The Short-Chain Fatty Acid Methoxyacetic Acid Disrupts Endogenous Estrogen Receptor-α–Mediated Signaling

Background Ethylene glycol monomethyl ether (EGME) exposure is associated with impaired reproductive function. The primary metabolite of EGME is methoxyacetic acid (MAA), a short-chain fatty acid that inhibits histone deacetylase activity and alters gene expression. Objective Because estrogen signaling is necessary for normal reproductive function and modulates gene expression, the estrogen-signaling pathway is a likely target for MAA; however, little is known about the effects of MAA in this regard. Methods We evaluated the mechanistic effects of MAA on estrogen receptor (ER) expression and estrogen signaling using in vitro and in vivo model systems. Results MAA potentiates 17β-estradiol (E2) stimulation of an estrogen-responsive reporter plasmid in HeLa cells transiently transfected with either a human ERα or ERβ expression vector containing a cytomegalovirus (CMV) promoter. This result is attributed to increased exogenous ER expression due to MAA-mediated activation of the CMV promoter. In contrast to its effects on exogenous ER, MAA decreases endogenous ERα expression and attenuates E2-stimulated endogenous gene expression in both MCF-7 cells and the mouse uterus. Conclusions These results illustrate the importance of careful experimental design and analysis when assessing the potential endocrine-disrupting properties of a compound to ensure biological responses are in concordance with in vitro analyses. Given the established role of the ER in normal reproductive function, the effects of MAA on the endogenous ER reported here are consistent with the reproductive abnormalities observed after EGME exposure and suggest that these toxicities may be due, at least in part, to attenuation of endogenous ER-mediated signaling.


Research
Methoxyacetic acid (MAA) is the primary metabolite of the industrial solvent ethylene glycol monomethyl ether (EGME), which has been used in a variety of coatings and as a jet fuel additive (Gargas et al. 2000;Miller et al. 1983). Interest in EGME and MAA stems from epidemiologic analyses and labo ratory studies that have linked exposure to these compounds with reproductive toxicity. In women, occupational exposure to ethylene glycol ethers has been associated with increased risks of spontaneous abortion and subfertility (Correa et al. 1996), whereas exposed males have decreased sperm counts (Welch et al. 1988). In laboratory studies, EGME has been shown to target the ovarian luteal cell, sup press cyclicity, and inhibit ovulation in female rats (Davis et al. 1997); in male rats, EGME has been reported to reduce testicular size and fertility (Rao 1971). Subsequent studies have shown that many of the untoward reproduc tive effects observed after EGME exposure can be reproduced by exposure to MAA alone, suggesting that MAA is primarily responsible for the compromised reproductive function associated with EGME exposure (Davis et al. 1997;Foster et al. 1984).
The chemical structure of MAA places it in the shortchain fatty acid family, which includes the anti epileptic drug valproic acid (VPA) and the intestinal bacterial product sodium butyrate (NaB). Interestingly, VPA is also associated with reproductive toxicity, including menstrual abnormalities and poly cystic ovaries (Isojarvi et al. 1993;Lofgren et al. 2007;O'Donovan et al. 2002), suggesting the shortchain fatty acids may share similar mechanisms of action that lead to reproductive toxicity. A common feature of MAA, VPA, and NaB is their ability to inhibit histone deacetylases (Boffa et al. 1978;Gottlicher et al. 2001;Jansen et al. 2004;Phiel et al. 2001;Sealy and Chalkley 1978), which suggests that one of their major mecha nisms of action may be to alter gene expression via histone hyper acetylation. Microarray analyses have con firmed that MAA and VPA, as well as other histone deacetylase (HDAC) inhibitors, alter gene expression profiles in human cell lines; however, the total number of genes regulated by these compounds is relatively low (Jansen et al. 2004;Reid et al. 2005). In addition to altering gene expression profiles through his tone hyper acetylation, several HDAC inhibi tors, as well as MAA and other shortchain fatty acids, have been shown through in vitro analyses to modulate intra cellular signal ing pathways such as the MAPK (mitogen activated protein kinase) pathway, which may contribute to their effects on gene expression and cell viability (Jansen et al. 2004;Jung et al. 2005;Michaelis et al. 2006;Rivero and Adunyah 1996;Witt et al. 2002).
Although the histone deacetylase inhibitory activity of MAA has been characterized, little is known regarding the effects of this com pound on estrogen signaling, which is critical to reproductive function in both the male and female (Hewitt et al. 2005). Estrogens use both genomic and non genomic mechanisms to alter the gene expression patterns and proliferative rates of target tissues and cells (Bjornstrom and Sjoberg 2005). Many of these effects are mediated by estrogen receptors ERα and ERβ, which are differentially expressed transcription factors that bind estrogens and transcription ally regulate the expression of numerous genes. In light of the role of estrogen signaling in nor mal reproductive function and gene expression and the reproductive toxicity associated with MAA, we sought to determine what effects MAA might have on estrogen signaling in vitro and in vivo to gain further insight into the molecular mecha nisms of action of MAA.

Materials and Methods
Mammalian cell culture. MCF7 cells (ATCC, Manassas, VA, USA) were cultured in Dulbecco's modified Eagle medium: Nutrient Mixture F12 (DMEM/F12; Invitrogen Corporation, Carlsbad, CA, USA) supplemented with 10% fetal bovine serum (FBS; Atlanta Biologicals, Lawrenceville, GA, USA), penicillin (100 U/mL), and strepto mycin (100 µg/mL) and incubated at 37°C in a humidified atmosphere containing 5% CO 2 . HeLa cells (ATCC) were cultured in DMEM (Invitrogen) supplemented with 10% FBS, penicillin (100 U/mL), and Background: Ethylene glycol monomethyl ether (EGME) exposure is associated with impaired reproductive function. The primary metabolite of EGME is methoxyacetic acid (MAA), a shortchain fatty acid that inhibits histone deacetylase activity and alters gene expression. oBjective: Because estrogen signaling is necessary for normal reproductive function and modulates gene expression, the estrogen-signaling pathway is a likely target for MAA; however, little is known about the effects of MAA in this regard. Methods: We evaluated the mechanistic effects of MAA on estrogen receptor (ER) expression and estrogen signaling using in vitro and in vivo model systems. results: MAA potentiates 17β-estradiol (E 2 ) stimulation of an estrogen-responsive reporter plasmid in HeLa cells transiently transfected with either a human ERα or ERβ expression vector containing a cytomegalovirus (CMV) promoter. This result is attributed to increased exogenous ER expression due to MAA-mediated activation of the CMV promoter. In contrast to its effects on exogenous ER, MAA decreases endogenous ERα expression and attenuates E 2 -stimulated endogenous gene expression in both MCF-7 cells and the mouse uterus. conclusions: These results illustrate the importance of careful experimental design and analysis when assessing the potential endocrine-disrupting properties of a compound to ensure biological responses are in concordance with in vitro analyses. Given the established role of the ER in normal reproductive function, the effects of MAA on the endogenous ER reported here are consistent with the reproductive abnormalities observed after EGME exposure and suggest that these toxicities may be due, at least in part, to attenuation of endogenous ER-mediated signaling. streptomycin (100 µg/mL) and incubated at 37°C in a humidified atmosphere containing 5% CO 2 .
Animals and treatments. All procedures involving animals were approved by the Animal Care and Use Committee of the National Institute of Environmental Health Sciences. All animals were treated humanely and with regard for alleviation of suffering. Tenweekold ovariectomized C57BL/6 mice (Charles River Laboratories, Raleigh, NC, USA) were housed in plastic cages in a temperaturecontrolled room (21-22°C) with a 12hr light/dark cycle. Mice were given NIH 31 mouse chow (Ziegler Bros. Inc., Gardner, PA) and fresh water ad libitum. Groups of mice (n = 3/group) were treated by intra peritoneal injection with saline, 1 µg/kg 17βestradiol (E 2 ), or 400 mg/kg MAA for 2 hr before necropsy. One additional group was treated with 400 mg/kg MAA 30 min before treatment with 1 µg/kg E 2 for 2 hr. Animals were killed using CO 2 , and uteri were collected and snapfrozen.
RNA isolation and real-time PCR analys is. Cells. MCF7 cells were plated into 6well plates (1 × 10 6 cells/well) and incubated over night in DMEM/F12 medium supplemented with 10% FBS. The next day, the media was aspirated from each well, cells were washed with phosphatebuffered saline (PBS), and fresh media [DMEM/F12 containing 10% charcoal/dextrantreated FBS (HyClone, Logan, UT, USA)] was added to each well. The cells were incubated overnight and then treated for 24 hr. At the end of the treatment period, the cells were harvested and total RNA was isolated using the RNeasy Mini Kit (Qiagen Incorporated, Valencia, CA, USA) according to the manufacturer's protocol.
Uterine tissue. Frozen uterine tissue was pulverized, and total RNA was isolated using TRIzol reagent (Invitrogen) according to the manufacturer's protocol.
Real-time reverse-transcriptase polymerase chain reaction (RT-PCR). Synthesis of complementary DNA (cDNA) and analy sis of genespecific cDNA concentrations by realtime PCR were performed as previously described (Deroo et al. 2004). Primers for real time PCR were designed with Primer Express software, version 2.0 (Applied Biosystems Incorporated, Foster City, CA, USA).
Western blots. MCF7 cells were cultured and treated as described above. After treat ment, cells were washed with PBS and lysed with MPER Mammalian Protein Extraction Reagent (Thermo Fisher Scientific, Rockford, IL, USA) containing Halt Protease Inhibitor Cocktail (Thermo Fisher Scientific) accord ing to the manufacturer's protocol to obtain total protein. Protein concentrations were determined using the BCA Protein Assay Kit (Thermo Fisher Scientific), and equal amounts of protein (20 µg) were separated on NuPAGE Novex 10% BisTris gels (Invitrogen). Proteins were transferred to nitro cellulose membranes and stained using the MemCode Reversible Protein Staining Kit (Thermo Fisher Scientific) to ensure equal protein transfer. Membranes were blocked and incubated with antibodies in Trisbuffered saline containing 5% milk and 0.1% Tween20. ERα protein levels were evaluated with a rabbit polyclonal antibody (sc7207; Santa Cruz Biotechnology Inc., Santa Cruz, CA, USA) and a horse radish peroxidase conjugated antirabbit antibody (NA934V; Amersham/GE Healthcare BioSciences Corporation, Piscataway, NJ, USA). ERα protein levels were visualized with ECL Plus (Amersham/GE Healthcare BioSciences) and BIOMAX MR film (Kodak/SigmaAldrich Corporation, St. Louis, MO, USA).
Transfections. HeLa cells were plated at a density of 1 × 10 5 cells/well into 24well plates in standard growth medium overnight. The following day the medium was changed to DMEM supplemented with 1% charcoal dextran stripped FBS (SFBS; Hyclone) and transfected using Fugene 6 (Roche Applied Science, Indianapolis, IN, USA) reagent according to the manufacturer's protocol. After transfection, the cells were incubated overnight in media supplemented with 10% SFBS. The cells were then treated for 24 hr and harvested and assayed for luciferase and βgalactosidase activities using the Luciferase Assay System (Promega Corporation, Madison, WI, USA) and the βGalactosidase Enzyme Assay System (Promega).
Statistical analysis. Data were analyzed for statistical significance using the MannWhitney non parametric test.

MAA potentiates exogenous ER-mediated signaling.
The reproductive toxicities associ ated with MAA exposure in both humans and animals are similar to some of the reproduc tive phenotypes observed in both ERα knock out mice (αERKO) and aromatase knockout mice (ArKO), which suggests that MAA may impart its untoward reproductive effects by compromising estrogenmediated signaling. Interestingly, despite the parallel phenotypes of MAA exposure and animal models of compro mised estrogen signaling, MAA has been shown to enhance exogenous nuclear receptor signal ing, including ER signaling (Jansen et al. 2004).
Because these in vitro data are incongru ent with the reproductive toxicity associated with MAA, we performed similar in vitro experiments to assess the effects of MAA on estrogen signaling. We transiently transfected ERnegative HeLa cells with an expression vector for human ERα or ERβ along with an estrogeninducible 3XERETATALuc firefly luciferase reporter plasmid and a constitutively active cytomegalovirus (CMV)βgalactosidase reporter plasmid and treated the cells for 24 hr with a solvent control or increasing concen trations of E 2 in the absence and presence of 5 mM MAA. Consistent with prior observa tions, our luciferase assay data show that MAA potentiates the activity of E 2 in HeLa cells transfected with the estrogenresponsive luciferase reporter and either human ERα or ERβ ( Figure 1A,B). The ER expression vec tor is necessary for this response, as identically treated HeLa cells transfected with the reporter in the absence of ERα and ERβ possessed mini mal luciferase activity that was unaltered by treatment with E 2 alone or E 2 plus MAA (data not shown). In contrast, when ERpositive MCF7 cells were transfected with only the 3XERETATALuc and CMVβgalactosidase reporter plasmids and treated identically to the HeLa cells, no potentiation of E 2 induced luciferase activity was observed with MAA cotreatment ( Figure 1C). However, MAA was able to potentiate estrogenstimulated luciferase activity in MCF7 cells cotransfected with an expression vector for human ERα ( Figure 1C). Interestingly, MAA alone increased luciferase activity relative to vehicle controls in HeLa cells transfected with either ERβ (~ 4.6fold) or ERα (~ 3fold) and in MCF7 cells trans fected with ERα (~ 6.5fold) ( Figure 1A-1C). Because the MAAmediated potentiation of E 2 stimulated signaling in both cell lines was observed only after transfection with ER expression vectors and because MAA alone was able to increase luciferase activity relative to vehicle controls, we examined the effect of MAA on exogenous ERα expression in HeLa cells. MAA dosedependently increased the expression of ERα protein in HeLa cells trans fected with the human ERα expression vec tor, which may account for the potentiation of estrogeninduced luciferase activity observed in the transfection experiments with exogenous ER ( Figure 1D). We observed this effect in the presence and absence of 10 nM E 2 (data not shown for E 2 ).
The expression vector for both ERα and ERβ is pcDNA3.1, which contains a CMV promoter. To determine if MAA was increasing ER expression by activating the CMV promoter within the expression vec tor, we transiently transfected HeLa cells with a renilla luciferase plasmid containing a CMV promoter (pRLCMV) and treated the cells with increasing concentrations of MAA. We observed a concentration dependent increase in luciferase activity, with maximum activation (~ 30fold relative to vehicle control) occurring after exposure to 20 mM MAA ( Figure 1E). In the same experiments 5 mM MAA induced an ~ 8fold increase in luciferase activity, indicating that this MAA concentration can significantly activate the CMV promoter. These data volume 117 | number 11 | November 2009 • Environmental Health Perspectives suggest that the MAAmediated potentiation of E 2 stimulated signaling in these transient transfection experiments is due to an increase in exogenous human ER expression via trans activation of the CMV promoter by MAA. Interestingly, MAA also trans activated the pRLtk and pRLSV40 promoters in a dose dependent fashion, demon strating 12fold and 10fold increases in luciferase activity, respectively, after exposure to 5 mM MAA (data not shown).

MAA treatment reduces endogenous ERα expression.
MAAinduced trans activation of the CMV promoter complicates the interpre tation of data obtained from in vitro experi mental systems incorporating exogenous ER. Therefore, we performed experiments to examine the effect of MAA on the endogenous expression of ERα in MCF7 cells. MCF7 cells were treated with increasing concentra tions of MAA, and endogenous ERα protein expression was detected by Western blot. We observed a concentrationdependent decrease in endogenous ERα protein expression, with maximal decreases occurring after treatment with 20 mM MAA, the highest concentra tion tested in these experiments (Figure 2A). To determine if the decrease in ERα protein levels corresponded with decreased steady state levels of ERα mRNA, MCF7 cells were treated with 5 mM MAA for 24 hr, and ERα expression was analyzed by realtime PCR. Treatment with 5 mM MAA decreased the expression of ERα mRNA by ~ 50% relative to vehicle controls ( Figure 2B), indicating that the decreased protein expression is due, at least in part, to diminished levels of ERα mRNA.
Further experiments were performed in mice to determine if this effect was observed in vivo. Ovariectomized C57/BL6 mice were treated for 2.5 hr with either saline or 400 mg/kg MAA, and uteri were collected for measure ment of steadystate levels of ERα mRNA by realtime PCR. The dose of MAA used in these experiments was based on a pre vious report showing that this dose affects nuclear receptor signaling in the mouse uterus (Jansen et al. 2004). MAA decreased ERα expression in the mouse uterus by ~ 30% rela tive to controls at 2.5 hr ( Figure 2C). Although this decrease was not statistically significant, the trend observed in these experiments indicates that MAA has similar effects on endogenous ERα expression in vitro and in vivo.
MAA treatment disrupts estrogen-mediated gene expression. We performed further experi ments to determine if decreased ERα expres sion after MAA treatment resulted in disrupted ERαmediated signaling. Toward this end, we treated MCF7 cells with either 1 nM E 2 , 5 mM MAA, or 1 nM E 2 plus 5 mM MAA for 24 hr, and evaluated estrogenregulated gene expression by realtime PCR. As shown in Figure 3A, the expression of pS2, MYC, GREB1, SPUVE, and MCM3 was potentiated by E 2 treatment; however, pretreatment with 5 mM MAA attenuated the estrogeninduced responses. Taken together, these data show that MAA attenuates endogenous ER signaling, resulting in disruption of estrogenmodulated endogenous gene expression in MCF7 cells.
To determine whether MAA has a similar effect on estrogenmodulated gene expression in vivo, ovariectomized mice were treated for 2 hr with 1 µg/kg E 2 , 400 mg/kg MAA, or 1 µg/kg E 2 plus 400 mg/kg MAA, and uteri were collected for analysis of gene expression by realtime PCR. The MAA dose used in these experiments is based on a previously published report showing that this dose affects nuclear receptor signaling in the mouse uterus (Jansen et al. 2004). The mRNA levels of the estrogen inducible genes Greb1, Inhbb, and Fos were increased after treatment with E 2 alone; however, a 30min pretreatment with MAA reduced the E 2 mediated stimulation of each   Figure 3B). Although statistical signifi cance was not reached in these experiments, the clear trend in MAAmediated attenuation of E 2 stimulated mouse uterine gene expres sion indicates that MAA has similar effects on in vitro and in vivo estrogen signaling.

Discussion and Conclusions
EGME exposure is associated with reproduc tive toxicity in both humans and animals, and the majority of these effects are attributed to MAA, the primary metabolite of EGME.
Despite an established role for estrogen signal ing in reproductive function, limited informa tion is available regarding the effects of MAA on estrogen action. Therefore, in the present study we examined the effects of MAA on estrogen signaling, as altered estrogen signal ing may be responsible for some of the repro ductive toxicity associated with MAA. Our results show that MAA exerts anti estrogenic effects in vitro and in vivo by reducing endogenous ERα expression and attenuating E 2 mediated gene expression. Members of the shortchain fatty acid fam ily such as MAA, VPA, and NaB elicit numer ous responses in cells and tissues. One such response that has been described for MAA is the inhibition of histone deacetylase activity (Jansen et al. 2004), which appears to be a common mecha nism of action for the short chain fatty acids (Boffa et al. 1978;Gottlicher et al. 2001;Phiel et al. 2001;Sealy and Chalkley 1978). Because HDAC inhibitors exert a variety of effects on cells and tissues, including altered gene expression, cell cycle arrest, and apoptosis, many of the responses elicited by the shortchain fatty acids are likely associated with their histone deacetylase inhibi tory activity. Our results show that, in MCF7 cells, MAA alone was able to decrease the steadystate mRNA levels of ERα (1.9fold decrease) and the estrogenresponsive genes pS2 (1.6fold), MYC (1.5fold), and SPUVE (2.7fold) compared with vehicle controls (Figures 2B and 3A), whereas it increased the expression of CDKN1C (4.2fold increase), a gene that was modestly down regulated by E 2 in our experiments (data not shown). However, some genes measured in this study, includ ing GREB1 and MCM3, were not altered by treatment with MAA alone ( Figure 3A). These results are consistent with those observed for other shortchain fatty acids and other HDAC inhibitors for which the expression levels of only a small number of genes are significantly altered. For example, treatment of MCF7 cells with either VPA or trichostatin A (TSA) results in ~ 6% and ~ 20% changes, respectively, in the number of genes whose expression is altered greater than 2fold as determined from micro array analyses (Reid et al. 2005). A comparison of the gene expression profiles after treatment of MCF7 cells with MAA, VPA, NaB, sub eroylanilide hydroxamic acid (SAHA), or TSA shows that many of the same genes are simi larly affected by each compound, including ERα, pS2, SPUVE, and CDKN1C ( Figures 2B  and 3A) (Reid et al. 2005). This suggests that most of these alterations in gene expression are due to a common mechanism, which is likely inhibition of histone deacetylase activity.
Because ERα plays an obligatory role in many aspects of estrogenmediated signaling, our observation that MAA reduces endoge nous ERα expression in vitro and in vivo sug gests that estrogenmediated signaling may be compromised. Indeed, our in vitro and in vivo analyses confirm that MAA inhibits estrogen mediated effects on gene expression, show ing for the first time that MAA antagonizes E 2 stimulated expression of ERα target genes. The shortchain fatty acids VPA and NaB have also been shown to reduce ERα expression in vitro (Reid et al. 2005;Stevens et al. 1984), suggesting that they may disrupt estrogen sig naling as well. Toward this end, VPA, in the absence of E 2 , has been shown to decrease the expression of ~ 90% of the genes that are upregulated by E 2 treatment in MCF7 cells (Reid et al. 2005) and to reduce E 2 stimulated MCF7 cell proliferation (Olsen et al. 2004). In addition, NaB, in the absence of estrogen, has been reported to inhibit MCF7 cell pro liferation (Abe and Kufe 1984). Furthermore, NaB attenuates E 2 stimulated expression of the known estrogen target genes progesterone receptor and pS2 (De los Santos et al. 2007).
Similar results have been observed for the HDAC inhibitors TSA and SAHA, suggesting that histone hyper acetylation may be respon sible for the anti estrogenic effects of the short chain fatty acids (De los Santos et al. 2007;Reid et al. 2005). Our in vivo data demon strate that MAA reduces ERα expression in the mouse uterus by ~ 30% compared with con trols ( Figure 2C) and attenuates E 2 stimulated gene expression in the uterus ( Figure 3B). The modest decrease in ERα expression in these studies suggests that the changes observed in E 2 stimulated gene expression may not be due solely to decreased ERα expression, but are likely due to additional mechanisms of action for MAA. This effect is not due to MAA acting as a competitive antagonist for ERα, as MAA does not compete with E 2 for binding to ERα (Jansen et al. 2004). Taken together, these data suggest that the anti estrogenic effects of the shortchain fatty acids are a class effect that may be due to their inherent HDAC inhibi tory activities, because MAA, VPA, and NaB have all been shown to reduce endogenous ERα expression and have been characterized as HDAC inhibitors.
Although MAA imparts anti estrogenic effects on endogenous ER signaling, it enhances estrogenstimulated reporter activity in the presence of exogenous ERα and ERβ in both HeLa and MCF7 cells. Similar results have been reported for MAA with respect to the exogenous ER as well as other exogenous nuclear receptors (Bagchi et al. 2009;Jansen et al. 2004). We observed these enhanced responses in MCF7 cells only when cells were cotransfected with ER expression vec tors, indicating that the presence of the ER expression vector is necessary for this effect. In contrast to endogenous ERα expression, which is decreased after MAA exposure, exog enous ERα expression is increased after MAA treatment in the presence and absence of E 2 , and this increase correlates with enhanced luciferase activity in our reporter assays. The under lying mechanism of MAAinduced increases in exogenous ER expression is activa tion of the CMV promoter, which is present in the ER expression vectors used in this study   and is frequently used in other expression vec tors. The shortchain fatty acids VPA and NaB and the HDAC inhibitor TSA have also been shown to activate the CMV promoter (Dion et al. 1997;Phiel et al. 2001), which sug gests that this is another shared feature of the shortchain fatty acid family and some HDAC inhibitors. The disparate results obtained in our experiments comparing the effects of MAA on endogenous and exogenous ER signaling highlight the importance of our observation that MAA activates the CMV promoter, as each set of results would lead to opposite con clusions regarding the effect of MAA on ER signaling. Based on this observation, careful considera tion should be given to experimental design when examining the effects of MAA and other shortchain fatty acids on nuclear receptor signaling to avoid errant conclusions based on experimental artifacts associated with CMVcontaining expression vectors. This observation extends to expression vectors con taining either TK or SV40 promoters, which were also potently transactivated by MAA in our experiments (data not shown). Similar data have been reported for VPA and NaB with respect to the SV40 promoter, again sug gesting a class effect for the shortchain fatty acids (Chen et al. 1999;Gorman et al. 1983).
We have demonstrated that MAA reduces endogenous ERα expression and that MAA treatment inhibits estrogenmediated endog enous gene expression in vitro and in vivo. Although extrapolation of the MAA doses used in this study to human exposure levels is challenging given the paucity of data that exists regarding EGME and MAA levels in exposed humans, the anti estrogenic effects of MAA we observed are consistent with the reproductive toxicities described for humans exposed to EGME (Correa et al. 1996;Welch et al. 1988;Welsch 2005). ERαmediated sig naling is critical to reproductive function in both males and females, as illustrated by the pheno types observed with αERKO mice. Male αERKO mice have reduced sperm counts, and both male and female αERKO mice are infer tile (Hewitt et al. 2005). Interestingly, these pheno types are similar to those observed in EGMEexposed men, who have reduced sperm counts, and women, who exhibit decreased fer tility (Correa et al. 1996;Welch et al. 1988). In a rat model, chronic EGME exposure sup pressed cyclicity and prolonged diestrus, pro viding further in vivo evidence consistent with attenuation of estrogenic responsiveness (Davis et al. 1997). Adverse reproductive effects have also been reported in men and women exposed to the shortchain fatty acid VPA, which pos sesses anti estrogenic properties similar to those of MAA (Isojarvi et al. 1993;O'Donovan et al. 2002). Taken together, these observations suggest that MAAmediated attenuation of ER signaling may play a role in the untoward reproductive effects observed in both males and females after EGME exposure.

correction
In the manuscript originally published online, the concentration of E 2 was given as 1 nM in Figure 1 and in the text referring to the figure. Also, the authors incorrectly noted that MAA potentiates E 2 activity in HeLa cells in a "dosedependent manner." These have been corrected here.