The gene encoding hGRα(hGR gene) is one locus on the long arm of chromosome 5 (q31.3), and consists of nine exons spanning over 150kb. Expressed hGRαis a panel of eight amino terminal translational isoforms of varying lengths, each of which consists of three subdomains: the N-terminal (NTD), the DNA-binding (DBD) and the ligand-binding (LBD) domains. In our expression and functional studies referred to here we have employed as representative the longest GRαisoform, comprising 777 amino acids.
The gene encoding hGRα(hGR gene) is one locus on the long arm of chromosome 5 (q31.3), and consists of nine exons spanning over 150kb. Expressed hGRαis a panel of eight amino terminal translational isoforms of varying lengths, each of which consists of three subdomains: the N-terminal (NTD), the DNA-binding (DBD) and the ligand-binding (LBD) domains. In our expression and functional studies referred to here we have employed as representative the longest GRαisoform, comprising 777 amino acids. The hGR gene also produces an equal number of hGRβisoforms by the use of an alternative 3’ exon 9β, which cannot bind glucocorticoids and exert a dominant negative effect upon the transcriptional activity of hGRα.1,5,6
In the absence of ligand, hGRαresides mostly in the cytoplasm of cells as part of a hetero-oligomeric complex, which contains chaperon heat shock proteins (HSPs) 90, 70, 23 and FKBP51, as well as other proteins.7 Upon ligand-induced activation, the receptor dissociates from this multiprotein complex and translocates into the nucleus through the nuclear pore with the energy-dependent mechanism that includes importin αand β. Inside the nucleus, hGRαbinds as a homodimer to glucocorticoid response elements (GREs) in the promoter regions of target genes and regulates their expression positively or negatively, depending on GRE sequence and promoter context.7–10 The ligand-activated hGRαcan also modulate gene expression independently of DNA binding by interacting, possibly as a monomer, with other transcription factors such as nuclear factor-κB (NF-κB), activator protein1 (AP-1), p53 and signal transducers and activators of transcription (STATs)11–14 (see Figure 1a).
To initiate the transcription, hGRαuses its transcriptional activation domains, activation function (AF)-1 and AF-2, located in NTD and LBD, respectively, as surfaces to interact with co-activators or co-repressors.15–19 The p160 co-activators such as the glucocorticoid receptor-interacting protein 1 (GRIP1) play an important role in the hGRα-mediated transactivation of glucocorticoid-responsive genes given that they interact directly with both the AF-1 of hGRαthrough their carboxyl-terminal domain and the AF-2 through multiple amphipathic LXXLL signature motifs located in their nuclear receptor-binding (NRB) domain.20 They also have histone acetyltransferase (HAT) activity, which promotes chromatin decondensation and facilitates initiation of transcription15–18 (see Figure 1b).
Alterations in the molecular mechanisms of hGRαaction may lead to alterations in tissue sensitivity to glucocorticoids, which may take the form of resistance or hypersensitivity and may be associated with significant morbidity.21–24 In this article we summarise the molecular mechanisms underlying primary generalised glucocorticoid resistance and secondary, inflammation-induced, generalised glucocorticoid resistance. Primary Generalised Glucocorticoid Resistance
Primary generalised glucocorticoid resistance is a rare condition characterised by generalised, partial, target-tissue insensitivity to glucocorticoids.25–30 This leads to activation of the HPA axis and compensatory elevations in circulating cortisol and adrenocorticotropic hormone (ACTH) concentrations, which maintain circadian rhythmicity and appropriate responsiveness to stressors. The excess ACTH secretion results in adrenal hyperplasia and increased production of adrenal steroids with mineralocorticoid and/or androgenic activity25–30 (see Figure 2).
The clinical manifestations of primary generalised glucocorticoid resistance are summarised in Table 1 and relate to the pathophysiological alterations depicted in Figure 2. The clinical spectrum of the condition is broad and ranges from most severe to mild forms, while a number of patients may be asymptomatic, displaying only biochemical alterations25–42 (see Table 1). The variable clinical phenotype of primary generalised glucocorticoid resistance may be due to: variations in the glucocorticoid, mineralocorticoid or androgen receptor signalling pathways; variations in the degree of tissue sensitivity to glucocorticoids, mineralocorticoids and/or androgens; variations in the activity of key hormone-inactivating or -activating enzymes, such as the 11β-hydroxysteroid dehydrogenase43 and 5αreductase;44 and other genetic or epigenetic factors, such as insulin resistance and visceral obesity.27
Molecular Mechanisms of Primary Generalised Glucocorticoid Resistance
The molecular basis of primary generalised glucocorticoid resistance has been ascribed to mutations in the hGR gene that impair the molecular mechanisms of hGR action and reduce tissue sensitivity to glucocorticoids. We have identified most hGR gene mutations associated with primary generalised glucocorticoid resistance and studied the molecular mechanisms through which the corresponding mutant receptors affect glucocorticoid signal transduction. More specifically, we determined: the transcriptional activity of the mutant receptors; the ability of the mutant receptors to exert a dominant negative effect upon the wild-type receptor; the affinity of the mutant receptors for the ligand; the subcellular localisation of the mutant receptors and their nuclear translocation following exposure to the ligand; the ability of the mutant receptors to bind to GREs; the interaction of the mutant receptors with the GRIP1 coactivator; and the motility of the mutant receptors within the nucleus of living cells.25–42,45,46 The molecular defects elucidated in the reported cases are summarised in Table 2, while the corresponding mutations in the hGR gene are shown in Table 2 and Figure 3.
All mutant receptors demonstrated variable reduction in their ability to transactivate the glucocorticoid-responsive mouse mammary tumour virus (MMTV) promoter in response to dexamethasone compared with the wild-type receptor, with the most severe impairment observed in the cases of R477H, I559N, V571A and D641V mutations.31–42 The mutant receptors hGRαI559N, hGRαF737L, hGRαI747M and hGRαL773P exerted a dominant negative effect upon the wild-type receptor that might have contributed to manifestation of the disease at the heterozygote state.31,35,38,40,42
All mutant receptors in which the mutation is located in the LBD of the hGRα showed a variable reduction in their affinity for the ligand, with the most severe reduction observed in the cases of I559N, I747M and V571A mutations.31–42 The decreased affinity of the mutant receptors for the ligand most likely reflects the location of the mutations in the LBD of hGRα. The mutant receptor hGRαR477H, in which the mutation is located at the DBD of the receptor, demonstrated normal affinity for the ligand41 (see Figure 2). In the absence of dexamethasone, the wild-type hGRα was primarily localised in the cytoplasm of cells. The pathological mutant receptors were also localised in the cytoplasm of cells in the absence of ligand, except for hGRαV729I and hGRαF737L, which were localised in both the cytoplasm and the nucleus of cells. The addition of dexamethasone (10-6M) resulted in translocation of the wild-type receptor into the nucleus within 12 minutes, but a much slower translocation of the mutant receptors into the nucleus, which ranged from 20 minutes (R477H) to 180 minutes (I559N and F737L).31–42 These findings suggest that all hGR mutations affect the nucleocytoplasmic shuttling of hGRα, probably through impairment of the function of the nuclear localisation signals NL1 and/or NL2.47–50 Defective mechanisms that may relate to delayed nuclear export, such as the calreticulin export pathway and certain motifs in the DBD that function as nuclear export signals, may account for the nuclear localisation of the unliganded hGRαV729I and hGRαF737L,51,52 an effect that may be similar to the nuclear retention mechanism of hGRα.35 The wild-type and all mutant receptors in which the mutations are located in the LBD of hGRα preserved their ability to bind to DNA.31–42 The only mutant receptor that failed to bind to DNA was hGRαR477H, in which the mutation is located at the C-terminal zinc finger of the DBD of the receptor.41 A major function of the C-terminal zinc finger of the DBD of hGRα is to contribute to receptor homodimerisation, a prerequisite for potent receptor binding to GREs and efficient transactivation of glucocorticoid-responsive genes.53,54 Point mutations in the DBD of the hGR may abolish DNA binding, resulting in silencing of transcriptional activation, although they may not affect the ability of the mutant receptors to transrepress AP-1-, NF-κB- and/or other target-genedependent transcription, possibly through protein–protein interactions and/or tethering of other co-factors to the transcriptional machinery.54–57
We next investigated the interaction between the mutant receptors and the GRIP1 co-activator in vitro. GRIP1 contains two sites that bind to steroid receptors: one site, the NRB site, is located at the central part of the protein and interacts with the AF-2 of hGRα in a ligand-dependent fashion, while the other site is located at the carboxyl-terminus of the protein and binds to the AF-1 of hGRα in a ligand-independent fashion.58–60 The wild-type and most mutant receptors bound to fulllength GRIP1 and the carboxyl-terminal fragment of GRIP1, but not to the NRB fragment of GRIP1, suggesting that they interact with the GRIP1 coactivator in vitro only through their AF-1. Exceptions were the receptors hGRαR477H, which interacted with both the AF-1 and AF-2 of hGRα, and hGRαI559N, which did not interact with either fragment of GRIP1.31–42 These findings suggest that the hGR mutant receptors may form a defective complex with GRIP1, which is partially or completely ineffective. Furthermore, the mutant receptors may also display an abnormal interaction with other AF-2-associated proteins such as the p300/CBP cointegrators and components of the vitamin D-receptor interacting protein (DRIP)/thyroid-hormone-receptor-associated proteins (TRAP) complex.15–18 Using fluorescence recovery after photobleaching (FRAP) analysis, we demonstrated that all hGR pathological mutant receptors had defective transcriptional activity and dynamic motility defects inside the nucleus of living cells.46 In the presence of dexamethasone, these mutants displayed a curtailed 50% recovery time (t1/2) after photobleaching and, hence, significantly increased intranuclear motility and decreased chromatin retention. The t1/2 values of the mutants correlated positively with their transcriptional activities and depended on the hGR domain affected. Therefore, the motility defect of the mutant receptors is a good overall index of functionality.46
In summary, mutations in the hGR gene impair the molecular mechanisms of glucocorticoid action and lead to generalised tissue insensitivity to glucocorticoids. A consequent increase in the activity of the HPA axis compensates for the reduced sensitivity of peripheral tissues to glucocorticoids at the expense of ACTH hypersecretion-related pathology. The study of the molecular defects of natural hGR mutants enhances our understanding of hGR action and highlights the importance of integrated cellular and molecular signalling mechanisms in maintaining homeostasis and preserving normal physiology.
Secondary Generalised Glucocorticoid Resistance
Associated with Systemic Inflammation and Chronic Inflammatory Diseases
Acute and chronic inflammatory processes and diseases such as sepsis, asthma, rheumatoid arthritis and inflammatory bowel disease are characterised by increased expression of multiple inflammatory genes. The latter are regulated by pro-inflammatory transcription factors such as NF-κB and AP-1, which play a critical role in amplifying and perpetuating the inflammatory process. Furthermore, inflammation-mediated activation of the nervous system and the HPA axis is an integral component in the regulation of the innate and adaptive immune response.61 Cytokines and inflammatory mediators activate the HPA axis both directly and indirectly. Activation of the HPA axis has profound inhibitory effects on the immune/inflammatory response.62 At the cellular level, the main anti-inflammatory and immunosuppressive effects of glucocorticoids include alterations in leukocyte traffic and function, decreases in production of cytokines and mediators of inflammation and inhibition of their action on target tissues by the latter. It is worth noting that hyperactivity or hypoactivity of the HPA axis leads to dysregulation of this neuroendocrine loop and may lead to systemic changes in inflammation and immunity.61,62
Glucocorticoids represent the most effective anti-inflammatory treatment for many of the above inflammatory conditions.63,64 However, failure to
respond to glucocorticoid therapy (glucocorticoid resistance) or need for chronic glucocorticoid treatment to maintain disease remission (glucocorticoid dependence) is common among patients treated with glucocorticoids and may be associated with significant morbidity and/or mortality. This variable response to glucocorticoid therapy is likely to be due to the process of inflammation, the disease itself or the genetic background of the patient.63,64
Anti-inflammatory Glucocorticoid Signalling Mechanisms
Glucocorticoids are important in suppressing several inflammatory pathways. For example, glucocorticoids inhibit prostaglandin production through three independent mechanisms: the induction and activation of annexin I, the induction of mitogen-activated protein kinase (MAPK) phosphatase 1 and the repression of transcription of cyclo-oxygenase 2. Annexin I (also called lipocortin-1) is an antiinflammatory protein that physically interacts with and inhibits cytosolic phospholipase A2 (cPLA2), which in turn blocks the release of arachidonic acid and its subsequent conversion to eicosanoids (i.e. prostaglandins, thromboxanes, prostacyclins and leukotrienes).62–64 Glucocorticoids also induce MAPK phosphatase 1, an anti-inflammatory protein that dephosphorylates and inactivates Jun N-terminal kinase (JNK), thereby inhibiting c-Jun-mediated transcription of inflammatory and immune genes.6
MAPK phosphatase 1 also dephosphorylates and inactivates all members of the MAPK family of proteins, including JNK, extracellular-signal-related kinase 1 and 2 and p38 kinase. Consequently, MAPK phosphatase 1 may also inhibit cPLA2 activity by blocking its phosphorylation by MAPKs and MAPK-interacting kinase. Furthermore, glucocorticoids and the hGR directly interfere with c-Junmediated transcription. The transcriptional interference between the hGR and c-Jun homodimers and/or c-Jun-Fos heterodimers occurs through protein–protein interactions and has proved to be a major antiinflammatory mechanism of glucocorticoids.64
Generalised Glucocorticoid Resistance
In addition to the above, the glucocorticoid–hGR complex physically interacts with NF-κB to block its transcriptional activity.65 NF-κB is a heterogeneous collection of dimers, the most common form consisting of a p65 and a p50 subunit. In its inactive state, NF-κB is sequestered in the cytoplasm by an inhibitory protein, the IκB. Tumour necrosis factor (TNF)-α, interleukin (IL)-1, microbial pathogens, viral infections and other inflammatory signals trigger signalling cascades that activate IκB kinases. Phosphorylation of IκB leads to its ubiquitination and degradation by the proteasome, thereby unmasking a nuclear localisation signal on NF-κB. In the nucleus, NF-κB binds DNA sequences, the NF-κB response elements, and stimulates the transcription of cytokines, chemokines, cell-adhesion molecules, complement factors and receptors for these molecules. NF-κB also induces the transcription of cyclo-oxygenase 2, an enzyme essential for prostaglandin production (see Figures 4a and 4b). A more detailed description of the interactions between the hGR and the NF-κB signalling systems are discussed in the following section.64,65
Glucocorticoids may also have rapid effects on inflammation that are not mediated by alterations in gene expression but, rather occur through membrane-associated receptors and second messengers (non-genomic actions)66 (see Figures 4a and 4b). Another mechanism of glucocorticoid-induced inhibition of inflammation involves decreased stability of messenger RNA (mRNA) for genes coding for inflammatory proteins, such as vascular endothelial growth factor and cyclooxygenase 2.67 Therefore, glucocorticoids act on diverse target tissues through multiple mechanisms to control inflammation.
Molecular Mechanisms of Glucocorticoid Resistance Associated with Systemic Inflammation and Chronic Inflammatory Diseases
Abnormalities of the Human Glucocorticoid Receptor
Research studies on T-lymphocytes and other target inflammatory cells have demonstrated several hGR abnormalities as potential mechanisms influencing response to glucocorticoid treatment in inflammatory conditions. Several studies have demonstrated that cytokines modulate glucocorticoid sensitivity. IL-1β, IL-2, IL-4, IL-6, IL-13, TNF-α and interferon (IFN)-γ alter hGR numbers and binding affinity,68–72 while IL-2 and IL-4 enhance the expression of hGRβ, a hGR splicing variant that acts as a dominant negative inhibitor of hGRα-mediated transactivation of glucocorticoid-responsive genes.73,74 Patients with steroid-resistant asthma may display decreased hGRβ binding affinity,69 abnormalities in hGR-AP-1 binding75 and increased expression of hGRβ.76 Increased hGRβ mRNA expression in peripheral lymphocytes has also been documented in patients with steroid-resistant ulcerative colitis (UC)77 and steroid-resistant rheumatoid arthritis.78
Inflammation and NF-κB
A substantial part of the anti-inflammatory actions of hGRα are mediated by its interference with the potent transcription factor NF-κB, which results in inhibition of the synthesis of cytokines and other gene products of the inflammatory cascade, such as chemokines, cell adhesion molecules, complement factors and receptors for these molecules. Glucocorticoids (GCs) penetrate the cell membrane readily and exert their effects through activation of the hGR, which translocates into the nucleus and influences NF-κB activity in five major ways: by physically interacting with the p65 subunit with formation of Adrenal Disordersan inactive (GC-hGRα/NF-κB) complex;79,80 by inducing the synthesis of the inhibitory protein IκBα via interaction with GREs in the promoter region of the IκB gene;79,80 by blocking the degradation of IκBα via enhanced synthesis of IL-10;79,80 by impairing TNF-α-induced degradation of IκBα;79,80 and by competing for limited amounts of hGRα co-activators, such as CREB-binding protein and steroid receptor co-activator-179,80 IκBα, in addition to maintaining NF-κB in an inactive state in the cytoplasm of cells also translocates into the nucleus, where it binds activated NF-κB complexes to induce their export to the cytoplasm65,79,80 (see Figure 4b).
On the other hand, NF-κB can repress the transcriptional activity of hGRα. In a regulated response, GC-hGRα activation is sufficient to maintain NF-κB levels in homeostasis and achieve a reduction in transcription of inflammatory mediators over time. Studies that examined the activities of GC-hGRα and NF-κB in patients with systemic inflammation and acute respiratory distress syndrome (ARDS) showed that those who improved on treatment with moderate doses of glucocorticoids had an excess activation of GC-hGRα compared to NFκB, as evidenced by the increased GC-hGRα binding to NF-κB and increased nuclear GC-hGRα binding (GC-hGRα-driven response). On the other hand, non-improvers demonstrated an excess of NF-κB activation compared with GC-hGRα, leading to protracted transcription of inflammatory mediators over time. In non-improvers, GC-hGRα binding to NF-κB was modestly increased, while nuclear NF-κB binding to its respective response elements increased substantially over time and nuclear GC-hGRα and cytoplasmic IκBα levels declined (NF-κBdriven response).79,80
Further studies that assessed the activity of NF-κB in biopsy specimens from steroid-resistant and steroid-sensitive patients with severe Crohn’s disease (CD) and UC showed that NF-κB activation was mainly noted in lamina propria macrophages in steroid-sensitive patients and in epithelial cells in steroid-resistant patients.81 Similar were the findings related to the activation of AP-1 and the upstream kinases p38 and JNK.82 The functional interference of these pro-inflammatory mediators with the glucocorticoid response was supported by reporter gene assays, which demonstrated that NF-κB, JNK and p38 inhibited the transcriptional activity of hGRα. These findings suggest that the generalised glucocorticoid resistance documented in patients with systemic inflammation, CD and UC is associated with constitutive activation of NF-κB and stress-activated protein kinases, which may inhibit the anti-inflammatory action of a limited number of hGRα molecules by reducing the transcriptional activity of hGRα. The Multidrug Resistance Gene
The multidrug resistance gene (MDR1) codes for a drug efflux pump P-glycoprotein-170, which is expressed on the apical surface of lymphocytes and intestinal epithelial cells and actively transports glucocorticoids and other drugs out of target cells, thereby reducing their efficacy. Increased expression of MDR1 has been demonstrated in peripheral T-lymphocyte and intestinal epithelial cells of patients with CD and UC who did not respond to glucocorticoid therapy.83 This suggests that a subset of patients with refractory CD and UC might escape effective immunosuppression by steroids and other immunosuppressive agents including cyclosporin because these drugs are MDR substrates and are effectively ‘pumped out’ of target cells. Increased T-lymphocyte MDR1 expression has also been demonstrated in patients with rheumatoid arthritis who require glucocorticoids,84 renal graft recipients who undergo graft rejection on cyclosporin therapy85 and patients with systemic lupus erythematosus.86 Specific MDR pump inhibitors (e.g. PSC 833) can significantly increase intracellular human intestinal epithelial and T-lymphocyte levels of cortisol and cyclosporin,87 and may have therapeutic applications in overcoming glucocorticoid resistance in many inflammatory conditions.
The repression of genes occurs through the reversal of histone acetylation that activates inflammatory genes.88 The ligand-bound hGRα may bind to CBP and other co-activators directly to inhibit their HAT activity,89 thereby reversing the unwinding of DNA around core histones and suppressing inflammatory genes. Furthermore, the activated hGRα may also recruit histone deacetylase 2 (HDAC2) to the activated transcriptional complex, which results in deacetylation of hyperacetylated histones and a decrease in inflammatory gene transcription.89 Several chronic inflammatory conditions such as asthma, rheumatoid arthritis, inflammatory bowel disease and chronic obstructive pulmonary disease (COPD) are characterised by a high degree of oxidative stress that decreases HDAC2 activity and leads to reduced glucocorticoid sensitivity.90 Indeed, there is a correlation between HDAC activity and the suppressive effects of glucocorticoids on cytokine release.90 The reduced HDAC2 expression in alveolar macrophages in patients with COPD was restored following overexpression of HDAC2 using a viral vector, and this effect was associated with restoration of glucocorticoid responsiveness in these cells. In contrast, transfection with an HDAC1 vector failed to restore corticosteroid responsiveness in COPD cells.91
Acetylation of Human Glucocorticoid Receptor
Non-histone proteins are also acetylated by histone acetyltransferases (HATs) and deacetylated by HDACs, and this may be an important mechanism of regulating their function.92 Acetylation of several nuclear receptors including the oestrogen and androgen receptors may affect binding to their ligands.93 Acetylation of the hGR occurs after ligand-binding and prior to nuclear translocation.91 The acetylated hGR is deacetylated by HDAC2 and this deacetylation is necessary for the hGR to be able to inhibit NF-κB activation of inflammatory genes.91 The site of acetylation of the hGR is the lysine-rich ‘hinge’ region 492–495 (sequence KKTK), which is analogous to the acetylation sites identified in other nuclear hormone receptors. Site-directed mutagenesis of the lysine residues K494 and K495 prevents hGR acetylation and reduces the activation of the SLPI gene by glucocorticoids, whereas repression of NFκB is unaffected.91
A reduction in HDAC2 activity prevents deacetylation of acetylated hGR so that glucocorticoids are no longer able to repress NF-κB-activated inflammatory genes, which required deacetylation of the liganded receptor.91 Moreover, this results in excessive acetylated hGR, which may then bind to GREs to induce genes responsible for the side effects of glucocorticoids. Therefore, a reduction in HDAC2 activity may not only lead to glucocorticoid resistance and poor disease control, but also to increased incidence of glucocorticoid side effects. Post-translational Modifications
Cytokines may also impair hGR action through post-translational modifications. Phosphorylation and dephosphorylation of the hGR plays. an important role in ligand binding, recycling, and turnover of the receptor. MAPKs, extracellular regulated kinases (ERKs), JNK, and p38 MAPK phosphorylate the hGR and inhibit hGRβ-mediated transacriptional activation.94,95 This may represent an early repression effect of mitogenic and pro-inflammatory signals on the expression of hGR-dependent genes.
In summary, glucocorticoid resistance associated with acute or chronic inflammation is characterized by insensitivity of the immune system to glucocorticoids. The underlying molecular mechanisms involve alterations in genomic and non-genomic glucocorticoid signaling mechanisms leading to impaired glucocorticoid signal transduction and inefficient suppression of the inflammatory process. The response to glucocorticoid treatment varies considerably among individuals depending on the type and severity of inflammation and the genetic background of the patient. Variations in the activity of the HPA axis and the stress response in general, as well as the presence of mutations or polymorphisms in the hGR gene, may also account for the variable response to treatment.
The literary work of this article was funded by grants from the Intramural Research Program of the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD), the National Institutes of Health (NIH), Bethesda, Maryland, 20892, US, the EU–European Social Fund, and the Greek Ministry of Development and General Secretariat of Research and Technology in Athens.