I. Introduction and Historical Development

Top Previous Next References

For more than 30 years steroids have been known to be involved in various physiological responses with a primary focus on the genomic aspects of action. According to the classic genomic theory of action, steroid hormones bind to specific receptors, which are intracellular transcription factors, and exert positive or negative effects on the expression of target genes (Beato et al., 1996; Beato and Klug, 2000). These effects are characterized by a specific delay and a sensitivity toward inhibitors of transcription and translation, e.g., actinomycin D and cycloheximide. Intracellular steroid receptors have been thoroughly characterized and, finally, cloned; they are composed of a ligand-binding domain, a DNA-binding domain, and several transactivation functions distributed along the molecule (Evans, 1988; Beato, 1989; Fuller, 1991). In addition to the delayed genomic steroid actions, increasing evidence for rapid, nongenomic steroid effects has been demonstrated for virtually all groups of steroids, and transmission by so far hypothetical specific membrane receptors is very likely. Nongenomic effects on cellular function involve conventional second messenger cascades, including phospholipase C (PLC²) (Civitelli et al., 1990), phosphoinositide turnover (Morley et al., 1992; Morelli et al., 1993), intracellular pH (Jenis et al., 1993; Wehling et al., 1996), free intracellular calcium ([Ca²⁺]_i) (de Boland and Norman, 1990b; Wehling et al., 1990), and protein kinase C (PKC) (Sylvia et al., 1993). They are clearly incompatible with the involvement of genomic mechanisms. Interestingly, the history of research on rapid responses of steroid hormones actually predates the knowledge on the existence of nuclear receptors. In 1942, Hans Selye was the first to describe a rapid effect of progesterone, which following intraperitoneal application induces a prompt onset of anesthesia in rats (Selye, 1942). In 1963, acute cardiovascular effects of aldosterone in men were demonstrated (Klein and Henk, 1963). Peripheral vascular resistance and blood pressure increased within 5 min, while cardiac output significantly decreased, suggesting a nongenomic mechanism of action because of the short time frame. Almost simultaneously, Spach and Streeten (1964) reported in vitro effects of physiological concentrations of aldosterone on Na⁺ exchange in dog erythrocytes. As a nucleus is absent, in vitro effects in these cells cannot be related to genomic mechanisms and, therefore, must be nongenomic in nature. Furthermore, rapid effects of glucocorticoids on isolated synaptosomes were demonstrated in the mid-1970s, being considered as cellular correlate for the long-known negative feedback mechanism between plasma cortisol and ACTH release occurring within a few minutes (Edwardson and Bennett, 1974). As another early example for a nongenomic steroid effect, Pietras and Szego (1975) showed rapid estrogen action on Ca²⁺ flux in endometrial cells in 1975. Despite these "old" observations, discussion about rapid, nongenomic mechanisms of steroid actions has emerged only recently (Gametchu, 1987; Wehling et al., 1987; Nemere and Norman, 1990).

In the following sections, important aspects of steroid action are summarized. Two sections on mechanisms of steroid actions in principle and the proteins that mediate them are followed by a condensed summary on nongenomic actions of each particular steroid group. Finally, clinical perspectives are discussed and an integrative model of steroid action is developed.

	II. How Do Steroids Act?

Top Previous Next References

According to the common theory of steroid action, steroids modulate gene transcription by interaction with intracellular, nuclear receptors, which act as ligand-dependent transcription factors. Steroids regulate expression of various genes in a network-like manner and initiate complex events involved in nearly every aspect of vertebrate development and physiological responses (Evans, 1988; Beato et al., 1996; Beato and Klug, 2000).

The detailed characterization of steroid actions and mechanisms involved was the result of intensive long-term research on steroid hormones. In the beginning of the 20th century, abnormalities of embryonal development and various diseases had been associated with defects in steroid and thyroid hormone action (Gudernatsch, 1912). The beginning of the modern era of steroid research is marked by a fundamental discovery by Clever and Karlson (1960), who investigated the puff reactions of chromosomes in larvae of insects. Injection of the steroid ecdysone induces changes in chromosomal structure within 2 h. These puff reactions disappear within 24 h, suggesting a link between steroid hormones and the activation of genes.

Subsequently, research has focused on the analysis of cellular and molecular mechanisms involved in related steroid actions on specific target tissues (for review see (Beato, 1989; Fuller, 1991; Beato et al., 1996; Beato and Klug, 2000). Steroids bind to cognate, intracellular receptors representing a superfamily of steroid/thyroid/retinoid/orphan receptors. Interestingly, orphan receptors may have no ligands or as yet undiscovered ligands (Funder, 2000); they seem to open a particularly challenging field of future research.

These receptors act as transcription factors to regulate gene expression by recognizing palindromic hormone response elements (HRE) at the DNA after homo- or heterodimerization of the ligand-receptor complex. Subsequently, transcription is initiated in conjunction with the basal transcription complex, different coactivators, repressors, and transcription regulators (Beato and Klug, 2000). The ligand-dependent modulation of transcription by the ligand-receptor complex has been termed "genomic" and is sensitive to inhibitors of transcription and translation. The expression of steroid-induced genes is modulated at the protein level some hours after stimulation with the steroid, although immediate early genes are differentially expressed after aldosterone stimulation within 1 h (Verrey, 1998).

Unlike intracellular steroid receptors, membrane-bound receptors of other agonists (such as peptide agonists, catecholamines, or platelet-derived growth factor) affect cellular function by modulation of intracellular second messenger levels. In addition to these direct effects of second messengers, agonist-induced changes of intracellular messengers modulate steroid-induced transcription by an intracellular cross-talk. Thus, activation of cells by peptide agonists may modulate steroid-induced nuclear transcription by second messengers induced with an intrinsic ability to modulate nuclear transcription [e.g., cAMP (Nordeen et al., 1994)]. Furthermore, intracellular cross-talk may even occur in the absence of the steroid ligand. Epidermal growth factor activates the estrogen receptor (ER)  by signaling through the MAPK pathway, suggesting that MAPK directly phosphorylates the critical serine 118 of ER (Bunone et al., 1996).

In the past two decades, a growing body of reports dealing with nongenomic steroid action has emerged, which reflects the increasing interest in this field. In these studies a variety of potential mechanisms thought to be involved in rapid steroid action has been described, suggesting that the mechanisms of rapid steroid signaling are not uniform. In this context, a classification of rapid steroid effects in distinct categories, relating to the mechanisms involved, has been proposed and discussed at the "First International Meeting on Rapid Responses to Steroid Hormones" in Mannheim, Germany, in 1998. This Mannheim classification scheme (Fig. 1) (Falkenstein et al., 2000) will help to adequately describe potential mechanisms involved in differential experimental settings and to facilitate the understanding of nongenomic steroid action. The scheme is divided into two major groups termed A (direct steroid action) and B (indirect steroid action), which are subsequently split into a nonspecific (I) and a specific (II) category. The latter is further divided into group a (classic steroid receptor involved) and b (nonclassic steroid receptor involved). In Section III., examples for the categories AI, AIIa, AIIb, and BIIb are given. For categories BI and BIIa, no examples are known to date.

View larger version (19K): [in this window] [in a new window]   Fig. 1.   Mannheim classification of nongenomic steroid actions. Solid arrows indicate examples for categories with examples given in the text. Dotted arrows indicate a hypothetical category with no example yet known. Reproduced with permission from Falkenstein et al. (2000).

Each of the steroid and thyroid hormones displays its own facets of signaling and modulation of cellular functions. Specific nongenomic responses seem to depend on the type of steroid, cells, tissues, or species used. Nevertheless, signaling cascades share large homologies with [Ca²⁺]_i, PKC, PLC, cAMP, pH, MAP kinase, and other traditional second messengers playing major parts of variable, but similar, scores.

	III. Steroid Receptors Mediating Genomic and Nongenomic Steroid Action

Top Previous Next References

The concept that steroids are involved in the regulation of cell function was originally triggered by the above-mentioned observation that the steroid hormone ecdysone induces puffs in giant chromosomes of insects (Clever and Karlson, 1960). All steroid hormones, which are mainly formed in the gonads and adrenals of mammals, regulate a variety of functions in target cells equipped with the cognate steroid hormone receptors. Although steroid hormones and retinoic acid, vitamin D₃, and thyroid hormones are neither structurally nor biosynthetically related, receptors for steroids, retinoic acid, vitamin D₃, and thyroid hormones have been characterized as nuclear receptors or a superfamily of steroid and thyroid hormone receptors due to their close structural homologies (Evans, 1988; Beato and Klug, 2000).

1. Structural Features of Steroid Hormone Receptors. The human glucocorticoid receptor (GR), which has been cloned and expressed as one of the first steroid hormone receptors in the early 1980s, exists as a 777-amino acid, ligand-binding GR displaying close homologies to the viral oncogen erbA and 742-amino acid -isoform, which differs in the last 115 amino acids and does not bind glucocorticoids (Hollenberg et al., 1985). The binding characteristics of the GR are consistent with pharmacological properties of glucocorticoid-induced effects shown previously by in vivo and in vitro studies describing a high-affinity binding for the synthetic glucocorticoid dexamethasone and low-affinity binding for mineralocorticoids (Lee et al., 1988; Gottschall et al., 1991; Lemberger et al., 1994). Subsequently, the mineralocorticoid receptor (MR) (Arriza et al., 1987) and receptors for estradiol (ER) (Greene et al., 1986; Krust et al., 1986), progesterone (PR) (Loosfelt et al., 1986; Misrahi et al., 1987), androgens (AR) (Chang et al., 1988; Lubahn et al., 1988), vitamin D₃ (VDR) (McDonnell et al., 1987), retinoic acid (Petkovich et al., 1987), and thyroid hormone (Weinberger et al., 1986; Giguere et al., 1988) have been cloned, sequenced, and functionally expressed. The -isoform of the GR has been regarded to be a cloning artifact for a long time; however, variants of steroid receptors and receptor isoforms generated by differential promoter usage have been described for nearly all steroid receptors (Kastner et al., 1990; Kuiper et al., 1996; Zennaro et al., 1997). However, the distinctive role of each of those variants is currently not known in detail.

View larger version (12K): [in this window] [in a new window]   Fig. 2.   General structure and functional organization of steroid and thyroid hormone receptors. Steroid receptors are structurally organized in different domains, which have been confirmed by results of cDNA cloning experiments: 1) a variable N-terminal region, which may have modulatory effects on transactivation (regions A/B, with a variable length of 50-500 amino acids). Most of the receptor antibodies described are directed against this part of the protein. 2) a well conserved cysteine-rich central domain, which exhibits cysteine residues compatible with the formation of zinc fingers (region C and D, with a length of ~70 and 45 amino acids). This part of the protein may be responsible for the DNA binding activity of the receptors (Evans, 1988). 3) a C-terminal domain, which is obviously responsible for hormone binding and nuclear translocation (region E, with a length of 220-250 amino acids). The functional assignment of amino acid sequences derived from cDNA cloning experiments is indicated below. NT, nuclear transcription; T, transactivation; D, dimerization; HSP, 90-kDa heat-shock protein binding. Modified according to Beato (1989).

2. Genomic Steroid Hormone Action. It is assumed that the lipophilic steroid hormones enter the respective target cells by simple diffusion, although the matter of active transmembrane transport is still under debate but unsettled (Allera and Wildt, 1992). Steroid hormone receptors are associated with a complex of chaperone proteins in the unliganded state (Pratt and Toft, 1997; Defranco, 2000). Upon binding of steroids to the cognate steroid hormone receptor in the cytosol, the heat-shock protein Hsp 90 and the immunophilin Hsp 56, which maintain the receptors in an inactive form with high affinity for the steroid hormones, dissociate from the receptors (Pratt and Toft, 1997). This transformation of the steroid hormone receptor is associated with an increased affinity of the receptor to DNA and a decrease in complex size. The chaperones are probably necessary to keep the steroid hormone receptors functional (Godowski and Picard, 1989).

3. Steroid Hormone-Responsive Elements. After translocation into the nucleus, the ligand-receptor complex binds to palindromic DNA sequences. The receptors for glucocorticoids, mineralocorticoids, androgens, and progestins bind to the same HRE, which have been originally described as glucocorticoid response elements. HRE are hexanucleotide halves arranged as inverted repeats and separated by three nonconserved base pairs [AGAACAnnnTGTTCT (Beato, 1989)]. The sixth base pair of each half-palindrome is not well conserved, and its identity is not essential for specific binding (Scheidereit et al., 1983). Initially it was believed that all members of the nuclear receptor family bind as homo- or heterodimers to the palindromic HREs in the promoter region of target genes, while each part of the HRE is recognized by one receptor monomer. However, the identification and characterization of the orphan estrogen-related receptors ERR and ERR revealed that those receptors bind the DNA as monomers and homodimers (Johnston et al., 1997; Vanacker et al., 1999). Furthermore, the predominant form of GR is monomeric in solution, whereas dimers of GR are formed after binding of the ligand-receptor complex to an HRE. That the strength of a weak dimerization region within the DBD and the LBD is modified by cooperativity observed during DNA binding is obviously responsible for this dimerization (Eriksson et al., 1995).

4. Steroid-Induced Initiation of Transcription. After binding of the respective DNA-recognition sites as homo- or heterodimers to the components of the basal transcriptional machinery and sequence specific coactivators, transcription starts or is down-regulated. Regulation of transcription not only depends on interaction with the consensus nucleotide sequence of the HRE, but also on interaction with the specific assembly of transcription factors and polymerases. Being sensitive to inhibitors of transcription and translation, related long-lasting physiological responses are classified as genomic actions of steroids. Among the earliest genomic steroid effects known is the increased rate of mouse mammary tumor virus long terminal repeat transcription in Ltk-aprt cells first seen within 7.5 min after glucocorticoid application (Groner et al., 1983). For mineralocorticoids, this effect starts within 30 min and peaks after 3 h in a feline renal cell line (Cato and Weinmann, 1988).

5. Alternative, Including Nontranscriptional Actions of Ligand-Steroid Hormone Receptor Complexes. In addition to the regulation of gene expression at the transcriptional levels, gene expression may be modulated by the interaction of nuclear receptors with sequence-specific transcription factors. For example, glucocorticoids affect the activity of NFB, an important modulator of cytokine-induced inflammation in at least two ways. Glucocorticoids genomically increase the expression levels of the inhibitor IB, which traps NFB in the cytoplasm (Auphan et al., 1995; Scheinman et al., 1995). In addition, GR interacts with p65, a transcriptionally active subunit of NFB, by protein-protein interaction (Ray and Prefontaine, 1994). Thus, glucocorticoids elicit distinct effects in different target tissues by direct actions at the transcriptional level and effects mediated by direct protein-protein interactions, which should be termed nontranscriptional activities of classic steroid receptors.

B. Receptors Responsible for Nongenomic Steroid Action

1. Classic Intracellular Receptors (Classification AIIa). In various studies it has been demonstrated that classic steroid hormone receptors may be involved not only in genomic steroid action, but also in rapid nongenomic steroid effects. Rapid signaling is inhibited by classic antagonists of these receptors, as demonstrated for the AR, the ER, and the GR. As an example for category AIIa (Fig. 1), data supporting an involvement of ER in nongenomic estrogen effects are given below.

8

3. Nonclassic Steroid ReceptorsCoagonist-Mediated Steroid Action (Classification BIIb). Over the last decade, substantial experimental work has been carried out that investigated the metabolism and effects of various steroids in the brain and the central nervous system (CNS). Most effects are not mediated through nuclear steroid hormone receptors but through ion-gated neurotransmitter receptors. The potential of neuroactive steroids to modulate the -aminobutyric acid (GABA)A receptor as allosteric coagonists or antagonists of GABA, or psychoactive drugs such as benzodiazepines and barbiturates, has attracted the most interest (category BIIb, Fig. 1). The mechanisms by which neuroactive steroids alter the excitability of GABAergic neurons depend on the specific structure of the GABA_A receptor with its subunits forming ligand-gated ion channels. Glycine-, nicotinic acetylcholine-, and 5-hydroxytryptamine type 3 (5-HT₃) receptors show remarkable homologies to the GABA receptors (Paul and Purdy, 1992; Lambert et al., 1995; Wetzel et al., 1998). GABA receptors are heterooligomeric proteins that contain a number of allosterically interacting binding sites for the neurotransmitter GABA, as well as for benzodiazepines and barbiturates. The first steroids shown to modulate the neuronal excitability by interaction with GABA_A receptors were 3,5 tetrahydroprogesterone (3,5-THP) and 3,5 tetrahydrodeoxycorticosterone (3,5-THDOC) (Majewska et al., 1986). These steroids are potent barbiturate-like ligands of the GABA receptor-chloride ion channel complex. At concentrations between 100 nM and 10 µM, both steroids inhibit binding of the convulsant t-butylbicyclo-phosphorothionate to the GABA receptor complex and, as coagonists at the GABA_A receptor, increase the binding of flunitrazepam. They also stimulate chloride uptake into isolated brain vesicles and potentiate the inhibitory actions of GABA in cultured rat hypothalamic neurons (Wetzel et al., 1999). In contrast to the pharmacological activity of benzodiazepines, which varies with the -subunit composition and requires the presence of a -subunit, the effects of neuroactive steroids do not depend on such strictly defined basic requirements for their structure-activity relationship (Puia et al., 1990). Studies investigating this relationship were able to delineate the presence of a 3-OH group within the A-ring of neuroactive steroids as the crucial determinant for a positive allosteric interaction at the GABA_A receptor to enhance GABA or benzodiazepine action (Gee et al., 1988; Paul and Purdy, 1992). All 3-hydroxysteroids that have been investigated so far seem to be inactive in increasing GABA_A receptor-mediated Cl conductance or flux (Purdy et al., 1990; Paul and Purdy, 1992), leading to the conclusion that 3-hydroxysteroids have a distinct stereoselectivity at the GABA_A receptors. In contrast, the 3-reduced pregnane steroids dehydroepiandrosterone sulfate (DHEA-S) and pregnenolone sulfate have been shown to exert GABA-antagonistic properties at the GABA_A receptor (Lambert et al., 1995; Rupprecht, 1997; Shen et al., 1999). This allosteric antagonism to GABA at the receptor, as well as the described coagonistic activity of other neuroactive steroids, confers a multitude of functional effects. These are briefly discussed below but have been extensively reviewed elsewhere (Paul and Purdy, 1992; Lambert et al., 1995). In addition to the 3-OH group within the A-ring of neurosteroids, there may be other components to the structure activity relationship of neurosteroids at the GABA_A receptor. It has recently been shown that 6-oxa analogs of the neurosteroid 3-hydroxy-5-pregnan-20-one, which do not possess the carbon atom 6 within the B-ring, have an approximately 100-fold reduced potency for modulating flunitrazepam binding to the GABA_A receptor compared to their natural carbon analogs (Nicoletti et al., 2000). Certainly, the field of structure-activity relationship is still wide open for neurosteroids, and the complex interaction of neurosteroids with the GABA_A receptor needs further in depth investigation.

4. No Receptor InvolvedDirect Nongenomic Action (Classification AI). In addition to the above-mentioned specific receptor-mediated actions, direct steroid-membrane interactions occurring without receptor involvement have been described that alter physicochemical membrane properties such as the fluidity and the microenvironment of membrane receptors. This intercalation of steroids in phospholipid bilayers may occur at high, nonphysiological steroid concentrations. The corresponding effects are termed as nonspecific, nongenomic steroid actions (classification AI, Fig. 1).

	IV. Steroid Groups

Top Previous Next References

A. Gonadal Steroids

1. Progesterone. a. Rapid Effects of Progesterone. Since the pioneering work of Selye in 1942, which has led to the development of a variety of steroidal anesthetics, a growing number of reports dealing with rapid, nongenomic actions of progesterone has been published.