Brandon Wainwright is Group Leader and Deputy Director (Research) at the Institute for Molecular Bioscience of the University of Queensland. His early research concentrated on the use of human genetic and genomic approaches to the mapping and isolation of the cystic fibrosis gene. Subsequently, he used the approaches developed in that work to isolate the patched gene which when mutated causes the most common form of cancer, basal cell carcinoma. Recently, his group and others have shown that the patched pathway is mutated or perturbed in most solid tumours, including tumours of the CNS, skin, pancreas, stomach, oesophagus, lung, muscle and prostate. It also appears that this pathway regulates the stem cell niche in a number of tissues. Given that the patched pathway likely controls regeneration in airway tissue then a developing focus of his laboratory is the intersection between lung regeneration and cell/repair-based therapies for airway diseases such as cystic fibrosis. Brandon has published over 130 peer-reviewed research articles and has received a number of awards and prizes, including the 1998 Gottschalk Medal from the Australian Academy of Science.

Airway repair with bone marrow-derived cells
by Professor Brandon Wainwright

Before we start, in case I run out of time, I would like to acknowledge the postdocs who have performed some of this work. I will just touch very generally on this work: Brendan McMorran and Elaine Costelloe, Steve Cronau in association with work done between our laboratory and that of David Hume at the Institute for Molecular Bioscience, and, if I get to some of the CF work, with Eric Alton and Steve Smith at Imperial College, in London. And I should put a disclaimer – if that is the correct word nowadays – that this work is actually supported by AMGEN, who have supplied us with money and growth factors, as you will see.

(Click on image for a larger version)

Martin Pera and others have given an absolutely brilliant introduction, and I don’t need to belabour many of these points. But every stem cell talk has a wish list of pathologies which, if this wonderful technology came through, we would be able to correct.

In the case of lung diseases, you can see that on the list there we have got hypoplasia; respiratory distress syndrome, both in adults and in neonates; pulmonary fibrosis (which I am going to talk a little bit about); CF, which is one of those diseases which are always raised in the context of gene therapy and therapies, it seems to be a newspaper-type disease such that when people have a supposed cell or genetic therapy they say, ‘Useful for diseases such as cystic fibrosis,’ so it has made the list here; and at the bottom of the list we come to lung cancer. It is interesting, when we talk a lot about being able to control stem cells for their ability to repair tissue, that there is a growing literature which informs a great deal of the efforts of my lab, that tumours may derive their proliferative potential from having embedded within them tumour stem cells. And in order to successfully treat tumours, we actually have to identify those resident stem cells within the tumours, not just the bulk of the tumour.

We have talked a lot in stem cell biology today about regeneration, but it is important to remember that perhaps if we can identify stem cells and molecules that not only don’t cause them to proliferate but also can remove them, then there may be therapeutic applications to cancer as well.

(Click on image for a larger version)

This is the traditional view that Martin Pera doesn’t really like, so I will skip over it – most of it has been covered already – except just to remind you that in terms of stem cell biology, the resident stem cells which sit in the tissue and are present there are very difficult to identify. They grow very slowly, and they may, hopefully, undergo asymmetric division. That, I think, is an absolutely fascinating topic that we just pass over, really, because it is long and complicated. It is the idea that we have a stem cell which sits in a niche within a tissue, and at some point it has got to be able to self-renew. But then it has got to divide into two, and then away it goes to become differentiated. But at some time this division has to become asymmetric. One of these cells has to go back and form a stem cell; another one has to go on and march down some other pathway. A lot of our work in the hedgehog signalling pathway works at this interface, and we take a signalling approach to this as well.

(Click on image for a larger version)

In terms of lung, the approaches which are shown here have already been illustrated today in several talks. If you are looking at the possibility of repairing a tissue, using a stem cell approach, then you have got a couple of choices. One is to try to identify the local cells, the stem cells that sit within that tissue, and try to find some magic molecule or stimulus which causes them to self-renew and repair your tissue of choice. The other one which we have heard about is to try to take some tissue which we know is rich in progenitor and stem cells, such as marrow stromal cells or hemopoietic stem cells, and to utilise those in some sort of strategy to introduce them to a tissue which is damaged, and cause some repair.

We have also noted today that there is a lot of controversial evidence which suggests that stem cells from one type of tissue may generate cells typical of other organs. The process under some circumstances is called transdifferentiation; in our lab, the people who work on this are called the Alchemy Section of the lab. And it is a bit like a modern alchemy – the idea that you can take a cell of a particular type and it is plastic and can receive signals from a new milieu, and it will become something else.

Of course, it can do that by this process of transdifferentiation and has also been touched on today that, particularly when we are using hemopoietic stem cells, these are naturally fusogenic cells. They migrate to the sites of damage and they will fuse with other cells. So perhaps all that we and others are seeing is the idea that the stem cells are finding a damaged cell and fusing to it, and they take on some character. To some extent it doesn’t really matter if one is taking this approach, because it is the repair that matters, not the actual mechanism – although it is good to go back and try to work out the mechanism.

So how did we get into this? We got into this because we have been working extensively in two model systems – the so-called hedgehog signalling pathway which we have shown controls the stem cell compartment of the epidermis, and increasingly the work we are doing with Perry Bartlett in the central nervous system. We first got into this particular pathway through patients who have Gorlin’s syndrome, and they also have lung defects. So, as part of our challenge to try to work out how this pathway gives pathologies in all of the affected tissues, we have become interested in whether the hedgehog pathway regulates the stem cell niche actually in the lung.

At the same time, we have had a historical interest in cystic fibrosis, which I will talk about at the end. One of the ways that have been put forward to try to treat cystic fibrosis if gene therapy ultimately fails is to use a stem cell approach, and it appeared to us in the lab that there are two possible ways of coming at it: from our interest in patched/hedgehog in the local niches – that is, the cells that sit in the lung – or this idea that cells can migrate into the lung and take on the characteristics of lung cells. So there are really two parts of the same story for us, albeit, as I said, operating in a small darkened corner of the lab with ‘Alchemy Division’ on it.

(Click on image for a larger version)

The lung is a complex organ, as everybody would know already. It has a proximal–distal axis, and you can see the cell types: the columnar and, as we progress down the airways from the large airways to the smaller airways, ultimately to the alveolar sacs where the oxygen exchange takes place, different cell types with different functions. Some are secretory, some are absorptive, obviously some are involved in gas exchange, some are involved in host defence. It is a complex organ made up of lineages from multiple embryonic compartments.

(Click on image for a larger version)

Shown here is some work from Anjos-Afonso published in the Journal of Cell Science. (There have been lots of publications of this ilk, with controversial findings.) I will explain this experiment, I hope without going into too much detail.

The experiment was to take a mouse, and take out its mesenchymal stem cells from the bone marrow. These were then treated: the cells were cultured on plastic dishes and a virus was infected into them which brought in a protein from jellyfish called green fluorescent protein. And it does just that, it fluoresces green – the problem is that you can’t really see it fluorescing green in lung because there is so much background.

These cells were then injected into a mouse and the researchers looked in the various tissues of the body to see whether there were cells that were green – on this slide, showing more as purple – and whether the cells stained for other markers typical of differentiated cells. In panel B4 you see a cytokeratin, and this is typical of an airway epithelial cell. So this is a little patch of cells which must have been derived from bone marrow which at some point, through either fusion or transdifferentiation – there are a number of other possibilities here – have become cytokeratin-expressing. Whether they participate in gas exchange we don’t know.

(Click on image for a larger version)

Previously to that there was a slightly different approach by Darrell Kotton, as we can see here. This is a similar type of experiment and I am going to show you briefly some other experiments that we have done in this area.

This is actually using bone marrow cells from a mouse which contained the bacterial protein LacZ, which is blue. In this case, the cells were taken from a transgenic mouse and were injected via a bone marrow transplant into another mouse, and then they just scanned all of [inaudible].

(Click on image for a larger version)

At the same time [inaudible] that, using an agent which I will describe in more detail in a moment, using bleomycin, which is an agent used to damage the lung – it is this issue about whether damage is required to promote this process – in mice where they previously had damaged the lung they could get a much greater uptake [inaudible].

So, with our particular interests, we set out to ask a number of questions. Most of these studies have been performed in distal airways – alveoli and the more distal airways, down towards the bottom of the lung – whereas a disease such as cystic fibrosis happens up in the proximal regions of the lung, as does patched/hedgehog signalling, and neuroendocrine cells.

We asked: can the proximal airway be repaired by bone marrow derived cells (a pretty simplistic notion); what is the role of damage in promoting repair by bone marrow derived cells; and, as we have heard from Bob Graham, what is limiting in this process? If you can cause cells to rush out of the bone marrow into your favourite organ, does that make this process much more efficient? Can mobilisation of bone marrow cells by growth factors influence the extent or frequency of repair by airway cells?

(Click on image for a larger version)

So we set up an experiment, as shown on this slide. I am not going to go into too much detail, but this is typically what one of these experiments looks like.

It is to take some female mice and lethally irradiate them so that without some form of treatment they don’t survive, and to take some male mice and remove their bone marrow. These mice are marked with a green fluorescent protein which expresses more or less in all of the cells in their body – although not quite, as I will show you. We then introduce those via a tail vein and perform a bone marrow transplant on these mice, so the only ones that survive are engrafted.

So we have here female mice with the bone marrow from male mice. You can test this by looking in the cells for the presence of a Y-chromosome, for the presence of green fluorescence protein.

And the experiment, laid out broadly, is like this: it is to not injure or to injure the airway, and we used a number of agents. I am just going to show you two today. (It didn’t happen in exactly this order. The growth factor treatment happened first.) We previously had done a lot of work on three growth factors: G-CSF, which has already been mentioned; stem cell factor, which has already been mentioned but didn’t work very well in our hands; and Flt3 Ligand. So all of these, the combination of G-CSF and Flt3L, caused a massive efflux of progenitor cells, CD45⁺ cells, into the lung. I am not going to show that today. You can see at the bottom of the slide the various treatment arms to try to answer these questions.

In order to do this you have to have a model of airway damage.

(Click on image for a larger version)

The airway is quite a nice system. It has got columnar cells; you can see here basal cells and little mucous-secreting goblet cells. Unlike the skin, this looks as if it is in layers, but the lung is a pseudostratified epithelium – every cell in this part of the lung is actually touching the basement membrane shown along the bottom of the slide. It just looks as if it has layers because the nuclei in the ciliated cells are at different levels in the cell. (The mucus sits on the top of the cilia.)

(Click on image for a larger version)

What we have been using is not bronchi but respiratory epithelium of the nasal septum – that is, on the nose. If you can cut through the nasal septum, you can see there is some cartilage, there are some blood vessels, and there is very nice respiratory epithelium in the nose. The reason we used mouse nose is that of all the parts of the respiratory epithelium in the mouse, the mouse nose most closely recapitulates the cellular context of a human respiratory epithelium. It is a much better model than bronchi and trachea.

(Click on image for a larger version)

Here is the model. In the left-hand panel you can see the nasal septum, with the head at the top; you can see where the jaw is. I have marked where the left and the right side would be, and where air would normally be. If we stick a mild detergent called polidocanol down into one nostril, after day one you can see that it has completely stripped off the airway epithelium, right back down to a series of flattened cells. (I won’t talk about the characterisation of these today.) And then, over a 21-day period, this airway regenerates. So we can force it to regenerate.

(Click on image for a larger version)

It is quite a nice system (the blue is marking the nuclei, the red is cytokeratin) because during that regeneration we can see respiratory epithelium at the top left of this section and olfactory epithelium to the right, and so we can tell within this section what is olfactory – being able to smell – and what is respiratory epithelium. And so it is quite a nice model.

We also use bleomycin, which is a glycopeptide isolated from Streptomyces sp. It is actually used in chemotherapy protocols for testicular cancer and some lymphoma. It is a DNA damage agent producing free radicals, and it produces damage equivalent to that of ionising radiation, but the major side effect is pulmonary fibrosis. (I will show you some of this in a moment.) It has a mortality of about 3 per cent and it is, effectively, untreatable.

So what we do is to introduce it into the murine airway. It causes extensive alveolar damage, inflammation and, at really high doses on its own, pulmonary fibrosis.

(Click on image for a larger version)

The other thing to do is to try to characterise the mouse that has the green fluorescence protein in it. You see here a typical section through the donor mouse, and in this case the fluorescent protein is green. To summarise all of this: we can see that within these distal air spaces of the lung, the GFP is expressed in most of the cell types in the lung.

(Click on image for a larger version)

However, in the upper airway – and this is important for interpreting our results – we see cytokeratin, we see the columnar cells marked with β-tubulin, we see the ‘green’ fluorescent protein now marked with red. (This is confusing; it is all computer generated and my lab swaps the colours just to try to trip me up!) In the donor mouse, not every airway epithelial cell is expressing the GFP – maybe about a quarter of them. And this becomes important. In other words, if we see a 1 per cent regeneration of the polidocanol-treated epithelial cells, then maybe it is four times higher, or something like that.

(Click on image for a larger version)

So this is the experiment.

(Click on images for larger versions)

We did the engraftment and they all engrafted very nicely. So we bone marrow transplanted over 100 animals.

(Click on image for a larger version)

This slide relates to an unusual finding that we got, pertaining to the idea of transdifferentiation, or stem cells. After bone marrow transplant we were treating the mice with growth factors, and then we damaged their distal airways – the alveolar spaces – with bleomycin. And, as you will see on the next slide, this does not look at all like a normal lung. Here we have an H&E section, with lots of dark purple areas of fibrosis and inflammatory foci. And if we look at collagen staining, which is a marker of fibrosis produced by fibroblasts, we see there is an incredible amount of fibrosis. These lungs are really, really sick.

(Click on image for a larger version)

It turns out that if we use growth factors – we can see here, this is just one series – after 21 days these mice have lost weight (a lot of weight loss for a mouse, it is only 30g) and ultimately they will die if we keep them going. They have a lot of fibrosis. And you can see here, with no growth factors, a beautiful, pristine lung.

So we had a growth factor model of pulmonary fibrosis.

It turns out that idiopathic pulmonary fibrosis is fibroblast foci, representing areas of active fibrogenesis. So they lay down so much extracellular matrix and collagen that the alveolar spaces get squeezed and then the lung function goes. It is unable to be treated effectively, and it is associated with a very poor prognosis.

It has always been assumed that the source of the cells is intrapulmonary – that is, they are fibroblasts coming from within the intrapulmonary epithelium – though there has been some suggestion in recent literature that there are circulating blood cells called ‘fibrocytes’ which may actually have a fibroblastic-type character.

In our experiments we see that 10 out of 10 (in this series) of the growth factor-treated mice developed extensive fibrosis, compared with only one out of nine control mice. So what does this mean?

(Click on image for a larger version)

This actually presents us with an opportunity. We have here another fluorescence picture, where the red fluorescence is marking bone marrow-derived cells (in this case it is red), and you can see a fibrotic lesion. Because we have done this bone marrow transplant we can actually look at the origin of this fibrosis in these animals. And it turns out that a lot of the fibrosis seems to be coming not from fibroblasts – or maybe they are fibroblasts – but originally from bone marrow-derived cells.

(Click on image for a larger version)

If we look really closely here, the green in this case is telling us they have come from bone marrow (the little arrows there), the red is telling us that these cells are expressing the characteristics – and we have got lots of these other things – of fibroblasts. When we overlay them we see that they have got an orange-y colour.

(Click on image for a larger version)

So this is saying that bone marrow-derived cells in these lesions which are producing collagen are derived from the bone marrow.

(Click on image for a larger version)

This is a heat shock protein which is expressed in fibroblasts which are producing collagen. We can see here a number of cells which are displaying both colours.

(Click on image for a larger version)

This is actually some evidence; it tells us two things. One is that the origin of pulmonary fibrosis is not necessarily just from fibroblasts, or fibroblasts that have come from within the lung. But it is telling us that there is a population of bone marrow-derived cells that are capable of migrating to the lung and either have or are capable of acquiring the characteristics of fibroblasts. Some may call this transdifferentiation and some may call it a cell type previously undiscovered, and it depends, in the stem cell field – that is why it is controversial. It could be interpreted as saying that this is evidence of transdifferentiation.

In any event, we can see that collagen-producing cells and pulmonary fibrosis can be derived from bone marrow progenitors. And, obviously, we are doing more accurate phenotyping at the moment.

(Click on image for a larger version)

This slide relates to the bleomycin model and shows what can be seen if we look hard enough in these animals, in other cell types, in other bits of the alveoli. In the left-hand panel, the arrow is pointing to a green cell, clearly an epithelial cell, and if we run other markers and look hard enough we can see evidence of cells of bone marrow origin that contribute largely to the type 2 cells in the alveoli or, in the lower airways, epithelial cells.

This is a very low frequency and there is a lot of work involved in this; I wouldn’t even want to try to estimate it.

Cystic fibrosis is a common genetic disease due to a defect in chloride ion transport in cells, largely in the upper airway. Individuals with cystic fibrosis have a defect in chloride secretion in some cells, and they have a concomitant increase in sodium, which rips into the cell. That chloride–sodium transport defect gives all of the characteristics of cystic fibrosis. The disease is characterised by a chronic infection with the pathogen Pseudomonas aeruginosa, which eventually leads to inflammation and ultimately death.

If you are going to try to manipulate this system to do anything with cystic fibrosis, you don’t need better alveoli, you don’t need better lower airway, you need better upper airway.

(Click on image for a larger version)

These two slides relate to the polidocanol model; we are back in the nose now. Looking at the non-damaged side you can see the little basal cells, columnar cells, stained red with cytokeratin. You can see where the air would normally be. And the damaged side is shown below the non-damaged side in each instance.

We can see pretty clearly, at some frequency, the presence of bone marrow-derived cells. If we do a 3D reconstruction of the relevant cell, using confocal microscopy and other markers, we can see that it is a ciliated columnar cell. And if we have a look right throughout this epithelium, we can see evidence of these.

(Click on image for a larger version)

These cells are also not positive, by and large, for the hemopoietic lineage marker CD45, although you have to be very careful. This slide shows GFP in red, DAPI for the nuclei, and untreated and treated epithelium with what looks like a very nice cell expressing cytokeratin that has come from the bone marrow, so it looks like an airway cell. And you can see what looks like an inflammatory cell that is also expressing cytokeratin. But if you go down deeply, if you reconstruct this using confocal microscopy, you find it is not a cell. So you have to be very careful. The literature is filled with single-dimension shots of this sort of stuff, where people are saying, ‘Well, we’ve repaired the airway.’

The conclusions are as follows. The good thing about the nose is that it has got a beginning and an end, and we can actually count the cells. And we can take an optical section on the confocal microscopy of five microns. So we can actually get a reasonably good estimate of how many cells within that are showing this phenomenon of having airway marker characteristics and GFP. Approximately 1 per cent of the polidocanol-treated epithelial cells are GFP+/cytokeratin+/CD45-. What that means is that approximately 1 per cent of them are bone marrow-derived – likely – airway cells.

The cell types are mostly those basal cells. If you were going to treat CF you would want them to be columnar cells.

The growth factor treatment in each class has resulted in no statistical difference in the frequency of transdifferentiation. So, unlike the data that Bob Graham showed earlier, when we put all the numbers together we can show no difference between growth factor treatment and not. However, with the number of animals we have got so far, all we can really say in terms of power is that we are not getting a ten-fold increase. If we were getting a two-fold increase, at the moment we wouldn’t even see it. We have to do lots more animals.

This is relevant because, in the case of cystic fibrosis, data show that your therapeutic goal in any of this is to get about 10 per cent of the cells in any epithelium to be functioning again. So how far away is this from a potential therapeutic goal? Well, if you say that this is 1 per cent, and we know that our GFP is only picking up about 25 per cent of those cells, you may conclude that we are almost at about 4 to 5 per cent already, that we are halfway there. And if the growth factor treatment doubled the frequency, we couldn’t tell yet. Then you may say that already this is at a therapeutic level for a disease such as cystic fibrosis.

How would it work? I don’t have time to go through it, but if you ever got that far it would involve, actually, an autologous bone marrow transplant in CF people, and the damage issue could be resolved – people with cystic fibrosis have a very damaged airway, and you would hope that over a period of time the damage that exists in the airway would cause a replenishment from their bone marrow, and replacement in that fashion.

I will not go any further, but just say that when we started this we thought that this may be possible, and I think that there is some suggestive evidence that if you look at extra timecourses you may get the right cell types. And if you look really hard, with a cynical eye, there still remain some cells in some of these tissues that really do look as though they have acquired a respiratory phenotype.

The mechanism we don’t know, but I think that if gene therapy for a disease such as CF doesn’t work, and if other, pharmacological therapies don’t work, then maybe in 20 years’ time we may be looking at this sort of approach. And, hopefully, we will be able to manipulate the resident stem cells using hedgehog.