Permanent oviduct posteriorization after neonatal exposure to the phytoestrogen genistein.
OBJECTIVE: We investigated the mechanisms underlying the contribution of the oviduct to infertility.
METHODS: Female mice were treated on postnatal days 1-5 with corn oil or Gen (50 mg/kg/day). We compared morphology, gene expression, and protein expression in different regions of the reproductive tracts of Gen-treated mice with those of control littermates at several time points.
RESULTS: Neonatal Gen treatment resulted in substantial changes in expression of genes that modulate neonatal oviduct morphogenesis, including Hoxa (homeobox A cluster), Wnt (wingless-related MMTV integration site), and hedgehog signaling genes. An estrogen receptor antagonist blocked these effects, indicating that they were induced by the estrogenic activity of Gen, Oviducts of adults treated neonatally with Gen had abnormal morphology and were stably "posteriorized," as indicated by altered Hoxa gene patterning during the time of treatment and dramatic, permanent up-regulation of homeobox genes (e.g., Pitxl, six1) normally expressed only in the cervix and vagina.
CONCLUSIONS: Neonatal exposure to estrogenic environmental chemicals permanently disrupts oviduct morphogenesis and adult gene expression patterns, and these changes likely contribute to the infertility phenotype.
KEY WORDS: developmental patterning, female reproductive tract, genistein, homeobox gene. Environ Health Perspect 119:1575-1582 (2011). http://dx.doi.org/10.1289/ehp.1104018 [Online 2 August 2011]
The effects of endocrine-disrupting chemica critically depend on the developmental timing of exposure and the concentration of active chemical that actually reaches the target tissue. Exposure to estrogenic chemicals is a well-documented risk to human health based on the experience with prenatal exposure to diethylstilbestrol (DES) and its effects on development of the female reproductive tract (FRT). Many estrogenic compounds are found as environmental contaminants at levels sufficient to have adverse effects on human health. Phytoestrogens in the diet or nutritional supplements are common sources of environmental estrogen exposure (Reinli and Block 1996). The most abundant dietary phytoestrogens are the isoflavones genistein (Gen) and daidzein, which are present in soybean-derived products (Reinli and Block 1996). Even in strict vegetarians, adult serum isoflavone levels are low ([less than or equal to] 0.5 [micro]M) and should have minimal estrogenic activity (Jefferson 2010; Peeters et al. 2007).
In contrast to adults, infants fed soy-based formula have relatively high serum isoflavone concentrations; in a recent study, Cao et al. (2009) found that the 25th, 50th, and 75th percentiles in serum were 1.5, 3.3, and 5.4 [micro]M, respectively, and the highest concentration was 13.2 jiM. In animal models these levels arc clearly estrogenic, as indicated by uterine wet weight bioassays (Jefferson and Williams 2010). Two prospective studies demonstrate that soy-based infant formula elicits estrogenic responses in the developing human female (Bernbaum et al. 2008; Zung et al. 2008). Limited epidemiological information is available regarding long-term effects of infant exposure to soy-based formulas, but in a recent large study, D'Aloisio et al. (2010) reported an increased incidence of physiciandiagnosed fibroids by 35 years of age. These studies suggest that neonatal exposure to phytoestrogens or other environmental estrogenic chemicals could have a long-term effect on adult female reproductive health.
FRT development is similar in most mammals, although the timing of morphogenesis differs among species. In the mouse, the FRT develops from the Mullerian duct beginning on gestation day (GD) 13 (Masse et al. 2009). Prenatal FRT development is controlled by several transcription factors and secreted signaling glycoproteins of the Wnt (wingless-related MMTV integration site) and hedgehog families. Anterior-posterior (AP) patterning of the Mullerian duct is regulated by positional information derived from the Hoxa (homeobox A cluster) family of homeobox transcription factors. Expression of these genes becomes spatially restricted before birth, and distinct FRT regions eventually develop that correlate with increased expression of specific Hoxa genes: Hoxa9 in the oviduct, HoxalO and Hoxall in the uterus and upper cervix, and Hoxa 13 in the lower cervix and vagina (Block et al. 2000; Ma et al. 1998; Taylor et al. 1997).
Epithelial cell differentiation in the mouse FRT begins after birth and continues until adulthood. At birth, the oviduct is a simple tubular structure with a single layer of columnar epithelium. During the neonatal period, this epithelium differentiates into secretory or ciliated columnar cells in response to cues received from the underlying mesenchyme that vary along the AP axis (Yamanouchi et al. 2010). These differences result in distinct patterns of epithelial morphogenesis such that the oviduct infundibulum and ampulla consist largely of ciliated epithelial cells, whereas the isthmus has more secretory cells. Similarly, uterine and vaginal mesenchymal cells generate cues that determine the epithelial cell types in those regions, with the adult uterus having mainly simple columnar cells and the vagina developing a stratified squamous epithelium (Kurita et al. 2001).
The neonatal FRT is highly sensitive to estrogenic disruption during morphogenesis and cellular differentiation. Numerous studies performed in animal models show that prenatal or neonatal exposure to estrogenic compounds causes dramatic alterations in FRT development and can lead to FRT cancers (Ma 2009). To determine whether exposure to environmentally relevant phytoestrogens has similar effects on FRT development and function, we used a mouse model of subcutaneous Gen exposure on postnatal days (PND) 1-5 (50 mg/kg/day). As long as the final serum Gen concentration is similar, the effects of Gen derived from oral exposure during the neonatal period are comparable to those observed when Gen is administered subcutaneously (Jefferson et al. 2009a). This model generates serum Gen levels [maximum concentration (Cmax) in female pups, 6.8 [micro]M] similar to those found in human infants fed soy-based infant formula (Cao et al. 2009; Doerge et al. 2002). As adults, these mice have multioocyte follicles, lack regular estrous cycles, and are infertile even after superovulation, and approximately 30% develop uterine cancer (Jefferson et al. 2005; Newbold et al. 2001). Abnormalities in both the oviduct and uterus contribute to the infertility phenotype. Although the mice become pregnant, approximately 50% of the pre-implantation embryos die in the oviduct after the two-cell stage, and the uterus does not support implantation (Jefferson et al. 2009b). The goal of the present study was to test the hypothesis that neonatal Gen treatment causes permanent changes in oviduct gene expression and morphology that could explain the oviduct's contribution to infertility.
Materials and Methods
Animals and treatments. All animal procedures complied with National Institutes of Health/ National Institute of Environmental Health Sciences animal care guidelines; animals were treated humanely and with regard for alleviation of suffering. Mice were fed NIH-31 mouse chow (Zeigler Brothers, Gardners, PA), which was assayed for phytoestrogen content as previously described (Jefferson et al. 2009b). Female CD-1 pups were injected subcutaneously on PND1-PND5 with corn oil (control) or Gen (50 mg/kg/day; Sigma, St. Louis, MO), ICI-182780 (ICI; 1 mg/kg/day; Sigma), or ICI (1 mg/kg/day) 30 min before Gen (50 mg/kg/day). Gen was prepared in a suspension of corn oil at 5 mg/mL, and ICI was dissolved in corn oil at 1 mg/mL. All treatments were delivered in 20 [micro]L cornoil, and doses were calculated based on a pup weight of 2.0 g. At 6-8 weeks of age, the female mice underwent superovulation [equine chorionic gonadotropin followed by human chorionic gonadotropin (hCG), both at 5.0 IU in 0.1 mL saline; both from Calbiochem, Gibbstown, NJ] or superovulation and mating as described previously (Jefferson et al. 2009b). Only plug-positive mated females were used in experiments requiring pregnant mice.
Real-time reverse-transcriptase polymerase chain reaction (RT-PCR). Samples were collected as follows (n:= 3-6 independent biological replicates per group). On PND1, oviducts, uteri, and cervix/vagina from five mice were pooled per biological replicate. On PND5, oviducts, uteri, and cervix/vagina were collected 4 hr after the last treatment; oviducts from three mice were pooled per biological replicate, but uterus and cervix/ vagina samples were from individual mice. For PND22 pups and adults, both oviducts from one mouse were pooled per biological replicate. For adult nonpregnant mice, oviducts of superovulated females were collected before (8 hr after hCG) and after (15 hr after hCG) ovulation, and cumulus masses were removed. For adult pregnant mice, oviducts of superovulated females were collected on GD2 or GD4 (48 hr or 96 hr after hCG, respectively), and embryos were removed. For untreated adult FRT samples, oviducts (pair), uterus, cervix, and vagina collected from one mouse per biological replicate.
Total RNA was isolated, and cDNA was generated from 1 [micro] RNA using standard protocols. Primers were designed to amplify exon junctions. RT-PCR was performed on 20 ng cDNA using SYBR[R] Green-based detection on an ABI 7200 HT sequence detector (Applied Biosystems, Foster City, CA). Relative gene expression was calculated by the [DELTA][c.sub.t] method (Pfaffl 2001) using cyclophilin A (Ppia) expression for normalization.
Microarray analysis. Oviducts were collected on GD2 (48 hr after hCG), and embryos were removed. RNA was isolated using the RNeasy Mini Kit (Qiagen, Valencia, CA). Gene expression analysis was conducted on four independent biological replicates for each group using Agilent Whole Mouse Genome 4 x 44 multiplex format oligo arrays (Agilent Technologies, Santa Clara, CA). Slides were scanned with an Agilent Scanner, and data were obtained using Agilent Feature Extraction software (version 9.5), which performed error modeling, adjusting for additive and multiplicative noise. The resulting data were processed using Rosetta Resolver[R] (version 7.2; Rosetta Biosoftware, Kirkland, WA) and Ingenuity Pathway Analysis (Ingenuity Systems, Redwood City, CA) and deposited in the National Center for Biotechnology Information (NCBI) Gene Expression Omnibus (accession no. GSE27639; NCBI, Bethesda, MD).
Histology and immunohistochemistry. Formalin-fixed tissues were paraffin embedded and sectioned at 5 [micro]m. Sections were stained with hematoxylin and eosin (H&E) or Masson's trichrome according to standard protocols. For Pitxl (paired-like homeodomain transcription factor 1) immunohistochemistry, sections were blocked with 5Percent bovine serum albumin (BSA) in Trisbuffered saline/Tween 20 (TBST). Rabbit polyclonal anti-Pitxl antibody (Tremblay et al. 1998) was diluted 1:150 in TBST containing 1Percent BSA and 1Percent milk. Biotinylated anti-rabbit IgG (1:500; Vector Laboratories, Burlingame, CA), ExtrAvidin (1:50; Sigma), and NovaRED (Vector) were used for detection. For E-cadherin immunohistochemistry, sections were blocked with 10Percent horse serum and avidin/biotin blocking reagent (Vector), and E-cadherin antibody (Sigma) was used at a 1:150 dilution. secondary antibody was from the Vectastain Elite Kit (Vector), and diaminobenzidine was used for visualization.
Immunoblotting. Total oviduct protein (5 [micro]g/mouse in each lane) was separated on Novex 4-12Percent Tris-glycine gels and transferred to polyvinyl difiuoride (Invitrogen, Carlsbad, CA). Blots were blocked with 5Percent milk in TBST and then incubated with either Sixl [sine oculis-related homeobox 1 homolog (Drosophila); 1:1,000; Abcam, Cambridge, UK] or [beta]-actin (0.2 [micro]g/mL; Sigma) primary antibodies followed by horseradish peroxidaseconjugated anti-rabbit or anti-mouse IgG (Amersham, Piscataway, NJ). Bands were visualized using West Fern to Reagents (Pierce, Rockford, IL).
Statistical analysis. We analyzed data using JMP software (version 9; SAS Institute Inc., Cary, NC). Student's t-tests were usedto compare control and Gen-treated samples at each time point.
Abnormal oviduct histology after neonatal Gen treatment. As a first step in determining why preimplantation embryos from neonatal Gentrcated mice had poor survival in the oviduct after the two-cell stage, we analyzed the histology of oviducts collected on GD2. Oviducts from Gen-treated mice retained the expected regional differences between the infundibulum, ampulla, and isthmus (Figure 1A--H); however, significant abnormalities were present in tissue and cellular morphology compared with controls. The stromal and muscle compartments were abnormally thickened in all oviduct regions, reminiscent of uterine myometrium and stroma (Figure 1A-H). The thickened areas mainlv consisted of additional muscle tissue, and many areas did not have the normal clearlv defined outer and inner smooth muscle layers (Figure 1J). In some areas of the ampulla, pseudoglands were near the outer epithelial surface of the oviduct, with no intervening muscle layers (Figure 1J). There were localized areas of excessive epithelial proliferation, and many regions had areas of epithelial disorganization that were particularly apparent after cell borders were labeled by immunostaining for E-cadherin (Figure 1K,L). This staining also revealed that the epithelial cells had enlarged, rounded nuclei and were cuboidal rather than the predominant tall columnar epithelial cells observed in controls.
Alterations in oviduct gene expression during neonatal Gen treatment. We hypothesized that the changes in oviduct morphology were a result of Gen-induced alterations in expression of genes that influence postnatal FRT morphogenesis. The effect of Gen treatment on oviduct expression of secreted signaling proteins and transcription factors previously documented to affect FRT development was examined by real-time PCR at the completion of Gen treatment on PND5 (Figure 2). Wnt genes required for FRT development include Wnt4, Wnt5a, and Wnt7a (Masse et al. 2009). There was no difference in Wnt4 expression, whereas Wnt5a was up-regulated and Wnt7a was almost completely repressed in Gen-treated mice compared with controls. Follis'tatin-like-1 (Fstl1), a diffusible oviduct mesenchymal factor, which promotes formation of ciliated epithelial cells (Yamanouchi et al. 2010), was significantly reduced in Gentreated mice compared with controls. Secreted proteins in the hedgehog signaling pathway can regulate postnatal FRT morphogenesis and adult FRT function, and Wnt genes can be targets of hedgehog signaling (Franco et al. 2010). Indian hedgehog (Ihh) and hedgehog interacting protein (Hhip) were highly up-regulated, and desert hedgehog (Dhh) was up-regulated 2-fold, after Gen treatment compared with controls, whereas sonic hedgehog (Shh) was not detected (data not shown). GLI-Kruppel family member 1 (Glil), a common downstream mediator of hedgehog signaling, was also up-regulated, indicating that Gen treatment caused activation of hedgehog signaling in the neonatal oviduct.
Transcription factors that regulate patterning and differentiation of the FRT include forkhead box a2 (Foxa2) and the posterior Hoxa genes (Jeong et al. 2010). Foxa2 expression was up-regulated on PND5 by Gen treatment. Expression of Hoxa9, Hoxal0 and Hoxa 11 was decreased in oviducts of Gen-treated mice compared with controls, whereas Hoxal3 was highly up-regulated. This Hoxa gene expression pattern indicated that the PND5 oviduct was posteriorized; that is, the oviduct expressed Hoxa genes in a pattern similar to that normally observed in the posterior cervical and vaginal tissues (Ma et al. 1998).
Alterations in oviduct gene expression on GD2. We hypothesized that genes important for FRT morphogenesis and perturbed by neonatal Gen treatment would continue to be altered in adult oviducts and that additional oviduct genes would be misexpressed around the time of preimplantation embryo loss. Rather than choosing candidate genes, we performed a microarray-based gene expression analysis of oviducts from control and Gentreated mice on GD2. We validated selected probes that covered a range of intensity values and fold changes by real-time RT-PCR [see Supplemental Material, Table 1 http://dx.doi.org/10.1289/ehp.1104018)]. Principal component analysis showed sample separation into distinct groups based on treatment; this was confirmed using hierarchical cluster analysis (see Supplemental Material, Figure 1).
Table 1. Most highly up- and down-regulated oviduct genes. Gene Gene name Fold change (gene/control) Up-regulated genes Pitxl paired-like 99.3 homeodomain transcription factor 1 Serpina3a serine (or cysteine) 10.3 peptidase inhibitor, clade A, member 3A Hao3 hydroxyacid oxidase 8.4 (glycolate oxidase) 3 Lipf lipase, gastric 7.6 Osap ovary-specific acidic 7.5 protein lgkv5-43 immunoglobulin kappa 7.4 chain variable 5-43 Prss35 protease, serine, 35 7.1 Cxcl15 chemokine(C-X-C 6.9 motif) ligand 15 Six1 sine oculis-related 6.8 homeobox 1 homolog (Drosophila) Serpina5 serine (or cysteine) 6.8 peptidase inhibitor, clade A, member 5 Cyp11a1 cytochrome P450, 6.7 family 11, subfamily a, polypeptide 1 Nkx3-1 NK-3 transcription 5.9 factor, locus 1 (Drosophila) Down-regulated genes Kcne1 potassium -4.9 voltage-gated channel, Isk-related subfamily, member 1 Serpina 11 serine (or cysteine) -4.9 peptidase inhibitor, clade A, member 11 Myh7 myosin, heavy -5.1 polypeptide 7, cardiac muscle, beta Myh7 myosin, heavy -5.2 polypeptide 6, cardiac muscle, alpha Coch Cochlin, coagulation -5.3 factor C homolog coagulation factor C homolog (Limulus polyphemus) Serpina 1d serine (or cysteine) -5.7 peptidase inhibitor, clade A, member 1D Olah oleoyl-ACP hydrolase -5.9 Spockl sparc/osteonectin, -6.0 cwcv and kazal-like domains proteoglycan 1 Serpina le serine (or cysteine) -6.2 peptidase inhibitor, clade A, member 1E Serpina lb serine (or cysteine) -6.2 preptidase inhibitor, clade A, member 1B Epha6 Eph receptor A6 -7.8 Ano2 anoctamin 2 -18.4
Of the > 41,000 probes on the array, 1,125 probes were significantly altered (analysis of variance, p < 0.01) in Gen-treated mice compared with controls. Duplicate probes and probes with average intensity < 100 in both treatment groups were removed from the analysis data set. Within the remaining probes, 335 genes were up- or downregulated [greater than or equal to] 2-fold; the most highly altered genes are shown in Table 1. Significantly altered Ingenuity biological function categories included development, immune response, and cell proliferation [see Supplemental Material, Table 2 (http://dx.doi.org/10.1289/ehp.l 104018)]. Four of the six secreted signaling proteins (Wnt7a Ibb Dhh, and Hhip) and one transcription factor (Foxa2) that were altered on PND5 after Gen treatment by PCR analysis were also significantly altered on GD2 by microarray analysis. Taken together, these findings indicate that neonatal Gen treatment permanently altered gene expression in the adult oviduct.
Three of the most highly up-regulated genes on GD2 were the homeobox transcription factors Pitxl, Nkx3-1 [NK-3 transcription factor, locus 1 (Drosophila)], and Sixl. These genes all have critical roles in embryonic development but are not reported to be expressed in FRT tissues. We hypothesized that misexpression of these early development transcription factors contributed to abnormal oviduct morphogenesis in Gen-treated mice. To test this idea and to further verify the microarray data, we performed realtime RT-PCR on independently collected biological replicates. Oviducts were collected on PND5, on PND22 (prepubertal) from superovulated adult mice before and after ovulation, and on GD2 and GD4. Pitxl and Nkx3-1 were highly up-regulated in Gen-treated mice compared with controls at all time points (Figure 3A,B). Pitxl protein was not detected in control oviducts on PND5 (Figure 3C,E). However, Gen treatment induced Pitxl expression in nuclei of sporadic oviduct epithelial cells in all areas of the PND5 oviduct (Figure 3D,F). As on PND5, Pitxl was not observed in adult control oviducts (Figure 3G,I), but it was abundantly expressed in focal areas throughout the oviduct epithelium of adult Gen-treated mice (Figure 3HJ). Like Pitxl, Sixl mRNA was highly up-regulated at all time points, and Sixl protein was significantly increased in the oviducts of Gen-treated mice compared with controls (Figure 3K,L).
Broadly posteriorized oviduct gene expression after neonatal Gen treatment. Because oviduct Hoxa gene expression was posteriorized in Gen-treated mice, we tested the idea that aberrant expression of Pitxl, Sixl, and Nkx3-1 was also related to FRT posterioriza-tion. To do this, we first needed to determine the timing of Hoxa gene pattern establishment in controls. Hoxa gene patterning in PNDl controls resembled that in adult controls; that is, the highest levels of Hoxa9 were in the oviduct, Hoxa10 and Hoxa11 in the uterus, and Hoxa 13 in the cervix/vagina (Figure 4A). These findings, together with alterations in the oviduct Hoxa pattern in Gen-treated mice on PND5 (Figure 2), indicate that Gen exposure disrupted maintenance of previously established Hoxa gene patterning.
To determine if genes highly up-regulated in the PND5 oviduct by Gen treatment are normally expressed in the posterior FRT, we compared gene expression levels in the oviduct, uterus, cervix, and vagina of control mice. Pitxl, Shcl, and Nkx3-1, like Hoxal3, were expressed exclusively in the cervix/vagina and not in the uterus or oviduct on PND1 (Figure 4A) and PND5 (data not shown). This expression pattern was still observed in control adults, when Pitxl, Sixl, Nkx3-1, and Hoxal3 were highly expressed in the vagina and cervix but had minimal expression in the uterus or oviduct (Figure 4B). There was no Pitxl protein in the adult uterus (Figure 4C). However, beginning in the cervixat the squamocolumnar junction, Pitxl was detected in the nuclei of squamous epithelial cells (Figure 4D), and the staining intensity increased in vaginal epithelium (Figure 4E). Expression of the secreted signaling factors Wnt4, Wnt5a, Wnt7a, and Fstll was similar in all regions of the FRT on PNDl and PND5 and in adults (data not shown). In contrast, several hedgehog pathway genes and Foxa2 were more highly expressed on PNDl in the cervix/vagina compared with both the oviduct (Ihh, 11.0-fold; Dhh, 5.8-fold; Hhip, 42.1-fold; and Foxa2, 6.1-fold increased) and uterus (Ihh, 9.4-fold; Dhh, 6.2-fold; Hhipy 63.1-fold; and Foxa2, 651.7-fold increased), suggesting that these genes are also normally restricted to the posterior FRT. Together, these data demonstrate that neonatal Gen treatment blocked the normal process of transcriptional silencing of many posterior developmental genes in the anterior FRT.
To determine if the effects of neonatal Gen treatment on the oviduct are mediated by its estrogenic actions (Jefferson et Al. 2002), we examined whether the estrogen receptor (ER) antagonist ICI (Wakeling and Bowler 1992) blocked these effects. Neonatal mice treated with both Gen and ICI did not have altered expression of Hoxa9 and Hoxa l3, nor did they have altered expression of the three most highly up-regulated homeobox transcription factors (Pitxl, Sixl, and Nkx3-1) observed in mice treated with Gen alone (Figure 5). These findings indicate that molecular posteriorization of the oviduct in response to Gen treatment is mediated by its estrogenic action.
The results presented here demonstrate that a brief neonatal exposure to the phytoestrogen Gen causes substantial permanent changes in oviduct morphology and gene expression in mice. The altered expression of Hoxa and Wnt genes in the oviduct combined with aberrant expression of homeobox transcription factors that are normally expressed only in the posterior FRT likely account for abnormal oviduct differentiation and subsequent abnormal tissue architecture and function later in life.
Cellular differentiation of the FRT occurs during postnatal development in rodents, ruminants, nonhuman primates, and humans (Gray et al. 2001). Based on the presence of ER[alpha] in the FRT epithelium at birth, it is not surprising that an estrogenic compound administered during this period disrupts FRT morphogenesis and cellular differentiation. Indeed, numerous reports have described the effects of the potent synthetic estrogen DES on both prenatal and postnatal FRT development. Most of these studies have focused on the uterus; however, a few reports described the effects of DES on oviduct histopathology (Newbold et al. 1985). Furthermore, the doses of DES used had much higher estrogenic activity than that of estrogenic chemicals typically encountered in daily life. The neonatal Gen dosing strategy used in the present study results in a serum Gen concentration similar to that in human infants fed soy-based infant formula (Jefferson and Williams 2010). This model is also likely relevant to human exposure to low levels of other estrogenic environmental chemicals.
Prenatal DES treatment alters the expression pattern of Wnt genes in the FRT on GD19 (Suzuki et al. 2007). Similarly, in the present study, neonatal Gen treatmentresulted in increased Wnt5a expression and significantly reduced Wnt7a expression inthe oviduct on PND5, indicating that bothprenatal and neonatal exposure to estrogenic compounds disrupts Wnt gene expression in the FRT. Stratified luminal epithelium is observed in the uterus of Wnt7a-null mice (Miller and Sassoon 1998), and in the present neonatal Gen model we found Wnt7a repression and a marked up-regulation in the oviduct of genes that are normally observed in stratified cervical/vaginal squamous epithelium. These findings support the idea that Wnt7a is required for maintenance of the columnar epithelium in the oviduct and uterus and that Gen-induced Wnt7a repression in the neonatal FRT may promote oviduct posteriorization. However, if reduction in Wnt7a were sufficient to promote a cervical/ vaginal phenotype, we would expect that in control mice, Wnt7a expression would be lower in these tissues than in the uterus and oviduct. Instead, in controls during the neonatal period, Wnt7a is highly expressed in the cervix and vagina and is not differentially expressed compared with the uterus and oviduct. For this reason, the oviduct posteriorization in Gen-treated mice cannot be attributed to reduction in Wnt7a alone.
Hoxa genes direct AP patterning of the FRT and maintain a distinct spatial distribution in adults. From GD16.5 through GD19, Hoxa9 is similarly expressed throughout the FRT, but Hoxa1O, Hoxal 11 and Hoxal3 are differentially expressed along the AP axis, with Hoxa1O and Hoxa 11 highest in the uterus and Hoxal3 highest in the cervix and vagina (Ma et al. 1998; Suzuki et al. 2007). On the day of birth (PNDl), we found that Hoxa10 Hoxa11, and Hoxal3 had an expression pattern similar to that on El9, whereas Hoxa9 was more highly expressed in the oviduct compared with the uterus and cervix/ vagina. This Hoxa gene expression pattern is permanently maintained from the day of birth (PNDl) forward because FRT tissues from PNDl4 and adult mice have a similar expression pattern (Taylor et al. 1997). Hoxa9 is not essential for FRT development because Hoxa9-null mice have normal fertility (Fromental-Ramain et al. 1996). In contrast, Hoxa1O- and Hoxal -null mice have abnormal FRT development and function (Benson et al. 1996; Gendron et al. 1997). Hoxal3null mice have agenesis of the caudal portion of the Mullerian duct, and Hoxdl3 (homeobox Dl3)-null mice that also lack one copv of Hoxal3 exhibit severe malformations of the vagina and external genitalia, suggesting critical roles for these Hox gene paralogs in posterior FRT development (Frornental-Ramain etal. 1996;Warot et al. 1997).
Neonatal Gen exposure alters Hoxa gene expression in the oviduct, with Hoxa9, Hoxa10, and Hoxall expression significantly reduced compared with controls and Hoxttl3 increased. Similarly, Hoxa10 and Hoxall are reduced in the uterus after neonatal DES exposure (Block et al. 2000; Ma et al. 1998; Suzuki et al.2007). Because the adult pattern of Hoxa gene expression is normally present at birth, thesedata suggest that Hoxa gene patterning is not irreversible once established but retains some plasticity in the neonatal period. This finding is consistent with the observation that in Wnt7a-null mice, Hoxa10 and Hoxall expression in the uterus is normal at birth but then gradually decreases (Miller and Sassoon 1998), suggesting that Wnt7a is required for maintenance of FRT Hoxa gene patterning.
Hoxal3 up-regulation may contribute to neonatal Gen-induced posteriorization of the oviduct by inducing expression of several genes normally restricted to the cervix and vagina. This finding is in agreement with a mouse knock-in study in which replacement of the Hoxa11 homeodomain with the Hoxal3 homeodomain [A11.sup.3hd/13hd] resulted in pos-teriorization of the uterus to resemble a cervix/ vagina-like structure (Zhao and Potter 2001). More than 30 genes normally expressed in the cervix and not in the uterus were expressed in the uterus of the Al 1[31ld/13hd mutants. One of these genes was the transcription factor Sixl, suggesting that Sixl is normally regulated by Hoxa 13 via a mechanism involving the homeodomain. In the present study, Sixl was highly up-regulated in the adult oviduct after neonatal Gen treatment, and we demonstrated that its up-regulation begins during the time of Gen treatment and remains stable through adulthood. This expression pattern correlates with the up-regulation of Hoxal3 in the oviduct, suggesting that Hoxal3 expression in the anterior FRT is a proximal factor in altering cellular differentiation and causing phenotypic and molecular posteriorization of the oviduct.
One explanation for the altered expression of Hoxa genes in the oviduct is that Geninduced activation of ER-mediated transcription disrupts global regulation of Hoxa genes in the FRT. Indeed, Hoxa10 and Hoxall are regulated bv estrogen in the adult FRT (Taylor et al. 1998). Regulation of Hox gene expression during mammalian development is complex and involves regulatory RNAs and proteins whose activities orchestrate the establishment and subsequent maintenance of the AP Hox expression patterns (for review, see Chopra 2010). Because Hoxa patterning in the FRT is already established at birch, alterations in initiation factors cannot explain Geninduced disruption. Instead, alterations in maintenance factors, such as polycomb and trithorax group proteins, that preserve the initial Hox gene expression pattern and help cells "remember" which regulatory domains to keep silent and which ones to keep active, are more likely candidates.
Although Hox genes may function upstream of Pitxl, Six1 and Nkx3-1 after Gen treatment, the converse may also be true. These three genes have critical roles in directing embryonic development of tissues outside the FRT. Sixl regulates development of numerous tissues, including the eye, skeletal muscle, and posterior limbs (Christensen et al. 2008; Laclef et al. 2003). Nkx3-1 is important for prostate development and marks a stem cell population that functions during prostate regeneration (Sciavolino et al. 1997; Wang et al. 2009). Nkx3-1 was previously reported to be absent in the FRT, but tissues from the cervix or vagina were not examined (Sciavolino et al. 1997). Pitxl is involved in development of anterior branchial arch structures and the hindlimb (Lamonerie et al. 1996; Lanctot et al. 1999; Szeto et al. 1999). Pitxl acts upstream of Hox genes during hindlimb specification because misexpression of Pitxl in the chick wing bud leads to increased expression of Hoxc10 and Hoxc111 and results in transformation into a more leglike structure (Logan and Tabin 1999). The permanent up-regulation of Pitxl, Sixl, and Nkx3-1 in the oviduct of Gen-treated mice, combined with the importance of these genes in development of many embryonic tissues, suggests that one or all of these homeobox transcription factors may contribute to oviduct posteriorization after Gen treatment. Because Sixl causes initiation and metastasis of several types of human cancer (Christensen et al. 2008), its permanent up-regulation could contribute to the observed increase in FRT cancers in neonatal Gen-treated mice (Newbold et al. 2001).
Infants fed soy-based formulas have high circulating levels of Gen (1.5-13.2 [micro]M) (Cao et al. 2009). The subcutaneous injection model of Gen exposure used in the present study produces serum circulating levels similar to those of human infants ([C.sub.max] = 6.8 [micro] in female pups). Because the primary exposure of humans is oral exposure to the glycosylated form, genistin, we previously compared a dose response of oral exposure to genistin with that of subcutaneous exposure to Gen in neonatal mice (Jefferson et al. 2009a). Oral exposure to genistin in neonatal mice on PND1-PND5 produces similar circulating levels of Gen as a subcutaneous injection (- 80Percent) and similar effects on the developing reproductive tract, including multioocyte follicles in the ovary, abnormal estrous cycles, and reduced fertility. All of these Gen effects are correlated with the circulating level of Gen and the estrogenic activity (uterine wet weight increase) during the time of treatment and are not dependent on the route of administration or form of Gen given. Therefore, the subcutaneous model used in the present study is a valid model for human infant exposures to Gen.
The findings reported here raise the possibility that exposure to low levels of estrogenic environmental chemicals or phytoestrogens during sensitive developmental windows can alter developmental programming of FRT tissues. These changes could contribute to improper function of the FRT later in life that manifests as an infertility phenotype. Should this phenomenon occur 1n humans, it would provide the mechanistic basis for some cases of unexplained infertility, which is the sole infertility diagnosis in a significant percentage of couples who undergo assisted reproduction treatment by in vitro fertilization. In addition, the permanent up-regulation of developmental transcription factors in the FRT could have long-term effects on its structure and function, including influencing the development of benign or malignant pathologies.
Benson GV, Lim H, Paria BC, Satokata I. Dey SK, Maas RL. 1996. Mechanisms of reduced fertility in Hoxa-10 mutant mice: uterine homeosis and loss of maternal Hoxa-10 expression. Development 122(9):2687-2696.
Bernbaum JC, Umbach DM, Ragan NB, Ballard JL, Archer Jl, Schmidt-Davis H, et al. 2008. Pilot studies of estrogenrelated physical findings in infants. Environ Health Perspect 116:416-420.
Block K, Kardana A, Igarashi P, Taylor HS. 2000. In utero diethylstilbestrol (DES) exposure alters Hox gene expression in the developing Miillerian system. FASEB J 14(9):1101-1108.
Cao Y, Calafat AM, Doerge DR, Umbach DM, Bernbaum JC, Twaddle NC, et al. 2009. Isoflavones in urine, saliva, and blood of infants: data from a pilot study on the estrogenic activity of soy formula. J Expo Sci Environ Epidemiol 19(2):223-234.
Chopra VS. 2010. Chromosomal organization at the level of gene complexes. Cell Mol Life Sci 68(6):977-990.
Christensen KL, Patrick AN, McCoy EL, Ford HL 2008. The six family of homeobox genes in development and cancer. Adv Cancer Res 101:93-126.
D'Aloisio AA, Baird DD, DeRoo LA, Sandler DP. 2010. Association of intrauterine and early-life exposures with diagnosis of uterine leiomyomata by 35 years of age in the sister study. Environ Health Perspect 118:375-381.
Doerge DR, Twaddle NC, Banks EP, Jefferson WN, Newbold RR. 2002. Pharmacokinetic analysis in serum of genistein administered subcutaneouslyto neonatal mice. Cancer Lett 184(11:21-27.
Franco HL, Lee KY, Rubel CA, Creighton CJ, White LD, Broaddus RR, et al. 2010. Constitutive activation of smoothened leads to female infertility and altered uterine differentiation in the mouse. Biol Reprod 82(5):991-999.
Fromental-Ramain C, Warot X, Lakkaraju S, Favier B, Haack H, Birling C, et al. 1996. Specific and redundant functions of the paralogous Hoxa-9 and Hoxd-9 genes in forelimb and axial skeleton patterning. Development 122(2):461-472.
Gendron RL, Paradis H, Hsieh-Li HM, Lee DW, Potter SS, Markoff E. 1997. Abnormal uterine stromal and glandular function associated with maternal reproductive defects in Hoxa--11 null mice. Biol Reprod 56(5):1097-1105.
Gray CA, Bartol FF, Tarleton BJ, Wiley AA, Johnson GA, Bazer FW, et al. 2001. Developmental biology of uterine glands. Biol Reprod 65(5):1311-1323.
Jefferson WN. 2010. Adult ovarian function can be affected by high levels of soy. J Nutr 140(12):2322S-2325S.
Jefferson WN, Couse JF, Padilla-Banks E, Korach KS, Newbold RR. 2002. Neonatal exposure to genistein induces estrogen receptor (ER)a expression and multioocyte follicles in the maturing mouse ovary: evidence for ER[5-mediated and nonestrogenic actions. Biol Reprod 67(4):1285-1296.
Jefferson WN, Doerge D, Padilla-Banks E, Woodling KA, Kissling GE, Newbold R. 2009a. Oral exposure to genistin, the glycosylated form of genistein, during neonatal life adversely affects the female reproductive system. Environ Health Perspect 117:1883-1889.
Jefferson WN, Padilla-Banks E, Goulding EH, Lao SP, Newbold RR, Williams CJ. 2009b. Neonatal exposure to genistein disrupts ability of female mouse reproductive tract to support preimplantation embryo development and implantation. Biol Reprod 80(31:425-431.
Jefferson WN, Padilla-Banks E, Newbold RR. 2005. Adverse effects on female development and reproduction in CD-I mice following neonatal exposure to the phytoestrogen genistein at environmentally relevant doses. Biol Reprod 73(4):798-806.
Jefferson WN, Williams CJ. 2010. Circulating levels of genistein in the neonate, apart from dose and route, predict future adverse female reproductive outcomes. Reprod Toxicol 31(3):272-279.
Jeong JW, Kwak I, Lee KY, Kim TH, Large MJ, Stewart CL, et al. 2010. Foxa2 is essential for mouse endometrial gland development and fertility. Biol Reprod 83(31:396-403.
Kurita T, Cooke PS, Cunha GR. 2001. Epithelial-stromal tissue interaction in paramesonephric (Mullerian) epithelial differentiation. Dev Biol 240(1):194-211,
Laclef C, Hamard G, Demignon J, Souil E, Houbron C, Maire P. 2003. Altered myogenesis in Sixl-deficient mice. Development 130(10)2239-2252.
Lamonerie T, Tremblay JJ, Lanctot C, Therrien M, Gauthier Y, Drouin J. 1996. Ptxl, a bicoid-related homeo box transcription factor involved in transcription of the pro-opiomelanocortin gene. Genes Dev 10(10):1284-1295.
Lanctot C, Moreau A, Chamberland M, Tremblay ML, Drouin J. 1999. Hindlimb patterning and mandible development require the Ptxl gene. Development 126(91:1805-1810.
Logan M, Tabin CJ. 1999. Role of Pitxl upstream of Tbx4 in specification of hindlimb identity. Science 283(5408); 1736-1739.
Ma L. 2009. Endocrine disruptors in female reproductive tract development and carcinogenesis. Trends Endocrinol Metab 20(7):357-363.
Ma L, Benson GV, Lim H, Dey SK, Maas RL 1998. Abdominal B (AbdB) Hoxa genes: regulation in adult uterus by estrogen and progesterone and repression in mullerian duct by the synthetic estrogen diethylstilbestrol (DES). Dev Biol 197(21:141-154.
Masse J, Watrin T, Laurent A, Deschamps S, Guerrier D, Pellerin I. 2009. The developing female genital tract: from genetics to epigenetics. Int J Dev Biol 53(2-31:411-424.
Miller C, Sassoon DA. 1998. Wnt-7a maintains appropriate uterine patterning during the development of the mouse female reproductive tract. Development 125(161:3201-3211.
Newbold RR, Banks EP, Bullock B, Jefferson WN. 2001. Uterine adenocarcinoma in mice treated neonatally with genistein. Cancer Res 61(111:4325-4328.
Newbold RR, Bullock BC, McLachlan JA. 1985. Progressive proliferative changes in the oviduct of mice following developmental exposure to diethylstilbestrol. Teratog Carcinog Mutagen 5(6):473-480.
Peeters PH, Slimani N, van der Schouw YT, Grace PB, Navarro C, Tjonneland A, et al. 2007. Variations in plasma phytoestrogen concentrations in European adults. J Nutr 137(5):1294-1300.
Pfaffl MW. 2001. A new mathematical model for relative quantification in real-time RT-PCR. Nucleic Acids Res 29(9):e45; doi:10.1093/nar/29.9.e45 [Online 14 March 2001].
Reinli K, Block G. 1996. Phytoestrogen content of foods--a compendium of literature values. Nutr Cancer 26(21:123-148.
Sciavolino PJ, Abrams EW, Yang L, Austenberg LP, Shen MM, Abate-Shen C. 1997. Tissue-specific expression of murine Nkx3.1 in the male urogenital system. Dev Dyn 209(11:127-138.
Suzuki A, Urushitani H, Sato T, Kobayashi T, Watanabe H, Ohta Y, et al. 2007. Gene expression change in the Mullerian duct of the mouse fetus exposed to diethylstilbestrol in utero. Exp Biol Med 232(4):503-514.
Szeto DP, Rodriguez-Esteban C, Ryan AK, O'Connell SM, Liu F, Kioussi C, et al. 1999. Role of the Bicoid-related homeodomain factor Pitxl in specifying hindlimb morphogenesis and pituitary development. Genes Dev 13(4):484-494.
Taylor HS, Arici A, Olive D, Igarashi P. 1998. HOXA10 is expressed in response to sex steroids at the time of implantation in the human endometrium. J Clin Invest 101(7)1379-1384.
Taylor HS, Vanden Heuvel GB, Igarashi P. 1997. A conserved Hox axis in the mouse and human female reproductive system: late establishment and persistent adult expression of the Hoxa cluster genes. Biol Reprod 57(61:1338-1345.
Tremblay JJ, Lanctot C, Orouin J. 1998. The pan-pituitary activator of transcription, Ptxl (pituitary homeobox 1), acts in synergy with SF-1 and Pitl and is an upstream regulator of the Lim-homeodomain gene Lim3/Lhx3. Mol Endocrinol 12(31:428-441.
Wakeling AE, Bowler J. 1992. ICI182,780, a new antioestrogen with clinical potential. J Steroid Biochem Mol Biol 43(1-31:173-177.
Wang X, Kruithof-de Julio M, Economides KD, Walker D, Yu H, Halili MV, et al. 2009. A luminal epithelial stem cell that is a cell of origin for prostate cancer. Nature 461(72631:495-500.
Warot X, Fromental-Ramain C, Fraulob V, Chambon P, Dolle P. 1997. Gene dosage-dependent effects of the Hoxa-13 and Hoxd13 mutations on morphogenesis of the terminal parts of the digestive and urogenital tracts. Development124(23):4781-4791.
Yamanouchi H, Umezu T, Tomooka Y. 2010. Reconstruction of oviduct and demonstration of epithelial fate determination in mice. Biol Reprod 82(31:528-533. Zhao Y, Potter SS. 2001. Functional specificity of the Hoxa13
homeobox. Development 128(16):3197-3207. Zung A, Glaser T, Kerem Z, Zadik Z. 2008. Breast development in the first 2 years of life: an association with soy-based infant formulas. J Pediatr Gastroenterol Nutr 46(2):191-195.
Wendy N. Jefferson, (1.) Elizabeth Padilla-Banks, (1.) Jazma Y. Phelps, (1.) Kevin E. Gerrish, (2.) and Carmen J. Williams (1.)
(1.) Reproductive Medicine Group, Laboratory of Reproductive and Developmental Toxicology, and (2.) Microarray Group, National Institute of Environmental Health Sciences, National Institutes of Health, Department of Health and Human Services, Research Triangle Park, North Carolina, USA
Address correspondence to C.J. Williams, NIEHS/ NIH/DHHS, P.O. Box 12233, MD E4-05, Research Triangle Park, NC 27709 USA. Telephone: (919) 541-2158. Fax: (919) 541-0696. E-mail: williamsc5[congruent to] niehs.nih.gov
Supplemental Material is available online (http://dx.doi.org/10.1289/ehp.1104018).
We thank S. Wagner and E. Ney for technical assistance, H. Kinyamu and H. Yao for critical reading of the manuscript, and J. Drouin (Institut de recherches cliniques de Montreal, Montreal, Quebec, Canada) for the Pitxl antibody.
This work was supported by the Intramural Research Program of the National Institutes of Health, National Institute of Environmental HealthSciences (z01-ES102405). The authors declare they have no actual or potentialcompeting financial interests.
Received 31 May 2011; accepted 2 August 2011.
|Printer friendly Cite/link Email Feedback|
|Author:||Jefferson, Wendy N.; Padilla-Banks, Elizabeth; Phelps, Jazma Y.; Gerrish, Kevin E.; Williams, Carmen|
|Publication:||Environmental Health Perspectives|
|Date:||Nov 1, 2011|
|Previous Article:||Benzo[a]pyrene reduces testosterone production in rat leydig cells via a direct disturbance of testicular steroidogenic machinery.|
|Next Article:||A toxicogenomic comparison of primary and photochemically altered air pollutant mixtures.|