Jmjd3 Controls Mesodermal and Cardiovascular Differentiation of Embryonic Stem CellsNovelty and Significance
Rationale: The developmental role of the H3K27 demethylases Jmjd3, especially its epigenetic regulation at target genes in response to upstream developmental signaling, is unclear.
Objective: To determine the role of Jmjd3 during mesoderm and cardiovascular lineage commitment.
Methods and Results: Ablation of Jmjd3 in mouse embryonic stem cells does not affect the maintenance of pluripotency and self-renewal but compromised mesoderm and subsequent endothelial and cardiac differentiation. Jmjd3 reduces H3K27me3 marks at the Brachyury promoter and facilitates the recruitment of β-catenin, which is critical for Wnt signal–induced mesoderm differentiation.
Conclusions: These data demonstrate that Jmjd3 is required for mesoderm differentiation and cardiovascular lineage commitment.
Post-translational modifications of histone proteins represent essential epigenetic control mechanisms that can either allow or repress gene expression.1 Trimethylation of H3K27 is mediated by Polycomb group proteins and represses gene expression.2 The JmjC domain–containing proteins, UTX (ubiquitously transcribed tetratricopeptide repeat, X chromosome) and Jmjd3 (jumonji domain–containing protein 3, Kdm6b), not only act as demethylases to remove the repressive H3K27me3 marks, but also exhibit additional demethylase-independent functions.3–6 Jmjd3 is induced and participates in Hox gene expression during development,7 neuronal differentiation,8,9 and inflammation,5,10–12 and recent data suggest that Jmjd3 inhibits reprogramming by inducing cellular senescence.13 Because previous studies suggest that H3K27me3 regulates endothelial gene expression in adult proangiogenic cells,14 we addressed the function of Jmjd3 in cardiovascular lineage differentiation of embryonic stem cells (ESCs).
In This Issue, see p 835
Editorial, see p 837
A detailed description of the experimental procedure is provided in the Online Data Supplement.
Mouse ESCs were cultured and maintained on mitotically inactivated mouse embryonic fibroblasts. Differentiation of ESCs was induced without leukemia inhibitory factor in static or rotary suspension condition. ESC differentiation into mesoderm was performed with a well-described OP9 coculture system.15,16 For cardiac differentiation, embryoid bodies (EBs) were derived using the hanging drop method.17 For endothelial differentiation, EBs were formed and cultured in medium supplemented with cytokine cocktail.18
Plasmid Construction and Stable Transfection
The full-length Jmjd3, the mutants, and Brachyury were cloned into pEF1 vector (Invitrogen). The linearized plasmids were transfected in Jmjd3−/− ESCs using the Amaxa nucleofection system (Lonza).
Chromatin immunoprecipitation assays were performed as previously described.14 Sheared chromatin was immunoprecipitated with antibodies as indicated. The purified chromatin was used in quantitative real-time polymerase chain reaction (Applied Biosystems).
Jmjd3 knockout ESCs were generated by 2 rounds of gene targeting (Online Figure IA and IB). We obtained 7 Jmjd3−/− ESC clones, which lacked Jmjd3 mRNA and protein expression. All of the clones showed slightly increased global H3K27me3, but the expression of pluripotency genes, the morphology, the growth kinetic, and survival was indistinguishable from wild-type (WT) ESCs (Figure 1A–1C; Online Figure IC–IF). No significant changes of repressive H3K27me3 marks at the promoters of pluripotency genes were detected in Jmjd3−/− compared with WT ESCs (Online Figure IH). When spontaneous differentiation was induced by leukemia inhibitory factor withdrawal, Jmjd3 expression increased in WT ESCs (Figure 1D). EBs derived from Jmjd3−/− ESCs were slightly smaller in size compared with WT EBs (Figure 1E). mRNA expression profiling of Jmjd3−/− and WT ESC clones at day 4 after induction of differentiation showed a distinct expression pattern of lineage-specific genes (Online Figure IIA). Gene ontology functional analyses revealed a significant repression of genes that are involved in mesoderm development (Figure 1F; Online Figure IIB). Moreover, repressed gene sets in Jmjd3−/− EBs were shown to be related to cardiac and vascular development, consistent with impaired mesoderm differentiation (Figure 1F; Online Figure IIB).
Validation of the microarray results showed a similar reduction of pluripotency gene expression after leukemia inhibitory factor withdrawal in Jmjd3−/− compared with WT ESCs (Figure 1G). However, depletion of Jmjd3 substantially compromised the induction of mesodermal genes (Figure 1G). Especially, the pan-mesoderm marker, Brachyury, and the early mesoendoderm marker, Mixl1, were profoundly increased at day 4 of differentiation in WT ESCs, but not in Jmjd3−/− ESCs (Figure 1G). Moreover, the mesoendodermal marker, Eomes, and endodermal markers, such as Sox17 and FoxA2, were significantly suppressed, which is consistent with a very recent study showing that Jmjd3 is required for endoderm differentiation.19 Ectodermal markers were not significantly changed in Jmjd3−/− ESCs when using the spontaneous differentiation protocol (Figure 1G).
Because Jmjd3−/− ESCs showed a prominent inhibition of mesodermal markers after leukemia inhibitory factor withdrawal, we next questioned whether this phenotype can also be observed when directing differentiation of mesoderm using 2 different protocols. Consistent with our findings, Jmjd3−/− ESCs showed a reduced expression of mesodermal marker genes when using the protocol for mesoderm differentiation described by Gadue et al20 (data not shown). Moreover, mesoderm differentiation was significantly suppressed when Jmjd3−/− ESCs were cultured on OP9 stromal cells, which support mesodermal differentiation21 (Figure 2A). Whereas WT ESCs showed the typical time-dependent increase in Brachyury+ cells, Jmjd3−/− ESCs generated significantly less Brachyury+ mesodermal cells (Figure 2B). Moreover, fluorescence activated cell sorting analysis revealed that fetal liver kinase (Flk)1+ vascular endothelial-cadherin− mesodermal cells were generated in WT ESCs but were reduced when Jmjd3−/− ESCs were used (Figure 2C). Interestingly, the formation of vascular endothelial-cadherin+ Flk+ cells was also significantly reduced by 96±1% and 88±3% in the 2 Jmjd3−/− ESC clones compared with WT ESCs (P<0.01), prompting us to explore the role of Jmjd3 in vascular differentiation further.
Endothelial differentiation was induced by a cytokine cocktail18 and was associated with a significant upregulation of Jmjd3 expression (Online Figure IIIA). Jmjd3−/− ESCs showed a marked reduction of endothelial differentiation as evidenced by significantly reduced mRNA levels of the endothelial marker vascular endothelial-cadherin and endothelial-specific receptor tyrosine kinase Tie2 (Figure 3A). The formation of endothelial marker expressing vascular structures after induction of endothelial differentiation was abolished in Jmjd3−/− ESCs (Figure 3B; Online Figure IIIB). The impaired endothelial differentiation of Jmjd3−/− cells was partially rescued by the overexpression of Brachyury (Online Figure IIIC and IIID), suggesting that the inhibition of mesoderm formation, at least in part, contributes to the impaired endothelial commitment.
Because genes involved in heart development and morphogenesis were significantly downregulated in Jmjd3−/− ESCs on differentiation (Figure 1F; Online Figure II), we additionally determined the capacity of Jmjd3−/− ESCs to generate cardiomyocytes by inducing cardiac differentiation.17 Expression of cardiac progenitor cell markers, Mesp1 and Pdgfra, was inhibited in Jmjd3−/− ESCs compared with WT ESCs (Figure 3C). Moreover, after plating on gelatin-coated dishes, the Jmjd3−/− ESCs showed an impaired formation of EBs and only 20% of EBs were contracting (Figure 3D). Consistently, expression of the cardiac transcription factor Mef2c, the marker of working myocardium Nppa, and cardiac structural proteins TnT2 and α-myosin heavy chain were downregulated in Jmjd3−/− ESCs (Figure 3E and 3F; Online Figure IIIE).
Next, we addressed whether the impaired mesoderm differentiation observed in Jmjd3−/− ESCs might be mediated by an increase of repressive H3K27me3 marks at the promoters of developmental regulators. Of the various promoters studied, only Brachyury and Mixl1 showed a significant augmentation of H3K27me3 marks in Jmjd3−/− ESCs on differentiation (Figure 4A; Online Figure IVA). Consistently, the recruitment of RNA polymerase II to the transcription start sites of the promoters of Brachyury and Mixl1 was also significantly reduced (Online Figure IVC). In addition, Jmjd3 deficiency repressed polymerase II recruitment to the Flk1 and Mesp1 promoter but the inactivation of these promoters was not associated with changes in H3K27me3 marks (Figure IVA and IVC). These data were confirmed using protocols that induce mesoderm differentiation by addition of Wnt3a (data not shown).20 Under these conditions, Jmjd3−/− ESCs showed a 1.81±0.23-fold (P<0.05) enrichment of H3K27me3 marks at the Brachyury promoter compared with WT ESCs.
To determine whether the demethylase activity of Jmjd3 controls Brachyury expression by reducing repressive H3K27me3 marks during differentiation, we overexpressed full-length Jmjd3, the carboxyl-terminal part, including the JmjC-domain (cJmjd3: amino acids, 1141–1641), and a carboxyl-terminal mutant construct, which includes a point mutation (cJmjd3H1388A) to inactivate demethylase activity. Overexpression of full-length Jmjd3 and the carboxyl-terminal part of Jmjd3 in Jmjd3−/− ESCs partially rescued the expression of Brachyury on differentiation (Figure 4B and 4C). However, the inactive carboxyl-terminal part of Jmjd3 failed to rescue the impaired Brachyury expression in Jmjd3−/− ESCs (Figure 4C), suggesting that the demethylase activity of Jmjd3 is required for the activation of the Brachyury promoter.
Because canonical Wnt signaling regulates the expression of Brachyury during development22,23 and Wnt/β-catenin–dependent genes were suppressed in Jmjd3−/− EBs compared with WT EBs (Online Figure V), we further explored whether Jmjd3 might interact with β-catenin signaling. Indeed, β-catenin recruitment to the Brachyury promoter was significantly suppressed in Jmjd3−/− ESCs (Figure 4D) and was rescued by Jmjd3 overexpression (Figure 4E). Similar results were obtained when using the protocol for direct mesoderm differentiation described by Gadue et al20 (data not shown). To determine whether Jmjd3 might interact with β-catenin, we performed coimmunoprecipitation studies and showed that Jmjd3 interacts with β-catenin in human embryonic kidney 293 cell and differentiated EBs (Figure 4F; Online Figure VI). To assess a direct effect of Jmjd3 on β-catenin responsive promoter activity, we used a luciferase reporter assay. Coexpression of lymphoid enhancer binding factor 1 and the constitutive active form of β-catenin harboring a nuclear localization signal resulted in the activation of lymphoid enhancer binding factor 1 luciferase reporter activity in WT ESCs, but this transcriptional activation was markedly impaired in Jmjd3−/− ESCs (Figure 4G).
The data of the present study demonstrate that deletion of Jmjd3 in ESCs does not affect self-renewal but significantly impairs the formation of mesoderm on induction of differentiation. The findings that Jmjd3 is not required for ESC maintenance are consistent with the dispensability of the Polycomb complex and the related demethylase UTX for self-renewal.1 The requirement of Jmjd3 for mesoderm differentiation was shown in spontaneous differentiation, as well as when more specifically inducing mesoderm differentiation by the OP9 coculture system or under serum-free conditions followed by Wnt3a stimulation. Jmjd3 deficiency profoundly suppressed the expression of Brachury, which is essential for mesoderm differentiation. In the absence of Jmjd3, repressive H3K27me3 marks at the Brachyury promoter are significantly increased, and the recruitment of β-catenin, which is a prerequisite for Wnt-induced mesoderm differentiation, is impaired. In addition, Jmjd3 is interacting with β-catenin and is contributing to β-catenin–dependent promoter activation. This is consistent with the recent findings that cofactors can form a complex with β-catenin/lymphoid enhancer binding factor 1 at Tcf/lymphoid enhancer binding factor 1 binding sites at β-catenin–dependent promoters and synergize with canonical Wnt signaling.24 Interestingly, a demethylase-independent regulation of β-catenin–dependent gene expression was recently described for UTX.25 However, our data provide evidence that Brachyury expression in Jmjd3−/− ESCs is only rescued by catalytically active Jmjd3, which has maintained the demethylase activity. On the basis of these findings, we propose a model in which Jmjd3 is recruited to the Brachury promoter to remove repressive H3K27me3 marks and on Wnt stimulation additionally promotes β-catenin–dependent promoter activation (Figure 4H). Such a model is similar to the recently described function of Jmjd3 in endoderm differentiation, whereby Jmjd3 associates with Tbx3 and is recruited to the poised promoter of Eomes, to mediate chromatin remodeling allowing subsequent induction of endoderm differentiation induced by activin A.19 The present study additionally demonstrates that Jmjd3 contributes to endothelial and cardiac differentiation. Particularly, endothelial differentiation was profoundly impaired, a finding that is consistent with previous findings in adult progenitor cells, showing a high H3K27me3 at endothelial genes.14 The relatively modest inhibition of cardiomyocyte differentiation in Jmjd3−/− ESCs may be, in part, explained by a compensatory effect of UTX which was shown to regulate cardiac development.26 Together, our study provides first evidence for the regulation of β-catenin–dependent Wnt target genes by Jmjd3 during differentiation of ESCs.
However, the in vivo relevance of the findings is still unclear. The Jmjd3−/− mice that we have generated out of the ESCs, used in the present study, showed embryonic lethality before E6.5, suggesting a crucial role of Jmjd3 in early embryonic development (Online Figure VII). This is consistent with the requirement of Jmjd3 for blastocyst development,27 but is in contrast to other studies which showed that Jmjd3-deficient mice are perinatal lethal.12,28 The discrepancy between the phenotypes of Jmjd3−/− mice is unclear but might be related to the different strategies that were used to generate the Jmjd3−/− mice. For example, the study of Burgold et al28 resulted only in a half maximal reduction of Jmjd3 expression, whereas Jmjd3 expression is fully absent in our mice. The early embryonic lethality of our mice, however, precluded the analysis of the effect of Jmjd3 on mesoderm development in vivo. Further studies involving conditional deletion of Jmjd3 using Sox2- and Brachyury-Cre lines are required.
We thank Marion Muhly-Reinholz, Susanne Heydt, Susanne Kreutzer, and Claudia Jourdan for excellent technical assistance.
Sources of Funding
S.D. and M.A.R. are supported by the German Research Foundation (SFB834). S.D. M.A.R., G.D., T.B. and A.M.Z. were supported by the LOEWE Center for Cell and Gene Therapy Frankfurt (IIIL4-518/17.004).
In June 2013, the average time from submission to first decision for all original research papers submitted to Circulation Research was 13.67 days.
Brief UltraRapid Communications are designed to be a format for manuscripts that are of outstanding interest to the readership, report definitive observations, but have a relatively narrow scope. Less comprehensive than Regular Articles but still scientifically rigorous, BURCs present seminal findings that have the potential to open up new avenues of research. A decision on BURCs is rendered within 7 days of submission.
This manuscript was sent to Benoit Bruneau, Consulting Editor, for review by expert referees, editorial decision, and final disposition.
The online-only Data Supplement is available with this article at http://circres.ahajournals.org/lookup/suppl/doi:10.1161/CIRCRESAHA.113.302035/-/DC1.
- Nonstandard Abbreviations and Acronyms
- embryoid body
- embryonic stem cell
- wild type
- Received June 18, 2013.
- Revision received July 12, 2013.
- Accepted July 15, 2013.
- © 2013 American Heart Association, Inc.
- Hong S,
- Cho YW,
- Yu LR,
- Yu H,
- Veenstra TD,
- Ge K
- Akizu N,
- Estarás C,
- Guerrero L,
- Martí E,
- Martínez-Balbás MA
- Ohtani K,
- Vlachojannis GJ,
- Koyanagi M,
- Boeckel JN,
- Urbich C,
- Farcas R,
- Bonig H,
- Marquez VE,
- Zeiher AM,
- Dimmeler S
- Nakano T,
- Kodama H,
- Honjo T
- Vittet D,
- Prandini MH,
- Berthier R,
- Schweitzer A,
- Martin-Sisteron H,
- Uzan G,
- Dejana E
- Gadue P,
- Huber TL,
- Paddison PJ,
- Keller GM
- Yamaguchi TP,
- Takada S,
- Yoshikawa Y,
- Wu N,
- McMahon AP
- Wang C,
- Lee JE,
- Cho YW,
- Xiao Y,
- Jin Q,
- Liu C,
- Ge K
- Canovas S,
- Cibelli JB,
- Ross PJ
Novelty and Significance
What Is Known?
Cell fate decisions require well-controlled changes in gene expression that are tightly controlled by epigenetic modulators.
The post-transcriptional modifications of histone proteins epigenetically regulate gene expression.
Trimethylation of lysine 27 at histone K3 (H3K27me3) silences gene expression and can be reversed by the demethylase Jmjd3.
What New Information Does This Article Contribute?
The histone demethylase Jmjd3 is required for mesoderm differentiation and cardiovascular lineage commitment of mouse embryonic stem cells.
This effect is partially mediated by a silencing of the mesodermal regulator Brachyury.
Ablation of Jmjd3 further reduces β-catenin recruitment to the Brachyury promoter, which interferes with Wnt signaling that is required for proper mesoderm differentiation.
The differentiation of stem cells to specific lineages requires a well-defined modulation of gene expression programs, which is often controlled by epigenetic mechanisms. Although several epigenetically active enzymes and complexes have been described, the function of the histone demethylase Jmjd3 for cardiovascular lineage commitment was unknown. Using mouse embryonic stem cells as a model, we now show that the demethylase Jmjd3 is required for mesoderm differentiation and for the differentiation of embryonic stem cells to the vascular and cardiac lineage. We further identified the mechanism and showed that ablation of Jmjd3 resulted in a silencing of the Brachyury promoter that is associated with an increase in H3K27me3 marks. In addition, Jmjd3 was shown to facilitate the recruitment of β-catenin to the Brachyury promoter, which contributes to the Wnt-dependent activation of mesoderm differentiation. Together these data describe a novel epigenetic mechanism that controls cell fate decision.