Retinoic acid induces NELFA‐mediated 2C‐like state of mouse embryonic stem cells associates with epigenetic modifications and metabolic processes in chemically defined media

Abstract Objectives Mouse embryonic stem cells (ESCs) are derived from the inner cell mass of blastocyst‐stage embryos and cultured in different culture media with varied pluripotency. Sporadically, a small population of ESCs exhibit 2‐cell stage embryonic features in serum containing medium. However, whether ESCs can transit into 2‐cell embryo‐like (2C‐like) cells in the chemically defined media remains largely unknown. Materials and Methods We established a robust in vitro induction system, based on retinoic acid (RA) containing chemically defined media, which can efficiently increase the subpopulation of 2C‐like cells. Further test the pluripotency and 2C features of ESCs cultured in RA. 2C reporter‐positive cells were selected by FACS; the level of protein was detected via immunofluorescence staining and western blot; the level gene expressions were measured by RNA‐seq. Results Retinoic acid drives a NELFA (negative elongation factor A)‐mediated 2C‐like state in mouse ESCs, characterized with 2C‐specific transcriptional networks and the ability to contribute trophectoderm (TE) when injected into developing embryos. In addition, RA treatment triggers DNA hypomethylation, active histone modification, suppressed glycolysis metabolism and reduced protein synthesis activity of ESCs. Conclusions We showed that RA has a broader role in 2C‐like cells state, not only is one of the upstream regulators of the 2C‐like state in chemically defined media but also illuminates genetic and epigenetic regulations that govern ESCs to 2C‐like transition.


| INTRODUC TI ON
Embryonic stem cells are thought of as a budding of early developmental stage embryos that retain the capacity to differentiate into nearly all cell types except extra-embryonic tissues in chimeric embryos. 1,2 In addition, in vitro cultured ESCs exhibit heterogeneous and retain small populations of cells harbours totipotent which expressing a set of genes are active in 2-cell stage embryos. [3][4][5] These 2-cell-like (2C-like) cells display several features of 2-cell embryos, including expression of 2-cell embryo specific transcripts, active histone modifications, DNA hypomethylation, and the capacity to contribute to both inner cell mass (ICM) and trophectoderm (TE). 6 In mice, the zygotic genome is activated at 2-cell stage by a process known as zygotic genome activation (ZGA) and followed by the gain of totipotency. 7 Thus, 2C-like state in ESCs is an invaluable tool to address the mechanisms underlying 2-cell stage embryonic development.  10 and Duxbl1 11 as well as developmental pluripotencyassociated 2 (Dppa2) and Dppa4. 12,13 Furthermore, previous studies indicated that 2C-like state is suppressed by ZGA repressor, such as tripartite motif-containing protein 28 (TRIM28; also known as KAP1), 14 long interspersed nuclear element-1 (LINE1), 14 lysine (K)specific demethylase 1A (LSD1/KDM1A), 15 chromatin assembly factor-1 (CAF-1) 16 and microRNA miR-34a. 17 In addition, these 2Clike cells spontaneously transit back into the pluripotent state. 18 This showed that ESCs and 2C-like cells are in keeping with a dynamic equilibrium state.
2-cell embryo-like cells of ESCs are induced in S/L medium, but the composition of serum is more complex, which factors play a key role on 2C-like cells activation remains largely undefined. Several studies reported that retinoic acid (RA) signalling leads to increase of 2C-like cells in S/L cultured ESCs (S/L-ESCs). 11,19 RA induces ESCs transition to 2C-like state through a coordinated expression of Dux and Duxbl1 or activation of prame family member Gm12794c. 20 Notably, it is well known that naïve ESCs maintenance medium fails to induce 2C-like state of ESCs. 10 Altogether, it remains unclear whether RA has a beneficial effect on 2C-like state of ESCs in 2i/L chemically defined media.
We previously investigated a negative elongation factor A (NELFA) whose expression in mouse ESCs is coupled to 2C-like state and expanded fate potential in vivo. 10 In this study, we found that RA induces 2C-like cells in mouse ESCs that are cultured in N2B27 based 2i/L chemically defined media. Notably, we demonstrated that RA activates NELFA-mediated 2C-like state of mouse ESCs and upregulates the 2C-specific genes expression in long-term in vitro culture as well as exhibits expanded developmental potency in vivo. In addition, our results indicated that epigenetic regulations and metabolic processes play an important role in RA-induced 2C-like state.

| Mice
The mice were housed in the animal facility of Inner Mongolia

| Flow cytometry
The MERVL::tdTomato and NELFA-GFP reporter ESCs were harvested by Accutase and sorting by BD LSRFortessa. No tdTomato and GFP ESCs were used for FACS gating negative control.
Measure fluorescence (detector 568 nm and 488 nm channel for tdTomato and GFP) by flow cytometer. Gating out of residual cell debris and measure diploid and tetraploid DNA peaks. A region representing tdTomato and GFP-positive cells were used to identify living cells.

| Real-Time PCR (qPCR)
The cells were collected, total RNA was isolated by RNeasy Plus Mini Kit (Qiagen), and cDNA was obtained by Reverse Transcription System (Promega). qPCR reaction was performed using KAPA SYBR FAST qPCR kit (KAPA Biosystems) and repeated with at least three biological replicates with similar results. The expression levels were plotted relative to Gapdh. All samples were conducted using a LightCycler ® 96 Instrument (Roche Molecular Systems). Relative expression levels were determined by the 2 −ΔΔCt method. Primer sequences used are given in Table S1.

| Immunostaining
For immunofluorescence staining, the cells or pre-implantation embryos were fixed with 4% paraformaldehyde (PFA) for 30 minutes, followed by blocking in PBS containing 1% BSA and 0.1% Triton X-100 for 30 minutes and then stained overnight at 4°C with the primary antibody against anti-ZSCAN4 (Abcam, ab106646, 1:500), anti-OCT4 (BD Biosciences, 611203, 1:200) and anti-CDX2 (BioGenes, AM392, 1:200). Next, the cells were washed three times with PBS and incubated for 1 hour at room temperature with secondary antibodies. The slides were then mounted in Vectashield with DAPI (Vector Laboratories) and imaged using an Olympus FV1000 confocal microscope.

| RNA extraction, library construction and sequencing
Total RNA was extracted using TRIzol reagent kit (Invitrogen) according to the manufacturer's protocol. RNA quality was assessed on an Agilent 2100 Bioanalyzer (Agilent Technologies) and checked using RNase free agarose gel electrophoresis. After total RNA was extracted, eukaryotic mRNA was enriched by Oligo(dT) beads, while prokaryotic mRNA was enriched by removing rRNA by Ribo Zero TM Magnetic Kit (Epicentre). Then, the enriched mRNA was fragmented into short fragments using fragmentation buffer and reverse transcripted into cDNA with random primers. Second strand cDNA was synthesized by DNA polymerase I, RNase H, dNTP and buffer. Then, the cDNA fragments were purified with QiaQuick PCR extraction kit (Qiagen), end repaired, poly (A) added and ligated to Illumina sequencing adapters. The ligation products were size selected by agarose gel electrophoresis, PCR amplified and sequenced using Illumina HiSeq 2500 by Gene Denovo Biotechnology Co.

| RNA-seq data analysis
For each transcription region, a FPKM (fragment per kilobase of transcript per million mapped reads) value was calculated to quantify its expression abundance and variations, using StringTie software.
After removal of the reads that contained adapters, poly-N strands and low-quality reads, the clean data were mapped to the genome using HISAT2. The FPKM method is able to eliminate the influence of different gene lengths and sequencing data amount on the calculation of gene expression. Therefore, the calculated gene expression can be directly used for comparing the difference of gene expression among samples. StringTie software was used to calculate the expression of all genes in each sample, and DESeq2 software 21 was used to obtain the differential genes among different groups. The  The basic unit of GO is GO term. Each GO term belongs to a type of ontology. Trend analysis of DEGs was performed using Short Timeseries Expression Miner (STEM) software. 23

| Gene set enrichment analysis (GESA)
We performed gene set enrichment analysis using software GSEA and MSigDB 24 to identify whether a set of genes in specific GO terms pathways. GO terms shows significant differences in two groups. Briefly, we input gene expression matrix and rank genes by Signal-to-Noise normalization method. Enrichment scores and P value were calculated in default parameters.

| Production of chimeric embryos
6-10 mCherry-labelled NELFA-positive ESCs were injected gently into the ICR mice eight-cell stage embryos using a piezo-assisted micromanipulator attached to an inverted microscope. The injected embryos were cultured in KSOM medium (Millipore) at 37°C in a 5% CO 2 atmosphere for 48 hours and further performed immunostaining.

| Retinoic acid induces stable 2C-like state of ESCs in chemically defined condition
Previous studies showed that 2C-like subpopulation could be cul- were significantly upregulated in RA-treated 2i/L-ESCs relative to those naïve 2i/L-ESCs ( Figure 1D). Therefore, we further asked whether RA treatment could increase the number of ZSCAN4positive 2C-like cells. As expected, the number of ZSCAN4positive cells is increased in the presence of RA ( Figure 1E,F and Figure S1B). These results indicated that long-term RA treatment (p15) could maintain ESCs self-renew and increase 2C-like cells in culture.  Figure 3A,B). In

| Epigenetic modifications enhance retinoic acid-induced 2C-like state of ESCs via NELFA
Globally, 2C-like cells possess hypomethylated genome, hyperacetylated histones and increased levels of chromatin accessibility. 3,27 Similarly, experimental perturbations lead global chromatin changes, including treatment with histone deacetylase (HDAC) inhibitors, knock-down of the chromatin assembly factors (such as Caf-1, G9a, Hp1, Kap-1 or Kdm1a) all trigger 2C-like state conversion. 16,28,29 However, acute DNA demethylation is not sufficient for MERVL network activation of 2C-like cells in S/L condition. 27 Here, we sought to investigate whether the RA-induced 2C-like state is associated with the global DNA hypomethylation. For this, we supplemented a well-known DNA methyltransferase inhibitor 5-aza-2'-deoxycytidine 2-cell embryo-like cells also display high core-histone mobility, suggesting that histone mobility may be linked to greater cellular plasticity. 16,33 For example, 2C-like cells display higher levels of the 'active' marks on H3K4me2, pan-acetylated H3 and pan-acetylated H4 compared with ESCs. 3,16 Therefore, we wondered whether RA treatment can alter histone modifications landscapes that are associated with 2C-like cells. Transcriptional analysis showed that majority histone deacetylation related genes were highly expressed in 2i/L-ESCs compared with RA-2i/L-ESCs ( Figure 4E). Expectedly, 'active' histone mark H3K27 acetylation (H3K27ac) level was significantly increased in RA-2i/L-ESCs compared with 2i/L-ESCs ( Figure 4F). Thus, retinoic acid-induced 2C-like state of ESCs mainly associated with epigenetic modifications such as DNA demethylation and histone acetylation.

| Retinoic acid regulates 2C-like state by metabolic process
Several studies have documented metabolic differences between 2C-like cells and ESCs, and specifically, 2C-like cells exhibit decreased glycolytic, respiratory activity and lower levels of reactive oxygen species. 10,34,35 Here, we wondered if the metabolic features of 2C-like cells in RA treatment might be altered. To uncover the role F I G U R E 2 Retinoic acid drives NELFA-mediated 2C-like state in ESCs. A, Experimental outline of the NELFA-GFP reporter ESCs were treated with RA. B, NELFA-GFP reporter ESCs were cultured in 2i/L and 2i/L+RA medium. 2i/L+RA condition was significantly increasing the NELFA-GFP-positive cell number. Scale bars, 100 μm. C, FACS quantification of the proportion of GFP-positive (GFP + ) cells upon activation by RA. D, qPCR of 2C-specific genes Nelfa,Zscan4c, Tcstv3, Mervl, Dppa2 and Dppa4 expression level in 2i/L and 2i/L+RA cultured NELFA-GFP reporter ESCs. Error bars are mean ± SD (n = 4). P values were calculated by two-tailed Student's t test, P < .05. E, Western blotting analysis for ZSCAN4 in 2i/L and 2i/L+RA cultured NELFA-GFP reporter ESCs and the band intensity of western blotting was quantified in control. F, Immunostaining of ZSCAN4 and OCT4 in 2i/L, 2i/L+RA cultured NELFA-GFP reporter ESCs. Scale bars, 50 μm. G, Quantification of the ZSCAN4-positive cells in 2i/L and 2i/L+RA cultured ESCs. Error bars are mean ± SD (n = 30). P values were calculated by two-tailed Student's t test, P < .05. H, mCherry-labelled NELFA-positive cells were injected into 8-cell stage embryos, which were then cultured 48 h in vitro. mCherry-labelled NELFA-positive RA-2i/L-ESCs were able to contribute to both TE or ICM, whereas 2i/L-ESCs were only contribute to ICM. Scale bars, 50 μm  Figure 5A and Figure S3A). In particular, we note that the glycolysis pathway-related genes (such as Hk1, Hk2, Pfkm and Pfkp) were significantly downregulated in RA-2i/L-ESCs ( Figure 5B). In addition, inhibition of glycolysis in S/2i/L condition can induce 2C-like state, but not in N2B27-based medium ( Figure S1A). This is congruous with our previous study where treatment with 2-deoxy-D-glucose (2-DG), a glycolysis inhibitor, significantly increases NELFA-positive cells in S/2i/L condition ( Figure S3B). 10 Our results suggest that RA regulates 2C-like cells through suppressing glycose metabolism in N2B27-based medium and partially substitute the function of serum and induce NELFA.
It has been reported that there are high protein turnover rates in MERVL + Zscan4 + cells. 3,16 We thus wondered if the inhibition of protein synthesis can induce 2C-like cells in vitro culture. We found that supplementing cycloheximide (CHX), protein synthesis inhibitor, 2i/L-ESCs and RA-2i/L-ESCs failed to activate 2Clike cells ( Figure 5C). This suggests that general protein depletion in MERVL + Zscan4 + cells may occurs after the 2C activation. In

| D ISCUSS I ON
Retinoic acid is member of the nuclear receptor superfamily that is discovered in 1987 as receptor for vitamin A metabolite. 41,42 The action of RA hinges on nuclear receptors, a family of ligand-regulated transcription factors that control a wide range of developmental processes. Recently, RA-dependent molecular mechanism for the maintenance and differentiation of ESCs and especially, for entry into 2C-like state has been reported. 11,19,20,43,44 2C-like transition is induced by serum containing media or shorttime culture in supplement with RA, but the defined function is remaining unclear. Here, to fill in this knowledge gap, we screened  Figure S3E). Notably, the level of 'active' mark H3K27 acetylation in ESCs are increased after RA and RA+5aza ( Figure 4F), and this is consistent with previous report. 20 Here, we also showed that reducing the expressions of Dnmt1, Dnmt3l and Uhrf1 in RA-2i/L-ESCs, which also in line with previous studies. 32 They indicated that Dnmt1 serves as a barrier mainly during 2C genes activation and the MERVL-LTR transcriptional network is not significantly upregulated in DNMT triple-knockout (TKO) ESCs. 27,32,45 However, they used S/L culture condition, it is difficult to determine which factors in serum play an important role in 2C-like cells regulation. In our study, retinoic acid activates 2C-like state in 2i/L medium to provide a new way to study 2C-like transition in ESCs.
In addition, except for transcriptome and epigenome-related features, 2C-like cells exhibit some distinctive metabolic features, and they display reduced oxygen consumption rates than do ESCs, which similar to early-stage embryos also display lower oxygen consumption compared with blastocysts. 34 In our study, RA-2i/L-ESCs have low glycolysis gluconeogenesis and low oxidative phosphorylation level and similar to our previous study and other reports. 10,34 Notably, in chemical-defined condition, 2-DG could not induce NELFA expression, which different with S/L condition. 10 We hypothesis 2i/L medium supplemented with RA provide 2C-like gene expression stable and plateau in ESCs. In conclusion, our study not only reveals retinoic acid induces ESCs to 2C-like transition by NELFA in defined culture condition, but also demonstrates that retinoic acid mediates 2C-like state by epigenetic and metabolic regulation. F I G U R E 6 Retinoic acid induces NELFA-mediated 2C-like state of ESCs in chemically defined media. 2C-like cells arise spontaneously in ESCs upon serum and leukaemia inhibitory factor (LIF) containing culture condition and some genetic modified ESCs, and therefore, 2C-like cells refer to totipotent. Intriguingly, naïve ESCs culture medium fail to induce 2C-like state of ESCs. In this study, we report a robust in vitro induction system, based on retinoic acid (RA) containing chemically defined media, which can efficiently increase 2C-like cells population in culture without any genetic modification and serum