Protein Kinase C Regulates Human Pluripotent Stem Cell Self-Renewal

Background The self-renewal of human pluripotent stem (hPS) cells including embryonic stem and induced pluripotent stem cells have been reported to be supported by various signal pathways. Among them, fibroblast growth factor-2 (FGF-2) appears indispensable to maintain self-renewal of hPS cells. However, downstream signaling of FGF-2 has not yet been clearly understood in hPS cells. Methodology/Principal Findings In this study, we screened a kinase inhibitor library using a high-throughput alkaline phosphatase (ALP) activity-based assay in a minimal growth factor-defined medium to understand FGF-2-related molecular mechanisms regulating self-renewal of hPS cells. We found that in the presence of FGF-2, an inhibitor of protein kinase C (PKC), GF109203X (GFX), increased ALP activity. GFX inhibited FGF-2-induced phosphorylation of glycogen synthase kinase-3β (GSK-3β), suggesting that FGF-2 induced PKC and then PKC inhibited the activity of GSK-3β. Addition of activin A increased phosphorylation of GSK-3β and extracellular signal-regulated kinase-1/2 (ERK-1/2) synergistically with FGF-2 whereas activin A alone did not. GFX negated differentiation of hPS cells induced by the PKC activator, phorbol 12-myristate 13-acetate whereas Gö6976, a selective inhibitor of PKCα, β, and γ isoforms could not counteract the effect of PMA. Intriguingly, functional gene analysis by RNA interference revealed that the phosphorylation of GSK-3β was reduced by siRNA of PKCδ, PKCε, and ζ, the phosphorylation of ERK-1/2 was reduced by siRNA of PKCε and ζ, and the phosphorylation of AKT was reduced by PKCε in hPS cells. Conclusions/Significance Our study suggested complicated cross-talk in hPS cells that FGF-2 induced the phosphorylation of phosphatidylinositol-3 kinase (PI3K)/AKT, mitogen-activated protein kinase/ERK-1/2 kinase (MEK), PKC/ERK-1/2 kinase, and PKC/GSK-3β. Addition of GFX with a MEK inhibitor, U0126, in the presence of FGF-2 and activin A provided a long-term stable undifferentiated state of hPS cells even though hPS cells were dissociated into single cells for passage. This study untangles the cross-talk between molecular mechanisms regulating self-renewal and differentiation of hPS cells.

Previously, we found that a proteoglycan, heparin promotes FGF-2 activity on the growth of undifferentiated hES cells in a minimal growth factor-defined culture medium, hESF9 [8], in which the effect of exogenous factors can be analyzed without the confounding influences of undefined components [8,[19][20][21][22][23] because insulin, transferrin, albumin conjugated with oleic acid, and FGF-2 (10 ng/ml) are the only protein components. Understanding cell signaling in undifferentiated hPS cells has lead to the development of optimal conditions for culturing hPS cells. However, manipulation of hPS cells still remains difficult because hPS cells as a single cell are unstable of self-renewal. Although Rho-associated kinase (ROCK) inhibitor (Y-27632) is quite effective to markedly diminish dissociation-induced apoptosis of single cells of hPS cells [24], the continuous use of the ROCK inhibitor increases differentiated cells [25]. For developing application using hPS cells, such as cell based therapy or toxicity screening tests, handling cell numbers would be beneficial. Even for basic research, handling cell numbers would be useful when the cells are dissociated for passages or differentiation. Presumably, if the culture conditions were able to fully support undifferentiated state, even single cells might maintain undifferentiated state. We suspected that there were unrevealed mechanisms to maintain undifferentiated state of single hPS cells. To further understand FGF-2 related molecular mechanisms regulating self-renewal would enhance understanding unclarified cell signaling in hPS cells. Therefore, we screened a kinase inhibitor library using a high-throughput alkaline phosphatase (ALP) activity-based assay in a minimal growth factor-defined culture medium, hESF9. We found that in the presence of FGF-2, an inhibitor of PKCs, GF109203X (GFX), increased ALP activity, suggesting that PKC reduces self-renewal of hPS cells. GFX inhibited FGF-2-induced GSK-3b phosphorylation. Addition of activin A increased phosphorylation of GSK-3b and ERK-1/2 synergistically with FGF-2 whereas activin A alone did not induce phosphorylation of GSK-3b. GFX negated differentiation of hPS cells induced by a PKC activator, phorbol 12-myristate 13-acetate (PMA) whereas Gö6976, a selective inhibitor of PKCa, b, and c isoforms did not counteract the effect of PMA. Functional gene analysis by RNA interference revealed that siRNA of PKCd, e, and f isoforms decreased phosphorylation of GSK-3b and also siRNA of PKCe and f isoforms decreased phosphorylation of ERK-1/2 in hPS cells. siRNA of PKCe decreased phosphorylation of AKT. On the basis of these results, we suggest that PKCd, e and f isoforms are FGF-2 downstream effectors, and they play various roles in regulating hPS cell self-renewal. This study helps to untangle the cross-talk between molecular mechanisms regulating self-renewal and differentiation of hPS cells.

PKC inhibitor increased ALP activity of hiPS cells
Previously, we detected the cell proliferative effect of heparin on hES cells without feeder cells in a minimal growth factor-defined culture medium, hESF9 [8], in which the effect of exogenous factors can be analyzed without the confounding influences of undefined components [8,[19][20][21][22][23]. In this culture condition using hESF9 medium (Table S1) on bovine fibronectin (FN), a highthroughput ALP activity-based assay was performed to evaluate a library of chemical kinase inhibitors to understand FGF-2 related molecular mechanisms regulating self-renewal of hPS cells. Nine compounds were found to increase ALP activity of the hiPS cell line 201B7 [26] (Fig. 1): Kenpaullone, which is a substitute for a reprogramming factor KLF-4 in mouse iPS cells [27]; Y-27632, which is a Rho-kinase (ROCK) inhibitor known to enhance hES cells survival [24]; HA-1004, H-89, and HA-1077, which are kinase inhibitors presumed to target ROCK [28]; GF109203X (GFX) [29], which is a inhibitor for PKC isoforms; and H-7, H-8, and H-9, which are also thought to target PKC [30]. These results suggest that FGF-2 induces PKC, and PKC acts downstream of FGF-2 to regulate self-renewal of hPS cells.
Addition of the PI3K inhibitor LY-294002 with FGF-2 completely inhibited AKT phosphorylation and significantly reduced GSK-3b phosphorylation (Fig. 2B, Fig. S1B). Addition of the MEK inhibitor U0126 with FGF-2 reduced ERK-1/2 phosphorylation and had little influence on GSK-3b phosphorylation. Addition of the GSK inhibitor BIO with FGF-2 signifi- Figure 1. An ALP activity-based high-throughput screening assay of chemical library for PKC inhibitors. The ALP activity using 4-methylumbelliferyl phosphate [59] in 201B7 hiPS cells in a 96well plate was measured by fluorometry. Each dot on the graph represents the fluorescent intensity for each compound of the kinase inhibitor library. Dotted line indicates the level for DMSO as a control. doi:10.1371/journal.pone.0054122.g001 cantly reduced phosphorylation of not only AKT, but also ERK-1/2 and GSK-3b.
Neither BMP-4 nor activin A in the absence of FGF-2 induced the phosphorylation of AKT, ERK-1/2, or GSK-3b in 201B7 iPS cells (Fig. 2C, Fig. S1C). From our previous report that activin A acts synergistically with FGF-2 in stimulating the phosphorylation of ERK-1/2 [20], we speculated that activin A may increase the phosphorylation of GSK-3b synergistically with FGF-2. Addition The cells were stimulated with FGF-2 (100 ng/ml) in fresh medium without insulin after overnight starvation. Fifteen minutes after FGF-2 addition together with each inhibitor as indicated on the panel. The data are represented as means 6 SE (n = 3). *P,0.05. (C) The cells were treated with FGF-2 (100 ng/ml), BMP-4 (100 ng/ml) or activin A (100 ng/ml) in fresh medium without insulin after overnight starvation. Fifteen minutes after the addition of each growth factor as indicated on the panel. The data are represented as means 6 SE (n = 3). *P,0.05. (D) The cells after growth factor starvation were stimulated with FGF-2 (10 ng/ml) and activin A (10 or 100 ng/ml) together with U0126 (5 mM) and GFX (5 mM) or Gö 6976 (5 mM) in fresh medium without insulin for 15 minutes. Fifteen minutes after the addition of each growth factor/inhibitor as indicated on the panel. The data are represented as means 6 SE (n = 3). *P,0.05. doi:10.1371/journal.pone.0054122.g002 of increasing concentrations of activin A with FGF-2 increased phosphorylation of both GSK-3b and ERK-1/2 in a dosedependent manner in H9 hES cells (Fig. 2D, Fig. S1D). Addition of U0126 with FGF-2 and activin A had little influence on phosphorylation of both AKT and GSK-3b, and completely inhibited phosphorylation of ERK-1/2. Addition of GFX together with U0126 in the presence of FGF-2 and activin A not significantly increased phosphorylation of AKT, while it completely inhibited phosphorylation of both ERK-1/2 and GSK-3b (Fig. 2D, Fig. S1D). A selective inhibitor of classical PKC (a, b, and c isoforms) [29], Gö6976 had little influence on phosphorylation of AKT and decreased phosphorylation of GSK-3b less than GFX. These results suggested that FGF-2-induced PKC stimulated phosphorylation of GSK-3b and that GFX inhibited the PKC-induced phosphorylation of GSK-3b, but it increased phosphorylation of AKT (Fig. S2).

Effect of GFX and PMA on colony morphology of the cells
To confirm the speculation that PKCs play roles in regulating self-renewal in hPS cells, the effect of the PKC activator PMA with several kinase inhibitors on the culture of 201B7 hiPS cells was determined (Fig. 3A). Treatment with PMA scattered the iPS cell colony dramatically. PMA-treatment with LY-294002, lithium chloride (LiCl, GSK inhibitor), Y-27632, or U0126 did not reverse the morphological change whereas GFX negated the effect of PMA on cultured 201B7 cells. Gö6976 did not negate the effect of PKC. The effect of Gö6976 was compared with that of GFX on ALP-activity of the cells: GFX with FGF-2 increased the ALPactivity of 201B7 iPS cells, while Gö6976 with FGF-2 had little effect on ALP-activity of the cells (Fig. 3B). GFX increased colony forming efficiency in hESF9 medium (Fig. 3C). Gö6976 did not increase the colony sizes of 201B7 cells and also cell numbers of H9 and 201B7 cells whereas GFX increased the colony sizes and also cell numbers (Fig. 3D, 3E, 3F). PMA activates PKCa, b, c, d, e, g, and h whereas GFX inhibits PKCa, b, c, d, e, and f isoforms. Gö6976 inhibits PKCa, b, and c isoforms. These results and findings suggested that PKCd or e isoforms regulate undifferentiated state of hPS cells.

Isoform-specific function of PKCs in FGF-2 signaling
To determine the isoform-specific function of PKCs on FGF-2 signaling, at first the expression of 11 PKC isoform genes in 201B7 iPS cells was determined by RT-PCR. The results showed that the cells expressed all of 11 PKC isoforms examined here (Fig. 4A). The PKC inhibitor results described above suggested that PKCd or PKCe might be responsible for GSK-3b phosphorylation but there is a possibility that PKCf might also be involved. Then, we examined whether FGF-2 stimulated phosphorylation of PKCd, PKCe or PKCf with or without GFX. Image analysis of western blotting showed that the phosphorylation of PKCd and PKCe was increased in a time-dependent manner after stimulation of FGF-2 and the phosphorylation of PKCf was increased in 15 min after stimulation of FGF-2 and then decreased, suggesting that activation mechanism of PKCf might be related with GSK-3b phosphorylation (Fig. 4B). GFX diminished the increased phosphorylation of all three PKCs. These result indicated that FGF-2 induced PKCd, PKCe, and PKCf in hPS cells.
We next examined the effects of short interfering RNA (siRNA) targeting PKCd, PKCe or PKCf on FGF-2 signaling in 201B7 iPS cells. The efficacy and specificity of siRNA was confirmed by quantitative RT-PCR (Fig. S3A). The expression of the targeted PKC genes was inhibited for at least 60%. The phosphorylation levels of AKT, ERK-1/2 and GSK-3b were measured in these PKCs-knockdown cells by AlphaScreenH SureFireH assay kit. The results showed that knockdown of PKCd, and PKCf did not affect FGF-2-induced AKT phosphorylation while knockdown of PKCe significantly reduced it (Fig. 4C). Knockdown of either PKCe or PKCf isoform significantly decreased FGF-2-induced ERK-1/2 phosphorylation. GFX which is reported to target PKCa, b, c, d, e and f isoforms did not change the level of FGF-2-induced ERK-1/ 2 phosphorylation, as shown above ( Fig. 2 and Fig. S1). These results implied that cross-interaction among PKC isoforms might affect on the level of FGF-2-induced ERK-1/2 phosphorylation. Then, the cells were treated with the inhibitory peptide cocktail for all isoforms (PKCa, b, c, d, e and f), or the inhibitory peptide cocktail for PKCd, e, and f. The inhibitory peptide cocktail for all isoforms did not affect on FGF-2-induced ERK-1/2 phosphorylation. On the other hand, the inhibitory peptide cocktail for PKCd, e, and f inhibited the ERK-1/2 phosphorylation (Fig. S4). These results suggested that inhibitions of all isoforms neutralized the reducing effect on FGF-2-induced ERK-1/2 phosphorylation by the inhibition of PKCe and f. GSK-3b phosphorylation was significantly reduced by the knockdown of all three PKC isoforms, compared with that by non-target siRNA. These results suggest that FGF-2 induced PKCs, followed by phosphorylation of ERK-1/2 and GSK-3b in hPS cells (Fig. S3B). From these results, we showed that FGF-2 induced PKCd, e, and f, resulting in stimulation of differentiation in hPS cells which might cause instability of the self-renewal state of hPS cells and that GFX targets these PKC isoforms in hPS cells, resulting in enhanced selfrenewal of hPS cells.

Stability of self-renewal of hPS cells in the presence of inhibitors of ERK-1/2 and PKC
Based on the results above, we hypothesized that inhibition of both PKC and ERK-1/2 might provide stable culture of hPS cells in our minimal defined medium hESF9 with activin A. Dissociated single hPS cells were inoculated on FN in hESF9 medium supplemented with activin A (10 ng/ml) [8,20], U0126 (5 mM) or GFX (5 mM). When dissociated single cells were cultured in hESF9, hESF9 + activin A, hESF9 + U0126, or hESF9 + activin A + U0126, many cells died or differentiated (Fig. 5A). On the other hand, when dissociated single cells were cultured in hESF9 + activin A + GFX, or hESF9 + activin A + GFX + U0126 (2i), cells could proliferate enough to be passaged. However, usually after 3 passages, epithelial-like cells appeared in the culture of hESF9 + activin A + GFX condition (Fig. S5A). Immunocytochemical analysis by image analyzer showed that ratio of OCT3/4-positive cell population in the culture of hESF9 + activin A + GFX + U0126 (2i) condition was slightly higher than that in the culture of hESF9 + activin A + GFX ( Fig. S5B and S5C). Gene expression in the cells cultured in these culture conditions was analyzed by realtime PCR (Fig. 5B). The expression of an endoderm marker, FOXA2, and a mesoderm marker, T were increased by activin A but it was significantly reduced by the addition of U0126. When the cells were cultured in hESF9 + activin A + U0126 + GFX, both FOXA2 and T were inhibited at lower level and also the undifferentiated makers, NANOG and OCT3/4 were maintained at higher ratio in the cells than those in other culture conditions. Next, the serial culture of dissociated single cells of hES H9, hES KhES4, hiPS 201B7 and hiPS Tic [33] cell lines were tested in hESF9 medium supplemented with activin A (10 ng/ml), U0126 (5 mM) and GFX (5 mM) (designated hESF9a 2i medium; Table  S1). Dissociated single hPS cells were grown on FN in hESF9a 2i medium for 3 passages. Phase-contrast image showed that cell morphology seemed undifferentiated although they did not form hPS typical cell colony. OCT3/4 expression profiles were confirmed by immunofluorescence analysis using image analyzer,  suggesting that the hPS cells maintained undifferentiated state. Another undifferentiated maker, TRA-1-60 expression was also confirmed in hPS cells grown in hESF9a 2i medium for 3 passages (Fig. S6).
The PKC family has been implicated as an intracellular mediator of several neurotransmitters, hormones, tumor promoters, a1-adrenergic agonists, and phorbol esters, and it is important in the regulation of growth, differentiation, cell death, and neurotransmission [38]. The PKC family comprises classical (PKCa, b, and c; activated by Ca 2+ and phorbol esters), novel PKC (PKCd, e, g, and h; activated by phorbol esters but not regulated by Ca 2+ ), and atypical PKC (PKCf and PKCi/l; not activated by Ca 2+ or phorbol esters). Different isoforms may perform distinct functions, as suggested by their differential pattern of localization, differences in condition of activation, and some differences in substrate specificity [39][40]. PKC has previously been implicated in GSK-3 regulation [41][42]. Fang et al. [43] showed that PKCa, bII, c, g, and d were capable of phosphorylating GSK-3b while PKCe and PKCf did not phosphorylate GSK-3 by in vitro kinase assays; also, expression of constitutively active PKCa, bI, c, g enhanced phosphorylation of cotransfected GSK-3b in HEK293 cells. On the other hand, Eng et al. [15] reported that negative construct of PKCe isoform prevented phosphorylation of GSK-3 in migrating fibroblasts._These pieces of evidence suggested that specific isoforms of PKC have different roles in different types of cells. Shuibing et al. [44] reported that activation of PKCa and/or b directs the pancreatic specification of hES cells. Recently, Feng et al. [45] reported that activation of PKCd induces extraembryonic endoderm differentiation of hES cells. These studies suggested that PKCs might be involved in differentiation of hPS cells. Our study showed that FGF-2 induced PKCd, e, and f, resulting in phosphorylation of GSK-3b, ERK-1/2, or AKT. Chou, et al. [46] reported that the phosphorylation of PKCf was regulated by PI3kinase and PDK-1 in NIH 3T3 fibroblasts. Intriguingly, PKCf can stimulate GSK-3 activity, by relieving PKB-imposed inhibition [47]. In mouse ES cells, it has been shown that PKCf plays an important role in inducing lineage commitment in mESCs through a PKCf-nuclear factor kappa-light-chain-enhancer of activated B cells signaling axis [48]. However, PKC inhibition does not change phosphorylation of ERK-1/2 or GSK-3b. In view of the fact that LIF mainly regulates self-renewal in mouse ES cells, isoform specific function might be cross-regulated by other signaling in the cells. Further, our study showed that the combination effect by inhibition of PKCa, b, c, d, e, and f was different from that by inhibition of PKCe and f, suggesting that each PKC might interact in different contexts and also PKCd, e, and f might have different activation mechanisms in hPS cells. It is needed further investigation in future.
GSK-3b is inhibited by phosphorylation stimulated by the canonical Wnt signal pathway, which is followed by the accumulation of b-catenin to the nucleus [49]. From the above findings, it follows that FGF-2 may activate Wnt signaling through PKC leading to differentiation of hPS cells. This conclusion contradicts the findings of previous studies demonstrating that canonical Wnt signaling supports self-renewal of stem cells [50][51][52]. However, it is consistent with a study showing that canonical Wnt signaling does not appear to promote stem cell maintenance, which prevents differentiation of stem cells [53]. On the other hand, some studies have shown a dual function for Wnt signaling in hES cells in that the pathways of self-renewal or differentiation are dependent on the presence of hES cell supporting factors [51][52]. Recently, Ding et al. [32] showed that FGF-2 modulates Wnt signaling through AKT/GSK-3b signaling and suggested that the differences in the results could be due to the culture platform. Our findings suggest that GSK-3b activity is regulated by FGF-2 through both PI3K/AKT and PKC pathways. AKT/GSK-3b signaling may support self-renewal whereas PKC/GSK-3b may promote cell differentiation of hPS cells. However, GFX decreased the phosphorylation level of GSK-3b to lower level than nontreatment. GSK-3b signaling might be stimulated also by other signal pathway in hPS cells. Target genes of these pathways and further regulation mechanisms in GSK-3b signaling should be analyzed in future.
TGF-b/activin/nodal pathways are thought to crosstalk with FGF signaling in regulating hPS cells. Vallier et al. [1][2]54] demonstrated that activin/nodal pathway in co-operation with FGF-2 is necessary for the maintenance of pluripotency in hES cells. We recently reported that activin A enhances FGF-2-induced ERK-1/2, which permits neural and mesendodermal differentiation of hES cells [20]. In this study we showed that activin A enhanced FGF-2-induced phosphorylation of not only ERK-1/2 but also GSK-3b. Inhibition of these pathways provided stable culture of hPS cells for long-term. In this study, we used both GFX and U0126 to inhibit these pathways. GFX targeting all of PKCa, b, c, d, e, and f had no inhibitory effect on ERK-1/2 pathway although siRNA targeting PKCe or PKCf decreased it. If more specific inhibitor is developed in future, it would be more useful.
analysis using an antibody detecting the phosphorylation or total protein amount of PKCd, PKCe, or PKCf. Protein content quantified from the gel blot images (n = 3). The values of the y-axis are the ratio of each phosphorylation to each total signal protein. (C) FGF-2 signaling in hPS cells with specific PKC isoforms-targeting siRNA. 201B7 iPS cells were transfected with specific PKCd, e, or f isoforms-targeting siRNA or non-targeting siRNA. The phosphorylation levels of the cells treated with FGF-2(100 ng/ml) after overnight starvation were measured by AlphaScreenH SureFireH assay kit. The values of the y-axis are the ratio of each phosphorylation to each total signal protein. The data are represented as means 6 SE (n = 3). *P,0.05. doi:10.1371/journal.pone.0054122.g004  Figure 5C were reseeded on a 6-well-plate and cultured for 5 days. The cells stained with anti-OCT3/4 antibody were visualized with Alexa Fluor 488 (upper panels). Nuclei were stained with Hoechst 33342 (blue). Scale bars, 200 mm. Whole cell images in whole plate were captured and OCT3/4 expression profiles were analyzed by Image Analyzer (lower panels). Antigen histogram (red); control histogram (green); Y axis is cell numbers and X axis is fluorescence intensity for anti-OCT3/4 antibody. doi:10.1371/journal.pone.0054122.g005 To maintain undifferentiated state, balancing among ERK-1/2, PI3K, SMAD, and PKC signal pathways may be required in any culture conditions. KSR of which components are not disclosed in public is known to have BMP-4-like activity [55]. Some components including BMP-4 in KSR together with secreting factors from mouse feeders might regulate PKC/ERK-1/2 signaling. Using our defined conditions, more molecules including growth factors would be screened to detect their accurate effects on hPS cells.
Human ES/iPS cell culture in feeder-free and growth factor defined serum-free medium Prior to culture in feeder-free conditions, the medium was changed from the KSR-based medium to a growth factor-defined serum-free hESF9 medium [8] (Table S1). Two days after the medium change, the cells were harvested with 1 mg/ml dispase or TrypLE (Invitrogen), and reseeded on plastic plates coated with bovine FN (Sigma, 2 mg/cm 2 ) [21]. For long-term culture, hPS cells were maintained on FN in hESF9 medium supplemented with 10 ng/ml human recombinant activin A (R&D Systems Minneapolis, MN, USA) in the presence of both 5 mM U0126 [20], and 5 mM GFX, designated hESF9a 2i medium. The medium was changed every day.

Single hPS cell culturing with two inhibitors
hPS cells were dissociated with TrypLE (Invitrogen) into single cells, and seeded on a 6-well plate coated with FN at the cell density of 1610 6 cells/well in hESF9, or supplemented with 10 ng/ml activin A, 5 mM U0126, or 5 mM GFX. The medium was changed every day.
Quantitative ALP activity-based high-throughput screening assay The hPS cells were dissociated with accutase into single cells and seeded at 5610 4 cells/well on a 96-well plate coated with FN (FN, 2 mg/cm 2 ) in hESF9 medium. Each compound in the chemical library was added at 2.5 mM to each well. After further 5 days-culture, the cells were washed with 3-[4-(2-Hydroxyethyl)-1piperazinyl] propanesulfonic acid (EPPA) buffer (30 mM, pH 8.2). Fluorescence ALP substrate (0.2 mM, 4-methylumbelliferyl phosphate) [59] in EPPS buffer was added into the wells. After incubation for 30 min at 37uC, EPPS buffer (100 mM, pH 7.7) supplemented with 1 M K 2 HPO 4 was add to terminate the enzyme reaction. The amount of 4-methylumbelliferone (4-MeU) produced via the enzyme reaction was measured with a fluorescence microplate reader (Gemini EM, Molecular Devices, Menlo Park, CA). The specific activity of ALP was quantified by reference to a standard fluorescence curve generated with known concentrations of 4-MeU (Sigma). Colony formation assay Dissociated single hPS cells were seeded at 10,000-250,000 cells/well on a 6-well plate coated with FN (2 mg/cm 2 ) in hESF9 medium supplemented with and without 1 mM GFX. After 5days-culture, the colonies were fixed in 4.5 mM citric acid, 2.25 mM sodium citrate, 3.0 mM sodium chloride, 65% methanol, and 3% formaldehyde for 5 min, and stained with ALP fastred substrate (Sigma) for 15 min at room temperature.

Immunocytochemistry
Immunocytochemistry was performed as described previously [20,60]. The image analysis was performed with In Cell analyzer 2000 and Developer tool box software (GE Healthcare, Little Chalfont, Buckinghamshire, UK), or a confocal microscope (Carl Zeiss). The primary and secondary antibodies used were listed in Table S2.

Western blotting
Western blots were performed as described previously [8,20,60]. Protein (2 mg/lane) was separated by 12.5% SDS-PAGE and transferred to polyvinylidene fluoride (PVDF) membranes (Millipore). The membranes were reacted with primary antibodies, peroxidase-conjugated secondary antibodies, and ECL Plus reagent (GE Healthcare). Protein bands were visualized using LAS-4000 imager (Fujifilm, Tokyo, Japan). The primary antibodies used were listed in Table S2.

AlphaScreen assay
AlphaScreenH SureFireH Cell-based Assay (Perkin-Elmer, Waltham, MA, USA) was performed to measure phosphorylation of AKT-1/2/3, ERK-1/2, and GSK-3b in the cells according to the manufacturer's instructions. Materials used were listed in Table  S2. The fluorescence signal was measured using an EnSpire TM plate reader (PerkinElmer).

Gene expression analysis
Total RNA extracted from cultured cells using RNeasy Mini kit (Qiagen, Valencia, CA, USA) were treated with DNase I to remove any genomic contamination, and reverse-transcribed using Superscript VILO cDNA synthesis kit (Invitrogen) according to the manufacturer's instructions. For RT-PCR, PCR products were amplified with AmpliTaq Gold DNA polymerase (Applied Biosystems, Foster City, CA, USA), following manufacturer's instruction. The DNA was separated by gel electrophoresis and visualized under ultraviolet light for photography. For quantitative real-time RT-PCR, PCR was performed based on the TaqMan or the SYBR Green gene expression technology in a 7300 Real Time PCR System (Applied Biosystems), following manufacturer's instruction. Threshold cycles were normalized to the housekeeping gene GAPDH and translated to relative values. Specific primers used are listed in Tables S3 and S4. For PCR-array, TaqMan lowdensity human stem cell pluripotency card PCR array (Applied Biosystems, Foster City, CA) was performed as previously described [61]. Expression levels were all normalized against the housekeeping gene b-actin. The relative expression levels of each gene in embryoid bodies were compared to the levels in H9 hES cells or 201B7 hiPS cells grown on feeders in KSR-based medium.

Transfections with siRNA
Transfections with siRNA were performed using Dharmafect1 (Dharmacon, Chicago, USA) as previously described [62]. Prior to transfection, the hiPS cells were incubated with ROCK inhibitor Y-27632 (10 mM) for 1 hour and dissociated with TrypLE (Invitrogen) and pelleted by centrifugation. To prepare siRNA/ lipid solutions, 50 pmol of siRNAs were diluted in 100 ml of hESF9 medium. In a separate tube, 6 ml of Dharmafect1 was diluted in 100 ml of hESF9 medium. The solution of the two tubes were mixed and incubated at room temperature for 20 mins. The resulting 200 ml of siRNA/lipid solution in hESF9 medium was used to resuspend the cell pelleted containing from 1610 4 to 1610 5 cells, and suspension incubated at room temperature for 10 min. After incubation, 1.5 ml of prewarmed hESF9 medium containing ROCK inhibitor (10 mM) was added and the suspension transferred into a FN-coated well of 24-well or 6-well plate, followed by culture for 24 hour. After recovery in fresh hESF9 medium, cells were transfected again at 24 hours. Total RNAs or proteins were extracted for analysis 72 hours after the fast transfection. siRNAs were listed as Table S4.

Live cell imaging analysis
After seeded on a 6-well plate coated with FN, the cells were incubated in a live cell imaging system, BioStation CT (Nikon Instruments Inc., Tokyo, Japan) at 37uC 10% CO 2 . The images were captured every 12 hours and analyzed by a soft ware CL-Quant (Nikon Instruments Inc.).

Cell Growth
The cells were inoculated on a 6-well plate coated with FN at the cell density of 250,000 cells/well in hESF9 medium including 10 ng/ml FGF-2, supplemented with 0.1% DMSO, GFX in H 2 O, or Gö6976 in DMSO. After 5 days culture, the cell numbers were counted by Coulter Counter (Beckman Coulter, Inc).

Flow cytometry
Flow cytometry was performed as described previously [61] with a FACS Canto flow cytometer (BD Biosciences). The primary antibodies used were listed in Table S2.

In vitro cell differentiation
In vitro differentiation was induced by the formation of embryoid bodies as described previously [61]. Floating embryoid bodies were maintained in DMEM with 10% FCS for more 14 days.

Teratoma formation
The cells were harvested by dispase treatment, collected into tubes, and centrifuged, and the pellets were suspended in DMEM supplemented ROCK inhibitor. The cells from a confluent onewell in 6-well plate were injected to the rear leg muscle or thigh muscle of a SCID (C.B-17/lcr-scid/scidJcl) mouse (CLEA japan, Tokyo, Japan). Nine weeks after injection, tumors were dissected, weighted, and fixed with 10% formaldehyde Neutral Buffer Solution (Nacalai tesque, Kyoto, Japan). Paraffin-embedded tissue was sliced and stained with hematoxylin and eosin. All animal experiments were conducted in accordance with the guidelines for animal experiments of the National Institute of Biomedical Innovation, Osaka, Japan.

Karyotype analysis
Log phase hPS cells (day 3-4 after subculture) were treated with metaphase arresting solution (Genial Genetic Solutions Ltd., Cheshire, UK) for 5 hr. The treated hPS cells were collected with 0.1% EDTA and processed according to the quality control protocol in the JCRB Cell Bank (http://cellbank.nibio.go.jp/ cellbank.html). Chromosome numbers were counted in 20 metaphases, and G-banding karyotype analysis was performed on 20 metaphase cells per sample. Figure S1 The phosphorylation of AKT, GSK-3b, and ERK-1/2 was confirmed by western blot analysis using an antibody to AKT, GSK-3b, and ERK-1/2 and their phosphorylated forms. Each gel image is a representative of independent three to five experiments. (A) Time course of phosphorylation level of AKT, GSK-3b, and ERK-1/2. H9 hES cells were stimulated with FGF-2 (100 ng/ml) with or without GFX (5 mM) for 180 minutes after overnight starvation of FGF-2 and insulin. (B) Effect of inhibitors on phosphorylation level of AKT, GSK-3b, and ERK-1/2. After starvation of FGF-2 and insulin overnight, 201B7 hiPS cells were stimulated with FGF-2 (100 ng/ml) for 15 min with LY294002, GFX, U0126, or BIO or without GFX (5 mM). (C) Effect of BMP-4 or activin A on phosphorylation level of AKT, GSK-3b, and ERK-1/2. After starvation of FGF-2 and insulin overnight, 201B7 hiPS cells were stimulated with with FGF-2 (100 ng/ml), BMP-4 (10 ng/ml) or activin A (100 ng/ml). (D) Effect of addition of activin A with and without inhibitors on phosphorylation level of AKT, GSK-3b, and ERK-1/2. After starvation of FGF-2 and insulin overnight, H9 hES cells were stimulated with FGF-2 (10 ng/ml) and activin A (10 or 100 ng/ml) together with U0126 (5 mM) and GFX (5 mM A + 2i (hESF9a 2i ) or hESF9 + activin A + GFX mediums at three passages, as described in Figure 5A and 5B. Scale bars, 200 mm. (B) Immunocytochemical staining for OCT3/4 expression of H9 cells cultured as described (A). The H9 hES cells stained with anti-OCT3/4 antibody were visualized with Alexa Fluor 488 (green). Nuclei were stained with Hoechst 33342 (blue). Scale bars, 50 mm. (C) Anti-OCT3/4 staining intensity profiles in the cell population grown in the hESF9 + activin A + 2i or the hESF9 + activin A + GFX conditions were analyzed by IN Cell image analyzer (lower panels). Antigen histogram (red); control histogram (green); Y axis is cell numbers and X axis is fluorescence intensity for anti-OCT3/4 antibody. (TIF) Figure S6 Immunocytochemical staining of H9, KhES-4, 201B7, and Tic hPS cells for TRA-1-60. The cells grown on FN in hESF9a 2i as described in Figure 5C were stained with TRA-1-60 antibody and Alexa Fluor 647-conjugated secondary antibody. Nuclei were stained with Hoechst 33342 (blue). Scale bars, 200 mm. . TaqMan low density PCR arrays (Applied BioSystems) were performed as previously described [61]. Expression levels were all normalized against b-ACTIN. The relative level of each gene expression were generated from the undifferentiated H9 hES cell or 201B7 hiPS cells cultured on mitomycin-inactivated mouse embryonic fibroblasts (MEF) in KSR-based medium (Sample No. 1-2). Heat-map colors (red for up-regulation, blue for down-regulation) depict gene expression. (C) Teratomas derived from H9 hES cells at passage 44 or 201B7 iPS cells at passage 26 maintained in hESF9a 2i conditions. (TIF)