IPSCs put to the test


233 Hyesoo Kim and Lorenz Studer are at the Center for Stem Cell Biology and the Developmental Biology Program, Sloan-Kettering Institute, New York, New York, USA. e-mail: studerl@mskcc.org to donor sex (higher motor neuron yield from female lines) will require independent confirmation, although it is interesting to speculate about potential Y-chromosome-linked or hormone-related factors that could affect differentiation. Three out of the 16 iPSC clones tested do not yield motor neurons under standard differentiation conditions despite expressing a normal pluripotency marker profile. However, these poorly performing clones can be rescued simply by using a stronger neural induction stimulus (a modified version of the dual SMAD–inhibition protocol developed in our laboratory10). What does this study teach us for future disease-modeling or drug-discovery efforts? The finding that iPSC lines do not perform statistically worse than a set of wellcharacterized ESC lines implies that any potential epi genetic memory related to the fibroblast origin of iPSCs does not negatively affect neural differentiation propensity. Furthermore, these data suggest that the higher variability in differentiation outcomes among iPSCs reported previously7 may be due to differences in protocols and cell-line choice rather than reflecting a fundamental limitation of iPSC technology. For example, the current study is limited to iPSC lines established using retroviral vector technology, whereas other studies included lines generated using lentiviral or episomal vector systems6. These issues stress the need for a reliable set of pluripotent stem cell lines that allows for meaningful crosslaboratory comparisons. The high reproducibility of motor neuron yield for any given cell line shown by Boulting et al.4 speaks well for using this test set of pluripotent stem cell lines as a robust research tool across the research community. The finding that neither karyotype status nor residual transgene expression affects motor neuron yield is surprising—hence, the rather bizarre result of ISL1+ putative motor neurons expressing OCT4 ESC line but rather one that is particularly efficient at neural differentiation. In fact, H9 has been used to optimize many neural differentiation protocols. Human ESCs are known to differ substantially in their propensity to differentiate to particular lineages8. For iPSCs, it is possible that the specific reprogramming technology used could affect differentiation results. For example, a recent study found that the retention or excision of a reprogramming cassette in human iPSCs led to reproducible changes in gene expression9. The paper by Boulting et al.4 represents the largest and most systematic study of these issues to date. The authors compare the efficiency with which 16 human iPSC lines and 6 ESC lines generate spinal motor neurons, a well-defined differentiated cell type (Fig. 1). The iPSC lines are selected to span differences in donor age, sex, genotype and health status (amyotrophic lateral sclerosis (ALS) patients versus healthy controls) as well as two reprogramming methods (with and without the c-MYC transgene). The authors also study the effects on differentiation of karyotypic instability and persistent transgene expression. Finally, all differentiation assays are performed in parallel in two different laboratories to control for variability in cell handling. The data demonstrate a remarkable reproducibility in motor neuron yield for any given line, suggesting that differentiation behavior is a stable trait for each human pluripotent stem cell line. Notably, a comparison of the motor neuron yield between human ESC and human iPSC lines shows no significant differences. And of all the possible sources of variability examined, none correlate with differentiation efficiency except possibly donor identity and donor sex. However, as the authors observe only one case for which donor identity significantly affects motor neuron yield, further studies will be required to address the important question of variability between lines from different individuals. Similarly, differences related The ability to reprogram adult skin fibroblasts into induced pluripotent stem cells (iPSCs) represents one of the most remarkable recent feats in the biological sciences. A particularly exciting development is the use of patientspecific iPSCs to capture aspects of human disease in a petri dish, as illustrated recently for several neurodevelopmental disorders1–3. However, for this approach to work, it is essential that disease phenotypes not be masked by technical variability associated with the derivation or differentiation of iPSCs. In this issue, Boulting et al.4 address this question head on by systematically comparing differentiation behavior across a well-designed set of human iPSC and embryonic stem cell (ESC) lines. Overall, the paper provides welcome news to the community, suggesting that the mean differentiation performance of human iPSC and ESC lines is indistinguishable and that poorly performing iPSC lines can be ‘rescued’ by simple changes in culture conditions. The advent of iPSC technology raises tantalizing questions about the molecular control and reversibility of differentiated cell fates. Of particular importance is the question of whether iPSCs represent a fully reprogrammed state that is functionally equivalent to ESCs. Recent studies in the mouse suggest that iPSCs retain a partial epigenetic memory characteristic of the somatic cell of origin, and that such memory can influence differentiation behavior5,6. Previous work on neural differentiation suggested that the performance of human iPSC lines may be more variable and on average inferior to the performance of wellcharacterized human ESC lines such as H9 (ref. 7). However, these results have been difficult to interpret as H9 may not be an average IPSCs put to the test

DOI: 10.1038/nbt.1805

1 Figure or Table

Cite this paper

@article{Kim2011IPSCsPT, title={IPSCs put to the test}, author={Hyesoo Kim and Lorenz P Studer}, journal={Nature Biotechnology}, year={2011}, volume={29}, pages={233-235} }