Maternal Aromatase Inhibition Effects on RFamide-Related Peptide-3 and Gonadotropin Releasing Hormone in Adult Murine Offspring With Polycystic Ovary Syndrome

Background: Despite the prevalence of polycystic ovary syndrome (PCOS) among childbearing women and the development of many animal models for this syndrome, information on its etiology is still scarce. Intrauterine hyperandrogenic environment may underlie changes at the levels of hypothalamus, pituitary and ovary organization in female offspring, and PCOS later in life. Letrozole, has been shown to mimic reproductive and metabolic characteristics of PCOS in adult rodent models. Therefore, the aim of this research was to assess the condition in a prenatal letrozole-treated rat model. Methods: Twenty-eight female rats from dams receiving letrozole at certain doses during late pregnancy were used in the trial. Pregnant Sprague-Dawley rats (n=21) received letrozole treatment on days 16-18 gestation at doses 1.25, 1.0, 0.75, 0.5, and 0.25 mg/kg body weight (BW). Results: Prenatal letrozole-treatment delayed parturition time and reduced the litter size in pregnant dams (P<0.0001). Late puberty onset, irregular ovarian cyclicity, increased anogenital distance (AGD), body weight gain, and serum testosterone concentration and reduced estradiol levels (P<0.0001) were observed in the female offspring of dams receiving 1.25 and 1 mg/kg BW letrozole. Furthermore, Letrozole at 1.25 and 1 mg/kg BW showed increased Rfrp and decreased Gnrh mRNA expression (P<0.0001). Conclusions: Letrozole treatment at doses 1 mg/kg BW and lower was not feto-toxic. It was concluded that 1 mg/kg BW letrozole may be suggested for prenatal PCOS induction.


Background
Polycystic ovary syndrome (PCOS) is a gynecological disorder along with clinical or biochemical manifestation of hyperandrogenism and menstrual problems (1).Despite abundant information, there is a lack of a precise de nition of its pathogenic mechanisms and animal models that accurately re ect the features of human PCOS.The underscored endocrine features of PCOS are hyperandrogenism and LH hypersecretion (2).According to the NIH, Rotterdam, and AE-PCOS conferences, hyperandrogenism is the most important diagnostic criteria of PCOS (3).Among androgenic animal models of PCOS, prenatal androgenized PCOS models displayed a set of reproductive and metabolic features of human PCOS (4).
Androgen exposure during intrauterine life created a phenocopy of reproductive and metabolic features of PCOS in adult female rats (5).In pregnant women with PCOS, higher circulating levels of testosterone than in normal pregnant women were reported (6).Notwithstanding the fact that differentiation of hypothalamic centers regulating reproductive and ovarian activity (folliculogenesis and ovulation) occurs during the fetal life (7), hormonal abnormalities during pregnancy can in uence reproductive regulating centers in the hypothalamus.
To create a hyperandrogenic environment during pregnancy, we used letrozole because it has the ability to mimic reproductive and metabolic characteristics as observed in PCOS, and also induces ovulation in PCOS women as therapeutic drug.In addition, letrozole can affect the neuronal pathways in the brain that control gonadotropin secretion by disrupting the mechanism through which testosterone is converted to estrogen, and these effects can be transmitted epigenetically to the female fetus.Therefore, the use of these drugs in pregnant PCOS women may have adverse effects on their offspring at adulthood.Letrozole, a non-steroidal aromatase inhibitor, created hyperandrogenism and polycystic ovarian morphology in rats (8), and has been introduced as the second drug line for PCOS.Letrozole lacks peripheral anti-estrogenic effects and has a higher pregnancy rate than clomiphene citrate.Furthermore, letrozole is less costly than other ovulation-stimulating preparations, with a therapeutic e cacy similar to gonadotropin agonists (9).In a retrospective cohort study, the offspring of mothers treated with letrozole and clomiphene citrate had 2.5 and 3.9% chromosomal abnormalities, respectively (9).Accordingly, we aimed to determine the optimal dose of letrozole for prenatal PCOS induction in female rats by assessing the reproductive, endocrine PCOS phenotypes and the hypothalamus-pituitary-gonad (HPG) axis alterations via evaluation of the mRNA expression of Gnrh and Rfrp genes.
Up to now, a number of studies have highlighted that prenatal testosterone exposure disrupted HPG axis differentiation (10)(11)(12).Recent evidence suggested that prenatal exposure with androgens led to insensitivity of GnRH neurons to the negative feedback of steroids resulting in elevated LH levels in rodents (10), Rhesus monkeys (10,11), guinea pig (10,11) and non-human primates (10).In addition, fetal androgen excess in rats affects the hypothalamic-pituitary axis and LH secretion (12).But there has been no model to induce PCOS using letrozole in the prenatal period so far.It is plausible that increased levels of maternal androgens by the mechanism of action of letrozole could cause changes in HPG axis differentiation including GnRH and RFRP-3 neuropeptides.Our previous study showed that by inducing PCOS in adult rats using 90-day constant light, along with crating a typical PCOS model, decreased RFRP-3 gene expression in the dorsomedial nucleus (DMN) (13).Based on this previous nding, the question arose as to whether letrozole was able to produce a typical prenatal PCOS model.Secondly, whether it would lead to changes in the hypothalamic level and major reproductive neuropeptides.Therefore, the present study was designed to determine the optimal dose of letrozole for prenatal PCOS induction in a rat model, and to evaluate a number of endocrine, neuroendocrine and reproductive features.

Ethics
All procedures were approved by the Ethics and Research Committee of Shiraz University (Approval ID: IR.SUMS.REC.1397.434;Approval date: 2018-08-01), and carried out in accordance to the health instructions for the care and use of animals.

Animals
Adult (8-week-old) female Sprague-Dawley rats (n=30) weighing between 130 to 180 g were obtained from the Center of Comparative and Experimental Medicine, Shiraz University of Medical Sciences.The animals were kept in normal experimental condition at 12-12 h light-dark period, 23±3 °C, and 25±5% humidity, having free access to the pelleted rat food and water.
After adaptation, vaginal smears were obtained from the rats by vaginal douching and the stages of the estrous cycle were determined based on the appearance of nucleated epithelial cells and corni ed cells.The rats in proestrous or estrous phases (n=21) were joined with sexually experienced adult male rats (375 g mean weighing) at 2:1 female/male ratio overnight.The next morning, the rats were examined for the presence of the vaginal plug which was recorded as the sign of mating and the rst day of pregnancy (gestation day: GD1).

Prenatal letrozole induction of PCOS
Pregnant rats were randomly divided into letrozole-treated, control and sham group.The optimal dose of letrozole was in a pilot trial designed to determine the safe doses of letrozole that showed no feto-toxic effects but could induce PCOS in the adult female offspring.Letrozole doses over 1.25 mg/kg resulted in parturition delay, fetal absorption, uterine infection, and pup mortality.
Due to the fact that, our main goal was to induce changes in ovarian, endocrine, and neuroendocrine levels in female offspring to develop PCOS phenotypes, safe and non-lethal doses were selected.Twentyone pregnant rats (n=3 for each group) were orally administrated with letrozole (L6545, Sigma-Aldrich, St. Louis, USA) dissolved in 1% carboxymethylcellolose (CMC, C5013, Sigma-Aldrich, St. Louis, USA) at concentrations of 0.25, 0.5, 0.75, 1.0, and 1.25 mg/kg BW during GDs 16, 17 and 18.In other words, letrozole doses were chosen to increase blood androgen levels higher than physiological levels that lead to PCOS development.Furthermore, GDs 16, 17 and 18 was chosen for letrozole administration owing to the brain differentiation for inactivation of the LH surge center occurs in the late pregnancy of rat (14), and is practically an aromatase peak occurs late in gestation and neonatal life (15).The sham group received vehicle only; 1% CMC dissolved in distilled water, and the control rats were untreated.Upon parturition, data including the parturition date, litter size, pup gender, and pup birth weight were recorded.
The pups were kept with their mothers until weaning.
At weaning on postnatal day (PND) 21, the pups were sexed and weighed again.The anogenital distance (AGD) was measured using a caliper.The AGD index (AGDI) was calculated in order to normalize the AGD for body weight at weaning.The AGDI was calculated as AGD/BW × 100 (16).After weaning, female offspring were separated in each group (n = 4).The female pups (n = 4 per group from more than one dam), maintained under standard conditions, were weighed every two days from weaning until the end of the study.After one-week post weaning (on PND 28), the female offspring were checked for vaginal opening as the sign of puberty.Vaginal smears were evaluated from 28 PND for 4 weeks to determine the phases of the estrous cycle (17).

Tissue and blood sampling
Four weeks after estrous cycle observation, due to the unavailability of other anesthetics, chloroform (Merck, KGaA, index No, 602-006-00-4) was used for momentary anesthesia.The rats in each group were euthanized temporary with a few drops of chloroform distributed on cotton in the desiccator.After that, blood was collected by cardiac puncture in tubes without anticoagulant, blood serum was prepared by centrifugation at 3000 rpm for 15 min.Blood serum was stored at -20 °C until evaluation of testosterone, estradiol, progesterone, follicle stimulating hormone (FSH) and luteinizing hormone (LH).
Immediately after blood sampling and then cervical dislocation, the brain was removed from the skull, and was sampled for real-time PCR analyses of GnRH and RFRP-3 relative gene expression.The mammillary bodies were separated from the inferior and then optic chiasm from the posterior parts of the brain.Finally, with third coronal cut, the diencephalon was removed and the pre-optic area (POA) and dorsomedial nucleus (DMN) was dissected in a piece of aluminum foil and immediately transferred into liquid nitrogen.Following decapitation, the ovaries were xed in 10% buffer formalin.The blood and tissue sampling were performed regardless of the stage of the estrous cycle.

Histological evaluation
The xed ovaries were washed in phosphate-buffered saline (PBS), dehydrated in ascending concentrations of alcohol and embedded in para n.Every ten serial sections (5 μm) were depara nized at 60 °C, transferred into ascending concentrations of ethanol, and then xylene, and stained with hematoxylin and eosin (H&E).The number of total follicles including: primary, secondary, tertiary or antral follicles and also atretic and cystic follicles, and corpora lutea were counted using a light microscope (CX21, Olympus, Japan) by one person blinded to the origin of the sections.For qualitative observations, three sections per animal from the beginning, middle and end of the ovarian tissue were evaluated using a digital camera (Nikon, NI_U 2013, Japan).

Expression levels of hypothalamic RFRP-3 and GnRH mRNA
Total RNA was extracted according to the manufacturer's extraction kit (Parstous RNA Isolation Kit, Parstous, Mashhad, Iran), and the pure extracted RNA was kept at -80 °C.The total puri ed RNA was measured by ultraviolet spectrophotometry (Nano-Drop, ND1000, USA).Removal of contamination was performed using the DNase I, RNase-free kit (Fermentas EN0521 DNaseI (RNase free)).
Based on the cDNA synthesis kit instructions (Smobio, Taiwan), the primers and dNTPs as well as rst strand cDNA buffer were mixed in a nal volume 20 µL, then incubated in 37 °C for 50 min and 85 °C for 5 min.Finally, for RNA removal, 1 µL RNase-H was added to each reaction medium and then incubated at 37 °C for 20 min.Prepared cDNA was preserved at -20 °C.To setup the annealing temperature, cDNA specimens were ampli ed by Thermocycler (StepOne™ Real-Time PCR, System Applied Biosystems, Carlsbad, CA).
The primers for target genes RFRP-3 and GnRH and the reference gene, β-actin, were designed by Primer3 Software (Table 1).For each reaction of real-time PCR, a mixture of 10 µL SYBR Green (TaKaRa, Dalian, China), 7 µL distilled water, 1 µL forward primer, 1 µL reverse primer and 1 µL cDNA was prepared in a nal volume of 20 µL.Eventually, the ampli cation process was carried out by StepOnePlus Real-Time PCR Systems (Bioneer, South Korea) and the CT values were recorded by real-time PCR software.For evaluation of relative mRNA expression of RFRP-3 and GnRH, the quantitative examination of CT values was carried out using the 2

Results
Oral administration of letrozole to rats on days 16, 17 and 18 of pregnancy impacted on the gestation length (P <0.0001), number of neonates born (litter size) (P <0.0001), and male/female ratio (P = 0.0369) (Figure 1-A, B, C).Longer gestation length was observed in the 1.25 mg/kg BW group compared to control and sham groups.Furthermore, in the 1 mg/kg BW group, an increase in gestation length was observed compared to the 0.5, 0.25 mg/kg BW and sham groups (P = 0.0093).In the pilot experiment evaluating the safe doses of letrozole in pregnant rats, doses higher than 1 mg/kg BW (1.5 to 3 mg/kg) on 16, 17 and 18 GDs delayed delivery, and needed cesarean section (data not shown).
A signi cant decrease in litter size was observed at 1.25 mg/kg BW (P <0.0001).The pilot experiment also revealed that higher doses also reduced the litter size, and increased neonatal mortality and fetal absorption (data not shown).
The offspring male/female ratio in the 1 mg/kg BW group was lower than the sham and 0.75 mg/kg BW groups (P <0.05).

AGD and AGDI and puberty onset
Letrozole treatment increased the anogenital distance (AGD) and AGDI compared to control groups (P <0.0001; Figure 2-A, B), and puberty was delayed (Figure 2-C) at 1.25 and 1 mg/kg BW letrozole doses (P <0.01) Body weight and ovarian weight Letrozole, time and their interaction signi cantly affected the weight gain in female rats (P <0.0001).
Weight gain changes at 6, 7 and 8 weeks were higher in letrozole-treated rats than in the control and sham groups (P <0.01; Figure 3).There was no effect of letrozole on the ovarian weight ((P = 0.14).

Estrous cycles
The pattern of estrous cycles in letrozole-treated groups was different (P<0.0001; Figure 4).Absence of cycles and irregular cycles were observed at all doses of letrozole, and the greatest delay in the onset of cycles ( rst proestrus) after puberty was observed at 1.25 and 1 mg/kg BW doses which also recorded a delay in puberty as reported earlier.
The higher percentage of proestrus in the 1 mg/kg BW group was only signi cant compared with 0.75 and 1.25 mg doses (P <0.01).The percentage of metestrus at 1 mg dose was lower than at 0.75 mg dose, and the percentage of diestrus at 1.25 mg dose was higher than 1 mg dose (P <0.01).There was no signi cant difference between the groups in the percentage of estrus (P = 0.37; Figure 5-A) or follicular phase and luteal phases (P = 0.9; Figure 5-B).The number of completed cycles in the control group was higher than other groups (P <0.05; Figure 5-C).

Ovarian histology
The total number of primary, secondary, antral and atretic follicles, and the number of corpora lutea in the 0.25 mg/kg BW group was higher than the other groups (P <0.0001; Figure 6-A-B-D).However, the number of cystic follicles in 0.25 and 0.75 mg/kg BW groups was higher than the sham and control rats (P <0.01; Figure 6-C).

Serum concentration of steroids and gonadotropins
Serum testosterone concentration was higher in the 1.25 mg group than the other groups (P <0.0001).
Also, in the 1 mg group, testosterone concentration was higher than the other groups except 0.75 mg group (P <0.01; Figure 7-A).Serum estradiol levels in 1.25 mg and 1 mg groups was lower than other groups (Figure 7-B).Serum progesterone level was higher in 0.25 mg group than in all groups but not in the 0.5 and sham groups (P <0.0001; Figure 7-C).

Gene expression of hypothalamic polypeptides RFRP-3 and GnRH
The relative expression of RFRP-3 gene increased with increasing dose of letrozole being the highest at 1.25 mg dose; however, GnRH gene expression decreased due to letrozole treatment (P <0.0001; Figure 8).

Discussion
The delayed parturition may be due to the reduced level of estrogen induced by letrozole treatment.In fact, high levels of estrogen are necessary for initiation of parturition (18).Letrozole administration at 0.002 or 0.02 mg/kg BW per day during 15-21 GDs delayed parturition in rat (19).Reduced litter size was observed in 1.25 mg/kg group, suggesting that letrozole at high doses has severe feto toxic effects.In a study by Zhang et al (20), co-treatment of pregnant rats with DHT (1.66 mg/kg BW per day from 7.5-13.5 GDs) and insulin resulted in fetal death and subsequently reduced litter size.Treatment of pregnant rats with 0.02 mg/kg letrozole at 15-21 GDs caused fetal mortality (19).Letrozole administration during organogenesis (6-16 GDs) in rats at 0.01, 0.02 and 0.04 mg per kg doses resulted in post implantation loss that included early and late resorption and decreased number of viable fetuses (21).These ndings suggested that embryo and feto toxic effects of letrozole depending on the time of administration, were dose-dependent (18).It seems these changes are caused by the letrozole action in reducing estrogen levels because simultaneous estrogen treatment at maximum dose of letrozole (0.04 mg/kg BW) restrained the feto toxic effects of letrozole (22).However, increased fetal mortality may be due to the direct effect of the androgen increment as a result of letrozole, as PCOS women have been reported to have higher perinatal fetal mortality rates than women with normal androgen levels (23).The shorter exposure time and higher letrozole dose was the main difference between our study and these studies; so it seems that the disrupting effects of letrozole are related to the speci c (organogenesis or fetal stage) time during pregnancy (18).and the administrated dose.In our study, the moderate effect of higher letrozole dose on fetal mortality could be due to administration at the late pregnancy, that is, letrozole had the greatest effect on fetal viability.Generally, the letrozole mediated embryo or feto toxic effects were dose dependent and the administered dose in pregnant rats in at least 1% of the maximum dose prescribed to humans on a daily basis resulted in fetal or embryonic death (24).Due to the fact that estrogen is greatly increased in the last third of rat pregnancy (24) and rodent ovaries are the prominent site of estrogen biosynthesis throughout pregnancy (18), inhibition of estrogen synthesis in the ovary in late pregnancy causes adverse effects on maternal and fetal aspects of pregnancy in rats as observed in our study with increasing dose.
This work highlighted delayed puberty as a result of 1.25 and 1 mg/kg BW prenatal letrozole treatment.Commonly, vaginal opening in rats occurs at 28-49 postnatal days (PNDs), but the age of puberty onset and beginning of sexual maturation is different according to the species and growth rate (25).Delayed puberty in our study, may be due to increased Rfrp expression.Similarly, in a study by Han et al (26) it was shown that intracerebroventricular injection of RFRP-3 between days 28 and 36, at the time of puberty onset, delayed vaginal opening in rats.Furthermore, Rfrp expression in DMN decreased during early pre-pubertal stages in the mouse.This reduces its inhibitory effect on GnRH neurons (27).Similarly, GPR147 (the RFRP-3 receptor) knockout on GnRH neurons caused delayed puberty in mice (27).The increased expression of Rfrp may be due to elevated androgen levels during the intrauterine life but this hypothesis should be investigated further by evaluating the AR on RFRP-3 neurons or by blocking the direct effects of androgen by utamide.
The letrozole androgenized female rats showed longer anogenital distance (AGD) at puberty than untreated groups.Furthermore, we observed increased AGD in all treated groups.The anogenital distance (AGD) and anogenital distance index (AGDI) are both main indicators of maternal hyperandrogenism (28).In general, AGD is longer in the male rats than in the female ones, so ADG in female re ects the degree of uterine hyperandrogenism that she experienced during the intrauterine life and the higher doses of androgen exposure during fetal life led to the longer AGD in female offspring.Consistent with our data, longer AGD caused by prenatal testosterone administration on 20 GD was reported (29).In addition, the long AGD in prenatal T and DHT (on 16-19 GDs) female rats was shown (30).Increased AGD re ects the androgenic effect of letrozole during the time of external genitalia differentiation.External genitalia differentiation in female rat occurs during late pregnancy (19 -20 GDs), underlying the direct effect of dihydrotestosterone (31).DHT is a non-aromatizable androgen and acts like an aromatase inhibitor, therefore, letrozole and DHT have the same effect on AGD changes mediated by androgen increments.These ndings suggest that androgen action during embryonic life, especially at the time of external genitalia differentiation, affects AGD, which is a good indicator of direct effect of androgens.
Prenatal letrozole administration at 1.25, 1.0 and 0.75 mg/kg BW signi cantly increased the body weight gain compared with other groups.These body weight gain increments were greater than control groups at 6 to 8 postnatal weeks, suggesting the metabolic effects of prenatal letrozole appeared at adulthood.Increased body weight gain as a metabolic feature of PCOS has been con rmed in various adult letrozole induced PCOS models in the rat (8, [32][33][34].and also in DHT PCOS induction model in mice (35).Arroyo et al. (36) reported that prepubertal letrozole model caused increased body weight gain in mice which did not improve with letrozole removal, contrary to the reproductive traits induced in this model that were completely recovered.However, in another study, weight gain caused by adult letrozole PCOS model was reversible by utamide in mice (37).These ndings suggested that metabolic alterations such as body weight gain by prenatal or prepubertal origins were more permanent.
In the current study, all letrozole treated groups exhibited acyclicity.Moreover, the delay in initiation of cycles ( rst proestrus) and reduced number of females that were able to complete one or more cycles were observed in all treated groups.In our study, the total number of growing follicles, atretic follicles, cystic follicles and corpora lutea was higher at 0.25 mg/kg BW letrozole.Irregular cycles and ovulation dysfunction are the usual effect of exposure to androgen excess during intrauterine life (38, 39).It seems that anovulatory cycles arise from androgen mediated functions, because utamide as an androgen receptor antagonist could recover ovulatory cycles in prenatal androgenized female mice (40).Irregular cycles, fewer corpora lutea and ovarian cysts formation were shown as a result of prenatal DHT treatment in wild-type mice, but these effects were not observed in AR knockout mice indicating the key role of androgen signaling in creating PCOS-like features (41).Increased number of antral and preantral follicles was observed in prenatal rats exposed to androgen on 16-19 GDs but not on 20 GD (29), indicating that the timing of androgen administration is important.Induction of testosterone and dihydrotestosterone at the end of gestation had no effect on polycystic ovarian development, although it induced irregular and anovulatory cycles (7,30).
In PCOS women, the high number of preantral and antral follicles results in antral cavity expansion, cyst formation, thin granulosa layers and thick theca layers (42).Similar to our results, fetal and postnatal treatment of testosterone in rat led to arrested follicular development in the small antral (preantral) follicle (43).In our prenatal letrozole model, the number of antral (small antral) follicles was higher.
Androgens stimulate the growth of small antral follicles ( 44), but at the advanced stages of follicular development they suppress the growth of follicles by granulosa apoptosis in the preovulatory follicles (45).Therefore, hyperandrogenism induced by letrozole severely impairs normal follicular development (46).Taken together, letrozole administration via increasing the levels of testosterone may in uence ovarian follicular development and nally disrupt ovarian cycles showing the direct effect of androgen action on the ovary.However, this effect of letrozole should be investigated using in vitro studies in which the ovaries of female fetuses are exposed to androgen excess by letrozole.It is also interesting to evaluate the recovery effect of utamide in prenatal letrozole models of PCOS to determine the de nitive effect of androgens.
The ndings of our research showed markedly elevated testosterone levels and reduced estradiol levels in 1.25 and 1 mg/kg BW groups.Serum FSH and LH levels were also higher in control groups showing that gonadotropins levels were not affected by prenatal androgenization via letrozole.However, progesterone concentration, in proportion to the higher number of corpora luteua in the 0.25 mg/kg BW group, was the highest in this group.Letrozole treatment in adult female rats for 14 days resulted in reduced estradiol levels (18).Increased LH levels, as a main feature of PCOS women, has also been observed in adult rodent letrozole models (47)(48)(49).In the present study, serum LH levels at 1.25 mg/kg BW were increased compared with other groups.In prenatal androgenized female rats, FSH levels were not affected by androgen action (29,30).In our study, Serum FSH concentration decreased in letrozole treated groups vs control groups.There have been inconsistencies in the baseline and pulse concentration of LH and FSH in various PCOS prenatal and clinical studies of PCOS women (7,40,50).These inconsistencies may be due to the different estrous phases of female rats during sampling, animal models and species, time of induction and other experimental conditions.In letrozole treated animals, increased endogenous androgen production by inhibiting aromatase function is unavoidable, and has been reported in different models of PCOS induction with letrozole (51)(52)(53).In our study, consistent with previous ndings, considerable increase in testosterone at 1.25 and 1 mg/kg BW was recorded, as well as decreases in estradiol levels.In daughters of women with PCOS, increased testosterone levels were reported during puberty but in general there is little information available (54).Serum estradiol levels did not change in PCOS induced by prenatal testosterone administration (29) and also adult letrozole model (33,48,53).
But in another studies, reduced estradiol levels were reported in letrozole treated PCOS adult rat model (32,52).Given that estradiol levels vary during the estrous cycle, and estradiol is at the lowest level in estrous phase, these differences are probably due to measurements taken at different times during the estrous cycle.On the other hand, variations in estradiol levels are the direct effects of letrozole due to reducing the conversion testosterone to estradiol.
In addition, decreased Gnrh gene expression as a result of prenatal letrozole treatment in female rats was proved.Letrozole induced mouse model of PCOS showed increased expression of gonadotropin releasing hormone receptor (Gnrhr) in the pituitary, that was not reversible by utamide treatment (37).Moreover, increased hypothalamic Gnrh and pituitary Gnrhr transcripts were observed in an adult rat model induced by 21 days 0.5 mg letrozole administration (53).In another adult letrozole PCOS model, no changes in Gnrh mRNA expression, but increased pituitary Gnrhr mRNA expression was observed (47).Moreover, an elevated number of AR and GnRH immunoreactive cells and AR mRNA expression were shown due to DHT induced PCOS in adult rats (55).The lack of prenatal androgenization PCOS studies evaluating Gnrh expression, and inconsistencies in studies of PCOS induction with letrozole in adulthood complicate the interpretation of results.However, the mechanism that indicates an increase in Gnrhr expression appears to be more in uential in the etiology of PCOS (47).On the other hand, puberty onset is controlled by high frequency GnRH neurons that affect FSH and LH release to trigger gonads for puberty initiation (26); therefore, the elevated Gnrh expression in our study in untreated group may be due to the normal neuroendocrine changes at puberty onset.At the same time, decreased Gnrh expression in letrozole treated groups led to late puberty.However, it would be better to investigate the direct effect of androgens using utamide during pregnancy on these neuroendocrine changes, which was not performed in our study due to budget constraints.
In addition, the impact of upstream mechanisms on GnRH control should not be overlooked.In prenatal androgenization mice models, the putative γ-aminobutyric acid GABAergic synaptic connections to GnRH neurons was elevated at adulthood; suggesting an increase in GnRH neurons pulse and frequency due to the effect of GABAergic neurons via mediating the negative feedback of steroids (40,(56)(57)(58).In addition, prenatal androgenized female rats by testosterone displayed decreased progesterone receptor (Pgr) mRNA expression in the hypothalamic POA indicating the capability of prenatal androgen in creating alteration in the GnRH neurosecretory system and neuroendocrine dysfunctions at adulthood such as infertility related to PCOS (39).Moreover, some abnormalities in KNDy neuropeptides secretion were shown in various animal models of PCOS; for instance, prenatal testosterone treated ewes showed KNDy expression abnormalities (59).In a study by Caldwell et al. (41), prenatal administration of DHT did not alter the KNDy neuropeptides.Furthermore, adult PCOS models are noteworthy in this eld.Letrozole treated adult female mice revealed an increased neuronal activation of Kiss1 (60).A DHT induced PCOS rat model resulted in decreased kiss1 gene expression, but the serum levels of testosterone, estradiol, LH, FSH were unaltered (61).Interestingly, in a letrozole induced PCOS model in adult rats, increased positivecell number of kisspeptin in the arcuate and decreased number of positive kisspeptin in anteroventral periventricular (AVPV) nucleus were reported (62).Kisspeptin neurons in the arcuate nucleus are involved in the negative feedback of estradiol on the GnRH/LH system.On the other hand, kisspeptin neuropeptides of the AVPV mediate preovulatory LH surge (63).Therefore, increased kiss1 gene expression in the arcuate nucleus can interfere with the PCOS pathology (62).In general, the effect of prenatal and adult administration of androgens on inhibitory and excitatory regulators of GnRH neurons is likely more important than the direct impact of androgens on these neurons.
Compared with other groups, increased expression of Rfrp gene in the hypothalamic DMN was observed at 1.25 and 1 mg/kg BW letrozole treatment.Prepubertal letrozole implants releasing 50 µg/day for 16 days before puberty did not impact on Rfrp expression in DMN; however, increased LH levels were found in adult female rats suggesting the role of other endogenous regulators of GnRH such as KNDy neuropeptides or GABAergic inputs rather than RFRP-3 neurons (60).Moreover, in our previous study, constant light induced PCOS rat model decreased Rfrp expression, along with unaltered FSH and LH serum levels in adult female rats (13).Furthermore, in neonatal testosterone treated female rats, decreased Rfrp mRNA expression was reported, without any effect on LH serum level (64).These studies revealed that serum LH concentration did not re ect the effect of Gnrh and Rfrp expression changes directly.In other words, the changes in the expression of these neuropeptides are probably not the main neuroendocrine mechanism for LH increase in PCOS women.On the other hand, it was shown that the intracerebroventricular injection of RFRP-3 decreased Gnrh mRNA expression in female rats (26), demonstrating the inhibitory effect of RFRP-3 on GnRH neurons.These results are consistent with our study.It has also been shown that the number of Rfrp expressing neurons in adulthood is lower than at birth in both sexes of mice (65).Therefore, alteration in Rfrp expression in control groups is normal for the ages after puberty onset, but in treated group, especially at 1.25 and 1 mg/kg BW, enhanced Rfrp expression may be as a result of excessive androgen production during intrauterine life.
In various models, alterations in Gnrh expression or basal or pulsatile levels of GnRH neuropeptide are different.This might be due to the fact that different neural pathways in uence the GnRH neurons (60).In spite of the fact that, there are limitations due to the lack of information on GnRH pulse frequency and amplitude, so the baseline value of serum LH levels cannot be attributed to the pulsatile GnRH secretion and also Gnrh gene expression.Given the fact GnRH secretion is affected by a set of stimulatory and inhibitory factors, including KNDy neuropeptides and gamma-aminobutyric acid (GABA)ergic neurons, further studies are needed to evaluate the changes in upstream GnRH regulators, especially in prenatal PCOS models.

Conclusion
In summary, the results of our study suggested that prenatal letrozole treatment at doses lower than 1 mg/kg BW is safe and has no detrimental embryotoxic and lethal effects on fetuses and mothers in Sprague-Dawley rats.A comprehensive dose response study was carried out to evaluate the alterations of major genes controlling reproductive phenomena, Gnrh and Rfrp, steroid hormones and gonadotropins and ovarian function in response to prenatal letrozole administration.We recorded irregularities in estrous cycles, follicular development changes, increased testosterone levels, decreased estradiol levels, decreased Gnrh, increased Rfrp expression at 1.25 and 1 mg/kg BW letrozole.Due to the greater effect of 1.25 mg/kg BW treatment on fetal mortality, we nally concluded that prenatal letrozole treatment at 1 mg/kg BW on 16-18 GDs in the rat, is probably optimal for PCOS induction by inhibiting aromatase and indirectly increasing endogenous androgen levels.In terms of future work, it would be interesting to examine the upstream and downstream pathways that control the GnRH and RFRP-3 neurons.Another possible area of future research would be to investigate whether there is a relationship between these neuroendocrine pathways and changes in androgen levels.
Tables Table 1.Real-time PCR primers for target and reference genes and annealing temperatures.

Figure 7 Prenatal
Figure 7 -ΔΔCT formula.Statistical analysisEvaluation of data normality was conducted by the Kolmogorov-Smirnov test.For multiple comparisons among groups one-way ANOVA and post-hoc Tukey was used.Binomial data were analyzed by the Chisquared test.Multiple comparisons were carried out by Kruskal-Wallis-H test for the data that were not normal.For weight gain values from letrozole-treated, control and sham groups were compared using repeated measure ANOVA.The differences among groups were considered statistically signi cant when P 0.05.For data analysis, the SPSS 22 for windows (IBM SPSS Statistics for Windows, version 22, IBM Inc., Chicago, Illinois) was used.The data are reported as mean ± standard error of the mean (SEM).Charts were created by GraphPad Prism version 5.01 for Windows (GraphPad Inc. Inc., San Diego, CA, USA).