Introduction
In recent decades, eating habits have changed with increased fat consumption, exceeding the World Health Organization recommendations of total fat ≤ 30% total calories and saturated fat <10% total calories per day.1 Associated with a low energy expenditure, excess fat intake contributes to the worldwide increase in metabolic dysfunctions such as overweight, obesity, dyslipidemia and diabetes.2–Reference Lin, Xu and Pan4 These chronic diseases now affect young populations in reproductive age. In women, the prevalence of preexisting type 2 diabetes and the incidence of maternal obesity at the start of pregnancy has increased worldwide.Reference Poston, Caleyachetty and Cnattingius5 Indeed, in the early 2010s, the proportion of obese women of childbearing age was over 30% in the United States and between 7 and 25% among European countries.Reference Poston, Caleyachetty and Cnattingius5–Reference Guelinckx, Devlieger, Beckers and Vansant7 Dietary changes and reduction in physical activity also affect low-income and middle-income countries, with the result that excess weight is now more common than underweight among women of reproductive age in most developing countries.Reference Poston, Caleyachetty and Cnattingius5 Thus, worldwide, a large proportion of pregnancies take place in a long-standing altered metabolic environment.
Overweight and obesity can be detrimental to fertility and have been associated with fetal defects and congenital abnormalities.Reference Poston, Caleyachetty and Cnattingius5 Moreover, altered maternal metabolic health can be detrimental to the lifelong offspring cardiometabolic health, increasing the likelihood that they will develop obesity and metabolic disease,Reference Palinski8–Reference Langley-Evans10 a phenomenon currently known as “Developmental Origins of Health and Disease” (DOHaD).Reference Langley-Evans10,Reference Gluckman, Hanson and Beedle11 Epidemiological and animal models studies have explored consequences of a variety of maternal metabolic alterations, deciphered critical windows of susceptibility and identified the discrepancy between males and females in terms of consequences.Reference Aiken and Ozanne12 Animal models were developed to discriminate effects of maternal metabolic parameters on offspring health independently of other maternal risk factors.Reference Palinski8,Reference Williams, Seki, Vuguin and Charron9 In rodents, a maternal high-fat diet during pregnancy was shown to induce sex-dependent susceptibility to develop obesity, dyslipidemia, cardiovascular deregulation, impaired liver lipid metabolism and glucose homeostasis in the adult offspring.Reference Williams, Seki, Vuguin and Charron9 In rabbits, a maternal hypercholesterolemic diet administered 2 weeks before mating and during gestation induced aortic lesions in offspring.Reference Palinski, D’Armiento and Witztum13 Moreover, a hypercholesterolemic and hyperlipidic diet administered 8 weeks before mating induced intrauterine growth retardation in both sexes with fetal dyslipidemia and led to adult offspring overweight associated with hypertension, with more effects on males than females.Reference Picone, Laigre and Fortun-Lamothe14,Reference Tarrade, Rousseau-Ralliard and Aubrière15 Using the same diet, lipid droplet accumulation has been observed as early as the preimplantation embryo, in the trophoblastic cells of the blastocyst.Reference Tarrade, Rousseau-Ralliard and Aubrière15
The periconceptional period, which includes ovogenesis, generation of a zygote from two gametes and first stages of embryonic development, represents a high vulnerability time to the maternal metabolic environment.Reference Watkins, Lucas and Fleming16,Reference Fleming, Watkins and Velazquez17 In the preimplantation embryo, epigenetic reprograming, embryonic genome activation, differentiation of cell lineages from totipotent cells, X inactivation and first sex-linked differential gene expression occur.Reference Duranthon, Watson and Lonergan18 Studies combining exposure during the periconceptional period until the end of the preimplantation period and subsequent embryo transfer into a control recipient female have been set up to discriminate the role of the periconceptional vs the post-implantation gestational period in the programing of offspring health. In rabbits, exposure to hyperglycemia during periconception only induced fetal hypotrophy, hyperglycemia and dyslipidemia as well as abnormalities in placental vascularization and nutrient transport functions, close to term.Reference Rousseau-Ralliard, Couturier-Tarrade and Thieme19 In sheep, exposure to maternal overnutrition during periconception increased body fat mass in female offspring and decreased the expression of insulin signaling molecules in liver and skeletal muscle in both male and female offspring.Reference Nicholas, Morrison, Rattanatray, Zhang, Ozanne and McMillen20 In contrast, maternal undernutrition during the periconceptional period altered placental and fetal growth dynamics and increased fetal arterial blood pressure.Reference Edwards and McMillen21 The mechanisms underlying the sex-specific programing of offspring phenotype by periconceptional maternal metabolic environment remains to be elucidated.
Previous studies in our laboratory have shown that a hypercholesterolemic and hyperlipidic diet administered to rabbit does from before puberty induced dyslipidemia and high adiposity but not obesity in females at mating ageReference Picone, Laigre and Fortun-Lamothe14 and induce long-term effects on offspring.Reference Picone, Laigre and Fortun-Lamothe14,Reference Tarrade, Rousseau-Ralliard and Aubrière15 The objectives of this study were to analyze the effects of the periconceptional maternal hypercholesterolemic and hyperlipidic diet, previously shown to induce long-term effects on offspring, on oocyte transcripts content, uterine fluid composition and preimplantation embryo gene expression in a rabbit model.
Methods
Animals
New Zealand White female rabbits (INRA 1077 Line) were housed individually in one building maintained at 18–20°C. At 10 weeks of age, does were fed ad libitum with either a lipid cholesterol-enriched diet (HH) or a control diet (C). The experimental HH diet, a control diet supplemented with 6% soybean oil and 0.2% cholesterol, contained quantitatively more fatty acids than the C diet (2% of lipids) from each fatty acid class and provided 16% more energy than the C diet. Nutrient and chemical composition of C and HH diets have been previously published.Reference Picone, Laigre and Fortun-Lamothe14,Reference Tarrade, Rousseau-Ralliard and Aubrière15
At 18 weeks of age, 8 C and 8 HH does were superovulated as previously describedReference Tarrade, Rousseau-Ralliard and Aubrière15 and subsequently mated with either a C vasectomized male for oocytes retrieval or a C male for blastocyst collection. At 16 hours post-coïtum (hpc), 4 C and 4 HH does were euthanized and freshly ovulated metaphase II oocytes were recovered from oviducts by flushing, removed from the cumulus by incubation in PBS/0.5% hyaluronidase and by mechanic treatment then frozen at −80°C until RNA extraction. At 144hpc, 2 C and 2 HH does were euthanized and blastocysts were recovered from uterus by flushing and subsequently dry frozen individually for sex determination, microarray analysis and RT-qPCR experiments. At 168hpc, 4 C and 4 HH does were anesthetized, then laparotomy, externalization of uterine horn and isolation of embryonic vesicle were realized as previously described.Reference Calderari, Daniel and Mourier22 The 70 MHz probe (MS-700) of a micro-ultrasound platform (Vevo2100, VisualSonics Inc., Toronto, Ontario, Canada) was positioned on the uterine horn opposite an embryonic vesicle to visualize it in real time and guide the micropuncture of blastocoelic fluid and of uterine fluids as described previously.Reference Calderari, Daniel and Mourier22 Blastocoelic and uterine fluids were snap frozen at -80°C until for metabolomics experiments.Reference Calderari, Daniel and Mourier22
Sex determination
DNA from each blastocyst was extracted using the AllPrep DNA/RNA Mini Kit (Qiagen) according to instructions of the manufacturer. Blastocyst' sex was determined by the sex-determining region of the Y chromosome gene (SRY) gene detection on genomic DNA using nested PCR as previously described.Reference Okamoto, Patrat and Thépot23 Sex determination was confirmed using detection of SRY gene using qPCR.Reference Jolivet, Daniel-Carlier and Harscoët24
Genome-wide gene transcription profiling
For the microarray analysis, total RNA was extracted from pools of oocytes (n = 4 pools of 20 oocytes C and n = 4 pools of 20 oocytes HH) and from individual blastocysts (n = 15 C and n = 12 HH) using the PicoPure RNA Isolation kit (Invitrogen). Prior to the elution, a purification procedure was performed using DNAse I (Qiagen) treatment at 25°C for 15 min. Total extracted RNAs were stored at − 80°C for further RNA labeling. Transcriptional profiling of oocytes and blastocysts were performed using customized rabbit microarrays (GEO accession GPL21733, Agilent-075973 Rabbit Microarray V3 020908 and GEO accession GPL18913, Agilent-042421 Rabbit BDR version 2, Agilent Technologies, respectively).Reference Rousseau-Ralliard, Valentino and Aubrière25 RNA amplification, labeling and hybridization was performed as previously described.Reference Tapponnier, Afanassieff and Aksoy26
After hybridization, the scanned images were analyzed using Feature Extraction software (v10.7.3.1; Agilent Technologies). The data were normalized using intra-array median subtraction and log2 transformation. The raw data intensity files were read into R (www.r-project.org). The identification of differentially expressed genes (DEG) from HH and C oocytes and blastocysts was achieved using the Limma package 67. Interaction between diet ad sex was assessed. P values obtained by this analysis were adjusted for multiple testing using the Benjamini-Hochberg procedure. Probes with an adjusted P value < 0.05 were considered significant.
The gene functional classification of DEG was carried out using DAVID bio-informatic database by identifying the first two terms of Gene Ontology (GO) in biological processes (BP) and molecular functions (MF) (DAVID Bioinformatics resources 6.8, NIAID/NIH) (https://david.ncifcrf.gov/).
Pathway analysis of gene transcription data
Transcriptomic data (normalized intensities of annotated probes) were analyzed by Gene Set Enrichment Analysis (GSEA version 4.1.0).Reference Subramanian, Tamayo and Mootha27 The GSEA approach was used on transcriptomic data to systematically identify biological pathways enriched using GO databases (BP and MF) and hallmark gene sets (v7.2), part of Molecular Signatures Database (MSigDB) collections. Hallmark gene sets comprise 50 gene sets that represent well-defined biological states or processes processes. The GSEA-derived normalized enrichment score (NES) was used for the visualization of pathway regulation. The NES was calculated for each gene set and reflects the degree to which a gene set is overrepresented at the top or bottom of a ranked list of genes created by GSEA for each gene set according to differential gene expression between HH or C blastocyst. Positive and negative ES indicated that the gene set was overrepresented or under-represented, respectively. Gene set was considered significantly enriched when FDR score was less than 0.25.
Proton nuclear magnetic resonance spectroscopy measurements
Proton nuclear magnetic resonance (1H-NMR) samples were prepared using 20 µl blastocoelic and uterine fluids diluted with 180 µl of 0.2 M potassium phosphate buffer in deuterium oxide. The spectra were recorded at 298 K on a Bruker Ascend 600-MHz spectrometer equipped with a TCI cryoprobe (Triple resonance Cryoprobe for Inverse detection) as previously described.Reference Calderari, Daniel and Mourier22 Metabolite quantifications using the Electronic Reference To access In vivo Concentrations peak as a quantitative reference were obtained by the specific subroutine of the Bruker TopSpin 3.2 program as previously described.Reference Calderari, Daniel and Mourier22 The metabolite data from blastocoelic and uterine fluids were analyzed by multivariate statistical data analysis, using MetaboAnalyst. Sample normalization was performed using the MetaboAnalyst procedure entitled “A pooled sample from group” using T values to normalize the data set. The data were scaled using auto scaling mode (mean-centered and divided by the standard deviation of each variable) of MetaboAnalyst. Data were analyzed using partial least squares discriminant analysis (PLS-DA). The overall quality of the models was judged by cumulative R 2 defined as the proportion of variance in the data explained by the model and cumulative Q 2, the class prediction ability of the model obtained by cross-validation. Comparison between groups was performed using a t-test. P < 0.05 was considered statistically significant and data are expressed as ratio.
Results
HH diet induces few differences in oocyte transcripts content
To determine the effect of HH maternal diet on oocyte transcripts content, 4 C and 4 HH pools of 20 oocytes were profiled using rabbit dedicated microarrays. Differential gene expression analyses between C and HH oocytes transcriptomes revealed 57 differentially expressed probes (Fig. 1a). These probes corresponded to 15 annotated genes, 12 being over-expressed and 3 under-expressed in HH as compared to C oocytes (Table 1 and Supplementary Table S1).
Figure 1. Volcano plot of microarray data of oocytes (a) and blastocyts (b). The data of all probes are plotted as log2 fold change (Log2FC) versus the -log10 of the adjusted p-value (-Log10(pvalue)). Probes selected as significantly down are highlighted as blue dots and probes selected as significantly up are highlighted as red dots.
Table 1. Differential expressed genes in oocytes of females under HH vs C diet
Among these DEGs, 3 were involved in meiosis and cell cycle: zinc finger CW type with PWWP domain 1 (ZCWPW1, related to “meiosis I” GO term), Ret finger protein-like 1 (RFPL1, related with “cell cycle” GO term) and Cyclin A1 (CCNA1, related to “G1/S transition of mitotic cell cycle” GO term) (Supplementary Table S1). Five DEGs were involved in the regulation of gene expression, transcription and translation processes: Arginyl-tRNA synthetase (RARS, related to “tRNA binding” GO term), Vav guanine nucleotide exchange factor 1 (VAV1, related to “regulation of transcription DNA templated” GO term), RIO kinase 1 (RIOK1, related to “rRNA processing” GO term), Mitochondrial ribosomal protein L55 (MRPL55, related to “translation” GO term) and ZCWPW1 (related to “methyl-CpG binding” GO term) (Supplementary Table S1). Collagen type XXVIII alpha 1 chain (COL28A1) and myosin heavy chain 4 (MYH4) were involved in extracellular matrix (Supplementary Table S1).
Despite these gene-by-gene statistical differences of 15 genes, global gene expression analyses using Gene Set Enrichment Analysis (GSEA) on hallmarks gene set collections and Gene Ontology Biological Process and Molecular Function databases did not identify any significant functional gene set enrichment between HH and C oocytes.
HH diet modifies uterine fluid composition
To determine the effect of HH diet on embryo’s microenvironment, uterine fluid in the vicinity of the blastocyst was collected in vivo using a real-time ultrasound biomicroscopy guided puncture. 1H-NMR metabolomic profiling was conducted on seven uterine fluids collected from four C females and on four uterine fluids collected from two HH females. PLS-DA analysis of metabolites concentrations displayed a slight separation between HH and C uterine fluids (Fig. 2a) (Accuracy = 0.73; R2 = 0.81; Q2 = 0.13). Table 2 shows the ratio between the metabolite concentration in HH and C uterine fluids. Pyruvate was identified more concentrated in HH than in C uterine fluid (Table 2).
Figure 2. Multivariate statistical data analysis of C (Δ) and HH (+) uterine (a) and blastocoelic (b) fluids. Supervised partial least squares discriminant analysis of metabolites concentrations.
Table 2. Metabolites concentration ratio of between HH and C uterine and blastocoelic fluids
nd, no detection.
* indicate significant differences between HH and C within the same column (p < 0.05).
HH diet modifies blastocoelic fluid composition
To determine the effect of HH diet on blastocyst’s metabolism, real-time ultrasound biomicroscopy guided puncture was used to collect blastocoelic fluid of seven embryos from four C does and five embryos from four HH does.
Metabolomic composition were determined using 1H-NMR. PLS-DA analysis of metabolites concentrations displayed a slight separation between HH and C blastocoelic fluids (Fig. 2b) (Accuracy = 0.67; R2 = 0.899; Q2 = 0.25). Table 2 shows the ratio between HH and C metabolites concentrations of blastocoelic fluids. Alanine, glucose and methionine were less concentrated in HH than in C uterine fluid (Table 2).
Maternal HH diet impaired blastocyst gene expression
The effect of HH maternal diet on blastocyst gene expression was determined by profiling individual blastocysts (15 C and 12 HH) using a rabbit dedicated microarray.
First, the expression data were analyzed regardless of independently of the sex oh the blastocyst sex. Gene-by-gene statistical analysis between C and HH blastocysts transcriptomes revealed 49 differentially expressed probes (Fig. 1b). These significantly differentially expressed probes corresponded to 26 annotated genes (Table 3 and Supplementary Table S2), 16 over-expressed and 10 under-expressed in HH as compared to C blastocysts. To identify coordinated gene expression changes, gene expression datasets were analyzed globally using GSEA on hallmarks gene set collections and Gene Ontology Biological Process and Molecular Function databases (Fig. 3 and Supplementary Table S3). In HH blastocysts, when sex was not considered, GSEA only identified the positive enrichment of one functional GO term and the negative enrichment of 22 functional gene sets (7 hallmarks and 15 GO terms).
Figure 3. Enrichment analysis of blastocysts gene transcription data. Gene sets enrichment analysis was performed on hallmark gene sets (H) and GO terms (GO) database. Gene set was considered significantly enriched when FDR score was less than 0.25. Normalized enrichment score (NES) was used for the visualization. Positive and negative NES reflects respectively over-representation and under-representation in HH blastocysts.
Table 3. Differential expressed genes in HH vs C blastocysts
Transcriptome analyses of HH blastocysts identified deregulation of metabolic processes. Gene-by-gene statistical analysis identified DEG implied in metabolism as the sterol carrier protein 2 (SCP2) and the phosphogluconate dehydrogenase (PGD) (Table 3, Supplementary Table S2). Functional analysis identified in HH blastocysts the negative enrichment of several gene sets involved in metabolism as “Adipogenesis” (NES = −1.56), “Fatty Acid Metabolism” (NES = −1.50) or “Mtorc1 Signaling” (NES = −1.71) hallmark gene sets and “Sterol Transfer Activity” (NES = −1.70), “Lipid Transfer Activity” (NES = −1.63) or “Neutral Amino Acid Transport” (NES = −1.97) GO terms (Fig. 3 and Supplementary Table S3).
Transcriptomes of HH blastocysts exhibited also deregulation of the expression of genes involved in transcriptional regulation. HH blastocysts exhibited over-expression of transcription factors: early growth response 1 (EGR1), early growth response 2 (EGR2), Jun proto-oncogene AP-1 Transcription Factor Subunit (JUN) and Fos proto-oncogene AP-1 transcription factor subunit (FOS) (Table 3, Supplementary Table S2). Global analysis pointed to deregulation of genes involved on DNA damage with under-representation of “DNA repair” (NES = −1.63) and “G2M Checkpoint” (NES = −1.48) hallmark gene sets (Fig. 3 and Supplementary Table S3).
A sex-specific response of the blastocyst to the maternal HH diet
The effect of HH maternal diet on gene expression in blastocysts was then explored according on sex. In the C group, we identified 8 males and 7 females and in the HH group, we identified 7 males and 5 females. The gene-by-gene statistical analysis did not detect significant differential diet-induced expression in males or females. GSEA analysis was performed as before, using hallmark gene sets and GO terms on male or on female blastocysts data sets.
In male HH blastocysts, GSEA identified the positive enrichment of one hallmark gene set and the negative enrichment of 2 GO terms (Fig. 4a and Supplementary Table S3). These 3 groups of genes were not identified as enriched in the sex-independent analysis (Fig. 3 and Supplementary Table S3).
Figure 4. Enrichment analysis of gene transcription data of blastocysts according to sex. Gene sets enrichment analysis was performed on hallmark gene sets (H) and GO terms (GO) database in male (a) and female (b) blastocysts. Gene set was considered significantly enriched when FDR score was less than 0.25. Normalized enrichment score (NES) was used for the visualization. Positive and negative NES reflects respectively over-representation and under-representation in HH blastocysts.
In female HH blastocysts, GSEA identified the enrichment of large number of gene sets: 13 hallmarks and 43 GO terms (Fig. 4b and Supplementary Table S3). Among them, 24 were under-represented (11 hallmark gene sets and 13 GO terms) (Fig. 4b and Supplementary Table S3). As in the sex-independent analysis, functional analysis identified an under-representation of metabolism gene sets such as “Fatty Acid Metabolism” (NES = −1.63), “Mtorc1 Signaling” (NES = −2.03), as well as “glycolysis” (NES = −1.52) or “cholesterol homeostasis” (NES = −1.48) hallmark gene sets in female HH blastocysts as compared to female C blastocysts (Fig. 4b and Supplementary Table S3). Functional analysis also identified under-representation of gene sets involved in transcriptional regulation response in female HH blastocysts such as “E2F targets” (NES = −1.60), “DNA repair” (NES = −1.60), “G2M Checkpoint” (NES = −1.45). Moreover, global analysis identified female-specific negative enrichment of gene sets involved in “histone ubiquitination” (NES = −2.31) gene sets and in translation such as “Ribosome Biogenesis” (NES = −2.15) and “Ribonucleoprotein complex biogenesis” (NES = −2.26) (Fig. 4b and Supplementary Table S3).
In addition, 32 gene sets were overrepresented in HH female blastocyst (2 hallmark gene sets and 30 GO terms). Of these gene sets, a large part was related to ion channel activity and to receptor signaling pathway as highlighted by the enrichment of gene set like "Ligand Gated ion Channel Activity" (NES = 2.30), “transmitter gated channel activity” (NES = 1.99), “transmembrane receptor protein tyrosine kinase activity” (NES = 1.73) and “Mitogen Activated Protein Kinase Binding” (NES = 1.82) (Fig. 4b and Supplementary Table S3).
Discussion
Objectives of this study were to decipher mechanisms involved in the sex-specific offspring programing previously observed in rabbit does fed a maternal high-fat diet.Reference Tarrade, Rousseau-Ralliard and Aubrière15 Transcripts content was weakly deregulated in the oocytes of females fed a HH diet, with the affected transcripts mainly involved in meiosis and translational control. The maternal HH diet had an impact on the composition of the uterine fluid in which increased pyruvate concentrations were observed. Lipid, glucose and amino acids transport and metabolism were also altered in the preimplantation embryos with a more pronounced effect in female than in male embryos.
The high-fat diet was given to rabbit does from the time of puberty, for a total period of 8 weeks. During this period in rabbits, hormonal dependent antral follicular maturation takes place, which is characterized by oocyte growth and maturation, in interaction with the surrounding cumulus cells.Reference Hutt, McLaughlin and Holland28,Reference Hennet and Combelles29 Oocyte growth and maturation have been shown to be particularly sensitive to changes in maternal nutritional and metabolic environment.Reference Cordier, Léveillé and Dupont30 Indeed, it has been previously shown in the same model that high-fat diet affects follicular growth with a decrease in the number of antral follicles and conversely an increase in the number of atretic follicles.Reference Cordier, Léveillé and Dupont30 During oocyte growth, transcriptional and translational activities and their post-regulations are critical for the oocyte accumulation and long-term storage of mRNA and proteins, which are essential to subsequently complete meiotic maturation and support early embryo development.Reference Hennet and Combelles29,Reference Brevini Gandolfi and Gandolfi31 Environmental conditions during maturation can influence the pattern of transcripts in matured oocytes.Reference Watson, De Sousa and Caveney32 In rodents,Reference Chang, Dale and Moley33–Reference Reynolds, Boudoures, Chi, Wang and Moley35 non-human primatesReference Chaffin, Latham, Mtango, Midic and VandeVoort36 and women,Reference Snider and Wood37,Reference Gonzalez, Robker and Rose38 maternal metabolic dysfunctions were shown to impair meiotic resumption and oocyte gene expression. Here, in agreement with these reports, a small number of genes involved in the regulation of meiosis were differentially expressed in HH oocytes. CCNA1 is known to control female meiotic cell cycle progression by blocking metaphase to anaphase transition.Reference Fuchimoto, Mizukoshi, Schultz, Sakai and Aoki39,Reference Radonova, Pauerova and Jansova40 ZCWPW1, a reader of histone (H3) modifications, is required to initiate the recombination of genetic information during meiosis.Reference Biot and de Massy41 In addition, transcriptomic variations observed in the present study also suggest that the HH diet affects the post-transcriptional regulation of oocyte gene expression, as several differentially expressed genes have been implicated in ribosomal processing like RARS,Reference Ibba and Soll42 RIOK1Reference Angermayr and Bandlow43 and MRPL55.Reference Cheong, Lingutla and Mager44,Reference Tselykh, Roos and Heino45 Thus, maternal HH diet altered oocyte transcripts content of genes involved in meiosis and translational control, suggesting difficulties to achieve meiosis and to support the embryo development. Global gene set enrichment analysis, however, did not identify enrichment of functional gene sets, suggesting a limited impact of HH diet on global oocyte transcripts content. The hypothesis of weak consequences of exposure to the HH diet during the pregestational window is reinforced by a recent study in which we showed that the biometric parameters as well as hepatic and placental gene expression were unaffected in term fetuses and placentas obtained after transfer of zygotes collected from HH females into control recipients, although fine differences in fatty acid profiles were observed compared to controls.Reference Rousseau-Ralliard, Aubrière and Daniel46
Consequences of HH maternal diet on the preimplantation embryo were first addressed through the study of its microenvironment, i.e., the uterine fluid. In mammals, uterine fluid composition is complex,Reference Calderari, Daniel and Mourier22,Reference Leese, Tay, Reischl and Downing47 varies according to the hormonal cycleReference Li and Winuthayanon48 and also in response to the presence of an embryo.Reference Yang, Wang and Chen49 Despite numerous studies on the impact of maternal diet on offspring, the consequences of altered maternal metabolism on uterine fluid composition remains poorly explored, partly because of sampling difficulties. Uterine fluid collected during surgery in women was shown to differ in branched-chain amino acid concentrations based on the patient’s healthy or unhealthy diet.Reference Kermack, Finn-Sell and Cheong50 In mice fed a low protein diet during the preimplantation period, the concentration of branched-chain amino acids was reduced compared to that of controls in uterine fluids collected postmortem.Reference Eckert, Porter and Watkins51 In ewes, a high protein diet was reported to induce increased ammonia and urea and decreased glucose concentrations in both oviductal and uterine fluids, also collected postmortem.Reference Tripathi, Farman, Nandi, Girish Kumar and Gupta52 Post-mortem hypoxia can lead to cell death, the degradation products of which can be found in the fluids.Reference Leese, Hugentobler and Gray53 Here, uterine fluid surrounding the embryo was collected in vivo under ultrasound bio-microscopic control, as previously developed in our laboratory.Reference Calderari, Daniel and Mourier22 Using 1H-NMR, we observed an impact of the HH diet on uterine fluid metabolomic profiles, more specifically on pyruvate concentration, that was increased in HH. Pyruvate is abundant in oviductal and uterine fluids in humans, mice and cows.Reference Aguilar and Reyley54 In mice embryo culture media, pyruvate is essential for the preimplantation embryo development as a nutrient and antioxidant.Reference Chi, Sharpley, Nagaraj, Roy and Banerjee55,Reference Gardner, Gardner, Harvey and Harvey56 Pyruvate regulates first steps of embryonic development such as the embryonic genome activation in miceReference Zhang, Yan and Sui57 and regulates levels of histone modification (H3K9) during embryonic genome activation in pigs.Reference Zhang, Zheng and Han58 In addition, while glucose has long been considered the preferred nutrient for blastocyst compaction and cavitation, it has recently been shown in mice that exogenous pyruvate contributes to the tricarboxylic acid cycle (TCA) and represents the main source of energy for the blastocyst.Reference Chi, Sharpley, Nagaraj, Roy and Banerjee55 Thus, the increased pyruvate concentration in uterine fluid of HH could have consequences on embryo development.
The impact of the maternal HH diet on the preimplantation embryo was assessed at the blastocyst stage through the analysis of metabolic profiles of blastocoelic fluids collected in vivo using ultrasound-guided puncture.Reference Calderari, Daniel and Mourier22 The blastocoel is a fluid-filled cavity, the blastocoelic fluid resulting from both uterine fluid influx and blastocyst cells' secretion.Reference Li and Winuthayanon48 Although playing a central role in embryonic development and being in direct contact with the inner cell mass and trophectoderm, little is known about the composition and function of the blastocoel. The impact of maternal metabolism on its composition is even less well understood. Here, a decrease in amino acids' and carbohydrates' concentrations was observed in HH blastocoelic fluids. Transcriptomic profiles of HH blastocysts demonstrated changes in the expression of genes involved in nutrient transport.
Genes involved in the transport of fatty acids were affected. The global functional gene sets analysis identified an under-representation of several pathways involved in lipid/fatty acid function. First, RBP4 was under-expressed. RBP4, is a fatty acid transporter,Reference Nono Nankam and Blüher59 regulated by the nutrient-sensitive kinase mTORC1,Reference Welles, Toro and Sunilkumar60 associated with insulin resistance, dyslipidemia, liver steatosis, type 2 diabetes and cardiovascular dysfunction.Reference Nono Nankam and Blüher59 In contrast, SCP2 was over-expressed in HH blastocysts. SCP2 is a lipid-binding protein that plays key roles a large variety of lipid trafficking and signaling.Reference Xu, Li and Tang61 SCP2 regulates lipids and fatty acids signaling pathways through lipid raft micro-domains of the plasma membrane in interaction with CAV1, which was also over-expressed in blastocysts from females exposed to high-fat diet of the present study.Reference Schroeder, Atshaves and McIntosh62 Moreover, SCP2 is described to enhance cholesterol transfer from intracellular membranes to mitochondria.Reference Schroeder, Atshaves and McIntosh62 OMA1, a zinc metalloprotease involved in the quality control system in the inner membrane of mitochondria, was also over-expressed.Reference Quirós, Ramsay and Sala63 OMA1 is required for mitochondrial metabolism in the blastocyst.Reference Zhou, Sun, Lee and Cui64 In preimplantation embryo, lipids and fatty acids metabolism serves as an energy source through fatty acid beta-oxidation and inhibition of beta-oxidation impaired blastocyst development.Reference Ye, Zeng, Cai, Qiao and Zeng65
In addition to fatty acid metabolism impairment, both transcriptomic and metabolomics analyses revealed an alteration in glucose metabolism in HH blastocysts. Reduced glucose concentrations were observed in HH blastocoelic fluids. Glucose is essential for the morula to blastocyst transition in mouse.Reference Chi, Sharpley, Nagaraj, Roy and Banerjee55 In mouse embryo, glucose is preferentially metabolized through the pentose phosphate pathway (PPP) to provide carbon for nucleotide formation.Reference Sharpley, Chi, Hoeve and Banerjee66 PGD, that converts 6-phosphogluconate into ribulose 5-P in the PPP,Reference Baardman, Verberk and Prange67 was under-expressed in HH blastocysts. The PPP is suggested to control the signals required for the trophectoderm differentiation occurring in the blastocyst development.Reference Chi, Sharpley, Nagaraj, Roy and Banerjee55 An impairment in glucose metabolism may impede the differentiation of the trophectoderm from the blastocyst stage. Impaired glucose metabolism may interfere with trophectoderm differentiation from the blastocyst stage and participate in defects in trophoblast function subsequently observed.Reference Tarrade, Rousseau-Ralliard and Aubrière15
Amino acid trafficking and metabolism were also impacted in HH blastocyst. Transcriptomic analysis identified negative enrichment of amino acid transport terms. The amino acid transporter SLC38A6 was under-expressed. SLC38A6 translocates small neutral amino acids, mostly glutamine and glutamate.Reference Gandasi, Arapi and Mickael68 Glutamate concentrations were not altered in HH blastocoelic fluids, but alanine concentrations were decreased. Alanine is one of the most abundant amino acids in blastocoelic and uterine fluids.Reference Calderari, Daniel and Mourier22 Alanine regulates pH and is a major player in the detoxification of ammonium generated by amino acid metabolism.Reference Orsi and Leese69,Reference Humpherson, Leese and Sturmey70 Alanine can be transported from uterine fluid across the trophectoderm,Reference Van Winkle, Tesch, Shah and Campione71 as well as be produced by the embryo, both by the inner cell mass and by the trophectoderm.Reference Gopichandran and Leese72 Pyruvate and glutamate can lead to the production of alanine and alpha-keto glutarate, respectively, through transamination. Thus, the observed decrease in alanine may reflect a decrease in alanine production from pyruvate to maintain sufficient pyruvate availability for the TCA cycle. A decrease in methionine concentrations was also detected in HH blastocoelic fluids. Methionine is an essential amino acid regular, central in methylation process, protein synthesis, lipid metabolism and oxidative stress regulation.Reference Martínez, Li and Liu73
Maternal HH diet-induced impairment in lipid, glucose and amino acids in blastocysts. The HH blastocysts were mainly affected by a decrease in nutrients and their metabolisms. Consistently with these decreases in glucose and amino acids metabolism, the central nutrient-sensing signaling pathway mTORC1 was identified as under-represented in HH blastocysts. Inhibition of mTORC is essential to maintain metabolic homeostasis under metabolic stress.Reference Liu and Sabatini74 Central to sense changes in nutrient supply in preimplantation embryo, mTORC1 signaling is promoted by maternal diabetes in rabbitsReference Gürke, Hirche and Thieme75 and repressed by low protein maternal diet in mice.Reference Fleming, Sun and Denisenko76 The key metabolic enzyme AK1 was also shown to be under-expressed in HH blastocysts in the present work. AK1 is involved in the synthesis, equilibration and regulation of adenine nucleotidesReference Dzeja and Terzic77 and in multiple energetic and metabolic signaling processes, partially via the AMP-activated protein kinase (AMPK) pathway.Reference Dzeja and Terzic77 AMPK is a sensor of glucose availability and energy status, acting in coordination with mTORC1.Reference Leprivier and Rotblat78 In parallel to the under-representation of the AMPK/mTORC1 pathway, an over-representation of the mitogen-activated protein kinase pathway was observed. DUSP2 was over-expressed. DUSP2 is a phosphatase known to dephosphorylate and control subcellular localization of MAPKs, such as extracellular signal regulated p38 proteins and Jun N-terminal kinases.Reference Chen, Chuang and Tan79 These kinases control a variety of cellular processes including proliferation, apoptosis, differentiation and signal transduction by activating transcription factors such as c-JUN.Reference Dunn, Wiltshire, MacLaren and Gillespie80 Transcription factors JUN and FOS, that, combined, form the transcription factor AP-1, were over-expressed.Reference Chinenov and Kerppola81 We also identified over-expression of the transcription factors EGR1 and EGR2, known to control ovarian function, embryo development and implantation.Reference Parfitt and Shen82–Reference Guo, Tian and Li84 These four transcription factors are defined as immediate early genes, known to rapidly and transiently be induced by diverse stimuli including nutrients, growth factor and stress to transduce signals to the downstream cascades involved in cell proliferation and apoptosis regulation.Reference O’Donovan, Tourtellotte, Millbrandt and Baraban85
Programming by an altered maternal metabolic environment differs according to offspring sex.Reference Aiken and Ozanne12 Evidence for sexual dimorphism to programing stimuli prior to gonadal differentiation and the appearance of sex-related hormonal differences is emerging.Reference Pérez-Cerezales, Ramos-Ibeas, Rizos, Lonergan, Bermejo-Alvarez and Gutiérrez-Adán86 Differences in sex chromosome dosage emerges with embryonic genome activation (EGA) and X chromosome inactivation, at a timing varying across species.Reference Okamoto, Patrat and Thépot23,Reference Bermejo-Alvarez, Rizos, Lonergan and Gutierrez-Adan87 In rabbits, major EGA occurs from the 8-cell stage and X chromosome inactivation initiates from blastocyst stage.Reference Okamoto, Patrat and Thépot23 The sex chromosome transcripts' expression regulates autosomal genes' expression leading to transcriptional sexual dimorphism before implantation.Reference Bermejo-Alvarez, Rizos, Lonergan and Gutierrez-Adan87 This transcriptional sexual dimorphism can lead to different susceptibilities to environmental stressors. Blastocyst sex-specific response to the maternal HH diet was explored here. Whereas gene-by-gene analysis did not identify differentially expressed genes, global analysis identified significant transcriptomic differences between male and female blastocysts. In female blastocysts, a large number of gene sets enrichment was induced by HH maternal diet, suggesting a coordinated response. As in the sex-independent analysis, functional pathways involved in fatty acid transport, glucose metabolism, nutrient sensing and transcriptional regulation were affected in female HH blastocysts. In addition, over-representation of gene sets involved in channel transporter and receptor signaling pathways was observed. Moreover, pathways involved in ribosomal RNA (rRNA) processing and more largely ribosome biogenesis were under-represented in female HH blastocysts, suggesting an impact on translational mechanisms. Synthesis of rRNA is regulated in response to metabolic and environmental changes.Reference Murayama, Ohmori and Fujimura88 In mice, maternal low protein diet increases rDNA transcription and RNA per cell content in offspring via the mTORC1 signaling pathway.Reference Fleming, Sun and Denisenko76 In contrast to females, the HH maternal diet had very little impact on the transcriptome of male blastocysts, which exhibited very few gene sets enrichments. Thus, at the blastocyst stage, females seem to be more sensitive than males to the maternal HH diet. Interestingly, we have previously shown in pregnant rabbit does fed the HH diet, that at the fetal stage, males were more metabolically affected than females by the maternal diet.Reference Tarrade, Rousseau-Ralliard and Aubrière15 It can be hypothesized that females, by adjusting their transcriptome in response to the maternal HH diet, were more successful in adapting to this altered environment than males.
In conclusion, the present work highlights the impact of a maternal high-fat diet on the embryo in its microenvironment. Metabolomics analyses revealed differences, notably an increased concentration of pyruvate, in the composition of uterine fluid surrounding the embryo from females on the HH diet. Thus, in the first stages of development, before the protective role of the placenta is established, the embryo is in direct contact with an altered environment. This result underlines the importance of exploring the impact of maternal metabolism alterations through in vivo exploration of uterine fluid composition in the DOHaD context. Further explorations, such as lipidomic studies, could improve our understanding of the consequences of the HH diet on the uterine fluid composition. Blastocysts that developed in this nutrient-rich environment were affected by a decrease in nutrients sensing and metabolism, partly through the mTORC pathway, that may represent protective mechanisms. The observation of a more altered transcriptome in female than in male embryos reinforces the hypothesis of the role of early sexual dimorphism in offspring programing.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/S2040174423000260.
Acknowledgments
We are grateful to the staff of the UE 1298 SAAJ (Livestock Physiology-Reproduction and Animal Diets Experimental Facility, https://doi.org/10.17180/MAQZ-V844), for housing the rabbits and for all the help given in the animal experimental procedures. We thank the MIMA2 platform (Microscopie et Imagerie des Microorganismes, Animaux et Aliments, https://doi.org/10.15454/1.5572348210007727E12) and particularly CIMA (Chirurgie et Imagerie Médicale chez l’Animal, https://www6.jouy.inrae.fr/mima2/Equipements/Imagerie-sur-l-Animal). We thank the @BRIDGe facility (GABI, AgroParisTech, INRAE, Université Paris-Saclay, Jouy-en-Josas, France, http://abridge.inrae.fr/). We thank Laura Hua, Luc Maillet and Clémentine Lapoujade for their help during their short-term internships.
Author contribution
Anne Couturier-Tarrade, Véronique Duranthon, and Pascale Chavatte-Palmer are contributed equally to this study.
Conceptualization: A.C.T., P.C.P., S.C., V.D.; Methodology; A.C.T., C.R., D.R.R, N.D., P.C.P., V.D. Validation: C.A., C.R., E.M., N.D., N.P., S.C.; Formal analysis: C.A., J.G., L.J., S.C.; Investigation: A.V.C., B.S.P., C.A., C.M., C.R., E.M., F.L., G.R., N.D., N.P., M.D., S.C.; Data curation: C.A., J.G., L.J., L.N.D.; Writing: A.C.T. P.C.P., S.C., V.D.; Visualization: C.R., P.C.P., S.C., V.D.; Supervision: A.C.T., P.C.P., S.C., V.D.; Project administration: A.C.T., P.C.P., S.C., V.D.; Funding acquisition: A.C.T., P.C.P, V.D.
Financial support
This work was supported by INRAE with dedicated help from the INRAE PHASE department (E.M., grant number CI_2014 and C.R., grant number CI_2015). B.S.P. was supported by a post-doctoral fellowship from the Agence Nationale de la Recherche (ANR, Plurabbit, PCS-09-GEM-08). A.V.C was supported by a post-doctoral fellowship from the European Commission (Fecund FP7-KBBE-2012-6).
Competing interests
None.
Ethical standards
The authors assert that all procedures contributing to this work comply with the ethical standards of the European regulations on animal welfare guides on the care and use of laboratory animals and has been approved by the institutional committee (Comethea, n°45 in French National register, experimental procedures n° 11/037 and n°2015042115447836_v2).
