Protective and sex‐specific effects of moderate dose folic acid supplementation on the placenta following assisted reproduction in mice

Abstract

Epigenetic defects induced by assisted reproductive technologies (ART) have been suggested as a potential mechanism contributing to suboptimal placentation. Here, we hypothesize that ART perturbs DNA methylation (DNAme) and gene expression during early placenta development, leading to abnormal placental phenotypes observed at term. Since folic acid (FA) plays a crucial role in epigenetic regulation, we propose that FA supplementation can rescue ART‐induced placental defects. Female mice were placed on a control diet (CD), a moderate 4‐fold (FAS4) or high dose 10‐fold (FAS10) FA‐supplemented diet prior to ART and compared to a natural mating group. ART resulted in 41 and 28 differentially expressed genes (DEGs) in E10.5 female and male placentas, respectively. Many DEGs were implicated in early placenta development and associated with DNAme changes; a number clustered at known imprinting control regions (ICR). In females, FAS4 partially corrected alterations in gene expression while FAS10 showed evidence of male‐biased adverse effects. DNAme and gene expression for five genes involved in early placentation (Phlda2, EphB2, Igf2, Peg3, L3mbtl1) were followed up in placentas from normal as well as delayed and abnormal embryos. Phlda2 and Igf2 expression levels were lowest after ART in placentas of female delayed embryos. Moreover, ART concomitantly reduced DNAme at the Kcnq1ot1 ICR which regulates Phlda2 expression; FAS4 partially improved DNAme in a sex‐specific manner. In conclusion, ART‐associated placental DNAme and transcriptome alterations observed at mid‐gestation are sex‐specific; they may help explain adverse placental phenotypes detected at term and are partially corrected by maternal moderate dose FA supplementation.

Abbreviations

ART

assisted reproductive technologies

CD

control diet

ddPCR

droplet digital PCR

DEGs

differentially expressed genes

DNAme

DNA methylation

FA

folic acid

FAS10

10‐fold folic acid supplemented diet or high dose

FAS4

4‐fold folic acid supplemented diet or moderate dose

IUGR

intrauterine growth restriction

NAT

naturally conceived

1. INTRODUCTION

With infertility affecting ~15% of couples, there has been a continuous increase in the use of assisted reproductive technologies (ART) in developed countries. ¹ Although most children born from ART are healthy, ² evidence suggests that ART pregnancies are at an increased risk of birth defects ³ , ⁴ and imprinting disorders. ⁵ , ⁶ , ⁷ In addition, ART is associated with adverse pregnancy outcomes such as preeclampsia and intrauterine growth restriction. ⁸ , ⁹ , ¹⁰ Data from human and animal model studies support the hypothesis that suboptimal placenta function may be responsible for several adverse pregnancy outcomes associated with ART. ¹¹

Animal models, particularly the mouse, that lack the confounder of parental infertility present in human studies, have been utilized to investigate the effects of ART on placental development. ¹² , ¹³ Similar to humans, the mouse placenta can be separated into three structures: (1) the outer maternal layer composed of decidual cells from the uterus and the maternal vasculature that brings blood to/from the implantation site; (2) a middle “junctional zone” (JZ) consisting of spongiotrophoblasts (SpTBs), glycogen cells (GCs), and trophoblast giant cells (TGCs) in contact with the maternal decidua; and (3) an inner layer known as the labyrinth, composed of syncytiotrophoblast I and II cells responsible for nutrient exchange. ¹² , ¹⁴ , ¹⁵ During pregnancy, the placenta ensures optimal fetal growth by mediating implantation, serving as the interface for nutrient and gas exchange, secreting hormones and protecting the growing fetus from the maternal immune system. ¹² , ¹⁵ , ¹⁶

In mice, the use of multiple ART procedures (e.g., superovulation, in vitro fertilization and embryo culture), similar to protocols used routinely for human ART, induces adverse placental phenotypes observed at the end of gestation, including placental vasculature defects, overgrowth, and increased levels of the pre‐eclampsia marker sFLT1. ¹⁷ , ¹⁸ Epigenetic alterations in the placenta have been suggested as a potential mechanism to explain abnormal placental phenotypes associated with ART. Indeed, ART procedures, particularly embryo culture, are performed during preimplantation development when DNAme erasure occurs across the genome except at imprinted loci and certain repeat elements. ¹⁹ , ²⁰ Imprinted genes in particular, which tightly regulate placenta development, must maintain their DNAme patterns during preimplantation development to ensure parent‐of‐origin specific expression. ²¹ Dysregulation of DNAme at imprinted loci due to ART may influence placentation, potentially modify nutrient transfer, and compromise offspring outcomes. ¹¹ , ¹⁸

The availability of methyl donors such as folate is critical to establish and maintain proper DNAme profiles during development through the one‐carbon metabolism pathway. ²² Adequate intake is achieved through consumption of folate‐containing foods or through supplementation using folic acid (FA). ²³ Prior to conception and during pregnancy, FA supplementation is recommended to reduce the risk of neural tube defects (NTDs). ²⁴ , ²⁵ Current guidelines recommend daily use of a multivitamin containing 0.4–1 mg FA for low to moderate risk pregnancies and up to 4 mg/day FA for women with high‐risk of NTDs. ²⁶ Mouse studies indicate that both folate deficiency and high dose folic acid supplementation can have deleterious effects on development. In mice, folate deficiency has been linked to craniofacial alterations ²⁷ whereas high dose maternal FA supplementation (10‐ and/or 20‐fold higher than recommended) was associated with disrupted embryonic development. ²⁸ , ²⁹ , ³⁰

To determine optimal FA doses for ART pregnancies, a previous study from our group investigated the effects of clinically relevant moderate 4‐fold (~1.6 mg/day in human) and high 10‐fold (~4.0 mg/day in human) FA‐supplemented diets on reproductive and epigenetic outcomes at mid‐gestation using a mouse model of ART. In this study, ART resulted in an increase in embryos with developmental delay along with genome‐wide DNAme abnormalities that were particularly marked in the placenta; both outcomes were partially corrected with moderate dose FA supplementation. ³¹ Whether the prominent placental epigenetic defects we found at mid‐gestation resulted in transcriptional changes that could impact placental function remained to be explored. In the current study, we hypothesized that ART alters DNAme programming in preimplantation embryos leading to altered gene expression during early placenta development in the post implantation period, subsequently resulting in the placental defects that have been reported at the end of gestation. ¹⁸ In addition, we postulated that FA supplementation can rescue these effects. We show that ART in mice has a notable impact on the placental transcriptome at midgestation and that FA supplementation acts to partially correct these changes in a dose‐dependent and sex‐specific manner. We reveal that many of the genes affected by ART are involved in early placenta development and differentially expressed in placentas from E10.5, delayed and abnormal embryos.

2. MATERIALS AND METHODS

2.1. Ethics

All animal experiments were performed in compliance with the guidelines established by the Canadian Council of Animal Care and approved by the Animal Care Committee at the Research Institute of the McGill University Health Centre (RI‐MUHC).

2.2. Mice and clinically relevant diets

This study is a follow‐up to the experiments described in Rahimi et al., 2019. ³¹ Briefly, mice were placed on a 12‐h light:12‐h dark cycle with access to food and water ad libitum. Hsd:NSA (CF1) outbred female mice (Envigo, Indianapolis, IN, USA) were fed one of three amino‐acid defined diets (Envigo) for 6 weeks prior to ART and throughout gestation: either a folic acid control diet (CD, 2 mg/kg diet, TD.130565) containing the recommended level of folic acid for rodents ³² ; a 4‐fold folic acid‐supplemented diet (FAS4, 8 mg/kg diet, TD.160058); or a 10‐fold folic acid‐supplemented diet (FAS10, 20 mg/kg diet, TD.160059). They were compared to a natural mating group fed the control diet. A six‐week diet period was chosen for two reasons: (1) It allows stabilization of long‐term intracellular folate storage, as reflected by red blood cell folate levels ³³ ; (2) It corresponds to three oocyte maturation cycles in mouse, equivalent to 12 months in humans, which is the usual time spent by couples trying to conceive before undergoing ART. B6SJLF1/J males (Jackson Laboratory, Bar Harbor, ME, USA) and CD1 vasectomized males (Charles River Laboratories, Senneville, QC, Canada) were fed a mouse chow diet (Teklad Global 18% Protein Rodent Diet, 4 mg folic acid/kg diet, Envigo).

2.3. ART procedures

Mouse ART procedures were adapted from the Jackson Laboratory method ³⁴ and published articles. ¹⁷ , ³⁵ Briefly, 9‐week‐old CF1 female mice were superovulated by IP injection of 5 IU pregnant mares serum gonadotropin (PMSG, 367222, EMD Millipore, Etobicoke, ON, Canada), followed by 5 IU human chorionic gonadotropin (hCG, 230734, EMD Milipore). ³⁶ Spermatozoa were isolated from the cauda epididymis and vas deferens of B6SJLF1/J males (12 to 14‐weeks‐old) and capacitated in a drop of EmbryoMax Human Tubal Fluid media (1X) (HTF, MR‐070‐D, EMD Millipore) covered by embryo tested mineral oil (M8410, EMD Millipore). IVF was initiated by dragging cumulus‐oocyte complexes using a 30G½ needle from both ampullae of superovulated CF1 females to the drop of HTF media containing capacitated spermatozoa. After 6 h, resulting zygotes were washed in HTF media then cultured in KSOM+½ AA (MR‐106‐D, EMD Millipore) under mineral oil at 37°C in a humidified, reduced oxygen environment (5% CO₂, 5% O₂, 90% N₂) for 4 days. Following embryo culture, ten blastocyst‐stage embryos (from a single CF1 superovulated donor) were transferred to a 2.5‐day post‐coitum (dpc) pseudopregnant CF1 recipient fed the same diet as the donor, using the Non‐Surgical Embryo Transfer (NSET) device (ParaTechs, Lexington, KY, USA). Day of blastocyst transfer was defined as E3.5 and embryos and placentas were collected 8 days later.

2.4. Natural mating protocol

Naturally cycling CF1 females were fed the CD for 6 weeks prior to conception and mated with B6SJLF1/J males. The morning of observance of a copulation plug was considered 0.5 dpc. Diets were continued though gestation until embryos and placentas were collected 10.5 dpc. These placentas served as controls forming the NAT_CD group.

2.5. Tissue collection, examination, and DNA extraction

Midgestation embryos, placentas and yolk sacs were collected as previously described. ³¹ Embryos were staged according to published criteria ³⁷ , ³⁸ and examined for developmental delay and malformations; they were considered developmentally delayed if they were staged at E9.5 or less. ³¹ Samples were snap frozen in dry ice and stored at −80°C until further use. Placental tissues were homogenized in a frozen tissue lysis buffer (Qiagen, Mississauga, ON, Canada) using mortars and pestles on dry ice. Placental genomic DNA was extracted using the DNeasy Blood & Tissue kit (Qiagen) for groups E10.5 and Abnormal, and the QIAamp DNA Micro kit (Qiagen) was used for smaller delayed placentas, according to manufacturer's protocol. Embryo sex was determined by PCR using yolk sac DNA as described previously, ³¹ and re‐confirmed in this study using the placental DNA.

2.6. RNA extraction and sequencing

Placentas chosen for transcriptome analysis were morphologically normal and associated with normal embryos staged at E10.5 (n = 4–6 per group/sex). Total RNA was extracted from placentas using the miRNeasy micro kit (Qiagen) according to the manufacturer's instructions. cDNA library construction, sequencing on Illumina NovaSeq PE150 (6G raw data per sample) and bio‐informatic analysis were performed by Novogene (Sacramento, CA, USA). Briefly, read mapping was performed with STAR_v2.6.1d, DESeq2_v1.26.0 was used for the differential analysis and genes with false rate discovery (FDR) adjusted p‐value <.05 and |log2(FoldChange)| > 0.3 were considered as differentially expressed. GO and KEGG enrichment analyses were performed using ClusterProfiler v3.8.1.

2.7. Gene expression analysis using droplet digital PCR (ddPCR)

Droplet digital PCR was chosen over qPCR for gene expression analysis because it is suitable for small samples (where 3–4 replicates/sample/gene are not required since each ddPCR result is the mean of ~10 000 reads), it can detect accurately low levels of expression and it is measured at reaction end point which eliminates the dependency on reaction efficiency (influenced by, for example, primer dimers and RNA contaminants). ³⁹ , ⁴⁰ The analysis was performed according to the manufacturer's protocol and published articles. ³⁹ , ⁴⁰ Briefly, 500 ng total RNA was converted to complementary DNA (cDNA) using the iScript Advanced cDNA Synthesis Kit (Bio‐Rad, Mississauga, ON, Canada) as per the manufacturer's instructions. Primers were designed in regions of the mRNA sequence transcribed for all protein coding transcripts using the NCBI primer Blast tool (mouse genome GRCm39): Forward and Reverse primers for‐ Phlda2: GCTCTGGGTCCGTGAAACG/GGGTTGGAAGCAGGTAACCA; EphB2: ACCATGACAGAAGCCGAGTA/CTGTTACATACGATGGCAATGAC; Igf2: ACACGCTTCAGTTTGTCTGTTC/AGTACGGCCTGAGAGGTAGAC; Peg3: GCACCAGCCGAGGTCTCAAA/GGTTGCGAGCCACATCCTTG; Rps18: GGGAAGTACAGCCAGGTTCTG/CAAAGGCCCAGAGACTCATTTC, Rpl13a: CTGCTGCTCTCAAGGTTGTTC/TGCCTGTTTCCGTAACCTCAAG.

A pool of 12 samples from all groups/sex was run on real‐time PCR (Bio‐Rad CFX96 Touch) to optimize annealing temperature, assess cDNA dilution, verify amplicon size and specificity for each primer set. Three representative samples per group/sex were used to determine optimal reference genes using predesigned M96‐well plate (Bio‐Rad): Rpl13a and Rps18 were identified as ideal reference genes by the GeNorm software (CFX Maestro Bio‐Rad) with M values <0.5 ⁴¹ (Datasets S1 and S2).

Serial dilutions of cDNA and primer sets were added to the QX200 ddPCR EvaGreen Supermix (Bio‐Rad) according to the associated protocol. The reaction mix was converted to droplets with the QX200 droplet generator (Bio‐Rad). Droplet‐partitioned samples were transferred to a 96‐well plate, sealed, and cycled in a C1000 Thermocycler (Bio‐Rad) according to the manufacturer's protocol. The cycled plate was then transferred and read using the QX200 reader (Bio‐Rad). Gene expression level measured as copies/μl was normalized to transcripts encoding Rps18 and Rpl13a using QX Manager v1.2 Standard.

2.8. DNA methylation analysis

For each midgestation placenta sample, 1 μg of DNA was treated with bisulfite using the EpiTect Bisulfite kit (Qiagen) and DNAme was quantified by bisulfite pyrosequencing at the Kcnq1ot1 ICR (For: AGGTTTTGGTAGGTGGTTT, Rev: Biot‐CCTAACTAAACCAAAATACACCATCATA, Seq: GTTAGGAGGAATAGTTGTTTTA, Ta:55°C) and the L3mbtl1 promoter region (For: GTTGTTTATGGGTGGGAAGATTGAG, Rev: Biot‐ACAAAAAAAACTACAAACCCTCATAC, Seq1: GTTAAGATATAATTTTTTTGGAA, Seq2: GTGGGTTTTAATAAAGTAGTG, Ta: 60°C) as previously described. ⁴²

2.9. Statistical analysis

Graphs and presented data were analyzed using the GraphPad Prism 9.4.1 software: means ± standard error of the mean (SEM) are shown. Statistical significance was set at p < .05 for all analyses. Two tailed unpaired student's t‐test was used to compare NAT_CD and ART_CD groups and determine the effect of ART. To assess the effect of FA supplementation following ART, means of ART groups (CD, FAS4 and FAS10) were compared using one‐way ANOVA with Tukey's correction for multiple comparisons. Percentage (%) correction of FA supplementation for individual DEGs shown in Figures 2C,D and [Link], [Link] was calculated using group means in Abbott's formula: $\left[1-\left(\frac{\mathit{NAT}\_\mathit{CD}-\mathit{ART}\_\mathit{FA}}{\mathit{NAT}\_\mathit{CD}-\mathit{ART}\_\mathit{CD}}\right)\right]*100$. DNAme variance was obtained by averaging variances at all CpGs within each group.

FIGURE 2.

Sex‐specific response of male and female DEGs to folic acid supplementation at midgestation and respective chromosomal distribution. (A) Proportion of female ART DEGs down‐ and upregulated also affected in male placentas. (B) Proportion of male ART DEGs down‐ and upregulated also affected in female placentas. (C) Correction of ART‐induced change in gene expression by the moderate (FAS4) and high dose (FAS10) of FA in E10.5 female and (D) male placentas (n = 4–6/group/sex). Black dots indicate correction of ART DEGs by FA and red dots indicate potentially adverse effects of FA. % Correction was calculated using group means in Abbott's formula: $\left[1-\left(\frac{\mathit{NAT}\_\mathit{CD}-\mathit{ART}\_\mathit{FA}}{\mathit{NAT}\_\mathit{CD}-\mathit{ART}\_\mathit{CD}}\right)\right]*100$. (E) Chromosomal distribution of DEGs caused by ART in male and female placentas. Dir., direction; F, female; M, male.

3. RESULTS

3.1. Effects of ART on the placental transcriptome at midgestation

To determine the effects of ART on the placental transcriptome, we performed RNA sequencing on whole placentas corresponding to normal E10.5 male and female embryos (n = 4‐6/group/sex). The ART_CD group was compared to the NAT_CD group and differential gene expression was determined when the FDR adjusted p‐value <.05. ART resulted in 41 and 28 differentially expressed genes (DEGs) in E10.5 female and male placentas, respectively (Figure 1A,B and Tables S1 and S2). In E10.5 female placentas, 17 genes were upregulated and 24 were downregulated (Figure 1A and Table S1) whereas in E10.5 male placentas, 21 genes were upregulated and 7 were downregulated (Figure 1B and Table S2).

FIGURE 1.

ART results in differential expression of genes in midgestation placentas of both sexes and folic acid partially rescues these effects. Heatmaps showing the relative expression of genes affected by ART, as revealed by RNA‐sequencing analysis (false discovery rate p < .05) in E10.5 normal (A) female and (B) male placentas and the effect of folic acid (FA). Each row represents an individual differently expressed gene (DEG), and each column represents one group (n = 4–6/group/sex). Blue color indicates lower levels of expression and yellow color indicates higher levels of expression. Circles indicate predominant imprinted regulation mechanism; confirmed non‐canonical genes are Slc38a4 and Smoc1 (C) Function distribution of placental male and female DEGs involved in early placenta development, angiogenesis, and spermatogenesis.

Gene Ontology (GO) analysis was performed to assess whether terms of biological significance were overrepresented among ART DEGs. Since ART resulted in very few DEGs, no significant GO enrichment could be detected. However, even though corresponding FDR values were not significant, the ten most enriched biological pathways (p < .05) among female DEGs were related to placenta development (i.e., regulation of vasculogenesis, spongiotrophoblast layer development, regulation of vasculature development, embryonic placenta development) (Dataset S3). In addition, the KEGG pathway analysis revealed that the JAK–STAT pathway, previously associated with recurrent implantation failures observed in ART pregnancies, ⁴³ , ⁴⁴ was identified as one FDR significant enrichment. These results provide evidence that ART‐induced DEGs are involved in several important functions including early placental development, angiogenesis, and spermatogenesis (Figure 1C). We also determined whether certain DEGs were common to males and females; we identified two upregulated DEGs (Susd2; 2410003L11Rik) and one downregulated DEG (Igf2os) shared between both sexes (Tables S1 and S2).

Interestingly, many DEGs identified in E10.5 female placentas were also affected in E10.5 male placentas, often exhibiting changes in the same direction. Out of 17 up‐ and 24 downregulated female ART genes, 16 and 22 were also up‐ and downregulated in male placentas, respectively (Figure 2A). Similarly, out of 21 up‐ and 7 downregulated male ART genes, 21 and 4 were undergoing the same change in expression in female placentas (Figure 2B). As shown in more detail at an individual‐placenta level, most DEGs identified in female placentas showed similar dysregulation in male placentas and vice‐versa: in both sexes, mean relative expression >1 for upregulated genes and <1 for downregulated genes (Figure S1).

3.2. ART‐associated DEGs respond to folic acid supplementation in a sex‐specific manner

Next, we assessed whether FA could rescue the ART‐induced gene expression alterations. Except for 2 DEGs which were not corrected by FAS4 (i.e., 2410003L11Rik and Nudt10) and 5 DEGs for FAS10 (i.e., Adm, Susd4, Ikzf4, Fam109b, Csf2rb); the two doses of FA supplementation exhibited partial correction of most female ART DEGs (Figures 1A, 2C and S2). FAS4 partially corrected 95% of the female ART DEGs, ranging from 2% to 103% (mean correction = 45%) while FAS10 corrected 90% of female ART DEGs to a lesser extent, from 1% to 74% (mean correction = 23%) (Figure 2C). The corrective effects of FA supplementation in female placentas were further demonstrated quantitatively as FAS4 and FAS10 significantly rescued the expression of female ART downregulated and upregulated genes, though FAS4 was more beneficial than FAS10 (Figure S3A,B).

In male placentas, FAS4 partially corrected 23 of the 28 ART DEGs (exception: Slc38a4, Galnt6, 2410003L11Rik, Pmaip1, Trpm2), exhibiting 0%–80% correction (mean correction = 18%) while FAS10 showed evidence of potentially adverse effects on the expression of 12/28 male ART DEGs, with a mean correction of −2% (Figures 1B, 2D and S4). Quantitatively, neither FA dose significantly rescued differential expression of male DEGs (Figure S3C,D). In male placentas, the corrective effect of FA on certain DEGs may be masked by the deleterious effects on others (Figure S3C,D).

Next, we set out to examine whether down‐ and upregulated DEGs responded to FA differently. We found that in female placentas, FAS4‐induced corrective effects were significantly different among DEGs, as FAS4 ameliorated upregulated genes (mean % correction = 66.8%) more than downregulated ones (mean % correction = 28.7%) (Figure S3E). Since increased gene expression often correlates with DNA hypomethylation of promoters, our observation that FA supplementation was more beneficial for upregulated genes is consistent with a possible correction of such hypomethylation.

Finally, we assessed whether FA supplementation exhibited sex‐specific effects in ART placentas. We found that for ART downregulated and upregulated genes, both FAS4 and FAS10 displayed better corrective effects of differential expression in females relative to males (Figure S3F,G). To explore this observation further, we looked in more detail at the effects of FA supplementation on female ART DEGs in male placentas and vice‐versa (Figures S2 and S4). Unexpectedly, we observed that FAS4 and FAS10 still exhibited better correction of most male ART DEGs in female placentas compared to male placentas (i.e., Cysltr2, Fgl2…) (Figure S4A).

3.3. Many ART‐associated DEGs are involved in genomic imprinting and implicated in early placenta development

Interestingly, many ART‐associated DEGs comprise canonical and non‐canonical imprints regulated by diverse epigenetic mechanisms including DNAme, long‐noncoding RNAs (lncRNA) and histone modifications (H3K9me/H3K27me) ⁴⁵ , ⁴⁶ , ⁴⁷ , ⁴⁸ (Figure 1A,B). Chromosomal location of male and female ART DEGs revealed that while DEGs were distributed across the genome, some DEGs clustered on distal chromosome 7 and 15 (Figure 2E). The mouse imprinted cluster on chromosome 7 was affected in both sexes and included syntenic genes located in human chromosome region 11p15.5, also known as ICR1 and ICR2 and associated with ART‐related syndromes such as the Beckwith‐Wiedemann syndrome. ⁴⁹ DEGs located in this cluster are well known to be primarily reliant on DNAme for parent‐of‐origin specific gene expression. ⁴⁵ As shown by arrows beside the gene names (Figure 1A,B), several imprinted genes with altered expression after ART in female and male placentas had been reported to have altered DNAme by Reduced Representation Bisulfite Sequencing (RRBS) in our previous study. ³¹

To gain a better understanding of how genes with altered expression at mid‐gestation following ART might impact placental function, we examined the timing of expression of affected genes in relation to placental cell lineage development (Figure 3). The early development of the three main mouse placenta tissue types is shown: the maternal decidua composed of the maternal vasculature; the JZ, first site of trophoblast (TB) invasion which serves an endocrine function; and the labyrinth responsible for nutrient and gas exchange. ¹⁵ Ephrin receptors including EphB2 have been suggested to play key roles in the remodeling and maturation of VEGF‐induced immature vessels in the placenta. ⁵⁰ Both Adm ⁵¹ , ⁵² ; a growth promoting angiogenic factor, and Ceacam1 ⁵³ , ⁵⁴ ; a cell‐adhesion molecule, promote trophoblast invasion and remodeling of uterine spiral arteries of the maternal decidua during placentation. Contrary to Adm and Ceacam1, Fos inhibits TB migration and invasion of the maternal decidua. ⁵⁵ , ⁵⁶ In addition, L3mbtl1 functions as a transcriptional repressor and chromatin compactor. ⁵⁷ It has been suggested that L3mbtl1 gene expression may regulate cell fate by preventing trophectoderm (TE) formation. ⁵⁷ After implantation, Nodal signaling from the inner cell mass promotes differentiation of polar TE cells into TB stem cells essential for JZ establishment. ⁵⁸ Interestingly, Phlda2 is a placenta‐specific imprinted gene responsible for regulating JZ growth: loss of Phlda2 in mice is associated with placenta overgrowth characterized by an expansion of the JZ ²¹ whereas over‐expression decreases JZ size and placental glycogen stores. ⁵⁹ , ⁶⁰ Similarly, Csf2rb is a subunit of Csf2 receptor, involved in the regulation of placental JZ structure: loss of Csf2 expands the JZ with increased levels of SpTBs and GCs. ⁶¹ , ⁶² In addition, imprinted genes Peg3 and Igf2 play an important role in optimal JZ establishment ⁶⁰ : loss of Peg3 results in reduced placental glycogen stores and decreased SpTBs and GCs ⁶³ while loss of Igf2 is associated with decreased SpTBs and GCs volume in female but not male placentas. ⁶⁴ Additionally, Igf2 is implicated in fetoplacental endothelial cell proliferation within the labyrinth. ⁶⁵ Given that circulating levels of Igf2 reflect fetal nutrient demands, the Igf2‐Igf2r axis is responsible for regulating placental microvasculature expansion within the labyrinth accordingly. ⁶⁵ , ⁶⁶ , ⁶⁷ Amino acid transporter Slc38a4 is also highly expressed by syncytiotrophoblasts in the labyrinth to support transfer of amino acids toward the fetal circulation. ⁴⁷ Additionally, Asb4 ubiquitin ligase promotes trophoblast differentiation in the labyrinth at midgestation. ⁶⁸ Finally, Klf14 is an imprinted transcription factor which acts to limit placenta growth in late gestation. ⁶⁹ Altogether, many ART‐induced DEGs play key roles in placenta cell lineage proliferation, migration, maturation, and differentiation. These early transcription changes may result in adverse placental phenotypes observed at the end of ART gestation in the mouse. ¹⁶ , ¹⁸

FIGURE 3.

ART‐associated DEGs are implicated in early placenta development. Upregulated genes are colored in green and downregulated ones in purple. E4.5, Embryonic day 4.5; TBs, trophoblasts; TGCs, trophoblast giant cells. Adapted from Rai and Cross ⁵⁸ and Tunster et al. ⁶³

3.4. Gene expression analysis of key genes affected by ART and involved in early placenta development

Mouse studies previously showed that ART is associated with late gestational placental overgrowth, mainly in the JZ. ¹⁶ , ¹⁸ To investigate potential early mechanisms at the transcriptome level, we selected two female DEGs (Phlda2 and EphB2) and two male DEGs (Igf2 and Peg3) that could lead to such placental phenotypes and performed gene expression analysis on an extended cohort. Previously, genome‐wide analysis (RRBS) ³¹ revealed that ART induced thousands of differentially methylated tiles (DMTs) in male and female E10.5 placentas, the majority of which were hypomethylated (Figure S5). Using these data, ³¹ we identified ART‐induced DMTs in EphB2, Igf2 and Peg3 DEGs (Figure 1A,B).

For analysis of gene expression, total RNA was isolated from individual E10.5 male and female placentas (n = 9–13 placentas/group/sex) and transcript levels were measured via droplet digital PCR (ddPCR) (Figure 4). Consistent with RNA‐sequencing data, Phlda2 and Igf2 were significantly downregulated in female and male E10.5 placentas following use of ART (p < .05; Figure 4A,C). EphB2 relative expression was significantly decreased in female, but not in male placentas while Peg3 displayed significantly greater abundance in ART_CD groups of both sexes compared to controls (p < .05; Figure 4B,D).

FIGURE 4.

Relative expression of key genes affected by ART and involved in early placenta development, in female and male midgestation placentas and impact of folic acid supplementation. Gene expression changes in E10.5 placentas for (A) Phlda2, (B) EphB2, (C) Igf2 and (D) Peg3. Sample sizes for each group are indicated in parenthesis on the x‐axis. Each circle and triangle correspond to an individual placenta in the NAT_CD and ART groups, respectively, black symbol representing a male placenta and red symbol, a female placenta. Gene expression was normalized to transcripts encoding Rps18 and Rpl13a. Data are presented as relative expression compared to controls with mean of NAT_CD E10.5 control group adjusted to 1. Means ± SEM are shown. Student's t‐test compared NAT_CD and ART_CD groups; *p < .05, ***p < .001, ****p < .0001. One‐way ANOVA with Tukey's correction for multiple comparisons was used to compare ART groups.

Next, we assessed whether FA supplementation could modify these ART‐induced expression alterations. While not significant, FAS4 and FAS10 slightly corrected Phlda2 relative expression levels in female, but not male placentas (Figure 4A). However, neither level of FA supplementation significantly affected mean relative expression levels of EphB2, Igf2, or Peg3 in either sex following ART. Together, these results from an extended cohort are consistent with the RNA‐sequencing data obtained from a subset.

3.5. Gene expression analysis of key genes affected by ART and involved in early placentation, in placentas from developmentally delayed/abnormal embryos

Previously, ART were associated with a significant increase in embryonic developmental delay; an adverse effect mitigated by moderate but not high dose FA supplementation (Figure S6). ³¹ Although not significant, ART were associated with female‐biased developmental delay, as the proportion of delayed female embryos (staged at E9.5 or earlier) was higher than males in the ART group (Figure S6A, Table S3). ³¹ In addition, embryos with abnormalities were observed in both sexes of the different groups, even though overall rates remained unaffected by ART or FA (Table S4, adapted from Rahimi et al., 2019 to detail sex differences). ³¹

For this reason, we first set out to examine the effects of ART on gene expression in placentas from ART‐conceived delayed/abnormal embryos (mutually exclusive), using the same male and female ART DEGs previously described (Figure 5). Note that placentas associated with abnormal embryos were morphologically normal looking, but this study will refer to them as “abnormal placentas” to shorten the text. ART‐downregulated Phlda2 and Igf2 also exhibited a significant decrease in female delayed and female abnormal placentas (Figure 5A,B,I,J). In contrast, EphB2 and Peg3 relative expression levels were significantly decreased and increased, respectively, as they were in normal placentas, in female abnormal placentas only (Figures 4B,D and 5F,N). Gene expression differences for male delayed and male abnormal placentas in the CD group are shown but were not tested for significance due to small sample size (Figure 5C,D,G,H,K,L,O).

FIGURE 5.

Relative expression of key genes affected by ART and involved in early placenta development, in placentas from delayed and abnormal embryos and impact of folic acid supplementation. (A) Phlda2 gene expression in female delayed, (B) female abnormal, (C) male delayed, (D) male abnormal placentas. (E) EphB2 gene expression in female delayed, (F) female abnormal, (G) male delayed, (H) male abnormal placentas. (I) Igf2 gene expression in female delayed, (J) female abnormal, (K) male delayed, (L) male abnormal placentas. (M) Peg3 gene expression in female delayed, (N) female abnormal, (O) male delayed, (P) male abnormal placentas. Sample sizes for each group are indicated in parenthesis on the x‐axis. Each white circle and black triangle correspond to an individual placenta in the NAT_CD and ART groups, respectively. Gene expression was normalized to transcripts encoding Rps18 and Rpl13a. Data are presented as relative expression compared to controls. The red line indicates mean of NAT_CD E10.5 control group which is adjusted to 1 as presented in Figure 4. Means ± SEM are shown. Student's t‐test compared NAT_CD and ART_CD groups; *p < .05, ***p < .001, ****p < .0001. One‐way ANOVA with Tukey's correction for multiple comparisons was used to compare ART groups; ^# p < .05, ^## p < .01, ^### p < .001.

Next, we assessed whether FA supplementation could rescue these ART‐induced changes. Despite lack of female delayed placenta samples in the ART_FAS4 group, we could observe that FAS10 and FAS4 corrected Phlda2 expression in female delayed and female abnormal placentas, respectively (Figure 5A,B). Similarly, FAS10 significantly corrected EphB2 expression in female abnormal placentas (Figure 5F).

Thus, using ddPCR, key ART DEGs involved in placenta development were similarly affected in normal, delayed, and abnormal placentas. For some genes in female placentas where sample sizes were adequate, folic acid supplementation showed evidence of significant correction of gene expression abnormalities in delayed and abnormal placentas that was not found in the normal placentas.

3.6. Sex‐specific effects of FA supplementation on Kcnq1ot1 ICR DNAme and positive correlation between Kcnq1ot1 DNAme and Phlda2 gene expression

ART has been suggested to affect gene expression through epigenetic dysregulation, particularly because ART procedures are performed during key periods of DNAme reprogramming. ¹⁹ To examine potential correlations between DNAme and expression in placentas from the normal, delayed, and abnormal placentas, we examined DNAme at the Kcnq1ot1 imprinting control region (ICR) which regulates Phlda2 expression, using bisulfite pyrosequencing (Table S5). ART was associated with female‐biased DNAme defects as female E10.5 and abnormal placentas exhibited more decreased mean methylation relative to males as evidenced by more significant p values between the NAT_CD and ART_CD groups in females relative to males (for females, p < .001 in Figure 6A,C; for males p < .05 in Figure 6G; p is not significant in Figure 6I). Interestingly, decreased DNAme levels were partially rescued by FAS4 but not FAS10 in female E10.5 placentas (Figure 6A). Observed sex differences were further evidenced by significantly increased methylation variance at the Kcnq1ot1 ICR in female placentas relative to males (Figure 6D–F,J–L). While FAS4 partially corrected methylation variance in female E10.5/abnormal placentas, FAS10 achieved a similar result in female delayed placentas. We could not test for significant changes of DNAme in delayed placentas in the FAS4 group due to the small sample size.

FIGURE 6.

Sex‐specific effects of folic acid supplementation on Kcnq1ot1 ICR DNA methylation and correlation between Kcnq1ot1 ICR DNA methylation and Phlda2 gene expression. (A) DNA methylation at the Kcnq1ot1 ICR in female E10.5, (B) female delayed, (C) female abnormal, (G) male E10.5, (H) male delayed and (I) male abnormal placentas. DNA methylation variance at the Kcnq1ot1 ICR in (D) female E10.5, (E) female delayed, (F) female abnormal, (J) male E10.5, (K) male delayed and (L) male abnormal placentas. Correlation analysis between Kcnq1ot1 ICR DNA methylation and Phlda2 gene expression in (M) female E10.5, (N) female delayed and abnormal, (O) male E10.5 and (P) male delayed and abnormal placentas. For DNA methylation and variance analysis, sample sizes for each group are indicated in parenthesis on the x‐axis. Each white circle and black triangle correspond to an individual placenta in the NAT_CD and ART groups, respectively. Means ± SEM are shown. Student's t‐test compared NAT_CD and ART_CD groups; *p < .05, ***p < .001, ****p < .0001. One‐way ANOVA with Tukey's correction for multiple comparisons was used to compare ART groups; ^# p < .05, ^## p < .01, ^#### p < .0001. For correlation analysis, each white circle represents an individual placenta in the NAT_CD group. White, gray, and black triangles indicate individual placentas in the ART_CD, ART_FAS4 and ART_FAS10 groups, respectively. Pearson correlation coefficient r is shown for each analysis; *p < .05, ****p < .0001. Strong correlation: r > 0.7; Moderate correlation: 0.5 < r < 0.7; Weak correlation: r < 0.3.

DNAme at the Kcnq1ot1 ICR results in expression of Phlda2 from the maternal allele via lncRNA. ⁷⁰ To examine this relationship in our samples, we performed a correlation analysis between Kcnq1ot1 ICR DNAme and Phlda2 gene expression, where delayed and abnormal placentas were combined. We confirmed a positive correlation for female E10.5 (r = 0.7; p < .0001), female delayed and abnormal (r = 0.61; p < .0001), male E10.5 (r = 0.64; p < .0001), and male delayed and abnormal placentas (r = 0.43; p = .02) (Figure 6M–P). Together, the results suggest that ART is associated with sex‐specific placental DNA hypomethylation at the Kcnq1ot1 ICR, which results in Phlda2 downregulation. Moderate but not high dose FA partially rescues some of these changes in a sex‐specific manner.

3.7. ART‐ associated hypomethylation of the L3mbtl1 promoter, correlation with transcription changes at midgestation and lack of response to folic acid supplementation

Although imprinted in humans, the ART‐upregulated transcription repressor L3mbtl1 (Figure 1A) is imprinted‐like in mice, i.e., it requires uninterrupted DNA methyltransferase activity at each early embryonic division to be accurately inherited, as we previously described. ⁷¹ Given that ART was associated with L3mbtl1 promoter hypomethylation in our previous RRBS study ³⁰ (Figure 1A) and considering the important role of this gene for extra‐embryonic lineage differentiation (E4.5), we examined the impact of ART on L3mbtl1 promoter 2 DNAme using bisulfite pyrosequencing (Table S5) and gene expression using ddPCR, in female placentas. ART resulted in a marked decrease of promoter 2 DNAme in female E10.5/abnormal placentas, as demonstrated by significant differences between the NAT_CD and ART_CD groups (Figure 7D,F). In addition, delayed female placentas displayed a trend toward a significant decrease of L3mbtl1 promoter DNAme (p = .06). In contrast, only female abnormal placentas showed a significant increase of L3mbtl1 relative expression. Neither level of FA supplementation affected mean DNAme or gene expression levels.

FIGURE 7.

Gene expression and DNA methylation profiles of L3mbtl1 transcription factor following ART with or without folic acid supplementation. (A) L3mbtl1 gene expression in female E10.5, (B) delayed and (C) abnormal placentas. (D) L3mbtl1 promoter DNA methylation in female E10.5, (E) delayed and (F) abnormal placentas. Sample sizes for each group are indicated in parenthesis on the x‐axis. Each white circle and black triangle correspond to an individual placenta in the NAT_CD and ART groups, respectively. Gene expression was normalized to transcripts encoding Rps18 and Rpl13a. Gene expression data are presented as relative expression compared to controls with mean of NAT_CD E10.5 control group equal to 1. Means ± SEM are shown. Student's t‐test compared NAT_CD and ART_CD groups; **p < .01, ****p < .0001. One‐way ANOVA with Tukey's correction for multiple comparisons was used to compare ART groups.

4. DISCUSSION

Evidence suggests that ART induces epigenetic defects which result in placental abnormalities; however, the underlying mechanisms are not well understood. ¹¹ Here, we examine mid‐gestation mouse placentas and show that ART results in transcriptional dysregulation of a number of genes involved in early placenta development, which could in turn potentially contribute to adverse placental phenotypes that have been reported to occur at the end of gestation. ¹⁸ We find evidence of sex‐specific effects of ART on gene expression profiles in placentas from normal, delayed, and abnormal embryos. We demonstrate that moderate FA supplementation partially rescues some of these effects more efficiently in female, while high dose FA exhibits evidence of male‐biased adverse effects.

Previously, we and others have investigated the impact of ART on fetal and placental epigenetic profiles. ¹¹ , ¹³ , ¹⁷ , ¹⁸ , ⁷² In our previous study, we found thousands of DMTs associated with ART in mid‐gestation placentas, ³¹ while here, relatively small numbers of genes were significantly differentially expressed overall. Potential explanations for the discrepancies between numbers of DMTs and genes affected by ART, include that the ART‐induced DNAme alterations detected were mostly of small magnitude (<10%) and occurred sporadically as well as in regions such as repeats that would not influence gene expression. Given that DEGs identified in the current study exhibit same‐direction changes in both sexes, we propose that they are critical for fetal and/or placental development as summarized in Figure 3. Some DEGs were also identified as placental enriched genes affected by embryo culture including imprinted genes Peg3 and Slc38a4. ⁷³ Since fetal growth is dependent on a functioning placenta, ART‐associated improper expression of genes which tightly regulate placenta development and/or function may plausibly result in adverse fetoplacental outcomes.

In mouse and human placentas, trophoblast invasion and remodeling of maternal spiral arteries are critical for placentation and embryo viability. ¹² , ⁷⁴ , ⁷⁵ Impaired or shallow invasion/remodeling has been implicated as a contributing factor to preeclampsia. ⁷⁶ Among DEGs identified, Fos (upregulated by ART) inhibits the invasion/spiral artery remodeling process whereas Adm/Ceacam1 (also upregulated by ART) promotes it. In agreement with these findings, after ART using the same protocol as the current study, preeclampsia marker sFLT1 expression was increased in E18.5 placentas, ¹⁸ as would be predicted by Fos upregulation. In contrast, concomitant upregulation of Adm/Ceacam1 may indicate an adaptive mechanism of the placenta in response to ART.

The mouse placental labyrinth, like the chorionic villi in humans, is the exchange interface for nutrient and gas between maternal and fetal circulations. ¹⁵ Imprinted genes, particularly Igf2, have a central role in the regulation of maternal‐fetal resource allocation. ⁷⁷ Igf2 expression in fetoplacental endothelial cells of the labyrinth promotes angiogenesis and endothelial cell proliferation. ⁶⁵ Within the labyrinth, Igf2 expression is driven by placenta‐specific promoter P0 and deletion of the P0 transcript results in proportionate reduction of both the labyrinth and junctional zones. ⁷⁸ Therefore, it can be inferred that downregulation of Igf2 following ART may contribute to the labyrinth abnormalities detected by Vrooman et al. at mid‐gestation, potentially interfering with gas and nutrient exchange. ¹⁸ In accordance with this hypothesis, Igf2 P0^+/− mutants display increased expression of System A amino acid transporter genes, including imprinted Slc38a4. ⁷⁷ Likewise, in the current study, ART produces a similar effect by inducing upregulation of the Slc38a4 amino acid transporter in mid‐gestation placentas. Notably, loss of imprinting of Slc38a4 also contributes to the placental enlargement phenotype. ⁴⁸ , ⁷⁹ Thus, canonical and non‐canonical imprinting dysregulation following ART may both contribute to placenta defects at E18.5.

The cluster of DEGs on distal mouse chromosome 7 is one of the most extensively characterized domains, controlled by the lncRNA Kcnq1ot1 transcription which recruits repressive machinery to silence the paternally unmethylated allele and express neighboring genes only from the maternal allele. ⁴⁵ , ⁸⁰ In the bovine ART model, loss‐of‐imprinting at certain imprinted loci within this domain has been associated with an overgrowth syndrome akin to Beckwith‐Wiedemann syndrome (BWS) in humans. ⁸¹ Our results indicate that within this domain, ART significantly downregulated placenta specific‐gene Phlda2 which acts to limit the placental JZ size. In previous studies, JZ overgrowth was observed at the end of gestation and associated with embryo culture and IVF. ¹⁷ , ¹⁸ Tunster and colleagues found that Phlda2 deficiency drives the expansion of the JZ, increases placental glycogen and induces fetal growth restriction. ⁶⁰ In contrast, Phlda2 overexpression drives a significant reduction of the JZ with limited glycogen stores and fetal growth restriction (FGR) at the end of gestation. ⁵⁹ We postulate that ART‐induced downregulation of Phlda2 during early placenta development contributes to JZ expansion observed at the end of gestation and causes alterations in JZ cell composition which may affect energy storage and endocrine function. This hypothesis is further supported by Igf2 downregulation in the placenta because a key study found that loss of Igf2 in the JZ alters cell composition and endocrine capacity. ⁶⁴ Gene expression changes detected at mid‐gestation may originate from the time of embryo culture as it was reported that even a 24 h culture from the 4‐cell to morulae stage leads to placenta overgrowth with the expansion of the JZ and increased parietal TGCs in mouse placenta. ⁷³ Moreover, a recent study also found that IVF and embryo culture lead to epigenetic changes already detectable in the inner cell mass of preimplantation embryos, with affected genes involved in developmental/regulatory processes and cardiac hypertrophy. ⁸²

In agreement with previous studies, our results support the observation that ART induces DNAme defects responsible for imprinted gene dysregulation. ¹¹ Beyond imprinted genes, many non‐imprinted and imprinted‐like DEGs, ⁷¹ which require continuous DNMT1 activity to maintain adequate methylation profiles, were associated with DNAme abnormalities in their respective sequences. While developmental delay, equivalent to FGR or IUGR in human, has been observed in the context of ART ³¹ , ⁸³ , ⁸⁴ , ⁸⁵ , to our knowledge, this is the first study to evaluate epigenetic and expression changes associated with ART in placentas from delayed and abnormal mouse embryos. We decided to include placentas from abnormal embryos in our study because placental defects often strongly correlate with abnormal embryonic development. ⁸⁶ Generally, placentas from E10.5, delayed and abnormal embryos were similarly affected by ART for the subset of DEGs selected, although numbers of affected embryos limited conclusions that could be made in males. Igf2 expression was considerably decreased in placentas from female delayed and abnormal embryos. Consistent with these findings, Igf2 expression was lower in trophectoderm cells after IVF compared with controls ⁸⁷ and deletion of placenta specific Igf2 (Igf2 P0^+/−) led to reduced placenta growth and FGR, with the former preceding the latter. ⁷⁸ In addition, IGF2 mRNA was also significantly decreased in human IUGR placentas. ⁸⁸ Therefore, adding to possible labyrinth abnormalities, disruption of Igf2 expression following ART may be a potential indicator of developmental delay. As observed in our previous study, moderate but not high dose FA corrected developmental delay, ³¹ which may explain the low numbers of female and male delayed embryos in the ART_FAS4 groups.

Altered Phlda2 expression at midgestation was accompanied by sex‐specific DNAme changes at the Kcnq1ot1 ICR with greater hypomethylation observed in female relative to male placentas. Recently, Mani et al. also identified female‐biased genome‐wide hypomethylation in mouse and human placentas associated with ART. ¹³ In support of our findings, PHLDA2 mRNA and protein expression were also significantly downregulated in term human ART placentas compared with placentas from natural pregnancies and associated with upregulated KCNQ1OT1 mRNA; DNA methylation levels at the KCNQ1OT1 ICR showed a decreasing trend in ART placentas. ⁸⁹ Furthermore, placentas from mouse embryos cultured from the 1‐cell to blastocyst stage exhibited hypomethylation at Kcnq1ot1, Peg3 and H19/Igf2 ICRs at E18.5. ⁷³ These results suggest that in humans and mice, ART may interfere with the DNAme maintenance machinery likely during the time of global DNA methylation erasure in preimplantation embryos. To explain the sex bias, we hypothesize that since loss of DNAme is already detectable at the blastocyst stage ⁹⁰ , ⁹¹ , ⁹² , ⁹³ and DNAme is required for X chromosome inactivation (XCI), ⁹⁴ , ⁹⁵ supply of methyl donors during embryo culture, when XCI takes place, has a sex‐specific role in females. During embryo culture, methyl donor availability for imprint maintenance may be limiting because of its use for XCI in female embryos. As a result, females may be more predisposed to DNAme alterations detectable during gestation. ⁴² Notably, placental sex‐specific ART‐induced DNAme alterations may result in long‐term sex‐specific health complications including metabolic outcomes, as reported by Narapareddy et al. ⁹⁶

During placentation, transcription factors act to control key cell fate decisions in the trophoblast (TB) cell lineages. ⁹⁷ The L3mbtl1 transcription factor inhibits differentiation of cells toward the trophectoderm of the blastocyst ⁵⁷ which is the precursor to the TB cell lineage of the placenta. ⁹⁷ Indeed, it is reported that overexpression of L3mbtl1 results in improper segregation and cytokinesis; thus, it may be important for adequate progression of cell division. ⁹⁸ , ⁹⁹ While DNAme changes associated with ART were observed at the L3mtbl1 promoter in mid‐gestation placentas, we did not observe significant L3mbtl1 upregulation except for female abnormal placentas. However, we noticed an increased variance in the DNAme profiles of the NAT_CD control groups. Though L3mbtl1 was detected in the mouse TE at E3.5, ¹⁰⁰ little is known about expression of this transcription factor in the different placental cell lineages at mid‐gestation and thus, it is difficult to interpret expression levels at this time. Given that L3mtbl1 acts before mid‐gestation, single‐cell RNA sequencing experiments in TB cell lineages before mature placenta formation may be more informative to examine ART effects related to this gene.

Overall, FA supplementation displayed partially corrective effects for ART‐associated DNAme and gene expression alterations in a dose dependent and sex‐specific manner. While neither dose significantly corrected differential expression at the individual gene level, except for certain genes in delayed and/or abnormal placentas, a corrective effect of moderate FA supplementation was more apparent on a larger scale from the RNA‐Seq results. As in our previous study on DNAme, ³¹ female placentas appeared to benefit more from FA than males, and males showed evidence of potentially adverse effects of FAS10. Given that female embryos have an increased need for methyl donors during XCI, FAS4 and FAS10 may have more beneficial effects in females relative to males. Thus, while FAS4 improves gene expression alterations associated with ART in males, FAS10 may be higher than the required threshold for males and exacerbate rather than correct ART‐induced changes.

Based on our results, we propose that supplementation of moderate dose FA prior to conception with ART and during early gestation benefits placenta development by enabling early adaptive placental responses and optimizing fetal growth. The fact that the correction of ART‐associated DNAme and gene expression observed in our two studies ³¹ is partial may be due to the experimental design necessitated by the embryo culture period we used to emulate current human ART practice. In our experimental paradigm, the folic acid supplements were given prior to conception, at a time that would optimize oocyte growth and counteract effects of superovulation, ¹² , ³⁶ as well as after implantation including the time of DNAme re‐establishment. For future studies, in addition to the FA regimen used here, adding methyl donors to the embryo culture medium should be considered, to cover the preimplantation time when methylation of key loci, including imprinted genes and those involved in XCI, need to be maintained.

AUTHOR CONTRIBUTIONS

Rita Gloria Ihirwe, Josée Martel and Jacquetta Trasler designed the study. Rita Gloria Ihirwe, Josée Martel and Sophia Rahimi performed the research experiments. Rita Gloria Ihirwe and Josée Martel analyzed the data. Rita Gloria Ihirwe, Josée Martel and Jacquetta Trasler interpreted the data and wrote the manuscript. Rita Gloria Ihirwe, Jacquetta Trasler and Josée Martel revised the manuscript.

DISCLOSURES

The authors declare no conflict of interest.

Supporting information

Figure S1

Figure S2

Figure S3

Figure S4

Figure S5

Figure S6

Table S1

Table S2

Table S3

Table S4

Table S5

Appendix S1

Ihirwe RG, Martel J, Rahimi S, Trasler J. Protective and sex‐specific effects of moderate dose folic acid supplementation on the placenta following assisted reproduction in mice. The FASEB Journal. 2023;37:e22677. doi: 10.1096/fj.202201428R

DATA AVAILABILITY STATEMENT

RNA‐seq data are available as supplementary information at FASEB journal online. RRBS data were first used by Rahimi et al., 2019 and are available at NCBI Gene Expression Omnibus accession number: GSE123143.