|Home | About | Journals | Submit | Contact Us | Français|
Environmental endocrine disruptors (EDs) are synthetic chemicals that resemble natural hormones and are known to cause epigenetic perturbations. EDs have profound effects on development and fertility. Imprinted genes had been identified as candidate susceptibility loci to environmental insults because they are functionally haploid, and because the imprints undergo epigenetic resetting between generations. To screen for possible epigenetic perturbations caused by EDs at imprinted loci, we treated pregnant mice daily between 8.5 and 12.5 days post coitum (dpc) with di-(2-ethylhexyl)-phthalate (DEHP), bisphenol A (BPA), vinclozolin (VZ) or control oil vehicle. After isolating RNA from the placenta, yolk sac, amnion, head, body, heart, liver, lung, stomach and intestines of 13.5 dpc embryos we measured the allele-specific expression of 39 imprinted transcripts using multiplex single nucleotide primer extension (SNuPE) assays. In this representative data set we identified only a small number of transcripts that exhibited a substantial relaxation of imprinted expression with statistical significance: Slc22a18 with 10% relaxation in the embryo after BPA treatment; Rtl1as with 11 and 16% relaxation in the lung and placenta, respectively after BPA treatment; and Rtl1 with 12% relaxation in the yolk sac after DEHP treatment. Additionally, the standard deviation of allele-specificity increased in various organs after ED treatment for several transcripts including Igf2r, Rasgrf1, Usp29, Slc38a4 and Xist. Our data suggest that the maintenance of strongly biased monoallelic expression of imprinted genes is generally insensitive to EDs in the 13.5 dpc embryo and extra-embryonic organs, but is not immune to those effects.
There are a great number of synthetic chemicals in the environment with structural and functional resemblance to natural hormones, and many of them are present in the body fluids of humans. Hormone-like chemicals, such as bisphenol A (BPA), di-(2-ethylhexyl)-phthalate (DEHP) and vinclozolin (VZ) can mimic the role of the hormone they resemble (Fig. 1). It is becoming clear that these chemicals are not harmless, but can perturb development and cause epigenetic aberrations.1,2
BPA (Fig. 1) is used as the monomer to manufacture polycarbonate. This chemical is produced at a worldwide capacity of over six billion pounds per year. Polycarbonate plastics are used as linings in most food and beverage cans, in baby bottles, and are components of many more commonly used consumer products. However, the polymer undergoes hydrolysis of the ester bond linkages releasing the monomer, in particular when the material is exposed to heat and/or to acidic or basic conditions (e.g., the presence of acidic or basic food in cans). Exposure of human populations to BPA is wide-spread as reported by Calafat et al. who have shown the presence of BPA in >95% of human urine samples tested.3 A number of recent reports in the literature have focused on low-dose reproductive effects of BPA.4 Ninety-four of 115 published studies have shown a significant low dose estrogenic activity of BPA. While none of the chemical industry-supported studies reported an effect of low dose BPA, more than 90% of government-supported studies reported such an effect.4 The effects of BPA, many of them observed after exposure during gestation or early development, include increased postnatal growth in both males and females,5,6 early onset of sexual maturation in females,6,7 increased prostate size in male offspring,8,9 reduced sperm production,10 stimulation of mammary gland development11 and altered immune and behavioral functions.4 Importantly, median levels of BPA found in human serum are higher than those causing adverse effects in mice.12 While much of the focus has been on the estrogenic activity of BPA,13 the chemical also has the potential to disrupt thyroid hormone action14 and acts as an anti-androgen.15 The reasons why low doses of BPA elicit effects not seen in the earlier high dose studies may be related to the phenomenon that low doses of hormones can stimulate a response whereas much higher doses inhibit the same response.16
Since the 1930s, plasticizers have been used to impart flexibility to polyvinylchloride (PVC). DEHP (Fig. 1) is the most widely used plasticizer present in PVC products, including PVC toys, vinyl flooring and drinking water ducts. However, DEHP leaks out from PVC over time and during use and consequently it is a ubiquitous environmental contaminant. Phthalate esters are also used in shampoos, soaps, cosmetics and paints. Phthalate metabolites have been detected in human urine at levels of over 3 micrograms per milliliter of urine,17,18 indicating significant exposure of human populations to phthalate esters. DEHP is considered to be an ED based on its hormone-like activity. DEHP is a carcinogen, which induces tumors in the liver and testes of rodents.19,20 DEHP, through its active metabolite mono-ethylhexyl phthalate (MEHP), exerts toxicity on the female reproductive system.21 In female rodents, it causes decreased serum estradiol levels, suppression of aromatase, and it activates peroxisome proliferator-activated receptors (PPARs).22 In male rat embryos, DEHP decreases testosterone synthesis to female levels during sexual differentiation; consequently, male fetuses have reduced anogenital distance.23 Phthalates act as anti-androgens and affect the development of the male reproductive system after in utero exposure.24,25 Although data on human developmental toxicity are limited, a recent report has shown a similar anti-androgenic effect in baby boys born to mothers who were exposed during pregnancy to higher levels of phthalate metabolites.26
VZ (Fig. 1) is a fungicide used in the wine-growing industry and in vegetable and fruit production. Since washing of produce is inefficient to remove the compound, humans are constantly exposed to VZ. VZ has anti-androgenic activity. It is biotransformed into at least two potent active metabolites that bind competitively to androgen receptors.27 Adult male rats have reduced anogenital distances, reduced seminal vesicle and ventral prostate weights and lower sperm counts after embryonic or perinatal exposure to VZ.28–30 When pregnant rats were exposed to VZ at the time of mid-gestation,1 this transient exposure caused an adult phenotype in male mice characterized by decreased fertility, low sperm count and sperm motility, and increased apoptosis in spermatogenic cells. Surprisingly, each generation of subsequently bred animals (up to four generations were tested) had the same disease state. The implication of this finding is that hormone-like chemicals could be causing population-wide heritable reproductive problems such as low sperm counts in men. This finding suggested a new paradigm for permanent reprogramming of the germ line. Although the mechanisms of the trans-generational effects of VZ have not yet been determined, an epigenetic effect was suspected. This epigenetic aberration would need to be stably transmitted to subsequent generations and resist epigenetic reprogramming during germ cell and pre-implantation development. Although it is thought that most epigenetic marks are erased during gametogenesis and early embryogenesis, some epigenetic modifications may not be completely erased.31,32 Indeed, transgenerational epigenetic inheritance was detected in the form of DNA methylation after in utero VZ treatment at a set of promoters and at some differentially methylated regions (DMRs) of imprinted genes.33,34
Imprinted genes of mammals are expressed from the maternally or paternally inherited chromosomes at least in one somatic or extra-embryonic organ.35,36 Correct allele-specific expression of imprinted genes in the embryo and in the placenta is essential for normal in utero development.37,38 The monoallelic expression of imprinted genes depends on DNA methylation and chromatin composition.39–41 DNA methylation marks at DMRs that originate in the sperm or oocyte are called gametic imprints42 and are critical for the allele-specific monoallelic expression of imprinted genes.43–47 The parental-specific methylation at DMRs is maintained in somatic cells during the life of the individual, but is reset (erased and reestablished) in the germ line between generations.48 Part of this imprint resetting occurs in the embryonic and fetal germ cells, which undergo global epigenetic remodeling.49 Because imprinted genes are functionally haploid, a single epigenetic hit by EDs could have serious consequences to development and health.50 Imprinted genes are therefore considered candidate susceptibility loci for environmentally induced diseases.51 Hormones are known regulators of transcription through their effect on DNA methylation and chromatin composition.52,53 Since EDs can affect DNA methylation patterns33,54,55 and gene expression,56–60 it is hypothesized that these hormone-like chemicals can alter the reciprocal epigenetic states of the two parental chromosomes at imprinted genes or their DMRs and those epigenetic changes would alter the allele-specific expression in the embryonic or extraembryonic organs. To test this hypothesis we generated multiplex allele-specific assays and measured the effects of BPA, VZ and DEHP on the maintenance of imprinted gene expression of a representative set of imprinted transcripts in the embryo and in extraembryonic organs after in utero exposure to these chemicals.
To assay for allele-specific expression of a representative number of imprinted genes we developed multiplex SNuPE assays on the Sequenom allelotyping platform. We included maternally and paternally expressed transcripts, aome with ubiquitous and others with organ-specific imprinted expression (www.har.mrc.ac.uk/research/genomicimprinting/maps.html and igc.otago.ac.nz/home.html). We utilized 41 SNPs between two mouse lines, JF1 and OG2,61 in multiplex assays. This allowed us to precisely quantify 38 transcripts, three with duplicate SNPs. We tested the validity of the three assays using a dilution set, where the OG2 allele's portion increased in the total alleles (Sup. Fig. 1). The measured portion of the OG2 allele showed linear response to the input portion in all of these assays. The four replicates in the dilution set verified reproducibility with an average standard deviation of 2.3% across 41 assays. The allele-specific expression of three transcripts, Asb4, Zac1 and Peg3, was confirmed using two independent SNPs (Sup. Fig. 2).
Animals were fed for several generations before the study and during the study with special verified casein diet I 1F (5K96) according to NIH 1996 recommendations to help unmask the effect of EDs from the variable levels of phytoestrogen compounds found in regular mouse food. Mice were kept in polypropylene cages to avoid the potential masking effect of ED contamination from polycarbonate cages.62 JF1 females were mated with OG2 males and oil vehicle was administered by gavage to the pregnant dams daily between 8.5 and 12.5 days post coitum (dpc). The particular cross of JF1XOG2 allows the allele-specific analysis of imprinted genes in the soma because SNPs exist between the two lines at many imprinted loci. The percent parental expression of 38 imprinted loci was measured in the oil-vehicle treated embryo samples at 13.5 dpc. Female embryos were analyzed to reveal potential effect of EDs on imprinted X chromosome inactivation in the extra-embryonic samples. Three biological replicates of head, embryo body (without internal organs), heart, liver, lung, stomach, intestines, placenta, amnion and yolk sac were analyzed. cDNA was prepared from total RNA using random hexamer primers and was submitted to PCR-single nucleotide primer extension (SNuPE) assays at locus specific SNPs. When designing the assays, care was taken to avoid regions where oppositely imprinted sense and antisense transcripts overlap at the Airn/Igf2r, Nesp/Nespas, Kcnq1ot1/Kcnq1 and Xist/Tsix transcripts, to make sure that each SNP will only detect one specific transcript. The overlap could not be avoided at the Rtl1/Rtl1as63 transcripts. This pair, therefore, is not resolved in experiments involving cDNA prepared using random hexamers. The average expression from the parental alleles was plotted as percent of total (maternal plus paternal or 100%) expression with standard deviations (Figs. 2–4). Ubiquitous maternal allele-specific expression was found for Asb4, Gtl2, H19, Igf2r, Phlda2, Slc22a18 and Zim1 (Fig. 2). Nesp and Rtl1/Rtl1as exhibited maternal allele-specific expression in each sample except in heart and placenta, respectively (Fig. 2). Placenta-specific maternal allele-specific expression was detected for Ampd3, Dcn and Osbpl5 (Fig. 3). Ascl2, Hprt1, Slc22a3, Tnfrs22 and Tsix exhibited maternal-allele-specific expression additionally in the yolk sac or amnion or both. Tsix has very high standard errors in every other organ. This is most likely due to the extremely low level of Tsix expression in those organs. No clear allele-specificity but a maternally biased expression was detected in placenta for Casd1 and Dhcr7 and in the amnion for Atp10a. Neurabin was maternally biased in most samples, especially in the yolk sac, Htra2 in the yolk sac and Cobl1 in the heart. No clear maternal bias was found for Pon3 at this developmental stage. Airn, Dlk1, Impact, Kcnq1ot1, Nespas, Peg3, Snrpn, Usp29 and Zac1 were exclusively expressed from the paternal allele or their expression was strongly biased toward the paternal allele in each sample (Fig. 4). The imprinted expression of some transcripts was relaxed in certain samples: Impact was slightly relaxed in the placenta, Ins1 in the stomach, and Slc38a4 in the yolk sac. Other transcripts exhibited organ-specific paternal allele-specific expression: Inpp5f_v2 was paternally expressed in the head, Rasgrf1 in the head, lung and stomach and was paternally biased in the carcass, heart, intestines and the amnion. Xist was paternally expressed in the placenta and yolk sac and was biallelically expressed in the other samples.
Embryos were exposed in utero to oil vehicle or one of the three different ED chemicals, vinclozolin (VZ) at 100 mg/kg/day, di-(2-ethylhexyl) phthalate (DEHP) at 750 mg/kg/day or bisphenol A (BPA) at 0.2 mg/kg/day. The selected doses had biological effects in rodents in previous studies. DEHP (750 mg/kg/day) reduced the anogenital distance in male rat offspring;23,24 VZ (100 mg/kg given as intra-peritoneal injections) has resulted in a trans-generational phenotype of reduced spermatogenetic capacity and sperm mobility1 and BPA at a ten-fold lower dose (0.02 mg/kg/day orally) has resulted in decreased sperm production after fetal exposure.10 Treatment was administered to pregnant dams daily by gavage between 8.5 and 12.5 dpc. Embryos were dissected at 13.5 dpc. Three biological replicates were used. The allele-specific expression was compared between ED-and vehicle treated samples for each of the transcripts. In this initial screen a total of 120 samples were analyzed at 41 SNPs each, resulting in 4,920 data points.
The effect of EDs on the allele-specific expression of selected maternally and paternally expressed imprinted transcripts are shown in Figures 5 and and66, respectively. The maternal bias of Slc22a18 decreased slightly but statistically significantly (89-79% MAT) in the BPA-treated embryo compared to vehicle-treated embryo (Fig. 5A). Igf2r was significantly but only slightly less maternal allele-specific (96-91% MAT) in DEHP-treated versus vehicle-treated yolk sac (Fig. 5B). VZ resulted in an increased standard deviation of Igf2r alleles in the head and the embryo. Osbpl5 became 11% more biased in the DEHP-treated the yolk sac, but in this organ a base line imprinted expression was not present (Fig. 5C). Rasgrf1 remained paternal allele-specific in the ED-treated head, lung and stomach, where it was strictly monoallelic in the vehicle-treated sample, but became slightly but significantly maternally biased (55–67% MAT) in the DEHP-treated placenta, where it had biallelic base-line expression (Fig. 6A). Impact became significantly more paternal allele-specific (91–98% PAT) in the DEHP-treated intestines, but the change was very small (Fig. 6B). For comparison, average and range of standard deviations in the dilution samples (Sup. Fig. 1) were as follows: Slc22a18 2.1 (0–4.2); Igf2r 1.4 (0–2.7); Osbpl5 1.15 (0–1.84); Rasgrf1 3.1 (0–6.6) and Impact 1.91 (0–3.83).
We noticed that for Rasgrf1 the standard deviation increased in the DEHP-treated liver and for each ED-treated amnion samples (Fig. 6A). Similarly, standard deviations increased for Usp29 in the BPA-treated placenta samples (Sup. Fig. 3A), and it was also found in the regular diet-fed animals (data not shown). Standard deviation also increased for Slc38a4 in several organs in the ED-treated samples (Sup. Fig. 3B) and for Xist in most ED-treated organs exhibiting biallelic base-line expression, but not in the placenta and yolk sac exhibiting strong base-line paternal bias (Sup. Fig. 3C).
Changes in allele-specific expression for 38 transcripts that were greater than 5% and also occurred with statistical significance (p value < 0.05) were tabulated (Table 1). In general, the larger changes occurred with less statistical significance and the highest significance accompanied smaller changes (Cobl1 in heart after VZ, Zim1 in heart after BPA, and Tnfrs22 in liver after BPA). These small changes 6–8% with (p < 0.01) may not be significant biologically, because the range of standard deviation in the dilution experiment (Sup. Fig. 1) was as follows: Cobl1 1.36 (0–2.5); Zim1 2.4 (0–5.3); Tnfrsf22 5 (0–12.3).
There were no tabulated changes for Gtl2, H19, Asb4, Phlda2, Dlk1 (Fig. 6C), Airn, Zac1, Ins1, Peg3, Snrpn, Kcnq1ot1, Nespas, Usp29, Hprt1, Ascl2, Slc22a3 Casd1 Neurabin and Dcn. The majority of these transcripts are those that exhibit very strongly biased ubiquitous imprinted expression in vehicle-treated samples (Figs. 2–4). Most tabulated changes occurred at transcripts with organ-specific imprinted expression in organs where there was no imprinted expression in the vehicle treated sample. Only a few of the tabulated changes occurred in organs where there was a base line maternal or paternal allele-specific expression. These results suggest that monoallelic expression of imprinted genes is generally very stable in the soma at 13.5 dpc.
Of those transcripts (Table 1) with base line imprinted expression, the expression became either more biased towards one parental allele or more relaxed. Nesp became more maternally biased in stomach after DEHP (82–100%), Zim1 became more maternally biased in heart after BPA (95–100%), and Impact became more paternally biased in the intestines after DEHP treatment (91 to 98% paternal allele). In comparison, the average and range of standard deviations of the dilution samples were as follows: Nesp 4.27 (0–8.2); Zim 2.38 (0–5.25); Impact 1.91 (0–3.83). The allele-specific change in Nesp expression in the heart (Fig. 7A) was accompanied by a highly variable increase in the level of Nesp expression (Fig. 7B). The increase occurred in the maternal allele, suggesting that some epigenetic marking must be present for monoallelic expression for Nesp even in the heart that exhibits biallelic Nesp expression at 13.5 dpc. The change in the allele-specific expression was also observed in BPA-treated heart under generic conditions with regular mouse food (Fig. 7C), but these differences were not statistically significant. Normal diet may, indeed, partially mask the effects of EDs in the embryo.
Only in a few instances, imprinted expression became relaxed—or less biased—in the ED-treated samples (Table 1). Igf2r became slightly relaxed (96-91% MAT) in the yolk sac after DEHP (Fig. 5B) and Slc22a18 became slightly relaxed (89-79% MAT) in the embryo after BPA (Fig. 5A). Among all imprinted genes examined, the sense-antisense transcript pair, Rtl1/Rtl1as, combined, appeared to exhibit the greatest change of imprinted expression (Table 1). Maternal allele-specific expression was reduced in the liver, lung and intestines after BPA (79-63; 82-71 and 90-81% MAT, respectively) and in the intestines after DEHP (90-85%) (Fig. 8B). The change was also observed in BPA-treated lung and intestines when mice were kept under generic conditions (Fig. 8A). These changes may have a biological significance: the average and range of standard deviations in the dilution set were as follows: Igf2r 1.4 (0–2.7); Slc22a18 2.1 (0–4.2) and Rtl1/Rtl1as 2.4 (0–4.8). The maternal-allele-specific combined expression of Rtl1/Rtl1as suggested that the maternal allele-specific Rtl1as may be more abundant than the paternal-allele specific Rtl1 in each organ except in the placenta, where Rtl1 may be more active. To test this we prepared Rtl1 and Rtl1as cDNAs using strand-specific primers and measured the relative expression levels of the strand specific transcripts in real-time PCR (Fig. 8C and D). Placenta, indeed, exhibited low level of Rtl1as transcription compared to other organs, but Rtl1 transcript levels were high in the placenta compared to other organs. We note that apart from the placenta, Rtl1 and Rtl1as transcript levels appear to be synchronously regulated in most organs. Applying SNuPE on the Rtl1 and Rtl1as cDNA samples we found that whereas Rtl1as exhibited ubiquitous maternal allele-specific expression, Rtl1 had a complex pattern of allele-specific expression (Fig. 9). Rtl1 was expressed from the paternal allele in the placenta, from the maternal allele in the yolk sac, and was biallelically expressed in the embryo, liver, lung and intestine. We found that the EDs affected these two transcripts in a complex way (Fig. 9 and Table 2). Rtl1 became significantly more paternally biased in the liver and lung after DEHP and BPA treatment and less maternally biased in the yolk sac after DEHP treatment. The maternal allele-specific expression of Rtl1as became significantly relaxed in the lung and intestines after BPA treatment. Interestingly, the transcript levels of both Rtl1 and Rtl1as were reduced after BPA and DEHP treatment in the head, embryo, liver, lung, stomach, intestines, amnion and yolk sac compared to vehicle treated matching samples (Fig. 8C and D). These two EDs, therefore, similarly affected the synchronous regulation of Rtl1 and Rtl1as in most organs except in the placenta.
To determine possible epigenetic perturbations caused by EDs in the somatic maintenance of imprinted gene expression, we treated embryos in utero with DEHP, BPA or VZ or oil vehicle and determined the change in allele-specific expression at this representative set of imprinted loci. The vehicle-treated samples provided a detailed and comprehensive analysis of allele-specific expression of 38 imprinted loci in the normal embryo at 13.5 dpc. We analyzed the body, head, heart, liver, lung, stomach and intestines and the extra-embryonic organs: placenta, yolk sac and amnion. Changes in expression occurred predominantly in organs that exhibited biallelic expression in the control samples. EDs generally did not have a larger than 5% effect on imprinted genes in the organs where they had strong monoallelic expression. There were, however, exceptions. BPA caused a 10% relaxation in imprinted expression for Slc22a18 in the eviscerated embryo and for the sum allele-specific output of the overlapping oppositely imprinted Rtl1/Rtl1as transcripts in liver, lung and intestines. When the effects of EDs on the combined allele-specific expression of Rtl1/Rtl1as were resolved between the sense and antisense transcripts, we found that both Rtl1 and Rtl1as were affected significantly in a complex way. This imprinted locus appears to be the most sensitive to EDs among all loci tested. Even small changes at the Rtl1 locus may have biological significance, because either the deletion or the overproduction of the Rtl1 transcript results in late fetal and/or neonatal lethality in mice.64 microRNAs out of Rtl1as negatively regulate Rtl1 by targeting it to degradation.65 The target organ of Rtl1 is the placenta. In placentae a significant relaxation of Rtl1as was detected after BPA treatment. However, a corresponding significant change in Rtl1 allele-specificity (such as seen in lung) was not observed in placenta. The explanation can be that Rtl1as expression is relatively low in the placenta, and a change in its low level or in its allele-specificity is not sufficient to alter the allele-specificity of Rtl1 in the placenta.
Our data suggest that the maintenance of strongly biased monoallelic expression of imprinted genes is generally insensitive to EDs in the 13.5 dpc embryo and extra-embryonic organs. But because there were several exceptions with statistically significant changes that exceeded the inherent errors of the method, we can conclude that imprinted gene expression in the embryo at 13.5 dpc is not immune to in utero exposure to EDs. The standard deviation of allele-specific expression was often increased after ED treatment, suggesting that a slight relaxation of imprinting may occur in response to the environmental toxicants and go undetected in the lack of statistical significance. One limitation of our screening method is that even if the allele-specificity of imprinted gene expression appears normal for most imprinted genes, the overall expression level may change for certain imprinted genes in response to EDs. Our study may, therefore, underestimate the effects of EDs on imprinted genes. It will be important to systematically characterize the potential adverse effects of BPA and DEHP on embryo and placenta development. Additional compounds and different doses must be tested. It will be also interesting to see in follow up studies, how EDs affect allele-specific epigenetic marks, DNA methylation and chromatin composition at imprinted loci in the soma. Recently, chromatin changes have been detected at a metastable epiallele.66 Our experimental system will also allow the purification of embryonic and fetal germ cells and the analysis whether the erasure of imprinted expression in primordial germ cells (PGCs),61 is affected by EDs. Indeed, PGCs that undergo global remodeling may be more sensitive to the effects of EDs than the embryonic soma and extra-embryonic compartment.
The experiments involving mice had been approved by the IACUC of the City of Hope under protocol ID 91,023. Housing and care of the animals has been consistent with Public Health Service Policy, the NIH “Guide for the Care and Use of Laboratory Animals” and the Animal Welfare Act.
It has been shown that BPA can be released from used polycarbonate animal cages into water at room temperature.62 Therefore, we used polypropylene animal cages that do not release EDs. The animals received a special verified diet, 5K96 (TestDiet), recommended by NIH for animal studies involving hormone-like chemicals. Drinking water was provided in glass bottles and was purified on a carbon filter (Filter Cartridge Hi-Cap Carbon 9-3/4 ID #: 2100-1970-102 from Edstrom Direct) just upstream of the bottle filler. Additionally, other control mice were kept for comparison under generic conditions. They were housed in polysulfone cages, received regular chow (5001 Rodent Diet from LabDiet) and drank tap water.
JF1 females were mated with OG2,61 males and were checked for copulation plugs the next morning, at 0.5 days post coitum (dpc). The females were gavaged daily for five days during pregnancy starting from 8.5 dpc to 12.5 dpc according to their body weight. On 13.5 dpc the treated pregnant dams were euthanized and fetuses were collected for organ isolation. The ED chemicals used were vinclozolin (ChemService Catalog no. PS-1049; Sigma, USA), bi-(2-ethylhexyl) phthalate (Selectophore, (DEHP), Catalog no. 80030; Fluka/Sigma Inc.,), and bisphenol A (Catalog no. 239658; Sigma Aldrich Inc.). All three were delivered in tocopherol-stripped corn oil vehicle (MPI Catalog no. 0290141584). The dosages for various chemicals were VZ (100 mg/kg/day), BPA (0.2 mg/kg/day), and DEHP (750 mg/kg/day). Control animals were treated with the vehicle alone, as was done in other studies.9,24
RNA was isolated from tissues using RNA-Bee (Tel-Test). Contaminating DNA was removed with the DNA-free Kit (Ambion).
Reverse transcription for Sequenom SNuPE assays (except when noted otherwise) and for Nesp real-time PCR was performed using equal amount of RNA and random hexamer primers using the SuperScript III First-Strand Synthesis System for RT-PCR (Invitrogen). Strand specific reverse-transcription was performed at the Rtl1/Rtl1as locus using 200 ng total RNA with M-MLV Reverse Transcriptase (Invitrogen) and one of the following oligos: Rtl1as cDNA L 5′-TAC TTG TCC TGC AAC TTG AAG AAA AG-3′ and Rtl1 cDNA L 5′-GGC AAA CCT CTC ATC CAT GTA GTT-3′.
Real-time PCR quantitation of the Nesp transcript was done using Universal ProbeLibrary probe #63 (Roche) and NespU 5′-GAG GGC CCT TAG ATC AGG A-3′ and NespL 5′-GGG CTG CCT GTT TTC CTC-3′ amplification primers. Beta-actin was quantified by real-time PCR using the Universal ProbeLibrary probe control conditions and serial dilution of cDNA from the specific organ/body part. The copy numbers of Nesp transcript in each sample were determined using serial dilution of cDNA and were standardized for beta-actin copy numbers. Rtl1 and Rtl1as expression was measured using equal aliquots of strand specific cDNA using the Universal ProbeLibrary probe #71 (Roche) and amplification primers Rtl1/Rtl1asU 5′-CCC AGA GAA GTG GAG GGT AA-3′ and Rtl1/Rtl1asL 5′-CGA TCA ATG TCT GGG TGG A-3′. Copy numbers in the individual samples were determined using serial dilution of strand specific cDNA standards that were prepared from a general mix of the oil-treated samples from each organ/body part. This way the relative RNA levels between organs can be also compared.
Allele-specific gene expression analysis was based on SNPs between of inbred JF1/Ms (JF1) and TgOG2 (OG2),61 mouse strains and was analyzed by reverse-transcription PCR SNuPE assays,67,68 except mass spectrometry quantified the extension primers (EP) based on molecular mass difference between alleles69,70 as we have done previously in reference 71 and 72. Primers were designed using MassArray Assay v3.1, and are listed in Supplemental Table 1. Amplified cDNA samples were spotted onto a 384 SpectroCHIP Array. Automated spectra acquisition was performed in a MassArray Compact mass spectrometer (Sequenom) using the Spectroacquire program (Sequenom) and was analyzed by MassArray Typer v3.4. We applied skew correction using a true heterozygote DNA sample to correct for any allelic imbalance in the SNP allele products. The percent expression of each allele in the total expression was calculated at each given SNP.
We thank Daobing Wang for technical assistance, Hector Rivera for DNA sequencing, and the Animal Research Center of the City of Hope for animal care. This work was supported by NIH grant RO1ES015185 to P.E.S. The funders had no role in study design, data collection and analysis, decision to publish or preparation of the manuscript.