Wolf Reik, Babraham Institute

Wolf Reik studies at the Babraham Institute at the University of Cambridge. In a recent review,1 he and colleagues discussed how pivotal epigenetic regulators nudge cells into lineage decisions. Nature Reports Stem Cells talked to him about his work and philosophy as a scientist.

How did you become a scientist?

I had already trained as a physician. I was interested in treating people, but it wasn't the kind of intellectual stimulation that I was looking for. It was more or less by accident that I went to a lecture by Rudolf Jaenisch, who was in Hamburg at the time. I was just totally mesmerized. So I visited his lab, and I asked, “Can I do a postgraduate degree here?” He said, “We don't want medics here”, knowing full well what he said because he's a medic himself, but after a little bit of talking he said, “Fine, you can try”.

And then you came to the UK to work with Azim Surani.

After finishing my degree, I wanted to learn more about mammalian embryology, so I decided to come to England because at that time that's where the major mammalian embryologists were working. I came across Azim and was hooked from the word go about this imprinting business, which was completely mysterious. He told me I had to figure out the molecular mechanism.

My impression is that Jaenisch is really excitable and Surani is really calm.

Absolutely.

What was it like working in their labs?

I was really fortunate because I learned science done in quite different ways. Rudolf gets excited. He immediately has the whole thing in his head. 'This is how it works, and these are the experiments that we need to do, and this is a big thing.' And it's great because it gives you that kind of energy to push forward, and I enjoyed that very, very much.

And then I experienced a very contrasting way of doing things. Azim is very, very thoughtful. He will think through things multiple times, and will do that with you. And that's a way of figuring out where you can go wrong. We have good ideas and bad ideas, and we have many more bad ideas than good ideas [laughs]. And one thing that's really important in doing good science is getting to the bottom of that. If you chase all your ideas equally hard, you can waste an awful lot of resources. And so the ability to talk yourself out of experiments can be very valuable.

So I think that I learned two very important things: get excited and think again.

Is that the advice you would give a new scientist?

Yes, but there's another principle. It's not all about good and bad ideas; it's about whether you like them or you don't like them, whether intuitively you agree with them. You have to keep your own excitement and motivation.

Okay, let's talk about the subject of your review: epigenetics in very early embryos. The differentiation mechanisms that are at play when the trophectoderm separates from the inner cell mass, are those the same mechanisms that are at play in further lineage decisions?

There is a paper that is very central to this review1, where our concepts come from, that [was] published last year in Nature Cell Biology by myself and Myriam Hemberger2; she works on placenta and extra embryonic lineages. When she arrived at the institute, we put our heads together.

Myriam had found that if you take mouse ES [embryonic stem] cells and you give them culture conditions so that they should develop into so-called trophoblast tissue, or placenta tissues, they won't do so. They have a block. They will only differentiate into embryonic-type tissues. It's as if they've undergone one type of decision already, which does not allow them to make trophoblast. This question started to fascinate Myriam and myself and also another collaborator, Wendy Dean.

And Myriam discovered if you used methylation-deficient ES cells, they can very happily differentiate into trophoblasts. And that made us think that it was DNA methylation in ES cells that's the epigenetic signal to lock them into that process. And we carried out a genomewide screen, MeDIP-chip [methylation DNA immunoprecipitation] kind of stuff, and we found to our great surprise and delight that there was a single gene that was methylated in ES cells and not in trophoblast cells.

During development, cells' potential to differentiate can be likened to water flowing downhill. This can occur both gradually and in cataracts, and these dynamics are influenced by epigenetic events. Credit: Wolf Reik, Nature Reviews Molecular and Cell Biology1

Was this Elf5?

This was Elf5. And our collaborative team went on to show that Elf5 is a very important part of a transcription factor network. You've heard of the Nanong-Oct4-Sox2 embryonic transcription network — Elf5 is part of a similar kind of network of extra-embryonic transcription factors, and what's unusual about Elf5 is that it's epigenetically regulated in a very black-and-white way by DNA methylation.

Is Elf5 the first in a class of gatekeepers for differentiation?

That's what we think. There could be a whole class of gatekeepers for different lineages. It's a very exciting concept I think.

How does this explain how cells decide to go into a lineage?

There's a huge amount of exciting work out there addressing that question, but we're still not completely sure how it works. It seems that in very early embryos, these master regulator transcription factors seem to be expressed stochastically. Sometimes there is a cell that expresses one of them, sometimes there's another combination of transcription factors [and so forth]. It looks almost — random is the wrong word — but we think that a stochastic process could actually be the outcome of major epigenetic reprogramming that is going on beforehand.

Because the demethylation and histone marks are being removed?

Exactly. So maybe there is a kind of ground-state situation where not many epigenetic marks are there, and so perhaps it's that ground state that lets these genes be expressed in a kind of stochastic, fluctuating way.

And then what happens? Do the right combinations somehow come up and things start falling into place?

That's the key question. I think some of the combinations, if they occur in the same cell, will reinforce each other in a kind of positive feedback circuitry.

But how does Elf5 get methylated in the first place?

This we don't know. That's the next big question. That's were the limits of the concept are at this time. How does the epigenetic system interact with all the other things that are going on at this time?

Is there a technology that the field really needs to develop?

One of the technologies that is desperately needed, which various labs are working on, including ourselves, is to be able to do the genomewide epigenetic profiling in small numbers of cells. We're using ES cells and TS [trophoblast stem] cells. We also use [these cell types] as a proxy for the the parts of a blastocyst: for inner cell mass cells, for trophectoderm cells. If you know how many cells are in a blastocyst, we're talking about tens of cells, less. The current epigenetic methods that we all use need a million cells.

Also needed is, the conditional knockout, where you deal with the wild-type situation for some time and then you hit it at the right time and in the right cell type with an enzyme that excises your gene, you go [from] a wild-type condition to a mutant condition in a very controlled, specific way. These are great, but the tools for manipulating them at exactly the right time in exactly the right cell type are still somewhat clunky. We still need to refine those tools.

Another area is live imaging. There are other very important things that go on in a cell. How are things arranged in the nucleus? How do genes move around in the nucleus? Maybe to transcription factories? And for this we need to be able to look into the cell nucleus with a very powerful microscope in a live cell.

Why don't pluripotent stem cells persist into adulthood?

There is no such thing, that we know [of], as a self-renewing pluripotent cell that is just sitting there in an embryo. They very quickly do things and they cease to be pluripotent, and so it's quite an interesting thing to ask: What's different about those cells compared to the cells that are established in culture, which have this property of limitless self-renewal?

The answer is probably epigenetic. ES cells, in order to adapt to culture and have this amazing property of self-renewal, must undergo epigenetic changes that we don't understand at this point in time. But they are probably locked into circuitry, whereas in vivo there isn't such a lock. In vivo, there are probably feed-forward loops that don't allow cells to stay in this state forever.