Dr Roser Vento-Tormo (5th from right) with her research group at the Sanger Institute. Credit: Wellcome Sanger Institute
Dr Roser Vento-Tormo is a Group Leader in the Cellular Genetics programme at the Wellcome Sanger Institute. Her lab’s latest work involves mapping the cells of the endometrium – the mucosal lining of the uterus – in space and time. She told us about the importance and insights of the project, and how it epitomises her love of collaboration in science.
The endometrium undergoes dynamic structural and functional changes through the reproductive cycle and is indispensable for pregnancy. It is also a key nexus of disease: endometrial dysfunction can lead to endometriosis and endometrial cancer, amongst various other conditions. Despite this, surprisingly little is known about its physiology or what goes wrong during disease. Roser’s team generated a single cell map of this poorly understood tissue in their latest research.
What got you started in science?
Growing up in Valencia I had always liked nature, and was very curious and endlessly asking questions. I guess that's why I also like research, because you get to develop your curiosity day in, day out. This early interest in nature meant I enjoyed biology in school and led me to study biotechnology as an undergraduate. Then I became fascinated by two topics – how the immune system works, and how the genome works – so I went to Barcelona and joined Esteban Ballestar's lab for my PhD, and worked on the epigenomic control of cell differentiation in the innate immune system. After my PhD I wanted to learn more computation and bioinformatics, so came here in 2016 to the Sanger to do a postdoc in Sarah Teichmann’s lab, which was a perfect place for me. Since 2019 I’ve had my own lab here.
How did you find the transition from postdoc to group leader?
It has been fun and I have learned a lot. In this transition, I think you need to learn to keep an open mind. For me, it was a challenge at first to accept something that’s now simple and obvious: people are different. You are used to working with yourself, and you think that everyone works and thinks like you, but then you realise that of course they are actually not. And it’s not bad that people are different: diversity is a strength. New perspectives help us to do new things. Here I really need to acknowledge my team, who have been fantastic from day one, and especially so in this project.
What are cellular maps and what do they tell us?
For many human organs, we still lack a lot of basic information: which cells are where in the tissue, what is the developmental origin of these cells, and how do they interact? Real world maps help you find out where you are and plan where you’re going, and it’s the same with cellular maps. To make one, we take an organ, separate the cells and individually sequence their mRNA. This tells you which genes are active, and we can use these sequences to generate a recipe of all the different cell types in the organ. Next we use an amazing new technique spatial transcriptomics, which allows us to map where these cell types are situated in the architecture of the tissue.
A good example of the power of cellular maps comes from our 2018 paper on pregnancy. We wanted to know how the function of the maternal immune system is affected by cells from the developing placenta, so we sequenced more than 70,000 single cells from the placenta and decidua of first trimester pregnancies. It revealed new cell states, and a more detailed understanding of the tissue’s layered architecture and its role in maternal immune tolerance of paternal antigens. This was the first cell map of a human organ at such a scale.
Why is the endometrium such a poorly understood tissue?
I think it’s a bit of everything. First of all, funding. For a long time, and even now, it's not been a high enough priority. But it should be, as it’s absolutely vital for pregnancy and human reproduction, and also underlies many diseases. Endometriosis, where endometrial-like tissue grows outside of the uterus, is a painful condition that affects one in ten women, and endometrial cancer is also quite a common cancer where treatment can cause infertility.
Second is the complexity of the tissue in time and space. During the menstrual cycle it undergoes dynamic cycles of shedding, regeneration and differentiation under the control of hormones, with dramatic changes in architecture happening over the span of days. The outermost lining of the uterus is an epithelium that initially exists as a flat layer of cells. It expands in the proliferative phase, and then invaginates to form glands in the secretory phase – where hormones are released - before shedding and starting again. In that sense the endometrium is like a developmental tissue that constantly changes, and that brings its own challenges but also makes it incredibly interesting for me.
Thirdly, it’s an inaccessible tissue, and you cannot use mice to study it. While we do have laboratory models like the endometrial organoids [human cells grown in the laboratory to represent the tissue], these are still very new. In fact the discoveries we made, relied on a combination of these organoids, single cell RNA sequencing technologies, and spatial transcriptomics. It might not have been possible even ten years ago.
Why did you set out to study it?
Right at the beginning of the project, when I started during my postdoc with Sarah Teichmann, we wanted to use single cell transcriptomics to profile the uterus before and after pregnancy. But we realised early on how dynamic it was, even day to day, and figured that to understand such a dynamic tissue we were going to need spatial information. That's where Omer Bayraktar and Kenny Roberts from his group came in for their spatial transcriptomics expertise. And then as we were generating this endometrial map, we wondered if we could use this reference to understand what’s going on in endometrial organoids. That’s where Margherita Turco and her PhD student Konstantina Nikolakopoulou came in.
It was quite a long road though. At the beginning, we had only a few donors and couldn't integrate the data. There's such huge heterogeneity in the tissue and you need to get to a certain number of samples just to be able to make sense of it, and also be very careful in the kinds of computational analysis you do. We are incredibly grateful to everyone who has been so generous and donated their tissue for this research. Even once we had enough data, analysing it was very challenging, as we had few clear markers to help guide what we were looking for.
But around the time of the start of the pandemic last year, everything really started coming together, mainly thanks to Luz Garcia-Alonso, the brilliant bioinformatician in my team. We spent a lot of time during that initial lockdown analysing the data.
What implications does your work have for diseases of the endometrium?
I think we can say more about the endometrial cancer, just because there's just so much more data than for endometriosis. There are hundreds of studies over decades that provided bulk RNA sequencing data of cancer samples, and we used this to ask which cell types are most enriched in cancer, including in more aggressive forms. We found one specific epithelial cell state was associated robustly with aggressiveness.
In contrast, for endometriosis we only had access to a much smaller amount of data. We are now looking for more endometriosis samples, in collaboration with a team in Oxford. It is a challenging project, but rewarding because we really know so little about it. We don't even know what the composition of the lesions in the endometrium are made up of. Simple facts about a disease that affects one out of 10 women are still unknown.
All of the data and tools you’ve produced will be publicly available – why is open science important to you?
It's important to generate knowledge, but also for people to be able to use it. The joy of a single cell data set like our endometrium one is that, depending on your background, you’ll see different things in it: I’ll immediately zoom in to the immune cells, someone else might be looking for metabolic pathways or whatever. Similarly, developing tools is great, but you have to help people learn how to use them. CellPhoneDB is our pipeline for discovering cell signalling interactions from single cell data sets, and we upgraded it for this project to include cell microenvironments. It is totally open, and while it takes a lot of work to maintain and to help people use and troubleshoot, when I see how many people have benefitted from it, wow! It’s been used for so many different problems, and often by people without computational backgrounds.
Your research is always highly collaborative – what are the benefits of co-operation?
Throughout my time here as a postdoc and Group Leader, collaboration has been key to my science. In my postdoc we worked with Ashley Moffett, a professor in Cambridge University whose perspective on the biology of pregnancy was really key to shape the questions we wanted to ask, putting our data into context [ref]. Similarly, with this current research, it’s been fantastic to collaborate with Margherita Turco, also from Cambridge and also an expert reproductive biologist, as well as Omer and Sarah.
I’ve seen the benefits of collaboration working as part of the Human Cell Atlas consortium: you’ll routinely generate data with five different groups, all of whom bring new ideas to the table. And in our team, we are very multidisciplinary: we have the people doing computational analysis, then those who are more into the tissue or cell biology, and those specialising in organoids; all complement each other. Finally, being at the Sanger really promotes collaboration: there’s a lot of freedom working here and not too much competition. You can talk to everyone about your project, openly, and you can share ideas around the campus. I do think that scientists should be more willing to collaborate than compete.
