Properties of four human embryonic stem cell lines maintained in a feeder-free culture system



Several laboratories have begun evaluating human ES (hES) cell lines; however, direct comparisons between different hES cell lines have not been performed. We have characterized the properties of four human cell lines maintained in feeder-free culture conditions. Quantitative assessment of surface markers, microarray analysis of gene expression patterns, expression of SOX-2, UTF-1, Rex-1, OCT3/4, CRIPTO, and telomerase activity demonstrated similar patterns in all hES cell lines examined. Undifferentiated hES cells do not respond to neurotransmitters such as acetylcholine, glutamate, and gamma-aminobutyric acid. In addition, the undifferentiated hES cells possess gap junctions. Although similarities in marker expression were observed, allotyping showed that all four lines have a distinct HLA profile, predicting differences in transplantation responses. These data provide the first detailed comparison of different hES cell lines and demonstrate remarkable similarities among lines maintained in identical culture conditions. Developmental Dynamics 229:243–258, 2004. © 2003 Wiley-Liss, Inc.


Embryonic stem (ES) cell lines were first generated from mouse blastocysts by culturing the inner cell mass (ICM) from preimplantation embryos on murine feeder layers. The resulting cultures contained populations of cells that grew as colonies, showed extensive replicative capacity and were pluripotent as demonstrated by their ability to contribute to multiple cell types in chimeric mice and differentiate in culture into ectodermal, endodermal, and mesodermal derivatives. The ability to generate mouse ES (mES) cell lines appears to be strain-specific, and the stability of lines varies depending on culture conditions (Smith, 2001a, b). Furthermore, mES cell lines derived from different mouse strains using different procedures differ in growth potential, differentiation ability, and the ability to contribute to chimeras (Kawase et al., 1994; Schoonjans et al., 2003). However, with careful culturing in appropriate conditions, mES cell lines can be maintained in continuous culture for years yet retain their pluripotent character and the ability to contribute to chimera formation (Suda et al., 1987).

More recently, primate and human ES (hES) cell lines have been described (Thomson et al., 1995, 1996, 1998). The first successful hES derivations were reported by Thomson and colleagues, who isolated five hES cell lines (Thomson et al., 1998). This initial success has been replicated by several laboratories, and there are over 60 different derivations listed in the NIH registry for research use. However, fewer than 10 cell lines are currently available in sufficient numbers for analysis and limited data on the fundamental properties of these lines is available. Some characterization has been performed on 26 of the 60 or so reported hES cell derivations (reviewed in Carpenter et al., 2003). Although similarities and some differences were noted, many of the lines are not clonal, and limited numbers of markers have been tested. Further comparisons have been hampered by dissimilar culture conditions and the necessity for growing hES cells on feeder layers. Direct comparisons of hES cell lines grown in feeder independent cultures will facilitate the assessment of similarities and differences among the hES cell lines.

Identifying reliable markers that can be used routinely and consistently to assess the state of hES cells is of great importance. Unfortunately, we cannot entirely depend on the wealth of data available from the derivation and maintenance of mES cell lines as even the limited analysis undertaken suggests differences between mouse and human ES cell lines. At least some of the human cell lines express surface markers (glycolipids and glycoproteins) that were originally identified on human embryonic carcinoma (EC) cells, and these surface antigens differ from those present on mES cells. Unlike mES cells, which can be maintained in the undifferentiated state by replacing the feeder layer with the addition of LIF to the medium, hES cells do not appear to require LIF for their derivation or propagation (Thomson et al., 1998; Reubinoff, 2000). Whether this indicates that the LIFR or CD130 are not expressed or whether LIF has different effects such as inhibition, as has been proposed for EC cells (Schuringa et al., 2002), remains to be determined. In addition, hES cells appear to have different differentiative capacity than the mES cells. In vitro, hES cells appear to be able to differentiate into trophoblast cells (Xu et al., 2002), whereas mES cells do not seem to have this capacity (Beddington and Robertson, 1989).

To assess the properties of hES cells and to begin to determine markers that could be used to identify the undifferentiated state, we examined and compared the properties of four hES lines by using quantitative analysis whenever possible. We have used flow cytometry, quantitative polymerase chain reaction (PCR), microarray analysis, electrophysiology, dynamic calcium imaging, and immunocytochemistry. In addition, we have generated an expressed sequence tag (EST) database using libraries generated from pooled samples of undifferentiated hES cells or embryoid bodies (EBs; Brandenberger et al., manuscript in preparation). We have interrogated this database to determine the relative abundance of specific gene products. By using these various techniques, we demonstrate that undifferentiated hES cells express receptors for SCF, Flt3L, and fibroblast growth factor (FGF) receptors, particularly FGF receptor-1 (FGFR1), but do not express CD130. hES cells can be distinguished from other cells by a constellation of markers and some of these markers differ from those used to identify mES cells. We show that the four hES cell lines have similar expression patterns, but differ in their allotype, suggesting that the immune response they provoke may differ in different patient populations.


hES Cells Morphologic and Phenotypic Features

hES cell lines have been generated in several laboratories by culturing the inner cell mass on mouse or human feeder layers (Thomson et al., 1998; Reubinoff et al, 2000; Amit and Itskovitz-Eldor, 2002; Richards et al., 2002). We have developed a feeder-free culture system in which the hES cells can be maintained over long-term culture using matrigel and conditioned medium (Xu et al., 2001). This system has been used to culture the H1, H7, H9, and H14 cell lines (Thomson et al., 1998; Xu et al., 2001) and under these conditions, the different cell lines show similar morphologies (Fig. 1A). To examine the properties of hES cells, we analyzed four different lines and determined the expression of surface markers by flow cytometry.

Figure 1.

A: Phase contrast photomicrographs of undifferentiated cultures of H1, H7, H9, and H14 human embryonic stem (hES) cells maintained in MEF-conditioned medium supplemented with basic fibroblast growth factor on growth factor reduced (GFR) matrigel for 6 days. Undifferentiated cells formed tight colonies composed of cells with high nucleus/cytoplasm ratios and containing prominent nucleoli (insets). Colonies were surrounded by stroma-like areas of differentiated cells. B: Flow cytometric analysis of day 7 H1 hES cells maintained in feeder-free conditions. Nonviable cells were identified using propidium iodide and excluded during analysis. The horizontal marker in each plot is positioned to represent the fluorescence of ∼99% of the appropriate fluorescein isothiocyanate- or phycoerythrin-conjugated isotype matched control; the vertical marker is positioned to represent the SSEA-4high population. Values in each quadrant indicate the percentage of positive cells. Scale bars = 300 μm in A, 50 μm in insets.[Color figure can be viewed in the online issue, which is available at]

Similar to hES cells maintained on feeders, hES cells in feeder-free conditions grow as compact colonies with a high nuclear cytoplasmic ratio and prominent nucleoli. The colonies are surrounded by differentiated cells that appear stromal or fibroblast-like (Fig. 1A). Whether maintained on feeders or in feeder-free conditions, hES cells derived in multiple labs have been demonstrated to express glycolipids and glycoproteins that were first identified on another source of human pluripotent stem cells, human EC cells. In general, these include the stage-specific embryonic antigens SSEA-3 and SSEA-4 and the tumor recognition antigens TRA-1-60 and TRA-1-81. Although the function of these glycolipids and glycoproteins is unclear, they show lower expression upon differentiation (Lebkowski et al., 2001; Draper et al., 2002). Although these markers are expressed consistently, it is unclear whether these markers are actually predictive of the status of differentiation of the hES cells.

Flow Cytometry Analysis

By using flow cytometry, we compared the expression of SSEA-4, TRA-1-60, and TRA-1-81 in all four hES cell lines and extended this examination to include evaluation of coexpression with other markers. Figure 1B shows two-color flow cytometric analysis of the H1 cell line at p27, showing the coexpression of SSEA-4 with other surface markers such as TRA-1-60, TRA-1-81, CD9, and CD133. We compared the expression of individual markers on the H1, H7, H9, and H14 cell lines and found that a high percentage of cells in all cell lines expressed SSEA-4, TRA-1-60, TRA-1-81 (Fig. 2). As shown in Figure 2, expression of SSEA-4, TRA-1-60, TRA-1-81, CD9, CD133, and SSEA-1 were assessed in many separate cultures from each cell line at various passages. Figure 2B presents a summary table of all the values (mean ± SEM, sample size and range of values) for each marker assessed in separate cultures of each cell line. SSEA-4, TRA-1-60, and TRA-1-81 were expressed by an average of 80–95% (n = 5–40 in four cell lines), 91–94% (n = 5–38 in four cell lines), and 88–93% (n = 5–36 in four cell lines) of the cells, respectively (Fig. 2B). Overall, the results from each cell line were quite similar. However, we did find statistically significant differences in SSEA-4 expression between H9 (79.9 ± 4.7%) and H1 (92.4 ± 1.2) and H7 (95.3 ± 1.0%) (Student's t-test: H9 vs. H1 P = 0.02, H9 vs. H7 P = 0.007) and between H7 and H14 (89.3 ± 4.0; P = 0.05). As shown in Figure 2A, we have observed the loss of SSEA-4 expression in the H9 line in several separate cultures from different users. However, upon closer evaluation, these cultures were traced back to a single parent culture. This loss of SSEA-4 was not correlated with the loss of expression of other surface markers except SSEA-3 (data not shown), which recognizes a different epitope on the same molecule, and no loss of pluripotency was observed.

Figure 2.

Surface marker expression in four human embryonic stem (hES) cell lines. A: Comparison of surface marker expression on H1 (circles), H7 (triangles), H9 (squares), and H14 (diamonds) hES cell lines. Each symbol represents measurement of a separate culture at different passages. Cells were maintained in MEF-conditioned medium supplemented with basic fibroblast growth factor on growth factor reduced (GFR) matrigel and harvested for flow cytometric analysis upon reaching confluence (day 7). Only viable cells (propidium iodide negative) were included in this analysis; positive events were defined as having greater fluorescence than 99% of the appropriate isotype matched control. Very similar expression of all markers tested was observed on viable hES cells. B: Table indicating means ± SEM for values presented in A. The range of values for each marker is indicated below the mean.

In addition, we examined the expression of CD9, a tetraspan transmembrane protein expressed by undifferentiated mouse ES cells, which appears to be regulated by activation of the LIF/STAT3 pathway and is decreased upon differentiation of the cells (Oka et al., 2002). We found that a large percentage of the cells in each hES cell line express CD9 (an average of 79–99%). Expression of CD9 was significantly less in H1 (78.7 ± 3.7) cells than in H9 (89.1 ± 1.7) and H14 (96.3 ± 0.7) cells (P = 0.008 and P = 0.04, respectively); however, the biological significance of this observation is unknown. CD133 is a pentaspan transmembrane glycoprotein expressed in primitive hematopoietic stem (HSC), neural, and endothelial cells (Miraglia et al., 1997; Yin et al., 1997; Uchida et al., 2000). We found that an average of 58–68% of the hES cells express CD133, and there was a small but significant difference in expression between H1 (57.9 ± 2.7) and H9 (65.6 ± 2.3; Student's t-test P = 0.04). We also detected CD90 (thy-1), a marker for HSC and neural cells, expression in undifferentiated hES cell cultures. CD90 was expressed by 100% of the cells from all four cell lines, similar to previous findings for the H1 (Draper et al., 2002) and H7 (Kaufman et al., 2001) lines.

Although in most cases almost all of the hES cells expressed SSEA-4, we observed two SSEA-4 expressing populations. A larger portion of the cells expressed SSEA-4 at high levels and was termed SSEA-4high and a smaller portion of the cells expressed SSEA-4 at lower levels and was termed SSEA-4low/neg (Fig. 1B). Between 65 and 90% of the hES cells showed SSEA-4high expression. The SSEA-4high expression appeared to correlate with the morphologic features of the undifferentiated hES cells; higher numbers of SSEA-4high cells were present in cultures with a greater number of undifferentiated colonies. TRA-1-60 and TRA-1-81 showed similar expression patterns, with portions of the populations showing high and low expression. By using two-color analyses, we examined the coexpression of SSEA-4 and the other hES cells markers described above (Figs. 1, 3). As described for the assessment of individual markers, many separate cultures at various passages from each cell lines were analyzed. Figure 3B shows summary values (mean ± SEM, sample size and range of values) of the percentage of SSEA-4high population coexpressing other markers. This analysis showed a direct correlation between the intensity of SSEA-4 and the intensity of TRA-1-60 or TRA-1-81. Moreover, TRA-1-60, TRA-1-81, and CD133 were preferentially expressed on the SSEA-4high cells. Small, but statistically significant differences in expression of SSEA-4high/TRA-1-60high expression were observed between the H1, H7, and H9 lines (Fig. 3A,B).

Figure 3.

Coexpression of surface markers on human embryonic stem (hES) cells. A: Comparison of surface marker expression in H1 (circles), H7 (triangles), H9 (squares), and H14 (diamonds) hES cell populations. Each symbol represents measurement of a separate culture at different passages. Cells were maintained in MEF-conditioned medium supplemented with basic fibroblast growth factor on growth factor reduced (GFR) matrigel and harvested for flow cytometric analysis upon reaching confluence (day 7). Only viable cells (propidium iodide negative) were included in this analysis. Positive events were defined as having greater fluorescence than 99% of the appropriate isotype matched control. Cultures of all four cell lines contained similar numbers of double positive cells. B: Table indicating means ± SEM for values presented in A. The range of values for each marker is indicated below the mean.

FGFRs Are Present in Undifferentiated hES Cell Cultures

Because basic FGF (bFGF) is known to be an important growth factor for the culture of hES cells, we examined the expression of FGF receptors in the different hES cell lines. We created an EST database using libraries generated from undifferentiated hES cells and EBs (see Experimental Procedures section) and used this database to estimate the abundance of gene products within the libraries. Assessment of the EST frequencies within the database indicated that the undifferentiated population of ES cells had more abundant FGFR1, FGFR2, FGFR3, FGFR4 transcripts than the differentiated EB population; however, only the differences seen for FGFR1 and FGFR3 were significant. FGFR expression was also evaluated using an antibody that reacts with the FGFR flg gene product (Biogenesis) and, therefore, identifies primarily FGFR1. We found that an average of 7–16% of the cells express FGFR (Fig. 2). Using two-color analysis, we further evaluated expression of markers in the SSEA-4high and SSEA-4low/neg populations. The SSEA-4high population showed dim expression of FGFR1 in a range of 2–60% of the cells, whereas the SSEA-4low/neg population was more variable, with some samples demonstrating very bright expression in greater than 70% of the cells and others showing dim expression similar to the SSEA-4high population (Table 1). An example of two such analyses is shown in Figure 4C,D. Table 1 summarizes findings from the assessment of SSEA-4high and SSEA-4low/neg expressing populations coexpressing growth factor receptors or other surface markers in many separate cultures from each hES cell line.

Table 1. Comparison of Surface Marker Expression on SSEA-4high and low/neg gated H1, H7, H9, and H14 hES cellsa
  • a

    Upper values indicate mean + SEM and lower values indicate the range of values for each marker. Sample sizes are the same as those shown in Figure 3B. hES, human embryonic stem.

  • *

    Significantly different from H1, P < 0.05.

  • **

    Significantly different from H7, P < 0.05.

Percentage of SSEA-4 high cell
 TRA-1-6098.0 ± 0.598.1 ± 0.898.3 ± 0.599.5 ± 0.2
 TRA-1-8196.9 ± 0.997.7 ± 1.195.0 ± 2.999.5 ± 0.5
 CD977.3 ± 4.691.9 ± 3.2*85.8 ± 3.697.9 ± 0.9
 CD13356.2 ± 6.865.9 ± 3.467.1 ± 3.476.0 ± 3.5
 SSEA-13.5 ± 0.84.8 ± 1.5*1.7 ± 0.3**2.0 ± 0.6
 FGFR15.6 ± 3.819.8 ± 3.910.5 ± 2.47.0 ± 3.3
 CD1300.1 ± 0.10.2 ± 0.11.0 ± 0.3**0.4 ± 0.2
 CD11718.6 ± 2.819.0 ± 4.024.0 ± 5.037.2 ± 5.2*,**
 CD13539.8 ± 6.246.2 ± 9.836.7 ± 7.533.8 ± 14.3
Percentage of SSEA-4 low/neg cells
 TRA-1-6061.3 ± 3.857.0 ± 5.371.1 ± 4.562.0 ± 7.0
 TRA-1-8151.1 ± 5.645.7 ± 6.158.8 ± 3.658.2 ± 9.1
 CD964.5 ± 6.256.8 ± 6.381.3 ± 2.6*,**76.9 ± 10.6
 CD13317.1 ± 2.022.0 ± 7.045.6 ± 4.1*,**38.2 ± 10.1
 SSEA-17.8 ± 1.911.9 ± 4.25.2 ± 1.16.1 ± 2.7
 FGFR25.0 ± 7.810.0 ± 4.1*9.0 ± 2.310.0 ± 2.3
 CD1305.3 ± 1.51.1 ± 0.51.6 ± 0.92.0 ± 1.3
 CD1179.0 ± 1.610.3 ± 2.322.1 ± 3.5*,**25.2 ± 5.2*,**
 CD1358.1 ± 5.12.8 ± 0.63.6 ± 0.75.4 ± 4.5
Figure 4.

CD130, fibroblast growth factor (FGF), FGF receptor (FGFR), c-kit, and Flt3L expression human embryonic stem (hES) cells. A: Genomic libraries of undifferentiated hES cell cultures and differentiated embryoid body (EB) cultures of hES cells were analyzed to compare the expressed sequence tag (EST) frequency of various fibroblast growth factors, their receptors, LIFR CD130, c-kit, and Flt3L. Very high EST counts of FGFR1 were detected in hES cells; no ESTs were detected in undifferentiated cells for CD130 or LIFR. Low EST counts for c-kit and Flt3L were detected in undifferentiated cells. Asterisks indicate significant difference of P ≤ 0.05. SSEA-4high-positive hES cell were further analyzed for expression of CD117, CD135 (B, H1 p27), CD130, and FGFR (C, H1 p27 and D, H7 p35) by flow cytometry. Cultures maintained in MEF-conditioned medium with basic FGF on growth factor reduced (GFR) matrigel were harvested for staining at confluence. Nonviable cells were identified by using propidium iodide and excluded during analysis. Differences in the expression of CD117, CD135, CD130, and FGFR1 were observed when hES cells were gated based on the intensity of SSEA-4 expression. SSEA-4high (undifferentiated) cells had no detectable CD130 expression and dimly expressed FGFR1, whereas SSEA-4low/neg cells had dim, but significant expression of CD130 and bright expression of FGFR1. In contrast, greater expression of CD117 and CD135 were observed on SSEA-4high cells.

hES Cells Express Receptors for Flt3L and SCF

hES cells express CD90 and CD133, markers closely linked with hematopoietic stem cell (HSC) populations. It is possible that different types of stem cells share expression of some markers and receptors. Therefore, we investigated hES cell expression of receptors for Flt3L and SCF, cytokines implicated in the maintenance of primitive HSCs, using antibodies against CD135 and CD117. We determined that all four hES cell lines expressed CD117 (average of 16–35%) and CD135 (average of 27–44%) (Fig. 2). Further analysis showed expression in the SSEA-4high, with 19–37% expressing CD117 and 34–46% expressing CD135, and less expression of these markers in the SSEA-4low/neg population (9–25% and 3–8%, respectively; Table 1).

hES Cells Express Low or Undetectable Levels of CD130

Mouse ES cells retain pluripotency by activation of STAT 3 by means of signaling through the CD130 receptor (Yoshida et al., 1994). In contrast, the hES cells do not appear to respond to LIF in the same manner. Therefore, we examined the expression of CD130 in our feeder-free cultures and found a low percentage of the population expressed CD130 (0 to 6% in all four cell lines; Figs. 2, 3). To determine whether this small number of cells coexpressed markers of the undifferentiated cells, we colabeled the hES cells with SSEA-4 and CD130 antibodies. The SSEA-4high population did not show any detectable expression of CD130, whereas the SSEA-4low/neg population showed dim expression in up to 5% of the cells (Fig. 4B; n = 4–13). This finding is consistent with the finding that EST counts of our sequenced library showed no ESTs for gp130 in the undifferentiated library (Fig. 4A). The EST counts for LIFR were also very low, with only 1 EST found in either the undifferentiated ES or the EB libraries. However, STAT3 appears to be present in both the undifferentiated and differentiated libraries. It is unknown whether STAT3 activation is necessary to maintain the hES cells in the undifferentiated state or whether STAT3 has a different function in the hES cells.

Transcription Factors Characteristic of Undifferentiated ES Cells Are Maintained in Feeder-Free Culture

We also investigated the expression of transcription factors known to be associated with pluripotent cells. Several transcription factors have been identified in undifferentiated murine ES and murine and human EC cells, which are down-regulated with differentiation, including telomerase reverse transcriptase (hTERT), OCT4, SOX-2, REX-1, and UTF1. In addition, Cripto is expressed in the ICM and trophoblast cells of the mouse blastocyst, mouse ES cells, and in human and mouse EC cells (reviewed in Adamson et al., 2002) as well as in hES cells (Stanton et al., manuscript in preparation). Expression of human TERT (hTERT), OCT3/4, and Cripto were evaluated in the H1, H7, H9, and H14 cell lines using quantitative real-time reverse transcriptase-PCR (RT-PCR) analysis (Fig. 5A). No detectable differences were observed between the cell lines. PCR analysis of SOX-2, REX-1, BCRP in all of the lines also demonstrated similar expression patterns between the different cell lines (Fig. 5B). These findings were further confirmed by the examination of the EST libraries. OCT-4 and SOX-2 were abundantly expressed in the undifferentiated library but were significantly decreased in the EB library (Fig. 5C).

Figure 5.

Comparison of molecular marker expression in H1, H7, H9, and H14 human embryonic stem (hES) cell lines. A: Expression of human telomerase reverse transcriptase (hTERT), OCT 4, and Cripto in hES cells maintained in MEF-conditioned medium with basic fibroblast growth factor (bFGF) on growth factor reduced (GFR) matrigel was measured by TaqMan analysis and showed no significant differences between cell lines. Relative fold difference in the detected signal is expressed on the vertical axis and in the table. Results were normalized to the average signal obtained from all cell lines. Values indicate mean ± SD for three replicate polymerase chain reaction (PCR) runs from individual biological samples. B: All cell lines showed expression of SOX-2, REX-1, BCRP, hTERT, and OCT 4 by reverse transcriptase-PCR. C: Comparison of expressed sequence tag frequencies of OCT 4, Sox-2, and Rex-1 in genomic libraries of undifferentiated hES cell cultures and differentiated embryoid body cultures of hES cells demonstrated greater expression of ES-associated markers in the undifferentiated hES cells. Asterisks indicates significant difference of P ≤ 0.05.

hES Cell Lines Show Little Electrical Activity or Response to Neurotransmitters

In addition to regulation by cytokines it has been suggested that neurotransmitters may regulate proliferation and differentiation of stem cells and precursor cells (Nguyen et al., 2001). It is possible that these agents also will have effects on ES cells. We chose to analyze the effect of glutamate, gamma-aminobutyric acid (GABA), glycine, substance P, acetylcholine (ACh), and ATP on the hES cell cultures. To assess the function of neurotransmitter receptors, the undifferentiated hES cultures were loaded with the Ca2+ indicator dye Fluo-4 am, and the cultures were perfused with various compounds. The undifferentiated H9 and H1 cells showed no response to any of the agonists tested (Fig. 6D). In addition to the undifferentiated cells in the colonies, we evaluated the surrounding differentiated stromal/fibroblast-like cells. Fewer differentiated H1 cells responded to neurotransmitters than H9 differentiated cells (20 vs. 47%). The majority of responsive differentiated H9 cells responded to 500 μM ACh (Fig. 6A,B). A subset of ACh-sensitive cells also responded to ATP (11%), and a small percentage of cells (3.9%) only responded to ATP.

Figure 6.

Undifferentiated H1 and H9 cells do not respond to agonists with increases in intracellular Ca2+, while a subpopulation of differentiated human embryonic stem (hES) cells show robust Ca2+ responses. Various neurotransmitters, peptides, and ions were applied through the bath perfusion. Regions of interest were placed on the images over individual cells and changes in intracellular Ca2+ were measured. A: Response of differentiated H9 cells to application of acetylcholine (ACh). Cells were loaded with the Ca2+ indicator dye Fluo-4 am, and nine images were taken before, during, and after ACh application. B: Responses to 500 μM ACh in 12 differentiated H9 cells. Baseline fluorescence has been subtracted off-line from each trace. C: Whole-cell voltage-clamp experiments assessed ion channel activity in the undifferentiated cells. Shown are data from one cell in response to 80-msec voltage steps from −60 to +60 mV. Top panel shows superimposed raw current traces; middle panel shows peak current–voltage relationship; bottom panel shows conductance–voltage relationship calculated by using an estimated K+ reversal of −81 mV. The red line indicates the Boltzmann fit of the data and yields a conductance at +60 mV of 1.5 nS and a V1/2 of 36 mV. The average peak current at +60 mV is 164 ± 57 pA, conductance is 1.1 ± 0.4 nS, capacitance is 5.2 ± 1.2 pF, and membrane resistance is 1.9 ± 0.5 gigaohms (n = 3, SEM). D: Table shows the number and percentage of H1 and H9 cells that responded to any of the neurotransmitters tested.

None of the cells responded to elevated K+, suggesting that they do not have voltage-gated Ca2+ channels. Patch-clamp studies supported this result, and no inward Ca2+ or Na+ currents were observed in the undifferentiated cells. In some cells, small voltage dependent K+ currents were present (Fig. 6C), but other cells had no currents at all (data not shown). Both H1 and H9 cells had large input resistances (>1 gigaohm), which is consistent with having few ion channels. Based on membrane capacitance measurements, the size of H1 cells were smaller than H9 cells (5 pF vs. 11 pF, respectively). The H1 cells appeared columnar with an average width of ∼5 μm and height of 16 μm (Fig. 7C). It remains to be determined whether these properties change with the stage of confluence or cell cycle.

Figure 7.

Human embryonic stem (hES) cells have gap junctions and express connexin 43. A: Expressed sequence tag frequency counts indicate abundant expression of connexin 43 in undifferentiated hES cells and embryoid bodies. B: Connexin (Cx) 43 and 45 are expressed by all the cell lines tested using reverse transcriptase-polymerase chain reaction. C: Undifferentiated H1 cells appear to have gap junctions. Patch recording pipettes were filled with 0.5% Lucifer Yellow. After electrical recording for 10 to 25 min, cells were injected with dye and fixed for 30 min in 4% paraformaldehyde, rinsed, mounted in Vectashield (Vector Labs), and visualized by using confocal laser scanning microscopy (Zeiss LSM 510) with excitation filter of 488 nm. A Z-stack was made at 0.8-μm intervals, and cell height was consistently 16 μm and width was 5 μm. D,E: hES cells were maintained in MEF-conditioned medium with bFGF on growth factor reduced (GFR) matrigel for seven days, fixed and stained with antibodies against connexin 43 (green), Tra-1-81 (red) and DAPI (4′,6-diamidine-2-phenylidole-dihydrochloride; blue). Scale bars = 50 μm in C, 30 μm in D,E.

hES Cells Possess Gap Junctions and Express High Levels of Connexins

Undifferentiated cells within the colonies appear tightly adhered. We previously have evaluated the expression of integrins on the surface of the cells (Xu et al., 2001). Our sequencing data show that there is abundant transcript expression of connexin 43, a component of gap junctions (reviewed in Rozental et al., 2000), in the undifferentiated cells (Fig. 7A). In addition, RT-PCR analysis also confirmed that connexin RNA, was present in undifferentiated hES cultures (Fig. 7B and data not shown), and we have found connexin 43–positive immunolabeling at the borders of the hES cells within the colonies (Fig. 7D,E). Furthermore, individual cells within H1 or H9 colonies injected with Lucifer Yellow showed dye transfer to surrounding cells (Fig. 7C), suggesting functional gap junctions.

Different hES Cell Lines Demonstrate Similar Gene Expression Patterns

Microarrays were used to characterize gene expression patterns in H1, H7, and H9 cell lines. A total of 12 RNA samples, four from each of three cell lines, were used on two dye-reversed arrays to look for differential expression relative to a common reference sample of the pooled 12 mRNAs. The data was filtered to remove genes with low expression or marginal data and log-transformed ratios were averaged from the two arrays (see Experimental Procedures section). A total of 2,802 cDNA clones, or 66%, passed the filtering criteria and were used for subsequent analysis.

Overall, remarkably few differentially expressed genes were detected among the 12 hES cultures tested. Scatter plots showing the average ratio (log transformed) relative to the reference sample as a function of signal intensity (log (Cy3*Cy5)) clearly demonstrate the similarity of overall gene expression profiles among these cell lines (Fig. 8). No genes showed 10-fold or greater differential expression between the highest and lowest expressing sample. Moreover, only three genes, COL3A1, IL1B, and LEFTB, showed between 5- and 10-fold differential expression between the highest and lowest expressing samples. None of the genes showed differential expression that was cell line-specific.

Figure 8.

Comparison of 12 human embryonic stem (hES) cultures by microarray. Ratio intensity plots are presented for each of 12 RNA samples. Each point represents the average of values from two dye reversed arrays. The plots show differential expression vs. intensity plotted as the log10 ratio of the sample relative to the reference sample reflecting differential expression (Y axis) vs. log10 of the product of Cy3 and Cy5 intensities reflecting overall signal intensity (X axis).

Expression of MHC and ABO Antigens on hES Cells

Undifferentiated hES cells have been shown to express low levels of MHC I and no MHC II (Drukker et al., 2002). The expression of MHC I and II in different types of differentiated cells has not been determined yet. By using DNA sequence analysis, we assessed the HLA typing and blood antigen expression of the different hES cell lines (Table 2). We found that the cell lines represent a range of HLA types and the blood antigen types O, A, and B.

Table 2. HLA and Blood Typing of H1, H7, H9, and H14 hES Cell Linesa
 H1 p31H7 p34H9 p29H14 p22
Allele 1Allele 2Allele 1Allele 2Allele 1Allele 2Allele 1Allele 2
  • a

    hES, human embryonic stem.

ABO blood groupO BOAOO 


Although many hES cell derivations have been performed, there are still few cell lines available for extensive characterization. Therefore, it has been difficult to determine what properties are characteristic of undifferentiated stem cells. Our initial examination showed that cells maintained in feeder-free conditions were apparently identical to their sibling cultures maintained on mouse feeder cells in their telomerase activity, expression of transcription factors and cell surface epitopes, and their ability to differentiate into ectodermal, endodermal, and mesodermal derivatives in vitro and in vivo (Xu et al., 2001). We, therefore, chose to examine hES cell lines maintained in feeder-free conditions as this reduced the variability from contaminating populations of feeder cells that may express some of the genes tested here. For example connexin 43, LIFR, and FGFR are expressed by many embryonic cell types that are present in the feeder cell population. In this report, we have compared the characteristics of four hES cell lines maintained in feeder-free conditions. We have used flow cytometric analysis, immunocytochemistry, TaqMan and quantitative RT-PCR, microarray analysis, electrophysiology, Ca2+ imaging, and allotyping to obtain a profile of the undifferentiated cells. Quantitative comparison of the transcription factors OCT-4, hTERT, and Cripto showed very similar expression between the cell lines. We have found that the expression of surface markers was overall quite similar, although with extensive analysis, we found some small but statistically significant differences in the expression of markers between cell lines. Although some statistically significant differences between cell lines were identified, the means for the different markers were quite similar. Overall, the cultures appeared more similar than different. It should be noted that hES cell cultures represent mixed populations of cells, and it may be extremely difficult to determine differences between cell lines. Therefore, it will require still further analysis to determine whether these differences will correlate with other characteristics of the cells such as the capacity to differentiate.

We have also demonstrated that multicolor flow cytometric analysis allows more sensitive evaluation of the hES cell populations. We found that a significant proportion of hES cells in culture expressed SSEA-4 very brightly (SSEA-4high population). The SSEA-4high population also expressed TRA-1-60, TRA-1-81, CD133, and CD9. This analysis is limited by the lack of directly conjugated reagents. As these reagents become available, a more extensive multicolor coexpression analysis should be undertaken and may reveal further differences between hES cell lines. In addition, isolation of coexpressing subpopulations by sorting will enable better understanding of the significance of variable marker expression on hES cells.

Analyzing a pooled EST database generated from the undifferentiated and differentiated cells and analysis by RT-PCR and flow cytometry provided an explanation for the lack of responsiveness to LIF and the apparent requirement for bFGF in hES cell cultures. We found that FGFR transcripts were quite abundant in the undifferentiated cell population, while CD130 transcripts were not detected. By using flow cytometry, we found expression of FGFR, CD117, and CD135 on undifferentiated cells. However, this analysis of FGFR expression showed relatively dim expression in undifferentiated cells. This finding may reflect specificity of the antibody used or may indicate that bFGF is only one of several factors required to maintain the hES cell in the undifferentiated state. Expression of CD117 and CD135 was limited to the SSEA-4high subpopulation of the cells within the culture. The expression of these receptors may suggest that hematopoietic cytokines are involved in maintenance and/or differentiation of the hES cells. We are currently investigating the effects of the ligands for these receptors, SCF and Flt3L, on undifferentiated hES cells. The lack of detectable expression of CD130 on the SSEA-4high population supports the observation that LIF and other members of the LIF family do not have direct impact on the undifferentiated cells. Indirect effects of LIF upon differentiated components of the hES cell culture, as suggested by the expression of CD130 on the SSEA-4low/neg population, will need further investigation.

In many cases, we found that subpopulations of the undifferentiated hES cells expressed markers and receptors, such as CD133, CD117, and CD135. These data may indicate that the undifferentiated hES cells represent a heterogeneous population of cells in different states of differentiation. Alternatively, this may indicate that the expression of particular markers is a dynamic process; expression patterns may change when cells are in different states of activation or cell cycle. More experimentation will be required to accurately assess these subpopulations.

The similarities between the hES cell lines extended to morphology and the expression of surface integrin receptors (Xu et al., 2001). We therefore examined potential modes of cell–cell communication by assessing neurotransmitter responses and the presence of gap junctions. The cells appear tightly adhered to each other, and we have now demonstrated that the cells have gap junctions by the presence of positive immunoreactivity to connexin 43 and the transfer of Lucifer Yellow dye between cells. This finding was further confirmed by the presence of RT-PCR products for connexin 43 and the presence of connexin 43 ESTs in our database.

Although the cells express integrins and gap junctions, they did not appear to have voltage activated ion channels. In addition, the undifferentiated ES cells did not show responses to neurotransmitters, while the differentiated surrounding cells responded to ACh. These differences allow further delineation of the undifferentiated and differentiated cells in the culture.

The three cell lines analyzed by microarrays showed no consistent differences. These cultures were all at relatively low passage (p28–p37) and contained undifferentiated colonies surrounded by stroma-like cells. When later-passage cultures are also analyzed, some small differences are detected (Rosler et al., 2004). This most likely results directly or indirectly from a loss of stroma-like cells in later-passage cultures.

The data presented here has focused on cell lines derived from one laboratory using similar procedures and cultures conditions. It is important to point out that these culture represent a heterogeneous population of cells. This heterogeneity makes it difficult to quantify the cells using standard growth curve assays. This type of information will be important to collect as cell culture conditions are optimized. By using the current conditions, all four cell lines show remarkably similar characteristics and the results were consistent over multiple passages (Rosler et al., 2004). Our results provide the first detailed comparison of four of the small number of lines currently available for analysis and provide a battery of markers that appear to be consistent over multiple lines. We caution, however, that although the markers we have identified and verified can serve as a useful starting point for assessing hES lines isolated by different groups, it is possible that not all markers will show consistent expression. hES cell lines have been derived all over the world with slightly different source material and at different stages of development, and cells are maintained in culture under varying conditions. It is important to note that strain differences in mouse ES cell lines have been described (Kawase et al., 1994; Schoonjans et al., 2003); interestingly, we have been unable to detect Fox-D3 or genesis expression in the four hES cell lines we have analyzed in contrast to reports from M. Pera that have suggested that this may be an important ES cell marker (M. Rao, unpublished observations). A careful comparison between lines will be necessary to establish which markers are most suitable for evaluating the status of the undifferentiated cells. Furthermore, it will be critical to establish which of these markers is predictive for the pluripotent status of the cells. It is unclear whether different culture conditions will affect the expression of markers or the pluripotency of the cells. To produce standard evaluation procedures and to replicate findings in different laboratories, using different cell lines, it will be critical to compare the fundamental characteristics of the different hES cell lines maintained in different conditions.


Human embryonic stem cell lines H1, H7, H9, and H14 (Thomson et al., 1998) were maintained and passaged under feeder-free conditions (MEF-conditioned medium + 8 ng/mL bFGF) as described (Xu et al., 2001). All determinations were performed at day 7 after passage when the cultures had reached confluence.

Flow Cytometric Analysis of Human ES Cells

Confluent cultures of hES cells were harvested for analysis by using 0.5 mM ethylenediaminetetraacetic acid (EDTA; Sigma, St. Louis, MO) in phosphate buffered saline (PBS). All staining was performed in staining buffer consisting of dPBS, Ca++-, Mg++-free (Gibco, Gaithersburg, MD) supplemented with 2% heat-inactivated (HI) fetal bovine serum (HyClone Laboratories, Inc., Logan, UT), 0.1% sodium azide (Sigma), and 2 mM EDTA (Sigma). Cells were blocked for 15 min at 4°C in 20% HI rabbit serum (Jackson ImmunoResearch Laboratories, Inc., West Grove, PA) in staining buffer, 5 × 105 cells were used per sample. Cells were incubated with appropriate primary antibodies, phycoerythrin (PE) -conjugated antibodies, or appropriate isotype matched controls (Southern Biotechnology Associates, Birmingham, AL, or Sigma or Becton Dickinson, San Jose, CA) for 30 min at 4°C. Primary antibodies used were MC480 (SSEA-1), 1:5; MC813 (SSEA-4), 1:5 (all from Developmental Studies Hybridoma Bank, University of Iowa, Iowa City, IA); Tra-1-60, 1:12, Tra-1-81, 1:20 (a gift from Dr. Peter Andrews, University of Sheffield, UK); and anti-FGFR, 1:25 (primarily recognizes flg gene product, Biogenesis, Brentwood, NH). PE-conjugated antibodies used were CD90-PE, 1:10; CD9-PE, 1:5 (BD PharMingen, San Diego, CA); CD133-PE, 1:10 (Miltenyi Biotec, Auburn, CA); CD135, 1:10 (Caltag Laboratories, Burlingame, CA); CD117, 1:10 (BD PharMingen); and CD130, 1:10 (R&D Systems, Minneapolis, MN). Cells were washed two to three times in staining buffer and incubated for 30 min at 4°C with fluorescein isothiocyanate (FITC) -conjugated goat F(ab′)2 anti-mouse IgG3, 1:100 and PE-conjugated goat F(ab′)2 anti-mouse IgM, 1:100 (Southern Biotechnology Associates) as appropriate. Cells were washed as before and resuspended for analysis in staining buffer containing 1 μg/ml of propidium iodide (PI; Sigma) to identify nonviable cells. Flow cytometric analysis was performed by using a FACSCalibur Flow Cytometer (Becton Dickinson). At least 10,000 PI negative events were collected. Acquired data was analyzed by using CELLQuest software (Becton Dickinson). Statistical analysis was performed by using a two-tailed Student's t-test with Prism software.

Physiological Recordings

Standard whole-cell voltage clamp and Ca2+ imaging techniques were used with the same acquisition protocols and solutions described in previous work (Piper et al., 2001).

cDNA libraries and EST Frequency Analysis

The generation of cDNA libraries for EST sequencing and EST frequency counts are described elsewhere (Brandenberger et al., manuscript in preparation). Briefly, cDNA libraries were generated by conventional means from pooled samples of undifferentiated hES cells (H1 p32, H7 p29, H9 p26) or from differentiated hES cells. The differentiated samples consisted of EBs generated from parallel cultures used for the undifferentiated samples. EBs were maintained in suspension for 4 days before plating onto gelatin-coated wells in differentiation medium for a subsequent 8 days before harvest. Each cDNA library had more than 100,000 primary clones with average insert size greater than 1.4 kb. Partial 5′ end sequences (an EST) were determined for independent clones derived from each cDNA library. Roughly 37,000 sequence reads were obtained from each cell type. Overlapping ESTs were assembled into conjoined sequences and compared with the Unigene database of human genes. ESTs that were at more than 98% identical, over a stretch of at least 150 nucleotides each, to a common reference sequence in Unigene, were assumed to be transcribed from the same gene, and placed into a common assembly. The frequency of ESTs for any particular gene correlates with the abundance of that mRNA in the cells used to generate the cDNA library (Okubo et al., 1992). Thus, a comparison of frequencies of ESTs among the libraries indicates the relative abundance of the associated mRNA in the different cell types. The Fisher exact test was used as a statistical method of comparison to identify only those differences that are likely to be real. The null hypothesis of a gene being equally represented in the undifferentiated hES cells and the EBs is rejected when probability P ≤ 0.05, where 0.05 is the level of statistical certainty. Thus, genes with P ≤ 0.05 are considered to be differentially represented.

Quantitative RT-PCR

Real-time PCR, TaqMan RT-PCR, was performed on the ABI 7700 under the following conditions: 1× RT-Master Mix (ABI), 300 nM for each primer, and 80 nM of probe, and 10 pg to 100 ng of total RNA in nuclease-free water. The reaction was conducted under default RT-PCR conditions of 48°C hold for 30 min, 95°C hold for 10 min, and 40 cycles of 95°C at 15 sec and 60°C hold for 1 min. RNA was isolated by a guanidinium isothiocyanate method according to manufacturer's (RNeasy kit, Qiagen) instructions and subsequently DNAse treated with the DNAfree kit (Ambion). Gene-specific primers and probes were designed by PrimerExpress software (version 1.5, ABI). Probe oligonucleotides were synthesized with the fluorescent indicators 6-carboxyfluorescein (FAM) and 6-carboxy-tetramethylrhodamine (TAMRA) at the 5′ and 3′ ends, respectively. The primers and probe for Oct 4 (GenBank accession no. NM_0020701) were as follows: forward primer (GAAACCCACACTGCAGCAGA), reverse primer (CACATCCTTCTCGAGCCCA), and probe (FAM-CAGCCACATCGCCCAGCAGC-TAM). The primers and probes Cripto (GenBank accession no. NM_003212) were as follows: forward (TGAGCACGATGTGCGC), reverse (TTCTTGGGCAGCCAGGTG), and probe (6 FAM-AGAGAACTGTGGGTCTGTGCCCCATG-TAM). The primers and probe for hTERT purchased from Applied Biosystems (PDAR, predeveloped assay reagent). Relative quantitation of gene expression between multiple samples was achieved by normalization against endogenous 18S ribosomal RNA (Applied Biosystems) by using the ΔΔCT method of quantitation. Fold changes were calculated as 2-ΔΔCT.

PCR Amplification

A 0.5-μl cDNA template was used in a 50-μl reaction volume with the RedTaq DNA polymerase (Sigma). The cycling parameters were 94°C, 1 min; 55°C, 1 min; 72°C, 1 min, for 30 cycles. The PCR cycle was preceded by an initial denaturation of 3 min at 94°C and followed by a final extension of 10 min at 72°C.

Connexin 43 Staining

Cultures of human ES cells were stained live with Tra-1-81 (1:20) or 10 μg/ml of mouse IgM (Southern Biotechnology Associates) in KO DMEM for 1 hr at 37°C. Cells were washed twice with dPBS and incubated with Texas Red–conjugated goat anti-mouse IgM (1:150; Jackson ImmunoResearch Laboratories, Inc.) for 30 min at 37°C. Cells were washed as before and fixed with acetone/methanol (1:1) for 20 min at −20°C. After blocking with 10% normal goat serum (Gibco) for 1 hr at room temperature, cells were incubated with 10 μg/ml of rabbit anti–connexin 43 antibody (Zymed Laboratories, South San Francisco, CA) or rabbit IgG (Chemicon, Temecula, CA) for 1 hr at 37°C. Cells were washed three times with PBS and incubated for 1 hr at 37°C with Biotin-SP–conjugated goat anti-rabbit IgG, 1:150 (Jackson ImmunoResearch Laboratories, Inc.), followed by additional washing and incubation with 20 μg/ml of FITC-conjugated streptavidin (Vector Laboratories, Burlingame, CA). Cells were washed, stained with DAPI (4′,6-diamidine-2-phenylidole-dihydrochloride), and mounted.

Microarray Analysis

The microarray data set analyzed in this manuscript is a subset of the data described elsewhere, and the methods are described in more detail there (Rosler et al., 2004) Briefly, microarrays containing 4,224 cDNA clones purchased from Research Genetics (Huntsville, AL) were prepared by using a GMS 417 Arrayer (Affymetrix, Santa Clara, CA) and GAPS amino silane-coated slides (Corning, Inc., Life Sciences, Acton, MA).

RNA samples were prepared from confluent cultures of hES cells grown for 6 days after subculture using feeder-free conditions as described (Xu et al., 2001). RNA was prepared by using RNeasy Midi kits as described by the manufacture (Qiagen). The reference sample used throughout this data set was a pool of undifferentiated hES RNA.

Probes were prepared from 10 μg of total RNA by direct incorporation of Cy3 or Cy5 dCTP (Amersham Biosciences, Piscataway, NJ) into oligo dT primed first-strand cDNA using SuperscriptII reverse transcriptase, as described by the dye manufacturer. RNA was removed by alkaline hydrolysis, then neutralized probes were purified by using Qiaquick columns as described by the manufacturer (Qiagen, Valencia, CA).

Hybridizations were carried out overnight at 42°C by using ArrayHyb Low Temp Hybridization Buffer (Sigma-Aldrich, St. Louis, MO) with the addition of blocking solution containing 15 μg of CotI DNA, 15 μg polydA, and 20 μg tRNA. The slides were then rinsed in 2× standard saline citrate (SSC), washed two times in 0.1× SSC, 0.1% SDS for 5 min each, two times in 0.1× SSC for 5 min each, and dried by centrifugation.

Arrays were scanned by using a GenePix 4000A microarray scanner (Axon Instruments, Fremont, CA) with 100% scan power and PMT voltages adjusted for each array. PMT voltages ranged from 710 to 850 V for scans at 635 nm and 610 to 680 V for scans at 532 nm. Image processing and data extraction were performed using GenePix Pro 3.0.6 (Axon Instruments).

Quality check calculations were preformed to eliminate signals outside of the linear detection range for each of the 48 individual scans. The intensity value for each spot was assigned a flag value of 1 if either less than 80% of measured pixels were above two times the standard deviation of the measured local background or more than 5% of the pixels were saturated, otherwise the flag value was 0.

Expression ratios were calculated for a spot on an array if a signal was detected in either the Cy3 or Cy5 channel. The ratio was calculated by subtracting the median local background from the mean pixel intensity of each feature then dividing the sample value by the reference value. The expression ratio was then balanced by multiplying the calculated expression ratio by the ratio of means for all data points on the same array. Log transformed, balanced ratios of background-subtracted intensities were used for subsequent calculations.

For each experimental RNA sample, duplicate arrays were run using dye reversal. Average log expression ratios were calculated from dye-reversed replicates when at least two of four measurements were not flagged, otherwise, the expression ratio was left blank. A measure of signal intensity was calculated by taking the average log transformed product of Cy3 intensity and Cy5 intensity from the two arrays. Data from the 12 experimental samples were then pooled and filtered to remove cDNAs in which any of the 12 expression ratios was blank or the sum of the 48 possible flags was greater than 5.


We thank John Irving and Michael Mok for technical assistance. We gratefully acknowledge the input of all members of our laboratory provided through discussions and constructive criticisms. Mahendra S. Rao was a consultant for Geron on this project and work constitutes approved outside activity. SSEA-1 and -4 antibodies were developed by D. Solter and B.B. Knowles and were obtained from the Developmental Studies Hybridoma Bank developed under the auspices of the NICHD and maintained by The University of Iowa, Department of Biological Sciences, Iowa City, IA, 52242.