Computer modelling of estrogenic transcriptional activation can account for different types of dose-response curves of estrogens

A model of estrogen-related gene expression reveals non-linear effects in transcriptional response to tamoxifen

BACKGROUND

Estrogen receptors alpha (ER) are implicated in many types of female cancers, and are the common target for anti-cancer therapy using selective estrogen receptor modulators (SERMs, such as tamoxifen). However, cell-type specific and patient-to-patient variability in response to SERMs (from suppression to stimulation of cancer growth), as well as frequent emergence of drug resistance, represents a serious problem. The molecular processes behind mixed effects of SERMs remain poorly understood, and this strongly motivates application of systems approaches. In this work, we aimed to establish a mathematical model of ER-dependent gene expression to explore potential mechanisms underlying the variable actions of SERMs.

RESULTS

We developed an equilibrium model of ER binding with 17β-estradiol, tamoxifen and DNA, and linked it to a simple ODE model of ER-induced gene expression. The model was parameterised on the broad range of literature available experimental data, and provided a plausible mechanistic explanation for the dual agonism/antagonism action of tamoxifen in the reference cell line used for model calibration. To extend our conclusions to other cell types we ran global sensitivity analysis and explored model behaviour in the wide range of biologically plausible parameter values, including those found in cancer cells. Our findings suggest that transcriptional response to tamoxifen is controlled in a complex non-linear way by several key parameters, including ER expression level, hormone concentration, amount of ER-responsive genes and the capacity of ER-tamoxifen complexes to stimulate transcription (e.g. by recruiting co-regulators of transcription). The model revealed non-monotonic dependence of ER-induced transcriptional response on the expression level of ER, that was confirmed experimentally in four variants of the MCF-7 breast cancer cell line.

CONCLUSIONS

We established a minimal mechanistic model of ER-dependent gene expression, that predicts complex non-linear effects in transcriptional response to tamoxifen in the broad range of biologically plausible parameter values. Our findings suggest that the outcome of a SERM's action is defined by several key components of cellular micro-environment, that may contribute to cell-type-specific effects of SERMs and justify the need for the development of combinatorial biomarkers for more accurate prediction of the efficacy of SERMs in specific cell types.

Design, synthesis, and evaluation of novel 2-substituted 3-hydroxypyridin-4-ones: structure-activity investigation of metalloenzyme inhibition by iron chelators

A range of novel 3-hydroxypyridin-4-ones with different R(2) substitutents has been synthesized for the investigation of the structure-activity relationship between the chemical nature of the ligand and the inhibitory activity of the iron-containing metalloenzyme 5-lipoxygenase. Results indicate that the molecular dimensions, together with the lipophilicity, have a critical impact on the ability of this class of chelator to inhibit 5-lipoxygenase. Hydrophilic ligands with a bulky R(2) substitutent tend to be weak inhibitors; thus 1,6-dimethyl-2-(4'-N-n-propylsuccinamido)methyl-3-hydroxypyridin-4(1H)-one (22b) which has the largest R(2) substitutent only caused 2% inhibition of the enzyme activity after 30 min incubation at 110 microM IBE (iron-binding equivalents), as compared with deferiprone which caused 40% inhibition of the enzyme activity, under the same conditions.

Monitoring of estrogen mimics by a recombinant yeast assay: synergy between natural and synthetic compounds?

Properties of mixtures of compounds exhibiting estrogenic potential have been questioned in the past. Synergistic effects of endocrine disrupters have been proposed, but could never be confirmed. In this study, the transactivational potential of xenoestrogens and phytoestrogens has been evaluated in a yeast test system. Pesticides such as endosulfan, dieldrin, atrazine, and the main metabolites, desethylatrazine and desisopropylatrazine, have been tested and their behavior as mixtures is compared to the behavior of the single compounds. Our results are in contrast to a report (Tran et al., 1996) on the inhibitive effects of xenoestrogens on 17 beta-estradiol-dependent transactivation. Phytoestrogens have been investigated in a similar manner. A synergistic effect could not be confirmed for both, xenoestrogens and phytoestrogens. These compounds are either weak estrogens or completely lack estrogenic potential. Their endocrine disrupting potential in more complex systems must be therefore attributed to other molecular mechanisms such as to metabolic modification or interference with steroidogenesis. This study shows that yeast systems are useful tools for monitoring pure estrogenic properties.

Regulation of human estrogen receptor by phytoestrogens in yeast and human cells

Phytoestrogens are defined as plant substances that are structurally or functionally similar to estrogen. They are present in many foods and their higher consumption in certain populations has been correlated with protection against many diseases including coronary heart disease, breast cancer and endometrial and ovarian cancer. In this report, ten phytoestrogens with diverse chemical structures were studied for their binding to the human estrogen receptor and their transcription activation properties in yeast and mammalian cells. Our results showed that some of these compounds bind with relatively high affinity to the estrogen receptor and activate the receptor in the yeast and mammalian cell system. In addition, none of these compounds showed anti-estrogenic activity. We conclude that the yeast system accurately predicts the estrogenic activity of compounds with diverse chemical structures in mammalian cells. In addition, our data with phytoestrogens that do not show transcription activation properties raise the possibility that these compounds may exert their biological effects through pathways different from the classical estrogen signalling mechanism.

Dissociation of 4-hydroxytamoxifen, but not estradiol or tamoxifen aziridine, from the estrogen receptor as the receptor binds estrogen response element DNA

Estradiol-liganded estrogen receptor (E2-ER) binds EREs with a stoichiometry of one E2-ER dimer per estrogen response element (ERE). In contrast, although 4-hydroxytamoxifen (4-OHT)-liganded ER (4-OHT-ER) binds EREs with high affinity, its saturation ERE binding capacity is consistently half that of E2-ER, giving an apparent stoichiometry of one 4-OHT-ER monomer per ERE. Here we show that one molecule of 4-OHT ligand dissociates from the ER dimer apparently during the process of binding to DNA. Under equilibrium conditions, the type I antiestrogen tamoxifen aziridine (TAz), covalently attached to ER (TAz-ER), binds a single ERE with high affinity (Kd = 0.27 nM), comparable to that of E2-ER and 4-OHT-ER. In contrast to 4-OHT-ER, the ERE binding stoichiometry of TAz-ER was identical to that of E2-ER: one dimeric receptor per ERE. By measuring [3H]ligand that was initially bound to ER, a significant loss of [3H]4-OHT from ER was detected after ERE binding, whereas all [3H]E2 or [3H]TAz remained ER-bound. These results confirm that one molecule of 4-OHT ligand dissociates from the ER dimer as a consequence of ERE binding. Binding of 4-OHT and TAz are likely to induce a conformation in ER dimers that alters their capacity for gene activation. Upon ER binding to DNA, this conformation reveals itself by allowing 4-OHT dissociation, and predictably would allow TAz dissociation were it not bound covalently.

Investigation of the conformational influences on the estrogenic activity of 1,2-bis(2,6-dichloro-4-hydroxyphenyl)ethylenediamines and of their platinum(II) complexes, II: Synthesis and studies on the estrogenic activity of cis- and trans[bis(2,6-dichloro-4-hydroxybenzylamine)]dihaloplatinum(II)-c omplexes

2,6-Dichloro-4-hydroxybenzylamine (1) and its N-methyl (2) and N-ethyl (3) derivatives were synthesized and tested for estrogen receptor affinity as well as for estrogenic activity. In contrast to their related highly active 1,2-bis(2,6-dichloro-4-hydroxyphenyl)ethylenediamines (meso-4 - meso-6) none of the benzylamines showed hormonal activity. The coordination of the benzylamine 1 to platinum did not lead to an estrogenic compound. The reasons for the different activity of [meso- 1,2(bis-2,6-dichloro-4-hydroxyphenyl)ethylenediamine]dichloroplatinum(II ) (meso-4-PtCl2) and cis[bis(2,6-dichloro-4-hydroxybenzylamine)]dichloroplatinum(II) (cis-1-PtCl2), the latter of which can be considered as a ring-opened counterpart of the highly active meso-4-PtCl2, are thoroughly discussed under inclusion of conformational facts. The results of this and the preceding work show, that the pharmacophore meso-1,2-bis(2,6-dichloro-4-hydroxyphenyl)ethylenediamine (meso-4) which is exclusively responsible for the estrogenic activity of meso-4-PtCl2 causes comparable hormonal effects in two different conformations with O-O distances of about 8 A (complex) and of about 12 A (diamine). Therefore, we discuss two binding sites for estrogens in their receptor.