Unlike these previous methods the
Unlike these previous methods, the 3S reprogramming system generates partially reprogrammed iPSCs in a reproducible and predictable manner by regulating the amount of KLF4. These orexin a resume reprogramming upon increased KLF4 expression and therefore are not aberrantly reprogrammed but have paused during iPSC generation. Different KLF4 levels yield paused iPSCs at a series of stages during iPSC generation, enabling the molecular analysis of orderly events that proceed during reprogramming. Moreover, the paused iPSCs are relatively homogeneous and phenotypically stable, allowing expansion of cells for further analyses. Last, this system is based upon a defective Sendai virus, which replicates exclusively in the cytoplasm (Nishimura et al., 2011), and thus is free from both the positional effects and the silencing of transgenes that may arise in the systems based upon the retrovirus or lentivirus (Jaenisch et al., 1981). Therefore, our system to generate paused iPSCs is a significant advance over the sorted cells and partially reprogrammed cell populations previously used for mechanistic analyses of iPSC generation. It remains to be determined how the low expression of KLF4, but not the other three factors, generates paused iPSCs. It is known that KLF4 represses somatic gene expression together with c-MYC at an early stage of iPSC generation and enhances pluripotency gene expression in cooperation with OCT4 and SOX2 at its late stage (Polo et al., 2012). Therefore, reduced levels of KLF4 may be sufficient for repressing somatic cell-specific genes, but not for activating pluripotency-related genes. A recent study also showed that KLF4 organizes long-range chromosomal interactions with the Oct4 locus by recruiting cohesion (Wei et al., 2013). These chromosomal interactions are observed not in SSEA-1− cells but only in SSEA-1+ cells during reprogramming, suggesting that KLF4 may mediate the chromosomal reorganization during the transition from the SSEA-1− to SSEA-1+ cells. Interestingly, this corresponds to a stage where SeVdp(fK-OSM) generates paused iPSCs in the absence of Shield1. It is thus possible that such dynamic chromosomal reorganization requires a high level of KLF4 and presents a roadblock for reprogramming with a low expression level of KLF4. Recently, Sui et al. reported production of partially reprogrammed iPSCs using trimethoprim (TMP)-regulated destabilizing domain to control the expression of reprogramming factors from a piggyBac transposon vector (Sui et al., 2014). In contrast to our 3S reprogramming system, SOX2 rather than KLF4 plays a predominant role in regulating the progression of reprogramming. Several reasons may account for this apparent discrepancy. First, each system employed a different combination of the vector and reprogramming factors for iPSC generation. Sui et al. used only OCT4, SOX2, and KLF4 in piggyBac transposon vectors whereas our system used OCT4, SOX2, KLF4, and c-MYC in SeVdp vectors. Second, unlike the FKBP-based domain used in our system, the destabilizing domain used by Sui et al. seems to promote more complete protein degradation. Third, reprogramming may proceed via distinct pathways in each system and therefore the role for each reprogramming factor may not be identical between different systems. In fact, recent studies indicate the presence of multiple pathways by which reprogramming proceeds toward pluripotency (O’Malley et al., 2013; Parchem et al., 2014; Rais et al., 2013). Given this possibility, the availability of different systems that may generate iPSCs stalled at alternative intermediate states should be invaluable for more comprehensive understanding of the reprogramming process. Regardless of the effect of the KLF4 expression level on the progression of reprogramming, the 3S reprogramming system described here provides a powerful tool for analyzing the mechanism of iPSC generation, particularly with respect to its ill-defined intermediate stages. One such example is MET, a reverse process of EMT that occurs during normal development. Consistent with the role for KLF4 in MET (Li et al., 2010), the paused iPSCs generated in the absence of Shield1 have stalled around MET and show gene expression profiles reminiscent of the ongoing MET. Another is the primitive streak-like mesendodermal (PSMN) state, where cells transiently display gene expression profiles resembling mesendoderm (Takahashi et al., 2014), which forms part of the primitive streak during early embryonic development. Paused iPSCs generated in the presence of 10–30 nM Shield1 express mesendoderm-related genes but have not reached the fully reprogrammed state. Because the addition of Shield1 permits the paused iPSCs to advance toward the fully reprogrammed state in a controlled manner, the 3S reprogramming system may be especially suitable for analyzing the MET stage, the mesendodermal stage as well as the late stage toward pluripotency. Finally, in addition to MEF-based iPSC generation, the current system also works for human somatic cells. Thus, the 3S reprogramming system will offer a unique opportunity to analyze transient and unstable stages of iPSC generation from mouse and human somatic cells.