We’ve just published a paper in eLife describing a new way to produce three-dimensional images of neurons and blood vessels in the mouse eye. We used a method called light-sheet fluorescent microscopy (LSFM) to quickly capture 3D images of live and static mouse retinas.
We’ve used the technique to show that abnormal blood vessels have a very different shape in a mouse model than what was previously described. Rather than appearing like bulbous blobs that had been flattened, being undistorted we saw they are knotted, interlaced vessels.
Image of mouse eyeball taken with light-sheet fluorescent microscopy, with the blood vessels shown in green.
- Prahst et al: eLife paper 2020It’s the culmination of seven years’ work for my lab and collaborators, and should be really useful for studying models of eye diseases and helping to uncover how they arise and develop. Ultimately, this could lead to a wealth of new treatments.
It started when I ran my first lab at the Beth Israel Deaconess Medical Center at Harvard Medical School.
I was studying abnormal vessel formation in mouse retinas, but was getting frustrated with the standard imaging technique, confocal microscopy. With that method you have to slice the retina and then mount it on to a flat slide. But because retinas are naturally curved, this slicing and flattening distorts them.
I’d heard about LSFM and how it can image whole tissue samples and thought we should give it a try. It was a risk, as it was so different to conventional practice, but it worked great! From there the project really took off.
We used LSFM to look at a well-studied model of eye disease in mice (oxygen-induced retinopathy). We saw that groups of abnormal blood vessels, which are called tufts, had a very different shape and structure to that currently described.
Image of whole mouse retina taken with light-sheet fluorescent microscopy, with the blood vessels shown in green.
- Prahst et al: eLife paper 2020This was a massive surprise! Under other imaging techniques, these tufts have been distorted due the need to be flattened to view from above on slide, which made them appear to be bulbous blobs flattened to a pancake. With LSFM, there’s so much more 3D detail as we can view them from all angles for the first time. We saw they appear to be knotted or swirled interlacing vessels, sometimes coiled like a snake, with many holes going through them, pointing to a very different mechanism driving their formation in eye disease than currently considered.
In human eye disease these tufts are known to pull on the retina leading to detachment and blindness. Therapies that can cure tufts are lacking; so understanding more about how they grow is critical to limiting blindness. For the first time, with this new 3D imaging method, we can study in full how and why these abnormal vessels form.
The new imaging method can also help future studies of abnormal structures growing in many other eye disease models - in particular enabling collaborations between specialists in blood vessels and specialists in neurons. This is because with LSFM, unlike other techniques, you’re able to image both vessels and neurons in the same sample, at the same time.
3D reconstruction of the vessels (red) and neurons (green) in a mouse retina.
- Prahst et al: eLife paper 2020It works by turning a laser into a sheet of light which reaches across a whole horizontal section of a sample, which could even be a whole eye, without needing to physically slice the sample. This makes it so much quicker than confocal imaging, which works more like an old scanner, going back and forth point by point to record a single section.
It’s been really collaborative in many ways. Firstly, it’s been an international effort, as I’ve worked between a few different institutes over the course of this project, including Harvard Medical School, Boston University, Uppsala University in Sweden and the Crick. Additionally, our collaborators at the Instituto de Medicina Molecular (iMM), in Lisbon were instrumental to the project.
New blood vessels formed into tufts in a mouse retina, taken with light-sheet fluorescent microscopy.
- Prahst et al: eLife paper 2020Secondly, it’s pulled together people from so many different disciplines including experts in LSFM, who have helped train us in the technique, neuronal eye disease experts at Mass Eye and Ear Institute working together with vascular biologists, as well as computational biologists, like myself, who worked on the image analysis and dataset challenges alongside performing the first imaging in our wetlab.
Within the Crick, the Electron Microscopy team and the Advanced Light Microscopy team were invaluable to the project and while my new lab was still establishing here they imaged a whole load of mouse retinas for us! Our in-house scientific illustrators also helped visualise the knotted, coiled tuft structures.
We’ve already started a follow-up study, to confirm exactly how these abnormal blood vessels form and grow. We’ve seen some interesting properties that we’d now like to understand, for example, as the abnormal groups grow from 20 cells to bigger groups, the number of connections to the vasculature beneath rapidly increases, suggesting smaller tufts merge to make the bigger ones at a critical size. We also saw very curved nuclei and want to understand why the vessels might be coiling upon themselves like a snake - and crucially explore ways to untangle these coils or prevent small tufts merging as a new therapies.
We’re also working hard on writing an automated image analysis toolkit that other researchers can use to analyse the big data 3D images of curved eye tissues that LSFM produces. We’re going to make this open-source so eye researchers everywhere can benefit from the advantages of LSFM.
