Peter Rathjen was educated at the University of Adelaide. He was awarded a D Phil from Oxford University in 1989 for research into 'jumping genes'. He returned to the Department of Biochemistry at Adelaide in 1990 as a lecturer, and was appointed to the Chair of Biochemistry in 1995. In 2000 he was elected Head of the Department of Molecular Biosciences and in 2002 was appointed Executive Dean of the Faculty of Sciences. His primary research interests are in early mammalian development and the mechanisms by which stem cells give rise to cell lineages during embryogenesis. Basic aspects of this research are carried out within the ARC Special Research Centre for the Molecular Genetics of Development where he is stream leader in cell differentiation. Application of his research to the emergent field of stem cell therapy has been a major undertaking that has included commercialisation of intellectual property in association with BresaGen. He is a member of the Australian Stem Cell Centre.

Stem cell therapies – from cutting-edge research to clinical application
by Professor Peter Rathjen

I have had trouble working out where to pitch this talk, for a number of reasons. When I was asked to give it, Bob Williamson said, ‘Look, it’s pretty easy. You’ll be the link talk between the scientists in the morning and the ethics and societal implications in the afternoon. So just find some strands that were common in the first few talks and pull them together, and help people to work out what it was that was talked about, to set up the next session.’ That has to be done in an environment where we have got probably most, or all, of Australia’s best stem cell biologists in the audience, but also other scientists who don’t work in this area, and people who are not scientifically trained at all.

I have tried to find a middle ground in terms of where I am going to go with the talk. I gave myself a very ambitious topic, ‘from cutting-edge research to clinical application’, and in fact most of what I am going to talk about today will be the first part of that. I want to look at the cutting-edge research which is being done; I want to look at why it is important; and I won’t talk terribly much about clinical application, in part because it was covered pretty well this morning by people that understand much more about it than I do.

(Click on image for a larger version)

The main thing that needs to be understood about stem cell therapies is that the technology and the biological understanding are at an extremely early stage.

Perhaps the easiest way I can try and show that is to point out that I no longer am quite certain what stem cell therapies really are. This is a point that came out very nicely from Melissa Little’s talk this morning.

There are a number of approaches, any one of which may turn out to be valid, which could equally be termed stem cell therapies. They are really united only by one common theme, that they are all attempts to treat diseases that are caused by some sort of cell deficiency – a cell population that has died, perhaps, or a cell population that is dysfunctional, maybe because of mutation. They are a set of diseases which are particularly attractive for treatment, if you like, because very often they are diseases which are more about quality of life than they are about perpetuating life that maybe has come to a natural end. If you get spinal cord injury, for example, you can be paralysed for decades. The other thing, of course, is that the sorts of diseases we are talking about – diseases associated with cell dysfunction – are increasing in frequency in Western civilisations, as populations age.

So the concept that diseases are caused by cells that are not working properly, or are absent, and that they should be able to be treated, I think is a fair description of what stem cell therapies are. But the ways in which we might go about treating those are many and various, and in many cases they pertain in some respect to our knowledge of stem cells somewhere around the place.

The previous talk, about transplantation biology, was clearly looking at what is shown at the bottom of this slide. It was about the idea of generating the cells somewhere and transplanting the cells back into the location in the body where they are deficient. For a Parkinsonian sufferer, for example, you would make neural cells and transplant them back into the brain.

Already, though, you will see that we are not even very certain where those cells are going to come from. There is more than one kind of stem cell that might turn out to be important here. The one I am going to talk about most in this talk – Martin Pera has already set it up for me this morning – is the embryonic stem cell, the granddaddy stem cell that can give rise to any of the cell types that are found in the body.

But other researchers are out there looking at another source of cells which might be equal or even better in terms of their promise. These are adult stem cells, the cells which seem to be cropping up in many different locations in the body and many different tissues, and which again have a developmental potential which suggests that you could use them to make cells which in turn could be transplanted.

But if you think about what Bob Graham was saying this morning, and if you think about what Perry Bartlett was getting at with the neural regeneration, you realise that there is a suggestion that perhaps we won’t be transplanting cells at all. It is becoming increasingly apparent that the body has rather more regenerative capacity than we originally gave it credit for – when tissues are damaged within the body, there are signals within the body that try to take advantage of the adult stem cells, or the endogenous stem cells, that are there, and use them to replace the cells that have gone missing or become dysfunctional.

So perhaps we won’t be transplanting cells from the outside into bodies at all. Perhaps we will be taking advantage of the body’s own endogenous stem cell population, using regenerative signals that are active in the body or even inductive signals that are just there during ordinary homeostasis – Perry told us that the cells in the brain are being replaced on an ongoing basis. The administration of those signals alone, perhaps in the form of proteins, might be sufficient to replace the cells that have gone missing here.

And there are all sorts of different in-between possibilities. One might imagine that a combination of cells with some sort of signals that act on those cells might also be a profitable approach to replacing cells that have gone missing.

So, to me, it is not even clear exactly what it is that we are trying to do. As it becomes less and less obvious which of these is likely to be the most profitable route to a future therapy, so it becomes more and more obvious that everything we do – all of the early results in the early clinical trials, some of which will look good and some of which won’t – all of it needs to be underpinned by the fairly steady, incremental accumulation of basic knowledge which in turn can act to inform the future therapies, which we would expect to be more and more successful.

(Click on image for a larger version)

I am going to talk about the basic properties of stem cells. For many of you who were here this morning, this will be repetitious, but it is important to be quite clear about what it is we require of the cells we are talking about.

Most of the examples that I will use – I won’t show great amounts of data in this talk – come from my own laboratory, which works with embryonic stem cells. In fact, conceptionally they are probably the easiest of the stem cells to understand, because we have a lot of knowledge of what it is that they normally do in the organism itself. One of the problems we have with the endogenous stem cells, or the adult stem cells, is that very often we are not quite sure what their true biological role is in the homeostatic organism.

This slide shows a very obvious thing, the first third or so of mouse embryogenesis, starting from the fertilised egg which you saw so beautifully in the previous talk. In fact, it is very important to know that if you superimposed a human embryo over this you would see something that looked really quite similar.

What you are looking at is the process by which the organism establishes itself. At a cellular level there are three things going on, really. There is a very carefully controlled increase in the number of cells, from a single cell to something of the order of several trillion cells in the adult human. There is an increase in cell diversity – we start life as a single kind of cell and we finish up with at least several hundred different kinds of cells in our body. The cells knew when and where to differentiate, or turn into another kind of cell.

Then there is a third process at work, which I will get to slightly in this talk: a process of organisation of the right number and the right kinds of cells into a structure which looks like and behaves like a human or, in this case, a mouse. Cells not only come from a particular lineage; they are not only of a certain cell type. They also, in certain respects, understand their position within the body.

What we have here is not a stem cell, and it’s not a stem cell because it is not able, or it doesn’t appear so at the moment, to continue to reproduce itself. But all of the things I have said about that cell and how it turns into an organism are equally true of this structure at the top right-hand of the slide. It is what we call a blastocyst, and in the mouse that is at about the stage where the embryo implants into the uterus of the mother.

You can see from the colours that there are already three cell types present in the blastocyst. Only one of them is interesting for this talk: the blue cells on the inside. We call them at this stage inner cell mass cells. As best we can tell, there is a small number of them, 10 to 20 at this stage; they appear to be identical; and any one of them can go on and give rise to any cell that will later be found in the embryo or the adult.

So the story of embryogenesis is in fact a story about how those 10 to 20 identical cells give rise to a complete organism. Those cells are the true founder cells of the mammal, and we call them the stem cells.

(Click on image for a larger version)

You will have seen from Martin Pera’s talk that we can isolate those cells and grow them in the laboratory. And when we grow them in the laboratory we no longer call them inner cell mass cells, which is what they are here (the blue cells inside the ES cells) but we call them embryonic stem cells.

I want to go through this slide, just quickly, because it enables us to look at the key properties of these cells which underpin how it is that we think we are going to be able to use them therapeutically. While this talk is about embryonic stem cells, adult stem cells that will be useful will also have very similar properties.

The first thing is that they are immortal. That is a rare property for a cell that we can grow in vitro. Cancer cells are immortal and not therapeutically very useful at the moment. Embryonic stem cells are apparently immortal, in terms of the length of time that we can grow them and in terms of the genes that they express and the properties that they display.

Why does it matter? It matters because if we are going to transplant cells to cure disease, we are going to need very large numbers of them, in many cases. Embryonic stem cells continue to divide – in the case of mouse embryonic stem cells, which are shown here, they continue to divide very rapidly – and so it is relatively straightforward to grow enormous numbers of these cells in the first place. If you just keep growing them, they will just keep dividing and will give you more and more embryonic stem cells.

The second thing that is important about embryonic stem cells, and arguably important about adult stem cells, is that they maintain the property of pluripotence. It is the exact property that they have in the embryo, and it simply means they can differentiate into every other kind of cell.

It is not very clear to me yet how this property pertains to adult stem cells. It sometimes looks as though this transdifferentiation concept might be correct. Perhaps you can take a stem cell from a tissue in the adult, and perhaps it can truly turn into all sorts of other cell types. It is also true to say that very often when these reports appear in the literature and are greeted with enormous excitement, because it would be wonderful if you could do such a thing, it also often turns out to be the case that the initial reports have missed something, and that what appears to be pluripotence in fact isn’t, and that the adult stem cells have a rather more restricted differentiation capability. In my mind, the jury is still out on these things.

But in terms of embryonic stem cells, those two simple properties give you the first part of what is stated at the bottom of this slide: the fact that an ES cell colony can be considered as an unlimited number – because it is immortal – of any kind of cell, because it is pluripotent.

There is a second point which actually doesn’t get talked about very much in terms of embryonic stem cells but is really quite important. That is, they are particularly suited to genetic modification. Genetic modification of mammalian cells is not trivial. It is not too bad in terms of bacteria in yeast, and things like that, but it is hard to change the genome of a mammalian cell with precision.

In the case of embryonic stem cells, decades of gene targeting have taught us that in fact we can do that – we can add genes at particular locations if we wish to do that; we can mutate genes, if that is what we want to do, to stop them working; and we can actually make quite subtle alterations to small numbers of nucleotides amongst the three billion-odd nucleotides that make up the chromosomes of the mammal.

In fact, a Fellow of this Academy, Bill Elliott, lectured me in Biochemistry II in 1982 (I hate saying that; it’s a long time ago now) and made the point as he was teaching us about sugar metabolism that yeasts, when they metabolise sugar under anaerobic conditions, give rise to ethanol, where we give rise to lactic acid, and that the main difference between those two is the presence of an enzyme in the yeast which causes that reaction to occur. And, trying to be fairly flippant at the time, he made the comment that if we could only find a way to get that gene into a human cell, then it should be possible to run up the stairs, get yourself exhausted, and get drunk at the same time. We found that amusing at the time. In fact, I think it is something which arguably could be tried, if that was what people wished to do, with embryonic stem cells.

(Click on image for a larger version)

The important point here is that the properties we seek for the starting material to develop stem cell technologies are relatively simple to understand. I am not going to go into any detail – in fact, Martin Pera, in the context of Australia, is the expert in this area – but I think it fair to say that we can now state with some confidence that there either are in existence, or will be in existence, human cells with the properties that we require to establish stem cell therapies.

The first is immortality, or at least an enormous ability to proliferate and give rise to large numbers of cells. And the ability to differentiate widely, preferably with pluripotence, appears to be characteristic of the human ES cell lines that we have.

Three or four years ago I would have put up any reference to stability and homogeneity with a little trepidation, because the cell lines were idiosyncratic and hard to grow, but as tends to happen in science, incremental advances have been made and the human stem cell lines that are available are now much more robust and one can say with confidence that they are likely to be the sorts of starting material that we wish, although I think we would all agree that more work remains to be done.

It is also now true that it has been shown that you can genetically manipulate human embryonic stem cells in much the same way as I was talking about earlier.

Again I will say that there is more than one potential source of these cells, and there are limitations associated with each.

Mouse embryonic stem cells we know about, and human embryonic stem cells are now relatively well characterised and, I think it not premature to say, will fulfil the key criteria that we wish to see fulfilled. But they come with some problems, as we saw in the previous talk. Immune rejection is a problem, and a lot of the approaches to overcoming immune rejection lead you to ethical constraints. Bob Williamson and Julian Savulescu will pick up some of those themes as we go forward.

Adult stem cells may or may not have those properties. It is not yet entirely clear to me that you can take adult stem cells, culture them in homogeneity, get extensive proliferation and control their differentiation properly. That may come, into the future. At the moment, I would say it is limited by technical constraints. It enables me to make the point that without knowing the answer to the technologies that we wish to develop, or the cures that we wish to develop, the important thing is to maintain a broad, scientifically based research endeavour which doesn’t eliminate any of the possibilities that may be important.

(Click on image for a larger version)

So the cells will exist, I think. I am comfortable that that is something we can say. They may start off as an embryonic stem cell, but it is not an embryonic stem cell itself that we are going to transplant to cure the disease. That point is another one which has been made, and again I will just stress that you can’t do that. It is an immortal cell, and if you transplant it, it will give rise to a tumour, called a keratoma.

To get from an embryonic stem cell to something which might be therapeutically useful, you are going to have to differentiate it into something. One of the key points for me is that I don’t believe we yet know what sort of cell we wish to produce for transplantation. To go back to Perry Bartlett’s talk this morning: he was talking about reseeding neurons in the brain, nerve cells. The question is obviously: well, if you want to replace nerve cells in the brain, do you transplant nerve cells, to take account of those that have been lost, or might you preferably transplant a precursor population that can give rise to nerve cells when it finds itself in the brain? I don’t think we know, and I think we are going to have to do some basic investigation which enables us to produce many, or all, of the cell types that might be useful for transplantation and then to test them systematically to find out which of those are most useful.

It is consistent with the fact that we are going to have to learn to control differentiation of these cells anyway, to achieve a therapeutic outcome. As we go through this, because we have targets there, we need to learn about the basic science so that we can undertake this kind of systematic analysis of all of the possibilities.

The other thing which will occur to you is that I talked somewhat airy-fairily about the fact that we might wish to identify regulatory signals or regenerative signals which we can inject into the body, and which might enable the body to repair itself, in a manner of speaking. Now, that is a nice concept, but there is a major flaw there: we don’t know the molecular identity of a lot of those molecules. Perry, I think, might have been suggesting that BDNF might be an early candidate for some of these things, but broadly I don’t think we know. And Melissa Little pointed out that one of the many approaches she is taking to kidney regeneration is to try and identify these sorts of things.

To identify regulatory signals, broadly, you need an assay in which you can show that they are present. And once you can show that they are present, you can try and identify them in a molecular sense. The ability to take these cells, turn them into other cells that might be useful, and then see how they respond to other signals, may give you the capacity to control their rates of proliferation and death, thereby controlling their number – or the signals that control their differentiation, thereby controlling their cell type.

In each case, what this says is that to achieve therapeutic usefulness we will need to understand not only the properties of the ES cell, so that we can grow it properly, but also the ways in which its differentiation can be controlled, so that we can firstly produce cell populations and then either transplant those cells or use them to identify these regenerative signals that are elusive.

(Click on image for a larger version)

So what is differentiation when you are in an embryo? Well, it is not a single-step process. You don’t start with a stem cell which in one hit gives you a red blood cell or in another hit gives you, say, a nerve cell. In fact, the process of differentiation during embryogenesis is very carefully controlled in terms of timing and location, and it is progressive. The inner cell mass cell, or the stem cell or the ES cell, does not jump straight into cells of the central nervous system, for example. It goes through a series of intermediate stages, each of which exists transiently and each of which may or may not have its own therapeutic potential, either as a transplantable population or as an assayable cell population. And, as we form these cell types that come closer and closer to physiologically useful cells, we have to keep taking decisions about how we are going to progress down the pathway of differentiation.

To form nerve cells, for example, you need to pass through intermediate stages which exclude endodermal lineages, and mesodermal lineages, in preference for ectodermal lineages. You have to take decisions not to form surface ectodermal lineages such as skin and hair but instead to follow down the sort of pathway shown here.

I want to talk about the properties of what happens when you differentiate embryonic stem cells in vitro, and I want to show you, in the next five or seven slides, four key features of how these cells differentiate in the laboratory. The key point about it all is the first point that I want to make. When you differentiate an embryonic stem cell in the laboratory, broadly speaking it follows the exact same pathway as the cell itself would have followed when it was in an embryo. That, in fact, came out of a slide that was shown previously today, looking at blood cell generation.

The first thing is that the cells in vitro differentiate the same way as the cells in vivo. The second thing is that you can differentiate your cells in homogeneity, so that you finish up with pure populations of each of these intermediates, or terminally differentiated cell populations. The third is that if you are careful with your culture conditions you can get the culture to differentiate synchronously, so that at any given point in time one of these transient intermediates is present. And the final point is that the cells that you produce are not positionally specified – they don’t think they come from the top of your body, or the bottom of your body or the back or the front – but they are responsive to the signals that would normally tell the cells what their location in the body is. And that itself also turns out to be important.

I am going to show a little bit of our data on generating these ectodermal lineages, particularly the nerve cell lineages. I do want you to know that, particularly as part of our work with the Australian Stem Cell Centre, we are starting to apply the sorts of principles I will talk about to mesodermal lineages and showing that the sorts of concepts that we have developed for these lineages apply equally to different differentiation decisions.

(Click on image for a larger version)

Intermediate cell populations are analagous [sic] to transient populations in the embryo]

I am not going to talk about how we achieved these sorts of things. I just want to show you that you can take a population of embryonic stem cells, as we saw in the previous talk, and that using the right sorts of differentiation conditions you can make them quite specifically choose to go down certain differentiation pathways. In this case, I talk about the formation of a population of cells equivalent to the neural tube. The neural tube is the precursor of the brain and the spinal column, which is found in the embryo early on – particularly the mouse embryo, in this case, because they are mouse cells we are talking about. But the key point about this slide is that the developmental progression of the cells in the embryo is the same as the developmental progression of the cells in vitro.

How do we know that the cells are following that developmental progression? Because we can monitor specific changes in the cells that tell us. What sorts of things? Gene expression tells us. The genes expressed at the primitive ectoderm stage are quite distinct from the genes expressed in the ES or inner cell mass stage, and distinct from the genes expressed when it forms neural plate, or folds neural plate to form neural tube.

The morphology of the cells is changed, and you can have a look at the morphology and see if it is consistent with the populations that you think you are forming. As we form neural tube, for example, we can section those bodies, have a look at the shape of the individual cells, and show that indeed it is consistent with the neural tube that is found in the embryo itself.

And you can assess the differentiation potential of the cells. What else do they go on and form? If this is genuinely neurectoderm, for example, one might expect it to form nerve cells but not, for example, to form muscle cells.

The key conclusion – without going through the details of how it is achieved – is that as you undergo that differentiation process you recapitulate embryogenesis, passing through each of those transient intermediates which you normally find difficult to identify, and very problematic to purify, from the embryo itself. How do you do that?

(Click on image for a larger version)

Well, at each step the cell has to decide whether it is going to take the left-hand route or the right-hand route, so to speak. In effect, all you do is to achieve this stepwise differentiation through these different cell states or cell types by sequential alteration of the culture environment. You add different cytokines and different factors to the mix at particular times, and the cells respond by adopting a particular state.

What sorts of signals are we using? Generally peptide based systems. They tend to work through receptors and activate fairly standard signal transduction systems. And, importantly, they seem to correlate with the signals that we believe to be important in regulating differentiation in vivo during embryogenesis.

So again what I am trying to point out here is that what goes on in the petri dish, in vitro, appears to be a very close mimic of what is occurring in the embryo itself.

(Click on images for larger versions)

This slide shows it in more complexity, and it shows the sorts of genes that one can monitor. We have now done an awful lot more work than is shown on this particular slide, but the point itself is the same: you can trace the formation of each transient intermediate cell population; the neural plate – a sheet of cells which goes from top to bottom – folds to give a neural tube; and the process, in terms of the molecular control, the signals that regulate it and the cells types formed, reflects what goes on in the embryo itself.

That is the first point.

(Click on image for a larger version)

The second point is that if you are careful and your conditions are good, you can achieve it in homogeneity. There are some data on this slide but they are pretty straightforward. This cell population at day 9 is the same cell population we were looking at before, when it is cross-sectioned. It is equivalent to the neural tube population that seeds the central nervous system and the peripheral nervous system. How can we recognise those cells? Because they switch on genes which are characteristic for the nervous system.

This slide shows two different genes, Sox1 and Sox2. You can tell when they are switched on, because of the purple staining. And what you can see is that in effect all of the cells in this culture – I can tell you that it is actually more 95 per cent of the cells in this culture – have adopted this particular fate. They have decided to move from being a stem cell to being a neural tube cell.

If you look at them under the scanning EM you can see that the outer surface of the epithelial layer – it is only one cell thick – is homogeneous and smooth: it is only a single-cell population. That is gene expression, that is morphology; it is also the case that you can look at differentiation potential of these cells. Ordinary ES cells, in uncontrolled differentiation, give rise to beating muscle, and that is what is shown with the blue bars you see here.

These structures which we are forming in homogeneity, which appear to be equivalent to the nervous system, do not form beating muscle because they have moved past the part in a differentiation pathway where that alternative is available to them.

(Click on image for a larger version)

We can show this homogeneity in another way. What does the neural tube do in the embryo itself? It differentiates into the central nervous system and the peripheral nervous system. Ron McKay has shown that there are conditions which can enrich for certain central nervous system cells, particularly glial cells. And if we apply those differentiation conditions – again these are proteins – to our precursor population, we get in effect 100 per cent cells that look glial by morphology and that stain with a poorly specific glial marker, GFAP.

The key point, though, is that if you take that exact same starting population and differentiate it in the presence of a rather grubby protein kinase inhibitor called staurosporine – one that we identified because Don Newgreen, at the Murdoch Institute, has shown that it induces the formal of neural crest from neurectoderm in a quail – if we use that inductive system on the same population of cells, we again get homogeneous differentiation but we get a completely different cell type formed. It is migratory, it expresses Sox10, and it appears to be equivalent to neural crest, which seeds the peripheral nervous system in vivo.

The fact that we can get alternate homogeneous terminal differentiation at the ends of these left and right arrows leads us once again to say that what we were dealing with in the first place must have been homogeneous.

We would like to be able to form homogeneous populations of neurons. We can’t yet do that. We can form neurons very efficiently with this as starting material, but we are lacking a factor which must tell these cells to become homogeneous populations of neurons. This enables me to make the point again that we lack the factor but, because we lack the factor, we have a chance to be able to discover what it is. Once we can find out what causes these to make neurons, then we can purify whatever it is that they are responding to.

(Click on image for a larger version)

So the differentiation occurs, as it does in the embryo – and that is not surprising because this is what the cells are programmed to do. It can be carried out in homogeneity. And the third point I said I wanted to look at was that it can be carried out in relative synchrony.

Here I want to look at the decision of the transient intermediate called ectoderm to either go on and give rise to the nervous system or to give rise to the surface ectoderm. Little is known about the control of that decision in the mammal, but there has been quite a lot of work done in the frog Xenopus and they know that a single growth factor can be sufficient to control that decision, and that in the presence of BMP4, ectoderm gives rise to surface ectoderm, and in the absence of BMP4, ectoderm gives rise to the nervous system.

(Click on image for a larger version)

We have talked about this system from ES cells through primitive ectoderm through definitive ectoderm to neurectoderm – and the neurons are shown on the right-hand side of the slide. We could have a look at what happens when we add BMP4 to these cultures at particular points in time. (And I will say that one of those intermediates exists for about 24 hours in our system.)

We can’t get any response of ES cells or neurectoderm to BMP4. If we add BMP4 to primitive ectoderm – and it doesn’t matter why, although this is expected – we get the formation of mesoderm, and many others have shown that. But if we add BMP4 some hours later to the population which has passed through the primitive ectoderm point and is now at the definitive ectoderm point, the first thing we see is a dramatic down-regulation in the ability of the cultures to form neurons, and we see the appearance of a cell type, not in homogeneity but in great abundance, a cell type that we don’t otherwise see, which is flat and squamous, it has tight cell-cell junctions, and it is not shown on this slide but it up-regulates genes like keratins, which are characteristic of surface ectoderm.

We pass through a cell population, an intermediate, for a short period of time (about 24 hours), which responds to a particular ligand by adopting a novel cell fate. Before that time we have in the culture a different cell type; after that time we have a different cell type. This enables us to say that the culture systems that we are using are synchronous in terms of the way they go about this.

(Click on image for a larger version)

The last point I want to make is about the fact that these populations are naïve and can respond to other signals and therefore be used as assays. Shown here is a picture which we ‘pinched’ from a famous Italian scientist – I don’t think he was a member of an Academy, being a bit before that kind of time – which shows the other key feature of embryogenesis. We don’t just specify cell number and cell type; we organise them. And we organise them with respect to the anterior-posterior axis, head-tail, and the dorsal-ventral axis, back-front. Again if you look at da Vinci’s pictures you can start to superimpose the morphological features of the nervous system, but the most anterior part of the brain is the forebrain and we pass through midbrain and hindbrain to spinal cord.

When a cell is first specified to form neurectoderm, it doesn’t know whether it is at the forebrain end or the spinal cord end, but that matters because if it gets told, for example, that it is at the anterior end then it switches on gene expression characteristic of those cells and differentiates into nerve cells that are normally found in the forebrain. If it is at the tail end it switches on genes that tell it that it is a posterior cell type, and it will differentiate into different kinds of neurons, for example those that are normally found in a spinal cord. The ‘address’ of the cell within the body determines the nerve type – or, more particularly, the nerve sub-type – that is formed as you differentiate the neurectoderm, which is the precursor for all of this material.

(Click on image for a larger version)

Looking at the cells that we have been forming, we can show that they don’t know where they come from. They don’t express any of the genes associated with forebrain, hindbrain or spinal cord; they only seem to express genes associated with midbrain, which we understand to be the default or naïve state for newly formed neurectoderm. I can also tell you that they do not express genes associated with ventralisation.

But we know a little bit about signals that control some of these things. We know signals that can switch ON posterior genes; we know signals that can switch ON ventral genes. So we can take this population and ask: do they respond in the way that they should? Can you tell them, for example, that they should try and form forebrain, or try and form spinal cord?

We know a fair bit about the signals that regulate ventralisation. The one that we have been working with is a protein signal called Sonic Hedgehog. It is not normally expressed by our neurectoderm, because it doesn’t think that it is ventral. But if we make Sonic Hedgehog and expose our cells to it, guess what? We switch ON markers that are characteristic of ventral neurectoderm. We are telling the cell not that it sits in the midbrain but specifically that it sits in both the midbrain and the ventral part of the midbrain. Why does that matter? Because that is the positional address which gives rise to dopaminergic neurons, which is a population of neurons that one might with to transplant to treat Parkinson’s disease.

As we differentiate these cells which have been ventralised, as we call it, we do find that they give rise to populations of neurons which are heavily enriched for cells that express tyrosine hydroxylase, which is a poor marker for dopaminergic neurons.

(Click on image for a larger version)

But it is not just ventralisation. That is a system that is understood. Equally, you can make the same starting cell population – I should say that this again is in homogeneity, 100 per cent of the cells – think that it should be posterior, that it should form spinal cord neurons. How do you do that? You use retinoic acid. We don’t normally express posterior markers like Hoxb1 and Hoxa7 with a starting population, but as we expose them to retinoic acid we up-regulate the genes.

The real reason I am showing this, though, is that I have talked about the need for us to identify novel regenerative-type signals in the environment. One of the elusive signals in embryogenesis is the one that specifies forebrain. People do not know what it is.

We have done an experiment, which I won’t go into any details about at all – it’s a cunning kind of experiment – and what we have been able to show is that we can achieve anteriorisation of the cell populations in response to a novel protein, which we haven’t yet identified in a molecular context. We have used our neurectoderm as an assay to identify a protein which can tell the cells that they should be there.

(Click on image for a larger version)

To summarise the differentiated populations that we are talking about, these are the points that I have made.

They are formed in a manner that reflects embryogenesis, including the progression through those transient differentiation intermediates; the cultures can be homogeneous and synchronous; and the cells can be further specified.

The potential applications might be the cells themselves, for transplantation; or it might be the cells themselves, as assay systems, that enable us to identify the regenerative or inductive signals that we might wish to inject at the same time as the cells.

Importantly, I think, we need to understand that different cells will do these things in different ways, and that until we know which of embryonic stem cells or various different adult stem cell populations are genuinely going to be of use in medicine, we are going to have to take the time to study each of them in terms of the factors that are required to enable them to proliferate and to turn into populations that we might otherwise find useful.

So I guess this is simply a plea to maintain the impetus of the basic research. We might be lucky, we or someone else might take some of these populations, put them into a patient and find that they work spectacularly. But that’s unlikely. We will probably see a heap of experiments where that is tried – they will look provocative, they may or may not stand up five or ten years out. At the same time, we need to maintain this basic effort, which will continue to inform us about the basic properties of these cells, thereby telling us how it is that we might use them for therapeutic benefit.

Questions/discussion

Question – We have found in plants that microRNAs are very important in determining differentiation and the developmental route. Has this been looked at?

Peter Rathjen – I am not aware that there has been a lot of work done on microRNAs and how they control these sorts of things. I think there is a growing feeling that they must be important in the mammalian genome, or regulation of expression, and John Mattick would be doing a lot of work on that. But, broadly speaking, I think that the ones that are known about in vivo are generally protein factors, and the ones that we are using in our systems are protein factors.

Question – When you have got to a certain stage in the differentiation of those embryonic stem cells, are the cells that you are then dealing with, with the factors that you have just described, like adult stem cells or like embryonic stem cells?

To put that a bit differently: you have caused some degree of development of the cells by allowing them to progress along the line of growth. Are those cells there still pluripotent cells, such that you could do something different with them, for example make a skin cell out of them? Or are they on the way now to becoming, as it were, nerve cells?

Peter Rathjen – The answer is that they are no longer pluripotent. The schema that I was showing indicates that primitive ectoderm is the last pluripotent cell population. Once they differentiate past that point, they are no longer pluripotent and they are committed to, in the case of what I showed, the ectodermal lineage – although alternatively we could shunt them down the mesodermal lineage.

I think where you might have been going with the question, though, is toward something that we have been interested in for a while but have done nothing about yet. At some stage we actually ought to see adult stem cells arise during that differentiation process. We haven’t yet looked for that, but it would be kind of nice if we did.

Question – When you look at hemopoietic cells versus the different lineages of progeny that progressively differentiate out from them, one thing that strikes you is the progressively increased fraction of chromosomes that are packed away in unexpressible heterochromin. Is that something that you observe in other lineages, or is it peculiar to the blood system?

Peter Rathjen – We are actually using the systems I have talked about up here to study precisely that question. We are doing it with a collaborator in the United States, so I can’t tell you too much because a lot of the results are his. But it does appear to be the case that as you move through progressive stages of differentiation, so you do change the chromatin status at particular sites in the genome – which I think is what you were asking about.

Question – I just wondered how important it was, from your perspective, to open up the possibility for the development of rare disease models for the use of somatic cell nuclear transfer for rare diseases.

Peter Rathjen – So this would be in line with the previous talk, where the idea is that you might have cells that are mutated for cystic fibrosis or something like that? I haven’t thought that through very carefully. Martin Pera might be able to answer that rather better than I can. I think he is a strong supporter of doing that kind of work.

One of the problems, of course, is that it is very hard to study humans. You can’t do experiments on them. And though I have talked fairly blithely as though a human is the same as a mouse, it does have particular physiological differences. So if we are really going to need to understand the molecular basis of defects that lead to human disease, then to have some sort of model, particularly a developmentally plastic model that you can actually manipulate in vitro, I would have thought would be a powerful tool.

Question – One may have got the impression, rightly or wrongly, on hearing talks this morning relating to the use of stem cell therapy in chronic renal disease, chronic lung disease and so on, that you are really trying to grow a new organ. I can understand stem cells being used to replace liver cells, neurons et cetera. But if you have to grow a complicated tissue with many different constituents, is that within the foreseeable future?

Peter Rathjen – I am not the one to talk about tissue engineering. You should ask Melissa Little that question; she could tell you better than I.

There is one slant to it, though, which impressed itself upon me. It was Bob Williamson that brought this up once. He asked me whether I thought it might be possible to use stem-derived cells to cure stroke. Being a scientist and relatively pure about these things, I thought no, you couldn’t possibly restore memory, et cetera, et cetera. He rephrased it and said, ‘Well, might you be able to achieve some effect in a stroke patient – 10 per cent, 15 per cent?’ And my answer as a scientist was, ‘Probably yes, but why would that matter?’ Of course, if you are paralysed, it could make an enormous amount of difference.

So whether we will be able to make proper organs that can be transplanted in toto, it seems to me that to make it in the lab is a very big step, and that perhaps systems which exploit the body itself might work. But to achieve some kind of therapeutic benefit by a partial outcome, that is something that I think might be possible.

Question – I was wondering if you could tell us to what extent animal models are predictive of the success of various techniques for humans. You said a moment ago that physiologically they are quite different. If you had success across a couple of species, how often would you expect – and I know it is hard to quantify – that you would have success in a human?

Peter Rathjen – I think ‘quite different’ is a bit stronger statement than I made. They can be quite different physiologically, but they are more similar than they are different.

Actually, the question you asked is a very important one, because a lot of the most advanced work has been done in the mouse system and people are interested in looking at how well it can be transferred to the human system. So when we sit down, at the Australian Stem Cell Centre, we tend to spend a fair bit of time wondering how similar are these cells.

There are clearly some differences. The growth factor that is used to maintain mouse embryonic stem cells, called LIF, does not appear to have an effect on human embryonic stem cells. Yet, if you look at the differentiation pathways that human embryonic stem cells follow, in terms of the gene expression, and also broadly speaking in terms of the way it can be regulated, then you’d reckon that broadly speaking they are differentiated in the same way as the mouse cells.

So I think the answer is a sort of intermediate one: there will be some peculiarities of the human system, but I think we all have confidence that other systems such as the mouse will broadly lead us in the right direction.