Which of the following describes how steroid hormones regulate gene expression?

Other Hormone-Response Elements

HREs for other steroid hormones have been identified, as have those for thyroid hormones, and for all of these cases the mechanisms involved appear to be similar to those for glucocorticoids. Various other hormones regulate gene expression, acting indirectly, via cell surface-associated receptors rather than directly at the nucleus. Details of the steps between the plasma membrane-associated receptor and the nucleus are incompletely understood, but rapid progress is being made. Thus, for hormones whose effects are mediated by the second messenger cyclic AMP, HREs [CREs] have been identified [using transfection experiments of the type described above] which mediate actions at the gene level, although the nature of the proteins that bind to these HREs is not fully understood. Some progress has also been made in identifying HREs for hormones that act by elevating Ca2+ levels and other second messengers.

C HORMONE RESPONSE ELEMENTS

Most HREs fall into the enhancer/silencer class of cis-acting elements since these DNA sequences, and their associated trans-acting factors, can function through heterologous promoters and in a relatively position- and orientation-independent manner [19]. However, in some cases HREs can also function as basal promoter elements. Much more is known about HREs acting as enhancers, although in certain instances HREs can act as silencers.

Biochemical [20] and genetic [21] evidence that glucocorticoid receptors bind to DNA was followed by data that showed that certain regions of DNA, containing specific glucocorticoid receptor-binding sites, confer regulation to a gene ordinarily unresponsive to glucocorticoids [19,22]. Numerous studies of this type, employing the promoter regions of several hormone-responsive genes ligated to a variety of reporter genes [forming so-called fusion genes] led to the identification of cis-acting elements that mediate the response to several hormones. The strategy used to define such elements has been reviewed in detail elsewhere [23].

Hormone response elements [HREs] for steroid, thyroid, sterol, and retinoid hormones have been identified, as has a cAMP response element [see Fig. 2]. The hormone receptor is the trans-acting factor in the case of the steroid-thyroid family of HREs. The observation that a GRE, MRE, PRE, and ARE all have the same consensus sequence appears to defy the rules of physiology since it implies that the same sequence mediates the effects of all these hormones. This consensus was established through the observation that the sequence GGTACAnnnTGTTCT will mediate the effect of the respective hormone-receptor complexes on the mammary tumor virus gene, but it is not likely to mediate the diverse effects these hormones have on mammalian genes. Various explanations may resolve this enigma. Subtle sequence differences may occur in specific HREs, receptor concentrations may be limiting in different cell types, ligand availability may be limiting, or HREs might not exist as isolated elements but in fact interact with other cis/trans elements to mediate their effect. Such complex hormone response units have been the subject of several recent publications [24–27].

Recently, a protein that interacts with the cAMP-response element [CRE], the CRE-binding protein [CREB], has been purified, and its cDNA has been cloned [28–30]. The identification of the CRE was a prerequisite in the purification and characterization of CREB, just as the identification of an insulin-response sequence [1RS] will be the first step in the identification and isolation of the transacting factor[s] involved in the regulation of an insulin-responsive gene.

Gonadotropin Subunit Genes

GnRH-responsive elements exist within the promoter regions of gonadotropin subunit genes, and the response of these genes to GnRH stimulation involves a large number of intracellular proteins.101,102 GnRH induces early response genes in gonadotropes, and these act to regulate transcription of the subunit genes.103 Studies in castrated, testosterone-replaced rats [in which gonadotropin secretion is suppressed] showed that administration of GnRH at frequencies of 8–30 min increased the mRNA levels of α subunit and LHβ subunit, but the levels of FSHβ subunit were increased at lower pulse frequencies of 120–480 min.104,105 Further regulation is through the feedback effects of gonadal steroids, but these differ both between species and in the way that the steroids regulate expression of the gonadotropin subunit genes. The means by which sex steroids may regulate gonadotropin subunit genes is complicated and appears to involve regulation of transcription factors, rather than the binding of liganded steroid receptors to consensussequences on the promoters of the genes for the subunits.106,107

The consensus sequence of the core motif

The TRE consensus sequence found is 5′-AGGTCA as for RARs, RXRs, PPARs or VDRs, which are all able to recognize the same DNA sequence [reviewed in references 14 and 15]. However, there is evidence for some differences in the [natural] response element repertoires of these receptors. It has been shown, for example, that TRα is able to bind to both 5′-AGGTCA and 5′-AGGACA motifs40–42. Such differences could be further enhanced by cooperative DNA binding with other promoter-bound factors and could contribute to the ability of a given target gene to respond preferentially to a particular signaling pathway.

6. Simple AuxREs

While animal HREs may take the form of composite elements, they may also consist of simple elements [i.e., simple HREs] with direct, inverted, or everted repeats of steroid hormone receptor binding sites [225]. With simple HREs, nucleotide composition, orientation, and spacing between the DNA binding sites determine which steroid hormone receptor recognizes and prefers to bind to a given HRE. Recent evidence indicates that, like simple HREs, simple AuxREs can be created by constructing direct and palindromic repeats of the TGTCTC element.

A synthetic construct referred to as P3[4X] that consisted of four palindromic repeats of the TGTCTC element functioned as a strong AuxRE in carrot protoplast transient assays [223]. Analysis of single copy palindromic constructs indicated that orientation and spacing of TGTCTC elements was important for auxin responsiveness, but that the nucleotide composition between the TGTCTC repeats was not particularly important. With everted repeats, spacing between half-sites was found to be crucial for AuxRE activity, and spacing of 7–8 bp was optimal. Tandem direct repeats of the TGTCTC element in either direction may also function as AuxREs when spaced appropriately [224]. These direct repeats may be simple elements, but it has not been ruled out that the direct repeats may represent composite elements with novel, undefined coupling elements [i.e., coupling elements that overlap the TGTCTC element] [224]. In another study, a construct containing twelve tandem direct repeats of the TGTCCCAT element fused to a minimal PS-IAA4/5 promoter-CAT reporter gene was shown to function as an AuxRE in transient assays with pea protoplasts [220]. It should be noted that the TGTCCCAT sequence used by Ballas et al. [220] conforms to the concensus TGTCNC element, and it has not been reported whether the 3’ terminal AT contributes to AuxRE activity.

TGTCTC AuxREs that were defined at the fine structure level in natural promoters [i.e., GH3 promoter] function as composite elements, and it had not been shown that simple AuxREs [i.e., TGTCNC elements with no coupling element] function in natural promoters. In this regard, Ulmasov et al. [223] pointed out that an everted repeat of the TGTCNC element is found in an auxin-responsive region of the PS-IAA4/5 promoter [220]. This everted repeat, TGTCACccctataagGAGACA, functioned as an AuxRE in carrot protoplast transient assays when fused to a minimal promoter-GUS reporter gene [223], suggesting that simple AuxREs may exist in natural promoters. Taken together, the results on direct and palindromic repeats of the TGTCTC element [or TGTCNC] indicate that if properly multimerized and spaced, this element may be sufficient to confer auxin inducibility in the absence of any constitutive or coupling element.

C THE CELL-SPECIFIC ELEMENTS [CSE] AND THE THYROID HORMONE RESPONSE ELEMENT [TRE] APPEAR TO COMPRISE AN ENHANCER-LIKE UNIT THAT CONFERS BOTH CELL-TYPE SPECIFICITY AND THYROID HORMONE REGULATION OF GROWTH HORMONE GENE EXPRESSION

The thyroid hormone response element contained in the −236/146 fragment functions most efficiently with the homologous rat growth hormone gene promoter and this is independent of the helical relationship between the two regions [Flug et al., 1987]. No l-T3 stimulation occurred when the −236/ −146 fragment was ligated [Flug et al., 1987] to the enhancerless SV40 promoter in pA10-cat2 [Laimins et al., 1984]. In the studies presented here no l-T3 stimulation was also found when DNA from −236 to −146 was placed upstream of an enhancerless RSV promoter [pRSVΔE-cat[2]]. However, hormone-regulated expression occurred when a fragment extending from −236 to −47 was ligated to the RSV promoter in either normal [pRSVΔE-cat[3]] or inverted [pRSVΔE-cat[4]] orientations but not 3’ of the CAT gene [pRSVΔE-cat[5]] [Figs. 11 and 12]. We have also recently found that both the thyroid hormone and the cell-specific elements are required to confer l-T3 regulated expression on the enhancerless SV40 promoter of pA10-cat2 [unpublished observation]. These results indicate that both the distal thyroid hormone response element and a cell-specific basal element are necessary for l-T3 regulated expression of the growth hormone gene. Since these sequences function in an orientation independent manner [Figs. 11 and 12], they appear to behave as a functional enhancer unit which can confer both cell-specific and thyroid hormone-regulated expression.

To account for our observations we propose a mechanism for regulation of the rat growth hormone gene where two elements are required for l-T3 stimulation in which hormone-receptor-cell-specific protein interactions occur. In this model we assume that the receptor binds to sequences in the −208/ −178 region [Fig. 20B] and that the l-T3-receptor complex acts to “stabilize” or “enhance” the protein-DNA interactions of the cell-specific basal element. This results in the formation of a more “active” or “stable” transcription complex which stimulates the level of gene expression. Although we have not verified that the functional element in the −208/ −178 region binds receptor, the 31 nucleotide sequence contains an interesting region of reverse dyad symmetry [Fig. 20B]. This region also shows some homology with 5′-flanking DNA of other rat genes which are stimulated by l-T3 after in vivo administration [Flug et al., 1987].

Recent studies on the mouse mammary tumor virus [MMTV] promoter support an analogous two element model for glucocorticoid hormone stimulation of gene expression. Inactivation of the binding site for the NF-I transcription factor [or a protein with a similar recognition sequence, e.g., TGGCA-binding protein] markedly lowers stimulation by glucocorticoid hormones without significantly altering the basal activity of the gene [Buetti and Kuhnel, 1986; Miksicek et al., 1987]. Furthermore, in vivo footprinting indicates that glucocorticoid hormone incubation increases the association of NF-I with its cognate sequence [Cordingley et al., 1987] suggesting that the glucocorticoid receptor enhances NF-I binding or interacts with NF-I to form a transcription complex which activates expression of the MMTV promoter. Since the control elements of other thyroid hormone responsive genes have not yet been identified, it is not possible to determine whether two elements are necessary for l-T3 regulated expression in other systems. However, it provides a mechanism to explain how thyroid hormone, presumably acting via the same receptor, can positively [e.g., rat growth hormone] or negatively [e.g., thyroid-stimulating hormone] [Kourides et al., 1984; Shupnik et al., 1985; Carr et al., 1987] regulate gene expression. Whether positive or negative regulation occurs would be dependent on the hormone-receptor complex acting in cis to enhance or suppress the effect of a second trans-acting regulatory protein[s] which plays a central role in determining the rate of expression of the gene.

Additional studies are required to support this hypothesis. For example, positive and negative regulation may be mediated by structurally similar but different thyroid hormone receptors. The observation that the human genome contains at least two erbA related genes on different chromosomes raises this possibility [Dayton et al., 1984; Spurr et al., 1984; Raines et al., 1985; Weinberger et al., 1986]. Significant progress has been made in the area of thyroid hormone action since the initial identification of thyroid hormone nuclear receptors in liver and kidney [Oppenheimer et al., 1972] and in cultured cells [Samuels and Tsai, 1973]. Future advances in this field will require the cloning of thyroid hormone receptor mRNAs from various tissues and cells to identify their structure and function. In addition, other thyroid hormone response genes are being isolated and sequenced. Identification of the thyroid hormone response elements of other genes and the regulatory proteins which mediate their expression should provide a comprehensive view of the detailed mechanisms involved in positive and negative regulation of thyroid hormone responsive genes in various cells and tissues.

Transcriptional Regulation

Studies of the MMTV HRE have been a paradigm for hormone-regulated gene expression. MMTV RNA levels increase c. 10–50-fold in the presence of glucocorticoids, progesterone, or androgens at the level of transcriptional initiation. Hormone inducibility is conferred by multiple binding sites to allow cooperative binding with the consensus TGTTCT in the region from –80 to –190 [Figure 4]. Glucocorticoids are believed to bind to their receptors [GRs] in the cytoplasm and to subsequently translocate to the nucleus where they exchange rapidly with their binding sites. The integrated MMTV LTR is occupied by six nucleosomes [A–F] [not shown] with the A nucleosome located at the transcription initiation site [+1]. The standard MMTV promoter contains a TATA element [−30 bp], which binds transcription factor IID [TFIID]. Hormone-bound GR is believed to recruit remodeling factors to increase accessibility at nucleosome B and to allow binding of transcription factors Oct1 and NF1 between the TATA box and the HRE.

The tissue distribution of MMTV expression is tightly linked to viral transcriptional control, yet steroid receptors are present in many tissues in which MMTV transcription is repressed. The importance of negative regulation in MMTV transcription and disease specificity is exemplified by the isolation of MMTV variants, such as type B leukemogenic virus [TBLV], which induce T-cell lymphomas rather than mammary tumors. Such variants lack all or part of the negative regulatory element [NRE] present upstream of the HRE and acquire a T-cell specific enhancer. The NRE binds at least two related homeodomain-containing proteins called special AT-rich binding protein 1 [SATB1] and Cut-like protein1/CCAAT displacement protein [Cutl1/CDP], which have different tissue distributions and act as repressors of MMTV transcription. The highest levels of SATB1 are expressed not only in T cells but also in B cells explaining transcriptional repression in most lymphoid tissues. Mutations of the promoter-proximal SATB1-binding site elevate MMTV expression of LTR-reporter constructs in cultured cells and in lymphoid tissues of transgenic mice. Interestingly, CDP expression is high in undifferentiated B cells and mammary cells and decreases during differentiation. In mammary epithelial cells, where SATB1 is also absent, CDP is cleaved during differentiation to yield a dominant-negative protein that interferes with full-length CDP binding to the MMTV LTR. Therefore, the levels of functional CDP repressor are lowest when MMTV particle production is highest at lactation, the period when virus transmission occurs in milk.

MMTV specifies several enhancer elements, including the HRE and a mammary gland enhancer [MGE] near the 5′ end of the LTR. The MGE contains binding sites for multiple nuclear factors, including MP4, NF1, AP2, Ets, and C/EBP. Synthesis of some factors is inducible by prolactin, epidermal growth factor, or tumor necrosis factor α. The absence of NRE-binding repressors and ligand-bound steroid receptors presumably cooperate with the HRE to maximize virus production in milk.

2.2.2 LHRE-Luciferase reporter assay

Principle: The lactogenic hormone response elements [LHRE]-luciferase reporter gene is an expression vector encoding the firefly luciferase under the control of an artificial promoter containing a six-repeat LHRE from the β-casein promoter, a classical PRL target gene. It is mainly mediated by Stat5 cascade, although participation of other signaling molecules cannot be formally excluded. The principle of this bioassay is thus to monitor the ability of any PRLR [WT or variant] to induce transcriptional activation of LHRE artificial promoter by measuring the level of luciferase produced by the cell. It should be used only as a “reporter” or receptor triggering, that is, that any pathophysiological interpretations should be avoided.

Bioassay: After trypsinization, cells are counted and aliquoted in 96-well plates at a density of 50,000 cells/100 μl/well. Although this assay has been classically run in low serum condition [0.5% FCS] to minimize luciferase background induced by any serum component [Bernichtein et al., 2003], growth medium [10% FCS] can be recommended as [i] it ensures better cell adhesion, which will minimize variability resulting from cell aspiration when medium is removed prior the lysis step, and [ii] bovine PRL [potentially present in FCS] was recently shown to not activate the human PRLR [Utama et al., 2009]. Eighteen hours after plating [overnight], 100 μl of twofold concentrated hormones diluted in same medium are added to each well [add only medium in negative control]. After another 18–24-h period, culture medium is aspirated and cells are lysed under gentle agitation for > 10 min in 50 μl of lysis buffer [Promega]. Ten to 20 μl of lysate is then transferred into white 96-well plates and luciferase activity is monitored in a bioluminescence multiplate reader [e.g., LB 940 Mithras, Berthold technologies]. When using transient transfection, firefly luciferase [PRLR-induced] and Renilla luciferase [PRLR-nonresponsive, used as transfection control] will be measured using Dual-Luciferase® Reporter Assay System [Promega]. If stable clones [receptor + LHRE-luciferase] are used, only the former is measured.

Data analysis: Raw data [relative light units, RLU] can vary depending on transfection efficiency, cell number, lysis efficiency, sensitivity of plate reader, etc. When using transient transfections, normalization of firefly/Renilla luciferase RLU is performed to normalize all values with respect to these variable parameters. Transcriptional activity can be expressed either as renilla-normalized raw data, or as fold induction, which is the ratio between RLU obtained for stimulated versus nonstimulated conditions. Fold induction presents the advantage to facilitate comparison [as nonstimulated conditions always has value = 1], but it masks part of the information as the efficacy of the system reflected in absolute values is not displayed [see discussion section below].

Regulation of the Oxytocin Gene

Regulatory sequences comprise a composite hormone response element [−164 to −145 and −110 to −79 upstream] through which nuclear receptors such as estrogen receptor β, thyroid hormone and retinoic acid receptors stimulate, and orphan receptors stimulate [SF-1; TR [testicular orphan receptor]4] or competitively repress [COUP-TF] expression. Control of hypothalamus-specific expression is exerted via the intergenic region between the oxytocin and vasopressin genes.

Progesterone, or its neurosteroid metabolite allopregnanolone, restrains oxytocin gene expression in pregnancy, and a transient post-partum increase in expression follows progesterone withdrawal and stimulation by estrogen. Progesterone action is indirect, as oxytocin neurons lack progesterone receptors, but allopregnanolone acts to amplify inhibitory GABA actions via allosteric modulation of GABAA receptors on oxytocin neurons. Some rostral inputs to oxytocin neurons [in the medial preoptic area and lamina terminalis] contain estrogen receptor-α, indicating an indirect regulation by estrogen.

In lactation, oxytocin gene expression, as reflected in cell body content of oxytocin mRNA, is increased only after a week of repeated stimulation by the suckling stimulus, when the posterior pituitary oxytocin store is depleted. Evidently, the level of oxytocin gene expression is set to provide ample oxytocin to meet sustained demand with a wide safety margin for normal reproductive purposes.

3 5α-Steroid Reductase Inhibitors

Intracellular androgen receptors bind to DNA hormone response elements that control cellular proliferation and apoptotic responses in transformed prostate cells. Dihydrotestosterone controls androgen receptor action by binding to the intracellular receptor. The 5α-steroid reductase types 1 and 2 isozymes catalyze the synthesis of dihydrotestosterone from testosterone, thus controlling intracellular androgen reception function. Targeting 5α-steroid reductase with finasteride, a selective, competitive inhibitor of type 2 5α-steroid reductase,133 deprives the transformed prostate cell of proliferation signaling. In the 3,2′-dimethyl-4-aminobiphenyl [DMAB], methylnitrosourea [MNU], and testosterone chemical carcinogenesis models in rats, finasteride reduces prostate tumor incidence by close to six fold. Finasteride appears to be more effective in the promotion phase of prostate carcinogenesis.134 A randomized, placebo-controlled cancer incidence endpoint risk reductive clinical trial of finasteride demonstrated that the finasteride-treated arm reduced the prevalence of prostate cancer by 24.8%. The initial analysis suggested that tumors of high Gleason's grade24–27 were higher in the finasteride arm [37%] compared to the placebo arm [22%]; however, subsequent analyses have suggested that the increase in high-grade prostate cancer was due to PSA and DRE prompted biopsies. For men with the greatest exposure to the drug [those with an end of study biopsy], there was no significant increase in risk. Sexual function side effects [erectile dysfunction, loss of libido, gynecomastia] were more common in the finasteride treatment arm.135,136 Recent analyses of longer-term follow-up from this trial demonstrate that finasteride significantly enhanced the ability of PSA to detect prostate cancer and high-grade prostate.137 Dutasteride inhibits both 5α-steroid reductase inhibitor types 1 and 2 isoforms, whereas finasteride inhibits only the type-2 isoform. In a cancer randomized, placebo-controlled prostate cancer risk reduction trial, dutasteride reduced the relative risk of prostate cancer by 22% and the absolute risk reduction of 5.1%. Similar to the finasteride study, in the final 2 years of observation, more high-grade [Gleasons 8 to 10] tumors were observed in the dutaseride-treated arm than in the placebo arm.138