pim-1 and pim-3 encode serine/threonine kinases involved in the regulation of cell proliferation and apoptosis in response to cytokine stimulation. We analyzed the regulation of pim-1 and pim-3 by the leukemia inhibitory factor (LIF)/gp130/signal transducer and activator of transcription-3 (STAT3) pathway and the role of Pim-1 and Pim-3 kinases in mouse embryonic stem (ES) cell self-renewal. Making use of ES cells expressing a granulocyte colony-stimulating factor:gp130 chimeric receptor and a hormone-dependent signal transducer and activator of transcription-3 estrogen receptor (STAT3-ERT2), we showed that expression of pim-1 and pim-3 was upregulated by LIF/gp130-dependent signaling and the STAT3 transcription factor. ES cells overexpressing pim-1 and pim-3 had a greater capacity to self-renew and displayed a greater resistance to LIF starvation based on a clonal assay. In contrast, knockdown of pim-1 and pim-3 increased the rate of spontaneous differentiation in a self-renewal assay. Knockdown of pim-1 and pim-3 was also detrimental to the growth of undifferentiated ES cell colonies and increased the rate of apoptosis. These findings provide a novel role of Pim-1 and Pim-3 kinases in the control of self-renewal of ES cells.
Disclosure of potential conflicts of interest is found at the end of this article.
The propagation of mouse embryonic stem (ES) cells is dependent on the presence of leukemia inhibitory factor (LIF), which engages a heterodimeric receptor complex consisting of LIF receptor (LIFR) and gp130. This complex activates the Janus-associated tyrosine kinases (JAKs), which phosphorylate the receptor chains. The phosphorylated tyrosines then act as docking sites for the signal transducer and activator of transcription-3 (STAT3) . Phosphorylation of STAT3 by JAKs, followed by its dimerization and nuclear translocation, plays a crucial role in self-renewal of ES cells . Hence, expression of a dominant-negative STAT3 mutant in ES cells forces differentiation , whereas expression of a hormone-dependent STAT3 mutant that can be activated directly by estradiol sustains self-renewal without the addition of LIF . Activation of the LIFR:gp130 receptor also induces JAK-dependent tyrosine phosphorylation of the Sarc-homology domain phosphatase-2 (SHP-2) phosphatase. However, SHP-2 recruitment opposes STAT3-dependent self-renewal and facilitates differentiation by promoting the activation of the Ras/mitogen-activated protein kinase pathway .
A number of genes whose expression is activated by LIF have been identified in ES cells [4, 6, , –9], although only two, c-myc and Krüppel-like factor (klf) 4, have been shown to contribute to the inhibition of differentiation. c-myc is a target gene of the LIF/STAT3 pathway. Forced expression of a hormone-dependent Myc-ER temporarily inhibits the differentiation of ES cells induced by withdrawal of LIF . Similarly, forced expression of Klf-4 delays differentiation induced by LIF deprivation .
The pim genes encode the serine/threonine kinases Pim-1, Pim-2, and Pim-3, involved in the regulation of cell growth and apoptosis . Pim-1 and Pim-2 both have the ability to cooperate with Myc in experimental T-cell lymphomagenesis [11, 12], a finding that established Pim-1 and Pim-2 as proto-oncogenes and important players in the process of malignant transformation. Pim kinases contribute to the regulation of cell proliferation in response to stimulation by cytokines and growth factors. The expression of the pim-1 gene can be induced via the activation of STAT3 and STAT5, which bind directly to the pim-1 promoter in eosinophils . In lymphoid cells, pim-1 is activated by interleukin-6 via STAT3 and regulates proliferation in cooperation with Myc . All three Pim kinases also play an important role in the prevention of cell death [15, 16].
In this study, we examined the regulation of the pim-1 and pim-3 genes by the LIF/STAT3 pathway, as well as their role in the control of proliferation and differentiation, in ES cells. The data presented show that Pim-1 and Pim-3 play an important role in maintaining the ES cell identity by inhibiting differentiation and apoptosis.
Materials and Methods
The hormone-dependent STAT3-ERT2 transcription factor was generated by fusing the coding sequence of the mouse STAT3 transcription factor to a 5′-XhoI/EcoRI-3′ fragment containing the entire ERT2 domain . The resulting cDNA was subcloned into the EcoRI site in pPCAGIZ  to generate pPCAGIZ-STAT3-ERT2. Full-length cDNA encoding mouse Pim-1 was subcloned into the blunted BstXI site in pPHCAG  and into the blunted XhoI site in pPHPGK  to generate pPHCAG-pim1 and pPHPGK-pim1, respectively. Full-length cDNA encoding mouse Pim-3 was subcloned into the blunted EcoRI site in pPCAGIZ  and into the XhoI site in pPHPGK to generate pPCAGIZ-pim3 and pPHPGK-pim3, respectively.
The BLOCK-iT lentiviral RNAi Gateway vector kit (ref. K4943-00; Invitrogen, Carlsbad, CA, http://www.invitrogen.com) was used to generate lentiviral vectors expressing small hairpin (sh) RNA directed to green fluorescent protein (GFP), pim-1, and pim-3, according to the manufacturer's instructions. In the pLenti6/BLOCK-iT backbone vector, the blunted KpnI (blunt)/XhoI fragment was replaced by a 1.3-kilobase ClaI (blunt)/XhoI fragment containing the PGK promoter and the neor selectable gene. Oligonucleotides encoding shRNAs for GFP, pim-1, pim-3, and stat3 were subcloned into the resulting pLenti6/BLOCK-iT-PGKneo vector to generate pLenti6/BLOCK-iT-PGKneo-GFP, pLenti6/BLOCK-iT-PGKneo-pim1, pLenti6/BLOCK-iT- PGKneo-pim3, and pLenti6/BLOCK-iT-PGKneo-stat3, respectively. The sequences of shRNA are as follows: GFP, 5′-GCCACAACGTCTATATCATTTCAAGAGAATGATATAGACGTTGTGGC-3′; pim-1, 5′-GCAAGACCTCTTCGACTTTATTTCAAGAGAATAAAGTCG-AAGAGGTCTTGC-3′; pim-3, 5′-CACGGTCTACACTGACTTTGA-TGTTCAAGAGACATCAAAGTCAGGTAGACCGTG-3′; stat3, 5′GGAGCTGTTCAGAAACTTATTCAAGAGATAAGTTTCTGA- ACAGCTCC-3′.
ES Cell Culture, Electroporation, and Infection
All ES cell lines were routinely cultured in Glasgow's modified Eagle's medium supplemented with 10% fetal calf serum and 1,000 U/ml LIF as described previously . To induce differentiation, cells were allowed to aggregate in hanging drops in ES cell medium without LIF . After 2 days, embryoid bodies were collected and further grown in suspension for 1–4 days. For episomal supertransfection, embryonic day 14 T (E14/T) cells  were electroporated with 20 μg of supercoiled vectors at 200 V and 960 μF in a 0.4-cm cuvette. Cells were plated at 5 × 104 cells per 10-cm dish and cultured in the presence of 100 μg/ml hygromycin B (Roche Diagnostics, Basel, Switzerland, http://www.roche-applied-science.com) or 1 μg/ml zeocine (Cayla-InvivoGen, Toulouse, France, http://www.cayla.com) for 7 days. Resistant colonies were pooled and further propagated in selection medium for 8–10 days prior to analysis.
Lentiviral vectors expressing shRNA were produced using the BLOCK-iT lentiviral RNAi expression system (ref. K4944-00; Invitrogen) according to the manufacturer's instructions. For lentiviral infection, CGR8 were plated at a density of 104 cells in 24-well plates in 1 ml of medium composed of 100 μl of ES cell medium and 900 μl of culture supernatant from virus-producer cells. After 48 hours, ES cells were trypsinized, replated at 104 cells per gelatin-coated 10-mm tissue culture dish, and further cultured in complete ES cell medium supplemented with 250 μg/ml G418 for 6 days.
Stimulation with LIF, Granulocyte Colony-Stimulating Factor, and Tamoxifen: Cell Lysates and Immunoblotting
ES cells were plated at a density of 2 × 106 cells per 10-cm dish and cultured for 24 hours in medium lacking LIF. The following day, cells were stimulated with 10,000 U/ml LIF, 30 ng/ml granulocyte colony-stimulating factor (G-CSF) (Calbiochem, San Diego, http://www.emdbiosciences.com), or 100 nM 4′-hydroxytamoxifen (4′OHT) (Calbiochem) for the times indicated. Cells were then washed and scraped off with ice-cold phosphate-buffered saline (PBS), centrifuged, and frozen at −80°C. Cell pellets were lysed in 20 mM Hepes (pH 7.4), 100 mM NaCl, 50 mM NaF, 1% Triton X-100, 10% glycerol, 1 mM dithiothreitol, 1 mM sodium orthovanadate, 1 mM phenylmethylsulfonyl fluoride, and cocktail of protease inhibitor (Roche Diagnostics) for 1 hour at 4°C. Protein lysates were then cleared by centrifugation (14,000 rpm for 20 minutes). For immunoblotting, 30 μg of total proteins were resolved by SDS-polyacrylamide gel electrophoresis and electroblotted onto nitrocellulose. After overnight treatment with blocking buffer (50 mM Tris-HCl, pH 7.6, 150 mM NaCl, 5% dry milk), the membranes were probed with specific monoclonal or polyclonal antibodies (anti-Pim1 [sc-13513; Santa Cruz Biotechnology Inc., Santa Cruz, CA, http://www.scbt.com] and anti-Cdc25A [sc-7389; Santa Cruz Biotechnology]). Blots were incubated with horseradish peroxidase-coupled anti-mouse, anti-rabbit, or anti-goat IgG and developed using enhanced chemiluminescence reagents (Amersham Biosciences, Piscataway, NJ, http://www.amersham.com).
Semiquantitative and Real-Time Polymerase Chain Reaction
RNA was extracted using RNeasy kits with on-column DNase digestion, and reverse transcription was carried out with Omniscript, according to the manufacturer's recommendations (Qiagen, Hilden, Germany, http://www.qiagen.com). Polymerase chain reactions (PCRs) were performed with a PerkinElmer thermal cycler (PerkinElmer Life and Analytical Sciences, Boston, http://www.perkinelmer.com), operating on a regimen of 96°C for 5 seconds, 55°C–57°C (according to primers) for 15 seconds, and 72°C for 60 seconds, for 35 cycles, followed by 72°C for 10 minutes. Primers and annealing temperatures used to detect Nanog, Oct-4, pim-1, pim-2, pim-3, cdc25A, rex-1, and β-actin by real-time PCR were as follows:
Nanog: cDNA length, 438 base pairs (bp); annealing temperature, 55°C; 5′ primer (5′ to 3′), TACCTCAGCCTCCAGCAGA; 3′ primer (5′ to 3′), CCTCCAAGTCACTGGCAG
Oct-4: cDNA length, 974 bp; annealing temperature, 56°C; 5′ primer (5′ to 3′), ATGGATCCTCGAACCTGGC; 3′ primer (5′ to 3′), TCAGTTTGAATGCATGCATGGGAG
pim-1: cDNA length, 231 bp; annealing temperature, 55°C; 5′ primer (5′ to 3′), GCCCTCCTTTGAAGAAATCC; 3′ primer (5′ to 3′), GGACCTGGAGTCTGGAATGA
pim-2: cDNA length, 212 bp; annealing temperature, 55°C; 5′ primer (5′ to 3′), ACATGGTCTGTGGGGACATT; 3′ primer (5′ to 3′), TCCTTTGGAGGAGTTGATGG
pim-3: cDNA length, 167 bp; annealing temperature, 57°C; 5′ primer (5′ to 3′), AGCAGTGACCTCTGACCCCT; 3′ primer (5′ to 3′), TCAAGTATCCACCCAGGGCA
cdc25A: cDNA length, 711 bp; annealing temperature, 52°C; 5′ primer (5′ to 3′), TCTGTCTAGATTCTCCTGG; 3′ primer (5′ to 3′), CAGATGCCATAATTTCTGGAG
rex-1: cDNA length, 130 bp; annealing temperature, 55°C; 5′ primer (5′ to 3′), CGTGTAACATACACCATCCG; 3′ primer (5′ to 3′), GAAATCCTCTTCCAGAATGG
β-actin: cDNA length, 200 bp; annealing temperature, 55°C; 5′ primer (5′ to 3′), TGAAACAACATACAATTCCATCATGAAGTGTGA; 3′ primer (5′ to 3′), AGGAGCGATAATCTTGATCTTCATGGTGCT.
Quantitative PCR was performed using the LightCycler 1.5 system and the LightCycler Fast Start DNA Master SYBR Green I kit (Roche Diagnostics) according to the manufacturer's instructions. Reactions were carried out in a total volume of 20 μl, comprising 0.4 μM each primer, 3.5 mM MgCl2, 2 μl of SYBR Green, and 2 μl of diluted cDNA. Amplification and online monitoring were performed using the LightCycler 1.5 system (Roche Diagnostics). Following 40 amplification cycles, melt-curve analyses were performed to verify that only the desired PCR product had been amplified. PCR efficiency of both the target and reference genes was calculated from the derived slopes of standard curves by the LightCycler software (version 4.0; Roche Diagnostics). These PCR efficiency values were used to calculate the relative quantification values for calibrator-normalized target gene expression by the LightCycler relative quantification software (version 4.0). In all cases, target genes were normalized to β-actin. A β-actin cDNA was used as normalization control to ensure constant β-actin levels in all samples regardless of experimental conditions.
Detection of Alkaline Phosphatase Activity
Dishes were fixed in methanol for 15 minutes and then stained for 15 minutes with a solution containing 1 mg/ml Fast Red TR salt (Sigma-Aldrich, St. Louis, http://www.sigmaaldrich.com) dissolved in 0.1 M Tris, pH 9.2, containing 200 mg/ml naphtol AS-MX phosphate [21, 22].
Detection of Stem Cell-Specific Embryonic Antigen-1 Expression
ES cells were trypsinized, and the cell suspension incubated in culture medium for 6 hours on a rocker platform. Cells were subsequently Fc-blocked by treatment with 1 μg of human IgG per 105 cells for 15 minutes at room temperature. Then, 105 cells in 25 μl of PBS supplemented with 0.5% bovine serum albumin were incubated for 30–45 minutes at 4°C with 50 μl (50 μg/ml) of the phycoerythrin-conjugated mouse monoclonal anti-stem cell-specific embryonic antigen-1 (anti-SSEA1) (ref. FAB2155P; R&D Systems Inc., Minneapolis, http://www.rndsystems.com). Cells were washed twice with PBS. Fluorescence intensity was determined by flow cytometry on a FACScan instrument (FACSCanto II; BD Biosciences, San Diego, http://www.bdbiosciences.com), and data acquisition was performed using the FACSDiva software (BD Biosciences).
The ApopTag Red In Situ Apoptosis Detection Kit (ref. S7165; Abcys, Paris, France, http://www.abcysonline.com) was used to identify apoptotic cells. All reactions were performed according to the manufacturer's instructions.
Influence of LIF on the Transcriptional Regulation of pim-1 and pim-3
Expression of pim-1, pim-2, and pim-3 was examined by semiquantitative RT-PCR in undifferentiated CGR8 cells and in 1–6-day-old embryoid bodies (Fig. 1A). ES cells expressed elevated levels of pim-1 and pim-3 transcripts, both of which declined abruptly during the first day of differentiation. By contrast, the pim-2 transcript level increased during the first 2 days of differentiation and dropped from third day onwards. Thus, pim-1 and pim-3 exhibit a unique expression pattern characterized by a drop in transcript levels within 24 hours following withdrawal of LIF. Downregulation of pim-1 and pim-3 expression takes place prior to the loss of pluripotency marker Oct-4 , occurring between day 3 and 4.
To examine the role of LIF in the regulation of pim-1 and pim-3 expression, CGR8 ES cells were LIF-deprived for 24 hours and then restimulated with 10,000 U/ml LIF for 1 hour. Withdrawal of LIF for 24 hours did not induce loss of ES cell pluripotency (supplemental online data 1). This stimulation protocol led to a strong activation of the LIF signaling pathway revealed by a 14-fold increase in the mRNA level of the LIF/STAT3 target gene junB [4, 6] (data not shown). pim-1 and pim-3 transcript levels increased 1.3 ± 0.1-fold (p < .06) and 1.74 ± 0.13-fold (p < .05) respectively, after LIF stimulation (Fig. 1B, left panel). Activation of pim-1 and pim-3 after stimulation by LIF was also observed in E14/T ES cells (pim-1: 2.4 ± 0.2-fold, p < .001; pim-3: 1.35 ± 0.05-fold, p < .01) (Fig. 1B, right panel). The level of pim-1 expression also increased at the protein level in response to LIF stimulation (2.1 ± 0.4-fold, p < .01). In contrast, the levels of Cdc25A phosphatase, a substrate of Pim-1 , remained unchanged (Fig. 1C). Thus, the levels of pim-1 and pim-3 mRNA and the level of Pim-1 kinase were upregulated as a consequence of LIF receptor stimulation in ES cells.
Contribution of STAT3 to the Transcriptional Regulation of pim-1 and pim-3
We made use of engineered ES cell lines stably expressing GRgp278, GRgp(Y118F), and GRgp(Y126-275F) receptors [3, 5] to study the role of STAT3 in the regulation of pim-1 and pim-3 expression. GRgp278 is a chimeric receptor consisting of the extracellular domain of the G-CSF receptor fused to the transmembrane and cytoplasmic region of gp130. Homodimerization of the GRgp278 receptor, mediated by G-CSF, induces recruitments of both STAT3 and SHP-2 and promotes self-renewal. GRgp(Y118F) is a mutated chimeric receptor in which phenylalanine has been substituted for tyrosine 118 in the intracytoplasmic domain, which prevents the recruitment of SHP-2 by the activated receptor. In the GRgp(Y126-275F) receptor, phenylalanines have been substituted for tyrosine 126 and 275, which strongly reduces recruitment of STAT3 by the activated receptor (supplemental online data 2). ES cells expressing wild-type, SHP-2 binding-deficient, and STAT3 binding-deficient chimeric receptors were LIF-deprived for 24 hours and restimulated for 1 hour either with 10,000 U/ml LIF, to activate the endogenous gp130 receptor, or with 30 ng/ml G-CSF, to activate the chimeric receptor. Stimulation of GRgp278 (wild-type) with G-CSF resulted in increased levels of pim-1 (1.6 ± 0.2-fold, p < .01) and pim-3 (2.8 ± 0.4-fold, p < .001) mRNAs (Fig. 2A, 2C). Similarly, stimulation of the GRgp(Y118F) cells with G-CSF resulted in increased levels of pim-1 (1.5 ± 0.25-fold, p < .05) and pim-3 (3.35 ± 0.45-fold, p < .001) mRNAs, indicating that recruitment of SHP-2 to gp130 receptor is not required for the transcriptional activation of pim-1 and pim-3 genes. In the STAT3 binding-deficient cell line (GRgp(Y126-275F)), LIF and G-CSF were equally efficient at activating the expression of pim-1, indicating that recruitment of STAT3 to the activated gp130 receptor is not essential to the transcriptional activation of pim-1. Examination of protein levels confirmed the activation of Pim-1 expression in response to LIF and G-CSF stimulation in the cell lines expressing the wild-type, SHP-2-deficient, and STAT3 binding-deficient receptors (Fig. 2B). By contrast, G-CSF failed to activate the expression of pim-3 in GRgp(Y126-275F), indicating that activation of STAT3 is required for the transcriptional activation of pim-3 (Fig. 2C).
pim-1 is a target gene of STAT3 that contributes to interleukin-6-dependent regulation of proliferation in lymphoid cells . This prompted us to examine further whether STAT3 activity could regulate pim-1 expression in the absence of gp130 receptor activation. To this end, a conditional mutant of STAT3 was generated by fusing the STAT3 coding sequence to ERT2, a mutated ligand-binding domain of the human estrogen receptor [17, 25]. The STAT3-ERT2 fusion protein was overexpressed by means of supertransfection of the pPCAGIZ-STAT3-ERT2 episomal expression vector into the E14/T-cell line in the presence of 4′OHT . pPCAGIZ-STAT3-ERT2 transfectants formed a uniform population of undifferentiated ES cells that was undistinguishable from the sister population of pPCAGIZ-STAT3-ERT2 transfectants propagated in the presence of LIF (data not shown). To assess the regulation of pim-1 and pim-3 by STAT3 activity, STAT3-ERT2 transfectants were deprived of 4′OHT for 24 hours and then restimulated with 100 nM 4′OHT for 2 hours (Fig. 3A). This resulted in increased levels of pim-1 (1.5 ± 0.1-fold, p < .01) and pim-3 (2.0 ± 0.1-fold, p < .001) mRNAs. Stimulation by 4′OHT in the presence of cycloheximide led to a 1.7-fold increase in mRNA levels, indicating that 4′OHT-induced transcriptional activation of pim-1 and pim-3 was not dependent on de novo protein synthesis. The observed effect is STAT3-specific since E14/T ES cells expressing a hormone-dependent Cre-ERT2 showed no increase in pim-1 and pim-3 expression following stimulation with 4′OHT. The increase of pim-1 mRNA level observed in response to 4′OHT was paralleled by a 2.5-fold increase in protein level (Fig. 3B). Together, these data indicate that the expression of pim-1 and pim-3 genes is regulated by STAT3 activity in ES cells.
Contribution of Pim-1 and Pim-3 Kinases to the Inhibition of ES Cell Differentiation
The role of Pim kinases in the inhibition of ES cell differentiation was examined by RNA interference, making use of lentiviral vectors encoding small hairpin RNA (shRNA) for pim-1 (pLenti6/BLOCK-iT-PGKneo-pim1), pim-3 (pLenti6/BLOCK-iT-PGKneo-pim3), and GFP (pLenti6/BLOCK-iT-PGKneo-GFP) and the neor selectable gene. Forty-eight hours after infection, CGR8 ES cells were replated at clonal density and further cultured in selection medium for 7 days to kill noninfected cells (approximately 50% of the population). The percentages of undifferentiated (alkaline phosphatase+ [AP+]), mixed (AP+/AP−), and differentiated (AP−) G418-resistant colonies were calculated (Fig. 4A). CGR8 ES cells expressing the shGFP maintained undifferentiated AP+ colonies at between 71% and 76% in this assay. Expression of shPim-1 or shPim-3 reduced, albeit moderately, the proportion of undifferentiated colonies compared with control (shPim-1, −20%, p < .01; shPim-3, −27.9%, p < .001). The decrease in percentage of undifferentiated AP+ colonies observed in ES cells expressing shPim-1 and shPim-3 was paralleled by an increase in the percentage of mixed AP+/AP− colonies (shPim-1, +17.3%, p < .01; shPim-3, +22.3%, p < .001) as well as of AP−, flat, differentiated colonies (shPim-1, +2.6%, p < .05; shPim-3, +5.5%, p < .01). Expression of rex-1, a marker of pluripotency , was analyzed by quantitative RT-PCR and showed 25% and 49% of reduction in transcript levels in cells expressing shPim-1 and shPim-3, respectively, compared with cells expressing shGFP (Fig. 4B). Cells expressing shSTAT3, a shRNA directed to STAT3, showed a 90% reduction in rex-1 transcript levels.
Gene expression analysis by real-time PCR showed that the levels of pim-1 and pim-3 transcripts were reduced only twofold in the infected cell population (data not shown), possibly explaining the relatively small differences observed in the proportions of undifferentiated versus mixed/differentiated colonies in the self-renewal assay. CGR8 cells expressing shGFP, shPim-1, and shPim-3 were subcloned, and the level of expression of pim-1 and pim-3 was analyzed to select clones showing the highest interference (Fig. 4C). Clones 1, 2, and 7 (shPim-1) and clones 1, 3, and 5 (shPim-3)—all six showing a residual expression of pim-1 and pim-3 between 20% and 38% with respect to levels measured in the shGFP control—were selected for further analysis. All shPim-1 and shPim-3 clones formed 3.3–16-fold fewer AP+ undifferentiated colonies and 3.2–4-fold more AP− differentiated colonies than the shGFP clone in the self-renewal assay (Fig. 4D). The stronger the interference of pim-1 and pim-3 expression is, the lower the percentage of undifferentiated colonies. Undifferentiated, mixed, and differentiated colonies were pooled and subsequently analyzed for the expression of pluripotency markers. All shPim-1 and shPim-3 clones displayed reduced levels of transcripts encoding Oct-4 (50%–66% reduction) and Nanog (44%–73% reduction) (Fig. 4E). They also displayed lower percentages of cells expressing the SSEA-1 antigen, a marker of pluripotent stem cells (shPim-1, 32%–50%; shPim-3, 25%–40%) compared with control cells (shGFP, 80%, p < .001) (Fig. 4F). In routine culture, all shPim-1 and shPim-3 clones displayed higher rates of spontaneous differentiation compared with the shGFP control. Interestingly, interference in the shPim-1 and shPim-3 clones was progressively lost during serial passagings, indicating that ES cells with low Pim-1 and Pim-3 content are counterselected (data not shown). These results demonstrate that expression of Pim-1 and Pim-3 kinases is required to inhibit differentiation of ES cells.
Overexpression of Pim-1 and Pim-3 Kinases Delays LIF-Induced Differentiation
Given that Pim-1 and Pim-3 kinases are necessary for self-renewal of ES cells, we asked whether overexpression of Pim-1 and Pim-3 would allow them to overcome spontaneous differentiation induced by LIF starvation. Pim-1 and Pim-3 kinases were overexpressed by means of supertransfection of pPHPGK-pim1 and pPHPGK-pim3 episomal expression vectors into the E14/T-cell line . Engineered ES cells were plated at clonal density and further cultured for 7 days in the presence of LIF. The percentages of undifferentiated, mixed, and differentiated colonies were calculated (Fig. 5A). ES cells overexpressing Pim-1 and Pim-3 formed significantly more undifferentiated colonies (Pim-1, +5.3%, p = .05; Pim-3, +17.9%, p < .001) and significantly fewer mixed colonies (Pim-1, −5.6%, p < .05; Pim-3, −17.8%, p < .001) than control cells. These data indicate that overexpression of Pim-1 and Pim-3 reduces the rate of spontaneous differentiation in the presence of LIF.
In a second step, Pim-1 and Pim-3 transfectants were subjected to a clonal assay combined with a LIF rescue test to evaluate the recovery of ES cells following a short period of LIF starvation . Cells were exposed to medium without LIF for 12, 24, 36, and 48 hours and were subsequently cultured for 5–7 days in medium supplemented with LIF. Then, the proportion of undifferentiated colonies was counted to assess the LIF rescue efficiency. In Pim-1 and Pim-3 transfectants, as in control cells, the proportion of undifferentiated colonies increasingly decreased with the duration of LIF deprivation, which indicates that overexpression of Pim-1 and Pim-3 is unable to fully sustain ES self-renewal in the absence of LIF (Fig. 5B). However, Pim-1 and Pim-3 transfectants showed a reduced sensitivity to LIF starvation. This was evidenced by a slower decrease in the proportion of undifferentiated colonies in ES cells overexpressing Pim-1 and Pim-3, compared with ES cells transfected with the empty vector (pim-1 vs. control, p < .001; pim-3 vs. control, p < .05). Together, these results are consistent with Pim-1 and Pim-3 overexpression partially counterbalancing the effects of LIF deprivation on induction of differentiation.
Contribution of Pim-1 and Pim-3 Kinases to the Regulation of ES Cell Proliferation and Apoptosis
In the LIF-rescue experiment reported in Figure 5B, we observed that the colonies generated from Pim-1 transfectants were larger than the control colonies (data not shown), suggesting that Pim-1 is also involved in the regulation of ES cell growth. We therefore asked whether pim-1 and pim-3 knockdowns are detrimental to ES cell growth, making use of the ES cells expressing pim-1-, pim-3-, and GFP-specific shRNA. Because knockdown of pim-1 and pim-3 expression increases the rate of spontaneous differentiation and ES cell differentiation is associated with increased cell-cycle duration , we thought to analyze the effect of pim-1 and pim-3 knockdown on the size of residual undifferentiated AP+ colonies in the clonal assay. We observed that AP+ colonies expressing the pim-1 and pim-3 shRNA were, respectively, 31% and 41% smaller than AP+ colonies expressing the GFP shRNA (p < .001) (Fig. 6A). Terminal deoxynucleotidyl transferase dUTP nick-end labeling analysis of the undifferentiated colonies showed 28% (pim-1) and 35% (pim-3) increases in the number of positive cells (p < .05), indicating that knockdown of pim-1 and pim-3 increases the rate of apoptosis (Fig. 6B, 6C). Together, these results are consistent with pim-1 and pim-3 contributing to the regulation of ES cell growth by inhibiting apoptosis.
First, we have shown that the expression of pim-1 and pim-3 can be activated by LIF in two different ES cell lines. We also demonstrated that pim-1 and pim-3 expression can be activated by hormone-dependent STAT3 in the presence of protein synthesis inhibitor. Together, these findings establish pim-1 and pim-3 as novel target genes of STAT3 in ES cells. The pim-1 promoter is known to contain a DNA sequence that potentially binds STAT oligomers . However, ES cells expressing the STAT3 binding-deficient G-CSF:gp130 chimeric receptor partially activate the expression of pim-1 in response to G-CSF. This observation suggests that in the absence of STAT3 recruitment, another pathway can activate pim-1. Candidate pathways are the LIF/PI3K and LIF/Yes pathways, both of which contribute to inhibiting ES cell differentiation [22, 28]. Furthermore, STAT3 is not the only factor regulating pim-1 and pim-3 expression in ES cells. It was observed that deprivation of LIF for 24 hours reduces the pim-3 RNA level to 50% of its original level, and pim-1 RNA level is not decreased by this treatment. This indicates that factors other than LIF and STAT3 contribute to the regulation of pim-1 and pim-3 expression. Loh et al. reported the mapping of Nanog-binding sites in ES cells . Mining their expression profiling data revealed that Nanog-binding sites are present in the promoter of pim-1. Furthermore, we observed that pim-1 and pim-3 expression is downregulated upon Nanog knockdown in CGR8 ES cells (unpublished data). Hence, the residual expression of pim-1 and pim-3 observed after withdrawal of LIF might result from a Nanog-dependent transcriptional activation mechanism.
Second, we demonstrated that overexpression of Pim-1 and Pim-3 decreases the rate of spontaneous differentiation in a clonal assay and enhances the resistance of ES cells to LIF starvation in a LIF rescue assay, whereas the knockdown of Pim-1 and Pim-3 increases the rate of spontaneous differentiation. These observations indicate that Pim-1 and Pim-3 kinases contribute to maintain ES cell pluripotency. Overexpression of Pim-1 and Pim-3 delayed the appearance of mixed and differentiated colonies in the LIF rescue assay but failed to block differentiation in the absence of LIF. This is also true of klf-4 and c-myc, both of which have been shown to contribute to the inhibition of ES cell differentiation [8, 9]. ES cells overexpressing Klf-4 demonstrated a trend toward reduced differentiation compared with wild-type cells but ultimately differentiated like their wild-type counterparts . Similarly, ES cells overexpressing a hormone-dependent Myc-ER and propagated in the presence of tamoxifen had the ability to self-renew in the absence of LIF, but this property was progressively lost with time in culture . Since none of these four genes are able to mimic the effect of LIF on ES cell self-renewal when they are individually overexpressed, it is likely that the four of them have to act cooperatively to block ES cell differentiation and maintain pluripotency. A synergistic action of Myc and Pim-1 would be particularly worthy of examination. The c-myc and pim-1 genes are activated at comparable levels in response to LIF or to 4′OHT in the STAT3-ERT2 ES cells (data not shown). Myc is a transcriptional regulator of the Cdc25A gene , and Pim-1 phosphorylates, and thereby activates, the phosphatase Cdc25A, which is a positive G1-specific cell cycle regulator [24, 31]. Such a synergistic action of Myc and Pim-1 has been evidenced in lymphoid cells, where it regulates interleukin-6/STAT3-dependent proliferation .
Third, we showed that Pim-1 and Pim-3 regulate the growth rate of the population of undifferentiated ES cells. The method used is based on the calculation of the size of undifferentiated colonies obtained in self-renewal assays after knockdown of Pim-1 and Pim-3. This method allowed us to measure the growth rate of undifferentiated ES cells and, therefore, to distinguish the alterations in growth rates resulting from alterations in the proliferation rate of undifferentiated cells and those indirectly resulting from differentiation . Inhibition of apoptosis is one mechanism by which Pim-1 and Pim-3 regulate the growth rate of undifferentiated ES cells. Pim kinases are known to play an important role in the prevention of cell death by inactivating the proapoptotic protein Bad [32, 33]. Antiapoptotic effects of Pim kinases have been demonstrated in several experimental systems [14, 16]. LIF has been shown to inhibit apoptosis in ES cells [6, 34]. Activation of Pim-1 and Pim-3 expression could thus be one mechanism by which LIF exerts its antiapoptotic effect.
To conclude, we provide evidence that the serine/threonine kinases Pim-1 and Pim-3 play an important role in the maintenance of the ES cell identity. This findings contribute to further deciphering the mechanisms by which the LIF/STAT3 pathway sustain self-renewal of mouse ES cells.
Disclosure of Potential Conflicts of Interest
The authors indicate no potential conflicts of interest.
We thank Dr. Austin Smith for the gift of pPCAGIZ, pPHCAG, and pPHPGK plasmids, as well as E14/T cells; Dr. Tom Burdon for the gift of ES cells expressing the G-CSF:gp130 chimeric receptors; and Dr. Anton Berns for the gift of mouse pim-1 and pim-3 cDNAs. This work was supported by research grants from Institut National de la Santé et de la Recherche Médicale (INSERM), Association Française contre la Myopathie and Vaincre La Mucovicidose (contract no. 4CS18H), INSERM AVENIR 2002 program, and the European Union sixth Framework Programme (FunGenES, contract no. LSHG-CT-2003-503494), Région Rhône-Alpes (Thématiques prioritaires 2003–2005), and Fondation Bettencourt-Schueller. C.S. and A.M. were recipients of a fellowship from the Association pour la Recherche contre le Cancer and the Ligue Nationale contre le Cancer. I.A., C.S., and P.-Y.B. contributed equally to this work.