AUSTRALIAN FRONTIERS OF SCIENCE, 2005
Walter and Eliza Hall Institute of Medical Research, Melbourne, 12-13 April
Molecular switches controlling cell death
Dr David Huang, Senior Research Fellow, Walter and Eliza Hall Institute of Medical Research
What I would like to do today is to
describe to you some of the work that we have been doing on the molecules and
the mechanisms by which a process called apoptosis, or programmed cell death,
is regulated, particularly in mammals.
(Click on image for a larger version)
The story really starts with work, including work by John Kerr in Queensland, using the electron microscope to look at the process by which cells die. What Kerr, Wyllie and Currie in the early 1970s were able to describe was a very stereotypic process by which cells apoptose, and they described this in terms of the morphology of how cells die.
This process occurs in our bodies all the time. If you think about it, even just within our blood system billions of cells are made every day, so in order to counter that huge number of cells that are generated in excess, there have to be processes such as programmed cell death in order to keep the numbers in check.
(Click on image for a larger version)
The next part of the story came in the mid-’80s, from people who were looking at genes that are involved in various cancers. One particular gene is called Bcl-2, which is the second gene that was associated with B cell lymphomas. Work by various groups, including a particular group in Philadelphia, identified a particular chromosomal translocation that moves the normal location of the Bcl-2 gene on chromosome 18 to an aberrant location on chromosome 14. It is a result of this chromosomal translocation in this particular form of lymphoma follicular lymphoma that causes the Bcl-2 gene and therefore the Bcl-2 protein to be overexpressed in this lymphoma.
(Click on image for a larger version)
The next step in the story came from work done by David Vaux, when he was a PhD student in this institution. David was trying to elucidate the normal function of Bcl-2. Up to that time a whole range of oncogenes had been discovered, and most of those oncogenes, or genes that can drive cancer formation, are thought to drive cellular proliferation, so that when they are overexpressed those genes cause the tumour cells to divide and proliferate much more quickly than their normal counterparts. What David showed instead was that Bcl-2 was not a gene that caused cellular proliferation; instead, what Bcl-2 did was to promote cell survival.
In this very elegant experiment, what David did was to overexpress this gene Bcl-2, mimicking the situation in patients with follicular lymphoma. What he discovered was that in this cell line, which normally dies very rapidly when the growth factor is withdrawn, because the growth factor is required to keep the cells alive, Bcl-2 is able to keep the cells alive for prolonged periods of time.
My major interest is really in trying to understand the molecular controls of programmed cell death, apoptosis, in mammals, particularly in relation to the function of Bcl-2. In short, what we can say is that Bcl-2 functions as an ‘ON’ switch for cell survival. So when you have plenty of Bcl-2 there, the cells survive. What forms the basis of much our work is really trying to understand the function of Bcl-2, what are the normal controls on Bcl-2 function, and how Bcl-2 overexpression can contribute to tumorigenesis.
(Click on image for a larger version)
Experimentally it has been proven that if you drive the pro-survival signal Bcl-2 by keeping the ‘ON’ switch, that can promote tumour formation. This is an experiment that Andreas Strasser did when he was a postdoc here. What he showed was that in a mouse model this is looking at the incidence of lymphomas in the mouse model if you combine a gene that causes cellular proliferation, such as Myc, with a gene that promotes cell survival, such as Bcl-2, you get a much more rapid onset of lymphomas than that that drives cellular proliferation alone.
It is now realised that many oncogenes such as Myc that drive cellular proliferation are not sufficient alone for tumour formation. What you often get is additional mutations that also promote cell survival. And that combination of genes that drive cellular proliferation with mutation in genes that promote cell survival has turned out to be a very potent combination in causing tumour development.
(Click on image for a larger version)
The key message is that the regulation of cell survival is essential for normal health. As I mentioned, there are billions of new cells made every day in the blood system alone. That has to be held in check. If you have excessive cell death compared with the number of cells that are produced, you can get degenerative conditions, and that is one consequence of deregulation of cell death, or apoptosis. On the other hand, in cancers, for example, where in addition to excessive proliferation you also get impaired cell death, that is a situation where you get an accumulation of abnormal, unwanted or damaged cells.
(Click on image for a larger version)
Overactivity of Bcl-2 not only contributes to tumour formation. I gave you the example right at the very beginning that Bcl-2 was first cloned as a gene that was overexpressed in follicular lymphoma, and in experimental models when you overexpress Bcl-2 tumours can form. In addition to that, Bcl-2, when it is overactive, can also contribute to chemoresistance. This is a very important clinical problem, because often the limitation to our current cytotoxic therapy is that the patient’s tumour samples become resistant to the conventional cytotoxics.
This is an experiment that was done in our lab. If you take a particular tumour, for example, that is overexpressed as just Myc, if you treat mice harbouring those tumours with a standard cytotoxic drug, those mice invariably recover. So almost all the mice that were treated by the standard cytotoxic drug now are cured of this experimental model of lymphoma. However, if you use exactly the same tumour, the only difference being that Bcl-2 is now overexpressed, these mice invariably still die of the tumour. So overactivity of Bcl-2 can confer chemoresistance.
(Click on image for a larger version)
Understanding Bcl-2 function is important for at least two reasons, because Bcl-2 overexpression can promote tumour formation, and overactivity of Bcl-2 can also blunt the response to many cytotoxic drugs. So there are many groups, including a number of pharmaceutical companies, who are interested in targeting Bcl-2 specifically as a novel target for cytotoxic therapy. As Bcl-2 acts as an ‘ON’ switch.
(Click on image for a larger version)
The next question we had was: what turns the activity of Bcl-2 off? As I say, the normal state in the body is that billions of cells are made every day. What turns the activity of Bcl-2 off in the dying cell in order to switch off cell survival?
(Click on image for a larger version)
One of these ‘OFF’ switches is a gene that we discovered in the mid to late ’90s called Bim. What Bim does is to antagonise the function of Bcl-2. In this experiment here, for example, you can see that this is a modern version of the David Vaux experiment. When you overexpress Bcl-2, which is the ‘ON’ switch for cell survival, it promotes cell survival compared with the control. If you now, on top of Bcl-2, antagonise this Bcl-2 action with an ‘OFF’ switch like Bim, you trigger cell death and counter the activity of Bcl-2.
(Click on image for a larger version)
Perhaps this was shown more elegantly when Philippe Bouillet created mice in our lab which lacked the Bim gene. Remember that Bim functions as an ‘OFF’ switch, so it promotes cell death and antagonises cell survival. What Philippe found was that if you knock out the Bim gene you get an accumulation of blood cells in the hematopoietic system consistent with the idea that Bim is an important negative regulator of cell survival.
(Click on image for a larger version)
These ‘ON/OFF’ switches of the Bcl-2 family of proteins belong to a large protein family. There are some, like Bcl-2 itself, that promote cell survival, and others that antagonise the action of Bcl-2. The two particular classes that I would like to focus on first of all are the pro-survival proteins such as Bcl-2 and the antagonist BH3-only proteins. They share a very limited sequence homology that is restricted to these regions called the Bcl-2 homology domains, and it is these signature homology regions that characterise these proteins.
(Click on image for a larger version)
The BH3-only proteins, of which Bim is an example one that we cloned here and that I introduced you to earlier on seem to act as damage sensors. What seems to happen is that the BH3-only proteins are normally held in check either by their different locations within cells or by transcriptional mechanisms, until damage signals converge upon the cells. These BH3-only proteins are then activated to bind to the pro-survival proteins. Structural work has shown that the homology regions fold into the groove, as it were the receptor, which the death ligand binds to.
(Click on image for a larger version)
Work that has been done in a number of labs, including our own, has established that a tight binding of these ‘OFF’ switches to the pro-survival ‘ON’ switches is what triggers apoptosis. For example, here the high-resolution helical structure that was described by Kappler and his colleagues shows that in the bound state the death ligand forms an alpha helical region that binds onto the pro-survival protein.
So, in the ON state, pro-survival molecules such as Bcl-XL promote cell survival. In the OFF state, the BH3-only molecules such as Bim antagonise the action of the pro-survival molecules to promote cell death.
(Click on image for a larger version)
One question that came from the whole variety of ‘OFF’ and ‘ON’ switches was: is there any binding selectivity, any interaction selectivity between these classes of molecules? One could imagine that all the death ligands, all the ‘OFF’ switches function in the same way so that they antagonise the whole array of pro-survival proteins, or they might act more selectively.
(Click on image for a larger version)
This is a question that three postdocs in the lab, Lin Chen, Simon Willis and Andrew Wei, recently addressed.
(Click on image for a larger version)
Contrary to what we anticipated, they found that in fact most of these ‘OFF’ switches act selectively so two examples here act very selectively whereas only a restricted number of them bind promiscuously.
(Click on image for a larger version)
What is even more interesting is that those promiscuous binders are very potent killers. In this experiment what they were comparing with a control was the effects of overexpressing some of these promiscuous death ligands. And what we found was that these promiscuous binders bind all the ‘ON’ switches, all the pro-survival proteins and cause death very potently.
(Click on image for a larger version)
In contrast to that, those ‘OFF’ switches that are highly selective are very poor killers. So these are some examples here. In particular, the two BH3-only proteins, Bad and Noxa, were very, very weak killers.
So the question we had in mind from a mechanistic point of view was what was the underlying for their weak activity. Were they weak because they only had selective targets? If so, then one would anticipate that combining something like a Bad or a Noxa would produce very potent killing that you would get with a very promiscuous binder. So a simple experiment was set up to test whether these two ‘OFF’ switches function in a complementary way.
(Click on image for a larger version)
You see here the result of the experiment, and it is quite clear that if you combine a selective ‘OFF’ switch like Noxa with another one that binds a complementary set of pro-survival targets, they can now potently induce cell death in contrast with the action of one or the other on their own.
(Click on image for a larger version)
What is important about these studies is that they allow us to unravel the complexity of the signalling network that regulates cell survival. What we are doing is to take a biological problem and try to figure out what are the interactions between the Bcl-2 family of proteins that control cell survival.
The key feature of this system appears to be that there are some selective ‘OFF’ switches that can work in concert with each other to induce cell death.
The implication of this work is that it may be possible to design small chemical entities that mimic the action of these more selective BH3-only mimetics. The reason that may be important from the practical point of view for drug discovery is that if you target all of these pro-survival proteins you are likely to cause a lot of toxicity both to tumour and to normal tissues. If you have a BH3 mimic that is much more limited in this action you might be able to spare toxicity to normal tissues, particularly if you target those pro-survival proteins that are expressed in particular tumour types.
So that is one of the important reasons why unravelling the signalling pathways to apoptosis might be important.
(Click on image for a larger version)
What it also suggests and is important for us is that the pro-survival ‘ON’ switches are not functionally equivalent. As I showed you in a previous slide, it appears that they fall into at least two distinct classes, and using this information about the selectivity we are really trying to tease apart what it is about these ‘ON’ switches that functions differently.
(Click on image for a larger version)
So how is it that the pro-survival proteins such as Bcl-2 can promote cell survival? We have got a good understanding of how ‘OFF’ switches, the BH3-only proteins such as Bim, antagonise the action of ‘ON’ switches, but how is it that Bcl-2 acts downstream of it in order to promote cell survival?
(Click on image for a larger version)
It looks as though the pro-survival proteins fall into two distinct classes, one shown here in blue, exemplified by Bcl-2, and another shown here in black, exemplified by Mcl-1. Do these two classes of protein control the same target, so that you need to knock each of them off in order to get cell death, or do they control distinct targets?
(Click on image for a larger version)
It turns out that the targets of the pro-survival proteins are a third class of proteins of the Bcl-2 related family. Structurally they, in fact, very closely resemble the pro-survival proteins, and one of the interesting questions is: what is it that makes one protein a pro-survival protein in contrast to one that promotes cell death?
It turns out that molecules of the pro-apoptotic class, particularly two particular ones called Bax and Bak, are the direct targets of the pro-survival proteins.
(Click on image for a larger version)
And it appears that the reason we get two functional classes in our killing assays is that one of each class in this case here represented in black, called Mcl-1, and another one in blue, Bcl-XL those two molecules control the same downstream target, Bak.
(Click on image for a larger version)
What you need is to antagonise the action of both of them in order to get efficient cell death. So if you have a selective ‘OFF’ switch, in the case of Noxa, that only targets Mcl-1, you don’t get efficient killing because there is a second brake on the activation of Bak. On the other hand, if you just go after Bcl-XL, you don’t get efficient cell death either because there is another brake on Bak activation. You need to antagonise both of these proteins in order to get efficient cell death.
(Click on image for a larger version)
What is important about this work as well is that it appears that Bcl-2, which is the molecule that I started with, and in fact is the molecule that is most commonly overexpressed in tumours, has no role in controlling the death mediator Bak. So it is possible now to bypass even Bcl-2 overexpression by going after two other pro-survival molecules.
What I have related to you today is some of our work on how we think the key switches to cell survival, particularly those of the Bcl-2 protein family, function, and I have provided evidence to you that their deregulation can contribute to diseases such as cancer. It is my belief that understanding how the interactions between them control cell survival and cell death would allow us to open possibilities for novel therapeutics in order to trigger killing, particularly, of cancer cells.
