The hESC line SA121 was transduced with ACTB-, CMV-, EF1α-, PGK- and UbC-eGFP self-inactivating lentiviral vectors (Fig. 1A). Efficiency of transductions was measured by flow cytometry (FACS) as percentage eGFP+ cells. To be able to quantitatively compare eGFP expression between the different promoters we aimed for similar copy numbers of integrated viral vectors. In addition, to avoid insertional mutagenesis, we transduced hESCs with low vector to target cell ratios. Initial test transduction experiments revealed that multiplicity of infection (MOI) 1 would generate transduction efficiencies up to 35% eGFP+cells for hESC line SA121 for the CMV-, EF1α and PGK-eGFP lentiviral vectors (Fig. S1A). Based on these experiments, hESC line SA121 was transduced at MOI 1 with ACTB-, EF1α-, CMV-, PGK and UbC-eGFP lentiviral vectors, which generated a maximum of 20% eGFP+ cells (Fig. 1B). Previous reports have showed that nonintegrated lentiviral vectors transiently expressed the transgene up to 10 days after transduction and thereafter gradually decreased as a result of dilution of the vector genome through cell divisions. [17], [18]. Therefore, transduction efficiencies were analyzed 10 days after transduction in order to avoid detection of transgene expression from nonintegrated vectors. qPCR on genomic DNA from FACS-isolated eGFP+ cells demonstrated that on average 1–2 viral vector copies per eGFP+ cell were integrated in ACTB-, CMV-, EF1α- and UbC-eGFP cells (Fig. 1C). PGK-eGFP transduced cells contained approximately 5 vector copies per eGFP+ cell. After 50 days of culture of eGFP+ isolated cells, copy numbers were detected at similar levels as immediately after FACS isolation with the exception of UbC promoters that decreased from average 2 vector copies to 1 copy per eGFP+ cell (Fig. 1C). Average gene copy numbers were measured by comparing the amount of eGFP and CDX2 amplified PCR products. Generation of a standard curve verified linear amplification of eGFP and CDX2 genomic DNA amplicons at similar efficiency (Fig. S1F).

To be able to quantitatively compare the activity of the different promoters, we isolated eGFP+ cells from the promoter-eGFP transduced populations by FACS sorting. Results for promoter-eGFP transduced populations are representative of three independent transductions that were FACS sorted separately and eGFP+ isolated cells were maintained as separate cell cultures in order to exclude possible variation in transductions or maintenance of cells. eGFP+ sorted cells exhibited characteristic hESC morphology and uniform expression of pluripotency markers OCT3/4, NANOG, and hES-Cellect (Cellartis AB) when cultured on Matrigel in mTESR cell culture medium, confirming that the transduction and FACS-sorting procedure did not affect hESC pluripotency (Fig. 2C–L). Promoter activity was measured 15, 30 and 50 days after FACS sorting and the sorting day was referred to as day 0. Percentage eGFP+ at day0 was approximately 98% for all promoters since reanalysis of sorted cells showed a purity of 97–100% eGFP+ cells (Fig. 2A and Table 1). The ACTB-, EF1α- and PGK promoters were found to be more efficient than CMV in driving long term expression of eGFP. In particular, the ACTB and PGK promoters mediated sustained eGFP expression (74,0±5,8% and 74,4±10,6%) after 50 days in culture (Table 1). Equal promoter activities were detected for ACTB, EF1α and PGK up to day 30 but thereafter EF1α activity decreased. The percentage of CMV-eGFP+ cells decreased to 6,7±2,9% at day 50. Already at day 7, the CMV promoter was rapidly downregulated and was expressed in approximately 30% of the cells (data not shown). Activity of the UbC promoter was observed in 62±4,5% of the total cell populations at day 15 and decreased to 24,8±10,4% after 50 days of culture (Fig. 2A and Table 1). At day 50, promoter activities for ACTB, EF1α and PGK were significantly higher than for the CMV promoter (p≤0.0001 students t'test).

To evaluate the strength of each promoter during maintenance of hESCs, the intensity of the eGFP fluorescent signal was compared at day 0 and day 50 (Fig. 2B). All promoters mediated eGFP expression with similar intensity on day0. After 50 days, the ACTB and PGK promoters maintained high eGFP expression levels, whereas expression levels dropped with CMV, EF1α and UbC and promoters.

To verify that the measured promoter activities were not specific to hESC line SA121, we repeated transduction experiments in hESC line Hues-4 [19], [20]. Transduction at MOI 1 resulted in low copy number integration and transduction efficiencies below 40% (Fig. S1A–C). The stability of the five constitutive promoters in Hues-4 was similar to that measured in SA121 (Fig. S1D). Thus, the ACTB-, EF1α- and PGK promoters maintained sustained activity up to day 30 (86,5±0,1, 80,0±10,1and 76,0±1,0% eGFP+ cells) whereas CMV promoter activity deceased strongly within 15 days (Table S1 and Fig. S1D). However, between day 30 and 50, the PGK promoter activity decreased more in Hues-4 than in SA121 (compare Fig. S1D with 2A). Like in SA121, activities at day 50 were significantly higher for ACTB-, EF1α- and PGK promoters than for CMV (p≤0.001 students t'test). In addition, intensity of the eGFP expression was comparable at day 0 for all promoters and showed similar pattern of stability as measured in SA121 (Fig. S1E). These observations reveal that the relative promoter stability and activity data are comparable between the two tested cell lines.

In summary, of all the tested constitutive promoters the EF1α, PGK and ACTB promoters were the most stable. The ACTB promoter was the most superior promoter in undifferentiated hESCs by maintaining transgene expression in 75–85% of the cells after 50 days in culture in both cell lines tested. In addition, the PGK promoter was found to express eGFP at high intensity up to 50 days of culture, whereas the intensities from the other promoters decreased to various extents.

To test the effectiveness of the promoters in differentiated hESCs, eGFP+ sorted cells were differentiated as embryoid bodies for 22 days. Differentiation into cell lineages of all three embryonic germ layers was verified by quantitative real time PCR (qPCR) showing an increase of gene expression levels of endodermal, mesodermal and ectodermal markers genes (Fig. 3A). In addition, the expression level of the pluripotency marker OCT3/4 decreased. FACS analysis performed at the end of differentiation demonstrated that promoter activities were less stable than in undifferentiated hESCs (Fig. 3B). EF1α was the most stable promoter during differentiation. Nevertheless, it was significantly downregulated during differentiation and was inactive in approximately 50% of the differentiated cells. In contrast to the stable PGK activity detected in undifferentiated cells, its activity was significantly downregulated during differentiation. The CMV promoter was active in only a small portion of the cells (approximately 15%) at start of the differentiation and was therefore not included in differentiation studies.

To evaluate the strength of each promoter during EB differentiation, intensity of eGFP fluorescent signal was measured by FACS analysis of eGFP+ cells. During EB differentiation, intensity of eGFP expression of ACTB- and EF1α was significantly reduced, whereas eGFP fluorescent signal within the PGK- and UbC-eGFP+ populations did not decrease during differentiation (Fig. 3C).

The observation that promoter-mediated transgene expression was downregulated during differentiation raised the question if promoter activities could be restricted to specific cell lineages. To address this, differentiated cells were separated into eGFP+ and eGFP-populations by FACS sorting. Gene expression analysis of eGFP+ and eGFP− populations were carried out for quantification of mRNA levels of marker genes representing the three embryonic germ layers. Notably, in cells where the EF1α promoter is active, mRNA expression levels of genes representing all three embryonic germ layers were similar to or higher than the levels in eGFP− cells (Fig. 4). In addition to these data, FACS analysis showed that the EF1α promoter was less prone to downregulation compared to other promoters during in vitro differentiation (Fig. 3B).

SOX17, which is expressed in the primitive streak and endoderm, is expressed at similar levels in EF1α -, PGK- and UbC-eGFP+ cells compared to eGFP− cells implying that these promoters are active during mesendoderm differentiation (Fig. 4A). In contrast, SOX17 expression was significantly reduced in ACTB-eGFP+ cells compared to ACTB-eGFP− cells. The hepatoblast marker ALBUMIN was expressed at significantly lower levels in the ACTB-, PGK- and UbC-eGFP+ populations compared to GFP−cells, suggesting that these promoters are inactive in hepatoblast cells (Fig. 4B).

In cells where the ACTB, EF1α, PGK or UbC promoters were active, higher or equal mRNA levels for neural marker genes PAX6 and NESTIN were detected compared to GFP− cells, indicating that all four promoters are active in neural progenitor cells (Fig. 4C–D). PPARγ is expressed in adipose tissues of mesodermal origin. Both the EF1α- and the PGK promoter exhibited similar mRNA expression levels of PPARγ in eGFP+ vs eGFP− cells, suggesting that these promoters are active in the PPARγ+ cells (Fig. 4E). However, ACTB-eGFP+ cell populations showed significantly lower levels of PPARγ expression than GFP− cells, indicating lower activity of ACTB promoter in PPARγ+ cells. Gene expression analysis of CD31, a marker for endothelial cells was significantly lower in ACTB and PGK-eGFP+ populations, suggesting that these promoters are inactive in endothelial cell types (Fig. 4F). In contrast, EF1α-eGFP+ cells showed similar levels of CD31 expression as EF1α -eGFP− cells, suggesting that the EF1α promoter efficiently drives eGFP expression in endothelial cells (Fig. 4F).

A number of different constitutive promoters have successfully been reported to maintain stable transgene expression in hESCs and are therefore good candidates in applications like cell lineage tracing, generation of fluorescent reporter cell lines and overexpression of transcription factors. Much effort has focused on techniques for stable integration of transgenes, but attempts to quantitatively compare the effectiveness of constitutive promoters to monitor and track cell fate determination in differentiating hESCs are limited. Here, we quantitatively compare the efficiency of ACTB, CMV, EF1α, PGK and UbC promoters to constitutively drive eGFP expression both in undifferentiated cells and in their differentiated progenies. To achieve this, we applied lentiviral gene delivery to ensure high transduction efficiency and stable transgene integration into the hESC genome. We reasoned that the lentiviral system was more favorable for stable integration of the transgene, than other approved transgene delivery methods in hESCs such as transfection or adenoviral transduction, which are more appropriate for transient expression [15], [16], [21], [22]. Moreover, HIV-1-derived lentiviral vectors are efficient tools for stable genetic modification of mammalian ES cells, since they are less prone to silencing than traditional retroviral vectors [12], [23]–[26]. Here, we provide relative data on promoter characteristics in hESCs using eGFP as reporter gene and thus results presented here are representative for constitutive promoter activities detected as eGFP expression. Nevertheless, it cannot be excluded that possible interference between the promoter and reporter gene might influence transgene expression and therefore future studies will have to resolve if different promoter/reporter combinations will result in other expression and stability profiles than those reported here.

The ACTB promoter was found to be the most stable promoter mediating stable transgene (eGFP) expression during long term culture (50 days) of undifferentiated hESCs. These observations were observed in two independent hESC lines. This promoter has previously been reported to stably express eGFP in both undifferentiated cells and derivatives of all three embryonic germ layers when inserted into to the Envy locus by transfection of bacterial plasmids[6]. Here, we used lentivirus as gene delivery tools for random insertion of the transgene and our data confirm that the ACTB promoter has the potential to generate sustained high level transgene expression in long term culture of undifferentiated hESCs.

The percentage of eGFP+ cells in the EF1α transduced populations decreased after 30 days in both hESC lines. Previous publications suggest that the EF1α promoter acts as a strong and stable promoter for transgene (eGFP) expression in hESCs [11]–[13]. Thus, the EF1α promoter has been used to generate stable eGFP expressing hESC lines with 95% of the cells maintaining eGFP expression up to four weeks [11]. This is in line with our data demonstrating stable eGFP expression up to 30 days in culture. In another study, hESCs transduced with low viral vectors copy numbers showed sustained high EF1α promoter activity for up to 60 days in culture [13]. After 30 days the EF1α promoter activity declined, albeit to a lower degree compared to in our study. The observed difference in EF1α promoter activity during long term culture of undifferentiated hESCs may be explained by lentiviral vector design or differences related to hESC culture techniques.

The percentage of eGFP+ cells in the PGK transduced populations decreased after 30 days in one of the two hESC lines studied. The PGK promoter has not been extensively studied in hESCs and quantitative information about its activity in relation to other constitutive promoters is lacking. Therefore, further studies are needed to elucidate long term activity of the PGK promoter in undifferentitated hESCs and possible variation of PGK promoter activity between hESC lines.

In both cell lines, the UbC and CMV promoters experienced a pronounced downregulation after 50 days in culture. This is consistent with other reports demonstrating substantial loss of UbC driven transgene expression in hESCs [9]. The rapid downregulation of CMV promoter activity during long term culture of undifferentiated hESCs is consistent with recent findings demonstrating that the CMV promoter is not stably expressed in undifferentiated mammalian ES cells [11], [27]. Furthermore, difficulties in obtaining stable CMV-eGFP expressing hESC lines support the inability of the CMV promoter to sustain stable and efficient transcriptional activity in undifferentiated hESCs [9], [11]. Thus, of the analyzed promoters the UbC and CMV are the least stable promoters during long term culture of undifferentiated hESCs (Fig. 2A and S1).

In the present study, analysis of transgene expression is performed on pools of transfected cells rather than isolated subclones. Therefore, it could be hypothesized that loss of transgene expression is caused by selective growth or survival of subclones with low expression levels of the transgene. However, we suggest that any negative effect of high transgene expression on hESC growth and survival would manifest itself equally in all promoter-eGFP transduced hESC cultures, rather than acting only on certain promoter-eGFP transduced cultures but not on others. Moreover, vector copy numbers remained constant throughout the 50 days culture period for all promoters except for the UbC promoter (Fig. 1C). Therefore, we propose that the variations in promoter stability reported here are mainly caused by promoter-dependent variations in transgene expression rather than variations in copy number and integration site in individual cells.

To successfully use reporter cell lines to follow and track cells it is important not only to ensure stable activity of the promoter but also to rely on strong promoter activity. The latter is necessary to ensure detectable levels of reporter gene expression. Thus, the fluorescent signal detected from ACTB-, CMV-, EF1α-, PGK- and UbC-eGFP expressing cells were used to assess strength of the different promoters. At day 0, eGFP levels were expressed at similar intensity for all five promoters. Notably, the ACTB and PGK promoters expressed eGFP at stable intensity levels during the 50-day observation period in both analyzed hESC lines (Fig. 2B and S1E), suggesting that these promoters ensures stable levels of transgene expression. In addition, the ACTB promoter stayed active in the majority of transduced cells over time, indicating that this is a strong and stable promoter in undifferentiated hESCs.

Ectopic expression of key regulatory genes is an important tool to study mechanisms of differentiation and to induce cell fate specification in hESCs. This emphasizes the need to identify constitutive promoters that remains active at high and stable levels not only in undifferentiated hESCs but also in their differentiated progenies.

Here, we provide quantitative data demonstrating that the activities of commonly used constitutive promoters decrease during hESC differentiation to various extend, whereby the EF1α promoter showed highest stability. Few attempts have been made to quantify constitutive promoter activity during EB differentiation of hESCs but it has been reported that in EF1α-eGFP transduced hESCs, the number of eGFP+ cells decreased from 84% to 78% eGFP+ cells during 4 weeks of EB differentiation [12]. During EB differentiation of mESCs, EF1α was shown as a superior promoter compared to the PGK promoter [8]. Moreover, during EB differentiation of EF1α-eGFP transduced mESCs, eGFP expression remained stable as observed by fluorescence microscopy, although quantitative qPCR analysis showed that eGFP mRNA levels decreased by approximately 40% [27].

HIV-1 based lentiviral vectors are known to be more efficiently expressed than their MLV gammaretrovirus counterparts that are often transcriptionally silent in both ES cells and in transgenic animals [26], [28]. Silencing have mainly been studied for retroviral vectors and encompasses several related phenomena including complete transcriptional silencing, which is observed shortly after infection, and variegation. The latter refers to the situation when genetically identical sister cells that inherit the same provirus either express or silence the provirus [29]–[31]. Finally, extinction refers to the progressive silencing of an initially expressed provirus during long-term culture or EB differentiation of mESCs [23], [30]. Little is known about silencing of SIN lentiviral vectors in ES cells but it has been shown that transgenes are efficiently expressed at multiple copy integrations but single copy integrations results in inconsistent expression [12], [13], [23], [28]. Detailed examination of mESC clones with single copy SIN lentivirus integrations suggests that lentiviral vectors are silenced by similar epigenetic modifications as their retroviral counterparts [32]. However, since we observed promoter-specific differences in the degree of eGFP inactivation, we conclude that the stability of lentiviral-mediated transgene expression in differentiating hESCs is at least partly dependent on the applied promoter.

Gene expression analysis of eGFP+ and eGFP− separated cells revealed that marker genes characteristic for the three germ layers were expressed in equal levels in EF1α-eGFP+ and eGFP−cells (Fig. 4). Thus, we conclude that activity of the EF1α promoter does not show any preference to endodermal, mesodermal or ectodermal hESC derivatives. The EF1α, PGK and UbC promoters were active in cells differentiating towards ectodermal lineages and in SOX17+ early endoderm. Notably, the EF1α promoter remained active during later stages of differentiation, here marked as ALBUMIN+ late endoderm/hepatoblast cells and in PPARγ+ and CD31+ late mesoderm while the ACTB promoter was not active in these populations.

In summary, the ACTB, EF1α and PGK promoters were the most stable promoters in terms of maintaining transgene (eGFP) expression during long term culture of undifferentiated hESCs. Furthermore, the intensity of eGFP expression from the ACTB and PGK promoter were expressed at stable levels during long term culture, whereas the intensities of eGFP expression from the other promoters decreased to various extents. In addition, our data demonstrate that during hESC differentiation, expression of constitutive promoters may be restricted to specific cell lineages and careful selection of promoters is thus important to ensure high transgene expression in differentiated hESC progenies. Our data provides a guideline to choose a suitable promoter to obtain stable gene expression in undifferentiated hESCs and when in vitro differentiation to certain germ layers is desired.

The hESC lines SA121 (Cellartis AB), previously adapted to enzymatic dissociation, and Hues-4 (D.A Melton, Howard Hughes Medical Institute, Harvard Institute, Cambridge, MA) were cultured according to protocols at http://www.mcb.harvard.edu/melton/HUES/ as previously described on mitotically inactivated mouse embryonic feeder cells (Lund Transgenic Core Facility, Lund University, Sweden) [19], [20].

eGFP+ cells were transferred to feeder free culture conditions using Matrigel Matrix (BD Biosciences) in mTESR.1 cell culture medium (Stemcell Technologies) according to manufacturer's instructions, and passaged every fifth to sixth day at 1∶3 split ratio.

For spontaneous differentiation of embryoid bodies, cells were dissociated with 0,05% tryspin-EDTA (Gibco) and cultured as suspension cultures in Knockout-DMEM (Gibco) supplemented with 20% Knockout-serum replacement (Gibco), 1% Non-essential amino acids (Gibco), 1% Glutamax (Gibco), 0.1% beta-mercaptoethanol (Gibco), 1% penicillin-streptomycin (Invitrogen) for 22 days with medium change every third day.

Cells were karyotyped by standard G-banding at Divison of Clinical Genetics, Linkoping University and Lund University, Sweden. SA121 were found to be karyotypically normal and Hues-4 was normal in 60% of the analyzed cells.

The backbone of the lentiviral construct, pTRIP, has been previously described [33]. The vector, pTRIP ΔU3.CMV-eGFP and pTRIP ΔU3.PGK-eGFP expresses the eGFP gene under the control of an internal cytomegalovirus (CMV) promoter and mouse phophoglycerate kinase promoter respectively kindly provided Alexis Pierre Bemelmans [34]. New lentiviral vectors pTRIP ΔU3.ACTB-eGFP, pTRIPΔU3.EF1α-eGFP and pTRIPΔU3.UbC-eGFP were constructed using the Gateway in vitro recombination system (Invitrogen). Briefly the RIP405 promoter was removed by MluI and BamHI restriction from pTRIP ΔU3.RIP405-eGFP [35]. Both extremities were filled by klenow polymerase and the RFA Gateway cassette was cloned to generate the pTRIP ΔU3.RFA (Gateway)-eGFP destination vector. All promoters were cloned by PCR into Gateway compatible Entry clones and finally inserted into the destination lentiviral vector by LR Clonase II recombination according to manufacturer's recommendations (Invitrogen). The elongation factor-1α (EF1α) promoter was amplified from vector pLOX/EWgfp (kindly provided by Dr. S. Karlsson Dept of Molecular Medicine and Gene Therapy, Lund University, Sweden) with primers forward 5′ GGGGACAAGTTTGTACAAAAAAGCAGGCT′3 and reverse 5′GGGGACCACTTTGTACAAGAAAGCTGGGTACTTTGAACCACTGTCTGAGGCTT ′3. The resulting PCR product was recombined into pDONR201 (Invitrogen) to generate EF1α entry clone. The human beta actin promoter (ACTB) was amplified from ACTB plasmid (kindly provided by Dr. E.G Stanley, Monash Immunology and Stem Cell Laboratories, Monash University, Australia) with primers: forward 5′ CACCCTTTCTAGAACTAGACT 3′ and reverse 5′ GTTAACCTCGACGTGAGCTGC 3′ and the resulting PCR product was cloned into pENTR/D-Topo vector. Human ubiquitinC (UbC) promoter was amplified from human genomic DNA using the following primers: forward 5′ GCCTCCGCGCCGGGTTTTGGC 3′ and reverse 5′ TCCACAACAAGAACCGCGAC 3′ and cloned into the pENTR/D Topo vector (Invitrogen).

Lentiviral vector stocks were produced by transient transfection of 293T cells with the p8.91 encapsidation plasmid [36], the VSV glycoprotein-G-encoding pHCMV-G plasmid [37], and the lentiviral recombinant vector as previously described [38]. Supernatants were treated with DNAseI (Roche Diagnostic) prior to ultracentrifugation and the resulting pellet was resuspended in Phosphate Buffered Saline, separated into aliquots and frozen at −80°C until use. The transduction efficiency of each vector stock was determined by FACS analysis as previously described [35].

hESCs were transduced with vector particles harboring ACTB-, CMV-, EF1α-, PGK- and UbC-EGFP, respectively, at MOI 1 previously determined to generate low number of integrated viral vector copies. Briefly, 500000 hESCs were dissociated to single cells dissolved in 200 µl cell culture medium, incubated with virus at MOI 1 under gentle shaking for 1 h at 37°C, thereafter seeded on MEF cells in 2 ml cell culture medium as described above. Medium was changed the next day and cells were cultured to confluence. eGFP expressing cells were isolated using FACS sorting and cultured on Matrigel matrix (BD biosciences) in mTESR.1 medium (Stemcell Technologies). hESC line SA121 was transduced in three separate experiments and at different passages for eGFP expression in undifferentiated cells and EB differentiated cells.

To isolate eGFPexpressing cells, trypsin dissociated cells were filtered through Filcon filter 50M (BD bioscieneces) to remove aggregated cell clumps and were sorted on FACSVantageSE DiVAOption (BD Biosciences) equipped with DiVa 5.0.3 software. Analysis was performed in FlowJo (Tree Star). Cells were collected in cell culture medium. Reanalysis of sorted cells reproducibly showed a high purity (>98%).

To analyze eGFP expression, transduced cells were dissociated to single cells and measured on a FACSCalibur equipped with CellQUEST software (BD biosciences). A population of 5000 cells was analyzed and 7-aminoactinomycin-D (7AAD) (Sigma) were used to exclude dead cells.

For quantification of copy number of integrated viral vectors, genomic DNA (gDNA) was extracted (Sigma GenElute Genomic Mammalian DNA Mini prep kit) from eGFP+ FACS sorted cells cultured without MEF cells. Quantification of eGFP was compared to a single copy reference gene CDX2 using PCR cycling conditions 50°C for 2 min, 95°C for 2 min followed by 40 cycles, denaturation at 95°C for 15 sec, annealing at 60°C for 25 sec, and extension at 73°C for 30 sec. The correct PCR-products were confirmed by agarose gel electrophoresis (2% w/v) and melting curve analysis. CDX2 was amplified with forward (fwd) 5′AGAGGGACTCAAGGGAAAGG′3 and reverse (rev) 5′ GGTCTGGGAAGGGAAGAGAA′3 primers and eGFP with fwd primer 5′ CTTGTACAGCTCGTCCATGCCG′3 and rev primer 5′AACATCGAGGACGGCAGCGT′3.

Briefly, eGFP and CDX2 were amplified from plasmid and genomic DNA, respectively, followed by purification of PCR products. Serial dilution of eGFP and CDX2 PCR products, mixed in 1∶1 molar ratio, were used for generation of standard curve by qPCR. Linear amplification of eGFP and CDX2 verified that the 2^−ΔΔCT method could be used to compare quantified PCR product of eGFP cDNA to the reference gene CDX2 [39].

mRNA gene expression analysis including mRNA extraction, cDNA synthesis and qPCR amplification were performed on eGFP+ EB isolated by FACS sorting, as described in [40]. The following primers were used for amplification; SOX17 fwd 5′AAGGGCGAGTCCCGTATC′3 and rev 5′TTGTAGTTGGGGTGGTCCTG′3, ALBUMIN fwd 5′GCAAGGCTGACGATAAGGAG′3 and rev 5′ TGGCTTTACACCAACGAAAA′3, PPARγ fwd 5′GCTGGCCTCCTTGATGAATA′3 rev 5′TTGGGCTCCATAAAGTCACC′3, CD31 fwd 5′ CCTGTCTTTCAGCCTTCAGC′3 and rev 5′CGCCTGTGAAATACCAACCT ′3, PAX6 fwd 5′GAACAGACACAGCCCTCACA′3 and rev 5′ATCATAACTCCGCCCATTCA′3 and NESTIN fwd 5′ AGCGTTGGAACAGAGGTTG′3 and rev 5′GCTGAGGGAAGTCTTGGAG′3.Ct values were normalized to GAPDH amplified with fwd 5′ GTTCGACAGTCAGCCGCATC′3 and rev 5′GGAATTTGCCATGGGTGGA′3 and plotted as relative mRNA expression. qPCR measurements were performed on three biological replicates, PCR-amplified as three technical replicates and plotted as standard deviation of the mean (± s.d.).

Cells were washed once in PBS and fixed in 4% PFA for 15 min, washed three times in PBS, permeabilized in 0,25 TritonX-100 for 15 min and blocked in 5% skim milk (Sigma) in 0,1% Triton X-100 (block buffer) (BDH). Primary antibodies mouse-α-Oct 3/4 (1∶500) (SantaCruz), mouse-α-Nanog (1∶500) (Sigma) and mouse-α-hES-Cellect (1∶500) (Cellartis AB) were incubated in block buffer 4°C over night. As secondary antibody, Cy3 donkey-α-mouse (Jackson ImmunoResearch) was added in a 1∶1000 dilution in PBS for 2 h at room temperature. Cell nuclei were stained with DAPI (Sigma). Immunostained eGFP expressing cells were visualized with Nikon Eclipse TE 2000-U Axioplan 2 fluorescence microscope and AxioVision LE software (Zeiss).