Author contributions: R.A.: conception and design, manuscript writing, collection and/or assembly of data, data analysis and interpretation; J.E.: collection and/or assembly of data, data analysis and interpretation; S.D.: conception and design, financial support, data analysis and interpretation.
Author contributions: R.A.: conception and design, manuscript writing, collection and/or assembly of data, data analysis and interpretation; J.E.: collection and/or assembly of data, data analysis and interpretation; S.D.: conception and design, financial support, data analysis and interpretation.
Disclosure of potential conflicts of interest is found at the end of this article.
First published online in STEM CELLS EXPRESS July 13, 2010.
Pluripotent stem cells are characterized by the capacity to self-renew and to differentiate into all the cell types of the body. To identify novel regulators of pluripotency, we screened cDNA libraries (>30,000 clones) in P19 embryonal carcinoma cells for factors that modulate the expression of a luciferase reporter driven by the promoter of the pluripotency master regulator Nanog. Ninety confirmed hits activated the reporter and 14 confirmed hits inhibited the reporter by more than two-fold. The identified hits were evaluated by gain- and loss-of-functions approaches. The reporter-activating hits Timp2, Hig2, and Mki67ip promoted embryonic stem (ES) cell self-renewal when episomally overexpressed in ES cells, whereas the reporter-inhibiting hits PU.1/Spi1, Prkaca, and Jun induced differentiation of ES cells. Conversely, the knockdown of the activating hits Timp2, Mki67ip, Esrrg, and Dusp7 in ES cells induced differentiation, whereas the knockdown of the reporter-inhibiting hit PU.1/Spi1 led to inhibition of differentiation. One of the novel hits, the RNA-binding protein Mki67ip was further characterized, and found to be overexpressed in ES cells and in early development and downregulated during differentiation. The knockdown of Mki67ip led to the differentiation of ES cells, decreased growth rate, reduction in pluripotency markers, and induction of lineage-specific markers. In addition, colocalization and coimmunoprecipitation experiments suggest that Mki67ip promotes ES cell self-renewal via a mechanism involving nucleophosmin, a multifunctional nucleolar protein upregulated in stem cells and cancer. STEM Cells 2010; 28:1487–1497.
Embryonic stem (ES) cells are pluripotent cells derived from the inner cell mass of the embryo, and they hold great promise for their remarkable ability to self-renew indefinitely as well as their potential to differentiate into any cell type of the body. Details regarding the genetic and epigenetic regulations that control these remarkable features of ES cells have started to emerge, and they include a molecular circuitry controlled by major transcription factors, chromatin modifications, regulation by small RNAs, and pathways of signal transduction . For example, the transcription factors Oct4, Sox2, and Nanog maintain ES cell pluripotency by forming a regulatory core of interconnected molecular circuitry that promotes the expression of self-renewal genes and suppresses lineage-specific developmental genes [2–4].
Much emphasis has been placed on the regulation of pluripotency by transcription factors, chromatin modifications, and miRNAs, but little is known about the roles of RNA-binding proteins and factors involved in protein synthesis. Nucleophosmin (Npm1) is a nucleolar RNA-binding protein that mediates multiple functions including regulation of cell growth, proliferation, and development. The disruption of Npm1 in vivo leads to several developmental defects and embryonic lethality at midgestation . Npm1 has also been implicated in transformation, and it is one of the most frequent targets of genetic alterations in hematopoietic tumors . Although Npm1 has been implicated in the maintenance of adult stem cells [7, 8], its functions in ES cells are poorly understood. Mki67ip is a newly described RNA-binding protein, and it has been proposed to regulate ribosome biogenesis . Mki67ip roles, if any, in stem cell growth and maintenance have not been tested.
Despite the recent advances, our understanding of the molecular mechanisms controlling pluripotency remains elusive. To address this issue, genome-wide gain-of-function screens in ES cells have been successfully performed [10, 11]. These studies, however, utilized libraries of pooled cDNA clones, a technique that limits the complexity and sensitivity of the performed screens. More recently, studies performing genome-wide loss-of-function screens in ES cells utilizing individually arrayed siRNAs have identified novel factors required for pluripotency [12, 13]. These studies have demonstrated the great potential in applying forward genetics in the study of stem cells and their functions. We report here the application for the first time of individually arrayed cDNA libraries for the identification of regulators of pluripotency in a genome-wide functional screen. We identified several hits with known roles in maintaining pluripotency or lineage specification, and other novel factors with previously unknown ES cell-related functions. Several novel hits were found to be required for pluripotency or for proper differentiation.
MATERIALS AND METHODS
ES Cell Culture and Differentiation
Mouse R1, OG2, and E14-T ES cells were grown on γ-irradiated mouse embryonic fibroblasts (MEFs) in 0.1% gelatin-coated plates in ES cell medium (Knockout-DMEM supplemented with 15% knockout serum replacement, 0.1 mM nonessential amino acids, 2 mM L-glutamine, 0.1 mM 2-mercaptoethanol, 103 unit/ml leukemia inhibitory factor (LIF) (Chemicon, Millipore Billerica, MA, http://www.millipore.com). To generate embryoid bodies (EBs), ES cells were trypsinized and seeded in low-attachment plates in N2B27 medium (DMEM/F12 supplemented with 0.5× N2, 0.5× B27 (without vitamin A), 50 μg/ml BSA fraction V, 0.1 mM nonessential amino acids, 2 mM L-glutamine, and 0.1 mM 2-mercaptoethanol). Media was replaced every other day.
High Throughput Screening
The Nanog luciferase reporter (Nanog 5/3P) was kindly provided by A. Cooney (Baylor College of Medicine, Houston, TX, http://www.bcm.edu), and it is flanked by an upstream fragment of 2.5 kb of the 5′ promoter region of the mouse Nanog gene, and a downstream fragment of 1.0 kb of the 3′ UTR region of the mouse Nanog gene in pGL3-basic reporter vector . The arrayed cDNA libraries were from the Genomics Institute of the Novartis Research Foundation (GNF, San Diego, CA, http://www.gnf.org/), and contained mouse (7249) and human (3820) cDNA clones originally from Mammalian Gene Collection (MGC, mgc.nci.nih.gov) and 19,823 unique human cDNA clones from OriGene Technologies (Rockville, MD, www.origene.com). The clones (62.5 ng/well) were in mammalian expression vector (under control of the CMV promoter), and have been arrayed in 384-well plates, ready for reverse transfection. As a plate control, 62.5 ng of p53 cDNA was spotted in four wells per plate. To each well, we dispensed 20 μl serum-free αMEM containing 3.2 μl FuGENE-6 (Roche Diagnostics, Indianapolis, Indiana, http://www.roche-diagnostics.us/) and 40 ng Nanog reporter. After incubating the DNA-FuGENE-6 mixture for 30 minutes at room temperature, 50 μl of 10% FBS MEMα media containing 3000 P19 embryonal carcinoma (EC) cells were plated in each well. Two days later, the media was removed and 20 μl Luciferase Assay Reagent of the Bright Glo-Luciferase Assay System (Promega, Madison, WI, http://www.promega.com/) was added. After 10 minutes, the firefly luciferase activity was measured using CLIPR (Molecular Devices, Sunnyvale, CA, http://www.moleculardevices.com), and the light reading was normalized to the plate average and the Z-score for each well was determined. For the secondary screen, the primary hits were reordered from GNF and arrayed into new 384-well plates in quadruplicates at 60 ng per well. Reverse transfection was performed as above with 40 ng Nanog reporter and 5 ng pTK-RL reporter (expresses Renilla luciferase from the minimum TK promoter). Two days later, the firefly and Renilla luciferase activities were measured using the Dual-Luciferase Assay System (Promega). Relative light readings were normalized to that of the control (empty vector), and genes were ranked according to their fold of activation or inhibition.
Cloning into Episomal Vector and Transfection of ES Cells
For evaluation of gene functions in ES cells, we cloned the cDNAs for 20 hits, Nanog (positive control), and yellow fluorescent protein (YFP; negative control) into pPyCAGIP, an episomal vector that contains the polyoma virus origin of DNA replication, a cloning site under the constitutive CAG promoter and the puromycin-resistance gene (pac) . The cDNAs were first verified to be full-length by sequencing, and they were amplified using high fidelity platinum Taq DNA polymerase (Invitrogen, Carlsbad, CA, http://www.invitrogen.com), forward primers with 1× Flag and XhoI or BstXI extensions, and reverse primers with NotI extension. The polymerase chain reaction (PCR) fragments were then inserted directionally between the NotI and XhoI/BstXI sites of pPyCAGIP. The generated clones were confirmed to be mutation-free by sequencing. For the transfection of E14-T ES cells, which express polyoma large T antigen, cells were seeded at 7.5 × 103 cell/cm2 in complete growth medium. After overnight incubation, the cells were transfected with the constructs individually using Lipofectamine 2000 Reagent (Invitrogen). Puromycin selection (1.5 μg/ml) was applied 36 hours later to quickly remove untransfected cells, as determined by monitoring cells transfected with YFP (Venus) in the pPyCAGIP vector. For the reporter-inhibiting hits, 7 days after selection, cells were stained for alkaline phosphatase (AP) using a staining kit (Millipore), according to the manufacturer instructions. For the reporter-activating hits, cells were split 4 days postselection and replated at 5 × 103 cell/cm2 in complete growth medium. Following an overnight incubation, medium was replaced with basal N2B27 medium and changed daily. After 7 more days, cells were fixed and stained for AP.
Five distinct DNA-based small hairpin RNAs (shRNAs; Sigma-Aldrich, St. Louis, MO, http://www.sigmaaldrich.com) in the pLKO.1-puro vector (containing the pac gene) were tested per gene for RNAi efficiency in P19 EC cells by reverse transcription (RT)-PCR, and the two shRNAs with the best RNAi efficiency were selected. As a negative control, we used a nontargeting (NT) shRNA with sequence that does not target any human or mouse gene. For the knockdown in ES cells, we seeded cells at 2 × 104 cells/cm2 in 6-well plates in complete growth medium. The shRNAs were delivered by either liposome-based transfection or by infection with lentiviruses (all experiments unless otherwise specified). After 36 hours, media was replaced with complete ES cell media with 1.5 μg/ml puromycin. Selection in complete media was maintained for 4–7 days before further manipulation. To examine induction of lineage-specific markers over extended period of time, lentivirus-infected ES cells were passaged three times under selection before isolation of total RNA for real-time RT-PCR analysis.
Viral Generation and Infection
To produce the lentiviruses, HEK-293T cells were transfected with a cocktail containing the lentiviral shRNA plasmid (pLKO.1-puro), the packaging plasmid (psPAX2), and the envelope plasmid (pMD2.G) using Lipofectamine 2000 reagent (Invitrogen). Viral supernatants were collected 24 and 48 hours after transfection, filtered through a 0.45-mm pore size filter, and supplemented with 4 μg/ml polybrene before infection.
Confocal Microscopy and Colocalization Analysis
E14-T ES cells and P19 EC cells were transfected with flag-tagged Mki67ip, and 2 days later were split and distributed to gelatinized chamber slides (feeder-free). Two days later, cells were fixed with 4% paraformaldehyde for 10 minutes at RT, and permeabilized with 0.1% triton X-100 in phosphate buffered saline (PBS) for 10 minutes at RT. Cells were blocked with 5% horse serum in PBS for 1 hour at RT, and incubated for additional 1 hour at RT with monoclonal anti-Npm1 antibody and polyclonal anti-flag antibodies (Sigma-Aldrich). Stained cells were then incubated for 1 hour with anti-mouse and anti-rabbit donkey IgG conjugated with Alexa Fluor 488 and Alexa Fluor 555, respectively (Invitrogen). Images were acquired using Laser Sharp 2000 software and then imported and further analyzed for quantitative colocalization using three independent software packages: Laser Sharp [Bio-Rad (Hercules, CA, http://www.bio-rad.com)–Zeiss (Dublin, CA, http://www.zeiss.com)], LSM examiner (Zeiss Dublin, CA, http://www.zeiss.com), and Image J (NIH Imaging, http://rsbweb.nih.gov/nih-image/).
Coimmunoprecipitation and Western Blotting
P19 cells or E14-T ES cells grown in 10-cm dishes were transfected with Flag-tagged Mki67ip. After three days, cells were washed twice with PBS and incubated with 1 ml lysis buffer (50 mM Tris HCl, pH 7.4, with 150 mM NaCl, 1 mM EDTA, 1% TRITON X-100, and protease inhibitor cocktail) for 30 minutes on a shaker at 4°C. Immunoprecipitation (IP) reactions were carried out overnight at 4°C using monoclonal anti-FLAG-M2 affinity gel or monoclonal anti-Npm1 bound to Protein G agarose gel (Sigma-Aldrich). The immunoprecipitants were eluted by adding 20–40 μl 2× SDS buffer (125 mM Tris HCl, pH 6.8, with 4% SDS, 20% (v/v) glycerol, and 0.004% bromphenol blue) and boiling for 3 minutes. Western blot was done using anti-Flag and anti-Npm1 antibodies (Sigma-Aldrich).
Real-Time RT-PCR and Expression Analyses
RNA was extracted from cells using the RNeasy Plus Mini Kit in combination with QIAshredder (Qiagen, Valencia, CA, http://www. qiagen.com). One microgram of RNA was converted to cDNA using iScript cDNA Synthesis Kit, and one-tenth of the reaction was used for Real-time PCR performed using iQ SYBR Green Supermix (BioRad). The RT-PCR was performed using the following program: initial denaturation at 94°C for 30 seconds, 40 cycles of 94°C for 10 seconds, and 64°C for 30 seconds, and melt curve at 65–95°C. The expression of genes of interest was normalized to that of glyceraldehyde 3-phosphate dehydrogenase (GAPDH) in all samples. Heatmaps and clustering were generated with dChip program. To give the values, a normal distribution they were increased by 1 and base 2 logged. Clustering was done with correlation as the distance metric and centroid linkage method. Red indicates increased expression and blue indicates decreased expression relative to the mean transcript expression. The key range of standardized values is ±3. For the microarray expression analysis of screen hits, Affymetrix .cel files were downloaded from GEO (NCBIs Gene Expression Omnibus), and each array .cel file was normalized by using standard Affymetrix MAS5.0 (scaling to a target intensity of 500) using Affymetrix Expression Console (Affymetrix, Santa Clara, CA, http://www. affymetrix.com).
RESULTS AND DISCUSSION
Genome-Wide Functional Screen Identifies Candidate Modulators of Pluripotency
Nanog is exclusively expressed in pluripotent cells and its overexpression is sufficient for the self-renewal of mouse ES cells in the absence of LIF [10, 15]. To identify novel regulators of pluripotency, we performed a genome-wide gain-of-function screen in pluripotent P19 EC cells, gauging the ability of overexpressed genes to modulate the expression of an exogenous luciferase reporter driven by the Nanog promoter . The luciferase reporter acts as a detector of factors that directly or indirectly promote pluripotency or induce differentiation (supporting information Fig. S1). P19 EC cells were used for the primary screen because they are more readily transfectable than ES cells, enabling the high-throughput screening of cDNA clones on a genome-wide scale. In addition, Nanog is expressed at lower levels in P19 EC cells relative to its level in ES cells, providing a dynamic range that facilitates the identification of both activators and inhibitors of Nanog expression (supporting information Fig. S1A). To screen the individually arrayed cDNA libraries consisting of 30,892 mouse and human clones (in 384-well format), P19 EC cells were reversely cotransfected with the spotted cDNAs along with the Nanog reporter (Fig. 1A). As a positive control, four empty wells were spotted with the cDNA of the strong Nanog inhibitor p53 in each plate of the library . The luciferase activity was measured 2 days post-transfection, a time-frame optimized to identify primarily activators, and to lesser degree inhibitors (supporting information Fig. S1), as we were more interested in identifying enhancers of pluripotency. Readings were normalized to plate averages, and the Z-score was calculated for each value to correct for variation among plates. Overall, 159 hits had a Z-score >2.5 and 70 hits had a Z-score of <−1.5 (supporting information Table S1). To confirm the primary hits, identified cDNA clones were reordered and individually arrayed in new 384-well plates, and a secondary screen was performed in identical fashion to the primary screen. To rule out toxicity or nonspecific effects, P19 EC cells were transfected with an additional plasmid constitutively expressing the Renilla luciferase as an internal reference. Two days post-transfection, firefly luciferase activity was normalized to that of Renilla, and all values normalized to the plate control (empty vector). Ninety hits led to >2-fold activation of the Nanog luciferase reporter and 14 hits led to >2-fold inhibition (Fig. 1B and supporting information Table S2).
Submitting the list of confirmed hits to the ingenuity pathways analysis revealed significant association between the hits and signaling pathways known to modulate pluripotency or stem cell differentiation [transforming growth factor beta (TGF-β), Wnt, p53, integrin, ERK/mitogen-activated protein kinase (MAPK), and PI3K/Akt signaling pathways; supporting information Table S3]. Furthermore, several Nanog reporter-activating hits are known to support pluripotency (Akt, Cnot3, Wdr5, and Pim1), and several reporter inhibiting hits are established inhibitors of pluripotency (p53) or regulators of lineage specification (e.g., Dlx2/3/5, Neurog1, and Foxa1). Functional classification of hits showed that the most significant functions associated with the hits were regulation of gene expression, cellular, organ and embryonic developments, cancer, cellular proliferation, and cell cycle (Fig. 2A). Exploring the published databases [2–4] of the promoter occupancy of major pluripotency transcription factors revealed that significant proportion of our hits are components of the regulatory pluripotency network. The promoters of 49% of the confirmed hits were reported to be bound in ES cells by at least one of the pluripotency factors Oct4, Sox2, Nanog, c-Myc, or Klf4 (Fig. 2B and supporting information Table S4), with the promoters of 23% of the hits being cobound by two or more factors (Fig. 2C). Examining the chromatin modifications corresponding to a particular gene can predict its expression status in the cell assayed . To explore the chromatin status of our hits in ES cells, we examined the corresponding H3 di- and tri-methylation (me2 and me3) and the association with the initiating and elongating RNA Polymerase II (Pol II) as reported in published data [18–20]. We anticipated that the identified activators of the Nanog reporters would be expressed in ES cells, and thus might have more active chromatin compared with the identified inhibitors. As expected, the activators associated more significantly in ES cells with hallmarks of transcription elongation (H3K79me2, initiating and elongating Pol II), compared with the inhibitors (Fig. 2D and supporting information Table S5). The H3K27me3, which marks gene suppression, associated more with the inhibitors than with the activators. No significant differences were observed in the association with H3K4me3, which marks the promoters of the majority of genes in both ES and differentiated cells .
To examine whether the associated epigenetic status of our hits reflect the expected expression profile in ES cells, we looked at the expression profiles of the confirmed hits as reported in the microarray databases from GEO (Gene Expression Omnibus) for five human ES cell-lines and two mouse ES cell-lines. As expected, the Nanog reporter-activating hits were more likely (p value > .05) to be expressed in ES cells than the inhibiting hits (supporting information Table S6). To confirm this further, we examined the expression of 12 activators associated with the H3K79me2 methylation and 10 inhibitors associated with the H3K27me3 methylation in ES cells, MEFs, and mouse embryo (E10) by quantitative reverse transcription polymerase chain reaction (RT-PCR). Expression analysis revealed that all the activators tested but one were expressed at higher levels in ES cells and mouse embryo relative to their expression in MEFs, whereas all the inhibitors but one were expressed at higher levels in MEFs and mouse embryo relative to their expression in ES cells (Fig. 2E). The relatively high expression of the reporter inhibitors in early development is consistent with their association with the bivalent domain (K4-K27) that defines developmental regulators . Collectively, these results confirmed that our Nanog reporter-based screen successfully identified candidate regulators of pluirpotency and lineage specification.
Gain- and Loss-of-Function Studies of Hits in ES Cells Reveal Factors Involved in the Self-Renewal and Differentiation of ES Cells
Reliance on gene expression analysis alone is not sufficient to predict gene involvement in pluripotency. For example, c-Myc and Klf4 are master regulators of pluripotency, but their expression is not restricted to pluripotent cells, whereas the embryonic stem cell (ESC)-specific gene 1 is dispensable for pluripotency of ESCs despite its specific expression in pluripotent cells. To functionally test the involvement of our hits in regulating pluripotency of ES cells, we randomly selected 16 activators and four inhibitors for episomal overexpression in ES cells (supporting information Table S6). We cloned the cDNAs for the hits, Nanog (positive control) and YFP (negative control) in the episomal pPyCAGIP vector [10, 22], and the constructs were transfected into the E14-T ES cells (which express large T antigen) in complete growth medium, and selection with puromycin was applied 36 hours later. Following selection, ES cells were cultured in either complete growth medium (for inhibitors) or serum-free differentiation N2B27 medium with no added cytokines (for activators). As expected, overexpressing Nanog blocked differentiation of ES cells cultured in N2B27 media, as ES cells maintained high levels of ES cell marker AP. Of the 16 activators tested, the overexpression of three novel hits (Hig2, Timp2, and Mki67ip) inhibited the loss of AP in ES cells, whereas the overexpression of the rest of the clones had no or little effect (Fig. 3A). Examining other ES cell-specific markers by quantitative RT-PCR confirmed that overexpression of Timp2 and Mki67ip, but not Hig2, inhibited ES cell differentiation, in particular as judged by the high-level expression of Nanog and Rex1 (Fig. 3C). On the other hand, overexpression of all the reporter-inhibiting hits tested (PU.1/Spi1, Prkaca, c-Jun, and, to a lesser degree, Pja1) induced rapid differentiation of ES cells cultured in complete growth media, as judged by the diminished expression levels of AP, Nanog, Oct4, Sox2, and Rex1 (Fig. 3B, 3D). In addition, overexpression of PU.1/Spi1 and Prkaca induced high levels of cell death in ES cells.
To gain better insight into the functional involvement of our hits in ES cell self-renewal and differentiation, we knocked down the expression of the screen hits in ES cells using shRNAs. ES cells expressing NT shRNA maintained compact colonies that expressed high levels of AP when cultured in complete growth medium. In contrast, ES cells expressing shRNA targeting Nanog differentiated extensively, became dispersed and expressed lower AP levels. Five of the reporter-activating hits and two of the reporter-inhibiting hits (PU.1/Spi1 and Prkaca) were targeted by shRNAs. The knockdown of four activators (Timp2, Mki67ip, Esrrg, and Dusp7) induced significant differentiation of ES cells grown in complete ES cell media (Fig. 4A). On the other hand, the knockdown of the inhibiting hit PU.1/Spi1 had a positive impact on ES cells grown in complete ES cell media and impaired the differentiation of ES cells when cultured in the differentiation N2B27 media (Fig. 4A, 4B). Surprisingly, the knockdown of the inhibiting hit Prkaca induced moderate level of differentiation, less than what is observed when overexpressed, suggesting that Prkaca is required at an intermediate level. To rule out off-target effects of the shRNAs, the knockdown results were confirmed using a different shRNA for each gene tested. The knockdown of hits in ES cells using the independent shRNAs led to similar results (supporting information Fig. S2).
The knockdown of the activating hits Mki67ip and Timp2, and the inhibiting hit PU.1/Spi1 led to the most dramatic phenotypes. To examine their loss-of-function effects further, we analyzed the expression of pluripotency markers in the knockdown ES cells using quantitative RT-PCR. As expected, the knockdown of Nanog in ES cells cultured in complete growth media led to significant reductions in the expression levels of Oct4, Sox2, Rex1, and Nanog itself. The knockdown of the activating hits Mki67ip and Timp2 also led to significant reductions in the levels of Nanog, Oct4, Sox2 and Rex1, whereas the knockdown of the inhibiting hit PU.1/Spi1 led to the induction of Oct4, Sox2, and Rex1 (Fig. 4D). The knockdown of PU.1/Spi1 led also to higher levels of the pluripotency markers in ES cells cultured in differentiation media (Fig. 4C). Collectively, the gain- and loss-of-function experiments were consistent with the results of our screen, and revealed important roles for some of the tested hits in the regulation of pluripotency.
Mki67ip Colocalizes and Interacts with Npm1 in Pluripotent Cells
Functional analysis of our hits revealed an enrichment of factors involved in RNA post-transcriptional modification and RNA trafficking (Fig. 2A). One of our hits, Mki67ip (Mki67 [FHA domain]) interacting nucleolar phosphoprotein) is a poorly characterized RNA-binding protein that has not been linked to ES cell-related functions before. We therefore aimed to further characterize its functions in ES cells. To gain insight into the function of Mki67ip in ES cells, we looked at other nucleolar proteins that have known stem cell-related functions and similar expression profiles to that of Mki67ip. One such protein is Npm1, a multifunctional nucleolar protein that is expressed at high levels in ES cells and in cancer, and is essential in development and organogenesis [5, 6, 23]. Npm1 has established roles in supporting the self-renewal and survival of adult stem cells [7, 8], but its functions in ES cells are less characterized. The expression profile of Npm1 was very similar to that of Mki67ip, as both factors were expressed at high levels in ES cells and at early developmental stages (supporting information Fig. S3). Furthermore, the expression levels of Mki67ip and Npm1 decreased significantly on retinoic acid-induced ES cell differentiation (Fig. 5A, 5B) and ES cell spontaneous differentiation (supporting information Fig. S4). The similarity in the expression profiles of Mki67ip and Npm1 prompted us to examine their cellular localization in pluripotent cells by confocal microscopy. Mki67ip and Npm1 were mostly localized to nucleoli during interphase and were highly dispersed during mitosis. Interestingly, we observed significant colocalization of Mki67ip and Npm1 in the nucleoli and nucleoplasm of EC cells and ES cells (Fig. 5C). We next aimed to confirm the cellular colocalization of Mki67ip and Npm1 by coimmunoprecipitation. First, we transfected P19 EC cells with the Flag-tagged forms of YFP or Mki67ip, and performed IP with anti-Flag or anti-Npm1 antibodies. Npm1 was coimmunoprecipitated efficiently with the Flag-tagged Mki67ip, but not with the Flag-tagged YFP, even though the amount of YFP exceeded that of Mki67ip (Fig. 5D). Similarly, the Flag-tagged Mki67ip, but not the Flag-tagged YFP, was efficiently coimmunoprecipitated with the Npm1 (Fig. 5E). We also confirmed the IP results in ES cells. The Flag-tagged Mki67ip, but not the Flag-tagged YFP, coimmunoprecipitated Npm1 from extracts of E14-T ES cells (Fig. 5F). Overall, our results demonstrate that Mki67ip and Npm1 share similar expression profiles, colocalize, and interact in pluripotent cells.
Mki67ip and Npm1 Are Required for Maintaining the Undifferentiated State of ES Cells
To better characterize the functions of Mki67ip and Npm1 in ES cells, we generated stable ES cell lines expressing shRNAs targeting Mki67ip and Npm1 with 70% and 90% knockdown efficiencies, respectively (Fig. 6B). The Mki67ip- and Npm1-depleted ES cells displayed a flat and dispersed morphology, and they expressed lower levels of the pluripotency factors Nanog, Oct4, Sox2, and stage-specific embryonic antigen-1 (SSEA1) as determined by immunefluorescence (Fig. 6A). Quantitative RT-PCR analysis also revealed decreased expression levels of Nanog, Oct4, Sox2, and Rex1 in the Mki67ip and Npm1 knockdown cells (Fig. 6B). The inhibition of the pluripotency markers by the knockdown of Mki67ip and Npm1 was accompanied by induction of lineage-specific markers (Fig. 6C). These results therefore indicate the involvement of Mki67ip and Npm1 in the maintenance of the pluripotent state and the suppression of genes associated with lineage determination.
When grown in suspension, ES cells form spheres (EBs) of differentiated cells, which might reflect early stages of embryogenesis. We tested the ability of the Mki67ip/Npm1 knockdown ES cell lines to form EBs. Interestingly, our initial attempts to generate EBs from the ES cells expressing the Mki67ip and Npm1 shRNAs were unsuccessful, even though ES cells expressing the NT shRNA efficiently formed large and healthy EBs. However, the ES cells expressing the Mki67ip and Npm1 shRNAs did form small EBs when seeded at higher cell density, but these EBs lost Oct4 expression at much faster rates compared with the control EBs (Fig. 6D). The knockdown of Mki67ip and Npm1 therefore does not only induce differentiation of ES cells but also impairs self-renewal of ES cells. Consistent with this observation, the ES cell lines expressing the Mki67ip and Npm1 shRNAs grew at slower rates compared with ES cells expressing the NT shRNA (Fig. 6E). Our results thus indicate that Mki67ip and Npm1 are required for maintaining the undifferentiated state of ES cells.
To identify novel regulators of pluripotency, we have systematically screened >30,000 cDNA clones for the ability to modulate the expression of an exogenous Nanog reporter in EC cells. Multiple lines of evidence validated our screening approach and confirmed the stem cell relevance of the identified hits. First, the screen identified several Nanog reporter-activating hits that are known to support pluripotency (Akt, Cnot3, Wdr5, and Pim1), and reporter inhibiting hits that are established inhibitors of pluripotency (p53) or regulators of lineage specification (Dlx2/3/5, Neurog1, and Foxa1). The screen, however, missed some known regulators of pluripotency, possibly due to technical limitations of the screen (e.g., the cDNA library lacked some known ES cell regulators). In addition, the differences in the expression profiles between P19 cells and ES cells (supporting information Fig. S1), suggest a possible P19 cell-specific bias that might have impacted the results of the primary screen. Nonetheless, the screen identified several components or interacting players in the major pathways known to control ES cell fate (such as BMP, Wnt, p53, PI3K/Akt, and ERK signaling pathways). Second, it was possible to generally predict the chromatin status and expression profiles of the identified hits in ES cells based on whether they activated or inhibited the Nanog reporter. Third, many of the identified hits are components of the pluripotency circuitry, as evident from examining the promoter occupancy of the pluripotency master regulators Nanog, Oct4, Sox2, Klf4, and c-Myc. Fourth, a quarter of our hits (n = 24) overlapped with the lists of stem cell-specific factors categorized by the systematic consolidation of data from multiple mouse and human stem cell studies . Gain- and loss-of-function experiments provided additional validation of our screening approach. Several hits were found to be required to maintain ES cells in the undifferentiated state (Mki67ip, Timp2, Esrrg, and Dusp7), whereas others (PU.1/Spi1, Prkaca, and c-Jun) induced differentiation of ES cells when overexpressed.
Mki67ip had one of the strongest effects when depleted by RNAi, reflecting its requirement for pluripotency. Mki67ip is a newly characterized RNA-binding protein that has been shown to interact with the mitotic marker pKi-67 following mitosis-dependent phosphorylation of Mki67ip . Mki67ip, like Npm1, has been suggested to function in ribosome biogenesis , but the details of this involvement and whether Mki67ip has other functions are largely unknown. Npm1 is essential for embryonic development and the maintenance of genomic stability . Npm1 knockout mice have defective organogenesis and die between embryonic day E11.5 and E16.5, likely due to defects in primitive hematopoiesis . Npm1 has also been implicated in regulating adult stem cell growth [7, 8], as well as cancer . Mki67ip shared a strikingly similar expression profile to that of Npm1, a multifunctional nucleolar protein with established roles in adult stem cells and cancer. Mki67ip and Npm1 colocalized to a large degree in the nucleoli and nucleoplasm of ES and EC cells, and coimmunoprecipitated in the same protein complex, suggesting that the two proteins might cooperate and mediate similar functions in pluripotent cells. Mki67ip and Npm1 are expressed at high levels in undifferentiated ES cells and the mouse embryo, and downregulated on ES cell differentiation. The knockdown of Mki67ip and Npm1 induced differentiation, indicating their requirement for proper maintenance of ES cells in the undifferentiated state. The knockdown of Mki67ip and Npm1 inhibited the expression of pluripotency markers, and induced the expression of differentiation markers. Furthermore, ES cells overexpressing Mki67ip/Npm1 shRNAs had impaired self-renewal capacity. Our results therefore identify a novel role for the nucleolar protein Mki67ip and extend the functions of the nucleolar protein Npm1 from supporting the self-renewal of adult stem cells and cancer to the regulation of ES cells functions. Our work highlights the significance of the nucleolus in the regulation of pluripotency, as evident also from the already established roles of the nucleolar protein nucleostemin in ES cells .
Interestingly, Mki67ip and Npm1 are upregulated in cancer, with Npm1 being one of the most frequent targets mutated in hematopoietic tumors [6, 26]. Npm1 has been shown recently to be essential for the transforming activity of c-Myc . In addition, Mki67ip and Npm1 are transcriptional targets of c-Myc, and both are implicated in ribosome biogenesis [28–30]. Although c-Myc acts globally to regulate gene transcription, groups of genes involved in ribosome biogenesis and protein synthesis are over-represented in the c-Myc target gene network [31, 32]. It has been proposed that c-Myc could promote tumorigenesis by upregulating genes involved in ribosome biogenesis [32, 33]. Accordingly, it is intriguing to propose that some factors with RNA-related functions such as c-Myc, Npm1, and Mki67ip might promote ES cell self-renewal and/or tumorigenesis through upregulation of ribosome biogenesis. Additional studies would be needed in the future to address some of these questions.
We thank Paul DeJesus for his technical support during screening, William Kiosses for his help with the confocal imaging and analysis, and David Robbins for his help with gene expression analysis.
Disclosure of Potential Conflicts of Interest
S.D. is a cofounder of Fate Therapeutics. The other authors have no financial interests to disclose.