Prenatal Dexamethasone Exposure Induced Ovarian Developmental Toxicity and Transgenerational Effect in Rat Offspring

Prenatal dexamethasone exposure (PDE) induces multiorgan developmental toxicities in offspring. Here we verified the transgenerational inheritance effect of ovarian developmental toxicity by PDE and explored its intrauterine programming mechanism. Pregnant rats subcutaneously received 0.2 mg/kg/d dexamethasone from gestational day (GD) 9 to GD20. A subgroup was euthanized for fetuses on GD20, and the other group went on to spontaneous labor to produce F1 offspring. The adult F1 females were mated with normal males to produce the F2 and F3 generations. The PDE fetal rats exhibited ovarian mitochondrial structural abnormalities, decreased serum estradiol (E 2 ) levels, and lower expression levels of ovarian steroidogenic factor 1 (SF1), steroidal synthetases, and insulinlike growth factor 1 (IGF1). On postnatal week (PW) 6 and PW12, the PDE F1 offspring showed altered reproductive behavior and ovarian morphology. The serum E 2 level and ovarian expression of SF1, steroidal synthetases, and IGF1 were also decreased. The adult F3 offspring showed alterations in reproductive phenotype and ovarian IGF1, SF1, and steroidal synthetase expression similar to those of F1. PDE induces ovarian developmental toxicity and transgenerational inheritance effects. The mechanism by which this toxicity occurs may be related to PDE-induced low-functional programming of fetal ovarian IGF1/SF1 and steroidal

D examethasone is a synthetic glucocorticoid that is widely used in therapy for threatened premature labor (1,2). Since the 1990s, to prevent prematurerelated diseases, single or repeated courses of treatment with dexamethasone have been recommended for women who are at risk for preterm delivery (3). A maternal and child health survey in 359 institutions from 29 countries launched by the World Health Organization showed that the dexamethasone treatment rate was 54% on average and could reach 91% in some countries (4). A growing number of studies indicate that prenatal dexamethasonetreated infants have lower birthweights (2,3,5), which is correlated with the frequency of dexamethasone treatment (6). Moreover, the low birthweight has been reported to be associated with noncommunicable diseases in adults (7,8). Animal experiments also indicate that prenatal dexamethasone administration might induce low birthweights, developmental toxicities in multiple organs, and chronic diseases in offspring (9).
As a common developmental toxicity, intrauterine growth retardation (IUGR) refers to embryonic (or fetal) growth and development restrictions during pregnancy caused by various adverse factors, mainly manifested as developmental disorders of multiple organs, growth retardation, and low birthweight. Intrauterine programming refers to the changes in morphology and function induced by damage in utero that can be sustained after birth; intrauterine programming mediates the developmental origin mechanism of susceptibility to chronic disease among IUGR offspring (10). As one of the most important female reproductive organs, the ovary provides oocytes and secretes sex hormones, including estrogens [estradiol (E 2 ) and estrone], progesterone, and a small amount of androgen. These functions are of importance in maintaining reproductive behavior. An epidemiological investigation showed that increased maternal testosterone levels cause reproductive function abnormality in female offspring (11,12). Studies have found that daughters of mothers who smoked during gestation had an earlier age of menarche (13)(14)(15). Studies in animals also showed that an adverse intrauterine environment, such as prenatal xenobiotic exposure (including dexamethasone) and prenatal stress, results in alterations in ovarian development and abnormalities in reproductive function in female offspring (16)(17)(18)(19)(20).
Dexamethasone treatment induces apoptosis of germ cells in the human fetal ovary (21), which suggests that dexamethasone may induce ovarian developmental toxicity. However, the underlying mechanism is unclear. E 2 is the most important and active sex hormone in females. By binding to the estrogen receptor (ERa and ERb), E 2 plays the physiological role of maintaining female sexual characteristics, follicular development, and ovulation. Steroidogenesis occurs in the granulosa cell layer under the catalysis of the steroidal synthetase system. Insulinlike growth factor 1 (IGF1) is an important factor in the regulation of the development of oocytes and the expression of steroid synthesis enzymes of estrogen (22). A lack of IGF1 activity in rat granulosa cells can lead to decreased expression of steroidogenic enzymes and lowered circulating estrogen levels as well as abnormal ovarian development (23). These results suggest that low IGF1 expression mediates the inhibition of estrogen synthesis and participates in ovarian developmental toxicity.
Transgenerational transmission refers to when a programming intervention is applied to a mother (the F0 generation) during pregnancy, and it will directly influence the offspring developing in utero (the F1 generation) (24). However, the germ cells (future gametes) that will form the F2 generation develop during this pregnancy and hence will also be directly exposed to the suboptimal environment. Therefore, it can be argued that only the effects on later generations (F3 and beyond) can truly be considered as transmitted across generations, namely, transgenerational inheritance (25). In this study, we applied dexamethasone during mid and late pregnancy using the clinical dose to establish the IUGR rat model. By observing the morphological and functional developments of the ovary at different periods in the F1 and F3 prenatal dexamethasone exposure (PDE) offspring, we aimed to confirm the ovarian developmental toxicity and transgenerational inheritance effect induced by PDE. Furthermore, we explored the intrauterine programming mechanism of the ovarian developmental toxicity at the level of expression and regulation of the ovarian steroidal synthetase system. This study helps explain the developmental origins of health and disease, guide rational clinical therapy, and improve quality of life. The antibody for cleaved-caspase 3 (R&D Systems catalog no. AF835; RRID: AB_2243952) was purchased from R&D Systems (Minneapolis, MN), and the antibodies for IGF1 (Santa Cruz Biotechnology; catalog no. sc-9013; RRID: AB_2122132) and 3b-hydroxysteroid dehydrogenase (3b-HSD) (Santa Cruz Biotechnology; catalog no. sc-30820; RRID: AB_2279878) were purchased from Santa Cruz Biotechnology Co. (Santa Cruz, CA). All primers were synthesized by Sangon Biotech Co., Ltd. (Shanghai, China). The enzyme-linked immunosorbent assay and radioimmunoassay kits for rat E 2 were obtained from R&D Systems and Beijing North Institute of Biological Technology (Beijing, China), respectively. TRIzol was purchased from Invitrogen Co. (Carlsbad, CA). Reverse transcription and real-time quantitative polymerase chain reaction (RT-qPCR) kits were purchased from Takara Biotechnology Co., Ltd. (Dalian, China). The SYBR Green dye was purchased from Applied Biosystems through Thermo Fisher Scientific (Foster City, CA). Other chemicals and reagents were of analytical grade.

Animals and treatment
Specific pathogen-free Wistar rats (no. 2012-2014, certification no. 42000600002258, license no. SCXK) weighing 209 6 12 g (females) and 258 6 17 g (males) were obtained from the Experimental Centre of the Hubei Medical Scientific Academy (Wuhan, China). Animal experiments were performed at the Centre for Animal Experiment of Wuhan University (Wuhan, China), which is accredited by the Association for Assessment and Accreditation of Laboratory Animal Care International. The Committee on the Ethics of Animal Experiments of the Wuhan University School of Medicine approved the protocol (permit no. 201719). All animal experimental procedures were performed in accordance with the Guidelines for the Care and Use of Laboratory Animals of the Chinese Animal Welfare Committee.
Rats were housed in metal cages with wire-mesh floors in an air-conditioned room under standard conditions (room temperature: 18°C to 22°C; humidity: 40% to 60%; light cycle: 12-hour light-dark cycle; 10 to 15 air changes per hour) and allowed free access to rat feed and tap water. All rats were acclimated 1 week before treatment, and two female rats were placed together with one male rat overnight in a cage for mating. The appearance of sperm in vaginal smears confirmed mating, and the day of mating was set as gestational day (GD) 0. From GD9 to GD20, the pregnant rats were injected subcutaneously with 0.2 mg/kg dexamethasone once per day, and the control rats were sham treated with the same volume of the vehicle (saline solution). On GD20, inhaled diethyl ether was applied to maintain an anesthetized status, while a subgroup of pregnant rats was euthanized. After anesthesia, the pregnant rats were euthanized by rapid exsanguination, and their blood and blood from embryos were collected. Pregnant rats with litter sizes of 12 to 14 (the male/female ratio was approximately 1:1) were considered qualified (n = 16 pregnant rats for each group). IUGR was diagnosed when the bodyweight of an animal was two standard deviations lower than the mean bodyweight of the control group (26). The female bodyweight and IUGR rate were obtained by dividing the number of IUGR rats in each litter by the total number of pups in the litter. The female fetal rats were randomly chosen from each litter for subsequent analysis. The fetal blood was collected and centrifuged for storage. For morphological purposes, right fetal ovaries were selected for fixing in freshly prepared Bouin's solution for 24 hours before being dehydrated in alcohol and embedded in paraffin. The remaining ovaries were immediately frozen in liquid nitrogen and stored at 280°C for further analysis.
The remaining pregnant rats were kept until normal delivery to produce first-generation offspring (the F1 generation). The number of pregnant rats in each group was 8 (the litter size was 12 to 14 at birth, and the male/female ratio was approximately 1:1). After weaning, all female F1 offspring were maintained on standard laboratory feed ad libitum. A subgroup of the F1 female offspring was euthanized via inhaled diethyl ether at postnatal week (PW) 6 and PW12, and n was set at 8 from eight different litters at each time point. At PW12, the remaining female offspring were mated with normal males to produce the second generation of offspring (the F2 generation). The F2 females were culled using the same protocol as the F1 generation to consecutively produce the female third generation (the F3 generation), which was killed via inhaled diethyl ether for blood and ovaries at PW12. The mating method is described briefly later (Fig. 1).
The vaginal patency of female offspring rats was examined every day after weaning until complete opening. The normal estrous cycle in the Wistar rat ranges from 5 to 7 days, which can be determined by cytological examination of vaginal smears. The estrous cycle was checked in the female offspring rats starting on PW10 and continued daily at 8:00 AM for 2 weeks and was determined based on vaginal cell morphology. Four phases were used for categorization: proestrous (predominance of epithelial cells, large and round), estrous (predominance of cornified epithelial cells, jagged shape), metestrous (coexistence of nucleated epithelial cells and cornified epithelial cells), and diestrous (predominance of leukocytes, small speckling).

Dexamethasone concentration detection for maternal and fetal serum
Five hundred-microliter serum samples mixed with ethanol (9:1, volume-to-volume ratio) and 4 mL of ethyl acetate were centrifuged for 5 minutes at 2000 3 g. The content of the upper solvent phase was collected in a rotary evaporator and redissolved in 100 mL of methanol. Qualities of dexamethasone control samples were prepared at concentrations of 155.4, 777.1, 3108.3, 7770.8, and 31,083.1 nmol/L. The experimental and quality control samples were stored at 2°C to 8°C for further detection. Experiments were carried out on an Agilent 1100 Series high-performance liquid chromatograph (Agilent Inc., CA). Separation was performed on a Waters C 18 column (Waters Co., Milford, MA) with the column temperature maintained at 30°C, and the sample injection volume was 20 mL (spectra wavelength: 240 nm).
Hematoxylin-eosin staining, immunohistochemistry, and transmission electron microscopic examination of the ovary Immediately after removal, one ovary from each animal was processed for sectioning. The 5-mm-thick paraffin histological sections were prepared and routinely stained with hematoxylineosin (HE). At each time point, the right-side ovaries from five rats per group were collected for morphological observation, and the largest cross-sectional areas were used for image analysis. Every fifth section of the series was saved, observed, and photographed with an Olympus AH-2 light microscope (Olympus, Tokyo, Japan).
For immunohistochemistry (IHC) analysis, the sections were incubated overnight at 4°C with the following antibodies: Ki67 (diluted 1:200), cleaved-caspase 3 (1:50), 3b-HSD (1:200), P450arom (1:100), and IGF1 (1:50). IHC analysis was performed using a diaminobenzidine staining kit to determine the expressional levels of Ki67 and cleaved-caspase 3 in fetal ovaries, and 3b-HSD, P450arom, and IGF1 in fetal and adult ovaries. Immunostaining for the negative controls was performed on parallel sections, in which the primary antibody was replaced with nonimmune rabbit immunoglobulin G. The signal was visualized using light microscopy and imaged, and positively stained areas were analyzed. The optical density of cleaved-caspase 3, 3b-HSD, P450arom, and IGF1 stained from each image was counted with the Nikon H550S photo imaging system (Nikon, Japan). The intensity of staining was determined by measuring the optical density in five random fields for each section, and five sections from different samples of each group were used for quantification. To determine the incidence of ovarian cell proliferation, fetal ovaries were stained for Ki67, which detects cells in the S-phase of the cycle. Ki67-stained nuclei localized primarily to granulosa cells, with some immunopositive theca, stromal, and endothelial cells. The number of Ki67-stained nuclei from each image was counted using the Nikon H550S photo imaging system. All images were captured using an Olympus AH-2 light microscope (Olympus). Analysis of the stained images was performed using Olympus software (version 6.1; Media Cybernetics, Silver Spring, MD).
Fetal ovaries were fixed with 2.5% glutaraldehyde in 0.1 M phosphate buffer (pH = 7.4) for 2 hours at 4°C and postfixed with 1% osmium tetroxide. The samples were dehydrated through a graded series of ethanol and embedded in epon 812. Ultrathin sections (;50 nm) were cut with an LKB-Vultra microtome (Bromma, Sweden), dually stained with uranyl acetate and lead citrate, and examined with a Hitachi H600 transmission electron microscope (Hitachi, Tokyo, Japan). We focused on the pregranulosa cells of fetal ovary and observed the changes in the subcellular fraction.
To classify follicles and measure the granulose layer thickness of the antral follicles, every five sections from five one-side ovaries (from five females) with the largest cross-sectional area were used for image analysis. Follicles were classified as primordial, primary, secondary, antral, and atretic follicles, and 20 to 30 follicles were counted in each section. In each section, five antral follicles were examined, and the thickness values were collected.

E 2 concentration detection in serum
The serum of female fetuses (from three to four litters) was pooled into one sample, and three to four samples were used in each group. The concentration of serum E 2 for fetal serum was measured with an enzyme-linked immunosorbent assay kit [minimum detectable dose: 2.14 pg/mL; intra-assay precision: coefficient of variation (CV) ,6.0%; interassay precision: CV , 7.1%]. The radioimmunoassay kit was used to measure the concentration of serum E 2 from postnatal samples (minimum detectable dose: 2.51 pg/mL; intra-assay precision: CV , 10%; interassay precision: CV , 15%). All assay procedures were performed according to the manufacturer's protocol.
Total RNA extraction, reverse transcription, and RT-qPCR of the ovary The total RNA was extracted from ovaries using TRIzol reagent (Invitrogen Co.) according to the manufacturer's protocol. The tissues of each littermate were pooled for homogenization as one sample. The concentration and purity of the total RNA were determined using a spectrophotometer (NanoDrop 2000; Bio-Rad Co., Hercules, CA), and the total RNA concentration was adjusted to 1 mg/mL. Single-strand complementary DNA was prepared from 1 mg of total RNA according to the protocol of the kit and stored at 220°C until use. All primers were designed using Primer Premier 5.0 (PREMIER Biosoft International, Palo Alto, CA). The sequences of each of the designed primers were queried using the National Center for Biotechnology Information BLAST database for homology comparison and are listed in Supplemental Table 1. Genes detected in this study included steroidogenic factor 1 (SF1), steroidogenic acute regulatory protein (StAR), cytochrome P450 cholesterol side chain cleavage (P450scc), 3b-HSD, steroid 17-hydroxylase (P450c17), 17b-hydroxysteroid dehydrogenase type 1 (17b-HSD1), 17b-HSD2, P450arom, and IGF1. RT-qPCR was performed using the ABI StepOne RT-PCR Thermal Cycler (Thermo Fisher Scientific, Waltham, MA) in a 10-mL reaction mixture. To quantify the gene transcripts more precisely, the messenger RNA (mRNA) level of a housekeeping gene, glyceraldehyde 3-phosphate dehydrogenase, was measured and used as the quantitative control. Each sample was normalized to the glyceraldehyde 3-phosphate dehydrogenase mRNA level.

Statistical analysis
Excel (Microsoft, Redmond, WA) and Prism (GraphPad Software, La Jolla, CA) were used to perform the data analysis. Quantitative data were expressed as the mean 6 standard error of the mean (SEM). For weights of the fetuses, the mean weight of each litter was used for statistical analysis. The IUGR rate for each litter was determined by arcsine square root transformation before t test evaluations (27). Student two-tailed t test was used to compare the mean values of two groups as applicable. Statistical significance was designated at P , 0.05.

Fetal birthweights and serum dexamethasone concentrations of maternal and fetal rats
First, we established the mid-and late-pregnancy dexamethasone-exposure rat model for IUGR (28). Compared with the control, the PDE fetus had a smaller body size ( Fig. 2A), a lower bodyweight (P , 0.01; Fig. 2B), and an elevated IUGR rate (P , 0.05; Fig. 2C). Moreover, the female fetal serum dexamethasone concentration was 267 nmol/L, which was 31.6% of that of the mother (846 nmol/L; Fig. 2D). The previous data indicated that the maternal dexamethasone could enter into the circulation of the fetuses, which was accompanied by a decreased bodyweight and IUGR.

Ovary morphology and steroid synthesis function in fetal rats
We observed the alterations in fetus ovarian morphology and steroidogenesis. Compared with the control, HE staining of the fetal ovary showed an unchanged number of primordial follicles per unit area (100 3 100 mm 2 ) ( Fig. 3A and 3B); the IHC results showed that the number of Ki67-stained nuclei per unit area (100 3 100 mm 2 ) was markedly reduced (P , 0.01; Fig. 3C and 3D). However, there was no significant change in the protein expression of cleaved-caspase 3 ( Fig. 3E and 3F); the fetal ovary ultrastructural observation revealed that the mitochondria in the PDE ovary appeared to have fewer cross-sections of the cristae and the vacuolelike changes (Fig. 3G). The previous results suggested that PDE induced the altered fetal ovary morphology and weakened cell proliferation.
Furthermore, we found that the serum E 2 level was lower than in the control (P , 0.01; Fig. 4A); the mRNA expression of the fetal ovary steroidogenic system (including SF1, StAR, P450scc, 3b-HSD, P450c17, 17b-HSD1, and 17b-HSD2) and IGF1 were significantly decreased in the PDE group (P , 0.05, P , 0.01; Fig. 4B and 4C). The IHC of the fetal ovary showed that P450arom and IGF1 expression were both reduced (P , 0.05, P , 0.01; Fig. 4F-4I), though there were no changes in the expression of 3b-HSD ( Fig. 4D and 4E). These results indicated that PDE caused the inhibition of ovarian IGF1 and downregulated the steroidogenic system and serum E 2 level in fetal rats.

Reproductive phenotype and steroid synthesis function during puberty in the F1 generation
To study whether the ovarian toxicity caused by PDE could be sustained into the postnatal period, we observed the reproductive function (vaginal patency and estrous cycle), follicular development, E 2 level, ovarian steroidogenesis, and IGF1 expression of the F1 generation starting at PW6, an important window of time during development (29). Compared with that of the control, the vaginal patency of the PDE group was advanced (P , 0.05; Fig. 5A). The PDE ovary weight and index showed no significant changes (Fig. 5B). After counting the follicles in the cross-section of the PW6 ovary, we found that the proportion of secondary follicles was elevated (P , 0.05; Fig. 5C), whereas other stages of the follicles remained constant. Furthermore, the serum E 2 concentration was reduced (P , 0.01; Fig. 5D), and the granulose layer thickness of the antral follicle was thinner than in the control (P , 0.01; Fig. 5E). The ovarian mRNA expression levels of P450scc and 3b-HSD were remarkably decreased (P , 0.05; Fig. 5F), whereas the expression levels of SF1, StAR, 17b-HSD2, and P450arom showed a decreasing trend (Fig. 5F). Moreover, the ovarian IGF1 mRNA expression was significantly reduced (P , 0.05; Fig. 5G). The IHC results revealed that the expression levels of IGF1, 3b-HSD, and P450arom were markedly decreased or showed a decreasing trend (P , 0.05, P = 0.067, P = 0.075, respectively; Fig. 5H-5M) in the PDE ovary. These results suggested that PDE induced early puberty and a follicle developmental abnormality. Moreover, the E 2 level, ovarian steroidogenesis, and IGF1 expression were inhibited in the F1 female offspring.

Reproductive phenotype and steroid synthesis function in adults from the F1 generation
To determine whether the ovarian toxicity caused by PDE could be sustained into adulthood, we examined the reproductive function, fertility (pregnancy rate, littermate number, neonatal bodyweight, and litter death rate at postnatal day 28), follicular development, E 2 levels, ovarian steroidogenesis, and IGF1 expression in the adult F1 offspring (PW12). The PDE ovary weight and index at PW12 showed no change compared with the controls (Fig. 6A), and for the classification of follicles, the proportion of secondary follicles was elevated, whereas the follicular atresia was reduced (P , 0.05; Fig. 6B). After 2-week testing of the estrous cycle, we found that the total duration in the F1 PDE group was longer than that in the control group (P , 0.05; Fig. 6C). When mating with normal male rats, the PDE group showed no changes in the pregnancy rate, littermate number, and litter neonatal bodyweight, but the litter death rate at postnatal day 28 rose significantly (P , 0.01; Fig. 6D-6G). Furthermore, the serum E 2 concentration was significantly decreased compared with the controls (P , 0.05; Fig. 6H), and the granulose layer thickness of the antral follicle was also reduced in the PDE group (P , 0.01; Fig. 6I). Moreover, the mRNA expression levels of SF1, 3b-HSD, 17b-HSD2, and P450arom as well as IGF1 were also inhibited (P , 0.05, P , 0.01; Fig. 6J and 6K). The IHC results indicated that the protein expression levels of 3b-HSD, P450arom, and IGF1 were significantly decreased  Fig. 6L-6Q). These results suggested that PDE induced a follicular abnormality and adverse pregnancy outcomes with a reduction of E 2 and a dramatic inhibition of ovarian steroidogenesis and IGF1 in the adult F1 offspring.

Reproductive phenotype and steroid synthesis function in adults from the F3 generation
To study the transgenerational inheritance effect of ovarian developmental toxicity caused by PDE, we measured the reproductive function, fertility, follicular   Fig. 7A). The ovary weights and indexes showed no differences between the PDE and control groups (Fig. 7B). Follicular atresia was increased (P , 0.05; Fig. 7C), but the estrous cycle duration remained unchanged in the PDE group  (Fig. 7D). While mating with normal male rats, the F3 females showed no changes in the pregnancy rate, littermate number, neonatal bodyweight, and litter death rate (Fig. 7E-7H). Next, we found that the serum E 2 concentration was significantly decreased (P , 0.01; Fig. 7I) and the granulose layer thickness of the antral follicle was unchanged in the PDE group (Fig. 7J). Moreover, the mRNA expression levels of  ovarian SF1, P450c17, 17b-HSD1, 17b-HSD2, P450arom, and IGF1 in the F3 offspring were significantly reduced (P , 0.05, P , 0.01; Fig. 7K and 7L). The IHC results indicated that the protein expression levels of 3b-HSD, P450arom, and IGF1 were decreased in the PDE group (P , 0.01; Fig. 7M-7R). The results suggested that earlier puberty, follicular abnormalities, and adverse pregnancy outcomes occurred in the F3 offspring of PDE, featuring lower E 2 levels, ovarian steroidogenesis, and IGF1 expression. , and IGF1 protein expression; five samples from each group were selected for protein detection, and five random fields from each section were scored with immunohistochemical measurements. Mean 6 SEM, *P , 0.05, **P , 0.01 vs control. GAPDH, glyceraldehyde 3-phosphate dehydrogenase.

Discussion
Dexamethasone is commonly used for the clinical treatment of premature labor. The National Institutes of Health recommends that routine perinatal dexamethasone treatment (therapeutic use) should be a 6-mg intramuscular injection every 12 hours (four times for one course) (30,31). To address threatened premature labor, multiple courses of dexamethasone are applied due to the difficulty of early diagnosis (32)(33)(34), and cases of repeated courses of dexamethasone treatment once per week at 25 weeks of gestational age until birth have been reported (35), which corresponds to our model. Using a dose conversion between humans and rats (human:rat 1: 6.17 by body surface area comparison) (36), PDE with 0.2 mg/kg/d dexamethasone in rats in the current study is equivalent to 12 mg of dexamethasone per day for a 60-kg person, which can mimic clinical use in pregnant women (0.3 to 0.5 mg/kg) (37). In the current study, the dexamethasone concentrations after exposure to 0.2 mg/ kg/d were 846 nmol/L (0.332 mg/mL) and 267 nmol/L (0.105 mg/mL) in maternal and fetal blood, respectively, which did not reach the maximal clinical dose (38). Based on this discussion, the applied dose of dexamethasone (0.2 mg/kg/d) and exposure time were of practical significance in studying ovarian developmental toxicity and the transgenerational inheritance effect. Previous studies have shown that dexamethasone exposure can lead to long-term reproductive dysfunction (39)(40)(41). Moreover, dexamethasone treatment can alter ovarian cyclicity and impair oocyte development potential in mice (42,43). Fetal exposure to dexamethasone on days 16 to 18 of pregnancy decreases the pool of nongrowing follicles in the neonatal ovary (18). The prenatal dexamethasone treatment reduces the pool of healthy primordial follicles through the induction of autophagy and/or apoptosis in spiny mouse offspring (44). In the current study, exposure to dexamethasone during mid and late pregnancy induced a decreased fetal bodyweight with an increased IUGR rate, an altered ultrastructure, and a weakened cell proliferation rate in the fetal ovary, as evidenced by the disappearance of mitochondrial cristae, vacuolelike changes in the mitochondria, and lower Ki67 expression. Vaginal patency in female rats is usually used as the marker of puberty initiation. At PW6 and PW12, the PDE female offspring showed earlier puberty, altered follicular development, reproductive dysfunction, and damaged fertility, characterized by earlier vaginal patency, an increased component of secondary follicles, a prolonged total duration of the estrous cycle, and an increased litter death rate at postnatal day 28. These results suggested that PDE induced alterations in ovarian morphology, reproductive function, and fertility in the prenatal and postnatal period.
Estrogen synthesis and secretion by the ovary are important for maintaining the reproductive phenotype in females. The steroidal synthetase system includes the rate-limiting enzymes StAR and P450scc and the steroid type-determining enzymes 3b-HSD, P450c17, 17bhydroxysteroid dehydrogenase, and P450arom. This system is regulated by various transcription factors, including SF1, DAX-1, GATA6, and AP-1. It has been confirmed that there are SF1 binding sites in the promoter regions of all steroid synthesis genes, and SF1 has been shown to play a central role in gonadal differentiation and development (45). Mice with a complete deletion of SF1 showed ovarian insufficiency and impaired fertility, and specific knockout mice with markedly decreased SF1 expression in the granulosa cells showed reduced ovarian expression of P450arom (46). The human fetal ovary successively initializes a primordial follicle at gestational week 20, which is formed by a single oocyte, a single layer of pregranulosa cells, and basement membrane cells. Although the ovarian and steroidogenesis development windows are different within species, many mammals (such as sheep, pigs, and rats) express steroidal synthetases and synthetize estrogen midpregnancy; development might be disturbed by xenobiotics (20,47,48).
Studies have indicated that dexamethasone can suppress steroidogenesis in the ovary. For example, dexamethasone has been shown to decrease StAR protein levels and estrogen production in rats and human granulosa cells (49,50). In the current study, the levels of serum E 2 and the expression of ovarian SF1 and steroidal synthetases were decreased during dexamethasone exposure. Based on the high levels of dexamethasone that were detected in the fetal serum, we assume that the inhibition of the steroidal synthetase system mediated the poor function of estrogen synthesis in the fetal ovary that was directly induced by dexamethasone. Furthermore, in both puberty and the adult stage (PW6 and PW12) in the PDE group, we found lower E 2 levels with thinner layers of granulosa cells in the antral follicle and lower expression levels of SF1 and steroidal synthetases (mainly 3b-HSD, 17b-HSD2, and P450arom) in ovarian tissue. These results suggested that the alteration in the reproductive phenotype in the offspring after birth might be associated with the consistent low function of ovarian estrogen synthesis caused by dexamethasone.
Several autocrine growth factors are known to regulate ovarian function, of which IGF1 performs a central role in ovarian cell proliferation and estrogen synthesis (51). By binding to the IGF1 receptor, the mitogenactivated protein kinase/extracellular signal-regulated kinase and/or phosphatidylinositol 3-kinase (PI3K)/Akt pathways are activated to promote cell proliferation and differentiation, respectively. In the developing primordial follicle, IGF1 intensively enhances the proliferation effect of follicle-stimulating hormone in granulosa and theca cells (51). Following treatment with selective inhibitors of the PI3K or extracellular signal-regulated kinase pathway, IGF1/IGF1 receptor stimulates ovarian expression of steroidal synthetase and E 2 synthesis, mainly through the PI3K/Akt pathway in bovine granulosa cells (52). Furthermore, female IGF1 mutant mice fail to ovulate even after administration of gonadotropins (53). The IGF1 expression level in fetal tissue is known to be negatively regulated by glucocorticoids, which inhibit IGF1 in multiple organs (54,55). We have previously found that prenatal xenobiotic (caffeine, nicotine, and ethanol) exposure induces IUGR and fetal "overexposure to maternal glucocorticoid" and inhibits the IGF1 signaling pathway in multiple fetal tissues (liver, skeletal muscle, growth-plate cartilage, and adrenal gland) (27,(56)(57)(58). Additionally, in mural granulosa cells, dexamethasone treatment is also associated with decreased mRNA expression of IGF1 (59). In this study, we found that the mRNA and protein expression levels of IGF1 were inhibited in the fetal ovary, and this alteration could be sustained into puberty and adulthood of the PDE offspring, suggesting that the abnormalities in ovarian morphology and steroidogenesis were related to low-expression programming of IGF1 caused by PDE.
Adverse external stimuli during the embryonic stage or childhood may result in life-long programming, which leads to tissue and/or organ functional or gene expressional changes during the developmental stage. These changes can be maintained from puberty to adulthood and even be passed to the next generations (60). As previously mentioned, only effects on later generations (F3 and beyond) can truly be considered as transmitted across generations, namely, as transgenerational inheritance. To determine the transgenerational inheritance effect of the ovarian developmental toxicity caused by PDE, we examined the alterations in the F3 generation in the current study. Interestingly, we found an earlier vaginal patency and an increased component of atretic follicles in the F3 generation, which were in accordance with the F1 generation. Furthermore, decreased levels of serum E 2 and inhibited expression of steroidal synthetases (including SF1, P450c17, 17b-HSD1, 17b-HSD2, and P450arom), as well as mRNA and protein expression of IGF1, were also observed in the F3 generation, in accordance with the abnormalities in the reproductive phenotype. These results confirmed that there was a transgenerational inheritance effect on ovary developmental toxicity and its associated reproductive phenotype alterations in offspring caused by PDE. Interestingly, in the F2 generation, the PDE female offspring showed increased expression of steroidal synthetases (including SF1, StAR, 3b-HSD, and P450arom) (data not shown), and a proper mechanism to explain this intergenerational difference remains elusive.
Although the exact mechanism underlying the transgenerational programming effects caused by adverse prenatal environment is currently unknown, epigenetic modification alterations cannot be ignored (8,61). Epigenetics is defined as heritable changes in gene expression that occur without alterations in the DNA sequence (62). In recent years, increasing data have indicated that a detrimental environment in utero causes a transgenerational inheritance effect of the disease phenotype, and the epigenetic marker in the germ cell plays a vital role (63). For example, exposure to the endocrine disruptor vinclozolin during pregnancy alters the expression of DNA methyltransferase in offspring and generates larger sets of differential DNA methylation regions in germ cells that are transmitted to multiple generations (64). Prenatal dioxin-exposed offspring rats exhibit polycystic ovarian disease in females and prostate disease in males that is sustained into the F3 generation; 50 differentially methylated DNA regions have been identified in the sperm epigenome of the F3 generation (65). The cognitive and psychiatric disorders induced by stress during early life in the F1 generation reoccurred in the F3 offspring, and the altered DNA methylation of the spermatogonium was believed to be the underlying mechanism (66).
Glucocorticoid is known to regulate epigenetic modifications and expression at multiple gene promoters by epigenetic modification enzymes (67). It has been reported that prenatal tris(1,3-dichloro-2-propyl) phosphate exposure induces transgenerational toxicity in Figure 8. Intrauterine programming mechanism of ovarian developmental toxicities in F1/F3 offspring rats induced by PDE.
zebrafish via the downregulation of IGF1 (68), and the transcriptional expression of IGF1 is under the control of epigenetic modification (69)(70)(71). Moreover, the expression of SF1 in endocrinal organs had been shown to be time dependent, cell specific, and modified by the methylation of promoter cytosinephosphate-guanine islands in a strict manner (72). Our previous study has indicated that a high level of maternal glucocorticoids induced by prenatal caffeine exposure may alter DNA methylation and histone modification of the IGF1 promoter in the liver of the IUGR fetal rats, resulting in lower expression of IGF1 (27). Using the same rat model, we also observed inhibition of the fetal adrenal IGF1/Akt signaling pathway, increased expression of DNA methyltransferase 1/3a/3b, and an augmented SF1 DNA methylation level, accompanied by inhibition of SF1 and StAR expression (73,74). In the current study, these results indicated that the low functional programming of ovary IGF1/SF1 and steroidal synthetases by PDE are important targets of transgenerational alterations in the F3 offspring. Therefore, we speculate that dexamethasone may regulate IGF1/ SF1 expression through epigenetic modifications and thus inhibit steroidal synthetases in the fetal ovary, even being transmitted transgenerationally. We regret that the amount of fetal ovarian tissue (especially in the PDE group) was too small for detection, and thus, it is difficult to confirm our speculation concerning the epigenetic mechanisms of ovarian low-functional programming of IGF1/SF1.
The current study demonstrated that PDE caused ovarian developmental toxicity and resulted in alterations of endocrine function and the reproductive phenotype, demonstrating evidence of its transgenerational inheritance effect. The underlying mechanisms might be associated with the dexamethasone-induced lowexpressional programming of IGF1/SF1 and steroidal synthetases (Fig. 8). This study provides some interesting evidence for the intrauterine origin, transgenerational inheritance, and possible biomarkers of ovarian developmental toxicity.