Maternal obesity causes a wide range of impairment in offspring, such as metabolic and reproductive dysfunctions. We previously demonstrated that female offspring of obese rats have increased serum estradiol levels during early postnatal life, probably because of decreased hepatic cytochrome P450 3A2 levels, which could lead to early onset of puberty and polycystic ovary condition in adulthood. Using metformin during pregnancy and nursing to improve the metabolic status of obese mothers could prevent the sequence of events that lead to an increase in postnatal serum estradiol levels in female offspring and, hence, reproductive dysfunction. We found that metformin prevented an increase in serum estradiol levels at postnatal day 14 in female offspring of obese mothers, which was associated with a restoration of hepatic cytochrome P450 3A2 levels to control values. Treatment using metformin could not prevent advanced puberty, but we observed that the number of antral follicles, follicular cysts and multi-oocyte follicles returned to control values in the female offspring of obese mothers treated with metformin. We also observed an increase in the levels of norepinephrine and the norepinephrine metabolite 3-methoxy-4-hydroxyphenylglycol in the ovaries, indicating increased sympathetic activity in female offspring induced by an obesogenic uterine environment. We found that this effect was prevented by metformin administration. From the results of this study, we concluded that metformin administration to obese mothers during pregnancy and nursing partially prevents ovarian dysfunction in female offspring during adulthood.
Maternal obesity is associated with various endocrine and metabolic disorders in offspring, including obesity (Paliy et al. 2014), hepatic steatosis (Oben et al. 2010, Mouralidarane et al. 2013), diabetes mellitus (Catalano & de Mouzon 2015) and cardiovascular disease (Taylor et al. 2014, Roberts et al. 2015). In rodents, maternal obesity induced by high-fat diet (HFD) is also associated with reproductive dysfunction in the offspring, such as early onset of puberty, impaired follicular development, altered estrous cyclicity and altered sex hormone profile (Sloboda et al. 2009, Connor et al. 2012, Cheong et al. 2014, Ambrosetti et al. 2016). These features in the female offspring of obese rats resemble the phenotype induced by estrogen exposure during early life. Administration of estrogenic compounds during the neonatal-to-infantile period leads to early onset of puberty, formation of ovarian follicular cysts and decreased follicular development in female offspring (Rosa et al. 2003, Sotomayor-Zarate et al. 2008, Cruz et al. 2012). Consistent with these studies, we observed increased serum estradiol levels in the offspring of rats fed an HFD before and during pregnancy and during nursing (Ambrosetti et al. 2016), where we demonstrated that female offspring of obese rats had increased serum estradiol levels from postnatal day (PND) 1 until early adulthood (PND 60) and that these elevated serum estradiol levels were probably due to decreased cytochrome P450 3A2 (CYP3A2) levels in the liver, reducing the liver’s capacity to metabolize serum estradiol, instead of increasing ovarian production (Ambrosetti et al. 2016). In addition, several studies showed that female offspring of obese rats develop hepatic steatosis, altering liver function (Oben et al. 2010, Mouralidarane et al. 2013, Ambrosetti et al. 2016). Therefore, an increase in serum estradiol levels due to hepatic dysfunction in female offspring of obese rats leads to long-term reproductive dysfunction.
Female rats exposed to high serum estradiol levels during early life show increased levels of norepinephrine (NE) in the ovaries, indicating an increased sympathetic tone related to the formation of follicular cysts during adulthood (Rosa et al. 2003, Sotomayor-Zarate et al. 2008, Cruz et al. 2012). Surgical denervation of the superior ovarian nerve, which carries sympathetic fibers to the ovary, reduces follicular cysts, improves ovarian follicular development and partially restores estrous cyclicity (Rosa et al. 2003, Sotomayor-Zarate et al. 2008). Serum estradiol can regulate tyrosine hydroxylase expression (Maharjan et al. 2005), the rate-limiting step in catecholamine synthesis, thus increasing NE levels. Serum estradiol increases the expression of neurotrophic peptide nerve growth factor (NGF), which could lead to higher sympathetic fiber density in the ovary (Lara et al. 2000, Dissen et al. 2009, Valladares et al. 2017). Persistently increased levels of serum estradiol during the neonatal-to-infantile-sensitive period (Cruz et al. 2012) in female offspring of obese mothers could increase NE levels and alter ovarian morphology and function.
Administration of metformin (a biguanide used to treat insulin resistance and diabetes mellitus 2) to obese pregnant rats reportedly reduces fetal liver inflammation (Harris et al. 2016), preventing liver alterations developing in offspring growing in an obesogenic uterine environment.
Therefore, we hypothesize that metformin administration to obese pregnant rats fed an HFD during pregnancy and nursing prevents the development of early liver dysfunction and increase in serum estradiol levels, thus preventing ovarian sympathetic hyperactivation and improving ovarian morphology and function in female offspring during adulthood. To test this hypothesis, we evaluated the effects of metformin administration to obese rats during pregnancy and nursing on their female offspring’s liver CYP3A2 expression, serum estradiol levels, ovarian sympathetic activity and follicular development during adulthood.
Materials and methods
Thirty female virgin Sprague–Dawley rats were weighed at least three times a week and their estrous cycle followed by taking vaginal smears daily for 3 weeks before the experiment (Fig. 1). Only rats with regular cycles were selected (N = 27). Selected rats were weight-matched and randomly divided into two groups: group 1 (controls; N = 8) was fed a control diet (CD), and group 2 (N = 19) was fed an HFD. After 3 weeks, the HFD-fed rats were weight-matched and randomly divided again into two groups: HFD (N = 10) and HFD plus metformin (HFD + MET; N = 9) (Table 1). After 4 weeks, on the afternoon of the proestrus stage, all female rats were placed with male rats with proven fertility. The next morning, successful mating was confirmed with sperm presence in the vaginal smears. Female rats that did not mate were again placed with male rats with proven fertility in the subsequent proestrus stage but for only one more attempt. If the female rats did not become pregnant after the second attempt, they were classified as ‘unsuccessful pregnancies’ and excluded from the study (N = 11). Pregnant rats were considered the experimental unit. The final number of pregnant rats was seven for the control group, five for the HFD group and four for the HFD + MET group. The control group was continuously fed a standard diet (27% of kilocalories in protein, 60% of kilocalories in carbohydrates and 13% of kilocalories in fat; LabDiet) for 1 month before pregnancy, during pregnancy and during nursing. The HFD and HFD + MET groups were fed an HFD (20% of kilocalories in protein, 20% of kilocalories in carbohydrates and 60% of kilocalories in fat; Research Diets Inc.) for 1 month before pregnancy, during pregnancy and during nursing. The HFD + MET group was administered metformin dissolved in drinking water from week 3 of the experiment (1 week before mating), during pregnancy and during nursing. Drinking water was measured each day to calculate and follow the exact doses of metformin intake. The actual dose was 160–200 mg/kg/day. After weaning, the male offspring were killed and not considered in this study. The female offspring of the three groups (control group, N = 39; HFD group, N = 27; HFD + MET group, N = 14) were separated from their mothers and divided into groups of 4–5 rats per cage. The offspring were fed a CD from weaning until killing. In some cases (in adult rats), we performed the experiments on two offspring from the same mother, wherein we averaged the values and considered the mean to create graphs and perform statistical analyses.
Body weight and fertility of obese and metformin-treated dams.
|Control||HF||HF + MET|
|Mothers body weight before diet (g)||222.8 ± 5.436||225.6 ± 5.995||224.4 ± 5.826||Control vs HF P = 0.7319 (ns)|
Control vs HF + MET P = 0.8430 (ns)
HF vs HF + MET P = 0.8880 (ns)
|Mothers body weight after 1 month of diet (g)||247.8 ± 6.654||274.6 ± 10.05||274.4 ± 7.710||Control vs HF P = 0.0322 (*)|
Control vs HF + MET P = 0.0375 (*)
HF vs HF + MET P = 0.9896 (ns)
|Number of pups||13.5 ± 0.500||11.6 ± 0.678||9.00 ± 2.16||Control vs HF P = 0.0467 (*)|
Control vs HF + MET P = 0.0151 (*)
HF vs HF + MET P = 0.1413 (ns)
All procedures were approved by the Institutional Committee for Bioethics and Animal Care, University of Valparaíso (CIBICA-UV, N°007-2013).
Euthanasia and tissue collection
All rats were killed by cervical decapitation at PND 14 or at first estrus after PND 57 (PND 60 ± 3); adult rats were killed after 6 h of fasting. At the time of killing, the ovaries, liver, retroperitoneal white adipose tissue (rpWAT) and trunk blood were collected. The blood was centrifuged at 850 × g for 10 min to obtain serum, which was stored at −80°C. All the collected tissues were stored at −80°C for further analyses.
Control of weight and estrous cycle
The body weights of all rats were recorded daily until the end of the experiment. The external genitalia were observed daily from PND 25 onward to determine the day of vaginal opening. The estrous cycle stages were determined by taking vaginal smears from the vaginal opening until killing. The percentage of cyclicity was calculated using the following formula: (No. cycles/(No. days since vaginal opening/4)) × 100, modified from Fernandois et al. (2012). This means that the cyclicity percentage would be 100% if a rat had 4-day continuous cycles from vaginal opening until killing.
Histology and morphometric analyses of the ovaries
At the time of killing of the adult offspring, 18 right ovaries (6 from each experimental group) were immersed in Bouin’s fixative. After fixation, the ovaries were embedded in paraffin, cut into 6 µm sections and stained with hematoxylin and eosin. Morphometric analysis of the follicles was performed, as previously reported (Cruz et al. 2012). We counted the total number of healthy antral follicles, atretic antral follicles, corpora lutea, follicular cysts and multi-oocyte follicles (MOFs). Antral follicles were counted when the nuclei of the oocytes were visible. The antral follicles were classified as atretic antral follicles if they had more than 5% of cells with pyknotic nuclei in the largest cross section and if they exhibited granulose shrinkage and occasional breakdown of the oocyte germinal vesicle. Follicular structures with more than one oocyte were classified as MOFs. Follicular structures were classified as follicular cysts if they met the following conditions: lack of an oocyte, absence of atresia criteria and a large antral cavity.
Plasma hormone and glucose determination
We determined serum estradiol, estriol, insulin and leptin levels by enzyme-linked immunoassays according to the manufacturer’s instructions: serum estradiol (11-ESTHU-E01, ALPCO Diagnostics, NH, USA), estriol (20-FE3HU-01, ALPCO Diagnostics), leptin (EZRL-83K, EMD Millipore) and insulin (EZRMI-13K, Merck Millipore). Intra- and interassay variations were <5% for serum estradiol, <4.7% for estriol, <6% for insulin and <3% for leptin. The minimum detectable concentrations were 10 pg/mL for serum estradiol, 0.075 ng/mL for estriol, 0.2 ng/mL for insulin and 0.05 ng/mL for leptin.
Serum glucose was determined by an enzymatic method (11504, Biosystem, Barcelona, Spain). The intra- and interassay coefficients of variation were <2.7%.
Western blot analyses
Hepatic CYP3A2 levels were detected by western blot at PNDs 14 and 60. For this, 15 mg of liver tissue were homogenized in radioimmunoprecipitation assay (RIPA) buffer: 50 mM Tris–HCl, pH 7.4; 1% NP-40; 0.1% SDS; 150 mM NaCl; 2 mM EDTA; 50 mM NaF and 1× protease inhibitor cocktail (Roche). Then, 30 µg of total protein were separated by SDS-PAGE in 10% polyacrylamide gel under reducing conditions. Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) was used as an internal control. The separated proteins were transferred to a nitrocellulose membrane, blocked with 5% milk for 1 h and probed with either rabbit polyclonal anti-CYP3A2 antibody (AB1276 Merck Millipore; 1:250, overnight incubation) or rabbit polyclonal anti-GAPDH antibody (G9545, Sigma-Aldrich; 1:40,000, 1-h incubation). The antibody complexes were detected using goat antirabbit immunoglobulin G (IgG) Fc (horseradish peroxidase (HRP)) (ab97200, concentration = 1:10,000; Abcam). We used the EZ-ECL Enhanced Chemiluminescence Detection Kit (Biological Industries, Israel) for complex detection; chemiluminescence was captured using the EpiChemi3 Darkroom system (UVP Inc., CA, USA). We analyzed the results by measuring the pixel intensities of bands using the semiquantification tool of the Image-Pro Plus 6.0 program (Media Cybernetics, Inc., Rockville, MD, USA). All western blots were performed at least thrice for each liver sample in independent blots. The coefficient of variation of the CYP3A2:GAPDH ratio for each individual sample was ≤25% among the independent blots.
Norepinephrine and 3-methoxy-4-hydroxyphenylglycol levels
We quantified NE and its metabolite 3-methoxy-4-hydroxyphenylglycol (MHPG) in the ovaries using high-performance liquid chromatography (HPLC) coupled with electrochemical detection by the EICOM ECD-700S electrochemical detector (Amuza Neuroscience, San Diego, CA, USA). In brief, we weighed the tissue samples using an analytical balance and homogenized them manually in a glass–glass homogenizer in 400 µL of 0.2 M perchloric acid (PCA). The samples were then centrifuged at 16,000 × g for 15 min at 4°C. The supernatant was collected and stored at −80°C until testing. Before the assay, we filtered an aliquot of 200 μL of the supernatant in 13 mm disposable PVDF membrane filters with 0.22 μm pores (Millex-GV, Merck Millipore Ltda). Then, 20 μL of the filtrate were injected into the JASCO PU-2089s plus HPLC system coupled with the JASCO LC-NetII/analog-to-digital converter digitizer (Tokyo, Japan); a Kromasil 100-3.5-C18 column (AkzoNobel N.V., Amsterdam, Netherlands) was used. To integrate the chromatograms, we used the JASCO ChromPass Chromatography Data System v1.7.403.1 software (Tokyo, Japan). The mobile phase consisted of 0.1 M NaH2PO4, 0.14 mM octyl sulfate, 0.02% EDTA and 1.5% acetonitrile (pH 2.6). The flow rate was set at 1 mL/min, and the electrochemical potential of the amperometric detector was set at +750 mV to simultaneously detect NE and MHPG. The retention time under these conditions was 5 min for NE and 10.5 min for MHPG.
All analyses were performed considering at least one rat per litter as a representative experimental unit. When two or more siblings were used for analysis, we considered the average of the values as a single data value. All data are expressed as mean ± standard error of mean (s.e.m.). One-way ANOVA (or two-way ANOVA for body weight curve analysis) followed by Fisher’s least significant difference test was used to compare the results between the control, HFD and HFD + MET groups. Statistical analyses were carried out with the GraphPad Prism v6.0 software (GraphPad Software), and P < 0.05 was considered statistically significant in all tests.
Effects of HFD + MET before and during pregnancy and during nursing on the mothers’ body weight and fertility
Before beginning the experiments, the female rats were randomly distributed in each experimental group; no differences were observed in body weight among groups (Table 1). As shown in Table 1, we found that after 1 month of being fed an HFD, the rats showed a considerable weight gain that metformin administration (administered 1 week before mating and 3 weeks after onset of the HFD) did not prevent. Regarding fertility, the HFD and HFD + MET groups showed a fewer number of offspring (Table 1). No differences in the female:male ratio was detected (data not shown).
Effects of HFD + MET before and during pregnancy and during nursing on the female offspring’s metabolic and reproductive parameters
We analyzed the body weight of female offspring daily from birth (PND 1) until PND 60. We found that the HFD and HFD + MET offspring were heavier than the control offspring from PND 1 until PND 60 (Fig. 2A). Surprisingly, the offspring of the HFD + MET offspring were heavier than the HFD offspring from PND 1 until ~PND 50 (P < 0.05 until PND50 for HFD vs HFD + MET) being magnified this difference prepubertally (Fig. 2B, C and D), but this difference disappeared at PND 60 (Fig. 2E). Interestingly, as shown in Table 2, at PND 60, the rpWAT weight increased in both HFD and HFD + MET offspring compared to control offspring, being more in HFD + MET offspring than in HFD offspring. Finally, as shown in Fig. 3, HFD offspring presented early onset of puberty, evidenced as early vaginal opening (Fig. 3A and B) and an early first estrus (Fig. 3C), an effect that was not prevented by metformin administration to the mothers during pregnancy and nursing. With regard to the estrous cycle, there were no differences in the number or percentage of estrous cycles among groups. However, we found an increase in permanency in the estrus stage and a decrease in permanency in the diestrus stage in HFD offspring compared to HFD + MET and control offspring (Table 3).
Body weight and metabolic parameters in offspring of obese dams treated with metformin.
|Control||HF||HF + MET|
|Liver weight PND14 (g)||0.6947 ± 0.02512||0.8892 ± 0.05811||0.8314 ± 0.1046||Control vs HF P = 0.0290 (*)|
Control vs HF + MET P = 0.1522 (ns)
HF vs HF + MET P = 0.5269 (ns)
|Liver weight PND60 (g)||7.477 ± 0.11783||8.044 ± 0.1921||8.013 ± 0.02574||Control vs HF P = 0.0283 (*)|
Control vs HF + MET P = 0.0473 (*)
HF vs HF + MET P = 0.9040 (ns)
|rpWAT PND60 (mg)||524 ± 44.22||818.7 ± 39.59||1003 ± 87.61||Control vs HF P = 0.0015 (**)|
Control vs HF + MET P < 0.0001(****)
HF vs HF + MET P = 0.0477 (*)
|Leptin PND14 (ng/mL)||1.573 ± 0.2589||4.117 ± 0.7462||3.498 ± 1.066||Control vs HF P = 0.024 (*)|
Control vs HF + MET P = 0.08895 (ns)
HF vs HF + MET P = 0.5614 (ns)
|Leptin PND60 (ng/mL)||1.293 ± 0.2092||1.385 ± 0.08292||2.051 ± 0.157||Control vs HF P = 0.6958 (ns)|
Control vs HF + MET P = 0.0096 (**)
HF vs HF + MET P = 0.0232 (*)
|Insulin PND60 (ng/mL)||1.030 ± 0.1158||1.553 ± 0.2441||1.601 ± 0.3055||Control vs HF P = 0.0473(*)|
Control vs HF + MET P = 0.0649 (ns)
HF vs HF + MET P = 0.4520 (ns)
|Glycemia PND14 (mmol/L)||5.438 ± 0.1603||5.718 ± 0.2380||5.363 ± 0.1626||Control vs HF P = 0.3173 (ns)|
Control vs HF + MET P = 0.7949 (ns)
HF vs HF + MET P = 0.2358 (ns)
|Glycemia PND60 (mmol/L)||5.249 ± 0.2349||4.635 ± 0.1637||5.239 ± 0.2242||Control vs HF P = 0.0836 (ns)|
Control vs HF + MET P = 0.9761 (ns)
HF vs HF + MET P = 0.1258 (ns)
Estrous cyclicity in offspring of controls, HF and HF + Met rats.
|Control||HF||HF + MET|
|Number of cycles||3.643 ± 0.2256||3.48 ± 0.1855||3.35 ± 0.2021||Control vs HF P = 0.5911 (ns)|
Control vs HF + MET P = 0.3717 (ns)
HF vs HF + MET P = 0.7074 (ns)
|Percentage of cyclicity||65.16 ± 3.648||63.18 ± 3.66||57.73 ± 3.633||Control vs HF P = 0.7044 (ns)|
Control vs HF + MET P = 0.1963 (ns)
HF vs HF + MET P = 0.3673 (ns)
|Percentage of permanency in each estrous stage||Proestrus||21.13 ± 2.023||19.62 ± 1.536||19.4 ± 0.991||Control vs HF P = 0.6237 (ns)|
Control vs HF + MET P = 0.5995 (ns)
HF vs HF + MET P = 0.9501 (ns)
|Estrus||23.36 ± 1.512||31.1 ± 2.824||22.6 ± 0.960||Control vs HF P = 0.0152 (*)|
Control vs HF + MET P = 0.8178
HF vs HF + MET P = 0.0197 (*)
|Diestrus||59.56 ± 2.892||48.14 ± 2.821||60.6 ± 2.056||Control vs HF P = 0.0006 (***)|
Control vs HF + MET P = 0.7511
HF vs HF + MET P = 0.0010 (***)
With regard to hormonal metabolic markers, we found that leptin levels increased at PND 14 in HFD offspring compared to control offspring, an effect that was not prevented by metformin administration to the mothers (HFD vs HFD + MET, P = 0.5614; Table 2). As shown in Table 2, at PND 60, leptin levels in HFD offspring were equal to control offspring. Surprisingly, HFD + MET offspring showed increased leptin levels compared to HFD and control offspring. Insulin levels were increased in HFD offspring compared to control offspring at PND 60 and only slightly increased in HFD + MET offspring compared to control offspring (controls vs HFD + MET, P = 0.0649).
Effects of HFD + MET before and during pregnancy and during nursing on the female offspring’s serum estradiol levels and hepatic CYP3A2 expression
We found that serum estradiol levels were higher at PND 14 in HFD offspring compared to control offspring, but no differences were observed at PND 60 (Fig. 4A and B). In addition, we found that metformin administered to the mothers prevented an increase in serum estradiol at PND 14 in the offspring (Fig. 4A). CYP3A2 expression in the liver at PNDs 14 and 60 was lower in HFD offspring than in control offspring of obese rats. However, metformin treatment to the mothers prevented the hepatic reduction of CYP3A2 induced by an HFD (Fig. 4G and H). To assess serum estradiol biotransformation by CYP3A2, we measured the plasma levels of serum estriol, a metabolite from direct and one-step serum estradiol biotransformation through hepatic CYP3A2. Figures 4C and D show that serum estriol did not decrease at PNDs 14 and 60 in HFD offspring compared to control offspring. However, when the serum estriol:serum estradiol ratio was analyzed (as an indicator of serum estradiol clearance by CYP3A2 metabolization), HFD offspring presented a decrease compared to control offspring (controls vs HFD, P = 0.047; HFD + MET vs HFD, P = 0.064). At PND 60, the serum estriol:serum estradiol ratio showed only a slight, but not significant, tendency to decrease (controls vs HFD, P = 0.0889; HFD vs HFD + MET, P = 0.0715).
Ovarian sympathetic activity in female offspring of obese mothers and effects of metformin administration during pregnancy and nursing
We measured NE and MHPG levels in the ovaries of female offspring as an index of sympathetic activity (Heal et al. 1989). Figure 5A shows that the ovarian NE levels increased in HFD offspring, while they returned to control values when the mothers were treated with metformin. The same pattern was observed for MHPG (Fig. 5B), indicating a decrease in NE release from ovaries in the HFD + MET offspring compared to HFD offspring.
Ovarian follicular development in female offspring of obese mothers and effects of metformin administration during pregnancy and nursing
Morphological analysis of the ovaries at PND 60 was conducted. Figure 6A shows a decrease in healthy antral follicles in the female offspring of obese mothers compared to control offspring. In addition, when healthy antral follicles were analyzed by size (Fig. 6B), we observed a decrease in the smallest antral follicles (<200 µm) in the ovaries of HFD offspring compared to control offspring. However, this effect was not prevented by administration of metformin to the mothers during pregnancy and nursing. With regard to atretic antral follicles, we found a decrease in HFD offspring compared to control offspring (Fig. 6C). Finally, we observed no significant changes in the number of corpora lutea (Fig. 6D). However, two abnormal structures, MOFs and follicular cysts, showed an increase in the ovaries of HFD offspring (Fig. 6E and F, respectively). Some of these MOFs were similar to oocyte nests, which are usually observed in neonates before follicular assembly (we found four oocyte nests in three different HFD offspring ovaries from three different litters). Figure 7 shows an oocyte nest–like structure (Fig. 7A and B), a secondary MOF (Fig. 7C), an antral MOF (Fig. 7D) and follicular cysts (Fig. 7E). All these effects (decrease in healthy and atretic antral follicles, increase in follicular cysts and increase in MOFs) were prevented in adult offspring when metformin was administered to the mothers during pregnancy and nursing.
Maternal obesity causes offspring to suffer from various metabolic and reproductive disorders. We previously demonstrated that female offspring of rats exposed to an HFD for 1 month before pregnancy, during pregnancy and during nursing have high serum estradiol levels during early life (Ambrosetti et al. 2016), which is associated with reproductive alterations such as early puberty, decreased follicular development and polycystic ovaries (Ambrosetti et al. 2016, Lin et al. 2017). We postulated that offspring of these rats have a hepatic dysfunction that leads to low hepatic serum estradiol metabolism through the enzyme CYP3A2. This low hepatic serum estradiol metabolism could cause high serum estradiol levels (Ambrosetti et al. 2016). In this study, we treated pregnant rats fed an HFD with the drug metformin during pregnancy and nursing in order to improve their metabolic profile and hence prevent ovarian and hormonal dysfunctions in the offspring.
Maternal obesity and maternal exposure to an HFD in rats causes inflammation, insulin resistance and increased accumulation of triglycerides in the fetal liver (Harris et al. 2016). This hepatic triglyceride accumulation is probably due to a hypercaloric fetal microenvironment caused by maternal insulin resistance, which causes high levels of glucose and free fatty acids to transfer from maternal blood to the fetal liver (Jansson & Powell 2007). The drug metformin could improve insulin sensitivity in pregnant rats and decrease the availability and hence the transfer of these substances into fetal blood, thereby preventing postnatal hepatic dysfunction. The results of this study showed that administering metformin to obese pregnant rats during pregnancy and nursing prevents a decrease in CYP3A2 levels in female offspring and prevents an increase in serum estradiol levels at PND 14, as we expected. However, although metformin restored serum estradiol levels to control values in the female offspring, it did not prevent early onset of puberty (measured as vaginal opening). This result suggested that maternal exposure to an HFD during pregnancy and nursing activates another mechanism that leads to precocious activation of the hypothalamic–pituitary–ovarian axis. Another reason for precocious puberty could be the increase in leptin levels, which previous studies have demonstrated as a causal factor for accelerating the onset of puberty (Ahima et al. 1997, Shalitin & Phillip 2003). In this study, we observed that metformin administration to obese mothers did not reduce leptin levels to control values in the female offspring at PND 14; in fact, the levels increased by PND 60. Therefore, increased leptin levels might induce precocious puberty in the female offspring of rats fed an HFD during pregnancy and nursing.
In this study, we did not find a significant decrease in serum estradiol levels at PND 60, as previously reported (Ambrosetti et al. 2016), even when the levels of CYP3A2 were still low and the plasma serum estriol:serum estradiol ratio showed a tendency to decrease (HFD vs controls, P = 0.077; HFD vs HFD + MET, P = 0.067). We believe that the mature hypothalamic feedback might be compensating for an increase in serum estradiol levels due to hepatic dysfunction during adulthood, causing plasma serum estradiol levels to be similar to control values or slightly increased, as shown by Ambrosetti et al. (2016). However, before puberty, with an immature reproductive hypothalamic axis (Picut et al. 2015), this increase in serum estradiol levels is not compensated by the hypothalamic feedback, which is demonstrated by the magnitude of serum estradiol increase.
On the other hand, despite being fed a CD after weaning, the HFD + MET offspring were heavier than the control offspring during adulthood. In addition, the HFD + MET offspring had an increased body weight at PNDs 1, 7 and 14, and until ~PND 50 compared to control and HFD offspring. Moreover, although we observed no differences in body weight between HFD and HFD + MET offspring at PND 60, the offspring of rats treated with metformin during pregnancy and nursing showed an increased rpWAT weight, which was also reflected in the leptin levels. This result showed that metformin administration to rats during pregnancy and nursing amplifies the effects of an HFD on body weight and fat accumulation in the offspring.
Interestingly, a recent study on humans showed that children of women with PCOS who were on metformin during pregnancy had a higher BMI from age 6 months until 4 years and an increased prevalence of obesity at age 4 years compared to children of mothers from the placebo group (Engen Hanem et al. 2018). The reason could be the transfer of metformin into the fetus through the placenta (Kovo et al. 2008, Salomaki et al. 2013). In this regard, the effects of metformin on the fetal liver are probably beneficial, as demonstrated by other studies (Salomaki et al. 2013, 2014, Harris et al. 2016); however, increasing insulin sensitivity in adipose tissue of the fetus could increase the storage of fat, favoring postnatal obesity.
In this study, we proposed that the phenotype observed in adult female offspring of rats exposed to an HFD before pregnancy, during pregnancy and during nursing resembles that observed in women with PCOS, an endocrine and metabolic disorder affecting a large number of women of reproductive age. The offspring of obese mothers have a higher prevalence of nonalcoholic fatty liver disease, insulin resistance and hypertension, as is usually seen in women with PCOS (Glintborg et al. 2016, Li et al. 2016, Orio et al. 2016). However, several authors have proposed a developmental origin of PCOS because of prenatal or early postnatal exposure to estrogenic or androgenic compounds. Early exposure to estrogenic compounds decreases antral follicles and increases follicular cysts (Rosa et al. 2003, Sotomayor-Zarate et al. 2008, Padmanabhan & Veiga-Lopez 2011, Cruz et al. 2012). We think that the increase in endogenous serum estradiol levels we observed in our study is similar to the endocrine disruption caused by estrogenic compounds. In addition, we and other researchers demonstrated that the adult female offspring of rats exposed to an HFD during pregnancy and nursing have a low number of antral follicles and a high number of follicular cysts (Cheong et al. 2014, Ambrosetti et al. 2016). The decrease in healthy and atretic antral follicles observed in this study was also observed by Ambrosetti et al. (2016) at PND 60, which is a short time after puberty and indicates low formation and growing of follicles. In addition, Ambrosetti et al. (2016) and Tsoulis et al. (2016) showed a decrease in healthy antral follicles and an increase in atretic antral follicles at PND 120, a long time after puberty. Because gonadotropin follicle-stimulating hormone (FSH), through action on its follicle-stimulating hormone receptor, is essential for both the formation/growth of follicles and the prevention of atresia (Hirshfield & Midgley 1978, Chun et al. 1996), it is possible that FSH sensitivity of follicles could be altered in HFD offspring. Tsoulis et al. (2016) showed an increase in anti-Mullerian hormone (AMH) expression in antral follicles. This finding supports the observed ovarian morphological changes, since AMH decreases follicular sensitivity to FSH, blocking the formation of follicles and increasing the rate of atresia (Visser et al. 2007, Pellatt et al. 2011). Similarly, when neonatal rats are exposed to serum estradiol valerate, there is an increase in AMH in the antral follicles at PND 60, indicating that the serum estradiol increase during this period (as was observed in our study) alters the genetic expression of AMH in follicles during adulthood (Martinez-Pinto et al. 2018). Reinforcing this finding is another feature we observed in our study and was also observed by Connor et al. (2012): an increase in permanency in the estrus stage, meaning slow follicular formation and growth. However, in this study, we found that metformin administration to mothers during pregnancy and nursing prevents the decrease in antral follicles and the increase in follicular cysts observed in HFD offspring. We think that metformin’s effect of preventing an increase in serum estradiol levels in female offspring of obese mothers during the first few days of life ultimately prevents a decrease in antral follicles and an increase in follicular cysts, as discussed later. Another feature of animals exposed to estrogenic compounds is long-term permanence of MOFs in the ovary (Jefferson et al. 2002, Kipp et al. 2007). In this study, we observed an increase in MOFs in female offspring of obese mothers that was prevented by administration of metformin to the mothers. A possible explanation for the appearance of MOFs is the effect of estrogenic compounds in preventing or delaying follicular assembly (Kezele & Skinner 2003, Pepling et al. 2010) and then altering the formation of primordial follicles with individual oocytes. This follicular assembly in rats starts before birth but ends around PND 4 (Pepling et al. 2010, Pepling 2012). In this study, we did not measure serum estradiol levels within this time window, but in a previous study, we demonstrated that endogenous serum estradiol levels increase at PNDs 1 and 7 in female offspring of obese mothers (Ambrosetti et al. 2016). Therefore, we hypothesized that an increase in serum estradiol levels during early postnatal life and even before birth might disrupt follicular assembly and increase the probability of MOF generation, which we observed in adult female rats. Moreover, upon histological evaluation, some of the MOFs we found were similar to oocyte nests, reinforcing the idea of aberrant follicular assembly. Another effect of exposure to estrogenic compounds in early life is a long-term increase in the activity of the sympathetic nervous system in the ovary. For example, early exposure to serum estradiol valerate increases NE levels in the ovary during adulthood (Rosa et al. 2003, Sotomayor-Zarate et al. 2008, Cruz et al. 2012). In this study, we demonstrated that female offspring of obese mothers have increased NE levels in the ovary. Therefore, increased endogenous serum estradiol levels during the neonatal-to-prepubertal period could be the cause of increased NE levels in the ovary during adulthood. This increase in NE levels was prevented by metformin administration to pregnant and nursing mothers, which also prevented an increase in serum estradiol levels at PND 14 in female offspring. The mechanisms explaining this increase in NE levels are not clear, but one plausible explanation is the ability of estrogenic compounds to increase NGF expression (Sotomayor-Zarate et al. 2008), which is strongly related to a higher density of nerve fibers. An additional factor is the neonatal-to-prepuberal hyperleptinemia occurring in offspring, as demonstrated by us and other researchers (Ambrosetti et al. 2016, Zambrano et al. 2016, Tellechea et al. 2017). The higher leptin levels during early development cause an increase in sympathetic outflow to several tissues, such as adipose tissue, the heart and kidneys (Rahmouni 2010, Samuelsson et al. 2010, Taylor et al. 2014). Conversely, a decrease in NE levels in plasma and peripheral tissues, particularly brown and white adipose tissue, was observed in obese and leptin-deficient (ob/ob) mice (Knehans & Romsos 1983, 1984) and obese and leptin receptor-deficient (fa/fa) rats (York et al. 1985, Rosenthal et al. 1996). On the other hand, hyperleptinemic diabetic (db/db) mice, which have a nonfunctional ObRb leptin receptor, have increased NE levels in the ovary and uterus (Garris 2004). Therefore, we hypothesized that an increase in leptin levels during early development might increase the sympathetic tone of the ovary during adulthood, as in other organs, but this idea needs to be investigated in the future. Interestingly, women with PCOS have a higher density of sympathetic fibers in the ovary and an increase in overall sympathetic activity (Heider et al. 2001, Lansdown & Rees 2012, Wojtkiewicz et al. 2014). This increased sympathetic nerve activity is usually determined by measuring NE metabolite levels (see Table 1 in Lansdown & Rees 2012). In this study, we measured the NE metabolite MHPG, which is formed from the enzymatic reaction through monoamine oxidase and catechol-o-methyl transferase (COMT). Since COMT has an extracellular distribution, MHPG is an indicator of NE release (Lookingland et al. 1991). We found increased MHPG levels in the HFD offspring compared to control offspring, reflecting an increase in NE release. Interestingly, administration of metformin to obese mothers prevented the increase in MHPG levels in the ovaries of offspring, suggesting that metformin decreases sympathetic activity in the ovaries during adulthood, probably by preventing an increase in serum estradiol levels during early life.
We conclude that female offspring of mothers exposed to an HFD before pregnancy, during pregnancy, and during nursing have a PCOS-like reproductive phenotype, including metabolic and reproductive dysfunctions. The administration of metformin to mothers fed an HFD during pregnancy and nursing partially prevents the PCOS-like reproductive phenotype but amplifies some metabolic impairments, such as fat accumulation and increased body weight, in the offspring. Finally, metformin can improve hepatic function in the newborn, mitigating reproductive dysfunction by improving the metabolism of serum estradiol during the most vulnerable period (PNDs 1–14) when serum estradiol disruption can occur during postnatal ovarian development in rats (Cruz et al. 2012).
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 work was supported by Fondecyt grant number 11130707 and DIUV-CL N° 01/2006 to Gonzalo Cruz.
This work was supported by Fondecyt Initiation Grant 11130707-CONICYT. Additional funds were provided by the Center for Neurobiology and Cerebral Plasticity (CNPC) – Universidad de Valparaíso. The authors thank Tania Cerda from the University of Valparaíso for technical support.
AmbrosettiVGuerraMRamirezLAReyesAAlvarezDOlguinSGonzalez-MananDFernandoisDSotomayor-ZarateRCruzG 2016 Increase in endogenous estradiol in the progeny of obese rats is associated with precocious puberty and altered follicular development in adulthood. Endocrine 53 258–270. (https://doi.org/10.1007/s12020-016-0858-0)
ConnorKLVickersMHBeltrandJMeaneyMJSlobodaDM 2012 Nature, nurture or nutrition? Impact of maternal nutrition on maternal care, offspring development and reproductive function. Journal of Physiology 590 2167–2180. (https://doi.org/10.1113/jphysiol.2011.223305)
CruzGBarraRGonzalezDSotomayor-ZarateRLaraHE 2012 Temporal window in which exposure to estradiol permanently modifies ovarian function causing polycystic ovary morphology in rats. Fertility and Sterility 98 1283–1290. (https://doi.org/10.1016/j.fertnstert.2012.07.1060)
DissenGAGarcia-RudazCParedesAMayerCMayerhoferAOjedaSR 2009 Excessive ovarian production of nerve growth factor facilitates development of cystic ovarian morphology in mice and is a feature of polycystic ovarian syndrome in humans. Endocrinology 150 2906–2914. (https://doi.org/10.1210/en.2008-1575)
Engen HanemLGStridsklevSJuliussonPBSalvesenORoelantsMCarlsenSMOdegardRVankyE 2018 Metformin use in PCOS pregnancies increases the risk of offspring overweight at 4 years of age; follow-up of two RCTs. Journal of Clinical Endocrinology and Metabolism 103 1612–1621. (https://doi.org/10.1210/jc.2017-02419)
GlintborgDPetersenMHRavnPHermannAPAndersenM 2016 Comparison of regional fat mass measurement by whole body DXA scans and anthropometric measures to predict insulin resistance in women with polycystic ovary syndrome and controls. Acta Obstetricia et Gynecologica Scandinavica 95 1235–1243. (https://doi.org/10.1111/aogs.12964)
HealDJProwMRBuckettWR 1989 Measurement of 3-methoxy-4-hydroxyphenylglycol (MHPG) in mouse brain by h.p.l.c. with electrochemical detection, as an index of noradrenaline utilisation and presynaptic alpha 2-adrenoceptor function. British Journal of Pharmacology 96 547–556. (https://doi.org/10.1111/j.1476-5381.1989.tb11852.x)
JeffersonWNCouseJFPadilla-BanksEKorachKSNewboldRR 2002 Neonatal exposure to genistein induces estrogen receptor (ER)alpha expression and multioocyte follicles in the maturing mouse ovary: evidence for ERbeta-mediated and nonestrogenic actions. Biology of Reproduction 67 1285–1296. (https://doi.org/10.1095/biolreprod67.4.1285)
KovoMKogmanNOvadiaONakashIGolanAHoffmanA 2008 Carrier-mediated transport of metformin across the human placenta determined by using the ex vivo perfusion of the placental cotyledon model. Prenatal Diagnosis 28 544–548. (https://doi.org/10.1002/pd.2026)
LaraHEDissenGALeytonVParedesAFuenzalidaHFiedlerJLOjedaSR 2000 An increased intraovarian synthesis of nerve growth factor and its low affinity receptor is a principal component of steroid-induced polycystic ovary in the rat. Endocrinology 141 1059–1072. (https://doi.org/10.1210/endo.141.3.7395)
LinYJTsaiCCHuangLTSheenJMTiaoMMYuHRChenCCTainYL 2017 Detrimental effect of maternal and post-weaning high-fat diet on the reproductive function in the adult female offspring rat: roles of insulin-like growth factor 2 and the ovarian circadian clock. Journal of Assisted Reproduction and Genetics 34 817–826. (https://doi.org/10.1007/s10815-017-0915-5)
LookinglandKJIrelandLMGunnetJWManzanaresJTianYMooreKE 1991 3-Methoxy-4-hydroxyphenylethyleneglycol concentrations in discrete hypothalamic nuclei reflect the activity of noradrenergic neurons. Brain Research 559 82–88. (https://doi.org/10.1016/0006-8993(91)90289-8)
MaharjanSSerovaLSabbanEL 2005 Transcriptional regulation of tyrosine hydroxylase by estrogen: opposite effects with estrogen receptors alpha and beta and interactions with cyclic AMP. Journal of Neurochemistry 93 1502–1514. (https://doi.org/10.1111/j.1471-4159.2005.03142.x)
Martinez-PintoJPiquerBTiszavariMLaraHE 2018 Neonatal exposure to estradiol valerate reprograms the rat ovary androgen receptor and anti-Mullerian hormone to a polycystic ovary phenotype. Reproductive Toxicology 75 127–135. (https://doi.org/10.1016/j.reprotox.2017.10.005)
MouralidaraneASoedaJVisconti-PugmireCSamuelssonAMPomboJMaragkoudakiXButtASaraswatiRNovelliMFusaiGet al. 2013 Maternal obesity programs offspring nonalcoholic fatty liver disease by innate immune dysfunction in mice. Hepatology 58 128–138. (https://doi.org/10.1002/hep.26248)
ObenJAMouralidaraneASamuelssonAMMatthewsPJMorganMLMcKeeCSoedaJFernandez-TwinnDSMartin-GronertMSOzanneSEet al. 2010 Maternal obesity during pregnancy and lactation programs the development of offspring non-alcoholic fatty liver disease in mice. Journal of Hepatology 52 913–920. (https://doi.org/10.1016/j.jhep.2009.12.042)
OrioFMuscogiuriGNeseCPalombaSSavastanoSTafuriDColarietiGLa SalaGColaoAYildizBO 2016 Obesity, type 2 diabetes mellitus and cardiovascular disease risk: an uptodate in the management of polycystic ovary syndrome. European Journal of Obstetrics and Gynecology and Reproductive Biology 207 214–219. (https://doi.org/10.1016/j.ejogrb.2016.08.026)
PellattLRiceSDilaverNHeshriAGaleaRBrincatMBrownKSimpsonERMasonHD 2011 Anti-Mullerian hormone reduces follicle sensitivity to follicle-stimulating hormone in human granulosa cells. Fertility and Sterility 96 1246–1251.e1241. (https://doi.org/10.1016/j.fertnstert.2011.08.015)
RosaESAGuimaraesMAPadmanabhanVLaraHE 2003 Prepubertal administration of estradiol valerate disrupts cyclicity and leads to cystic ovarian morphology during adult life in the rat: role of sympathetic innervation. Endocrinology 144 4289–4297. (https://doi.org/10.1210/en.2003-0146)
SalomakiHVahataloLHLaurilaKJappinenNTPenttinenAMAilanenLIlyasizadehJPesonenUKouluM 2013 Prenatal metformin exposure in mice programs the metabolic phenotype of the offspring during a high fat diet at adulthood. PLoS ONE 8 e56594. (https://doi.org/10.1371/journal.pone.0056594)
SalomakiHHeinaniemiMVahataloLHAilanenLEerolaKRuohonenSTPesonenUKouluM 2014 Prenatal metformin exposure in a maternal high fat diet mouse model alters the transcriptome and modifies the metabolic responses of the offspring. PLoS ONE 9 e115778. (https://doi.org/10.1371/journal.pone.0115778)
Sotomayor-ZarateRDorfmanMParedesALaraHE 2008 Neonatal exposure to estradiol valerate programs ovarian sympathetic innervation and follicular development in the adult rat. Biology of Reproduction 78 673–680. (https://doi.org/10.1095/biolreprod.107.063974)
TellecheaMLMensegueMFPirolaCJ 2017 The association between high fat diet around gestation and metabolic syndrome-related phenotypes in rats: a systematic review and meta-analysis. Scientific Reports 7 5086. (https://doi.org/10.1038/s41598-017-05344-7)
TsoulisMWChangPEMooreCJChanKAGohirWPetrikJJVickersMHConnorKLSlobodaDM 2016 Maternal high-fat diet-induced loss of fetal oocytes is associated with compromised follicle growth in adult rat offspring. Biology of Reproduction 94 94. (https://doi.org/10.1095/biolreprod.115.135004)
VisserJADurlingerALPetersIJvan den HeuvelERRoseUMKramerPde JongFHThemmenAP 2007 Increased oocyte degeneration and follicular atresia during the estrous cycle in anti-Mullerian hormone null mice. Endocrinology 148 2301–2308. (https://doi.org/10.1210/en.2006-1265)
ZambranoEIbanezCMartinez-SamayoaPMLomas-SoriaCDurand-CarbajalMRodriguez-GonzalezGL 2016 Maternal obesity: lifelong metabolic outcomes for offspring from poor developmental trajectories during the perinatal period. Archives of Medical Research 47 1–12. (https://doi.org/10.1016/j.arcmed.2016.01.004)