Author contributions: S.S.: conception and design, collection and assembly of data, data analysis and interpretation, and manuscript writing; M.S.: conception and design, collection and assembly of data, and data analysis and interpretation; E.C.: collection and assembly of data and data analysis and interpretation; C.S.: conception and design and manuscript writing; R.B.: conception and design, data analysis and interpretation, manuscript writing, and financial support. S.S. and M.S. contributed equally to this article.
Disclosure of potential conflicts of interest is found at the end of this article.
First published online in STEM CELLSEXPRESS January 10, 2013.
The pluripotency of mouse embryonic stem cells (mESCs) is controlled by a network of transcription factors, mi-RNAs, and signaling pathways. Here, we present a new regulatory circuit that connects miR-335, Oct4, and the Retinoblastoma pathway to control mESC self-renewal and differentiation. Oct4 drives the expression of Nipp1 and Ccnf that inhibit the activity of the protein phosphatase 1 (PP1) complex to establish hyperphosphorylation of the retinoblastoma protein 1 (pRb) as a hallmark feature of self-renewing mESCs. The Oct4-Nipp1/Ccnf-PP1-pRb axis promoting mESC self-renewal is under control of miR-335 that regulates Oct4 and Rb expression. During mESC differentiation, miR-335 upregulation co-operates with the transcriptional repression of Oct4 to facilitate the collapse of the Oct4-Nipp1/Ccnf-PP1-pRb axis, pRb dephosphorylation, the exit from self-renewal, and the establishment of a pRb-regulated cell cycle program. Our results introduce Oct4-dependent control of the Rb pathway as novel regulatory circuit controlling mESC self-renewal and differentiation. STEM CELLS2013;31:717–728
The unlimited proliferative potential and pluripotent character of mouse embryonic stem cells (mESCs) is characterized not only by specific core transcriptional regulatory circuits comprising the pluripotency transcription factors Oct4, Sox2, and Nanog but also by an atypical cell cycle structure [1–3]. In addition to transcription factors, several microRNAs (miRNAs) were identified to modulate mESC self-renewal by directly targeting pluripotency transcription factors or cell cycle regulators [4–9]. Rapid cell division of self-renewing mESCs is mediated by a muted oscillation of the Cdk2-Cyclin E complex resulting in an increase in overall Cdk2 activity that holds Retinoblastoma family proteins in a hyperphosphorylated and biochemically inactive state, thereby ensuring rapid cell cycle progression with remarkably short G1/G2 phases [2, 10–14]. Induction of cell differentiation results in the repression of core self-renewal transcription factors and is paralleled by the establishment of a classic Retinoblastoma pathway-dependent cell cycle program [2, 10, 12, 13]. Recent studies indicate a role for the Rb pathway in the control of mESC self-renewal: constitutive pRB hyperphosphorylation due to the targeted deletion of the Cdk2 inhibitor Cdk2ap1 results in an impaired cell differentiation potential that can be rescued by expressing nonphosphorylateable pRb . In line with this, mESC lacking a functional Rb pathway due to the combined targeted deletion of all Retinoblastoma protein family members display differentiation defects in teratoma formation experiments . Together, this suggests that the inactivation of the pRb pathway antagonizes the initiation of mESC differentiation and consequently enhances the self-renewal potential of mESC.
The concurrent repression of self-renewal transcription factors and switch in cell cycle structure at the transition from mESC self-renewal to differentiation suggest the existence of specific pathways that connect developmental potential with cell cycle regulation. However, functional evidence for such a connection in mESCs is still missing.
Here, we demonstrate that Oct4 controls the cell cycle program of self-renewing mESCs by protecting pRb from dephosphorylation by the protein phosphatase 1 (PP1) complex. The self-renewal promoting Oct4-pRb axis is under tight control of miR-335 that targets conserved sequence motifs in the Untranslated Regions (3′UTRs) of Oct4 and pRb. At the onset of mESC differentiation, transcriptional repression of Oct4 in conjunction with post-transcriptional silencing of Rb/Oct4 by miR-335 causes the collapse of the Oct4–PP1 self-renewal axis, leading to a rapid dephosphorylation of p-pRb, the exit from self-renewal, and the establishment of a cell cycle program of differentiated cells.
MATERIALS AND METHODS
Feeder-independent mouse ES cells were cultured on 0.2% gelatin-coated plates in mESC self-renewal medium (Dulbecco's modified Eagle's medium supplemented with 15% knockout serum replacement) (Gibco, Grand Island, NY, http://www.invitrogen.com), 1% nonessential amino acids (Gibco), 1 mM sodium pyruvate (Invitrogen, Carlsbad, CA, http://www.invitrogen.com), 1% L-glutamine (Invitrogen), 0.1 mM β-mercaptoethanol, 1% penicillin/streptomycin (Invitrogen), and 1,000 U/ml mouse leukemia inhibitory factor. ES cell differentiation schemes are indicated in supporting information. Self-renewing mESCs were transfected with pMSCV-miR-335 (supporting information section) and subjected to selection using Blasticidine (Invitrogen, 2 μg/ml) to generate mESCs stably overexpressing miR-335. To generate mESCs with reduced miR-335 levels, cells were transfected with TW3-decoy-miR-335 (supporting information) and subjected to selection using puromycin (Invitrogen, 3 μg/ml).
Transient Transfections of miRNAs or siRNA Oligos
Mouse ES cells were transiently transfected with hsa_miR-335 mimic (C-300708-05, Thermo Scientific Dharmacon), antagomiR-335 hairpin miR-IDIAN inhibitor (IH-300708-07, Thermo Scientific Dharmacon) according to the manufacturer's suggestions. MiR-IDIAN miRNA mimic negative control small interfering RNA (siRNA) (CN-00100-01-05, Thermo Scientific Dharmacon) was used as a negative control. Mimic miRNAs resemble siRNA molecules that recapitulate miRNA function. Total RNA and protein were prepared 72 hours after transfection. ES cells were transiently transfected with siRNAs indicated in supporting information according to the manufacturer's suggestions.
For fluorescence-activated cell sorting (FACS) analysis, self-renewing and differentiating mESCs were collected and fixed in ethanol. After rehydration, cells were resuspended in 1× phosphate buffered saline (PBS), 0.1% Nonidet P40 and treated with 200 μg/ml RNase A for 10 minutes; propidium iodide was added to a final concentration of 50 μg/ml. Cells were analyzed on a flow cytometer (FACSCalibur, Becton Dickinson, Franklin Lakes, NJ, http://www.bd.com). FACS data were analyzed using ModFit LT 3.1.
Analysis of miRNA Expression
Total RNA was extracted and purified using TRIzol reagent (Gibco/BRL) and reverse transcribed using the miR-335 and sno-135 TaqMan MicroRNA Assay system (Applied Biosystems, Foster City, CA, http://www.appliedbiosystems.com). The mouse snoRNA sno-135 was used as a reference. The quantitative stem-loop real-time polymerase chain reaction (RT-PCR) was performed according to conditions suggested by Applied Biosystems. Quantitative miRNA expression data were analyzed using a StepOnePlus Sequence Detection System (Applied Biosystems).
Analysis of mRNA Expression
Total RNA from indicated cell lines was purified using TRIzol reagent (Gibco/BRL). 1.5 μg of total RNA was treated with DNAse (RQ1, Promega, Madison, WI, http://www.promega.com) and subjected to reverse transcription using random primers (Promega) and SuperScriptII Reverse Transcriptase (Invitrogen) according to the manufacturer's suggestions. Quantitative PCR was performed using the SYBR Green Master Mix (Applied Biosystem) and analyzed with a StepOnePlus real-time PCR machine (Applied Biosystem). mRNA levels were normalized with gapdh. qRT-PCR primer sequences are indicated in supporting information.
Immunoprecipitation, Western Blot Analysis, and Antibodies
mESC lysates were prepared with ice-cold lysis buffer containing 10 mM Tris-HCl, pH 7.4, 100 mM NaCl, 1 mM phenylmethanesulfonylfluoride (PMSF), 20 mM β-glycerophosphate, 1 mM Na3VO4, 1 mM Na2F, 5 mM EDTA, and protease inhibitor mixture (Sigma) plus 1% Triton X-100. After centrifugation and preclearing, lysates were incubated at 4°C for 4 hours with 25 μl of protein A-Sepharose CL-4B (Amersham Biosciences, Piscataway, NJ, http://www.amersham.com) plus specific antibodies. The resin was washed, and bound proteins were eluted in SDS-PAGE sample buffer. Western blot analysis was performed according to standard procedures using the primary antibodies indicated in supporting information.
ES cells were grown to a final count of 1 × 108 cells for each chromatin immunoprecipitation (ChIP) analysis. ChIP was performed as previously described . A rabbit polyclonal anti-Oct4 (ab19857; Abcam, Cambridge, U.K., http://www.abcam.com) and a rabbit polyclonal Human influenza hemagglutinin (HA)-probe (Y-11; sc-805; Santa Cruz Biotechnology, Santa Cruz, CA, http://www.scbt.com) were used in ChIP experiments. The oligos used for quantitative RT-PCR on chromatin immunoprecipitates are specified in supporting information. Quantitative PCR was performed using the SYBR Green Master Mix (Applied Biosystem) and analyzed with a StepOnePlus real-time PCR machine (Applied Biosystem).
Luciferase Reporter Constructs and Dual Luciferase Assay
The 3′-UTR constructs are described in supporting information. In luciferase reporter assay, 2 × 104 U2OS cells were transiently transfected with 200 ng of UTR pGL3Control (Promega) and 20 ng of reference Renilla luciferase expression vector using FUGENE 6 (Roche, Basel, Switzerland, http://www.roche-applied-science.com) transfection reagent following the manufacturer's protocol. The next day miRIDIAN miRNA mimics (Thermo Scientific Dharmacon) were transfected with Lipofectamine RNAiMAX Transfection Reagent (Invitrogen) following the manufacturer's protocol at a concentration of 30 nM. Firefly/renilla luciferase activity was assayed 72 hours after transfection using the Dual Luciferase Assay kit (Promega).
Alkaline Phosphatase Assays and Colony Reformation Assay
mESC colonies were stained with StemTAG Alkaline Phosphatase Staining Kit (CBA-302; Cell Biolabs, San Diego, CA, http://www.cellbiolabs.com) following manufacturer's instruction. The self-renewal colonies stain positive for alkaline phosphatase (AP), but differentiated colonies stain less or negative for AP. For the photospectrometric measurement of AP activity in experimental cells, p-nitrophenol levels were determined using StemTAG Alkaline Phosphatase Activity Assay Kit (CBA-302; Cell Biolabs) following the manufacturer's protocol.
For colony replating efficiency assays, ES cells were transiently transfected with miRNA mimics and trypsinized and counted 1 day after transfection. 5 × 103 cells were replated in triplicate on gelatinized cell culture 30 mm petri dishes and allowed to grow for 5 days. For staining, colonies were washed in PBS and fixed in 4% PFA for 20 minutes. After washing, colonies were incubated with 0.05% crystal violet in H2O for 30 minutes in the dark. Subsequently, plates were washed in H2O, and the replating efficiency of colony reformation was determined by counting colony number.
Flow Cytometry Analysis of Self-Renewal and Immunofluorescence Analysis
For self-renewal analysis, mES cells were trypsinized into single cell suspensions (5 × 105), washed with PBS, and fixed 1% paraformaldehyde in PBS for 10 minutes at room temperature. Cells were then washed in PBS, resuspended in PBS with 5% bovine serum albumin (BSA), and incubated with primary antibody mouse anti-SSEA1 (550079; BD Pharmingen, San Diego, CA, http://www.bdbiosciences.com/index_us.shtml) at 10 μg/ml final concentration for 30 minutes at room temperature. After washing, cells were resuspended in anti-mouse Fluorescein isothiocyanate (FITC)-conjugated secondary antibody (Sigma) in PBS with 0.5% BSA for 30 minutes, washed, and analyzed using a FACS flow cytometer (FACSCalibur, Becton Dickinson).
For immunofluorescence stainings, cells grown on coverslips were pretreated with 20 μg/ml polylysine (Sigma). After 12 hours, cells were fixed with 4% PFA for 10 minutes, washed with PBS, and incubated in PBS with 0.1% Triton X-100 for 5 minutes. After washing, PBS with 5% BSA was used to block for 30 minutes before the addition of primary antibody against Oct4 (ab19857; Abcam), Rb1 (554136; BD Pharmingen), or SSEA1 (550079; BD Pharmingen). Cells were washed twice in PBS and incubated with FITC- or Tetramethylrhodamine-5-(and-6)-isothiocyanate (TRITC)-conjugated secondary antibodies (Sigma) for 1 hour. For SSEA1 staining, cells were not treated with Triton X-100. Slides were analyzed on a Leica DM 4000B fluorescence microscope.
200,000 ES cells resuspended in Matrigel Basement Membrane Matrix (BD Bioscience, San Diego, http://www.bdbiosciences.com) were injected subcutaneously into the flanks of nude mice (Charles River Laboratories). Six injections per cell line were performed. Teratoma formation was followed during 3 weeks before xenograft excision from the animal.
A t test was performed so that p values for the difference between the means of the experimental conditions and control could be calculated. Error bars represent SD. Each finding was confirmed by three independent biological replicates, unless specified.
Oct4 and pRb Co-Operate to Ensure mESC Self-Renewal
The loss of self-renewal potential of mESCs is paralleled by an acquisition of a Retinoblastoma family protein-regulated cell cycle program [12–14]. To identify a possible link between Oct4-dependent transcriptional self-renewal circuits and pRb-dependent cell cycle regulation, we used RNAinterference (RNAi) to deplete Oct4 or Rb (encodes pRb protein) from self-renewing mESCs. Interestingly, Oct4 (siRNA) provoked a rapid reduction of total and hyperphosphorylated pRb (pRb and p-pRb-T356, respectively) without altering total Rb mRNA levels (Fig. 1A). In line with this result, ectopic overexpression of HA-tagged Oct4 in self-renewing mESCs increased pRb hyperphosphorylation, resulting in pRb stabilization and augmented total pRb levels (Fig. 1B). Of notice, knockdown of pRb does not affect Oct4 mRNA or protein expression (Fig. 1A). Together, these results suggest that Oct4 has a role in controlling pRb levels in self-renewing mESCs (Fig. 1A). To test whether this pathway has a relevance for mESC self-renewal, we studied self-renewal features in the context of Rb siRNA. RNAi-mediated depletion of Rb from mESCs growing under self-renewal conditions reduced AP activity to a similar extent like the depletion of Oct4 (Fig. 1c). Of notice, the observed reduction of Alkaline phosphatase (AP) activity in Rb siRNA-treated cells is similar to that observed after 3 days of ES cell differentiation, confirming that pRb has a role in promoting mESC self-renewal (supporting information Fig. 1). Impaired self-renewal was also observed on the level of gene expression. In particular, knockdown of Rb in mESCs provoked a downregulation of self-renewal marker genes Nanog, Sox2, and Gdf3 but did not affect Oct4 or Klf4 expression (Fig. 1D). Consistent with reduced self-renewal potential, pRb-depleted cells display reduced mESC colony size, altered cell morphology, and reduced cell proliferation rate (Fig. 1E, 1F). The induction of mESC differentiation is characterized by an extension of G1 and G2 phases of the cell cycle (Fig. 1G, [5, 10]). Consistent with reduced cell proliferation, depletion of Rb from self-renewing mESCs resulted in a reduction of cell numbers in S phase and increase of cell numbers in G1 and G2/M phase of the cell cycle, reflecting a shift toward the cell cycle status of differentiated cells (Fig. 1G; supporting information Table 1). We conclude that high pRb levels promote mESC self-renewal potential.
Oct4 Drives the Expression of Nipp1/Ccnf to Maintain pRb Hyperphosphorylation
Our data anticipate a functional link between Oct4 and the phosphorylation status of pRb in self-renewing mESCs. We found that Oct4 does not act as a transcription factor for Rb and does not physically interact with pRb (Fig. 1A; supporting information Fig. 2A, 2B). We consequently validated whether Oct4 impacts on pRb function by modulating its phosphorylation status. pRb is a substrate for PP1 complex in proliferating cells . At mitotic exit, all three PP1 isoforms, α, γ1, and β bind to pRb and dephosphorylate its Ser/Thr sites in a sequential and site-specific way [19–22]. Importantly, PP1 activity is regulated by the PP1 subunit Nipp1 (nuclear inhibitor of PP1); in addition Ccnf (cyclin F), a member of the Skp, Cullin, F-box containing complex (SCF) ubiquitin ligase complex was associated with the maintenance of pRb in an inactive status [23, 24]. In line with a potential role in differentiation, Nipp1 and Ccnf were found to be efficiently repressed during mESC differentiation (supporting information Fig. 2C). To functionally validate a role for Oct4 in controlling Nipp1/Ccnf expression, we depleted Oct4 from self-renewing mESCs and measured Nipp1 and Ccnf mRNA expression. As expected, Oct4 knockdown caused a significantly reduced expression of Ccnf and Nipp1 mRNA (Fig. 2A). Importantly, ectopic expression of HA-tagged Oct4 in differentiating mESCs that do not express endogenous Oct4 was able to not only rescue Ccnf and Nipp1 expression (Fig. 2B) but also restored hyperphosphorylation of pRb and increased total pRb levels (Fig. 2B, right panel). Finally, ChIP verified the presence of Oct4 at Ccnf and Nipp1 promoter regions (Fig. 2C). Altogether, our results demonstrate that Oct4 promotes the expression of the PP1 inhibitors Nipp1 and Ccnf in self-renewing mESCs.
To validate the relevance of Nipp1/Ccnf for controlling the post-translational modification of pRb, we tested whether alterations of Nipp1 or Ccnf expression impact on the phosphorylation status of pRb and mESC self-renewal potential. RNAi-mediated reduction of Ccnf/Nipp1 resulted in an efficient reduction of total pRb and phospho-pRb-T356 levels, the accumulation of cells with differentiated cell morphology and reduced cell proliferation rates, altogether recapitulating the effect of combined Oct4 and Rb depletion (Fig. 2D–2F). These effects were paralleled by a significantly reduced AP activity and expression of self-renewal marker genes Nanog, Sox2, Gdf3, and Klf4 (Fig. 2G, 2H). Consistent with the role of PP1 in mediating pRb hypophosphorylation, RNAi-mediated depletion of PP1β significantly increased the abundance of hyperphosphorylated pRb and total pRb (Fig. 2D). In order to better address the relation between Oct4-dependent pRb-phosphorylation and total pRb levels, we depleted Oct4 and PP1β from self-renewing mESCs. We found that siRNA-mediated knockdown of PP1β efficiently rescues pRb-phosphorylation levels in Oct4-depleted mESCs (supporting information Fig. 2D). The rescue of pRb-phosphorylation in siOct4/siPP1 knockdown cells is paralleled by a restoration of total pRb protein to expression levels observed in control mESCs (supporting information Fig. 2D). These results underline that PP1 not only controls pRb phosphorylation levels in mESCs but also indicate that phosphorylation has a central role in controlling total pRb levels in self-renewing mESCs.
Altogether, this set of experiments demonstrates that Oct4 drives the expression of Nipp1/Ccnf to protect hyperphosphorylated pRb from dephosphorylation by PP1 in self-renewing mESCs. The novel Oct4-Nipp1/Ccnf-PP1-pRb axis promotes mESC self-renewal potential and establishes a direct link between self-renewal transcription circuits and the Rb pathway.
MiR-335 Controls Oct4 and pRb Expression on the Post-Transcriptional Level
Repression of Oct4, as occurring during mESC differentiation, is expected to result in the collapse of the Oct4-Nipp1/Ccnf-PP1 axis, leading to pRb hypophosphorylation, loss of self-renewal gene expression patterns, and the acquisition of a cell cycle profile of differentiated cells. In line with this Oct4, pRb, p-pRb-T356, and Nipp1 protein levels were rapidly repressed during mESC embryoid body (EB) differentiation (Fig. 3A; supporting information Fig. 3C). However, the persistence of considerable levels of Oct4 and Rb mRNA levels at day 3 and 6 of EB differentiation suggest that post-transcriptional regulatory mechanisms executed by miRNAs could accelerate the collapse of the Oct4-pRb axis at the exit from mESC self-renewal (Fig. 3A, right panel). Based on the presence of conserved target sites for miR-335 in the 3′UTRs of Rb and Oct4, we speculated that miR-335 could regulate the Oct4-Nipp1/Ccnf-PP1-pRb axis by targeting Oct4 and Rb in mESCs (supporting information Fig. 3A) . Consistent with the tight regulation of miRNAs controlling mESC function [26, 27], we found that miR-335 and its hosting gene Mest were efficiently upregulated in diverse mESC differentiation schemes (Fig. 3B; supporting information Fig. 3B) . As expected, the pluripotency transcription factors Sox2, Nanog, and Klf4 together with Oct4 were downregulated during mESC differentiation (supporting information Fig. 3C). To validate a possible interaction between miR-335 and its predicted target sites in the 3′UTR of Oct4 and pRb, we performed classic luciferase reporter assays. Ectopic miR-335 efficiently reduced the expression of a luciferase reporter fused to the 3′UTR of Oct4 or Rb in transient cotransfection experiments using U2OS cells (Fig. 3C). Deletion of the miR-335 target sites (Rb-3′UTRΔ; Oct4-3′UTRΔ) renders the Oct4 and Rb control reporter constructs resistant to ectopic miR-335 (Fig. 3C). We conclude, that miR-335 targets conserved sequence motifs in the 3′UTR of Oct4 and Rb to control Oct4 and pRb protein levels in mESCs.
Transient transfection of mESCs with miR-335 mimic siRNAs efficiently reduced pRb and Oct4 protein levels under self-renewal conditions whereas antago-miR-335 siRNAs had an opposite effect (Fig. 3D). Importantly, ectopically elevated miR-335 levels antagonized pRb hyper-phosphorylation, recapitulating the effect of Oct4 siRNA-mediated pRb dephosphorylation (Figs. 1A and 3D). We conclude that miR-335 can control the expression of Oct4 and pRb on the post-transcriptional level and impacts on the phosphorylation status of pRb.
MiR-335 Antagonizes mESC Self-Renewal
The sharp upregulation of miR-335 during mESC differentiation anticipates an important role for miR-335 in antagonizing the activity of the self-renewal promoting Oct4-Nipp1/Ccnf-PP1-pRb axis. To test this hypothesis, we performed miR-335 gain and loss of function experiments in self-renewing mESCs (supporting information Fig. 3D). mESCs transfected with mimic miR-335 siRNAs showed reduced cell proliferation rates and formed smaller colonies that contained a significant fraction of cells with a morphology of differentiated cells (Fig. 4A). Importantly, when experimental cells were replated in self-renewing promoting mESC medium 3 days post-transfection, miR-335 transfected cells displayed a severe impairment in the formation of colonies containing self-renewing mESCs (Fig. 4B). In line with increased cell differentiation, we found that ectopically increased miR-335 levels reduced the number of cells in S phase, indicating a shift toward the cycle profile of differentiated cells (Fig. 4C; supporting information Table 2). Reducing endogenous miR-335 levels by transfecting antago-miR-335 did not have an impact on cell morphology or cell cycle regulation (Fig. 4A–4C; supporting information Table 2).
To validate whether miR-335 antagonizes the expression of molecular markers of mESC pluripotency, we transfected self-renewing mESCs with mimic-miR-335 siRNAs and antago-miR-335 siRNAs. Introducing ectopic miR-335 into self-renewing mESCs reduced AP activity, SSEA-1 expression, colony re-plating efficiency, and significantly reduced the expression of the self-renewal marker genes Sox2, Oct4, Gdf3, and Klf4. (Fig. 4D, 4E; supporting information Fig. 4A, 4B). Of interest, ectopic introduction of miR-335 also induced the expression of primitive ectoderm marker genes Nestin and Otx2 (supporting information Fig. 4C). This suggests that reduced self-renewal potential in the context of elevated miR-335 levels lowers the barriers to the induction of cell differentiation programs. As expected, reducing cellular miR-335 levels by introducing antago-miR-335 did not impact on any of the investigated markers for mESC self-renewal potential (Fig. 4D, 4E; supporting information Fig. 4A, 4B).
After discovering a role of miR-335 in opposing mESC self-renewal, we evaluated the separate contribution of the pRb and Oct4, both targets of miR-335, to the observed alterations in mESC self-renewal potential. Individual siRNA-mediated knockdown of Oct4 or Rb significantly reduced AP activity (Fig. 4F). This effect was enhanced by the cotransfection of si-Oct4 and si-Rb oligonucleotides or mimic miR-335 siRNAs (Fig. 4F). This demonstrates that concurrent targeting of Oct4 and Rb by miR-335 is an efficient mechanism to reduce self-renewal potential. Importantly, ectopic expression of pRb or HA-tagged Oct4 rescued the impaired self-renewal potential of miR-335 overexpressing mESCs as shown in AP assays and pluripotency marker gene expression analysis (Fig. 4G; supporting information Fig. 4D–4F). Consequently, we exclude that alternative miR-335 targets drive the observed effects in self-renewal assays.
Together, this set of experiments demonstrates that miR-335 is a novel antagonizer of mESC self-renewal that limits pluripotency potential by concomitant targeting of Oct4 and pRb at the post-transcriptional level.
Targeting of Oct4 and Rb by MiR-335 Promotes the Exit from mESC Self-Renewal
The dramatic increase of miR-335 expression at the onset of mESC differentiation indicates that miR-335 dependent repression of Oct4 and pRb is an important step to facilitate the exit from self-renewal (Fig. 3B). To further validate this model, we ectopically altered miR-335 expression during mESC differentiation and studied the kinetics of mESC self-renewal potential, cell cycle structure, and the upregulation of differentiation markers.
Ectopic introduction of miR-335 accelerated the appearance of cells with differentiated morphology, further underlining that miR-335 promotes mESC differentiation (Fig. 5A). In contrast, antagonizing the upregulation of miR-335 during mESC differentiation by introducing antago-miR-335 oligonucleotides prior to the induction of mESC differentiation enabled mESCs to retain an undifferentiated cell morphology (Fig. 5A). In line with this, mESCs treated with antago-miR-335 maintained a self-renewal cell cycle program even under differentiation conditions; in contrast, ectopic introduction of miR-335 accelerated the reduction of cell numbers in S phase during mESC differentiation (Fig. 5B; supporting information Table 3). The differentiation promoting effect of miR-335 was also confirmed on the molecular level. The induction of differentiation in control miRNA transfected cells induced a 30% reduction of AP activity. Whereas, the introduction of antago-miR-335 under the same conditions significantly delayed the reduction of AP activity during mESC differentiation, ectopic introduction of miR-335 accelerated the differentiation associated loss of AP activity (Fig. 5C). Similar results were obtained for the expression of SSEA-1, a classic marker for self-renewing mESCs (supporting information Fig. 5A). Expression analysis of marker genes for mESC pluripotency and early differentiation events confirmed the relevance of miR-335 at the exit from self-renewal. The introduction of miR-335 prior to the induction of mESC differentiation efficiently reduced the expression of self-renewal markers such as Oct4, Sox2, Klf4, and Gdf3 and increased the expression of marker genes for primitive ectoderm (Fgf5, Otx2, Nestin) and mesoderm (Bmp4) when compared to control cells (Fig. 5D, 4E). In contrast, antagonizing the upregulation of miR-335 during mESC differentiation by transfecting antago-miR-335 significantly delayed the downregulation of self-renewal marker genes and resulted in an impaired induction of early differentiation marker genes at day 3 of mESC differentiation (Fig. 5D, 5E). In line with a role of miR-335 in promoting cell differentiation, miR-335 overexpression promoted the formation of contractile cardiomyocyte structures during EB differentiation; in contrast expressing a decoy-miR-335 construct significantly impaired the formation of pulsating cardiomyocytes (Fig. 5F). In addition to altered in vitro differentiation, we observed that mESC overexpressing miR-335 form EB aggregates with reduced growth potential in long-term suspension cultures and teratomas with reduced size when subcutaneously injected into immune-compromised mice (supporting information Fig. 5B, 5C). Together, this suggests that a tight control of miR-335 expression is essential to execute long-term cell differentiation events.
Together, our data indicate that upregulation of miR-335 promotes the exit from self-renewal toward cell differentiation by targeting the expression of Oct4 and pRb. To investigate whether miR-335 plays a role in controlling the Oct4-Nipp/Ccnf-PP1-pRb axis at the exit from self-renewal, we used mESC stably overexpressing miR-335 or a decoy-miR-335 and followed the expression of Oct4, pRb, pRb-T356, and Nipp1 during EB differentiation.
Under self-renewal conditions, control mESCs (day 0) express high levels of total pRb (pRb) and hyperphosphorylated pRb (pRb-T356) (Fig. 5G). Consistent with reduced Oct4 protein levels, we found decreased Nipp1 expression and a reduced abundance of total and hyperphosphorylated pRb in miR-335 overexpressing cells. Reducing endogenous miR-335 levels by overexpressing a decoy-miR-335 construct increases Oct4, Nipp1, and pRb expression and leads to augmented pRb hyperphosphorylation in self-renewing mESCs (Fig. 5G, lanes 1–3). Upon induction of differentiation, transcriptional repression combined with targeting by miR-335 efficiently reduces Oct4 expression, leading to low Nipp1 protein permitting the dephosphorylation of pRb by PP1. Destabilization of pRb by dephosphorylation and targeting of Rb by endogenous miR-335 finally reduces total pRb expression levels (Fig. 5G, lane 4). Of notice, this effect is aggravated upon miR-335 overexpression (Fig. 5G, lane 5). In contrast, antagonizing miR-335 upregulation by expressing a decoy-miR-335 construct partially rescues Oct4 and Nipp1 expression leading to increased phosphorylation of pRb at day 3 and 6 of mESC differentiation, indicative for an inefficient exit from mESC self-renewal programs (Fig. 5G, lanes 6, 9).
Altogether, our data demonstrates that Oct4 in conjunction with miR-335 controls the activity of the pRb pathway in mESCs. In self-renewing mESCs, high Oct4 expression drives the expression of Nipp1 and Ccnf that antagonize the enzymatic activity of the PP1 complex, leading to pRb hyperphosphorylation as hallmark feature of self-renewing mESCs (Fig. 6A). Upon induction of cell differentiation, transcriptional repression of Oct4 in conjunction with targeting of pRb and Oct4 by upregulated miR-335 causes the collapse of the Oct4-Nipp1/Ccnf-PP1-pRb axis promoting the exit from self-renewal and establishment of a Retinoblastoma regulated cell cycle program (Fig. 6B).
Self renewing mESCs are characterized by the coordinated expression of core pluripotency transcription factors and the establishment of a specific cell cycle program that allows rapid cell proliferation [1–3, 13, 14]. Here we present a novel pathway that links Oct4 and miR-335 with the Retinoblastoma pathway in self-renewing and differentiating mESCs.
In self-renewing mESC, cell cycle regulation is mainly executed by Cdk2 that mediates the hyperphosphorylation and biochemical inactivation of Retinoblastoma family protein [12–14]. The induction of cell differentiation is paralleled by a massive downregulation of total Oct4 and pRb levels and the acquisition of a Retinoblastoma pathway-regulated cell cycle program [11–14]. We show that RNAi-mediated knockdown of pRb from self-renewing mESCs reduces self-renewal potential, recapitulating effects normally observed upon Oct4 depletion or the induction of mESC differentiation. This indicates that high pRb protein levels improve the self-renewal potential of mESCs. In line with these data, mESC derived from a mouse model with constitutive activity of Cdk2 accumulate total pRb and hyperphosphorylated pRb and show a severely impaired differentiation potential . Together, this implies that pRb does not only impact on cell cycle regulations but also enhances the self-renewal potential of mESCs by limiting the initiation of cell differentiation programs. We further show in gain and loss of function experiments that Oct4 has a critical role in ensuring the hyperphosphorylation of pRb under self-renewal conditions by impairing its dephosphorylation by the PP1. In mammals, PP1 is the major pRb cell cycle-related phosphatase that binds to pRb and dephosphorylated pRb at the mitotic exit . Several PP1 interacting proteins have been reported that either increase or inhibit phosphatase activity [18, 23]. Here, we provide evidence that Oct4 drives the expression of the reported PP1 inhibitors Nipp1 (nuclear inhibitor of PP1) and Ccnf (Cyclin F) in self-renewing mESCs. This results in the protection of pRb from dephosphorylation by PP1. In line with our data, Nipp1 and Ccnf expression were previously shown to correlate with Oct4 expression in somatic cells . Nipp1 is a negative regulatory subunit of the PP1 complex, and Ccnf is anticipated to modulate the stability of PP1 via its presence in the SCF ubiquitin ligase complex [23, 29, 30]. Importantly, consistent with a role in inhibiting PP1 activity, depletion of Nipp1 and Ccnf results in efficient dephosphorylation of pRb that is associated with a loss of mESC self-renewal potential, pRb dephosphorylation and the acquisition of a cell cycle program of differentiated cells. The function of Nipp1 in inhibiting pRb dephosphorylation also provides an explanation for unsuccessful attempts in establishing mESC lines from Nipp1-deficient blastocysts . In line with a role in modulating early differentiation events, mice carrying a targeted deletion of Ccnf or Nipp1 die at day E6.5–E-8.5 or E10 of embryogenesis, respectively [31, 32].
Altogether, our data indicate that in self-renewing mESCs high Oct4 expression levels drive the expression of Nipp1 and Ccnf to inhibit the activity of PP1, thereby establishing pRb hyperphosphorylation as a key feature of mESC pluripotency.
Several miRNAs have been identified to control self-renewal or differentiation of mESCs [6–9, 33]. Here, we demonstrate that miR-335 is an important regulator of the Oct4-Nipp1/Ccnf-PP1-pRb axis that targets Oct4 and pRb on the post-transcriptional level at the onset of mESC differentiation. miR-335 is hosted by the imprinted Mest gene and we show that expression levels of both RNAs are dramatically upregulated during mESC differentiation. Impairing miR-335 upregulation during differentiation leads to ectopic Oct4, Nipp1, and pRb expression that finally leads to a reappearance of hyperphosphorylated pRb, impaired silencing of self-renewal gene expression patterns, and a delayed upregulation of genes involved in cell differentiation. In contrast, ectopically increased miR-335 levels efficiently reduced mESC self-renewal potential. Together, this identifies miR-335 as a novel self-renewal antagonizing miRNA. Interestingly, the fact that Mest is subjected to genomic imprinting could suggest that monoallelic expression of miR-335 could be important for early embryogenesis. A mouse model carrying a lacZ insertion into the Mest locus has been reported; however, the applied targeting strategy does not affect the miR-335 coding sequence . Consequently, new mouse models are required to investigate the relevance of miR-335 for the regulation of the Oct4-Nipp1/Ccnf-PP1-pRb axis during early embryonic development in vivo.
On the basis of our data, we propose a model where high Oct4 expression drives the expression of inhibitors of the major pRb phosphatase PP1 to ensure the hyperphosphorylation of pRb as hallmark-feature of self-renewing mESCs. The induction of differentiation leads to a contemporary repression of Oct4 and upregulation of miR-335 leading to the collapse of the self-renewal promoting pRb-Nipp1/Ccnf-PP1-pRb axis, the exit from self-renewal and the establishment of pRB-regulated cell cycle profile of differentiated cells. Our work not only integrates mESC cell cycle control into the functional repertoire of Oct4 but also anticipate an important role for Oct4 in the inactivation of the Retinoblastoma tumor suppressor pathway in human cancers with embryonic stem cell gene expression signatures.
We thank Dr. Anton Wutz for carefully reading the manuscript and helpful suggestions and Sabrina Germoni for help with subcutaneous injections and teratoma isolation. We thank Dr. Ramiro Mendoza Maldonado (ICGEB, Trieste, Italy) for help with FACS analysis. This work was supported by the Italian Association for Cancer Research (AIRC) grant Rif 42/08, 6352, a Regional Grant LR26/06; D. 2007/LAVFOR/1461, a Young Investigator Grant, Ministry of Health GR-2007-683407 to R.B and an Italian Association for Cancer Research (AIRC) grant Rif 10299 to S.S. M.S. and E.S. are the participants of the PhD program in Molecular Biomedicine, Università degli Studi di Trieste, Italy.
DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
The authors indicate no potential conflicts of interest.