Cellular Differentiation and Response to Stress: Modeling Cancer Initiation and Progression – August 29 to September 1 2010, Sedona
Cellular Differentiation and Response to Stress: Modeling Cancer Initiation and Progression
The Sedona Workshop was held under the auspices of the Arizona State University Physical Sciences Oncology Center, and was the third in the series of focused workshops on different aspects of cancer. The core focus of this workshop was the differential response of tumor cells to environmental stress. This central issue was approached by three groups: i) experts in the cancer field, ranging from cell biologists to bioinformaticists to pathologists, ii) biologists and bioengineers with expertise in cell differentiation in the context of stem cells and developmental biology, and iii) a cadre of biological modelers, with backgrounds in physics, engineering, and mathematics.
Listen to Audio Interviews and Read Transcripts
The meeting kicked off with a reception on Sunday evening and a plenary lecture by Paul Davies, the Director of the ASU PSOC. Dr Davies’ lecture was provocatively entitled “Where is the epigenome?” and examined the fundamental issues of reductionism and causation in complex biological systems, especially cancer. This lecture set the tone of the workshop, namely to emphasize fundamental issues and question received wisdom in the hope of uncovering rich new directions of research.
The workshop proceeded through Monday and Tuesday, and was partitioned into four sessions, each with a different theme. The first was concerned with heterogeneity at different scales. Hal Berman presented a pathologist’s view, with a long series of histological images of different breast cancer samples, emphasizing the immense diversity of cancer phenotypes, and heterogeneity in structure at the scale of hundreds of microns – which represents the microenvironment of cancer cells. Beverly Emerson then examined the intracellular processes of transcription regulation, and how aberrant cells may use alternative regulatory pathways for transcription to survive the harsh environment created by chemo-therapeutics. This session was wrapped up by presentations by Rod Smallwood and Sal Torquato, presenting contrasting views of complex modeling of cancer. Dr Smallwood presented the “epitheliome project” in which computer scientists use cutting-edge algorithm design to connect multiple scales, from gene pathways, to cells, to tissues. Dr Torquato presented the complementary approach of phenomenological model building, starting with simple assumptions and building complexity step by step.
The second session had microevolution as its primary theme. Michael Barrett presented his front-line work on uncovering genomic signatures of cancer patients, and how these signatures change over time in individual patients, and how these signatures differ from one patient to another. His talk emphasized the extraordinary genomic diversity of cancer, along with the remarkable genomic technologies that are available to uncover this information. Yosef Yardin introduced the idea that evolution has created robustness at the scale of gene networks rather than individual genes, and presented work on the gene pathways connected to the EGF receptor to illustrate his hypothesis. Sandy Anderson presented a multiscale modeling perspective on cancer dynamics, in particular the metaphor of ecosystem dynamics as a way of understanding the maintenance of diversity in tumors and their environments. Jeff Trent presented a very different perspective on heterogeneity and evolvability of cancer – that of the last 12 weeks of cancer patients: a period in which patients opt to take part in phase I trials with promising new drugs and therapies, and the increased benefit such trials provide.
The third session focused on plasticity of cancer cells, and lessons on cell differentiation which could be taken from studies of embryonic systems. David Shaffer presented a bioengineering perspective on ex vivo measurements and reprogramming of stem cells through creating artificial biochemical and biomechanical microenvironments. Timothy Newman presented a general perspective of modeling multicellular systems, contrasting modeling approaches to embryo development and cancer. Paul Kulesa discussed his fascinating experiments on introducing melanoma cells into the neural crest of developing avian embryos, and the resulting response of the cancer cells which closely follow the normal neural crest migration phenotype. Luis Cisneros shifted the discussion to collective behavior in prokaryotic cells, namely swimming bacteria. Movies of coordinated pattern formation in highly dense populations of bacteria showed streaming patterns reminiscent of eurkaryotic cell cultures. The dependence of the coordinated cell motion on gradients of oxygen provided an interesting new perspective on nutrient competition within tumors.
The final session of the workshop was concerned with the microenvironment of the tumor. Charles Little presented his striking results on the highly dynamic nature of the extracellular matrix in avian embryos. In particular, the matrix was imaged using fluorescent markers, and observed to undergo coordinated movement patterns almost identical to those produced by the embryonic epithelial-like cells. This observation brings to the fore the complex coupling which exists between cells and their surrounding matrix, both in structural and dynamical contexts. Nastaran Kuhn provided a bioengineering perspective on how the mechanical properties of the cell’s microenvironment can prompt the cell to radically change its behavior. For example, cells on stiffer substrates were observed to be more susceptible to apoptosis following stress. This session, and indeed the meeting, was brought to a close by Thea Tlsty, who provided a wide-ranging view of the phenotypic plasticity of tumor cells in response to various signals from the microenvironment, and how cell phenotype can change markedly during tumor progression. The fascinating mediating role of certain stromal cells – named cancer associated fibroblasts – was discussed in detail.
A key concept that was reinforced from the discussions in this meeting is that cancer must be viewed as a set of diseases of cells within a variety of microenvironments, rather than as a set of key mutations in individual or isolated cancer cells. The heterogeneity of the disease was emphasized with compelling experimental data at the genomic scale, the cellular scale, the cell-matrix scale, and the tissue scale. This provided the physical scientists present with a stark sense of the challenges facing them as they attempt quantitative measurements and modeling of cancer, and emphasized the necessity to forge intimate collaborations with cancer biologists.
The organizers would like to thank all the participants for traveling to Sedona at a busy time in the academic year, and for taking part in a successful and highly interactive workshop. The organizers would also like to thank Paul Davies for supporting this workshop, both with funding and keen intellectual insights. Finally, the organizers would like to thank the PSOC coordinator Vanessa Baack, and her team, for their tireless efforts in ensuring the smooth running of this workshop – it simply would not have taken place without them.
Timothy Newman, Professor of Physics, (http://biodyn.physics.asu.edu)
Audio Interviews from the Workshop
- Interview with Professor Salvatore Torquato. Professor of Chemistry, and Princeton Institute for the Science and Technology of Materials. Salvatore discusses how computer models can simulate the growth of tumors. The models are (of course) informed by biology but now the best models are suggesting new pathways for cancer biologists to explore.
More Info: http://cherrypit.princeton.edu/sal.html
- Interview with Professor Beverly Emerson. Regulatory Biology Laboratory, The Salk Institute. Beverly Emerson discusses the epigenetic control of cancer and how a cell’s response to stress is regulated. More Info: http://biology.ucsd.edu/faculty/emerson.html
- Interview with Professor David V. Schaffer. Chemical and Biological Engineering, Bioengineering and Neuroscience and co-director of the Berkeley Stem Cell Center. David explores how neuronal stem cells in the brain can be manipulated into becoming differentiated cell types useful for the treatment of Alzheimer’s and Parkinson’s diseases. He speculates that analogous manipulations can be applied to cancer cells. More Info: http://cheme.berkeley.edu/faculty/schaffer/index.php
Audio Interview Transcripts
Interview with Salvatore Torquato
Salvatore Torquato I’m Salvatore Torquato from Princeton University.
Pauline Davies Salvatore, you’re working on a model of cancer, tell me a little bit about that?
Salvatore Torquato Well it’s trying to develop computationally the basic features of cancer growth. Starting from the incipient tumor cell all the way up to very macroscopic sizes.
Pauline Davies What is special about your model?
Salvatore Torquato Well it’s a minimalist model in the sense that it incorporates just a few parameters and it’s able to capture the basic features that we see in real tumor growth.
Pauline Davies We were looking at a model that Tim Newman was describing, so there are lots of modelers out there, how should we distinguish between different peoples models, what makes them worthwhile?
Salvatore Torquato That’s a great question. It depends on the focus of the particular research. So for example in Tim’s model he was specifically considering how cells actually respond to stress. So that’s a very focused problem, on the other hand you can be interested in what happens across larger length scales and so that will involve many more cells. So it depends really on the specific question that one asks and depending upon what that question is I guess you can do a better job than certain more difficult questions. So it really is a choice in some sense, since cancer is an extremely complicated disease that we know very little about, it’s a medical mystery, we’re really in the infancy of trying to model cancer but it definitely appears that it will be a fruitful approach because clearly the medical community has been at a loss to be able to really understand the mysteries of cancer. On the other hand the physics based approach the more reductionist type of an approach could add to the mix something that people never considered before. So I think that a physics-based modeling approach to understanding cancer will be extremely profitable in the future, although we’re really in the infancy stage.
Pauline Davies Are we talking about decades away, do you think, before a model can actually incorporate all of the relevant physical parameters?
Salvatore Torquato That’s a great question. I think the only way that this can really happen, for example from a modeling point of view, which is what I do, the theoretical computational aspect, you really need to have constant feedback from experiments. So without a model being informed continually by experiments a model will never be successful. So to the extent that one is linked with cell biologists, oncologists, clinicians, that will really determine whether or not we are going to be very successful. So, that’s not actually an easy task right because you have different communities, you have different languages, people speak and have a certain vocabulary and you have to be able to go beyond those barriers and that takes quite a bit of energy, so this can only happen if its really truly interdisciplinary and that’s what I’m hopeful will happen.
Pauline Davies Will we have enough computing power to actually really model what’s going on?
Salvatore Torquato With increasing computing power, the computing power is constantly increasing, certainly that’s going to be less of an issue. I think really right now the bottleneck is the fact that we really don’t understand the mechanisms of cancer. The biologists don’t understand the mechanisms of cancer truly. They might understand very very specialized mechanisms but clearly we do not understand all of the mechanisms and that I think is going to be more of the bottleneck than it is the computational power that we have. So it’s certainly decades away I would say because we need to really do the experiments, but by the way I would add that the computational modeling suggests new experiments and will suggest perhaps experiments that will enable us to find, discover new mechanisms.
Pauline Davies Yes I was going to actually ask you that because it seems at the moment that you’re following the biologists, but will there be a time when the biologists actually follow what you say?
Salvatore Torquato Absolutely. I think there’s absolutely no question about it because with a computational model, the beauty of it is that you can actually conduct various scenarios that for example would be difficult to do in the laboratory environment. So by having that freedom that could suggest interesting new experiments, no question.
Pauline Davies Has that happened yet?
Salvatore Torquato I believe that that has happened, for example for some of the work that I have done, the group at Princeton that is doing experimental work will be trying to in fact validate some of the work that we have obtained, some of the results that we’ve obtained.
Pauline Davies Such as?
Salvatore Torquato Well for example we have this cellular automaton model that is able to take into account how tumor cells respond to chemotherapy and there are various scenarios that arose from that that can now be easily tested experimentally, which have not been done so far.
Pauline Davies Well that must be very exciting for you?
Salvatore Torquato Yeah it is. It is quite exciting. Well one of the reasons why it’s particularly exciting is that if by doing this we can extend a patience life, especially with the most malignant cancers, by 50% some figure like that, that’s extremely rewarding, from just the humanitarian point of view.
Pauline Davies Is it a great strain for you to keep up with the biology as well as doing the physical modeling?
Salvatore Torquato Great question. I tend to actually, not to be worried too much about that because I believe that the model is sophisticated enough, in fact to be able to take into account, we know some of the basic mechanisms and its not like we don’t know some of the basic mechanisms. It doesn’t intimidate me as much not to know the biology. Of course one has to become steeped in the biology at some point. I myself have not been working on this full time but with this new center that’s going to change. So I probably will be immersing myself more in the biology but it doesn’t intimidate me that I don’t know it as well as I should.
Pauline Davies At the start of your lecture, I loved the questions you posed, the unknown questions that biologists should be asking.
Salvatore Torquato Right. Yeah some of those questions for example, why does a whale, which has many more cells than the human being, still a mammal, have a much lower incidence of cancer than humans. We don’t understand that, why is it that reptiles and amphibians rarely get cancer an its especially true of amphibians, we really don’t know the answer to that.
Pauline Davies Well I think it’s to do with infection.
Salvatore Torquato It could very well have to do with infection. These are I think basic questions that we should really try to answer and I think that from a physicist point of view these are intriguing questions. Why not get a group of biologists and physicists and so on to try to understand these very interesting questions because if you could understand why these other species have a lower incidence of cancer then you might be able to understand better what the underlying mechanisms are. So for example, can we actually turn a cancer, this to me is a very intriguing question, can we understand it well enough to be able to make cancer actually do something that it normally does not do. So to ask that question requires a really sophisticated understanding of what is going on in cancer. So if you could target behavior for particular cancer to make it do something that it normally doesn’t do, if you can get to that point that is really an amazing result and those are the kinds of questions that intrigue me.
Interview with Beverly Emerson
Beverly Emerson I’m Beverly Emerson. Professor at The Salk Institute and I study transcriptional regulation, gene regulation and chromatin structure and epigenetics.
Pauline Davies What is the excitement of epigenetics for you?
Beverly Emerson The excitement is that states of cellular programming to me transcription, gene regulation are reversible and so that means you can go from a normal cell to a cancer cell and back or you can go from a stem cell to a differentiated cell and presumably back so its all reversible they’re all insomatic processes and we can hopefully manipulate these insomatic processes. They’re insomatic and they’re targeted so you can take enzyme and target them to specific genes and cause an event to happen at those genes or block that targeting. So it’s the reversibility.
Pauline Davies This is all pre-cancer?
Beverly Emerson Yes the excitement is in pre-cancers, early cancers because by the time you have a full-blown cancer you’ve got a genetic mess and its best for the cell just to die.
Pauline Davies Is it true that most pre-cancers would never actually turn into cancer, there’s a bit of coming and going?
Beverly Emerson I would say all of us are riddled with pre-cancers but we have lots of protective mechanisms that prevent them from developing into full-blown malignancies. So that’s why one’s balance is very important.
Pauline Davies Your lecture was related to stress and cancer cells?
Beverly Emerson That’s right. So there were three fundamental points of my talk. The first was how are genes abhorrently silent, so this is a very common feature in almost all cancer, where tumor suppressor genes are just turned off and its associated with DNA methylation, suddenly they become methylated and they’re no longer expressed. So this had been known for a long time and we wanted to understand the mechanism of it and the mechanism turns out to be very simple. All of these structural changes take place in the gene, which silences the gene due to the fact that the neighborhood of the gene has changed because the boarders surrounding the gene have been lifted. So you lose your fence posts and that means that the neighborhood around your genes which, are usually bad neighborhoods, meaning condensed chromatin, just spread right through, so the condensation spreads through, condenses the formerly active gene and silences it, its very simple.
Pauline Davies That’s bad news?
Beverly Emerson That is bad news. So you lose fences, the boarders if you will, because you lose the binding of a protein. You lose the binding of the protein not because anything is wrong with the protein, its not mutated or anything its still working on its other target genes in the cell, but it no longer functions as a boundary because its lost a certain post translational modification and so now we’re trying to determine called ADT ribosylation, that’s the modification, it’s a chemical modification of the protein that allows it to act as a fence post.
Pauline Davies So that’s how cancer might progress?
Beverly Emerson That’s how it might initiate. So and very early event in genomic instability leading to a pre-cancer is the loss, specific loss of certain tumor suppressor genes, their functions.
Pauline Davies Now can anything be done to actually reverse that process?
Beverly Emerson Well currently people use DNA demethylating agents. So they reason that okay the gene is turned off because it’s methylated, so I’ll use this compound that removes DNA methylation, that does reactivate silenced genes. The problem is it demethylates everything else in your genome and that’s not good and even though it has reactivated the silenced gene it will be remethylated because nothing has fixed the primary problem, which is the loss of the fence allowing the bad neighborhood in. So what we want to do is to find why the fence post has lost this modification and then try to screen for compounds that will restore the modification, I’m sure its due to a signaling event, some problem in a signaling pathway. If we can restore the function to the protein, its called CTCF, then we can possibly rebuild the fence so that then when you demethylate, if you have to maybe you don’t have to, then that will result in a stable reactivation of silenced genes.
Pauline Davies Is this one of the main initiators of cancer?
Beverly Emerson I have no idea but I will say that it’s an extremely common phenomenon; epigenetic silencing. That’s the reason why epigenetics, you hear that word so much, that’s why its important in cancer its based on changes in DNA methylation patterns and then later its become, it has incorporated changes in histone modification patterns, that’s what epigenetics is.
Pauline Davies Okay, so you’re picking up some ideas about what’s going on and how to possibly change things. Do you think eventually that the work you’re doing will have some clinical significance?
Beverly Emerson I really hope so because I think that especially if you can reverse pre-cancers that’s a much easier thing to do that killing cancer cells. We actually work at both ends of the spectrum. So the pre-cancer to us means the epigenetic silencing of these tumor suppressor genes which is due to the loss of the fence post, the other end of the spectrum is our work with p-53 and there you want to restore p-53 function because you want to actually kill, if you’ve got a full blown tumor cell you want to kill it. If you’ve got a precancerous cell you want to erase this abhorrent epigenetic phenomenon.
Pauline Davies Okay, is there anything else you want to mention?
Beverly Emerson Only that one part of my talk dealt with the fact that you could kill RNA polymerase, which is the enzyme that reads all of your genes, that you can poison it and that causes apoptosis but interestingly p-53 regulated genes are still on so they are immune to being inactivated by stomping on the major enzyme that reads genes and so we were interested in understanding how are these genes on when you’ve just killed the engine and in fact you can still use the engine but now doesn’t need the same co-factors as or machinery that it did need for the vast majority of cellular genes. So it tells us that there is a stealth enzymatic pathway in cancer cells that can be utilized by genes under extreme duress. So while this is good for the p-53 pathway because that’s what protects us, it can also be used by prometastatic genes and so that might be something that one might target.
Interview with David Schaffer
David Schaffer I’m David Schaffer a professor in chemical and biological engineering, bioengineering, and neuroscience at Berkeley. I’m also the Co-Director of the Berkeley Stem Cell Center.
Pauline Davies You were talking today about stem cells. Tell me what you were saying?
David Schaffer Sure, so the entire field is extremely interested in the question of how stem cells are controlled, both inside the body as well as potentially outside the body so that we could try to harness them or tap into them to cure disease. So it’s pretty clear that stem cells are regulated by signals from outside, signals that end up communicating information to them to tell them what to do. We’re extremely interested in formalizing this idea to understand how a niche that surrounds the stem cell impacts its behavior.
Pauline Davies In fact I think you were saying that stem cells left to their own devices would not be stem cells. You really need the surrounding material to support them.
David Schaffer Yes. If you take a stem cell or pluck it out of its tissue inside the adult body or at various time during development and simply put it on a plate without the right support, its going to die. So it needs very complex support in the form of soluble proteins as well as matrix proteins.
Pauline Davies So it’s really the system working together.
David Schaffer That’s right. So we’ve been trying to really make the argument that stemness or this property of a stem cell being able to continuously maintain itself in an immature state or differentiate into a specialized state is a collective property not just to the cell but its matrix, its surrounding, its niche.
Pauline Davies I seem to remember that you were actually concentrating on some cells from the brain.
David Schaffer Yes there is a really interesting, at least to us, population of stem cells that resides within the adult brain within a particular structure called the hippocampus, there are only a couple regions within the adult brain that have active stem cells that are continuously dividing throughout our lives. This population of stem cells continuously divides and creates greater than 10,000 brand new neurons in each one of our brains everyday and that region of the brain integrates these cells and they become part of the circuitry that enables us to remember to think and to learn. So we’re extremely interested in understanding how those cells are controlled. They have choices available to them, once the stem cell decides to differentiate, it can turn into a neuron, which it does most of the time, it can turn into other cell types such as an aster site, which is a major support cell within the hippocampus. We’re extremely interested in studying how bi-physical properties of that niche, in which that stem cell lives, can flip a fate, one direction versus another. We found for example that the stiffness of the environment the mechanical stiffness or hardness of the microenvironment can almost completely flip that stem cell from turning almost exclusively into neurons to turning almost exclusively into aster cites.
Pauline Davies Well that’s fascinating. So you can actually play with the system and change the cells yourself?
David Schaffer Exactly. So we had done most of this work so far in cell culture where we have very tight control over the changing the physical properties of a niche or a microenvironment in which the cells are living, but the next step of course is to really prove that it happens within our own brains. Our own hippocampus has a gradient of stiffness, so the stiffness of the hippocampus varies somewhere around 5-fold from location to location and also stiffness is really important in other regions of the brain because grey matter and white matter have different stiffnesses, tumors are extremely stiff within the brain, sites of ischemic injury or stroke become stiffer so it raises the question of whether or not this mechanical property is changing, not just what happens normally in our brains on a day to day basis but can also be a contributor to disease.
Pauline Davies What sorts of implications can that have for cancer?
David Schaffer Well for cancer, and cancer is a field that I haven’t worked in as much, so I’m enjoying myself learning about it at this conference, but cancer is also a disease that involves a complex interaction between the tumor cells themselves as well as your local microenvironment and this microenvironment is not just soluble proteins that bind to the surface of these cancer cells. It’s the other neighboring cancer cells, neighboring normal cells as well as extracellular matrix that surrounds the cancer cells and makes up a tumor.
Pauline Davies Are you studying some properties more than others?
David Schaffer Well our group in particular recognizes the fact that the field has made a lot of progress in…or identifying the soluble signals that are part of the niche but in reality cell are of course enmeshed in a solid phase; neighboring cells, matrix etcetera, the things we’ve been discussing and they are very subtle but never the less very important bi-physical aspects of this neighboring niche that play a large and important role in regulating the biology of the cells. So its not just soluble liquid phase factors, there’s actually very important information that’s contained within the solid phase of that niche. So this includes the mechanical properties of the solid phase, the stiffness, as well as nanostructural organization, the various components within that solid phase, order and the presentation of bio-chemical signals to the cell as well as shape, topographical features of the matrix, how fibrous it is how porous it is.
Pauline Davies If you can change the properties of the matrix can you somehow influence the course of cancer?
David Schaffer Yes our own work has been exploring the ways in which you could change the properties of a matrix to change what a stem cell does and change how a stem cell ticks. So you could make into lets say a dopaminergic neuron to try to treat patience who have Parkinson’s and by the same idea others within this field have been trying to manipulate bi-physical properties of tumors and in that case you’re trying to make the tumor harmless or more benign by changing really key signals bi-physical signals within the matrix or within the niche that would otherwise cause these cells to behave more malignant.