Pdx1 Is Transcriptionally Regulated by EGR-1 during Nitric Oxide-Induced Endoderm Differentiation of Mouse Embryonic Stem Cells

The transcription factor, early growth response-1 (EGR-1), is involved in the regulation of cell differentiation, proliferation, and apoptosis in response to different stimuli. EGR-1 is described to be involved in pancreatic endoderm differentiation, but the regulatory mechanisms controlling its action are not fully elucidated. Our previous investigation reported that exposure of mouse embryonic stem cells (mESCs) to the chemical nitric oxide (NO) donor diethylenetriamine nitric oxide adduct (DETA-NO) induces the expression of early differentiation genes such as pancreatic and duodenal homeobox 1 (Pdx1). We have also evidenced that Pdx1 expression is associated with the release of polycomb repressive complex 2 (PRC2) and P300 from the Pdx1 promoter; these events were accompanied by epigenetic changes to histones and site-specific changes in the DNA methylation. Here, we investigate the role of EGR-1 on Pdx1 regulation in mESCs. This study reveals that EGR-1 plays a negative role in Pdx1 expression and shows that the binding capacity of EGR-1 to the Pdx1 promoter depends on the methylation level of its DNA binding site and its acetylation state. These results suggest that targeting EGR-1 at early differentiation stages might be relevant for directing pluripotent cells into Pdx1-dependent cell lineages.


Introduction
Early growth response-1 (EGR-1) is a zinc-finger transcription factor that regulates the expression of numerous differentiation and growth genes in response to environmental signals. EGR-1 expression is rapidly induced through stimulation with many environmental signals including growth factors, hormones, neurotransmitters, cellular stress, injury, mitogens, and cytokines [1,2]. EGR-1 activity is modulated in part through interaction with its co-repressors NGFI-A binding protein 1 (NAB1) and NGFI-A binding protein 2 (NAB2). It is known that EGR-1 induces the expression of Nab2, thereby preventing permanent transactivation of EGR-1 target genes [3].
EGR-1 is reported to be involved in a multitude of signaling pathways, standing out in the response to oxidative stress and apoptosis [4] and in reactive oxygen species (ROS)mediated signaling and inflammation [5][6][7][8][9][10][11]. Additionally, in several types of tumor cells, EGR-1 exhibits suppressor gene activity [12][13][14]. Furthermore, EGR-1 is known to play a key role in cell differentiation. Thus, it has been described that EGR-1 can promote differentiation of bovine skeletal muscle-derived satellite cells [15,16] and osteogenic differentiation in human ligament periodontal stem cells [17]. It has also been described that EGR-1 expression favors white adipocyte differentiation and represses white adipose tissue browning [18].
Interestingly, EGR-1 has recently been reported as a critical factor to promote β cell identity, maintain β cell characteristics, and repress non-β cell programs [19]. This study describes how the loss of EGR-1 uncouples metabolic stress from the transcriptional cascade essential for the β cell compensatory response, which underscores EGR-1 as a critical factor in the pathogenesis of pancreatic islet failure. In fact, it was previously reported that EGR-1 contributes to the glucose responsiveness and function of pancreatic β cells through the regulation of insulin and pancreatic and duodenal homeobox 1 (Pdx1) expression [20,21]. These studies revealed the mechanisms by which EGR-1 regulates the expression of Pdx1 in pancreatic β cells. They identified EGR-1 recognition sites within the Pdx1 promoter (regulatory areas referred as III and IV by Gerrish et al. [22]) that contribute to its responsiveness to regulation by EGR-1. It has been proposed that the suppression of Egr-1 expression may be an important factor for the differentiation of embryonic stem cells (ESCs) into pancreatic cells [23]. In the same way, it has recently been reported that the inhibition of Egr-1 in the early phase induces the differentiation of pluripotent stem cells (PSCs) into pancreatic endoderm and insulin-producing cells; in contrast, the inhibition of Egr-1 in the late phase suppresses the differentiation process [24]. These results suggest that Egr-1 may play a dual role in β cell development, playing a repressive role in early development and then engaging in positive activity at a later stage. However, the mechanisms underlying Egr-1 and β cell development at early stages are not fully elucidated.
We previously reported that exposure to high concentrations of the chemical nitric oxide (NO) donor diethylenetriamine nitric oxide adduct (DETA-NO) causes ESC differentiation by downregulating pluripotency genes Nanog and Oct4, and upregulating differentiation genes such as Pdx1 [25]. We later elucidated the repressor role of polycomb repressor complex 2 (PRC2) and histone acetyltransferase P300 (P300) on Pdx1 gene expression. Additionally, epigenetic study of the Pdx1 promoter reveals that NO induces changes in the methylation pattern and leads to a shift in the balance of H3K27me3 and H3K4me3 occupancy [26]. In the present study, we investigate the involvement of EGR-1 in the regulation of Pdx1 expression in mouse ESCs (mESCs) after exposure to DETA-NO. The results show that EGR-1 represses Pdx1 gene expression in mESCs. Mechanistically, we propose that DETA-NO modifies the methylation degree of the EGR-1 binding site to the Pdx1 promoter and EGR-1 acetylation state, and consequently, reduces the EGR-1 binding capacity to the Pdx1 promoter, thus allowing Pdx1 expression.

EGR-1 Binds to Pdx1 Promoter in mESCs
We previously described how Pdx1 expression is significantly increased after a high dose of DETA-NO exposure in mESCs. Now, we aimed to investigate the role of EGR-1 in Pdx1 expression at early β cell development. It has previously been described that EGR-1 is bound to regulatory areas III and IV, located at around 650 and 6000 base pairs (bp), respectively, upstream of the translation starting site of the Pdx1 promoter in mouse and rat β cells (MIN6 and INS-1 cell lines, respectively). In this study, we used the JASPAR database [27] to elucidate the EGR-1 consensus binding sites in the Pdx1 promoter. Analysis of 2000 bp of Pdx1 promoter, upstream of its translation starting site, was carried out, and transcription factors with relevant roles in the control of pluripotency core and endodermal transcription factors were identified. Interestingly, the JASPAR database predicted with high confidence scores (total score 10.403 and relative score 0.8688), as not described before, as far as we know, an EGR-1 consensus binding site at 1064 to 1054 bp upstream of the translation starting site of the Pdx1 promoter (TGCGGGGGGGG) ( Figure 1A). Thus, we aimed to investigate if EGR-1 could bind to this novel site in mESCs and insulin-producing cells (INS-1E). We first verified that EGR-1 mRNA and protein levels do not vary in NO-exposed mESCs ( Figure 1B,D,E), which indicates that the potential role of EGR-1 in mESC differentiation is not due to its differential expression after NO exposure. We also confirmed that cells cultured in the presence of leukemia inhibitory factor (LIF) (control condition) and treated with NO showed a significant increase in Pdx1 expression, as reported in our previous studies [25,26] (Figure 1C,D,E). After these validations, chromatin immunoprecipitation (ChIP) studies were undertaken to study if EGR-1 binds to the Pdx1 promoter at the novel position proposed by the JASPAR database. The results showed that EGR-1 binds to the Pdx1 promoter in a control condition, and it is released after NO treatment ( Figure 1F). To test, if EGR-1 also binds to the Pdx1 promoter in the reported regulatory area III of insulin-producing cells (MIN6 and INS-1 cell lines), we performed ChIP assays in this same area in mESCs. The results showed that EGR-1 also binds to regulatory area III in mESCs, but no significant changes in EGR-1 binding to the Pdx1 promoter after DETA-NO treatment were observed ( Figure S1A). As a control, the occupation of EGR-1 over the Pdx1 promoter in INS-1E cells was performed, showing that EGR-1 is bound to regulatory area III of the Pdx1 promoter in a functional β cell line ( Figure S1B).

EGR-1 Has a Repressor Role on Pdx1 Expression in mESCs
To elucidate the action of EGR-1 on Pdx1 expression at an early stage of β cell development, EGR-1 loss and gain of function experiments were conducted in mESCs cultured under pluripotent conditions and DETA-NO treatment. The EGR-1 loss of function study was carried out by transfecting mESCs with a pool of short hairpin RNA (shRNA) targeting EGR-1 (shEGR1). This assay led to a decrease in EGR-1 expression of around 40% (Figure 2A), which allowed a significant increase in Pdx1 expression ( Figure 2B). This result suggests that EGR-1 acts as a repressor of Pdx1 expression in an early stage of differentiation of mESCs. Conversely, EGR-1 overexpression ( Figure 2C,D) did not affect Pdx1 expression, neither under control nor DETA-NO conditions ( Figure 2E,F). Thus, the repressor role of EGR-1 in Pdx1 expression was not observed in the former assay. This result suggests that other mechanisms are involved in Pdx1 regulation by EGR-1 in NO-exposed cells. Result shows the means ± SD of three independent experiments. Y-axis corresponds to the percentage input relativized to IgG binding. Data with ** p < 0.01 or *** p < 0.001 were considered statistically significant.

EGR-1 has a Repressor Role on Pdx1 Expression in mESCs
To elucidate the action of EGR-1 on Pdx1 expression at an early stage of β cell development, EGR-1 loss and gain of function experiments were conducted in mESCs cultured under pluripotent conditions and DETA-NO treatment. The EGR-1 loss of function study was carried out by transfecting mESCs with a pool of short hairpin RNA (shRNA) targeting EGR-1 (shEGR1). This assay led to a decrease in EGR-1 expression of around 40% (Figure 2A), which allowed a significant increase in Pdx1 expression ( Figure 2B). This re- Figure 1. EGR-1 binds to Pdx1 promoter in mESCs. (A) Pdx1 promoter scheme shows CpG sites (vertical gray lines), CpG islands (horizontal gray rectangles), transcription start site (TSS) and translation start site (ATG) represented by vertical gray rectangles, and EGR-1 consensus sequence (TGCGGGGGGGG) on Pdx1 according to the JASPAR database and regulatory area III (Area III). Analysis of (B) Egr-1 and (C) Pdx1 expression after DETA-NO treatment by real time-PCR. These values were normalized to expression values of β-actin, used as the loading control and analyzed using the ∆∆Ct algorithm. They represent the average of five independent experiments. Data are means ± standard deviation (SD). (D) PDX1 and EGR-1 expression were analyzed by Western blotting in control and DETA-NO conditions. β-actin was used as loading control. The image shown is the most representative of three independent experiments. (E) Western blotting quantification of three independent experiments of PDX1 and EGR-1 proteins, relativized to β-actin expression, using ImageJ software. (F) Chromatin immunoprecipitation (ChIP) assays of EGR-1 on Pdx1 promoter at JASPAR database-predicted location in mESCs. Result shows the means ± SD of three independent experiments. Y-axis corresponds to the percentage input relativized to IgG binding. Data with ** p < 0.01 or *** p < 0.001 were considered statistically significant.
sult suggests that EGR-1 acts as a repressor of Pdx1 expression in an early stage of differentiation of mESCs. Conversely, EGR-1 overexpression ( Figure 2C,D) did not affect Pdx1 expression, neither under control nor DETA-NO conditions ( Figure 2E,F). Thus, the repressor role of EGR-1 in Pdx1 expression was not observed in the former assay. This result suggests that other mechanisms are involved in Pdx1 regulation by EGR-1 in NO-exposed cells.  . Data were normalized to the expression values of β-actin, used as the loading control and analyzed using the ∆∆Ct algorithm. Bar graphs represent the means ± SD of four independent experiments. Data with * p < 0.05 or ** p < 0.01 were considered statistically significant.

EGR-1 Binding to Pdx1 Promoter Is Dependent on DNA Methylation and Its Acetylation Status
Since binding of EGR-1 to its consensus sequences is reported to be affected by DNA methylation [28], and mESCs exposed to DETA-NO showed changes in DNA methyltransferases' (Dnmts') expression at RNA levels ( Figure S2), we aimed to elucidate if Pdx1 promoter methylation is affected by DETA-NO exposition in mESCs. Thus, Pdx1 promoter methylation was analyzed by bisulfite sequencing PCR (BSP) and pyrosequencing. To this end, we studied, by BSP, the methylation levels of 11 CpG sites located around 1000 bp upstream of the translation starting site of the Pdx1 promoter. BSP results showed that exposure to DETA-NO increased the methylation level of the CpG site of the EGR-1 binding site (CpG site referred to as 7), from 60% in cells cultured under pluripotent conditions to 90% after NO treatment ( Figure 3A and Figure S3). Moreover, pyrosequencing of bisulfiteconverted DNA was carried out to validate this result. It was confirmed that NO exposure increases methylation of the EGR-1 binding site and neighboring CpG sites. Specifically, bisulfite pyrosequencing results showed that DNA methylation of CpG site 7 increased from 59.37% to 80.88% after DETA-NO treatment ( Figure 3B). Thus, it is proposed that the increase of DNA methylation level of the EGR-1 binding site CpG may interfere with the EGR-1 binding to the Pdx1 promoter after NO treatment.
Furthermore, the repressor role of EGR-1 has been described to be modulated by P300/CBP-dependent acetylation [29,30]. Moreover, our previous study revealed that the acetyltransferase P300 is bound to the Pdx1 promoter in mESCs cultured in control condition, and DETA-NO treatment decreases this binding. In addition, we demonstrated that treatment with a p300 inhibitor (P300i) induces significant Pdx1 expression [26]. Thus, we set out to identify if EGR-1 could be implicated in this mechanism. For this purpose, cells were treated with a p300 inhibitor (P300i) and valproic acid (VPA), a histone deacetylase (HDAC) inhibitor, alone or in combination with DETA-NO, and the Egr1 and Pdx1 expression levels were tested. The results showed that the addition of p300i significantly reduces Egr1 expression (p = 0.0023), while it significantly induces Pdx1 (p = 0.0023). The combination of DETA-NO and p300i exhibited a cooperative effect on Pdx1 expression (p = 0.0363 compared to DETA-NO condition). Conversely, VPA treatment significantly decreased Pdx1 expression (p = 0.0021), but no expression changes were observed in cells exposed to DETA-NO ( Figure 3C,D).
Finally, to elucidate if these changes in Pdx1 expression were correlated with EGR-1 binding to its promoter, we analyzed the EGR-1 occupation on the Pdx1 promoter by ChIP assays. The results showed that treatment with P300i tends to release EGR-1 from the Pdx1 promoter, while a higher occupancy was observed in cells exposed to VPA in the presence of DETA-NO, compared to DETA-NO alone, although no significant changes were observed ( Figure 3E). We also revealed that these changes are accompanied by epigenetic changes in the H3K27me3/H3K4me3 balance at the Pdx1 promoter in the EGR-1 binding site region ( Figure 3F).
These results indicate that the EGR-1 binding capacity to the Pdx1 promoter is dependent on P300 activity, and DNA methylation of its consensus binding site may affect EGR-1 occupation, but other factors might be also involved in Pdx1 regulation by the EGR-1 transcription factor (Figure 4). These results indicate that the EGR-1 binding capacity to the Pdx1 promoter is dependent on P300 activity, and DNA methylation of its consensus binding site may affect EGR-1 occupation, but other factors might be also involved in Pdx1 regulation by the EGR-1 transcription factor (Figure 4).  Question marks (?) symbolize that EGR-1/P300 interaction and the role of other proteins in regulating Pdx1 expression remain to be elucidated. Ac (acetylation); diethylenetriamine nitric oxide adduct (DETA-NO); early growth response-1 (EGR-1); pancreatic and duodenal homeobox 1 (Pdx1).

Discussion
It has recently been reported that EGR-1 plays a crucial role in promoting β cell identity, maintaining β cell characteristics, and repressing non-β cell programs [19]. Moreover, EGR-1 is known to contribute to the glucose responsiveness and function of pancreatic β cells through the regulation of insulin and Pdx1 expression [20,21]. Pdx1 is a transcription factor essential for pancreatic development and β cell maturation, and appropriate regulation of its expression is instrumental to the generation of insulin-producing cells from PSCs. We previously defined that exposure to high concentrations of DETA-NO causes mESCs' differentiation by downregulating pluripotency genes such as Nanog and Oct4, and upregulating differentiation genes including Pdx1 [25]. We also described some of the mechanisms involved in Pdx1 upregulation after DETA-NO treatment. Specifically, we revealed that PRC2 and P300 negatively regulate Pdx1 gene expression in mESCs cultured under pluripotent conditions, and we observed that NO induces changes in the methylation pattern of the Pdx1 promoter and leads to a shift in the balance of H3K27me3 and H3K4me3 occupancy at the proximal CpG island [26]. Thereafter, the present study was focused on elucidating the mechanisms by which EGR-1 regulates the expression of Pdx1 during endoderm differentiation triggered by exposure to DETA-NO in mESCs.
The studies carried out by Eto et al. reported that EGR-1 binds to conserved regulatory area III [21], defined by Gerrish et al. [22] of the Pdx1 promoter in the MIN6 cell line, describing EGR-1 as a positive regulator of this gene. Our results are in agreement with this finding as we reported EGR-1 occupancy in regulatory area III in INS-1E cells. Nevertheless, the present study has also identified a novel EGR-1 binding site located at −1064 to −1054 bp upstream of the translation starting site of the Pdx1 promoter in mESCs. We have shown by EGR-1 loss and gain function assays that EGR-1 plays a repressor role in Pdx1 expression during NO-induced differentiation of mESCs. In the same context, it has been reported that EGR-1 inhibition may be relevant for the differentiation of PSCs into insulin-producing cells [23]. In addition, Tsugata et al. recently described how the inhibition of EGR-1 in an early phase induces the differentiation of human PSCs into pancreatic endoderm and insulin-producing cells, but the inhibition of Egr-1 in a late phase suppresses the differentiation process [24]. Altogether, these results seem to indicate that EGR-1 could play a different role with the Pdx1 promoter depending on the differentiation state. Figure 4. Overview of Pdx1 regulation by EGR-1 and P300 in mESCs. This graph represents the Pdx1 promoter in control and DETA-NO conditions. Control condition: mESCs cultured in medium supplemented with LIF. DETA-NO condition: mESCs cultured in medium supplemented with LIF and exposed to 500 µM DETA-NO for 19 h. I, II, III, and IV represent the conserved regulatory areas in the Pdx1 promoter. The circles represent the CpG site of the EGR-1 binding site studied on the Pdx1 promoter. Methylation grade is shown by the circle color (white and black: 40-60%; black: 80-100%). Question marks (?) symbolize that EGR-1/P300 interaction and the role of other proteins in regulating Pdx1 expression remain to be elucidated. Ac (acetylation); diethylenetriamine nitric oxide adduct (DETA-NO); early growth response-1 (EGR-1); pancreatic and duodenal homeobox 1 (Pdx1).

Discussion
It has recently been reported that EGR-1 plays a crucial role in promoting β cell identity, maintaining β cell characteristics, and repressing non-β cell programs [19]. Moreover, EGR-1 is known to contribute to the glucose responsiveness and function of pancreatic β cells through the regulation of insulin and Pdx1 expression [20,21]. Pdx1 is a transcription factor essential for pancreatic development and β cell maturation, and appropriate regulation of its expression is instrumental to the generation of insulin-producing cells from PSCs. We previously defined that exposure to high concentrations of DETA-NO causes mESCs' differentiation by downregulating pluripotency genes such as Nanog and Oct4, and upregulating differentiation genes including Pdx1 [25]. We also described some of the mechanisms involved in Pdx1 upregulation after DETA-NO treatment. Specifically, we revealed that PRC2 and P300 negatively regulate Pdx1 gene expression in mESCs cultured under pluripotent conditions, and we observed that NO induces changes in the methylation pattern of the Pdx1 promoter and leads to a shift in the balance of H3K27me3 and H3K4me3 occupancy at the proximal CpG island [26]. Thereafter, the present study was focused on elucidating the mechanisms by which EGR-1 regulates the expression of Pdx1 during endoderm differentiation triggered by exposure to DETA-NO in mESCs.
The studies carried out by Eto et al. reported that EGR-1 binds to conserved regulatory area III [21], defined by Gerrish et al. [22] of the Pdx1 promoter in the MIN6 cell line, describing EGR-1 as a positive regulator of this gene. Our results are in agreement with this finding as we reported EGR-1 occupancy in regulatory area III in INS-1E cells. Nevertheless, the present study has also identified a novel EGR-1 binding site located at −1064 to −1054 bp upstream of the translation starting site of the Pdx1 promoter in mESCs. We have shown by EGR-1 loss and gain function assays that EGR-1 plays a repressor role in Pdx1 expression during NO-induced differentiation of mESCs. In the same context, it has been reported that EGR-1 inhibition may be relevant for the differentiation of PSCs into insulin-producing cells [23]. In addition, Tsugata et al. recently described how the inhibition of EGR-1 in an early phase induces the differentiation of human PSCs into pancreatic endoderm and insulinproducing cells, but the inhibition of Egr-1 in a late phase suppresses the differentiation process [24]. Altogether, these results seem to indicate that EGR-1 could play a different role with the Pdx1 promoter depending on the differentiation state.
This report also revealed some of the mechanisms by which EGR-1 regulates Pdx1 in mESCs. We showed that EGR-1 occupies the Pdx1 promoter in mESCs cultured under conditions that preserve pluripotency, and it releases the promoter after NO exposure. Our results suggest that these changes to EGR-1 binding on the Pdx1 promoter are dependent on the level of CpG methylation status of the EGR-1 consensus binding site. The methylation studies shown here indicate that NO modifies the methylation pattern of the Pdx1 promoter, specifically increasing the methylation of the EGR-1 binding site and neighboring CpG sites. These results are in agreement with our previous findings. We observed a significant increase in the methylation level of the proximal and distal CpG islands of the Pdx1 promoter after NO treatment [26]. Moreover, several studies have previously reported that the methylation state of the EGR-1 binding site affects its ability to bind to DNA [31][32][33]. Additionally, the finding that pharmacological inhibition of the histone P300 acetyltransferase activity led to enhanced expression of Pdx1 suggests that EGR-1 acetylation by P300 could play a role in this regulatory mechanism. Besides this, VPA exposure decreased Pdx1 expression in PSCs. These results are consistent with our previous report showing that P300 is downregulated after DETA-NO exposure in mESCs, while Pdx1 is negatively regulated by P300 [26]. Furthermore, cells exposed to p300i and VPA treatments showed slight changes in EGR-1 occupancy in the Pdx1 promoter, which reinforces our proposal that the EGR-1 acetylation status may regulate its DNA binding capacity. However, the EGR-1/P300 interaction and EGR-1 acetylation status remain to be elucidated. These results also agree with the study by Liu et al., who described how P300 and other proteins are essential for EGR-1 activation and cell proliferation and survival [32].
Some studies have previously demonstrated a dual action of EGR-1. Yu et al. reported that EGR-1 binds directly to p300 regulatory sequences, and they form an acetylated EGR-1/P300/CBP complex activating growth and survival. They demonstrated that after an exogenous stimulus, EGR-1 is deacetylated and the p300/CBP promoter is repressed, leading to apoptosis [30]. Furthermore, EGR-1 has been described to act as a transcriptional repressor in melanoma cells and as an activator in breast, colon, and prostate cancer cells of the heparanase gene [31]. Our results also suggest that EGR-1 plays a dual role in Pdx1 regulation during β cell development, playing a repressive role at an early stage and engaging in positive transcriptional activity at a later stage.
In conclusion, this study reveals a novel EGR-1 binding site on the Pdx1 promoter and identifies EGR-1 as a repressor transcription factor of the Pdx1 gene in mESCs. We have defined how EGR-1 occupancy in the Pdx1 promoter depends on the methylation level of the EGR-1 binding site and histone P300 acetyltransferase activity. These results provide support for EGR-1 inhibition as a relevant target for efforts to direct PSCs into functional insulin-producing cells.

RNA Isolation, Reverse Transcription, PCR, and Real-Time PCR Analysis
Total RNA was extracted using Easy Blue ® reagent (Intron Biotechnology, Gyeonggi-do, South Korea) and following the chloroform/isopropanol purification procedure. cDNA synthesis was performed with 1 µg total RNA using Moloney murine leukemia virus reverse transcriptase (M-MVL RT) (Promega, Madison, WI, USA) and random primers according to the manufacturer's instructions. For real-time PCR analysis, endogenous mRNA levels were measured based on SYBR Green (Applied Biosystems, Foster City, CA, USA) detection with an ABI Prism 7500 machine (Applied Biosystems). Results were normalized with the β-actin expression. The real-time PCR primers used are shown in Table S1.

Bisulfite Sequencing PCR (BSP)
A region of approximately 2000 base pairs (bp) of the Pdx1 promoter was analyzed with the software Methyl Primer Express v1.0. (Applied Biosystems) to identify CpGrich regions. Primers designed for these regions are listed in Table S1. Then, genomic DNA from 7.5 × 10 4 cells was converted with sodium bisulfite using a Cells-to-CpGTM Bisulphite Conversion Kit (Applied Biosystems, Waltham, MA, USA). Converted DNA was amplified by PCR using MyTaqTM HS Red DNA Polymerase, and then PCR products were purified and cloned into the pGEM-T vector to obtain Escherichia coli (E. coli) colonies. Ten colonies per condition were analyzed by PCR and later sequenced in a DNA Analyzer 3730 (Applied Biosystems). As the technical control, we treated the isolated genomic DNA of mESC cultured under a control condition with the CpG methyltransferase, M.SssI (New England BioLabs, Ipswich, MA, USA) according to the manufacturer's instructions. The methylation of the proximal CpG island of the Pdx1 promoter was studied. The results were analyzed by BiQ Analyzer Software (Max-Planck-Institute and Saarland University, Saarbrücken, Germany).

Bisulfite Pyrosequencing
Pdx1 promoter methylation status results obtained by BSP were confirmed by pyrosequencing. Sodium bisulfite modification of genomic DNA of 7.5 × 10 4 cells was carried out as described above. Converted DNA was eluted in 15 µL and 2 µL for each PCR cycle. The primers used for PCR and sequencing were designed using PyroMark assay design software, version 2.0.01.15 (Qiagen, Hilden, Germany). The pyrosequencing primers are shown in Table S1. These primers were designed to hybridize with CpG-free sites to ensure methylation-independent amplification. PCR was performed with biotinylated primers to convert the PCR product to single-stranded DNA templates, using the Vacuum Prep Tool (Biotage, Uppsala, Sweden), according to the manufacturer's instructions. Pyrosequencing reactions and methylation quantification were performed using a PyroMark Q24 System, version 2.0.6 (Qiagen).

Egr-1 Loss and Gain of Function
In loss of function experiments, R1/E cells were transfected after two days in culture with a pool of short hairpin RNA (shRNA) (Sigma-Aldrich), while in gain of function experiments, EGR-1 overexpression was achieved by transfecting R1/E cells with the plasmid pCG-HA-EGR-1, kindly provided by Dr. Shigeru Taketani. Cells were transfected using Fugene HD transfection reagent (Promega, Madison, WI, USA). In both cases, before the transfection, a Fugene:DNA complex was created with a ratio of 4:1, incubated for 1 h at RT in Opti-MEM medium (Gibco, Waltham, MA, USA). Then, the complex was added to the cells, covering the plate with Opti-MEM medium, then incubating for 3 h at 37 • C and 5% CO 2 . Finally, the plate's volume was completed with the corresponding culture medium. After leaving overnight, the medium was changed and DETA-NO was added, and after 19 h, the cells were collected. R1/E cells transfected with the shEGR1 pool were selected with 2 µg/mL of puromycin 48 h after the transfection. The degree of silencing and overexpression was assessed by real-time PCR. The primers used are listed in Table S1.

Statistical Analyses
The data represented are the means ± standard deviation (SD) of at least three independent experiments (unless specified otherwise in the figure caption). Comparisons between values were analyzed using Student's t-test with GraphPad Prism v7 (GraphPad Software, San Diego, CA, USA). p-values < 0.05 were considered statistically significant. . Along with this, there was support to PAIDI group CTS576, and by the European Regional Development Fund (FEDER) and the Consejería de Economía, Conocimiento, Empresas y Universidades de la Junta de Andalucía, within the framework of the operational program FEDER Andalucía 2014-2020. Specific Objective 1.2.3 "Promotion and generation of frontier knowledge and knowledge oriented to the challenges of society, development of emerging technologies" led the reference research project (UPO-1381598) of J.R.T.