the community site for and by developmental biologists

2 thoughts on “A glimpse of a dynamic cell fate decision”

  1. Very stimulating reading!
    I would however “disagree” on what you described as “phenotypic plasticity of cells”, which is based mainly on the “transcriptomics” plasticity captured by single-cell RNAseq. You’ll agree that, functionally, cells are phenotypically very robust, in particular what we call “differentiated” cells (that’s why “we” function!). Whether the “observed” transcriptomics variability is a technical problem or an irrelevant “problem” for the cells, is still an open question. Does it make any difference (energetically) for the cell to produce 100 or 300 mRNA copies of a given gene? Even if it does, buffering at the protein level will make this irrelevant, as the ratio protein/mRNA is very high (see 10.7717/peerj.270). Because we are limited to “transcriptomics” and cannot do (yet) proteomics at single-cell level, we should be careful to prevent misleading generalizations.

    And yes, stem cells are more “plastic”, transcriptionally and “phenotypically”, and that is what makes it very difficult to isolate these cells as a single “entity”. Your work contains some of the best experimental evidence for this unique stem cell features. The challenge is to understand how the transcriptional plasticity repeatedly found in stem/progenitor cells is central to their identity/function, and how is this “noise” generated/regulated? I guess that my question is whether stem cells know who they are, or whether they are an extreme case of multiple personality (dis)order?


  2. You are indeed raising fundamental questions! Although we have no definite answer for them, I think that we can say a little more than what you suggest. It is absolutely clear now that the observed transcriptomic variability is real and not just a technical artefact. We are also know, that the proteomic variability is not less. Just consider a cytometry plot: the difference between the protein levels in cells of the same type may reach up to two or more logs! There are many different reports that demonstrate and quantify transcriptomic and proteomic variability. May I suggest you one of ours?
    You can see that the protein levels may change several times during the same cell cycle!

    A very interesting recent study has demonstrated (Lestas et al. Fundamental limits on the suppression of molecular fluctuations Nature 467, 174–178 (09 September 2010) doi:10.1038/nature09333) that the cell’s capacity to suppress these fluctuations is very limited. Since it is impossible to get rid of it, the “noise” is presumably part of the normal functioning. I think that the best interpretation of our observation is precisely this: external stimulations induce a kind of non-specific “ready to respond” activated state where the increased intracellular noise generates an essentially random gene expression pattern. This is indeed a kind of “multiple identity state”. The relaxation from this “activated” state to a less noisy stable phenotype is gradual and accompanied by slow fluctuations. A number of papers came to similar conclusion. I suggest two recent reports:
    The key question therefore is to understand the mechanisms of “noise modulation” as a basis of “plasticity”. However, I think that there are some conceptual issues need to be settled before we can consider cell plasticity: What is a cell phenotype? Can it be described by the transcriptome? By the proteome? By cell morphology? How long a phenotype has to be maintained to be considered as stable?



Leave a Reply

Your email address will not be published. Required fields are marked *