There are two methods for reprogramming mature cells to pluripotent stem cells, which can give rise to all cells of the body. The first direct comparison of the methods reveals that both can cause subtle molecular defects. See Article p.177
Pluripotent stem cells hold promise for disease modelling and therapeutics, because they have the potential to differentiate into almost all cell lineages. In particular, there is much interest in patient-derived pluripotent stem cells, which are genetically matched to the patient's own cells, minimizing the risk of rejection by the immune system. In the past decade, two cell-reprogramming methods have been successfully used to generate patient-derived pluripotent stem cells: cloning1 and direct reprogramming of differentiated cells to induced pluripotent stem cells, through the addition of a defined transcription-factor cocktail2. However, the molecular differences between cells derived using each method remain unclear. In this issue, Ma et al.3 (page 177) define the molecular differences between these two cell types, and set the stage for further improvements to reprogramming methodology.
Derivation of induced pluripotent stem (iPS) cells is an appealing technology, because iPS cells can be reproducibly derived from patient samples4. But comparison of iPS cells with the pluripotent embryonic stem (ES) cells generated during normal embryogenesis shows that human iPS cells are not completely reprogrammed, and reveals epigenetic differences between the two cell types5 (epigenetic marks are lingering genomic modifications that affect gene expression without changing DNA sequence). Understanding the biological origins of such differences is necessary if we are to continue to optimize iPS-cell reprogramming.
Cloning — also called somatic cell nuclear transfer (SCNT) — involves transfer of the nuclear material from a mature donor cell into an egg from which the nucleus has been removed. Pluripotent cells, called nuclear transfer ES (NT ES) cells, which are genetically matched to the donor, then arise as the egg begins to develop into an embryo. Generation of patient-specific NT ES cells from adult human cells is now feasible6. Although SCNT does not involve introducing transcription factors that have the potential to cause cancer (which is a problem with iPS cell generation), the protocol is technically difficult. It also requires donation of eggs, which creates health, safety and ethical challenges.
Ma and colleagues conducted a comprehensive molecular analysis of iPS cells and NT ES cells that were derived from the same mature cell line, and so were genetically matched (Fig. 1). They compared the reprogrammed cells to in vitro fertilized (IVF) ES cells7 — stem-cell lines derived from embryos produced by IVF, which have therefore not undergone reprogramming. The eggs used to generate IVF ES cells and those used for SCNT were provided by the same person.
The authors first focused on identifying differing large genetic duplications or deletions (known as copy number variations) in the three cell types. On average, they detected two copy number variations per iPS cell line, but this was not significantly higher than in NT ES or IVF ES cells. Furthermore, Ma and co-workers found that up to 30% of iPS cells had no identifiable genetic abnormalities, indicating that copy number variations do not always occur during iPS cell reprogramming, and that such lines can be obtained with reasonable efficiency. Although the occurrence of genetic mutations in individual nucleotides was not analysed in the current study, these results reiterate previous indications8 that deep sequencing should be used to ensure that iPS and NT ES cell lines are mutation free.
Next, Ma et al. profiled DNA methylation, an epigenetic modification that regulates gene-expression patterns9. Perfectly reprogrammed ES cells would be expected to carry exactly the same methylation marks as those generated by IVF ES cells. In genome-wide analyses, the authors found that methylation in NT ES cells was more similar to that in iPS cells when compared with IVF ES cells. The iPS cells did not carry any specific type of abnormality that was not found in NT ES cells, but rather retained many more altered methylation patterns.
Finally, Ma and colleagues studied transcriptional differences in the three cell types. They observed that incomplete demethylation patterns correlated with abnormal gene transcription in iPS cells. Again, NT ES cells were more similar to IVF ES cells, although some transcriptional alterations were apparent in both reprogrammed cell types.
What is the basis for the epigenetic differences between reprogrammed cell types? Compared with iPS cell generation, the reprogramming cues dictated by the egg in SCNT are more natural. This probably explains the relatively more normal DNA-methylation patterns in NT ES cells. Notably, however, the authors detected DNA-methylation defects that were specific to NT ES cells at some small genomic regions. Most instances of abnormal methylation in NT ES cells were attributable to incomplete reprogramming, leading to epigenetic traces of the mature cell being retained. By contrast, most methylation aberrations found in iPS cells arose owing to abnormal addition of methyl groups during the course of reprogramming. Clearly, neither reprogramming method is perfect.
To build on these intriguing insights, future studies must analyse the reprogramming of mature cell types in a larger cohort of donors. This will be necessary to better understand to what extent epigenetic variance is brought about by differing reprogramming techniques. Testing alternative transcription-factor cocktails may also improve iPS cell quality — it has been found that the combinations or relative amounts of transcription factors used during the generation of mouse iPS cells influence the cells' epigenetic properties10.
Importantly, one possible explanation for the improved reprogramming outcome observed with SCNT is that the protocol routinely involves using inhibitors of the enzyme histone deacetylase, which facilitate reprogramming by preventing the removal of certain epigenetic marks and so promoting chromosome decompaction. Finally, it will be interesting to test whether recently devised11 non-conventional cell-growth conditions (which alter the epigenetic state of human pluripotent cells, leading them to adopt a more naive state) might alleviate certain defects.
In summary, there is an abundance of factors that can be used to reprogram cells and expand them in vitro, and each can influence the epigenetic and functional properties of reprogrammed cells in distinct ways. This complexity disrupts simplistic attempts to define and obtain 'perfect' stem cells. Ma and colleagues' valuable findings may prompt scientists and clinicians to redefine their quest for perfection as a hunt for the adequate, cost-effective and safe.