| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Receptor Biology Section (B.J.D., S.C.H., K.S.K.) and Biostatistics Branch, Environmental Diseases and Medicine Program (S.D.P.), National Institute of Environmental Health Sciences, National Institutes of Health, Research Triangle Park, North Carolina 27709
Address all correspondence and requests for reprints to: Dr. Bonnie J. Deroo, MD E4-01, Laboratory of Reproductive and Developmental Toxicology, National Institute of Environmental Health Sciences, 111 Alexander Drive, Research Triangle Park, North Carolina 27709. E-mail: deroo{at}niehs.nih.gov.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
(ER), we demonstrate that these E2-mediated changes require ER, but not ERß. The repression of Txnip by E2 was also demonstrated in vitro in MCF-7 human breast cancer cells. This repression was blocked by treatment with the histone deacetylase inhibitor, trichostatin A, suggesting that repression by E2 may involve regulation of histone acetylation. We conclude that the rapid E2-mediated activation of the Txn pathway is an important step in the response of the mammalian uterus to estrogen. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The Txn pathway regulates the levels of intracellular reactive oxygen species, which are required for normal cellular proliferation, but are toxic to the cell at excessive levels. The core of the Txn system is Txn, which carries out numerous essential roles, such as protein disulfide reduction; protein repair, folding, and stability; regulation of apoptosis and cellular proliferation; regulation of transcription factor activity; and protection against oxidative stress (2). Oxidized Txn is inactive, but is converted to the active reduced form by Txn reductase (Txnrd), and in turn, active Txn is inhibited by Txn-interacting protein (Txnip; Fig. 1
) (3, 4). The Txn pathway functions in several subcellular locations (Fig. 1), including the cytosol (Txn1, Txnrd1, and Txnip), mitochondria (Txn2 and Txnrd2), and endoplasmic reticulum (Txnrd3).
|
The reported expression pattern of the Txn pathway in female reproductive tissues suggests that estrogen may play a role in its regulation. In female mice, Txn is expressed in the uterus, vagina, oviduct, corpus lutea, and ovary (8). Txn is also expressed in normal human endometrium, and in primary cultures of human uterine stromal cells, estrogen treatment increases thioredoxin mRNA and protein expression (9, 10). Although there is evidence to support estrogen regulation of Txn1, it is not known whether Txnrd1 or Txnip levels are similarly regulated or whether estrogen regulates the mitochondrial and/or endoplasmic reticulum Txn system.
In our original microarray study we observed an E2-mediated increase in Txn2 levels as well as an increase in the expression of Txnrd1 and a decrease in Txnip. These data along with those from other studies suggested that the Txn pathway might play a critical role in the response of the uterus to estrogen. Therefore, we characterized which components of the cytosolic and mitochondrial Txn pathways were regulated by E2 in the mouse uterus. We also investigated the mechanistic basis for this regulation. Herein, we report the characterization of the expression of these Txn system components in ovariectomized (ovx) mouse uterus after acute E2 treatment. We found that E2 modulates the levels of cytosolic, endoplasmic reticulum, and mitochondrial Txn pathways, and that this regulation requires functional estrogen receptor (ER).
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
knockout (ERKO) or ßERKO or wild-type (wt) littermates were obtained from Taconic Farms (Germantown, NY), and were ovx and housed for at least 10 d before the studies to allow endogenous ovarian steroids to decrease. Groups of animals (six mice per group) were treated with sesame oil vehicle (Sigma-Aldrich Corp., St. Louis MO) or with 1 µg E2 (Steraloids, Newport, RI), dissolved either in 100 µl sesame oil and injected sc (6, 12, and 24 h groups) or in 100 µl normal saline and injected ip 30 min and 2 h before necropsy. In some cases, animals (four per group) were treated with 45 µg ICI 182,780 (ICI; Zeneca Pharmaceuticals, Cheshire UK) which was dissolved in 50 µl dimethylsulfoxide and injected ip, 30 min before E2 injection. Animals were killed at the indicated times using CO2, and uteri were collected and frozen.
Cell culture
MCF-7 human breast cancer cells were maintained in MEM with 2 mM glutamine, 10% fetal bovine serum, and 10 mM HEPES, pH 8.0, and incubated at 37 C under 5% CO2. BG-1 ovarian cancer cells were maintained under similar conditions in DMEM/F-12 medium with 10% fetal bovine serum.
RT and real-time PCR analysis
Uterine tissue.
Frozen tissue was pooled, pulverized, then homogenized in TRIzol (Invitrogen Life Technologies, Inc., Carlsbad, CA), and RNA was prepared according to the manufacturer’s protocol.
Cells.
Cells were plated (800,000 and 650,000 cells/60-mm dish for MCF-7 and BG-1 cells, respectively) in medium containing 10% fetal bovine serum and allowed to incubate overnight. The next day, the original medium was removed and replaced with medium containing 5% charcoal-stripped serum. The cells were then incubated for at least 48 h before treatment and harvesting. RNA was prepared using TRIzol reagent according to the manufacturer’s protocol.
RT.
To remove genomic DNA, RNA samples from either tissue or cells were incubated with 1 U deoxyribonuclease I (DNase I; Invitrogen Life Technologies, Inc.)/µg RNA for 15 min at room temperature. DNase I was then inactivated by addition of 2.5 mM EDTA (pH 8.0) and heating at 65 C for 10 min. RT of RNA (2 µg) using Superscript II (400 U) was carried out according to the manufacturer’s instructions using oligo(deoxythymidine) primers or random hexamers (Invitrogen Life Technologies, Inc.). The resulting cDNA was treated with 4 U ribonuclease H (Invitrogen Life Technologies, Inc.) for 20 min at 37 C to remove RNA:DNA hybrids. As a negative control, a sample containing RNA but no reverse transcriptase was also included. The resulting samples were diluted with DNase-free water. One hundred nanograms of cDNA were used per well for cell culture studies, and 1–100 ng of cDNA was used for uterine tissue. cDNA levels were detected using real-time PCR with the ABI PRISM 7700 Sequence Detection System (Applied Biosystems, Foster City, CA) and SYBR Green I dye. Primers were created using Applied Biosystems Primer Express software version 2.0 (Table 1). For cDNA amplification, 10 µl cDNA were combined with 40 µl of a mixture containing 1x SYBR Green Master Mix (Applied Biosystems, catalog no. 4304886) in addition to 200 nM reverse and forward primers. Samples were analyzed in duplicate or triplicate, and a sample without reverse transcriptase was included with each plate to detect contamination by genomic DNA. Amplification was carried out as follows: 50 C for 2 min, 95 C for 10 min (denaturation), 95 C for 15 sec, and 60 C for 30 sec (denaturation/amplification). Dissociation curves were also created by adding the following steps to the end of the amplification reaction: 95 C for 15 sec (denaturation), 60 C for 20 sec, then gradually increasing to 95 C over 20 min, finally holding at 95 C for 15 sec. Fold expression or repression was determined by quantitation of cDNA from target (treated) samples relative to a calibrator sample (vehicle). For all samples, the gene for ribosomal protein 7 (Rpl7) or 18S was used as the endogenous control for normalization of initial RNA levels. Expression ratios were calculated based on the mathematical model described by Pfaffl (11): ratio = (Etarget)
Ct(target)/(ERpl7)Ct(Rpl7), where E is the efficiency of the primer set, calculated from the slope of a standard curve of log nanograms of cDNA vs. the threshold cycle (Ct) value for a sample that contains the target according to the formula E = 10–(1/slope) and Ct = Ct(vehicle) – Ct(treated sample).
|
-32P]CTP using T3 RNA polymerase and the MAXIscript kit (Ambion, Inc.). A riboprobe for Rpl7 was similarly prepared using T7 polymerase and a construct containing the cDNA for Rpl7. These probes were hybridized to the filter using Ultrahyb buffer (Ambion, Inc.) according to the manufacturer’s protocol.
Western blots
Protein extracts were prepared by homogenization of uteri with a Polytron homogenizer (Brinkmann Instruments, Inc., Westbury, NY) in buffer containing 50 mM Tris-HCl (pH 7.5), 150 mM NaCl, 1% Triton X-100, 0.25% sodium deoxycholate supplemented with protease inhibitors (20 µg/ml aprotinin, 20 µg/ml leupeptin, and 4 µg/ml phenylmethylsulfonylfluoride), and phosphatase inhibitor mixture (Sigma-Aldrich Corp.). Thirty micrograms of protein were loaded into each lane of 10% Tris-glycine Novex gels (Invitrogen Life Technologies, Inc.), and proteins were separated and transferred to a polyvinylidene difluoride membrane according to the manufacturer’s protocol. Antimouse Txnip polyclonal antibody was a gift from Dr. Richard Lee (Harvard University, Cambridge, MA) and used at a 1:500 dilution. After transfer, the membrane was blocked in 5% milk in Tris-buffered saline-Tween 20 (TBST), then incubated with primary antibody diluted in 5% milk/TBST overnight at room temperature. The membrane was washed three times for 15 min each time in TBST, then incubated for 1 h with horseradish peroxidase-conjugated secondary antibody (antirabbit horseradish peroxidase, Amersham Biosciences, Piscataway NJ) diluted 1:10,000 in 5% milk/TBST. After additional washing, bands were visualized with ECL-Plus reagent (Amersham Biosciences) and Hyperfilm (Amersham Biosciences).
Statistical analysis
In Fig. 2, data were tested for trends using the Williams test statistic (12). Because the sample sizes were small, the P values were determined using a nonparametric bootstrap procedure (13). Individual pools were resampled so that the dependence between the repeated measurements was preserved. A log transformation was first performed on the data so that variability between various time points, within each gene, was approximately homogeneous. In Fig. 4, the bootstrap P values were obtained using 100,000 bootstrap samples. To compare the differences between the genotypes and the differences between the ovx and non-ovx animals a mixed effects ANOVA was performed. The individual pools were taken as random effects, because multiple measurements were made on each pool, which may result in correlations between the repeated measurements. The two fixed effects in the model were genotype of an animal and whether an animal was ovx. The interaction term between genotype and ovx was nonsignificant (P = 0.2620). All analyses in Fig. 4 were performed using the procedure PROC MIXED in the statistical software package SAS (release 8.2, SAS Institute, Inc., Cary, NC). In Fig. 5 for MCF-7 and Txnip, the four time points were compared with the vehicle control using the standard ANOVA methodology. All analyses for Fig. 5 were performed using PROC GLM (with LSMEANS) in SAS. A statistically significant interaction was found between experiments and time groups (P = 0.0011). Hence, comparisons were made between the time points and the time zero control in each experiment separately. In each experiment a one-sided alternative hypothesis was tested; namely, that the vehicle control mean was greater than the mean at a given time point. For MCF-7 and c-Myc, analysis was conducted as described for Txnip. However, a statistically significant interaction was not found between experiments and time groups (P = 0.1345). Hence, a main effects analysis was performed for the four time groups relative to the control group. For BG-1 and Txnip, after ANOVA a statistically significant interaction was not found between experiments and time points (P = 0.8441). Hence, a main effects analysis was performed for the four time groups relative to the control group. For BG-1 and c-Myc, after ANOVA a statistically significant interaction was found between experiments and time points (P = 0.0003). Hence, comparisons were made between the time points and the control by each experiment separately. In each experiment a one-sided alternative hypothesis was tested; namely, that the vehicle control mean was less than the mean at a given time point. In Fig. 6 for A (ICI), B [trichostatin A (TSA)], and C [actinomycin D (ActD)], the data were log-transformed so that the resulting data were approximately homoscedastic between treatment groups and were approximately normally distributed. In each case, the data were analyzed using the standard ANOVA methodology and the PROC GLM (with LSMEANS) in SAS software. In Fig. 6, A–C, a statistically significant interaction was found between experiments and treatment groups (P < 0.01). Hence, separate comparisons were made between the treatment groups and the vehicle control in each of the three independent replicate experiments. For ICI, a one-sided alternative hypothesis was used to determine whether the vehicle control mean was significantly higher than E2, the mean for ICI, and the mean for ICI plus E2. Also, a one-sided hypothesis was used to determine whether the mean E2 was significantly lower than the mean for ICI and the mean for ICI plus E2. Using residual plots, a potential outlier was found in experiment 3 vehicle control data (0.832). For TSA, a one-sided alternative hypothesis was used to determine whether the vehicle control mean was higher than the mean E2, but lower than mean TSA and mean TSA plus E2 values. A one-sided hypothesis was used to determine whether the mean TSA plus E2 value exceeded the mean E2 value. For ActD, a one-sided alternative hypothesis was used to determine whether the vehicle control mean was significantly higher than the mean E2, the mean ActD, and the mean ActD plus E2 values. A one-sided hypothesis was used to test whether the mean ActD plus E2 value was lower than the mean ActD value. A potential outlier was found in experiment 1 (ActD plus E2) data (0.410), and an outlier was found in experiment 3 (ActD) data (0.440).
|
|
|
|
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
). In this present study we followed the regulation of these nuclear-encoded genes by real-time PCR over a time course following E2 treatments at 0.5, 2, 6, 12, and 24 h (Fig. 2). We also extended the study to investigate other genes involved in the thioredoxin family, namely (cytosolic) Txn1, (mitochondrial) Txnrd2, (endoplasmic reticulum) Txnrd3 (Fig. 1), and a related member of the pathway, (mitochondrial) peroxiredoxin III (Prdx3). Prdx3 is not a Txn family member, but it is involved in the mitochondrial thioredoxin pathway.
We observed E2 regulation of Txn pathway genes whose protein products are localized to all three subcellular locations. RNA was prepared from whole tissue and analyzed by RT, followed by real-time PCR. Consistent with our original microarray observations, we observed a time-dependent decrease in Txnip mRNA levels (Fig. 2), with a concomitant increase in Txnrd1. Txnip mRNA levels dropped to approximately 20% of levels in untreated mice as early as 2 h, and this level was maintained up to and including 24 h. Txnrd1 levels gradually increased over the time course, with a maximal induction of approximately 2.5-fold at 24 h (Fig. 2). To expand our analysis to other members of the cytosolic pathway and to investigate genes involved in the pathway in the endoplasmic reticulum, we then determined the mRNA levels of Txn1 and the Txnrd3 genes and found that these levels also increased after E2 treatment, with approximately 4-fold induction at the 24-h point for Txn1 and 2.5-fold induction for Txnrd3 (Fig. 2). Analysis of genes encoding mitochondrial Txn proteins indicated that E2 treatment also led to a 4.5-fold increase in the expression of Txn2 (Fig. 2), but not its associated reductase, Txnrd2 (data not shown). Lastly, we investigated the expression of a third nuclear-encoded, mitochondrial-specific, Txn-related gene, Prdx3, which breaks down hydrogen peroxide and uses mitochondrial Txn2 as the electron donor for its peroxidase activity (14). Interestingly, the expression of Prdx3 was also regulated by E2 and was elevated 5-fold at 24 h compared with the vehicle-treated control value, indicating that E2 increases the mRNA levels of several antioxidant enzymes that are located in the mitochondria. For each of these E2-regulated genes, ICI inhibited the repression or activation induced by E2, indicating that the ER is required for the regulation we observed (Fig. 2).
Among the genes assayed, Txnip was unique in that E2 caused its mRNA level to decrease. In general, the mechanisms of estrogen-mediated repression have not been well characterized. This fact led us to investigate the regulation of Txnip expression by E2. First, we determined whether the loss of Txnip mRNA also resulted in a rapid reduction in Txnip protein. The loss of Txnip protein correlated with the loss of Txnip mRNA, and as expected, this loss was prevented by treatment with ICI (Fig. 3). The ability of ICI to inhibit the repression of Txnip by E2 suggested that the ER was involved in the mechanism of Txnip repression. Therefore, to determine whether ER or ERß was involved in this mechanism, we investigated the levels of Txnip mRNA in mice lacking ER (ERKO mice). As predicted, the E2-mediated reduction in Txnip mRNA levels observed in wt mice did not occur in ERKO mice after 2 and 24 h of E2 treatment (Fig. 4A, compare lanes 2 and 3 with lanes 5 and 6). Txnip protein levels also were not reduced by E2 in the ERKO mice (Fig. 4B, compare lanes 2 and 4). In mice lacking ERß (ßERKO mice), the E2-mediated reduction in Txnip mRNA levels was comparable to the reduction observed in wt mice (data not shown). These results indicate that in uterine tissue ER, but not ERß, is required for Txnip repression by E2.
|
is required for this repression, then Txnip expression should be higher in untreated ERKO mice than in untreated wt or ßERKO mice. In addition, given that E2 treatment reduces Txnip mRNA levels, ovariectomy should lead to increased Txnip mRNA levels in wt mice. Therefore, we compared the basal levels of Txnip mRNA in intact and ovx wt, ERKO, and ßERKO animals (Fig. 4C). Uteri from intact wt mice were collected at diestrus, the stage of the estrous cycle at which estrogen levels are lowest, and Txnip levels should be highest. Ovariectomy of wt mice resulted in a 22-fold increase in Txnip mRNA levels and a significant increase in both ERKO and ßERKO uterine Txnip mRNA levels. As predicted, ERKO mice expressed significantly higher endogenous levels of Txnip compared with ovx wt or ßERKO animals (Fig. 4C). These results confirm that E2 rapidly and robustly reduces Txnip mRNA levels in the mouse uterus via a mechanism requiring ER.
To elucidate the mechanism of ER-mediated Txnip repression, we determined whether E2 repression of Txnip might also occur in vitro. Mammary carcinoma MCF-7 cells, which express ER, were treated with an E2 time course, and Txnip mRNA levels were determined by real-time PCR. E2 treatment again resulted in a rapid reduction in Txnip mRNA levels of approximately 50% as early as 1 h after treatment (Fig. 5). In contrast, c-myc mRNA increased with E2 treatment (Fig. 5), as previously reported (15). To determine whether this repression would be observed in another ER-containing cell line, we similarly treated BG-1 ovarian carcinoma cells and found that E2 also significantly reduced Txnip mRNA levels in this cell line, although the repression was weaker, with maximal repression approximately 20% compared with untreated levels (Fig. 5). As a positive control, we again analyzed c-myc levels in BG-1 cells, which increased after E2 treatment with a profile similar to that observed in MCF-7 cells. To determine whether the E2-mediated reduction in Txnip expression in MCF-7 cells also required the ER, we treated MCF-7 cells with E2 in combination with ICI (Fig. 6A). As observed in the uteri of mice, ICI blocked the repression of Txnip by E2, indicating that in vitro the ER is required in this mechanism.
Transcriptional repression by steroid receptors has been associated with targeting of corepressors to the promoter region of genes and a reduction in the level of histone acetylation in the chromatin of the promoter region (16). The histone deacetylase inhibitor, TSA, has been used extensively to study the effect of histone acetylation on promoter activity. We wanted to determine whether increasing the levels of histone acetylation on the promoter by TSA treatment might alter the repression of Txnip by E2. We treated MCF-7 cells with E2 in the absence or presence of TSA (Fig. 6B). TSA treatment alone enhanced Txnip mRNA levels and blocked the E2-mediated decrease in these levels, suggesting that the reduction in Txnip expression by E2 may be regulated by the levels of histone acetylation on the Txnip promoter. Finally, it is possible that E2 treatment reduces the stability of Txnip mRNA rather than reduces the rate of transcription. However, E2 treatment did not dramatically alter Txnip mRNA stability in MCF-7 cells in the presence of the transcription inhibitor, ActD (Fig. 6C). Although E2 treatment resulted in a statistically significant reduction in Txnip mRNA levels compared with ActD alone (P < 0.05 in two of three experiments), this difference was small (7–19%). These results indicate that the reduction of Txnip mRNA levels in response to E2 treatment is due to reduced Txnip transcription rather than a reduction in mRNA stability.
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
There is evidence that active Txn plays a critical role in mammalian reproduction. Txn is present in early pregnancy factor, components that appear in maternal serum within hours after fertilization (5). In mice, Txn1 expression is localized to the simple columnar epithelium and the uterine glands of the uterus (8). In humans, Txn1 is highly expressed in cytotrophoblast cells, which invade the endometrium during formation of the placenta, and in stromal cells of the decidualized endometrium (17, 18). Both Txn1 and Txn2 knockout mice die before or shortly after implantation (6, 7). Hyplip1 mice, which express 8-fold lower levels of Txnip mRNA (19) than wt controls, demonstrate reduced reproductive fitness compared with wt mice and produce infrequent litters (Lusis, A. J., unpublished observations), suggesting that Txnip plays a critical role in fertility.
Consistent with our data, there is evidence that both Txn and Txnrd are regulated by E2 in various tissues. E2 regulation of Txn1 levels has been demonstrated in the uterus, brain, bone, and cardiovascular system, whereas evidence for E2 regulation of Txnrd1 has been limited to rodent bone and cultured bovine endothelial cells (8, 9, 10, 20, 21, 22, 23, 24, 25). Mouse uterine Txn1 levels fluctuate during the estrous cycle, but peak at estrus (8), and treatment of ovx rats with E2 increases uterine Txn1 mRNA levels (20, 21). In the human endometrium, Txn1 is present in the glands of the late proliferative phase and peaks during the early secretory phase (22). E2 also increases the levels of Txn1 in cultured human endometrial stromal cells and human neuroblastoma cells (10, 23). Whereas E2 regulation of Txn1 in human and rodent uterus has been reported, ours is the first study to show similar uterine regulation of its cytosolic partner, Txnrd1, and regulation of the mitochondrial and endoplasmic reticulum Txn systems. Therefore, we have demonstrated that E2 regulates the uterine Txn system in several different cellular compartments, suggesting that Txn may play multiple roles in E2-mediated stimulation of the mouse uterus.
Our study suggests that increased levels of Txn are important for the uterine response to estrogen. In the rodent this response includes a series of well characterized physiological changes, including water imbibition, infiltration of immune system cells, and increased transcription and DNA synthesis, resulting in increased cellular proliferation and an increase in uterine weight (26, 27). Txn may be involved in several of these processes. With respect to cellular proliferation, Txn is known to act as both a mitogen and a growth factor in many cellular models (28, 29, 30, 31, 32, 33, 34). Txn increases cytokine production in vitro and synergizes with other cytokines to enhance proliferation (35, 36). Several types of immune cells release Txn (29, 30). Thus, in addition to enhancing proliferation in the uterus, Txn may act as a chemoattractant for immune cells, because it has demonstrated chemotactic activity both in vitro and in vivo (37). Txn may also have a role in the estrogenic stimulation of uterine gene transcription, because Txn is known to regulate the activity and DNA binding ability of a number of transcription factors, including ER, nuclear factor-
B, activating protein-1, cAMP response element-binding protein, Myb, p53, and the glucocorticoid receptor (38). Because Txn is present in many tissues, it would also be interesting to determine whether the ER regulates this pathway in other estrogen-responsive tissues, such as the cardiovascular system, or in other estrogen-responsive cell lines.
Although the mechanisms by which E2 increases the transcription of genes have been well characterized, relatively little is known about how estrogen down-regulates gene expression. However, recent microarray analyses from our laboratory and others indicate that E2 reduces the expression of a significant number of genes. Of the genes regulated by E2 treatment of MCF-7 cells, 70% were down-regulated (39). Of the E2-regulated uterine genes identified by microarray analysis, 12% were repressed after 2 h, and by 24 h the repressed proportion had increased to 28% (1). In our current study uterine Txnip mRNA levels were rapidly reduced to 20% of vehicle-treated levels after E2 treatment, and this reduced level was maintained up to 24 h after treatment. This mechanism required ER, because an ER antagonist, ICI, blocked this reduction in vivo and in vitro. In addition, although repression of Txnip occurred in wt and ßERKO mice, repression was lost in ERKO mice. Therefore, our data indicate that ER, but not ERß, is required for Txnip repression. Furthermore, TXNIP levels were decreased by E2 treatment of cells cultured in vitro. This reduction was blocked by both ICI treatment and treatment with the histone deacetylase inhibitor, TSA, suggesting roles for ER and chromatin structure in the regulation of this gene by E2. Treatment with IGF-I and epidermal growth factor also dramatically reduces Txnip mRNA levels in the mouse uterus (Hewitt, S. C., J. Collins, S. Grissom, B. Deroo, and K. S. Korach, manuscript in preparation). Other factors are reported to regulate Txnip expression, including an increase after 1,25-dihydroxyvitamin D3 treatment in human promyelocytic leukemia cells and a rapid reduction in aortic smooth muscle cells after treatment with growth-promoting compounds such as platelet-derived growth factor, H2O2, thrombin, lysophosphatidic acid, growth factor, TNF, and IL-1ß (40). Our studies investigating how E2 treatment reduces Txnip gene expression will shed light on the common, but poorly understood, mechanisms by which E2 represses, rather than activates, gene expression.
We also found that ERKO mice express higher uterine Txnip mRNA levels than wt controls, again suggesting that ER is involved in Txnip repression. A significant increase in Txnip mRNA levels was observed due to ovariectomy in all three genotypes. In the wt and ßERKO animals, this increase was expected because the animals express ER. However, although the approximately 1.5-fold increase in ERKO animals was much smaller than the 22-fold increase observed in wt animals, the ERKO response was unexpected. There are several explanations for this increase. First, ovariectomy results in a loss of other steroids and factors in addition to E2, such as progesterone and testosterone, inhibin, and activin. Thus, these other factors may also repress Txnip, such that their loss due to ovariectomy would result in increased levels of Txnip. For example, in ovx rats, testosterone administration increases the levels of uterine Txn, as does E2 treatment (21); thus, it is possible that testosterone may also repress the Txn inhibitor, Txnip, in our system. Additionally, as previously reported by our laboratory, the ERKO contains a mutant functional ER splice variant that retains residual E2 binding and functional activity in the uteri of intact ERKO mice approximately 3–9% that of wt activity (41). Given that ERKO females exhibit chronically elevated E2 and testosterone levels compared with wt females, it is possible that Txnip is repressed in the intact ERKO via the ER splice variant and/or the androgen receptor. This repression would then be alleviated by removal of E2 and testosterone by ovariectomy and would be expected to be a much weaker increase in Txnip due to ovariectomy than in the wt animals, which we observed.
It remains to be determined by what mechanism the ER regulates Txn1 and Txnip expression, given that neither the Txn1 nor the Txnip promoter contains a canonical estrogen response element. However, the human Txn promoter and mouse Txnip promoters contain half-estrogen response elements (AGGTCA) and several SP1 sites, which have been shown to synergize with the ER on various promoters (8, 21, 42). In our study, although E2 slightly reduced Txnip mRNA stability compared with actinomycin D alone, this reduction was minor and is probably of limited biological significance compared with the approximately 50% reduction in Txnip mRNA levels due to E2 treatment. In addition, we analyzed the effect of E2 treatment on a reporter construct containing 1.7 kb of the Txnip proximal promoter attached to luciferase (not presented), but were unable to observe any reduction in luciferase in response to E2 in several cell lines, suggesting that the elements required for this repression are either not present in the cloned promoter or are masked by positive elements. More detailed promoter studies will be required to examine these mechanisms.
Our studies have led us to hypothesize that the rapid and profound E2-mediated reduction of the Txn inhibitor, Txnip, may be an important step in the proliferation of uterine cells in response to E2. Overall, our study demonstrates that E2 regulates the key proteins of the mammalian Txn system in the rodent uterus by simultaneously increasing the expression of Txn and Txnrd and reducing expression of Txnip by mechanisms requiring ER. These results indicate that activation of the Txn pathway is an important process in the response to estrogen in the mammalian uterus.
|
Footnotes |
|---|
ERKO, estrogen receptor knockout; ICI, ICI 182,780; ovx, ovariectomized; Prdx3, peroxiredoxin III; TBST, Tris-buffered saline-Tween 20; TSA, trichostatin A; Txn, thioredoxin; Txnip, thioredoxin-interacting protein; Txnrd, thioredoxin reductase; wt, wild-type. Received April 13, 2004.
Accepted for publication August 24, 2004.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  M. Almeida, L. Han, M. Martin-Millan, L. I. Plotkin, S. A. Stewart, P. K. Roberson, S. Kousteni, C. A. O'Brien, T. Bellido, A. M. Parfitt, et al. Skeletal Involution by Age-associated Oxidative Stress and Its Acceleration by Loss of Sex Steroids J. Biol. Chem., September 14, 2007; 282(37): 27285 - 27297. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. Zidek, J. Hellmann, P.-J. Kramer, and P. G. Hewitt Acute Hepatotoxicity: A Predictive Model Based on Focused Illumina Microarrays Toxicol. Sci., September 1, 2007; 99(1): 289 - 302. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Gong, P. Guo, Y. Zhai, J. Zhou, H. Uppal, M. J. Jarzynka, W.-C. Song, S.-Y. Cheng, and W. Xie Estrogen Deprivation and Inhibition of Breast Cancer Growth in Vivo through Activation of the Orphan Nuclear Receptor Liver X Receptor Mol. Endocrinol., August 1, 2007; 21(8): 1781 - 1790. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Satoh, C. M. Matter, H. Ogita, K. Takeshita, C.-Y. Wang, G. W. Dorn II, and J. K. Liao Inhibition of Apoptosis-Regulated Signaling Kinase-1 and Prevention of Congestive Heart Failure by Estrogen Circulation, June 26, 2007; 115(25): 3197 - 3204. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. O'Lone, K. Knorr, I. Z. Jaffe, M. E. Schaffer, P. G. V. Martini, R. H. Karas, J. Bienkowska, M. E. Mendelsohn, and U. Hansen Estrogen Receptors {alpha} and {beta} Mediate Distinct Pathways of Vascular Gene Expression, Including Genes Involved in Mitochondrial Electron Transport and Generation of Reactive Oxygen Species Mol. Endocrinol., June 1, 2007; 21(6): 1281 - 1296. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 I. M. Kooter, J. L. A. Pennings, P. H. B. Fokkens, D. L. A. C. Leseman, A. J. F. Boere, M. E. Gerlofs-Nijland, F. R. Cassee, J. A. C. Schalk, T. J. H. Orzechowski, M. M. Schaap, et al. Ozone induces clear cellular and molecular responses in the mouse lung independently of the transcription-coupled repair status J Appl Physiol, March 1, 2007; 102(3): 1185 - 1192. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. V. Henley, N. Lipson, K. S. Korach, and C. A. Bloch Prepubertal Gynecomastia Linked to Lavender and Tea Tree Oils N. Engl. J. Med., February 1, 2007; 356(5): 479 - 485. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S Morimoto, C A Mendoza-Rodriguez, M Hiriart, M E Larrieta, P Vital, and M A Cerbon Protective effect of testosterone on early apoptotic damage induced by streptozotocin in rat pancreas J. Endocrinol., November 1, 2005; 187(2): 217 - 224. [Abstract] [Full Text] [PDF]
|
|
|
|
) or presence (
) of the estrogen receptor antagonist, ICI. mRNA was isolated and cDNA was prepared by RT, then analyzed by real-time PCR for members of the cytosolic, endoplasmic reticulum, and mitochondrial Txn pathways. Results were normalized to the Rpl7 transcript and expressed relative to the vehicle control, which was assigned a value of 1.0. Each data point is the average ± SE of values obtained using three independent RNA samples (E2 treatment) or two independent samples (ICI treatment), and each sample was analyzed in triplicate. A significant decreasing trend due to E2 treatment was detected with a significance of P < 0.0001 for Txnip, and a significant increasing trend due to E2 was detected with a value of P < 0.0001 for all other genes, by Williams test.
