Perinatal exposure to high dietary advanced glycation end products affects the reproductive system in female offspring in mice

Abstract

Maternal nutrition and the intrauterine environment are important in determining susceptibility to reproductive and metabolic disturbances. Advanced glycation end products (AGEs) are widely consumed in Western diet. The purpose of this study was to determine whether perinatal exposure to a high levels of dietary AGEs affect metabolic and reproductive parameters in female mice offspring. Female CD1 mice, 7 weeks old, were placed on either a diet low (L-AGE) or high (H-AGE) in AGEs before mating and then during pregnancy and lactation. All offspring were weaned onto the L-AGE diet and studied through to 16 weeks of age; they were counted and weighed at birth and then every week for a total of 11 weeks. Vaginal opening, litter size, growth curve, liver and abdominal fat weights, serum levels of anti-Mullerian hormone, leptin and adiponectin, as well as insulin and glucose tolerance tests were compared. Ovaries were harvested for follicular count and gene expression by real-time polymerase chain reaction. Compared to perinatal exposure to the L-AGE diet, perinatal exposure to the H-AGE diet caused lower body weight at birth, and adult offspring exhibited delayed growth, lower serum leptin and adiponectin levels, delayed vaginal opening, irregular oestrous cyclicity, arrested follicular development and significant alterations in the expression of genes involved in folliculogenesis (Amh and Amhr2) and steroidogenesis (Cyp19a1). These results indicate that perinatal exposure to a diet elevated in AGEs causes deficits in perinatal growth, pubertal onset, and reproductive organ development in female mice. Whether these findings translate to humans remains to be determined in future studies.

Introduction

Studies have shown that maternal nutrition and the intrauterine environment are important in determining susceptibility to reproductive and metabolic disturbances (Barker, 1997; Dicken et al., 2012). For instance, maternal obesity or consumption of a high-fat diet during pregnancy and lactation increases the risk of metabolic diseases in offspring (Williams et al., 2014) and maternal vitamin D3 deficiency could alter puberty in female offspring (Dicken et al., 2012).

The typical Western diet, high in protein and fat, contains high amounts of advanced glycation end products (AGEs), which are highly reactive inflammatory molecules formed when lipids and proteins become glycosylated following exposure to glucose (Horiuchi et al., 1991; Edelstein and Brownlee, 1992; Goldberg et al., 2004). Data have shown that this dietary pattern results in reproductive disturbances (Merhi et al., 2019), increased postprandial glucose excursion, oxidative stress and inflammation in humans (Goldberg et al., 2004). Maternal exposure to a diet containing large amounts of AGEs during pregnancy might predispose mice offspring to manifestation of metabolic and behavioural disturbances later in life; for instance, perinatal exposure to high dietary AGEs have been shown to predispose the male progeny to weight gain and to affect their glucose homeostasis (Csongova et al., 2018). Additionally, reducing exposure to dietary AGEs throughout gestation, lactation and early postnatal life has been shown to benefit pancreatic islet secretion and immune infiltration in mice (Borg et al., 2018).

In this study, we aimed to evaluate the effect of exposure to high dietary AGEs in utero and during lactation on metabolic disturbances and reproductive potential in female littermate wild-type mice. The results of this study showed that exposure to a maternal diet rich in AGEs could adversely affect body weight, onset of puberty, oestrous cyclicity and ovarian function in the offspring. These data highlight the intergenerational effect of AGEs on metabolic and reproductive potential in females.

Methods

Animals and experimental design

All the experimental methods involving mice were approved by the Albert Einstein College of Medicine and conducted in accordance with the Animal Welfare Act guidelines approved by the Animal Care Institute. Animals were housed in a barrier facility and maintained on a 14-h to 10-h light-dark cycle. As previously described, 7-week-old female CD1 mice (Charles River Labs, Worcester, MA) were placed ad libitum on either a diet low in AGEs (L-AGE), or diet high in AGEs (H-AGE) for 2 weeks before mating to CD1 males (Charles River Labs), then for another 6 weeks throughout pregnancy and lactation. The L-AGE diet was purchased from BioServ AIN-93-G- Modified Product #F3542 (chemical composition: 16.6% fat, 18.8% protein, 64.6% carbohydrate, 3.73 kcal/g). H-AGE was produced by subjecting the L-AGE diet to heating at 125°C for 30 min as previously described (Lin et al., 2003; Peppa et al., 2003; Guo et al., 2012). The heating process increases the levels of AGEs, in particular N(ε)-(carboxymethyl) lysine (CML) (Uribarri et al., 2010; Li et al., 2012). In order to ensure that heating the L-AGE diet increased the levels of AGEs, we used an ELISA kit for the quantification of CML (Cell Biolabs, Inc.) in both diets, which is the most abundant dietary AGE and most frequently selected compound as a marker for AGEs in clinical and laboratory studies in both humans and animals (Li et al., 2012; Tantalaki et al., 2014; Abate et al., 2015). An approximate 10 times higher level of CML was measured in the H-AGE diet compared to the L-AGE diet (605 μg/g vs. 62 μg/g; respectively). After placing singly housed dams into clean cages, the total amount of food consumed each day was recorded in the week preceding mating the mice and done so for a period of five consecutive days, and the average daily food intake was calculated. Body weight of the dams was measured on the day of mating and again on embryonic days 0.5, 9.5 and 19.5, and the total weight gain during pregnancy was calculated. Male and females offspring were born with the expected Mendelian frequency. The male offspring remained with the dam throughout lactation. Offspring from both L-AGE (n = 10) and H-AGE(n = 13) dams were all weaned onto the L-AGE diet at postnatal day 21 and studied through to 16 weeks of age.

Litter size, growth curve, liver and abdominal fat weights, insulin tolerance test, and glucose tolerance test

Offspring were counted and weighed at birth and then every week for a total of 11 weeks. At 10 weeks of age, overnight fasted mice were administered glucose (1.5 g/kg body weight) via intraperitoneal (i.p.) injection. Tail blood glucose levels were obtained using a glucometer (Precision Xtra; MediSense, Waltham, MA) at 0 (baseline), 15, 30, 60 and 120 min after glucose loading. For insulin tolerance test (ITT), 6 h fasted offspring were i.p. injected with a bolus of human insulin at 0.75 units/kg of body weight (Novolin R; Novo Nordisk, Denmark). Blood glucose levels were determined in tail vein blood at the indicated times (0, 15, 30, 60 and 120 min) with a glucometer. The area under the curve (AUC) for both glucose and insulin levels obtained by each test were calculated. At the end of the experiments, at 16 weeks of age, all animals were sacrificed at the dioestrus stage by cervical dislocation then body weights were recorded, and ovaries were dissected and frozen for real-time polymerase chain reaction (RT-PCR) or fixed for follicle counts. The liver and the abdominal fats were retrieved and weighed.

Puberty onset and characterisation of oestrous cyclicity

Pubertal onset was assessed by recording the age at vaginal opening. The oestrous staging was determined with daily vaginal lavage. Oestrous stages were defined as pro-oestrus (80–100% epithelial cells), oestrus (100% cornified epithelial cells), dioestrus (∼50% cornified epithelial cells and 50% leukocytes) and metoestrus (80–100% leukocytes) (Cohen et al., 2002). Oestrous cycle length was defined by the number of days required for a mouse to transition from one pro-oestrus event to the next. Oestrous cycle length and percentage of time mice spent in each stage of the oestrous cycle were recorded during a 3-week period.

Serum analysis for anti-Mullerian hormone, leptin and adiponectin

At the time of sacrifice, i.e. 16 weeks of age, whole blood was collected by retroorbital bleed and centrifuged. The serum was stored at −80°C for future quantification for anti-Mullerian hormone (AMH), leptin and adiponectin. ELISA kits were used according to the manufacturer’s instructions to measure AMH (Ansh Laboratories [Webster, TX] for Rat and Mouse), leptin (R&D Systems [Minneapolis, MN] Quantikine Mouse/Rat) and adiponectin (R&D Systems [Minneapolis, MN] Quantikine ELISA Mouse). The sensitivity for each was 0.003 ng/ml, 0.41 ng/ml and 22 pg/ml for adiponectin, AMH and leptin; respectively. The intra-assay coefficients of variation were 3.2%, 8.5% and 4.3% for adiponectin, AMH and leptin, respectively.

RT-PCR on ovaries

At the time of sacrifice, one ovary per mouse (n = 10–13 per group) was snap frozen in liquid nitrogen and stored at −80°C for RNA extraction and gene expression while the other ovary was fixed in formalin and later used for follicle counting (see below). RNA extraction was performed using Trizol reagent (Invitrogen, Carlsbad, CA), chloroform extraction and RNeasy mini kit (Qiagen, Valencia, CA), according to the manufacturer’s instructions following lysis and homogenisation using a Tekmar homogeniser. RNase-free DNase set (Qiagen, Valencia, CA) was then used according to the manufacturer’s instructions to digest DNA from the solution. After DNA digestion, the solution underwent RNA purification using Qiagen’s RNeasy mini kit according to the manufacturer’s instructions. The concentration was determined using an ND-2000 spectrophotometer. RNA quality was confirmed by electrophoresis and an optical density (OD) 260/280 ratio of greater than 1.8. First-strand cDNA was created from 500 ng of total RNA by reverse transcriptase (Invitrogen SuperScript III First-Strand Synthesis System, Carlsbad, CA). The mRNA levels were measured using FastStart SYBR Green Master (Roche, Indianapolis, IN) for genes involved in folliculogenesis and steroidogenesis: Amh, Amhr2, Fshr, Lhcgr and Cyp19a1 (aromatase enzyme). The dissociation curve was checked at the end of every PCR run in order to ensure homogeneity of each product. Primer sets were designed using Universal Probe Library Assay Design Center. Primers were synthesised by Fisher (Pittsburgh, PA), confirmed using BLAST, and are shown in Table I. For quantitative analysis, all samples were normalised to cyclophillin and relative mRNA expression levels were determined by the ΔΔ-CT method and expressed as arbitrary units. Samples were measured in triplicate for each gene to assess technical variability.

Table I.

Primer sequences used for qRT-PCR.

Gene	Forward	Reverse
Amh	CTGGCTAGGGGAGACTGGA	AGGTGGAGGCTCTTGGAACT
Amhr2	CCAACATCCCATCCACTTG	CTGCGTCCCAGCAATCTT
Fshr	CCAGCCTTACCTACCCCAGT	CAAATTGGATGAAGTTCAGAGGT
Lhcgr	GGGACGACGCTAATCTCG	CCTGGAAGGTGCCACTGT
Cyp19a1	CCACTCCTGCTGATCATGG	TCCCAGACAGTAGCCAGGAC
Cyclophilin	CCG GGA CAA GCC ACT GAA	GGC GAA GGG TTT CTC CAC TT

Ovarian morphology for follicle count

Paraffin-embedded, formalin fixed, ovaries (n = 6 per group) were sectioned using a rotary microtome at 5 μm and stained with haematoxylin and eosin (H&E). Follicles at different stages of development were counted every 10th section under a light microscope (Olympus BX51), according to previously established definitions (Myers et al., 2004): primordial follicle, primary follicle, secondary follicle, antral follicle and corpus luteum. Slides were coded so the examiner was blinded to the group of mice.

Statistics

All statistical analyses were performed using Prism 8.0 software. Data were reported as mean ± SEM. Mann–Whitney U test was used because the data were not normally distributed. Repeated measures ANOVA was performed for the growth curve to assess changes over time. For GTT and ITT, glucose concentrations were plotted over time and the total AUC for 120 min for glucose concentrations in milligrams per decilitre was calculated. The AUC was calculated using Prism 8.0 software, which provides the data following the trapezoid rule, with the formula ΔX*([(Y1 + Y2)/2]-Baseline]. Experimental results were considered significant when P < 0.05.

Results

Maternal body weight and daily food intake

As seen in Table II, the daily food intake of the dams was similar between both groups. Additionally, there was no statistically significant difference in dam body weight on the days of mating or on embryonic days 0.5, 9.5 and 19.5 (P > 0.05). Weight gain during pregnancy was also similar between both groups (P > 0.05).

Table II.

Total food intake and body weight of dams on L-AGE and H-AGE diet.

	Food intake (g/day)	Weight on day of mating (g)	Weight on e0.5 (g)	Weight on e9.5 (g)	Weight on e19.5 (g)	Total weight gain (g)**
Dams on L-AGE diet (n = 5)	3.2 ± 0.1	25.5 ± 1.5	26.1 ± 1.2	32.3 ± 0.8	56.0 ± 1.4	29.9 ± 1.1
Dams on H-AGE diet (n = 5)	3.4 ± 0.1	24.8 ± 1.5	24.8 ± 1.0	31.1 ± 2.2	52.0 ± 6.0	27.2 ± 5.9
P-value*	0.3	0.7	0.4	0.6	0.5	0.7

Repeated measures ANOVA for body weight from day of mating until e19.5 was performed (P = 0.5). e, embryonic day.

^*Mann–Whitney U test between both groups.

^**Total weight gain is difference between weight on e19.5 and weight on e0.5.

Exposure to perinatal H-AGE diet affects body weight but not glucose or insulin tolerance

At birth

Although the number of pups per litter was similar for dams that consumed the H-AGE (n = 13) or L-AGE (n = 10) diets (13.00 ± 1.82 vs. 14.00 ± 1.82; respectively, P > 0.05; Fig. 1A), mice exposed in utero to the H-AGE diet were born 1 day earlier than mice exposed in utero to the L-AGE diet (data not shown). Additionally, body weight at birth of pups born to dams that consumed the H-AGE diet (n = 13) was significantly lower compared to the body weight of pups born to dams that consumed the L-AGE diet (n = 10) (1.38 ± 0.10 vs. 1.58 ± 0.16, P < 0.05; Fig. 1B).

Figure 1.

Litter size, body weight at birth and growth curve of pups following perinatal exposure to H-AGE or L-AGE diet. (A) Litter size per animal was similar in dams that consumed the H-AGE (n = 13) or L-AGE (n = 10) diet for 2 weeks prior to mating and then for 6 weeks during pregnancy and lactation (P > 0.05). (B) Body weight at birth of pups born to dams that consumed the H-AGE diet (n = 26) were significantly lower compared to the body weight of pups born to dams that consumed the L-AGE diet (n = 26) during pregnancy and lactation. (C) Developmental growth of mice shows higher body weight in pups born to dams that consumed the L-AGE diet at birth, and at weeks 3, 6, 7, 8, and 14 of age. Data are presented as mean ± SEM. Mann–Whitney U test was used for litter size and body weight at birth. Repeated measures ANOVA was performed for growth curve (*P < 0.05).

Postnatally and at sacrifice

Offspring from both L-AGE and H-AGE dams were all weaned onto the L-AGE diet at postnatal day 21 and studied through 16 weeks of age. Growth curve showed that pups born to dams maintained on the H-AGE diet had significantly lower body weight (P < 0.05) at 3, 6, 7, 8 and 14 weeks of age compared to pups born to dams maintained on the L-AGE diet (Fig. 1C). Pups from dams that consumed the H-AGE diet had significantly lower abdominal fat pad weight (P = 0.009; Fig. 2A), but similar liver weight (P = 0.7; Fig. 2B) compared to pups of dams that consumed the L-AGE diet (n = 10–13 per group). There was no significant difference in AUC for GTT (P = 0.7; Fig. 3A) or AUC for ITT (P = 0.2; Fig. 3B) between pups exposed perinatally to the H-AGE or L-AGE diets.

Figure 2.

Body, abdominal fat pad and liver weights at sacrifice of mice following perinatal exposure to high dietary AGEs. Pups of dams that consumed the H-AGE diet with weaning to the L-AGE diet had significantly lower gonadal fat pad weight (A; P = 0.009) but similar liver weight (B; P = 0.7) compared to pups of dams that consumed the L-AGE diet with weaning to the L-AGE diet, when they were sacrificed at 16 weeks of age (n = 10–13 per group). Data are presented as mean ± SEM. Mann–Whitney U test was used for comparison.

Figure 3.

Glucose tolerance test (GTT) and insulin tolerance test (ITT) in mice following perinatal exposure to high dietary AGEs. There was no significant difference in AUC for GTT (P = 0.7; A) or AUC for ITT (P = 0.2; B) between pups born to dams that consumed the H-AGE or L-AGE diets with weaning to the L-AGE diet for 7 weeks (n = 10–13 per group). Data are presented as mean ± SEM. Mann–Whitney U test was used for comparison.

Perinatal exposure to H-AGE diet delays puberty onset and disrupts oestrous cyclicity

To determine whether perinatal exposure to the H-AGE diet affects pubertal onset, we assessed the age at vaginal opening in pups born to dams that consumed the H-AGE or L-AGE diet. Mice of dams who consumed the H-AGE diet during pregnancy and lactation had significantly delayed vaginal opening (Fig. 4B) compared to mice of dams who consumed the L-AGE diet during pregnancy and lactation (Fig. 4A). As seen in Fig. 4A and B (black arrows), by the seventh day, 100% of pups born to dams that consumed the L-AGE diet displayed vaginal opening compared to only 66.6% of pups born to dams that consumed the H-AGE diet (P < 0.05).

Figure 4.

Perinatal exposure to H-AGE diet delays the first vaginal opening of pubertal offspring. Pups of dams that consumed the H-AGE diet during pregnancy and lactation had significantly delayed vaginal opening (B) compared to pups of dams that consumed the L-AGE diet during pregnancy and lactation (A). Black arrows show that by postnatal day 7, 100% of pups born to dams that consumed the L-AGE diet had vaginal opening compared to only 66.6% of pups born to dams that consumed the H-AGE diet (n = 10–13 per group).

After completion of the pubertal transition, female mice typically exhibit 5-day oestrous cycles. To determine whether perinatal exposure to the H-AGE diet effects was sustained beyond puberty, we used daily vaginal smears to monitor oestrous cycle length and the percentage of time spent in each stage of oestrous over a 3-week interval (n = 10–13 per group). Pups of dams that consumed the H-AGE diet during pregnancy and lactation spent less time in the pro-oestrous phase and more time in the metoestrus phase compared to pups of dams that consumed the L-AGE diet during pregnancy and lactation (P = 0.04; Fig. 5).

Figure 5.

Perinatal exposure to H-AGE diet disrupts the oestrous cycle of pubertal offspring. Daily vaginal smears monitored oestrous cycle length and the percentage of time spent in each stage of oestrous over a 3-week interval (n = 10–13 per group). Pups of dams that consumed the H-AGE diet during pregnancy and lactation spent less time in the pro-oestrus phase and more time in the metoestrus phase compared to pups of dams that consumed the L-AGE diet during pregnancy and lactation (*P = 0.04). Data are presented as % time. Mann–Whitney U test was used for comparison. P, pro-oestrus; E, oestrus; D, dioestrus; M, metoestrus.

Exposure to maternal H-AGE diet affects serum markers of ovarian reserve and adiposity

Serum AMH is known to be one of the best markers of ovarian reserve (van Rooij et al., 2005), thus we measured serum AMH in all pups. Additionally, the two adipokines, leptin and adiponectin, play an important role in the onset of puberty (Nieuwenhuis et al., 2020), thus we measured serum leptin and adiponectin in all pups. Pups from dams that consumed the H-AGE diet during pregnancy and lactation had significantly lower serum AMH (Fig. 6A), leptin (Fig. 6B) and adiponectin (Fig. 6C) levels compared to pups of dams that consumed the L-AGE diet during pregnancy and lactation (P < 0.05 for all).

Figure 6.

Perinatal exposure to H-AGE diet lowers serum markers of ovarian reserve and lowers serum leptin and adiponectin in pubertal offspring. At 16 weeks of age, pups of dams that consumed the H-AGE diet during pregnancy and lactation had significantly lower serum anti-Mullerian hormone (AMH) (A), leptin (B) and adiponectin (C) levels compared to pups of dams that consumed L-AGE diet during pregnancy and lactation (*P < 0.05, n = 10–13 per group). Data are presented as mean ± SEM. Mann–Whitney U test was used for comparison.

Ovarian folliculogenesis and gene expression are altered by perinatal exposure to H-AGE diet

Pups from dams that consumed the H-AGE diet during pregnancy and lactation had similar ovary sizes compared to pups of dams that consumed the L-AGE diet during pregnancy and lactation (data not shown). However, they had significantly fewer corpora lutea (CL) and exhibited arrested folliculogenesis, with most follicles in the secondary follicle stage (Fig. 7) compared to pups of dams that consumed the L-AGE diet during pregnancy and lactation. Additionally, they had significantly lower ovarian mRNA expression levels of Amh (P = 0.005), Amhr2 (P = 0.02) and Cyp19a1 (P = 0.03) compared to pups of dams that consumed L-AGE diet during pregnancy and lactation (Fig. 8). The mRNA expression levels of ovarian Fshr (P = 0.3) and Lhcgr (P = 0.07) were similar between groups (Fig. 8).

Figure 7.

Effect of perinatal exposure to H-AGE diet affects follicular development in pubertal offspring. Pups of dams that consumed the H-AGE or L-AGE diets during pregnancy and lactation, with weaning to the L-AGE diet, had their ovaries extracted at 16 weeks of age. Paraffin-embedded ovarian tissues were sectioned at 5 μm and stained with haematoxylin and eosin (H&E). Pups of dams that consumed the H-AGE diet during pregnancy and lactation had significantly fewer corpora lutea (CL) and exhibited arrested folliculogenesis, with most follicles in the secondary follicle stage (*P < 0.05, n = 6 per group). Data are presented as mean ± SEM. Mann–Whitney U test was used for comparison.

Figure 8.

Perinatal exposure to H-AGE diet affects the relative expression of ovarian genes involved in folliculogenesis and steroidogenesis in pubertal offspring. Pups of dams that consumed H-AGE or L-AGE diets during pregnancy and lactation, with weaning to L-AGE diet, had their ovaries extracted at 16 weeks of age and subjected to real-time polymerase chain reaction (RT-PCR). Pups of dams that consumed the H-AGE diet during pregnancy and lactation had significantly lower ovarian expression of Amh, Amhr2 and Cyp19a1 mRNA levels compared to pups of dams that consumed the L-AGE diet during pregnancy and lactation (*P < 0.05, n = 10–13 per group). Data are presented as arbitrary units of the mean ± SEM. Mann–Whitney U test was used for comparison.

Discussion

The present study demonstrates that perinatal exposure to a H-AGE diet causes delayed weight gain, delayed puberty onset, and disruption in oestrous cyclicity characterised by showing extended periods of metoestrus and reduced frequency of pro-oestrus. These phenotypes were not associated with disturbances in glucose metabolism but were associated with lower serum levels of the two adipokines, leptin and adiponectin. In addition, perinatal exposure to H-AGE diet caused lower transcription of the expression of ovarian genes involved in folliculogenesis (Amh and its receptor Amhr2), and steroidogenesis (Cyp19a1, aromatase enzyme responsible for conversion of testosterone to oestradiol), as well as lower serum AMH levels, a marker of ovarian reserve (van Rooij et al., 2005).

Puberty depends on adiposity (Abou El Ella et al., 2020) and on activation of all the components of the hypothalamic–pituitary–gonadal axis (Leka-Emiri et al., 2017). The onset of puberty is triggered by neuroendocrine events characterised by dynamic changes in input onto GnRH neurons in the hypothalamus, with leptin being one of the main triggers (Han et al., 2002, 2005; Egan et al., 2017; Abou El Ella et al., 2020). These changes ultimately induce GnRH release that leads to activation of the pituitary-gonadal axis characterised by an oestrogen surge, and subsequently vaginal opening (Dicken et al., 2012). It is possible that perinatal exposure to the H-AGE diet delays puberty either by delaying weight gain, as reflected by a delayed growth curve and lower abdominal fat weight on necropsy, which leads to low leptin levels observed in the serum, and/or by compromising the trans-synaptic excitatory and/or inhibitory afferent input required for peripubertal activation of GnRH neurons. Future studies should assess body composition at early time points to determine whether perinatal exposure to H-AGE diet delays white abdominal fat development or whether it alters lipogenesis/lipolysis. Interestingly, a recent study demonstrated that maternal intake of a diet with a moderately increased content of AGEs from puberty to weaning of offspring modulates neurodevelopment such as manifestation of physiological reflexes like hind-limb placing, fore- and hind-limb grasp, negative geotaxis and surface righting as well as a behavioural phenotype (such as working memory and anxiety-like behaviour) in later life (Csongova et al., 2019). Thus, we hypothesize that perinatal exposure to high dietary AGEs could alter the peripubertal activation of GnRH neurons. Given the rising preference of ingestion of diets containing large amounts of AGEs, more studies are needed to determine, independent of obesity, whether the hypothalamic-pituitary axis of peripubertal females is susceptible to the adverse effects of AGEs.

The completion of the pubertal transition and attainment of reproductive competence depend upon functional and gonadotropin-responsive gonads (Safranski et al., 1993). Within 1 week of vaginal opening, a second oestrogen surge occurs, which is followed by ovulation and first oestrus (Safranski et al., 1993). Vaginal opening and first oestrus signify the completion of puberty and the potential to reproduce. Perinatal exposure to the H-AGE diet disrupted oestrous cyclicity by showing extended periods of metoestrus and reduced frequency of pro-oestrus potentially reflecting hypothalamic, pituitary, or ovarian dysfunction. We sought to determine whether abnormal oestrous cyclicity reflected a primary ovarian dysfunction. We then assessed gross histologic evaluation of the ovaries as well as expression of genes specifically expressed in the ovaries and we found lower transcription of Amh, Amhr2 and Cyp19a1. In addition, we found that ovaries collected from mice exposed to the perinatal H-AGE diet exhibited fewer CL with a significantly higher number of ovarian follicles arrested in the secondary stage of development. This arrested follicular development was further demonstrated by a significantly lower transcription of Amh and Amhr2 genes, both of which are important in folliculogenesis (Umer et al., 2019). Finally, the lower transcription of Cyp19a1, which is an aromatase enzyme responsible for conversion of testosterone to oestradiol (Dewailly et al., 2016), could explain potentially lower serum oestradiol levels, which are important in both vaginal opening and regular oestrous cyclicity. Studies that assess serum oestradiol levels are needed to investigate this possibility.

The average intake of AGEs in healthy adults was recently found to be approximately14,700 kU/day (Uribarri et al., 2010). This has been tentatively used to define in humans a H-AGE or L-AGE diet, depending on whether the estimated daily AGE intake is significantly greater or <15 000 kU; respectively. Nearly 10% of dietary AGEs are absorbed in the gastrointestinal tract and is delivered to liver and to other tissues including the ovaries (Semba et al., 2012). One-third of dietary AGEs are excreted in the urine, and the remaining is accumulated in the body, including the ovaries, leading to oxidative and inflammatory damage (Semba et al., 2012). Studies in humans have shown that H-AGE diets contain at least three times the CML levels than those observed in L-AGE human diets (Tantalaki et al., 2014; Abate et al., 2015).

It is noteworthy to state that studies in diabetic men have demonstrated that AGEs are deposited in the male reproductive tract such testis and epididymis, and on spermatozoa, ultimately leading to male subfertility (Mallidis et al., 2009). The receptor for AGEs (RAGE) is expressed in the testis, epididymis and sperm acrosomes. Additionally, the number of sperm displaying RAGE and the overall RAGE protein levels found in sperm and seminal plasma were found to be significantly higher in diabetic men (Mallidis et al., 2007). These results suggest that RAGE could play an important role in sperm dysfunction especially in diabetic men where their levels are elevated.

Even though there were no significant changes in blood glucose or insulin, perinatal exposure to H-AGE diet in transgenic NOD8.3 mice has been shown to lead to pancreatic beta-cell dysfunction whereby insulin, proinsulin and glucagon secretion were greater in pancreatic islets isolated from offspring of the L-AGE diet mice compared to H-AGE diet mice (Borg et al., 2018). Similarly, our results showed no changes in the ITT or GTT in the offspring exposed to the perinatal H-AGE diet, however, we did not assess whether the differences in the number, or the specific type, of cells in pancreatic islets are present in H-AGE offspring compared to L-AGE offspring.

Jinno et al. (2011) quantified the levels of toxic AGE (TAGE), pentosidine, and CML in the blood and the follicular fluid of 157 infertile women undergoing in-vitro fertilisation. The results showed an accumulation of TAGE, pentosidine, and CML in follicular fluid and that accumulation of TAGE in serum is significantly negatively correlated with follicular growth (as observed on ultrasound), fertilisation and embryonic development. The most significant predictors for ongoing pregnancy were lower concentrations of pentosidine in follicular fluid and TAGE in serum, independent of age and ovarian reserve (as indicated by serum day 3 FSH levels). These data provide clinical evidence of an important role of AGEs' accumulation in ovarian dysfunction as well as adverse outcomes for infertile women with elevated AGEs.

Conclusion

In conclusion, this study provides novel evidence that perinatal exposure to high dietary AGEs affects the growth as well as the pubertal transition and establishment of regular oestrous cyclicity in the female offspring. Additionally, it causes arrests in ovarian follicular development and reduces ovulatory events. These findings indicate that perinatal exposure to high dietary AGEs, even despite weaning to L-AGE diet postnatally, can lead to deficits in perinatal growth, pubertal onset and reproductive organ development in female mice. We also propose that perinatal exposure to high dietary AGEs could impair female reproductive function by affecting hypothalamic function, and thus, future studies focusing on the neuroendocrine axis are necessary to further define the mechanisms by which dietary AGEs influence female reproduction.