"The OBS (purity of 98%) was provided by Ningbo Yongshen Trading Co., Ltd. (Zhejiang, China). During the experimental period, OBS was dissolved in acetic acid, and the volume (acetic acid): the volume (drinking water) = 1:20,000 in every group.

All mouse experiments abided by Guiding Principles of Zhejiang University of Technology. The animals were treated humanely and every effort was made to minimize animal suffering. We purchased the 8-week-old C57BL/6 female (n = 30) and 10-week-old C57BL/6 male mice (n = 30) from the China National Laboratory Animal Resource Center (Shanghai, China). All mice were housed in animal facilities for 1 week prior to experimentation with a 12 hours light cycle before the beginning of experiments. During experimental period, OBS was administered through the drinking water to female mice at dietary levels of 0.0 mg/L (CON), 0.5 mg/L (OBS-L) and 5.0 mg/L (OBS-H), food and water were available ad libitum.

After 1 week, we randomly divided the female animals (F0, n = 30) into three groups (10 mice each) with similar body weights. Females were paired with males for fertilization. After detection of the vaginal plug in females, the males were removed from the cage, and females were exposed to a specific concentration of OBS until weaning of F1 generation mice. During the experiment, two 10-day-pregnant female mice were randomly selected from each group to detect whether maternal exposure resulted in OBS entry into the embryo through the placental barrier. The offspring (F1) were weaned at 3 weeks of age. Then, the animals were fed a standard diet and water. We divided the three groups of weaned mice into two groups (including a control group and two experimental groups of 8 mice each) and housed the animals in cages containing no more than three mice. Subsequently, a group of F1 generation mice was housed to 8 weeks of ages, and another group was housed to 24 weeks of ages. After weaning of F1 generation mice, the F0 mice (n = 8/group) were sacrificed and examined. The F1 generation (n = 8/group) was sacrificed at 8 and 24 weeks of age (F1-8w and F1-24w, respectively). All blood samples were collected retro-orbitally as reported previously [37], and the testes, ovaries and uterus were rapidly removed and weighed. All tissues were rapidly frozen in liquid nitrogen and stored at –80°C until use.

To detect the concentration of OBS from the testis (male) and uterus (female), the sample was spiked with an internal standard (potassium L-PFOS; 10 ng) first, and then ultrasonically extracted with 500 μL of acetonitrile for 5 minutes. To remove lipids, we selected approximately 450 ul of extract to pass through an Oasis PRiME HLB column (1 cc/30 mg, Waters, MA). After processing the collected samples, we performed ultra performance liquid chromatography-tandem mass spectrometry (UPLC-MS/MS). We used an Agilent 1290 Infinity Series and Triple Quadruple 6420 from Agilent Technology (Santa Clara, CA, USA) to quantitative analysis OBS. Quantification was performed by the internal standard approach. And the average recovery rate of OBS was between 95.8% and 116.0% for tissues.

A testicular, randomly selected 3 mice from F1 generation 8 male mice in each group, was fixed directly in 4% formaldehyde or Bouin’s fixative for pathological observation. The samples were embedded in paraffin and cut into 5 μm thick, stained with hematoxylin and eosin (H&E). According to previous studies, we analyzed the spermatogenesis in the testicular sections [38-41]. Immunohistochemical analysis in the testis- The testis of three mice from each group were randomly selected for immunohistochemical analysis, and paraformaldehyde-fixed testis tissues were cut into 4- μm-thick sections and embedded in paraffin. After deparaffination with xylene and hydration through a graded ethanol series, the slices were incubated with a peroxidase–blocking solution for 10 min. Antigen retrieval was performed according to the manufacturer’s instructions. The optimal effect dilution for the poly‐ clonal ERα (GB13205, Servicebio) and 17β-HSD (Bs-3905R, Bioss) antibody was determined to be 1:200 and 1:300. Place the slices flat in a wet box and incubate overnight at 4°C. Then, incubate the slices with the secondary antibody (1:200) of the corresponding species for 50 min at room temperature. The slices were developed by DAB color liquid. Finally, sections were counter‐ stained with hematoxylin and Alcian blue, dehydrated and mounted for microscope observation [42].

All reagents were prepared before starting assay procedure. The serum TC, HDL-C and LDL-C were detected according to the manufacturer’s instructions. Then, the level of estrogen (CK-E20462M, Biosamite) and androgens (CK-E92576M, Biosamite) in the serum of each mouse were detected using commercial enzyme-linked immunosorbent assay (ELISA) kits (Jijin Chemical Technology Co., Ltd., Shanghai, China) according to the corresponding instructions. The optical density (O.D.) at 450 nm was measured on a microplate reader (Power wave XS, Bio-TEK, USA) by using the specific 96-well microplate. Among them, the standard curve was generated by plotting the average O.D. (450 nm) obtained for each of the six standard concentrations. The standard curve was used to determine the hormone levels in an unknown sample.

Testicular tissues were washed with PBS, lysed with RIPA lysis buffer (Merck, USA). Protein concentrations were quantified using BCA Kits according to the manuscript’s instructions. Tissue lysates were run on SDS/PAGE and transferred to PVDF membranes (Millipore, USA). Before immunoblotting, PVDF membrane were blocked with TBS-T containing non-fat milk (5%) for 1 h and then blotted with an antibody against ERα (GB111843, Servicebio) and 17β-HSD (Bs-3905R, Bioss) overnight at 4°C. Among, the optimal effect dilution for the antibodies ERα and 17β-HSD in western blotting was determined to be 1:500 according to the corresponding instructions. After extensive washing with TBS-T, membranes were further incubated with horse‐ radish peroxidase (HRP)-labeled secondary antibody (1:10,000) for 1 h at room temperature. Finally, mem‐ branes were incubated with ECL detection kits and exposed with Luminescent Imaging Workstation (Tanon, China). The relative protein content was quantified by ImageJ.

Total RNA was harvested from tissues (testis, ovary) in BioZOL reagent. The concentration of total RNA was measured by absorbance at 260 nm using an ultraviolet-visible spectrophotometer (Bio-TEK, USA). The cDNA preparation from 500 ng of RNA was performed using qScript first-strand cDNA synthesis kit (Vazyme, China). RT-qPCR was performed using SYBR Green (Vazyme, China) on an Eppendorf MasterCycler ep RealPlex2 thermocycler machine (Wesseling-Berzdorf, Germany). The quantification of relative gene expression was accomplished as previously described [43, 44]. The sequences of the primers of genes related to estrogen and androgen biosynthesis, steroidogenesis, cholesterol transport in mice were used according to previous reports [41, 45] and the levels of 18s RNA (Forward: 5'-CGAACGT CTGCCCTATCAACTT-3'; Reverse: 5'-CCGGAATCGA ACCCTGATT-3') transcripts were determined, which served as a housekeeping gene [46].

Data were analyzed using GraphPad Prism version 7 (GraphPad Software) and SPSS 13.0 (SPSS, Chicago, Illinois) between the control group and the treated group by one-way ANOVA followed by Dunnett’s test, two-way ANOVA followed by Tukey’s test, and Kruskal-Wallis test. They were evaluated as the mean ± standard error of the mean (SEM). The results were considered significantly changed when p-value ≤.05."

F1 Male C57BL/6 Mice