Elizabeth K. Stancik, Ivan Navarro-Quiroga, Robert Sellke, Tarik F. Haydar
The recent discovery of short neural precursors (SNPs) in the murine neocortical ventricular zone (VZ) challenges the widely held view that radial glial cells (RGCs) are the sole occupants of this germinal compartment and suggests that precursor variety is an important factor of brain development. Here, we use in utero electroporation and genetic fate mapping to show that SNPs and RGCs cohabit the VZ but display different cell cycle kinetics and generate phenotypically different progeny. In addition, we find that RGC progeny undergo additional rounds of cell division as intermediate progenitor cells (IPCs), whereas SNP progeny generally produce postmitotic neurons directly from the VZ. By clearly defining SNPs as bona fide VZ residents, separate from both RGCs and IPCs, and uncovering their unique proliferative and lineage properties, these results demonstrate how individual neural precursor groups in the embryonic rodent VZ create diversity in the overlying neocortex.
|
|||||
Heterogeneity in Ventricular Zone Neural Precursors Contributes to Neuronal Fate Diversity in the Postnatal Neocortex6 comments to Heterogeneity in Ventricular Zone Neural Precursors Contributes to Neuronal Fate Diversity in the Postnatal Neocortex |
|||||
|
Copyright © 2013 The Third Reviewer - All Rights Reserved |
|||||
I find this article and the previous one upon which this is based (Gal et al, 2006) extremely frustrating. Both suffer from similar flaws of poor images, poor figure design, low slice counts, lack of comparable controls and poor experimental design. Since this discussion is in regards to the 2010 paper I will restrict more detailed comments to this paper.
Figure 1. – The choice of time points baffles me. Going by Takahashi & Caviness 1995 numbers on cycle times in the telencephalon, the time to label cells in m-phase along the ventricular zone should be 3-6 hours prior to the time of the elecroporation not from 2 hours prior to 6 hours after (s-phase ~ 4 hours; G2+M = 2 hours).
The authors intended to address this time discrepancy in a supplemental figure showing that not all cells are electroporated in m-phase. I like the idea of using BrdU labeled plasmid to determine where the plasmid is going immediately after electroporation(Supplemental fig). However, these pictures are horrible, of low magnification and do not indicate where the cell bodies of the electroportated cells are. Radial glial fibers extend throughout the cortical wall. Just because BrdU labeling is found in the upper vz does not mean that cells were electroporated there. For all we know the cell bodies of those cell could all be along the ventricle. It may be that electroporations can reach deeper into the vz than thought, but this paper does not provide adequate evidence of that.
Furthermore, regarding Fig 1, these pictures are too low of a magnification to convincingly show coexpression of anything. For some reason this group feels no need to display all of the channels in black in white and then showing the colored composite. This makes data extremely difficult to interpret.
Figure 2. I’m puzzled the CAG-RFP isn’t on the same graph with the other two traces. The only effect this serves is to prevent comparing the Talpha1 and GLAST promoters to the control. Shouldn’t the GLAST and Talpha1 traces average to look like the CAG-RFP? They don’t. It’s not even close.
Quantifying electroporated slices is dangerous. Three slices is enough to get you a star or statistical “significance” but quite often this statistical significance will go away after quantifying multiple slices from multiple electroporations.
Figure 3. After looking at hundreds of electroporated slices I’m very confident in saying there is no meaningful difference between these images aside from density of labeling. Again with an n of 3-4 slices you can convince yourself of anything.
Figure 4. I want to see these in the same slice. There is too much variability in timing these litters. A mating at the beginning of the pairing vs the end can easily switch from predominantly layer 2/3 to 4 fate. For this figure only 2 slices were quantified. Nice!
Figure 5. Not only are we not allowed to see the grey-scales for each channel, the imaging settings between these images is vastly different. Forget the gfp+ cells. Look at the slice, which should be identical between X and X’. These images are not comparable.
Figure 6. There is no evidence for non-symetric divisions in the svz. It’s 2 progenitors or 2 neurons. Also if I am to believe figure 2 then what neural progenitors would still be BrdU labeled in Talpha1 30hrs post electroporation if they produce post mitotic neurons?
On snp’s in general: Stephen Noctor and Arnold Kriegstein have done beautiful work time-lapsing hundreds of individual neurons (not bulk electroporations) and have yet to see one without a radial glia. In a many of their movies the radial glia narrows as the cytoplasm bulges and pulls down towards the soma, but a thin radial glia remains(see Noctor 2008). I don’t believe the quality of the images taken in this and the 2006 paper would resolve it, especially when they put arrows right over where the fiber would be (see 2006 paper). SNP’s might exist but this and the 2006 paper have provided no reason for me to believe in them. I would certainly welcome someone else’s opinion on this.
Agree or Disagree
0 
0
I agree that most, if not all of the experiments could have been done better. You raise several excellent questions about the methodology as well as the quantification. However, there must be some systemic bias in the experiments to generate significant results after a very few n. Perhaps the bias could come from over-eager researchers or perhaps there really is something there (or they could have the extraordinary bad luck to have it be truly random). Basically I think there are too many positive results to be random. This should not be construed as a robust defense of the paper. The best I can manage is that maybe what they say could be true. The researchers absolutely must design, analyze, and present their data more convincingly.
Regarding your skepticism about SNPs, I would make a couple of points. First, you refer to Noctor 2008 and point out that they never see a neuron without radial glia. These time-lapse imaging studies are impressive but are done in slice culture, a decidedly artificial situation. Perhaps SNPs are not generated in slice culture, not a very difficult thing to imagine. Second you are not convinced of the resolution in the images presented in Gal 2006. But they did serial section EM reconstruction of a cell without radial glia. How can you get better resolution than that?
-Slightly less pessimistic anon.
Agree or Disagree
0 
0
Thank you for your reply. It’s very nice to be able to discuss this with someone.
Regarding electroporations: I’ve noticed a large degree of variability between different electroporations even with the same construct. If one electroporation is just a little bit more medial or lateral than another, this can result in a very different “migration pattern.” This is dangerous because you can quantify 2000 cells in a slice and then test for significance against 2000 cells in another slice. With this many cells you are very likely to get statistical significance, but this significance would average out if you quantified more slices. Even within the same electroporation results can vary depending upon whether one looks at the center or edge of the electroporation. I’ve been heart broken many times with results that were significant after 3 slices (thousands of cells counted), but went away after repeated testing.
Regarding EM: Perhaps my opinion of the rest of the figures is increasing my skepticism of these images. However, I find it suspicious that the SNP they reconstructed was only done with DAB, while the radial glial progenitor they show was first labeled with DiI and then DAB. Also when I look at figure 5Bi (Gal et al. 2006) there appears to be a radial glial fiber coming out of the top of the SNP. If you are more familiar with looking at EM images, maybe you could ease my paranoia a bit. Any ideas on what that is on top of the cell?
Regarding Noctor’s slice culture: Perhaps there is a shift in cell fate in slice culture, but I’m skeptical that progenitors accounting for 50% of the population (figure 4 Gal et al. 2006) disappear completely. The Shen et al. 2006 article from Sally Temple’s Lab would seem to indicate otherwise as dissociated cortical neurons in vitro appear to generate the same types of neurons in the same order as in vivo. Maybe these SNP’s don’t transfect as well? That would explain the discrepancy between the two groups.
I would find their fluorescent images more convincing if they did use in vivo transfection instead of electroporation. It would be much easier to distinguish single cells in the VZ.
Thanks again for the input.
Agree or Disagree
0 
0
Thanks for your reply. The more we discuss these papers, the more I’m coming to your point of view. Going back and reading the papers more carefully, I’ve become more skeptical.
Your experiences with electroporation are illuminating. To some extent, I always expected that there would be the variability that you describe. What is surprising is the way that statistics are done for these experiments. They are counting each cell as an individual “n” and doing their tests on that. Going back and looking at their figures now, that is very obvious, given the tight error bars. I thought they would average the cells and calculate each slice as an “n.” Statistically there’s a huge difference. Doing each cell as an “n” will give a lot of false positives, as you rightly point out. Still it seems incredibly reckless to publish these data unless they were sure of what they were seeing.
About the EM, I think you just have to trust the authors, somewhat. It’s true that the methods used for the RGCs and SNPs were different, although I think at least one RGC was labeled with only DAB. It’s also true that photoconversion can be variable. However I don’t think it’s worth it to look at one panel in a figure and try to decide if they did their EM right. As a side note, it would be good if they specified that the EM was done blind (they do say that the cell counting in Gal 2006 was done blind).
Regarding the Shen 2006 paper, you could say the same thing applies as for the Noctor 2006 which is that something required for the appearance of SNPs is missing in culture. However going back and reading the Gal paper again, they actually do some slice culture experiments where they see SNPs. So they are pretty much in direct conflict with the Noctor papers. Of course, it’s entirely possible that the different results are due to differences in the culturing protocol between labs, since even small differences can have huge effects, especially in the composition of the culturing media. Still the fact that one group sees these SNPs and another does not under similar conditions is alarming.
I’m in complete agreement about doing some in vivo transfection, or infection, or whatever. Anything other than electroporation. Do you think they’ve tried other methods and didn’t see SNPs or are they just so enamored of this method that they don’t think anything else is worth it?
Also how would you explain their experiments showing that the reporter constructs driven by Talpha1 and GLAST seem to identify two morphologically different populations? Do you think they are just labeling two subpopulations of radial glia progenitors? One thing that’s troubling to me is the possibility that the reporter itself is affecting cell morphology. They are adding exogenous Tubulin-alpha1 promoters that are competing with the endogenous promoter. Presumably this decreases the expression of, um, tubulin, a cytoskeletal element that might be required for the radial glia. Still that can’t be the whole story since they see SNPs with straight EGFP reporters.
Thanks again for the discussion.
Agree or Disagree
0 
0
? ??? ??? ???????? ???
Agree or Disagree
0 
2
???????? ? ????? ??????? ??? ??? ? ????
Agree or Disagree
0 
1