SCIENCE AT THE SHINE DOME canberra 4 - 6 may 2005

Symposium: Recent advances in stem cell science and therapies

Friday, 6 May 2005

Professor Perry Bartlett
Director, Queensland Brain Institute, The University of Queensland

Perry Bartlett was appointed Foundation Chair in Molecular Neuroscience at the University of Queensland in 2002 and Director of the newly established Queensland Brain Institute and Federation Fellow in 2003. He previously headed the Neurobiology Division at the Walter and Eliza Hall Institute in Melbourne, where his group was the first to identify a neuronal precursor in the adult brain, leading to a paradigm shift in our understanding of the brain and underpinning the burgeoning field of neuroregeneration. In Queensland, he is continuing his research into how the production of new neurons influences the function of the adult brain, such as how memory and learning, and neurogensis, are regulated. Finding the mechanisms that regulate these processes is the primary focus of the Queensland Brain Institute and will lead to a new generation of therapeutics to treat the avalanche of neurological and mental illnesses.

The brain – are stem cells required for a healthy brain?

I am going to tantalise you, I hope, with some evidence that making new neurons might be very important for our normal functioning brain.

There has been a paradigm or a profound shift in the way we view the brain. I guess most of us grew up thinking the brain was pretty stagnant, pretty hard-wired, and it wasn’t capable of change. The last dozen or 20 years have profoundly changed that view.

We know the adult brain is not hard-wired, it can change. And while we have known that synaptic connections between brain cells could change, the degree and magnitude of that change has only recently been verified.

Perhaps the more profound change in our view has resulted from the discovery that we might be able to change the repertoire of nerve cells in your and my brain at any stage of life, not just during very early development.

(Click on image for a larger version)

Although the idea about making new neurons in the adult has a long history – dating back to the turn of the century with Allen’s work, where they have shown that there were cells in the brain that were dividing – most of this was not recognised or was ignored for some reason, mainly because markers were not readily available that were unequivocally determinant of neurons.

There were two papers in 1992: a paper from our lab by Linda Richards, who I am happy to say has returned from overseas to the Institute, and a Canadian group headed by Reynolds and Weiss. And I am happy to say that Brent Reynolds has also joined the Institute. So we now have the founders of the neural stem cell isolation within Australia.

We were able to identify a cell, grow it in vitro and show that it could give rise to neurons. When we did that, of course, it met with some scepticism that it could be just a remnant of development and perhaps had no role.

I want to talk about three things today. Firstly, is there a functional role for those stem cells in making neurons? Secondly, is there evidence that, in fact, that selection of making new neurons, and that stimulation of making new neurons, is impacted to a large degree by the environment? That is something we never thought might be so. The third thing is to think about how this might be regulated, and to tell you something about the way I view the regulation to occur as a two-step process: one which is generating neurons and another one which involves selection of those neurons, or neuronal selection.

(Click on image for a larger version)

Everything in this field is dependent on an assay system, and I am going to show you a little bit more about developing new assay systems. The power of being able to dissect these fields, as Don Metcalf and Ray Bradley showed many years ago, is to have a robust, potent assay system by which you can determine the frequency, the identity and the output of these cells.

Panels A–F here show the assay that was used, a neurosphere assay, where a single cell could be taken out of a mouse or rat brain and incubated with epidermal growth factor or fibroblast growth factor. Over a period of a week, large balls of cells would generate – tens of thousands of cells. You could show, when you flattened these out, that they formed the red neurons, the blue astrocytes and the yellow oligodendrocytes that you see in panel G, the cells that cause myelination in the body.

(Click on image for a larger version)

The feature of this system – and here is a cross-section – is that they are actually quite organised balls of cells. You have neurons on the outside and astrocytes on the inside, and highly proliferative, undifferentiated cells on the inside. It gives us an opportunity to actually look in three dimensions and model how this differentiation occurs, but I won’t go into that today.

(Click on image for a larger version)

The beauty of these cells was that not only could you form the neurospheres but you could disaggregate them and, within those aggregates, there were black cells which would form secondary neurospheres. And you could, in fact, do this almost ad infinitum in the mouse. You can passage these cells 50 to 100 times, you can grow something like 10²⁰ cells out of a single cell of an adult brain of a mouse. So the ability to make cells from stem cells within the brain is almost unlimited, at least in the mouse.

The other characteristic of stem cells, which Martin Pera talked about, is that they must be able to proliferate, they must be able to self-renew – this ability to make, in a clonal fashion, a secondary neurosphere – and they must be able to generate functional, differentiated progeny.

This assay, at least superficially, meets all those hallmarks in terms of an assay for a stem cell in the brain.

(Click on image for a larger version)

Following that identification in 1992, a plethora of papers came out to show that these cells weren’t remnants and in fact they were giving rise to neurons in the adult brain, both in humans and in the mouse. Perhaps the most florid example of this is in the olfactory system – at the left of this slide you see a cartoon of the olfactory bulb in a mouse, where you are receiving sensory input from the cells at the back of the nose, as Brandon Wainwright talked about. As well as other types of cells you see the olfactory mucosal cells, each one of which has a single receptor for a single odorant. Remember that in the mouse about 2 per cent of the genome is dedicated to coding for olfactory receptors.

They bundle together in glomeruli, and cells migrate in from the lateral wall, the lateral ventricle. The diagram at the right of this slide depicts a sagittal section through a mouse brain. The cells migrate en masse, cheek by jowl, from the forebrain into the olfactory bulb. In a mouse, many thousands of cells a day come in; in you and me it is estimated that perhaps even up to 10,000 cells a day are marching into those olfactory bulb areas.

We do have some preliminary data, mainly from a colleague of mine, Jeff Macklis, of Harvard, about the function of these cells. He has been able to look at individual glomeruli and look at those neurons that migrate in to those glomeruli and modify olfactory signalling. What he has shown is that normally 95 per cent of these cells die within three weeks. However, if they are stimulated with the appropriate cognate ligand, the olfactant, in fact you can rescue about 20 per cent of those cells. This is unpublished data which he won’t mind my telling you about – it is the first evidence that environmental or, here, ligand-stimulated electrophysiological input can select the survival of neurons as they make their way into the olfactory bulb.

(Click on images for larger versions)

For all of us, I guess, much more important was the discovery that even areas of the forebrain associated with memory formation – in fact, the hippocampus, which is very predominantly involved in spatial memory and working memory – also were turning over. It was shown by a number of groups that in this structure the geniculate area, seen in the cross-section sketch at the left of this slide, was turning over not as quickly as the olfactory epithelium but perhaps a few hundred cells per week, in the mouse. We don’t know how much that is in humans. It has been estimated, though – this is not just a perfunctory finding – that in the mouse perhaps this whole area turns over every two to three months.

Some studies done by Elizabeth Gould some years ago, where she ablated the ability of animals to make cells in the hippocampus during this period and showed some loss of memory formation, are one of those provocative bits of evidence that point to the idea that the ability to make new neurons in the hippocampus might be important in memory formation. (We will come back to that later.)

(Click on image for a larger version)

The first bit of data that pointed to the idea that environmental stimuli, a bit like what I just told you about in the olfactory bulb – that is, that an external odorant can select the survival of neurons – came in the hippocampus from a study in which mice were put in what is called an enriched environment, with a few black tubes that they can crawl through and sniff around; a running-wheel environment; or just their normal cage pattern.

(Click on image for a larger version)

The animals run up to 3 to 5 kilometres between the hours of 11pm and 1am or 2am, and it was found that only under these conditions do they make more cells as indicated by uptake of BrdU. But the most interesting finding was that three weeks later, if you looked at what cells were surviving that were labelled with BrdU, in fact it was the enriched crossbar group where most of those cells were rescued from dying, whereas just as many of the running-wheel mice died as did the other populations.

So it says that you can drive neurogenesis by one means, but rescuing that invokes a totally different paradigm of stimulation. In this case it is an enriched environment.

Now, ‘enriched environment’ means a lot of things. The hippocampus receives enormous amounts of input from olfactory, visual, cortical stimulation, et cetera. So it is going to be much more difficult to dissect what those secondary signals are, but it is provocatively of interest to think that they are coming through a neurophysiological source.

(Click on image for a larger version)

So what I have said is that one of the things we have to be able to do is to ablate the ability to make new neurons and look at the effect of this on the animal’s behaviour. That is probably the best way for us to understand what the functions of these stem cells are.

There is a model in which we already have ablation of, or a deficit in, neurogenesis. That is in Huntington’s disease, the familial, genetically inherited disease where you have long strings of glutamine repeats on the Huntington protein, resulting in quite severe motor, cognitive and psychiatric syndromes.

I am going to show you some provocative data – I am sorry, I am using the word ‘provocative’ a lot, but this is quite provocative in terms of what happens in these animals if you expose them to environmental stimuli. This work I am going to tell you about was done by Tony Hannan, my colleague at the Howard Florey.

(Click on image for a larger version)

Here is an enriched environment again, on the right. It is a bit more colourful, a bit more exciting, and it is changed every three to four days for the mouse to explore.

(Click on image for a larger version)

What Tony showed originally was that, as you can see in terms of BrdU uptake, in the enriched versus the non-enriched there is an increase, especially in the Huntington’s disease, in the number of cells that are labelled with BrdU.

(Click on image for a larger version)

The most profound thing found was in relation to the running-wheel, looking at motor skills of these Huntington’s disease animals, because this is one of the first things that go. After five months, which is getting towards the end of this disease in these animals which have been engineered with a human defect – in the enriched population you can see that the wild-type and the Huntington’s disease performed almost entirely the same.

(Click on image for a larger version)

In fact, when you look at the data long term, you can see the performance of normal unenriched cells versus the enriched Huntington’s disease, which is very close to wild-type.

This is an incredible change in the outcome of a disease process which results in quite large loss of motor and memory functions in the animal.

(Click on image for a larger version)

I am showing you lots of tantalising bits of data, none of it foolproof but some of it pretty interesting. The other bit of data in this regard was a paper published by Santarelli in Science, August 2003, when he was looking at the effect of Prozac – fluoxetine serotonin uptake inhibitors. What he showed, and what had been shown before him, as you can see in panel A, was that in fact the fluoxetine, the F here, when given to an animal, causes a fairly massive increase in the number of dividing cells in the hippocampus, almost a doubling of cells, but it doesn’t occur till about 11 days post fluoxetine. It is interesting that in patients Prozac usually takes 11 to 28 days to work.

So he asked what would happen if you knocked those cells out. Again this was a flawed experiment, because he used irradiation to irradiate the hippocampus, obviously doing many other things as well. Nevertheless, when he irradiated and showed that there was no hippocampal neurogenesis, the ability of fluoxetine or Prozac to work in an animal model of depression was totally ablated.

So here is another bit of evidence that not only is neurogenesis perhaps going on, but it is important to maintain a healthy brain. Perhaps – drawing a long bow – it says that if you don’t have that, the response to depressive illness will not be the same.

(Click on image for a larger version)

Let me show you some unpublished data from Tony Hannan, which we have been doing some work on over the last six months, using fluoxetine in the Huntington’s model. You see here the outline when fluoxetine was given over quite a lengthy period of time, and then several things were done: the animals were culled but also performance indicators were carried out on them.

(Click on image for a larger version)

This, then, is a model of spatial memory: a T-maze where an animal makes a choice of going left or right, and if he has gone left once he will tend to go right the next time. In the Huntington’s disease animal you can see that the number of times they do this is quite low, whereas the fluoxetine-treated animals are back up to wild-type levels, so almost a total inhibition of decline in these animals, using a model of spatial memory.

(Click on image for a larger version)

But it is even more dramatic to look at the volume of dentate gyrus, that area of the brain that is turning over. This is in panel A, and it is something I found incredibly hard to believe when we first saw this. In fact, in the fluoxetine animal the level of dentate gyrus, the number of cells in the dentate gyrus, was back to or had been maintained at wild-type level. That is, there was absolutely no degeneration in these animals on fluoxetine, such as occurred normally. There was also a concomitant increase, obviously, in the number of NeuN, a marker for neurons in the hippocampus, and an increase in the wild-type.

So here we have more neurons being made, and a maintenance of the volume of dentate gyrus – a pretty incredible finding, that you can overcome this incredibly dangerous, toxic disease with environmental input, and that this seems to mediate through the hippocampus in the production of neuron replacement.

(Click on image for a larger version)

Now, before you all rush out and buy fluoxetine – especially the older members of the audience – I should mention that we have been doing some studies directly looking at a whole range of these neurotropic and antipsychotic drugs. It turns out that they all do some interesting things. Dhanisha Jhaveri, a postdoc in the lab, has been looking at this directly by incubating slices of tissues which include the stem cells from the sub-ventricular zone, those that go to the olfactory bulb and those that go to the hippocampus.

(Click on image for a larger version)

She has shown that, although in some cases you can get an increase with things like inhibitors of serotonin receptors, in one area, like the SVZ, you get inhibition of the number of neurospheres or the number of precursors, whereas in the hippocampus [inaudible].

It seems that we have to be very aware that these agents, in terms of serotonin level, are going to be doing different things to different populations of precursors, and while you might be driving a lot of neurogenesis you may be depleting the progenitor or stem cell pool. So that’s just a word of caution.

(Click on image for a larger version)

Here is the model I wanted to get to – I’ve taken a long time getting here. This is what we think is going on. This is what we are trying to test.

Neuronal production is regulated, and it is happening, usually deviated to the production of astrocytes, this other type of cell in the brain. However, we do know that environmental stimuli, as I have talked about, can work through factors – and I will talk about what these factors might be – and do stimulate the production of neurons. We thought that was all there was to the story.

I now know that the most interesting part of the story is up here at the top right of the slide, under the heading ‘Neuron Selection’. That is, making neurons is really the easy part; selecting those neurons to be able to integrate and function is really where the prize is, and finding out exactly what the regulators of this activity are, which must involve synaptic activity, physiological inputs, to prevent cell death that occurs in most of the neurons that are being made. I think you are going to hear a lot more about neuronal selection in this model over the next five years.

(Click on image for a larger version)

So let’s just talk a bit about what does regulate that neuronal production.

Rod Rietze1, a graduate student when we were at the Walter and Eliza Hall Institute (WEHI), managed to purify the cell using this assay and using cells sorted from the sub-ventricular zone – those precursors of the olfactory interneurons.

(Click on image for a larger version)

He showed, using two markers which are unrelated to function, that there was a population of about 0.2 per cent in the sub-ventricular zone of the lateral ventricle that could give rise to neurospheres, almost on a one-to-one ratio – about 80 per cent of those cells were neurospheres.

(Click on image for a larger version)

However, we have been going on to ask the question: are all those cells that give rise to neurospheres truly stem cells? That is the question that haunts everyone: what are we actually measuring here?

In the diagram here you can see that although there probably is a neural stem cell, in the assay that we are looking at perhaps all three of these populations, the progenitors and the blastic cells, can give rise to neurospheres.

So how do we go about looking at that?

(Click on image for a larger version)

This is the assay that we normally use, but you can see the sizes of neurospheres in fact vary quite dramatically. Brent Reynolds, the young Canadian guy who invented the neurosphere assay some 10 years ago, has been working on another assay in the lab to look at this, using semi-solid media.

(Click on image for a larger version)

We have called this a colony forming assay, à la Metcalf and Bradley. Again, obviously, having other forms of assays drives the discovery of these different populations.

We plate these cells out in semi-solid collagen assays, and we find a very interesting thing. One is that there are different-sized populations after 10 days in vitro. And it turns out (I am not going to show you the data) that the only population of cells that passes the stem cell test – and the only really good stem cell test we have is that you can continually passage them for about 10 to 12 generations – is this population that gives rise to greater than 2-millimetre colonies in collagen. All the rest peter out after two to three passages.

Of interest is that you can see, in this figure at the top right of the slide, that in the stem cells from the hippocampus there are virtually none of these colonies. So the hippocampus appears not to have the characteristic of a stem cell.

(Click on image for a larger version)

And that is true, because when you go to passage the hippocampal cells, neurospheres, in fact they only go through two passages, as compared with these large sub-ventricular zone cells.

(Click on image for a larger version)

So the question really is: what is the cell that gives rise to the hippocampal progenitors? Where is the replenishment coming from in the hippocampus that is giving rise to the progenitors that give rise to the neurons? That is a question we are very avidly pursuing at the moment.

(Click on images for larger versions)

One of the links between what regulates the environmental signals and the production of neurogenesis has been around for a while. These are slides put together by Dhanisha Jhaveri, who has got a wonderful colour sense about things going up and down (yellow means that it is going down, red that it is going up).

She found that these things have been shown in the literature to cause the reduction in BDNF. So everything that seems to repress neurogenesis, suppresses the production of brain-derived neurotrophic factors – one of those early growth factors described in the nervous system. And that leads to down-regulation of neurogenesis. When BDNF goes up, by counterpoint in all the things I have been talking about, you get increased neurogenesis. So could BDNF be the local mediator of these external stimuli that give rise to neurogenesis?

(Click on image for a larger version)

It is clear that many of the antidepressants, in fact, work by regulating BDNF, including the serotonin regulators norepinephrine and dopamine, although they look as if they might directly regulate neurogenesis.

(Click on image for a larger version)

One of the things that came out of the cell sorting exercise, where we killed 600 animals to isolate enough stem cells out of a normal mouse brain – we should have killed 6000! – was that we found a series of receptors on the surface. And of course these are the receptors that should be regulated in situ to cause the production of neurons. One of those receptors was the p75 neurotrophin receptor, the first receptor isolated that bound both NGF – nerve growth factor, the first growth factor ever isolated – and also DDNF, NT4 and NT-3. Kaylene Young, a student who has been a student both at WEHI and at the Queensland Institute, has done the majority of these studies.

(Click on image for a larger version)

We went back and asked whether it was really true that this molecule was on the surface of the progenitor population, or the stem cell population. You can see, just looking at the right-hand panel, that the red cells stained with anti-p75 antibody are in fact in the right place in the adult brain.

(Click on image for a larger version)

She showed that when she sorted for the expression of p75, the same way as Rod had done earlier, about 0.3 per cent of these cells were positive.

(Click on image for a larger version)

And when she did the assay she found that 50 per cent of the cells that were sorted for p75 gave rise to neurospheres. So again the beauty of combining purity of cell sorting with Affymetrics gene array is that we have been able to pull out quite cognate receptors on the surface of these stem cells, and that these cells, when sorted, have those properties.

(Click on image for a larger version)

The most interesting finding, though, is that in fact this cell population, when incubated with BDNF (remember BDNF goes up and down according to neuron production in vivo) the high p75 stem cell population, when you titrated-in BDNF, gave rise to neurons – neurons being produced in neurospheres – whereas the medium and low populations of stem cells without p75, of which there was a large number, in fact did not respond by neurogenesis. So we think Kaylene has discovered the population of precursors within the sub-ventricular zone, at least, that is respondent to BDNF and gives rise to neurogenesis and might be the major population of precursors being regulated.

(Click on images for larger versions)

So that is our model. We think BDNF is a major regulator of neuron production. We know very little about the neuronal selection side of the model; that is something that is going to have to wait to be done.

(Click on image for a larger version)

Shown here are the people who have done all the work at the Institute. As I said, Kaylene Young is both a QBI and a WEHI recent PhD student.