| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Department of Molecular and Cell Biology, Ligand Pharmaceuticals, San Diego, California 92121
Correspondence: Address all correspondence and requests for reprints to: Jeffrey N. Miner, Department of Molecular and Cell Biology, Ligand Pharmaceuticals, 10275 Science Center Drive, San Diego, California 92121. E-mail: jminer{at}ligand.com
| Abstract |
|---|
|
|
|---|
| I. Introduction |
|---|
|
|
|---|
The GR pathway is well-suited to a drug discovery effort because the key target in the pathway, the GR itself, binds to and is regulated by endogenous small molecule glucocorticoids. In addition, the molecular mechanisms through which the GR acts have recently become better understood; and lastly, drugs are already on the market that target the receptor. The receptor is expressed in a wide variety of tissues, including bone (osteoblasts and osteocytes), liver, brain, T and B cells, and macrophages (4). Cortisol in man and corticosterone in rodents are the major glucocorticoids that act through the GR to mediate numerous physiological responses (5). The unliganded receptor is associated with chaperone proteins (heat shock proteins 70, 90, 54, and others) in an inactive state in the cytoplasm of cells (6, 7, 8, 9). After interaction with hormone, the response of the GR is quite rapid and follows a well-defined signal transduction pathway (10). Glucocorticoids bind the receptor and produce a conformational change in the receptor that results in dissociation of the heat shock proteins, nuclear translocation, and DNA binding activity. This conformational change also results in the formation of various interaction surfaces on the receptor for multiple regulatory factors required by the receptor for activation and repression of gene expression (11). Once bound at a given gene promoter, the receptor can regulate gene expression either positively or negatively. We present our efforts to understand and utilize the GR as a target for the discovery of an improved class of antiinflammatory drugs.
| II. Use of Glucocorticoids |
|---|
|
|
|---|
Immunosuppressive therapy for transplant rejection and autoimmune disorders often includes short-term high-dose steroid treatment to rapidly reduce the cell-mediated response to transplanted foreign tissue. This is typically followed by chronic lower dose steroid treatment. Glucocorticoids inhibit the immune response to transplanted tissue by several mechanisms, including both the general killing of T cells and the induction of a known immunomodulator, lipocortin. These effects are likely tied to the other antiinflammatory effects of glucocorticoids. The development of acute graft vs. host disease is a major problem associated with bone marrow transplantation. This disease often determines the success or failure of allogeneic marrow transplantation. Glucocorticoids are extremely useful in the treatment of graft vs. host disease (15, 16, 17).
Additionally, a number of cancers, such as multiple myeloma and certain lymphomas and leukemias, respond well to combination therapies that include the glucocorticoids prednisone or dexamethasone. In these diseases, the cancer cells are killed through glucocorticoid-mediated induction of apoptosis. Thus, the impressive efficacy of steroids in the treatment of multiple diseases has made them one of the most commonly prescribed classes of drugs. Unfortunately, their utility can be severely limited by a wide spectrum of side effects.
A. Side effects associated with steroid use
The side effects of glucocorticoids have been shown to be strictly dose dependent. Thus, as the dosage is escalated to improve efficacy, the side effects also increase. In addition, some side effects are known to be age and sex-dependent. The side effects of glucocorticoid therapy show different degrees of severity, likely due to the wide variety of physiological contexts in which glucocorticoids act (18, 19, 20, 21). The list of side effects from long-term steroid use is long and includes suppression of the production of endogenous glucocorticoids (adrenal suppression) and other steroids (testosterone and estrogen), dermal atrophy due to lack of remodeling of the skin, and impacts on behavior and mental state. A number of the more common side effects are detailed below.
1. Osteoporosis.
Long-term glucocorticoid treatment often results in some degree of osteoporosis. Susceptibility to fractures and the chance of aseptic necrosis of the femoral head increases within months of starting glucocorticoid therapy (18, 22). Steroids reduce the quality of trabecular bone, resulting in an increase in fracture rate (23, 24, 25). Detrimental bone effects have been documented in several disease settings after glucocorticoid treatment, including RA (26, 27), chronic obstructive pulmonary disease (25), asthma (28), and transplantation (29, 30). Bone loss is highest in the first 6 months of therapy, after which patients continue to lose bone, but at a slower rate. When taken off steroids, patients appear to partially regain bone (18, 23, 31).
2. Muscle wasting.
Glucocorticoid-induced myopathy, resulting in decreased strength and muscle mass, likely contributes to the high fracture rate caused by steroids due to an increased likelihood of falls. The mechanism by which glucocorticoids affect muscle mass is partially due to hypogonadism observed in many patients. This is manifested as a decline in levels of the sex steroids estrogen and testosterone, two hormones that normally contribute to the maintenance of both muscle and bone mass (32, 33).
3. Hypertension.
Excess glucocorticoids can lead to increased blood pressure. These effects contribute to increased risk of heart-related illness and other complications. Glucocorticoids and mineralocorticoids exert effects at several different points critical for regulation of blood pressure. Glucocorticoids are in vast excess relative to mineralocorticoids in serum. Normally, the kidney is protected from the effects of these high cortisol levels through the oxidizing action of 11ß-hydroxysteroid dehydrogenase 2, a tissue-specific enzyme capable of converting cortisol to the weaker 11-ketosteroid cortisone. However, aldosterone, with an aldehyde group at C18, as well as synthetic steroids such as dexamethasone (with a 9
-fluoro group) are not susceptible to this activity and have major effects directly on the kidney through both the mineralocorticoid and GRs. The effects in this tissue include increases in both transepithelial sodium transport and sodium reabsorption in the proximal tubule as a result of increased sensitivity to angiotensin II (34). A similar system may operate in brain, best characterized in the rat; 11ß-hydroxysteroid dehydrogenase 2 is expressed along with mineralocorticoid receptor (MR) in a few select areas involved in central regulation of salt, water balance, and blood pressure (35, 36). There are however areas of the brain where MR is likely unprotected and may be exposed to cortisol.
4. Glucocorticoid-mediated insulin resistance.
The glucocorticoid effect on glycemic control is thought to target insulin signaling (37, 38, 39). Glucocorticoids affect insulin-mediated increases in blood flow to muscles (40, 41), and they decrease key insulin receptor signaling molecules and increase glucose output by increasing the rate-limiting enzyme in gluconeogenesis, phosphoenol pyruvate carboxy kinase (42, 43, 44).
5. Truncal obesity and fat redistribution.
Glucocorticoids induce fat redistribution and accumulation; fat is shed from limbs and accumulates in truncal and visceral areas. Facial, supraclavicular, and posterior cervical fat depots are particularly sensitive to glucocorticoids, resulting in the moon face and buffalo hump characteristic of long-term glucocorticoid treatment. This significantly affects the quality of life for glucocorticoid-treated patients by negatively impacting their appearance and by predisposing them to obesity-related health issues.
6. Inhibition of wound repair.
Glucocorticoids increase the risk of infection by hindering wound healing (45). These effects are dependent on both the dose and timing of glucocorticoid administration. Glucocorticoids affect wound healing by several mechanisms. Inflammation itself is a natural and critical part of the wound healing process and as a consequence, the antiinflammatory effects of glucocorticoids are detrimental to wound repair (46). In addition, glucocorticoids inhibit both collagen synthesis and cross-linking, directly affecting the structural components of a healing wound (47, 48).
| III. Mechanism of Glucocorticoid Receptor Action |
|---|
|
|
|---|
A. Transcriptional activation
During transcriptional activation, the receptor makes contact with a variety of proteins that either directly or indirectly mediate transcriptional regulation. The GR has been shown to directly contact TFIID and TFIIB and other components of the basal transcriptional machinery (56, 57). In addition, the receptor interacts with members of the steroid receptor coactivator (SRC)/p160 family of coactivators (58, 59). These proteins in turn interact with other protein families with histone acetylase activity (60). Histone acetylation causes a "loosening" of the nucleosomal structure in the vicinity of the promoter and can be accompanied by other posttranscriptional modifications of histones bound nearby (e.g., methylation, phosphorylation). This more accessible promoter is now capable of interacting with transcriptional activators more readily and results in an increased rate of transcription (61). It should be noted that there is also evidence that the GR can clear chromatin of nucleosomes without the use of acetylase-containing cofactors (62, 63, 64, 65, 66, 67, 68). The stabilization of the transcriptional machinery through direct contact, recruitment of histone-modifying enzymes to the promoter, and the removal of existing chromatin by other means are key elements contributing to transcriptional activation.
B. Transcriptional repression
Transcriptional repression activity is central to the glucocorticoid-mediated antiinflammatory and antiproliferative effects (3, 51, 52, 69, 70, 71, 72). Unlike the GR, most nonsteroidal nuclear receptors like peroxisome proliferator-activated receptor and retinoic acid receptor can interact with corepressors and repress transcription in the absence of ligand or in the presence of antagonists. These corepressors in turn have histone deacetylase activity that trims acetyl groups off nucleosomes, compacting and silencing the promoter to which unliganded nuclear receptor is bound (73, 74, 75, 76). In contrast, repression by steroid receptors occurs only in the presence of ligand. Fundamentally, this suggests a unique mechanism of action by these proteins. In fact, numerous mechanisms for repression have been suggested. Table 1
describes some, but not all, of the different mechanisms of GR-mediated repression. Two main types involve either specific DNA binding by GR or specific protein-protein interactions by GR. This demarcation is somewhat fuzzy because, of course, the ultimate activity of GR likely depends on both protein-protein and protein-DNA interactions, but here we are referring specifically to the proximal interaction that GR undergoes to initiate the repression process. Within these interactions, several general classes of targets can be identified. These include, but are not limited to, negative GREs, kinase interactions, and corepressor interactions. The mechanism of action within each class varies significantly and ranges from effects at the level of DNA to effects on RNA polymerase and transcription factors directly. Examples for each are given, along with references. Posttranslational repression as well as other interesting mechanisms have been left off the list only because of the relatively small amount of literature available on the topic (77, 78).
|
B (NF
B) and activator protein-1 (AP-1)] (79). There are also reports that glucocorticoids can induce inhibitor of NF
B (I
B) (51, 52, 80, 81). I
B induction appears to be cell line dependent (2) and as such may not be a universal mechanism. There have also been descriptions of direct interactions between GR and other transcription factors resulting in repression (1, 71, 82, 83, 84). In addition to classical GREs, two other functional types of GREs have been described. At composite GREs, both the receptor and the targeted transcription factor bind DNA side by side (1, 50). At tethering GREs, the receptor does not bind directly to the DNA but is tethered to the promoter through its interaction with transcription factors such as AP-1 or NF
B (50, 85). Both AP-1 and NF
B promote the transcription of proinflammatory genes. Thus, their regulation by GR is a key element in the immunoregulatory effect of steroids and appears to be mediated through several mechanisms. Direct binding between c-Jun N-terminal kinase and GR has been shown to negatively regulate AP-1 activity (86). Alternatively, direct tethering interactions between the receptor and NF
B or AP-1 at specific promoters results in decreased transcription and is presumed to result in the recruitment of a repression complex that inhibits transcription (71, 87, 88). Once a promoter is repressed, the receptor can dissociate from that promoter and bind to and regulate another gene promoter. This is consistent with recent fluorescence quenching data measuring the residence time of the receptor on a given gene in the cell. These green fluorescent protein fusion experiments have indicated that the residence time for intracellular receptors is less than 1 sec (89). This means that some mechanism must exist that keeps a promoter repressed in the absence of receptor. Compaction of promoter nucleosomal structure could explain this phenomenon. Interestingly, recent data indicate that the corepressor in question may, in fact, be a coactivator (90, 91). These experiments indicate that the coactivator GR-interacting protein-1 (GRIP-1)/SRC-2 is recruited to the promoter of the repressed collagenase gene by GR. They also show that there are changes in the phosphorylation state of RNA polymerase II during repression of NF
B (85). Other mechanisms have also been described that include specific TATA box occlusion by DNA-bound GR (92). Recently, a protein mediator for repression by GR has been proposed. Thyroid receptor interacting protein 6 (TRIP6), a LIM domain-containing protein, is implicated in repression by binding to AP-1, NF
B, and GR (93). These authors suggest that repression by GR is mediated by TRIP6. TRIP6 normally functions as a coactivator for NF
B, but when GR interacts, it becomes a corepressor. The current notions of how steroid receptors repress transcription suggest that the mechanism of repression may differ between genes. Clearly, transcriptional activation and repression are critical regulatory steps in mediating the physiological and pharmacological effects of glucocorticoids. | IV. Novel Glucocorticoid Receptor Ligands |
|---|
|
|
|---|
Current efforts to identify novel GR ligands have resulted in a number of divergent terminologies. To prevent misunderstandings due to potentially similar terms with what we believe are different meanings, we provide our definitions of some of the key terms below:
A. Compounds that test the activation/repression hypothesis: dissociated glucocorticoids
Truly dissociated glucocorticoids have been a major goal for a number of groups, including ours. The first attempt at testing the activation/repression hypothesis came with the publication by Vayssiere et al. (55) wherein the authors describe several steroidal compounds that are capable of separating transcriptional activation from repression (RU24858, RU40066, and RU24782). This group was able to show significant differences between their compounds and commonly used steroids in a number of in vitro assays. These molecules retained the very efficient inhibition of both AP-1- and NF
B-mediated gene induction and were strong antiinflammatory agents in vivo. However, they showed a reduction in transactivation activity on several genes. Subsequent work by Belvisi et al. (99, 100) demonstrated that although in vitro the compounds were dissociated for transcriptional activation and repression, in vivo, they still had the same side effect profile as steroids. Thus, no therapeutic advantage can be attributed to the dissociation between activation and repression using these compounds. It is unclear whether this is a problem where the compounds were dissociated in vitro, but fully activating in vivo or whether, in fact, the repression activities of these compounds were sufficient to cause both efficacy and side effects. It will be of interest to continue the analysis of these compounds for differential effects on gene regulation in vivo and in other possible side effect areas. Interestingly, compounds with the opposite profile, strong activation and weak repression, are unable to block inflammation (99). In addition, using GR dimerization mutants that prevent activation by GR but do not affect repression, it was shown that the antiinflammatory activity of steroids was maintained. Together, these results suggest that repression may be sufficient for antiinflammatory activity (101).
We have found that many of the compounds we originally viewed as dissociated were in fact, gene selective instead. In hindsight, this makes sense given the fact that the mechanisms of activation and repression are extremely diverse. This does not mean that these molecules are less valuable. In fact, they are extremely interesting compounds with strongly improved therapeutic indexes. They are not, however, a good test of the activation/repression model. The existence of these compounds does suggest that it is possible to achieve therapeutic benefit without complete separation between activation and repression. Vayssiere et al. (55) show separation in vitro but not in vivo. In contrast, Schacke et al. (95) demonstrate interesting separation in vivo between the nonsteroid ZK216348 and the glucocorticoids prednisolone and dexamethasone. These in vivo studies reveal significant separation using sc and topical dosing in a number of inflammation models with ZK216348. These data indicate that the compound exhibits efficacy comparable to prednisolone with reduced effects on blood glucose, tyrosine aminotransferase (TAT) enzyme induction, and skin thinning. The compound may be a useful tool to understand how to achieve selectivity in vivo, although the PR and MR antagonist activity associated with this compound must be factored into the interpretation of the data. Interestingly, this compound does not show good separation between activation and repression in vitro; this may be due to some unexplained low activity in cellular assays.
B. Activation vs. repression: true dissociation
To effectively test the activation/repression hypothesis requires a compound that truly separates all activation and repression. Unfortunately, such activity has not been shown for any molecule to date using a large number of genes. Eventually, we believe that microarray studies will likely reveal that each ligand differentially regulates gene expression. Determining the correct and most beneficial profile for a novel SGRM will require multiple iterations of the process of producing new compounds and determining the precise spectrum of their transcriptional profile.
Are truly dissociated compounds possible? Given the array of genes and the clear multitude of potential regulatory mechanisms, the likelihood of finding a compound that actually separates all activated genes from all repressed genes seems highly unlikely. It is also unclear whether such a compound would be truly desirable because activation of antiinflammatory genes may also play a role in the treatment of inflammatory diseases. Of the proposed dissociated compounds that have been published, all have been shown to differentially regulate one or sometimes two genes. This is not the same as demonstrating that the compound is dissociated on all glucocorticoid target genes. Despite these caveats, it is our belief that the activation/repression hypothesis has provided a very useful framework to find novel compounds with potential utility, and some success has already been achieved at least preclinically.
C. Avoiding the deflazacort trap
The steroid deflazacort was originally believed to have less impact than classical steroids on bone. This compound is a D ring substituted steroid otherwise similar to cortisol. The initial clinical data on deflazacort were quite encouraging and suggested decreased impact on both bone and glucose metabolism (102, 103, 104). Multiple randomized clinical trials appeared to indicate that deflazacort indeed has less severe side effects (105). The difficulty lies in demonstrating equivalent antiinflammatory efficacy between prednisone and deflazacort. Equivalence trials require large numbers of patients to ensure that a small but significant difference could be detected (106, 107, 108). Many of these clinical trials relied on the original determination of a 1:1.2 relative potency ratio described by the manufacturer (5 mg of prednisone = 6 mg of deflazacort). Subsequent trials that adjusted the steroid dose to maintain equivalent antiinflammatory efficacy usually needed higher levels of deflazacort than the ratio of 1:1.2 (109). Thus, trials comparing side effects may not have used biologically equivalent doses of deflazacort. Unfortunately, at these higher doses, the advantages of deflazacort were minimized (102).
To establish true clinical selectivity, it is important to avoid the problem encountered when deflazacort was tested by clearly demonstrating that the compound being tested can achieve full efficacy and that side effects are characterized at the fully efficacious dose or concentration. The deflazacort "trap" could occur when comparisons are made across species or across cell lines without regard for metabolic, pharmacokinetic, or receptor sequence differences.
D. Antagonists
The steroids prednisone (after conversion to prednisolone by 11ß-hydroxysteroid dehydrogenase 1 in liver) and dexamethasone are full agonists of GR, able to induce all the activities of the receptor. It is possible to prepare ligands that entirely block receptor activity by competing with endogenous ligand. Selective antagonists of GRs could be useful in treating hypercortisolemia associated with Cushings syndrome and other conditions in which the endogenous GR is hyperactivated through either higher glucocorticoid levels or increased receptor sensitivity (110). Other possible uses include a reduction of the immunosuppression associated with ongoing HIV infection and treatment of depression (111) and other stress-associated phenomena (112, 113, 114). Interestingly, patients suffering from obesity and diabetes associated with syndrome X may benefit from the use of a selective glucocorticoid antagonist (115). Antagonist activity in the liver (inhibiting gluconeogenesis), muscle (decreasing insulin resistance), and fat (reducing obesity) would all be potentially beneficial tissue activities. Receptor selectivity would be crucial to any useful GR antagonist. Currently, compounds like RU486 are available that are powerful antagonists of GR. Unfortunately, these molecules also effectively bind PRs, preventing their widespread use. In addition, tissue selectivity would be critical because the antagonist must be neutral on other critical tissues and systems such as the hypothalamic-pituitary-adrenal axis, immune system, and inflammatory response.
E. Agonists
The steroidal glucocorticoid agonists show some differences in their biological half-life and possibly brain penetration (116), but generally the most important difference in the antiinflammatory activity of a given steroid and its side effects is directly related to its affinity for the receptor (e.g., dexamethasone > prednisolone > cortisol) (117). The development of new and more selective steroids has been stymied by the apparent link between efficacy and side effects of this class of molecules (102). However, the relative cross-reactivity with the MR for several of the synthetic glucocorticoids (dexamethasone and prednisolone) has improved. These newer steroids have less interaction with MR than the endogenous glucocorticoid cortisol (118).
There is a report that the antibiotic rifampicin is a ligand and an activator for GR (119). This does not appear to be a general phenomenon, based on our own data and other published reports (119, 120, 121, 122, 123, 124). Although it seems likely that no direct connection between GR and rifampicin exists, it is true that both glucocorticoids and rifampicin can activate other nuclear receptors (pregnane X receptor and steroid and xenobiotic receptor) involved in the increased expression of specific cytochrome P450 enzymes, including 3A4 (125, 126). This connection may have therapeutic consequences because these compounds may affect metabolism for each other and, consequently, the relative serum concentrations of both compounds in patients if coadministered.
| V. Drug Discovery Efforts |
|---|
|
|
|---|
A. Discovery and characterization of a novel nonsteroidal antagonist
A compound library of approximately 100,000 chemicals was run in the GR-VP16 assay. Based upon this screen, numerous compounds capable of increasing transcriptional activity were identified. After demonstrating that the compound activity was specific to GR-VP16, we assayed the compounds in a cotransfection assay with wild-type GR and an MMTV-luciferase reporter in the presence and absence of dexamethasone to determine whether the compounds were agonists, antagonists, or partial agonists. Virtually all were antagonists. One nonsteroidal compound, AL082D06 (D06), was selected for further characterization, and this work will be described next. In addition, a nonsteroidal partial agonist was also identified. A significant amount of medicinal chemistry was applied to this lead compound to improve its profile. The compound resulting from this work is described in Section V.B.
The activities of D06 were characterized in detail at the molecular and cellular level (129). Using a cotransfection assay with wild-type GR, D06 acts as an antagonist of dexamethasone not only with the MMTV promoter, but also with the 3-kb TAT promoter and less complex promoters comprised of isolated GREs. D06 binds to the GR with an affinity of approximately 200 nM but shows virtually no affinity for other steroid receptors based upon competitive binding assays. In cotransfection assays, D06 again demonstrates virtually no activity with any tested receptor other than the GR. Further characterization was performed to determine the effects of D06 in cell-based models of transcriptional activation. D06 strongly antagonizes steroid-induced transcription of both the TAT enzyme in human skin fibroblasts and glutamine synthetase RNA in MG63 bone-derived cells. Similarly, D06 is able to reverse the dexamethasone-mediated suppression of TNF
and IL-1ß-induced expression from the E-selectin promoter. Furthermore, unlike the antagonist RU486, but like the antagonist ZK-299, D06 does not induce DNA binding by GR in vitro and can inhibit both dexamethasone- and RU-486-induced DNA binding. The ability of these compounds to induce DNA binding in cells was further characterized using an assay that measured the ability of ligand-bound wild-type GR to inhibit the DNA binding of a constitutively active mutant of GR. In this assay, unlike either RU-486 or ZK-299, D06-bound GR does not compete with the constitutively active GR.
Thus, through a nested set of assays, we were able to identify a novel, nonsteroidal GR antagonist with future potential value in the treatment of Cushings syndrome and its sequelae.
B. Discovery and characterization of a nonsteroidal SGRM
When we began our efforts to identify a SGRM, we, like many others, attempted to find dissociated glucocorticoids that were qualitatively different in terms of their ability to activate gene expression (undesirable) vs. transrepression activity (desirable). The basis for this hypothesis was the realization that many genes involved in metabolic pathways that lead to undesirable side effects are up-regulated, including enzymes in gluconeogenesis and lipid and muscle metabolism, whereas many proinflammatory genes (e.g., IL-1, TNF-
, IL-6, Cox-2, and E-selectin) are repressed. As described above, up-regulation of these genes is generally through an activated receptor dimer binding to GREs within a promoter. Down-regulation is generally more complex and occurs through an indirect mechanism through which the ligand-bound GR binds to and inactivates other transcription factors such as AP-1 (1, 87) or NF
B (130). That these activities are distinct functions of the receptor was demonstrated genetically in experiments that showed that GR mutants that lack the capacity for transcriptional activation maintain their ability to repress (72).
There are a number of examples of dissociated glucocorticoid modulators in the literature (131, 132, 133). However, in our experience, many of these compounds do not adequately separate efficacy from side effects in vivo, and those that do have clear in vivo separation do not simply separate activation from repression, but have a more complex gene-selective action.
Due to the subtle nature of our desired endpoint, the search for novel SGRMs has led us to develop a much more complex series of assays than was required for identification and characterization of the antagonist, D06. To establish the flow scheme used in this process, an iterative series of in vitro and in vivo characterization studies was performed. Compounds were "binned" based upon their in vitro activities, including transactivation, transrepression, and cell and promoter selectivity. These assays included competitive ligand binding, cotransfection assays with wild-type GR and the MMTV promoter in a neutral cell background in both agonist and antagonist mode, induction of the TAT promoter in liver cells, and repression of cytokine-induced E-selectin promoter activation. Compounds typical of several bins were examined in vivo. This included compounds with strong or weak activity in the E-selectin repression assay with varied activity in the activation assays. As expected, strong repression in the E-selectin promoter assay correlated with antiinflammatory activity. However, both agonists and partial agonists in the MMTV cotransfection assay had antiinflammatory activity in the rat carrageenan paw edema model.
These screening activities led to the identification of numerous compounds capable of antagonizing GR-mediated activation. In addition, a small number of compounds were found that were partial agonists of GR. These also had some repression activity. One of these compounds was chosen as the scaffold for medicinal chemistry optimization. From that effort, AL-438, a compound that we determined to have binding affinity for GR virtually identical to that of prednisolone, partial agonist activity in the MMTV activation assay, and repression efficacy in the E-selectin assay equivalent to prednisolone, was synthesized and profiled in detail. Our goal was to demonstrate that this nonsteroidal GR modulator had an improved side effect profile relative to clinically used steroids (96).
C. In vitro characterization of AL-438
In the TAT promoter assay in the HepG2 liver cell line, AL-438 was a full agonist, despite being a partial agonist in the MMTV assay. However, when its ability to induce aromatase in human skin fibroblasts was measured, it was a partial agonist. This partial activity is probably not due to metabolism or differential uptake of AL-438 because in the same cell background in the IL-6 repression assay, AL-438 is fully efficacious. In addition, AL-438 can inhibit prednisolone-induced activation of aromatase as well as MMTV, indicating that it is binding to the receptor. We tested the cross-reactivity of AL-438 for a number of nuclear receptors and found low but significant MR antagonist activity. This suggests that because 11ß-hydroxysteroid dehydrogenase type 2 does not protect MR from AL-438, there might be some detectable MR inhibition, although it has not been confirmed in vivo.
In addition to the E-selectin promoter assay performed in HepG2 cells, in which AL-438 fully repressed the IL-1ß and TNF
mRNA induction, its activity in a second repression assay was also measured. In this assay, IL-6 protein in human skin fibroblasts was induced by IL-1ß, and the ability of AL-438 to suppress this expression was determined. Again, AL-438 had efficacy equal to that of prednisolone.
Finally, we examined the activity of AL-438 in an in vitro osteocalcin expression assay in the MG63 osteosarcoma cells. Because glucocorticoids may inhibit bone formation partially through repression of genes involved in bone turnover and formation, we felt that this assay might yield insight into whether AL-438 may have reduced negative effects in bone. In this assay, prednisolone suppresses osteocalcin mRNA levels approximately 5-fold, whereas AL-438 had no effect on osteocalcin mRNA levels. Furthermore, AL-438 exhibits only partial inhibition (60%) of osteoprotegerin, a bone formation marker in MG-63 cells, whereas prednisolone strongly inhibited osteoprotegerin production. These two assays in a bone cell background suggest that AL-438 may have a significantly reduced impact on bone. Based on these in vitro assays, it was gratifying to note that not only does AL-438 demonstrate selective gene activation, but also selective gene repression. We then tested how this differential regulation of both activation and repression would translate into potential therapeutic benefits in vivo (96).
D. In vivo activity of AL-438: efficacy
AL-438 was tested in both acute and chronic models of inflammation. The carrageenan-induced paw edema assay in the rat is a standard model for acute inflammation. Rats are treated, and carrageenan is injected into the right hind paw, causing the development of acute edema. Paw volume is measured 3 h after carrageenan injection. In this model, prednisolone inhibits edema with an efficacy of 77% relative to the volume of the left hind paw that was not injected with carrageenan. In the same assay, AL-438 had an efficacy of 64%, demonstrating that it had similar activity to prednisolone in this acute inflammation assay.
To demonstrate longer term efficacy in a more stringent model that involves complex changes in both soft tissue and bone, the rat adjuvant-induced arthritis model was employed (Fig. 1A
). Glucocorticoids such as prednisolone or dexamethasone demonstrate desirable effects on joint swelling, synovitis, and periosteal new bone formation. The assay is performed by injecting the right hind paw with Freunds complete adjuvant. The effects of the injection become clearly evident at d 14 with maximal soft tissue injury but before the onset of changes in bone. At this time animals are culled and randomized, and daily treatment is initiated and continued for an additional 14 d. When left hind paw swelling is determined (the injected right hind paw cannot be used for this measurement due to substantial bone and necrotic lesions that complicate the analysis), AL-438 has an efficacy equal to that of prednisolone at 30 mg/kg·d, although its potency is somewhat lower (ED50, 9 vs. 1 mg/kg). Surprisingly, although efficacy based on reduction of paw edema was equivalent, AL-438-treated animals showed grooming behavior and overall activity equivalent to nonadjuvant-treated controls, whereas the behavior of the prednisolone-treated rats was similar to that of the vehicle-treated animals, all of which exhibited signs of stress and disease, including lack of grooming and minimal activity. In other words, the behavior of AL-438-treated animals is similar to that of the healthy control animals, whereas the prednisolone-treated rats still show signs of stress. This result may be due to the more rapid onset of improvement in edema for AL-438 relative to prednisolone. The results of the acute and chronic inflammation models clearly show that AL-438 has full antiinflammatory activity despite its gene-selective activation and repression profile (96).
|
In the glucose study, fasted rats were treated with compound, and glucose levels were measured (Fig. 1B
). Prednisolone causes a clear increase in glucose at 10 mg/kg, whereas neither AL-438 (at 30 mg/kg) nor the GR antagonist, RU-486, change glucose levels relative to vehicle controls. Furthermore, predosing with either RU-486 or AL-438 was able to inhibit the increase in glucose caused by prednisolone. This antagonism of prednisolone clearly demonstrates that AL-438 is not a glucocorticoid-like agonist at this endpoint and that the differences in efficacy are not simply due to levels of exposure to the compound.
The bone assay was performed using rats treated with vehicle, prednisolone at 10 mg/kg, or AL-438 at 30 mg/kg. As in the glucose study, these levels were chosen to compensate for the slightly weaker potency of AL-438 relative to prednisolone in the inflammation assays. Mineralizing bone was labeled with calcein and tetracycline at specific times to measure bone formation rates. Prednisolone clearly reduces the cancellous mineral apposition rate measured in the tibia, whereas AL-438 had no suppressive effect. When cortical bone was examined, AL-438 was distinctly weaker than prednisolone at inhibiting the periosteal bone formation rate. These results demonstrate that at doses with equal efficacy in inflammation models, AL-438 has a distinct and improved side effect profile.
F. AL-438 mechanism of action
Evidence is mounting that selectivity among nuclear receptors is based upon changes in cofactor interactions of the ligand-bound receptor. These in turn are driven by the exact conformation of the receptor determined by the specific ligand bound. Using the ER as an example, Shang and Brown (134) elegantly demonstrated how coregulator recruitment could differentiate between the activities of the two selective ER modulators, tamoxifen and raloxifene. In this study, they show that both tamoxifen and raloxifene recruit corepressors to target gene promoters in mammary cells, where they both act as antagonists. An explanation for the different activities of the two compounds in the uterus was demonstrated in endometrial cells where tamoxifen, but not raloxifene, acts like estrogen by recruiting coactivators to a subset of genes. This difference required a high level of a specific coactivator, SRC-1, in uterine cells.
We examined GR cofactor recruitment induced by AL-438 using two cofactors: peroxisome proliferator-activated receptor-
coactivator-1 (PGC-1), which is involved in hepatic glucose production; and GRIP-1, which appears to play a role in GR-mediated transcriptional repression of proinflammatory genes (135, 136). Using a two-hybrid assay, prednisolone is able to efficiently induce the interaction of GR with both PGC-1 and GRIP-1. AL-438 was able to induce the interaction with GRIP-1 with an efficacy equal to that of prednisolone, but recruited PGC-1 with an efficacy significantly reduced relative to that of prednisolone. The difference was even more striking when measured in a biochemical pull-down experiment. This result may explain the maintenance of antiinflammatory activity, whereas hyperglycemia is significantly reduced. We propose that the structural changes induced by AL-438 are different from those induced by full agonists, such as prednisolone, and that these differences are responsible for altered cofactor interactions, and thus, altered pharmacology (Fig. 2
).
|
A safer glucocorticoid should have full efficacy in antiinflammatory activity, but reduced efficacy and potency in one or more side effects. The development of new safer antiinflammatory agents that target the GR is now gaining momentum after years of work on steroids and, more recently, nonsteroidal molecules. The molecular details behind the action of the newer compounds being described may point the way to more effective assays capable of detecting novel antiinflammatory agents.
The detection of a tissue selective or a functionally selective ligand for the GR will be difficult, and there is no guarantee, once such a ligand is found, that it will have the necessary profile in vivo. However, recent reports of SGRMs with equal efficacy and improved side effect profiles compared with steroids together with molecular discoveries of the receptor mechanism of action provide fertile ground for additional efforts. Thus, despite the difficulties associated with developing a novel glucocorticoid, progress in this area would be a major benefit to the large number of patients suffering from the side effects of steroids, but needing the antiinflammatory and anticancer activity to maintain their quality of life.
| Acknowledgments |
|---|
| Footnotes |
|---|
Abbreviations: AP-1, Activator protein-1; ER, estrogen receptor; GR, glucocorticoid receptor; GRE, glucocorticoid response element; GRIP-1, GR-interacting protein-1; I
B, inhibitor of NF
B; MMTV, mouse mammary tumor virus; MR, mineralocorticoid receptor; NF
B, nuclear factor
B; PGC-1, peroxisome proliferator-activated receptor-
coactivator-1; PR, progesterone receptor; RA, rheumatoid arthritis; SeGRA, selective GR agonist; SGRM, selective GR modulator; SRC, steroid receptor coactivator; TAT, tyrosine aminotransferase; TRIP6, thyroid receptor interacting protein 6.
| References |
|---|
|
|
|---|
B
-independent downregulation of NF-
B activity by glucocorticoid receptor. EMBO J 16:46984707[CrossRef][Medline]
B-dependent transcription involves direct interference with transactivation. Proc Natl Acad Sci USA 94:1350413509
B and the steroid hormone receptors: mechanisms of mutual antagonism. Mol Endocrinol 12:4556
B and conserved lymphokine element 0. Am J Respir Cell Mol Biol 30:555563
B: what language? Trends Biochem Sci 23:233235[CrossRef][Medline]
B activity by glucocorticoids. Mol Endocrinol 12:355363
B and the glucocorticoid receptor. Proc Natl Acad Sci USA 91:752756
B and the glucocorticoid receptor. Chest 107:139S
B by interfering with serine-2 phosphorylation of the RNA polymerase II carboxy-terminal domain. Genes Dev 14:23142329
B-regulated promoters. Genes Dev 18:25182528
B repression in anti-inflammation and immunosuppression by glucocorticoids. Trends Endocrinol Metab 8:152157[CrossRef][Medline]
B activity through induction of I
B synthesis. Science 270:286290
B by activated glucocorticoid receptors. Mol Cell Biol 15:943953[Abstract]
B signaling by CREB-binding protein and steroid receptor coactivator-1. J Biol Chem 273:2929129294This article has been cited by other articles:
![]() |
K. A. Muzikar, N. G. Nickols, and P. B. Dervan Repression of DNA-binding dependent glucocorticoid receptor-mediated gene expression PNAS, September 29, 2009; 106(39): 16598 - 16603. [Abstract] [Full Text] [PDF] |
||||
![]() |
Z. Zhang, Z.-Y. Zhang, and H. J. Schluesener Compound A, a Plant Origin Ligand of Glucocorticoid Receptors, Increases Regulatory T Cells and M2 Macrophages to Attenuate Experimental Autoimmune Neuritis with Reduced Side Effects J. Immunol., September 1, 2009; 183(5): 3081 - 3091. [Abstract] [Full Text] [PDF] |
||||
![]() |
N. Niu, V. Manickam, K. R. Kalari, I. Moon, L. L. Pelleymounter, B. W. Eckloff, E. D. Wieben, D. J. Schaid, and L. Wang Human Glucocorticoid Receptor {alpha} Gene (NR3C1) Pharmacogenomics: Gene Resequencing and Functional Genomics J. Clin. Endocrinol. Metab., August 1, 2009; 94(8): 3072 - 3084. [Abstract] [Full Text] [PDF] |
||||
![]() |
N. Nader, G. P. Chrousos, and T. Kino Circadian rhythm transcription factor CLOCK regulates the transcriptional activity of the glucocorticoid receptor by acetylating its hinge region lysine cluster: potential physiological implications FASEB J, May 1, 2009; 23(5): 1572 - 1583. [Abstract] [Full Text] [PDF] |
||||
![]() |
A. N. Gerber, K. Masuno, and M. I. Diamond Discovery of selective glucocorticoid receptor modulators by multiplexed reporter screening PNAS, March 24, 2009; 106(12): 4929 - 4934. [Abstract] [Full Text] [PDF] |
||||
![]() |
H. Gong, M. J. Jarzynka, T. J. Cole, J. H. Lee, T. Wada, B. Zhang, J. Gao, W.-C. Song, D. B. DeFranco, S.-Y. Cheng, et al. Glucocorticoids Antagonize Estrogens by Glucocorticoid Receptor-Mediated Activation of Estrogen Sulfotransferase Cancer Res., September 15, 2008; 68(18): 7386 - 7393. [Abstract] [Full Text] [PDF] |
||||
![]() |
Y. Liu, Y. Nakagawa, Y. Wang, L. Liu, H. Du, W. Wang, X. Ren, K. Lutfy, and T. C Friedman Reduction of hepatic glucocorticoid receptor and hexose-6-phosphate dehydrogenase expression ameliorates diet-induced obesity and insulin resistance in mice J. Mol. Endocrinol., August 1, 2008; 41(2): 53 - 64. [Abstract] [Full Text] [PDF] |
||||
![]() |
F. J. Lopez, R. J. Ardecky, B. Bebo, K. Benbatoul, L. De Grandpre, S. Liu, M. D. Leibowitz, K. Marschke, J. Rosen, D. Rungta, et al. LGD-5552, an Antiinflammatory Glucocorticoid Receptor Ligand with Reduced Side Effects, in Vivo Endocrinology, May 1, 2008; 149(5): 2080 - 2089. [Abstract] [Full Text] [PDF] |
||||
![]() |
K. Suino-Powell, Y. Xu, C. Zhang, Y.-g. Tao, W. D. Tolbert, S. S. Simons Jr., and H. E. Xu Doubling the Size of the Glucocorticoid Receptor Ligand Binding Pocket by Deacylcortivazol Mol. Cell. Biol., March 15, 2008; 28(6): 1915 - 1923. [Abstract] [Full Text] [PDF] |
||||
![]() |
P. Dewint, V. Gossye, K. De Bosscher, W. Vanden Berghe, K. Van Beneden, D. Deforce, S. Van Calenbergh, U. Muller-Ladner, B. Vander Cruyssen, G. Verbruggen, et al. A Plant-Derived Ligand Favoring Monomeric Glucocorticoid Receptor Conformation with Impaired Transactivation Potential Attenuates Collagen-Induced Arthritis J. Immunol., February 15, 2008; 180(4): 2608 - 2615. [Abstract] [Full Text] [PDF] |
||||
![]() |
K S Fegan, M T Rae, H O D Critchley, and S G Hillier Anti-inflammatory steroid signalling in the human peritoneum J. Endocrinol., February 1, 2008; 196(2): 369 - 376. [Abstract] [Full Text] [PDF] |
||||
![]() |
J. N. Miner, B. Ardecky, K. Benbatoul, K. Griffiths, C. J. Larson, D. E. Mais, K. Marschke, J. Rosen, E. Vajda, L. Zhi, et al. Antiinflammatory glucocorticoid receptor ligand with reduced side effects exhibits an altered protein protein interaction profile PNAS, December 4, 2007; 104(49): 19244 - 19249. [Abstract] [Full Text] [PDF] |
||||
![]() |
S. G Hillier Diamonds are forever: the cortisone legacy J. Endocrinol., October 1, 2007; 195(1): 1 - 6. [Abstract] [Full Text] [PDF] |
||||
![]() |
A. McMaster and D. W. Ray Modelling the glucocorticoid receptor and producing therapeutic agents with anti-inflammatory effects but reduced side-effects Exp Physiol, March 1, 2007; 92(2): 299 - 309. [Abstract] [Full Text] [PDF] |
||||
![]() |
Y. Liu, C. Yan, Y. Wang, Y. Nakagawa, N. Nerio, A. Anghel, K. Lutfy, and T. C. Friedman Liver X Receptor Agonist T0901317 Inhibition of Glucocorticoid Receptor Expression in Hepatocytes May Contribute to the Amelioration of Diabetic Syndrome in db/db Mice Endocrinology, November 1, 2006; 147(11): 5061 - 5068. [Abstract] [Full Text] [PDF] |
||||
![]() |
D. M. Presman, E. Hoijman, N. R. Ceballos, M. D. Galigniana, and A. Pecci Melatonin Inhibits Glucocorticoid Receptor Nuclear Translocation in Mouse Thymocytes Endocrinology, November 1, 2006; 147(11): 5452 - 5459. [Abstract] [Full Text] [PDF] |
||||
![]() |
T. Hiragun, Z. Peng, and M. A. Beaven Cutting Edge: Dexamethasone Negatively Regulates Syk in Mast Cells by Up-Regulating Src-Like Adaptor Protein J. Immunol., August 15, 2006; 177(4): 2047 - 2050. [Abstract] [Full Text] [PDF] |
||||
![]() |
J.-C. Wang, N. Shah, C. Pantoja, S. H. Meijsing, J. D. Ho, T. S. Scanlan, and K. R. Yamamoto Novel arylpyrazole compounds selectively modulate glucocorticoid receptor regulatory activity. Genes & Dev., March 15, 2006; 20(6): 689 - 699. [Abstract] [Full Text] [PDF] |
||||
![]() |
K. De Bosscher, W. V. Berghe, I. M. E. Beck, W. Van Molle, N. Hennuyer, J. Hapgood, C. Libert, B. Staels, A. Louw, and G. Haegeman A fully dissociated compound of plant origin for inflammatory gene repression PNAS, November 1, 2005; 102(44): 15827 - 15832. [Abstract] [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |