Nature Biotechnology
22, 42 - 43 (2004)
doi:10.1038/nbt0104-42
Unnatural selection of cultured human ES cells?Martin F PeraMartin F. Pera is at the Monash Institute of Reproduction and Development, Monash University, 246 Clayton Road, Clayton, Victoria 3168, Australia. martin.pera@med.monash.edu.au Human embryonic stem (ES) cells maintained in culture can develop chromosomal abnormalities.Human ES cells, unlike almost all cells derived from adult or fetal human tissues, have the potential to provide an indefinitely renewable source of a wide range of normal cell types for use in research and regenerative medicine. The many revolutionary applications envisioned for human ES cells are based not only on their demonstrated properties of immortality and pluripotentiality, but also on the evidence to date that they maintain a normal genetic makeup during extensive propagation, expansion and manipulation in vitro. The report of Draper et al.1 in this issue raises a significant warning concerning the genetic stability of these cells. Several widely used ES cell lines—representing a quarter of those available for use in research funded by the US National Institutes of Health (Bethesda, MD, USA)—repeatedly developed specific karyotypic abnormalities in vitro when grown under established culture regimens by experienced workers (Fig. 1).
 | | |
Though the subject has not been systematically studied, it is known that mouse ES cells can develop karyotypic abnormalities during passage in culture, and that these abnormalities are, not surprisingly, associated with a decreased capacity of the cells for germ line colonization in chimeras obtained after blastocyst injection2,
3. Published results on the derivation of human ES cells have been reassuring. It is actually surprising, given the high rates of aneuploidy observed in human embryos cultured in vitro4, that no cell lines with abnormal karyotypes have been reported, and indeed, most studies have indicated that euploidy, at the level of resolution of a G-banded karyotype, is maintained for at least 30−40 passages in vitro5,
6,
7,
8,
9.
Now, Draper et al. have shown that specific abnormalities resulting in overrepresentation of chromosomes 17q and 12 appeared on multiple occasions in cultures of four separate ES cell lines; the changes emerged in most cases after extended passage (beyond passage 30 from derivation). Overrepresentation of these regions of the genome is also characteristic of the spontaneous development of germ cell tumors of the testis in man. The stem cells of these tumors are malignant counterparts of pluripotent stem cells, and it is possible that chromosomes 17q and 12 contain genes whose overexpression confers some survival or proliferative advantage to pluripotent stem cells.
The Draper et al. study does not enable us to draw strong conclusions about how general their findings may be. First, most of the data were obtained on only two lines of human ES cells. Second, the specific karyotypic abnormalities that predominated in the cultures differed between the two laboratories involved in the study. Finally, because the data do not allow us to make even a coarse estimate of the frequency per cell generation at which the abnormalities occurred, it is hard to assess how common the problem is, or to compare these results with other studies, or even to make comparisons between the two laboratories involved in this work.
It is important not to draw the conclusion that these results reflect a high intrinsic level of genetic instability of human ES cells grown in vitro. Apart from the caveats regarding the generality of the findings, it may be that the results stem from particular aspects of the cell culture methods or the period for which the cells were cultured under these conditions. Although both laboratories participating in the study used culture media incorporating a proprietary serum replacement and fibroblast growth factor 2, there were differences; in one laboratory, the cells were grown without a feeder layer and were subjected to clonal derivation, whereas in the other laboratory, a feeder cell layer was employed and cells were passaged at relatively higher density. Most, though not all, reports of human ES cell derivation have used feeder cell layers and serum-containing medium, and have not routinely attempted low-density passage of the cells.
Further systematic, interlaboratory comparison of the genetic stability of different ES cell isolates, grown under several carefully defined conditions at defined passage levels, will be required to identify both the scope of the problem of karyotypic instability and the culture conditions, if any, that predispose to it. If a culture system is in any way suboptimal, it presents a selective barrier to long-term maintenance of cells in vitro; genetic variants arising spontaneously with a growth or survival advantage may overcome this barrier and come to dominate the cell population. It is not yet known whether the cytogenetic changes observed in this study actually do confer a selective advantage on ES cells in vitro, but further study of this issue might yield important clues about genes involved in the maintenance of pluripotent cells in vitro.
The present report highlights some other very important issues. In this study, and other published descriptions of human ES cells, cell lines were examined for genetic lesions using chromosome banding or, less frequently, fluorescence in situ hybridization. Microdeletions, or minor rearrangements or amplifications, would escape detection by these techniques, as would point mutations in genes regulating growth, survival or differentiation. It will be important to apply higher resolution methodology, such as comparative genomic hybridization, to assess the genetic health of human ES cells during long-term culture, particularly as novel culture methods are developed and introduced into practice.
Another important lesson is that the karyotypically abnormal cell lines in this study still expressed canonical ES cell markers and retained the ability to differentiate in vitro, as indeed do certain cell lines derived from human embryonal carcinoma. Unlike the case of mouse cells, to which the bright-line criterion of germline chimerism can be applied, with human cells we lack robust biological markers that allow us to distinguish between normal ES cells and ES cells that have undergone minor genetic changes that might confer growth or survival advantages, or reduce their capacity for differentiation. Identification of such markers will be an important aspect of the evolution of standards for human ES cells.
Finally, no study to date has assessed the epigenetic stability of human ES cells. Changes in methylation status or in chromatin structure during culture in vitro may strongly affect the cells' capacity for differentiation. Though these issues may seem daunting, it is reassuring that mouse ES cells have been used throughout the world for over 20 years to produce normal live chimeric offspring, in the absence of any detailed or systematic assessment of the occurrence of spontaneous genetic or epigenetic alterations and their impact on developmental potential.
REFERENCES
- Draper, J.S. et al. Nat. Biotechnol. 22, 53–54 (2004).
- Longo, L., Bygrave, A., Grosveld, F.G. & Pandolfi, P.P. Transgenic Res. 6, 321–328 (1997).
- Liu, X. et al. Dev. Dyn. 209, 85–91 (1997).
- Wilton, L. Prenat. Diagn. 22, 512–518 (2002).
- Thomson, J.A. et al. Science 282, 1145–1147 (1998).
- Reubinoff, B.E., Pera, M.F., Fong, C.Y., Trounson, A. & Bongso, A. Nat. Biotechnol. 18, 399–404 (2000).
- Mitalipova, M. et al. Stem Cells 21, 521–526 (2003).
- Hovatta, O. et al. Hum. Reprod. 18, 1404–1409 (2003).
- Amit, M. et al. Biol. Reprod. 68, 2150–2156 (2003).
|
