DDT and its metabolites alter gene expression in human uterine cell lines through estrogen receptor-independent mechanisms.

Endocrine-disrupting organochlorines, such as the pesticide dichlorodiphenyltrichloroethane (DDT), bind to and activate estrogen receptors (ERs), thereby eliciting estrogen-like effects. Although ERs function predominantly through activation of transcription via estrogen-responsive elements, both ERs, alpha and ss, can interact with various transcription factors such as activator protein-1 (AP-1). Additionally, estrogens may regulate early signaling events, suggesting that the biological effects of environmental estrogens may not be mediated through classic ER (alpha and ss) activity alone. We hypothesized that known environmental estrogens, such as DDT and its metabolites, activate AP-1-mediated gene transactivation through both ER-dependent and ER-independent means. Using two Ishikawa human endometrial adenocarcinoma cell line variants that we confirmed to be estrogen responsive [Ishikawa(+)] and estrogen unresponsive [Ishikawa(-)], we generated stably transfected AP-1 luciferase cell lines to identify the role of an estrogen-responsive mechanism in AP-1-mediated gene expression by various stimuli. Our results demonstrate that DDT and dichlorodiphenyldichloroethane (DDD) were the most potent activators of AP-1 activity; 2,2-bis(p-chlorophenyl) acetic acid failed to activate. Although stimulated in both Ishikawa(+) and Ishikawa(-) cells by DDT and its congeners, AP-1 activation was more pronounced in the estrogen-unresponsive Ishikawa(-) cells. In addition, DDT, DDD, and dichlorodiphenyldichloroethylene (DDE) could also stimulate AP-1 activity in the estrogen-unresponsive human embryonic kidney 293 cells using a different promoter context. Thus, our data demonstrate that DDT and its metabolites activate the AP-1 transcription factor independent of ER (alpha or ss) status.

Extracellular signals such as growth factors, hormones, and environmental compounds elicit an array of effects on a cell that include proliferation, differentiation, stress, and death (1)(2)(3). In many cases, external stimuli signal a cell, resulting in the activation or inhibition of transcription factors such as activator protein-1 (AP-1) and nuclear factor κB (NFκB), which form complexes and regulate gene expression (4,5). The resultant change in gene expression leads to an overall effect on the cell, dictating its final interpretation of the external cue. Determining what signals target transcription factors and how they do so is key to understanding the specificity of gene expression and resultant biological responses.
AP-1 is a generic term used to describe transcription factors that bind specifically to a DNA enhancer sequence [TGA(G/C)TCA] called the 12-O-tetradecanylphorbol-13acetate responsive element (TRE), which is also referred to as the AP-1 site. In addition, AP-1 components can bind and potentiate transcription from AP-1-related DNA elements (e.g., cAMP response element) (6). Members of the Jun and Fos families of proteins dimerize to preferentially bind AP-1 sites with high affinity; hence, each dimer combination makes up an AP-1 protein. Upon stimulation, Jun and Fos proteins recruit p300/CBP coactivators that recruit other coactivators such as the p160s, which directly bind nuclear hormone receptors like the estrogen receptor (ER) (7). p300/CBP helps stimulate gene expression by two major mechanisms: making direct interactions with core RNA polymerase machinery and using their intrinsic histone acetylase activity to unravel chromatin structure. Both mechanisms work toward ultimately recruiting the RNA polymerase complex to the sites of AP-1-regulated genes (8). Conversely, AP-1-mediated gene expression is ablated when Jun or Fos family members associate with co-repressors such as silencing mediator of retinoic acid and thyroid hormone receptor, which is known to interact with Sin3 and recruit histone deacetylases, condensing chromatin and terminating transcription (9). AP-1 is a ubiquitous protein that can be induced by multiple stimuli, leading to diverse cellular effects. For example, proliferation, differentiation, cellular stress, and death have all been associated with elevated AP-1 activity (5,(10)(11)(12)(13)(14)(15). Hence, AP-1 appears to play a diverse role in the regulation of the cell cycle. Cell type, promoter context, associative proteins, and stimuli are factors that cumulatively will determine the effect AP-1 has on a cell.
Although estrogen-like molecules are thought to function predominantly through ER-mediated activation of transcription via estrogen-responsive elements (EREs), both ERs, α and β, can interact with various cellcycle transcription factors such as AP-1. The nature of the ER-AP-1 interaction depends on the ligand, ER subtype, and AP-1 components (37)(38)(39). Additionally, the ER can regulate AP-1 at the level of its gene expression by promoting transcription of the c-fos gene in the presence of the major endogenous estrogen, 17β-estradiol (E 2 ) (40)(41)(42)(43)(44). The cross-talk between the pathways modulating the ER and those modulating the AP-1 transcription factor introduces another level of complexity in determining the effects of hormonally active compounds.
Several environmental compounds that have estrogenic activity affect various signal transduction pathways. For example, at low concentrations the isoflavone genistein can bind to and activate ERs, but at high concentrations it binds to and inhibits receptor tyrosine kinase activity (2). Organochlorine compounds, such as DDT, the polychlorinated biphenyls 2´,4´,6´-trichloro-and 2´,3´,4´,5´-tetrachloro-4-biphenylol, β-hexachlorohexane, and heptachlor, activate multiple kinase pathways (45)(46)(47)(48)(49)(50)(51)(52). The weakly estrogenic polychlorinated biphenyl Arochlor 1254 has been reported to induce AP-1 activity (53,54). In neurons, the endocrine disruptor β-hexachlorohexane up-regulates expression of c-fos (55). Methoxychlor (a member of the DDT family), catechol estrogens, and the organochlorine kepone can stimulate expression of estrogen-responsive genes in ERαknockout mice treated with ICI 182,780 (56,57). Additionally, catechol estrogens promote the formation of breast tumors in ERαknockout mice that do not express ERβ in the mammary tissue, indicating the involvement of a non-ER (α or β) mechanism (58). Thus, compounds that can induce estrogenic responses may alter kinase signal transduction pathways via ER-dependent and/or ER-independent mechanisms, working through multiple mitogenic pathways. Hence, endocrine disruptors may cause adverse health effects due primarily to their pleiotropic nature rather than a single endocrine-altering event.
Given the ability of the ER to interact with AP-1 signaling and estrogenic compounds to elicit early signaling events, we hypothesized that known estrogenic organochlorines, such as DDT and its metabolites, could induce AP-1-mediated gene expression. To do this, we wanted to develop a cell culture system that would enable us to quickly and efficiently determine what role, if any, the ER has in AP-1-mediated signaling. This would allow us to explore whether the estrogenicity of certain compounds was responsible for their ability to regulate AP-1 activity or ER-independent mechanisms were additionally involved.
Here, we show that the persistent pesticide DDT, along with its metabolites, activates AP-1-mediated gene expression at environmentally relevant concentrations (i.e., 10-50 µM) (59)(60)(61). Using a novel, stablereporter gene-cell system developed in our laboratory that includes the use of estrogenresponsive and estrogen-unresponsive cell variants, we show the activation is independent of classic ER-mediated mechanisms. This effect is also seen in other cell lines under different AP-1 promoter contexts. Collectively, this demonstrates another level of complexity regarding the nuclear signaling of environmental chemicals. Although hormonal activity may explain some of the adverse health effects of xenobiotics, the total effect of these exogenous chemicals will likely be the end result of multiple signaling pathways that are stimulated due to the pleiotropic nature of many environmental compounds.

RT-PCR analysis of ERα and ERβ expression.
Reverse-transcriptase polymerase chain reaction (RT-PCR) analysis was performed as previously described (62). Briefly, RNA was extracted using an Ultraspec RNA isolation kit (Biotecx Lab, Houston, TX). Frozen cell pellets (1 × 10 6 cells) were lysed in 1 mL Ultraspec RNA solution. Then 0.2 mL chloroform was added to the cell lysate and shaken. After centrifugation, the colorless upper aqueous phase was transferred to a new tube. An equal volume of isopropanol was added, and the RNA was precipitated by centrifugation. The RNA was washed with 75% ethanol and recovered in water treated with diethylene pyrocarbonate. We checked RNA purity by gel electrophoresis and optical density ratio (data not shown). RT-PCR was done using a Perkin-Elmer AmpliTaq Gold with GeneAmp RNA PCR kit following the manufacturer's instructions. Primers were synthesized from Invitrogen (Carlsbad, CA). Primers used to amplify the ERα were: 5´-TGC CAA GGA GAC TCG CTA-3( nucleotides 894-912) and 5´-TCA ACA TTC TCC CTC CTC-3´ (nucleotides 1139-1157), giving an amplified product of 263 bp. For ERβ, primer sequences were: 5´-TTC CCA GCA ATG TCA CTA ACT-3´(nucleotides 33-53) and 5´-TCT CTG TCT CCG CAC AAG G-3´ (nucleotides 539-558), giving an amplified product of 525 bp. Fragments of glyceraldehyde phosphate-3-dehydrogenase (GADPH) were amplified in parallel to serve as an internal control. The primers for GADPH were: 5´-TC ACC ATC TTC CAG GAG C-3´ and 5´-CAA GAA GGT GGT GAA GCA G-3´, giving a PCR product of 571 bp. PCR reactions were run for 30 cycles, which were in the linear range of the reaction (data not shown). PCR products were verified by subcloning and sequencing (data not shown).
Cell culture, transient transfection, and reporter gene assay. Ishikawa endometrial adenocarcinoma cells were grown in Iscove's Modified Dulbecco's Medium: Ham's F12 (1:1) supplemented with 10% fetal bovine serum (FBS), Basal Medium Eagle (BME) amino acids, Minimum Essential Medium Eagle (MEM) nonessential amino acids, sodium pyruvate, penicillin-streptomycin, and 1 × 10 -10 M insulin. Human embryonic kidney (HEK) 293 cells were grown as previously reported in DMEM supplemented with 10% FBS, BME amino acids, MEM nonessential amino acids, sodium pyruvate, penicillin-streptomycin, and 1 × 10 -10 M insulin (63). Cultures of cells were transferred to phenol red-free DMEM supplemented with 5% dextran-coated charcoal-treated FBS (DCC-FBS), BME amino acids, MEM nonessential amino acids, sodium pyruvate, and penicillin-streptomycin for 48-72 hr before plating. Cells were plated at a density of 2 × 10 6 cells/well and maintained for an additional 24 hr in DMEM with DCC-FBS. For estrogen dose-response assays, cells were then transfected for 5 hr with 200 ng of pERE2x-luciferase plasmid with or without 50 ng of pcDNA3.1-hERα or pcDNA3.1-hERβ (485 bp) using Effectene transfection reagent (Qiagen Inc., Valencia, CA) according to the manufacturer's protocol. For all luciferase assays, we then incubated cells for 18-24 hr in DMEM with DCC-FBS in the presence of vehicle or various chemicals as previously described (64). When double treatments were performed, E 2 or o,p´-DDT was added 20 min after the ICI 182,780 treatment, or various concentrations of DDT metabolites were treated 20 min before adding 20 ng/mL PMA. We used PMA as a positive control because we have previously shown that 20 ng/mL PMA activates protein kinase C, downstream mitogen-activated protein kinases, and AP-1 (65). The concentration of ICI 182,780 used inhibits E 2 -stimulated gene expression (66). In our results we have shown the data from treatments using 10-50 µM DDT and its metabolites, which gave significant activity and are environmentally relevant concentrations (59)(60)(61). Cells were then harvested and luciferase activity was measured using 30 µL of cell extract and 100 µL of luciferase assay substrate (Promega, Madison, WI) in a monolight 2010 luminometer. The data shown are an average of at least three independent experiments with 2-4 replicates.
Generation of stably transfected cell lines. Cells were transfected for 5 hr with either 10 µg of pAP-1(PMA)-luciferase (Clontech, Palo Alto, CA; Ishikawa cell variants) or with pAP-1-luciferase (Stratagene, La Jolla, CA; HEK 293 cells) along with 1 µg of the neomycin-resistance gene expression vector pcDNA3.1 plasmid using Lipofectamine lipofection reagent (Life Technologies, Gaithersburg, MD) in 75-cm 2 flasks. Fortyeight hours after transfection, we switched cells to a medium containing 400-1,200 µg/mL G418 (Mediatech, Herndon, VA) and maintained them in this medium for 3 weeks until the disappearance of all cells in a control, nontransfected flask and appearance of colonies in the transfected flasks. Colonies were pooled, grown, and tested for luciferase activity by incubation with 20 ng/mL PMA for 24 hr.
Statistical analysis. We analyzed data using one-way analysis of variance and post-hoc Tukey's multiple comparisons with GraphPad Prism, version 3.02 (GraphPad Software, Inc., San Diego, CA). We separated the data of each Ishikawa stable cell line into two groups to compare fold activation over negative control (DMSO) and potentiation/inhibition of PMAinduced activity to explore the effects of organochlorines on a tumor promoter-stimulated cascade. Because of large differences in activation in some cases, we used a natural log transformation to obtain normal distribution. Statistically significant changes were determined at the p < 0.05 or < 0.001 level as indicated for each figure or table.

Ishikawa(+) cells, unlike Ishikawa(-) cells, demonstrate estrogen-mediated gene expression.
The difference in estrogen response was measured using a reporter gene assay in which an ERE-luciferase reporter plasmid was transiently transfected into the two Ishikawa cell variants alone or in combination with an hER(α or β) expression plasmid ( Figure 2). Increasing concentrations of E 2 stimulated luciferase activity in Ishikawa(+) cells ( Figure  2A), but not in Ishikawa(-) cells, even at 10 nM ( Figure 2B). In all experiments where E 2 stimulated luciferase activity, the pure antiestrogen ICI 182,780 (100 nM) inhibited the effects of 10 nM E 2 (data not shown). The responsiveness of the Ishikawa(-) cells could be restored by cotransfection with either an hERα or hERβ expression plasmid, although, overexpression with hERβ consistently produced a lower activity profile. On the other hand, hERβ expression slightly inhibited the endogenous response of Ishikawa(+) cells (Figure 2A). This phenomenon is consistent with what other laboratories have reported and is a result of the difference in endogenous ER expression (70)(71)(72)(73). The endogenous low ERβ expression was unable to transmit an estrogenic response ( Figure 2B). Hence, the difference in estrogen response between the two Ishikawa cell variants is caused by a loss of ERα expression.
DDT and its metabolites stimulate AP-1mediated gene expression in both Ishikawa cell variants. To examine the effects of the ER on AP-1-mediated gene expression by environmental compounds, we generated from our two Ishikawa cell variants stable cell lines containing an AP-1-responsive promoter linked to a luciferase reporter gene. This allowed us to quickly and efficiently test related cells that varied in their response to estrogen for AP-1-induced activity by various stimuli. Treatment of the estrogen-responsive Ishikawa(AP-1)+ stable cells produced a significant induction of AP-1 activity by all the DDT, DDD, and DDE compounds in a dose-dependent manner at the environmentally relevant concentrations 25 µM and 50 µM ( Figure 3A) with 50 µM o,p´-DDT being the most potent (4.2-fold). In addition, ICI 182,780 (100 µM) potentiated o,p´-DDT (50 µM)-induced AP-1 activity 2.6 ± 0.97-fold in Ishikawa(AP-1)+ cells (data not shown), consistent with the findings that ICI 182,780 potentiates AP-1-mediated gene expression (38). Surprisingly, however, when the estrogenunresponsive Ishikawa(AP-1)-cells were treated with o,p´-DDT and o,p´-DDD, we observed a potent activation of AP-1 of roughly 15.5-fold ( Figure 3B). This result and the fact that o,p´-DDD has virtually no ER binding activity (19,21), indicates a non-ER-mediated AP-1 signaling mechanism. p,p´-DDA, a metabolite found in humans, had no effect on AP-1 activity. When organochlorine compounds were administered in conjunction with the phorbol ester PMA and compared to percent activation by PMA, potentiation of AP-1 activity was seen in both cell variants, with the Ishikawa(AP-1)-cells once again demonstrating higher-fold stimulation ( Figure 3C, D). o,p´DDD in conjuction with PMA gave the greatest AP-1 stimulation at 50 µM in both the Ishikawa(AP-1)+ (7.6fold) and Ishikawa(AP-1)-(38.1-fold) variants. Again, p,p´-DDA did not significantly increase AP-1 activity.

DDT and its metabolites stimulate AP-1mediated gene expression in HEK 293 cells under a different AP-1-responsive promoter
context. Next, we further tested this effect in an additional estrogen-unresponsive cell line using a different AP-1-responsive promoter context. HEK 293 cells, which have been shown to be estrogen unresponsive (63,64,74)

Articles • DDT and its metabolites induce AP-1 activity
Environmental Health Perspectives • VOLUME 110 | NUMBER 12 | December 2002 were used to make stable cell lines with a different AP-1-luciferase reporter construct integrated into the genome as confirmed above.
As with both Ishikawa cell lines, the DDT, DDD, and DDE metabolites in conjunction with PMA all stimulated AP-1 significantly versus control, whereas p,p´-DDA could not induce an effect (Figure 4). Additionally, significant increases were detected at 10 µM for all activating compounds tested, raising the possibility that further testing of multiple cell lines or AP-1-responsive promoters could reveal effects of these compounds at yet lower concentrations.

Discussion
Organochlorines represent a class of environmental compounds characterized by a chlorinated hydrocarbon backbone motif. These chemicals are found in pesticides, plastics, and industrial wastes and are ubiquitous environmental pollutants (1,2). DDT, one of most widely used pesticides until 1973, was banned because of its adverse effects on wildlife (1,75). However, this chemical is still used in many developing countries today and exists at high concentrations along with its metabolites and contaminants throughout the world, including the United States, because of long half-lives in soil, water, and the adipose tissue of animals (24,28,30,(75)(76)(77).
Although DDT and some of its metabolites bind ERs (16,19,21), induce egg-shell thinning in raptors (78,79), and induce estrogen-like effects in exposed animals and humans (1,22,25), the mechanisms of action of this class of synthetic compounds is still not completely understood. Here we have shown that DDT can directly affect the activity of AP-1, an important cell regulatory factor.
To determine whether DDT and its metabolites induced AP-1 activation through an ER-mediated mechanism, we created human endometrial adenocarcinoma Ishikawa stable cell lines containing an AP-1-responsive promoter linked to a luciferase reporter gene. This novel cell system allowed us to evaluate rapidly the signaling effects of a number of environmental compounds. Our Ishikawa variants provide us with the cell culture equivalent of a comparison system similar to the wild-type versus ER-knockout mice. Treatment of the Ishikawa(AP-1)+ cells revealed that the DDTs, DDDs, and DDEs could all significantly stimulate AP-1 activity, with the DDTs and DDDs having the greatest activity on average. Although the DDTs have been demonstrated to have estrogenic activity (1,19,21), and hence, may work though an ER-AP-1-dependent pathway, the nonestrogenic DDDs also stimulated AP-1mediated gene expression, indicating an alternative signaling mechanism for this class of compounds. Also, not only did treatment of the Ishikawa(AP-1)-cells activate AP-1, but the DDTs appeared to be more potent in this cell line, indicating that DDT and its metabolites can signal AP-1 in an ER-independent mechanism. Again, DDTs and DDDs appeared most potent, whereas in both cell lines the metabolite p,p´-DDA gave no significant induction. p,p´-DDA, unlike the DDT, DDD, and DDE isoforms, possesses no central chlorines but only a carboxylic acid motif, suggesting the chlorine residues are necessary.   Often, late-stage endocrine cancers grow in the absence of hormones (80)(81)(82), indicating the potential up-regulation of some other cell survival/cell signaling mechanisms. Because our estrogen-unresponsive cells give a stronger AP-1 activation, it is possible that the AP-1 signaling mechanisms have been up-regulated to compensate for the loss of responsiveness to estrogen. Regardless of which variant is tested, treatment of cells in combination with the tumor promoter PMA revealed that DDT and its metabolites potentiate AP-1 activity well beyond PMA alone. This effect could also be seen in HEK 293 cells using a different AP-1responsive promoter, indicating that DDTs additionally stimulate AP-1 in both other estrogen-unresponsive cell lines and under different promoter contexts, demonstrating the prevalence and diversity of this environmental signaling process. Significant increases in AP-1 activity were seen at lower concentrations in the HEK 293 cells, suggesting that different cell or AP-1 promoter contexts would reveal that the DDTs could signal AP-1 at more widespread concentrations. The potentiation could signify the involvement of a DDT-AP-1 signaling cascade different from tumor promoter-induced AP-1 activation. Future experiments will determine the kind of signal (i.e., proliferation, stress, etc.) these organochlorines are sending to the cell and how they are signaling AP-1.
Recently, compounds of the DDT family have been reported to affect a number of AP-1 regulated genes in vivo. Diel et al. (83) reported that ovariectomized rats treated with o,p´-DDT reduced the expression of clusterin, which plays an important role in apoptosis, similar to AP-1. The promoter region of the clusterin gene contains two functional AP-1 sites (84). In this situation, o,p´-DDT could potentiate members of the Jun and Fos families of proteins that have more inhibitory effects toward AP-1-regulated gene expression (e.g., JunB, Fra-1, Fra-2) and hence inhibit expression of the clusterin gene, leading to a more proliferative phenotype. This would help explain the results demonstrating the ability of o,p´-DDT to significantly increase uterine growth similar to E 2 , despite poorly binding to the ER and displaying a different gene expression pattern in the uterus than E 2 (83,85). Ghosh et al. (56) demonstrated in ERα-knockout mice that methoxychlor, but not E 2 , could still stimulate estrogen responsive lactoferrin and glucose-6phosphate dehydrogenase mRNAs in the mouse uterus despite the addition of the pure antiestrogen ICI 182,780, indicating a non-ER (α or β) mechanism. In addition, Das et al. (86) demonstrated that mice treated with p,p´-DDD, which has no ER-binding ability, also increases lactoferrin expression, regardless of the presence of ICI 182,780. Both the lactoferrin and glucose-6-phosphate dehydrogenase genes contain complex promoters that are responsive to multiple stimuli, such as epidermal growth factor, PMA, transforming growth factor-α, that also potentiate AP-1 activity (87). In fact, the yeast homologue of the human c-Jun, YAP-1/PAR1, potentiates glucose-6-phosphate dehydrogenase activity (88). Further support for this comes from the fact that the lactoferrin promoter contains an AP-1-like element that has been shown to bind the AP-1 protein (6). Collectively, these in vivo observations combined with our in vitro data strongly suggest that organochlorine compounds such as DDT regulate gene expression through multiple elements present in the promoter regions of various genes.
AP-1 activation has become a marker for a number of effects elicited initially at the plasma membrane. External stimuli such as growth factors, hormones, ultraviolet irradiation, heavy metals, and even mechanical stress have been implicated in the induction of AP-1 (5,7,(89)(90)(91). For example, epidermal growth factor binds to its tyrosine kinase receptor located at the plasma membrane, initiating a signaling cascade that results in the up-regulation of AP-1 activity (3). Estrogen has been implicated in a number of early signaling effects including activation of G-protein-coupled receptors and Ca 2+ spikes (92)(93)(94). Cadmium stimulates AP-1-mediated gene expression through an ER-associated process (95,96), further hinting at the role of AP-1 as a potential environmental sensor. Additionally, treatment with cadmium, or other environmental estrogens such as bisphenol A and diethylstilbestrol, has been linked to alterations at the membrane level, leading to activation of specific signaling cascades (97,98). These responses are thought to be mediated through plasma membrane receptors. The effect we have reported here suggests that organochlorine compounds, such as DDT and its metabolites, may stimulate early membrane responses, similar to other extracellular stresses, leading to gene expression.
Here we demonstrated a novel effect of the DDTs using a new, stable-cell reporter-gene system. The pleiotropic nature of these organochlorines is revealed in their ability to stimulate expression from a classic ERE promoter element (2,19), activate receptor tyrosine kinases as well as multiple kinase pathways (45,47,48,50,51), inhibit L-type Ca 2+ channels (99,100) and, as shown here, their capacity to stimulate AP-1 activity through non-ER-mediated mechanisms. Thus, just as the ER appears to be a convergent point for multiple cellular signals such as ligands (38,101), growth factors (102)(103)(104), and cell survival signals (105), AP-1 may function as a transcriptional mediator of various signaling pathways. Environmental compounds may adversely affect the physiology of a cell by mimicking aspects of its molecular endocrine system. Likewise, compounds may co-opt the cell's physiology by altering other cell signal regulators. Collectively, this may represent a new approach to environmental toxicology in which we go beyond cell damage to molecular modulators of cell responses.