Deletion of activating transcription factor 3 (Atf3) promotes cardiomyocyte differentiation from embryonic stem cells

Abstract

Myocardial infarction is associated with massive cardiomyocyte loss, and embryonic stem cells (ESCs), owing to their cardiomyocyte differentiation ability, have emerged as a promising therapy. One potential factor involved in regulating ESC-to-cardiomyocyte differentiation is activating transcription factor 3 (Atf3), and this study clarifies its involvement. Atf3 was knocked out (ko) in mouse ESCs (mESCs) with CRISPR/Cas9. Both wild-type (WT) and Atf3-KO mESCs formed embryoid bodies (EBs) over 6 days, followed by adherent culture to induce cardiomyocyte differentiation from days 7–12; there, Atf3-KO had more beating EBs, at higher frequencies, from Day 8 of differentiation than WT. They also had increased expression of mesodermal markers on Days 3–6 (ex. T), cardiac progenitor markers on Days 6–9 (ex. Pdgfra), and cardiomyocyte differentiation marker on Days 9–12 (ex. cardiac troponin T), as measured by RT-qPCR and flow cytometry. Furthermore, cardiomyocyte differentiation-associated differentially-expressed genes were significantly upregulated in Atf3-KO EBs, compared to WT, under Gene Set Enrichment Analysis of RNA sequencing. All these effects in Atf3-KO mESCs were reversed upon “rescue”, where Atf3 expression was restored in these cells. Therefore, Atf3-KO in mESCs promotes differentiation into mesodermal lineages, which further differentiate into cardiac progenitors, serving as a target for cell-based cardiac regeneration therapies.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1038/s41598-026-58837-9.

Introduction

Adult cardiomyocytes are terminally differentiated cells; thus, they are unable to regenerate themselves once damaged¹. As a result, the most effective treatment for end-stage heart failure (HF), such as those caused by myocardial infarction (MI), is still cardiac transplantation. However, the current donor shortage is a significant impediment from meeting clinical needs, necessitating the development of effective alternative treatment approaches². One potential approach involves embryonic stem cells (ESCs), which, owing to their pluripotency, have often attracted widespread interest for potential myocardial stem cell therapies. In fact, a previous study demonstrated that ESC-derived cardiomyocytes were similar, both morphologically and functionally, to adult myocardial cells, resulting in ESCs being considered the most optimal choice for myocardial regeneration therapies³. Despite the numerous studies conducted on those ESC-derived cardiomyocytes, as well as the development of multiple myocardial differentiation procedures, overall efficiency for ESC differentiation into myocardial cells is relatively low, with uneven cell quality⁴. Consequently, obtaining a homogeneous mature cardiomyocyte population from ESCs is of great interest for clinical applications, which could potentially be facilitated by identifying the key regulatory factors involved in promoting ESC differentiation into mesodermal cells, as the mammalian heart has been noted to develop from the mesodermal germ layer⁵. One such regulatory factor is the transcription factor activating protein (AP)-1, which plays critical roles in cell growth and differentiation⁶, as shown by Beisaw et al., in which it was able to promote sarcomere disassembly and myocardial cell proliferation during zebrafish cardiac regeneration⁷. AP-1 comprises a highly conserved homo- or heterodimer, formed from c-Jun, c-Fos, Jun dimerization protein (JDP), and activating transcription factor (ATF)⁶.

In particular, Atf3, an important component of AP-1, has been shown to play a variety of roles in embryonic development and pathogenesis⁸. It is normally expressed at low levels in resting cardiomyocytes, but its expression significantly increases when the heart is subjected to stress stimulation, in turn contributing to subsequent cardiomyocyte changes. For instance, Wang et al. demonstrated that overexpressing Atf3 in mouse disease models could result in myocardial fibrosis, ventricular hypertrophy, and HF⁹, illustrating that Atf3 expression was closely related to cardiac pathophysiological changes. Based on those observations, we postulated that Atf3 could serve as a potential therapeutic target to prevent adverse cardiac remodeling. However, its exact role in cardiac development and disease pathogenesis is still largely unknown. As a result, in-depth, systematic analyses on the role of Atf3 in cardiac development and differentiation are necessary. This study aimed to provide further clarity on Atf3 involvement in cardiomyocyte differentiation, in which Atf3 was knocked out in mouse ESCs (mESCs) with CRISPR/Cas9¹⁰. Embryoid bodies (EBs) were first formed with the hanging drop method for up to 6 days, followed by adherent cell culturing, with continued differentiation, for Days 7–12. Large numbers of beating cells, expressing mesodermal lineage, cardiac progenitor, as well as cardiomyocyte markers, such as actin-α cardiac muscle 1 (Actc1), α-actinin 2 (Actn2), and cardiac troponin T (cTnT), were found in Atf3-KO EBs on Day 9. These cells were also enriched for differentially-expressed genes (DEGs) associated with cardiomyocyte differentiation under RNA-sequencing (RNA-seq). By contrast, wild-type (WT) EBs, even beyond Day 9, only contained weak-beating cells, with lower expression of cardiomyocyte differentiation-associated genes. Similar findings to that of WT, for cardiomyocyte marker gene and cTnT protein expression, were also observed in Atf3-KO mESCs where Atf3 expression was “rescued”. Therefore, Atf3 serves as a negative regulator against ESC differentiation into cardiomyocyte differentiation in its early stages, as its silencing leads to the promotion of ESC differentiation into mesodermal lineages, and subsequently cardiomyocytes.

Materials and methods

mESC culture and Atf3-KO

mESCs (AisenGene Bioscience) were cultured at 37 °C in 5% CO₂, in high-glucose Dulbecco’s Modified Eagle Medium (DMEM; Hyclone), containing 15% fetal bovine serum (FBS), 100× penicillin-streptomycin (BasalMedia), 1 mM sodium pyruvate (GIBCO), 1 mM non-essential amino acids (GIBCO), 1mM GlutaMAX (GIBCO), 0.1 mM β-mercaptoethanol (GIBCO), 1000 U/ml leukemia inhibitory factor (LIF; Millipore), as well as 3 mM CHIR99021 (Sigma) and 1 mM PD0325901 (Sigma), both of which were 2i inhibitors. The culture medium was changed daily, and the resulting mESC colonies were passaged every 3 days to maintain undifferentiated cell states.

To knock-out the Atf3 gene in mESCs, CRISPR/Cas9 was used; briefly, mESCs were transfected with a pair of PX330 CRISPR-Cas9 expression plasmids (PX330-2 A-PuroR, Addgene, 110403), as well as PXP plasmids carrying the puromycin resistance gene (Adgene). The PX330 plasmids targeted regions both up- and downstream of the Atf3 coding sequence, leading to Atf3 gene deletion via non-homologous end joining. Puromycin was then used for mESC selection, followed by identifying homozygous Atf3-KO cell lines using RT-qPCR and Western blot. Primer sequences for L- and R-sgRNA were provided in Table S1.

EB formation and in vitro mESC differentiation

Both WT and Atf3-KO mESCs were encouraged to form EBs, in which adherent mESCs were dissociated with 0.25% trypsin-EDTA (GIBCO), then resuspended in differentiation medium comprising of DMEM (Hyclone) with 20% FBS, 1 mM sodium pyruvate (GIBCO), 1 mM non-essential amino acids (GIBCO), 1 mM GlutaMAX (GIBCO), 0.1 mM β-mercaptoethanol (GIBCO), and 100× Penicillin-Streptomycin Solution (BasalMedia). To form EBs, the hanging drop method was applied, as described in our previous study⁷. In short, hanging droplets containing mESCs (800 cells/30 µL differentiation medium) were dropped onto the lid of a 10 cm Petri dish; the lid was placed back onto the dish, containing 10 mL PBS (BasalMedia), and mESCs cultured for 6 days. The resulting uniformly sized EBs that have formed were transferred to a 24-well plate, containing 2 drops/well of differentiation medium, and differentiation continued for 6 days (Days 7–12), under 37 °C and 5% CO₂.

Identification of cardiomyocyte differentiation from mESCs with immunofluorescence staining and flow cytometry

To identify whether mESCs differentiated into cardiomyocytes, WT and Atf3-KO EBs, at Day 9 of differentiation, were washed 3 times with PBS, fixed with 4% para-formaldehyde at room temperature for 30 min, permeabilized with 0.2% Triton X-100 for 10 min, and blocked with 5% goat serum for 30 min. EBs were incubated with the following primary antibodies overnight at 4 °C: cTnT (BD Biosciences, 565744; 1:300) and Atf3 (Abcam, ab254268; 1:300), followed by AlexaFluor 594-conjugated secondary antibodies (Abcam, ab150116; 1:500) at room temperature for 1 h. Nuclei were stained in 4’,6-diamidino-2-phenylindole (DAPI), and cells observed under a confocal microscope (LSM-880, Zeiss).

For flow cytometry, the same procedure as for immunofluorescence staining was used, but only for cTnT. Data were acquired using a flow cytometer (Cytoflex LX, Beckman Coulter) and analyzed with FlowJo_v10.8.1_CL software.

RT-qPCR

RT-qPCR was conducted by extracting total RNA from mESCs on Days 0, 3, 6, and 9 of differentiation with the FastPure RNA rapid extraction kit, following the manufacturer’s instructions. RNA was reverse transcribed into cDNA, using HiSCRIPT IV cDNA reverse transcription kit, and qPCR performed with ChamQ Blue SYBR Green qPCR Master Mix (all from Vazyzme, China). β-actin was used as a housekeeping gene, and the primers used were presented in Supplemental Table 1. Relative expression was calculated using the 2^−ΔCt method.

Western blot

Total protein was extracted from mESCs by RIPA lysis buffer, and protein concentration quantified with Pierce bicinchoninic acid protein assay kit (Thermo Fisher). Proteins were denatured by boiling in loading buffer for 10 min, and equal amounts of protein for each sample were loaded onto an SDS-polyacrylamide gel, separated by electrophoresis, transferred to PVDF membranes, and blocked with 5% skim milk for 1 h at room temperature. Membranes were subsequently incubated with Atf3 primary antibody (Cell Signaling, 3528 S; 1:1000) overnight at 4 °C, washed in PBS-0.1% Tween-20, then incubated with horseradish peroxidase-conjugated secondary antibody (Abcam, ab254268; 1:300). Protein expression was detected with ECL detection reagent and visualized with Qinxiang imaging system (Clinx Science Instruments Co. Ltd.). Protein expression levels were normalized to the housekeeping gene GAPDH (Proteintech, 60004-1-ig).

RNA-seq analysis of WT and Atf3-KO-differentiated cardiomyocytes

To investigate differences in gene expression levels between WT and Atf3-KO cell groups, RNA-seq was conducted, and the raw image files obtained were subjected to base recognition and error filtering to obtain clean reads, via FastQc for quality control, and Trimmomatic to remove low-quality reads and adaptor sequences. Low-quality reads were defined as having > 10% unknown nucleotides, or > 50% low-quality bases, based on a cut-off of Q-score ≤ 20. Furthermore, sequencing accuracy was assessed using Q-score metrics Q20 and Q30. The resulting clean reads were aligned to the mouse reference genome (mm10) using HISAT2 software, and gene expression levels were quantified as fragments per kilobase of transcript per million mapped reads values. Differential expression analyses between WT and Atf3-KO samples, at each differentiation stage, were performed with DESeq2, and DEGs were defined as |log2fold change|≥1 and adjusted p (false discovery rate) < 0.05. The unique mapped reads identified were then extracted for subsequent analyses, such as cluster analysis of DEGs, in which the relationships and expression differences between different groups were visualized using volcano plots. Furthermore, Gene Set Enrichment (GSEA) and Kyoto Encyclopedia of Gene and Genomes (KEGG) pathway enrichment analyses were conducted on DEGs to explore their potential biological significance. The statistical significance (P value) for each pathway identified by GSEA and KEGG was calculated by Fisher’s test, and significant pathways were identified for additional analyses.

Rescuing Atf3 gene expression in Atf3-KO mESCs to exclude off-target effects from CRISPR/Cas9

To confirm that the cardiomyocyte differentiation effects were Atf3-KO-specific, we performed rescue experiments, involving the construction of an Atf3 overexpression plasmid (pcDNA3.1(+)mAtf3[NM_007498.3]+EGFP, Paizhen Biotechnology, C-21139), which also co-expressed green fluorescence protein (GFP). Plasmids were transfected into Atf3-KO mESCs, in which 2.5 × 10⁵ mESCs were seeded into each well, in 6-well cell culture plates. Upon reaching 70% cell confluence, transfection was carried out in Atf3-KO mESCs using the Lipo3000 liposome transfection kit, following the manufacturer’s instructions (Shanghai Yuanye Biotechnology, R33136). The culture medium was changed after 12 h of transfection, followed by RT-qPCR to identify Atf3-KO cell lines with restored Atf3 expression.

Statistical analysis

All statistical analyses were performed using GraphPad Prism (v10.0). Data are expressed as mean ± standard error of the mean (SEM). Comparisons between 2 groups were carried out with the unpaired 2-tailed Students t-test for normally distributed data, while for data that was not normally distributed, the Wilcoxon rank-sum test was used. For repeated measures over time, 2-way analysis of variance (ANOVA) was conducted. As for comparisons among 3 or more groups, 1-way ANOVA, followed by Tukey’s post hoc test, was applied. P < 0.05 was considered statistically significant.

Results

Atf3-KO in mESCs does not affect their functions

To identify the role that Atf3 plays in the early stages of cardiomyocyte differentiation from ESCs, we knocked out the Atf3 gene in mESCs with CRISPR/Cas9. As shown in the schematic diagram in Fig. S1A, we designed a pair of sgRNA primers to target the 3 exons comprising the coding sequence of Atf3, and successful deletion was confirmed by the resulting PCR products, in which primer set 1, which yields a 912 base-pair (bp) sequence covering exons 1 and 2 in the WT mESCs, was missing in the Atf3-KO mESCs (Fig. S1B). Furthermore, primer set 2, which covers all 3 exons, was 7442 bp in WT, but 1527 bp in Atf3-KO (Fig. S1B). Atf3-KO was further confirmed at both the mRNA and protein levels, by respectively, RT-qPCR (Fig. 1A), as well as immunofluorescence staining and Western blot (Fig. 1B,C; Fig. S1C). Additionally, gene sequencing results verified the deletion of the 3 coding exons, in Atf3-KO (Fig. S1D). All these results thus demonstrated that Atf3 was successfully knocked out in mESCs.

Fig. 1.

Confirmation of activating transcription factor 3 (Atf3) gene knockout in mouse embryonic stem cells (mESCs). (A) Relative gene expression levels for Atf3 between wild-type (WT) and Atf3-KO mESCs. (B) Representative immunofluorescence image of Atf3 protein expression between the 2 groups. (C) Representative Western blot image of Atf3 protein expression between the 2 groups. (D) Atf3-KO did not affect the morphology of monoclonal cells in the 2 groups. (E) Relative expression levels for pluripotency genes Oct4, Sox2, Nanog, and estrogen-related receptor-β (Esrrb), between the 2 groups. (F) Representative images of alive and dead embryoid bodies (EBs), as well as (G) EB formation percentages, between the 2 groups. Data represented as mean ± standard error of the mean (SEM). n = 3/group for all experiments. ***p < 0.001, ns is no statistical significance. Statistical analysis was conducted using the unpaired 2-tailed Students t-test.

To determine whether Atf3-KO affected mESC differentiation and pluripotency, we cultured both WT and Atf3-KO mESCs, and found no significant differences in their morphologies, as they both formed unique round clones (Fig. 1D). We also measured expression levels for pluripotency-related genes Oct4, Sox2, Nanog, and estrogen-related receptor-β (Esrrb), and found no significant differences between WT and Atf3-KO (Fig. 1E).

With respect to EB formation, no significant differences were present between WT and Atf3-KO in terms of the number of EBs formed (Fig. 1F,G). Furthermore, to determine whether the EB phenotype was the result of Atf3-KO, rather than off-target effects, 2 different Atf3-KO clones, 23 and 30, were examined, in which EBs from both clones, from Days 3–9, had the same morphological appearance (Fig. S2A), as well as the percentage of beating EBs on Day 9 (Fig. S2B). Therefore, Atf3-KO did not affect mESC pluripotency or EB formation.

Atf3-KO mESCs yield more beating EBs with higher beating frequencies

To investigate whether Atf3-KO affected cardiomyocyte differentiation from mESCs, both WT and Atf3-ko mESCs, via the hanging drop method, formed EBs on the lid of a 10 cm Petri dish from Days 0–6 of culturing. Afterwards, from Days 7–12, EBs were then cultured as adherent cells in a 24-well plate (Fig. 2A). We observed that up to Day 6, no significant morphological differences were present between WT and Atf3-KO EBs under the hanging drop method (Fig. 2B). However, from Day 8 (day 2 of adherence), significant differences were present between the 2 groups (Fig. 2B), in which Atf3-KO had a higher percentage of beating EBs (Fig. 2C), as well as a higher beating frequency (Fig. 2D), compared to WT. These differences were even more evident in a video of Day 9 cell beating (Videos of WT vs. Atf3-KO). Therefore, we postulated that Atf3-KO promoted cardiomyocyte differentiation, and those differentiated cells were responsible for the beating observed in Atf3-KO EBs.

Fig. 2.

Atf3-KO mESCs yield more beating EBs with higher beating frequencies. (A) Schematic diagram of the timeline for EB formation and mESC differentiation. (B) Images of EBs for both WT and Atf3-KO groups in Days 3, 6, 9 and 12 of differentiation. Red dotted line indicated areas of beating cardiomyocytes. (C) Percentages and (D) frequencies of beating EBs for WT and Atf3-KO mESCs in Days 7–12 of differentiation. Data represented as mean ± SEM. n = 3/group for all experiments. *p < 0.05, **p < 0.01, ***p < 0.001, ns is no statistical significance. Statistical analysis was conducted using 2-way analysis of variance (ANOVA).

Atf3-KO promotes cardiomyocyte differentiation from mESCs

To further confirm whether Atf3-KO promoted cardiomyocyte differentiation, cells from both WT and Atf3-KO groups were obtained on Days 0, 9, and 12, and expression levels for cardiomyocyte markers Actc1, Actn2, Tnnt2 (for cTnT), myosin heavy chain, α isoform (Myh6) and GATA binding protein 4 (Gata4) were measured by RT-qPCR, in which on Day 9, expression levels for those markers were significantly higher in Atf3-KO than WT (Fig. 3A). On Day 12, overall expression levels for these genes were lower for both groups, but they were still higher in Atf3-KO (Fig. 3A). To determine whether cardiac progenitor cells were present prior to cardiomyocyte differentiation, we also examined the expression of cardiac progenitor marker platelet-derived growth factor receptor-α (Pdgrfa) and found that its levels significantly increased in Atf3-KO versus WT from Day 6–9 (Fig. 3B), which was prior to the peaking of cardiomyocyte marker expression on Day 9.

Fig. 3.

Atf3-KO promotes mESC differentiation into cardiomyocytes. (A) Relative gene expression levels for cardiomyocyte markers actin-α cardiac muscle 1 (Actc1), α-actinin 2 (Actn2), Tnnt2 (cardiac troponin T [cTnT]), myosin heavy chain, α isoform (Myh6) and GATA Binding Protein 4 (Gata4), between WT and Atf3-KO mESCs, in Days 0, 9, and 12 of differentiation. (B) Relative gene expression levels for cardiac progenitor marker platelet-derived growth factor receptor-α (Pdgrfa) between the 2 groups, in Days 0, 3, 6, and 9 of differentiation. (C) Representative immunocytochemical images for cTnT, between the 2 groups, in Day 9 of differentiation. Insets show high-magnification views, in which Atf3-KO had obvious sarcomere structures. (D) Flow cytometry analyses and (E) quantification of cTnT⁺ cells, between the 2 groups, in Day 9 and 12 of differentiation. Data represented as mean ± SEM. n = 3/group for all experiments. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001, ns is no statistical significance. Statistical analysis was conducted using 2-way ANOVA.

We then examined whether Atf3-KO affected the differentiation of other cardiac-related cell lineages, particularly endothelial and smooth muscle cells, and found that Atf3-KO, compared to WT, had significantly higher expression levels for endothelial markers VE-cadherin and von Willebrand factor (vWF) on Day 9 (Fig. S3A). However, no significant difference was present between the 2 groups on Day 12 (Fig. S3A). As for the smooth muscle markers smooth muscle myosin heavy chain (SM-Mhc) and Caldesmon, no significant differences were present between the 2 groups on Days 9–12 (Fig. S3B). Thus, Atf3-KO promoted not only cardiomyocyte differentiation but also endothelial differentiation, but it did not affect smooth muscle cell differentiation.

Owing to Day 9 potentially representing the peak of cardiomyocyte differentiation, we further verified this process among WT and Atf3-KO groups via immunofluorescence staining for cardiomyocyte marker cTnT, where it was observed that cTnT⁺ was significantly greater in Atf3-KO versus WT (Fig. 3C); beaded sarcomere structures were also observed in Atf3-KO under high-power magnification, but they were not present in WT (Fig. 3C). The promotion of cardiomyocyte differentiation by knocking out Atf3 was further confirmed by flow cytometry analyses of WT and Atf3-KO cells on Days 9–12, in which Atf3-KO had significantly higher percentages of cTnT⁺ cells than WT (Fig. 3D,E).

Atf3-KO promotes mesodermal germ layer development

As part of the cardiomyocyte differentiation process, mESCs initially differentiate into 3 germ layers: ectoderm, mesoderm, and endoderm, and cardiomyocytes are mainly derived from the mesoderm. Consequently, we investigated whether Atf3-KO affected the development of those germ layers and observed that mesoderm marker T-box transcription factor T (T) was significantly higher in Atf3-KO than WT from Days 3–6 (Fig. 4A). Furthermore, the other mesoderm markers fibroblast growth factor 8 (Fgf8) and Flk-1 had significantly higher expression levels in Atf3-KO than WT on Day 6 (Fig. 4A). As for the endoderm markers SRY-box transcription factor 17 (Sox17), Amn1, and tetraspanin 7 (Tm4sf2; Fig. 4B), as well as ectoderm markers Nestin, SOX1, and paired box 6 (Pax6; Fig. 4C), no significant differences were present between WT and Atf3-KO groups for Days 0–6. Therefore, Atf3-KO promotes mesoderm development in the early stages of differentiation, serving as a potential basis for subsequent cardiomyocyte differentiation.

Fig. 4.

Atf3-KO promotes mesodermal germ layer development. (A) Relative gene expression levels for mesoderm markers T-box transcription factor T (T), fibroblast growth factor 8 (Fgf8), Flk-1, (B) endoderm markers SRY-box transcription factor 17 (Sox17), Amn1, tetraspanin 7 (Tm4sf2), and (C) ectoderm markers Nestin, SOX1, and paired box 6 (Pax6), between WT and Atf3-KO mESCs in Days 0, 3, and 6 of differentiation. Data represented as mean ± SEM. n = 3/group for all experiments. *p < 0.05, ***p < 0.001, ****p < 0.0001, ns is no statistical significance. Statistical analysis was conducted using 2-way ANOVA.

Atf3-KO is enriched in cardiomyocyte differentiation-associated DEGs and pathways under RNA-seq

To further verify that Atf3-KO promotes cardiomyocyte differentiation, we examined the mRNA expression profiles of Atf3-KO versus WT cells, via RNA-seq analyses at Days 0, 3, 6, and 9. Figure 5A shows volcano plots with DEGs identified between the 2 groups, where similar numbers of down- and up-regulated DEGs were found in Atf3-KO versus WT on Days 0 (55 down-, 27 up-regulated) and 3 (51 down-, 11 up-regulated; Fig. 5A). However, Day 6 had a significant increase in the number of down- and up-regulated DEGs for Atf3-KO, at, respectively, 482 and 260, while Day 9 numbers were similar to that of Day 6 (353 down-, 307 up-regulated; Fig. 5A). The top 29 DEGs with the most significant expression changes in Atf3-KO versus WT (15 up-, 14 down-regulated), at days 0, 3, 6, and 9, were then included in heat maps, in which on Day 9, cardiomyocyte markers Actc1 and Actn2 were significantly upregulated in Atf3-KO versus WT, further supporting that Atf3-KO promoted cardiomyocyte differentiation (Fig. 5B).

Fig. 5.

RNA-sequencing (RNA-seq) analysis of WT and Atf3-KO mESCs at Days 0–9 of differentiation. (A) Volcano plots of differentially-expressed genes (DEGs) between WT and Atf3-KO groups, at Days 0, 3, 6, and 9 of differentiation. Downregulated DEGs were blue, upregulated were red, and those with no statistically significant expression changes were grey. (B) Heat maps of the 29 DEGs (15 upregulated, 14 downregulated) with the most significant expression level changes between WT and Atf3-KO groups, at Day 0, 3, 6, and 9 of differentiation. Data represented as mean ± SEM. Statistical analysis was conducted using the unpaired 2-tailed Students t-test.

Furthermore, KEGG analysis of DEGs on Days 6 and 9 identified their top 10 most enriched signaling pathways in Atf3-KO versus WT^11,12, in which on Day 6, these pathways were steroid biosynthesis, Hedgehog signaling, and basal cell carcinoma, while on Day 9, they were dilated and hypertrophic cardiomyopathy, as well as cardiac muscle contraction (Fig. 6A,B). These findings were further supported by GSEA enrichment analysis of KEGGs on Day 9, in which the same 3 pathways identified by KEGG on Day 9 were also found to be significantly upregulated (Fig. 6C–E). Overall, these discoveries demonstrated that knocking out Atf3 was associated with the upregulation of genes associated with cardiomyocyte differentiation from mESCs.

Fig. 6.

RNA-seq analysis revealed that the signaling pathways related to myocardial differentiation were upregulated after Atf3-KO. Kyoto Genes and Genomes (KEGG) enrichment analysis of DEGs (p < 0.05) at (A) Days 6 and (B) 9 of differentiation. (C) Bubble, (D) pathway, and (E) ridge plots of GSEA enrichment analysis of DEGs on Day 9 of differentiation. Data represented as mean ± SEM. Statistical analysis was conducted using the unpaired 2-tailed Students t-test.

Rescuing Atf3 expression in Atf3-KO mESCs confirmed that the cardiomyocyte differentiation effects were Atf3-KO-specific

To exclude the possible off-target effects from Atf3-KO with CRISPR/Cas9, rescue experiments, entailing transfection of Atf3-KO mESCs with an Atf3 overexpression plasmid (Fig. S4A,B), were performed, yielding the Atf3-KO+Rescue group. Successful transfection of the plasmid was confirmed by GFP expression, both in the initial mESCs and EBs (Fig. S4C). Furthermore, under RT-qPCR, mRNA expression levels for Atf3 in Atf3-KO+Rescue were significantly higher than Atf3-KO, and similar to that of WT, indicating that the rescue experiments restored Atf3 gene expression (Fig. 7A).

Fig. 7.

Restoration of Atf3 expression in Atf3-KO mESCs reversed the effects of Atf3-KO towards that of WT cells. (A) Relative mRNA expression levels for Atf3 among WT, Atf3-KO, and Atf3-KO mESCs who had Atf3 expression restored (Atf3-KO+Rescue). Relative gene expression levels for (B) Actc1, (C) Actn2, (D) Tnnt2, (E) Myh6, and (F) Gata4, among the 3 groups, on Day 9 of differentiation. (G) Representative immunocytochemical images for cTnT, among the 3 groups, on Day 9 of differentiation. (H) Flow cytometry analyses and (I) quantification of cTnT⁺ cells, among the 3 groups, on Day 9 of differentiation. Data represented as mean ± SEM. n = 3/group for all experiments. *p < 0.05, **p < 0.01, ***p < 0.001, ns is no statistical significance. Statistical analysis was conducted using 1-way ANOVA, followed by Tukey’s post hoc test.

With respect to cardiomyocyte differentiation, mRNA expression for cardiomyocyte markers Actc1 (Fig. 7B), Actn2 (Fig. 7C), Tnnt2 (Fig. 7D), Myh6 (Fig. 7E), and Gata4 (Fig. 7F), on Day 9 of differentiation, returned to comparable levels as WT in Atf3-KO+Rescue, further confirming that Atf3-KO promoted cardiomyocyte differentiation. These mRNA expression results were supported by immunofluorescence staining for cardiomyocyte marker protein cTnT, where cTnT⁺ was significantly greater in Atf3-KO versus WT, but were lowered towards that of WT in Atf3-KO+Rescue (Fig. 7G). This pattern was also present for flow cytometry analyses of WT, Atf3-KO, and Atf3-KO+Rescue cells on Day 9, in which Atf3-KO had significantly higher percentages of cTnT⁺ cells than WT (Fig. 7H-I). However, in Atf3-KO+Rescue, these levels were reduced to similar levels of WT, signifying that restoring Atf3 expression in Atf3-KO cells counteracted the effects of knocking out ATF3 (Fig. 7H-I). All these findings thereby validate the notion that promoting cardiomyocyte differentiation was specific to Atf3-KO, without off-target effects from CRISPR/Cas9.

Discussion

Atf3 is a member of the ATF/cAMP response element binding family of transcription factors, and its expression is induced by a variety of stress responses⁶. Likewise, multiple studies have shown that its expression has been closely related to cardiovascular disease pathogenesis, but its precise roles in cardiac development have not been clearly defined⁸. Consequently, identifying the roles that Atf3, as well as other key regulatory factors, play in ESC differentiation into cardiomyocytes could aid in fostering the development of regenerative cell therapies for cardiovascular diseases. In this study, we aimed to further elucidate the role that Atf3 plays in cardiomyocyte differentiation by Atf3-KO in mESCs. We found that Atf3 served as a negative regulator of cardiomyocyte differentiation, as Atf3-KO, compared to WT mESCs, yielded larger numbers of beating EBs from Days 8–9 of differentiation (days 2–3 of adherence). Furthermore, Atf3-KO cells had significant up-regulation of mesodermal and cardiomyocyte differentiation markers, compared to WT. This was further verified by RNA-seq analyses showing that on Day 9, the cardiomyocyte markers Actc1 and Actn2 were significantly upregulated in Atf3-KO versus WT, as well as Atf3-KO DEGs being enriched for dilated and hypertrophic cardiomyopathy, as well as cardiac muscle contraction, under both KEGG and GSEA analyses. All of these effects, particularly for cardiomyocyte differentiation gene and protein marker expression, were reversed upon restoration of Atf3 expression in Atf3-KO mESCs. Therefore, Atf3-KO in the early stages of ESC differentiation could promote additional mesodermal germ layer development, and subsequently, cardiomyocyte differentiation, serving as a potential approach to obtain a more homogeneous ESC-derived cardiomyocyte population for cell therapies.

It has long been noted that during embryonic development, cardiac tissues are mainly derived from the mesodermal germ layer¹³. As a result, we initially postulated that Atf3-KO would promote mesodermal lineage development, which was confirmed by our observations that Atf3-KO cells, from Days 0–6 of differentiation, had upregulated expression of mesodermal markers, while no significant changes from WT were present for the other 2 germ layers. Moreover, Pdgfra, a marker for cardiac progenitor cells, which are a critical cell type for cardiomyocyte differentiation, was found to be upregulated from Days 6–9 in Atf3-KO, compared to WT. Furthermore, RNA-seq analyses of the regulatory pathways also revealed that Atf3-KO cells had significant upregulation of cardiomyocyte differentiation and development-related genes and pathways, compared to WT. All these findings collectively indicated that Atf3-KO results in the promotion of cardiac cell lineages from the mesodermal germ layer, and subsequently cardiac progenitors for cardiomyocyte differentiation.

Various stimuli, including hypertension, diabetes, pressure overload, ischemia, hypoxia, hyperlipidemia, and inflammatory factors, could induce cardiovascular Atf3 expression¹⁴. Atf3, as a transcription factor, in turn interacts with multiple signaling factors, such as TLR4 and the kinases PKA, AKT, ERK, p38, and JNK¹⁵, subsequently regulating the expression of multiple target genes. With respect to its role in cardiovascular disease, though, it is still controversial whether Atf3 plays a pro-pathogenic or protective role¹⁶. For instance, Zhou et al. found that 4 weeks post-aortic ligation, Atf3-KO mice, compared to WT, had more prominent myocardial hypertrophy and fibrosis, along with increased ERK and JUNK pathway activation, indicating that Atf3 may play a protective role against myocardial hypertrophy, via negative feedback on ERK and JNK pathways¹⁷. Similarly, Liu et al. observed that Atf3 exerted protective effects against post-MI ventricular remodeling, in which Atf3 knock-down aggravated, while Atf3 overexpression inhibited ferroptosis-induced ventricular remodeling¹⁸. Likewise, Shao et al., noted that GAS6 activated Atf3, which in turn prevented MerTK⁺ macrophage apoptosis post-myocardial ischemia-reperfusion injury, via its inhibition of genes involved in type I interferon expression (Ifih1 and Ifnb1) and apoptosis (Apaf1)¹⁹. On the other hand, Eiselein et al. observed that TGF-β1 activated the non-Smad pathway, involving the Atf3-JNK transcription factor network, to induce the expression of proinflammatory cytokines (IL-6, NFKBIA, NFKB1) and cell apoptosis (caspase-3), suggesting that Atf3 could play a pro-inflammatory and apoptotic role in the cardiovascular endothelium²⁰. These contradictory findings regarding the role of Atf3 in cardiovascular disease may be due to differences in disease models, as well as the type of stimuli and stimulation duration. However, all of these findings only relate to the role Atf3 could potentially play in cardiovascular disease, rather than in cardiomyocyte differentiation, which our findings remediate by demonstrating that Atf3-KO in mESCs increased their differentiation into cardiomyocytes, indicating that it served as a negative regulator of the early stages of cardiomyocyte differentiation. Further research, though, is still needed to more fully elucidate the roles of Atf3 in cardiomyocyte differentiation, as well as its applicability in treating cardiovascular diseases.

Another factor involved in regulating cardiomyocyte differentiation is c-Jun, which we have identified in our previous study as being a key negative regulator in this process¹⁰. Concurring with our findings, Zhong et al. noted, using single-cell sequencing, that c-Jun was able to inhibit cardiomyocyte differentiation, via its regulation of chromatin accessibility and modification of H3K4me3²¹. In line with these previous observations, Atf3, which, like c-Jun, is also part of the AP-1 protein family, was also shown in this study to play a key regulatory role in cardiomyocyte differentiation. This may also involve modifying chromatin accessibility, as a prior study by Ben-Yair et al. demonstrated that AP-1 family transcription factors were able to promote zebrafish cardiac tissue regeneration by altering chromatin accessibility²². Based on those observations, we postulate that this mechanism could serve as the basis behind Atf3 potentially being able to affect cardiomyocyte differentiation and regeneration. Yet another potential mechanism behind Atf3 being able to regulate cardiomyocyte differentiation is through its involvement in glucose and lipid metabolism^23,24, which has been found to be able to subsequently coordinate cardiac and hematopoietic progenitor cell differentiation²⁵. Additionally, recent findings demonstrated that inhibiting fatty acid oxidation could lead to mouse myocardial tissue regeneration^27,28. Future studies, though, will be required to fully elucidate whether Atf3-KO could promote cardiomyocyte differentiation by regulating metabolic and chromatin accessibility changes.

There are a number of limitations to this study, one of which is the small sample size of 3 in the RNA-seq analysis, leading to discrepancies in DEG expression levels among the samples. Consequently, future studies will involve larger sample sizes to obtain more representative DEG profiles. Moreover, the observations of putative increases in cardiomyocyte differentiation in Atf3-KO mESCs, in terms of marker expression, were not able to be fully correlated with functional analyses, such as Ca²⁺ transient detection or patch clamp technique to identify action potentials. Future studies will thus involve functional analyses to further substantiate the notion of Atf3-KO being associated with increased cardiomyocyte differentiation in mESCs. Lastly, proliferation levels of non- and myocardial cells over time should be tracked using cell proliferation analyses, such as the 5-bromo-2′-deoxyuridine (BrdU) pulse-chase assay, coupled with staining for non- (ex. smooth muscle, endothelial) and myocardial cell markers, to determine whether the decrease in mature myocardial cells in Days 9–12 of differentiation could be owed to increases in non-myocardial cell proliferation.

Conclusion

In this study, we demonstrated that Atf3 serves as a key negative regulator of cardiomyocyte differentiation from mESCs, in which its deletion promoted mESC differentiation into mesodermal germ lineages, followed by cardiac progenitors and cardiomyocytes. This was demonstrated by Atf3-KO having larger numbers of beating EBs from Days 8–9 of differentiation, compared to WT EBs, which was associated with up-regulated expression of cardiomyocyte markers, such as cTnT, Actc1 and Actn2. The underlying mechanisms were identified by RNA-seq, in which Atf3-KO was enriched for cardiomyocyte differentiation DEGs; these DEGs were associated with dilated and hypertrophic cardiomyopathy, as well as cardiac muscle contraction pathways under KEGG and GSEA. On the other hand, restoration of Atf3 expression in Atf3-KO cells reversed the cardiomyocyte differentiation-promoting effects of Atf3-KO, indicating that they were Atf3-KO-specific, rather than from the off-target effects of CRISPR/Cas9. Our findings thus provide new insights in further promoting cardiomyocyte differentiation from ESCs by knocking out Atf3, serving as a potential therapeutic target for cardiac regeneration cell therapies.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

L.S. and S.L. conceived and designed the project. Z.J., R.H., Y.S. and M.W. analyzed the data. L.S. and Z.J. wrote the manuscript with the help of all authors. S.L., Z.L. and C.L. had full access to all data in the study and took responsibility for the integrity of the data, as well as for the manuscript.

Funding

This work is supported by Fujian Provincial Natural Science Foundation of China (2024J08006).

Data availability

The datasets generated and/or analysed during the current study are available in the Gene Expression Omnibus database, [GSE306049]. The data in the current study are available from the corresponding authors upon reasonable request.

Declarations

Competing interests

The authors declare no competing interests.