An established body of knowledge and clinical practice has argued in favor of the use of glucocorticoids in various chronic inflammatory and autoimmune diseases. However, the very well-known adverse effects associated with their treatment hampers continuation of therapy with glucocorticoids. Analyses of the molecular mechanisms underlying the actions of glucocorticoids have led to the discovery of several mediators that add complexity and diversity to the puzzling world of these hormones and anti-inflammatory drugs. Such mediators hold great promise as alternative pharmacologic tools to be used as anti-inflammatory drugs with the same properties as glucocorticoids, but avoiding their metabolic side effects. This review summarizes findings about the molecular targets and mediators of glucocorticoid function.
Glucocorticoid hormones (GCs) are the mainstay therapy for several inflammatory and autoimmune diseases: bronchial asthma, chronic obstructive pulmonary disease, Crohn’s disease, ulcerative colitis, immune-mediated glomerulonephritis, rheumatoid arthritis (RA), systemic lupus erythematosus, and multiple sclerosis [1–5]. The effective suppression of the inflammatory effects by GCs causes strong attenuation of symptoms, which, in some cases, may be life-saving. However, GCs are associated with deleterious side effects such as hypertension, hyperglycemia, osteoporosis, Cushing’s syndrome, and mood disorders [1,6,7]. All these adverse effects reflect the physiologic role of endogenous GCs in regulating metabolism, cell responses to stress, and the homeostatic control of the growth and survival of cells.
The synthetic GCs used most in therapy (prednisolone, methylprednisolone, betamethasone, dexamethasone) have negligible mineralocorticoid activity. However, they retain an inhibitory action on the hypothalamus–pituitary–adrenal (HPA) axis and the various metabolic effects that are responsible for the serious side effects that occur during anti-inflammatory therapy with GCs.
Although the use of GCs in clinical practice has been consolidated for a long time, most of the molecular mechanisms of actions of these hormones have been defined only recently. In this review, we present the anti-inflammatory effects mediated by GCs in the cells of the immune system.
Molecular mechanisms of GCs
Virtually all the cellular effects induced by the physiologic and pharmacologic concentrations of GCs are mediated by interactions with the glucocorticoid receptor (GR), a transcription factor that belongs to the nuclear receptor superfamily . Upon ligand binding, the GR migrates to the nucleus and binds to GC-response elements (GREs) in the promoter regions of target genes, thereby leading to the positive or negative regulation of the transcription of hundreds of genes . The binding of activated GR to positive GRE results in gene activation; on the other hand GR can also repress transcription by binding to negative GRE (nGREs) [10,11]. To note, an emerging complexity on GR binding pattern is due to the fact that, despite GR is present in almost all tissues of our body, it has a widely distinct cell type-specific effects on different tissues depending also on variation of epigenetic mechanisms such as chromatin accessibility on GRE . Furthermore, GR may also associate with other transcription factors, thus regulating their transcriptional regulation. Long-lasting efforts to separate activation of beneficial genes from harmful genes by modulating GR activity has not yielded a definitive solution, and GC toxicity remains a major issue in clinical practice. It is important to state that the complexity of the transcriptional activity of GR, depending on activation or repression of target genes, is object of intensive research to identify selective GR agonists (SEGRAs). In this way, ‘dissociated’ forms can be formed to carry out the repression of transcription factors such as AP-1, with consequent anti-inflammatory effects, without modulating gene activation, which is probably responsible for the side effects of GCs [13,14]. The in vivo efficacy of SEGRA compounds with reduced side effects has been studied for the treatment of inflammatory diseases such as RA, asthma, and colitis [15,16]. However, the idea that GR transrepression compared with transactivation activities may be related to immunosuppressive or metabolic effects (side effects), respectively, is clearly an oversimplification of the molecular mechanisms that regulate anti-inflammatory effects of GC and despite some SEGRA molecules have been shown to have anti-inflammatory properties with reduced side effects, their actual potential as therapeutic agents is still a matter of debate.
A huge number of GR transcriptional effects contribute to the anti-inflammatory and immunoregulatory actions of GCs [1,5,17]. For example, GCs up-regulate expression of anti-inflammatory molecules such as lipocortin-1, glucocorticoid-induced leucine zipper (GILZ), and Ik-B, whereas they down-regulate transcription of genes coding for pro-inflammatory cytokines and adhesion molecules such as tumor necrosis factor (TNF)-α, interleukin (IL)-1, and intercellular adhesion molecule (ICAM)-1. However, these mechanisms might not be involved in the more rapid anti-inflammatory/immunoregulatory effects of GCs. Indeed, several rapid, non-transcription-dependent (non-genomic) events have been suggested to mediate (at least in part) the pharmacologic effects of GCs. Non-genomic actions are characterized mainly by a short delay in action and often involve generation of intracellular second messengers and various signal-transduction cascades, such as: phospholipase activation; modulation of cAMP; protein kinase pathways; Ca2+ mobilization [18–20]. Actinomycin-D and cycloheximide, which are inhibitors of protein synthesis, do not reverse these non-genomic effects. These effects require GC/GR interaction, suggesting that most probably the GR has other functions apart from its classic role as a transcription factor .
Two human isoforms of the GR have been described: α and β . GRα (the classical form of the receptor) is made up of 777 amino-acid residues and can bind steroids; GRβ consists of 742 amino-acid residues and, unlike GRα, it does not bind steroid ligands and cannot target the genes regulated by GCs [23,24]. GRβ is induced by inflammatory pathways such as TNF-α and nuclear factor-κ B (NF-κB), and GRβ overexpression is related to immunologic diseases such as ulcerative colitis, asthma, and chronic sinusitis [25,26]. In contrast, other studies have shown that GC sensitivity cannot be ascribed to the variation in the GRα and GRβ expression [27,28]. Thus, it is not clear yet how GRα/GRβ expression ratios in tissues and cells may predict GC sensitivity or resistance.
The GR is expressed ubiquitously, but GCs in vivo regulate transcription in a promoter- and tissue-specific manner. The DNA sequence in the promoter region and the possible binding of other transcription factors linked to the same promoter define the specificity for the transcriptional modulation mediated by the GR . Moreover, the regulation mediated by the GR also contributes to the modulation of other transcription factors . For example, the GR can bind transcription factors such as AP-1 , NF-κB , signal transducer and activator of transcription (STAT)-3 , STAT-5 , and cAMP response element-binding protein (CREB) , thus preventing activation of these factors and repressing expression of the genes placed under their control. Alternatively, competition for cofactors such as p300/CREB-binding protein (CBP)  or binding with factors SMRT/NCoR  have been proposed as additional mechanisms of modulation of GR transcriptional activity. The effects of GC-mediated up- or down-regulation of the target genes involved in the control of inflammatory processes are discussed below.
Mechanistically, another level of complexity of GC/GR activities is given by post-translational modifications affecting the GR, which contains several phosphorylation sites, at least one ubiquitination site, and several sumoylation sites, responsible for selective interaction with transcriptional cofactors and/or subcellular localization . This capacity to generate dozens of unique GRs in a single cell and tissue presents enormous potential for signaling diversity [38,39] in terms of transcription regulation as well as non-genomic signals .
GCs in the regulation of inflammatory mediators
Inflammation is a complex process that occurs in response to noxious stimuli such as pathogenic infection or extracellular damage. The acute response that follows must heal the inflamed site by recruiting cells from immune and vascular systems. If this process is not regulated strictly, the inflammation can become chronic and favor the development of autoimmune diseases, chronic inflammatory diseases, and even cancer.
During the early onset of inflammation, GCs are released in the circulation as a result of triggering of the HPA axis by the pain signals that reach the brain, and regulate the cytokines produced as part of the immune response [41–44]. Thus, GCs can ‘extinguish’ the inflammatory process by controlling the transcription of several hundreds of genes, including many pro-inflammatory genes. This is the reason why they are considered the most efficacious and widely used anti-inflammatory drugs, and represent the first line of treatment in many autoimmune and chronic inflammatory diseases.
However, long-term treatment with GCs entails not only health-damaging side effects and metabolic problems, which are the result of their complex mechanism of action, but also the onset of GC resistance. Hence, it is important to unravel how GCs function at the molecular level, even considering the heterogeneity of GR isoforms, to uncouple the clinical benefit from their detrimental side effects. This is extremely necessary taking into account that inflammation continues to be a burden on general health, with consequent healthcare costs for governments.
Inflammation is characterized by the recruitment of neutrophils and macrophages to the site of injured tissue and the production of pro-inflammatory cytokines and chemokines such as TNF-α and IL-1. These cytokines are released by macrophages upon engulfing a particle or by mast cells in damaged skin and, together with IL-6 (released by cells of the innate immune system), fibroblasts and endothelial cells, induce the liver to begin the acute-phase response . Furthermore, IL-6 can regulate the transition from neutrophil recruitment to monocyte recruitment during inflammation . TNF-α, IL-1, and IL-6 harbor in their gene promoter regions the sequences for NF-κB binding that can activate these genes. GCs, once bound to the GR, can bind directly to nGREs in the promoter regions of pro-inflammatory genes and even bind to NF-κB to prevent the transcription of TNF-α, IL-1 and IL-6, in a process referred to as ‘tethering’ [32,46]. Importantly, IL-1 and IL-6 are involved in an oncogenic process that follows a prolonged inflammatory status, as discussed below .
Recently, an additional mechanism of NF-κB inhibition by GCs was found in the interaction with the p53 molecule, which can control NF-κB repression by the GR . Similar to NF-κB, the GR/GC complex can bind to AP-1 heterodimers, formed by c-Fos/c-Jun, to repress transcription of target pro-inflammatory genes by direct binding to the c-Jun subunit or to the promoter region of c-Jun [13,31]. A common additional mechanism of inhibition of NF-κB and AP-1 is represented by GILZ, which can bind to NF-κB or AP-1, thereby mediating the GC-based inhibition of downstream pro-inflammatory genes [49–51]. Chemokines such as CXCL-1, IL-8, and CXCL-2 are repressed by GCs similar to cytokines, thus inhibiting the recruitment of mononuclear cells at the inflammation site and contributing to prevention of an overexuberant inflammatory response [52,53].
The reduction in NF-κB activity by GCs can be mediated by histone deacetylases (HDACs) such as HDAC2, which reduces histone acetylation at the activated inflammatory target gene. In addition, the GR becomes acetylated upon ligand binding, thus resulting in GRE binding . Therefore, very complex regulation at the transcription level occurs if the GR interacts with other transcription factors and regulators for repression of expression of pro-inflammatory genes.
An important player in inflammation is the inducible cyclooxygenase (COX)-2 enzyme whose products, prostaglandins (PGs), enhance vasodilation, edema formation, and vascular permeability. PGE2 is a potent pyretic agent. GCs can bind to c-jun proteins and suppress IL-1-mediated expression of COX-2 mRNA as well as regulate expression of COX-2 protein by modifying the stability of COX-2 mRNA [55,56]. Recently, another mechanism of suppression by GCs on COX-2 has been described: the inhibition of hypoxia-induced COX-2 by GILZ in human pulmonary epithelial A549 cells . COX-2 is not the only enzyme regulated by GCs and involved in the inflammatory process. Inducible nitric oxide (NO) synthase (iNOS), the expression of which is triggered by pro-inflammatory cytokines (TNF-α, IL-1β, interferon (IFN)-γ), releases NO in tissues, which takes part in the ‘resolution phase’ of inflammatory cells by inducing apoptosis. Conversely, NO helps to enrich the microenvironment of oxidants, thus creating persistent oxidative stress that can drive the development of some cancers [58,59]. Similar to COX-2, iNOS expression is suppressed by GCs via NF-κB inhibition .
Another important player in the cascade of the lipid mediators that take part in the inflammatory process is phospholipase-A2. This enzyme releases the arachidonic acid (AA) stored in the membrane phospholipids when they are activated by a microorganism or any other injury . AA is the substrate of distinct enzymes that produce eicosanoids, including COX. GCs can down-regulate phospholipase-A2 expression to dampen the inflammatory cascade .
The mitogen-activated protein kinase (MAPK) signaling pathway plays a major part in regulation of the inflammatory process. It contributes to control of the production of pro-inflammatory cytokines such as TNF-α and IL-6, and inflammatory mediators such as PGs and NO. The dephosphorylation of enzymes belonging to the MAPK pathway, specifically p38 MAPK, c-Jun N-terminal kinase (JNK), and extracellular signal-regulated kinase (ERK), takes place thanks to mitogen-activated MAP kinase phosphatases (MKPs). Amongst these proteins, MKP-1 inhibits p38 MAPK and JNK, thereby blocking the entire pathway. GCs can control MKP-1 expression by two mechanisms: (i) directly by up-regulating MKP-1 via direct binding to the promoter, thus inducing MKP-1 transcription; (ii) indirectly by up-regulating GILZ, which in turn can inhibit ERK, p38, and JNK MAPKs and induce MKP-1, as demonstrated in endothelial cells . Up-regulation of MKP-1 expression prevents activation of all the downstream pro-inflammatory effectors of the MAPK pathway .
An opposite (though equally important) mechanism in the control of the inflammatory response by GCs at the transcriptional level is up-regulation of anti-inflammatory genes. IL-10 is one of the most important anti-inflammatory cytokines that plays a pivotal part in immunosuppression. IL-10 has been found to be induced by GCs in cultures of whole human blood, airway cells of asthmatic patients, and in polyclonal-stimulated T cells, although the exact mechanism of how it occurs is not known [65,66]. There may be a pathway involving activation of p38 MAPK (which can up-regulate IL-10) and account for the indirect mechanism of up-regulation of IL-10 by GCs, or direct binding of the GR on GREs in the IL-10 promoter (human and murine) . An additional (though indirect) contribution comes from the induction by GCs of regulatory T cells (Tregs), which in turn secrete IL-10, thus facilitating the onset of an immunosuppressive environment [68–71]. GCs can also promote expression of a non-signaling, type-II IL-1 receptor antagonist which inhibits the responses to IL-1 by acting as a ‘decoy’, thus contributing to inhibition of the inflammatory reactions that depend on IL-1 activation .
In addition to the genomic mechanisms elicited classically by GCs via transcriptional repression or tethering in combination with other transcription factors, novel GR-mediated post-transcriptional regulations have been discovered. Initially, modulation of expression of the tristetraprolin protein, which can bind mRNA in AU-rich elements in the 3′-UTR and promote its degradation, was found to account for an additional means to repress pro-inflammatory cytokines such as TNF-α via mRNA destabilization . A plethora of other players in inflammation were found to be deregulated through this mechanism: vascular endothelial growth factor, IFN-β, IL-1α, IL-1β, IL-6, iNOS, and COX-2 [67,73]. More direct regulation of mRNA stability was discovered for the mRNA of monocyte chemoattractant protein (MCP)-1 independently of the AU-rich regions of 3′-UTRs. MCP-1 mRNA was demonstrated to be bound specifically by the GR at the 5′-end, thus undergoing degradation, in arterial smooth muscle cells . This is another anti-inflammatory mechanism because MCP-1 plays an important part in attracting monocytes to inflammation sites.
More recently, a new mechanism of mRNA stability regulated by the GR in a ligand-dependent manner was discovered and termed as ‘GR-mediated mRNA decay’. This efficient degradation of mRNA requires a protein complex containing GR, PNRC2, UPF-1, and DCP1A. The exact mechanism through which this complex promotes degradation of the target mRNA is not known. Importantly, one of the target mRNAs is represented by CCL2 (a member of the CC chemokine family), which mediates the migration of monocytes at inflammation sites .
Recently, another way to control the inflammatory status by GCs was discovered that adds a degree of complexity and diversity to the mechanisms of actions of these compounds. Regulation of gene expression occurs via several pathways, but the last decade has witnessed growing interest in a family of non-coding RNAs: miRNAs . Amongst the distinct biologic functions regulated by miRNAs is the control of inflammation. GCs were found to control expression of various miRNAs, and particularly to inhibit expression of miR-155, an miRNA with pro-inflammatory properties, in response to lipopolysaccharide in a macrophage cell line . As another example, a potent down-regulator of TNFR1, miR-511, is strongly induced by GCs, and thus protects against TNF in mice overexpressing miR-511 and appears to be an anti-inflammatory agent . A very recent report described up-regulation by GCs of the mature miRNA miR-511-5p (located at the 5′-end of the pre-miRNA) in monocytes, which made miR-511-5p a direct target responsible for GC-mediated inhibition of pro-inflammatory cytokines in endotoxin-tolerant human monocytes . A new regulatory function of GC-controlled miRNA in the context of inflammation is emerging, thus complicating understanding of the gene regulation operated by GCs.
To avoid the progression from acute to chronic inflammation, the inflammatory response has evolved a finely tuned, active process: ‘resolution of inflammation’ . Several players take part in this process, but the most characterized is the protein Annexin-a1 (Anxa-A1), a classically GC-induced anti-inflammatory protein [81,82]. Anxa-A1 is known to specifically inhibit neutrophil adhesion to endothelial layers, thus acting as a ‘brake’ for excessive transmigration of cells . Until very recently, the mechanism by which GCs up-regulate AnxaA1 was not known. However, studies by our research team identified GILZ as the mediator of Anxa-A1 up-regulation by GCs via an interaction with the transcription factor PU.1 . The control of neutrophil migration by GCs through this mechanism joins the puzzling and complex context in which GCs exert their anti-inflammatory actions. A summary of the modulatory effects of GCs on described targets is given in Figure 1.
A simplified schematic of GC multiple effects on the transcription of distinct target genes and miRNAs
Endogenous GCs, chronic inflammation, and cancer
The consequence of an acute inflammatory stimulus is a surge in serum levels of adrenocorticotropic hormone (ACTH) and cortisol released from the HPA axis in response to circulating inflammatory cytokines such as TNF-α, IL-1β, and IL-6. This effect is necessary for permitting the availability of energy stores for mobilization of the cells of the immune system to the inflammation site. Experimentally, if the presence of inflammatory cytokines persists, the HPA axis cannot release additional amounts of hormones. This effect seems to be related to the risk of sepsis and inadequate immune response to infections that may arise due to the strong immunosuppression elicited by released GCs [85–87]. Therefore, the persistence of chronic stimuli in inflammatory diseases such as RA and inflammatory bowel disease (IBD) causes inappropriate and insufficient HPA responses, so that patients with these diseases respond symptomatically to exogenous GCs . Even though adrenal insufficiency has not been observed in such patients, the circulating cortisol level could be below a critical level . In this context, local production or degradation of cortisol plays an important part. In patients with RA, the enzyme 11β-HSD1, which activates cortisol, has been found to be up-regulated potently in fibroblasts whereas, in macrophages and lymphocytes, higher 11β-HSD2 activity has been observed (i.e. cortisol degradation) . Finely tuned regulation of local production of cortisol can, therefore, drive the fate of inflammation independently of low circulating levels of hormones.
One out of four cancers develop because of a persistent inflammatory stimulus, whatever the cause, either infection or chronic inflammation. The unresolved inflammatory status generates oxidative stress with consequent reactive oxygen species (ROS) production that causes DNA damage and increases the risk of developing cancer . Moreover, ROS can inhibit cell-to-cell communication and increase cell proliferation by inducing changes in several transcription factors, including NF-κB, STAT3, hypoxia-inducible factor (HIF)-1α, AP-1, and genes such as p53 . Several other inflammatory players take part in carcinogenesis, including cytokines, PGs and COX2; their mechanisms of actions are beyond the scope of this review but have been described exhaustively [92–94]. In this context, in which the local inflammatory microenvironment establishes a ‘cross-talk’ with systemic inflammatory players, whether the effects of GCs favor or inhibit the growth of non-hematologic cancers is controversial . Nonetheless, activation of the HPA axis and CNS with consequent release of GCs and catecholamines has been studied with regard to the link between stress and cancer onset . Few studies have reported the causal relationship between stressful conditions and the growth and progression of cancer. Nevertheless, a link has been found involving the release of stress-induced cytokines such as IL-6 and the recruitment of inflammatory cells in the tumor microenvironment, and the consequent activation of the transcription factors of NF-κB and STAT3 favors the survival, proliferation, invasion, angiogenesis, and metastasis of tumor cells .
As drugs, GCs are administered usually as an aggressive anti-inflammatory treatment upon chemotherapy and to counteract the toxic effect of the latter. However, inhibition of a single inflammatory player by non-steroidal anti-inflammatory drugs is more efficacious and safer than GCs for the treatment of some cancers, such as colorectal cancer. Conversely, the ability of GCs to induce cell death in hematologic cancer cells remains first-line treatment for these tumors, although the development of resistance is a problem [17,98].
GCs and cells of the immune system
Innate immune cells
The GR is expressed by almost all nucleated cells, but the effects of GCs differ according to the cell type. This is the case for cells of the innate immune system, which are differentially sensitive to the effects of GCs. As a part of the first line of defense, dendritic cells (DC) are sensitive to GC effects throughout their lifecycle, and they are sensitive to GC-induced apoptosis only before maturation induced by an encounter with an antigen. Interestingly, GCs stimulate antigen uptake by DCs, thus helping to mount a defensive response upon infection. This action makes DCs tolerogenic, thereby down-regulating expression of MHC-II molecules, co-stimulatory molecules, and cytokines such as IL-1, IL-6, and IL-12 [99,100]. GCs can also inhibit DC maturation induced by activation of toll-like receptor (TLR)7 and TLR8 .
Monocytes and macrophages are influenced by the action of GCs. High doses of GCs suppress the intracellular signaling cascade of MAPKs in monocytes and macrophages, thus exerting their anti-inflammatory effects by inhibiting the transcription of pro-inflammatory cytokines such as IFN-γ, IL-1α, and IL-1β. Moreover, suppression of these pro-inflammatory cytokines occurs via the direct interaction of GC/GR with transcription factors such as AP-1 and NF-κB at the promoter regions of target genes [100,102,103].
Interestingly, GCs can drive the differentiation of monocytes (human and murine) toward a specific anti-inflammatory subtype, which can migrate rapidly into inflamed tissues to help resolve inflammation [104,105]. A contribution to inflammation resolution is also given by the protection from apoptosis GCs exert on macrophages via prolonged activation of the ERK/MAPK pathway, which results in inhibition of caspase activities . Such a mechanism provides further means for the anti-inflammatory phenotype of macrophages. However, low doses of GCs have an immunostimulatory effect on macrophages by inducing the production of pro-inflammatory cytokines and NO, thereby exerting distinct effects on macrophage function dependent on their concentration [107,108]. Recently, another physiologic role of GCs on macrophages was found: GCs induce differentiation of human monocytes toward macrophages that can interact with erythroid cells to form clusters which recapitulate aspects of ‘erythroblastic islands’ and resemble bone marrow and fetal liver-resident macrophages . The understanding of GC-specific function in macrophages has been further revealed by studies on mice with a myeloid cell-specific GR deletion. GRlysM mice, in which myeloid cell types lack a functional GR, including macrophages, have been used in mouse models of acute GVHD, DSS-induced colitis, and contact hypersensitivity (CHS). These studies demonstrated that endogenous GCs are necessary (i) to induce repression of macrophage-produced inflammatory cytokines elicited in the GVHD disease, (ii) to be important in myeloid cells for the induction of tissue repair mechanisms after intestinal tissue damage in DSS-induced colitis, and (iii) to down-regulate the inflammatory response by acting on macrophages in a model of CHS [110–112]. Therefore macrophages appear to be pivotal players in the control of the inflammatory status by therapeutic GCs.
GCs can control granulocyte trafficking by reducing expression of selectins on neutrophils, and of integrin receptors on endothelial cells, thus limiting the inflammatory response. In addition, neutrophils are prevented from migrating into inflamed tissues by GCs via the up-regulation of Anxa-a1. The control of Anxa-a1 expression occurs with the intermediation of GILZ, as described above . Furthermore, neutrophils are ‘privileged’ cells because they are protected from the apoptosis induced by GCs, differently from all other immune cells. Their survival is a stress-response first-line of defense which helps the organism fight against an infectious threat (as discussed below), whereas other immune cells die due to the effects of released GC.
Conversely, granulocytes such as eosinophils and basophils undergo GC-induced apoptosis, and it occurs in a Fas-mediated pathway . In particular, in basophils, GCs inhibit release of histamine and IL-4, which is another means to limit the inflammatory process, especially in allergic reactions . In such a context, mast cells are also central targets of the effects of GCs. IL-4 and IL-5, which are essential for IgE production and eosinophilia, are released by mast cells in allergic conditions, but can be reduced markedly by GCs . Furthermore, GCs have been shown to inhibit degranulation of airway mast cells significantly in an experimental model of allergic asthma, and to inhibit IgE-mediated exocytosis and histamine release of these cells rapidly via a non-genomic mechanism .
Natural killer cells
Amongst innate cells, natural killer cells (NKs) exert important functions, such as induction of the lysis of viral-infected or cancer cells. Recently, a ‘double control’ role of NKCs by GCs was described: GCs enhance the proliferation of primary human NKCs stimulated with IL-2 plus IL-12, while protecting NKCs from cytokine-induced death. Simultaneously, IFN-γ production is suppressed initially but can be increased in a ‘secondary recall’ response depending on the local production of cytokines . Similar to these findings, treatment with a low concentration of GCs was found to reduce the cytolytic activity of NKC cells but, simultaneously, to increase histone acetylation in the regulatory regions of IFN-γ and IL-6 . This is another example of the dichotomous effect of GCs which, besides possessing classical anti-inflammatory functions, can ‘prime’ and stimulate immune cells to also exert pro-inflammatory effects for a prompt response of the immune system, as discussed below.
Adaptive immune cells
GCs decide if a thymocyte survives or dies. This decision is the result of the predominant stimulus because TCR engagement (which is how thymocytes are positively or negatively selected in the thymus) can antagonize GC-induced apoptosis if both stimuli are present [46,119]. Conversely, it has been demonstrated that GR signaling is not essential for the development or selection of T cells in the thymus because GR-deficient thymocytes can develop and mature in the absence of GR signaling . Understanding the complex function and regulation of GCs in the control of thymocyte maturation requires appropriate in vivo mice models with specific GR mutations. Studies on transgenic GR overexpressing mice demonstrated that thymocytes with an increased GC expression are highly sensitive to GC-induced apoptosis, and peripheral T cells are reduced in number , whereas GR knockout mice show a normal number of thymocytes with resistance to GC-induced apoptosis .
Mature T cells can undergo or be protected from GC-induced apoptosis depending on the time of steroid exposure and their activation phase. One of the mechanisms of protection from apoptosis elicited by GCs involves prevention of expression of the FasL, which usually induces cell death by binding to its cognate receptor, Fas, on T cells soon after activation . Another powerful effect on T cells is the selective induction or suppression of specific T-cell subtypes by GCs. Amongst the distinct T-cell populations, T-helper (Th)1 cells are susceptible to GC-induced apoptosis, whereas Th2 and Th17 cells are resistant. The resistance of Th17 to GC action depends on the expression of BCL-2, and it was demonstrated in a murine severe asthma model. However, Th17 cells were found to be sensitive to GC-induced apoptosis, thus the exact effect GCs elicit on these cells is still controversial [46,124]. Interestingly, Tregs have been demonstrated to be sensitive and resistant to GC-induced apoptosis. The resistance of Tregs to GC-induced apoptosis is a means of immunosuppression because GCs can also increase the number of Tregs and their ability to produce the immunosuppressive cytokine IL-10 . In thymocytes, GC-induced GILZ expression inhibits Fas/FasL expression and protects against activated-induced cell death (AICD) . In mice, GILZ is an intermediate player in the GC-induced increase in Tregs because it can co-operate with TGF-β in the induction of FoxP3 expression . Furthermore, the interaction between the FoxP3 transcription factor and the GR is responsible for the increase in Tregs number in the thymus and accounts for the production of the immunosuppressive cytokines IL-10 and TGFβ . Therefore, several mechanisms that can control Treg number by GCs exist and, if this complex regulation fails, autoimmune diseases or cancer can develop.
Besides deciding the life or death of T-cell subtypes, GCs can drive their differentiation by selectively suppressing or promoting secretion of specific cytokines. GCs cause a shift from Th1 to Th2 immunity at physiologic doses by suppressing the production of IL-12, IFN-α, IFN-γ, and TNF-α by antigen-presenting cells (APCs) and Th1 cells, but inducing the production of IL-4, IL-10, and IL-13 by Th2 cells, thus triggering their differentiation [1,127]. This is one of the reasons why GCs are used in the treatment of diseases in which the Th1/Th2 ratio is unbalanced, such as asthma or IBD [128,129].
The effects of GCs on B cells have become clearer only recently. B cells, like T cells, and throughout their lifespan, are sensitive to GC-induced apoptosis, as well as to transformed B cells. Conversely, B-lymphoblastic leukemia cells are resistant, due to enhanced expression of B-cell lymphoma-2 protein [130,131]. Interestingly, GCs appear to use an intermediate protein to exert their effects in B lymphocytes: GILZ is the mediator because, in GILZ-deficient mice, an accumulation of B cells was found in the bone marrow, blood, and lymphoid tissues, along with reduced apoptosis of B cells . Therefore, GCs play an important part in the regulation of B-cell survival and also in the regulation of their main function: the humoral response. Indeed, GCs can increase IgE production in conjunction with IL-4 depending on the increased transcription of cluster of differentiation-40L [133,134]. Moreover, physiologic concentrations of GCs can reduce activation-induced levels of cytidine deaminase (AICDA) mRNA. AICDA is the principal regulator of Ig gene somatic hypermutation and class-switch recombination in B cells, so GCs can regulate humoral responses by down-regulating this protein, though the exact mechanism of this GR-mediated effect is not known .
Pro-inflammatory properties of GCs
As opposed to the known and anti-inflammatory effects of GCs described above, several evidences suggest that GCs are pro-inflammatory. Some studies have demonstrated that these hormones favor TLR expression, in contrast with other studies that had found down-regulation of TLRs by GCs via MAPK-1 induction [136,137]. TLRs are important sentinel receptors and have pivotal roles in protecting the organism against invading pathogens or noxious molecules released endogenously. With enhanced expression of TLRs (including TLR2), GCs prepare the body to fight against a hazardous stimulus. In this context, GCs are also involved in priming of the inflammasome, which has a key role in IL-1β maturation, by up-regulating NLRP3, a central component of the inflammasome. NLRP3 in turn sensitizes macrophages to ATP, with the resulting production of pro-inflammatory cytokines such as IL-1β and TNF-α . Such a pro-inflammatory role seems to function physiologically in the central nervous system, in which stress-induced GCs can shift the neuroimmune microenvironment toward a microglial activation state, thus priming the neuroinflammatory response (Figure 2A) .
Pro-inflammatory properies of GCs
The shift from acute to chronic stress causes a shift from activation of the immune system to its suppression. This is because GCs respond to external threats, such as pathogens, by initially activating the immune system and then by repressing this activation, with the aim of limiting the potential damage derived from a prolonged immune reaction . Hence, acute stress-released GCs reduce blood leukocyte counts but this is not an apoptotic effect, instead a redistribution of circulating leukocytes to tissues where they may be needed . Although not completely understood, the mechanisms that regulate immune cells and inflammation by GCs are finely regulated, and GCs act depending on the time of the exposure or the nature of the stimulus (physiologic or pathologic), resulting in a GC response that is ‘tailored’ to face any situation (Figure 2B). Unraveling these mechanisms will help find the appropriate pharmacologic tools to take advantage of the properties of GCs to be pro-inflammatory and anti-inflammatory.
Resistance to GCs: molecular mechanisms
The lack of susceptibility to the effect of GCs is a problem concerning the treatment of chronic inflammatory diseases such as asthma, RA, and systemic lupus erythematosus . Given the clinical relevance of this phenomenon, it is important to delineate the possible molecular mechanisms underlying this resistance to study new approaches for the treatment of these diseases.
Several studies have shown that a direct correlation between the GR concentration in a cell and susceptibility to steroid treatment exists [143,144]. For example, GRβ expression is increased in inflammatory cells (particularly T cells, eosinophils, and macrophage) in nasal polyps, and this overexpression of the GRβ isoform correlates with a poor response to treatment with fluticasone propionate . In patients with asthma, the ratio between the α and β isoforms of the GR is altered in favor of the GRβ isoform in lymphocyte cultures of patients with generalized steroid resistance . In other cases, the GC resistance was dependent upon variation in expression of ‘co-chaperone molecules’ that bind the GR, such as FK506-binding protein-51 .
Conversely, genetic alterations of the GR have been described, and involve mainly inactivating mutations of the GC-binding domain, which implies generalized resistance to GC activity compensated by hyper-reactivity of the HPA axis . The excessive production of ACTH, cortisol, cortisol precursors with mineralocorticoid (especially corticosterone), and adrenal androgen (especially androstenedione) activity can cause various diseases in these individuals, such as hypertension and hyperandrogenism. A negative, pathologic mutant of the human GR has been described due to substitutive mutation of Asp559 with isoleucine. Patients with this mutation show marked resistance to steroid treatment, with a five to ten fold increase in levels of free cortisol in the urine . In IBD, GC resistance is associated with polymorphisms or down-regulated expression of the GR [149,150].
In summary, the resistance to steroid treatment, studied very extensively for asthma but less extensively for other inflammatory diseases, may be due to five main molecular mechanisms: (i) variation in GRα expression; (ii) variation in GRβ expression; (iii) post-transcriptional modifications of the GR; (iv) alterations of the ability of the GR to dissociate from the cytoplasmic macromolecular complex (receptosome) and/or to move at the nuclear level; (v) interaction of the GR with proteins with transcriptional and non-transcriptional activity (cytoplasmic and nuclear).
In conclusion, there are numerous mechanisms responsible for steroid activity and the actual contribution of each of them in determining the GC effect in inflammation needs further studies.
The authors declare that there are no competing interests associated with the manuscript.
activator protein 1
activation-induced levels of cytidine deaminase
B cell lymphoma 2
chemokine (C-C motif) ligand 2
central nervous system
cAMP response element-binding protein
hemokine (C-X-C motif)ligand
dextran sulfate sodium
mRNA-decapping enzyme 1A
extracellular signal-regulated kinase
Forkhead box P3
glucocorticoid-induced leucine zipper
graft vs host disease
inflammatory bowel disease
inhibitor of kB
inducible nitric oxide synthase
c-Jun N-terminal kinase
mitogen-activated protein kinase
monocyte chemoattractant protein
MAP kinase phosphatase
nuclear receptor corepressor 1
natural killer cell
NOD-like receptor family pyrin domain containing 3
proline-rich nuclear recptor coactivator 2
reactive oxygen species
selective GR agonist
silencing mediator of retinoid and thyroid receptors
signal transducer and activator of transcription
tumor necrosis factor
tumor necrosis factor receptor 1
regulatory T cell
regulator of nonsense transcripts 1