Combined Overexpression of JARID2, PRDM14, ESRRB, and SALL4A Dramatically Improves Efficiency and Kinetics of Reprogramming to Induced Pluripotent Stem Cells

Abstract

Identification of a gene set capable of driving rapid and proper reprogramming to induced pluripotent stem cells (iPSCs) is an important issue. Here we show that the efficiency and kinetics of iPSC reprogramming are dramatically improved by the combined expression of Jarid2 and genes encoding its associated proteins. We demonstrate that forced expression of JARID2 promotes iPSC reprogramming by suppressing the expression of Arf, a known reprogramming barrier, and that the N‐terminal half of JARID2 is sufficient for such promotion. Moreover, JARID2 accelerated silencing of the retroviral Klf4 transgene and demethylation of the Nanog promoter, underpinning the potentiating activity of JARID2 in iPSC reprogramming. We further show that JARID2 physically interacts with ESRRB, SALL4A, and PRDM14, and that these JARID2‐associated proteins synergistically and robustly facilitate iPSC reprogramming in a JARID2‐dependent manner. Our findings provide an insight into the important roles of JARID2 during reprogramming and suggest that the JARID2‐associated protein network contributes to overcoming reprogramming barriers. Stem Cells 2016;34:322–333

Significance Statement

Induced pluripotent stem cells (iPSCs), which show embryonic stem cell‐like properties, are generated by forced expression of transcription factors including OCT4, SOX2, KLF4, and c‐MYC (OSKM). The identification of a gene set that can achieve proper reprogramming to iPSCs is essential for understanding the molecular mechanism underlying reprogramming and for cell‐based therapies in regenerative medicine. Here, we show that JARID2 promotes iPSC reprogramming via regulation of gene expression and physical interaction with pluripotency factors including PRDM14, ESRRB, and SALL4A. Furthermore, our data suggest that the combined expression of OSKM and these four factors can drive highly efficient and rapid reprogramming to iPSCs.

Introduction

Somatic cell reprogramming can be achieved by cell fusion with embryonic stem cells (ESCs) or by nuclear transfer into enucleated oocytes 1, 2. In addition, induced pluripotent stem cell (iPSC) reprogramming has been shown to occur via poorly resolved stochastic and epigenetic events in a small number of cells overexpressing multiple transcription factors, such as Oct4 (also known as Pou5f1), Sox2, Klf4, and c‐Myc (OSKM), which are essential for ESC identity 1, 3. Reprogramming by OSKM is a less efficient and slower process than that by cell fusion or nuclear transfer 2, 3. Furthermore, higher rates of aberrant DNA methylation were observed in pluripotent stem cells (PSCs) generated by OSKM than by nuclear transfer 4. These findings suggest that OSKM by themselves are not sufficient for induction of proper somatic cell reprogramming compared with that attained by cell fusion or nuclear transfer, and a rapid and highly efficient generation of iPSCs may be achieved by the addition of some other factors expressed in ESCs and/or oocytes. Recently, several studies have demonstrated that iPSC reprogramming by OSKM can be improved, both in efficiency and kinetics, by the addition of pluripotency genes, including Kdm2b (also known as Jhdm1b and Fbxl10) and Nr5a2 (also known as Lrh‐1) 5-8, or the repression of genes that function as barriers against iPSC reprogramming including Arf, Tp53, Tcf3, Jmjd3, Mbnl1, and Mbnl2 9-14. However, only a limited number of studies have revealed the promoting factors associated with oocytes 15-17.

Polycomb repressive complex 2 (PRC2), which plays an important role in transcriptional repression of many developmental regulatory genes via trimethylation of histone H3 lysine 27 (H3K27me3), is essential for embryonic development 18-20, ESC pluripotency 21-23, and iPSC reprogramming 14, 24-27. One of the unresolved aspects regarding PRC2 function is the mechanism by which PRC2 core proteins (SUZ12, EED, and EZH2) are recruited to specific gene target sites. Recent findings suggest that PRC2 accessory proteins, long noncoding RNAs (lncRNAs), and DNA‐binding transcription factors function as a scaffold for PRC2 core proteins 28. Jumonji, AT rich interactive domain 2 (JARID2; also known as JUMONJI), one of PRC2 accessory proteins, is the founding member of the Jumonji C domain‐containing protein family of histone demethylases. Despite being catalytically inactive, JARID2 can repress target gene expression by regulating H3K27 methyltransferase activity and target gene occupancy of PRC2 in ESCs 29-33. Similar to the impairment of PRC2 function, knockdown (KD) or knockout of Jarid2 has been shown to impair reprogramming from mouse and human fibroblasts to iPSCs 14, 27. Accumulating evidence has revealed that JARID2 plays important roles in regulating PRC2 function in ESCs. However, the biological role and significance of elevated expression of Jarid2 during iPSC reprogramming remain elusive.

In this study, we show that forced expression of JARID2 accelerates iPSC reprogramming in a cell proliferation‐independent manner. Our data showed that JARID2 suppressed the expression of genes, including Arf and Meox2, which function as a barrier against the reprogramming process 34, 35. Furthermore, JARID2, together with OSKM, rapidly induced Nanog promoter demethylation and facilitated silencing of retroviral transgenes. Finally, we demonstrate that OSKM‐mediated iPSC reprogramming can be dramatically improved, both in the efficiency and kinetics, by the combined expression of JARID2 and the genes encoding its associated proteins. Taken together, this report provides evidence that a JARID2‐associated protein network plays a critical role in overcoming barriers against reprogramming of somatic cells to iPSCs.

Materials and Methods

Cell Culture

Mouse embryonic fibroblasts (MEFs) and tail‐tip fibroblasts (TTFs) were isolated from 13.5 days postcoitum embryos and adult mice with a Nanog promoter‐driven EGFP‐IRES‐Puromycin resistance gene reporter cassette (STOCK Tg (Nanog‐Gfp, Puro) 1Yam, RIKEN BioResource Center, Japan), respectively 36. Two packaging cell lines, PLAT‐GP and PLAT‐E 37, were kindly provided by Dr. Kitamura (The University of Tokyo, Japan). MEFs, TTFs, PLAT‐E, PLAT‐GP, and 293FT (Invitrogen, Carlsbad, CA, http://www.invitrogen.com) cells were cultured in Dulbecco's modified Eagle's medium (DMEM; Nacalai Tesque, Kyoto, Japan, http://www.nacalai.co.jp) supplemented with 10% heat‐inactivated fetal bovine serum (Invitrogen) and 1% non‐essential amino acids (NEAA; Invitrogen). Human dermal fibroblasts (HDFs) were purchased from Toyobo (Osaka, Japan, http://lifescience.toyobo.co.jp) and cultured in fibroblast basal medium with growth supplements (Toyobo). Mouse ESCs and iPSCs were maintained on gelatin‐coated dishes/plates in mouse ESC medium (DMEM containing 20% knockout serum replacement (KSR; Invitrogen), 1% NEAA, 1% GlutaMAX (Invitrogen), 0.1 mM 2‐mercaptoethanol (2‐ME; Invitrogen), and leukemia inhibitory factor‐containing conditioned medium supplemented with two chemical inhibitors (2i), 3 μM CHIR99021 (Axon Medchem BV, Groningen, The Netherlands, http://www.axonmedchem.com/) and 1 μM PD0325901 (Axon Medchem BV). Human iPSCs were maintained on mitomycin C‐treated SNL feeder cells in human ESC medium (DMEM/F12 (Invitrogen) supplemented with 20% KSR, 2.5 μg/ml basic fibroblast growth factor (PeproTech, Rocky Hill, NJ, http://www.peprotech.com), 1% NEAA, 1% GlutaMAX, and 0.1 mM 2‐ME). All cells were cultured in a 5% CO₂ atmosphere at 37 °C.

Mouse iPSC Reprogramming

Mouse iPSC reprogramming was performed with the KSR‐based ESC medium that supports the reprogramming process and pluripotency of iPSCs. One day before transduction, MEFs and TTFs were seeded onto gelatin‐coated six‐well plates at a density of 4 × 10⁴ cells per well. The cells were transduced with a pMXs retrovirus mixture (OSK or OSKM combined with individual genes) containing 4 µg/ml polybrene (Sigma‐Aldrich, St. Louis, MO, http://www.sigmaaldrich.com). The following day, media were replaced with mouse ESC medium. Reseeding of MEFs at 4 days after transduction was performed only for primary screening of 39 selected genes. Puromycin selection (1.5 µg/ml) was started at the indicated time points until 12 days post‐transduction (dpt), and then the number of green fluorescent protein (GFP⁺) colonies was counted under a fluorescence microscope. To establish mouse iPSC clones (#A1–12), puromycin‐resistant GFP⁺ colonies were picked and expanded on a gelatin‐coated dish in mouse ESC medium containing 2i and 1.5 µg/ml puromycin. To obtain homogeneous colonies, we picked individual GFP⁺/PuroR⁺ colonies derived from each primary GFP⁺ colony that was replated at cloning density (shown as clone #1–6). For reprogramming using OSK‐inducible secondary MEFs 38, the cells were treated with 2 µg/ml doxycycline at the time of transduction, and then alkaline phosphatase (AP) staining was performed at 12 dpt. Other methods conducted in this study, including vector construction, virus preparation, human iPSC reprogramming, and statistical analysis are provided in the Supporting Information Materials and Methods.

Results

JARID2 Promotes Mouse and Human iPSC Reprogramming

To identify potent reprogramming‐promoting factors, we selected highly expressed genes in both mouse ESCs and oocytes, compared with those in somatic tissues, by using the BioGPS microarray database (Mouse GeneAtlas GNF1M). Box plots of the expression levels for the 39 selected genes showed an obvious trend of high expression specifically in ESCs and oocyte, compared with most of the other cell types (Supporting Information Fig. 1). We then evaluated whether transduction of these selected genes promotes iPSC reprogramming by OSKM from MEFs carrying a Nanog promoter‐driven GFP reporter gene and puromycin resistance gene (PuroR). Our data revealed that JARID2, DDX4 (also known as MVH), DNA2l, MSH2, MYST4, KDM2B 6, 8, SALL4A 39, and RING1B (also known as RNF2) significantly increased the number of GFP⁺ colonies from OSKM‐transduced MEFs, compared with that from empty vector control plus OSKM‐transduced MEFs (Supporting Information Fig. 2A–2C). Of note, JARID2 40-44, DNA2l 45, and MSH2 46, have been proposed to suppress tumor formation. Considering that the promoting activity for iPSC reprogramming by JARID2 was consistently observed across all three different experimental settings (Supporting Information Fig. 2A–2C) and the effect of JARID2, but not those of DNA2L and MSH2, was apparent even at the earlier time point (7 dpt) (Supporting Information Fig. 2B), we decided to further investigate the significance of forced expression of JARID2 on iPSC reprogramming.

To begin, we showed that endogenous Jarid2 mRNA was gradually upregulated from early to late stages during reprogramming to iPSCs (Supporting Information Fig. 2D). We confirmed that forced expression of JARID2 promoted GFP⁺/PuroR⁺ colony formation not only with MEFs but also with adult mouse TTFs (Fig. 1A–1C). In addition, an increase in the ratio of GFP⁺ cells to total cells was demonstrated by flow cytometry using MEFs (Fig. 1D). Furthermore, the promoting effect of JARID2 on iPSC reprogramming was observed in MEFs transduced with OSK plus N‐MYC, L‐MYC, or several mutants of c‐MYC (Fig. 1E, 1F). Two different types of experiments showed that JARID2 accelerated OSK‐mediated iPSC reprogramming, but the promoting effect was much less than that observed in OSKM‐mediated iPSC reprogramming, implying that there are c‐MYC‐dependent effects of JARID2‐induced iPSC reprogramming (Fig. 1G, 1H). These data were slightly different from the conclusion reported by Zhang et al. 27 in which forced expression of JARID2 failed to facilitate OSK‐mediated iPSC reprogramming. However, considering the Dox‐inducible secondary MEFs reprograming system (Fig. 1H) demonstrated that the promoting effect by forced expression of JARID2 was statistically significant, we believe that JARID2 can accelerate OSK‐mediated iPSC reprogramming. Of note, Zhang et al. 27 also suggested some promoting activity of JARID2 in OSK‐mediated reprogramming, although that had been regarded as a negligible change. Using a set of deletion mutants of JARID2, we found that the N‐terminal half of JARID2 (J2ΔC) was sufficient and required to enhance iPSC reprogramming with MEFs (Fig. 1I, 1J). This region contains a nuclear localization signal (NLS), a transcriptional repressor domain that participates in binding to PRC2 core proteins (Supporting Information Fig. 2E) 31, and a region potentially required for binding to nucleosomes and lncRNAs 32, 33, 47, 48. Furthermore, both wild type (WT) and ΔC mutant JARID2 also facilitated iPSC reprogramming with HDFs (Fig. 1K).

Figure 1

JARID2 facilitates mouse and human induced pluripotent stem cell (iPSC) reprogramming. (A): Scheme depicting induction of reprogramming from somatic cells containing a Nanog promoter‐driven GFP and PuroR cassette to iPSCs. (B): Relative number of GFP⁺/PuroR⁺ colonies derived from OSKM‐transduced MEFs with puromycin treatment from 6 or 8 dpt to 12 dpt. *, p < .05 compared with empty control (mean + SD; n = 3). (C): Relative number of GFP⁺/PuroR⁺ colonies derived from mouse TTFs. *, p < .05 compared with empty control (mean + SD; n = 3). (D): FCM analysis showing the ratio of GFP⁺ cells to total cells at 6 and 8 dpt with or without GSK‐3 inhibitor CHIR treatment. *, p < .05 compared with empty control either with or without CHIR (mean + SD; n = 3). (E): Relative number of GFP⁺/PuroR⁺ colonies derived from MEFs transduced with OSK plus N‐Myc or L‐Myc. *, p < .05 compared with empty control (mean + SD; n = 3). (F): Upper: Scheme of deletion mutants of HA‐tagged c‐MYC; Lower: Relative number of GFP⁺/PuroR⁺ colonies derived from MEFs transduced with OSK plus each c‐MYC mutant. *, p < .05 compared with empty control; NS, not significant (mean + SD; n = 3). W136E, V394D, and L420P c‐MYC mutants have been reported to be deficient for transformation activity, Miz1 binding, and Max binding, respectively 75. (G): Relative number of GFP⁺/PuroR⁺ colonies derived from OSK‐transduced MEFs. *, p < .05 compared with empty control (mean + SD; n = 3). (H): Relative number of AP⁺ colonies derived from MEFs containing a doxycycline‐inducible OSK cassette. *, p < .05 compared with empty control (mean + SD; n = 3). (I): Upper: Scheme of deletion mutants of Flag‐tagged JARID2; Lower: Relative number of GFP⁺/PuroR⁺ colonies derived from MEFs transduced with OSKM plus each indicated Jarid2 mutant. *, p < .05 compared with empty control (mean + SD; n ≥ 3). (J): Upper: Scheme of constructs for 2A peptide sequence‐linked c‐Myc and Jarid2 coexpression system; Lower: Relative number of GFP⁺/PuroR⁺ colonies derived from MEFs transduced with OSK plus 2A peptide‐linked c‐MYC and JARID2. *, p < .05 compared with c‐MYC alone (mean + SD; n = 3). (K): Number of TRA‐1‐60⁺ colonies derived from HDFs. *, p < .05 compared with empty control (mean + SD; n = 3). Abbreviations: ARID, AT‐rich interactive domain; DMEM, Dulbecco's modified Eagle's medium; Dox, doxycycline; FCM, flow cytometry; GFP, green fluorescent protein; HDFs, human dermal fibroblasts; ; JmjN, Jumonji N domain; JmjC, Jumonji C domain; KSR, knockout serum replacement; LIF, leukemia inhibitory factor; mESC, mouse embryonic stem cell; NBR, nucleosome‐binding region; NLS, nuclear localization signal; OSKM, OCT4, SOX2, KLF4, and c‐MYC; RBR, RNA‐binding region; TTFs, tail‐tip fibroblasts; TR, transcriptional repressor; WT, wild type; ZF, zinc finger.

The J2ΔC is known to be sufficient to regulate PRC2 occupancy and activity through physical interactions with core proteins 31, 47, 48. PRC2 plays important roles in counteracting cellular senescence of mouse and human fibroblasts induced by INK4a/ARF, which is a protein that functions as a potent reprogramming barrier 24, 49. Thus, we investigated whether JARID2 regulates Arf expression. We found that Arf expression was elevated by KD of endogenous Jarid2 expression in MEFs at 3 dpt (Fig. 2A). Consistent with this finding, reverse transcriptase quantitative polymerase chain reaction (RT‐qPCR) showed that forced expression of JARID2 suppressed Arf expression in MEFs (Fig. 2B). Because ARF is known to stabilize p53 protein by antagonizing MDM2‐mediated degradation 50-52, we examined Trp53 gene expression at mRNA and protein levels. As expected, p53 protein level was significantly decreased by forced expression of WT or ΔC mutant JARID2, without obvious effect on p53 mRNA level (Supporting Information Fig. 3A, 3B). Moreover, KD of endogenous Jarid2 increased the p53 protein level in spite of its decline in mRNA level (Supporting Information Fig. 3C, 3D). The expression level of Mdm2 mRNA was not noticeably modulated by JARID2 overexpression (OE), but Mdm2 mRNA was upregulated by Jarid2 KD (Supporting Information Fig. 3B). Notably, JARID2 inhibited cell proliferation during iPSC reprogramming despite its ability to decrease Arf expression (Fig. 2C). We also found that Jarid2 KD in MEFs resulted in a decline in the efficiency of iPSC reprogramming by OSK with N‐MYC, L‐MYC, or c‐MYC (Fig. 2D). This inhibitory effect by Jarid2 KD was rescued by ectopic expression of MDM2 or a dominant‐negative form of p53 (p53DD), as well as WT or a ΔC mutant of JARID2 (Fig. 2E). We also demonstrated that the reprogramming‐promoting effect of JARID2 was much less under OE of MDM2 or p53DD (Fig. 2F). These results indicate that one of the mechanisms for the promotion of iPSC reprogramming by JARID2 is inhibition of the ARF‐MDM2‐p53 pathway.

Figure 2

Promotion of induced pluripotent stem cell (iPSC) reprogramming by JARID2 is associated with suppressing the expression of Arf and development‐associated genes. (A): Relative expression levels of Jarid2, Arf, p15, and p21 in mouse embryonic fibroblasts (MEFs) transduced with two different shRNAs against Jarid2 at 3 dpt. *, p < .05 compared with shEmpty (mean + SD; n = 3). (B): Relative expression levels of Arf, p15, and p21 in MEFs transduced with WT or a ΔC mutant of Jarid2 at 3 dpt. *, p < .05 compared with empty control (mean + SD; n = 3). (C): Relative number of MEFs transduced with OSKM plus JARID2 at 4, 6, and 8 dpt. *, p < .05 compared with empty control; NS, not significant (mean + SD; n = 3). (D): Number of GFP⁺/PuroR⁺ colonies derived from MEFs transduced with OSK with or without each MYC family member plus each shRNA against Jarid2. Lentiviral shRNA expression vectors were added in a small volume (1×, 10 µl per well) or a large volume (10×, 100 µl per well). Data are represented as the mean + SD; n = 3. (E): Number of GFP⁺/PuroR⁺ colonies derived from MEFs transduced with shJarid2 in combination with OSKM plus each JARID2 mutant, p53DD, or MDM2. *p < 0.05 compared with empty control with shEmpty (mean + SD; n = 3). (F): Relative number of GFP⁺/PuroR⁺ colonies derived from MEFs transduced with OSKM plus WT or a dominant negative mutant of p53 (p53DD) and MDM2 (mean + SD; n = 3). (G): GO analysis of microarray data for downregulated genes in JARID2‐transduced MEFs at 3 dpt. The top 10 GO terms are shown. (H): Reverse transcriptase quantitative polymerase chain reaction showing relative expression levels of development‐associated genes (Meox2, Sox6, and Mfap2) selected from genes identified in the GO analysis in (G). *, p < .05 compared with empty control (mean + SD; n = 3). (I): Relative number of GFP⁺/PuroR⁺ colonies derived from MEFs transduced with OSKM plus Eed, Ezh2, Suz12, Mtf2, or Phf1 (mean + SD; n = 4). Abbreviations: GFP, green fluorescent protein; GO, gene ontology; OSKM, OCT4, SOX2, KLF4, and c‐MYC; shRNA, short hairpin RNA; WT, wild type.

Next, we conducted gene ontology (GO) analyses of microarray data to gain an insight for the biological consequence of JARID2 OE in MEFs. The GO analyses revealed an association of genes downregulated by JARID2 OE with embryonic development (Fig. 2G; Supporting Information Table 2). RT‐qPCR confirmed that JARID2 OE repressed the expression of development‐associated genes (Meox2, Sox6, and Mfap2) in MEFs (Fig. 2H). These results suggest that the effect of JARID2 on promoting iPSC reprogramming is also mediated by inhibiting the expression of development‐associated genes. Unlike JARID2, forced expression of other PRC2 components did not promote iPSC reprogramming (Fig. 2I). Thus, the promoting activity for reprogramming is not a general feature among PRC2 components, but rather is a specific function of JARID2.

JARID2 Facilitates Silencing of Klf4 Expression from Retroviral Transgene in Mouse iPSCs

To assess the effect of JARID2 on iPSC quality, we obtained six mouse iPSC clones from MEFs by retroviral transduction of OSKM plus Jarid2 (Jarid2‐miPSCs) (Supporting Information Fig. 4A, 4B). We did not observe any noticeable differences in the expression levels of Gtl2, Nanog, and Rex1 genes, nor the intensity of Nanog‐promoter‐driven GFP fluorescence between Jarid2‐miPSC and empty‐control iPSC (Emp‐miPSC) clones (Fig. 3A; Supporting Information Fig. S4C). The global expression profile of Jarid2‐miPSC clones was highly similar to that of ESCs, but not to that of MEFs (Fig. 3B, 3C). Silencing of retroviral transgenes a hallmark of completely reprogrammed iPSCs. To clarify the effect of JARID2 OE on silencing of retroviral transgenes, we next examined the expression levels of total (endogenous plus retroviral) and endogenous OSKM in Emp‐miPSC and Jarid2‐miPSC clones by RT‐qPCR. Retroviral transgene expressions of OSM appeared to be highly repressed in all Emp‐miPSC and Jarid2‐miPSC clones examined (Fig. 3D). Intriguingly, we observed silencing of the retroviral Klf4 transgene in all of the six Jarid2‐miPSC clones, but not in four out of the six Emp‐miPSC clones (Fig. 3D, 3E). A significant difference in exogenous Klf4 silencing was recapitulated in other clones (Supporting Information Fig. 4D, 4E). Previous studies have demonstrated that residual expression of retroviral transgenes in early passage iPSCs was silenced after several passages, which we believe might explain why exogenous Klf4 expression was silenced after six additional passages in two Emp‐miPSC clones (Fig. 3F).

Figure 3

miPSC clones generated by OCT4, SOX2, KLF4, and c‐MYC (OSKM) plus JARID2 show early transgene silencing. (A): Relative expression level of Nanog, Rex1, and Gtl2 (mean + SD; n = 3). (B, C): Scatter plots of global gene expression comparing Jarid2‐miPSC clone #3 or #4 with mouse ESCs (B) or MEFs (C). Gene expression profiles were analyzed by DNA microarray. (D): Relative expression levels of total and endogenous OSKM genes in each Emp‐ and Jarid2‐miPSC clone #1–6 (mean + SD; n = 3). (E): Relative ratio of endogenous to total Klf4 expression in miPSC clones shown in (D). (F): Relative expression level of Esrrb and the relative ratios of endogenous to total Oct4 and Klf4 expression after an additional six passages. (mean + SD; n = 3). (G, H): Reverse transcriptase quantitative polymerase chain reaction (G) and immunofluorescence staining (H) demonstrating that cells derived from EBs of Jarid2‐miPSC clones expressed each marker for the three germ layers. Data are represented as the mean + SD; n = 3. Scale bar = 100 µm. (I): Chimeric mice generated from Jarid2‐miPSC clone #2, #4, and #5. Abbreviations: EB, embryoid body; ESC, embryonic stem cell; MEF, mouse embryonic fibroblast; miPSC, mouse induced pluripotent stem cell; Un, undifferentiated.

Similar to Emp‐miPSC clones, Jarid2‐miPSC clones possessed the capacity to differentiate into all three germ layers in vitro (Fig. 3G, 3H). To validate the pluripotency of Jarid2‐miPSC clones more stringently, we conducted in vivo differentiation analyses (i.e., chimeric mouse and germline transmission analyses) and found that Jarid2‐miPSCs contributed to generating chimeric and germline‐transmitted mice (Fig. 3I). Taken together, these results suggest that JARID2 enhances the reprogramming to genuine iPSCs and/or may improve the quality of iPSCs.

JARID2 Physically Interacts with and Regulates the Activity of c‐MYC

To understand the molecular function of JARID2 on iPSC reprogramming, we next examined whether JARID2 physically interacts with OSKM and NANOG proteins and found that JARID2 associated with SOX2, KLF4, and c‐MYC (Fig. 4A; Supporting Information Fig. 5A–5C). Physical interaction between JARID2 and c‐MYC supports the notion that JARID2 effectively accelerated iPSC reprogramming in a c‐MYC‐dependent manner. Thus, we focused on the relevance of JARID2 to c‐MYC function. Coimmunoprecipitation combined with Western blot analyses (CoIP‐WB) showed that at least two domains, the N‐terminal NLS and the C‐terminal half, of JARID2 were independently involved in the interaction with c‐MYC (Supporting Information Fig. 5D). Conversely, the C‐terminal basic‐rich and helix‐loop‐helix domains and the central portion of c‐MYC were required for the interaction with JARID2. (Supporting Information Fig. 5E). Unexpectedly, the total amount of HA‐tagged c‐MYC protein was reduced in the presence of J2ΔC mutant for an unknown reason (Supporting Information Fig. 5D, lane 4). As a result, we were unable to show an apparent band of c‐MYC protein in the immunoprecipitated fraction with J2ΔC (Supporting Information Fig. 5D, lane 13). Therefore, we repeated the CoIP‐WB analyses with a larger amount of cell extracts and confirmed the interaction between c‐MYC and J2ΔC mutant (Supporting Information Fig. 5F). Additionally, we found that c‐MYC interacted with OCT4, NANOG, and SUZ12 (Fig. 4B; Supporting Information Fig. 5G), implying a protein interaction network between c‐MYC, SUZ12, and JARID2.

Figure 4

JARID2 physically interacts with c‐MYC and regulates the expression levels of c‐Myc‐regulated transcripts. (A): CoIP‐WB analysis of HA‐tagged JARID2 with Flag‐tagged c‐MYC. (B): CoIP‐WB analysis of Flag‐tagged c‐MYC with HA‐tagged SUZ12. (C): Venn diagram showing co‐occupancies of JARID2 (blue), c‐MYC (red), and SUZ12 (green) in mouse ESCs. (D): Occupied regions of JARID2 (blue), SUZ12 (green), and c‐MYC (red) from chromatin immunoprecipitation deep sequencing data in the Ink4a/Arf locus. (E, F): GO term (E) and KEGG pathway (F) enrichment analyses of 274 genes co‐occupied by JARID2, c‐MYC, and SUZ12 in mouse ESCs. (G): Left: Relative expression levels (log₂ ratio) of JARID2/c‐MYC/SUZ12‐occupied genes in ESCs compared with those in MEFs. Red and blue lines represent genes with greater than 1.75‐fold higher (ESC‐high) or lower (ESC‐low) expression in ESCs, respectively. Right: Relative expression levels of ESC‐high, ESC‐low, and other genes among JARID2/c‐MYC/SUZ12‐occupied genes in MEFs transduced with either c‐Myc alone or both c‐Myc and Jarid2 compared with those in the empty control. (H): Box plots showing relative expression levels of ESC‐high, ESC‐low, and other genes among JARID2/c‐MYC/SUZ12‐occupied genes (all or selected genes represented by GO term and KEGG pathway with development and cancer, respectively) in MEFs transduced with both c‐Myc and Jarid2 in comparison with c‐Myc alone. (I): Relative expression level of Arf in MEFs transduced with the indicated genes (mean + SD; n = 3). Abbreviations: ESC, embryonic stem cell; GO, gene ontology; HA, hemagglutinin; IP, immunoprecipitation; KEGG, Kyoto Encyclopedia of Genes and Genomes; MEF, mouse embryonic fibroblasts; WB, Western blot.

Using publicly available chromatin immunoprecipitation‐deep sequencing datasets 30-32, 53, we found that approximately 5% (274 genes including Arf) of JARID2 targets were co‐occupied by c‐MYC and SUZ12 in mouse ESCs (Fig. 4C, 4D; Supporting Information Table 3). Notably, JARID2/c‐MYC/SUZ12‐co‐occupied genes that exhibited enrichments for GO terms and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways associated with development and cancer, respectively (Fig. 4E, 4F). Bioinformatics analyses of microarray data showed that forced expression of c‐MYC caused either activation or suppression of JARID2/c‐MYC/SUZ12‐co‐occupied gene transcription in a target gene‐dependent manner in MEFs (Fig. 4G). ESC‐low genes showing lower expression in ESCs than that in MEFs tended to show lower expression in MEFs after simultaneous expression of c‐MYC and JARID2, compared with that in MEFs with forced c‐MYC expression alone (Fig. 4H). Likewise, co‐expression of JARID2 with c‐MYC tended to upregulate expression levels of ESC‐high genes showing higher expression in ESCs than that in MEFs (Fig. 4H). These findings raise the possibility that JARID2 is the regulator determining whether JARID2/c‐MYC/SUZ12‐co‐occupied genes show relatively higher or lower expression in ESCs than that in MEFs. It is noteworthy that the expression levels of Arf gene (one of the ES‐low genes) were elevated owing to forced expression of c‐MYC in MEFs, which is consistent with the notion that Arf is one of the representative genes subjected to positive regulation by c‐MYC 52. However, JARID2 alleviated Arf expression induced by c‐Myc OE (Fig. 4I). These results suggest that JARID2 may function as a fine regulator of c‐MYC activity via a physical interaction.

To understand the molecular feature of JARID2‐mediated regulation of expression levels of the co‐occupied target genes, we conducted gene set enrichment analysis (GSEA) between the JARID2/c‐MYC/SUZ12‐co‐occupied genes and a repressive histone mark H3K27me3 in ESCs or MEFs. The GSEA showed that the ESC‐low genes among the co‐occupied genes were associated with H3K27me3 in ESCs, but not MEFs, whereas the ESC‐high genes among the co‐occupied genes were associated with H3K27me3 in MEFs, but not in ESCs (Supporting Information Fig. 6A, 6B). Together with the evidence that Jarid2 gene is highly expressed in ESCs, we speculate that the ESC‐low genes might be repressed via H3K27me3 by the PRC2 complex with JARID2, despite the presence of c‐MYC, in a cell context manner, such as with ESCs and JARID2‐transduced MEFs. In contrast, the ESC‐high genes might be activated by c‐MYC due to inhibiting PRC2‐mediated H3K27me3 in such cell context via an undetermined factor. One possible mechanism modulating JARID2 function to promote or inhibit PRC2 activity in a target gene‐dependent manner is the methylation of JARID2 protein at lysine 116, which promotes PRC2 enzymatic activity, as shown by a recent report 54, although it remains to be determined in future studies.

Combined Transduction of Jarid2, Prdm14, Esrrb, and Sall4a Dramatically Enhances the Efficiency and Kinetics of iPSC Reprogramming

To achieve a more robust enhancement of iPSC reprogramming, we tested potential additional effects of novel JARID2‐associated proteins on reprogramming by OSKM and JARID2. We found that JARID2 and ESRRB co‐occupied many genomic sites between −2,000 bp upstream and 500 bp downstream of the transcription start sites in mouse ESCs (Supporting Information Fig. 7A), and that JARID2 physically associated with ESRRB (Fig. 5A). Consistent with a previous report 55, ESRRB improved iPSC reprogramming in combination with SALL4, NANOG, and LIN28 (Supporting Information Fig. 7B). We also demonstrated that JARID2 physically associated with SALL4A and PRDM14 (Fig. 5A; Supporting Information Fig. 7C) 56. These JARID2‐associated proteins (PRDM14, ESRRB, and SALL4A; PrEsS) reciprocally coprecipitated each other (Supporting Information Fig. 7C) and synergistically promoted OSKM‐mediated reprogramming (Supporting Information Fig. 7B, 7D). The combination of PrEsS with OSKM or OSK plus 2A peptide‐dependent coexpression of c‐MYC and J2ΔC (OSKM‐J2ΔC) strongly increased the GFP⁺ cell population to 25%–31% and 43%–49% of total cells at 8 dpt, respectively (Fig. 5B). We also found that the promoting effect of PrEsS on the generation of the GFP⁺ cell population was augmented by a glycogen synthase kinase‐3 inhibitor CHIR99021 (CHIR) or 2i condition using a mitogen‐activated protein kinase kinase inhibitor PD0325901 (PD03) in addition to CHIR (Fig. 5B). Moreover, our data suggest that PRDM14, ESRRB, or SALL4 alone do not have the ability to enhance OSKM‐ or OSKM‐J2ΔC‐mediated reprogramming, compared with the combined expression of these three factors (Supporting Information Fig. 7D, 7E). To clarify the effect of PrEsS on the reprograming kinetics, we next examined the time point at which Puro⁺ cells appears after transduction of PrEsS together with OSK plus c‐MYC, M‐J2ΔC, or L‐MYC. Our data revealed that PrEsS with either OSKM‐J2ΔC (Fig. 5C, 5D) or OSKM (Fig. 5D; Supporting Information Fig. 7F) accomplished iPSC reprogramming at a much earlier time point (within 2 days) compared with the controls, OSKM (6 days) or OSKM‐J2ΔC (4 days). Similarly, together with OSK plus L‐MYC, PrEsS facilitated the reprogramming process with respect to both efficiency and kinetics, although the promoting effect was relatively lower compared with that mediated by OSKM or OSKM‐J2ΔC (Fig. 5D; Supporting Information Fig. 7G), suggesting that c‐MYC, but not L‐MYC, has some domains that further facilitate iPSC reprogramming together with PrEsS.

Figure 5

JARID2‐associated proteins together with OSKM and JARID2 drive rapid and highly efficient induced pluripotent stem cell (iPSC) reprogramming. (A): CoIP‐WB analysis of HA‐tagged JARID2 with Flag‐tagged PRDM14, ESRRB, or SALL4A. (B): Flow cytometry showing the ratio of GFP⁺ cells to total cells at 8 dpt. *, p < .05 (mean + SD; n = 3). CHIR: 3 μM CHIR99021; 2i: CHIR and 1 μM PD0325901. (C): Number of GFP⁺/PuroR⁺ colonies generated by OSKM‐J2DC with or without PrEsS. Puromycin selection was started at the indicated days (1, 2, 3, 4, 5, or 6 dpt), and then GFP⁺/PuroR⁺ colonies were counted at 12 dpt. Data are represented as the mean ± SD; n = 3. (D): Schematic representation indicating the appearance of GFP⁺/PuroR⁺ cells based on the day of puromycin administration that allowed GFP⁺/PuroR⁺ cells to specifically grow to form colonies after puromycin selection as shown in (C) and Supporting Information Fig. 7F, 7G. (E): Hierarchical clustering of microarray data using the expression level of the ES core module in ESCs and MEFs transduced with the indicated genes at 3 dpt. (F): Relative expression levels of early‐ and late‐phase marker genes in MEFs transduced with OSKM or OSKM‐J2ΔC plus PrEsS at 3 dpt. *, p < .05 (mean + SD; n = 3). (G): Relative expression levels of early‐ and late‐phase marker genes in Jarid2 KD MEFs transduced with OSKM or OSKM‐J2ΔC plus PrEsS at 3 dpt. *, p < .05 (mean + SD; n = 3). (H): Bisulfite sequencing showing the DNA methylation status of the proximal Nanog promoter in MEFs transduced with OSKM or OSKM‐J2ΔC plus PrEsS at 3 and 5 dpt. White and black circles represent unmethylated and methylated cytosines, respectively. The percentage of cytosine methylation is shown at the bottom. Abbreviations: ESC, embryonic stem cell; GFP, green fluorescent protein; HA, hemagglutinin; IP, immunoprecipitation; MEF, mouse embryonic fibroblasts; OSKM, OCT4, SOX2, KLF4, and c‐MYC; PrEsS, PRDM14, ESRRB, and SALL4A; WB, Western blot.

Hierarchical clustering revealed that PrEsS regulated expression of the ES Core module 57 to a level similar to that in ESCs (Fig. 5E; Supporting Information Fig. 8A). RT‐qPCR confirmed that PrEsS not only accelerated the regulation of early‐phase genes (Arf, Thy1, and Fut9) but also accelerated the reactivation of some, but not all, pluripotency genes, including late‐phase genes (Rex1, Nanog, and endogenous Sox2) in MEFs with OSKM or OSKM‐J2ΔC at an early time point (3 dpt) (Fig. 5F; Supporting Information Fig. 8B). We further showed that iPSC reprogramming by OSKM‐J2ΔC plus PrEsS was even more augmented by the addition of chemicals (either CHIR or forskolin) or forced expression of genes that are highly expressed in ESC/oocyte (either Kdm2b, Ncoa3, or Wdr5) (Supporting Information Fig. 8C, 8D). The effect of PrEsS on the expression of early‐ and late‐phase genes was alleviated by Jarid2 KD, especially the levels of endogenous Nanog and Sox2 (Fig. 5G), suggesting that the reprogramming‐promoting activity of PrEsS is dependent, at least in part, on endogenous JARID2.

DNA methylation at regulatory regions of pluripotency genes, including Nanog, functions as a strong barrier against reprogramming of somatic cells to iPSCs 1. Ectopic expression of OSKM alone cannot reactivate pluripotency genes until the late stage of iPSC reprogramming, mainly because pluripotency gene promoters are fairly resistant against demethylation during the iPSC induction process (Fig. 5H) 1, 2. Since we found that PrEsS, together with OSKM, were able to induce Nanog expression, even at the early stage of induction (Fig. 5H), we investigated the possibility that PrEsS augments Nanog gene expression by facilitating demethylation of the Nanog promoter. Unexpectedly, we found that PrEsS by themselves did not reduce, but elevated to some extent, methylation levels of the Nanog promoter at the early phase of OSKM‐mediated reprogramming (3 and 5 dpt) (Fig. 5H; Supporting Information Discussion). However, our data suggest that, unlike PrEsS, JARID2 facilitate the demethylation process at the early phase of OSKM‐ or OSKM+PrEsS‐mediated reprogramming (Fig. 5H; Supporting Information Discussion). Finally, we demonstrated that human iPSCs generated by PrEsS in combination with OSKM or OSKM‐J2ΔC exhibited a human ESC‐like morphology, pluripotency marker expression, AP activity, and differentiation properties (Supporting Information Fig. 9A–9D), indicating that human iPSCs, like mouse iPSCs, generated by OSKM or OSKM‐J2ΔC in combination with PrEsS, also have the appropriate pluripotent properties conferred. These data suggest that the combined expression of JARID2 and PrEsS is important to overcome barriers in OSKM‐mediated reprogramming.

Discussion

iPSC technology is a promising tool for cell‐based therapies in regenerative medicine. A challenge regarding standardization of the technique for generation of iPSCs is the identification of a gene set that can achieve rapid and efficient iPSC reprogramming. In this study, we found that JARID2 is capable of promoting iPSC reprogramming in a proliferation‐independent manner and the N‐terminal half of JARID2 is essential and sufficient to promote iPSC reprogramming. Importantly, this essential region for the promoting activity contains the functional domains required for binding to PRC2, nucleosomes 48, and lncRNAs 47, but not potential DNA‐binding domains 30, 58, 59. A recent study suggests that lncRNAs stimulate the interaction between JARID2 and PRC2, leading to the facilitated occupancy of these proteins at some gene targets 47. Therefore, JARID2 may promote iPSC reprogramming via a functional interaction with lncRNAs and PRC2.

To generate iPSCs for clinical application, factors capable of overcoming reprogramming barriers without oncogenic activity are desirable. Forced expression of the c‐Myc oncogene promotes the efficiency, kinetics, and quality of iPSC reprogramming by supporting genome‐wide occupancy of OSK on target genes, while also causing unwanted effects, including transformation and senescence 1. Thus, other than a factor capable of replacing c‐Myc, a regulator of c‐Myc function is assumed to be a candidate to generate safer iPSCs. In addition to their oncogenic properties, we and others have shown that c‐MYC and N‐MYC play crucial roles in sustaining ESC pluripotency and self‐renewal 60-64. Intriguingly, previous reports show that EZH2 physically interacts with c‐MYC and N‐MYC in human tumor cells 65, 66. Deregulation of EZH2 and SUZ12 have been found in many solid and hematological tumors 67. Moreover, c‐MYC has been demonstrated to regulate the expression of PRC2 components in tumors and pluripotent stem cells 68. These observations highlight the mechanistic link between MYC and PRC2 functions in tumorigenesis and maintenance of stem cell pluripotency. Supporting this concept, tumor suppressor functions of PRC2 have been demonstrated in a mouse model of Myc‐driven B‐cell lymphoma 69. Here, we showed that JARID2 also interacts with c‐MYC and effectively promotes iPSC reprogramming in the presence of c‐MYC. Additionally, JARID2 modulated expression changes by c‐Myc OE in MEFs. These findings imply that JARID2 also contributes to regulation of c‐MYC function via physical association together with PRC2 core proteins. A mutation or deletion in the JARID2 gene has been found in human nonsmall cell lung carcinoma and myelodysplastic/myeloproliferative malignancies 42-44. It has also been shown that Jarid2 is a target of oncogenic microRNAs 40, 41. These findings are of interest because our analysis showed KEGG pathway enrichment related to “cancer” among JARID2/c‐MYC/SUZ12‐occupied genes in ESCs. Taken together, we speculate that JARID2 might accelerate iPSC reprogramming by regulating c‐Myc‐driven transformation and Arf expression. Further studies regarding the role of JARID2 in c‐MYC function will provide important findings to control the tumorigenicity of ESCs and tumor development by deregulation of c‐Myc.

Our data also demonstrated that combined transduction of Jarid2, Prdm14, Esrrb, and Sall4a robustly enhanced iPSC reprogramming. PRDM14 is a PRC2‐ and JARID2‐binding protein that plays essential roles in maintenance of ESC identity, specification of primordial germ cells, and generation of iPSCs 56, 70-73. PRDM14 regulates the expression of gene targets that are co‐occupied by pluripotency factors, including ESRRB and NANOG, in mouse and human ESCs 56, 71, 72. ESRRB and SALL4A are OCT4‐interacting proteins that connect multiple pluripotency factors and pathways in ESCs 74. We showed the physical interaction of JARID2 with ESRRB and SALL4A, but not NANOG. Notably, PRDM14, ESRRB, and SALL4A were found to interact with one another. Furthermore, the relevance of this finding was supported by our bioinformatics analyses showing the high levels of co‐occupancy of JARID2 with SALL4 and ESRRB. Endogenous expression of genes encoding PrEsS does not become apparent until the late stage of OSKM‐mediated reprogramming processes. Therefore, the reprogramming‐promoting effect of these factors is not expected at the early phase of iPSC reprogramming. Combined expression of these genes drove rapid and efficient reprogramming with early induction of late‐phase genes, including endogenous Sox2 55, in a JARID2‐dependent manner, whereas expression of each factor alone had no or little effect on the efficiency of reprogramming compared with that induced by co‐expression of PrEsS. Similar to JARID2, PrEsS also effectively promoted iPSC reprogramming in combination with c‐MYC and repressed Arf expression in OSKM‐transduced MEFs, implying a possible connection between PrEsS and c‐MYC function. Therefore, elevated expression of the factors of a previously unrecognized JARID2‐associated protein network may be required to induce proper reprogramming to iPSCs.

Summary

Our findings provide evidence that forced expression of JARID2 improves the efficiency and kinetics of iPSC reprogramming in a proliferation‐independent manner, which may be explained by suppression of the ARF‐MDM2‐p53 pathway and developmental regulators, and the physical interaction of JARID2 with c‐MYC oncoprotein. In addition, we showed that JARID2 participates in an interaction network with PRDM14, ESRRB, and SALL4A, and that these four factors in combination with OSKM achieve rapid and efficient somatic cell reprogramming. Therefore, JARID2 and its associated factors may be promising candidate genes to overcome reprogramming barriers and shed light on the mechanisms underlying somatic cell reprogramming.

Acknowledgments

We thank Toshio Kitamura for providing the retroviral vector pMXs and packaging cell lines; Hiromitsu Nakauchi and Tomoyuki Yamaguchi for providing MEFs containing a doxycycline‐inducible OSK cassette; and Chiharu Shimizu, Yukari Tanizawa, Yukiko Yatsuka and Miyuki Katano for technical and material assistance. This work was supported in part by JST, CREST, and by MEXT, Support Project of Strategic Research Center in Private Universities. T.H. is currently affiliated with Gene Expression Laboratory, Salk Institute for Biological Studies, La Jolla, California, USA.

Author Contributions

H.I.: conception and design, collection and assembly of data, data analysis and interpretation, and manuscript writing; Y.N., T.H., Y.Y‐S., M.H., A.U., Y.T., S.I., F.S., K.‐i.Y., S.T.: collection and assembly of data and data analysis and interpretation; A.O.: data analysis and interpretation and financial support; Y.O.: data analysis and interpretation, financial support, administrative support, and final approval of manuscript

Disclosure of Potential Conflicts of Interest

The authors indicate no potential conflicts of interest.