Glucocorticoids (GCs) have a central role in the treatment of many inflammatory diseases because of their antiphlogistic and immunosuppressive properties. However, GC signaling also affects metabolic pathways. As a result, long-term treatment can result in severe adverse effects, including but not limited to muscle wasting, osteoporosis, and insulin resistance. The classic mechanisms of GC actions are mediated through the GC receptor (GR), which belongs to the nuclear receptor family and is located in the cytosol. In the absence of GC hormone, GR is retained in the cytoplasm in a complex with other proteins, poised to bind its ligand. Upon hormone ligation, this complex dissociates, inducing nuclear translocation of the receptor. Upon arrival in the nuclear compartment, GR can mediate genomic regulation of target genes by 2 different mechanisms.
The first mechanism by which the GR can mediate genomic regulation is through direct interaction with regulatory regions on DNA, termed GC response elements (GREs). GREs can be positive, indicating that binding of the GR to the GRE elicits messenger RNA (mRNA) transcription called transactivation. GREs can also be negative, leading to transrepression. The second mechanism by which the GR can mediate genomic regulation is by interacting with other transcription factors, such as proinflammatory NF-κB and activator protein 1, via protein–protein interactions, thus preventing them from interacting with their binding site on DNA, also leading to transrepression (Figure 1, left side). It is thought that most of the metabolic effects of GCs are mediated through the first pathway, while most of the antiinflammatory effects are mediated through the second pathway. An example of the metabolic actions caused by transactivation is expression of the urea cycle enzyme carbamoylphosphate synthetase, in which the GR can physically interact with its binding region on the gene (1). Repression of the expression of the interleukin-6 gene serves as an example of the antiinflammatory properties of GCs through transrepression by interaction with NF-κB signaling (2).
The process of genomic signaling involves transcription factor binding to its regulatory region, followed by the initiation of transcription, splicing, translation, and posttranslation modification. Given the fact that nuclear translocation of liganded GR requires 10–30 minutes, and transcription requires 5–120 minutes, the entire process will take at least 15 minutes (and in most instances hours) to complete, indicating a relatively time-consuming process (3).
Nongenomic GC signaling
It has long been recognized that some GC signaling events occur much faster and are of shorter duration than would be expected on the basis of genomic signaling. It has been postulated that besides the genomic mechanism, GCs could also act on diverse downstream targets, bypassing nuclear signaling (Figure 1, right side). Several of these targets have been identified, among which are intracellular proteins such as kinases, including MAP kinase (MAPK), phosphatidylinositol 3-kinase (PI 3-kinase), and protein kinase C (PKC) (4, 5). A different mechanism is used by altered function of membrane-bound proteins such as ion channels (6). Furthermore, GCs have been shown to act on membrane proteins including protein G–coupled receptors (7). The notion of yet another possible nongenomic mechanism, namely a membrane-bound GR, was introduced ∼25 years ago (8). It is now realized that a pool of membrane-bound GRs exists that could at least in part mediate nongenomic effects. Despite efforts, little is known pertaining to the identity, origin, and function of this membrane-bound GR and its role in GC signaling.
In this issue of Arthritis & Rheumatism, Strehl and colleagues report new findings concerning membrane-bound GRs (9). The aim of their study was to elucidate the origin and functional activity of membrane-bound GRs. The authors demonstrate that both cytosolic GR and membrane-bound GR are encoded by the same human GR gene. To address the possibility of a common origin of membrane-bound and cytosolic GRs, they generated stable knockdown of GRs, using short hairpin RNAs (shRNA) against 3 different GR exons in HEK 293T kidney cells. This resulted in a significant reduction of GR expression at both the protein and mRNA levels. For the assessment of membrane-bound GRs, Strehl et al relied on a liposome-based enhancer system and observed a reduction in mean fluorescence intensity, although the number of positive cells did not decrease. Because the shRNA-mediated knockdown resulted in a decrease in both cytosolic and membrane-bound GRs, the authors concluded that both receptor types originate from the same gene.
After establishing that bovine serum albumin (BSA)–coupled dexamethasone (DEX-BSA) was indeed acting only on membrane-bound GRs, the authors next tried to tackle the mechanism of action of membrane-bound GRs by looking at the kinome. The effect of lipopolysaccharide stimulation and subsequent treatment with DEX-BSA on different kinase activation of CD14+ monocytes was examined by PepChip analysis. Thereby, 43 substrates with reduced phosphorylation and 8 substrates with increased phosphorylation were observed. P38 MAPK, PKA, PKB, and casein kinase 2 were identified as possible upstream kinases by the Human Protein Reference Database, and functional annotation clustering was performed using the PANTHER database. Further analysis using the Bio-Plex phosphoprotein detection assay showed slight but not significantly increased p38 MAPK phosphorylation in DEX-BSA–treated cells. Interestingly, BSA alone was also able to increase the degree of phosphorylation of p38 MAPK, and phosphorylation data for other kinases still need to be generated.
A significant role for kinases in relaying the nongenomic GC signal has also been described by other investigators. Bartis and colleagues showed that nongenomic GCs induced ZAP-70 tyrosine phosphorylation in T cells (10). Furthermore, DEX has been shown to induce serine phosphorylation and membrane translocation of annexin I through the GR, MAPK, PI 3-kinase, and calcium-dependent PKC pathways (5). Löwenberg et al demonstrated rapid GC-induced suppression of insulin signaling independent of actinomycin D (thus nongenomic signaling) through suppression of phosphorylation of the insulin receptor and its downstream substrates (4). The current study by Strehl et al takes the investigation of nongenomic GC signaling one step further by identifying the common origin of membrane-bound and cytosolic GRs. The next challenge is to determine the mechanisms that allow the membrane-bound GR pool to be formed and unravel the downstream signaling cascade, because these mechanisms may provide novel therapeutic targets.
When GC treatment is used, classic genomic pathways are involved in the inhibitory effects of GCs on inflammation. Especially when higher dosages of GCs are used, nongenomic mechanisms may also be involved to dampen ongoing inflammation. In general, nonspecific interpolation of GC molecules into cell membranes may alter cell function by influencing cation transport and increasing mitochondrial proton leakage. The resulting inhibition of calcium and sodium cycling across the plasma membrane of immune cells is thought to contribute to the observed rapidly induced immunosuppression and reduced inflammation.
In the acute phase of diseases such as systemic lupus erythematosus and rheumatoid arthritis, patients are sometimes treated with high-dose GCs. It is thought that the immunosuppressive properties of GCs at that moment are mediated by both genomic and nongenomic pathways, although it is not clear which GC properties should be attributed to either pathway. For instance, high-dose hydrocortisone (10−4M) may inhibit neutrophil degranulation within 5 minutes (11). In addition, T cell receptor stimulation may be inhibited by rapid nongenomic GC action through downstream kinases (4). Further elucidation of the nongenomic signaling pathways may unveil novel therapeutic targets in the treatment of rheumatic diseases and aid in distinguishing therapeutic effects from adverse effects.
As a rule of thumb, it is believed that the nongenomic actions of GCs are especially elicited by higher dosages (equivalent to a prednisone dosage of ≥30 mg/day). If specific targeting of the GR pools responsible for the antiinflammatory properties becomes feasible, it might be possible to obtain the above-described combined genomic and nongenomic effects without using higher dosages of GCs. This may bring us one step forward in effectively treating our patients, because higher dosages of GCs have a more serious spectrum of adverse effects compared with lower dosages. The new findings on GRs reported by Strehl et al may open the way to decreasing the high dosages of GCs that are presently still used to treat patients with different rheumatic diseases.
All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published.