁ cell cycle arrest (198). Moreover, GR interacts with CEBPB (C/EBPβ), potentiating its transcriptional induction. Reciprocally, CEBPB potentiated transcriptional activation by the DNA-bound GR LBD (199). The bZIP region of CEBPB is adequate for interaction with GR (200). In addition, GR was found to cooperate with CEBPB to enhance vitamin D receptor (VDR)-mediated 24-hydroxylase transcription by a mechanism related to vitamin D metabolism (201). Finally, GR binds to CEBPD, but this interaction seems to have no additional effect on GR activity (199).

GR interacts with proteins that can affect its intermolecular conformation and subsequently facilitate the recruitment of coactivators. Binding of Jun dimerization protein 2 (JDP2) within the DBD of GR leads to the acquisition of a functionally active secondary/tertiary structure in its AF-1 activation domain. This induced structure in AF-1 creates interaction surfaces for other coactivators such as TATA box binding protein (TBP) and CBP for the GR (202). Additionally, it was pointed out that TBP binding to AF-1 induces a significantly higher helical content at the expense of a random-coil configuration in the GR AF-1, resulting in a more accessible surface for coactivator proteins like SRC-1 (203).

Apart from the p160 family and the HATs mentioned above, there are members of the switch/sucrose-nonfermentable (SWI/SNF) family of proteins that also enhance transcription by alleviating the repressive effects of histone-DNA contacts, and GR has been shown to interact with these family members. BRG1 (SMARCA4), the ATPase subunit of the SWI/SNF complex, is essential for GR-mediated transrepression of the Pomc promoter. As described previously, GR drives this inhibition upon interaction with the orphan receptor Nur77. Brg1 is required in vivo to stabilize interactions between GR and Nur77 as well as between GR and HDAC2 in the promoter of Pomc (204). The members of the SWI/SNF family BRG1-associated factor 250 (BAF250) (205), BAF60a, and BAF57 associate with GR and enable the receptor to remodel the chromatin and induce transcriptional enhancement (206). The factor BRM (SMARCA2), by interaction with GR and retinoblastoma (RB), is involved in GC-induced apoptosis (205). The SWI/SNF complex was used in a previous study to show the mode of chromatin occupancy by the GR. Both the in vitro and in vivo results are consistent with a dynamic model (hit and run) in which GR first binds to chromatin after ligand activation, triggers remodeling activity, facilitates transcription factor binding, and is simultaneously lost from the template (207). Although these proteins modulate GR activity, there is no evidence for their direct involvement in GR actions under inflammatory conditions.

In the realm of infections, two HIV-1 proteins, namely, Vpr (virion-associated accessory protein) (208) and Tat (transactivator of transcription) (209), have been identified as coactivators of GR. Vpr affects both viral replication and differentiation, while it contains the classic coactivator signature motif, LXXLL. Vpr interacts directly with GR, enhancing the activity of the ligand-bound receptor on the MMTV promoter in lymphoid and muscle-derived cell lines, while it can create additional contacts with known GR coactivators, i.e., p300 or SRC-1 (210). Hence, it is speculated that the GR-Vpr interaction may contribute to the development of symptoms in patients with AIDS, such as muscle wasting, in the absence of increased GC levels. Additionally, Vpr overexpression enhances the GR-dependent suppression of IL-12 secretion by human monocytes and macrophages. Thus, Vpr may contribute to the suppression of innate and cellular immunities of HIV-1-infected individuals and AIDS patients (211). Tat is the most potent virally encoded transactivator of HIV-1, which utilizes members of the p160 family of coactivators, such as GRIP1, to enhance its transcriptional activity on the NF-κB promoter in concert with the host cell protein cyclin T1 complexed with P-Tefb. On the other hand, cotransfection of Tat, GRIP1, and cyclin T1 enhanced the GR-mediated stimulation of the MMTV promoter, suggesting that Tat can also attract the P-Tefb complex to the MMTV promoter through GRIP1. Thus, Tat may function as an adaptor molecule, efficiently stimulating the processes of transcription initiation and elongation through the recruitment of p160 coactivators and the P-Tefb complex (209). On the contrary, Dex inhibits Tat- and Vpr-enhanced luciferase activities in a tissue-specific, dose-dependent, and GR-mediated manner, thus acting protectively for the host (212).

GR Interaction with Classical Corepressors

The search for cofactors with opposite actions from coactivators led to the discovery of the nuclear receptor corepressors (NCoRs) and their homologue known by the complex name of silencing mediator of retinoic acid receptor and thyroid receptor (SMRT). The NR interaction site in NCoR is remarkably similar to the NR boxes in the p160 coactivator family. The corepressor motif is L/I-XXI/V-I, which differs from the p160 coactivator motif, LXXLL. The corepressor motif is referred to as a CoRNR box. The corepressors NCoR and SMRT bind to HDACs (213).

GRα also interacts with NcoR and SMART, which are actually macromolecular docking platforms for nuclear receptors and many transcription factors and repress the transcriptional activity of GRα by attracting HDAC/Sin3 complexes (14).

The glucocorticoid antagonist RU486 interferes with steroid-mediated activation, and in this process corepressors have been implicated. RU486 does not inhibit GR from DNA binding but enhances the recruitment of corepressors, leading to decreased GR activity. Upon RU486 treatment, GR can interact with the corepressors NcoRs and SMRT, and this interaction is antagonist dependent. For this interaction, both the N- and C-terminal domains of GR are important, and three NCoR regions have been described to be essential (positions 1745 to 1954, 1954 to 2215, and 2239 to 2453) (214). Within the GR steroid binding pocket, tyrosine 735 makes hydrophobic contact with the D ring of the steroid ligand. The substitutions Tyr735Phe, Tyr735Val, and Tyr735Ser resulted in impaired SRC-1 interaction but enhanced NCoR1 recruitment, basally and after RU486 treatment. Both SRC-1 and NCoR have been predicted to recognize a common hydrophobic cleft in the GR, and it seems that changes favorable to one interaction are detrimental to the other, thus identifying a molecular switch (215).

Apart from exhibiting a dampening function toward GR-dependent transcription, NCoR1 is also involved in the regulation of GR gene expression itself (NR3C1 gene), together with GR. This process is orchestrated by the recruitment of agonist-bound GR to exon 6, followed by the assembly of a GR/NCoR1/HDAC3 repression complex at the transcriptional start site of the GR gene. A functional nGRE in exon 6 of the GR gene and a long-range interaction occurring between this intragenic response element and the transcription start site appear to be instrumental in this repression. Consequently, the chronic nature of inflammatory conditions involving long-term GC administration may lead to the constitutive repression of GR gene transcription and thus to GC resistance (216).

The corepressor RIP140 binds GR as well and suppresses its actions. The negative effect of RIP140 on GR activity can be eliminated by the positive GR cofactor 14-3-3. This GR partner interacts with the corepressor RIP140 and antagonizes its binding to GR. 14-3-3 can additionally export RIP140 out of the nucleus and, interestingly, can change its intranuclear localization, thereby minimizing its negative effect on GR activity (217).

Importantly, the inflammatory environment can restrict the pool of available coactivators to bind to hormone receptors, leading to diminished steroid responsiveness. According to this notion, Leite et al. demonstrated that treatment of cultured human uterine smooth muscle cells (UtSMCs) with TNF significantly reduced the mRNA levels of the coactivators SRC-1 and SRC-2 (42% and 47% reductions, respectively) and diminished the respective protein levels but did not significantly alter the mRNAs encoding the corepressors NCoR and SMRT, suggesting this as a mechanism obtained by inflammatory cytokines to functionally block steroid hormone action (218).

Except for the coactivation effects that viral proteins may exert on GR, as mentioned above, they may blunt the receptor's transcriptional activity. More specifically, it has been reported that the molluscum contagiosum (MC) poxvirus protein MC013L suppresses the receptor's transcriptional activity via its direct interaction with GR. MC poxvirus, the only poxvirus circulating in the human population, causes epidermal lesions on the trunk and the limbs. Since GCs act as potent inhibitors of keratinocyte proliferation, it is speculated that MC013L may promote efficient virus replication by blocking the differentiation of infected keratinocytes (219). Similarly, overexpression of RSV nonstructural protein 1 (NS1) can also repress GR activity, as shown by decreased Gilz mRNA levels. These findings suggest that the inhibition of viral nonstructural proteins may be a viable target for therapy against RSV-related disease. However, a direct interaction of NS1 with GR has not yet been described (220).

GR INTERACTIONS INVOLVED IN T-CELL DIFFERENTIATION AND SIGNALING

T-cell receptor (TcR) signaling is the central pathway regulating T-cell biology. TcR signaling allows T cells to recognize antigens that are displayed on infected or professional antigen-presenting cells, typically in complexes with major histocompatibility complex (MHC) class I or II proteins.

Modulation of T-Cell Signaling via Nongenomic GR Effects

Apart from the classic genomic effects, it is well known that GCs can also exert fast nongenomic effects through specific interaction with cytosolic GR (cGR), nonspecific interactions with cellular membranes, and specific interactions with membrane-bound GR (mGR) (15). mGR participates in the rapid nongenomic effects of the GR, which have been correlated many times with the anti-inflammatory properties of the receptor. In patients with rheumatoid arthritis, the number of mGR-positive monocytes correlates with disease activity. A similar relationship with disease activity, as well as an inverse correlation between the number of mGR-positive monocytes and GC dosage, has been demonstrated in patients with other rheumatic diseases (12).

Ligation of the TcR leads to a cascade of signaling events, ultimately resulting in T-cell effector function (reviewed in references 221 and 222). Here we present the main mediators of TcR signaling that are associated with GR actions through PPIs. The primary event after TcR ligation is the induction of tyrosine phosphorylation by Src family kinases, especially LCK (lymphocyte-specific protein tyrosine kinase) and FYN. Active LCK phosphorylates tyrosine activation motifs on the TcR ζ chain and the γ, δ, and ε chains of CD3. Consequently, ZAP-70 is recruited in the complex, which directly phosphorylates the linker for activation of T cells (LAT). LAT is a transmembrane protein with 10 sites of potential tyrosine phosphorylation. Phosphorylation of the C-terminal four sites allows binding and activation of a set of adaptor molecules, including phospholipase C γ1 (PLC-γ1), growth factor receptor-bound protein 2 (Grb2), and other adaptor molecules.

The binding and activation of PLC-γ1 and Grb2 lead to the recruitment of additional molecules and/or the triggering of downstream signaling events. Activated PLC-γ1 cleaves phosphatidylinositol biphosphate (PIP2) to diacylglycerol (DAG), followed by Ca²⁺ release, and this event in turn activates protein kinase C (PKC) and consequent nuclear translocation of the transcription factor NF-κB (Fig. 6). Recruitment of Grb2 to LAT leads to the activation of Ras and MAPKs, including ERK. The combination of calcium flux, MAPK activation, and cytoskeletal rearrangements leads to T-cell activation, including the activation of transcription factors, cytokine synthesis, cell cycle entry, and cytotoxic activity.

FIG 6.

Modulation of TcR signaling by GR. In the absence of GCs, GR resides in a complex with LCK, FYN, and Hsp90, permitting the activation of the TcR cascade. Upon GC binding to GR, the complex is dissociated, and TcR signaling is impaired via a fast GR-dependent nongenomic effect. Additionally, GR interferes with TcR signaling via genomic mechanisms that include the formation of PPIs with T-bet, NF-ATc, and NF-κB, resulting in the suppression of their transcriptional activity.

LCK and FYN were identified as cellular targets involved in nongenomic GC immunosuppression in vitro and in vivo in T cells. Unliganded GR is part of a TcR-linked multiprotein complex containing Hsp90, LCK (binds CD4 and CD8 cell surface receptors), and FYN (binds CD3 coreceptors), and unligated GR is essential for TcR-dependent LCK/FYN activation. GR small interfering RNA (siRNA) transfection in T cells revealed that GR is crucial for mediating efficient TcR signaling (223). In vitro short-term GC treatment induces the dissociation of this complex, resulting in impaired TcR signaling as a consequence of abrogated LCK/FYN activation. Application of GCs in healthy individuals confirmed the suppression of LCK/FYN in T cells within 1 h in vivo. Moreover, Hsp90 is essential for this multiprotein complex formation since its silencing abrogates assembly (Fig. 6). The suppression of TcR signaling after GC treatment was evaluated by subsequent suppressed phosphorylation of TcR signaling intermediates (p38 MAPK, ERK, JNK, PKB, and PKC) (224). Therefore, drugs that selectively target membrane-bound GR might represent a novel immunosuppressive approach.

Among the most important properties of GCs that render them an efficient treatment for malignancies are their antiproliferative and apoptotic actions. The apoptotic effects of GCs have been evaluated in natural killer cells (225), neutrophils and eosinophils (226), leukemic T cells (227), and osteoblasts (228). In the case of thymocytes, the apoptotic actions of GCs are attributed to GR nongenomic effects, and for this outcome, the interaction of GR with the protooncogene tyrosine-protein kinase Src is essential. It has been proposed that GC-dependent thymocyte apoptosis requires a sequence of events, including binding of GR with Hsp90, Src, and PLC-γ1; binding of GR to its ligand; activation of PLC-γ1, acidic sphingomyelinase (aSMase), and endonuclease; ceramide production; Ca²⁺ mobilization; cytochrome c release; and proteasome and caspase activation (229). In untreated thymocytes, GR, Src, PLC-γ1, and Hsp90 are found in the same complex. Dex-bound GR mediates the activation of Src, which in turn rapidly activates PLC-γ1, and this has been proposed as a crucial event in the apoptotic process (230).

Modulation of T-Cell Differentiation through GR's Genomic Effects

Moreover, GR is involved in the differentiation of T-cell lineages by interaction with master transcription factors of these processes. The cytokines produced after pathogen infection contribute to the differentiation and expansion of CD4⁺-derived T cells, which play a critical role in mediating adaptive immunity. They are also involved in autoimmunity, asthma, allergic responses, as well as tumor immunity. During TcR activation in a particular cytokine milieu, naive CD4 T cells may differentiate into one of several lineages of Th cells, including Th1, Th2, and Th17 cells and regulatory T cells (Tregs), as defined by their pattern of cytokine production and function (231).

GR can exert its immunosuppressive effects by repressing cytokine production by T lymphocytes by suppressing their differentiation. GCs downregulate the production of cytokines produced by Th1 and Th2 lymphocytes by interfering with the activity of their master regulators T-bet and GATA-3, respectively. Suppression of the Th2-mediated response by the GCs does not involve a direct interaction of the GR with GATA-3 but is caused by decreased phosphorylation of GATA-3, which decreases its DNA binding (232). On the other hand, GC-mediated Th1 immunosuppression is based on the direct interaction of GR with T-bet, as shown by A. C. Liberman et al. (233). The interaction of GR with T-bet occurs in vitro and in vivo in primary splenocytes and leads to mutual inhibition of both transcription factors (Fig. 6). Furthermore, it was shown that GR-mediated inhibition of T-bet is dependent on the first zinc finger of GR and involves reduced DNA binding ability as well as reduced mRNA and protein levels of T-bet (233). Additionally, GR inhibits the production of IL-4 in T cells through its interaction with NF-AT (nuclear factor of activated T cells) (Fig. 6). IL-4 gene expression is regulated at the transcriptional level by increases in intracellular calcium concentrations and by the calcium-activated phosphatase calcineurin. The promoter of IL-4 contains binding sites for GR and NF-AT, but after combinatorial treatment of Jurkat cells with calcium and Dex, NF-AT is coimmunoprecipitated with GR in nuclear extracts and is unable to bind to the promoter of IL-4 and induce its expression (234).

Apart from regulating white blood cells, GCs also determine the fate of red blood cells. Treatment of mouse erythroleukemia (MEL) cells with hexamethylene bisacetamide induces erythrocyte differentiation, as judged by an increase in the synthesis of globins and other erythroid-specific products whose expression is mediated by the transcription factor GATA-1. GR binds GATA-1 and interferes with its function before any interaction with DNA, but the presence of a GRE near a GATA response element augments the GR effect. The N-terminal domain (106 amino acids) of the GR was found to be essential for this effect, possibly by binding to GATA-1. Since GATA-1 is autoregulatory, the interference of activated GR with GATA-1 function may explain how GCs inhibit the entire program of erythroid differentiation in MEL cells (235).

GR INTERFERENCE WITH THE TGF-β PATHWAY

From the time when growth factor beta was first described to have the ability to inhibit macrophage activation, transforming growth factor beta (TGF-β) has been analyzed for its role in regulating immune responses to a variety of pathogens, including viruses, bacteria, yeast, and protozoa. The majority of these studies have been performed with protozoan pathogens, including Trypanosoma cruzi (236) and a variety of Leishmania species (237, 238), while other studies have focused on mycobacteria and viruses, including HIV (239, 240). Investigators have reported that TGF-β has a negative influence on host responses and a beneficial effect on the survival and growth of intracellular pathogens by blocking the activation of lymphocytes and monocyte-derived phagocytes. However, later studies found that TGF-β may have a positive or beneficial effect for the host in some models of infection (241).

SMAD proteins are involved in signaling processes elicited through cell surface receptors of TGF-β signaling. The TGF-β and GC signaling pathways interact both positively and negatively in regulating a variety of physiological and pathological processes. Interaction of GR with SMAD3/4 represses the TGF-β transcriptional activation of the type 1 plasminogen activator inhibitor (PAI-1) gene in a ligand-dependent manner. GR can interact with SMAD3 and SMAD4 in vitro and in vivo through its C-terminal part, which is indispensable for this interaction in Hep3B cells (242). However, upon TGF-β stimulation, the SMAD3 interaction with GR enhances GC induction in reporter assays (243) (Fig. 7). Furthermore, SMAD6 interacts with the N-terminal domain of the GR through its Mad homology 2 domain and suppress GR activity in vitro. SMAD6 overexpression inhibits GC action in rat liver in vivo by attracting HDAC3 to DNA-bound GR and by antagonizing the acetylation of histones H3 and H4 induced by p160 histone acetyltransferase. SMAD6-mediated GR repression prevents Dex-induced elevation of blood glucose levels and hepatic mRNA expression of phosphoenolpyruvate carboxykinase (PEPCK), a well-known rate-limiting enzyme of liver gluconeogenesis. Collectively, it appears that the anti-GC actions of SMAD6 may contribute to the neuroprotective, anticatabolic, and pro-wound-healing properties of the TGF family of proteins (244) (Fig. 7).

FIG 7.

Interaction of GR with members of the TGF-β pathway. TGF-β mediates its transcriptional effects via SMAD and Daxx proteins. Daxx represses the activity of DNA-bound GR, yet PML antagonizes the negative actions of Daxx on GR. SMAD6 interacts with GR while recruiting HDAC3 in the same complex to cause histone deacetylation, thereby effectively antagonizing p160 actions and ultimately decreasing the transcription of GR target genes. Interaction of GR with DNA-bound SMAD3 and SMAD4 represses the transcription of PAI, while the interaction of SMAD3 with GR enhances the activity of GR.

Additionally, TGF-β mediates its effects through Daxx (death-associated protein). When cells are treated with TGF-β, Daxx becomes phosphorylated and activated and in turn activates the JNK pathway. The interaction of Daxx with GR in different tissues has been described by different research groups, and in all cases, this interaction is correlated with Daxx-mediated decreased GR activity (195, 245). It has been demonstrated that the interaction of the C-terminal region of Daxx with GR in the promoter region of GR target genes mediates decreased GR transcriptional activity, leading to reduced mRNA levels of Gilz and Slc19a2. This negative effect of Daxx could be alleviated by the coexpression of PML (promyelocytic leukemia protein), which is also an interaction partner of the GR (246) and mediates the translocation of Daxx to PML oncogenic domains (PODs), leading to an enhancement of GR transactivation. As described below, sumoylation generally regulates GR's activity positively but can also act as attraction site for negative regulators to be recruited to GR. According to this assumption, Daxx was found to preferentially bind to sumoylated GR. However, sumoylated GR and sumoylated PML share the same domain for interaction with Daxx, and upon treatment with arsenic trioxide, the sumoylation status of PML is increased, and it binds more efficiently to Daxx (Fig. 7). The result is that Daxx binds less efficiently to GR-bound promoters, leading to an alleviation of Daxx-mediated inhibition and increased Gilz levels (247).

GR INTERACTIONS WITH MODIFYING ENZYMES

The transactivation and transrepression properties of GR are dependent partially on interactions with modifying enzymes that can cause posttranslational modifications to the GR protein. Specific amino acids of GR can undergo modifications such as phosphorylation, acetylation/deacetylation, sumoylation, or ubiquitination, with significant effects on its actions. Therefore, we describe crucial GR interactions with modification enzymes that can alter the transactivation of important anti-inflammatory genes or the GR-dependent transrepression pathway.

GR Phosphorylation

Phosphorylation is the reversible addition of phosphate groups on a protein. Many residues along the GR protein, mainly in human, mouse, and rat GRs, have been characterized to undergo phosphorylation, which influences the GR ligand and DNA binding abilities, subcellular localization, GR PPIs, and protein half-life, ultimately affecting GR transactivation and transrepression properties. In human GR, 10 serine residues have been identified as phosphorylation targets, including S45, S113, S141, S203, S211, S226, S234, S267, S404, and, more recently, S134 and threonine 8 (T8) (Fig. 1). Of note, all phosphorylation sites are localized in the N-terminal domain of GR. In mouse GR, eight phosphorylation sites have been discovered, S122, S150, S212, S220, S234, S315, S412, and T159, and in rat, eight residues have been found, S134, S162, S224, S232, S246, S327, S424, and T171 (11, 248, 249).

The first kinases reported to interact with GR as a target for phosphorylation belong to the family of cyclin kinases. Phosphorylation of hGR S203 and S211 (corresponding to rGR S224 and S232) can be mediated by Cdk2/cyclin A kinase complexes, whereas Cdk2/cyclin E targets only hGR S203 (250). Mutational studies proved that cyclin-dependent kinase function is necessary for full receptor-mediated transcriptional enhancement (251). In general, phosphorylation at S203 can be cytoplasmic and is correlated with decreased GR activity, while S211 phosphorylation is predominantly nuclear and is related to enhanced transactivation of GRE-containing promoters.

The GR phosphorylation pattern is species specific, and there are differences among human, mouse, and rat GRs regarding the positions and the kinases that interact with the receptor. According to this notion, glycogen synthase kinase 3 (GSK-3) mediated the phosphorylation of rGR at T171 in vitro, leading to decreased GR transcriptional activity without affecting the transrepression GR properties (on the AP-1-dependent reporter system). Despite its effect on rGR, GSK-3 does not affect phosphorylation levels at A150 of hGR, which is the corresponding residue. Interestingly, human S404 was identified as a target of GSK-3b. Nuclear phosphorylation at this residue results in GR nuclear export, enhanced GR downregulation and reduced transactivation and transrepression.

The central nervous system (CNS)-specific cyclin-dependent kinase 5 (CDK5) also phosphorylates human GRα at multiple serines, including S203 and S211, and modulates GR-induced transcriptional activity by physically interacting with the receptor through its activator component p35 and by changing the accumulation of transcriptional cofactors on GRE-bound GRα (252).

Recently, it was reported that binding of GR to a protein with isomerase activity alters its phosphorylation status. The two N-terminal GR serines S203 and S211 are targets for the prolyl isomerase Pin1, leading to the potentiation of GR transactivation (as determined for the Gilz promoter). Pin1 physically interacts with the GR and mediates its transactivation but not its transrepression properties. Pin1 expression is augmented in both cancer and inflammation, thus determining the GR-dependent outcome in these pathophysiological states (253). Interaction of the DNA binding site for the Ku antigen/DNA-dependent protein kinase (DNA-PK) catalytic subunit with GR negatively regulates the receptor's transcriptional activity. In order for DNA-PK to exert its negative effects on GR-mediated transactivation, the colocalization of the two proteins on DNA in cis is critical. Under these circumstances, the DNA-PK kinase phosphorylates GR at S527 located in the hinge region of the receptor, and it is speculated that the inhibitory effect of this phosphorylation on the transactivation of the GR could be due to a subsequent dimerization defect or an impaired nuclear localization signal (254). The serine/threonine protein phosphatase PP5 has been correlated with in vitro cytokine-induced GC resistance in ASM cells. When cells were treated with IFN-γ and TNF in combination with fluticasone, the fluticasone-induced phosphorylation of GR was greatly suppressed on S211 but not on S226. Knockdown of PP5 partially prevented the cytokine-suppressive effects on GR phosphorylation and mediated transactivation, suggesting that proasthmatic cytokine-induced corticosteroid insensitivity in ASM cells is due, in part, to PP5-mediated impairment of GR's S211 phosphorylation (255).

GR O-GlcNAcylation

O-linked N-acetylglucosamine (O-GlcNAc) transferase (OGT) catalyzes the addition of a single N-acetylglucosamine in an O-glycosidic linkage to serine or threonine residues of intracellular proteins, a mechanism known as O-GlcNAcylation. Since both phosphorylation and O-GlcNAcylation compete for similar serine or threonine residues, the two processes may compete for sites. O-GlcNAc transferase interacts with the ligand-bound GR and potentiates the GR transrepression pathway, as indicated for NF-κB and AP-1 reporter systems. The recruitment of OGT by GR leads to increased O-GlcNAcylation and decreased phosphorylation of RNA polymerase II on target genes. Functionally, overexpression of OGT enhances GC-induced apoptosis in resistant cell lines, while knockdown of OGT prevents sensitive cell lines from apoptosis (256).

GR Acetylation/Deacetylation and Interaction with HDACs

GR is known to be acetylated after ligand binding at K494 and K495 of the KXKK motif in the hinge region (Fig. 1). Furthermore, Nader et al. found that the acetyltransferase Clock (circadian locomotor output cycles kaput) interacts with GR through the LBD in a ligand-dependent manner and acetylates GR not only at lysines 494 and 495 but also at lysines 480 and 492. ChIP assays showed that GR acetylation by the Clock/Bmal1 complex reduces the binding of GR to GREs in selected target genes, including Gilz and glucose-6-phosphatase (G6p), and impairs GR transcriptional activity (257).

Transcriptional repression of IL-5 by GCs involves the GR-mediated transcriptional inhibition of GATA-3, the key-specific determinant of expression of Th2 cytokines. GATA-3 inhibition is dependent on HDAC1 recruitment in a complex with the GR (258).

Furthermore, Qiu et al. showed that HDAC1 acts as a GR coactivator. GR activation by GCs leads to progressive acetylation of HDAC1 in vivo, which in turn inhibits the deacetylase activity of the enzyme, and this acetylated HDAC1 form is enriched in a GR complex containing the p300 protein. Chromatin binding analysis based on the activity of a GR-dependent reporter revealed that HDAC1, acting as a coactivator, is required for efficient induction of some genes by GR (259).

Furthermore, GCs act on natural killer cells to inhibit immune response genes through epigenetic processes. This mechanism was shown to involve the association of HDAC1 and SMRT with ligand-activated GR in the nuclei of natural killer cells. These cofactors are participants in the histone deacetylation and transrepression processes that accompany the GC-mediated decrease in the function of natural killer cells (260).

Commonly, protein deacetylation leads to decreased transcriptional activities, but this not the case for GR. GR binding to HDACs can lead to either increased or decreased GR-dependent activities. GR deacetylation, following its interaction with HDAC2, plays an important role in the repression of NF-κB. GR becomes acetylated after ligand binding, and HDAC2-mediated GR deacetylation enables GR binding to the p65 subunit of NF-κB. Site-directed mutagenesis of K494 and K495 reduced GR acetylation, and the ability to repress NF-κB-dependent gene expression becomes insensitive to histone deacetylase inhibition. In addition, it was shown that overexpression of HDAC2 in GC-insensitive alveolar macrophages from patients with COPD is able to restore GC sensitivity. Thus, HDAC2 plays a critical role in GC sensitivity to repress NF-κB but not GRE-mediated gene expression (261). GR can also inhibit IL-1β-enhanced GM-CSF transcription in a similar mechanism by association with HDAC2. In the promoter of GM-CSF, it was found that GR acts both as a direct inhibitor of CBP-associated HAT activity and as a recruiter of HDAC2 to the p65-CBP HAT complex (262).

LCoR was identified as a corepressor of agonist-bound nuclear receptors, including GR. The suppressive effect of LCoR on GR activity was mediated through the recruitment of HDAC3 in the same complex. By using the HDAC inhibitor trichostatin, the negative effect on GR activity was completely abolished (263).

GR Sumoylation

The activity of GR can be modulated by sumoylation, which is known to target three residues of the human GR protein: lysine 277 (K277) and K293 in the N-terminal domain and K703 within the ligand-binding domain (K297, K313, and K721 in rat, respectively) (264) (Fig. 1). Sumoylation of the NTD sites mediates the negative effect of the synergy control motifs of GR on promoters with closely spaced GR binding sites. RSUME (RWD-containing sumoylation enhancer) interacts with rat GR, mediating its sumoylation at K721 and resulting in positive regulation of the target genes FKBP51 and S100P. Both mutation of K721 and silencing of RSUME diminish GRIP1 coactivator activity. RSUME expression is induced under stress conditions, and it acts as a key factor in heat shock-induced GR sumoylation (264). By using isogenic HEK293 cells expressing wild-type GR or sumoylation-defective GR, it was shown that GR-dependent genes affected by GR sumoylation are significantly associated with pathways of cellular proliferation and survival and that sumoylation influences genome-wide chromatin occupancy of the GR (265).

Additionally, SUMO4 (small ubiquitin-like modifier 4) interacts with GR. Following EMSAs involving TNF treatment, it was shown that SUMO4-induced GR sumoylation enhanced GR DNA binding activity (266).

GR Interaction with Ubiquitin-Conjugating Enzymes

Ubiquitin-conjugating enzymes are important components of the ubiquitin-proteasome pathway, facilitating the transfer of activated ubiquitin from ubiquitin-activating enzymes to target proteins with the help of ubiquitin-protein ligases.

K419 of GR has been described as a target residue for ubiquitination (Fig. 1). Reports on the effect of the ubiquitin-conjugating enzyme Ubch7 on GR activity are contradictory, possibly due to the tissue specificity of the actions of Ubch7. In HeLa cells, Ubch7 interacts with GR and modulates its transcriptional activity in a hormone-dependent manner. The ubiquitin conjugation activity of Ubch7 is required for its ability to potentiate transactivation by GR, and its coactivation function is dependent on SRC-1 (267). Conversely, in COS7 cells, Ubch7 reduces GR-dependent reporter activity by causing both a significant reduction in maximal activation and a significant increase in the 50% effective concentration (EC₅₀). It was shown that the ubiquitin ligase activity of Ubch7 is responsible for this negative effect, suggesting a role for the 26S proteasome and GR protein stability in mediating the UbcH7 effect (268).

Ubc9 also binds to GR and displays no intrinsic transactivation activity but modifies both the absolute amount of the induced gene product and the fold induction by GR. With high concentrations of GR, added Ubc9 also reduces the EC₅₀ of agonists and increases the partial agonist activity of antagonists in a manner that is independent of the ability of Ubc9 to transfer SUMO1 to proteins. This new activity of Ubc9 requires only the LBD of GR and part of the hinge region (269).

E6-AP (encoded by the Ube3a gene) is an E3 ubiquitin ligase containing a HECT (homologous to E6-associated protein carboxy-terminal) domain and three consensus receptor-interacting LXXLL motifs which interact with GR in a ligand-dependent manner (270). Mutations of maternal E6-AP are associated with Angelman syndrome (AS). It has been reported that E6-AP regulates GR transactivation and that the GR signaling pathway is disrupted in maternal Ube3a-deficient mouse brain, and AS mice show a significantly higher level of blood corticosterone and selective loss of GR. The chronic stress due to alterations in the GR signaling pathway might lead to anxiety-like behavior in a mouse model of AS (271).

CONCLUDING REMARKS

Glucocorticoids have been widely used for over 60 years for the treatment of numerous inflammatory and autoimmune diseases due their potent ability to suppress inflammation and their antiproliferative effects. However, frequent administration results in serious adverse effects. The specificity of the GR outcome depends on diverse factors, including its physical contacts with modulatory proteins. As discussed in this review, the GR interactome consists of a plethora of protein interactions in different cell compartments that affect the GR activity. More specifically, GR has been described to undergo various posttranslational modifications, and for this reason, it interacts with kinases, phosphatases, acetylases/deacetylases, ubiquitin-conjugated enzymes, and sumoylating enzymes. Furthermore, its binding to other nuclear receptors constitutes cross talk of high complexity. GR interactions with coactivators and corepressors as well those with transcription factors like NF-κB and AP-1 are of immense significance due to their potent anti-inflammatory and antiproliferative implications. The use of conventional PPI techniques, such as coimmunoprecipitation, affinity electrophoresis, and chemical cross-linking, and also the development of high-throughput technologies such as yeast two-hybrid assays and mass spectrometry enable researchers to reveal complexes involved in various regulatory pathways in health and in disease. For instance, the correlation of the cofactor SRC-1 with the pathophysiology of breast cancer as well as its involvement in resistance to endocrine therapy opened new ways for drug development (272). Furthermore, the implication of histone deacetylases in promoting the anti-inflammatory actions of the GR constitute a milestone in understanding the mechanisms related to asthma (111, 262). Apart from describing the PPIs of GR related to its transrepressive pathway, which is well documented for its GC-mediated anti-inflammatory actions, our goal was to report the interactions that affect the transactivation of GR as well. Since GR homodimerization and subsequent gene activation have been described as equally important in the battle against inflammation, we provide evidences for the crucial role of the GR binding proteins in this pathway. Many of these proteins have been reported to physically interact with GR and affect its actions, but the mechanisms underpinning these effects have not always been clearly characterized. Interestingly, the physical contacts of GR are extended with nonprotein molecules, such as noncoding RNAs, which have emerged relatively recently, and the current data pinpoint these interactions of high importance for the actions of the receptor. A characteristic example is represented by the noncoding RNA Gas5 (growth arrest-specific 5), which is postulated as a marker for GC sensitivity (273). Gas5 binds to the DBD of GR, acting as a decoy GRE and rendering GR unable to bind to DNA and mediate its transactivation program. In this way, Gas5 sensitizes cells to apoptosis by suppressing the expression of GR-dependent genes, including cellular inhibitor of apoptosis 2 (274). Noncoding RNAs appear as promising interactors and regulators of GR, yet research is still ongoing to decipher their actions.

In this review, we have summarized the known GR interactors, with an emphasis on those that have been described in the context of inflammation and microbial infection. The interactome of GR is indeed extensive, and it is involved in many physiological and pathological processes. Unraveling the GR interactome will lead to a deeper understanding of the far-reaching and complex actions of the receptor and will pave the way to the development of more effective and safer treatments for infectious, autoimmune, and inflammatory diseases as well as malignancies.

Biographies

Ioanna Petta is a Ph.D. student in a shared project between the Lab of Mouse Genetics in Inflammation (MGI) of Prof. Claude Libert in the Inflammation Research Center (IRC) and the Cytokine Receptor Lab (CRL) of Prof. Jan Tavernier, University of Ghent-VIB (Vlaams Instuteit voor Biotechnologie). She obtained her bachelor's degree in Biology from the University of Patras (Greece), after which she conducted her master's in Biotechnology in the same university. She started her Ph.D. project in 2010 after obtaining a fellowship from VIB for international Ph.D. students. Her research is focused on the identification of novel protein-protein interactions of glucocorticoid receptor and their importance in the receptor's anti-inflammatory actions.

Lien Dejager finished her Ph.D. in Biotechnology from the University of Ghent in 2010 under the guidance of Prof. Claude Libert, IRC, VIB (Belgium). Afterwards, she became a postdoctoral researcher at FWO-Vlaanderen in the same group and worked for 6 months in the laboratory of Prof. Jan Tuckermann (Germany). Her major research interests are elucidating the anti-inflammatory mechanisms of glucocorticoids and the mechanisms underlying glucocorticoid resistance, aiming to design more efficient glucocorticoid-based therapies.

Marlies Ballegeer is a Ph.D. student in the Lab of Mouse Genetics in Inflammation (MGI) of Prof. Claude Libert in the Inflammation Research Center (IRC), University of Ghent-VIB (Vlaams Instituut voor Biotechnologie). She obtained her bachelor's and master's degrees in Biomedical Sciences from the University of Ghent (Belgium). She started her Ph.D. project in 2011, and her research is focused on the role of GR dimers in inflammation and the anti-inflammatory functions and regulation of Gilz in acute inflammation.

Sam Lievens obtained a Ph.D. on the molecular biology of plant-rhizobium symbiosis at the VIB/Ghent University Department of Plant Systems Biology in 2001. In 2002, he joined the Cytokine Receptor Lab for a postdoc that mainly focused on the development of the MAPPIT technology for protein interaction analysis. Since 2007, he has been the technology/business development manager of a university-wide consortium aimed at technology development and valorization, centered on drug discovery platforms.

Jan Tavernier obtained his Ph.D. in 1984 on the cloning of interferon and interleukin genes. After an extended stay at Biogen and later at Roche, he returned in 1996 to Ghent University at the VIB Department of Medical Protein Research. He founded the Cytokine Receptor Laboratory that currently hosts 40 researchers. His main areas of expertise are cytokine receptor activation and signal transduction, also linking to pathways involved in innate immunity, and the analysis of protein-protein interactions, including interactome mapping, pathway walking, and molecular description of interdomain interactions (http://www.mappit.be/). Prof. Tavernier has published more than 220 refereed manuscripts, 20 of which have been cited over 100 times. He also holds 23 patent applications. He is Chair of the Department of Biochemistry at the Faculty of Medicine and Health Sciences and is Codirector of the VIB Department of Medical Protein Research. He is a member of Royal Belgian Academy of Sciences and the Arts.

Karolien De Bosscher obtained her Ph.D. at UGent on the molecular mechanisms of glucocorticoids in 2000. Following her Ph.D., she studied TGF-β-signaling pathways at Cancer Research UK. From 2003 onwards, she coordinated nuclear receptor research at UGent. Since 2013, she has held a full professorship at the medicine faculty of UGent. Within the Receptor Research Laboratories as part of the UGent-VIB department of biochemistry and medical protein research, she runs the Nuclear Receptor laboratory. Prof. De Bosscher has a longstanding focus on anti-inflammatory action mechanisms of nuclear receptors (GR and PPARs). In 2007, she won the prestigious Pfizer Prize, and in 2012, she received the Prize of the Academy for Fundamental Research in Medicine as well as the Belgian Endocrine Society Lecture Award. To date, Prof. De Bosscher has 67 original publications in PubMed and contributions to several book chapters. Her h-index factor is 30, with an average number of citations of 52/article.

Claude Libert has a master's and Ph.D. in Sciences, both obtained at the University of Ghent under the guidance of Walter Fiers. After a postdoc in Italy, he became a group leader with VIB (Flanders Institute for Biotechnology) in 1997 and a professor at the University of Ghent in 2003. His major research interest is acute inflammation and the cross-talk between several important players in inflammation, with a focus on TNF, IFN, matrix metalloproteinases, and glucocorticoids.