EMBRYONIC STEM CELLS/INDUCED PLURIPOTENT STEM CELLS miR-290 Cluster Modulates Pluripotency by Repressing Canonical NF-jB Signaling PATRICK LU¨NINGSCHRO¨R,a BENEDIKT STO¨CKER,a BARBARA KALTSCHMIDT,a,b CHRISTIAN KALTSCHMIDTa a
Department of Cell Biology and bLaboratory of Molecular Neurobiology, University of Bielefeld, Bielefeld, Germany Key Words. NF-jB • Epithelial to mesenchymal transition • Pluripotency • MicroRNA-290 cluster • Embryonic stem cells
ABSTRACT Embryonic stem cell (ESC)-specific microRNAs (miRNAs) play a critical role in the maintenance of pluripotency and self-renewal but the complete network between these miRNAs and their broad range of target genes still remains elusive. Here we demonstrate that miR-290 cluster, the most abundant miRNA family in ESCs, targets the NF-jB subunit p65 (also
known as RelA) by repressing its translation. Forced expression of p65 causes loss of pluripotency, promotes differentiation of ESCs, and leads to an epithelial to mesenchymal transition. These data define p65 as a novel target gene of miR-290 cluster and provide new insight into the function of ESC-specific miRNAs. STEM CELLS 2012;30:655–664
Disclosure of potential conflicts of interest is found at the end of this article.
INTRODUCTION Recently, an important role has emerged for distinct embryonic stem cell (ESC)-specific microRNAs (miRNAs) in the control and induction of pluripotency [1–6]. Within this specific subset of miRNAs, the murine miR-290 cluster, identified as a 2.2-kb region on chromosome 7, is the most abundant miRNA family in ESCs [7]. The single primary transcript produced at this locus generates 14 mature miRNAs. Among these, miR-290-3p, miR-291a-3p, miR-291b-3p, miR-292-3p, miR-294, and miR-295 share the hexamer seed ‘‘AAGUGC.’’ The other miRNAs of the miR-290 cluster (miR-290-5p, miR291a-5p, miR-291b-5p, miR-292-5p, miR-293, miR-293*, miR-294*, and miR-295*) differ in their seed but are still highly expressed in ESCs with the exception of the hardly detectable [8] minor forms of miR-293, miR-294, and miR-295 (miR-293*, miR-294*, and miR-295*). Recent studies reported that miR-290 cluster regulates various important cellular processes by post-transcriptional inhibition of a broad range of different target genes [9–12]. Deficiency of this miRNA cluster in developing mice leads to a partially penetrant embryonic lethality and defective germ cells, indicating an important role for miR-290 cluster in early mouse development [13]. Here we report a translational repression of the nuclear factor kappa B (NF-jB) subunit p65 by miR-291b-5p and miR-293, which is mediated by targeting the coding sequence (cds) of p65. Overexpression of p65 resulted in a loss of pluripotency and differentiation of ESCs by an upregulation of markers for an epithelial to mesenchymal transition (EMT).
MATERIALS
AND
METHODS
Cell Culture, Differentiation, and Transfection J1 and Oct4-green fluorescent protein (GFP) ESCs were cultured on Mitomycin C-inactivated mouse embryonic fibroblasts (MEFs) in ESC medium consisting of Glasgow-minimum essential medium (Gibco, Grand Island, NY, http://www.invitrogen.com) supplemented with 15% fetal calf serum (FCS; Gibco), 1,000 U/ml leukemia inhibitory factor (LIF) (Millipore, Billerica, MA, http:// www.millipore.com), 1 nonessential amino acids (PAA Laboratories, Linz, Austria, http://www.paa.at), 2 mM L-glutamine (PAA), 1 penicillin/streptomycin (PAA), and b-mercaptoethanol. 46c ESCs (Sox1-GFP) were cultivated on gelatin-coated plates in ESC medium. For differentiation, ESCs were cultured in suspension as embryoid bodies (EBs) and treated at day 4 and day 6 after formation of EBs with retinoic acid (RA) (5 lM, Sigma-Aldrich, St. Louis, http://www.sigmaaldrich.com) [14]. For overexpression of p65, c-Rel, constitutive active form of IKK2 (IKK2-ca), or IjBa-AA1, ESCs were transfected with Lipofectamine 2000 (Invitrogen, Carlsbad, CA, http://www. invitrogen.com) using 800 ng of each plasmid. The empty vector was used as control. Briefly, 1 day prior to transfection, cells were plated at a density of 4 104 per centimeter square on gelatin-coated dishes without MEFs and transfected overnight. One day after transfection, 500–1,000 cells per centimeter square were plated on gelatin-coated plates and selected for the indicated time with puromycin (Sigma-Aldrich) at a concentration of 1.5 lg/ml. Mir-inhibitor (Qiagen, Hilden, Germany, http://www1.qiagen. com) experiments were performed using Hyperfect (Qiagen) with different concentrations (0, 10, 50, and 100 nM) of each inhibitor. One day prior to transfection, 5 104 cells per centimeter square were
Author contributions: P.L.: conception and design, collection and/or assembly of data, data analysis and interpretation, manuscript writing, and final approval of manuscript; B.S.: collection and/or assembly of data, data analysis and interpretation, and final approval of manuscript; B.K.: conception and design, data analysis and interpretation, and final approval of manuscript; C.K.: financial support, conception and design, data analysis and interpretation, manuscript writing, and final approval of manuscript. Correspondence: Christian Kaltschmidt, Prof. Dr., Department of Cell Biology, University of Bielefeld, Universit€atsstr. 25, 33501 Bielefeld, Germany. Telephone: þ495211065625; Fax: þ495211065654; e-mail:
[email protected] Received September C AlphaMed Press 15, 2011; accepted for publication December 19, 2011; first published online in STEM CELLS EXPRESS January 9, 2012. V 1066-5099/2012/$30.00/0 doi: 10.1002/stem.1033
STEM CELLS 2012;30:655–664 www.StemCells.com
miR-290 Cluster Represses NF-jB
656
plated on gelatin-coated 12-well plates without MEFs. For transfection, 6 ll Hyperfect (Qiagen) was used. For removal of MEFs, ESCs were preplated on nongelatinized plates for 45 minutes. HEK 293FT were cultivated in Dulbecco’s modified Eagle’s medium (DMEM, PAA) containing 10% FCS and transfected with Turbofect (Fermentas, http://www.fermentas.com, St. LeonRot, Germany). For coexpression of flag-tagged p65 (600 ng) with each miR-expression vector (2,400 ng), 1 105 cells per centimeter square were plated on six-well plates and directly transfected after plating. To study the effect of increasing amounts of miR-291b and miR-293, flag-p65 (100 ng) plus different amounts of the miR-expression vector (500, 1,000, 2000, and 3,000 ng) were cotransfected. NIH/3T3 cells were cultivated in DMEM containing 10% FCS and transfected with Amaxa Nucleofactor (Lonza, Cologne, Germany, http://www.lonzabio.com) using the Cell Line Kit R (program: U-30). Briefly, 1 106 cells were used per sample with a total amount of 5 lg DNA, which contained 2.5 lg of each miRNA expression vector. In all transfections, the total amount of DNA was kept constant by addition of the empty vector.
Reverse Transcription Polymerase Chain Reaction (RT-PCR) and Quantitative Polymerase Chain Reaction (qPCR) For quantification of mRNA, total RNA was prepared using Nucleo Bond RNA Kit (Macheray & Nagel, Dueren, Germany, http://www.mn-net.com/). Total RNA (1 lg) was treated with DNase I (Fermentas) and subsequently reverse transcribed with First Strand cDNA Synthesis Kit (Fermentas). Glyceraldehyde 3phosphate dehydrogenase (GAPDH) was used for normalization. For quantification of mature miRNAs, RNA was prepared using standard phenol–chloroform extraction. Total RNA (1 lg) was treated with DNase I (Fermentas) and reverse transcribed using miScript Reverse Transcription Kit (Qiagen). Amplification of mature miRNAs was performed using miScript Primer Assays (Qiagen), and data were normalized to U6 non coding RNA (ncRNA). For both applications, 1 ll of 1:5 diluted cDNA was used as template per reaction. All qPCR reactions were performed as triplicate using Platinum SYBR Green qPCR Super-Mix UDG (Invitrogen, Carlsbad, CA, http://www.invitrogen.com) and assayed with a Rotor Gene 6000 (Qiagen).
Plasmid Construction For overexpression of NF-kB subunits in ESCs, cDNAs for p65, c-Rel, inhibitor of kappa B kinase (IKK2)-ca, and ikBa-AA1 were PCR amplified from RcCMV expression vectors and cloned into pEF-Flag-Puro. BamHI and XbaI restrictions sites were used for cloning of p65 and ikBa-AA1. IKK2-ca and c-Rel were cloned into the EcoRV restriction site. miRNA expression vectors were generated by PCR amplification of miRNAs from their genomic locus. The amplified fragment contained the miRNA precursor and 150–200 bp of native genomic sequences flanking each side of miRNAs without overlap to other miRNA precursor. The PCR product was inserted into the EcoRV restriction site of pcDNA3.1(þ) (Invitrogen) and verified by sequencing. For construction of the p65-30 UTR-psicheck-2 reporter-construct, the murine p65 30 untranslated region (UTR) was PCR amplified and cloned into XhoI and NotI restriction sites of psiCHECK-2 (Promega, Madison, WI, http://www.promega.com) vector. For pLuc-MRE constructs, the sequences corresponding to predicted miRNA responsive elements (MREs) were synthesized as sense and antisense oligos, annealed, and cloned into XhoI and NotI restriction sites of psiCHECK-2 (Promega) directly 30 downstream of Renilla luciferase as described elsewhere [15]. The doxycyclin inducible vector FUW-tetO-p65-FPRed is based on FUW-tetO-hOct4 [16] (Addgene plasmid 20726). FUWtetO-hOct4 was EcoRI digested followed by a polylinker ligation
to exchange the hOct4 cds with a polylinker. The resulting vector, FUW-tetO-multiple cloning site (MCS) containing a set of unique restriction sites, was used to generate FUW-tetO-p65FPRed. The cds for p65:FPRed fused to FPRed was cloned into BamHI and XbaI sites of FUW-tetO-MCS.
Western Blot Cells were extracted in lysis buffer (1% SDS, 5 mM EDTA, complete protease inhibitor cocktail [Roche Diagnostics, Basel, Switzerland, p://www.roche-applied-science.com]) directly in the cell-culture-dish on ice. Subsequently, extracts were boiled for 5 minutes at 95 C. Protein amounts were determined by measurement of A280 using a Nanodrop Spectrophotometer. Equal amounts of protein extracts were separated on a 10% sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to nitrocellulose membranes (PALL). Membranes were blocked in phosphate buffered saline (PBS) with 5% milk powder and 0.05% Tween for 1 hour at 37 C. Blots were probed with primary antibodies overnight at 4 C and incubated with horseradish peroxidase-conjugated secondary antibodies the next day for 1 hour at room temperature. Western blots were carried out using p65 (Santa Cruz Biotechnology Inc., Santa Cruz, CA, http://www.scbt.com; sc-8008, 1:500), GAPDH (Santa Cruz; sc-32233, 1:4,000), Oct4 (Santa Cruz; sc-5279, 1:200), Pax6 (DSHB, Developmental Studies Hybridoma Bank; http://dshb.biology.uiowa.edu; 1:200), or Flag (SigmaAldrich; F7425, 1:8,000) primary antibodies. Western blots were developed using enhanced chemiluminescence. Densitometric quantification of Western blot films was carried out using the Image J (NIH) gel analysis routine (http://lukemiller. org/journal/2007/08/ quantifying-western-blots-without.html).
Immunohistochemistry For immunohistochemical stainings, cells were fixed with 4% paraformaldehyde for 15 minutes at room temperature, permeabilized with 0.1% Triton X-100 for 10 minutes, and blocked with 2% (Bovine serum albumin; Sigma-Aldrich) for 1 hour at room temperature. Incubation with the primary antibody was performed in 0.2% BSA for 2 hours at room temperature, followed by incubation with the appropriate Alexa-conjugated secondary antibody. Immunohistochemical stainings were carried out using smooth muscle actin (SMA) (Sigma-Aldrich; clone 1A4, 1:100), stage specific embryonic antigen 1 (SSEA1) (DSHB; MC-480, 1:100), and vimentin (DSHB; 40E-C, 1:200) primary antibodies. Pictures were captured at a Zeiss Axio Observer D1 microscope.
Lentivirus Production and Transduction Lentivirus production was performed by cotransfection of 293FT cells using calcium-phosphate precipitation. One day before transfection, 1 107 cells were plated on a 15-cm dish. The next day, cells were transfected with 50 lg of the transfer vector indicated, 37.5 lg D8.91 and 15 lg vesicular stomatitis virus-glycoprotein (VSV-G) helper plasmids [17]. After 16–24 hours, medium was changed, and 60–72 hours after transfection, the supernatant was harvested and stored at 80 C or used immediately for concentration by ultracentrifugation (50,000g, 2 hours, 4 C). One day prior to transduction, 2.5 104 cells per centimeter square were plated on gelatin-coated dishes without feeders. ESCs were transduced with concentrated lentivirus supernatants and an equal amount of ESC medium. 16 hours after transduction, fresh ESC media was added to each plate, and 24 hours after transduction, cells were passaged and expanded on MEFs.
Luciferase Reporter Assays For luciferase reporter assays, HEK 293FT cells were plated at a density of 2 105 cells per 24-well and transfected directly after plating. p65 30 UTR-psicheck-2 (50 ng) (Promega) was transfected along with 800 ng of the indicated miR-expression vector. Each assay was repeated a minimum of three times. For screening of potential miRNA binding to the p65 cds, 4 104 cells were plated 1 day before transfection in 96-well
Lu¨ningschrO¨r, StO¨cker, Kaltschmidt et al.
657
Figure 1. RA upregulates p65 expression. (A–C): Comparison of mRNA and protein expression of Oct4 (A), Pax6 (B), and p65 (C) during RAinduced differentiation of embryonic stem cells. Relative changes in the mRNA levels were measured by qPCR. Changes in protein levels were measured by quantification of Western blots. In both procedures, GAPDH was used for normalization. Mean 6 SEM of three independent experiments. Abbreviations: GAPDH, glyceraldehyde 3-phosphate dehydrogenase; qPCR, quantitative polymerase chain reaction; RA, retinoic acid.
plates. miRNA expression vector (60 ng) was cotransfected with 2 ng of the indicated pLUC-MRE vector. Each assay was repeated at least three times with four culture replicates. Luciferase activity was determined 24 hours after transfection with the Dual Luciferase Assay System (Promega) and normalized by the coexpressed firefly luciferase.
RESULTS Translational Repression of p65 by miR-290 Cluster During RA-induced differentiation of ESCs as EBs, changes in the mRNA and protein levels of Oct4, Pax6, and the NF-jB subunit p65 were assayed (Fig. 1A–1C). After treatment with RA, we were able to detect a marked decrease in the expression of Oct4 (Fig. 1A). Conversely, Pax6 and p65 expression increased during RA-induced differentiation of ESCs (Fig. 1B, 1C). In all cases, changes in the protein level correlated to the changes of the corresponding mRNA. Notably, the increase in the expression of p65 was more prominent at the protein level. This may suggest a translational repression of p65, which is probably accomplished by ESC-specific miRNAs. At the transcriptional level, other components of the NF-jB signaling pathway and markers for an EMT were also investigated (Supporting Information Fig. 1). Consistent with the upregulation of p65 during ESC differentiation, an increase in the expression of IKK2, p50, and IjBa could be observed after 8 days of differentiation (Supporting Information Fig. 1A). In contrast to the upregulation of components of the canonical NF-jB signaling pathway, p100/p52 and RelB, which functions in the noncanonical pathway, are downregulated during RAinduced differentiation (Supporting Information Fig. 1A). This is consistent with a recent report, describing potentially opposing roles for the canonical and noncanonical NF-jB signaling pathway in human ESCs [18]. The strong reduction in the expression of RelB and p100/p52 during differentiation observed here may indicate similar roles in mouse ESCs. Notably, an induction of the EMT markers Slug, Zeb1, Twist, and Vimentin was also detectable upon RA treatment, consistent with a massive decrease in the expression of E-cadherin (Supporting Information Fig. 1B). These data indicate that RA is able to induce an EMT during ESC differentiation. www.StemCells.com
In search of potential miRNAs targeting p65, candidate miRNAs have to be expressed in ESCs and must be downregulated upon differentiation. Various large-scale sequencing and expression profile studies produced a large dataset [8, 19], enabling bioinformatic exclusion of miRNAs not matching these criteria. Recently, miRNA binding to the cds without a canonical seed match was reported [20], prompting us to select a set of miRNAs that was expressed in ESCs with potential binding sites within the 30 UTR and cds of p65. Each of the selected miRNAs was coexpressed in HEK293 cells together with flag-tagged p65 (Fig. 2A) to assay the cds (Fig. 2B, 2C) or together with a luciferase-gene fused to the p65 30 UTR (Fig. 2A) to assay potential miRNA binding to the 30 UTR (Fig. 2D). A significant reduction of flag-p65 protein levels were caused by miR-291b and miR-293 (Fig. 2B, 2C), whereas no miRNA caused remarkable changes in the activity of the p65-30 UTR reporter (Fig. 2D). After coexpression of increasing amounts of miR-291b and miR-293 with flag-tagged p65 in HEK-293 cells, both miRNAs were able to reduce p65 expression in a dose-dependent manner (Fig. 2E, 2F). In contrast, miR-27b had no effect on the p65 expression, even at the highest plasmid ratio (Fig. 2G). For miR-293, potential MREs within the cds were validated. Binding site prediction by three different tools (RNA-Hybrid [21], RNA 22 [22], and probability of interaction by target accessibility (PITA) [23]) showed only a poor overlap in their prediction. Thus, three predictions of each tool and two overlapping predictions were tested (Supporting Information Table 1). Recently, it was reported that MREs from the cds are still functional within an artificial UTR [15, 22]. Potential MREs were therefore cloned directly downstream of a firefly-luciferase gene and tested for their ability to repress luciferase activity. Three of these MREs, which are conserved among different species (Supporting Information Table 2), were able to cause a significant downregulation of luciferase activity by miR-293 when compared with control miR-27 (Fig. 2H). Silent mutations introduced into these MREs were partially able to rescue the effect. Although all mutated MREs showed a diminished reduction of the luciferase activity, only the mutated form of MRE-102 and MRE-249 had a significant effect when compared with the wtMRE. The mutations introduced into the different MREs only affected the seed sequence within MRE-249 and MRE-600 by one and two substitutions, respectively. We cannot exclude that additional or other mutations would be able to completely
658
miR-290 Cluster Represses NF-jB
Figure 2. p65 is targeted by miR-290 cluster. (A): Scheme of the constructs used to screen for potential miRNA targeting of p65. (B): Representative Western blot analysis of flag-p65 protein levels after cotransfection of 293FT cells with the miRNA expression vectors indicated. (C): Quantification of Western blot analysis. Intensity of the flag signal is shown relative to the empty vector, normalized to GAPDH. Data represent mean 6 SEM of three independent experiments. (D): Cotransfection of 293FT cells with the miRNA expression vector indicated and a reporter construct encoding the Renilla luciferase cds fused to the p65 30 UTR. Renilla luciferase activity was normalized to firefly luciferase. Mean 6 SEM of three independent experiments. (E–G): Western blot analysis of p65 protein levels. miR-293 (E) and miR-291b (F) are able to reduce p65 protein levels in a dose-dependent manner after ectopic coexpression with flag-tagged p65 in 293FT cells. miR-27 (G) had no effect even at the highest plasmid ratio. (H): Validation of potential MREs within the p65 cds. Numbers indicate the position of MREs within the p65 cds in base pairs. Mean 6 SEM of three independent experiments each performed with four culture replicates. Two-tailed paired t test, *, p < 0.05; ***, p < 0.001. (I–K): Effect of the mutated forms of MRE 102 (I), MRE-249 (J), and MRE-600 (K) as measured by luciferase activity after cotransfection with miR-293. Two-tailed paired t test, **, p < 0.01; ***, p < 0.001. (L): Sequences of MRE-102, MRE-249, and MRE-600. Mutated nucleotides are highlighted in gray. Abbreviations: CMV, cytomegalovirus; cds, coding sequence; GAPDH, glyceraldehyde 3-phosphate dehydrogenase; miRNA, microRNA; MREs, miRNA responsive elements; pA, polyadenylation signal; SV40, Simian Virus 40; UTR, untranslated region; wt, wild type.
rescue the reduced luciferase activity. Nevertheless, our data suggest a direct binding by miR-293 to multiple, low affinity sites within the cds of p65. We also want to emphasize that the
poor overlap between the different prediction tools inclines us to assume that there are even more functional MREs within the cds, which were not identified by this study.
Lu¨ningschrO¨r, StO¨cker, Kaltschmidt et al.
659
Figure 3. Translational repression of endogenous p65 by miR-290 cluster. (A): Changes of p65 mRNA and protein levels after overexpression of the miRNA expression vectors indicated. Changes of mRNA levels were quantified by qPCR. Protein levels were quantified by Western blotting. One representative Western blot is shown. Two-way ANOVA; Bonferoni post hoc test. **, p < 0.01; ***, p < 0.001. (B, C): Expression of miR-291b-5p (B) and miR-293 (C) decreases during RA-induced embryonic stem cell (ESC) differentiation as assayed by qPCR. Mean 6 SD of two independent experiments. One-way ANOVA, Dunnett’s post hoc. **, p < 0.01; ***, p < 0.001. (D–F): p65 and Oct4 protein levels of J1 ESCs were assayed by Western blot after transfection of increasing amounts of miRNA inhibitors. A scr inhibitor (n ¼ 2) was used as control (D) and had no significant effect. Inhibition of miR-291-5p (n ¼ 3) (E) and miR-293 (n ¼ 3) (F) resulted in an upregulation of p65, whereas Oct4 protein levels decreased. Changes in protein levels were measured by quantification of Western blots. Mean 6 SD. Two-way ANOVA, Bonferoni post hoc. *, p < 0.01; **, p < 0.01; ***, p < 0.001. (D) Expression of miR-291b-5p and miR-293 decreases during RA-induced ESC differentiation as assayed by qPCR. Mean 6 SD of two independent experiments. One-way ANOVA, Dunnett’s post hoc. **, p < 0.01; ***, p < 0.001. Abbreviations: ANOVA, analysis of variance; GAPDH, glyceraldehyde 3-phosphate dehydrogenase; miRNA, microRNA; qPCR, quantitative polymerase chain reaction; RA, retinoic acid; scr, scrambled.
To test the ability of miR-291b and miR-293 to reduce endogenous p65 protein levels as well, miR-291b and miR293 expression vectors were transfected into NIH/3T3 cells. The empty vector and miR-27 were used as controls. The level of both pre-miRNAs in each sample was measured by qPCR to confirm the expression of miR-291b and miR-293 (Supporting Information Fig. 2). Overexpression of miR-291b and of miR-293 or of both miRNAs together resulted in a significant reduction of p65 protein levels, whereas mRNA levels almost remained unaffected (Fig. 3A). These results strengthen the suggestion that the interaction of the p65 mRNA with both miRNAs results in a translational repression rather than a degradation of the mRNA. The results so far were obtained using HEK 293FT and NIH/3T3 cells. Both cell lines provide an experimental environment devoid of the complex interaction network of ESCs and lacking the expression of ESC-specific miRNAs [19]. This supports our assumption of a direct inhibition of p65 by the miRNA-290 cluster but has no direct relevance www.StemCells.com
for ESCs. To address this issue, we quantified the expression of both miRNAs during ESC differentiation and applied miRNA inhibitors to undifferentiated ESCs. Since none of the miRNAs sharing the canonical AAGUGC seed were predicted to target p65, we investigated the mature form of miR-291b-5p, which does not share the canonical seed. Consistent with an increase in p65 protein levels during RA-induced differentiation (Fig. 1C), the expression of both miRNAs decreased significantly after treatment with RA (Fig. 3B, 3C). The effect of the indicated miRNA inhibitors (Fig. 3D–3F) was measured at the protein level for p65 and Oct4 by Western blot analysis. A scrambled control-inhibitor caused no significant changes (Fig. 3D), whereas an increase of endogenous p65 protein levels was observed after applying inhibitors for miR-291b-5p and miR-293 (Fig. 3E, 3F). This increase correlated with the amount of the inhibitor applied. In contrast, Oct4 protein levels decreased weakly upon miRNA inhibition (Fig. 3E, 3F).
660
miR-290 Cluster Represses NF-jB
Forced Expression of p65 Causes Loss of Pluripotency Our observations so far imply the question, whether forced expression of p65 could influence pluripotency of ESCs. To address this question, we used a doxycycline (DOX)-inducible lentivirus system to overexpress p65 fused to FPRed (p65FPRed). A vector expressing FPRed alone was used as control (Fig. 4A). ESCs with an Oct4::GFP genotype [24] were transduced with the vectors indicated and subsequently cultivated in the presence of DOX or left untreated (Fig. 4B, 4C). Cells expressing p65-FPRed lost their Oct4-driven GFP expression (Fig. 3F), whereas FPRed alone had no effect. In contrast to FPRed, we only observed a partial expression of p65-FPRed within individual colonies. This may reflect an inhibitory effect of the miR-290 cluster, which is partially able to repress p65-FPRed expression. Nevertheless, these data indicate that NF-jB signaling promotes differentiation of ESCs and therefore has to be repressed in the pluripotent state. Recent studies indicated that pluripotency is closely related to an epithelial phenotype [3, 25, 26]. Furthermore, it was reported that a mesenchymal to epithelial transition (MET) is required for reprogramming of fibroblasts. Notably, a function of NF-jB in promoting the opposite process, an EMT, was described in cancer progression [27–29]. This prompted us to investigate the expression of Snail, Slug, Twist-1, and vimentin, all markers for an EMT, and Pax6, the major neuroepithelial marker, as potential NF-jB target genes. Besides Pax6, all markers showed a significant upregulation after p65-FPRed overexpression (Fig. 4D), suggesting a potential role of NF-jB signaling in inducing an EMT during ESC differentiation and thereby causing loss of pluripotency.
Canonical NF-jB signaling causes differentiation of ESCs by promoting an EMT To further assess the impact of canonical NF-jB signaling on pluripotency and the epithelial character of ESCs, elongation factor 1 alpha (EF1a) driven, puromycin selectable vectors for the overexpression of flag-tagged p65, c-Rel, IKK2-ca, and IjBa-AA1 were constructed. Functionality of all vectors could be confirmed by NF-jB-driven luciferase expression (Supporting Information Fig. 3A), and expression of the different NF-jB subunits in ESCs could also be confirmed by Western blot analysis with an anti-Flag antibody (Supporting Information Fig. 3B). Oct4::GFP ESCs were transfected with the vectors indicated and selected for 7 days with puromycin (Fig. 5A). The empty vector was used as negative control. At the time points indicated, the morphology and GFP expression of individual colonies was measured (Fig. 5B). The empty vector had no effect on the cell morphology and GFP expression. The majority of colonies appeared densely packed with sharp edges and a strong expression of GFP which did not decrease during puromycin selection. The NF-jB inhibitor IjB-AA1 also had no effect on the Oct4::GFP expression and on the cell morphology (Fig. 5B, 5C). In contrast, most colonies overexpressing p65, c-Rel, and IKK2-ca start to lose their densely packed morphology along with a reduction of Oct4-driven GFP (Fig. 5C) after 3 days of puromycin selection. Seven days after overexpression of p65, c-Rel, and IKK2ca, most cells are GFP negative with a flatten morphology. Within these colonies, we still observed patches of cells that remained GFP positive with a densely packed appearance (Fig. 5C). The effect of NF-jB was not as strong as RA, which resulted in a faster drop of GFP expression (Fig. 5B) and a complete change of the cell morphology (Fig. 5C).
Figure 4. p65 promotes differentiation of ESCs. (A): Lentivirusbased, DOX-inducible vector system for overexpression of p65FPRed. (B, C): Embryonic stem cells, with GFP under control of the Oct4 promoter (Oct4::GFP), were either transduced with FPRed or p65-FPRed. After 5 days of addition of DOX, cells were examined. (D): Four days after addition of DOX, qPCR analysis of the markers indicated was performed. Mean 6 SD of two independent experiments. One-way ANOVA, Dunnett’s post hoc. *, p < 0.05; **, p < 0.01; ***, p < 0.001. Abbreviations: DOX, doxycycline; GFP, green fluorescent protein; hUbiQ, human Ubiquitin C promoter; LTR, long terminal repeat; M2R-tTA, reverse tetracycline-dependent transactivator; TRE, TET responsive element; WPRE, woodchuck post-transcriptional regulatory element.
Lu¨ningschrO¨r, StO¨cker, Kaltschmidt et al.
661
Figure 5. Canonical NF-jB signaling causes loss of pluripotency by inducing an epithelial to mesenchymal transition. (A): Scheme of the experimental procedure. (B): Quantification of Oct4::GFP fluorescence at the time points indicated. (C): Morphology and GFP expression of cells transfected with IjBa-AA1, p65, c-Rel, IKK2-ca, or the empty vector 7 days after puromycin selection. (D–H): Expression levels of Oct4, E-cadherin, Slug, Zeb1, and miR-290 cluster at the time points indicated after transfection with the empty vector (D), IjBa-AA1 (E), p65 (F), c-Rel (G), IKK2-ca (H), or the empty vector treated with RA (I). Mean 6 SEM of two independent experiments. One-way ANOVA, Dunnett’s post hoc. Abbreviations: ANOVA, analysis of variance; GFP, green fluorescent protein; IKK2, inhibitor of kappa B kinase; IKK2-ca, constitutive active form of IKK2; NF-jB, nuclear factor kappa B; RA, retinoic acid.
To assess the impact of canonical NF-jB signaling at the transcriptional level, we assayed the expression of Oct4 and the miR-290 cluster as markers for pluripotency as well as www.StemCells.com
Slug, Zeb1, and E-cadherin as markers for an EMT (Fig. 5D– 5I). Transfection with the empty vector resulted in a moderate but significant upregulation of Oct4, E-cadherin, and the
662
miR-290 Cluster Represses NF-jB
Figure 6. NF-jB overexpression promotes differentiation of embryonic stem cells (ESCs) toward a mesenchymal phenotype. (A–R): Immunohistochemical stainings of ESCs transfected with the vectors indicated 7 days after puromycin selection. Cells were stained for SSEA1 (A–F), SMA (G–L), and Vimentin (M–R) 7 days after puromycin selection. Arrowheads indicate remaining patches of cells still expressing Oct4 driven GFP after overexpression of p65 (C, I), c-Rel (D, J), and IKK2-ca (E, K). Abbreviations: GFP, green fluorescent protein; IKK2-ca, constitutive active form of IKK2; RA, retinoic acid; SMA, smooth muscle actin; SSEA1, stage specific embryonic antigen 1.
miR-290 cluster (Fig. 5D), indicating the maintenance of pluripotency during puromycin selection. The expression of Slug and Zeb1 was not altered by the empty vector (Fig. 5D). Overexpression of IjB-AA1 caused a moderate but not significant upregulation of E-cadherin, whereas Slug was significantly downregulated (Fig. 5E). In contrast, transfection with p65, c-Rel, IKK2-ca, and treatment with RA (Fig. 5F–5I) resulted in an increase of the expression of Slug and Zeb1, along with a decrease in the expression of the miR-290 cluster. With the exception of IKK2-ca, activation of canonical NF-jB signaling and treatment with RA also caused a moderate but significant downregulation of Oct4, whereas the expression of E-cadherin was not affected. The increase in the transcriptional level of Slug and Zeb1 indicates that NF-jB signaling positively regulates an EMT. This suggestion is supported by changes of the cell morphology, although we did not observe changes in the E-cadherin expression, even after treatment with RA. In contrast, ESCs, cultivated in suspension as EBs without LIF, showed a marked downregulation of E-cadherin after treatment with RA (Supporting Information Fig. 1). A recent study [26], demonstrating that a MET is required for reprogramming of mouse fibroblasts, showed that E-cadherin was upregulated by Klf4, which is directly downstream of the Jak/Stat3 axes of the LIF pathway [30]. This indicates that LIF is able to maintain E-cadherin and, in part Oct4 expression at the transcriptional level, although the epithelial phenotype and Oct4driven GFP expression was lost.
cells still expressing SSEA1 and Oct4::GFP (Fig. 6C–6E). These were not present after treatment with RA (Fig. 6F). Consistent with these observations no SMAþ cells were detectable after transfection of the empty vector (Fig. 6G) and IjBa-AA1 (Fig. 6H). In contrast, transfection with p65 (Fig. 6I), c-Rel (Fig. 6J), and IKK2-ca (Fig. 6K) resulted in a huge amount of flat and widespread SMAþ cells. Treatment with RA had the strongest effect and converted most cells toward a SMAþ phenotype. The appearance of SMAþ cells suggests that NF-jB signaling can induce a mesodermal cell fate under these conditions. To exclude that canonical NF-jB signaling is also able to drive pluripotent ESCs toward a neuronal direction, ESCs expressing Sox1-driven GFP were transfected with the indicated vectors (Fig. 6M–6R) [31]. During 7 days of puromycin selection, none of the vectors induced GFP expression. Transfection with the empty vector (Fig. 6M) and IjBa-AA1 (Fig. 6N) did not alter ESC morphology. As described above, transfection of p65 (Fig. 6O), c-Rel (Fig. 6P), and IKK2ca (Fig. 6Q) resulted in a predominant spread, Vimentinþ phenotype (Fig. 6O–6Q). Again, the strongest effect was observed after treatment with RA (Fig. 6R). Vimentin was also detectable within the densely packed colonies after transfection with the empty vector (Fig. 6M) and IjBa-AA1 (Fig. 6N), but these cells did not display an altered, spread morphology. Our data demonstrate that canonical NF-jB signaling promotes differentiation of ESCs toward a mesenchymal phenotype, which partially resembles the phenotype we observed after treatment with RA.
NF-jB Overexpression Promotes Differentiation of ESCs Toward a Mesenchymal Phenotype To further explore ESC differentiation after overexpression of NF-jB, Oct4::GFP- and Sox1::GFP-expressing ESC lines were stained for SSEA-1, SMA, and Vimentin (Fig. 6A–6R). After transfection with the empty vector (Fig. 6A) and IjBaAA1 (Fig. 6B), ESC colonies were still positive for SSEA1 and Oct4::GFP, whereas most cells are negative for SSEA1 after expression of p65, c-Rel, and IKK2-ca (Fig. 6C–6E). Within these colonies, we also found remaining cluster of
DISCUSSION In summary, we provide evidence that ESCs are characterized by a low protein expression of the NF-jB subunit p65. In the pluripotent state of ESCs, p65 translation is inhibited by members of the miR-290 cluster, namely miR-291b-5p and miR-293. Both miRNAs target the cds of p65, which might be accomplished by multiple binding to low-affinity MREs
Lu¨ningschrO¨r, StO¨cker, Kaltschmidt et al.
without a canonical seed match. This is in agreement with the notion that miRNA targeting goes beyond a canonical seed match within the 30 UTR [20]. Upon RA-induced differentiation of ESCs, expression of miR-291b and miR-293 is downregulated, causing an increase of p65 protein levels. In contrast to the upregulation of the canonical NF-jB signaling pathway, we observed a strong reduction of the expression of the noncanonical NF-jB subunits p52/p100 and RelB, which may suggest a role for noncanonical NF-jB signaling in the pluripotent state of ESCs. A recent report, investigating the role of NFjB signaling in human ESCs, also suggests putative opposing roles for the canonical and noncanonical NF-jB pathway [18] in human ESCs. Consistent with our data, this study proposed the involvement of the canonical pathway in regulating human ESC differentiation, whereas noncanonical signaling is involved in the maintenance of human ESC pluripotency. To what extent these two distinct branches of NF-jB signaling equal in the murine and human ESC system needs to be explored further. Notably, in the murine system it was also reported [32] that Nanog inhibits NF-jB signaling by directly interacting with the NF-jB subunits p65, RelB, and c-Rel. In the case of p65, this suggests two levels of inhibiting the activation of NF-jB signaling: first, the translational repression of p65 by miR-290 cluster to maintain low p65 protein amounts and second, the inhibition of translated p65 by physical interaction with Nanog. This inhibitory network might be reflected in the fact that we usually observed that some cells still remain pluripotent after overexpression of NF-jB. In a more general view, such a model might indicate how ESC-specific miRNAs may overcome the problem of a global transcriptional hyperactivity caused by an ‘‘open’’ chromatin structure of ESCs [33], which may lead to the expression of transcription factors (e.g., p65) that promote differentiation of ESCs and therefore have to be repressed. Another aspect of our study is the inductive effect of canonical NF-jB signaling on an EMT, resulting in the loss of the epithelial phenotype and pluripotency. Even after treatment with RA, whereupon no remaining patches of cells with an epithelial phenotype could be observed, we did not detect a decline in the expression of E-cadherin. The presence of LIF is known to maintain the epithelial character of ESCs by activation of Klf4 [30], which in turn drives expression of E-cadherin [26]. Nevertheless, the massive changes in the cell morphology toward a spread phenotype and the upregulation of Slug and Zeb1 incline us to believe that additional mechanisms may prevent E-cadherin translation or function and thereby accelerating an EMT.
REFERENCES 1 2 3 4 5 6 7
Gangaraju VK, Lin H. MicroRNAs: Key regulators of stem cells. Nat Rev Mol Cell Biol 2009;10:116–125. Anokye-Danso F, Trivedi CM, Juhr D et al. Highly efficient miRNAmediated reprogramming of mouse and human somatic cells to pluripotency. Cell Stem Cell 2011;8:376–388. Liao B, Bao X, Liu L et al. MicroRNA cluster 302–367 enhances somatic cell reprogramming by accelerating a mesenchymal-to-epithelial transition. J Biol Chem 2011;286:17359–17364. Melton C, Judson RL, Blelloch R. Opposing microRNA families regulate self-renewal in mouse embryonic stem cells. Nature 2010;463:621–626. Wang Y, Keys DN, Au-Young JK et al. MicroRNAs in embryonic stem cells. J Cell Physiol 2009;218:251–255. Judson RL, Babiarz JE, Venere M et al. Embryonic stem cell-specific microRNAs promote induced pluripotency. Nat Biotechnol 2009;27: 459–461. Marson A, Levine SS, Cole MF et al. Connecting microRNA genes to the core transcriptional regulatory circuitry of embryonic stem cells. Cell 2008;134:521–533.
www.StemCells.com
663
Our data strongly support the notion that the pluripotent state of ESCs is closely related to an epithelial character. Elsewhere it was already reported that ESC-specific miRNAs are involved in inhibiting an EMT and thereby participate in the maintenance and induction of pluripotency [3, 34, 35]. Further mechanistic insight into these interactions may provide a better understanding of the relationship between pluripotency and an epithelial cellular phenotype. These data add NF-jB to the map of EMT inducers in amniotes during development [27] as has been shown for the NF-jB ortholog dorsal during mesoderm invagination in Drosophila [36]. Taken together, we identified p65 as a new target gene of miR-290 cluster regulating pluripotency of ESCs by indirectly inhibiting an EMT transition.
CONCLUSSION Our study demonstrates that p65 is targeted by miR-290 cluster within its coding sequence. MiR-291b-5p and miR-293, two members of the miR-290 cluster without the canonical ‘‘AAGUGC’’ seed sequence, inhibit NF-jB signaling by a translational repression of p65. Activation of canonical NF-jB signaling promotes differentiation of ESCs by inducing an EMT resulting in a mesenchymal phenotype. MiR-290 cluster thereby indirectly participates in inhibition of an EMT and maintenance of pluripotency. Our data provide strong evidence for a close relationship between pluripotency and an epithelial phenotype.
ACKNOWLEDGMENTS We are grateful to A. Kralemann-K€ohler for excellent technical assistance; E.M. Fu¨chtbauer for providing the Oct4-GFP ESC line; A. Smith and R. Heinen for providing Sox1-GFP ESC line; D. Baltimore for providing the lentivirus system. We thank P. Heimann and H. Sch€oler for helpful discussion and critical comments on the manuscript. This study was supported by the University of Bielefeld and Deutsche Forschungsgemeinschaft (DFG).
DISCLOSURE
OF OF
POTENTIAL CONFLICTS INTEREST
The authors declare no potential conflicts of interest.. 8 9
10 11 12 13 14
Landgraf P, Rusu M, Sheridan R et al. A mammalian microRNA expression atlas based on small RNA library sequencing. Cell 2007; 129:1401–1414. Benetti R, Gonzalo S, Jaco I et al. A mammalian microRNA cluster controls DNA methylation and telomere recombination via Rbl2-dependent regulation of DNA methyltransferases. Nat Struct Mol Biol 2008;15:998. Sinkkonen L, Hugenschmidt T, Berninger P et al. MicroRNAs control de novo DNA methylation through regulation of transcriptional repressors in mouse embryonic stem cells. Nat Struct Mol Biol 2008;15:259–267. Wang Y, Baskerville S, Shenoy A et al. Embryonic stem cell-specific microRNAs regulate the G1-S transition and promote rapid proliferation. Nat Genet 2008;40:1478–1483. Zheng GX, Ravi A, Calabrese JM et al. A latent pro-survival function for the mir-290–295 cluster in mouse embryonic stem cells. PLoS Genet 2011;7:e1002054. Medeiros LA, Dennis LM, Gill ME et al. Mir-290–295 deficiency in mice results in partially penetrant embryonic lethality and germ cell defects. Proc Natl Acad Sci USA 2011;108:14163–14168. Bibel M, Richter J, Schrenk K et al. Differentiation of mouse embryonic stem cells into a defined neuronal lineage. Nat Neurosci 2004;7: 1003–1009.
miR-290 Cluster Represses NF-jB
664
15 Tay Y, Zhang J, Thomson AM et al. MicroRNAs to Nanog, Oct4 and Sox2 coding regions modulate embryonic stem cell differentiation. Nature 2008;455:1124–1128. 16 Hockemeyer D, Soldner F, Cook EG et al. A drug-inducible system for direct reprogramming of human somatic cells to pluripotency. Cell Stem Cell 2008;3:346–353. 17 Lois C, Hong EJ, Pease S et al. Germline transmission and tissue-specific expression of transgenes delivered by lentiviral vectors. Science 2002;295:868–872. 18 Yang C, Atkinson SP, Vilella F et al. Opposing putative roles for canonical and noncanonical NFkappaB signaling on the survival, proliferation, and differentiation potential of human embryonic stem cells. Stem Cells 2010;28:1970–1980. 19 Houbaviy HB, Murray MF, Sharp PA. Embryonic stem cell-specific microRNAs. Dev Cell 2003;5:351–358. 20 Thomas M, Lieberman J, Lal A. Desperately seeking microRNA targets. Nat Struct Mol Biol 2010;17:1169–1174. 21 Rehmsmeier M, Steffen P, Hochsmann M et al. Fast and effective prediction of microRNA/target duplexes. RNA 2004;10: 1507–1517. 22 Miranda KC, Huynh T, Tay Y et al. A pattern-based method for the identification of microRNA binding sites and their corresponding heteroduplexes. Cell 2006;126:1203–1217. 23 Kertesz M, Iovino N, Unnerstall U et al. The role of site accessibility in microRNA target recognition. Nat Genet 2007;39: 1278–1284. 24 Kim JB, Zaehres H, Wu G et al. Pluripotent stem cells induced from adult neural stem cells by reprogramming with two factors. Nature 2008;454:646–650. 25 Redmer T, Diecke S, Grigoryan T et al. E-cadherin is crucial for embryonic stem cell pluripotency and can replace OCT4 during somatic cell reprogramming. EMBO Rep 2011;12: 720–726.
26 Li R, Liang J, Ni S et al. A mesenchymal-to-epithelial transition initiates and is required for the nuclear reprogramming of mouse fibroblasts. Cell Stem Cell 2010;7:51–63. 27 Thiery JP, Acloque H, Huang RY et al. Epithelial–mesenchymal transitions in development and disease. Cell 2009;139:871–890. 28 Pantuck AJ, An J, Liu H et al. NF-kappaB-dependent plasticity of the epithelial to mesenchymal transition induced by Von Hippel-Lindau inactivation in renal cell carcinomas. Cancer Res 2010;70:752–761. 29 Huber MA, Azoitei N, Baumann B et al. NF-kappaB is essential for epithelial-mesenchymal transition and metastasis in a model of breast cancer progression. J Clin Invest 2004;114:569–581. 30 Niwa H, Ogawa K, Shimosato D et al. A parallel circuit of LIF signalling pathways maintains pluripotency of mouse ES cells. Nature 2009;460:118–122. 31 Aubert J, Stavridis MP, Tweedie S et al. Screening for mammalian neural genes via fluorescence-activated cell sorter purification of neural precursors from Sox1-gfp knock-in mice. Proc Natl Acad Sci U S A 2003;100(Suppl 1):11836–11841. 32 Torres J, Watt FM. Nanog maintains pluripotency of mouse embryonic stem cells by inhibiting NFkappaB and cooperating with Stat3. Nat Cell Biol 2008;10:194–201. 33 Orkin SH, Hochedlinger K. Chromatin connections to pluripotency and cellular reprogramming. Cell 2011;145:835–850. 34 Subramanyam D, Lamouille S, Judson RL et al. Multiple targets of miR-302 and miR-372 promote reprogramming of human fibroblasts to induced pluripotent stem cells. Nat Biotechnol 2011;29: 443–448. 35 Miyoshi N, Ishii H, Nagano H et al. Reprogramming of mouse and human cells to pluripotency using mature microRNAs. Cell Stem Cell 2011;8:633–638. 36 Roth S, Stein D, Nusslein-Volhard C. A gradient of nuclear localization of the dorsal protein determines dorsoventral pattern in the Drosophila embryo. Cell 1989;59:1189–1202.
See www.StemCells.com for supporting information available online.
