Background
Most estrogen actions are mediated by specific nuclear estrogen receptors (ERs), which classically regulate transcription by binding as dimers to specific estrogen-response elements (ERE) found in promoters of estrogen-responsive genes (ERGs) [
1]. This transcriptional activity is dependent on conformational changes of the ERs two activation functions (AF): the N-terminal (ligand-independent) AF-1 and the C-terminal (ligand-dependent) AF-2, which function independently or synergistically to recruit and interact with coregulator (coactivator or corepressor) proteins leading to changes in the rate of gene transcription [
2]. Two ER subtypes (α and β) are present in most vertebrates, although in teleost fish one ERα and two ERβ genes (βa and βb) have been identified [e.g. [
3‐
5]].
There is also substantial evidence that estrogens also function via non-classical mechanisms [
1]. These include indirect transcriptional activation through interaction of ligand-bound ER with other transcription factors, ER ligand-independent activation in response to intracellular signaling cascades, and rapid non-genomic actions initiated at the plasma membrane. However, it is not clear if these are mediated by a subset of nuclear ERs that localize to the plasma membrane, or by novel membrane receptors unrelated to ER or through both [
1,
6,
7].
ERs are known to accept a wide range of ligands, including natural estrogens, synthetic estrogens or antiestrogens, phytoestrogens and a variety of xenoestrogens [
8]. While many behave as estrogen agonists, other compounds may act either as agonists or antagonists depending on the species, tissue, promoter or ER subtype (the selective ER modulators, SERMs; e.g. tamoxifen) [
9]. The tissue-selective effects of SERMs have been exploited to develop new drugs for the treatment of estrogen-related diseases, although some have unwanted side effects or generate resistance to treatment, in part due to their agonist effects in some tissues [
10].
ICI 182,780 (trade names Faslodex, Fluvestrant) belongs to a new class of antiestrogens developed to have no agonistic effects, and besides its therapeutic potential demonstrated in clinical trials, it has been used as an alternative and efficient means to "knock-out" ER effects in studies of estrogen functions, and to establish the contribution of nuclear ERs to particular estrogen actions [
11‐
14]. In mammals, ICI 182,780 appears to act at several levels to block estrogen actions [reviewed by [
11,
15]], but little is known about its effects and mechanisms of action in fish. ICI blocked E
2-induced interstitial cell proliferation in immature rainbow trout (
Oncorhynchus mykiss) testis [
16] and the production of zona radiata proteins and vitellogenin in Atlantic salmon (
Salmo salar) hepatocytes [
17] and of vitellogenin in channel catfish (
Ictalurus punctatus) [
18] and Siberian sturgeon (
Acipenser baerii) [
19] hepatocytes. Agonistic actions have been identified in Atlantic croaker (
Micropogonias undulatus), in which both estradiol-17β (E
2) and ICI decreased gonadotropin-stimulated 11-ketotestosterone production in testicular fragments
in vitro, although these rapid effects appeared to be mediated by interaction with membrane-bound receptors [
20,
21].
The objective of this study was to investigate the effects and mechanisms of action of ICI 182,780 (ICI) on several typical
in vivo estrogenic responses in the teleost fish sea bream (
Sparus auratus), in order to evaluate the potential of ICI as an agent to knock-out estrogen effects in fish. In a preliminary experiment with the more readily available tilapia (
Oreochromis mossambicus) we established that ICI potentiated the calciotropic effect of E
2 [
22‐
24]. The effects of E
2 and ICI on plasma calcium and on hepatic and testicular gene expression of the three ER subtypes and the estrogen-responsive genes vitellogenin II and choriogenin L (egg yolk and eggshell precursors, respectively) were then analyzed in sea bream, for which these molecular markers were available [
5,
25].
Discussion
This study demonstrated that, at least in fish, ICI does not always function as an anti-estrogen since it did not block the effects of an E2 challenge. Indeed, prior administration of ICI potentiated the response to E2. Furthermore, the agonistic response to ICI could also be detected at the level of gene expression and was different in liver and testis.
The level of total calcium in plasma is known to correlate with Vg protein and E
2 plasma levels in females during vitellogenesis and in males in response to E
2 exposure, and it is thus used as a vitellogenesis marker [
22‐
24]. The lack of a statistically significant calcium elevation with E
2 treatment alone in sea bream was probably due to the low doses and/or exposure time compared to previous sea bream experiments (>4 days, 10 mg/kg) [
27] and to the tilapia experiment (>48 h, 10 mg/kg). ICI alone did not change total plasma calcium levels or may have slightly lowered calcium within 48 h (Figures
1 and
3), consistent with an antagonistic action. However, pretreatment with ICI synergistically potentiated the hypercalcemic effect of E
2 in both sea bream and tilapia. This observation seems to indicate that the initial binding of ICI to ER effectively blocks ER binding to target genes (antiestrogenic action) in the liver, but subsequently E
2 triggers a disproportionate agonistic response. Whether ICI acts by stimulating ER synthesis or at the level of ER responsiveness is not clear. In support for the first possibility is the fact that ERα levels in the liver of fish treated with ICI are upregulated and at similar levels to the E
2-treated fish. However, it is surprising that ERα levels in E
2 challenged fish after pretreatment with ICI are no higher than fish treated with ICI only (Figure
4). Analysis of the early time-course changes in this response is required to clarify the possible mechanism involved.
The ERα, VgII and ChgL up-regulation by E
2 in liver is in accordance with our previous observations [
25]. ERα autoregulation in liver is a common characteristic of oviparous animals [e.g. [
29‐
32]] that has been attributed to ER involvement in the production of egg yolk and egg shell precursors vitellogenins and choriogenins, respectively, in the liver of mature females in response to E
2. In contrast to ERα, the regulation of ERβ genes by estrogen is poorly investigated and appears to be much more variable, with these genes being either slightly up- or down-regulated depending on the species and ER subtype [
33,
34]. In this study, the expression of both ERβa and ERβb are strongly down-regulated by E
2 in liver in the first sea bream experiment, and only slightly down-regulated in the second experiment in which lower doses are used, possibly indicating that their regulation is dose-dependent (and less sensitive to E
2 than ERα). These results indicate a differential estrogen regulation of sea bream ERα and ERβ genes in liver and support the hypothesis that the role of the ERβ forms, in the transcriptional regulation of genes associated with reproduction in fish liver is probably less important than that of the ERα subtype and may depend on the life stage of the fish and/or the species. In contrast to the liver, a slight up-regulation by E
2 of both ERα and ERβb but not ERβb is detected in testis, suggesting the regulation of ER subtypes varies among tissues, while the up-regulation of VgII and ChgL confirms its recent identification as ERGs in this tissue [
25].
ICI mimicked the E
2 effects in the liver, up-regulating ERα, VgII and ChgL in sea bream liver, but not in the testis and unlike E
2 it did not down-regulate the two ERβ subtypes in liver, supporting tissue- and gene-specific effects for this compound. The simultaneous administration of ICI with E
2 did not block the E
2 effects on the expression of any of the genes, neither did the ICI pretreatment in the E
2-induced up-regulation of ChgL and VgII, suggesting that ICI did not act as an antagonist. However, ICI pretreatment synergistically potentiated the E
2 down-regulation of the ERβb gene, while it appeared to have an inhibitory effect on the E
2 up-regulation of ERα (Figure
4), at least in the time-frame and doses analyzed (see above).
Taken together, these results contrast with the conventional classification of ICI as a "pure estrogen antagonist", which has been reported to block the effects of E
2 and some partial agonists (e.g. tamoxifen) with no detected agonistic activities in several
in vivo and
in vitro models of estrogen action in different mammalian species [reviewed by [
11,
35]]. However, some recent
in vitro studies have also reported agonistic or partial agonistic activities for ICI [
36‐
40], which appear to depend on the species, the tissue, the ER subtype and the promoter, as reported for other SERMs.
The mechanisms in place for the agonistic effects in fish are as yet unknown. Most antiestrogens act through competitive binding to ERs and induction of an inactive conformation of the ligand-dependent AF-2 function of ERs, and their context-specific agonistic activities have been mainly attributed to a tissue- or promoter-specific activation of the ligand-independent AF-1 function or to the induction of a partially active AF-2 conformation [
41]. ICI appears to act at several levels to completely block ER-mediated actions (better studied for the ERα subtype), including the competitive inhibition of agonist binding to ERα, the inhibition of ER dimerization, nuclear translocation and transcription activation through both AF-1 and AF-2, and increased ER protein degradation [reviewed by [
11,
15]]. While estrogens are known to rapidly down-regulate the ERα and ERβ protein levels in several mammalian cell types but up regulate its mRNA levels [
42], ICI has been shown to cause ERα protein degradation without affecting the ERα mRNA levels [
11], thereby leading to an effective reduction of the ER protein levels. Possible explanations for the partial ICI agonism are: 1) lack of ERα protein down-regulation, as observed in cells of the sheep uterus or in human breast cancer cells; 2) species-specific differences in the N- or C-terminal regions of ERs, which could influence ligand discrimination; 3) ER activation via non-classical mechanisms (e.g. non-genomic actions and indirect activation at AP-1 promoters); 4) ICI activation of other ER subtypes (nuclear ERβ or membrane ERs) or ER variant proteins whose relative expression depends on the cell type or species [
36‐
40]. In addition, it was recently reported that ICI was able to promote human ERα interaction with the CBP/p300 but not the p160 family of coactivators in HeLa cells, although this was insufficient to promote transcription from the pS2 (an ERG) promoter [
43].
In fish, unlike in mammals, estrogens have been shown to increase both ERα mRNA (through increased transcription and enhanced stability) and ER protein levels in liver [
31,
44,
45]. The ICI up-regulation of ERα in liver detected in the present study could contribute to the observed agonistic effects, and the potentiation effects observed for the ICI pretreatment may be due to an increased responsiveness of the tissue at the time of E
2 administration through sbERα up-regulation by ICI. Whether the ERα mRNA ICI up-regulation is followed by an increase in ERα protein level in liver, as occurs with E
2, must be investigated in future studies.
The inability of ICI to inhibit the up-regulation of ERα, VgII and ChgL by E
2 in fish liver (Figure
4) could also be interpreted as evidence for an ER-independent mechanism, as suggested for other ICI-insensitive actions [e.g. [
46]]. However, this appears not to be the case, since ICI alone was able to up-regulate these genes (Figure
2) and because the E
2-induced transcriptional activation of both ERα and VgII genes have been demonstrated to involve binding of the ER proteins to specific response-elements in their promoters [
33,
47,
48], while the stabilization of their mRNAs has also been shown to be mediated by E
2/ER complexes [
44]. The dependence on ERα has also been demonstrated in some studies reporting ICI agonism in mammals by using ER-specific siRNA [
39].
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
PISP planned and carried out the sea bream experiments, some of the RT-PCRs, gene expression quantification, statistical analysis and discussion of results, and wrote the manuscript. PBS and JBC planned and carried out the tilapia experiments and calcium measurements. HRT carried out RNA extractions, RT-PCRs and calcium in the sea bream experiments. DMP participated in the discussion of results and wrote the manuscript. AVMC devised the study, participated in the planning of all experiments, statistical analysis and discussion of results, and wrote the manuscript. All authors read and approved the final manuscript.