AUSTRALIAN FRONTIERS OF SCIENCE, 2005

Walter and Eliza Hall Institute of Medical Research, Melbourne, 12-13 April

Session 2: Discussion

Chair – I will start the ball rolling by asking, mainly, David but certainly Trevor too: has there been any practical spin-off from all of this research on cell death? Have any of these antagonists been tested in clinical trials?

David Huang – For the Bcl-2 protein family themselves, there is a good candidate compound made by one of the major pharmaceutical companies that is fairly close to clinical trials, and that to me is the most promising one at the moment. There are other approaches being taken, including targeting the expression of Bcl-2, for example, but I think the data from that is much more equivocal for us to evaluate.

My background is clinical. One of the interesting things for me in studying this system is really that we have all these ideas that the Bcl-2 protein family are important for tumour formation and chemoresistance, and it would be nice to be able to (a) prove that in a clinical setting, and (b) actually make a difference for patients. That, to me, would be the most important end point for that.

Certainly five years ago I would have been very sceptical, but looking at our data and our evaluation of the lead compound from one of the pharmaceutical companies, I am much, much more optimistic.

Question – You have described what I guess are generic machines. Is there a tissue-specific control here? You get transcription differences in different tissues, and presuming you do that – to give tumours or whatever in different tissues – do you see it that rather than operating at the protein–protein interaction level where drugs might operate, you might operate at the transcription level and knock out one of the key components?

David Huang – I think at a transcriptional level for the Bcl-2 protein family as a whole is probably going to be more difficult to target, especially those proteins that promote cell survival, because we know from experiments in mice where these genes have been knocked out that you don’t usually get a phenotype unless you have a homozygous knock-out. So the problem I have with targeting the expression of genes transcriptionally is that I suspect the knock-down will not make a significant difference, because most of the heterozygous knock-out animals are complete normal.

In terms of cancer therapy, of course, with the BH3-only proteins you want the flip side. You want to upregulate the expression. And I don’t think we know enough yet to target them. So at the moment I think our best hope is with understanding protein–protein interactions, and modulating protein–protein interactions. I think the selectivity and the signalling pathways are actually very amenable to manipulation in that regard. I just feel that with the transcription targets, the genetics is telling us that it is going to be tough to do it by that route.

Question – I am intrigued by the process of chemoresistance. I didn’t quite understand the mechanism, but my question concerns: what is its role in a natural situation without the cytotoxic drugs?

David Huang – Obviously, in a tumour cell situation, a tumour cell as we understand it acquires many mutations that will allow it to survive even under very dire circumstances. I think there are a number of features of tumour cells that make them much more prone to acquire additional mutations. So you can now select for tumour cells that are more chemoresistant, more resistant to standard cytotoxic drugs, than when it started off.

In a clinical situation, when you first start treating patients, usually in the first instance the tumour often responds very well to the cytotoxic treatment. And what almost always happens is that you select out those cells that have acquired additional mechanisms that will allow you to counter the action of the cytotoxic drugs.

The important reason for understanding chemoresistance is that if you know what the mechanisms are, or you can predict which are the likely important chemoresistant mechanisms for that particular tumour, that will, hopefully, allow you to pre-empt that situation so that, in the case of Bcl-2, you could design specific molecular targets for those mechanisms that might arise as you select the standard cytotoxic drugs.

Chair – I would like to add a couple of comments to that. All cells – cancer cells and normal cells – are mortal. So you can kill them, if you inhibit some process that is required for their ongoing survival. (Tautological, but true.) Now, the process that has been discussed today is not the process by which cells are killed but it is the mechanism that has evolved so that cells can commit suicide.

It turns out that, as well as cells committing suicide during morphogenesis and to maintain a constant cell number, the cell suicide mechanism is also activated when cells become stressed. If practically any signal or toxin that can kill a cell by interfering with an essential metabolic pathway is added at a lower dose, a sublethal dose, or added slowly, that will generate stress in the cell. The cell will detect that some pathway is being interfered with and it will respond by what is called the stress response. It may try to repair the damage, but a very common stress response is for a cell to activate its cell suicide mechanism.

This cell suicide in response to stress is why, when you look at cells killed by so many different agents – all of the chemotherapeutic agents, irradiation, many different things, all these different toxins – even though they had different mechanisms of action, the morphology of cell death looks like apoptosis. It looks like a suicidal cell death. We don’t yet know, or are uncertain, when chemotherapy is used to treat a cancer, how much of the cancer cell death is due to the direct toxic activity of the drug inhibiting a vital metabolic pathway, and how much of it is due to a sublethal insult that nevertheless causes cell death because the cell commits suicide.

The reason that overexpression of genes such as Bcl-2 can give a general resistance to a wide range of chemotherapy is not that it inhibits the direct activity of all these different drugs; it is that it stops the cell suicide process. If you could get cancer cells to do the right thing and kill themselves by activating their suicide mechanisms, you might be able to do it with a much lower dose of drug than you would need to kill a cell outright. If you need the higher doses of drugs to kill the cells directly, then you will probably kill all of the normal cells as well as all of the tumour cells.

That is a long way of going round it, but it explains why you see a general drug resistance in cells expressing lots of Bcl-2.

Question – A number of the Bcl-2 family members have been proposed to be expressed on other organelles, like the nucleus and the endoplasmic reticulum. Do you think that the same sorts of processes are going on for the expression and processing of those molecules on those organelles, or is there something else that is going on there?

Trevor Lithgow – The short answer to your question is that we don’t know. it is very clear, as you said, that some members of the Bcl-2 family of proteins are formed on other subcellular membranes. In the case of Bcl-2 itself, in fact, it is found on the ER (the endoplasmic reticulum) and also on the mitochondrial outer membrane. Some of them are restricted only to the mitochondrial outer membrane, and I guess there are some that might be found only on the endoplasmic reticulum membrane.

Whether those proteins – at least from my reading of the literature, because it is not really my field – are acting in the same way at a different location, or whether they are doing something a bit different at the different location, I think is still relatively open. There is some evidence that all of these molecules are affecting permeability of membranes, and of course the permeability of all of these membranes is an important factor for cells staying normal.

There is also the possibility that they are actually communicating with each other – that is, that the Bcl-2 family over somewhere else is communicating, via some other signalling molecule, with the other members of the family, though I guess that part is a less attractive model.

Question – Trevor, you talked about programmed cell death, I guess, in the animal kingdom. I was just trying to broaden it out a little bit. I am just wondering whether plants make use of the other organelle occurrence of that. Also I wonder if there are any bacterial homologues which don’t have the advantage of that organelle.

Trevor Lithgow – At least as far as I am aware, there are no Bcl-2 family members that are found in plants. Certainly plants, as an organism, do make use of cell death programs – I will choose my words carefully here, because I am not so sure that they are necessarily functionally or mechanistically completely the same sort of process – and certainly there are molecules in plants that must be mediating the process and may or may not be present in animals.

But in terms of the animal systems where we know the molecules – we know which ones they are and are starting to understand what they do – I think in the case of the Bcl-2 family, for example, the plant genomes don’t encode proteins like that. That is also true in fungi. Animals and fungi are much more closely to each other than they are to any other sort of eukaryote, and fungi also don’t have Bcl-2 type proteins and almost certainly don’t undergo any sort of programmed cell death.

As for bacteria, we have had a few discussions and I know there has been some speculation in the literature. Maybe single-cell microbes have reasons for committing altruistic hara-kiri but I’m not really convinced about that. Certainly, again, bacteria don’t have mitochondria and they don’t have Bcl-2 type proteins. And I am not even really sure that they have serious programmed cell death.

David Huang – I think the only thing I would like to add is that in plants, as far as I am aware, there is a phenomenon called a hypersensitivity response, in which selective parts, particularly in the leaves, can die off in response to particular toxins. Although there are no Bcl-2 homologues in terms of sequence and structures, people have done experiments where they have overexpressed mammalian Bcl-2, a worm homologue of Bcl-2, in plants and have shown that that suppressed the hypersensitivity response. Now, what is the molecular relationship between that hypersensitivity response in plants and the programmed cell death in mammals is really not clear.

The only other family of proteins of which there are some orthologues in plants is enzymes that are more downstream of Bcl-2, particularly the caspases. There are some very distant relatives of them in plants, but the direct relationship of that is really unclear at the moment.

Question – Trevor, I was interested in your mention of the evolutionary relationships in the mitochondria of different animals. It is supposedly reported that sharks, for instance, don’t often get cancer. Is there any usefulness in looking at, perhaps, shark mitochondria in particular, to see if there is anything going on there?

Chair – Can I just interrupt and say it’s a myth. But you can talk about the mitochondrial aspect, Trevor.

Trevor Lithgow – Ah well, I don’t think there is anything to say now!

It is definitely true that getting mitochondria was either the defining event or one of the two defining events in eukaryotic evolution. All eukaryotic organisms have mitochondria and many of the mitochondrial functions are conserved, and so you expect that shark mitochondria, in most respects would be similar – very, very similar, I would suggest, although I have not read anything about this – to our own mitochondria in terms of what they are capable of doing.

As for programmed cell death in fish, I don’t know that it has been studied but they are vertebrates still so I guess they would have the sorts of molecules that we have in our own bodies, and that mice and worms have.

Chair – In broad terms, the mechanisms of programmed cell death are conserved amongst all the metazoans, all of the animals. You can take, say, a human Bcl-2 and express it in a nematode that is roughly a billion years separate in evolutionary terms and it will still function to inhibit cell death in a worm. Now, in a lot of other phyla – sponges, hydra – there have been components of the cell death mechanism that have been recognised, such as Bcl-2 family members, caspases and similar adaptor proteins.

Sharks, being vertebrates, are very similar – they are practically our cousins. So cell death pathways are very similar in them.

Most work in cell death in fish has been done in the zebra fish, and many of the homologues of mammalian cell death components have been cloned and analysed in zebra fish.

Question – Trevor, does the size of the mitochondrial pore mean that most proteins that pass through it have to unfold?

Trevor Lithgow – Yes.

Question (continued) – If they do, what things have you got on the other side as chaperones to help them get back together?

Trevor Lithgow – I definitely skipped over all of these details. The thousand proteins that have to get into the mitochondria are made outside the mitochondria, and when they are made, they are made as a linear molecule. And most of the proteins that function somewhere else in a cell will rapidly fold into a tight and globular three-dimensional conformation.

What the questioner is referring to is the fact that to get across a membrane the pore is actually quite restrictive. In fact, we know from the diameters it is 20 Ångströms at minimum diameter across the pore. What it literally means is that the protein molecules that are being imported have to be still in a linear conformation, more or less completely extended. That is a dangerous situation to be in if you are a protein molecule inside a cell, and there are professional helpers that help these proteins to get in.

There are HSP70 type proteins on the outside. On the inside, we suspect – and this is still under debate, I guess – that the machines actually come together as modules, so that as a protein is going across the outer membrane it would literally immediately engage the next machine in the series and continue to be unfolded until the final machine, whichever one that should be (it is dictated by the individual protein sequences) has had done with it. And then a chaperone on the other side will rapidly coil the protein up and fold it, compact it back down. There are suitable chaperones in each of the compartments of the mitochondria to do that, so depending on which of the machines have come together to channel the protein to where it should go, at the other end there will be factors waiting that will put it together again, or put it together in the first place.