Nutritional Requirements of Pancreatic Cancers - A Conversation with Dafna Bar-Sagi
, by NCI Staff
Editor's Note: Dafna Bar-Sagi is Professor in the Departments of Biochemistry & Molecular Pharmacology and Medicine, and Senior Vice President and Vice Dean for Science at New York University’s School of Medicine. In 1986 Dr. Bar-Sagi and her postdoc advisor, Dr. James Feramisco, observed that injection of oncogenic KRAS protein into fibroblasts caused morphological changes, including pronounced membrane ruffling and the accumulation of pinocytic vesicles. Recent investigations into the metabolism of cancers, especially cancers driven by mutant KRAS, have transformed macropinocytosis from a phenotypic curiosity to a key player in the nutrition of pancreatic cancer cells. RAS Central editors recently interviewed Dr. Bar-Sagi about her science and her career.
Can you describe where you were, what you were doing, and what your position was when you first observed macropinocytosis induced by RAS protein? I was a fairly new postdoc in Jim Feramisco's lab at Cold Spring Harbor. It was a cell biology lab that had essentially transformed itself into a cancer biology group with all of the excitement about the discovery of oncogenes, particularly RAS. We wanted to understand what RAS was doing—how the transfection of a piece of DNA encoding RAS into NIH3T3 cells caused them to become transformed. The challenge we faced was that it took up to 3 weeks from the time NIH3T3 cells were transfected with a mutant RAS gene to when we saw them form foci. Lots of other things might have happened over that time, many of which were probably secondary or tertiary consequences of having RAS in the cells.
I was drawn to Jim’s lab because he was an excellent biochemist who could purify proteins and then introduce them into cells by microinjection—a technique that was then in the domain of only very few people in the scientific community. We reasoned that microinjection of RAS protein would be a more effective approach to understand the events most proximal to RAS itself.
We discovered that microinjection of RAS caused two phenotypes. First, when microinjected into PC12 cells, rather than transforming them, RAS induced their differentiation into neuronal cells. This was a big surprise: as an oncogene, RAS was supposed to promote proliferation. Instead, it induced cell cycle exit and a differentiated phenotype. Second, when we microinjected RAS protein into normal fibroblasts, the very first phenotype that we saw (maybe after 5 minutes) was intense membrane ruffling activity and the accumulation of phase-bright vesicles, which we now know are macropinosomes. These different responses to RAS were an early indication of the importance of context—in this case cellular background—as a guiding principle to understand the role of RAS in cancer development.
Was watching the cells under the microscope always part of the plan? Or had you planned to do something else if you hadn't seen anything happen in the first five minutes or in the first hour?
Watching the cells was the intent of most of the experiments simply because the technologies for doing any kind of other analyses on small numbers of cells were very limited. But sometimes, being limited can be an advantage. Not being able to use many sophisticated methods that generate a lot of data forced us to focus on this single observation.
I remember something else when I look at the 1986 paper we published. The first submission of the paper, to a top-tier journal, did not go well because the editor asked me whether this would explain how RAS causes cancer. My answer was: “At this point, to be quite honest with you, I don't know”. The editor thanked us and suggested that we resubmit when we had the answer to this pressing question. So we became smarter in the next submission, which was to Science. The final paragraph of the Science paper speculates about the potential significance of macropinocytosis for a cancer cell. There, Jim Feramisco and I invoked a couple of potential mechanisms, including nutrient delivery.
So in essence I was frustrated that an editor forced me to provide evidence for the significance of this finding for cancer. But in retrospect, it was an important lesson for me, and one that I pass on to my trainees. I always challenge them to ask questions and design experiments that will change fundamentally how we think about an important problem.
Development of resistance to targeted cancer therapies seems to be nearly universal. Is resistant to therapies that target metabolic vulnerabilities less likely?
It depends on the type of therapy. But irrespective of the specifics, it is quite clear that we are not going to be able to rely on a single magic bullet, and that the successful treatments of RAS cancers will involve the simultaneous targeting of multiple vulnerabilities. To design effective therapeutic strategies, we will need to better understand context-specific features of RAS cancers. For example, epithelial cells that give rise to pancreatic cancer differ from epithelial cells that give rise to colon cancer. They are wired to perform distinct physiological functions, and therefore the acquisition of a RAS mutation would likely have different consequences. Also, the time course for cancer development is probably important. Unfortunately we don't have a good way of studying it, because what we do experimentally is somewhat artificial. We drive the process of cancer development very quickly, and we end up with a situation where the cells adapt to oncogenic changes in a way that might be very different from what happens when the changes occur very gradually. And finally, we need to think about the entire organism. Although we have to start with the tumor-centric view, considering the physiology of the entire organism will play a critical role in designing effective cancer therapies.
You authored a recent paper with several of your cancer metabolism colleagues showing that RAS-transformed pancreatic cells depend on macropinocytosis to import proteins from their environment. Is the use of albumin and other proteins from the environment maintained as pancreatic tumors become larger and metastatic?
We are only starting to look at this question, but I predict that it is going to be present in the metastatic site perhaps even at a higher level than in the primary site. The reason is that the cells in the metastatic site tend to be much more mesenchymal in nature, and macropinocytosis increases as the cells become more mesenchymal.
We really need a way to score macropinocytosis in vivo so that we can understand the extent to which tumors in people rely on this process. Our invasive approach to address this question in animals doesn't lend itself to any follow-up studies, because it's usually an end-point of the experiment. Hopefully, we or others can come up with a non-invasive imaging approach or a biomarker for macropinocytosis. We also want to understand the metabolic profile of tumor cells that feed on albumin, or for that matter, any other protein that the cells can take up from their extracellular environment. Do they differ from cells that use free amino acids to meet their energetic needs? I should mention that besides the role we discovered for macropinocytosis as a source of nutrients, macropinocytosis has long been known as an effective delivery pathway for drugs that are packaged in nanoparticles. So although one would want to inhibit macropinocytosis in the context of targeting the metabolic vulnerabilities of tumor cells, there might be value in thinking about ways to stimulate macropinocytosis to improve drug uptake by tumor cells.
How did you get into science and cancer research?
I grew up in Israel. My parents, who were both Holocaust survivors, came from Europe. My father was a chemical engineer and my mother was a homemaker. My upbringing was certainly part of what drew me to science as a discipline. I remember my father being genuinely interested and completely immersed in his job. Growing up, I spent many weekends with him at his workplace. So as a child, my dream was to have a profession that would consume me in terms of intellectual interest and engagement. After high school, I served for 2 years in the army. That also influenced my career path. I was 20 when I left the army and enrolled in college. Serving in the army made me more mature than your average 20-year-old. I felt that I did not have much time to kill, and wanted to settle on something that I'm would pursue for the rest of my life.
I studied neurobiology during my undergrad and masters in Israel, and then came to the US to pursue my PhD at Stony Brook—still in neurobiology. I credit much of how I conduct my research to my undergraduate and graduate education. I was privileged to be mentored by individuals who took the time to train me how to pose interesting questions and design experiments that would yield meaningful answers.
How did you get from Stony Brook to Cold Spring Harbor?
When I was looking for postdoc opportunities, I hoped to do something different from what I did in my PhD. But I also searched for exciting but underexplored areas. Oncogenes were just making it onto the scene; we knew very little about them and there were very few tools to study them. That's what led me to join Jim’s lab; his use of microinjection to study oncogenes was unique.
How do you advise your grad students and postdocs about their employment prospects? Do you give them any advice on how they can make themselves stand out when it's time to leave your lab?
Taking a job in academia is not necessarily the only way to capitalize on the skills acquired during one’s pre-doc or post-doc training. Graduate students and postdocs need to realize that they have many other options after they finish their training. Based on my experience, I always tell trainees that a research position in academia can be extremely gratifying. But one has to be really passionate about it. Besides running my lab, I have a very large administrative job as the vice dean for science at NYU Langone Medical Center. But no matter what happens during the day in my administrative life, the last thing I think about before I go to sleep is an intriguing set of data I saw during the day. And the first thing that I think about when I get up in the morning is what's going to happen in the lab today! This is what I mean when I say that research has to be driven by passion.
The beauty of scientific research is that there are many, many different ways in which it can be accomplished. That is why it is important to attract to the field people who are very different from each other and who contribute diverse skill sets. There is no gestalt - no particular route - to becoming successful. The fun and the challenge when you train individuals is to figure out the secret sauce that can make them thrive. For each individual there is a different recipe.