Effects of thyroid hormone on mitochondria and metabolism of human preimplantation embryos

Abstract Thyroid hormones are regarded as the major controllers of metabolic rate and oxygen consumption in mammals. Although it has been demonstrated that thyroid hormone supplementation improves bovine embryo development in vitro, the cellular mechanisms underlying these effects are so far unknown. In this study, we investigated the role of thyroid hormone in development of human preimplantation embryos. Embryos were cultured in the presence or absence of 10−7 M triiodothyronine (T3) till blastocyst stage. Inner cell mass (ICM) and trophectoderm (TE) were separated mechanically and subjected to RNAseq or quantification of mitochondrial DNA copy number. Analyses were performed using DESeq (v1.16.0 on R v3.1.3), MeV4.9 and MitoMiner 4.0v2018 JUN platforms. We found that the exposure of human preimplantation embryos to T3 had a profound impact on nuclear gene transcription only in the cells of ICM (1178 regulated genes—10.5% of 11 196 expressed genes) and almost no effect on cells of TE (38 regulated genes—0.3% of expressed genes). The analyses suggest that T3 induces in ICM a shift in ribosome and oxidative phosphorylation activity, as the upregulated genes are contributing to the composition and organization of the respiratory chain and associated cofactors involved in mitoribosome assembly and stability. Furthermore, a number of genes affecting the citric acid cycle energy production have reduced expression. Our findings might explain why thyroid disorders in women have been associated with reduced fertility and adverse pregnancy outcome. Our data also raise a possibility that supplementation of culture media with T3 may improve outcomes for women undergoing in vitro fertilization.

(T3) till blastocyst stage. Inner cell mass (ICM) and trophectoderm (TE) were separated mechanically and subjected to RNAseq or quantification of mitochondrial DNA copy number. Analyses were performed using DESeq (v1. 16.0 on R v3.1.3), MeV4.9 and MitoMiner 4.0 v2018 JUN platforms. We found that the exposure of human preimplantation embryos to T3 had a profound impact on nuclear gene transcription only in the cells of ICM (1178 regulated genes-10.5% of 11 196 expressed genes) and almost no effect on cells of TE (38 regulated genes-0.3% of expressed genes). The analyses suggest that T3 induces in ICM a shift in ribosome and oxidative phosphorylation activity, as the upregulated genes are contributing to the composition and organization of the respiratory chain and associated cofactors involved in mitoribosome assembly and stability. Furthermore, a number of genes affecting the citric acid cycle energy production have reduced expression. Our findings might explain why thyroid disorders in women have been associated with reduced fertility and adverse pregnancy outcome. Our data also raise a possibility that supplementation of culture media with T3 may improve outcomes for women undergoing in vitro fertilization.

| INTRODUCTION
Thyroid hormones, thyroxine (T4), and its biologically active form triiodothyronine (T3) play vital roles in regulating homeostasis and metabolic rate of human cells and tissues. They are essential for physical and mental development; insufficient production of thyroid hormones before birth or during childhood can lead to reduced growth and mental impairment. In adults, hypothyroidism causes reduced metabolism, poor memory, depression, and reduced fertility.
In women of fertile age, hypothyroidism frequently results in menstrual irregularities; ovulation and conception are possible, though resulting pregnancies are liable to miscarry. In addition, women with severe hypothyroidism have been reported to have diminished libido and ovulation failure. 1 Thyroid hormone is detected in follicular fluid and has a role in ovarian follicle cycle and egg maturation. 2 Evidence is now emerging on the importance of thyroid hormone signaling during implantation. 3 Thyroid hormone receptors α (THRA) and β (THRB) transcripts and protein levels were detected in developing bovine preimplantation embryos up to the blastocyst stage. 4 The exposure of preimplantation embryos to T3 in culture medium had a positive effect on their development-a greater viability, higher hatching rate, significant increase in total cell number of blastocysts, as well as a significantly lower number of apoptotic cells. 5,6 Exposure to T3 of human embryonic stem cells (hESCs) derived from the inner cell mass (ICM) of the blastocyst stage embryos and induced pluripotent stem cells triggered expression of multiple genes linked to regulation of gene transcription, cell cycle, morphology, apoptosis, cell viability, and cellular and embryonic development. 7 Here, we report investigations into the effect of thyroid hormone

| Human embryo culture
The embryos were thawed using Quinn's Advantage thaw kit according to the manufacturer's instructions (Sage). After thawing, the embryos were cultured in 40-μL microdrops of Quinn's medium supplemented with 10% synthetic protein serum substitute (Sage) under mineral oil (Sage) at 37 C in 5% CO 2 and 5% O 2 in a humidified atmosphere. Treated embryos were exposed from thawing till harvesting to 10 −7 M T3 (Sigma). The medium was refreshed on a daily basis.
ICM and TE were separated mechanically as described. 8,9 Significance statement Thyroid hormones are regarded as the major controllers of metabolic rate and oxygen consumption in mammals. Little is known about the effects of the thyroid hormones in the earliest stages of human development. The results suggest that thyroid hormones affect mitochondrial function in human embryos: stimulate mitochondrial replication and energy production within mitochondria by switching metabolism from glycolytic pathway to more efficient oxidative phosphorylation. The findings shed a light on metabolic switch in early embryo development and might explain why thyroid disorders in women have been associated with reduced fertility and adverse pregnancy outcome. Data in the present study also suggest that supplementation of culture media with T3 may improve outcomes for women undergoing in vitro fertilization.

| RNAseq and data analysis
The recovered ICM and TE fractions were stored at −80 C till processing. cDNA was prepared using SMARTer Ultra Low Input RNA for Illumina Sequencing-HV (Clontech) directly from cells. All experiments were carried out in a PCR clean workstation using protocols provided by the manufacturer. The purification of the first-strand cDNA was carried out using SPRI Ampure Beads (Beckman Coulter).
The amplification of the ds cDNA was carried out in a PCR Clean workstation using an Advantage 2 PCR Kit as per the manufacturer's protocol. The PCR-amplified cDNA was purified using immobilization onto SPRI beads. One microliter of the amplified cDNA was used for the validation using the Agilent 2100 BioAnalyzer and the High Sensitivity DNA Chip (Agilent) per the manufacturer's instructions. The GenoSplice Technology performed quality control (QC), processing, and further analyses of the data. For each gene present in FAST DB v2014_2, reads aligning on constitutive regions (that are not prone to alternative splicing) were counted. Based on these read counts, normalization and differential gene expression were performed using DESeq (v1. 16.0 on R v3.1.3). 10 Sample clustering was performed with MeV4.9 11 using correlation of normalized read counts of expressed genes and average linkage clustering.

| Mitochondrial gene network analysis
Genes identified to have significant differential expression (n = 1178; fold change ≥1.5, P ≤ .05; GenoSplice, EASANA) were filtered for mitochondrial identity in MitoMiner 4.0 v2018 JUN (accessed December 2018) as defined by the integrated mitochondrial protein index as "KNOWN" mitochondrial. 12 Uniprot identifier of "KNOWN" mitochondrial proteins were imported into STRING version 10.5 (last accessed December 2018), and a kmeans network was built using an interaction score set at a confidence of 0.7 using the whole human genome as background statistical reference. 13

| Quantification of mtDNA copy number
Quantification of mtDNA copy number was performed as described. 14

| Mitochondrial bioenergetic differential events in silico
In an attempt to obtain an integrated picture of T3 effect on energy metabolism in ICM, we filtered the 1178 genes identified to have significant differential expression for mitochondrial identity. Among these genes, 42 mitochondrial genes were upregulated and 17 downregulated ( Table 1). The k-means built network based on these mitochondrial proteins is illustrated in Figure 4A (Table 1).
To understand the change in mitochondrial bioenergetic profile further, we have divided k-means network into an UP ( Figure 4B-D) and a DOWN ( Figure 4E, F). A shift in OXPHOS activity appears to be occurring on T3 exposure, as all the mitochondrial related proteins identified in silico F I G U R E 3 Effects of T3 exposure on transcriptome of human preimplantation embryos. Human 2pn embryos were cultured to the blastocyst stage either in the standard culture medium or in the medium supplemented with 100 nM T3. Inner cell mass was separated mechanically from trophectoderm (TE) (Figure 1), and the transcriptome has been analyzed with RNAseq using Illumina's HiSeq platform ( Figure 2). A, Principal component analysis of samples: untreated control (CTL) ICM (n = 3) and TE (n = 4) and T3-treated ICM (n = 4) and TE (n = 6). Cumulative proportion of variance: 82.4%. B, Analysis of differential gene expression between ICM of embryos cultured in a standard medium and in the medium supplemented with 100 nM T3 has identified 1178 genes (fold change ≥1.5, P ≤ .05). Kyoto Encyclopedia of Genes and Genomes database singled out 11 pathways regulated by T3. The most significantly upregulated are genes in Ribosome and Oxidative phosphorylation pathways (red arrow heads) and downregulated genes in Glyoxylate and dicarboxylate metabolism and Citrate (TCA) cycle (green arrowheads). Taken together, the data suggest a metabolic switch from glycolysis to oxidative phosphorylation accompanied with an increase in protein synthesis (Ribosome pathway). C, Analysis of differential gene expression between TE of embryos cultured in a standard medium and in the medium supplemented with 100 nM T3 has identified only 38 genes (fold change ≥1.5, P ≤ .05) suggesting that TE cells do not respond to thyroid hormone in the same way as the cells from ICM   (Table 1, Figure 4E, F).

| T3 exposure is associated with higher levels of mtDNA in ICM
Since mitochondrial respiratory chain subunits are encoded by both nuclear and mitochondrial genomes, communication between nuclear and mtDNA is essential for mitochondrial function. To determine whether the replication of these genomes is coordinated following exposure to T3, we analyzed expression levels of mtDNA encoded genes.
Mitochondrial DNA is circular in shape and contains 16 569 base pairs. 15 The entire human mitochondrial genome encodes 37 genes, 22 of them code for transport RNA (tRNA), 2 code for ribosomal RNA (rRNA, RNR1, and RNR2), and 13 code for proteins that are necessary for cellular energy production: cytochrome b (CYTB), cytochrome  Figure 5A). Among these, we found that only MT-ND3 was significantly increased in cells of ICM exposed to T3. This suggests that, in the case of T3, the primary effect of T3 was on the nuclear, not the mitochondrial genome.
The T3-induced shift in OXPHOS activity in ICM may not simply be a result of upregulation of the genes that are contributing to the composition and organization of the respiratory chain; the shift could also indicate an increase in the number of mitochondria. 16 Indeed, mtDNA copy number as determined with qPCR was significantly increased in ICM and not in TE of T3-treated embryos ( Figure 5B), suggesting a crosstalk between the nuclear genome and mitochondrial replication. The nature of the crosstalk is currently unknown and remains to be investigated.

| DISCUSSION
Thyroid hormones are major regulators of metabolic rate. Although the mechanism is not well understood, they have a profound effect on mitochondrial energetics, adenosine triphosphate (ATP) production, and OXPHOS. Studies on metabolism of bovine embryos have shown that ICM and TE have pronounced differences in carbohydrate metabolism, which may reflect their diverging roles in embryological development. 17 The ICM had lower pyruvate consumption per cell than did the isolated TE, which is indicative of very low oxidative metabolism. Glycolytic index of TE indicated that all the glucose consumed was likely converted to lactate.
Transmission electron microscopic analyses of mitochondria in human preimplantation embryos have shown that dramatic changes in mitochondrial morphology occur during the differentiation of blastomeres in TE and ICM. Rounded to oval "immature" mitochondria are transformed into elongated tubular forms with well-defined transverse cristae, a sign of increased metabolic activity. 18 Mitochondria have been shown to be the major sites of T3 accumulation in cells, potentially having a direct effect on mitochondrial activity. 19 Indeed, subsequent studies have demonstrated that multiple isoforms of THRA and THRB are present in mitochondria. 20 Although our data suggest that T3 induced mitochondrial biogenesis in ICM, we did not detect an increase in the expression of these genes, which suggests that our observed increase in number of mitochondria was mediated through a different pathway or pathways. For example, recently, a group of small non-coding RNAs (sncRNA), known as P-element-induced wimpy testis (piwi)-interacting RNAs (piRNA), corresponding to mtDNA, have been described. 26 Nuclear DNA-encoded piwi-like RNA-mediated gene silencing 1 (PIWIL1) protein was detected in both nuclear and mitochondrial fractions suggesting a novel crosstalk between nuclear and mitochondrial genomes. 26 mtDNA-encoded piRNA was quite abundant in germ cells, gametes, and zygotes. 27 Mitochondrial respiration contributes to energetic requirements during embryonic development. In the cleavage stage only 10% of available glucose is metabolized through aerobic respiration, whereas this is increased at the blastocyst stage to 85%. 28,29 Oxygen consumption has been linked to reproductive competence in mouse. suggested that successfully implanted blastocysts tended to contain lower mtDNA levels than those failing to implant. 33 The study also suggested that the quantity of mtDNA was significantly higher in embryos from older women and in aneuploid embryos, unrelated to age.
In contrast, mitochondrial transplantation has recently emerged as a potential treatment of various diseases, including restoration of fertility. 34 The first babies born following autologous mitochondria  45 Whether and how T3 is involved in any switch from ground stage of pluripotency of naive hESC to primed pluripotency of hESC remains to be determined.

| CONCLUSION
Little is known about the effects of the thyroid hormones in the earliest stages of human development. However, women of fertile age with underactive thyroid glands are prone to suffer from menstrual irregularities and, although ovulation and conception are possible, resulting pregnancies are prone to miscarriage.
Our results indicate that thyroid hormones affect mitochondrial function in human embryos; although transcription of mitochondrial DNA (mtDNA) genes was not affected by T3, mitochondrial replication was stimulated, as was energy production within mitochondria by switching metabolism from the glycolytic pathway to more efficient OXPHOS. This effect was restricted to the ICM, the part of the embryo that gives rise to the body of the fetus.
Our report suggests a new mechanism of metabolic switching in human preimplantation embryos and highlights an important role of thyroid hormones in the early stages of pregnancy. It is plausible to hypothesize that supplementing commercial media with physiological concentrations of thyroid hormone might have a beneficial effect on embryo development and improve implantation and pregnancy rate.
Understanding of the basic biochemistry and roles of the ICM and TE in homeostasis of the blastocyst is still limited. One hypothesis is that the TE is a transporting epithelium, sparing nutrients for use by the ICM. 46 More recent studies in mouse embryos have demonstrated that the TE consumes significantly more oxygen, produces more ATP, and contains a greater number and more mature mitochondria than the ICM. 47,48 The data from our study, although limited, suggests that such a balance in energy production between pluripotent ICM and dif-

DATA AVAILABILITY STATEMENT
The data have been deposited in the Gene Expression Omnibus (GEO) under accession number GSE131499.