Sir Martin J.  Evans  (2011) - The Lability of the Differentiated State

I'm going to talk to you about the lability of the differentiated state. Well normally and I hope I will during the course explain that title, but normally of course lineages of cells and stability of the differentiated, the stability of the differentiated state has for long been a dogma in developmental biology. And of course I wouldn't be standing here, staying upright if all the cells in my body were continuously changing. The structure and function of the body depends on the integrated but autonomous action of a large number of different cells all of which are specialised, i.e. differentiated. Of course all of them developed from single egg and the proliferation is accompanied by the differentiation. Now during development and in the adult, cells do not typically change from one type to the other. Now of course having said that, development is all about cells changing from one type to the other. So immediately we can put down a marker that actually the differentiated state is not stable, it is something which changes during development. I can give you an example of this, an example from very early mouse embryo. At the top here we have an egg dividing and dividing to a clump of 16 or so cells which are loosely attached, the so-called morula, they then compact down, they start to form a cavity and eventually become this structure on the middle right which is the blastocyst, which comprises an outer layer of cells which pump liquid, different and an inner layer, so called the inner cell mass. And if we take that compacted or take the morula before it compacts, de-compact it, every cell is the same and can go back to form part of the inner cell mass and indeed part of all of the mouse. The cells on the other hand after compaction, there are inside cells and outside cells and the inside cells can do that chimerisation but the outside cells have become already specialised. If we put them into a blastocyst nothing happens. If we put them into a uterus they will implant as though they were an embryo. The inside cells on the other hand, if we put them into a uterus will just die. So you're already getting differentiation. And just trying to explain this, I took 2 very early examples, the other example is there to explain that this differentiation is due to the interaction. Now normally here the little green bit is the normal differentiation. But if you take half the embryo off as is being done here and shown in the film you will see that the green is spreading artifactually and that is a demonstration that it's the rest of the embryo that is interacting and controlling where the differentiation takes place. Now the very early stages of development it's quite clear isn't it that there must be cells whose progeny will give rise to the entire animal. What isn't self-evident is that such cells can be isolated and grow and it was the biology of mouse teratocarcinomas that led us to this discovery of such cells which were called or are now called ES cells in the teratoma, the embryonal carcinoma or EC cells. Teratomas or teratocarcinomas are really quite peculiar tumours, most tumours, most cancers are a sort of caricature of the tissue they came from, a sort of wash of rapidly growing and invading and uninteresting in many cases, cells. The teratocarcinomas on the other hand are also rapidly growing tumours, they're malignant tumours, they can be transplanted from animal to animal. But they have within them all a wide variety of different tissues, cartilage, bone, nervous tissue, skin etc, etc. And these came from work that was done by Leroy Stevens who found mice, an inbred strain of mice with these tumours. He also did work looking at where the tumours came from and showed that he could induce them, either by transplanting the primordial germ cells in the early mouse testis to another site in the mouse. Or he could do that by taking an early embryo and putting it in a different site, thus raising the possibilities that these are tumours of cells that are yet still able to differentiate. And he quite pleasantly said, following repeated serial transplantations these tumours have retained their pleomorphic character, pluripotent embryonic cells appear to give rise to both rapidly dividing cells and the differentiated cells. So he, that's as far as I know in the literature, the first description of pluripotential embryonic stem cells, it is actually a definition of a stem cell if you look at it. And Barry Pierce who was not, I should say Leroy Stevens was a mouse geneticist, Barry Pierce was a human pathologist and he was particularly interested in that observation that Leroy Stevens had made that the growing cells appear to be malignant but the differentiated cells appear to have become benign. And he did, with one of his students, the vital experiment that demonstrated that indeed it wasn't just a matter of looking at these in the histology and assuming that there was a stem cell population in there but there really was. What he did was clone the tumours, he transplanted not a lump of tissue but a single cell. And when he transplanted a single cell he was able to recover tumours with a wide variety of differentiation, hence proving that that single cell was an example of a pluripotential stem cell. Now that's roughly when I came into the picture because I was interested in seeing if I could get a developmental system in tissue culture. And of course in those days you could make cultures relatively easily from tumours. It was actually in those days very difficult to make good tissue cultures from normal cells. So I though these are tumours, no problem, we can get them and indeed I did that and very rapidly was able to isolate colonies of cells which, they could be repeatedly cloned but on reintroduction into a mouse produced teratomas. So we had the stem cells in culture. Now clearly we want to see, this was an effort to get a differentiation system, that we could study differentiation in culture. They develop and differentiate in a tumour as I've just said. Also most dramatically and in collaboration with Richard Gardner who is seen here and Virginia Papaioannou we injected some of these cells, from the tumour, grown in tissue culture, into mouse blastocysts. I told you very briefly just now that you can test the ability to differentiate a cell by putting it into a blastocyst. This was a technique that was really pioneered by Richard. And when we did this with these cells they indeed made some pretty normal mice. Pretty normal mice in which really most of the tissues had quite large contributions from the cells in culture. So these cells were able to differentiate in the context of an embryo. We'd also be very pleased if they could differentiate in tissue culture and indeed after a bit of trial and error we found techniques where they could. So you could either grow a clone of these cells and get it get extremely big and globby and overgrown in culture and then suddenly to your surprise you'd see beating muscle and you'd see nerve and you'd see magnificent differentiation. Or you could grow a mass culture, let it go through a crisis and in that crisis would produce these little clumps of cells which we knew from the tumours were called embryonic bodies, very like an embryo. And those put down again in another Petri dish would differentiate. And the picture on the right there is a section of a tumour, or actually it's a section of cells on a Petri dish. It might as well be a section of a tumour because histologically it's almost identical, you can see all the different types of tissue in it. So they did their stuff. But how did they do it. Now one of the huge conceptual breakthroughs, and it seems very simple now, absolutely you wouldn't call it a surprise but it was a real breakthrough. These were tumour cells, these were abnormal cells, these were cells that were probably pooping off suddenly into differentiated state. Suddenly we saw that actually they were going through a normal pathway. On the left here is a diagram of what happens in the normal mouse embryo, a little bit later on, the inner cell mass that I have mentioned to you before, will form all of the mouse. And if you isolate it, it will form an embryoid body like this on the bottom. Our tissue cultures formed similar embryoid bodies. Oh, now we understand that these cells are trying to be embryo cells, they're going through an embryo form of differentiation. Nowadays as I say not a surprise, of course that's the program in the genes is to go through embryo stages and go through all the interactions to differentiate. But then a surprise, oh, these are actually embryo cells, not tumour cells. Here they are differentiating. And we had a situation where we now had cells, we could get cells from the tumours into culture, we could get culture cells back into tumours, they would go through an embryoid differentiation. They could be put into a normal embryo and go into a mouse. And of course those tumours could actually be formed by transplanting an embryo ectopically. So all of those interrelationships were there and we could grow them in culture very well, we had the conditions going. So why couldn't we grow them directly from the embryo. It was about, nearly 5 years before we managed to do it. And we managed to do it together with Matt Kaufman in the picture there, it's a rather poor grainy little picture but it was, I choose to show this because this was taken at a celebratory dinner with Matt 25 years after we first isolated the cells. What we did was to take these blastocysts which are actually larger than normal, they've been delayed in implantation. Just a trick, just an embryo trick, it was helpful but it's not actually necessary. And when you put them out you can see them starting to grow out in culture and the knob of cells in the middle is the inner cell mass, that's going to grow out, I can pick it and transplant it and I will get cultures of it like that, little knobbly cultures to start with but the cells will soon flatten out rather better and become established in culture. Ok so we had those cells. Now the interesting thing is that of course we already knew that if these cells coming out from the embryo really were embryonic stem cells, really were the stem cells that we knew and loved from teratocarcinomas then all of the properties were predicted. And this is a page from my notebook at the time, where I wrote down what we needed to do, thoughts on embryo cells, it's entitled. We need to inject them into mice, well they produced beautiful teratomas when we did. It would be nice to know are they normal karyotypically, the cells from tumours typically have nearly normal karyotype, nearly normal chromosome constitution but usually a few translocations and very often slightly aneuploid. These on the other hand were, as far as we could see 100% normal. This is a male, we also had females, although the female cells, the xx cells turn out to be slightly unstable because the X chromosome is maintained active in embryonic stem cells at this stage. And that means that it's unbalanced chromosomally and there's effectively a somatic cell genetic way of getting rid of this unbalance and that is by either losing one of the 2 x chromosomes or most interestingly partially deleting up to the X enactivation centre. Can we clone them, yes no problem. In vitro differentiation, fantastic, much better than before. Here's some nerve cells, you can see beating muscle, you can see everything and there's another example of the section that we can take. So what if we do the experiment of putting them into the blastocyst. This picture is Liz Robertson I believe whose skilled hands are operating the pipette here. There they are in the blastocyst, put it into the uterus, let it grow up and here you have a chimeric mouse. This is a little male mouse that is made from a blastocyst from a pure breeding white mouse line. You can see already that you've got both black and brown coat colour and a darkened eye colour, those are just visual markers from the cells injected. But if we look internally everything is mixed. Everything is chimeric. And most importantly he and some of his colleagues also have chimerism in the testis and are able to form functional sperm. And therefore we can go from the culture right through to the normal animal and then we can breed from it. So we started off with the S cells as a development concept. Pluripotential cells in culture. They've given us at least 2 opportunities. We can look at their differentiation, study differentiation here, the original aim. We also have the opportunity to do a genetic engineering in mice as I shall show you in a moment. And what are they, I'd like to just question what are these cells. And some recent work we've done, it's taken us a long time but we have been doing it. Are there cells an artefact of culture, or are they absolutely normal cells that happen to be able to grow in culture under the conditions we've got. We've chosen, there are a number of ways of answering this but we've chosen to look at expression array of cDNA. And I just point out that Matt and I originally when we were discussing what we were doing, pointed out that actually we didn't think that we would get cells from the inner cell mass, we thought that we would probably get cells from the, immediately post implanted embryo. Numbers of markers that suggested that. So we took these stages, the bits in purple are the samples we took. A delayed blastocyst over there, diagrams of embryos here. Did a cDNA array. Very carefully set up so it's actually arrayed against a mixture of the same samples. So we're looking very carefully for differences. And if you ask the computer the result you get this sort of cladistic tree, it's very difficult to see exactly why but I notice that the, here we are, the IMT11 which is actually the ES cell line here, is closer to 51/2 day embryonic epiblast than to anything else. Now it's a bit difficult to know quite how robust the computer is but there's a very human way of looking at it through these scatter diagrams, where you're looking at log ratio against log intensity of each spot and looking for significant differences from normal or from a match. And you'll see that although here the ES cell against delayed ICM, 136 days, is not bad. So the delayed is a bit of a match. You see the most remarkable match and it really is remarkable, very few significantly different spots and many of those, I won't show you, but many of those are actually we know. And they're a bit up-regulated, some of them, because they are actually involved in maintenance of pluripotency. So probably we've been selecting a bit in culture, these are long term cultures. But we're comparing them against absolutely fresh out of the mouse embryos. So where do they come from, they actually come from there. Now the genetic engineering we can make germ line chimerism. So unlike in the normal embryo where there may be some 1 or 2 dozen stem cells for the embryo, we can now grow in culture 10's or indeed if we really pushed it, 100's of millions of these cells. Each one of which, and we can maintain them in a situation where every one of those could go through into the germ line of a mouse. That means that you can select rare variance, particularly genetic variance. The first one we did was HPRT, then that's pretty easy because you can select it, very few you can select in culture. What about insertion mutagenesis. When we started this, somebody yesterday or the day before, I think it was Oliver said, oh, you must remember that there was a time when we didn't have the gene sequences, we didn't have cloning. When we were doing this we had cloning, we had very few gene sequences. Very little genetic gene sequence of the mouse was available. So the way to go forward to find mutation was to do a mutation, mutation in these cells and then see if could see something in the mice. Here's an example where we looked for insertions that could not breed to homozygosity. We found one which is actually, which we now know as nodal, very early expression which organises the axis of the embryo. We can look for overt phenotypes. What we call bulgy eye here, I could tell you what it's all about but I. We can look for it in traps where we're looking at the expression. But of course along came Oliver Smithies, Mario Capecchi with their ability to change, target a change in the chromosome genes in cells and in tissue culture. That's the thing which was needed, we can now target anything we like, absolutely anything we like. So I'm going to give you one example out of several that we've done and of course probably, I think probably about 5,000 that have been done across the world now of experimental knock outs and designer gene changes in mice. One of the things we did was cystic fibroses. As you know cystic fibroses is an inherited disease in man and there's one particular locus, one particular gene delta F508 which is deletion of one triplet and hence one amino acid. We made mice modelled and we showed in the mice both in the null mice and in the delta F508 that we could restore function in the mouse by a fairly simple gene therapy, putting back cDNA. So can we do that in humans? The answer is yes we can and if you look at this top diagram here, the green line, there are a series of normal individuals and these are voltages across the membrane in the nose, one of them is me, other people as well. The red line is the number of cystic fibrosis patients. And the bottom diagram is the 2 sides of a patient's nose, both before treatment, the lower lines and after treatment the red lines. And you can see that it does actually work. A long way off. Now I chose this because I wanted to stretch your minds and say why is there cystic fibrosis, why do we have it, is there a Darwinian explanation, is it resistance to disease. Well possibly it could be resistance to typhoid. Typhoid gives a huge flux, if you're minus, minus CF you don't have that flux. But you have to look at selection in the heterozygous, not in the wild type. And we could find no difference, physiologically between a heterozygous and the wild type. However salmonella typhi, the human pathogen uses CFTR as the binding site and interestingly the delta F508 is expressed as a protein, goes to the cell surface and the salmonella needs a cooperative binding to bind. And here we can show, this is in our mice that of course there's no binding in the minus, minus, homozygous but it's down by about 70% in the heterozygous. Very likely, right. Now how about the human ES cells, the human ES cells have come in because really there's opportunity of using this in vitro differentiation to get any cell type you want and probably any intermediate cell type you want, possibly for cell therapies. So the scenario is that you have a patient, my patient here has had a heart attack, maybe we could take from her some cell, take from her an oocyte, put them together by cell nuclear transplantation, get an ES cell culture, not yet being done in humans, done in mice. Of course that will have to be differentiated then into a suitable precursor, heart muscle is very easy to do. And then maybe she can be cured. Well let's just look at that again, sorry did you see that. We need a bit of magic, we can't go through these embryo manipulations. And so we need to be able to reprogram cells. And there for a long time we talked about this, this is a diagram I published a long time ago, showing all the various things that we knew that were stabilising the differentiated state. And of course one of them is transcription factors, Yamanaka has shown that a cocktail of 3 transcription factors can differentiate back. Now this is interesting because we now have the situation on my diagram of IPS, induced pluripotent cells, that's taking absolutely from the end right back to the beginning. It's a colossal change in cell phenotype. Very good but what about the other small change that I suggested was blocked, that maybe can be induced. And we shouldn't be so surprised that maybe there can be trans-differentiation. We have known about trans-differentiation in development biology for a long, long time. This is a paper of 1973, a beautiful paper by Eguchi and Okada showing that you can clone the pigmented retinal cells and you can see the clone growing up, pigmented but later this culture will spontaneously change into eye lens, with all the crystal lens. You can induce differentiation, once again the older one was c-myc, you could treat fibroblasts and this was Weintraub's work, you could treat fibroblasts with a cDNA which would make them become muscle cells. There are more recent direct reprogramming, some with a single transcription factor but more with a collection of transcription factors which is really Yamanaka's big breakthrough, using numbers of them at the same time. So I wanted to finish off by saying to you the differentiated state is actually labile. It's labile in development, we now know that it's manipulable artificially. Therefore we shouldn't be talking about the stability of the differentiated state, we should be talking about stabilisation or change in the differentiated state. And I think this is bringing us a new idea into biology and ideas I think are so important. Now some advice to you. You mustn't necessarily believe anything, nothing at all, don't believe anything that any of us tell you, anything you read. You must find out. And this of course is the Royal Society's motto 'Nullius in verba', it's no good talking about it, go out and look at the evidence. And I would just like to finish, I don't know if you've ever read J. P. Donleavy's book, his eponymous hero, Balthazar B, had the motto 'expect the unexpected'. Expect the unexpected and of course Louis Pasteur said and I've got the translation here, Are your minds prepared? Have you expected the unexpected? Did any of you see that that was PC but it was not properly spelt? There, come on wake up, it was a test to see if you were awake after my talk. Thank you for listening and just wake up in the future.

Sir Martin J. Evans (2011)

The Lability of the Differentiated State

Sir Martin J. Evans (2011)

The Lability of the Differentiated State

Abstract

Many classical studies have shown that cell fates become progressively restricted during development and that this restriction is typically irreversible. This has led to the dogma of the Stability of the Differentiated State: cells cannot typically move from one differentiated state to another.
The recent reports of experimental manipulation of cell differentiation and fate by use of transcription factors must lead to a reevaluation of this concept. Retrospectively these results should not surprise us as there are many well-established examples of „trans-differentiation“, both in spontaneous regeneration and in particular experimental situations.
Differentiation is a metastable state and as we start to understand mechanism simple dogmas prove unhelpful guides.

Cite


Specify width: px

Share

COPYRIGHT

Cite


Specify width: px

Share

COPYRIGHT


Related Content