Although polycystic ovary syndrome (PCOS) is among the most common endocrine disorders in women of reproductive age, its etiology remains poorly understood. From the perspective of developmental origins of health and disease, some studies have investigated the relationship between low birth weight and the prevalence of PCOS and/or PCOS phenotypes in humans; however, the results of these studies were inconclusive. Here, we evaluated the effects of prenatal undernutrition on the metabolic and reproductive phenotypes of dihydrotestosterone-induced PCOS model rats. The PCOS model rats showed increased body weight, food intake, fat weight, adipocyte size and upregulation of inflammatory cytokines in adipose tissue; prenatal undernutrition exacerbated these metabolic changes. Prenatal undernutrition also increased the gene expression of hypothalamic orexigenic factor and decreased the gene expression of anorexigenic factor in the PCOS model rats. In addition, the PCOS model rats exhibited irregular cyclicity, polycystic ovaries and disrupted gene expression of ovarian steroidogenic enzymes. Interestingly, prenatal undernutrition attenuated these reproductive changes in the PCOS model rats. Our results suggest that in dihydrotestosterone-induced PCOS model rats, prenatal undernutrition exacerbates the metabolic phenotypes, whereas it improves the reproductive phenotypes and that such phenotypic changes may be induced by the alteration of some peripheral and central factors.
Polycystic ovary syndrome (PCOS) is among the most common endocrine disorders in women of reproductive age, with an estimated prevalence of 5–16% depending on the ethnic and diagnostic criteria (Lauritsen et al. 2014, Rosenfield & Ehrmann 2016, Ding et al. 2017). Although this syndrome is characterized by a heterogeneous and complex presentation, the key symptoms are chronic anovulation, hyperandrogenism and polycystic ovaries (PCO) (Rotterdam ESHRE/ASRM Group 2004). PCOS is commonly associated with obesity and insulin resistance, which frequently induce metabolic disorders, such as type 2 diabetes mellitus and hyperlipidemia (Dunaif 1997, Moran et al. 2010a). The etiology of PCOS remains poorly understood, although the interaction of genetic and environmental factors may play a pivotal role in its pathogenesis (Franks et al. 2006, Rosenfield & Ehrmann 2016). Some experimental studies have indicated that an excess of prenatal androgen is a possible environmental factor related to the etiology of PCOS (Abbott et al. 2005, Shi & Vine 2012, Walters et al. 2012, Abbott et al. 2013), although it remains unknown whether this applies to human PCOS (Hickey et al. 2009).
Epidemiological and experimental studies have suggested that prenatal undernutrition and decreased birth weight affect the development of physiological and metabolic functions after birth and that these alterations may be linked to the pathogenesis of metabolic-related disorders in adulthood, especially under high nutrient conditions (Godfrey & Barker 2000, Breier et al. 2001). This concept, referred to as developmental origins of health and disease (DOHaD), may have important medical, biophysical and socioeconomic implications for the prevention of future diseases (Gluckman & Hanson 2004). Over the past decade, it has been proposed that the concept of DOHaD may be applicable to the evaluation of PCOS etiology because metabolic derangement with developmental origins is a cardinal feature of PCOS (Dunaif 1997, Moran et al. 2010a). Some human studies have investigated the relationship between birth weight and the prevalence of PCOS and/or PCOS phenotypes in women; however, the results of these studies were inconclusive. Some of the studies showed that a low birth weight was associated with PCOS phenotypes and hyperandrogenism in adulthood (Melo et al. 2010, Davies et al. 2012), whereas other epidemiological studies could not demonstrate any association between birth weight and reproductive or metabolic phenotypes in PCOS patients or their relatives (Laitinen et al. 2003, Legro et al. 2010). In addition, no study has yet evaluated the association between prenatal undernutrition and the pathogenesis or phenotypes of PCOS in animal models.
The physiological, metabolic and reproductive changes related to DOHaD can be reproduced in some animal models. In mice and rats, the offspring of undernourished mothers tend to exhibit obesity and insulin resistance in adulthood (Yura et al. 2005, Bieswal et al. 2006, Breton et al. 2008), and female offspring also show delayed puberty and premature reproductive senescence (Iwasa et al. 2010, Khorram et al. 2015). Various rodent models of PCOS have been generated by genetic modifications and exposure to androgens, estrogen, antiprogestins or aromatase inhibitor (Shi & Vine 2012, Walters et al. 2012). Among these models, chronically 5α-dihydrotestosterone (DHT)-administered rodents are commonly used because they present the typical features of PCOS, including anovulation, increased body weight (BW) and body fat, enlarged adipocytes and atypical follicles (Manneras et al. 2007, van Houten et al. 2012, Kim et al. 2013, Osuka et al. 2017).
To evaluate the epidemiological hypothesis that DOHaD might be related to the etiology of PCOS, we evaluated the effects of prenatal undernutrition on the phenotypes of DHT-induced PCOS model rats. We further investigated the mechanism behind such alterations by measuring the levels of serum hormones and gene expression in peripheral and central tissues.
Materials and methods
Eight pregnant Wistar rats (Charles River Japan, Kanagawa, Japan) were housed individually under controlled lighting (12-h light, 12-h darkness; lights were turned on at 08:00 and turned off at 20:00) and temperature (24°C) conditions with free access to food (type MF; Oriental Yeast Co., Ltd., Tokyo, Japan) and water. This study was approved by the institutional animal care and use committee of the University of Tokushima (No. T28-47) and all animal experiments were conducted in accordance with the ethical standards. The pregnant rats were divided into normally nourished (NN; n = 4) and undernourished (UN; n = 4) groups. From gestational days 14 to 21, the undernourished dams were restricted to 50% of the daily food intake (FI) of the NN dams of the previous day; subsequently, they were allowed to feed ad libitum during the lactation period. The day on which the pups were delivered was defined to be postnatal day (PND) 0, and the pups were randomly assigned to dams (11–12 animals per dam) and fostered until weaning at PND 21. After weaning, the pups were separated by sex and housed with three to four animals per cage. On PND 29, the female offspring from normally nourished dams (NN) and those from undernourished dams (UN) were randomly divided into either the chronically DHT-administered group or the non-DHT-administered (control) group for a total of four groups, i.e., the NN control, NN-DHT, UN control and UN-DHT groups (n = 9 or 10 per group). In the DHT-administered groups, rats were implanted with a silastic tube filled with crystalline DHT (As One Co., Ltd., Tokyo, Japan; inner diameter, 3 mm; outer diameter, 5 mm; length of the filling part, 5 mm) under sevoflurane-induced anesthesia. In the control group, rats were implanted with an empty tube. Rats were individually housed after surgery.
Assessment of BW and FI
The total BW of the NN and UN groups was measured just after birth and thereafter measured every week until PND 28. The data on BW at birth include the data from both males and females and those after PND 7 include only data from the females. After implantation, the total BW and FI of rats in the NN control, NN-DHT, UN control and UN-DHT groups were measured every week until they were killed. BW change (% of initial BW) was calculated as BW/initial BW × 100, and feed efficiency was calculated as total BW gained/cumulative FI.
Assessment of the estrous cycle
Vaginal epithelial smears were obtained over 12 days starting from 4 weeks after implantation, and the estrous cyclicity was determined by light microscopic analysis. The estrus stage was characterized by the presence of mostly cornified cells; the metestrus stage was characterized by the presence of cornified epithelial cells and leukocytes and the proestrus stage was characterized by the presence of mostly nucleated and some cornified epithelial cells.
Collection of blood and tissue samples
On PND 72 (6 weeks after implantation), rats were fasted for 16 h, anesthetized with sevoflurane, weighed and then the brain, blood, visceral fats, subcutaneous fats, liver and ovary were collected. The blood glucose concentration was measured with a glucometer (Glucose Pilot, Aventir Biotech, Carlsbad, CA, USA). The weights of the visceral fats (parametrial, perirenal and mesenteric depots), subcutaneous fats (inguinal depots) and ovaries (both sides) were measured immediately after removal. Tissue samples with approximately 300–400 mm3 of visceral (parametrial) and subcutaneous (inguinal) fats and liver were dissected. An ovary from one side and sections of visceral and subcutaneous fats and liver were fixed in 4% paraformaldehyde, while the ovary from the other side and other sections of visceral and subcutaneous fats and liver were immediately immersed into RNAlater (Ambion; Thermo Fisher Scientific) and stored at −80°C until used for quantitative PCR. Brains were flash-frozen using liquid nitrogen and stored at −80°C. Serum was separated by centrifugation and stored at −20°C.
The fixed ovaries and visceral and subcutaneous fats were dehydrated with ethanol and xylene and sliced into 4 μm thick sections after being embedded in paraffin. Sections were stained with hematoxylin and eosin, and histological findings were captured by a Zeiss Imager M2 microscope with AxioVision version 4.8 acquisition software (Zeiss). The mean adipocyte area of 50–100 randomly selected adipocytes per specimen was determined using ImageJ software. Cystic follicles and corpora lutea were counted in randomly selected sections from the middle of the ovaries.
Serum insulin levels were measured using a radioimmunoassay (rInsulin [I-125] RIA kit; Institute of Isotopes Co., Ltd., Tokyo, Japan); the sensitivity of the assay was 0.06 ng/mL. Serum leptin levels were measured using another radioimmunoassay kit (multi-species leptin RIA kit; Linco Research, Inc., St. Charles, MO, USA); the sensitivity of the assay was 1.0 ng/mL, and the inter- and intra-assay coefficients of variation (CVs) were 3.2% and 7.8%, respectively. Serum estradiol (E2) and testosterone levels were measured by a commercial laboratory (SRL, Tokyo, Japan) using an electrochemiluminescence immunoassay (Roche Diagnostics GmbH). Serum luteinizing hormone (LH) levels were measured using a radioimmunoassay (rLH [I-125] RIA kit; Institute of Isotopes Co., Ltd.); the sensitivity of the assay was 0.8 ng/mL, and the inter- and intra-assay CV were 7.7% and 6.5%, respectively. Serum follicle-stimulating hormone (FSH) levels were measured using a radioimmunoassay (rFSH [I-125] RIA kit; Institute of Isotopes Co., Ltd.); the sensitivity of the assay was 0.09 ng/mL and the inter- and intra-assay CV were 10.3 and 8.4%, respectively.
Quantitative real-time PCR
Whole hypothalamic explants were dissected from the frozen brains, as described previously (Iwasa et al. 2018). Briefly, the target region of the brain was dissected out via an anterior coronal cut at the posterior border of the optic chiasm, a posterior cut at the posterior border of the mammillary bodies, parasagittal cuts along the hypothalamic fissures and a dorsal cut 2.5 mm from the ventral surface. This hypothalamic block contained the arcuate nucleus and the lateral hypothalamic area. Total RNA was isolated from the hypothalamic explants, ovaries, visceral and subcutaneous fats and livers using a TRIzol reagent kit (Invitrogen) and an RNeasy mini kit (Qiagen GmbH). cDNA was synthesized with oligo (deoxythymidine) primers at 50°C using the SuperScript III first-strand synthesis system for real-time PCR (Invitrogen). The PCR analysis was performed using the StepOnePlus real-time PCR system and FAST SYBR green (Applied Biosystems). The PCR conditions were as follows: initial denaturation and enzyme activation at 95°C for 20 s followed by 45 cycles of denaturation, annealing and extension for 30 s. The primer sequences, product sizes and annealing temperatures used are shown in Table 1. The mRNA expression level of each molecule was normalized to that of glyceraldehyde 3-phosphate dehydrogenase, 18S rRNA or β-actin. Dissociation curve analysis was also performed to ensure that only one single product was present.
Primer sequences and annealing temperature.
|Primer||Sequence||Annealing T (°C)|
|IL-1β forward||GCT GTG GCA GCT ACC TAT GTC TTG||61|
|IL-1β reverse||AGG TCG TCA TCA TCC CAC GAG|
|TNF-α forward||AGC CCT GGT ATG AGC CCA TGT||65.5|
|TNF-α reverse||CCG GAC TCC GTG ATG TCT AAG T|
|IL-6 forward||TCC TAC CCC AAC TTC CAA TGC TC||67|
|IL-6 reverse||TTG GAT GGT CTT GGT CCT TAG CC|
|P450arom forward||CTC CTC CTG ATT CGG AAT TGT||63|
|P450arom reverse||TCT GCC ATG GGA AAT GAG AG|
|StAR forward||TCC GAG TAA ACG GTC TTA GTC GT||63|
|StAR reverse||GCG TTC CAC GTT GTT CTG TTC|
|P450scc||GAA CGA CCT GGT GCT TCG TAA||65|
|P450scc||GTG CGT GGT GTT TTG GCT TT|
|P450C17 forward||ACC TAG AGG CCA CAA CTA ACA TCC||64|
|P450C17 reverse||GAG GCA CTG GGA CTA GCA CCT|
|SF-1 forward||GCT GTG TGT TTG GGA TGA TG||63|
|SF-1 reverse||AGA CGG AGG AAG GAG TGG TT|
|3β-HSD forward||ATG CCC AGT ACC TGA GGA GA||62|
|3β-HSD reverse||TTG AGG GCC GCA AGT ATC A|
|17β-HSD forward||TCA CGA TTG GAG CTG AAA CCT||64|
|17β-HSD reverse||TCG CCA AGA TTT CAT GAG CA|
|AR forward||GGG GCA ATT CGA CCA TAT CTG||66|
|AR reverse||CCC TTT GGC GTA ACC TCC CTT|
|LHr forward||AAC TCA GTG GCT GGG ATT ATG||62|
|LHr reverse||CCA AAT CAG GAC CCT AAG GAA G|
|NPY forward||GGG GCT GTG TGG ACT GAC CCT||66|
|NPY reverse||GAT GTA GTG TCG CAG AGC GGA G|
|pp-orexin forward||GCC GTC TCT ACG AAC TGT TG||60|
|pp-orexin reverse||CGA GGA GAG GGG AAA GTT AG|
|OBRb forward||GCA GCT ATG GTC TCA CTT CTT TTG||63|
|OBRb reverse||GTT CCC TGG GTG CTC TGA|
|5HT2CR forward||AGA AAG AAA AGC GTC CCA GAG||63|
|5HT2CR reverse||CCA CAA AGA ACC GAC AGG ATA|
|Kiss1 forward||ATG ATC TCG CTG GCT TCT TGG||65|
|Kiss1 reverse||GGT TCA CCA CAG GTG CCA TTT T|
|GnRH forward||GCA GAA CCC CAG AAC TTC GA||64|
|GnRH reverse||TGC CCA GCT TCC TCT TCA AT|
|GAPDH forward||ATG GCA CAG TCA AGG CTG AGA||70|
|GAPDH reverse||CGC TCC TG GAA GAT GGT GAT|
|18S rRNA forward||GAC GGA CCA GAG CGA AAG C||64|
|18S rRNA reverse||AAC CTC CGA CTT TCG TTC TTG A|
|β-Actin forward||CCG TAA AGA CCT CTA TGC CAA CA||66|
|β-Actin reverse||GCT AGG AGC CAG GGC AGT AAT|
Data are presented as the mean ± standard error (s.e.). Statistical analysis was performed using one-way or two-way ANOVA followed by the Tukey–Kramer post hoc test or Student’s t test. P values <0.05 were considered to be statistically significant in all statistical comparisons.
Maternal BW change, BW and BW change of female offspring
The BW of UN dams was significantly lower than that of the NN dams on gestational day 21 (Fig. 1A). Similarly, the BW change of UN dams was significantly smaller than that of NN dams from gestational days 15 to 21 (Fig. 1B). The BW (data of both sexes included) of the UN group was significantly lower than that of the NN group (Fig. 1C), indicating that our protocol of maternal food restriction might successfully induce the prenatal undernourished condition and subsequent intrauterine growth retardation of offspring. Although the BW (only female data included) of the UN group was significantly lower than that of the NN group from PND 7 to 21, the BW of these two groups were not different at PND 28.
BW change, cumulative FI and feed efficiency during DHT administration
To evaluate whether prenatal undernutrition affects the metabolic phenotypes of PCOS model rats, BW, BW change, cumulative FI and feed efficiency were measured during DHT administration. Although DHT administration increased the BW and BW change in both NN and UN groups, these changes were greater in the UN group than those in the NN group (Fig. 2A and B). Similarly, DHT administration increased the cumulative FI in both NN and UN groups; however, these changes were not different between these two groups (Fig. 2C). In contrast, DHT administration increased feed efficiency, which is the marker of metabolic profiles, in both NN and UN groups, and these changes were greater in the UN group than those in the NN group (Fig. 2D). These results suggest that prenatal undernutrition might exacerbate the BW gain in PCOS model rats by the increment of the conversion from nutrients to body fat composition.
Visceral and subcutaneous fat weights and adipocyte size
To evaluate whether exacerbated BW gain in prenatally undernourished PCOS model rats is attributed to increased adiposity, body fat weight and adipocyte size were measured. DHT administration increased visceral fat weight in the UN group, whereas it did not affect the NN group (Fig. 3A). In contrast, although DHT administration increased subcutaneous and total fat weight in both NN and UN groups, these changes were greater in the UN group than in the NN group (Fig. 3A). The effects of DHT administration on the visceral and subcutaneous adipocyte sizes were similar to those on the fat weights. Namely, DHT administration increased the adipocyte size of visceral fat in the UN group, whereas it did not affect the NN group (Fig. 3B). In contrast, DHT administration increased the adipocyte size of subcutaneous fat in both NN and UN groups; however, these changes were not different between these two groups (Fig. 3C). These results indicate that exacerbated BW gain in prenatally PCOS model rats was caused by the accumulation of body fats and that these changes might be caused, at least in part, by the enlargement of adipocyte size.
Serum leptin and insulin levels and gene profiles in peripheral tissues
It has been revealed that insulin resistance, decreased sensitivity to leptin and chronic inflammation of peripheral tissues are related to the metabolic disturbance of PCOS. Thus, to evaluate whether prenatal undernutrition affects metabolic and inflammatory conditions in PCOS model rats, serum hormonal levels and gene profiles of inflammatory cytokines in peripheral tissues were measured. DHT administration increased serum leptin level in the UN group, whereas it did not affect the NN group (Fig. 4A). However, this increase in serum leptin level, BW gain and adiposity was not attenuated in DHT-administered rats in the UN group, indicating that sensitivity to the anorectic action of leptin might be decreased in these rats. In contrast, DHT administration did not affect fasting insulin and glucose levels in both NN and UN groups (Fig. 4A). DHT administration increased the mRNA levels of interleukin (IL)-1, tumor necrosis factor (TNF)-α and IL-6 in visceral fat in the UN group, whereas it did not affect the NN group (Fig. 4B). Similarly, DHT administration increased the IL-1 and IL-6 mRNA levels in subcutaneous fat in the UN group, whereas it did not affect the NN group (Fig. 4C). Conversely, DHT administration did not affect the hepatic IL-1, TNF-α and IL-6 mRNA levels in both NN and UN groups (Fig. 4D). These results suggest that prenatal undernutrition might elicit chronic inflammation in adipose tissue and attenuate the sensitivity to leptin in PCOS model rats.
Estrous cyclicity and ovarian morphology
To evaluate whether prenatal undernutrition affects the reproductive phenotypes of PCOS model rats, estrous cyclicity during DHT administration and ovarian morphology were measured. Almost all non-DHT-administered (control) rats showed a regular 4- to 5-day estrous cycle in both NN and UN groups (Fig. 5A). In contrast, 80% of DHT-administered rats in the NN group showed an acyclic or irregular cycle, whereas only 22.2% of DHT-administered rats in the UN group had an irregular cycle (Fig. 5A). DHT administration decreased and increased the number of estrus stage and diestrus stage during 12 days, respectively, in the NN group, whereas it did not affect the UN group (Fig. 5B). The ovaries of the non-DHT-administered (control) rats had normal morphology in both NN and UN groups, whereas those of the DHT-administered rats exhibited polycystic morphology (Fig. 6A). DHT administration decreased the ovarian weight and the number of corpora lutea and increased the number of cystic follicles in the NN group, whereas it did not affect the UN group (Fig. 6B). These results indicate that prenatal undernutrition improves the reproductive phenotypes, such as abnormal cyclicity and cystic change of ovaries, observed in PCOS model rats.
Serum gonadotropin and gonadal hormone levels and ovarian gene profiles
To evaluate the underlying mechanism by which prenatal undernutrition improves the reproductive phenotypes in PCOS model rats, serum hormonal levels and gene profiles of ovarian steroidogenic enzymes were measured. DHT administration decreased serum LH levels in the UN group, whereas it did not affect the NN group (Fig. 7A). In contrast, DHT administration did not affect serum FSH, testosterone and E2 levels in both NN and UN groups. DHT administration decreased the ovarian steroidogenic acute regulatory protein (StAR) mRNA level in the NN group, whereas it did not affect the UN group (Fig. 7B). Similarly, DHT administration decreased the ovarian cholesterol side-chain cleavage enzyme (P450scc) mRNA levels in both NN and UN groups; however, these changes were greater in the NN group than in the UN group (Fig. 7B). In contrast, DHT administration increased the ovarian steroidogenic factor-1/nuclear receptor subfamily 5, group A, member 1 (SF-1/NR5A1) and 17β-hydroxysteroid dehydrogenase (17β-HSD) mRNA levels in the NN group, whereas it did not affect the UN group (Fig. 7B). DHT administration increased the ovarian luteinizing hormone receptor (LHR) mRNA level in the UN group, whereas it did not affect the NN group (Fig. 7B). These results indicate that prenatal undernutrition might attenuate DHT-induced disturbances of ovarian steroid biosynthesis and that these alterations might improve the reproductive phenotypes in PCOS model rats. In contrast, pathophysiological roles of the effects of prenatal undernutrition on DHT-induced changes of serum LH and ovarian LHR mRNA levels remain unclear.
Hypothalamic gene profiles
Although some hypothalamic factors play pivotal roles in the regulation of metabolic, appetite and reproductive functions, these factors have not been fully examined in the PCOS model. To evaluate whether prenatal undernutrition affects hypothalamic factors of PCOS model rats, gene profiles of hypothalamic orexigenic factors (NPY and pp-orexin), anorexigenic factors (OBRb and 5HT2CR) and reproductive factors (Kiss1 and GnRH) were measured. Although DHT administration did not affect the hypothalamic pp-orexin mRNA levels, such levels were significantly higher in the UN group than in the NN group (Fig. 8). DHT administration increased the hypothalamic OBRb mRNA level in the NN group, whereas it did not affect the UN group. Although DHT administration did not affect hypothalamic 5HT2CR mRNA levels in both NN and UN groups, its level in DHT-administered rats in the UN group was significantly lower than that in DHT-administered rats in the NN group. DHT administration attenuated hypothalamic Kiss1 mRNA levels in both NN and UN groups, whereas it did not affect hypothalamic GnRH mRNA levels. These results indicate that prenatal undernutrition tends to increase orexigenic factors and attenuate or prohibit the increase of anorectic factors in PCOS model rats.
PCOS is a complex disorder that includes reproductive and metabolic dysfunction, and the phenotypes of human PCOS vary among patients. Although the etiology of PCOS remains poorly understood, the interaction of genetic and environmental factors may play a pivotal role in its pathogenesis (Franks et al. 2006, Rosenfield & Ehrmann 2016). Some epidemiological studies have evaluated the relationship between BW and the prevalence and/or phenotypes of PCOS in women; however, the results of these studies were inconclusive (Laitinen et al. 2003, Legro et al. 2010, Melo et al. 2010, Davies et al. 2012). Because the effects of bias and confounding variables cannot be completely eliminated from epidemiological studies and because reproductive and metabolic organs cannot be explored in human studies, animal studies are useful for the evaluation of the etiology and pathophysiological mechanisms of PCOS. Here, we examined, for the first time, the effects of prenatal undernutrition on the phenotypes of PCOS and their underlying mechanisms using DHT-induced PCOS model rats. We found that prenatal undernutrition exacerbated the metabolic phenotypes, whereas it improved the reproductive phenotypes, in the PCOS model rats. In addition, prenatal undernutrition affected the gene expression levels of adipose inflammatory cytokines, ovarian steroidogenic enzymes, and hypothalamic appetite regulation factors in the PCOS model rats.
Effects of prenatal undernutrition on the metabolic phenotypes
PCOS is commonly associated with metabolic perturbations, such as obesity and insulin resistance (Dunaif 1997, Moran et al. 2010a,b), and hyperandrogenemia has been shown to play potential roles in the development of these metabolic disorders. Metabolic disorders are reported to occur predominantly in PCOS patients with hyperandrogenism and oligomenorrhea, regardless of BW (Barber et al. 2007, Moghetti et al. 2013). In addition, hyperandrogenemia has been shown to associate positively with insulin resistance in PCOS patients of all ages (Livadas et al. 2014). Accordingly, chronically androgen-treated animals that have such metabolic and reproductive features have been used as PCOS animal models (Manneras et al. 2007, Shi & Vine 2012, van Houten et al. 2012, Walters et al. 2012, Kim et al. 2013, Osuka et al. 2017). In the present study, the metabolic phenotypes of prenatally NN PCOS model rats were similar to those of PCOS patients, i.e., they showed increased BW gain and FI, increased fat weight and enlarged adipocytes. In addition, the gene expression levels of adipose inflammatory cytokines, IL-1, TNF-α and IL-6, were increased in the PCOS model rats. Because inflammation has been shown to play pivotal roles in the pathogenesis of PCOS (Rosenfield & Ehrmann 2016), chronic low-grade inflammation might be related to the adiposity seen in our PCOS model.
Intriguingly, in the present study, the metabolic phenotypes and adipose inflammation were exacerbated by prenatal undernutrition in our PCOS model rats, whereas they remained unchanged in the control rats. Furthermore, the gene expression levels of some hypothalamic factors were also altered by prenatal undernutrition. The gene expression level of pp-orexin, an orexigenic factor, was increased, whereas that of 5HT2CR, an anorexigenic factor, tended to be decreased in the prenatally UN PCOS model rats, suggesting that these changes may contribute to increased FI and BW gain in this group. Recently, it was reported that central leptin resistance was observed in PCOS model rats, i.e., central injection of leptin could not reduce the FI and BW of letrozole-induced PCOS model rats (Lian et al. 2016). Similarly, although the serum leptin level was increased in prenatally UN PCOS model rats, their gene expression levels of hypothalamic neuropeptide Y and OBRb remained unchanged, indicating that hypothalamic sensitivity for leptin might be decreased in this group. In contrast, metabolic phenotypes in prenatally UN control rats were not different from those in prenatally NN control rats.
Taken together, these results suggested that prenatal undernutrition is not directly involved in the onset of PCOS metabolic phenotypes, but it may accelerate the development of the metabolic phenotypes in PCOS patients. In other words, hyperandrogenism plays a role as an aggravating factor of metabolic derangement with developmental origins and may increase the risk of onset of diseases related to DOHaD.
Effects of prenatal undernutrition on the reproductive phenotypes
Chronic anovulation and PCO are typically required for the clinical diagnosis of PCOS (Rotterdam ESHRE/ASRM Group 2004). Elevated LH levels and relatively decreased FSH levels are also frequently observed in PCOS patients (Blank et al. 2006, Iwasa et al. 2009). Some of these abnormal reproductive features, but not all, can be reproduced in PCOS model rats (Walters et al. 2012, Abbott et al. 2013). Specifically, it has been reported that chronic anovulation and PCO, but not the serum gonadotropin levels, can be reproduced in chronically DHT-administered PCOS models (Caldwell et al. 2014, Osuka et al. 2017). In the present study, the reproductive phenotypes of the prenatally NN PCOS model rats were in agreement with those reported in previous studies, i.e., they showed an irregular estrous cycle, fewer corpora lutea, more cystic follicles and a decreased ovary size. In addition, the gene expression levels of ovarian steroidogenic enzymes were altered in the PCOS model rats. Specifically, the gene expression levels of P450scc and StAR, both of which play roles in the conversion of cholesterol to pregnenolone, were decreased, whereas the gene expression level of 17β-HSD was increased in the PCOS model rats. Thus, in PCOS model rats, the initial steps of ovarian steroid biosynthesis may be disturbed by the decreased levels of P450scc and StAR, and the upregulation of 17β-HSD gene expression may be a compensatory mechanism for maintaining proper steroidogenesis. Because SF-1/NR5A1, a member of the orphan nuclear receptor transcription factor family, positively regulates the expressions of P450scc and StAR (Sunagawa et al. 1996, Caron et al. 1997, Parker & Schimmer 1997, Lalli et al. 1998) and a heterozygous point mutation in SF-1/NR5A1 has been reported in a PCOS woman (Calvo et al. 2001), we speculated that disturbances of ovarian steroidogenic factors observed in PCOS model rats have been induced by the reduction of SF-1/NR5A1 action. However, contrary to our expectation, the gene expression level of SF-1/NR5A1 was increased in the PCOS model rats, indicating that this change might be a compensatory response to recover the expressions of P450scc and StAR genes. Although some past studies have shown that the gene expression level of ovarian aromatase changes in letrozole- or DHEA-induced PCOS models (Kauffman et al. 2015, Yuan et al. 2016), such changes could not be reproduced in our DHT-induced PCOS model rats. Because it has been well established that the androgenic effects of DHT are much stronger than those of DHEA, discrepancies in results between our DHT-induced PCOS model rats and DHEA-induced PCOS model rats used in previous studies may be attributed to the difference in the kinds of injected androgens. In addition, the hypothalamic Kiss1 gene expression level was decreased, whereas the serum LH, FSH, testosterone and E2 levels remained unchanged in our PCOS model rats. These results partially differed from those reported in a past study, which showed that there were no changes in the hypothalamic kisspeptin level in a DHT-induced PCOS model (Osuka et al. 2017). In the present study, we used DHT-filled silicon tubes for chronic DHT administration, whereas Osuka et al. used the commercially available DHT pellet. Thus, it is possible that the amount of DHT released from devices may be different between our and their study and that these differences may explain the discrepancies in the results.
In contrast to the metabolic phenotypes, the onset of abnormal reproductive phenotypes and changes in ovarian steroidogenic enzymes were attenuated by prenatal undernutrition in the PCOS model rats. Although the polycystic ovarian morphology was observed in the prenatally UN PCOS model rats, ovarian weight, estrous cyclicity and gene expression levels of ovarian steroidogenic enzymes did not statistically differ from those of the non-DHT-administered control rats. In addition, gene expression of SF-1/NR5A1 in the prenatally UN PCOS model rats did not differ from that of the non-administered control rats.
Taken together, these results suggested that prenatal undernutrition may affect the reproductive phenotypes of PCOS patients, and it might play a role in the preservation of their fertility, even in hyperandrogenic conditions. Although it has been reported that elevated serum testosterone levels are commonly observed in PCOS patients (McCartney & Marshall 2016), such changes were not observed in our PCOS model rats, regardless of the prenatal nutritional conditions. These results of the present study correspond with those of a previous study using the DHT-induced PCOS model rats (Osuka et al. 2017). Thus, administration of DHT, which cannot be converted to testosterone, may directly affect the metabolic and reproductive phenotypes in PCOS model rats, and changes of the phenotypes in prenatally UN PCOS rats might be caused by the alteration of the sensitivities to administered DHT.
From an evolutionary perspective, some studies have indicated that PCOS might be a compensatory factor that provides a survival advantage under prehistoric conditions of undernourishment (Fessler et al. 2016, Unluturk et al. 2016), namely, insulin resistance protected from excessive protein loss and increased the possibility of long-term survival during prolonged starvation (Cahill 2006). In addition, decreased fecundity due to chronic ovulatory disorders might have provided a survival advantage by decreasing the number of children produced and mouths to feed, thus enabling more food to be allocated to the mother and existing children during conditions of starvation (Escobar-Morreale et al. 2005, Corbett & Morin-Papunen 2013). Genetic studies have demonstrated that PCOS may have existed since approximately 50,000 years ago, suggesting that the genetic factors involved in PCOS might be beneficial for survival (Unluturk et al. 2016). It has been assumed that these defense mechanisms were unable to adapt to the rapid improvement in nutritional conditions that has occurred over the past few decades, and they became pathogenic factors of metabolic and reproductive diseases. The data from the present study indicate that not only long-term evolutional changes but also changes with developmental origins that are induced by the short-term prenatal nutritional condition may affect the phenotypes of PCOS in adulthood. It is possible that the exacerbated metabolic phenotypes observed in prenatally UN PCOS model rats were types of adaptations for expected conditions of starvation after birth. However, we could not fully explain why prenatal undernutrition improves the reproductive phenotypes in the PCOS model rats, because, from an evolutionary perspective, fecundity might be disadvantageous for survival in conditions of undernourishment. One possibility is that the effects of reduced fecundity on survivability might differ on population and individual levels.
In summary, we demonstrated for the first time that prenatal undernutrition exacerbated the metabolic phenotypes and improved the reproductive phenotypes of DHT-induced PCOS model rats and that such phenotypic changes may be induced by the alteration of some metabolic- and reproductive-modulating factors in peripheral and central tissues. These results indicated that the concept of DOHaD, in addition to evolutionary perspectives, should be taken into account when evaluating the etiology and pathophysiology of PCOS.
Declaration of interest
The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the research reported.
This study was supported by JSPS KAKENHI Grant Number 18K09292.
Author contribution statement
T Iwasa and T Matsuzaki designed the study; T Iwasa and Y Mayila performed the experiments; T Iwasa, K Yano, A Kuwahara, M Irahara, Y Yamamoto and R Yanagihara analyzed the data and T Iwasa wrote the paper.
BarberTMWassJAMcCarthyMIFranksS 2007 Metabolic characteristics of women with polycystic ovaries and oligo-amenorrhea but normal androgen levels: implications for the management of polycystic ovary syndrome. Clinical Endocrinology 66 513–517. (https://doi.org/10.1111/j.1365-2265.2007.02764.x)
BretonCLukaszewskiMARisoldPYEnacheMGuillemotJRiviereGDelahayeFLesageJDutriez-CastelootILaborieC, 2008 Maternal prenatal undernutrition alters the response of POMC neurons to energy status variation in adult male rat offspring. American Journal of Physiology-Endocrinology and Metabolism 296 E462–E472. (https://doi.org/10.1152/ajpendo.90740.2008)
CahillGFJr 2006 Fuel metabolism in starvation. Annual Review of Nutrition 26 1–22. (https://doi.org/10.1146/annurev.nutr.26.061505.111258)
CaldwellASMiddletonLJJimenezMDesaiRMcMahonACAllanCMHandelsmanDJWaltersKA 2014 Characterization of reproductive, metabolic, and endocrine features of polycystic ovary syndrome in female hyperandrogenic mouse models. Endocrinology 155 3146–3159. (https://doi.org/10.1210/en.2014-1196)
CalvoRMAsuncionMTelleriaDSanchoJSan MillanJLEscobar-MorrealeHF 2001 Screening for mutations in the steroidogenic acute regulatory protein and steroidogenic factor-1 genes, and in CYP11A and dosage-sensitive sex reversal-adrenal hypoplasia gene on the X chromosome, gene-1 (DAX-1), in hyperandrogenic hirsute women. Journal of Clinical Endocrinology and Metabolism 86 1746–1749. (https://doi.org/10.1210/jcem.86.4.7424)
HickeyMSlobodaDMAtkinsonHCDohertyDAFranksSNormanRJNewnhamJPHartR 2009 The relationship between maternal and umbilical cord androgen levels and polycystic ovary syndrome in adolescence: a prospective cohort study. Journal of Clinical Endocrinology and Metabolism 94 3714–3720. (https://doi.org/10.1210/jc.2009-0544)
IwasaTMatsuzakiTMurakamiMShimizuFKuwaharaAYasuiTIraharaM 2009 Reproducibility of luteinizing hormone hypersecretion in different phases of the menstrual cycle in polycystic ovary syndrome. Journal of Obstetrics and Gynaecology Research 35 514–519. (https://doi.org/10.1111/j.1447-0756.2008.00998.x)
IwasaTMatsuzakiTMurakamiMFujisawaSKinouchiRGereltsetsegGKuwaharaAYasuiTIraharaM 2010 Effects of intrauterine undernutrition on hypothalamic Kiss1 expression and the timing of puberty in female rats. Journal of Physiology 588 821–829. (https://doi.org/10.1113/jphysiol.2009.183558)
IwasaTMatsuzakiTYanoKYiliyasiMIraharaM 2018 Effects of dihydrotestosterone administration on the expression of reproductive and body weight regulatory factors in ovariectomized and estradiol-treated female rats. Gynecological Endocrinology 34 73–77. (https://doi.org/10.1080/09513590.2017.1337096)
KauffmanASThackrayVGRyanGETolsonKPGlidewell-KenneyCASemaanSJPolingMCIwataNBreenKMDulebaAJ, 2015 A novel letrozole model recapitulates both the reproductive and metabolic phenotypes of polycystic ovary syndrome in female mice. Biology of Reproduction 93 1–12. (https://doi.org/10.1095/biolreprod.115.131631)
KimJYXueKCaoMWangQLiuJYLeaderAHanJYTsangBK 2013 Chemerin suppresses ovarian follicular development and its potential involvement in follicular arrest in rats treated chronically with dihydrotestosterone. Endocrinology 154 2912–2923. (https://doi.org/10.1210/en.2013-1001)
LaitinenJTaponenSMartikainenHPoutaAMillwoodIHartikainenALRuokonenASovioUMcCarthyMIFranksS, 2003 Body size from birth to adulthood as a predictor of self-reported polycystic ovary syndrome symptoms. International Journal of Obesity and Related Metablic Disorder 27 710–715. (https://doi.org/10.1038/sj.ijo.0802301)
LauritsenMPBentzenJGPinborgALoftAFormanJLThuesenLLCohenAHougaardDMAndersenAN 2014 The prevalence of polycystic ovary syndrome in a normal population according to the Rotterdam criteria versus revised criteria including anti-Mullerian hormone. Human Reproduction 29 791–801. (https://doi.org/10.1093/humrep/det469)
LegroRSRollerRLDodsonWCStetterCMKunselmanARDunaifA 2010 Associations of birth weight and gestational age with reproductive and metabolic phenotypes in women with polycystic ovarian syndrome and their first-degree relatives. Journal of Clinical Endocrinology and Metabolism 95 789–799. (https://doi.org/10.1210/jc.2009-1849)
LianYZhaoFWangW 2016 Central leptin resistance and hypothalamic inflammation are involved in letrozole-induced polycystic ovary syndrome rats. Biochemical and Biophysical Research Communications 476 306–312. (https://doi.org/10.1016/j.bbrc.2016.05.117)
LivadasSKolliasAPanidisDDiamanti-KandarakisE 2014 Diverse impacts of aging on insulin resistance in lean and obese women with polycystic ovary syndrome: evidence from 1345 women with the syndrome. European Journal of Endocrinology 171 301–309. (https://doi.org/10.1530/EJE-13-1007)
MoghettiPTosiFBoninCDi SarraDFiersTKaufmanJMGiagulliVASignoriCZambottiFDall’AldaM, 2013 Divergences in insulin resistance between the different phenotypes of the polycystic ovary syndrome. Journal of Clinical Endocrinology and Metabolism 98 E628–E637. (https://doi.org/10.1210/jc.2012-3908)