Nature Biotechnology 25, 803 - 816 (2007)
Published online: 17 June 2007 | doi:10.1038/nbt1318

Characterization of human embryonic stem cell lines by the International Stem Cell Initiative

Oluseun Adewumi1, Behrouz Aflatoonian2, Lars Ahrlund-Richter3, Michal Amit4, Peter W Andrews5, Gemma Beighton6,7, Paul A Bello8, Nissim Benvenisty9, Lorraine S Berry1, Simon Bevan6, Barak Blum9, Justin Brooking6,10, Kevin G Chen11, Andre B H Choo12, Gary A Churchill13, Marie Corbel14, Ivan Damjanov15, Jon S Draper16, Petr Dvorak17, Katarina Emanuelsson18, Roland A Fleck1, Angela Ford5, Karin Gertow3,19, Marina Gertsenstein16, Paul J Gokhale5, Rebecca S Hamilton11, Ales Hampl17, Lyn E Healy1, Outi Hovatta20, Johan Hyllner18, Marta P Imreh21,22, Joseph Itskovitz-Eldor4, Jamie Jackson5, Jacqueline L Johnson8, Mark Jones5, Kehkooi Kee23, Benjamin L King13, Barbara B Knowles24, Majlinda Lako25, Franck Lebrin26,27, Barbara S Mallon11, Daisy Manning28,29, Yoav Mayshar9, Ronald D G Mckay11, Anna E Michalska30, Milla Mikkola31, Masha Mileikovsky16, Stephen L Minger32, Harry D Moore2, Christine L Mummery26, Andras Nagy16, Norio Nakatsuji33, Carmel M O'Brien8, Steve K W Oh12, Cia Olsson31, Timo Otonkoski31, Kye-Yoon Park11, Robert Passier26, Hema Patel1, Minal Patel34, Roger Pedersen14, Martin F Pera35,36, Marian S Piekarczyk28, Renee A Reijo Pera23, Benjamin E Reubinoff37, Allan J Robins38, Janet Rossant39, Peter Rugg-Gunn14,40, Thomas C Schulz38, Henrik Semb41, Eric S Sherrer38, Henrike Siemen23, Glyn N Stacey1, Miodrag Stojkovic25,42, Hirofumi Suemori33, Jin Szatkiewicz13, Tikva Turetsky37, Timo Tuuri31, Steineke van den Brink26, Kristina Vintersten16, Sanna Vuoristo31, Dorien Ward26, Thomas A Weaver6, Lesley A Young1 & Weidong Zhang13 for The International Stem Cell Initiative*

The International Stem Cell Initiative characterized 59 human embryonic stem cell lines from 17 laboratories worldwide. Despite diverse genotypes and different techniques used for derivation and maintenance, all lines exhibited similar expression patterns for several markers of human embryonic stem cells. They expressed the glycolipid antigens SSEA3 and SSEA4, the keratan sulfate antigens TRA-1-60, TRA-1-81, GCTM2 and GCT343, and the protein antigens CD9, Thy1 (also known as CD90), tissue-nonspecific alkaline phosphatase and class 1 HLA, as well as the strongly developmentally regulated genes NANOG, POU5F1 (formerly known as OCT4), TDGF1, DNMT3B, GABRB3 and GDF3. Nevertheless, the lines were not identical: differences in expression of several lineage markers were evident, and several imprinted genes showed generally similar allele-specific expression patterns, but some gene-dependent variation was observed. Also, some female lines expressed readily detectable levels of XIST whereas others did not. No significant contamination of the lines with mycoplasma, bacteria or cytopathic viruses was detected.

Since 1998, when the first human embryonic stem (hES) cell lines were reported, a plethora of genetically diverse cell lines have been derived from human blastocysts. Techniques for their derivation and subsequent cell culture have varied from laboratory to laboratory1. Combined with the intrinsic genetic variation in the human samples, any of these additional sources of variation could lead to the selection of hES cell lines with substantially different properties. The available data also suggest differences between human and mouse ES (mES) cells2, 3, 4. Despite the opportunity for genetic and environmental influences affecting the phenotype of hES cell lines isolated by different investigators, few lines have been subjected to in-depth analysis, and new isolates are often published without detailed characterization data. As a consequence it is uncertain whether the different isolates of hES cells are indeed very similar, or whether significant differences exist between the various lines. To address this issue, The International Stem Cell Initiative (ISCI)5 was established by the International Stem Cell Forum ( to carry out a comparative study of a large and diverse set of hES cell lines derived and maintained in different laboratories worldwide. The goals of this study, the results of which we now report, were to assess the similarities and differences in the expression of commonly used markers of hES cells among as many as possible hES cell isolates, and to identify a set of well-validated markers to establish hES cell identity.

Several approaches have been used to characterize hES cells, but the most widespread are analyses of cell surface–antigen phenotype, often by flow cytofluorimetry, and gene expression studies, commonly assessed by RT-PCR and, increasingly, by microarray analyses, for example6, 7, 8, 9, 10, 11. Many cell-surface antigens used to identify hES cells were first detected using antibodies prepared against preimplantation mouse embryos and/or against mouse or human embryonal carcinoma (hEC) cells3, 12. Likewise, several gene products that are functionally associated with maintenance of the undifferentiated ES cell state were first recognized in mES cells, most notably POU5F1 (refs. 13–15), NANOG (refs. 16,17) and SOX2 (ref. 18). Other developmentally regulated genes, such as TDGF19, and many identified in microarray studies have been postulated to play key functions in hES cells, but there is less evidence to support their role in maintenance of the undifferentiated state. Nevertheless, they may represent useful markers for identification of pluripotent stem cells. Several studies have compared patterns of gene expression in various hES cell lines, with some reporting similarities6, 7, 8, 20, 21 and others emphasizing differences10, 11, 22 between different hES cell lines. However, these studies used only small numbers of hES cell lines.

ES cells are derived from early embryos at a time when epigenetic reprogramming is occurring23, and it is unclear whether their epigenetic status is stable or subject to perturbation. Imprinted genes, whose parent-of-origin–dependent expression is epigenetically regulated, have variable allele-specific expression and methylation patterns in some mES cells24, 25. In contrast, studies of a limited number of hES cells suggest that genomic imprinting in hES cells is rather more stable26, 27, but the generality of this conclusion remains in question.

To provide a more comprehensive assessment of the hES phenotype, the present ISCI study analyzed 59 independently derived hES cell lines, from 17 laboratories in 11 countries, for expression of 17 cell-surface antigens and 93 genes, chosen as potential markers of the undifferentiated stem cells or their differentiated derivatives. In addition, the allele-specific expression status of ten imprinted genes was examined in those cell lines with distinct alleles. The microbiological status of the lines was also assessed, and the histology of a sample of xenograft teratomas was reviewed. Overall, the data show that all the independent hES cell lines studied exhibited a common expression pattern for a specific set of marker antigens and genes, despite their different genetic backgrounds and their derivation in different laboratories using different techniques. Although more subtle differences between the lines, either in phenotype or in behavior upon differentiation, cannot be excluded, this study provides no evidence for markedly distinct subsets of hES cell lines.



Cell-surface antigens

We examined the expression of a panel of cell-surface antigens widely used as markers of hES and hEC cells (Fig. 1a, Supplementary Tables 1 and 2 online). An example of the expression patterns in one line, HUES1, is provided in Supplementary Figure 1 online. These are: the globoseries glycolipid antigens SSEA3 (ref. 28) and SSEA4 (ref. 29); the keratan sulfate antigens TRA-1-60, TRA-1-81 (ref. 30), GCTM2 and GCTM343 (ref. 31); and a miscellaneous set of protein antigens comprising the two liver alkaline phosphatase antigens TRA-2-54 and TRA-2-49 (ref. 32), Thy1 (ref. 33), CD9 (ref. 34), and HLA class 1 antigens35, 36. Several antigens commonly expressed following hES differentiation33, SSEA1, A2B5, CD56 (NCAM), GD2 and GD3, were also analyzed. These all showed substantially lower expression than those commonly used as hES markers (Fig. 1a). In addition, expression of the pan-human antigen TRA-1-85 (Oka)37 was assessed.

Figure 1: Surface-antigen expression patterns of undifferentiated hES cells.

Figure 1 : Surface-antigen expression patterns of undifferentiated hES cells.

(a) Box and whisker plots illustrating the range of percent antigen positive cells in cultures of undifferentiated human ES cells. This analysis was carried out on data from those cell lines (44 cell lines from 12 laboratories), which met FACS quality control criteria defined using the reference standard hEC cell line, 2102Ep (Supplementary Table 2). The open circles indicate outlying data points. The antigens detected by each antibody are shown in parentheses: P3XAg8 (negative control), A2B5 (ganglioside GT3), B159 (NCAM), F15-14-1 (Thy-1), GCTM2 (GCTM2), GCTM343 (GCTM343), MC480 (SSEA1), MC631 (SSEA3), MC813-70 (SSEA4), TG30 (CD9), TRA-1-60 (TRA-1-60), TRA-1-81 (TRA-1-81), TRA-2-49 (liver alkaline phosphatase), TRA-2-54 (liver alkaline phosphatase), VIN2Pb22 (ganglioside GD2), VINIS56 (ganglioside GD3), W6/32 (HLA-A, B, C), TRA-1-85 (Oka). (b) Clustering of cell-surface antigen expression versus cell line. The heat map shows a two-way cluster analysis, using Euclidean distance measures, carried out on the data from those cell lines (44 cell lines from 12 laboratories), which met FACS quality-control criteria defined using the reference standard hEC cell line, 2102Ep, and for which two time points were available (Supplementary Table 2). Two cell lines (CA1 and CA2) were grown in both standard and nonstandard (alternate) conditions (indicated by an asterisk). For each antibody, the average percent cells scored positive (logit transformed) in two independent assays for each cell line is colored from green (low % positive) to red (high % positive); white indicates missing data. (c) Immunofluorescence detection of surface antigens in hES cells. Fixed colonies of hES cells were stained with antibodies TRA-1-60 (HES-2), TG30 detecting CD9 (HES-2), SSEA4 (H1), and GCTM-2 (HES-2). Nuclei were counterstained with DAPI (TRA-1-60 and GCTM2, blue; TG30 and SSEA4, red). Scale bar, 85 μm.

Full size image (97 KB)


The relationship between the different cell lines was examined by cluster analysis (Fig. 1b). Overall, a high proportion of cells in all the hES cell lines examined expressed cell-surface antigen markers previously associated with undifferentiated stem cells, whereas few cells expressed markers associated with differentiated derivatives. The former group of antigens includes SSEA3 and SSEA4, which appear to be expressed on a lower proportion of the hES cells than the keratan sulfate–associated antigens, a finding consistent with the notion that that SSEA3 and SSEA4 expression is more rapidly downregulated on differentiation than the keratan sulfate antigens33, 38. Only subtle quantitative differences were found between the various hES cell lines, most likely reflecting variation in culture conditions; for example, the cell lines CA1 and CA2, when assayed after culturing in two different media, appeared in separate clusters, but these were based upon only small quantitative changes in levels of expression and not on any significant qualitative differences. We could discern no significant clustering linked to passaging technique. One cell line, HES5, clustered separately from all the others, most likely the result of spontaneous differentiation in this culture; that is, a higher proportion of HES5 cells expressed the differentiation-marker antigens and a lower proportion, the stem cell antigens. Thus, the panel of antigens used in this study allows robust identification of undifferentiated hES cells.

The same antigens were also assayed by indirect immunofluorescence of fixed hES cell cultures (Fig. 1c). In general the results qualitatively matched the observations from flow cytometry, with most cells in a given culture expressing the antigens associated with the undifferentiated state, but with isolated cells or clusters of cells expressing antigens associated with differentiated derivatives. For example, antigens showing high levels of mean fluorescence intensity by flow cytofluorimetry (e.g., SSEA4, TRA-1-60, GCTM2 and CD9) were readily detected by in situ immunostaining, whereas detection of SSEA3 reactivity was more variable. Some antigens, such as SSEA4 and CD9, showed staining patterns typical of integral membrane components, whereas others, such as the keratan sulfate antigens recognized by GCTM2 and TRA-1-60, showed a granular pattern of reactivity around the cell and on the culture surface characteristic of proteins deposited in the pericellular matrix (Fig. 1c).

Gene expression

Quantitative RT-PCR was used to assay the expression of a series of genes in stock cultures of undifferentiated hES cells and in differentiated embryoid bodies derived from them (Supplementary Table 3 online). These genes had been previously suggested as markers of undifferentiated hES cells or of their differentiated derivatives. Among them, NANOG16, 17, POU5F1 (ref. 13) and TDGF39 are well known to be closely associated with the pluripotent state in both mES and hES cells, and to be strongly downregulated upon their differentiation. They can be regarded as archetypal pluripotent stem cell markers. The level of expression of these three genes was remarkably consistent across all the undifferentiated hES samples tested (Fig. 2a, red lines, and Supplementary Table 4 online). Slightly higher levels (that is, lower DeltaCt values), particularly of NANOG and TDGF, were evident in the two undifferentiated hEC cell lines, 2102Ep and NTERA2, included for comparison, probably a reflection of the greater tendency for spontaneous differentiation in the hES cultures (Fig. 2a, bar charts on right hand side). Rather more variation in the levels of these three genes was seen in the hES embryoid body samples (Fig. 2a, black lines), although their continued expression suggested the persistence of undifferentiated stem cells in these preparations. For reference, the differentiated NTERA2 cells induced by retinoic acid, and a human fibroblast line, expressed substantially lower levels of all three genes compared with the undifferentiated hEC cells. In undifferentiated hES and differentiated embryoid body samples analyzed separately (data not shown), and in combined datasets (Fig. 2b), the levels of expression of the three archetypal stem cell genes, NANOG, POU5F1 and TDGF, were strongly positively correlated whereas their expression was negatively correlated with the levels of expression of AFP and NEUROD1, markers commonly associated with differentiation to extraembryonic endoderm and neurectoderm, respectively. This is consistent with variable levels of expression of the pluripotency genes among different cell lines being a result of variable levels of spontaneous differentiation.

Figure 2: Expression profiles of the archetypal hES genes, NANOG, POU5F1(OCT4) and TDGF.

Figure 2 : Expression profiles of the archetypal hES genes, NANOG, POU5F1(OCT4) and TDGF.

(a) The expression of NANOG, POU5F1 and TDGF was estimated by quantitative RT-PCR and is shown as DeltaCt values (average of the two time points for each cell line), relative to β-actin, for both undifferentiated hES cells (red line) and differentiated embryoid bodies derived from them (black line). Cell lines for which only single samples were available were excluded from the analysis, though they, too, showed comparable levels of gene expression (see Supplementary Tables 3 and 4). The results for several lines grown under both 'standard' and 'alternate' (*) conditions are included. For comparison, the corresponding bar charts indicate gene expression in the undifferentiated hEC cell lines 2102Ep and NTERA2, as well as in differentiated NTERA2 cells induced by exposure to retinoic acid (NTERA2.RA) for 7 d, and in the human foreskin fibroblast cell line, HS27. Note that an increase in DeltaCt value of 1.0 denotes a twofold decrease in mRNA level. (b) Correlation of gene expression (DeltaCt values) for NANOG expression with that of POU5F1 and TDGF, genes expected to mark undifferentiated hES cells, and with that of AFP and NEUROD1, genes that mark differentiation into endodermal and neural cell types, respectively. Because the undifferentiated hES cultures and the embryoid bodies derived from them both appeared to contain variable proportions of undifferentiated cells and their differentiated derivatives, data from both sets were combined in each scatter plot to estimate the coefficient of correlation for the expression of each gene with NANOG. In this analysis, the data from the two time points were not averaged; undifferentiated and differentiated samples are indicated by blues crosses and green triangles, respectively. Note the strong positive correlation of NANOG with POU5F1 and TDGF, consistent with the downregulation of all three genes upon differentiation, and the negative correlation of AFP and NEUROD1 expression with that of NANOG, consistent with their up-regulation upon differentiation.

Full size image (91 KB)

We next tested pair-wise correlations in expression with other genes in the panel, using NANOG as a reference, and the combined data from 'undifferentiated' and 'differentiated' (embryoid body) samples (Fig. 3). This combination provides a dataset containing a wider range of stem-to-differentiated cell proportions than either dataset alone, and so a more powerful test to discriminate between genes that mark the undifferentiated and differentiated cells. In this analysis expression of five genes showed a high correlation with that of NANOG (r ≥ 0.75): TDGF, POU5F1, GABRB3, GDF3 and DNMT3B. All of these genes have been previously identified as expressed in undifferentiated hEC and hES cells8, and as showing developmental regulation upon hES differentiation40. We suggest that these six genes constitute a core set of markers that could be used to define undifferentiated hES cells. Among the 59 independent hES cell lines from the primary laboratories and the four duplicates from the secondary laboratories studied here, none failed to express these genes.

Figure 3: Pairwise correlation of expression of NANOG with that of all tested genes.

Figure 3 : Pairwise correlation of expression of NANOG with that of all tested genes.

The pairwise correlation coefficient of NANOG expression with the expression of each gene analyzed was calculated from the combined data for the undifferentiated hES samples and differentiated embryoid body samples, and plotted in descending order. In this analysis, the data from the two time points were not averaged. Expression of five genes, TDGF, POU5F, GABRB3, GDF3 and DNMT3B, showed very strong correlations with NANOG expression (r ≥ 0.75), whereas 14 additional genes, FGF4, GAL, LEFTB, IFITM1, NODAL, TERT, UTF1, FOXD3, EBAF, LIN28, GRB7, PODXL, CD9 and BRIX showed strong but lower correlation (0.75 > r ≥ 0.5).

Full size image (48 KB)

Expression of an additional 14 genes (FGF4, GAL, LEFTB, IFITM1, NODAL, TERT, UTF1, FOXD3, EBAF, LIN28, GRB7, PODXL, CD9, BRIX) showed a marked, though weaker correlation (r < 0.75; r ≥ 0.5) with that of NANOG. Each hES line studied here expressed each of these genes. Most of these have also been previously associated with undifferentiated hES cells (Supplementary Table 3), but the lower correlations with expression of NANOG most likely reflect their more widespread expression in derivatives differentiating from hES cells. A notable instance is SOX2, which is not included in this set of genes as its expression showed an even slightly lower correlation with that of NANOG (r = 0.47); SOX2 is thought to play a role in undifferentiated hES cells, and is closely linked in a regulatory loop with NANOG and POU5F1 (refs. 41,42), but it is also expressed in the neurectoderm lineage43.

Apart from the expression of a number of genes showing weak positive correlations with that of NANOG, the expression of several exhibited strong negative correlations, consistent with the view that they are not expressed by the undifferentiated cells and are good markers of differentiation. Indeed these genes had been selected, a priori, as candidates to identify the appearance of particular differentiated lineages. For example, genes whose expression showed a negative correlation with that of NANOG (r < −0.25) included CDX2 and CGB (encoding the beta chain of human chorionic gonadotropin, βHCG), associated with trophectoderm differentiation44; GATA6 (ref. 45) and AFP46, associated with extraembryonic endoderm; and PAX6 (ref. 47) and NEUROD1 (ref. 48), associated with the neural lineage.

We next examined whether a two-way cluster analysis of gene expression in the undifferentiated samples would reveal any significant groupings of genes or significant differences between groups of cell lines (Fig. 4 and Supplementary Table 4 online). Indeed, this analysis revealed a substantial cluster of genes expressed by essentially all the undifferentiated hES cell samples, and included those genes that were downregulated upon differentiation, identified in the above correlation analysis. It also included a set of genes that are commonly associated with differentiated cell lineages, such as FN1, ACTC, LAMA1, AFP, EOMES, GATA4, FOXA2, GATA6, SOX17, FLT1, SYP, COL2A1 and COL1A1. Whether these are expressed in the undifferentiated stem cells themselves or in a set of differentiated derivatives that are always present in association with the undifferentiated cells as a result of spontaneous differentiation cannot be ascertained from the present data. It is notable that many of these genes are thought to be expressed by extraembryonic endoderm, which might be an inevitable early derivative of ES cells to appear on differentiation. In this regard, CGB (HCG-β), GCM1 and CDX2, which are markers of trophectodermal cells, the other extraembryonic lineage produced by hES cells, were not inevitably present.

Figure 4: Two-way cluster analysis of hES cell lines with respect to gene expression.

Figure 4 : Two-way cluster analysis of hES cell lines with respect to gene expression.

Hierarchical clusters were generated for the cell lines and gene expression levels from the undifferentiated hES cell samples; when data for two time points for a given cell line were available the results were averaged (Supplementary Table 4). The levels of gene expression relative to β-actin are indicated by the color change from red (high expression levels; low DeltaCt) to green (low expression levels, high DeltaCt); white indicates missing data. The results for several lines grown under both 'standard' and 'alternate' (*) conditions, or grown in separate primary and secondary laboratories (+) are included. Note cell line CF-1 is now renamed KCL-003-CF-1.

Full size image (96 KB)

However, this cluster analysis provided no evidence of any discrete subsets of the hES cell lines in which the undifferentiated cells differed qualitatively with respect to the expression of the core stem cell genes identified above. Some clustering of lines was apparent, but inspection suggested that this was largely due to subtle quantitative differences in levels of gene expression and could be most easily explained by varying amounts of spontaneous differentiation occurring in the cultures from which RNA was isolated. This was most marked in a cluster comprising H14, H7, H13 and H9. These showed somewhat higher expression of genes associated with differentiation, most strikingly AFP, and lower levels of various stem cell associated genes, suggesting that the particular cultures analyzed contained a higher proportion of differentiated cells than the average across all the samples. When we examined cases where the same lines were grown in standard and alternate conditions (e.g., the two lines CA1 and CA2 grown with and without plasminate, and the HES1-5 lines grown in KSR or FCS), there were some quantitative differences in gene expression levels, but no consistent clustering that could be ascribed to variations in the media. More likely, these differences also reflected random fluctuations in the levels of spontaneous differentiation.

We finally examined a two-way cluster analysis of gene expression in the differentiated samples (Fig. 5). All expressed those genes identified above as characteristic of undifferentiated hES cells, indicating the continued presence of undifferentiated stem cells in these embryoid body preparations. In addition, there were significant variations in expression of a number of genes associated with distinct pathways of differentiation. For example, PAX6, a neural lineage marker, was expressed strongly in the embryoid bodies of some lines but not others. However, given the stochastic nature of differentiation in embryoid bodies and the difficulties of standardizing this method, these results cannot be taken to indicate intrinsic differences between the hES cell lines in their capacity for differentiation. More detailed studies would be required to test such a hypothesis.

Figure 5: Two-way cluster analysis of differentiated hES cell lines with respect to gene expression.

Figure 5 : Two-way cluster analysis of differentiated hES cell lines with respect to gene expression.

Hierarchical clusters were generated for the cell lines and gene expression levels from the differentiated hES cell samples; when data for two time points for a given cell line were available, the results were averaged (Supplementary Table 4). The levels of gene expression relative to beta-actin are indicated by the color change from red (high expression levels; low DeltaCt) to green (low expression levels, high DeltaCt); white indicates missing data. The results for several lines grown under both 'standard' and 'alternate' (*) conditions, or grown in separate primary and secondary laboratories (+) are included.

Full size image (95 KB)

Imprinting and X inactivation

To address the epigenetic status of hES cell lines, we assessed the allele-specific expression of ten genes subject to genomic imprinting (Table 1). Of the 59 independent hES cell lines submitted to the ISCI, 46 provided informative heterozygous polymorphisms within one or more of these imprinted genes. Occasionally, allele-specific imprinted gene expression patterns were different between samples of the same hES cell line, perhaps reflecting epigenetic changes between different time points. However, because the structure of the study did not permit detailed follow-up, it was not possible to exclude technical problems in individual cases. Accordingly, conclusions have been drawn from the data as a whole on the basis of samples analyzed rather than from individual cell lines (Table 1 and Supplementary Table 5 online).

Overall, monoallelic expression of imprinted genes was the most common result observed. Thus, of the six paternally expressed genes studied, four (IPW, SNRPN, KCNQ1OT1 and PEG3) showed monoallelic expression in all samples of all the cells analyzed. Of the four maternally expressed genes studied, two (H19 and NESP55) were predominantly monoallelically expressed, although in each case one sample from a single cell line was scored biallelic (Table 1). Of the remaining genes, the paternally expressed gene IGF2 and the maternally expressed gene MEG3 were mostly monoallelically expressed, although a substantial proportion of samples (~20–30%) were scored as partial- or bi-allelically expressed. On the other hand, the paternally expressed gene MEST and the maternally expressed gene SLC22A18 were predominantly bi-allelically expressed, albeit with a substantial proportion (>20%) of mono- and partial-allelic samples.

As a further assessment of epigenetic status, we looked for evidence of X chromosome inactivation in female cell lines. In both hES and mES cells, the absence of X inactivation in the undifferentiated cells, with subsequent inactivation upon differentiation, has been reported49, 50. However, this is not invariably the case, and instances of inactive X chromosomes in undifferentiated cells have also been reported40, 50, 51. We therefore tested the expression of XIST in all the hES samples submitted to the ISCI (Fig. 6). With the exception of the undifferentiated samples of one line, H13, all the karyotypically male hES lines expressed low or undetectable levels of XIST (DeltaCt > 13) (Fig. 6a). Within the context of the current screen, the individual significance of the one outlier cannot be assessed. By contrast, the female hES lines fell into two broad groups (Fig. 6b), one group in which XIST levels were low or undetectable (DeltaCt > 13), matching the male lines, and a second group in which XIST was readily detectable (DeltaCt < 10). These results indicate marked heterogeneity between the female hES cell lines with respect to X-chromosome inactivation, a conclusion reflected by the contrasting results previously reported40, 50, 51. The reasons for this variability might relate to the X-chromosome inactivation state of the particular embryonic cells from which individual hES lines were derived, but it may be related to either cellular differentiation within the hES cell culture50 or adaptation to culture conditions40.

Figure 6: XIST expression in male and female hES cell lines.

Figure 6 : XIST expression in male and female hES cell lines.

(a,b) The level of XIST expression (estimated as DeltaCt values relative to β-actin) is shown separately for those karyotypically male (a) and female (b) undifferentiated hES cell lines. When results from two time points were available, these have been averaged. Note that an increase in DeltaCt value of 1.0 denotes a twofold decrease in mRNA level. Note cell line CF-1 is now renamed KCL-003-CF-1. The results for several lines grown under both 'standard' and 'alternate' (*) conditions, or grown in separate primary and secondary laboratories (+) are included.

Full size image (61 KB)

Microbiology of feeder cells and hES cell lines

Sixteen human and mouse feeder cell samples were submitted from different laboratories, representing the feeder cells on which the hES cells were maintained. All tested negative for bacterial, fungal and mycoplasma contamination. Also, in cell line inoculation tests, with subsequent testing for hemadsorption of chick and guinea pig red blood cells, no evidence of viral infection was revealed, and these same cells showed no evidence of viral particles, inclusion bodies or aligned or unaligned viral nucleoprotein by transmission electron microscopy. Of the 59 independent hES cell lines studied, samples from 58 were submitted from the primary laboratories for microbiology, together with an additional four from the secondary laboratories. One sample was positive for fungal contamination; none were positive for bacteria or mycoplasma. In a minority of cases (eight cell lines; Supplementary Table 6) antibiotics had been used in growth media sampled for sterility tests. Consequently, in these few cases, we cannot exclude the presence of low-level contamination, which could have been masked by the presence of antibiotics. None of the hES lines showed any evidence of viral infection when assayed, as described above for the feeder cells.

Xenografted tumors

A systematic study of the conditions that affect the development of xenografted human ES cell–derived tumors was not planned within the present study. Nevertheless, we were able to obtain a review by a well-qualified clinical pathologist, with long experience of teratoma biology, of 37 histological slides from xenograft tumors produced from 15 independent hES lines in immunocompromised mice by the participating laboratories. In almost all cases the tumors were teratomas, benign tumors composed of somatic tissues derived from each of the three embryonic germ layers—ectoderm, mesoderm and endoderm (Supplementary Fig. 2). Ectodermal and mesodermal tissues predominated, most with the appearance of mature somatic tissues. However, neural tissue was often present in the form of immature embryonic neural rosettes. Mesodermal tissues were found in all grafts, and included fibroblasts and capillaries, smooth muscle, striated muscle, cartilage, bone and fat cells. Endodermal tissues included gland-like structures lined with columnar or cuboidal epithelium, which occasionally contained mucus-secreting goblet cells or ciliated cells, and even cells resembling intestinal Paneth cells. Three grafts contained foci of undifferentiated cells, which had the morphologic features of hEC cells. The nature of these EC-like cells could not be determined, but they might represent either residual, undifferentiated hES cells or hES cells that had undergone malignant transformation into true embryonal carcinomas. In one case, the investigator had explanted the tumor and observed outgrowth of karyotypically abnormal cells with a phenotype consistent with both hES and hEC cells52.



The current study is an initial attempt to provide a systematic and comparative survey of the many different human ES cell lines established and maintained by various groups around the world. Previously, some smaller-scale comparisons of lines maintained in single laboratories have been undertaken22. It is difficult to know exactly how many lines have so far been derived, although estimates range from 200 to >300 (refs. 53,54). The current panel of 59 independent lines provides a representative sample of those readily available worldwide. Overall, the microbiological assays pointed to the high standards of cell culture in the different laboratories and the absence of significant contamination from adventitious agents.

In such a survey, detailed studies of individual lines are not possible and, inevitably, obtaining complete sets of data for each submitted line is subject to the vagaries of experimental variation. Thus, fluorescence-activated cell sorting (FACS) or gene expression data were not obtained for all time points for every cell line. In particular, difficulties in standardizing FACS assays of surface-antigen expression in different laboratories became apparent as we analyzed the data collected, despite the provision of common batches of antibodies and protocols. This points to a need to develop standards for these tests, which will be addressed in the second ISCI that is currently being planned.

Certainly, there were differences between the hES cell lines included in the current ISCI study, for example, variable XIST expression in the female lines, and differences in the imprinting status of some genes. Nevertheless, a global analysis of all the data collected indicates a similar pattern of expression of a series of characteristic surface antigens and genes by hES cells that have been derived by different laboratories, using various protocols, from different gene pools of donated embryos. The similarities were evident despite the small variations in culture media and conditions used by some laboratories. All the lines examined expressed a comparable spectrum of genes and surface marker antigens characteristic of hES cells, suggesting that there is a common set of markers that can be used in general to monitor the presence of pluripotent stem cells. Contrary to a recent suggestion that SSEA3 and SSEA4 may not be acceptable markers for hES cells because they are not essential for human ES cell pluripotency55, the expression of these antigens by all tested hES cells makes them valuable operational markers of this cell type. Of course, rare variants may emerge as more lines are studied. For example, the known polymorphism for SSEA4 expression on red blood cells56 suggests that about 1% of hES cells derived from populations of European ancestry may prove to be SSEA4(−), although for most purposes SSEA4 is a good indicator for undifferentiated hES cells. Also, the results do not exclude more subtle variations among the cell lines, either with respect to quantitative levels of expression of specific antigens or genes, or in differentiation. One observation that might impinge on such potential variability is the evident potential for culture adaptation and genetic instability of hES cell lines during prolonged passage57, 58.

The results of the current study demonstrate the existence of both similarities and differences between hES and mES cells, as well as the closely related molecular properties of hES and hEC cells. Thus, the SSEA1(−)/SSEA3(+)/SSEA4(+) phenotype of hES cells versus the SSEA1(+)/SSEA3(−)/SSEA4(−) phenotype of mES cells is confirmed3, 12, 59. Also the presence of markers of trophectodermal differentiation in hES cell cultures agrees with the apparent increased ability of hES cells to adopt this lineage of differentiation in contrast to their murine counterparts60, although the failure to detect trophectodermal differentiation in the hES cell xenograft tumors that were examined is notable. On the other hand, of the group of genes whose expression correlates closely with that of NANOG and appears to be characteristic of the hES cells (NANOG, POU5F1, TDGF, GDF3 and GABRB3), all are also expressed by mES cells, with the possible exception of GABRB3 (refs. 13,16,17,61,62).

It was notable that several of the genes that showed a positive correlation with NANOG expression (EBAF, PODXL, NODAL, ZFP42, LIN28, EOMES and SFRP2) were identified in ChIP on CHIP experiments as potentially direct targets of NANOG42. Also, several other genes identified in that study as potential NANOG targets (ISL1, PAX6, CDX2) showed a negative correlation in the present study. Thus, combining the correlation data from gene expression experiments with target-binding assays for pluripotency-specific targets may provide insight into positive and negative control in the regulatory networks that affect pluripotency in hES cells.

The results provided additional insight into the epigenetic status of hES cells, for which relatively few lines have previously been studied26, 27, 63. Overall, the results suggest that genomic imprinting appears to be relatively invariant in hES cells, at least at the loci examined here. Nevertheless, imprinting stability, as indicated by the presence or absence of monoallelic expression, appears to be gene dependent, with some imprinted genes showing a consistent pattern in all hES cells and others being variable. In the case of one such variable gene, IGF2, contrasting results were obtained between samples of the same line, I3 (TE03), submitted by two separate laboratories (that is, it was consistently monoallelically expressed in samples from one laboratory and biallelically expressed in samples from the other). This result suggests that changes in imprinting patterns of at least the IGF2 gene and the nearby H19 gene26 can occur upon prolonged culture. One possibility is that overexpression of IGF2, a likely growth factor for hES cells, might provide variant cells with a selective advantage. However, any conclusions must be tempered by the recognition that detailed information about the allele-specific expression of imprinted genes in the early human embryo is mostly lacking, and extrapolation from mouse embryo data might be misleading. Of note in this connection is the variable genomic imprinting of SLC22A18 in humans64, 65. Derivation of hES cells could perpetuate any such variability in imprinting if it were present in the inner cell mass cells of the blastocyst, from which the hES lines were derived.

The studies of XIST expression by the hES cells examined provided a mixed picture. Variation in X inactivation in the female lines, inferred from XIST expression, matched variation evident from contrasting reports of specific lines. One factor pertinent to this variability is the apparent selective advantage offered by the presence of multiple active X chromosomes in hES cells, suggested both by the loss of X-chromosome inactivation40 and the appearance of multiple X-chromosomes in late-passage, culture-adapted hES lines58. Alternatively, it may be that in hES cells, like in mES cells, X inactivation occurs only after differentiation, and the differences between the various samples result from partial differentiation within the cultures.

The conclusions of the histology studies are based on a limited number of donated samples and must be considered tentative, but nevertheless represent independent and direct comparative analysis of tumors generated in different laboratories. Most xenografts produced from hES cells in immunocompromised mice were teratomas composed of fully differentiated somatic tissues. Strikingly, however, no extraembyonic tissue, such as trophoblast, was identified, although yolk sac–like tissue was tentatively identified around foci of hEC-like cells. Some of the somatic tissues, most notably the neural tubes, were immature, fetal-like, and proliferating, suggestive of a malignant potential. This and also the apparent persistence of undifferentiated hEC-like cells in a small number of xenografts raises the issue of the malignant potential of hES cells and highlights the need for detailed retransplantation experiments coupled with analysis of genomic stability to assess its potential significance.

Studies of the biology of hES cells have developed rapidly over the past eight years since the first reports of their derivation. They clearly offer enormous potential, not only for regenerative medicine, but also for drug discovery and toxicology, human developmental biology and cancer research. Advances in all of these fields require internationally co-ordinated work by numerous researchers to establish common standards and procedures that will allow the ready cross-verification of results between laboratories. The results of the present ISCI study provide a step in this direction, and indicate the commonality of many features of the hES cell lines currently in use. However, they do not exclude the possibility of more subtle differences between lines, for example, with respect to their differentiation potential. The present study also does not address the extent to which the phenotype of hES cells may be influenced by progressive genetic and epigenetic changes that can occur upon prolonged culture under suboptimal conditions. A new initiative (ISCI2) is now planned to assess the nature of the genetic changes that occur in hES cells on prolonged passage and to compare the performance of various new media that have been proposed for the long-term maintenance of hES cells under more defined conditions. ISCI2 will also continue to maintain and expand the International Stem Cell Forum's Registry of validated hES lines established in ISCI1 (, by providing arrangements for validation and inclusion of additional lines into the future.



Study design.

Members of the International Stem Cell Forum were asked to nominate laboratories in their countries that had derived hES cell lines. These 'primary' laboratories were invited to participate and to submit their hES cell lines for comparison in the study. Initially 75 lines were registered with the ISCI but, for practical reasons, material from only 63 lines was eventually analyzed (Supplementary Table 6). Of these, four were duplicated, with material being submitted both from a primary laboratory, and separately from cultures of the same cells maintained in one of two 'secondary' laboratories. Thus the study comprised 59 independent primary lines, together with the four duplicates included to provide some insights into the stability of lines between laboratories.

Owners of the different cell lines were asked to sign a participant's agreement, indicating that they agreed to inclusion of their hES cell lines in the study, and confirming that the cell lines had been derived in accordance with ethical standards for informed consent to embryo donation for research comparable with those required for hES cell derivation in the UK. These lines are generally available to the scientific community and contact details are provided in Supplementary Table 7 online. The intention in the ISCI was that analyses should be conducted in as consistent and reproducible a manner as possible. Inevitably in a multilab study of this type, for quality control and other reasons, not every line was represented in every assay. For logistical and other reasons it was not practicable to transfer live cell cultures between laboratories. Accordingly, flow cytofluorimetric analyses of antigen expression were conducted in the individual laboratories of the ISCI participants. All other analyses, however, were carried out by single reference laboratories using samples prepared by the participating laboratories from their own cell cultures. Participants were asked, where possible, to culture their cells using a standard protocol and medium ('Standard' conditions), and to note any significant deviations (see Supplementary Table 6). In some cases, they also provided additional data and materials from cells grown under the conditions routinely employed in their laboratories ('Alternate' conditions). Each laboratory analyzed and provided material from two cultures of each of their hES cell lines, separated by at least one month, denoted as Time Point 1 (TP1) and Time Point 2 (TP2). They also carried out a simple differentiation protocol by culturing clumps of ES cells (embryoid bodies66) in suspension for 10 d, then providing material from these for gene expression studies (see below), again on two occasions. Because the embryoid bodies were only allowed to develop for 10 d, they were expected to retain a significant population of persisting hES cells, as well as their newly differentiated derivatives. Conversely, experience suggests that hES cell cultures themselves also contain some differentiated cells.

Data collection and analysis.

An archive stock of hybridomas, comprising 17 monoclonal antibodies commonly used to study hES cells, was established at the UK Stem Cell Bank (National Institute for Biological Standards and Control, South Mimms, UK), with agreement from the different organizations that own them (Supplementary Table 1). Single batches of antibody were prepared, titrated and distributed to the participating laboratories for use in flow cytofluorimetric analyses of antigen expression according to a standard protocol. A well-characterized hEC cell line 2102Ep67 was provided and cultured by each laboratory as a reference standard; for detailed cluster analysis, data from assays that did not meet specified criteria with respect to antibody reactivity with 2102Ep cells were excluded. Cells were also grown on glass slides, fixed and dispatched to a single laboratory (M. Pera, Monash University, Australia) for in situ immunofluorescence. DNA and RNA were prepared centrally from cell lysates submitted by participating laboratories (The UK Stem Cell Bank). Samples were then provided to single reference laboratories for gene expression and DNA fingerprinting (Geneservice) and for imprinting studies (P. Rugg-Gunn and R. Pedersen, University of Cambridge). Flow cytofluorimetric and gene expression data were subjected to statistical analysis (Computational Sciences Unit, The Jackson Laboratory). Culture samples were tested for mycoplasma, bacteria, fungi and viruses (The UK Stem Cell Bank, at National Institute of Biological Standards and Control), and preexisting slides from xenografts were analyzed by a single, experienced pathologist (I. Damjanov, University of Kansas).

Cells and cell culture.

Individual participating laboratories chose which cell lines they would contribute to the initiative. Sufficient Knockout Serum replacement (KSR), bFGF and 'Knock-out' DMEM was donated and distributed to each laboratory by Invitrogen from a single lot of each reagent. The lots of KSR and bFGF were selected following a prescreen of two lots of each tested for their ability to support culture of H7 hES cells. Each laboratory was asked, where possible, to culture their cells in a 'Standard' medium, as previously described68. Full details of this medium are provide in Supplementary Table 6, which also records any significant deviations from this formulation adopted by particular laboratories. Cells were maintained on inactivated feeder cells as used by each group; in most cases these were mouse embryo fibroblasts but a few laboratories used human fibroblast feeders.

Generally, cultures were passaged using collagenase and scraping68, but exceptions are noted in Supplementary Table 6. Several laboratories also provided samples of cells cultured in conditions normally used in those laboratories, denoted 'Alternate' conditions. Generally, the most significant difference between Standard and Alternate conditions was the use of FBS rather than KSR to supplement the medium; details are provided in Supplementary Table 6.

Cells were assayed and cell lysates prepared from two such undifferentiated cultures of each hES cell line, passaged for approximately a month between each sampling, denoted Time Point 1 and Time Point 2 (UT1 and UT2, respectively). Two separate batches of embryoid bodies, (denoted DT1 and DT2), were also prepared a month apart from each line by harvesting clumps of hES cells from stock cultures using collagenase and scraping, followed by culture in suspension for ~10 d in the same medium as used to culture the stock, undifferentiated hES cells.

For in situ antigen staining, each cell line was cultured in chambers (QuadriPerm dishes, Sartorius) on glass slides (ethanol resistant multitest slide with 12 wells, ICN, cat no. 604120E). After growth for several days the slides were fixed in 70% ethanol vol/vol, dried and shipped to a single central laboratory (M. Pera, Monash University) for immunofluorescence analysis.

For control purposes, each participating laboratory was provided with a culture of 2102Ep cl.4D3 hEC cells to be cultured at high density (>5 × 106 cells per 75 cm2 tissue culture flask) as previously described67. Limited differentiation of 2102Ep cells was induced by culture at low cell densities (seeding cells at 105 cells per 75 cm2 tissue culture flask). Undifferentiated NTERA2 cl.D1 hEC cells were cultured, and differentiated NTERA2 cells were induced with retinoic acid as previously described69.

Antibody production and antigen assays.

Culture supernatants from each hybridoma were collected and the resulting preparation for each antibody was titrated by flow cytofluorimetry on 2102Ep and NTERA2 EC cells and NTERA2 cells induced to differentiate with retinoic acid, as previously described70. Monoclonal antibody MC480 was tested on a mouse EC cell line, P19 (ref. 71).

Aliquots from single batches of each antibody were distributed to each participating laboratory, together with a detailed protocol for flow cytofluorimetry to test for antigen expression (Supplementary Methods online). In all cases supernatant from the parental myeloma cell line P3X63Ag8 (ref. 72) was used as a negative control, whereas reactivity with the pan-human antibody TRA-1-85 (anti-Oka)37 was used as a positive control.

As a quality control for flow cytofluorimetry, each laboratory tested all antibodies on the hEC cell line, 2102Ep67 and results from assays on hES cells were only included in detailed further analyses if the laboratory scored 2102Ep cells strongly positive (>70%) for reactivity with at least three out of the four antibodies, TRA-1-85 and the anti-keratan sulfate antibodies TRA-1-60, TRA-1-81 and GCTM2, and negative (<10% positive) for reactivity with the negative control antibody from P3X63Ag8. Undifferentiated hEC cells, like 2102Ep, are typically strongly positive for TRA-1-85, TRA-1-60, TRA-1-81 and GCTM2 reactivity and negative for reactivity with P3X63Ag8 (ref. 73). From the remaining assays, data were excluded if the hES samples were scored <50% positive for reactivity with TRA-1-85. Aliquots of the same antibody preparations were provided to the Central Laboratory (M. Pera), responsible for in situ staining of fixed cultures of hES cells.

Gene expression, imprinting and DNA fingerprinting.

DNA and RNA were prepared at the UK Stem Cell Bank from the Trizol lysates submitted by participating laboratories. Samples were then provided to single reference laboratories for gene expression and DNA fingerprinting (Geneservice) and for imprinting studies (Rugg-Gunn and Pedersen, University of Cambridge). For gene expression studies, Applied Biosystems (ABI) prepared a batch of custom low density arrays (LDAs) using their proprietary microfluidics Genecard system. All of primers used were designed by ABI to perform with a measured efficiency of 100% ± 10% ( (see Supplementary Table 3). These were analyzed using Taqman based reagents74. Gene expression is reported as a DeltaCt value. For each individual gene the number of amplification cycles for the fluorescent reporter signal to reach a common threshold value (Ct) was estimated, and then normalized by subtracting the Ct value obtained from the same sample for a positive control transcript (in this case beta-actin), to give the DeltaCt value. Thus an increase in DeltaCt of 1.0 represents a twofold decrease in relative gene expression level.

The LDAs allowed for the simultaneous, triplicate analysis of gene expression levels by quantitative PCR, for 96 genes (90 test genes and six internal controls, see Supplementary Table 3). In addition quantitative PCR for three additional genes was carried out by individual assays using primers provided by ABI. Genes were selected for inclusion in the study on the following criteria: (i) expression in hEC, primordial germ cells or hES cells, (ii) known role in maintenance of pluripotency in hES cells including NANOG, SOX2, POU5F1, (iii) identification in microarray-based experiments as being correlated with the stem cell state, (iv) expression known to be characteristic of specific differentiation lineages in the post implantation embryo and mature tissues. The primers used were designed by ABI (see Supplementary Table 3). For two genes, NANOG and TDGF, it was not possible to design primers that could work on the LDA card to give unambiguous results. Primers for these genes were therefore designed for use in individual assays. In addition a third individual assay for XIST was included in the study due to data published post design and manufacture of the LDA cards indicating potential variability in XIST expression even between cell lines of the same sex.


Samples were analyzed for allele-specific expression patterns as previously described26: for primers see Supplementary Table 8 online. Additional material from HES-3 and HES-4 cell lines cultured in BTI, Singapore was also included in the imprinting study. For each gene, the contribution of the 'minor' allele to the total expression of the gene was calculated by measuring the area under each SNP chromatogram peak using NIH ImageJ software. Expression was considered to be monoallelic if the minor allele contributed 0–14% to the total expression of the gene; partial allelic expression was defined as the minor allele expression contributing to 15–29% of the total expression; and expression was defined as biallelic if the minor allele contributed 30–50% of the total expression.


Culture samples were tested for the presence of mycoplasma, bacteria and fungi at the UK Stem Cell Bank using in-house protocols. Samples were also tested for induction of a cytopathic effect when inoculated onto mouse embryo fibroblasts and MRC-5 human fibroblasts75 and also for hemadsorption of chick and guinea pig red blood cells Selected samples, including any showing apparent cytopathic effects, were also reviewed by EM to assess the presence of any virus-like particles.

Histology of xenograft tumors.

Participating laboratories submitted slides from xenografts they had already produced for analysis to a single, experienced pathologist. Some xenografts were produced from cloned lines of hES cells and some were treated and or submitted to special growth conditions that were not known to the pathologist who examined the grafts microscopically. All slides were stained, routinely, with hematoxylin and eosin. Some slides were also stained with the periodic Shiff's reaction, or with antibodies to smooth muscle actin or neurofilament proteins.

Statistical analysis.

General. Flow cytofluorimetric and gene expression data were subjected to statistical analysis by The Statistics and Analysis Group of the Computational Sciences Unit, The Jackson Laboratory, Bar Harbor, Maine, USA., using R2.3.0 ( software.

FACS data analysis.

Boxplots were generated from the average values of data which passed the quality control criteria (see antibody production and antigen assays section). Data were considered to be outlying when falling greater than 1.5 times the interquartile range and are indicated by open circles. The distribution of percentage positive cells in the flow cytofluorimetric assays was skewed to the right, so that a Logit transformation was applied to all the data. The transformed data were used for statistical analysis. Data for both the control cell line (2102Ep) and control antigens (P3X, TRA-1-85) were examined. Data points failing to meet quality control criteria as described above (Antibody Production and Analysis) were excluded. As a result, data were analyzed from 44 cell lines provided by 12 laboratories. Measurements at two time points were averaged and the averaged values were used in clustering analysis. Two-way hierarchical clustering analysis with Euclidean distance as a dissimilarity measure was performed. A heat map plot was generated to show clusters for both antigens and cell lines.

Gene expression analysis.

Quantitative PCR expression data for 99 genes from 61 cell lines (57 primary lines and 4 duplicate lines) were recorded in this part of the study. Six out of the 99 genes were control genes. For both differentiated and undifferentiated cell lines, all genes were measured at two time points. At each time point, triplicates were recorded to minimize measurement error and gene expression was normalized to β-actin (ACTB); the s.d. for each triplicate was calculated. Outliers were excluded from the triplicates with a s.d. >2. Averages of the triplicate gene expression data were used for further statistical analysis. Statistical analysis: Two-way hierarchical clustering with Euclidean distance was performed. Heat map plots were generated to present clusters for both cell lines and genes. Pearson correlations among genes were calculated.

The correlations between NANOG and every other gene were calculated taking into account the interaction effects of cell lines, differentiation conditions and time points. Genes that have greater correlations with NANOG are expected to have the same expression pattern as NANOG across the combination level of cell line, differentiation and time point.

Cell line availability. Supplementary Table 7

online provides contact information and conditions of availability for the cell lines described in this study. Materials transfer agreements indicated in this table are provided in Supplementary Notes online.

Note: Supplementary information is available on the Nature Biotechnology website.

The International Stem Cell Initiative

Oluseun Adewumi1, Behrouz Aflatoonian2, Lars Ahrlund-Richter3, Michal Amit4, Peter W Andrews5, Gemma Beighton6, 7, Paul A Bello8, Nissim Benvenisty9, Lorraine S Berry1, Simon Bevan6, Barak Blum9, Justin Brooking6, 10, Kevin G Chen11, Andre B H Choo12, Gary A Churchill13, Marie Corbel14, Ivan Damjanov15, Jon S Draper16, Petr Dvorak17, Katarina Emanuelsson18, Roland A Fleck1, Angela Ford5, Karin Gertow3, 19, Marina Gertsenstein16, Paul J Gokhale5, Rebecca S Hamilton11, Ales Hampl17, Lyn E Healy1, Outi Hovatta20, Johan Hyllner18, Marta P Imreh21, 22, Joseph Itskovitz-Eldor4, Jamie Jackson5, Jacqueline L Johnson8, Mark Jones5, Kehkooi Kee23, Benjamin L King13, Barbara B Knowles24, Majlinda Lako25, Franck Lebrin26, 27, Barbara S Mallon11, Daisy Manning28, 29, Yoav Mayshar9, Ronald D G Mckay11, Anna E Michalska30, Milla Mikkola31, Masha Mileikovsky16, Stephen L Minger32, Harry D Moore2, Christine L Mummery26, Andras Nagy16, Norio Nakatsuji33, Carmel M O'Brien8, Steve K W Oh12, Cia Olsson31, Timo Otonkoski31, Kye-Yoon Park11, Robert Passier26, Hema Patel1, Minal Patel34, Roger Pedersen14, Martin F Pera35, 36, Marian S Piekarczyk28, Renee A Reijo Pera23, Benjamin E Reubinoff37, Allan J Robins38, Janet Rossant39, Peter Rugg-Gunn14, 40, Thomas C Schulz38, Henrik Semb41, Eric S Sherrer38, Henrike Siemen23, Glyn N Stacey1, Miodrag Stojkovic25, 42, Hirofumi Suemori33, Jin Szatkiewicz13, Tikva Turetsky37, Timo Tuuri31, Steineke van den Brink26, Kristina Vintersten16, Sanna Vuoristo31, Dorien Ward26, Thomas A >Weaver6, Lesley A Young1 & Weidong Zhang13

Project coordination: Peter W Andrews5

Coordination of data collection and analysis: Paul J Gokhale5

Coordination of material distribution and sample collection: Lyn E Healy1

The scientific management of the ISCI project was provided by a steering committee comprising: Peter W Andrews5, Nissim Benvenisty9, Barbara S Mallon24, Ronald D G Mckay11, Martin F Pera35, Janet Rossant39, Henrik Semb41, Glyn N Stacey1



The International Stem Cell Initiative was sponsored and supported by the International Stem Cell Forum. Additional support from individual participants, and for elements of the project, was also provided by: A*STAR; Academy of Sciences of the Czech Republic; The Academy of Finland; Applied Biosystems; Axordia Ltd.; Bayer; Biotechnology and Biological Sciences Research Council, Centre for Cell Therapy and Tissue Repair, Charles University, Prague, Czech Republic; Chemicon now part of Millipore; Dutch Platform for Tissue Engineering; The Ellison Medical Foundation; Engineering and Physical Sciences Research Council; Embryonic Stem Cell International; GE Healthcare; The Helsinki University Central Hospital Research; Invitrogen Corporation; The Juvenile Diabetes Research Foundation; The Medical Research Council; The Ministry of Education, Culture, Sports, Science, & Technology of Japan; The National Institute for Biological Standards & Control; National Institutes of Health grants 9R24RR021314-04, 5R24RR018405, 5R24RR017498-04; One North East Regional Developmental Agency; The Shelby Cullom Davis Foundation; Stem Cell Network Canada; Stem Cells in Development and Disease; Swedish Research Council. We would like to thank the following for their invaluable assistance with the project: Marga van Rooijen, Hubrecht Laboratory; Christine Pigott, University of Sheffield; Ludmila Ruban, University of Sheffield; Katie Amps, University of Sheffield; Nicholas Jenkins, University of Sheffield; Jeff Jones, WiCell Research Institute; Nancy Block, WiCell Research Institute. We are also grateful to D. Melton and C. Cowan, Harvard University/Howard Hughes Medical Institute, respectively, for making available the human ES cell lines HUES1-17 and to the ES Cell International Pte Ltd. for human ES cell lines HES1-6.

Competing interests statement

The author declares competing financial interests.

Received 8 January 2007; Accepted 31 May 2007; Published online 17 June 2007.



  1. Allegrucci, C. & Young, L.E. Differences between human embryonic stem cell lines. Hum. Reprod. Update 13, 103–120 (2006). | Article | PubMed |
  2. Pera, M.F., Roach, S. & Ellis, C.J. Comparative biology of mouse and human embryonal carcinoma. Cancer Suveys 9, 243–262 (1990). | ChemPort |
  3. Andrews, P.W. From teratocarcinomas to embryonic stem cells. Phil. Trans. R. Soc. Lond. B 357, 405–417 (2002). | Article |
  4. Ginis, I. et al. Differences between human and mouse embryonic stem cells. Dev. Biol. 269, 360–380 (2004). | Article | PubMed | ChemPort |
  5. Andrews, P.W. et al. The International Stem Cell Initiative: toward benchmarks for human embryonic stem cell research. Nat. Biotechnol. 23, 795–797 (2005). | Article | PubMed | ISI | ChemPort |
  6. Brimble, S.N. et al. Karyotypic stability, genotyping, differentiation, feeder-free maintenance, and gene expression sampling in three human embryonic stem cell lines derived prior to August 9, 2001. Stem Cells Dev. 13, 585–597 (2004). | Article | PubMed | ISI | ChemPort |
  7. Dvash, T. et al. Temporal gene expression during differentiation of human embryonic stem cells and embryoid bodies. Hum. Reprod. 19, 2875–2883 (2004). | Article | PubMed | ChemPort |
  8. Sperger, J.M. et al. Gene expression patterns in human embryonic stem cells and human pluripotent germ cell tumors. Proc. Natl. Acad. Sci. USA 100, 13350–13355 (2003). | Article | PubMed | ChemPort |
  9. Brandenberger, R. et al. MPSS profiling of human embryonic stem cells. BMC Dev. Biol. 4, 10 (2004). | Article | PubMed | ChemPort |
  10. Rao, R.R. et al. Comparative transcriptional profiling of two human embryonic stem cell lines. Biotechnol. Bioeng. 88, 273–286 (2004). | Article | PubMed | ChemPort |
  11. Abeyta, M.J. et al. Unique gene expression signatures of independently-derived human embryonic stem cell lines. Hum. Mol. Genet. 13, 601–608 (2004). | Article | PubMed | ISI | ChemPort |
  12. Pera, M.F., Reubinoff, B. & Trounson, A. Human embryonic stem cells. J. Cell Sci. 113, 5–10 (2000). | PubMed | ISI | ChemPort |
  13. Nichols, J. et al. Formation of pluripotent stem cells in the mammalian embryo depends on the POU transcription factor OCT4. Cell 95, 379–391 (1998). | Article | PubMed | ISI | ChemPort |
  14. Matin, M.M. et al. Specific knockdown of Oct4 and beta2-microglobulin expression by RNA interference in human embryonic stem cells and embryonic carcinoma cells. Stem Cells 22, 659–668 (2004). | Article | PubMed | ISI | ChemPort |
  15. Scholer, H.R., Balling, R., Hatzopoulos, A.K., Suzuki, N. & Gruss, P. Octamer binding proteins confer transcriptional activity in early mouse embryogenesis. EMBO J. 8, 2551–2557 (1989). | PubMed | ISI | ChemPort |
  16. Mitsui, K. et al. The homeoprotein Nanog is required for maintenance of pluripotency in mouse epiblast and ES cells. Cell 113, 631–642 (2003). | Article | PubMed | ISI | ChemPort |
  17. Chambers, I. et al. Functional expression cloning of Nanog, a pluripotency sustaining factor in embryonic stem cells. Cell 113, 643–655 (2003). | Article | PubMed | ISI | ChemPort |
  18. Avilion, A.A. et al. Multipotent cell lineages in early mouse development depend on SOX2 function. Genes Dev. 17, 126–140 (2003). | Article | PubMed | ISI | ChemPort |
  19. Ciccodicola, A. et al. Molecular characterization of a gene of the 'EGF family' expressed in undifferentiated human NTERA2 teratocarcinoma cells. EMBO J. 8, 1987–1991 (1989). | PubMed | ChemPort |
  20. Bhattacharya, B. et al. Gene expression in human embryonic stem cell lines: unique molecular signature. Blood 103, 2956–2964 (2004). | Article | PubMed | ISI | ChemPort |
  21. Richards, M., Tan, S.P., Tan, J.H., Chan, W.K. & Bongso, A. The transcriptome profile of human embryonic stem cells as defined by SAGE. Stem Cells 22, 51–64 (2004). | Article | PubMed | ISI | ChemPort |
  22. Ware, C.B., Nelson, A.M. & Blau, C.A. A comparison of NIH-approved human ESC lines. Stem Cells 24, 2677–2684 (2006). | Article | PubMed | ISI | ChemPort |
  23. Monk, M., Boubelik, M. & Lehnert, S. Temporal and regional changes in DNA methylation in the embryonic, extraembryonic and germ cell lineages during mouse embryo development. Development 99, 371–382 (1987). | PubMed | ISI | ChemPort |
  24. Dean, W. et al. Altered imprinted gene methylation and expression in completely ES cell-derived mouse fetuses: association with aberrant phenotypes. Development 125, 2273–2282 (1998). | PubMed | ISI | ChemPort |
  25. Humpherys, D. et al. Epigenetic instability in ES cells and cloned mice. Science 293, 95–97 (2001). | Article | PubMed | ISI | ChemPort |
  26. Rugg-Gunn, P.J., Ferguson-Smith, A.C. & Pedersen, R.A. Epigenetic status of human embryonic stem cells. Nat. Genet. 37, 585–587 (2005). | Article | PubMed | ISI | ChemPort |
  27. Sun, B.W. et al. Temporal and parental-specific expression of imprinted genes in a newly derived Chinese human embryonic stem cell line and embryoid bodies. Hum. Mol. Genet. 15, 65–75 (2006). | Article | PubMed | ISI | ChemPort |
  28. Shevinsky, L.H., Knowles, B.B., Damjanov, I. & Solter, D. Monoclonal antibody to murine embryos defines a stage-specific embryonic antigen expressed on mouse embryos and human teratocarcinoma cells. Cell 30, 697–705 (1982). | Article | PubMed | ISI | ChemPort |
  29. Kannagi, R. et al. Stage-specific embryonic antigens (SSEA-3 and -4) are epitopes of a unique globo-series ganglioside isolated from human teratocarcinoma cells. EMBO J. 2, 2355–2361 (1983). | PubMed | ISI | ChemPort |
  30. Andrews, P.W., Banting, G., Damjanov, I., Arnaud, D. & Avner, P. Three monoclonal antibodies defining distinct differentiation antigens associated with different high molecular weight polypeptides on the surface of human embryonal carcinoma cells. Hybridoma 3, 347–361 (1984). | PubMed | ChemPort |
  31. Pera, M.F. et al. Analysis of cell-differentiation lineage in human teratomas using new monoclonal antibodies to cytostructural antigens of embryonal carcinoma cells. Differentiation 39, 139–149 (1988). | Article | PubMed | ISI | ChemPort |
  32. Andrews, P.W., Meyer, L.J., Bednarz, K.L. & Harris, H. Two monoclonal antibodies recognizing determinants on human embryonal carcinoma cells react specifically with the liver isozyme of human alkaline phosphatase. Hybridoma 3, 33–39 (1984). | PubMed | ChemPort |
  33. Draper, J.S., Pigott, C., Thomson, J.A. & Andrews, P.W. Surface antigens of human embryonic stem cells: changes upon differentiation in culture. J. Anat. 200, 249–258 (2002). | Article | PubMed | ISI | ChemPort |
  34. Oka, M. et al. CD9 is associated with leukemia inhibitory factor-mediated maintenance of embryonic stem cells. Mol. Biol. Cell 13, 1274–1281 (2002). | Article | PubMed | ISI | ChemPort |
  35. Andrews, P.W., Trinchieri, G., Perussia, B. & Baglioni, C. Induction of class I major histocompatibility complex antigens in human teratocarcinoma cells by interferon without induction of differentiation, growth inhibition, or resistance to viral infection. Cancer Res. 47, 740–746 (1987). | PubMed | ChemPort |
  36. Drukker, M. et al. Characterization of the expression of MHC proteins in human embryonic stem cells. Proc. Natl. Acad. Sci. USA 99, 9864–9869 (2002). | Article | PubMed | ChemPort |
  37. Williams, B.P. et al. Biochemical and genetic analysis of the OKa blood group antigen. Immunogenetics 27, 322–329 (1988). | Article | PubMed | ChemPort |
  38. Fenderson, B.A., Andrews, P.W., Nudelman, E., Clausen, H. & Hakomori, S. Glycolipid core structure switching from globo- to lacto- and ganglio-series during retinoic acid-induced differentiation of TERA-2-derived human embryonal carcinoma cells. Dev. Biol. 122, 21–34 (1987). | Article | PubMed | ChemPort |
  39. Baldassarre, G. et al. Transfection with a CRIPTO anti-sense plasmid suppresses endogenous CRIPTO expression and inhibits transformation in a human embryonal carcinoma cell line. Int. J. Cancer 66, 538–543 (1996). | Article | PubMed | ChemPort |
  40. Enver, T. et al. Cellular differentiation hierarchies in normal and culture-adapted human embryonic stem cells. Hum. Mol. Genet. 14, 3129–3140 (2005). | Article | PubMed | ISI | ChemPort |
  41. Kuroda, T. et al. Octamer and Sox elements are required for transcriptional cis regulation of Nanog gene expression. Mol. Cell. Biol. 25, 2475–2485 (2005). | Article | PubMed | ISI | ChemPort |
  42. Boyer, L.A. et al. Core transcriptional regulatory circuitry in human embryonic stem cells. Cell 122, 947–956 (2005). | Article | PubMed | ISI | ChemPort |
  43. Uwanogho, D. et al. Embryonic expression of the chicken Sox2, Sox3 and Sox11 genes suggests an interactive role in neuronal development. Mech. Dev. 49, 23–36 (1995). | Article | PubMed | ISI | ChemPort |
  44. Niwa, H. et al. Interaction between Oct3/4 and Cdx2 determines trophectoderm differentiation. Cell 123, 917–929 (2005). | Article | PubMed | ISI | ChemPort |
  45. Koutsourakis, M., Langeveld, A., Patient, R., Beddington, R. & Grosveld, F. The transcription factor GATA6 is essential for early extraembryonic development. Development 126, 723–732 (1999). | ISI | ChemPort |
  46. Dziadek, M.A. & Andrews, G.K. Tissue specificity of alpha-fetoprotein messenger RNA expression during mouse embryogenesis. EMBO J. 2, 549–554 (1983). | PubMed | ISI | ChemPort |
  47. Bishop, K.M., Goudreau, G. & O'Leary, D.D. Regulation of area identity in the mammalian neocortex by Emx2 and Pax6. Science 288, 344–349 (2000). | Article | PubMed | ISI | ChemPort |
  48. Lee, J.E. et al. Conversion of Xenopus ectoderm into neurons by NeuroD, a basic helix-loop-helix protein. Science 268, 836–844 (1995). | Article | PubMed | ISI | ChemPort |
  49. Martin, G.R. et al. X-chromosome inactivation during differentiation of female teratocarcinoma stem cells in vitro. Nature 271, 329–333 (1978). | Article | PubMed | ISI | ChemPort |
  50. Dhara, S.K. & Benvenisty, N. Gene trap as a tool for genome annotation and analysis of X chromosome inactivation in human embryonic stem cells. Nucleic Acids Res. 32, 3995–4002 (2004). | Article | PubMed | ChemPort |
  51. Hoffman, L.M. et al. X-inactivation status varies in human embryonic stem cell lines. Stem Cells 23, 1468–1478 (2005). | Article | PubMed | ChemPort |
  52. Andrews, P.W. et al. Embryonic stem (ES) cells and embryonal carcinoma (EC) cells: opposite sides of the same coin. Biochem. Soc. Trans. 33, 1526–1530 (2005). | Article | PubMed | ISI | ChemPort |
  53. Civin, C.I. & Rao, M.S. How many human embryonic stem cell lines are sufficient? A U.S. perspective. Stem Cells 24, 800–803 (2006). | PubMed |
  54. Abbott, A., Dennis, C., Ledford, H. & Smith, K. The lure of stem-cell lines. Nature 442, 336–337 2006). | Article | PubMed | ChemPort |
  55. Brimble, S.N. et al. The Cell Surface Glycosphingolipids SSEA-3 and SSEA-4 are not Essential for Human ES Cell Pluripotency. Stem Cells 25, 54–62 (2006). | PubMed |
  56. Tippett, P., Andrews, P.W., Knowles, B.B., Solter, D. & Goodfellow, P.N. Red cell antigens P (globoside) and Luke: identification by monoclonal antibodies defining the murine stage-specific embryonic antigens -3 and -4 (SSEA-3 and SSEA-4). Vox Sang. 51, 53–56 (1986). | PubMed | ChemPort |
  57. Draper, J.S., Moore, H.D., Ruban, L.N., Gokhale, P.J. & Andrews, P.W. Culture and characterization of human embryonic stem cells. Stem Cells Dev. 13, 325–336 (2004). | PubMed | ChemPort |
  58. Baker, D.E. et al. Adaptation to culture of human embryonic stem cells and oncogenesis in vivo. Nat. Biotechnol. 25, 207–215 (2007). | Article | PubMed | ISI | ChemPort |
  59. Henderson, J.K. et al. Preimplantation human embryos and embryonic stem cells show comparable expression of stage-specific embryonic antigens. Stem Cells 20, 329–337 (2002). | Article | PubMed | ISI | ChemPort |
  60. Nagy, A. et al. Embryonic stem cells alone are able to support fetal development in the mouse. Development 110, 815–821 (1990). | PubMed | ISI | ChemPort |
  61. Xu, C., Liguori, G., Adamson, E.D. & Persico, M.G. Specific arrest of cardiogenesis in cultured embryonic stem cells lacking Cripto-1. Dev. Biol. 196, 237–247 (1998). | Article | PubMed | ISI | ChemPort |
  62. Levine, A.J. & Brivanlou, A.H. GDF3, a BMP inhibitor, regulates cell fate in stem cells and early embryos. Development 133, 209–216 (2006). | Article | PubMed | ISI | ChemPort |
  63. Mitalipov, S.M. Genomic imprinting in primate embryos and embryonic stem cells. Reprod. Fertil. Dev. 18, 817–821 (2006). | Article | PubMed | ISI | ChemPort |
  64. Cooper, P.R. et al. Divergently transcribed overlapping genes expressed in liver and kidney and located in the 11p15.5 imprinted domain. Genomics 49, 38–51 (1998). | Article | PubMed | ISI | ChemPort |
  65. Dao, D. et al. IMPT1, an imprinted gene similar to polyspecific transporter and multi-drug resistance genes. Hum. Mol. Genet. 7, 597–608 (1998). | Article | PubMed | ISI | ChemPort |
  66. Itskovitz-Eldor, J. et al. Differentiation of human embryonic stem cells into embryoid bodies compromising the three embryonic germ layers. Mol. Med. 6, 88–95 (2000). | PubMed | ISI | ChemPort |
  67. Andrews, P.W., Goodfellow, P.N., Shevinsky, L.H., Bronson, D.L. & Knowles, B.B. Cell-surface antigens of a clonal human embryonal carcinoma cell line: morphological and antigenic differentiation in culture. Int. J. Cancer 29, 523–531 (1982). | Article | PubMed | ChemPort |
  68. Amit, M. et al. Clonally derived human embryonic stem cell lines maintain pluripotency and proliferative potential for prolonged periods of culture. Dev. Biol. 227, 271–278 (2000). | Article | PubMed | ISI | ChemPort |
  69. Andrews, P.W. Retinoic acid induces neuronal differentiation of a cloned human embryonal carcinoma cell line in vitro. Dev. Biol. 103, 285–293 (1984). | Article | PubMed | ChemPort |
  70. Andrews, P.W., Oosterhuis, J.W. & Damjanov, I. Cell lines from human germ cell tumours. in Teratocarcinomas and Embryonic Stem Cells: A Practical Approach. (ed. Robertson, E.J.) 207–248, (IRL Press, Oxford, 1987).
  71. McBurney, M.W. & Rogers, B.J. Isolation of male embryonal carcinoma cells and their chromosome replication patterns. Dev. Biol. 89, 503–508 (1982). | Article | PubMed | ChemPort |
  72. Kohler, G. & Milstein, C. Continuous cultures of fused cells secreting antibody of predefined specificity. Nature 256, 495–497 (1975). | Article | PubMed | ISI | ChemPort |
  73. Andrews, P.W. et al. Comparative analysis of cell surface antigens expressed by cell lines derived from human germ cell tumours. Int. J. Cancer 66, 806–816 (1996). | Article | PubMed | ISI | ChemPort |
  74. Heid, C.A., Stevens, J., Livak, K.J. & Williams, P.M. Real time quantitative PCR. Genome Res. 6, 986–994 (1996). | Article | PubMed | ISI | ChemPort |
  75. Jacobs, J.P., Jones, C.M. & Baille, J.P. Characteristics of a human diploid cell designated MRC-5. Nature 227, 168–170 (1970). | Article | PubMed | ISI | ChemPort |
  1. UK Stem Cell Bank, Division of Cell Biology and Imaging, National Institute for Biological Standards and Control, South Mimms, Herts., EN6 3QG, UK.
  2. Centre for Stem Cell Biology, Department of Molecular Biology and Biotechnology, The University of Sheffield, Sheffield S10 2TN, UK.
  3. Unit for Molecular Embryology, Department for Laboratory Medicine, Karolinska Institutet, 141 57 Stockholm, Sweden.
  4. Rambam Medical Center, Technion–Israel Institute of Technology, POB 9602, Haifa 31096, Israel.
  5. Centre for Stem Cell Biology, Department of Biomedical Science, The University of Sheffield, Sheffield S10 2TN, UK.
  6. Geneservice Limited, No.2 Cambridge Science Park, Cambridge, CB3 0FE, UK.
  7. Cancer Research UK Genotyping Facility, St James's University Hospital, Leeds, LS9 3LP, UK.
  8. Stem Cell Sciences Ltd., Building 75, Monash University, Wellington Rd., Clayton, Victoria 3800, Australia.
  9. The Hebrew University of Jerusalem, Institute of Life Sciences, Department of Genetics, Edmond J. Safra Campus, Givat Ram, Jerusalem 91904, Israel.
  10. Matrix Technologies, Thermo Fisher Scientific, Unit 2 Lower Meadow Road, Handforth, SK9 3LP, UK.
  11. NIH Stem Cell Unit, Bldg 35/2B-213, MSC 3703, National Institutes of Health, 9000 Rockville Pike, Bethesda, Maryland 20892, USA.
  12. Bioprocessing Technology Institute, 20 Biopolis Way, #06-01 Centros, Singapore 138668.
  13. The Jackson Laboratory, 600 Main Street, Bar Harbor, Maine 04609, USA.
  14. Department of Surgery, University of Cambridge, Addenbrookes Hospital, Hills Road, Cambridge, CB2 2XY, UK.
  15. Department of Pathology, The University of Kansas, School of Medicine, 3901 Rainbow Blvd., Kansas City, Kansas 66160, USA.
  16. Samuel Lunenfeld Research Institute, Room 881, Mount Sinai Hospital, 600 University Avenue, Toronto, Ontario, Canada, M5G 1X5.
  17. Department of Biology, Faculty of Medicine, Masaryk University, Brno, Czech Republic, and Department of Molecular Embryology, Institute of Experimental Medicine, Academy of Sciences of the Czech Republic, Brno, Czech Republic.
  18. Cellartis AB, Arvid Wallgrens Backe 20, 413 46, Göteborg, Sweden.
  19. Monash Immunology and Stem Cell Laboratories (MISCL), Level 3, STRIP–Building 75, Monash University, Clayton, Victoria, 3800 Australia.
  20. Karolinska Institutet, Dept. of Clinical Science, Intervention and Technology, Div. of Obstetrics/Gynecology, Karolinska University Hospital, Huddinge, SE-141 86 Stockholm, Sweden.
  21. Department of Laboratory Medicine, Clinical Research Centre, Karolinska Institutet, 141 57 Stockholm, Sweden.
  22. Uppsala University, Department of Development and Genetics, SE-7523236, Sweden.
  23. Center for Human Embryonic Stem Cell Research and Education, Institute for Stem Cell Biology & Regenerative Medicine, Department of Obstetrics and Gynecology, Stanford University, Palo Alto, California USA.
  24. The Jackson Laboratory & the Institute for Molecular Biophysics, 600 Main Street, Bar Harbor, Maine 04609, USA.
  25. Institute of Human Genetics, International Centre for Life, Central Parkway, University of Newcastle, Newcastle upon Tyne, NE1 3BZ, UK.
  26. Hubrecht Laboratory, Uppsalalaan 8, 3584CT Utrecht, The Netherlands.
  27. INSERM UR 525, Faculté de Médecine Pitié-Salpêtrière, Paris, France.
  28. WiCell Research Institute, PO Box 7365, Madison, WI 53707-7365, USA.
  29. WiCell Research Institute, 1298 S Mason Road, St. Louis, Missouri 63131, USA.
  30. Laboratory of Embryonic Stem Cell Biology, Monash Institute of Medical Research, Monash University & the Australian Stem Cell Centre, Wellington Road, Clayton 3800, Victoria, Australia.
  31. Biomedicum Stem Cell Center, Biomedicum Helsinki, PO Box 63 FIN-00014, University of Helsinki, Helsinki, Finland.
  32. Wolfson Centre for Age-Related Diseases, King's College London, London SE1 1UL, UK.
  33. Institute for Frontier Medical Sciences, Kyoto University, Kyoto 606-8507, Japan.
  34. King's College London/University College London, Regenerative Medicine Bioprocessing Unit, The Advanced Centre for Biochemical Engineering, Department of Biochemical Engineering, University College London, Torrington Place, London, WC1E 7JE, UK.
  35. Centre for Reproduction and Development, Monash Institute of Medical Research, Monash University & the Australian Stem Cell Centre, Wellington Road, Clayton 3800, Victoria, Australia.
  36. Center for Stem Cell and Regenerative Medicine, Keck School of Medicine, University of Southern California, c/o Zilkha Neurogenetic Institute, Room 543, 1501 San Pablo Street, Los Angeles, California 90089-282, USA.
  37. Hadassah Human Embryonic Stem Cell Research Center, The Goldyne Savad Institute of Gene Therapy & The Department of Obstetrics and Gynecology, Hadassah University Medical Center, Ein Kerem POB 12000 Jerusalem 91120, Israel.
  38. Novocell, 111 Riverbend Rd., Athens, Georgia 30606, USA.
  39. Developmental Biology Program, The Hospital for Sick Children, TMDT Building, Room 13-305, 101 College Street, Toronto, Ontario, M5G 1L7, Canada.
  40. Program for Developmental Biology. The Hospital for Sick Children, 555 University Avenue, Toronto, Ontario, M5G 1X8, Canada.
  41. Stem Cell Center, Lund University, BMC, B10, Klinikgatan 26, SE-221-84 Lund, Sweden.
  42. Centro de Investigación Príncipe Felipe, Valencia, Spain.

Correspondence to: e-mail:


These links to content published by Nature Research are automatically generated.