Beta-hydroxybutyrate alters bovine preimplantation embryo development through transcriptional and epigenetic mechanisms^†

Abstract

Developing embryos are susceptible to fluctuations in the nutrients and metabolites concentrations within the reproductive tract, which can lead to alterations in their developmental trajectory. Ketotic dairy cows have diminished fertility, and elevated levels of the ketone body beta-hydroxybutyrate (BHB) have been associated with poor embryonic development. We used an in vitro model based on either in vitro fertilization (IVF) or parthenogenesis to investigate the effects of BHB on the preimplantation bovine embryo development, epigenome, and transcriptome. Embryo culture medium was supplemented with BHB at a similar concentration to that present in the blood of cows suffering with severe ketosis, followed by analysis of blastocysts formation rate, diameter, total number of cells, levels of H3K9 beta-hydroxybutyrylation (H3K9bhb), apoptosis, and transcriptional alterations. As a result, we observed that BHB reduced the blastocysts rates, the diameter and the total number of cells in both parthenotes and IVF embryos. Exposure to BHB for either 3 or 7 days greatly increased the H3K9bhb levels in parthenotes at the 8-cells and blastocyst stages, and affected the expression of HDAC1, TET1, DNMT1, KDM6B, NANOG, and MTHFD2 genes. Additionally, culture of IVF embryos with BHB for 7 days dramatically increased H3K9bhb and reduced NANOG in blastocysts. RNA-seq analysis of IVF blastocysts revealed that BHB modulated the expression of 118 genes, which were involved with biological processes such as embryonic development, implantation, reproduction, proliferation, and metabolism. These findings provided valuable insights into the mechanisms through which BHB disrupts preimplantation embryonic development and affects the fertility in dairy cows.

Graphical Abstract

Open in new tabDownload slide

Introduction

Ruminants grazing in grasslands and pastures worldwide suffer from scarcity of food depending on the rainy season, the grass-growing cycle and other climate-related challenges [1]. There are other physiological or pathological conditions in which mammalians suffer energy restrictions: (1) ewes and goats by the end of pregnancy due to the increased energy demand caused by the growing fetus; [2] (2) Humans adopting ketogenic diet, time-restricted feeding regimens, practicing intense exercises, and suffering from diabetic ketoacidosis [3, 4]; (3) negative energy balance in cows following the parturition and early lactation [5].

Curiously, a molecule that increases in all of these situations is the ketone body beta-hydroxybutyrate (BHB hereafter). In mammals, during periods of food scarcity, BHB is produced in the liver, released into the bloodstream, and terminally oxidized in extrahepatic mitochondria from peripheral tissues to generate energy [6]. In the context of animal production, ketosis poses a significant challenge for the dairy industry. High-producing dairy cows which fail to maintain an adequate energy balance after parturition and beginning of lactation develop ketosis, a condition marked by elevated circulating levels of BHB [7]. Ketosis has been associated with poor oocyte competence, failure to conceive after artificial insemination [8], slower embryonic development, and transcriptional and epigenetic alterations [9]. However, several mechanisms leading to all of these alterations are unknown.

Beyond its canonical role as an energy provider, emerging studies have demonstrated that BHB is a potent epigenetic modifier. It has been shown to: 1- function as a histone deacetylase inhibitor [10]; 2- indirectly modulate various other epigenetics modifications (e.g. DNA and histone methylation, lysine succinylation, and reduction in other acylations) [11], and most importantly, 3- directly cause an epigenetic modification known as the histone lysine β-hydroxybutyrylation (Kbhb) [12].

The role of β-hydroxybutyrate in driving Kbhb during ketogenesis provides a dramatic example of an endogenous short-chain fatty acid that is used for histone acylation [12]. Ketogenesis is a physiological response to low blood glucose levels. It can be caused by nutrient deprivation or excessive glucose utilization, as observed in early lactating cows [13]. Under ketogenic conditions, serum levels of BHB can increase from micromolar to millimolar concentrations. In bovines, the early embryos develop the first days in the oviduct and reach the uterus ~5 days after fertilization. The oviduct and the uterus support the embryo development through secretion of nutrients, which are influenced by the stages of the estrous cycle and the metabolic condition of the animal [14, 15]. Consequently, early embryos developing in an undernourished or ketotic animal might be exposed to elevated levels of BHB in the oviductal and uterine secretion [16].

The fact that BHB has emerged as a potent epigenetic modifier and is increased in all dietary paradigms involving undernutrition makes it an ideal candidate to investigate the connection between nutrition, epigenetics and reproduction. In vivo, BHB increases only following bulk nutrient restriction, exposing developing embryos to a complex hormonal and nutritional environment. Conversely, in vitro, it is possible to add a single molecule to the culture medium and track its effects on embryonic development.

Building upon this foundation, the objective of this study was to examine whether exposing preimplantation embryos to pathophysiological concentration of BHB during in vitro culture could interfere with its development, epigenome and transcriptome. To address this, we carried out a series of experiments to investigate whether: (1) BHB causes histone lysine H3K9 beta-hydroxybutyrylation (H3K9bhb) accumulation in embryos; (2) BHB affects embryo development in vitro; (3) BHB interferes with the embryonic growth, cell proliferation and apoptosis; (4) BHB modulates the embryonic gene expression.

Material and methods

Chemicals and reagents used were purchased from Merck KGaA (Darmstadt, Germany) and ThermoFisher Scientific Inc. (Waltham, MA, USA), unless otherwise stated.

Ethical considerations

The present study was approved by the Animal Experimentation Ethics Committee of the University of São Paulo—Faculty of Animal Science and Food Engineering. We conducted the experiments in accordance with the International Guiding Principles for Biomedical Research Involving Animals, as it is recommended by the Society for the Study of Reproduction to its members.

Oocyte recovery and in vitro maturation

Ovaries were collected from a local slaughterhouse, washed, and transported to the laboratory in an insulated container filled with saline solution (0.9% NaCl) at ~30°C. Oocytes were aspirated from visible antral follicles (~2 to 8 mm) using an 18 G needle connected to a syringe. Cumulus-oocyte complexes (COCs) with compact cumulus cells layers and homogeneous cytoplasm were selected and matured in groups of 50 COCs in 500 μl of M199 with Earle’s salts (ThermoFisher cat. # 11150–059) supplemented with 0.5 μg/ml FSH (Folltropin; Ourofino Saúde Animal, Cravinhos, Brazil), 5 U/ml hCG (Vetecor; Ourofino Saúde Animal), 50 μg/ml gentamicin sulfate, 0.2 mM sodium pyruvate, and 10% fetal bovine serum. In vitro maturation (IVM) was performed for ~22 h in a humidified atmosphere of 5% CO₂ in air at 38.5°C.

Stock solution of beta-hydroxybutyrate

A stock solution (1 M) of Beta-hydroxybutyrate (Sodium (R)-3-hydroxybutyrate, Sigma cat. # 298360) was prepared diluting the salt in ultra-pure water. BHB was added to the embryo culture medium to reach a final concentration of 6 mM. To ensure the effects were not caused by changes in the medium osmolality or pH, we measured these parameters after medium supplementation. The BHB addition to the SOF medium marginally increased the osmolality to ~289 mOsm/kg (Osmo1—Advanced Instruments) compared with control SOF medium (~282 mOsm/kg), but still within the normal physiological range (270 to 290 mOsm/kg in the plasma). The pH marginally increased from 7.4 to 7.52. However, after the culture dish was prepared, it was placed back in the incubator to re-equilibrate the pH before the time the embryos were cultured.

Production of parthenogenetic embryos and treatment with BHB

After IVM, COCs had their cumulus cells removed by incubation in 400 μl of trypsin (Tryple Express, ThermoFisher) and gentle pipetting. The denuded oocytes were washed with M199 containing 25 mM HEPES (ThermoFisher, cat. # 12350–039) and 10% FBS, and selected for the presence of the first polar body using a stereomicroscope. Matured oocytes were artificially activated 26 h after the beginning of in vitro maturation by incubation with 5 μM ionomycin diluted in M199 + 1 mg/ml BSA. The presumptive zygotes were washed 3 times in M199 + 30 mg/ml of BSA and incubated in SOF medium, containing 2 mM 6-DMAP for 3 h to complete the activation protocol. After thorough washing, the presumptive zygotes were evenly divided into two groups. One group was cultured in SOFaa (Control group), and the other in SOFaa supplemented with 6 mM BHB (BHB group) for 7 days (168 h) in a humidified atmosphere of 5% CO₂, 5% O₂ in air at 38.5°C. For the experiments involving 8-cells embryos, parthenotes were cultured for a duration of 72 h.

In vitro fertilization and embryo treatment with BHB

Oocytes matured as described above were inseminated in vitro with frozen–thawed semen from a fertile bull. Capacitated sperm were obtained after Percoll gradient (45% and 90%) separation. The fertilization medium was composed of Tyrode’s lactate stock, 50 μg/ml gentamicin, 22 μg/ml sodium pyruvate, 40 μl/ml PHE (2 mM D-penicillamine, 1 mM hypotaurine and 245 μM epinephrine), 5.5 IU/ml heparin, and 6 mg/ml bovine serum albumin (BSA). Sperm and matured COCs were co-incubated for 6 h in droplets of 100 μl of fertilization medium covered with mineral oil and maintained in an incubator at 38.5°C and 5% CO₂. The resulting presumptive zygotes were denuded by gentle pipetting in a 400 μl drop of trypsin solution (Tryple Express, ThermoFisher), washed in SOFaa medium, and randomly and evenly split in two groups. One group was cultured in SOFaa medium (control group) and the other in SOFaa medium supplemented with 6 mM BHB (BHB group). The blastocysts were cultured for 7 days, with the day of COCs insemination considered as Day 0. The culture was carried out in an incubator with a humidified atmosphere of 5% CO₂ in air at 38.5°C for 7 days. The cleavage rate and the blastocyst rate were evaluated at 72 h and 168 h after in vitro fertilization, respectively. Blastocysts were harvested and used for subsequent analysis.

Blastocyst diameter measurement

Blastocysts (i.e. parthenotes or in vitro fertilized) were recovered 168 h after in vitro fertilization (IVF) or artificial activation (Day 7) and placed in a drop of SOFaa. Brightfield images were acquired at 4x magnification using an inverted microscope (Nikon Eclipse TI) equipped with a Nikon photo documentation system. A line was manually traced in the equatorial region of the embryo spanning the zona pellucida from end to end. The diameter was calculated by comparing the length of this line with a reference bar of known length. The blastocysts were considered a perfect sphere, despite some distortions in their shape caused by the expansion or pre-hatching activities.

Total number of cells in blastocysts

Blastocysts (i.e. parthenotes or in vitro fertilized) were fixed in 4% paraformaldehyde (PFA) in PBS for 12 min, washed three times with PBS + 10 mg/ml BSA + 0.2% Triton X-100 (v/v), permeabilized with PBS + 1% Triton X-100 (v/v) for 20 min. The nuclei were stained by incubating with Hoechst 33342 for 30 min. The blastocysts were mounted on a glass slide, covered with a coverslip using ProLong Gold Antifade Mountant (Life Technologies, cat. # P36935). Images from the blastocysts were captured using an epifluorescence microscope (Zeiss AxioPlan 2, Zeppelingstrasse, Germany), equipped with a digital camera (Zeiss MC 80 DX), and using a 20x objective. To estimate the total number of cells in the blastocysts, the nuclei labeled with Hoechst 33342 were counted using the ImageJ plugin Cell Counter, by manually clicking on each nucleus.

In situ cell death detection in bovine parthenote blastocysts

To detect apoptotic cells in parthenote blastocysts, we used the In Situ Cell Death Detection Kit (Roche cat# 11684795910) following the manufacturer’s instructions with minor modifications. Briefly, parthenote blastocysts were collect after 7 days of culture, washed in PBS + 0.1% PVP, and fixed in 4% PFA in PBS for 15 min. Embryos were permeabilized in 1% Triton X-100 + 0.1% sodium citrate solution in PBS for 20 min. Blastocysts were washed in PBS + 0.1% PVP and divided in four groups before the TUNEL reaction: (1) Control (untreated embryos), (2) 6 mM BHB-treated embryos, (3) negative control for the reaction, and (4) positive control for the reaction. The TUNEL reaction mixture was prepared first by removing 100 μl from the Label solution (vial 2) for the negative control. Then 50 μl of the Enzyme solution (vial 1) was added to the remaining 450 μl Label solution in vial 2 to obtain a 500 μl TUNEL reaction mixture. The embryos used for the positive control were treated with DNase I amplification grade (Thermo cat# 18068015) for 15 min at RT to induce DNA strand breaks prior to labeling procedures. To label the embryos to detect apoptotic cells, embryos from the experimental control and 6 mM BHB-treated groups, and the positive control for the reaction embryos were incubated in 100 μl drops of TUNEL reaction mixture at 37°C in a humidified atmosphere in the dark for 1 h. The negative control embryos were incubated in 100 μl of Label solution only in similar conditions. After labeling, embryos were incubated with Hoechst 33342 for DNA staining, washed thrice in PBS/PVP and mounted on glass slides covered with coverslips in ProLong Gold mounting medium (Thermo cat# P36934).

DNA synthesis and cell proliferation assay in IVF blastocysts

The rate of DNA synthesis/cell proliferation was measured using the nucleoside analog of thymidine EdU (5-ethynil-2′-deoxyridine), incorporated into DNA and detected via a click reaction. The Click-iT EdU imaging kit (Thermo cat# C10337) was used following the manufacturer’s instructions. Briefly, 162 h after IVF, we added 10 μM EdU directly to the embryo culture drops and returned the dish to the incubator for 6 h to allow the embryos to incorporate the EdU, totaling 168 h of culture, which was the endpoint for all experiments involving blastocysts. The blastocysts were fixed in 4% PFA in PBS for 15 min, followed by a 1% Triton X-100 permeabilization step for 20 min. The permeabilization solution was removed, and the embryos were washed twice with 1 ml of 3% BSA in PBS. EdU was detected by incubating the embryos in a 200 μl drop of Click-iT reaction cocktail. The embryos were incubated for 30 min at RT, protected from light. The embryos were then removed from the reaction cocktail, and washed twice with 3% BSA in PBS. Subsequently, the embryos were stained with Hoechst 33342 for 15 min, washed in PBS/BSA, and mounted on a glass slide with a coverslip using ProLong Gold (Thermo cat# P36934). The embryos were imaged using a Leica Mica Widefocal Live Cell microscope.

Immunostaining of 8-cell embryos and blastocysts

Embryos were fixed in 4% paraformaldehyde (PFA) in PBS for 12 min, washed three times with PBS + 10 mg/ml BSA + 0.1% Tween 20 (PBST), and permeabilized with PBS + 1% Triton X-100 (v/v) for 20 min. Samples were washed with PBST and blocked in PBS + 3% BSA (w/v) + 0.3 M glycine +0.1% Tween 20 solution for 1 h at RT. The samples were incubated overnight at 4°C on a rocking platform with the respective primary antibodies: 1- anti-H3K9bhb (1:1000; PTM biolabs cat. # 1250); 2- anti-NANOG (1:1000; PeproTech cat. # 500-P236); 3- anti-acetyl lysine (1:1000; Abcam cat. # AB21623) diluted in PBST. After extensive washing (5 × for 10 min) with PBST, the samples were incubated with the Alexa Fluor 488-conjugated goat anti-rabbit IgG secondary antibody (ThermoFisher, cat. # A11008), diluted in PBST at 1:1500 for 1 h at room temperature protected from light and under agitation. Appropriate negative controls were obtained by substituting the primary antibody with a rabbit polyclonal IgG isotype control (Abcam, Cambridge, MA, USA, cat # AB27478) or omitting the primary antibody (secondary-only control). Samples were counterstained with Hoechst 33342 (1 μg/ml) and mounted on a glass slide in ProLong Gold Antifade Mountant.

The samples were visualized, and images were captured using either a confocal microscope (TCS-SP8 AOBS; Leica, Wetzlar, Germany) or the epifluorescence microscopes (Zeiss AxioPlan 2, Zeppelingstrasse, Germany) or (Leica Thunder Imager 3D Assay). The same microscope, settings and parameters were applied for all images within the same experiment, thereby enabling direct comparison of the signal intensities among the samples. To quantify the pixel intensities of the images, nuclei were individually outlined using the manual tool, and the average signal intensity was calculated using the ImageJ software (National Institutes of Health, Bethesda, MD, USA).

Analysis of gene expression in parthenotes embryos

To analyze gene expression, we collected (1) seven pools of twenty 8-cell parthenotes embryos 72 h after artificial activation; and (2) five pools of 10 parthenotes blastocysts 168 h (7 days) after artificial activation. We washed them in PBS + 1 mg/ml BSA, transferred into RNase/DNase-free 1.5-ml microcentrifuge tubes, flash-frozen in liquid nitrogen, and kept them at −80°C until analysis. Total RNA was extracted using TRIzol reagent (Thermo Fischer) following the manufacturer’s instructions with a few modifications. Briefly, 1.33 μl of GlycoBlue Copreciptant (Thermo Fischer Scientific) was added to the aqueous phase, and after the precipitation step with isopropanol, samples were centrifuged at 20,000 g for 30 min to increase the RNA yield. The extracted RNA was dissolved in ultra-pure water and treated with DNAse I (Invitrogen) to eliminate any potential genomic DNA contamination. RNA was converted to cDNA using the High Capacity RNA-to-cDNA Kit (Applied Biosystems, Foster City, CA) following the manufacturer’s instructions.

For qRT-PCR analysis, 20 target genes in 8-cell embryos and 14 genes in parthenotes blastocyst were investigated. To normalize the gene expression data, the geometric mean of 3 previously validated reference genes (ACTB, PPIA, and TBP) were used [17]. The gene identification, the primer sequences, and the accession numbers can be found in Supplementary Table 1. The primers were designed using the Primer-BLAST (NCBI) software based upon sequences available in Ensembl or GenBank databases.

Relative quantification of gene-specific mRNA transcripts was performed in 10-μl reactions containing 200 nM of each primer, 5 μl of PowerUp SYBR Green PCR Master Mix (Applied Biosystems), and 2 μl of 8-fold diluted cDNA. The qRT-PCR conditions were determined based on pilot experiments involving different cDNAs dilutions and primers concentrations. All gene-specific cDNAs from a given sample were run in duplicate on the same qRT-PCR plate. To monitor for contamination, a non-template control containing 2 μl of ultra-pure water instead of cDNA was always run in parallel with samples. The following cycling conditions were applied for amplification: initial denaturation at 95°C for 10 min, followed by 40 cycles consisting of 95°C for 15 s (denaturation) and 60°C for 1 min (annealing and extension). SYBR Green fluorescence was read at the end of each extension step (60°C). The amplification of a single PCR product was confirmed by analysis of the melting curves. The geometric mean was calculated by using the Ct values of housekeeping genes (ACTB, PPIA, and TBP). The geometric mean of the housekeeping genes was used to obtain the delta Ct (dCt) values of the target genes, and to calculate the relative expression. Next, the data were transformed by the 2^-dCt method and are presented in the figures. Gene expression data from seven biological replicates for 8-cell embryos and five biological replicates for parthenotes blastocysts are presented.

Western blot in bovine parthenotes blastocysts

Twenty parthenote blastocysts were collected, washed in PBS, transferred to a 0.6 ml tube, and flash-frozen in liquid nitrogen. Protein were extracted by adding 7.5 μl of RIPA buffer directly to the tube followed by vigorous pipetting. The protein extracts were kept on ice for 1 h to allow the proteins to solubilize. Next, 2.5 μl of 4X Laemmli buffer (Bio-Rad cat# 161–0747) were mixed with the samples and boiled at 95–100°C to denature the proteins. Proteins were loaded in a 4–15% precast gel (Bio-Rad cat# 456–1086) and run at a constant 100 V for 90 min. After electrophoresis, proteins were transferred to a PVDF membrane using a semi-dry transfer system (TransBlot Turbo Bio-Rad) using a transfer pack (Bio-Rad cat# 170–4156). The membranes were blocked for 1 h at room temperature (RT) using a 3% BSA solution in Tris-buffered saline with 0.1% Tween-20 detergent (TBS-T). The NANOG antibody (Peprotech cat# 500-P236) was added directly to the blocking solution (1:3000) and the membranes were incubated overnight at 4°C under agitation. The next day, the membranes were washed 5 times for 3 min in TBS-T solution and incubated for 1 h RT with HRP-conjugated anti-rabbit secondary antibody (Sigma cat# A0545; diluted 1:5000). Membranes were washed 5 times for 3 min and proteins were detected through incubation with Clarity ECL substrate (Bio-Rad cat# 170–5060). Images were taken in ChemiDoc MP imaging system (Bio-Rad), and the band intensity was quantified using the Image Lab 5.1 software. After NANOG detection, membranes were incubated with Restore WB stripping buffer (Thermo cat# 21059) to dissociate and strip the primary and secondary antibodies, blocked and re-probed with Anti-GAPDH antibody (Sigma cat # G9295) to serve as a loading control for the amount of loaded proteins.

RNA extraction from IVF blastocysts and RNA-seq library preparation

Three pools of seven blastocysts, either cultured under control condition or treated with 6 mM BHB, were used for RNA-seq. Total RNA was isolated using the Arcturus PicoPure RNA Isolation Kit (ThermoFisher Scientific cat # KIT0204), following the manufacturer's instructions. The RNA integrity and quantity were assessed using the RNA 6000 Pico Kit and the 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA), and all the samples presented RIN values ranging from 9.3 to 9.9. After the quality check, we proceeded with cDNA synthesis and amplification using the SMART-Seq HT Kit (Takara Bio Inc, Kusatsu, Japan). The cDNA quantity and profile were checked using the Qubit dsDNA High Sensitivity kit (ThermoFisher Scientific) and the High Sensitivity DNA Kit (Agilent Technologies), following by library preparation using the Nextera XT DNA Library Prep (Illumina Inc, San Diego, CA, USA). RNAseq library quantity and size were again checked using the Qubit dsDNA High Sensitivity kit, pooled and sequenced on a NextSeq 2000 (Illumina Inc) with 100 bp paired-end reads.

RNA-seq pipeline

The data were processed and passed a quality check by FASTQC software (http://www.bioinformatics.babraham.ac.uk/projects/fastqc). Reads were filtered according to data quality using TrimGalore/cutadapt pipeline (https://www.bioinformatics.babraham.ac.uk/projects/trim_galore/), with a minimum Phred quality score of 24 at the start and end of the read, and a minimum read size of 30 bps. After quality filtering, we repeated the QC analysis with FASTQC software. Samples that passed the quality screening were then aligned with the Bos taurus genome (assembly ARS-UCD1.3) using Star with standard parameters [18]. The alignment quality was then verified with a final report generated using the MULTIQC software (https://multiqc.info/). After alignment, the reads were counted using the function feature counts implemented in Rsubread using the parameters “strandSpecific = 0” and “isPairedEnd = T”.

Differential gene expression analysis

First, we performed a principal component analysis (PCA), followed by differential gene expression analysis using the DESeq2 R package [19]. To compare the two groups, i.e. Control × BHB, we considered a gene expressed when it had at least 10 counts in at least three samples, and it was considered as differentially expressed gene when adjusted p-value (False Discovery Rate—FDR) < 0.05 (Benjamini-Hochberg method [“BH”]) and absolute log2FoldChange > = 0.6. For global analysis such as heatmap and PCA, we used the variance normalization transformation (vst, “varianceStabilizingTransformation” function from DESeq2 package). For enrichment analysis, we submitted the lists of differentially expressed genes to the ClusterProfiler [20, 21] and DAVID Bioinformatics Resources 6.8 using default parameters [22].

Additional analysis

Additional graphics were performed and constructed using R (https://www.r-project.org). Genes were translated from symbol to entrez id using the reference “org.Bt.eg.db” (https://bioconductor.org/packages/release/data/annotation/html/org.Bt.eg.db.html) through bitr function from ClusterProfiler package [21].

Additional packages for colors for all heatmaps and graphics were select and created using the R package “RColorBrewer” (https://cran.r-project.org/web/packages/RColorBrewer/index.html) and the function “colorRampPalette” from grDevices package (https://rdocumentation.org/packages/grDevices/versions/3.6.2).

Statistical analysis

Statistical analysis was performed using GraphPad Prism 10 (GraphPad Software, San Diego, California, USA). Data were tested for normality of residuals and homogeneity of variances using the Shapiro–Wilk test and analyzed as described in the text. When normality and homoscedasticity were not met, non-parametric test Mann–Whitney (Wilcoxon Rank Sum Test) was applied. Parametric variables were analyzed with a two-tailed Student’s t-test for independent samples. Frequency data were analyzed by chi-square test. Values are presented as means, means ± the standard error of the mean, median, or percentages.

Results and discussion

Exposure to BHB during the preimplantation period impaired parthenogenetic embryonic development and growth

Undernutrition during the periconceptional period causes several reproductive and metabolic problems in both fetuses and adult animals [23]. In rats, maternal undernutrition during the preimplantation period caused abnormalities in blastocysts, altered birth weight, postnatal growth rate, hypertension, and organ/body-weight ratios [24]. In ewes, undernutrition during the period of early embryonic development reduced the total and the viable number of embryos, possibly mediated by disruption of endocrine homeostasis, oocyte quality, and the oviduct microenvironment [25]. BHB is a molecule that increases in situations where there is scarcity of nutrients and it has been demonstrated to serve as an energy source for preimplantation embryos [26, 27]. However, its effects on embryo quality, morphology, number of cells, and epigenome have not been investigated in cattle thus far.

To test the viability of BHB as a supplement for embryo culture, we investigated whether: (1) parthenogenetic embryos can develop normally in culture medium supplemented with high concentration of BHB; and, (2) it affects the embryo morphology, size and total number of cells. Considering the circulating levels of BHB in severely ketotic cows [28, 29], 6 mM BHB was chosen for embryo treatment during the entire culture period (7 days, from oocyte parthenogenetic activation to the blastocyst stage).

We carried out 38 biological replicates, and activated 2085 oocytes in the control group and 2087 in the treated group. As a result, BHB treatment did not affect the cleavage rate (Figure 1B). In the control group, 1845 embryos cleaved, and the rate was on average 88.57 ± 0.86%, while 1854 embryos cleaved (88.28 ± 0.90%) for the BHB-treated (P = 0.72 from a chi-square). Regarding the blastocyst rate, the BHB moderately impaired the embryo development. In the control group, 50.06 ± 1.42% (n = 1049) of the activated oocytes developed into blastocysts, whereas in the BHB-treated, n = 974 reached the blastocyst stage (47.08 ± 1.38%), P = 0.019 from a chi-square, Figure 1C). These data demonstrated that BHB slightly impaired the development of bovine embryos. When we examined the blastocysts under a stereomicroscope, we noticed that the embryos seemed slightly smaller (as can be seen in the brightfield pictures from five different replicates; Figure 1A)—suggesting they were developing at a slower pace than the control group. This observation prompted us to measure the diameter to check whether they were indeed smaller. We measured the diameter of 897 embryos from the control group and 854 from the BHB-treated group. On average, BHB-treated embryos had 182.8 ± 1.10 μm of diameter, which was ~3.3% smaller than the control embryos (189.1 ± 1.21 μm), P = 0.0006 from Mann–Whitney (Figure 1D). The diameter of the embryo is somewhat correlated with the total number of cells [30]. We counted the total number of cells in n = 595 blastocysts from the control group and n = 581 from the BHB-treated group. The control embryos had an average 131.4 ± 2.44 cells, while the BHB-treated embryos had ~9 fewer cells (122.1 ± 2.08), P = 0.0326 from a Mann–Whitney test (Figure 1E).

Due to the reduced cell number in BHB-treated blastocysts, there was a possibility that the BHB was increasing the number of cells undergoing apoptosis in the embryos. To detect apoptosis, we carried out a TUNEL assay and observed that the percentage of apoptotic cells (Figure 2A and 2B) did not differ between the control (23.77 ± 1.22%, n = 63) and BHB (25.17 ± 1.48%, n = 53) groups (P = 0.45, from Mann Whitney test).

It is well established that maternal nutrition influences offspring size [31]. Several dietary paradigms, during different periods throughout the pregnancy, have been shown to affect the newborn weight, postnatal growth rate and metabolism [32]. In addition to bulk nutrient restrictions, micronutrients or molecules such as iron, iodine, calcium, vitamins (e.g. A, C, B₁₂) have also been implicated in fetal growth [33, 34]. Pioneer studies have demonstrated that the addition of BHB alone to the whole embryo culture medium caused growth retardation, reduction in DNA synthesis, and neural tube closure defects [35], hinting that BHB can be responsible for many phenotypes observed in nutrient-deprived animals. The molecular mechanisms behind these observations might involve alterations in embryonic metabolism, epigenome and transcriptome due to the pleiotropic roles of BHB [27, 35].

H3K9bhb is elevated, and global acetylation is reduced in 8-cell parthenotes embryos cultured with high levels of BHB

Bovine embryos rely on stored maternal RNAs to develop up to the 8-cell stage, a period during which they undergo a major wave of embryonic genome activation (EGA) [36]. During this event, an intense chromatin remodeling process occurs, mediated by several genes and transcription factors operating in concert [37]. In the case that the Kbhb is deposited in the chromatin, it can disturb the EGA by: (1) physically blocking other marks (e.g. histone acetylation) that should be in place on the histone tails; (2) deregulating the generation of novel embryonic-specific transcripts; and (3) altering the activity of chromatin remodeling enzymes, which are sensitive to fluctuations in metabolites concentrations [38, 39]. To assess whether BHB affected genes involved in the EGA and the epigenetic marks H3K9bhb and Pan-acetylation (Pan-Kac), we cultured the embryos with BHB for 72 h and analyzed the global levels of Kbhb (H3K9bhb) and Pan-Kac, as well as the expression of several transcription factors and epigenetic enzymes implicated in this process.

As a result, the embryos cultured with BHB (n = 30, Figure 3A) had ~3-fold (Figure 3B) increase in Kbhb immunoreactivity (15.11 ± 0.50 arbitrary units) compared to the controls (n = 23) embryos (5.04 ± 0.32 a.u.), P < 0.0001 obtained with a Mann–Whitney test.

Regarding the global levels of acetylation, embryos treated with BHB (n = 25) had decreased global levels of Pan-Kac (44.76 ± 1.93 a.u., Figure 3C and D) compared with the control (n = 26) ones (51.07 ± 1.67 a.u.), P = 0.0007, Mann–Whitney test.

With respect to the genes involved in EGA, none of them were affected by BHB (Figure 4A and B). With respect to genes involved in epigenetic remodeling, BHB decreased the expression of TET1 (P = 0.0133 from a Paired Student’s T-test) and HDAC1 (P = 0.0356), while increasing the expression of KDM6B (P = 0.0022), Figure 4C.

HDACs are essential for preimplantation development as they act by removing acetylation marks, which prevent the premature expression of developmental genes, thereby safeguarding the EGA [40]. HDAC1 is likely a major deacetylase in preimplantation embryos. RNAi-mediated reduction of HDAC1 resulted in hyperacetylation of histone H4 and caused a developmental delay [41]. Our data showed reduction in HDAC1 expression. We would than expect an increase in the global levels of acetylation. However, we observed the opposite effect. While these results seem controversial, given that BHB has been demonstrated to cause hyperacetylation in some contexts by inhibiting HDACs, recent work has identified a non-canonical function for HDACs. This research demonstrated that Class I HDACs unexpectedly catalyze protein lysine modification with BHB [42]. Thus, the reduction in transcription may be a regulatory mechanism to avoid excessive Kbhb accumulation. Additionally, since Kbhb and Kac compete for the same sites on histone tails, an increment in one modification can lead to a decrease in the competing modification.

The TET enzymes are also essential for early embryogenesis and completion of EGA. Knockdown of all three Tet enzymes in mice arrested the embryonic development and disturbed the EGA [43]. In bovine parthenogenetic embryos, TET inhibition reduced the rate of blastocysts development and the expression of NANOG and POU5F1 [44]. Similarly, we observed reduced TET1 expression, decreased blastocyst development and NANOG expression (see below Figure 6C), indicating conserved roles among mammalian embryos.

Finally, we observed an increase in KDM6B expression. The histone demethylation activity mediated by KDM6B is essential to remove H3K27me3, reduce its global levels, and allow genes essential for EGA to activate properly. Deregulation in KDM6B levels prevented the global reduction of H3K27me3 and compromised the embryonic development to the blastocyst stage in bovines [45]. The KDM6B was also demonstrated to be essential for proper cell allocation during the blastocyst development and to promote adequate trophectoderm formation [46].

Overall, these data emphasize that the three epigenetic enzymes altered after BHB are critical regulators of EGA and blastocyst development. The misregulation of genes caused by BHB exposure during the first 3 days of culture is pointing to a possible molecular mechanism by which this ketone body can compromise embryo development.

Treatment of parthenogenetic embryos with BHB dramatically increases the levels of H3K9bhb at the blastocyst stage

To inquire whether in vitro treatment with BHB increases Kbhb levels in blastocysts, we supplemented the embryo culture medium with 6 mM BHB from the presumptive zygotes stage until the blastocyst stage (7 days). After staining, we observed that blastocysts from the control group (n = 19) showed only faint signals (Figure 5A), and the measured fluorescence intensity was 2.130 ± 0.14 a.u. On the other hand, the H3K9bhb was strongly induced (Figure 5B) in treated blastocysts (n = 15), and they had ~10-fold more fluorescence intensity (22.03 ± 0.54 a.u.), P < 0.0001 from a Mann–Whitney test (Figure 5C).

Blastocyst parthenotes cultured with BHB during the preimplantation period have diminished levels of NANOG protein and alterations in the expression of genes involved in pluripotency, metabolism, and epigenetic regulation

Given the key alterations in crucial epigenetic remodeling enzymes involved in EGA in 8-cell embryos, we postulated that these early alterations could persist and also modify gene expression in blastocysts. To explore the mechanisms by which BHB affects the parthenogenetic embryonic development, we selected and investigated genes involved in epigenetic regulation, pluripotency, trophectoderm fate, and metabolism (Figure 6A).

Among all genes investigated in blastocysts (Figure 6B), we observed that BHB increased the expression of DNA methyltransferase 1 (DNMT1) (P = 0.0120), and reduced the expression of NANOG (P = 0.0049) and MTHFD2 (P = 0.0291), from a Paired Student’s T-test.

DNMT1 is the enzyme responsible for propagating the DNA methylation patterns during DNA replication in somatic tissues and in preimplantation embryonic development [47]. Curiously, periconceptional diet persistently influenced the DNA methylation signature in humans exposed to famine during the Dutch Hunger Winter [48]. In cows, embryos collected from animals in negative energy balance and with elevated serum levels of BHB also presented abnormal methylation patterns [9]. Specifically, the blastocysts exhibited: 1- decreased methylation at subtelomeric regions in every chromosome, along with increased methylation in shore (~0 to 2 kb from the CpG islands) and shelf (~2 to 4 kb from the CpG islands) regions in BHB-stressed animals; 2- higher levels of methylation within CpG islands in embryos that started their development in a low-BHB maternal environment [9]. Interestingly, CpG shore methylation is associated with transcription, and dysregulation in these regions may affect cell differentiation [49]. In human famine or negative energy balance in cows there are massive metabolic alterations, and it seems that those alterations allow the embryos to reprogram their gene expression patterns and adapt to the metabolic stresses impinged by these situations [5, 9, 50].

The MTHFD2 gene is expressed in the developing embryo but absent in most healthy adult tissues, even those that are proliferating [51]. MTHFD2 is integral to mitochondrial one-carbon metabolism, a metabolic system recently implicated in rapid cancer cell proliferation, and it is the enzyme most consistently expressed in tumors. MTHFD2 is particularly responsive to extracellular stimuli, and its expression is repressed upon deprivation of growth signals. In the nucleus, it is involved in DNA replication and translation. The suppression of this enzyme in cancer cells caused enhanced glycolytic flux [52]. Curiously, mouse embryos treated with BHB have enhanced glycolytic flux [27], and one of the mechanisms of teratogenicity caused by the BHB is the decreased rates of DNA synthesis [53]. The BHB-treated blastocysts have reduced expression of MTHFD2 and are smaller in size, indicative of reduced cell proliferation. Additionally, since MTHFD2 is involved in one-carbon metabolism, its misregulation can contribute to the methylation alterations observed in nutrient-deprived embryos.

NANOG is required to form the epiblast and maintain pluripotency in the bovine embryos [54]. Since the epiblast form the embryo proper, alterations in NANOG expression might disturb the entire ontogenesis. The reduction in NANOG might be a consequence of reduced TET expression in 8-cells embryos. Tet1 has an important role in maintaining the expression of Nanog in mouse embryonic stem cells (ES), thus contributing to ES cell maintenance. Downregulation of Nanog via Tet1 knockdown correlates with methylation of the Nanog promoter, supporting a role for Tet1 in regulating DNA methylation status. Furthermore, knockdown of Tet1 in pre-implantation embryos results in a bias towards trophectoderm differentiation [55]. We observed a decrease in the expression of NANOG in blastocysts and TET1 in eight cell-embryos. The misregulation of epigenetic enzymes during the time of EGA, as observed here, can affect the genes expressed later in development. Supporting this view, another study has demonstrated that downregulation or inhibition of TET activity in embryos affected the levels of 5-hmec, promoter methylation, and decreased NANOG expression at the blastocyst stage [56]. To further support our claims, we investigated the levels of NANOG protein in embryos, and we observed a reduction of ~35% in BHB-treated embryos (P = 0.0022, Paired t-test, Figure 6 C–D). The fact that the exposure to BHB alone can modulate several genes related to epigenetic regulation, cell fate specification, and embryo metabolism hints that it probably is a key molecule in mediating the nutrient deprivation phenotype.

Culture of in vitro fertilized embryos with BHB increased H3K9bhb, impaired preimplantation development, and generated embryos smaller and with fewer cells

Treatment of IVF mouse embryos with BHB during culture slowed embryonic development and reduced the total number of cells in the blastocysts [27]. We observed similar findings in bovine parthenogenetic embryos, denoting the usefulness of this model. However, IVF mouse embryos exposed to BHB showed delayed morphokinetic development beginning at syngamy [27] as well as sex-specific effect [57] on development and gene expression—features that cannot be measured in parthenotes. Additionally, parthenotes are limited in their developmental potential and exhibit an aberrant pattern of gene expression beyond the expected alterations in the imprinted genes expression. In contrast, IVF embryos have potential to develop to term and possess a transcriptome more similar to that of their in vivo counterparts. Thus, from this point onward, we opted for IVF embryos to further investigate the effects of BHB on early embryonic development, growth, total cell number, global levels of H3K9bhb, DNA synthesis, and the transcriptome by RNA-seq.

As a result, we observed that, in contrast to parthenotes, the BHB treatment affected the cleavage rates of in vitro fertilized embryos. In the control group, 70.60% ± 1.90 of the embryos cleaved at 72 h after insemination, while in the BHB this rate was reduced to 66.91% ± 1.74 (P = 0.0186 from a chi-square, Figure 7B). Regarding the number of inseminated oocytes reaching the blastocyst stage, the culture with BHB reduced the production of blastocysts by ~14%. In the control group 27.38% ± 1.61% reached the blastocyst stage 7 days (168 h) after insemination, while in the BHB group, 23.47% ± 1.80% (P = 0.0043 from a chi-square test, Figure 7C). Morphologically, the BHB-treated IVF embryos seemed to have a smaller size when compared with the controls (Figure 7A). We measured the diameter and counted the total number of cells. After measurement, blastocysts from the BHB group (n = 326) had, on average, a diameter of 167.5 ± 1.76 μm, and they were ~3.5% smaller (P = 0.0016) when compared with the control (n = 396) group (173.5 ± 1.62 μm), Figure 7D. Regarding the total number of cells, the BHB blastocysts had ~10% fewer cells (128.9 ± 4.225) than the controls (143.2 ± 4.145), P = 0.0167 from an unpaired t test (Figure 7E). Similar to the parthenotes, we also carried out immunofluorescence on the IVF-blastocysts to demonstrate the epigenetic effect of BHB. As a result, we observed that blastocysts from the control group (n = 14) exhibited only faint signals (Figure 7G), and the measured fluorescence intensity was 22.76 ± 0.69 a.u. In contrast, the treated blastocysts (n = 21) showed a strong induction of H3K9bhb (Figure 7H), with ~6-fold higher fluorescence intensity (138.70 ± 0.86 a.u.), P < 0.0001 from a Mann–Whitney test (Figure 7F), corroborating our findings with parthenogenetic embryos.

These data clearly show that the addition of BHB alone to the culture medium greatly increased the Kbhb levels and impaired the embryo development, the growth, and the total number of cells. In agreement, mouse embryos cultured with BHB at all concentrations within a physiological range tested exhibited slowed development, a reduction in the total number of cells, and epigenetic and metabolic perturbations [27]. Furthermore, embryos collected in vivo through uterine flushing from cows with high circulating levels of BHB presented a higher proportion of morulae and a lower proportion of blastocysts, indicating a developmental delay occurring in the first days of gestation [9]. The phenotypes observed in all of these different scenarios are concordant, suggesting a conserved mechanism of adaptation triggered by BHB in mammalian embryos, probably reflecting adaptations to this molecule that signals nutrient scarcity [58].

Transcriptional responses of the IVF-embryos cultured with BHB

Microarray analysis carried out using in vivo-produced bovine embryos collected from cows with low or high-BHB levels have identified ~800 genes of potential interest. Many of these genes were involved in metabolism [9], suggesting that embryos developing in cows with high levels of serum BHB undergo alterations in a specific subset of genes, allowing the embryos to respond and adapt to the “energetically-deficient environment” signaled. Herein, we identified transcriptional alterations in chromatin modifying enzymes, pluripotency and metabolism-related genes in parthenotes, hinting at possible mechanisms by which BHB disturbs embryonic development. To gain a comprehensive view of the transcriptional alterations caused by the treatment, we leveraged the power of RNA-seq to unravel additional altered genes and pathways in IVF embryos (Figure 8A).

Principal component analysis (PCA) indicated that 91% of variance between the two principal components could be explained by the BHB treatment (Figure 8B), suggesting that the treatment led to a substantial alteration in the gene expression profile. Next, we proceeded to differential expression analysis. We identified a total of 12,826 genes expressed in the blastocysts, and 118 genes were differentially expressed genes (DEGs) between the groups (adjusted p-value <0.05) as visualized in the heatmap (Figure 8C), which represent 0.92% of the expressed genes. From those DEGs, 54 were more expressed in the control group (Control ↑, BHB ↓) and 64 in the embryos cultured with BHB (Control ↓, BHB ↑).

Next, we carried out KEGG and Gene Ontology analyses to identify the enriched pathways and GO terms categorized into Biological Process (BP), Molecular Function (MF), and Cellular Component (CC). We found several pathways and biological processes, such as one carbon pool by folate, angiogenesis, muscle and tissue development, stem cell maintenance, maintenance of cell number (Figure 8D and8E), which can explain some of the abnormalities commonly observed in nutrient-deprived embryos.

In all dietary paradigms and epidemiological studies investigating undernutrition thus far, common phenotypes are always present, such as delayed embryonic development, reduced placenta and fetal growth, stunted offspring, abnormal skeletal muscle and cardiovascular development, metabolic alterations, and aberrant DNA methylation patterns [24, 25, 59–61].

We observed enrichment in the BP stem cell maintenance and maintenance of cell number, which could explain the fewer number of cells observed in the blastocysts and, in some way the ontogenesis. The genes involved in this BP are CDH2 and NANOG. CDH2 has been shown to act as a regulator of stem cell fate decisions [62]. NANOG is required to maintain the pluripotency in bovine embryos and to form the epiblast, which, in turn, gives rise to the embryo proper [54]. Since we found reduced transcript levels in both parthenotes and IVF embryos, and protein expression in parthenotes, we envisioned that BHB treated embryos might have fewer cells expressing NANOG. Indeed, we observed that BHB-treated embryos averaged ~12 cells, whereas the Control ones averaged ~17 cells per embryo expressing NANOG (Figure 9A and B), indicating ~29% fewer positive cells (P = 0.0188, Mann–Whitney test). Given the critical role of epiblast cells in embryo formation, beginning development with a reduced cell population could severely impact fetal size. This is because early in embryonic development the cells double with each round of cell division, and later on, the growth becomes exponential. Thus, even a slight initial difference in cell numbers after 40–50 rounds of cell division, necessary for producing a mature organism, can have a substantial effect.

Another possible explanation for the observed phenotype (i.e. reduced cell numbers) is that embryo culture with BHB reduces cell proliferation, thereby delaying embryonic development. DNA synthesis can serve as a proxy for measuring cell proliferation. To test this hypothesis, we cultured embryos with EdU (5-ethynyl-2′-deoxyuridine) and BHB, and measured the rate of DNA synthesis using fluorescent detection of EdU-labeled DNA. We analyzed 28 embryos per group and found no significant differences in the proportion of cells incorporating EdU between the groups (P = 0.488, chi-square test). In both groups, ~78% of cells were positive for EdU (Figure 9C). Specifically, embryos cultured with BHB had 2033 cells out of 2576 (78.92%) incorporating EdU, while in Control embryos had 3139 out of 4014 (78.20%) positive for EdU (Figure 9D). Additionally, among EdU-positive cells, we measured the mean fluorescence levels per nucleus and observed similar DNA synthesis levels between the groups (P = 0.5927, Mann Whitney test), with median fluorescence values of 28.66 for the Control group and 27.30 for the BHB (arbitrary units, Figure 9E). It is possible that BHB slightly stalls DNA synthesis, and the short incubation period (6 h) used here was insufficient to detect this effect. However, extending the incubation period throughout the preimplantation period is not feasible due to the toxic effects of EdU on cell proliferation.

Another critical pathway altered was the “One-carbon pool by folate”. It supports multiple physiological processes such as biosynthesis of nucleic acids (thymidine and purines), redox defense, amino acid homeostasis (glycine, serine, and methionine), and epigenetic maintenance [63]. BHB embryo treatment decreased the expression of the early-embryogenesis-specific mitochondrial dehydrogenase (MTHFD2) and Methionine Synthase (MTR). Given their critical roles, the alteration can provide clues about the mechanisms by which we observed smaller embryos herein. Early embryos and cancers have high usage of the folate pathway; thus, reduction in this pathway can cause several phenotypes (e.g. smaller size, reduced DNA synthesis, abnormal DNA methylation) commonly observed in undernourished or embryos exposed to BHB. Supporting these claims: (1) MTHFD2^−/− embryos are smaller and paler than their WT littermates [63, 64]. (2) Mouse embryos cultured with high BHB have decreased rate of DNA synthesis, and supplementation with ribose reduced the incidence of ketone body-induced neural tube defects [35, 53]. (3) Embryos growing in vivo under high BHB concentrations have abnormal methylation patterns [9]. (4) Humans exposed to famine in utero possess abnormal DNA methylation signatures in genes involved in growth and metabolism [48].

Further supporting the idea that BHB can disturb the expression of genes involved in cardiovascular development and placentation, we found several enriched GO terms related to these processes (Figure 8E), such as muscle development, myofibril assembly, cellular components involved in morphogenesis, tissue development, and angiogenesis. The placenta and cardiovascular systems are the first organs to form during mammalian development. Interestingly, the gene HAND1, altered by BHB, is critical for the development of both [65]. HAND1 starts to be expressed during pre-implantation development and it is essential for differentiation of trophoblast and cardiomyocytes. Mutations in this gene cause abnormal looping in the heart tubes and ventricular myocardial differentiation. Another altered gene encodes the sarcomeric protein CSRP3. Mutations in CSRP3 have been implicated in Hypertrophic Cardiomyopathy and mild skeletal muscle diseases [66]. These findings may provide some clues into the mechanisms underlying the cardiovascular issues commonly observed in individuals or livestock animals who have experienced famines during their development.

Some of the retrieved pathways and GO terms are involved in cancers and other human diseases (e.g. Hepatitis C, Graft-versus-host disease, regulation of viral genome replication). This occurs because gene annotation is based on the cataloged roles that these genes play in humans and mice [67]. That said, the interpretation can be misleading and sometimes confusing when we are investigating the mechanisms by which bovine embryos are being affected. As an example, among the enriched pathways in genes increased after embryo treatment with BHB, we found enrichment for Hepatitis C. Looking into that pathway, we observed that BHB downregulated the genes RSAD2 and ISG15. Curiously, these are interferon-stimulated genes (ISGs) and have been demonstrated to be highly induced by the presence of the conceptus in pregnant endometria [68]. However, little is known about the role these ISGs play during embryo development. ISG15 knockdown decreased the proportion of hatching blastocysts, the diameter of blastocysts, and the cell number per diameter of blastocysts when compared with control embryos. Moreover, ISG15 knockdown also inhibited interferon-tau (IFNT), the pregnancy recognition signal [69]. Mouse embryos cultured in vitro with BHB showed reduced post-transfer viability, with implantation rates being 50% lower compared with controls [27]. These alterations in genes critical for pregnancy recognition can hint about the mechanisms causing the reduced fertility in ketotic dairy cows. By analyzing the genes and pathways modulated by BHB, it is possible to observe that the embryos are reshuffling their transcriptome to cope with the “adverse” environment signaled by the ketone body. From an evolutionary standpoint, these responses reflect the embryos' ability to quickly and dynamically respond to any type of insults in order to adapt and survive. However, in terms of animal production, these responses may be detrimental because (1) they can generate stunted offspring; (2) they can reduce the pregnancy rates and compromise the dairy industry; (3) based on epidemiological studies in other mammals, it may predispose the adult animals to cardiovascular and metabolic diseases [70]; (4) the effect on genes regulating epigenetic marks can affect the health status of the next generation.

Limitations of the study and conclusions

The study has some limitations that should be considered when interpreting these findings, such as: (1) it was carried out only on in vitro embryos, and the end point was at the blastocyst stage. Therefore, it is unknown whether these alterations will affect the developmental potential of the embryos to establish pregnancies and produce a healthy offspring; (2) The in vitro culture provides a static environment, where BHB is added to the medium, and the embryos consume it, causing the concentration decrease over time. In contrast, in vivo, the liver constantly produces and releases ketone bodies into the bloodstream, maintaining the concentration constant. Moreover, the exposure can be chronic and last for months; (3) The concentration of BHB used herein was based on levels measured in the blood of severely ketotic cows [28, 29], as the concentration in the oviductal and uterine fluids has not been precisely measured. Usually, in well-managed herds around the world, at the moment of artificial insemination, cows with subclinical ketosis have BHB blood levels around 1.2 to 2.9 mM [8, 71, 72], which are far lower than the concentrations used herein. However, in an epidemiological survey conducted by our group (unpublished data) and others [73, 74], elevated concentrations of BHB (i.e. >2.9 mM) were found. (4) There is a possibility that in vivo embryos are never exposed to such a high level of BHB. Supporting this claim, in mice, a maternal ketogenic diet elevated the BHB levels in the oviductal fluid, but no correlation was observed between the blood and the oviductal fluid levels in KD-fed mouse [16]. These data suggest that the oviduct can dampen the exposure of the embryo to high levels of BHB; (5) BHB can affect oocytes during their growth and leave epigenetic sequels, which might cause alterations in embryo development, even in healthy cows conceiving months later. (6) In vivo, the embryos growing in a ketotic cow are exposed to a variety of molecules altered concomitantly with the BHB, such as hormones (e.g. insulin, glucagon, IGF-1), lipids, sugars, and metabolites, creating a complex biochemical environment. All of those molecules working together contribute to creating the ketotic embryo phenotype.

Regarding the wider implications of the findings, we cite: (1) Supplementation with the BHB alone, a metabolite commonly present circulating in all mammalian facing nutrient deprivation (i.e. undernutrition, famine, starvation, poor pasture for livestock, ketosis in cattle, pregnancy toxemia in sheep and goats, diabetic ketoacidosis in humans) was capable to affect the embryo development, expression of enzymes critical for epigenetic reprogramming in early embryos, and substantial transcriptional alterations in genes involved in embryo growth, implantation and metabolism. (2) The model/approach used herein can be used to test the effect of a plethora of dietary nutrients and metabolites with strong epigenetic effects (e.g. lactate, butyrate, propionate, acetate) on the embryo metabolism and development. (3) early IVF embryos are very sensitive to fluctuations in the concentrations of all metabolites present in the artificial culture medium, implicating that the modulation of a given component can be used to “artificially program the embryo epigenome and metabolism” in a desired way to increase animal production.

Acknowledgment

Authors thank the slaughterhouse for ovaries’ donation. Members from Laboratory of Molecular Morphophisyology and Development for technical assistance and critical reading. We would also like to thank BioRender.com for helping create some illustrations.

Authors contributions

J.R.S., P.J.R., and F.V.M. Conceptualization; J.R.S., R.P.N., A.B., M.R.C., H.F.R.A.S and R.C.B. Investigation, Formal analysis; J.R.S., R.P.N. and D.R.A. Software; J.R.S., F.P., J.C.S., P.J.R., and F.V.M. Supervision, Funding acquisition; J.R.S. and F.V.M. Writing – Original draft.

Conflict of Interest: The authors have declared that no conflict of interest exists.

Data availability

Data are available from the corresponding author upon reasonable request. The RNA-seq data were deposited in the National Center for Biotechnology Information Gene Expression Omnibus (GEO) and are accessible through GEO series accession number GSE271458.

Footnotes

References