Cancer Metabolism: A Conversation with Jason Locasale
, by NCI RAS Initiative Staff
Editor's note: Jason Locasale is at Duke Univeristy in the Duke Cancer Institute, the Duke Molecular Physiology Institute and the Department of Pharmacology and Cancer Biology. He has made important contributions to our understanding of metabolism in cancer biology, including the identification of serine synthesis as a key pathway in cancer metabolism and the discovery of a link between metabolism and epigenetics through histone methylation and one carbon metabolism.
How did you get into cancer metabolism? Beginning in college, I was thinking I would study molecular biology and perhaps go into medicine. But very early on I became more interested in physical sciences. Mostly I was drawn to the problem-solving aspect of the courses and the ability to use reductionist approaches to understand complex problems which I found lacking in the biology curriculum. And so I transitioned from a biology to a chemistry major, but I was always drawn to the complexity of biology, and also the potential to impact health and disease so I took biochemistry and genetics courses as electives. In grad school I was focused on computational modeling of T-cell activation, the same pathways like RAS and PI3K that are aberrant in cancer. My mentor was Arup Chakraborty and he was using a lot of unconventional tools from physical chemistry, statistical physics and mathematical modeling to understand what was going on. So metabolism eventually turned out to be a good way to bring together a lot of this interdisciplinary training. So was the possible connection between growth factor signaling and metabolism which I also got into later. My undergraduate research was mostly wet lab biochemistry, and my Ph.D. was computational. In retrospect, I think the key piece of training from that time that shaped my current research was using advanced methods since most labs in chemistry and computational biology were mostly interested in technology development whereas Arup was more interested in addressing biological questions.
I started out at Berkeley but Arup moved from Berkeley to MIT and I went with him. I was working on my thesis at MIT and I was familiar with Lew Cantley's work because he was one of the major figures in growth signaling and he had discovered the PI3K pathway. So I had a bunch of conversations with Lew over a period of about a year where I was in Cambridge and Lew was in Boston in the medical campus and I would take a train over there and meet with him for half an hour or an hour, and he was bouncing a lot of ideas off me based on work that a graduate student at the time, Heather Christofk, had done. Basically she had done an unbiased screen for binders of phosphorylated tyrosine proteins and one of her major hits was an enzyme involved in glycolysis. That enzyme is known as pyruvate kinase and it catalyzes one of the end steps involved in glucose metabolism. There was a clinical fellow in the lab, Matt Vander Heiden, and Matt and Heather were working through this study on pyruvate kinase. One of the things that really resonated with me was that outside of my conversations with Lew I knew of almost no literature on metabolism occurring downstream of growth signaling but the idea was in the back of my mind for a few years. There were certain proposals here and there I had read about when studying immunology. For instance there was there this 2-signal hypothesis in T cell activation which I later learned was proposed by Craig Thompson and others. The idea was that T cell proliferation required two signals, and the reason it required two signals was that one of the signals was to activate ERK, and initiate transcriptional programs relevant for cell proliferation through ERK, and the other signal, through CD28, was to go through the PI3 kinase pathway and do something about metabolism to potentially meet the metabolic demands during cell proliferation. There wasn't a lot of solid data at the time but it was a very attractive hypothesis I remember when I was reading about it. So several times after discussing with Lew about his research, I remember coming back to the lab at MIT one day and telling my lab members and classmates how exciting and transformative this guy Lew’s work was. This was what I wanted to work on and the next step was to figure out what sort of technologies we could bring to bear on this problem.
Is there a central dogma to cancer metabolomics? Is there a big idea, or are there a few big ideas?
I guess one way to think about it is that cancer cells use their metabolism for three major functions and there are a lot of nice reviews that have come out on this recently [Science Advances 2016, 2, e1600200, PMID: 27386546; Cell Metabolism 2016, 23, 27, PMID: 26771115]. Each of these processes is important and each of them is different in the context of cancer. One involves catabolism, the production of energy. This involves the Warburg Effect, the observation that cancer cells take in large amounts of glucose and carry out fermentation [incomplete oxidation of the glucose] even when lots of oxygen is available. Historically this was thought to happen because the mitochondria were defective but this turned out not to be true in large part. Instead it looks like the TCA cycle [tricarboxylic acid cycle in in the mitochondria that fully oxidizes the glucose for energy] appears to be used differently. As one example an enzyme pyruvate carboxylase that feeds the TCA cycle which is usually thought to function mostly in the liver is highly active in cancers including RAS-driven lung adenocarcinomas. So together this is called central carbon metabolism and it's a major area of drug development in cancer. Another function involves anabolism, or the production of biomaterials that support cancer cell maintenance and growth – nucleotide, lipid and amino acid biosynthesis are all altered in cancer and some of the most effective chemotherapies we know of target anabolic metabolism with a surprisingly level of specificity. And the third one is that there are essential signaling functions for cancer metabolism and cellular metabolism in general. These signaling functions are there to communicate nutritional states and metabolic states to other aspects of the cellular machinery. One of the messengers here are the reactive oxygen species which can oxidize cysteine residues to change protein conformations and activities. Another area that we’re really excited about is the link between metabolism and chromatin status. Some of the pathways we’ve discovered to be important for cancer metabolism, like serine, methionine and one carbon metabolism, directly influence chromatin status by supplying the methyl and acetyl groups for histones which can have pervasive effects on gene activity [Cell Metabolism 2015, 22, 861, PMID: 26411344; Cancer Metabolism 2015, 3, 10, PMID: 26401273].
Can you make some general observations about the redox states of metabolites?
There is a general flow of reducing power from catabolic states to anabolic states, redox is how you get from one point to another since it performs the chemistry. The redox status in metabolism we’re referring to is about the carbon. A static, energy rich metabolic state like the starch in a piece of corn is very reduced as are the end products of anabolic metabolism, nucleic acids, cell membranes, protein. Redox involves letting go of and adding on the electrons in order to do the chemistry needed to convert sugar to biomass. For example fully oxidized carbon that's highly present in mammals is CO2 which is made in the TCA cycle, that's the most oxidized you can get while the food you eat is in the most reduced form. And everything you can accomplish near that center of metabolism is to add electrons whether you're going back to make glucose or whether you're making nucleic acids or proteins. There are layers of biological complexity associated with how a cell senses the redox state, and how that is communicated to all the rest of the cellular machinery, even things that affect the activity of RAS. That's the general idea, we think about this redox state in terms of "where is metabolism going". We’ve done some mathematical analysis to try to clarify this a bit in the context of the Warburg Effect [Biophysical Journal 2016, 111, 1088, PMID: 27602736].
What are the prospects for affecting metabolism in cancer therapy?
The really interesting, and difficult, aspect of metabolism and its potential for therapy, is that it's much harder to define the context in which a specific metabolic therapy is likely to be efficacious. When you have a mutated oncogene you have a very well defined biomarker, and the success of the therapy is predicated on the existence of that mutation. In the case of cancer metabolism what we're finding is that there are context dependencies and cancer metabolism is very heterogenous, but they are much harder to define. They're not simply coded by the presence of a single mutation and it’s a complex convolution between both genetic and environmental factors such as tissue of origin, tumor microenvironment and host metabolism that is determined by both genetics and environment such as diet and exercise. For instance, Matt Vander Heiden's recent paper [Science 2016, 353, 1161, PMID: 27609895] where he looked at RAS cancer in the pancreas as well as RAS cancer in the lung, basically showed that you get entirely different metabolic pathways and requirements, depending in that case on the tissue of origin. But other studies like Maria Yuneva’s [Cell Metabolism 2012, 15, 157, PMID: 22326218] have shown that different oncogene-driven tumors from the same tissue environment in the same mouse strain eating the same diet can create different metabolic dependencies which contrasts Matt’s findings. So it’s turning out to be very complex. We've been working on a study where we think we have a drug that selectively targets cells with the Warburg Effect, and there's a huge spectrum of responses in both culture and in vivo that ultimately can be predicted by measuring the extent of the Warburg Effect. But here there isn't a simple mapping between a specific genetic event and a metabolic dependency even in cell culture, and in this case the biomarkers for the drug may not be contained in the DNA but seem be encoded in the metabolomics. So, we know that RAS can alter metabolism, we know that every other oncogene and tumor suppressor gene that we study also alters metabolism, as does almost every environmental factor such as diet or in cell culture, media composition, and all this together seems to create liabilities. But figuring out whether there is a defined mechanism with an input and a readout that actually predicts what the dependency is, that's a major challenge for the cancer metabolism field and it’s proving to be way more complex than looking at one or two oncogenic events such as RAS and p53.
There are also two extremes in cancer metabolism. We know that targeting metabolism is a successful strategy, since some of the common cancer chemotherapies actually target metabolism. Pemetrexed in lung cancer, 5-FU in colon cancer, gemcitabine in pancreatic cancer, these are all anti-metabolite agents. Historically these agents have been thought to be non-specific and act non-selectively against cancer by killing all dividing cells. But we know these agents can often kill tumor cells in ways that are selective over other normal proliferating cells in the body. The unfortunate part however is that these therapies are still very toxic and they're not always efficacious as well. So a major focus of our lab is understanding the context in which some of these therapies might be selective, with the underlying hypothesis that there is some science about the metabolism that is behind how these antimetabolite agents work. Essentially we're trying to make antimetabolite therapy more precise and especially avoiding it when it’s not likely to be effective [Cell Reports 2016, 15, 2367, PMID: 27264180].
On the other end there are milder more tolerable interventions that affect metabolism and appear to have some anti-cancer effects. There's all kinds of literature on how life style limits breast cancer recurrence, and there's tons of literature on metformin, a drug that targets central carbon metabolism and is widely used to treat type II diabetes, how it prevents and in some cases possibly treats cancer. But we have no idea which patients we should give it to so it's back to the question of the biomarkers and understanding the context in which these mild therapies might be useful. But this is something we’re trying to understand [Cell Metabolism 2016, 24, 728, PMID: 27746051]. It's the inherent challenge in the translational aspect of this science that we know that there's a lot of heterogeneity in metabolism in how these milder interventions might work in a minority of situations.
Is there someplace in the middle, between the very mild interventions and the highly toxic interventions, where metabolomics might lead to either smarter chemotherapy or something intermediate which is less toxic but more efficacious?
There is no question it exists. Whether we have the right models and the right tools to understand it is our current problem. For example, methotrexate, which is a commonly used chemotherapy at high doses in certain blood cancers and is a fairly decent drug, but is also incredibly toxic, methotrexate is also used at far lower doses for the management of other diseases such as rheumatoid arthritis. The doses that are used to manage arthritis are doses where people can go to work and have a reasonably active day. If there were certain cancers where you could take that low dose and manage your cancer, that would be a profound accomplishment in cancer therapy. Whether it's methotrexate or whether it's the drugs that we and others groups are developing, like those that target the Warburg effect, what we're seeing is that there's a whole spectrum of concentrations required to achieve anti-tumor efficacy which gets even more complicated when we think about what administrations have the needed tolerability.
What are your principal wet lab tools? Do you do a lot of stable isotope studies?
Usually what we do involves mass spectrometry. We developed some methods to profile a lot of metabolism at once, and then when isotope tracing is coupled to that profiling we can measure how things are flowing in metabolic pathways. That's integrated with the computational modeling that we do as well as some model systems that we use to probe these pathways.
Can any of this be done in humans?
Yes, in fact we've actually published some work on this. We have a paper recently out [Cell Metabolism 2016, 24, 728, PMID: 27746051], a collaboration with a group of physician scientists at University of Chicago led by Ernst Lengyel, and the basic idea is that we took patients who were taking metformin right before their surgery for ovarian cancer, and we could immediately profile some of the metabolomics of the tissues and infer fluxes from them. The other aspect is that humans can actually take isotope-labeled nutrients. Ralph DeBerardinis, and also Theresa Fan and Andrew Lane at University of Kentucky, showed that if patients drank something on the order of 2 cans of Coke worth of stable-isotope labeled C13 glucose before their surgery, you could actually label the tumor to levels that were informative of flux throughout the tumor [Cell 2016,164, 681, PMID: 26853473; J Clin Invest 2015, 125, 687, PMID: 25607840].
Only with PET [positron-emission tomography, done with 18F-fluorodeoxyglucose] positive tumors though?
That's another complicated aspect of this field because the PET signal is associated both with the tumor's intrinsic metabolism and also with the tumor microenvironment. Pancreatic tumors tend to be PET negative likely because the tracer just simply doesn't get into the tumor cells. One thing we’ve learned as a field over the recent years is that there is a lot more to metabolism than a positive or negative signal from a PET image.
What do you tell your students about their future careers, and how studying metabolism of cancer cells is going to lead to productive employment and helping cancer patients in the next ten years?
We're fortunate to be in a very active field, and as a result I think there are plenty of career opportunities. I’ve found that many of the students I’ve mentored, and I was the same way when I was younger, tend to place too much emphasis on the perceived competition from others. But the advice I try to give is that the real competition comes from those that think what you’re working on is not important, so it’s really important to be collegial and friendly to those working in close areas and to support one another which ends up growing the field as a whole and giving more opportunities for everyone. Another thing I usually emphasize in my lab is that there are a lot of technologies and expertise we use which are less common across the broader community which places them in further demand, like for instance some of the mass spectrometry and the computational biology we do, so it’s important to master these at the expense of outsourcing other more common techniques. But ultimately, something that I learned from Lew is that the way to advance your career and help others is to take the path of least resistance. In my case, of course it didn’t hurt that metabolism was an emerging area in biomedical research when I was looking for jobs as a postdoc but the fact that the field provided a union of many of the things I was passionate about is what ultimately led to a satisfying career.
To impact patients, the thing I emphasize is to do your best to understand how cancer therapy works practically in addition to how it theoretically works in a publication in research settings – talk to clinicians and try to learn what they are facing, attend the grand rounds at the cancer center, read the leading clinical journals to see what is considered the latest research, pay attention to what the new drugs and therapies are and changes in practice and healthcare that are coming out. From there keep those perspectives in mind as you develop your own research questions. This doesn't necessarily mean that you now have to add an immune therapy or a CDK inhibitor to your study but I think having these experiences informs how you think about your basic science in ways that aren’t immediately foreseeable. One example is our recent metformin paper where the paper can stand alone as basic science investigating mitochondrial biology, but since we used a clinically relevant drug as the probe to understand the biology, the study now has a chance to also have an impact on patients. Basic and translational science are not necessarily in contrast to each other. There’s a lot of basic science to learn by focusing on the important clinical questions which is something the RAS signaling field has seen in recent years but I expect the same to be true for metabolomics.
You have published on KRAS cancers. Do you have anything in particular to say about how oncogenic KRAS in pancreatic or lung or colon cancers, which is where most of the grief comes, most of the human tragedy comes, what's happening there in terms of metabolism?
I think there's still a lot of basic science left to address. Five or ten years ago cancer metabolism was generally thought to be a function of a single oncogenic event and manifested in the observation of a single parameter such as a PET signal. Now we're learning that the interaction between the oncogene and the environment in which the tumor establishes can be just as influential in determining the metabolic requirements of that tumor. So I think where the future is going to lie is mapping out the interaction between the oncogene, say KRAS or NRAS, and the metabolic processes that the oncogene influences, and the environment of the host tumor such as the nutritional status. One of the common questions cancer patients ask their docs is what can I change in my lifestyle to help, like diet or exercise. What I think can immediately make the most impact in studying cancer metabolism is that we will have better scientific explanations to these questions very shortly. Imagine if we defined diets and exercise regimens for cancer patients in ways that had precise science surrounding them based on metabolomics approaches, that would be so profound.