Modulation of glucocorticoid-regulated transcription by purines: novel characteristics and implications for tissue specificity of steroid responses.

Treatment of the T47D(A1-2) mammary carcinoma cell line with the nonspecific kinase inhibitor 2-aminopurine (2AP) has an unusual effect on induction of the mouse mammary tumor virus promoter by glucocorticoids. 2AP initially abrogates the dexamethasone-mediated increase, but after about 16-20 h, this effect is reversed, and continued incubation with 2AP potently stimulates the hormone induction. The biphasic kinetics of 2AP action displayed cell specificity. No inhibitory phase was seen in fibroblasts. Cell type specificity is also seen upon treatment of cells with isobutylmethylxanthine (IBMX). Dexamethasone action is inhibited in mammary cells, but potentiated in fibroblasts. Unlike 2AP treatment, IBMX continues to suppress the hormone response through at least 48 h of treatment. IBMX is commonly used as a phosphodiesterase inhibitor, and indeed, it stimulates a cAMP-dependent promoter in both mammary carcinoma cells and fibroblasts. Because elevating cAMP will potentiate dexamethasone-mediated induction in mammary cells, IBMX must be interventing in another pathway that elicits IBMX-dependent suppression of the hormone response. Pharmacological studies suggest that this interaction is not mediated through adenosine receptors, histamine receptors, or imidazoline receptors. The five-member imidazole ring plays a critical role in the suppressive activity; substitutions at the 7 position inhibit suppression. These results give further indication of coupling of steroid receptor action to other cellular signaling pathways. This complex spectrum of interactions may play a central determining role in the tissue specificity of the response to steroid hormones.

The two faces of a steroid antagonist: when an antagonist isn't.

Activation of protein kinase A potentiates the transcriptional response mediated by the glucocorticoid receptor in responsive fibroblasts and in mammary carcinoma cells. This potentiation is ligand-dependent and occurs in responsive fibroblasts and in mammary carcinoma cells. This potentiation is ligand-dependent and occurs without detectable change in the phosphorylation of receptor. The transcriptional response to glucocorticoid or progestin agonists can be blocked by potent antagonists like RU 486. However, upon activation of protein kinase A, the antagonist action of RU 486 on both receptors is blunted. Indeed, RU 486 can itself activate transcription of a hormone-responsive promoter. The conditional agonist activity is observed with type II antagonists, those which recapitulate many of the early steps of ligand-dependent receptor activation, but not type I antagonists, which do not. These studies have now been extended to antimineralocorticoids. In COS-1 cells transfected with a mineralocorticoid receptor expression vector, treatment with 8-BromocAMP potentiates the response to the agonist aldosterone and elicits additional agonist activity in mineralocorticoid antagonists. A model is proposed wherein type II antagonist-receptor complexes occupy receptor binding sites on the genome. The antagonist, however, fails to promote a receptor conformation that can interact productively with a coactivator mediating the communication between receptor and the basal transcription apparatus. Activation of protein kinase A results in the recruitment or activation of a coactivator that permits recovery of receptor-mediated activation function.(ABSTRACT TRUNCATED AT 250 WORDS)

The coupling of multiple signal transduction pathways with steroid response mechanisms.

In a human breast carcinoma-derived cell line engineered to contain a hormone-responsive luciferase reporter gene, manipulation of cell growth conditions or cellular signal transduction in a variety of ways can enhance or impair glucocorticoid-mediated induction of a target gene. Induction may be enhanced as much as 10-fold or inhibited 90% by different treatments. For example, two different inhibitors of protein phosphatase-1 and -2A potentiated the hormone-dependent induction of luciferase. Activation of protein kinase-A via addition of 8-bromo-cAMP or forskolin also potentiated the hormonal induction, whereas 8-bromo-cGMP was ineffective. In contrast, activating protein kinase-A by inhibiting cAMP turnover with the phosphodiesterase inhibitors isobutylmethylxanthine or Ro20-1724 inhibited the hormone response rather than potentiated it. The inhibitory activity of isobutylmethylxanthine was evident even when activators of protein kinase-A are administered simultaneously. Isobutylmethylxanthine must, therefore, activate a signal transduction pathway in addition to the protein kinase-A pathway. Activation of protein kinase-C potentiated the hormone response in a cell-specific manner. Treatment with epidermal growth factor and imposition of cell stress by heat shock or inhibition of protein synthesis also enhanced the glucocorticoid response. Thus, our results suggest an elaborate coupling of the steroid response pathway with other cellular signal transduction mechanisms that permits an additional layer of control to be imposed on hormone-mediated transcriptional responses. It is proposed that cell-specific phosphorylation events influence steroid receptor interaction with the basal transcription apparatus, thereby altering receptor-mediated induction mechanisms.

Modulation of cell signaling pathways can enhance or impair glucocorticoid-induced gene expression without altering the state of receptor phosphorylation.

We have stably introduced expression vectors for the glucocorticoid receptor and a sensitive, hormone-responsive reporter (mouse mammary tumor virus-luciferase) into a human breast carcinoma-derived cell line. Employing this cell line, we have conducted a detailed examination of the induction of glucocorticoid-regulated genes and the phosphorylation of glucocorticoid receptor following pharmacologic manipulation of cell signaling pathways. The hormone response can be enhanced from 2 to 10-fold by activators of protein kinase A, protein kinase C, and inhibitors of protein phosphatase. Forskolin and 8-bromoadenosine 3':5'-cyclic monophosphate (BrcAMP), but not BrcGMP, enhance the hormone effect, yet surprisingly, phosphodiesterase inhibitors, isobutylmethylxanthine and Ro20-1724, strongly inhibit hormone-mediated induction of the reporter gene. These treatments do not alter cellular receptor content, dexamethasone binding, nor hormone-mediated receptor down-regulation. Tryptic peptide analysis of 32P-labeled receptor reveals that neither BrcAMP, isobutylmethylxanthine, nor the tumor promoter and protein kinase C activator, 12-O-tetradecanoyl-phorbol-13-acetate, detectably alter the state of glucocorticoid receptor phosphorylation. The only agent which alters receptor phosphorylation is the protein phosphatase inhibitor okadaic acid, but only at concentrations higher than required for maximum effects on glucocorticoid receptor transactivation. We propose that these effectors do not modify receptor directly but alter its interaction with transcription complexes.

The steroid antagonist RU486 exerts different effects on the glucocorticoid and progesterone receptors.

To determine whether the steroid antagonist RU486 mediates its antiglucocorticoid and antiprogestin activities by the same or different receptor mechanisms, a direct comparison of RU486 interaction with glucocorticoid (GR) and progesterone (PR) receptors was made. The effects of RU486 on transformation of GR and PR 8-10S complexes in the intact cell and in vitro were analyzed by sucrose density gradient centrifugation, and the in vitro stability of receptor-heat shock protein-90 interactions was analyzed by coimmunoprecipitation. Compared to agonist, RU486 binding produced a reduction in the amount of GR converted from 8S to 4S and stabilized the GR-heat shock protein-90 complex. By contrast, PR-RU486 complexes were transformed both in vitro and in the intact cell to the same extent as receptor-agonist complexes. PR-RU486 complexes sedimented at 5-6S, whereas PR-R5020, GR-RU486, and GR-agonist complexes sedimented at 4S. The portion of GR that undergoes nuclear transformation when bound to RU486 was examined for binding to the glucocorticoid-progesterone response element of the mouse mammary tumor virus by an immunoprecipitation assay. The nuclear-transformed GR-RU486 complex bound the glucocorticoid-progesterone response element with the same affinity as the nuclear-transformed GR-triamcinolone acetonide complex. The electrophoretic mobilities of GR-RU486 complexes and GR-agonist complexes were the same, as determined by gel retardation assay. These results suggest that RU486 exerts its antiglucocorticoid activity at two levels of receptor action: prevention of complete GR transformation and alteration of a step subsequent to GR-DNA binding. As an antiprogestin, RU486 action is exerted predominantly at a post-DNA-binding step.

Latent agonist activity of the steroid antagonist, RU486, is unmasked in cells treated with activators of protein kinase A.

RU486 is a glucocorticoid and progesterone antagonist. In glucocorticoid-responsive fibroblasts, it mediates little or no induction of a truncated, hormone-responsive mouse mammary tumor virus promoter; moreover, it abrogates the induction mediated by the glucocorticoid agonist, dexamethasone. However, when the fibroblasts are treated with activators of protein kinase A, 8-Br-cAMP or forskolin, along with RU486, the steroid now acts as a partial agonist, capable of mediating an induction of hormone-responsive reporter genes. In addition, the ability of RU486 to block the action of the glucocorticoid agonist, dexamethasone, is compromised by concomitant treatment with 8-Br-cAMP. Activators of protein kinase C fail to elicit these phenomena. Induction of gene expression in the presence of 8-Br-cAMP is dependent on the dose of RU486 over a range consistent with a glucocorticoid receptor-mediated mechanism. An antagonist, ZK98 299, which unlike RU486 is not thought to permit receptor binding to DNA, is not activated by 8-Br-cAMP. The elicitation of RU486 agonist activity cannot be attributed solely to idiosyncrasies of the cell line or the promoter. Similar phenomena are observed in another glucocorticoid-responsive fibroblast line. Furthermore, RU486 can induce a minimal promoter bearing two copies of a synthetic receptor target site. However, we have identified at least one promoter toward which RU486 still behaves as an antagonist despite 8-Br-cAMP treatment. These observations suggest that the unmasking of latent agonist activity in a type II antagonist is not an isolated phenomenon and may, therefore, be seen with other receptors and antagonists. The finding that modulation of cellular signal transduction pathways can unmask agonist activity in an otherwise effective steroid antagonist has significant implications for the use of steroid antagonists in the clinical setting and could represent a heretofore unrecognized mechanism for the development of steroid resistance.

The constitution of a progesterone response element.

Progesterone receptors (PRs) recognize and bind to DNA in a sequence-specific manner. To define the sequence-specific determinants of PR binding to DNA, we employed a detailed series of target elements and analyzed both PR binding in vitro and PR-mediated activation of a reporter gene in vivo. These elements represent point substitution or insertion mutants of a progesterone response element derived from the mouse mammary tumor virus (MMTV). Substitution of a G for the first T in the 15-base pair (bp) MMTV progesterone response element core sequence, GTTACAAACTGTTCT, increases PR-DNA binding in vitro. All other tested mutations within this sequence either had no effect or decreased PR binding. From these analyses, we infer an optimal PR recognition sequence, -7RGNACANRNTGTNCY+7, composed of hexameric half-sites separated by precisely 3 bp and exhibiting dyad symmetry. Limited substitution of a suboptimal base for an optimal base, as in the wild type MMTV element, can be tolerated, but further suboptimal substitutions significantly decrease binding by PR. The identity of 1 bp within the hexameric half-site, position +/- 5, is unconstrained. Transition mutations, but not transversions, are permitted at position +/- 7. Here the hydrogen bond acceptor of the N-7 position of the purine ring may be involved in receptor recognition, since this feature is shared by the permitted substitutions. Insertion of a single base pair between the half-sites abolishes detectable binding, suggesting that the spatial orientation of the DNA binding domains of the monomeric receptor subunits are fixed by dimerization of the receptor. This pattern of sequence-specific recognition parallels that previously inferred for the glucocorticoid receptor, indicating that the two receptors may be unable to distinguish individual target sites. Each of the mutant response elements was also assessed for its ability to confer progestin responsiveness to a truncated MMTV promoter when introduced into a PR-containing cell line. Elements exhibiting strong receptor binding in vitro were fully inducible, whereas poorly inducible or uninducible elements displayed little or no recognition by receptor in vitro. However, some elements, though poorly bound in vitro, still activated transcription in vivo in response to hormone. In certain cases activation was as good as that seen with the wild type element. Further examination of in vitro receptor binding by this class of mutant elements using higher levels of baculovirus-expressed receptor revealed that many were indeed recognized by receptor, albeit with a lower affinity than the wild type sequence.(ABSTRACT TRUNCATED AT 400 WORDS)

Heat shock alters the composition of heteromeric steroid receptor complexes and enhances receptor activity in vivo.

Under normal cellular conditions, human progesterone receptors (PR), immune-isolated from cytosols of T47D breast cancer cells, associate with two heat shock proteins (hsps), hsp 90 and hsp 70. Receptors activated by hormone binding in vivo and extracted from nuclei with 0.5 M NaCl no longer associate with hsp 90 but retain association with hsp 70. We have examined the effect of heat shock treatment of cells on hsp-receptor interactions and on receptor function. Heat shock resulted in a partial reduction in cellular levels of PR, but receptors that remained were functional for both steroid and DNA binding activities. By steady-state [35S]methionine labeling prior to heat shock treatment, it was determined that heat shock did not affect the composition or maintenance of preexisting cytosolic PR.hsp 90.hsp 70 complexes. By contrast, immune isolation of PR complexes from cells pulse-labeled with [35S]methionine showed that heat shock altered the composition of newly synthesized hsps associated with PR. After heat shock, both the highly inducible form of hsp 70 (72K hsp) and a 100K hsp were bound to cytosol PR, and inducible 72K hsp remained bound with the nuclear-activated PR. Neither of these hsps were associated in detectable amounts with PR under normal cellular conditions. With respect to receptor function, heat shock treatment substantially enhanced the activity of PR in vivo as determined by measuring hormone-dependent PR-mediated transcription of a target reporter gene (MMTV-CAT) that was stably transfected into T47D cells. Heat shock treatment alone, in the absence of hormone, did not stimulate MMTV-CAT expression nor did it affect transcription from a control reporter gene, pSV2-CAT, suggesting that enhanced receptor activity was due to an effect on PR-mediated processes and not to a general effect on transcription. Induction of the heat shock response by a related chemical stress (sodium arsenite) also enhanced PR activity in vivo. Interestingly, sodium arsenite produced both a greater induction of hsp 90 and hsp 70 synthesis and a greater fold enhancement of PR-mediated gene transcription than did heat shock. This suggests that enhancement of PR activity is related not only to induction of hsp synthesis but also to the severity of the stress response. The present results provide an indication that in certain cells there may exist an interrelationship between the activation pathways by which cells respond to stress and to steroid hormones. Possible mechanisms responsible for heat shock effects on PR activity are discussed.