Retinoic Acid Specifically Enhances Embryonic Stem Cell Metastate Marked by Zscan4

Pluripotency confers Embryonic Stem Cells (ESCs) the ability to differentiate in ectoderm, endoderm, and mesoderm derivatives, producing the majority of cell types. Although the majority of ESCs divide without losing pluripotency, it has become evident that ESCs culture consists of multiple cell populations with different degrees of potency that are spontaneously induced in regular ESC culture conditions. Zscan4, a key pluripotency factor, marks ESC subpopulation that is referred to as high-level of pluripotency metastate. Here, we report that in ESC cultures treated with retinoic acid (RA), Zscan4 ESCs metastate is strongly enhanced. In particular, we found that induction of Zscan4 metastate is mediated via RA receptors (RAR-alpha, RAR-beta, and RAR-gamma), and it is dependent on phosphoinositide-3-kinase (PI3K) signaling. Remarkably, Zscan4 metastate induced by RA lacks canonical pluripotency genes Oct3/4 and Nanog but retained both self-renewal and pluripotency capabilities. Finally we demonstrated that the conditional ablation of Zscan4 subpopulation is dispensable for both endoderm and mesoderm but is required for ectoderm lineage. In conclusion, our research provides new insights about the role of RA signaling during ESCs high pluripotency metastate fluctuation.


Introduction
Embryonic stem cells (ESCs) are derived from the inner cell mass (ICM) of blastocyst and are characterized by two main peculiarities, namely self-renewal and pluripotency: self-renewal is defined as the symmetrical division of ESCs into identical undifferentiated daughter cells; pluripotency confers ESCs the ability to produce the majority of cell types upon appropriate determinants. It has become evident over the past few years that ESCs fluctuate among different levels of potency as a consequence of paracrine effects and cell-to-cell interactions that are not homogeneously regulated within current in vitro culture conditions [1,2,3]. The ESCs culture heterogeneity may be revealed by the expression of genes that distinctly mark ESCs subpopulations that are commonly defined metastates. Recent evidence reported that the ESCs metastate marked by Zscan4 (zinc finger and SCAN domain containing 4) is characterized by the remarkable potential to produce both embryonic and extra-embryonic cell lineages, therefore it is referred to as totipotent ESCs metastate [4,5,6,7]. The Zscan4 metastate is activated in about 3-5% of the ESC population at any given time and it is required to prevent senescence thus improving the quality of long-term culture. The Zscan4 metastate is enriched in high pluripotency culture conditions by the coordinated actions of extrinsic regulators, signaling pathways, and transcription factors. In particular, the cytokine Leukemia Inhibitory Factor (Lif) is implicated in the establishment of the Zscan4 ESCs metastate, while Phosphoinositide-3-kinase (PI3K) signaling has been reported to regulate Zscan4 expression [8]. Although Zscan4 ESC fluctuation is stabilized by high-pluripotency culture conditions obtained through Extracellular signal Regulated-Kinase (ERK) and Glycogen synthase kinase-3 (Gsk-3) signaling inhibition (2i) [9], the Zscan4 metastate is heterogeneously marked by Gm12794 that defines a notcanonical pluripotency signature required for ESCs maintenance [10,11].
In the current study, we investigated how differentiation stimuli such as Dimethylsulfoxide (DMSO), Lif removal, and Retinoic Acid (RA) could affect Zscan4 metastate balance. Interestingly, we found that the Zscan4 metastate was negatively affected by Lif removal and by DMSO treatment, meanwhile it was significantly increased in RA culture condition. We have extended these results by evaluating RA induced Zscan4 metastate through global expression profile, and both self-renewal, and pluripotency capabilities. Our data showed that although the Zscan4 metastate retains self-renewal and pluripotency capabilities, it is also characterized by key markers of ectoderm lineage such as cadherin20 (Cdh20), and brain derived neurotrophic factor (Bdnf). Consistently, the conditional ablation of the Zscan4 subpopulation is dispensable for both endoderm and mesoderm but is required for ectoderm lineage. In our opinion, the Zscan4 metastate enhanced by RA is primed to early ectoderm differentiation and represents a suitable opportunity to characterize key molecular signaling underlying the fluctuation between pluripotency maintenance, and early specification.
Flow cytometry and sorting pZscan4-Emerald and pGm12794-Strawberry ESCs pZscan4-Emerald/pGm12794-Strawberry cells were fed at least 2 hours before harvesting by Trypsin (Gibco) and resuspended in complete ES medium containing 25 mM HEPES buffer. The cells were then FACS-sorted according to the fluorescent intensity of EMERALD or STRAWBERRY into complete ES medium containing HEPES. For qPCR analysis of FACSsorted cells, total RNAs were collected immediately after sorting by TRIzol (Invitrogen) according to the manufactures instruction.
One microgram of total RNA was reverse-transcribed by Quantitec reverse transcription kit (Qiagen) according to the manufacturer's instructions. qPCR analyses were performed using 7.5 ng cDNA per well in duplicate with the SYBR green master mix (Applied Biosystems) according to the manufacturer's instructions. Reactions were run on QuantStudio 7 Flex system (Applied Biosystems). Fold induction was calculated and normalized by the ΔΔCt method. The gene-specific primers are available in S1 Table. RNA extraction, labeling and hybridization on the NIA 22K 60-mer oligo microarray ESCs were sorted based on the presence and absence of Emerald fluorescence at culturing day 5, and mRNA was extracted from each subset using a Quickprep micro poly-A RNA Extraction Kit (Amersham Biosciences, Piscataway, NJ, USA) with linear acrylamide as a carrier (Ambion, Austin, TX, USA). The mRNA aliquots from each subset were labeled with Cy3 and Cy5 dyes by two-round linear amplification labeling reactions to make each cRNA target using a Fluorescent Linear Amplification Kit (Agilent Technologies, Palo Alto, CA, USA). A complete list of annotated gene content of the microarray can be found at the NIA mouse cDNA project web site (http://lgsun.grc.nia.nih.gov/cDNA/cDNA.html). Data are being submitted on Gene Expression Omnibus.

Microarray data analysis
Intensity values of 21939 gene features per array were extracted from scanned microarray images using Feature Extraction 5.1.1 software (Agilent Technologies), which performs background subtractions and dye normalization. Text output was processed using an application developed in-house to perform ANOVA analysis (http://lgsun.grc.nia.nih.gov/ANOVA/). Intensities measured with a>50% error were replaced with missing values except for features with very low intensity. Surrogate values equal to mean error were inserted for values that were negative or less than the probe error. Pair-wise mean comparison was done using t-statistics and the false discovery rate (FDR = 15%) method. The small number of biological replications typical in expression profiling experiments results in a highly variable error variance, and this problem is usually addressed by log-ratio thresholds that require subjective decisions about biological significance, or by Bayesian adjustment of error variance that may still underestimate error variance and result in false positive results. To reduce false-positives, we opted for a very conservative error model in which the error variance used for estimating F-statistics for a feature is the maximum of the actual error variance and the average error variance in 500 features with similar average intensity. Statistical significance was determined using the false discovery rate FDR (= 15%) method. A scatter plot was created using the NIA microarray analysis tool (http://lgsun.grc.nia.nih.gov/ANOVA/).

Cell signaling inhibition and Western Blot analysis
Specific inhibitors were used to probe the inhibition and attenuation effects on phosphorylation. ES cells were pre-treated for 90 minutes with 20 μM MEK1 inhibitor PD0325901. RA was added after 90 minutes of inhibitors for 24h. ES cells were pre-treated for 12h with 5 μM LY294002 and maintained in culture for 24h in presence of RA and LY294002. Because inhibitors were dissolved in dimethylsulfoxide (DMSO), control cells were similarly treated with DMSO.

Cell ablation strategy
To generate the pGm12794-TK vector, a DNA fragment containing the thymidine kinase (TK) coding sequence completed of a poly(A) signal was amplified with the couple of primers HpaI-TKf (5'-agcgcgtatggcttcgtacc-3') and SalI-TKr (5'-cttgataccccacgcaa cgc-3'). This fragment was inserted in the pGm12794-Strawberry digested EcoRV-SalI, replacing the Strawberry-pA sequence. All the passages of the plasmids construction were verified by sequence analysis.

Zscan4 is induced in ESC cultured with RA
We investigated how the Zscan4 metastate is affected during ESCs differentiation. For this purpose, we cultured ESCs in three differentiating culturing media: first, we used regular medium (hereafter RM) without Leukemia inhibitory factor (hereafter Lif-), a cytokine required for ESC self-renewal; the second culture medium was obtained adding 1% of Dimethylsulfoxide (hereafter DMSO), which is a chemical inducer of ESCs differentiation across all three germ layers [12]; the third culture condition was RM supplemented with Retinoic Acid at 1.5 μM (hereafter RA), which is a crucial signaling molecule during embryonic morphogenesis and ESC differentiations [13,14]. ESCs differentiation induction was evaluated based on the ability of culturing media to reduce expression of Oct3/4 that it is canonical marker of pluripotency. Oct3/4 expression was compared among the three differentiating conditions both at 36 hours (h), and at 72h through quantitative PCR (qRT-PCR) analyses. As expected Oct3/4 expression was downregulated in all culturing media (Fig 1A). The Zscan4 ESCs metastate was evaluated based on the expression of Zscan4. Compared to RM, Zscan4 expression was reduced in Lif-(about 2 folds at 36h, and about 8 folds at 72h), and in DMSO (about 74 folds at 36h and about 300 folds at 72h). The expression of Zscan4 remarkably it was upregulated in RA (about 9 folds at 36h, and about 20 folds at 72 h) ( Fig 1A). We evaluated whether Zscan4 induction reflects an increase of Zscan4 metastate percentage within ESC cultures through RNA in situ hybridization (ISH) (Fig 1B). RNA ISH results showed "spotted" patterns of Zscan4 expression on ESCs colonies that are consistent with the increase of the Zscan4 metastate. To quantitatively evaluate the Zscan4 metastate increase, we employed transgenic ESCs line (ES Zscan4_Em cells) in which the expression of Zscan4 is marked by green fluorescence of Emerald reporter (S1A Fig) enabling us to characterize RA-Zscan4 + in living cell cultures [5]. In particular, we cultured ES Zscan4_Em cells both in RM and in RA media, and compared the percentage of fluorescent cells (Zscan4 + ) through cytofluorimetry analyses. The percentage of green cells measured in ESCs culture at 72h was about 5% in RM (hereafter RM-Zscan4 + ), and it was about 14% in RA (hereafter RA-Zscan4 + ) (Fig 1C and S1 Video). Furthermore, we performed RNA ISH also on AF067063, BC061212, Eif1a, Gm12794, Gm4340, and Tcstv1 because altogether those genes define the Zscan4 metastate molecular signature [10]. The results revealed mosaic in colony expression patterns (Fig 1D) that were consistent with the Zscan4 metastate enrichment, and strengthen our conclusion that RA culture condition enhances ESC Zscan4 metastate fluctuation [6,7].

RA induction of Zscan4 metastate is RAR dependent
In order to characterize RA signaling involved in Zscan4 induction, we assayed the requirement of RA receptors (RARs). First, we antagonized the three RAR isotypes (RARα, RARβ and RARγ) by the addition of BMS493 in RA culturing medium, a compound known to antagonize the three RAR isotypes [15]. We evaluated the effects of BMS493 on the RA-Zscan4 + metastate by comparing the percentage of Zscan4 + cells in Emerald transgenic cell line cultured in RA with and without RAR antagonists. The ES Zscan4_Em cells fluorescence was measured by flow cytofluorimetry analyses on day3, and, showed that the RA-Zscan4 + metastate percentage was about 18% in RA and it was reduced to about 10% in RA with BMS493 ( Fig 2AI). These data were further confirmed by qRT-PCR analysis conducted on the genes that mark Zscan4 metastate (Fig 2AII). In particular, RA inductions of AF067063, BC061212, Eif1a, Gm12794, Gm4340, and Tcstv1 were strongly reduced by BMS493 treatment. Second, we further dissected the RA signaling requirement using RAR agonists [16]. BMS753 is a selective agonist of RARα, and UVI2060 is an agonist of both RARβ, and RARγ. In particular, we compared the Zscan4 + cell percentage among ESCs cultured in: RM, RM+BMS753, and RM+UVI2060 by cytofluorimetry analyses. We found that Zscan4 + percentage in RM (about 6%) was increased up to 15% in RM+BMS753, and up to 14% in RM+UVI2060 (Fig 2AI). In order to validate those data, we also assayed the expression of the Zscan4 metastate signature (AF067063, BC061212, Eif1a, Gm12794, Gm4340, and Tcstv1). Consistently with cytofluorimetry results, we found that both BMS753 and UVI2060 treatments induced the Zscan4 metastate gene signature (Fig 2BI and  2BII). The assay efficiency was measured based on Stra8 expression, since it was previously reported to be specifically induced by RA [17]. In conclusion, our data showed that RA-Zscan4 + induction was mediated through RARs signaling.

RA-Zscan4 activation is PI3K dependent
Recent reports show that phosphoinositide-3-kinase (PI3K) signaling, known to be crucial for self-renewal maintenance, is required to establish the Zscan4 + pluripotent metastate in RM [8,18]. Here, we investigate whether PI3Ks is also involved in RA-dependent Zscan4 + induction. ESCs were cultured for 12 hours with LY294002, an inhibitor of PI3K. We validated the effectiveness of PI3K inhibition treatment by addressing NANOG expression, which is a known self-renewal marker positively regulated by PI3K [8]. The immunoblot analysis using NANOG antibody confirmed that continuous exposure to LY294002 suppressed the expression of the pluripotency markers NANOG (Fig 3A). In the same culture conditions, we evaluated the effects of PI3K inhibition on the expressions of the Zscan4-related molecular signature through qRT-PCR. Expressions of AF067063, Eif1a, Gm12794, Gm4340, Tcstv1, and Zscan4 were dramatically dropped in ESCs cultures with RA with LY294002 ( Fig 3B). We also examined involvement of ERK signaling that, in contrast to PI3Ks, promotes ESCs differentiation. As the Mek/Erk pathway is also a regulatory of early differentiation and was first considered for self-renewal due to its constitutive activity in ESC cultures, we assessed its impact on Zscan4 expression [19,20]. In particular, to evaluate the correlation between ERKs and RA-Zscan4 + induction, we cultured ESCs in RA with and without PD0325901, a selective ERKs signaling inhibitor. The Western Blot analysis showed that RA activated ERKs by increasing phosphorylation of ERK1, and ERK2 (Fig 3C). Although ERKs activation was efficiently suppressed by PD0325901 treatment, Zscan4 expression did not change (Fig 3D). Finally, we demonstrated that PI3Ks and not ERKs signaling activation was required in RAinduced Zscan4 + activation.

RA-Zscan4 + retains high pluripotency molecular signature and selfrenewal capabilities
Since retinoic acid is a known cellular morphogen it was an intriguing question to examine whether RA-Zscan4 + subpopulation represents a differentiation state. In particular, we cultured ES Zscan4_Em cells in RA for 5 days, collected homogeneous RA-Zscan4 + and RA-Zscan4populations using fluorescence activated cell sorting (FACS), and then we compared their global gene expression profiles through DNA microarray hybridization methodology. Bioinformatic analyses revealed that RA-Zscan4 + expressed 129 genes 3-folds more than RA-Zs-can4 - (Fig 4A, S2 Table). To assess how similar the Zscan4 metastate induced in RA was to the Zscan4 metastate in RM, we compared 129 transcripts enriched in RA-Zscan4 + versus 68 genes specifically enriched in RM-Zscan4 + . Remarkably, the RA-Zscan4 + signature contained 54 out of 68 genes enriched in RM-Zscan4 + corresponding to about 80% of molecular signature (Fig 4B). These analyses suggested that RA induced subpopulation marked by Zscan4 expression was closely related to known high pluripotent ESCs metastate based on their transcription profile. The further analysis of known genes enriched only in RA-Zscan4 + and not in RM-Zscan4 + revealed that only 34 out of 75 have known gene symbols and include among others key factors that are known to be specifically required for neuronal lineage specification (Bdnf, and Cdh20) [21,22]. On the other hand among genes that are specific of RM-Zscan4 + and not revealed in RA-Zscan4 + , 9 out of 14 have known gene symbols and among these there was Tbx3 that was recently described in published work as playing a role in promoting transition of ESC toward Zscan4 + metastate [23,24]. Altogether these data prompted us to evaluate whether RA-Zscan4 + could be primed to differentiation. In particular, we homogeneously collected RA-Zscan4 + and RA-Zscan4on day 5 through FACS, then we plated in RM, and characterized both the expressions of pluripotency markers, and the morphology. First, the qRT-PCR expression analyses revealed that RA-Zscan4 + cultured in RM were enriched in pluripotency markers compared RA-Zscan4 -(S1B Fig). Second, RA-Zscan4 + plated in RM gradually restored RA-Zs-can4cell population, going from 100% on day 0 to about 12% on day 4 (Fig 4C), giving rise to ESC colonies that were expressing high level of alkaline phosphatase (AP), a known marker of colony pluripotency (Fig 4D). Altogether, these data suggested that Zscan4 + , although induced by RA, maintained pluripotency potential.

RA induced Zscan4 metastate transition toward ectoderm derivatives
The RA-Zscan4 + molecular signature included Gm12794 is a novel member of the Prame family, known to be involved in RA signaling [25,26,27,28]. To characterize the effects of RA on Gm12794, we generated a transgenic ESC line in which Zscan4 and Gm12794 promoters regulated the expressions of Emerald and Strawberry fluorescent reporters respectively (hereafter named ESC Gm12794_St / Zscan4_Em ) (Fig 5A). This ESC cell line enabled us to quantify the percentage of ESCs expressing either Gm12794 and/or Zscan4 in the same culture conditions. The percentage measures were done in ESCs cultured in RM on day3 and in RA on day3 through fluorescent cytofluorimetry. The results showed that compared to RM culture conditions, RA induced Zscan4 + /Gm12794metastate less than 3 folds (%10% in RM, and %25% in RA; n = 3), Zscan4 -/ Gm12794 + metastate was induced about 3 folds (%0.5% in RM and %1.5% in RA; n = 3), and Zscan4 + / Gm12794 + more than 30 folds (%0.3% in RM and %10% in RA; n = 3) (Fig 5B). To elucidate whether the Zscan4 + /Gm12794 + metastate increase depends on Zscan4 + metastate induction, we set up time-lapse live imaging visual inspection on ESC Gm12794_St / Zscan4_Em cell line. We observed that the Gm12794 + metastate derived from the Zscan4 + subpopulation (Fig 5C, and S1 Video). Altogether these data revealed that RA enhance the Zscan4 metastate transition to Gm12794 metastate.
We investigated whether Gm12794 metastate transition is associated to RA induced ESCs differentiation. For this purpose we selectively ablated Gm12794 + cell using herpes simplex 1 virus thymidine kinase (HSV-TK) system. Briefly, HSV-TK expression renders cells sensitive to the nucleoside analog ganciclovir (GCV). We used an ESC transgenic line in which the expression of HSV-TK was under the control of Gm12794 promoter. The ESC Gm12794_HSVTK were cultured in RA culturing media with GCV (GCV + ), and without GCV (GCV -) and on day 3, and further characterized for the expression of ectoderm (Sox1), endoderm (Sox17, and Foxa2), and mesoderm (Brachyury, and Gata2) markers. The quantitative marker's expression analyses between GCV + and to GCVrevealed that Brachyury, Foxa2, Gata2, and Sox17 were enriched, meanwhile Sox1 was strongly downregulated (Fig 5D and S1C Fig). In conclusion, Gm12794 + ESC metastate ablation hampered early ectoderm ESCs differentiation induced by RA.

Discussion
Embryonic Stem Cells (ESCs) cultured in standard conditions exhibit expression heterogeneity for several transcription factors functionally linked to both pluripotency, and differentiation state of individual cells. An emerging theme is that cellular heterogeneity represents ESC fluctuations among metastates that play an important role in self-renewal maintenance. In particular, ESCs metastate marked by Zscan4 expression define high pluripotent ESC metastate, known also as 2C-like state because it is marked by a panel of genes that are specifically expressed during 2 cell preimplantation development stage [5]. The Zscan4 metastate represents a fluctuating subpopulation that in average is about 5% of the ESC population in regular culture conditions [5], and is characterized by the potential to produce both embryonic and extra-embryonic cell derivatives [6]. In this work, we evaluated the Zscan4 metastate upon ESC differentiating treatments obtained by modifying regular culture mediums by removal of Lif, or by the addition of two morphogens such as DMSO, and RA.
Our data showed that RA culture condition is very strongly inducing the enrichment of Zscan4 metastate up to about 20% of the cell population. The RA effect on Zscan4 metastate enrichment is not dosage dependent and it occurs at low concentration (S1D Fig). In our study, we used RA specific agonists, and antagonists to investigate the roles of retinoid receptors in the induction of Zscan4 metastate. Our results revealed that induction of Zscan4 metastate by retinoids is mediated by all three RA receptors (RARs). The most common mechanism of action of these receptors involves their binding to RA response elements (RARE) in the promoters of retinoid-responsive genes. However, we found that the promoters of the genes that characterize the global expression profile of RA-Zscan4 metastate did not present RARE (data not shown), and therefore may be induced either by RARE-independent mechanism or as response to early RARE-dependent genes.
The global gene expression comparison between RM-Zscan4 and RA-Zscan4 revealed significant transcriptome overlap (Fig 4B) that could be explained suggesting a common origin. Consistently with this hypothesis, the cell live imaging analysis showed that upon the addition of RA, RM-Zscan4 cells were induced to active division. Based on genes that are specifically enriched either in RM-Zscan4 or in RA-Zscan4, we summarized a cellular transition model in Fig 6. Briefly, Tbx3, known to be a direct inducer of the Zscan4 is more expressed in RM-Zscan4 metastate than in the RA-Zscan4; while Gm12794 a member of the Prame family [27], Bdnf and Cdh20 that are key markers of ectoderm lineage [23,24] are among the genes significantly enriched in RA-Zscan4. Altogether these data suggest that RA-Zscan4 metastate is primed to differentiate toward ectoderm lineage.
This hypothesis is also further supported by ablation cell lineage assessments. In particular, our data revealed that removing the RA-Zscan4 population during RA differentiation specifically hampers ectoderm, but not endoderm or mesoderm derivatives (S1C Fig). Remarkably, RA-Zscan4 homogeneously isolated population once plated in regular medium revert to ESC that express canonical markers of pluripotency, and retained self-renewal capability by giving rise to colonies positive for the pluripotency marker AP. Altogether, we conclude that although the Zscan4 metastate enhanced by RA is primed to ectoderm lineage it is not fully differentiated. In our opinion, this research provides insights to extend the molecular