This Workshop will focus on understanding the nature of cancer in terms of the flow and control of information, and the activities of various regulatory and signaling networks in a systems context.

Topics will include GRNs, network theory applied to signaling pathways, chromosomal reorganization, systems biology, attractors, epigenetic landscapes, and thermodynamic and information theoretic aspects of stressed biosystems.

• Workshop Agenda
• Participant List

Group Photo of the workshop participants

Paul and Pauline Davies

Dr. Sara Walker and family

Audio Interviews and Transcripts from the Workshop

• Interview with Tim Newman, a professor of Biophysics and SULSA Research Professor of Systems Biology at the University of Dundee in Scotland. His talk on Buffered Qualitative Stability, which he calls BQS, generated a lot of excitement at the workshop Professor of Biophysics and SULSA Research Professor of Systems Biology at the University of Dundee in Scotland. Interview Transcript
  
• Interview with Christoph Adami, professor of Microbiology and Molecular Genetics and a Professor of Physics and Astronomy at Michigan State University.

  He discusses his experience of the workshop. Interview Transcript
  

• Interview with Sui Huang, professor and cancer researcher with a medical degree working at the Institute or Systems Biology in Seattle. He tries to understand cancer from very basic principles using a theoretical ‘landscape’ model first created by Conrad Waddington, the founder of systems biology. Interview Transcript
  
• Interview with Carol MacKintosh, Professor of Molecular Signalling at the University of Dundee in Scotland. She has discovered that spare copies of signalling pathways genes left over from evolutionary history are scattered throughout the genome and these make pathways that operate in teams. Over half of all cancers utilize a pathway in one of these teams and the essential pathways for cancer in these teams are non-mutated. These could be new targets for anti-cancer therapies. Carol’s fascinating story begins 500 million year s ago with a little animal that lived in the sea. Read more on Carol MacKintosh | Interview Transcript
  

Interview with Tim Newman

(Back to Audio)

Tim Newman: Well, we’re very excited about this project, so it’s probably the most exciting thing I’ve been involved in in my career. So, essentially, it started off with a coffee room discussion about what is robustness and my colleagues, I should just say who’s is involved in this; myself and then Julian Blow, who is a molecular biologist and Luca Albergante, who is a post doc with a background in computer science.
We were having a coffee one day a few years ago and got into an argument about what is robustness and how does it connect to evolvability, and Julian had this wonderful insight that somehow because cells are so complex and be described by so many parameters, the fact that they are robust means they can’t really, their robustness can’t depend on parameters, but he said the magic word that somehow it’s qualitative and then the whole project went off from there. And Luca found this theory from the 1960s that was written down by economists Ruppert and Quirk, called ‘Qualitative Stability,’ which basically says if you have a network – and they were thinking about a network of companies – is their particular shape they can have, so that it’s always stable, no matter how strong or weak the links are. And so we took this idea and applied it to gene networks and found that Ruppert and Quirk’s theory from economics actually holds in gene networks, which was a remarkable sort of application of one area of science to another. But because in biology, things evolve, we can take their theory a lot further, so that it’s not just a question of these gene networks obeying Ruppert and Quirk’s theory about having certain shapes that make them very stable, they also have to buffered against losing that stability through mutations and evolution.
So, we essentially have this buffered qualitative stability theory – in one sentence which is that gene networks are designed to be free of long loops and also to be buffered against getting long loops – and just from that one statement, you can make lots of predictions about what gene networks should look like and then we can test those predictions against gene networks that have been measured for bacteria and yeast and human cell lines and the remarkable thing is all the predictions that we make from this very simple theory are found to be true. So that’s very exciting.
Pauline Davies: Well, that’s just incredible because you know just superficially, who would ever have suspected it, but of course it does make sense, doesn’t it to give that stability?
Tim Newman: Absolutely and the reason I’m at this conference is because we did one final thing which was to then ask, “If we take a cancer cell line from a human cell line, does it satisfy the rules or not?” And we find, that whereas for e-coli and yeast and the non-human cancer cell line which obey all of these different predictions for stability, actually the cancer cell line breaks every rule. So, it’s really sort of the bad kid on the block and it’s clearly through evolution and the body giving up all of these very carefully created patents in the gene network, throwing them all away and becoming actually non-robust, and we think what this means is that it allows the cancer cells to be phenotypically plastic, so they can respond to all kinds of different environments in the body which would be a very strong survival strategy for cancer cells.
But I think the thing I’d like to stress is that this is a theory which has no parameters whatsoever and all the predictions are either just right or wrong, and so it’s been extremely exciting to find you can make a myriad of predictions where there’s absolutely no room for error, they’re either right or wrong, and we find in wild type cell lines and e-coli and yeast and so on, they’re all seem to be confirmed, so we really think we’re on to something about how gene networks are made.
Pauline Davies: Can there in the long term be any implications for treatment or prevention?
Tim Newman: I think it has actually because what I should say is that the, even though the cancer cell line that we studied breaks all of the rules, there’s only a handful of genes that are misbehaving in the network of hundreds of thousands of genes that actually cause the network to break these rules, and so by looking at buffered qualitative stability, you can immediately zoom in to a handful of genes that are sort of key players in disabling the cancer cells robustness, so this gives you a completely new way of thinking about targets. And what’s even more exciting is by removing just one or two genes from this handful; you can essentially resurrect the robustness of the cancer cell again. So, I think if one can look at the gene networks of cancer cells from different kinds of cancers, identify – assuming that they have these loops and break BQS – identify the genes that are misbehaving, it gives you a completely independent new way of identifying strong targets for gene therapy.
Pauline Davies: Well, you’d need to study even more networks, don’t you?
Tim Newman: So, yes, so basically this is an example now where the theory is ahead of the game. We’ve studied all of the data that is out there, so we just have to wait now for a year or two, hopefully experimentalists will get excited about this theory and accelerate the pace at which they are generating these transcription networks. But the theory is sitting there waiting to be wielded, and we’re just waiting now for these large teams to generate more networks.
Pauline Davies: Well congratulations.
Tim Newman: Thanks Pauline, it’s great to see you.

Interview with Christoph Adami

(Back to Audio)

Christoph Adami is a Professor of Microbiology and Molecular Genetics and a Professor of Physics and Astronomy at Michigan State University. He discusses his experience of the workshop.
Chris Adami: Even though I’m in the department of Microbiology and Molecular Genetics, I work mostly in evolution and mostly inside of the computer. So I’m a theorist and if you want to make progress in disease, for example, you need to know lots of details and there are some people here that do fantastic things with actual experimental systems, but also with computational and theoretical systems, and so for me, I can learn a lot of things here that I could not learn in my normal environment.
Pauline Davies: Can you pick anything out that has meant a lot to you?
Chris Adami: One of the talks today made the point that you can have adaptation in cancer cells and normal cells without any mutations and that’s something that I was kind of aware of but that I also tend to forget, because I work with mutations all the time. All this stuff that I do is with evolution is with mutations and that makes me forget about the fact that cells have the capacity to change their state as a function of environmental influences and then therefore that many of the things in the changes that we see in cancerous cells may actually not be mutation-related, so that was a good reminder for me.
Pauline Davies: Right. I know you were instrumental in some ways in setting up this PS-OC Network. Some of the initial conversations I believe were with you. So looking back at this, as an outsider really, because you haven’t’ been involved directly in the last five years, what do you think?
Chris Adami: I think first of all, that this was a fantastic idea to start these things. I wasn’t aware that throughout 30 years of research in cancer, so little progress has been made. As a theorist, it turns out that I like to think in fundamental terms and that not many people in cancer biology were doing that. And so, putting together cancer biologists and physicists and engineers and chemists, I think was a fantastic idea because the different cultures needed to interact. On the one hand, the physicists possibly did not have access to all the kind of clinical data that was necessary in order to make interesting theories and on the other hand, the clinical people did not have the kind of broad thinking that physicists would bring to the table. So, I think in the long run, it’s going to pay off handsomely. I think five years is probably too short a period to measure success. My hope is that these centers or other centers will go on.
Pauline Davies: Yes, I notice myself because I’ve been involved for the last five years, just seeing the level of discussion now that we have the physicists and their computers specialists and they’re giving some fairly very high-powered talks, but they’re bringing in biology that they might not have known about those years ago.
Chris Adami: Yes, I think that’s the key. When I talk to biologists, I have a very good biologist friend, and he always tells me, “I like talking to you because you are the kind of physicist that actually cares about details.” There are two kinds of physicists, those that care about the biological details and those that don’t. Those that don’t are useless for the biologists, so I mean these details matter. You cannot really build these types of theories in a vacuum; you need to have access to facts and very often, very, very detailed facts, and it is very often that minute, small details that perhaps even an experimentalist would overlook that end up making big breakthroughs, where somebody just looks at a fact and says, ‘Hmm, I wonder how we could understand that.’ It is in my view always the experiments and the facts that bring about new thinking. And in the absence of those, you’re just sitting in a room trying to imagine how the world is, and I don’t think that’s very successful.
Pauline Davies: Actually, that’s how major breakthroughs in physics have also occurred.
Chris Adami: That’s right. I don’t think what we’re doing here is dramatically different from how we’ve been doing it in physics, you know, when I was a graduate student in physics, I had the same type of approach. I always said ‘I need to look at data.’ Because in the absence of data, I don’t know what to even think about. I want to see data that nobody else has seen, data that is new, so that I can take a look at it, see the things that I understand and then see something that makes me stumble, and it is these type of things that open up new directions, because once you take a look at them and then you mull them over and you discuss them with other people and then at some point you might say, “You know what? I think I know where this comes from!” And then that’s where the beginning of research is. The beginning of research starts in my view, with data, with data that you don’t understand, but in order to have access to this kind of data, you have to have these collaborations that are being successfully put together by these PS-OC centers.
Pauline Davies: Thank you very much Chris.

Interview with Sui Huang

(Back to Audio)

Sui Huang is a professor and cancer researcher with a medical degree working at the Institute or Systems Biology in Seattle. He tries to understand cancer from very basic principles using a theoretical ‘landscape’ model first created by Conrad Waddington, the founder of systems biology.
Sui Huang: I think the biggest question I ask myself is “Why is it so hard to treat cancer?” I have given up trying cure or eradicate it, but first we need to understand why it is so hard to treat cancer. It is exactly the same as why is it so hard to fight the war against terrorists, because we never kill all of them – those that survive, they adapt very fast, not through a Darwinian evolution but predisposed to learn. You know when we kill 80% of terrorists in some way or cancer cells, the surviving ones are not innocent bystanders that just happened just to survive; they are not those that have mutations that make them more stronger and fitter but they actually adapt actively to the perturbation that we use to kill the other cells. But essentially it is a price we pay for a very complex development that we have evolved, and so it’s kind of a leftover, things that are not perfect.
Pauline Davies: And you have a very interesting way of describing it – you’re using a landscape, aren’t you, a landscape of possibilities with valleys and peaks and cancer evidently occupies some of the previously unoccupied valleys in that landscape.
Sui Huang: Yes, so that’s a metaphor that goes back to Waddington who uses landscape to describe development.
And the landscape gives us a lot of clues about that – the potential behaviors and the idea that when you treat cancer you have all those surviving cells and those are actually more malignant, and that’s all predicted by the landscape model. And then we move more and more towards new valleys, side valleys that are more and more away from the normal phenotype and so I think that’s a problem. It’s almost like entropy; we diversify them, we disperse them through this landscape and then they go through new states and new phenotypes. And we know now that this landscape can be computed based on the gene network and if we do so, then you will see you have many more potential cell types and we think that these potential never-occupied occupied cell types, when they are occupied, then they become cancer.
Pauline Davies: How did you come to find your model?
Sui Huang: It was a systematic search and I’m not the first. You know people previously suggested that cell types are these valleys in these landscapes and one can mathematically show that, and Stuart Kauffman proposed a long time ago that cancer cells could be those unoccupied attractors. And now we know, given what we know given about gene network, that there must be probably maybe many unoccupied attractor states. And that we also know that the beautiful thing about this is that it opens the possibility that you could get cancer without mutations and we have now a few examples of that, but mutations dramatically accelerate cancer.
Pauline Davies: How can you get cancer without mutations according to the landscape model?
Sui Huang: Essentially development is trying to go down to the lowest valleys and evolution has carved optimal paths to go down but, through some accidents you can get stuck further up in side valleys, so to speak. And to get to those side valleys there are all kinds of ways like, like toxins, carcinogens, mutagens and chronic injuries and stress can lead to cells being deviated from these highways down to the normal phenotypes. But mutations would just be a brutal deviation, so it’s much more likely to achieve that sort of mutation.
Pauline Davies: Does this landscape model have any predictive power?
Sui Huang: Yes, well it explains a lot of things. It is supposed to predict, and that was the old idea by Stuart Kauffman, that one should be able to apply differentiation therapy, try to differentiate cancer cells back to these normal valleys or high ways down to the normal state. And that has now become really popular with the idea of cancer stem cells, but it turns out that it is probably more difficult because once you’re in the side valleys, the path to the normal valleys is very, very rugged and so it’s probably harder than one thinks. So the qualitative properties of this is predicted this landscape model.
Pauline Davies: So do you think the main power in this landscape model is to show people from other disciplines just, where cancer might lie and the possibilities?
Sui Huang: So that was one original idea of visualization of very abstract ideas, but then it turns out that it’s much more useful than that, because if you have a landscape then that is a class of objects with its own properties, morphologies and that predicts a lot of constraints as to whether or not something is possible. So it’s more than a pedagogical visualization tool.
Pauline Davies: And what about any use in, giving clues for treatment in the future?
Sui Huang: The landscape would be useful if you know everything about molecular details of the gene network. Once you have that, then the idea of landscape will be very useful. As of now, we have just a more qualitative model that stimulates new ideas, for example, one is that a lot of this chemotherapy or attempts to kill tumors actually makes it worse, so you go farther away from the normal valleys.
Pauline Davies: Well thank you very much.
Sui Huang: Thank you.

Interview with Carol MacKintosh

(Back to Audio)

Interview with Carol MacKintosh

(Back to Audio)

Carol MacKintosh: So, we don’t have a name for the ancestor, but a modern day animal that’s similar to it is called ‘amphioxus’ or ‘lancelet’ and the remarkable thing is that the ancestor, which lived in the Cambrian Oceans, underwent a dramatic genetic event when all of the DNA duplicated and then later it all happened again. And somehow, these duplications give rise to a new species of animal and that new species became the ancestor of all of us, all of the back-boned animals; the fish, the birds, mammals, including humans.
Pauline Davies: And the animal, what does it look like?
Carol MacKintosh: Well, you would recognize it as a relative, so it’s like us in terms of having a neural tube, muscle blocks, it has structures called pharyngeal arches, where we have facial structures and it has a post anal tail. But, of course it doesn’t have all of our fine features, it doesn’t have a big brain, it has no bones, it’s heartless and so, that is how it differs from the vertebrae animals
Pauline Davies: Okay and so this very primitive animal, it started doubling its chromosomes?
Carol MacKintosh: Yes, actually, doubling of chromosomes is quite common in nature, particularly well-known in plants, but it has only recently realized that such an event happened in our own ancestry and the remarkable things as far as I’m concerned is that the progeny survived, you know, despite this massive change in doubling of the DNA , and it all happened again in a later generation.
Somehow, the babies, survived and became our ancestor.
Pauline Davies: So I think you’re saying we, well the animals that followed this primitive organism that used to live in the sea, had eight copies of each chromosome, am I correct?
Carol MacKintosh: The original ancestor had a pair of each chromosome and then that doubled, so they had four, and that doubled so they had eight. But, then of course, during this process and afterwards, there were lost of losses of parts of chromosomes and individual genes, so in modern day humans, this has been incorporated so that we’re back to two, but we have copies that are scattered throughout the genome, that are what was retained from that time.
Pauline Davies: It’s really a remarkable story.
Carol MacKintosh: Well the discovery was made, I guess, in 2008 when the sequence of the modern day amphioxus that resembles the ancestor, its genome was compared with that of humans and that was when it really became established. Before that, various lines of research had indicated that this was a possibility.
Pauline Davies: And now you think that these spare copies of genes that we’ve got are somehow implicated in cancer?
Carol MacKintosh: Well, yes, so what we discovered in my research group is that the extra copies, so having two or three or four of particular genes, what this has really done is its given a real boost to the communication systems in our bodies.
So what we’ve discovered is that these copies form little teams, families. and these families operate in networks inside our cells and because of the way that they work in this team way, what this means is that cells in our bodies are able to transmit lots of different messages from the hormones that control our cells, and they do this in a way, it’s kind like Smartphones. You know, what’s special about smartphones is that they can integrate multiple messages and make sense of them and cells in our body are better at integrating multiple messages than even the smartest of smartphones.
Pauline Davies: And that’s because we’ve got the spare capacity, the duplication.
Carol MacKintosh: Yes! Having extra copies means you can pick and choose and make up different transmission routes for messages, but the cancer issue comes when, because of these sophisticated systems that were required to make the vertebrae animals, the down-side of that is when the communication systems go wrong – because of the sophistication of these systems it can go wrong – that can give rise to cancer.
Pauline Davies: And what is the purpose in the normal individual in having so many signaling pathways?
Carol MacKintosh: Well, cancer becomes reliant on signals that tell the cancer to grow and divide and proliferate. Now these same pathways are required at certain points during our development, so for example, as an embryo, its essential that these pathways that tell cells to grow and proliferate, they have to be switched on in a very controlled way and the right cells at the right time to allow the organism to grow and develop. But in the adult, for example, what happens is the cancer cells become very dependent on these pathways, whereas rest of the body only requires these pathways at certain times, for regenerating, you know, various organs and so on.
Pauline Davies: Wound healing?
Carol MacKintosh: Yes, wound healing for example. But so the cancers have a greater reliance on the signaling pathways that tell it to do that particular thing, whereas the rest of the body is operating using different signaling pathways that are telling different cells in our bodies to do different things at different times.
Pauline Davies: And I think that you are implying, or saying, that 60% of cancers had one of these families implicated, where you have multiple copies.