Mapping embryonic development

26 February 2019

Interview with

Bertie Gottgens, Cambridge University


a blue Computer Generated outline of an embryo


At conception, a single sperm and egg meet and unite their DNA. And this triggers a developmental programme, controlled by our genes, that causes the fertilised egg - and the cells it turns into - to begin to divide. Next, the ball of cells this produces begins to specialise, with certain groups of those cells ultimately turning into different bits of the body. Sometimes this goes wrong, but because we don't know what genetic programmes are running in which cells, we don't know why or therefore how to fix it. On the flip side, if scientists want to grow replacement body parts in a dish, at the moment we don't know precisely what instructions to feed to the cells to make that happen. But now scientists at Cambridge University have done the painstaking job of reading the genetic instructions that are active in every one of the 100,000 cells that form right at the beginning of the development of a mouse embryo, including at the crucial time when those cells are deciding what to turn into. Bertie Gottgens…

Bertie - What happens is that the embryo grows from a really small number of cells. It's less than 1000 cells. In 48 hours it grows to over 100,000 cells and there is this explosion of diversity. When we begin the cells are unspecified they can turn into any cell in your body whether it's muscle, heart, blood, brain etc., and then within a period of just 48 hours they make decisions of what they want to become.

Chris - Huge, huge challenge though. Embryologist have been grappling with that very issue for about 100 years, so how did you attack it?

Bertie - The opportunity arose through new technology called single cell genomics. And what that means is from a single cell we can make really comprehensive measurements of what goes on in this single cell. Before we had to use millions of cells to do the same types of measurements. Now we can do it on single cells and this technology has really only become available in the last five years.

Chris - Talk me through then what it is you're measuring and how you're measuring it?

Bertie - Each of our cells has about 20,000 genes. These are the bits of our DNA that determine the function of the cells, and what we're measuring is the activity of all of these genes. And the amazing thing is that we can measure the activity of 20,000 genes in each individual cell, and the dataset that we generated has done exactly this in over 100,000 single cells.

Chris - Essentially, it boils down to then you are looking at very early stages of development? Looking at the cells and saying what repertoire of genes are switched on and by how much in these cells and how does that change as these cells grow, proliferate and also, critically, start to turn into things, make decisions about what bits of the future body plan they're going to be?

Bertie - Yes, and this is important for two reasons. What activity profile characterizes a cell directly tells us something about the function of the cell and how this function arises from an unspecified precursor. The second point is it also tells us if we are looking at a situation where there might be a developmental disorder, what's wrong with these cells now compared to normally. Now that we have these very detailed molecular profiles we can ask those questions that before were completely inaccessible to us.

Bertie - The other issue is we want to know how does this particular cell or population of cells know in inverted commas to say become an arm or become an intestine or become a future liver, and what messages are they passing among themselves to fix them to that fate but also tell them not to become something else? So does your system now give us a clue as to what some of those messages and signals and control pathways might be?

Bertie - In essence, not yet, because what our study has provided is a baseline of reference to, in future, ask exactly those questions. Because I think in order to get a solid answer to those questions we do actually have to look at mutations where, let's say, an arm can't be formed and then say what is now different specifically in terms of gene activities? So it is an essential and very vital reference point for us to then move on.

Chris - Now mice are very similar to us but there are also important differences. So to what extent can we take the sort of reference set that you've created and say well that's how a human works?

Bertie - You're absolutely right. And there are differences between mice and human. This period of development, which in the mouse is between six and a half and eight and a half days after fertilisation, translates to between 14 and 20 days after fertilization in human. This stage of human development is inaccessible to us. We can't study this. So we have no choice, at this stage anyway, than to turn to model systems. Mouse is a good model because we have access to a lot of these genetic mutations and they're often really good copies of human developmental mutation. The second important point is where this is then directly useful is if scientists want to grow in the lab organs such as muscle etc., they need to have a reference point of how does the animal make it, and this is what our reference map landscape provides. Let's look how it happens in the animal and then design our in-the-lab protocols based on that.


Add a comment