Caren Norden heads the Cell Biology of Tissue Morphogenesis lab at the Instituto Gulbenkian de Ciência, where she also serves as Deputy Director of Science. She obtained her Ph.D. at the ETH Zurich and did her postdoc with William Harris at Cambridge University’s Institute of Physiology, Development and Neuroscience. Dr. Norden, first-author Rana Amini, and their collaborators have recently published a paper in eLife entitled “Amoeboid-like migration ensures correct horizontal cell layer formation in the developing vertebrate retina”. The manuscript was posted as a preprint and submitted to the journal through Review Commons. We spoke to Dr. Norden about preprints, peer review, imaging, and how she became interested in neuronal migration as a research topic.
Caren Norden: Our initial interest was not in neuronal migration per se. It grew out of what I’d done during my postdoc, when I started working with zebrafish retinal development. I looked at progenitor cells and interkinetic nuclear migration, which is the movement of the nuclei during the cell cycle. We’ve kept on studying this in my lab for quite a while. But then, as is normal when you have your own lab, new questions come up.
The neuroepithelium has cells that look the same. They are attached at the apical and basal sides and the only different thing is that the nucleus is at different positions. When you have the onset of neurogenesis, the structure falls apart. You have a phase of what I call chaos because all neurons in the retina – at least in the zebrafish retina – are generated more or less at the same time. They move in different directions with different modes. The question that drives still a big portion of the lab is how you get from this chaos back to an ordered state which is the laminated retina, where all the neurons are at the right place to perform their proper function. An important part of this lamination is neuronal migration. So it was more of this question of how you get from chaos to order. To understand how you get there, you have to understand how the different neurons move through the tissue.
One thing that I really enjoyed when reading this paper is that the introduction has a deep historical background, going all the way back to Angevine and Sidman’s 1961 study of migration in the mouse cerebral cortex. To someone outside the field like me, it seems clear that radial migration is a much more tractable problem. Did you choose to work on multipolar migration, or did this come about because of some particularity of the system?
It’s a particularity of the system and our overall goal to understand the retina as a whole. In the brain, you have this relatively simple inside-out making of the layers, while here (in the retina) everything happens at the same time. This is of course much more fascinating, but also much more difficult to approach. As always in biology, you only see what you are looking for. The Miyata and Murakami labs in Japan, among others, also do multipolar migration studies. We still do one neuron at a time, but as I said we want to put them together. Our systems allows us to understand and investigate these migration modes, whereas in the cortex or other areas, you use organotypic slices, which are not so easy to get from the parts in which you have multipolar migration – and it is much harder to do meaningful quantitative live imaging.
Scrolling down the list of historical papers in your introduction is like strolling through a zoo. There are rats, mice, opossums, macaques… Why is the zebrafish a good model to address this question?
Live imaging.
Quantitative live imaging. We don’t do one movie following one cell. There is actually a prelude to this paper that was published in Development in 2019, where we established the phenomenon and show that these cells more or less move like snowflakes. We had not yet found a parameter that we can say is always the same no matter which cells reach the apical type site underneath the photoreceptors. They all get there in a different way. Then, of course, we wanted to understand the phenomenon and the interactions, and the advantage of the zebrafish is the long-term live imaging over many embryos, getting dozens of cells, and finding features that are the same versus features that differ.
One question that we now follow more and more in the lab is this question of how you make a stereotypical layered retina, and is it always in the same way. In the horizontal cells, it is definitely not always in the same way. I would bet you that if we had the ability to image and track all the horizontal cells, over five different retinas, we would not find one cell that always moves the same way. But they still always make this one neuronal layer beneath the photoreceptors.
Sorry, I diverted. The answer to your question is very simple: quantitative live imaging. You can only do it in the fish, and there you can best do it in the retina.
The live imaging in this paper is amazing. Is there anything that you wish you could do, technically or methodologically, that is currently not possible?
At some point in the relatively close future, we want to put it all together. We want to look at the single cell, but we also want to understand how the single cell interacts with the surrounding cells in the tissue and how tissue properties influence the migration of single cells or groups of cells.
The image analysis and deep learning algorithms still need to progress a lot for us to be able to really compare this within one embryo – or even between the two eyes of the same embryo. That would be my dream: to understand how all these cells move and how much variability versus how much “stereotypicity” there is. We can do it cell by cell, but not over the whole embryo. That is a technical advance that would be much appreciated.
In my dream world, we would get an even better light sheet system. But that is not the technical barrier anymore. The technical barrier is getting the data out of the movies.
Caren Norden
It sounds like the key technical advance would be more at the data processing and analysis end than in the imaging technology itself.
The imaging we can do over twenty-four to forty-two hours with the light sheet system. In my dream world, we would get an even better light sheet system. But that is not the technical barrier anymore. The technical barrier is getting the data out of the movies.
Years ago, I saw a presentation by Jennifer Lippincott-Schwartz with data from a single cell collected over a very long period. The sheer volume of data was completely absurd…
We couldn’t store all the raw data from this paper. We have the analyzed data, and that’s okay with the journals. Otherwise, it would be impossible. To give you a ballpark figure, this paper probably generated between 50 and 80 terabytes of raw data.
In the discussion of your paper there is a theme of the path of least resistance in cell migration – and an expression that I liked very much: “invade not degrade”. Why is this important in retinal development do you expect to see it in other parts of the body?
We know that this type of migration is very prevalent for cells of the immune system. We also know it happens in the brain, where it is exhibited by microglia, immune cells in the brain, which undergo very similar migration patterns. There are other areas of the brain, like the hippocampus, where there has been shown to be multipolar migration. Whether it’s the same or a little bit different we don’t know. But I don’t think that this path of least resistance is a particularity of the retina. We were just lucky enough that we could look at it in so much depth and explore it in detail. But once people start looking, I am convinced that they will also find it elsewhere.
He was a third-year Ph.D. student and I didn’t fire him, I took his advice. Since then, every paper we have we put on bioRxiv.
Caren Norden
A quick look at bioRxiv shows that you’ve been an avid preprinter. How did you start posting preprints?
Can I tell an anecdote? I started when my Czech Ph.D. student, Jaroslav Icha – the one who challenged me the most and to whom I am forever grateful – wanted to put his paper on bioRxiv. I was still on the dark side and said, “oh no then we will get scooped, or it will have flaws yada yada yada…”. I only wanted this on bioRxiv once we had the reviews, because maybe there’s something not good in the paper, or something that we overlooked and then he said, and I will never forget this, “Caren if you that think the paper is not ready, send me back to the bench. If it’s ready to submit to a journal it’s ready to submit to the archive. You have to decide if you want to be in front of the wave or follow the wave”. He was a third-year Ph.D. student and I didn’t fire him, I took his advice. Since then, every paper we have we put on bioRxiv. This happened in 2015. Since 2015, we’ve put everything on bioRxiv at the latest when it is under review at a journal.
What did you think of the quality of the reviews at Review Commons?
I think the quality of the reviews was similar to the reviews that I’ve gotten throughout my career. I don’t think they were there were any better or particularly worse. I have published a bit more than a dozen last-author papers by now, it didn’t stick out in any particular direction.
Were there any problems with the Review Commons process for you?
I don’t think there are any problems with the process at Review Commons. We got what I found to be relatively constructive and addressable comments from three reviewers. We didn’t do this thing where we first revise and then send it to the journal – we took the other road where we wrote a rebuttal to the comments. I thought that the process at Review Commons took relatively long. I know that it is very hard to get reviewers at the moment, but I think we waited six or seven weeks for the reviews. It’s within the margin but it was not particularly fast. We sent it to the Journal of Cell Biology and this was the one step that I was not very happy with because it took a very long time. It took four weeks to get an editorial rejection. Then we sent it to eLife and it was smooth sailing. Which is fine, because this is exactly where I would have sent the paper in the beginning but the first author and I were initially still undecided, so I said look let’s see what happens if we try Review Commons. But it didn’t make the process faster and the outcome was the same as if we would have had it submitted to eLife in the first place. That’s just my personal N-equals-one experience.
I was a little bit disappointed in the process because if, based on the reviewer comments and the contents of the paper, the editor had said “we don’t think it is suitable for JCB” we would have been okay with it. But getting another review by the editor which was coming up with completely different issues, I didn’t think this was the point of the whole. That being said, I’ve published very happily in JCB before! I’m a huge fan of the journal, it’s just that this experience was not particularly fruitful.
I believe in the peer-review process with all its flaws. My experiences were more often good than bad. Maybe I’m an exception. Peer review has really improved I would say all my papers.
Caren Norden
If we now have refereed and revised preprints, what is the role of journals in scientific publishing?
I’m not sure there is a real one. However, the majority of the scientific community, even though we have DORA – I don’t know how to say this in English and you’ve not lived in Germany long enough that you might know how to say this in English: Lippenbekenntnis – when you say something with your lips but not with your brain or heart.
Lip service. Paying lip service.
Exactly, lip service. Of course, we Germans have one word for it. I think there’s a lot of lip service to DORA, but I think most of us still look at the journal name. I am now a medium-career researcher. It will become important when I need to renew my grants. But for the people in my group, especially postdocs who will hit the job market, it is still seen as a part of the process. I think we are beyond the ‘get a Cell/Nature/Science paper or you don’t get a job’ stage, but we still need the quality stamp. By this, I mean Development, eLife, Journal of Cell Biology – all the journals that are under the umbrella of Review Commons can give you a quality stamp. I think the community still perceives this quality stamp as important. I am not saying it is good or bad, I just believe it still exists.
I believe in the peer-review process with all its flaws. My experiences were more often good than bad. Maybe I’m an exception. Peer review has really improved I would say all my papers, some more than others, but overall all of them.
