Completed on 30 Jan 2017 by Alfonso Martinez Arias . Sourced from http://biorxiv.org/content/early/2017/01/23/102376.
Login to endorse this review.
Author response is in blue.
Very interesting study that complements well the work from the Siggia and Brivanlou groups (e.g Warmflash et al. 2014) referred to in the text. What catches my eye is the requirement for Nodal which is not reported in Warmflash et al. though in that work they use conditioned medium which may contain some TGF-ß. The more controlled culture conditions of this study allow details like this to come out. As indicated by the authors the need for Nodal is likely to be related to its requirement for T/Bra expression; but there is also a requirement for BMP i.e. the full pattern requires both Nodal and BMP. I wonder if the authors have tested a requirement for Wnt in the process.
The current views of the patterning of the mouse epiblast (the main reference we have for human embryos) suggests that BMP initiates the expression of Nodal and Wnt3 in the epiblast and that Wnt3 increases the expression of Nodal, as well as its activity by regulating the expression of the coreceptor Cripto. Then, both Nodal and Wnt signalling lead to the expression of T/Bra. Observations in hESC differentiation in culture suggest that this relationships might be transferable to human embryos or gastruloids. So, just wonder if the observation of the interdependence of BMP and Nodal does not have Wnt signalling in the middle. Perhaps, in the experiments BMP signalling is not strong enough to activate Nodal expression, thought it activates enough Wnt3 to get the system ready for Nodal. Nodal would then kick in its autoregulation which would be sustained by the Wnt induced by BMP. If this were the case, inhibition of Wnt signalling will alter, and perhaps abolish, the patterning event. Also, Wnt signalling might be able to substitute for BMP i.e. the pattern could be driven by Wnt and Nodal. Just wonder if the authors have looked at this. Alternatively, it could be that the circuitry differs between mESC and hESCs as suggested by Etoc et al (2016) for the case of BMP and Noggin.
A second issue concerns the evidence provided in Figure 5E for periodicity as predicted by a change of the length scale in a Turing system. Is that spotty pattern rather than concentric rings, what one could predict in two dimensions?
Thank you very much for your comments Dr. Martinez Arias!
We found the involvement of Nodal in this fate patterning interesting as well. Although Nodal signaling is necessary for the induction of BRA expression we found it was unable to induce the differentiation, and the morphogenetic changes we observe in the colonies in the absence of BMP. We note that Warmflash et al did show Nodal inhibition mediated abrogation of BRA in their study (Fig 5d in their report) – although they didn’t explore the ability (or inability) of Nodal to induce the morphogenetic changes observed within these colonies. Furthermore, in the following study from the Brivanlou and Siggia groups, Etoc et al also implicate a need of SMAD2 in the fate patterning.
Your comments about Wnt induction are extremely perceptive. We do, in fact, see a requirement for Wnt in these patterning events. Inhibiting Wnt activity by either stabilizing Axin or inhibiting Porcupine abrogates the BRA positive region. Additional studies relative to this observation are underway.
With respect to the larger patterns, we have not observed concentric rings. We note that our simulations suggest that the rings evolve to a spotted pattern over “simulation time”. How this time relates to the molecular and cellular dynamics of the system requires further investigation.