Derivation, Characterization, and Differentiation of Human Embryonic Stem Cells



The derivation of human embryonic stem (hES) cells establishes a new avenue to approach many issues in human biology and medicine for the first time. To meet the increased demand for characterized hES cell lines, we present the derivation and characterization of six hES cell lines. In addition to the previously described immunosurgery procedure, we were able to propagate the inner cell mass and establish hES cell lines from pronase-treated and hatched blastocysts. The cell lines were extensively characterized by expression analysis of markers characteristic for undifferentiated and differentiated hES cells, karyotyping, telomerase activity measurement, and pluripotency assays in vitro and in vivo. Whereas three of the cell lines expressed all the characteristics of undifferentiated pluripotent hES cells, one cell line carried a chromosome 13 trisomy while maintaining an undifferentiated pluripotent state, and two cell lines, one of which carried a triploid karyotype, exhibited limited pluripotency in vivo. Furthermore, we clonally derived one cell line, which could be propagated in an undifferentiated pluripotent state.


The inner cell mass (ICM) of the preimplantation blastocyst contains a core of cells, termed the epiblast, that have the potential to generate somatic and germ cells of the embryo. It has previously been demonstrated that human embryonic stem (hES) cell lines, exhibiting a stable developmental potential to form derivatives of the three germ layers after prolonged culture in vitro, can be generated by the isolation and culturing of the human ICM [1, 2]. So far, the available knowledge of conditions for deriving, characterizing, and culturing undifferentiated hES cells is largely based on a relatively few successfully isolated cell lines. Furthermore, the underlying mechanisms that control the developmental decisions of hES cells in culture remain essentially unknown. It has been suggested that the poor success rates in developing ES cells in species other than mice is either due to fundamental biological differences between the species or simply due to technical factors associated with the derivation and culture conditions [3, 4]. For example, unless spontaneous differentiation of hES cells is prevented, the cells become gradually restricted and lose the characteristics of ES cells. Frequent passaging and optimization of the quality of the culture conditions and feeder cells may overcome this problem. However, many problems remain with regard to the optimal maintenance and expansion of hES cells.

Due to the limited number of available hES cell lines, there is an urgent need for the generation and characterization of more cell lines, as each line may have its own characteristics and advantages for different applications. Furthermore, the availability of more hES cell lines for comparison will aid in defining criteria for bona fide hES cells and the establishment of appropriate and robust methods for maintenance and expansion of hES cells.

Here, we describe the successful establishment of hES cell lines from the ICM by immunosurgery, from spontaneously hatched blastocysts, and from blastocysts after pronase-mediated removal of the zona pellucida. Three of the hES cell lines have been maintained in culture for more than 1 year, during which time high levels of telomerase activity, stable karyotype, and expression of markers characteristic for undifferentiated hES cells were maintained. The cells could be cryopreserved by vitrification without any effect on their ability to re-establish pluripotent hES cell colonies. The pluripotent qualities of these cell lines were demonstrated in several ways. Most importantly, the cells were able to differentiate into cell types originating from each of the three embryonic germ layers (endoderm, mesoderm, and ectoderm) in vitro as well as in vivo. In addition, we subcloned one of our cell lines and showed that it could be propagated in an undifferentiated state while maintaining its pluripotency both in vitro and in vivo, as was previously shown for other hES cell lines [5].

Materials and Methods

Establishment and Culture of Human Embryonic Stem Cell Lines

Surplus human embryos from clinical in vitro fertilization (IVF) treatment were donated after informed consent and approval of the local ethics committees at Göteborg University and Uppsala University. Donated embryos were cultured to blastocysts until the age of 6–7 days as previously described [6, 7]. Blastocysts were graded (grade shown within brackets) according to Dokras et al. [8] and randomly selected for either pronase treatment (lines Fertilitetscentrum [FC]018 [B], Akademiska sjukhuset [AS]034 [A], and AS038 [B]) or pronase treatment followed by immunosurgery (lines Sahlgrenska [SA]121 [A], SA181 [B]) (see below). Spontaneously hatched blastocysts (line SA002 [B]) were placed directly in VitroHES supplemented with 4 ng/ml human recombinant basic fibroblast growth factor (hrbFGF) (GIBCO/Invitrogen; Paisley, UK;, and 125 μg/ml hyaluronic acid (HA) (Ophthalin, CIBA Vision Nordic AB; Askim, Sweden; on a layer of mitotically inactivated early passage mouse embryonic fibroblasts (MEFs) (F1 hybrid CD1xC57BL/6, Charles River Laboratories; Sulzfeld, Germany; VitroHES was composed as previously described [5] and manufactured by Vitrolife AB (Kungs-backa, Sweden; where it was subjected to physical and functional tests as part of a quality control program to increase the final medium quality. The MEF cells were derived and cultured as previously described [9]. Briefly, the mitotic activity of the MEFs was abolished by an incubation with 10 μg/ml mitomycin C (Sigma-Aldrich Sweden AB; Stockholm, Sweden; for 3 hours at 37°C, after which the cells were seeded at a density of 130,000 cells/ml in IVF cell culture dishes (Falcon 3653, Becton Dickinson; Franklin Lakes, NJ; in MEF medium (Dulbecco's-modified Eagle's medium [D-MEM]) supplemented with 10% fetal calf serum (FCS), 100 U/ml penicillin G, and 1× Glutamax (Invitrogen, Sigma-Aldrich). Prior to the addition of treated blastocysts or hES cells, the MEF medium was changed to VitroHES.

Blastocysts with intact zona were treated in pronase for 1–3 minutes (Sigma-Aldrich:10 U/ml in ICM-2 [Vitrolife AB] 1–3 minutes in three subsequent drops), washed three times in ICM-2 and plated on MEFs in hES medium supplemented with 125 mg/ml HA and 4 ng/ml hrbFGF. ICM-2 is a blastocyst culture medium containing recombinant human albumin and HA. Pronase-treated zona-free blastocysts selected for immunosurgery were washed twice in antihuman serum antibody (Sigma-Aldrich, 1:5 in ICM-2). After the second wash the blastocysts were placed in a new drop of the antibody and incubated for 30 minutes. The blastocysts were then extensively washed three times in ICM-2 medium followed by three washes in guinea-pig complement serum (Invitrogen, 1:5 in ICM-2). The blastocysts were then incubated for 10 minutes in guinea-pig complement serum and followed by three washes in ICM-2 and placed on MEFs in VitroHES supplemented with 125 μg/ml HA and 4 ng/ml hrbFGF.

ICM outgrowths were passaged to plates with fresh medium and MEF cells by mechanical dissection using Stem Cell Tool (Swemed Lab International AB; Billdal, Sweden). Established hES cell lines were routinely passaged every 4–5 days. The hES cell colonies were mechanically cut into pieces, 200 × 200 μm, and removed from the culture dish and transferred to a new culture dish with fresh MEF cells and VitroHES supplemented with 4 ng/ml hrbFGF.

Telomerase Activity

It has been demonstrated that high telomerase activity in ES cells correlates with their ability to divide indefinitely in culture [10]. Degradation of telomeric sequences or end-to-end fusion of chromosomes can lead to genomic instability [11]. To analyze telomerase activity in hES cells a polymerase chain reaction-based enzyme-linked immunosorbent assay was used (Roche Diagnostics GmbH; Mannheim, Germany; The cells were harvested and lysed according to the manufacturer's instructions.

Karyotyping and FISH

hES cells designated for karyotyping were cultured in hES medium supplemented with 0.1 μg/ml colcemid (Invitrogen) for 1–3 hours. The cells were subsequently trypsinized, fixed, and mounted on glass slides. The chromosomes were visualized by using modified Wright's staining. For fluorescence in situ hybridization (FISH) analysis, a commercially available kit containing probes for chromosome 13, 18, and 21 and the sex chromosomes (X and Y) was used (MultiVysion PB Multicolour Probe Panel; Vysis, Inc.; Downers Grove, IL; according to the manufacturer's instructions. For each cell line at least 200 nuclei were analyzed. The slides were analyzed in a fluorescence microscope equipped with appropriate filters and software (CytoVision; Applied Imaging; Santa Clara, CA;

In Vitro Differentiation

The cells were kept on MEFs without passaging up to 14 days. Medium (hES medium without hrbFGF) was changed every second day. Alternatively, incubating clumps of hES cells in VitroHES without hrbFGF in suspension culture for 4–9 days generated both simple and cystic embryoid bodies (EBs).

Xenografting of hES Cells

Severe combined immunodeficient (SCID) mice, [12] (C.B-17/lcrCrl-scidBR; Charles River Laboratories) were used as animal hosts for the xenografted hES cells. Four- to five-week-old animals were anesthetized with intraperitoneal injections of ketamine hydrochloride (Ketalar; Warner Lambert Nordic AB; Solna, Sweden;; 75 μg/g mouse) and medetomidine hydrochloride (Domitor; Orion Pharma Corporation; Espoo, Finland;; 1 μg/g mouse). hES cell colonies were mechanically cut into 200 × 200-μm pieces, washed once in phosphate-buffered saline (PBS) containing 4 mg/ml human serum albumin and penicillin (Cryo-PBS; Vitrolife AB), and 20 cell clusters were injected under the kidney capsule or in the testicular lumen using a 200-μm lumen glass transfer pipette (Swemed Lab International AB). The number of cells transferred was approximately 20,000 to 40,000 per organ. Control animals were either injected with Cryo-PBS or grafted with primary brain cells from a littermate. The mice were resuscitated with intraperitoneal injections of atipamezol (Antisedan; Orion Pharma; 1 μg/g mouse), and kept on a heated pad until consciousness. Palpable tumors started to appear 3 weeks after transplantation. The tumors were allowed to develop for 8 weeks before the animals were sacrificed by cervical dislocation. All animals appeared healthy during the 8-week period, and no animal died due to illness. The tumors were excised and immediately fixed in 4% paraformaldehyde, incubated for 24 hours, washed or stored in 70% ethanol, and processed in a Tissue-Tek paraffin infiltrator (Sakura Fine Technical; Tokyo, Japan;, and paraffin-embedded. The tumors were subsequently sectioned in 6–8 μm sections with a Microm HM 360 (Microm GmbH; Walldorf, Germany). Samples were stained for morphology using hematoxylin and eosin with a Sakura DRS-601 stainer. To determine the origin (human, mouse) and tissue type within teratomas, we combined histopathological and marker expression analysis. To evaluate whether the tissues were of human origin we used several human-specific antibodies, e.g., anti-E-cadherin and anti-human nuclei (see below). To confirm the presence of tissues derived from all three germ layers, we focused on tissues that can be easily distinguished by histopathological methods, e.g., neuroectoderm, cartilage, kidney tubuli, and gut-like epithelium. To strengthen these conclusions, we also used antibodies against markers characteristic for derivatives of the germ layers, e.g., α-smooth muscle actin, desmin, nestin, β-III-tubulin, α-fetoprotein, and HNF3β.


The cells were washed inside the wells twice with cloning medium, 150 μl, 500 μl, and 1,000 μl for the 96-, 48-, 24-well plates, respectively. The inner part of the colonies was cut with a 300-μm Stem Cell Tool and subsequently incubated with 0.5 mM EDTA for 20 minutes at 37°C. The cells were triturated carefully with a pipette and diluted either in knockout (KO)-DMEM medium (GIBCO) supplemented with 15% concentrated conditioned medium, 3.5 mM glucose, 1 mM Glutamax (invitrogen-sigma-albumin), 1% NEAA (GIBCO), and 4 ng/ml bFGF, KO-DMEM medium supplemented with 15% FCS, 3.5 mM glucose, 1 mM Glutamax, 1% NEAA, and 4 ng/ml bFGF, or KO-DMEM medium supplemented with 20% serum replacement (GIBCO) SR, 3.5 mM glucose, 1 mM Glutamax, 1% NEAA, and 4 ng/ml bFGF. Single cells were picked and put into individual wells with MEF-coated plates. To confirm the colony-forming ability of the cells, positive controls were performed (e.g., 10, 100, 1,000 cells/ml or smaller clusters) as well as negative controls (wells without dissociated hES cells). Subclones were preferentially obtained in 15% concentrated conditioned medium of hES cells grown in presence of FCS.

Histochemical Staining for Alkaline Phosphatase

Histological staining for alkaline phosphatase was carried out using a commercially available kit (Sigma-Aldrich) following the manufacturer's instructions.


The cells were fixed in 4 % paraformaldehyde for 15 minutes at room temperature, washed in PBS, and exposed to the primary antibodies overnight at 4°C. As secondary antibodies, we used fluorescein isothiocyanate (FITC)- and Cy-3-conjugated antibodies (1:50, Southern Biotech; Birmingham, AL; The monoclonal antibodies (mAb) directed against SSEA-1, SSEA-3, and SSEA-4 (Developmental Studies Hybridoma Bank, The University of Iowa; Iowa City, IA) were used at 1:200, whereas the TRA-1-60 and TRA-1-81 mAbs (Santa Cruz Biotechnology; Santa Cruz, CA; were used at 1:500. Neuroectodermal precursor cells and neurons were detected by a nestin mAb (BD Biosciences; Stockholm, Sweden; 1:100), and β-tubulin-III mAb (Sigma-Aldrich; 1:100), respectively. Endodermal cells were recognized by a mAb against Cdx2 (gut endoderm, visceral endoderm) (BioGenex, Nordic BioSite; Täby, Sweden; 1:200;, and polyclonal antibodies against α-1-fetoprotein ([AFP] Sigma-Aldrich; 1:2000) and HNF3β (Santa Cruz Biotechnology; 1:500). Mesodermal cells were detected by a desmin antibody (Chemicon; Temecula, CA;; 1:200). hES cell-derived cells were detected by mAbs against human nuclei (Chemicon; 1:100) and human E-cadherin (Zymed Laboratories; South San Francisco, CA;; 1:500). Some cultures were double stained with DAPI (4′-6′Diamidino-2-phenylindole, final concentration 0.1 μg/ml, Sigma-Aldrich) for 5 minutes.


Methods for Deriving hES Cell Lines

The most commonly used method for deriving hES cell lines is by immunosurgical isolation of the ICM from the human blastocyst [1, 2, 13]. Here, we show that hES cell lines can be established from pronase-treated and hatched blastocysts as well (Fig. 1). Due to the limited number of embryos used in this study, we can, however, not conclude anything about the relative efficiencies of the different methods. Whereas cell line SA002 was derived from a spontaneously hatched blastocyst, cell lines FC018, AS034, and AS038 were established from pronase-treated blastocysts (Table 1). Immunosurgery was used for establishing cell lines SA121 and SA181. One to two weeks after plating, the expanded ICM was transferred to a fresh MEF-coated IVF-cell culture dish by mechanical dissection. Successful propagation of the ICM was associated with the appearance of ES-like cells in the outgrowth, whereas differentiated cells, presumably representing primitive endoderm and trophectoderm, either died or disappeared upon repeated passaging (Fig. 1).

Table Table 1.. Summary of hES cell characterization
  1. a

    *NIH-eligible cell line

  2. b

    n. d. = not determined

hES lineIsolation-methodSSEA-1SSEA-3SSEA-4TRA-1-60TRA-1-81ALPOct-4
*SA002Spontaneous hatching++++++
hES lineKaryotypeTelomeraseFISH In vitro differentiation Teratoma 
*SA00247 XX+++13 XX Three germ layers Three germ layers 
FC01869 XXY++Trip. XXY Three germ layers Fluid-filled cysts 
AS03446 XY++2n XY Three germ layers Three germ layers 
AS03846 XY+2n XY Three germ layers Fluid-filled cysts 
SA12146 XY++2n XY Three germ layers Three germ layers 
SA18146 XY++2n XY Three germ layers Three germ layers 
AS034.146 XYn.d.2n XY Three germ layers Three germ layers 
Figure Figure 1..

Human ES cell derivation from pronase-treated and hatched blastocysts.A) Blastocyst before pronase treatment. B) Outgrowth of pronase-treated blastocyst shown in (A) 6 days after pronase treatment. C) hES colony (passage 6) derived from the ICM in (B). D) Spontaneously hatched blastocyst. E) Outgrowth of the blastocyst shown in (D) 5 days after plating. F) hES colony (passage 3) derived from the ICM in (E).

Characterization of hES Cell Lines

To analyze the long-term pluripotency and replicative immortality of the six newly established hES cell lines SA002, FC018, AS034, AS038, SA121, SA181, and subclone AS034.1 (see below), we used previously defined criteria for in vitro and in vivo characterization of hES cells by examining the morphology, marker expression, telomerase activity, karyotype, and pluripotency in vitro and in vivo.

In the presence of mouse feeders and human recombinant bFGF, all six hES cell lines gave rise to large compact multicellular colonies of cells with the characteristic hES cell morphology, i.e., a high ratio of nucleus to cytoplasm and prominent nucleoli. Some of these lines have been passaged more than 120 times. Initially, the different cell lines could not be discriminated morphologically from each other except for cell line AS038, which never developed a clear distinguishable border towards the mouse feeders (data not shown). However, with time the morphology of AS038 became indistinguishable from the other cell lines.

Karyotype analyses carried out at different passages (from passage three to passage 76) indicated a normal stable karyotype in four of the cell lines (Fig. 2A, Table 1). In two of the cell lines chromosomal aberrations were apparent, trisomy 13 in cell line SA002 (Fig. 2B), and triploid karyotype in FC018 (Table 1).

Figure Figure 2..

Karyotype of hES lines (see also Table1). A) Normal 46XY karyotype from cell line SA181. B) XX karyotype from line SA002 with a trisomy 13. The arrow in (B) points to the trisomic chromosome 13.

Similar to undifferentiated pluripotent cultures of human germ cells [14] and previously established hES cell lines, all our cell lines possessed high levels of alkaline phosphatase (AP) activity (Fig. 3B, Table 1). The overall percentage of visible AP-positive cells within a colony varied from 60% to 90% in all cell lines. The hES cells lines were further characterized by expression analysis of five cell surface markers: SSEA-1, SSEA-3, SSEA-4, TRA-1-60, and TRA-1-81, and the intermediate filament protein nestin [1, 1417]. Figure 3 depicts examples of the expression of these markers in undifferentiated colonies from cell line SA002. The results of the expression analysis are summarized in Table 1. Whereas all cell lines expressed SSEA-3, SSEA-4, TRA-1-60, and TRA-1-81, they were, with a few exceptions (see below), negative for nestin and SSEA-1.

Figure Figure 3..

Immunohistochemical marker expression analysis of cell line SA002 (passage 21).The colonies were analyzed 5 days after passaging. A) Morphology, (B) Alkaline Phosphatase (AP), (C) SSEA-1, (D) SSEA-3, (E) SSEA-4, (F) TRA-1-81, (G) TRA-1-60, and (H) Oct-4. Scale bar: 25 μm.

Oct-4 is a POU-domain transcription factor that is essential for establishment of ES cells from the ICM [18]. Similar to previous reports [19, 20], we found that undifferentiated hES cells expressed Oct-4 (Fig. 3H, Table 1), whereas Oct-4 was downregulated concomitant with differentiation (data not shown).

Nestin is expressed in a variety of stem/precursor cell populations of neuroectodermal and mesodermal origin, and we found it useful for detecting differentiated hES cells that could not be recognized by morphological criteria. Generally, nestin-positive colonies often appeared during suboptimal growth conditions and seemed to result in irreversible commitment. Interestingly, whereas TRA-1-60 and TRA-1-81 were initially expressed normally in cell lines FC018 (passage 25) and AS038 (passage 41) (Fig. 4C and 4D, Table 1), a patchy expression pattern of SSEA-3 and SSEA-4 was observed (Fig. 4F, Table 1). Moreover, additional signs of cell differentiation were the appearance of SSEA-1- and nestin-expressing cells (Fig. 4B and 4E, Table 1). However, with time both line FC018 (passage 118) and AS038 (passage 72) expressed SSEA-4 and Oct-4 uniformly within the colonies (Fig. 4G and 4I), whereas nestin was no longer expressed (Fig. 4H).

Figure Figure 4..

Immunohistochemical marker expression analysis of cell line FC018.The colonies were analyzed 5 days after passaging. A-F: passage 25, G-I: passage 118. (A) Morphology, (B) SSEA-1, (C) TRA-1-60, (D) TRA-1-81, (E, H) nestin, (F, G) SSEA-4, and (I) Oct-4. Scale bar: 25 μm.

Finally, all six hES cell lines expressed high levels of telomerase activity that were maintained even after more than 50 passages (Table 1).

In summary, three of the six characterized hES cell lines, AS034, SA121, and SA181, exhibited the morphology, genotype, telomerase activity, and marker expression characteristics for previously reported pluripotent stem cell lines with a normal karyotype [1, 2, 13].

Analysis of In Vitro Pluripotency of hES Cell Lines

Similar to mouse ES cells, hES cells spontaneously form three-dimensional aggregates of differentiated cells known as EBs when grown in suspension. Upon continued in vitro culture of EBs, a variety of ectodermal, endodermal, and mesodermal germ layer derivatives, such as hematopoietic, endothelial, cardiac, skeletal muscle, and neuronal cell lineages appear [21]. We could show that all hES cell lines are capable of generating both simple and cystic EBs. Marker expression analysis and morphological examination of plated EBs revealed derivatives of all three germ layers, including areas of beating heart muscle-like cells (data not shown). However, EB formation is not an exclusive pathway for initiating hES cell differentiation. An alternative efficient and timesaving method to induce spontaneous differentiation of hES cells is simply by keeping the colonies on mouse feeders for more than 7 days without passaging. Similar to EB formation, this method gives rise to a variety of cell types derived from all three germ layers. The vast majority of cells within the differentiated colonies expressed neuroectodermal cell markers, such as nestin and β-III-tubulin (Fig. 5A and 5B). These markers were preferentially expressed within typical rosette-like structures during early stages of differentiation (data not shown). Derivatives of mesoderm were confirmed by desmin stainings, and the appearance of synchronously beating cardiomyocyte-like cells (Fig. 5C, data not shown). Endodermal derivatives appeared later during differentiation in the periphery of the colonies and were identified by the expression of AFP, Cdx2, and HNF3β (Fig. 5D–5FFigure 5.). In summary, based on in vitro differentiation, all cell lines displayed the potential to form derivatives of all three embryonic germ layers. Importantly, these characteristics remained the same after repeated freezing-thawing cycles (data not shown).

Figure Figure 5..

In vitro differentiation of cell line SA002.The colonies were analyzed after 12 days after passaging. A) Nestin-positive neuronal precursors, B)β-III-tubulin-positive postmitotic neurons, C) desmin-positive mesodermal cells, D, E, F)α-fetoprotein-positive (D), HNF3β-positive (E), and Cdx2-positive (F) endodermal cells. Scale bar: 25 μm.

Since none of the described hES cell lines were clonally derived, it cannot be excluded that multiple precursor or stem cells committed to different lineages may coexist within a population of homogeneously appearing cells. Theoretically, this would imply that, until proven, it cannot be stated that a single hES cell is capable of forming derivatives of all three embryonic germ layers. In our initial efforts to subclone cell line AS034, we obtained one clone, AS034.1. To promote cell survival, we used concentrated conditioned medium from hES cells grown in presence of FCS as cloning medium (Materials and Methods). However, the overall yield was low; on average from approximately 103 dissociated single cells one colony resulted. Characterization of subclone AS034.1 revealed that it behaved comparably to the other cell lines in terms of the expression of SSEA-4, SSEA-3, TRA-1-60, and TRA-1-81 (Fig. 6). Importantly, SSEA-1 and nestin were not detected in undifferentiated colonies (Fig. 6B, data not shown). Furthermore, the subclone was capable of differentiating into ectodermal, mesodermal, and endodermal cell types both in vitro and in vivo (Table 1).

Figure Figure 6..

Immunohistochemical marker expression analysis of subclone AS34.1.The colonies were analyzed 3 days after passaging. A-F: passage 12, G: passage 34. A) Morpholgy; B) SSEA-1; C) SSEA-3; D) SSEA-4; E) TRA-1-60; F) TRA-1-81, and G) Oct-4. Scale bar: 25 μm.

Analysis of In Vivo Pluripotency of hES Cell Lines

When clusters of hES colonies are xenotransplanted to SCID mice, they form teratomas consisting of cell types derived from ectoderm, mesoderm, and endoderm. To analyze the in vivo pluripotency of the derived hES cell lines, we transplanted clusters of hES cells under the kidney capsule of SCID mice. Eight weeks after the transplantation the mice were sacrificed and teratomas were analyzed. To distinguish human cells from mouse cells, we used antibodies specific for human-specific nuclear antigen [22] and human E-cadherin. Xenografting of hES cluster resulted in two morphologically distinct structures in the kidneys. From a majority of the cell lines (SA002, AS034, AS034.1, SA121, and SA181) solid teratomas, consisting of highly differentiated cells and tissues derived from all three germ layers, such as gut epithelium; glandular epithelium (endoderm); cartilage, bone, smooth muscle, striated muscle, and kidney glomeruli-like structures (mesoderm); and pigment epithelial cells, neural epithelium, hair follicles, and stratified squamous epithelium (ectoderm) formed (Fig. 7, Table 1, data not shown). However, two of the cell lines (FC018, AS038) consistently formed fluid-filled cyst-like structures composed of hES cell-derived connective tissue and epithelial cells (Fig. 7B, Table 1). Infrequently, these cysts contained small solid teratomas consisting of cell types derived from several germ layers (data not shown). Thus, in summary, whereas all cell lines, including lines with an abnormal karyotype, exhibited pluripotent differentiation qualities in vitro, pluripotency in vivo was only consistently observed in lines SA002, AS034, AS034.1, SA121, and SA181

Figure Figure 7..

Teratoma analysis.A, B) Low-power view of a solid tumor from cell line SA121 (A) and a fluid-filled cyst generated from cell line FC018 (B). C, D) Derivatives of the ectoderm. Pigmented epithelium (arrowhead) and neuroepithelium (arrow) are shown in C and hair follicles (arrow) in D. E, F) Derivatives of the mesoderm. Cartilage (arrowhead) and bone (arrow) are shown in E, and skeletal muscle in F. G, H) Derivatives of the endoderm. Gut-like epithelium with mucous-containing cells (arrowhead) and glandular epithelium are shown in G and H, respectively. Scale bars: 100 μm.


Here we report the derivation and characterization of six new hES cell lines. With a few exceptions the cell lines behaved like previously reported blastocyst-derived pluripotent stem cell lines. We believe that the phenotypic discordance in vitro and in vivo of some of the characterized cell lines is of value both for evaluating present characterization tools and for their further improvement in order to set up robust criteria for analyzing hES cells.

In addition to the previously reported immunosurgery protocol [1, 23], we report the successful derivation of hES cell lines from a spontaneously hatched blastocyst, and from blastocysts after enzymatic removal of the zona pellucida by pronase. A possible drawback of only removing the zona is overgrowth by the trophectoderm and the possible generation of trophectoderm stem cell lines. However, this should also be taken into account when using immunosurgery since this method does not guarantee complete removal of trophectoderm cells. Nevertheless, due to the morphology of trophoectoderm cells, i.e., flattened and polarized as they mature [24], they can be discriminated from the ICM once plated on MEFs. The fact that all of our cell lines expressed Oct-4, which normally is downregulated during differentiation and in trophectoderm stem cells [18, 19], substantiates our conclusion that the derived pluripotent stem cells were derived from the ICM.

Although four of our cell lines appeared phenotypically similar to previously derived hES cell lines [1, 2, 13], two of the cell lines (FC018 and AS038) were phenotypically different. Instead of generating solid teratomas upon xenografting, they developed into cyst-like structures. In contrast, when differentiated in vitro these lines generated cell types from all three germ layers. Except for their morphology in culture, which was similar to the “normal” hES cell lines, FC018 and AS038 also exhibited aberrant expression patterns of some of the markers characteristic for undifferentiated and differentiated hES cells, respectively. However, with time, the expression of these markers became reminiscent of normal hES cells. Whether the temporal shift in marker expression pattern in these cell lines can be explained by an initial mixed population of cells which went through selection, or whether cells were more prone to spontaneously differentiate initially, is presently unclear.

Our results indicate that adding early differentiation markers, such as nestin, to the list of markers whose expression patterns are analyzed in undifferentiated hES cells improves the detection of early hES cell differentiation. We show that the expression of nestin, a marker for stem/precursor cells of neuroectodermal and mesodermal origin, precede visible morphological signs of differentiation. We also found that the frequent spontaneous commitment to the neuroectodermal pathway of cell lines FC018 and AS038 correlate with loss of developmental potential to form derivatives of the three germ layers in vivo.

By definition, clonal expansion of hES cells is a prerequisite for the strict definition of pluripotent cell lines. Currently, the culture conditions for clonal expansion of hES cells are suboptimal. Unlike mouse ES cells, hES cells die at a high rate when they are dissociated into single cells [5]. Depending on which cell line was used, only 0.1%-1% of plated single cells was able to generate colonies that could be propagated. Among the few clones that survived, the majority were lost due to irreversible differentiation. We found that concentrated conditioned medium from hES cells grown in the presence of FCS promoted cell survival and maintenance of an undifferentiated fate. In general, our experience is that culture conditions that may be rate limiting for maintaining undifferentiated growth of hES cells include MEF quality and density, changes in the osmolarity, pH, and temperature of the medium, as well as the presence of supplements, such as β-mercaptoethanol.

The stable maintenance of diploid chromosome number in the majority of our cell lines indicates that our cell lines maintained a stable karyotype in vitro after extensive passaging and repeated freezing/thawing cycles. The fact that two of the stem cell lines were chromosomally abnormal is not surprising taking into account that Hardarson et al. [25] recently found that only 42% of surplus IVF embryos were chromosomally normal at the blastocyst stage.

Consistent with previous reports, xenografting of most of our cell lines (SA002, AS034, AS034.1, SA118, and SA121) generated solid teratomas consisting of endodermal, mesodermal, and ectodermal cell type derivatives [1, 2, 5]. However, some of the cell lines (FC018, AS038) preferentially generated fluid-filled cyst-like structures of human origin. We have yet to determine the cellular identity and origin of these cysts. Potentially, these cells may be of trophoectoderm origin, since it was recently shown that hES cells are capable of differentiating into extraembryonic derivatives, such as trophoblast cells [26]. Alternatively, FC018 and AS038 may represent primitive endoderm lineages. However, the facts that undifferentiated colonies of these cell lines expressed significant levels of Oct-4, and that they lacked expression of primitive endoderm markers, such as Gata4 and AFP [27, 28], suggest that it is unlikely that they are of extraembryonic or primitive endoderm origin (data not shown).

Needless to say, more knowledge is needed about basic hES cell biology, such as the regulatory pathways that govern self-renewal and differentiation, before it will be possible to appreciate the potential applications of hES cells in basic science and cell-replacement therapy.


We acknowledge Katarina Andersson, Karin Axelsson, and Angelica Niklasson for assistance with derivation, characterization, and propagation of hES cells; Anita Sjögren and Thorir Hardarson for culture of human embryos; Gunilla Caisander for karyotyping and FISH analysis; Inger Bryman and staff at the IVF clinic at the Sahlgrenska University Hospital; Monalill Lundqvist at IVF clinic at the Uppsala University Hospital; and Matts Wikland at Fertility Center Scandinavia for supplying embryos. We thank Marie Rehnström and Ulrika Karlsson for assistance in subcloning; Gabriella Brolén for support in stainings and cell culture; and Peter Sartipy for general assistance. The work was supported by grants from the Cell Therapeutics Scandinavia AB, Swedish Research Council (H.S.), Juvenile Diabetes Research Foundation (H.S.), and Inga Britt och Arne Lundbergs Forskningsstiftelse (H.S.).