Nature Biotechnology 25, 207 - 215 (2007)
Published online: 7 February 2007 | doi:10.1038/nbt1285

Adaptation to culture of human embryonic stem cells and oncogenesis in vivo

Duncan E C Baker1,4, Neil J Harrison2,4, Edna Maltby1, Kath Smith1, Harry D Moore2, Pamela J Shaw3, Paul R Heath3, Hazel Holden3 & Peter W Andrews2

The application of human embryonic stem cells (HESCs) to provide differentiated cells for regenerative medicine will require the continuous maintenance of the undifferentiated stem cells for long periods in culture. However, chromosomal stability during extended passaging cannot be guaranteed, as recent cytogenetic studies of HESCs have shown karyotypic aberrations. The observed karyotypic aberrations probably reflect the progressive adaptation of self-renewing cells to their culture conditions. Genetic change that increases the capacity of cells to proliferate has obvious parallels with malignant transformation, and we propose that the changes observed in HESCs in culture reflect tumorigenic events that occur in vivo, particularly in testicular germ cell tumors. Further supporting a link between culture adaptation and malignancy, we have observed the formation of a chromosomal homogeneous staining region in one HESC line, a genetic feature almost a hallmark of cancer cells. Identifying the genes critical for culture adaptation may thus reveal key players for both stem cell maintenance in vitro and germ cell tumorigenesis in vivo.

The inner cell mass (ICM) of the mammalian embryo can give rise to all somatic cell types. HESC lines are derived from the ICM, and these cells seem to maintain in vitro the same pluripotent properties as ICM cells in vivo. However, these two systems are not equivalent, since the moment that HESCs are transferred from the embryo to the culture dish they are subject to selective pressures from their new environment. Indeed, a recent report comparing epigenetic control in murine ICM and ES cells noted differences in histone methylation and acetylation1, perhaps indicating a change in cell regulation in culture. The facility for indefinite self-renewal is a key feature for HESCs, yet this is not a property of ICM cells in vivo and must be a characteristic selected during initial outgrowth in culture. As such, the establishment of an HESC line must involve some form of adaptive process, most likely epigenetic in origin as, at least in the mouse, established ES cells can revert to an ICM-like state when replaced in a blastocyst2.

On initial derivation HESCs have shown diploid karyotypes, which may remain stable for extended periods3. However, several reports indicate that they may acquire chromosomal abnormalities during prolonged culture (Table 1). For dividing cells, the occurrence of genetic aberrations is not an unusual phenomenon; yet the nonrandom survival of such variants and their ability to overtake the normal diploid cells in a culture implies that the chromosome abnormalities observed in HESCs can impart a growth advantage to those cells in which they arise. As such, we consider cells that are karyotypically abnormal and show an increased growth rate as 'culture adapted', and the process through which this occurs to be 'adaptation'. Such adaptation may also result in enhanced cloning efficiencies after plating single cells4; a reduced tendency for apoptosis in adapted HESCs has also been reported5. Another expectation is that culture-adapted cells may show a reduced capacity for differentiation, but this is difficult to assess quantitatively and has not yet been studied systematically, although the retention of undifferentiated stem cells in a xenograft tumor of a culture-adapted HESC line has been reported6. The culture adaptations discussed below are all based on the observation of chromosomal changes, though adaptation could also involve other genetic and epigenetic changes that are not observed by standard cytogenetic analyses4, 7. Here we discuss the karyotypic changes observed in culture adaptation, the parallels between adaptation and germ cell tumorigenesis and the implications of this process in the maintenance and therapeutic use of HESCs.


Genetic change and germ cell transformation

For many years genetic change has been known to be associated with neoplasia8. In tumor development, genetic abnormalities, which can affect apoptotic pathways, differentiation control or cell cycle, arise in precancerous cells, resulting in an uncontrolled increase in growth. Such a process may be considered similar to adaptation in HESCs, in which a mutation can provide a growth advantage that allows the abnormal cell to dominate a culture over a series of passages. Thus the functions of the same genes might be subject to similar change in both situations, though, since the growth conditions of a cancer cell in vivo and a stem cell in vitro are not the same, not all of the progressive changes are likely to be congruent.

The concept that HESCs adaptation and malignancy are related has most direct relevance to testicular germ cell tumors (TGCTs). These tumors develop from pluripotent germ cells9, by way of intermediate carcinoma in situ (CIS) cells10, and are the most common malignancy in young men11. TGCTs consist of two histologically distinct groups: seminomas and nonseminomas. Seminomas are generally uniform and resemble the primordial germ cells and CIS from which they arise. Nonseminomas contain a number of different components, which include embryonal carcinoma (EC), yolk sac tumor, choriocarcinoma and teratoma12. Confusingly, they may also contains elements of seminoma13. The EC cells of the nonseminomas may be pluripotent, with the ability to differentiate into a wide range of cell types10, and show a phenotypic resemblance to ICM cells. Thus, they present a malignant caricature of ES cells, and one can envisage a similarity between adapted HESCs that have been selected during continuous passage and EC cells, which are also pluripotent but are primed for invasive expansion rather than differentiation. Many EC cells from TGCTs do show limited developmental potency, or even nullipotency, which might be regarded as the ultimate consequence of adaptation occurring in tumors that have developed over a period of twenty or more years.


Chromosomal abnormalities in HESCs and TGCT cells

Karyotypic abnormalities have been reported by many HESC laboratories (Tables 1 and 2; Figs. 1 and 2). Over the past 18 months we have subjected 30 cell lines and sublines to routine cytogenetic analysis, and 16 of them have developed abnormal karyotypes, though they were initially diploid (methods of cell culture and cytogenetic analysis can be found in Supplementary Methods online). In a number of cases, diploid cells were often present together with karyotypically abnormal cells (Table 2). Though there have been relatively few systematic studies of progressive changes in single sublines of abnormal HESCs, it seems likely that such cases of mosaic cultures reflect the 'snapshot' nature of most reports, and that the abnormal cells will generally overtake the normal population on continued culture14 (Fig. 3).

Figure 1: Representative cytogenetic data from a HESC line maintained in Sheffield.

Figure 1 : Representative cytogenetic data from a HESC line maintained in Sheffield.

G-banded karyotype of chromosome abnormalities in an example cell line, HUES 5 (passage 19), in which gains of 12 and 17 are frequently seen.

Full size image (42 KB)

Figure 2: Ideogram of all reported chromosome abnormalities in HESCs.

Figure 2 : Ideogram of all reported chromosome abnormalities in HESCs.

The bars represent gains (green) and losses (red) of chromosome regions. Abnormalities observed in the authors' laboratory are highlighted in dark green and dark red.

Full size image (48 KB)

Figure 3: Interphase FISH analysis of H14.s3 cell line hybridized with an iso17q probe (Kreatech Biotechnology) specific for the p53 (17p13) and MPO (17q23) genes.

Figure 3 : Interphase FISH analysis of H14.s3 cell line hybridized with an iso17q probe (Kreatech Biotechnology) specific for the p53 (17p13) and MPO (17q23) genes.

(H14.s3 is a subline of H14, provided by J. Thomson77, University of Wisconsin, USA). On the x axis, passage number of the culture is shown, and on the y axis, the percentage of cells showing trisomy 17 at each passage number is plotted. A total of 300 cells were examined at each passage. The graph shows the abnormal HESCs overtaking the culture.

Full size image (14 KB)


Most studies of genetic change in HESCs have used the well established cytogenetic techniques of G-banding and fluorescent in situ hybridization (FISH)14, 15, 16, 17, 18, though a few reports have used array-based comparative genomic hybridization (CGH) and single-nucleotide polymorphism (SNP) techniques4, 19, 20. The latter methodologies offer a much higher level of resolution in detecting changes, though they suffer from rapid loss of sensitivity when applied to mosaic cultures. By contrast, G-banding techniques can readily identify small subpopulations of abnormal cells. In practice, reliable identification of 5% abnormal cells can be readily achieved, whereas interphase FISH techniques to identify cells carrying specific chromosome amplifications allow recognition of 0.5% abnormal cells. Nevertheless, the approaches are complementary and will no doubt provide valuable insights in due course.

The most frequent karyotypic changes observed were gains of chromosomes 12, 17 and to a lesser extent X, again invoking direct comparison to TGCTs, in which similar nonrandom changes are also seen. The gain of chromosome 12 material, most often as isochromosome 12p (i(12p)), is so prevalent in TGCT that it can be used as a diagnostic marker of this form of malignancy21, and the gain of chromosome 17 has been observed in a number of TGCTs22; in particular, gain of 17q material was associated with nonseminomas23. In another study of the EC element of TGCT, gain of material from chromosomes 12, 17 and X, as well as 1 and 7, were the most frequent imbalances24. Other chromosomal gains and losses have been reported25, and TGCTs are typically highly aneuploid with many chromosomal aberrations, yet the nature of cancer may make the malignant cells progressively unstable and prone to acquire further changes that may have no functional significance. However, the repeated gain of 12, 17 and X in both TGCTs and abnormal HESCs indicates that these chromosomes may contain genes critical for cell growth and potentially tumorigenesis.


Gain of chromosome 17

Of the HESC abnormalities that we have observed, gain of material from 17q predominated (present in 15 of 16 abnormal karyotypes). Gain of chromosome 17 was the sole abnormality in 4 of 16 karyotypes, though it was mostly seen with additional gains of chromosomes 12, X or both. The gain of chromosome 17 in HESCs has also been reported by other groups15, 19; Maitra and co-workers19 noted a gain of 17q using an oligonucleotide-based array assay (Table 1). In addition, a study of abnormal mouse ESC (MESC) lines found trisomy of chromosome 11 (syntenic to human chromosome 17) in almost 20% of these lines26.

Derivative chromosomes involving unbalanced translocations resulted in only a partial gain of 17q material in four cases. In two cases the breakpoints were in 17q11–q12, whereas in one, a HUES 14 subline with a der(10;17), the breakpoint was between 17q21 and 17q23.2 (Fig. 4a,b). In the fourth case, involving the HESC line H1, only the terminal region of chromosome 17, 17q25–qter, was gained. This points towards a minimal amplicon in the terminal half of the q arm of chromosome 17, and possibly only in the region 17q25–qter. A similar amplicon has also been observed in TGCTs and neuroblastomas, with the shortest regions of gain identified as17q24–qter (ref. 23) and 17q23.1–qter (ref. 27) respectively. Nonrandom gain of 17q has also been reported in other cancer types (for example, breast cancer28), indicating that this region may house a number of gene(s) that are important to the malignant cell. One candidate gene located in this region is BIRC5. BIRC5 is an antiapoptotic gene, which is expressed in neuroblastomas and is particularly associated with the highest-risk tumors29.

Figure 4: Evidence of a minimal region of amplification on chromosome 17q in culture-adapted HESCs.

Figure 4 : Evidence of a minimal region of amplification on chromosome 17q in culture-adapted HESCs.

(a) G-banded chromosomes from H7.s6, HUES6 and HUES13 lines (see Table 2 for karyotypes), demonstrating unbalanced chromosomal translocations that result in gain of 17q. (b) G-banded chromosomes with an unbalanced translocation in the HUES14 cell line resulting in a partial 17q gain. The translocation breakpoint was ascertained using FISH, with locus specific probes showing its location between 17q21 and 17q23 (see Supplementary Methods). We thus suggest a minimal amplicon on chromosome 17 for adapted HESCs that is likely to be in the region 17q21–qter.

Full size image (38 KB)


Gain of chromosome 12

We observed gain of chromosome 12 in 9 of 16 abnormal karyotypes; on one occasion it was the sole abnormality, as i(12p). Other groups5, 15, 16, 30 have reported the occurrence of trisomy 12 in HESC cultures more often than gain of 17. This implies that culture adaptation may be achieved in different ways, and that subtle differences in culture conditions between laboratories may exert differential selective pressures for alternative routes to adaptation.

Amplification of chromosome 12p in TGCTs is most commonly achieved by the formation of i(12p), with multiple i(12p) sometimes present9. Among the HESCs, gain of chromosome 12 material was mostly as trisomy for the entire chromosome, but i(12p) has been observed twice by ourselves and also in another laboratory18. The pluripotency gene NANOG is located on 12p in the 12p13 band, and cytogenetic study of TGCTs in which translocations of chromosome 12 fragments occur has shown this to be a frequently amplified region in these tumors31, 32. Clearly, overexpression of a gene such as NANOG, which tends to promote self-renewal and prevent differentiation, could provide cells with an advantage in culture, and so such genes are potential players in culture adaptation. The other stem cell–associated genes, DPPA3 and GDF3, are also located in the 12p13 band, as is the cell-cycle regulator CCND2, and so these are also candidates behind the recurrence of this amplification.

A second frequently amplified region on chromosome 12p,12p11.2–12, has also been reported in TGCT33, 34. Genes of potential interest in this region include the oncogene KRAS, and also SOX5, which is involved in the determination of cell fate. Malignant cells with this amplification show reduced apoptosis compared to those cells without it33. The reduced apoptosis associated with this amplification may be particularly noteworthy, as a recent report indicates this as a feature of transformed HESCs5. The presence of two separate regions of selected amplification in 12p is striking, but over-representation of either or both of these neighboring regions on 12p may provide cells with a selective advantage for different reasons, which may explain the prevalence of this particular chromosomal gain.


Gain of X chromosome

We have also found additional copies of chromosome X in 5 of 16 HESC lines, but each time with trisomy 17. Other groups have also observed aneuploidy for chromosome X in cells that were trisomic for chromosomes 12 and 17 (refs. 15,16). Trisomy X has once been reported as the sole abnormality35, though a partial gain was seen by Inzunza et al. 36, who noted an Xp gain and a partial Xq gain resulting from an isodicentric X with a breakpoint at Xq21.

An effective gain of X can also be achieved by failure of X inactivation if the XIST gene is not expressed. Sperger et al.37 detected XIST in only two out of three different female HESCs, indicating that there may be no X inactivation in the XIST-negative cell line. In addition Dhara et al.38 showed that the undifferentiated HESCs of line H9 possessed two active X chromosomes, whereas X inactivation followed differentiation. However, Enver et al.4 found loss of X-inactivation and overexpression of X-linked genes in undifferentiated stem cells of a culture-adapted derivative of the H7 HESC line, with a 47,XX,+1,der(6)(t(6;17)(q27;q1) karyotype, in comparison to the parental diploid line, which did show X inactivation. Unlike the case in the study by Dhara et al.38, X inactivation failed in the differentiated derivatives. Like HESCs, TGCTs often have more than one X chromosome39, and Kawakami et al.40 suggested that such additional X chromosomes are also mostly active in these tumors, regardless of XIST expression.

Oncogenes ELK1 and ARAF are cell signaling molecules41, 42 that are present on chromosome X and are candidates for adaptation. In addition, the genes for the androgen receptor (AR) and its interacting protein NONO43 are also located on this chromosome. AR is involved in cell cycle progression44, with influence over cell division and apoptosis, making its gene a possible target. In TGCT a tumor susceptibility gene (TGCT1) has been implicated at Xq27 (ref. 45), though the nature of the protein encoded is yet to be determined.


Other abnormalities

Other observed HESC abnormalities include trisomies for chromosomes 1, 3, 7, 8, 9, 14 and 20 (Figs. 1 and 2). In addition to the gain of the whole of chromosome 1, we have also noted the involvement of the p arm in structural rearrangements, such as translocations in sublines of Shef5 and H7. Notably, however, in neither HESCs nor TGCTs have translocations with specific breakpoints been seen, comparable, for example, to the Philadelphia chromosome that yields the ABL-BCR gene fusion responsible for malignancy in chronic myeloid leukemia46. Mostly, the additional changes that have been seen occur in cells that also carry gains of chromosomes 12, 17 or both, indicating that their effects may be secondary. For example, a subline of H7, H7.s6, accumulated abnormalities such as gains of 8 and 20, but only after first acquiring a gain of 17q. In this case, the gain of chromosome 8 in H7.s6 cells that already carried gains of chromosomes 17q and 1 occurred during growth in a severe combined immunodeficiency mouse as a xenograft teratoma, from which undifferentiated ES-like cells were recovered by explantation back to culture6. Of the less frequent HESC changes, gain of whole or parts of chromosomes 1, 7, 8 and 9 have also been reported in TGCTs22, 23, 47. Given the result with the H7.s6 xenograft, it may be that gain of chromosome 8 has pertinence to some aspect of tumor growth, but it is also possible that after the initial chromosomal changes the HESCs are more genetically unstable and may by chance acquire random changes that confer little benefit.

In addition to chromosomal gains, nonrandom loss of chromosome regions has also been seen in TGCT23, 25. Chromosome deletion is much less frequent than chromosome gain in HESCs, though loss of chromosome 13 has been reported. This abnormality is not uncommon in TGCTs, with decreased copy number at 13q21–31 observed by Looijenga et al.48 in four of five nonseminomas. Microarray studies comparing normal and adapted cells4 indicate downregulation of certain genes in the adapted cells, so loss of small regions of chromatin (undetectable by G-banding) may occur. It is also possible that gross chromosomal losses may occur later in tumor development, and further adaptation of the ES cells may be necessary before frequent DNA loss is seen.


Cancer characteristics in an HESC line

Recently we have found a homogeneous staining region (HSR) derived from chromosome 17 in an adapted subline of HESC line, H14 (Fig. 5). HSRs are terminal or interstitial additions to a chromosome that usually stain uniformly with banding techniques; they are the cytogenetic representation of gene amplification49.They are found almost exclusively in cancer and cultured tumor cell lines50, and to our knowledge this is the first report of such an abnormality in HESCs. Further investigation of the HSR by CGH and FISH (see Supplementary Methods) showed that it was derived by amplification of a region originating from 17p11.2 (Fig. 5). We also found a subline of the HESC line Shef5 that developed an abnormality of 17p involving the triplication of a segment from 17p11.2 (Supplementary Fig. 1 online), the same region from which the HSR was derived in the H14 cells. Only five cases of HSR in TGCTs have been reported22, 51, and the origins of this rare abnormality have only been confirmed in two cases: one derived from 1p32 and, strikingly, one derived from 17p11 (ref. 22).

Figure 5: Characterization of the HSR in the abnormal HESC line H14.

Figure 5 : Characterization of the HSR in the abnormal HESC line H14.

(a) G-banding showing two normal chromosomes 17 and one abnormal 17 with an HSR. (b) Chromosomes 17 from a multicolor FISH analysis, showing that the HSR paints only with chromosome 17–specific probes. (c) Comparative genomic hybridization using DNA probes prepared from the adapted HSR-carrying H14 cells (test DNA) and from karyotypically normal cells (reference DNA). The hybridized chromosome 17 shows a strong yellow band on proximal 17p, indicating amplification. The ideogram and CGH profile for chromosome 17 is shown; the green and red lines represent gain and loss of DNA respectively, and the black line represents the normal diploid copy number. The pink line indicates the ratio of test to reference DNA and is showing both trisomy of 17q and amplification of proximal 17p (n, number of chromosomes analyzed). (d) FISH with the SMS probe (Vysis, Abbott Molecular), which is specific for the critical region in 17p11.2 that is commonly deleted in Smith Magenis syndrome78. The SMS signal (red) is amplified on the HSR compared to the signal from the RARA probe (17q21, green) which was included as a control. These studies show that the HSR amplicon is derived from proximal 17p and includes the region specified by the SMS probe.

Full size image (47 KB)

Identifying the amplified gene(s) in the 17p11.2 HSR responsible for tumorigenic-like growth remains a challenge. We performed a microarray study comparing the HSR-containing H14 subline with its karyotypically normal parent line and observed marked overexpression of several genes located in 17p11.2, consistent with the origin of the HSR from this region of chromosome 17 (see Supplementary Methods). Of the genes upregulated in the abnormal H14 cells, TOP3A, COPS3 and MAPK7 are candidates, as they have been implicated as potentially oncogenic in osteosarcomas52.

Candidate genes from chromosome 17p were also highlighted in an analysis of TGCTs in which regions showing increases in DNA copy number and also gene expression were identified53. For those tumor cells trisomic for chromosome 17, eighteen genes were identified as having obvious increases in copy number and expression, and two of these were located in 17p11.2. These two candidate genes are FAM18B, a protein of unknown function, and GRAP, a Grb2-related adaptor protein54. In MESCs, GRB2 has been implicated in ERK signaling that encourages differentiation55; yet in HESCs, ERK signaling downstream of bFGF may maintain cells in an undifferentiated state56. Further, GRAP has also been shown to interact with oncogenic proteins such as KIT (ref. 54) that could transform HESCs such that they possess an advantage in culture. However, in the microarray expression study of the HSR-carrying H14 subline, FAM18B, but not GRAP, seemed to be overexpressed.


Transformation of HESCs: mutation and selection

The appearance of karyotypically abnormal HESCs inevitably involves the operation of two quite separate processes, mutation and selection. At present, however, there are no data to assess the relative importance of mutation rate and selective pressure in the appearance of variant cells. As the ICM cells of the blastocyst give rise to all cell types in the body, any mutations that occur in them would be perpetuated throughout the developing embryo, with potentially catastrophic effects. Hence, one might expect ICM cells, and the ES cells derived from them, to be more efficiently protected against genetic instability than later somatic cells. Indeed, MESCs do show a lower mutation frequency than somatic cells; the spontaneous mutation rate at an APRT reporter gene is reported as 10−6 in MESCs, compared to 10−4 in somatic cells57, 58. In somatic cells, DNA damage may be repaired following cell cycle arrest at the G1 checkpoint, though with some risk of errors arising in that process, too. However, mouse and primate ES cells lack a G1 checkpoint, and variant cells are proposed to be funneled directly toward an apoptotic fate58, 59, which may explain why they have a low rate of mutations.

Nevertheless, aberrations clearly do occur during HESC replication. The most frequent karyotypic changes so far reported involve chromosomal trisomy, indicating that chromatid separation during mitosis may be prone to error in HESCs. Indeed, the G2 decatenation checkpoint, which delays entry into mitosis from G2 if the chromosomes have not been sufficiently disentangled or decatenated60, has been reported as highly inefficient in MESCs61. Mis-segregation of chromatids may also result from defects in the mitotic spindle checkpoint. This checkpoint ensures the correct alignment of sister chromatids before their separation62, and faults in its action can cause aneuploidy if segregation is attempted while the chromatids are linked to only one spindle pole. In preimplantation-stage embryos, the high frequency of aneuploidy, estimated between 30–65%, resulting from mitotic as well as meiotic errors63 indicates that poor mitotic control may be a general feature of embryonic cells. Many of these changes seem to be incompatible with embryo survival, or may be eliminated by apoptosis of abnormal cells in the blastocyst64. By contrast, most reports of early-passage HESCs indicate that they are diploid and so derived from surviving karyotypically normal ICM cells. It may be that an apoptotic mechanism, coupled with the transient nature of the ICM and its limited expansion in vivo, reduces the opportunity for the appearance and survival of abnormal cells in the embryo to a much lower level than subsequently encountered during prolonged propagation of HESCs in vitro.

The aneuploidy observed in HESCs is most likely driven by the stresses induced by the more variable environmental conditions to which the cells are exposed in culture. At present, it is difficult to disentangle the roles of mutation and selection, although one notion is that oxygen tension in routine culture does not match the low oxygen tension in vivo. This high oxygen tension could promote abnormalities by causing damage to chromosomes (shortening telomeres) and mitochondrial DNA65. Maitra et al.19 found alterations in mitochondrial DNA in late-passage HESC lines cultured under high oxygen tension, implying that the high oxygen tension could be responsible for promoting chromosome abnormalities.

Nevertheless, the repetitive nature of the aberrations seen in HESCs is most likely to reflect the role of selection processes acting in culture. A high rate of mutation without selection would lead to cultures with quite random variation; the nonrandom nature of the observed karyotypic change in HESC cultures points strongly instead toward selection as having a critical role in fixing variants in the population. Mechanisms are likely to exist in vivo for the selective destruction of abnormal embryos, but alternative pressures may exist in vitro leading the abnormal cells to thrive.

A wide range of selective pressures could drive the appearance of variant cells in HESC cultures. One major consideration is the difference in techniques for the subculture of the cells employed by different HESC laboratories. Some use manual dissection and transfer of colonies, whereas others use bulk techniques involving nonenzymatic (cell dissociation buffer) or enzymatic (collagenase or trypsin) methods, which disaggregate cultures to small clumps or single cells. In one study comparing these approaches15, those lines propagated by manual dissection retained normal karyotypes at passages up to and including passage 105, whereas lines propagated by bulk methods acquired chromosomal abnormalities (including +12, +17, +X) between passages 23 and 45. This effect of passaging technique was supported by studies of genetic change assessed by an SNP analysis16, 19. One argument is that bulk passaging methods might either increase the stress upon the cells and increase the underlying mutation rate, or impose specific selective pressures that are not exerted by manual dissection techniques.

Although these ideas are plausible, the available evidence does not permit a clear test of these hypotheses. For a given mutation rate, the probability of a genetic change occurring in a group of cells depends upon population size. Inevitably, manual passaging techniques involve the transfer of smaller numbers of cells at each passage than techniques based upon mass dissociation, and so they may inevitably result in a lower rate of acquisition of genetic change than other techniques. In any case, comparison of the data from different laboratories is difficult because passage number, the datum ordinarily reported, has very little validity as a measure of the number of cell generations attained. Further, in manual passaging, conscious or unconscious selection of colonies with specific morphology means that the cells transferred at each passage are not a random set of the cells in the culture, whereas bulk passage following disaggregation techniques is more likely to lead to random transfer of cells. Thus, whether or not manual passaging proves a more reliable technique in practice for maintaining diploid HESC stock cultures, the reasons for its advantage in this respect remain obscure.

Other factors, apart from the technique used to harvest HESCs for passage, may also induce stresses through which abnormal cells could be selected. High cell density in culture may result in rapid depletion of nutrients in the medium and select for those cells that can survive on lower concentrations of these factors. Conversely, low-density cultures may also be deleterious to HESC survival: EC cells maintained at low density show a greater propensity to differentiate compared to those kept at higher density, and as such they are less likely to be perpetuated through extended passage66, 67; and the low plating efficiency of HESCs is notorious68. These observations imply that the human cells, in contrast to their mouse counterparts, are dependent upon intracellular signaling, perhaps mediated by cell-cell contact, for maintenance of self-renewal. The increased cloning efficiency observed in culture-adapted HESCs4 indicates that they have, at least to some degree, overcome their dependency upon such signaling.

Another key variable in HESC culture is the medium in which the cells are grown, which may also provide an avenue for the selection of abnormal cells. Mutations that result in more effective use of medium components or development of pathways that can supplement the components already present are likely to confer a growth advantage. Recent work by Pyle et al.69 demonstrated that standard HESC medium (as described by Amit et al.68) is not optimal for stem cell maintenance, as addition of neurotrophins increased HESC clonal survival, thus showing the potential to increase cell growth beyond that obtained using standard medium. It is noteworthy that a gene encoding one of the neurotrophin receptors (NGFR) is located on chromosome 17q. In addition to the factors contained in the medium, the feeder cells on which HESCs are grown also provide support necessary for maintenance of undifferentiated stem cells. A systematic trial of HESC survival on different types or batches of feeder cells has not yet been performed, and it may be that different feeder stocks support some cells more effectively than others, again providing a selective pressure. Freezing and subsequent thawing is another pressure point that HESCs must endure; it once again affords the possibility of selection for those cells which can best survive these processes.


Implications for the practical applications of HESCs

The potential for long term genetic instability in HESCs, and their tendency to acquire genetic changes during prolonged culture, raises obvious concerns about their potential use in regenerative medicine. Akin to cancerous transformation, these changes provide growth advantages, and it has already been shown that the abnormal HESCs give rise to terato carcinomas when inoculated into severe combined immunodeficient mice6. However, it seems unlikely that the undifferentiated cells would ever be used directly in therapy: it is their differentiated derivatives that would be most likely to be transplanted into patients to replace diseased or damaged tissue, and these cells are quite distinct, having a finite lifespan with different growth characteristics and control mechanisms.

It is by no means clear that a mutation, or other genetic or epigenetic change that provides a selective advantage for the undifferentiated stem cells, would necessarily have any effect on the behavior of the specific differentiated derivatives. Clearly, this has to be addressed experimentally, yet by understanding the processes of genetic change and the factors that drive selection of specific variants, it is likely that we will be able to develop culture conditions that minimize the appearance of abnormal cells. Certainly MESCs are no less subject to the generation of adapted cells than HESCs70, 71, but this has not seriously hampered their use to produce transgenic mice. Nevertheless, one practical consequence of culture adaptation is its potential effect on efforts to develop defined media for the culture of HESCs. It will be essential to establish that newly defined medium formulations are able to support the growth of HESCs generally, and not only that of the HESC lines used in its development.

The use of culture-adapted HESCs might in some circumstances offer particular advantages. Although detailed studies of the effect of adaptation on the differentiation of HESCs have yet to be described, it is clear that adapted cells may still show extensive pluripotency, and that they may also be adapted to culture in simple media without the presence of feeders72. Such adapted cells might be more convenient and cost-effective for use in high-throughput assays for drug screening and toxicology than genetically normal cells that are more difficult to maintain and expand. Nevertheless, as for their potential use in regenerative medicine, the behavior of culture-adapted HESCs will have to be characterized with reference to specific applications.

Regardless, the HESC cytogenetic data demonstrates that regular monitoring of the available lines will be essential for future experimental and therapeutic work. Analyses using standard cytogenetics, multicolor FISH or metaphase CGH offer whole-genome scans at relatively low resolution (4–10 Mb), whereas array CGH and SNP arrays afford analyses at increased resolution. However, based on our present knowledge, FISH with probe sets specific for the common abnormalities (gain of material from chromosomes 12, 17 and X) could provide a rapid screening technique for routine monitoring of cultures at frequent intervals, although this would not obviate a need for more detailed screening, perhaps every 20–30 passages.


Final comments

That genetic change occurs in HESCs is now well established. It is also clear that the karyotypic changes observed are nonrandom and commonly affect only three chromosomes in both HESCs and TGCTs, consistent with only a few genes and regulatory pathways being affected. Combining the available data from TGCTs as well as HESCs, parsimony suggests that considerable adaptive advantages in these cells can be generated by amplification of any one out of five chromosomal regions: two on the short arm of chromosome 12 (12p11.2–12 and 12p13), two on chromosome 17 (one in the terminal half of the long arm (17q21–qter) and one in the short arm (17p11.2)) and one on the X chromosome. Because sometimes these changes occur independently of each other, it further indicates that a similar number of alternative pathways may provide substantial adaptive advantages for such pluripotent stem cells dependent upon precise local growth conditions. The simplistic hypothesis would be that only five genes are involved, one from each region, but of course the situation may be more complex, the latter possibility perhaps being supported by the historical difficulty in identifying single genes that underlie common chromosome amplifications in various cancers, not only TGCT. Nevertheless, these ideas serve as working hypotheses with which to begin to address the mechanisms that control the proliferation of HESCs and the decisions they make among self renewal, commitment to differentiation or apoptosis. In an iterative manner, identification of the genes commonly affected may provide insights into these mechanisms.

Apart from its nonrandom nature, the other striking feature of the karyotypic changes in HESCs is their similarity to those arising in EC cells from TGCTs. In many tumors, an origin from a normal tissue stem cell can be inferred, but direct experimentation with such normal stem cells is generally difficult. Arguably the ES/EC cell pair, and the phenomenon of culture adaptation of ES cells, provides a rare, if not the only, circumstance in which one can readily access the normal counterpart of a tumor stem cell and observe progressive changes as the 'normal' stem cell converts to a malignant state. In this respect, deciphering the mechanisms of culture adaptation of HESCs may give insights not only into testicular cancer, but to the development of cancers more generally.

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



We are grateful to C. Cowan and D. Melton for providing the HUES1-17 HESC lines, and to J. Thomson for the H1, H7 and H14 HESC lines. We are grateful to our colleagues, especially J. Jackson, K. Amps, G. Bray and G. Bingham for culture of the HESCs. In addition, we would like to acknowledge B. Aflatoonian and L. Ruban for their assistance in the derivation and proliferation of the Shef cell lines. This work was supported by grants from the Medical Research Council, Yorkshire Cancer Research, The Engineering and Physical Sciences Research Council and The Juvenile Diabetes Research Foundation.

Competing interests statement

The authors declare no competing financial interests.



  1. O'Neill, L.P., Vermilyea, M.D. & Turner, B.M. Epigenetic characterization of the early embryo with a chromatin immunoprecipitation protocol applicable to small cell populations. Nat. Genet. 38, 835–841 (2006). | Article | PubMed | ISI | ChemPort |
  2. Bradley, A., Evans, M., Kaufman, M.H. & Robertson, E. Formation of germ-line chimaeras from embryo-derived teratocarcinoma cell lines. Nature 309, 255–256 (1984). | Article | PubMed | ISI | ChemPort |
  3. Hoffman, L.M. & Carpenter, M.K. Characterization and culture of human embryonic stem cells. Nat. Biotechnol. 23, 699–708 (2005). | Article | PubMed | ISI | ChemPort |
  4. 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 |
  5. Herszfeld, D. et al. CD30 is a survival factor and a biomarker for transformed human pluripotent stem cells. Nat. Biotechnol. 24, 351–357 (2006). | Article | PubMed | ChemPort |
  6. 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 |
  7. Liu, Y. et al. Genome wide profiling of human embryonic stem cells (hESCs), their derivatives and embryonal carcinoma cells to develop base profiles of U.S. Federal government approved hESC lines. BMC Dev. Biol. 6, 20 (2006). | Article | PubMed | ChemPort |
  8. Boveri, T. The origin of malignant tumours (Williams and Walkin, Baltimore, 1914).
  9. Oosterhuis, J.W. & Looijenga, L.H. Testicular germ-cell tumours in a broader perspective. Nat. Rev. Cancer 5, 210–222 (2005). | Article | PubMed | ISI | ChemPort |
  10. Damjanov, I. Teratocarcinoma stem cells. Cancer Surv. 9, 303–319 (1990). | PubMed | ISI | ChemPort |
  11. Moller, H. Clues to the aetiology of testicular germ cell tumours from descriptive epidemiology. Eur. Urol. 23, 8–13; discussion 14–15 (1993). | PubMed | ChemPort |
  12. Damjanov, I. Pathogenesis of testicular germ cell tumours. Eur. Urol. 23, 2–5; discussion 6–7 (1993). | PubMed |
  13. Andrews, P.W. From teratocarcinomas to embryonic stem cells. Phil. Trans. R. Soc. Lond. B 357, 405–417 (2002). | Article |
  14. Draper, J.S. et al. Recurrent gain of chromosomes 17q and 12 in cultured human embryonic stem cells. Nat. Biotechnol. 22, 53–54 (2004). | Article | PubMed | ISI | ChemPort |
  15. Mitalipova, M.M. et al. Preserving the genetic integrity of human embryonic stem cells. Nat. Biotechnol. 23, 19–20 (2005). | Article | PubMed | ISI | ChemPort |
  16. 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 |
  17. Buzzard, J.J., Gough, N.M., Crook, J.M. & Colman, A. Karyotype of human ES cells during extended culture. Nat Biotechnol. 22, 381–382; author reply 382 (2004). | Article | PubMed | ISI | ChemPort |
  18. Imreh, M.P. et al. In vitro culture conditions favoring selection of chromosomal abnormalities in human ES cells. J. Cell. Biochem. 99, 508–516 (2006). | Article | PubMed | ChemPort |
  19. Maitra, A. et al. Genomic alterations in cultured human embryonic stem cells. Nat. Genet. 37, 1099–1103 (2005). | Article | PubMed | ISI | ChemPort |
  20. Josephson, R. et al. A molecular scheme for improved characterization of human embryonic stem cell lines. BMC Biol. 4, 28 (2006). | Article | PubMed | ChemPort |
  21. Reuter, V.E. Origins and molecular biology of testicular germ cell tumors. Mod. Pathol. 18 (suppl. 2), S51–S60 (2005). | Article |
  22. Samaniego, F. et al. Cytogenetic and molecular analysis of human male germ cell tumors: chromosome 12 abnormalities and gene amplification. Genes Chromosom. Cancer 1, 289–300 (1990). | PubMed | ChemPort |
  23. Kraggerud, S.M. et al. Genome profiles of familial/bilateral and sporadic testicular germ cell tumors. Genes Chromosom. Cancer 34, 168–174 (2002). | Article | PubMed | ISI | ChemPort |
  24. Summersgill, B.M. et al. Definition of chromosome aberrations in testicular germ cell tumor cell lines by 24-color karyotyping and complementary molecular cytogenetic analyses. Cancer Genet. Cytogenet. 128, 120–129 (2001). | Article | PubMed | ChemPort |
  25. von Eyben, F.E. Chromosomes, genes, and development of testicular germ cell tumors. Cancer Genet. Cytogenet. 151, 93–138 (2004). | PubMed | ISI | ChemPort |
  26. Sugawara, A., Goto, K., Sotomaru, Y., Sofuni, T. & Ito, T. Current status of chromosomal abnormalities in mouse embryonic stem cell lines used in Japan. Comp. Med. 56, 31–34 (2006). | PubMed | ChemPort |
  27. Lastowska, M. et al. Breakpoint position on 17q identifies the most aggressive neuroblastoma tumors. Genes Chromosom. Cancer 34, 428–436 (2002). | Article | PubMed | ISI | ChemPort |
  28. Tirkkonen, M. et al. Molecular cytogenetics of primary breast cancer by CGH. Genes Chromosom. Cancer 21, 177–184 (1998). | Article | PubMed | ISI | ChemPort |
  29. Azuhata, T. et al. The inhibitor of apoptosis protein survivin is associated with high-risk behavior of neuroblastoma. J. Pediatr. Surg. 36, 1785–1791 (2001). | Article | PubMed | ISI | ChemPort |
  30. Cowan, C.A. et al. Derivation of embryonic stem-cell lines from human blastocysts. N. Engl. J. Med. 350, 1353–1356 (2004). | Article | PubMed | ISI | ChemPort |
  31. Korkola, J.E. et al. Down-regulation of stem cell genes, including those in a 200-kb gene cluster at 12p13.31, is associated with in vivo differentiation of human male germ cell tumors. Cancer Res. 66, 820–827 (2006). | Article | PubMed | ISI | ChemPort |
  32. Skotheim, R.I. et al. Novel genomic aberrations in testicular germ cell tumors by array-CGH, and associated gene expression changes. Cell. Oncol. 28, 315–326 (2006). | PubMed | ChemPort |
  33. Roelofs, H. et al. Restricted 12p amplification and RAS mutation in human germ cell tumors of the adult testis. Am. J. Pathol. 157, 1155–1166 (2000). | PubMed | ISI | ChemPort |
  34. Zafarana, G. et al. 12p-amplicon structure analysis in testicular germ cell tumors of adolescents and adults by array CGH. Oncogene 22, 7695–7701 (2003). | Article | PubMed | ISI | ChemPort |
  35. Ludwig, T.E. et al. Derivation of human embryonic stem cells in defined conditions. Nat. Biotechnol. 24, 185–187 (2006). | Article | PubMed | ISI | ChemPort |
  36. Inzunza, J. et al. Comparative genomic hybridization and karyotyping of human embryonic stem cells reveals the occurrence of an isodicentric X chromosome after long-term cultivation. Mol. Hum. Reprod. 10, 461–466 (2004). | Article | PubMed | ChemPort |
  37. 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 |
  38. 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 |
  39. Sandberg, A.A., Meloni, A.M. & Suijkerbuijk, R.F. Reviews of chromosome studies in urological tumors. III. Cytogenetics and genes in testicular tumors. J. Urol. 155, 1531–1556 (1996). | Article | PubMed | ISI | ChemPort |
  40. Kawakami, T. et al. The roles of supernumerical X chromosomes and XIST expression in testicular germ cell tumors. J. Urol. 169, 1546–1552 (2003). | Article | PubMed | ISI |
  41. Yang, S.H., Jaffray, E., Hay, R.T. & Sharrocks, A.D. Dynamic interplay of the SUMO and ERK pathways in regulating Elk-1 transcriptional activity. Mol. Cell 12, 63–74 (2003). | Article | PubMed | ISI | ChemPort |
  42. Wu, X., Noh, S.J., Zhou, G., Dixon, J.E. & Guan, K.L. Selective activation of MEK1 but not MEK2 by A-Raf from epidermal growth factor-stimulated HeLa cells. J. Biol. Chem. 271, 3265–3271 (1996). | Article | PubMed | ISI | ChemPort |
  43. Ishitani, K. et al. p54nrb acts as a transcriptional coactivator for activation function 1 of the human androgen receptor. Biochem. Biophys. Res. Commun. 306, 660–665 (2003). | Article | PubMed | ChemPort |
  44. Bai, V.U., Cifuentes, E., Menon, M., Barrack, E.R. & Reddy, G.P. Androgen receptor regulates Cdc6 in synchronized LNCaP cells progressing from G1 to S phase. J. Cell. Physiol. 204, 381–387 (2005). | Article | PubMed | ChemPort |
  45. Rapley, E.A. et al. Localization to Xq27 of a susceptibility gene for testicular germ-cell tumours. Nat. Genet. 24, 197–200 (2000). | Article | PubMed | ISI | ChemPort |
  46. Daley, G.Q. & Baltimore, D. Transformation of an interleukin 3-dependent hematopoietic cell line by the chronic myelogenous leukemia-specific P210bcr/abl protein. Proc. Natl. Acad. Sci. USA 85, 9312–9316 (1988). | Article | PubMed | ChemPort |
  47. Wang, N., Trend, B., Bronson, D.L. & Fraley, E.E. Nonrandom abnormalities in chromosome 1 in human testicular cancers. Cancer Res. 40, 796–802 (1980). | PubMed | ChemPort |
  48. Looijenga, L.H. et al. Comparative genomic hybridization of microdissected samples from different stages in the development of a seminoma and a non-seminoma. J. Pathol. 191, 187–192 (2000). | Article | PubMed | ISI | ChemPort |
  49. Biedler, J.L. & Spengler, B.A. A novel chromosome abnormality in human neuroblastoma and antifolate-resistant Chinese hamster cell lives in culture. J. Natl. Cancer Inst. 57, 683–695 (1976). | PubMed | ChemPort |
  50. Benner, S.E., Wahl, G.M. & Von Hoff, D.D. Double minute chromosomes and homogeneously staining regions in tumors taken directly from patients versus in human tumor cell lines. Anticancer Drugs 2, 11–25 (1991). | PubMed | ChemPort |
  51. Albrecht, S., Armstrong, D.L., Mahoney, D.H., Cheek, W.R. & Cooley, L.D. Cytogenetic demonstration of gene amplification in a primary intracranial germ cell tumor. Genes Chromosom. Cancer 6, 61–63 (1993). | PubMed | ChemPort |
  52. van Dartel, M. & Hulsebos, T.J. Amplification and overexpression of genes in 17p11.2 approx p12 in osteosarcoma. Cancer Genet. Cytogenet. 153, 77–80 (2004). | PubMed | ChemPort |
  53. Skotheim, R.I. et al. Novel genomic aberrations in testicular germ cell tumours by array-CGH, and associated gene expression changes. Cell. Oncol. 28, 315–326 (2006). | PubMed | ChemPort |
  54. Feng, G.S. et al. Grap is a novel SH3-SH2-SH3 adaptor protein that couples tyrosine kinases to the Ras pathway. J. Biol. Chem. 271, 12129–12132 (1996). | Article | PubMed | ISI | ChemPort |
  55. Burdon, T., Smith, A. & Savatier, P. Signalling, cell cycle and pluripotency in embryonic stem cells. Trends Cell Biol. 12, 432–438 (2002). | Article | PubMed | ISI | ChemPort |
  56. Kang, H.B. et al. Basic fibroblast growth factor activates ERK and induces c-fos in human embryonic stem cell line MizhES1. Stem Cells Dev. 14, 395–401 (2005). | PubMed | ChemPort |
  57. Cervantes, R.B., Stringer, J.R., Shao, C., Tischfield, J.A. & Stambrook, P.J. Embryonic stem cells and somatic cells differ in mutation frequency and type. Proc. Natl. Acad. Sci. USA 99, 3586–3590 (2002). | Article | PubMed | ChemPort |
  58. Hong, Y. & Stambrook, P.J. Restoration of an absent G1 arrest and protection from apoptosis in embryonic stem cells after ionizing radiation. Proc. Natl. Acad. Sci. USA 101, 14443–14448 (2004). | Article | PubMed | ChemPort |
  59. Fluckiger, A.C. et al. Cell cycle features of primate embryonic stem cells. Stem Cells 24, 547–556 (2006). | PubMed | ChemPort |
  60. Downes, C.S. et al. A topoisomerase II-dependent G2 cycle checkpoint in mammalian cells. Nature 372, 467–470 (1994). | Article | PubMed | ISI | ChemPort |
  61. Damelin, M., Sun, Y.E., Sodja, V.B. & Bestor, T.H. Decatenation checkpoint deficiency in stem and progenitor cells. Cancer Cell 8, 479–484 (2005). | Article | PubMed | ISI | ChemPort |
  62. Amon, A. The spindle checkpoint. Curr. Opin. Genet. Dev. 9, 69–75 (1999). | Article | PubMed | ISI | ChemPort |
  63. Wilton, L. Preimplantation genetic diagnosis for aneuploidy screening in early human embryos: a review. Prenat. Diagn. 22, 512–518 (2002). | Article | PubMed | ISI |
  64. Hardy, K. Cell death in the mammalian blastocyst. Mol. Hum. Reprod. 3, 919–925 (1997). | Article | PubMed | ChemPort |
  65. von Zglinicki, T., Pilger, R. & Sitte, N. Accumulation of single-strand breaks is the major cause of telomere shortening in human fibroblasts. Free Radic. Biol. Med. 28, 64–74 (2000). | Article | PubMed | ISI | ChemPort |
  66. Andrews, P.W. et al. Pluripotent embryonal carcinoma clones derived from the human teratocarcinoma cell line Tera-2. Differentiation in vivo and in vitro. Lab. Invest. 50, 147–162 (1984). | 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. Pyle, A.D., Lock, L.F. & Donovan, P.J. Neurotrophins mediate human embryonic stem cell survival. Nat. Biotechnol. 24, 344–350 (2006). | Article | PubMed | ISI | ChemPort |
  70. Liu, X. et al. Trisomy eight in ES cells is a common potential problem in gene targeting and interferes with germ line transmission. Dev. Dyn. 209, 85–91 (1997). | Article | PubMed | ISI | ChemPort |
  71. Longo, L., Bygrave, A., Grosveld, F.G. & Pandolfi, P.P. The chromosome make-up of mouse embryonic stem cells is predictive of somatic and germ cell chimaerism. Transgenic Res. 6, 321–328 (1997). | Article | PubMed | ISI | ChemPort |
  72. 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 |
  73. Caisander, G. et al. Chromosomal integrity maintained in five human embryonic stem cell lines after prolonged in vitro culture. Chromosome Res. 14, 131–137 (2006). | Article | PubMed | ChemPort |
  74. Rosler, E.S. et al. Long-term culture of human embryonic stem cells in feeder-free conditions. Dev. Dyn. 229, 259–274 (2004). | Article | PubMed | ISI | ChemPort |
  75. Xiao, L., Yuan, X. & Sharkis, S.J. Activin A maintains self-renewal and regulates fibroblast growth factor, Wnt, and bone morphogenic protein pathways in human embryonic stem cells. Stem Cells 24, 1476–1486 (2006). | Article | PubMed | ISI | ChemPort |
  76. Kim, S.J. et al. Efficient derivation of new human embryonic stem cell lines. Mol. Cells 19, 46–53 (2005). | PubMed | ChemPort |
  77. Thomson, J.A. et al. Embryonic stem cell lines derived from human blastocysts. Science 282, 1145–1147 (1998). | Article | PubMed | ISI | ChemPort |
  78. Smith, A.C. et al. Interstitial deletion of (17)(p11.2p11.2) in nine patients. Am. J. Med. Genet. 24, 393–414 (1986). | Article | PubMed | ISI | ChemPort |
  1. Sheffield Regional Cytogenetics Service, Sheffield Children's Trust, Western Bank, Sheffield S10 2TH, UK.
  2. Centre for Stem Cell Biology, Department of Biomedical Science, University of Sheffield, Western Bank, Sheffield S10 2TN, UK.
  3. Academic Neurology Unit, University of Sheffield, School of Medicine and Biomedical Science, Beech Hill Road, Sheffield S10 2RX, UK.
  4. These authors contributed equally to this work.

Correspondence to: Peter W Andrews2 e-mail: