In vivo estrogenicity of p-phenoxyphenol and p-pentyloxyphenol

p-Alkoxyphenols (AOPs) are a class of ethers that are widely used in industrial and agricultural productions and daily necessities. p-Phenoxyphenol (PhOP) and p-pentyloxyphenol (PeOP) belong to this class and have been reported to be estrogenic in vitro. However, their in vivo estrogenic activities have rarely been of concern. In this study, we performed an immature mouse uterotrophic assay and studied the estrogenic effects of these two compounds in mice. The results revealed that the uterine weights of the animals treated with PhOP significantly increased at doses of 30 and 300 mg kg-1 bw day-1 for 3 days (P < 0.05), while no significant uterotrophic effects were observed in the mice treated with PeOP. Using next-generation transcriptome sequencing (RNA-seq), we also analyzed the gene expression in the uterine tissue of mice treated with PhOP and PeOP. The observed gene regulation patterns of the PhOP- and PeOP-treated specimens were similar to those of the 17β-estradiol (E2)-treated specimens. In particular, some estrogen-responsive genes, such as the Sprr2 gene family, Apoa1, Prap1, and Ahsg, displayed a regulation trend similar to that of E2. In addition, molecule docking analysis revealed that both PhOP and PeOP could be well docked into the active site of hERα, with potential of mean force (PMF) values of − 58.68 and − 52.67 kcal mol-1 for PhOP and PeOP, respectively. The results of this study indicate that PhOP exhibits relatively strong in vivo estrogenic activity, which could be of future concern.

Scientific Reports | (2020) 10:17305 | https://doi.org/10.1038/s41598-020-73271-1 www.nature.com/scientificreports/ Notably PhOP and PeOP, are widely used in industrial and agricultural productions and daily necessities, and PhOP monomers may be released from PhOP-containing products as metabolites and degradants. Thus, because humans could become inevitably exposed to these chemicals, it is necessary to study the in vivo estrogenicity of these two compounds.

Results and discussions
Uterotrophic effects of PhOP and PeOP in immature CD-1 mice. The uterotrophic assay is a useful in vivo approach to determine the estrogenic activities of chemicals 30 and has been incorporated into the USEPA screening and testing program for EEs 31 . To better understand the PhOP and PeOP estrogenic activities, we performed an uterotrophic assay using immature CD-1 mice to identify their in vivo estrogenic potentials. There was no mortality during the treatment period. Our previous in vitro study showed that the estrogenic activities of PhOP and PeOP were about 2-3 orders of magnitude lower than that of E 2 8 . So we selected doses of 1-3 orders of magnitude higher than E 2 for uterotrophic assay. Figure 2 reveals an increase in the relative uterine weights of the CD-1 mice in all the test groups as compared to that of the mice in the vehicle control group. The relative uterine weights of the mice treated with E 2 (0.4 mg kg -1 bw day -1 ) increased to 349% that of the control (P < 0.01), indicating that the animals and experiment were reliable. In the mice treated with 30 and 300 mg kg -1 bw day -1 doses of PhOP, the relative uterine weights were significantly (P < 0.05) increased to 287 and 218% that of the control, respectively. In the groups treated with 3 mg kg -1 bw day -1 PhOP, the relative uterine weights also increased to 157% that of the control; however, no statistical significance was observed when compared to the control results (P > 0.05). In the PeOP-treated mice, the average uterine weights increased in a dose-dependent pattern and again, no statistical significance (P > 0.05) was observed when compared to the control results. The results of the uterotrophic assay indicated that PhOP exhibited marked estrogenic activity in animals at low doses, while the in vivo estrogenic activity of PeOP was weak. These observations were consistent with previously published in vitro data 7 , whereby the PhOP estrogenic activity in a recombinant yeast assay was reported to be ~ 66-fold higher than that of PeOP 7 . Several alkylphenol chemicals are known as classic environmental estrogens and have been reported to have uterotrophic effects in immature rodents. p-Nonylphenol at doses of 100 and 200 mg kg -1 bw day -1 has been reported to significantly increase the uterine weights of the treated rats 32 ; p-cyclohexylphenol and p-(phenylmethyl)phenol at dose of 200 mg kg -1 bw day -1 have significantly increased  Uterotrophic effects of p-phenoxyphenol (PhOP) and p-pentyloxyphenol (PeOP) in immature CD-1 mice. All the treated mice (n = 8) were administered three days of oral gavage from PND 21 and the relative uterine weight was calculated. Data were expressed as percentages of control treated and the error bars indicated the standard errors of mean (SEMs) from the average values; *P < 0.05; **P < 0.01.

Scientific Reports
| (2020) 10:17305 | https://doi.org/10.1038/s41598-020-73271-1 www.nature.com/scientificreports/ the uterine weights of the treated rats, respectively; and p-amylphenol at dose of 800 mg kg -1 bw day -1 have been reported to significantly increase the uterine weights of the treated rats 33 . The in vivo estrogenicity of PhOP observed in this study is similar to that of the alkylphenols.
Transcriptome analysis of mice uteri treated with PhOP and PeOP. RNA-seq is a useful tool to study the mechanisms underlying the toxicity of chemicals in vivo 34 . In this study, we used RNA-seq to analyze the transcriptome of mice uteri treated with PhOP or PeOP at dosages of 3 and 30 mg kg -1 bw day -1 . DEGs as well as DEG-enriched GO terms and KEGG pathways [35][36][37] were also studied. Figure 3A illustrates the gene regulation profiles by E 2 , PhOP, and PeOP. The most evident feature in the gene profiles was that the DEG regulation patterns in the uteri of the PhOP-and PeOP-treated mice were similar to that of the E 2 group. Out of all the DEGs, some genes that have been previously reported as estrogen-response markers were of great concern. The small proline-rich 2 (Sprr2) gene family comprises 11 genes that encode the cross-linked envelope proteins of keratinocytes. The Sprr2 genes were reported to be positively regulated by estrogen receptor-dependent path- www.nature.com/scientificreports/ ways in the mice uteri 38 . In this study, the Sprr2 family genes, including Sprr2a2, Sprr2a3, Sprr2b, and Sprr2f, were significantly up-regulated by PhOP and PeOP at the dose 30 mg kg -1 bw day -1 . Of them, Sprr2b was the most sensitive gene with fold changes of 183 in the E 2 group, 105 in the 30 mg kg -1 bw day -1 PhOP group, and 16 in the 30 mg kg -1 bw day -1 PeOP group. Apolipoprotein A I (Apoa1) encodes the major protein of plasma high density lipoprotein 39 . Study has revealed that estrogens induce Apoa1 gene expression 40 . In this study, Apoa1 was significantly up-regulated by PhOP and PeOP with fold changes of 142 in the E 2 group, 80 in the 3 mg kg -1 bw day -1 PhOP group, 85 in the 30 mg kg -1 bw day -1 PhOP group, 8 in the 3 mg kg -1 bw day -1 PeOP group, and 118 in the 30 mg kg -1 bw day -1 PeOP group. Proline-rich acidic protein 1 (Prap1) was also reported as an estrogen up-regulated gene, which is considered to play a crucial role in the maintenance of gestation 41 . In this study, Prap1 was also up-regulated by PhOP and PeOP with fold changes of 51 in the E 2 group, 2 in the 3 mg kg -1 bw day -1 PhOP group, 57 in the 30 mg kg -1 bw day -1 PhOP group, 3 in the 3 mg kg -1 bw day -1 PeOP group, and 16 in the 30 mg kg -1 bw day -1 PeOP group. In addition, alpha-2-HS-glycoprotein (Ahsg), a glycoprotein gene that was up-regulated by estrogen in women 42 , was also up-regulated in the mice uteri in this study with fold changes of 172 in the E 2 group, 103 in the 3 mg kg -1 bw day -1 PhOP group, 97 in the 30 mg kg -1 bw day -1 PhOP group, 9 in the 3 mg kg -1 bw day -1 PeOP group, and 159 in the 30 mg kg -1 bw day -1 PeOP group. These expression data of the estrogen-response marker genes indicated that both PhOP and PeOP displayed estrogenic activities at the transcriptome level. The regulation patterns of more estrogen-responsive marker genes are presented in Fig. 3B. The Venn diagram of the gene overlap shows the comparison of the DEG numbers in the uteri of E 2 -, PhOP-, and PeOP-treated mice (Fig. 3C): 708 DEGs in E 2 group, 39 and 687 DEGs in 3 and 30 mg kg -1 bw day -1 PhOP groups, and 62 and 133 DEGs in the 3 and 30 mg kg -1 bw day -1 PeOP groups, respectively. In the GO enrichment analysis, DEGs were annotated to the GO database and enriched in three main categories: biological processes (BP), cellular components (CC), and molecular functions (MF) (Fig. 4). DEGs in the 0.4 mg kg -1 bw day -1 E 2 group significantly enriched 304 terms in the BP, 43 terms in the CC, and 78 terms in the MF. DEGs in the 3 mg kg -1 bw day -1 PhOP group significantly enriched 41 terms in the BP, nine terms in the CC, and 17 terms in the MF. DEGs in the 30 mg kg -1 bw day -1 PhOP group significantly enriched 263 terms in the BP, 63 terms in the CC, and 87 terms in the MF. DEGs in the 3 mg kg -1 bw day -1 PeOP group significantly enriched 11 terms in the BP, three terms in the CC, and seven terms in the MF. DEGs in the 30 mg kg -1 bw day -1 PeOP group significantly enriched 52 terms in the BP, 14 terms in the CC, and 28 terms in the MF. Notably, the "response to estrogen" and the "response to estradiol" in the BP category were enriched by 0.4 mg kg -1 bw day -1 E 2 group, 3 and 30 mg kg -1 bw day -1 PhOP groups, and 30 mg kg -1 bw day -1 PeOP groups. The "estrogen receptor binding" and the "estradiol 17-beta-dehydrogenase activity" in the MF category were enriched by 0.4 mg kg -1 bw day -1 E 2 group and 30 mg kg -1 bw day -1 PhOP group.
KEGG pathway analysis was also performed to determine the pathways in which the DEGs were significantly enriched (P < 0.05). DEGs in the 0.4 mg kg -1 bw day -1 E 2 group were significantly enriched in 43 pathways including "drug metabolism-cytochrome P450", "proteoglycans in cancer", "estrogen signaling pathway", and "endometrial cancer". DEGs in the 3 mg kg -1 bw day -1 PhOP group were significantly enriched in "PPAR signaling pathway", "fat digestion and absorption", "cytokine-cytokine receptor interaction", and "asthma". DEGs in the 30 mg kg -1 bw day -1 PhOP group were significantly enriched in 41 pathways, most of which were consistent with those of the E 2 group. DEGs in the 3 mg kg -1 bw day -1 PeOP group were significantly enriched in five pathways, including "type I diabetes mellitus" and "autoimmune thyroid disease". DEGs in the 30 mg kg -1 bw day -1 PeOP group were significantly enriched in 11 pathways, including "complement and coagulation cascades", "intestinal immune network for IgA production", and "PPAR signaling pathway". Figure 5 illustrates the top pathways with the smallest p-values that the DEGs were significantly enriched in. Only 0.4 mg kg -1 bw day -1 E 2 -and 30 mg kg -1 bw day -1 PhOP-treated groups were enriched in "estrogen signaling pathway".
To validate the RNA-seq results, expressions of 10 estrogen-responsive genes were measured by RT-qPCR. The results demonstrated good consistency between the RNA-seq and RT-qPCR data (Fig. 6).  Fig. 7. The PMF values of PhOP and PeOP were − 58.68 and − 52.67 kcal mol -1 , respectively, indicating that both chemicals could fit well into hERα. The main interactions between the chemicals and hERα active sites were hydrogen bonds and hydrophobic interactions. The phenolic hydroxyl groups in PhOP and PeOP formed hydrogen bonds with the side chains of Glu353 and Arg349 in the active pocket, while their hydrophobic groups formed hydrophobic interactions with the hydrophobic amino acids (Leu346, Leu387, Leu391, Phe404, Met 421, Ile424, and Leu525) in the active pocket (Fig. 7A,B). Figure 7C

Conclusions
In conclusion, PhOP presented marked estrogenic activity in vivo. It significantly increased the uterine weights of immature mice at doses of 30 and 300 mg kg -1 bw day -1 and induced the expression of the estrogen-response genes in mice uteri. In contrast, PeOP displayed a weak uterotrophic effect; however, it significantly affected the expression of the estrogen-response genes in mice uteri.

Materials and methods
Chemicals. PeOP (> 97%), PhOP (> 98%), and E 2 (> 98%) were purchased from Aladdin Reagent Co., Ltd. After acclimatization, the mice were assigned to treatment and control groups using a completely randomized design (n = 8). All the animals were kept in identical environments with continuous food and water availability. PhOP and PeOP were dissolved in peanut oil (vehicle) to prepare dosages of 3, 30, and 300 mg kg -1 bw day -1 , respectively. Mice treated with peanut oil only were used as the peanut oil vehicle control, while those treated with 0.4 mg kg -1 bw day -1 E 2 43,46 were used as the positive control. The mice were treated for three days by oral gavage from PND 21. The mice in each cage were marked by shaving hairs on different parts of the body, and each group was treated by oral gavage in turn, where by in each turn only one mouse in each group was treated. Each group of mice was then weighed and sacrificed by cervical dislocation after 24 h from the last treatment. The uteri were isolated, weighed, and frozen in liquid nitrogen until use for next-generation transcriptome sequencing (RNA-seq). The relative uterine weight, represented as the uterine-to-final-body weight (bw) ratio, was determined to evaluate the uterotrophic activities of the chemicals. To reduce the risk of bias in the data collection, three researchers were responsible for different experimental steps: the first researcher was responsible for catching mice and recording data; the second was responsible for body weighing, killing by cervical dislocation, and dissecting the abdominal cavity; the third was responsible for cutting, blotting, and weighting of the uteri. The latter two researchers were blind to the group of the mice. www.nature.com/scientificreports/ Next-generation sequencing-based transcriptome analysis. The uteri of the immature CD-1 mice from the 0.4 mg kg -1 bw day -1 E 2 , 3 mg kg -1 bw day -1 PhOP, 30 mg kg -1 bw day -1 PhOP, 3 mg kg -1 bw day -1 PeOP, 30 mg kg -1 bw day -1 PeOP, and peanut oil vehicle control groups were selected for RNA-seq. The total RNA was extracted from the uteri of each group (n = 3) using TRIzol reagent (Invitrogen, Carlsbad, CA, USA) and the concentrations were measured using a NanoDrop 2000 spectrophotometer (Thermo Scientific, Waltham, MA, USA). RNA contamination and degradation were estimated by agarose gelation gel electrophoresis, the integrity of which was checked using an Agilent 2100 bioanalyzer (Agilent Technologies, Inc., Santa Clara, CA, USA). The Hiseq-PE150 sequencing platform was used to construct a raw data library with a data volume of 6G. The total RNA from each pool of samples (a mixture from three animals) was used for high-throughput sequencing according to the Illumina transcriptome sequencing method (Illumina, San Diego, CA, USA; Beijing Novogene Co., Ltd.). RNA-seq analysis was performed by CLC Genomics Workbench 12 with reference gene model annotations (Mus musculus GRCm38. p6). The number of readings per kilobase length from a gene per million readings

Real-time quantitative reverse transcriptase PCR (RT-qPCR).
RT-qPCR with SYBR green fluorescence detection was performed to verify the gene expression using an Agilent Mx3005P real-time PCR machine (Agilent Technologies). Ten estrogen responsive genes were selected for validation. The primers were designed using Primer Express Software v3.0 (Applied Biosystems, USA) and are listed in Table 1. Actin beta (Actb) was used as an endogenous control to normalize the data. The 2 −ΔΔCt method was used to determine the relative gene expression levels 48 .

Molecular docking.
Scigress software (Ultra Version 3.4.0; Fujitsu Ltd., Tokyo, Japan) was used for in silico molecular docking analysis according to previously reported methods (Zhang et al. 2017). The threedimensional (3D) protein structure of the hERα LBD in complex with TIF2 NRBox3 (PDB ID: 1GWR) was downloaded from the Protein Data Bank (PDB, https ://www.rcsb.org/pdb). The 3D structure was reduced to a monomeric structure. Water and small molecules (except those that were considered important to the receptor pocket) were deleted and the refined protein structure was preserved. The molecular structures with PhOP and PeOP were then drawn, hydrogenated, and energy optimized by the PM3 mode of the Scigress-integrated procedures. The hERα active site was selected for molecular docking with PhOP and PeOP, whereby the ligand was set to be flexible and the active site was set to be rigid. The grid was evaluated using a 15 × 15 × 15 Å grid frame with the grid spacing 0.375 Å. The program was set to 60,000 generations with an initial population size of 50, accuracy of 5, a crossover rate of 0.8, and a mutation rate of 0.2. The potential of mean force (PMF), a knowledgebased approach that extracts pairwise atomic potentials from the structure information of the receptor-ligand complexes, was used to evaluate the binding potential of the chemicals with hERα 49 .
Data analysis. The data were analyzed using the statistical program SPSS (v.18.0; Chicago, IL, USA) and presented as means and standard errors of mean (SEMs) unless otherwise indicated. The group differences were evaluated using one-way analysis of variance and Fisher's least significant difference tests, with P < 0.05 considered of statistical significance (Supplementary information).  TGA CCC AGA TCA TGT TTG AGA  CAC AGC CTG GAT GGC TAC GT   Sprr2b  GTG TCC ACC CAA GAA TAA ATGAG  AGG ACA GGC GTT CAA AGG AG   Sprr2a2  GGT CAC TGC TGT TTC ATT TCCT  ATT AGA CCA TCA CCA AAG GGG   Sprr2a3  TCT CCA CCC TTC ATC CTC CAT  GAG ATC AGC CTG AGA GCA ATGC   Sprr2f  ATG GGT CTT GTT CCA TTG TTCA  AAC AGT AAC AAC TAC CCT GCT CAA G   Car2  TCT GCT CTG CCC CAA TCA C  TCT GGT CCG TTG TGC TTG  www.nature.com/scientificreports/