Final Report: Methoxychlor and Environmental "Estrogen" Receptors in ER-Minus MiceEPA Grant Number: R825295
Title: Methoxychlor and Environmental "Estrogen" Receptors in ER-Minus Mice
Investigators: Lubahn, Dennis B. , Welshons, Wade V.
Institution: University of Missouri - Columbia
EPA Project Officer: Hahn, Intaek
Project Period: February 10, 1997 through February 9, 1999
Project Amount: $190,000
RFA: Endocrine Disruptors (1996) RFA Text | Recipients Lists
Research Category: Endocrine Disruptors , Economics and Decision Sciences , Health , Safer Chemicals
Objective:The objective of the research project was to provide evidence for and clone the putative "methoxychlor" receptor.
Effects of Estrogens in ER -KO Mice. Experiments were performed to verify the initial in situ hybridization results of estrogen responsiveness in ER -KO mice with methoxychlor by quantitative RT-PCR (Endocrinology [Ghosh paper discussed briefly below) and 4-hydroxy estradiol (PNAS 1997;94:12786-12791) in the ER -KO mice. We developed a competitive RT-PCR protocol for confirming the quantity of glucose 6-phosphate dehydrogenase and lactoferrin mRNA induction by environmental estrogens in the uterus. These confirmed our earlier results that there is a novel estrogen response mechanism, which is not ER OR ER present in the ER -KO mice.
Technical difficulties because of limited tissue quantities particularly when we begin to characterize unique mRNA responses have been dealt with as we have gained expertise. However, we have continued parallel experiments in our estrogen-treated animals on alternative tissues that may be responsive, in particular liver, which is present in relatively larger quantities and is known to respond to estradiol. We examined lactoferrin, malic enzyme and glucose-6-phosphate dehydrogenase mRNAs responsiveness to estrogens in the liver in ER -KO mice without much success. We had hoped that our encouraging efforts with differential display would provide better liver markers. This has not worked and in future experiments we plan to utilize microarray cDNA techniques. Our problems with limited amounts of uterine tissue to examine methoxychlor responses would then be solved.
Methoxychlor Stimulation of Lactoferrin and Glucose 6 Phosphate Dehydrogenase mRNAs in ER -KO Mice. In our original proposal we only presented lactoferrin mRNA induction by methoxychlor at a single time point and concentration with a single technique, in situ hybridization. The results presented below and in the attached paper provide strong indications that methoxychlor is likely to be acting through its own putative receptor, which is distinct from ER and also the recently discovered ER .
Adult wild type (+/+) or homozygous (-/-) ER -KO sibling mice of the same mixed genetic background (129/C57BL/6J) were ovariectomized and rested for two weeks before treatment. All treatments were given as two injections of 6 hrs apart of oil (vehicle control), Estradiol-17 (10 µg/kg) and Methoxychlor (3.75, 7.5, 15, 30 and 60 mg/kg) and animals were sacrificed 12 hrs after the final injection. In a separate group of mice Methoxychlor (15 mg/kg was injected together with Estradiol-17 (10 µg/kg). All of the test agents were dissolved in olive oil and injected (0.1 ml/mouse) subcutaneously. A time course study (2, 6, 12, 18, 24 hrs for Methoxychlor and 12, 18 and 24 hrs for estradiol) was done with a single injection of Methoxychlor (15 mg/kg) or estradiol (10 µg/kg). Competitive RT-PCR Assays for lactoferrin and glucose-6-phosphate dehydrogenase mRNAs were developed.
The following conclusions were derived from our research:
1. The classical estrogen, estradiol-17 , induced increased lactoferrin and glucose-6-phosphate dehydrogenase RNA levels only in wild type mice not in ER - knockout type. In contrast, the responses of methoxychlor in ER - knockout mice reveal that its action is not being mediated solely through ER
2. The dose and time dependent response to methoxychlor supports its action through receptor binding and saturation. A clear biphasic nature of response was observed both in wild type and ER -knockout mice suggesting more than one response mechanism in the ER -KO mice, i.e. ER AND "methoxychlor" receptors. The differences in the shapes of the curves imply that these "receptors" may be acting differently in their activation of the lactoferrin and glucose-6-phosphate dehydrogenase genes.
3. Glucose-6-phosphate dehydrogenase is an early responsive gene. In the time course study in the wild type and ER -knockout mice an early response to estradiol-17 in wild type mice was observed and this pattern of response was also seen in the methoxychlor response also.
Progress and Comments on Each of Initially Proposed Aims.
4a. Test for methoxychlor binding and responses through Gustaffson's ER . [In brief our working model now is that ER and ER appear to have varying binding and transcriptional activation ability dependent upon whether they are homodimers or heterodimers and dependent upon the cell line, indicating a role for additional co-activators in steroid receptor binding affinity and perhaps specificity.]
4b. Screen for receptors that change concentration by differential display PCR analysis in the steroid receptor families using anchored oligonucleotide primers. [Aim was abandoned because of irreproducibility.]
4c. Screen uterine ER-minus mouse cDNA library with probes from conserved sequences of the steroid receptor super family via low stringency hybridization. [We continue to do this as new sequences become available, particularly when three ER's are found in one species (Mary Beth Hawkins, Joseph W. Thornton, David Crews, James K. Skipper, Alyssa Dotte, and Peter Thomas. Identification of a third distinct estrogen receptor and reclassification of estrogen receptors in teleosts. PNAS 2000;97:10751-10756). [So far without success.]
4d. Test expressed candidate orphan receptor cDNAs for binding to methoxychlor or metabolites. [We have now cloned human ER , ER , ERR1, and ERR3 (ERR2 has resisted our efforts and several other labs to clone until quite recently, partly because the initial human sequence reported was from another species!!) None of these to date appear to be the methoxychlor receptor, at least in the combinations and in the cell lines studied. These studies are still ongoing.]
4e. Expression cloning using [3H]-methoxychlor. [We elected not to use this approach because of methoxychlor's apparent low binding affinity (and subsequent likely high non-specific binding resulting in high false positives) because of the high concentrations of methoxychlor needed in vivo. We have been able to find nM binding affinity to 3H-4-OH-E2 in many tissues that is not competed by 100 fold excess estradiol, but unfortunately this catechol estrogen binding activity is not competed by either methoxychlor or its metabolite HPTE. Thus disproving a major hypothesis in the lab that the 4-OH-E2 and methoxychlor receptors were one and the same.]
Effects of Estrogens in ER -KO Mice. We continue to be intrigued by the potential of ER to be involved in the methoxychlor response and have begun initial attempt to express human ER and ER together to see if together they will bind and respond to methoxychlor better. However, in general, ER appears so far to be very much like ER in its ligand-binding domain specificity.
Our hypothesis was that various ER /ER heterodimerizations might affect ligand specificity and produce a "methoxychlor receptor" with the correct combination. So we began a series of studies combining the classic ER and full length ER . In our initial set of control experiments we found some very surprising results. We recently presented these data (Yuan, et al., 2000) in a poster at the 2000 Toronto Endocrine meeting. We found that ER could greatly affect the ligand binding ability of ER presumably via a heterodimerization mechanism where helix 12 from one receptor is changing the binding pocket of the other receptor. Additionally, that the ratio of homodimer binding affinities of ER and ER (Kds for estradiol) could vary 60-fold between cell lines (C0S and MCF-7) and also when compared to in vitro transcribed and translated ERa and ERb binding ability. These is very disturbing because all published ligand binding (not the better than 20 published ER transcriptional activation studies in response to various ligands, which can not now be assumed to be interchangable!) studies with ER have been done outside of cell lines and have used recombinant in vitro transcribed/translated ER receptor protein. Clearly this suggests to us that we must now combine the various ERs (and alternately spliced forms!) in all combinations but also must be concerned about the cell line being used because of variations in co-activators which are know to also interact with helix 12. (See Westin S, Kurokawa R, Nolte RT, Wisely GB, McInerney EM, Rose DW, Milburn MV, Rosenfeld MG, Glass CK. Interactions controlling the assembly of nuclear-receptor heterodimers and co-activators. Nature 395(6698):199-202, where a model of helix 12 interactions between interacting heterodimers is proposed for the retinoid receptors binding to their cognate ligands.)
This whole story has become even more complicated with the report by a Pfizer group [Endocrinology 1998;139:1082-1092] that there exist alternately spliced forms of ER . These alternately spliced forms contain an 18 amino acid insert located in the ligand-binding domain, resulting in about 10 fold lower binding affinity for 17 -estradiol. Data were presented comparing the binding affinities of these splice variants to other estrogens, including methoxychlor. All these variants also bound DES and estradiol thus ruling out our novel ?methoxychlor" receptor. Additionally several groups have reported ER and ER can form heterodimers. No steroid-binding characterization has yet been reported for any of the 12 possible heterodimers. Combined, these two new reports raise the testable possibility of altered ligand specificities with these heterodimers. The possibility remains that the responses we are observing with methoxychlor in ER - KO mice could be mediated by novel combinations and variants of known ER genes.
In our system, because of the lack of response with 17 -estradiol and competition studies of responsive estrogens with ICI-182, 780 and 17 -estradiol, it is clear to us that the methoxychlor responses we are seeing in ER -KO mice are not being mediated by a ligand-binding domain of either ER or . This supports the existence of a unique non- ER or receptor, "ER- ", if together as heterodimers they do not possess an altered specificity. A ligand specific for ER or is still not available either an agonist or an antagonist would be very useful. If it were available, it might be possible to design an experiment to distinguish them and then confirm an "ER " using in vitro and in vivo non-ER -KO models. We learned that Karobio, a company in Sweden, has such a compound and we are still attempting to obtain it.
In any case, it is likely that ER will mediate some methoxychlor responses. We have arranged with two labs currently making ER -KO mice that when they become available without major "strings" we will breed a double Er - and ER -KO mouse. In this way, we will have a cleaner system to determine if there are additional ERs capable of mediating methoxychlor and other estrogens' actions.
Our efforts to express ER were originally interrupted when it became apparent that all published ER sequences lacked 46 amino acids on their N-terminal ends because of a sequencing error in the original rat ER paper. We have cloned this region in humans and have made a now complete human ER expression vector. We still continue to screen mouse genomic DNA by PCR and mouse mRNA from several tissues by RT-PCR with various fish primers to pull out the mammalian homolog. Many fish ERs have now been reported including the very exciting results of three different ERs in one species (PNAS 2000;97:10751-10756). We continue to design primers and to sequence degenerate PCR products. After 2 years of effort differential display efforts have been abandoned by us because of irreproducibility. We plan to use cDNA microarray analysis to more exhaustively search for genes that are uniquely regulated by either methoxychlor or the catechol 4-OH-estradiol. In interesting aside, and to my surprise and disappointment, another group has not been able to find responses to find uterine mouse genes that are uniquely regulated by 4-OH-E2 alone (Sanjoy K. Das, et al. J Biol Chem 2000;275(37):28834-28842). In every case, estradiol also regulates them, but to my great surprise this novel set of five genes is not inhibited by the ER and ER antagonist, ICI182,780. This suggests there are four ERs. Two ERs because there are uterine genes regulated by E2 through ER and ER , which are inhibited by ICI. Then another set of genes that are upregulated by 4-OH-E2 and methoxychlor but not competed by ICI182,780 (lactoferrin and G6PD) [perhaps this is an ER ], and then finally a fourth set of genes regulated by both E2 and 4-OH-E2 that is not inhibited by ICI182,780 [perhaps this is an ER ?]. Life becomes very complicated if there are four ERs, but the data are compelling.
We continue to screen for novel "methoxychlor" receptors. All in all progress is being made, but while we now believe there is firm evidence for a "methoxychlor" receptor we have not yet cloned it but have learned important new things about how ERs interact with one another to affect both their ligand binding and transcriptional activation capabilities.
Journal Articles on this Report : 1 Displayed | Download in RIS Format
|Other project views:||All 5 publications||2 publications in selected types||All 2 journal articles|
||Ghosh D, Taylor JA, Green JA, Lubahn DB. Methoxychlor stimulates estrogen-responsive messenger ribonucleic acids in mouse uterus through a non-estrogen receptor (Non-ER)alpha and Non-ER beta mechanism. Endocrinology 1999;140(8):3526-3533||