As a differentiated cell proceeds toward the pluripotent state, markers turn on and off in an orderly fashion.
What happens inside a cell that is being reprogrammed into an induced pluripotent stem (iPS) cell? Investigating this question is especially difficult in the human system, where it is unclear how to prospectively identify the rare cells that are on the correct path to pluripotency or even cells that are fully reprogrammed. On page 1033, Chan et al.1 begin to tackle the question by carefully characterizing in real time several markers that are silenced and activated during the dedifferentiation of human fibroblasts to iPS cells. They find a combination of cell-surface markers and cell-growth characteristics that can be used to predict fully reprogrammed iPS cells.
Reprogramming of differentiated cells to iPS cells is a stochastic process2, but those cells that reach the pluripotent state (the capacity to make all cell types in the adult body) seem to progress through several milestones in an orderly fashion. This order has been most clearly studied in mouse cells, where there is a larger set of markers that can be followed in live cells3,4. These markers include reporters introduced into the genome at the Fbx15, Oct4 and Nanog loci and cell-surface markers such as SSEA-1, which were chosen based on their known expression in mouse embryonic stem cells2.
In human cells, the markers of pluripotency that can be followed in live cells are less well studied. Lowry et al.5 showed that one could identify promising colonies using the TRA-1-81 antibody on living cells. However, most groups continue to use colony morphology to identify potential human iPS cells. Unfortunately, many morphologically promising colonies turn out not to be fully reprogrammed when evaluated molecularly (by gene-expression and epigenetic profiles) or functionally (by in vitro differentiation or teratoma assays). Therefore, having additional surrogate markers that could more accurately identify fully reprogrammed colonies would be very useful to the field. Moreover, a detailed understanding of the order of marker activation could lead to methods for isolating cells at intermediate stages to begin to characterize the 'process' of reprogramming.
In their paper, Chan et al.1 perform the herculean task of characterizing a series of markers by following thousands of cells over the lengthy course of reprogramming (3–4 weeks) with a combination of live cell staining and real-time microscopy (Fig. 1). They follow the three cell-surface markers CD13, SSEA-4 and TRA-1-60. CD13 is expressed by fibroblasts, and SSEA-4 and TRA-1-60 are expressed by human embryonic stem cells. As the markers are on the cell surface, the cells do not need to be fixed and permeabilized, enabling live cell staining. They also assess the intensity of nuclear staining with Hoechst dye and silencing of the four reprogramming transgenes (OCT4, SOX2, KLF4, c-MYC) using viral vectors that co-express GFP, which similarly can be followed without destroying the cells.
Their analysis of these markers reveals a consistent pattern among those cells that go on to be fully reprogrammed. CD13 is silenced early; next, GFP expression is silenced and SSEA-4 and TRA-1-60 are activated. Hoechst staining, which is bright early, is transiently reduced at late stages in the reprogramming process, providing an additional landmark. This latter finding presumably occurs because Hoechst dye is actively pumped out by stem cells. However, these markers are not enough to identify iPS cells, as some cells that are CD13−, GFPdim/−, SSEA-4+, TRA-1-60+ and Hoechstdim fail to expand. But these markers combined with characteristic growth are sufficient to identify those colonies that are highly likely to go on to produce 'bona fide' iPS cell lines, as confirmed by expression of multiple pluripotency genes, demethylation of the OCT4 and NANOG promoters and the ability to form well-differentiated teratomas when injected into immunocompromised mice.
The authors call these colonies “type III colonies.” Two other colony types that are morphologically similar to the type III colonies come out of their experiments (Fig. 1). Like type III colonies, type I and type II colonies are expandable, but neither silences GFP. Furthermore, type I colonies do not activate SSEA-4 or TRA-1-60, and type II colonies express SSEA-4 but not TRA-1-60. Molecular and functional analysis of type I and II colonies show that they are not fully reprogrammed. For example, neither produces the well-differentiated teratomas one normally sees with human embryonic stem cells.
The authors conclude that type I and II cells are stuck in incompletely reprogrammed states, although type II cells do rarely spontaneously convert to type III cells. Why do these cells get locked into incompletely reprogrammed states? Hints come from previous work done in the mouse system. Mikkelsen et al.6 found that incompletely reprogrammed mouse colonies continue to express lineage-specific genes and show abnormal DNA methylation of the promoters of key pluripotency genes. Manipulation of ongoing lineage-specific gene expression (by siRNA) and of DNA promoter methylation (by the DNA methyltransferase inhibitor 5-azaC) reverted some of the incompletely reprogrammed cells to 'true' iPS cells. Interestingly, Chan et al.1 also observe inappropriate hypermethylation of the endogenous promoters of OCT4 and NANOG, two key pluripotency transcription-factor genes, in their type I and II colonies. Therefore, inhibition of DNA methyltransferases might similarly help convert these colonies to type III colonies.
What enables incompletely reprogrammed cells to self-renew (that is, grow indefinitely while maintaining their identity)? One culprit is the exogenously introduced c-MYC. c-MYC is a well-known oncogene that can induce many cell types to inappropriately self-renew. Indeed, removal of c-MYC from the cocktail of reprogramming factors dramatically decreases the number of incompletely reprogrammed colonies in the mouse and human systems7,8,9. Therefore, it might be expected that the relative number of type I and II colonies would also be significantly reduced in the absence of c-MYC.
The ability to use combinations of markers to identify cells at different stages along the road to pluripotency is a valuable resource in efforts to understand reprogramming at the molecular level. Although we know something about the beginning and end of reprogramming, the process in between remains largely a black box. Using specific combinations of markers, one can potentially sort out cells at various intermediate stages and determine their transcriptional and epigenetic profiles. Such studies will no doubt shed light on the molecules and pathways involved and help identify the barriers that must be overcome in the making of an iPS cell.
Chan, E.M. et al. Nat. Biotechnol. 27, 1033–1037 (2009).
Hochedlinger, K. & Plath, K. Development 136, 509–523 (2009).
Brambrink, T. et al. Cell Stem Cell 2, 151–159 (2008).
Stadtfeld, M., Maherali, N., Breault, D.T. & Hochedlinger, K. Cell Stem Cell 2, 230–240 (2008).
Lowry, W.E. et al. Proc. Natl. Acad. Sci. USA 105, 2883–2888 (2008).
Mikkelsen, T.S. et al. Nature 454, 49–55 (2008).
Nakagawa, M. et al. Nat. Biotechnol. 26, 101–106 (2008).
Marson, A. et al. Cell Stem Cell 3, 132–135 (2008).
Judson, R.L., Babiarz, J.E., Venere, M. & Blelloch, R. Nat. Biotechnol. 27, 459–461 (2009).
About this article
Cite this article
Subramanyam, D., Blelloch, R. Watching reprogramming in real time. Nat Biotechnol 27, 997–998 (2009). https://doi.org/10.1038/nbt1109-997