You are here

BCRF Conversations: A Discussion with Drs. Daniele Gilkes and Paul Macklin

How applied mathematics and tumor biology are coming together to drive new research


Subscribe to BCRF Coversations here: 

Today's conversation proves a simple math problem: two is better than one. Actually today's conversation also tackles some really complicated math and science. Specifically, how applied mathematics and tumor biology are coming together to drive important new research in breast cancer and metastasis.

Doctor Daniele Gilkes is assistant professor of oncology. As well as an assistant professor in chemical and molecular engineering at the Johns Hopkins University School of Medicine. This is her first year as a BCRF grantee. Doctor Paul Macklin is an associate professor of intelligent systems engineering, and a member of the Melvin & Bren Simon Cancer Center at Indiana University. He's been a BCRF grantee since 2014. Together, Gilkes and Macklin are apart of a new partnership between the BCRF and the Jayne Koskinas Ted Giovanis Foundation for Health and Policy that links computational scientists with biologists.

This partnership includes three multi-institutional collaborations jointly funded to provide new insights into tumor growth, metastasis, and the mechanics of drug resistance. These ground breaking studies will have the potential to substantially advance understanding of cancer biology, and improve clinical outcomes. You can tell from the conversation, the inner play, between Dr. Gilkes and Macklin this is a true partnership. You can also tell they're doing incredibly important and hard work, and they are having fun.

"I found myself wondering and I asked both of them, "Whether a cross discipline approach like theirs just might be the future of research?" Host, Chris Riback

Read the transcript of the conversation below: 

Chris Riback: Dr. Gilkes, Dr. Macklin thanks for joining me. Let's start with an explanation of what each of you does? To really over simplify. You're the ultimate mash up of math and biology. It's not everyday we hear about a tumor biologist and applied mathematician teaming up. Dr. Gilkes lets start with you. What do you do for a living?

Dr. Gilkes: Okay, well our group mainly studies breast cancer metastasis and metastasis is the spread of cancer cells. In this case, from the breast to distant organs. We're interested in brain, bone and lung 'cause these are common sites of metastasis. The reason why we are focused of metastasis is because this is the main cause of death for breath cancer patience. So, the uniqueness of our research is that we're looking at cancer cells as they continue to grow and form a tumor. What happens is, that the interior of the tumor begins to run out of oxygen. This is because the tumor's being feed by blood vessels which carry blood and transfer oxygen to the tissue, but it can't reach, sort of, the interior of the tumor. So, the lack of oxygen then activates a chemical signal and this chemical signal has been termed hypoxia-inducible factors. So, these factors are important because the control proteins that switch-on many genes that can help the cells cope under these low oxygen conditions.

So, basically, they initiate a chain of events that allow the tumor cells to survive under these low oxygen conditions. We and others have found that this can allow cells to spread to other parts of the body. So, our recent work has been looking at survival advantage under these hypoxia conditions and showing that cells can transform from being, sort of, ridged and stationary to become more mobile and invasive. So, these are traits that are necessary for tumors to metastasize. That's really the main focus of our work.

Chris Riback: And you personally, you're an assistant professor of oncology, but also in chemical and molecular engineering. So, not only ... We'll get the applied mathematician in a moment, but how do you mash up oncology, chemical and molecular engineering?

Dr. Gilkes: I guess I have a little bit of a dirty background. So, I ... My PHD training has been in cancer biology, but prior to that I was chemical engineer, material scientist. For my post-doctoral training I begin working in really physical sciences and looking at how physical attributes of cancer, such as [inaudible 00:04:46] , can really promote tumor progression to metastasis. So, that's where that all, kind of, fits in.

Chris Riback: Okay and once you master the oncology portion you'll move onto another section of science, and just keep going I assume?

Dr. Gilkes: Let's hope we can master this problem.

Chris Riback: Yes, yes.

Dr. Gilkes: So, we can do that.

Chris Riback: Yes, understood. No, I understand and Dr. Macklin tell me please about your background and applied mathematics, and I guess thee Intelligent Systems Engineering Group, which has recently started at Indiana? Is that ... Do I understand that correctly? That it's a, kind of, recent group and if I got it wrong I'm sure you'll fix that right now.

Dr. Macklin: No, you've absolutely got that right. My background is in mathematics, both through my undergraduate and graduate degree. Although I've had a smattering of other fields like, physics, and economics and even a bit of German. Just trying to build up as a multi-disciplinarian. Here in cancer, what we have been trying to do is say, "Can we draw on expertise from fields like physics, mathematics, computer science, and data sciences?" Drawing these expertise to build models that help us understand cancer and not just understand individual cancer cells, but understand cancer as a system.

So, a big part of my work is to develop big computation frameworks that string together these different processes each to which contributes to how cancer progress. You know, just as how Daniele said, "When cancer gets to big it out ships it supply and requires androgenesis", and we know that hypoxia is a big driver in how cancer advances. So, we've developing systems that say, "Well let's start with a network of blood vessels", and we will actually simulate the way oxygen and drugs and signal factors move around in the tissue and reship in cells.

We've also been developing a big part of that system as to model how individual cells move around in tissues and re-boost different chemical and physical and other cues to change the behavior. Then, the cells themselves actually are like the ultimate rogue engineers. They themselves send out signals and re engineer their environment. They'll call to blood vessels and say, "Hey, I need new supplies. Let's build new blood vessels in a system", in a process called angiogenesis.

So, we've been developing these systems that seem like both the environment and the cells, and how they combined and how those systems interact, to form cancer. In this new Intelligent Systems Engineering Department we're actually building on that approach across many fields including cancer. To say, "These are complex intelligent systems, they adapt in some way", and their ultimate goal is to take knowledge from different fields and learn how to intervene and change that system from cancer state to a healthy state. We can't do it without great experimentalist to tell us what to do and we hope we provide a key part in that puzzle too.

Chris Riback: When you say we do this across different systems including cancer ... That was one of the areas I was, kind of, curious about. So, you have an approach and a discipline and a profession in a way you think and things that you do. You are in this current state applying that to the study of cancer. Is that for you personally ... Is that the area ... Is that the interest section where you always work or do you have great benefit of having done your work in different disciplines and right now we just happen to be talking about the work that you are doing with Dr. Gilkes and with breast cancer, and cancer generally?

Dr. Macklin: Well that's a really insightful question. I mean, part of why I went into applied math is because math is sort of the universal language of science and engineering. When I entered math and choose a problem, well over ten years ago, in cancer modeling I said, "Well this is interesting and I'm just going to do it until either I feel like I can't make an impact anymore or I just get bored and move on." Here I am 15 years later and the more I learn about cancer the more I realize there are huge problems that need to be answered. The more you learn the more I realize there is a lot more to learn.

So, I foresee myself in cancer as my main thrust for a long, long time and I'm grateful to be part of that. With that said, the cool thing about it is the same thing computational models, the same techniques can also make an impact in related areas like, tissue engineering and  cryobiology, and systems biology. All these different areas that,  immunology, can take the same tools and we can make some advances and vice versus. We can form a sort of bridge between disciplines and say, "You know, if there is an advance in cryobiology or tissue engineering sometimes we can take that insight, put it back through or model and apply it back to cancer." So, we can, kind of help facilitate this exchange of insights between different fields and bring it back to cancer. Which will always be, kind of, my driving passion.

Chris Riback: Yeah, I can feel that. I also love that line. You've now helped me next time one of my kids say, "Dad, I just don't understand math", I can reply, "You know well, it's easy. Math is the universal language of science and engineering. What don't you understand?"

Dr. Macklin: Well, thanks. No, it's funny. Until I took calculus, I wanted to go into piano. I mean, math got me passionate about the world and the way it describes things it changed.

Chris Riback: Wow. Excellent. You know, my iTunes collection isn't as good then as, maybe it might of been, but ... You'll you know. You'll certainly, we all hope, help solve this major issue.

So, what brought the two of you together? Did you come across each other years go? Have you just come together on this, for this, effort and this research project? Is there, you know, a that you guys were connected in your studies? How did you guys come together?

Dr. Macklin: Daniele, do you want to take that one?

Dr. Gilkes: Sure, why not. We, I guess we both, applied to a program called the Amigos Program, which had to do with applied mathematics and oncology. It was a program that was trying to facilitate cancer biologist meeting with mathematicians, engineers, just kind of a bunch of different disciplines in the same room. So, both Paul and I, and several other people applied for this event, this sort of workshop and Paul can correct me if I'm wrong. We met at the National Cancer Institute for a meeting, but prior to that we have several phone conversations and web-chats where we could share data, through out different strange ideas, to try to see how our individual strengths could synergize to address a problem in cancer.

So, it was not ... I wouldn't exactly call it speed dating, but when we got the Amigos Conference people were sitting at different tables and kind of knew each other only from E-introduction, but we kind of went around and talked, showed our data, presented a little bit of data and Paul had already done computational modeling with hypoxia. I really wasn't familiar with his work because I hadn't ready a lot of mathematical descriptions, but then when we met and got to together we, sort of, had this perfect storm where we are experimental looking at tumors in animals and can identify regions of hypoxia. So, we can really serve him data on a platter and he can really work with that to expand our horizons because he can get his answers much more quickly than we can in our day-to-day work by basically generating a model built off of our experiment.

Chris Riback: And how ... Go ahead, Dr. Macklin. Sorry it sounds like you were about to say something.

Dr. Macklin: No, I think it's funny that Daniele mentioned speed dating because matchmaker came to my mind when you asked your question. This is a really innovated program put together by the BCRF and the JKTG Foundation and kind of facilitated by the MCI, where they had this self selecting group. Unlike a lot of workshops where they just toss you in a room and hope for the best ... They really facilitated a lot of discourse before we ever arrive just like Daniele mentioned. So we were primed and ready to go.

From my point of view, I'd been doing hypoxic modeling, but the work that Daniele has done was coming up with these incredible experimental systems where the cells change color and kind of remembered their history. That's something I'd never seen before and I was just absolutely fascinated by the new things we can learn by the system. Then our job is to kind of say, "The problem you run into experimental is can only image so many times." You know, every few days, in one mouse, in a different mouse and modeling provides a way to fill in the missing times and just start forming hypotheses on what happens in the parts of the system you can't see and maybe we can intervene.

Discussion like that were only possible in this really fantastic program that was put together by the MCI, BCRF and the JKTG Foundation.

Chris Riback: Yes, the Jayne Koskinas Ted Giovanis Foundation ... The work that they are doing with the BCRF. It's such an interesting way to think about. So, here you are doing the modeling and I assume, thinking, "Gosh, I really wish I could get more data, fresh data, more data. How can I get new imaging?" You know, Dr. Gilkes on that side of the equation is doing imaging as you just indicated. There is only so much imaging you can do. There's only so fast that the cells are moving within one body, and I would assume that as Dr. Gilkes is finding various images she's thinking ... You're thinking Dr. Gilkes, "Boy, how could extract like this? How can I move faster? How can I take the information that I'm gaining through this imaging and extrapolate it out and really start to move things?" And it's like, "Wait a minute this makes total sense merging the math and the science." And math is a science of course, but you know what I mean.

Is that a far characterization, Dr. Gilkes?

Dr. Gilkes: Yeah I think it's an excellent characterization. I mean, I think we, as biologist, ... Our field it moves a little slower than what we would like and if you can add in mathematics and we can teach at the computer ... Or Paul would have a better way to explain it, but he could actually begin to predict things now that we can't do. We can test his prediction, but we can't necessarily come up with a good prediction.

Chris Riback: So, take me through the actually research. I know both of you have talked about it a little bit already in this conversation, but you're looking at ... For this specific study you're looking at the low oxygen cells and these low oxygen cells to the extent to which they move through the body and metastasize. So, Dr. Gilkes you're, I assume, getting the imaging and getting the actually information from human cells and from actually cases. Then, Dr. Macklin you're taking that data and with your modeling extrapolating against various hypotheses that you both may have collectively, I assume collectively, or maybe you have individual hypotheses and you're able to test against each of them.

Take me through, kind of, ... Maybe we start with Dr. Gilkes and at the hand off, where it would hand off, in your professional scientific interaction ... Maybe hand that off in the conversation right now. As you tell it to me, if I am a person I would hear ... Listening to this, you know, I am a survivor or I have breast cancer or I have a loved one with it or I'm thinking of how can I make a difference in ... Really help me understand how can I benefit, how can it really make a difference in my life or someone who I love?

Dr. Gilkes: So, I think we are again aimed at looking at the cells that can metastasize. Those are the cells that we would like to go after. First of all we have to figure out what cells in that tumor have that ability. So, our experimental model, one of the experimental things that we're trying to do, is to basically ... We've come up with a little trick that can make cells change color when they don't have enough oxygen. So, using these trick ... If a cell is red it means it didn't experience these low oxygen conditions. If it's green it did.

So, by having these two different color changes we can form tumors in an experimental mouse model. Actually in the breast doctor, in the breast fat pad of the mouse. We can monitor these tumors as they grow over time and identify whether cells that were red had a better preference to metastasize  or whether those cells that were exposed to hypoxia had a better chance to metastasize. We do that both by monitoring the primer tumor or we can take sections over time and we can also look at the blood to identify those red or green cells. We can also look at distant organs.

So, from that data and with Paul's help we can ask within the tumor, "How quickly do these individual cell type proliferate? How quickly do they die? Are they localized near a blood vessel that would give them, maybe, a preference to enter the blood stream to begin that metastatic process?" We can, basically, collect all of the data. So, on our side we would be doing the actually experimental model of generating mice that will go on to develop metathesis so that we can generate all these tissue sections. Then with Paul's help we can ... Well we can do the imaging and Paul can help us analyze the tissue that we have. Then he can also use the data and I think he can talk about his digital pathology, where he can take that data and begin to teach the computer and generate a model, at that point.

Chris Riback: So, talk to us about the digital pathology, Dr. Macklin.

Dr. Macklin: Okay. Well this is something we are doing jointly, but digital pathology means to take the digital microscope slides and ... Daniele has been doing a tremendous amount on this. To take the pathology slides and you slice them very thin and you start staining for different things, like the colored markers that Daniele had mentioned. What you can do is use image processing algorithms to go and look for objects in those slides to start counting them. So you can count how many cells are hypoxic, how many cells are not, how many cells use to be hypoxic and still have the colored marker?

The other thing we can analyze is where those cells happen. Are those cells mainly in the middle of the tumor? Are they on the outer edge of the tumor? We can start, you know, getting this idea of how to correlate what's going on, versus where the cells are in the tumor and use those to start forming our hypotheses.

Then, in the computer modeling side those observations, those hypotheses, become rules in our computer program. So, we build the computer programs based upon the rules that we've observed together and we chat it over and say, "Oh, okay we see this, where they see this. We're going to guess this should be the following model rule." We program those into the computer models. We try to collaborate them as well as we can. Then, we simulate and see what happens and see if it matches our other observations, and dependent observations.

The other thing we are doing is we're really, kind of, building this into two stages. First, we have to build a model of just the primary tumor, the first tumor that pops up, the one that Daniele implants. Get the perimeters right for things like how the option moves through the tissue, how the cells consume it, how they become hypoxic, you know, these kinds of things. Once that part of the model is build then we build the next part. We say, "Now I need to build a model of metastasis."

So, we are building ... We're hoping is going to be a very efficient framework to say, "We can model a whole bunch of metastases all at the same time", and the missing piece of that puzzle is to say, "How quickly did the cells spread from the main site to other sites?" And "How often do they successfully seed a new metastasis?" By combining the data and the different parts of the model, they start getting this extra window into the dynamics of how these tumors change and evolve. That's really the biggest goal here, I think. For this first year here, is to really build this framework that says, "How can we get a new window into the processes? Based on combining state of the art observation and a model to fill in the pieces."

Once, that's built we have a base to build new investigations.

Chris Riback: What's the hope? Maybe it's new investigations, but even beyond that. As you open that new window, that you just described, what's the hope out of what you will discover or hope to discover about metastasis?

Dr. Gilkes: I think one of our main goals is first to decrease the time and the expense to get new treatments to patients. So, we would ... The synergizing of our output would be, potentially, if we could test different treatment strategies in his computational model. If we were able to build the model up to that level we could essentially do testing prior to any mouse studies. We'd have the opportunity to test a lot more perimeters and then really focus on just the more promising one.

Chris Riback: It could really cut down. So, not only could it really cut down the time ... Maybe it cuts down the time by rejecting, kind of, false avenues or paths that might not be as promising. Is that the time save? Obviously you're talking about a world in which every minute counts. You know, if you're talking about ones own life or a family member or a loved one time matters. Is that how you are cutting it down or are there other ways?

Dr. Gilkes: Well, I think we can test out a lot more perimeters then what us biologist could do in the lab. So, I think we would be able to test out many more strategies and hopefully come up with more hits. So, by testing more things.

Chris Riback: Got it. Dr. Macklin what has surprised you so far?

Dr. Macklin: Together we're building this system that gets beyond looking at individual parts of cancer because we've learned in the past that if you just target cancer cells and kill them as fast as you can you get unintended consequences like resistance. If you just target the blood vessels with hope of starving the cancer, that works in the short term, but again you get unintended consequences. The cancers usually adapt and evolve. Together we are building a framework that let's us attack cancer as a system. Hopefully by attacking cancer as a system we can avoid these unintended consequences and get more effective treatments.

I think that's the overarching goal at this platform. As we continue to build it over the years we're going to add more and more components that capture that complex system and let us target cancer more comprehensively.

Chris Riback: What has surprised you so far, Dr. Macklin and then maybe Dr. Gilkes you can answer it as well? Are you far enough into to have been surprised?

Dr. Macklin: Daniele, do you want to take this first? Or do you want me to?

Dr. Gilkes: I think I've just been surprise that we can work together so well having different backgrounds, but being able to communicate still, with the same goals of trying to improve patient outcomes.

Chris Riback: Dr. Macklin?

Dr. Macklin: Yeah, I'd buy that. I think also through the course of our discussions ... You know, mathematicians, by nature, try to simplify the things the things we're looking at. We try to abstract and form generalizations that try to get rid of the messiness and get to the heart of the matter. Sometimes we over simplify ... And the discussions we had together have really changed the way I think about cells, and cancer cells, and biology, and I can tell it's already making a big impact on how we do our models and proceed in our work. In traditional science, where the biologist work a while and then we work for a while, then we trade notes at a conference ever couple years is very slow. We've been able iterate on that in a matter of weeks rather than a matter of years. So, that's something that's been great about the program.

Chris Riback:Talk to me about that 'cause I'm trying to wonder myself ... One the one hand, what your describing is a perfect example of what we see in so many industries. We'll see it in technology, where you'll bring together a data programs and psychologist and let's say, business people or ... You know, they'll all sit in a room and that cross discipline approach works. So, you're in an environment right now with BCRF and with JKTG where they really are bringing together, there are few of these, research projects ... Where they are bringing together science and math and trying to create new solutions.

So, the scalability and the efficacy, let's say, of this model ... You know, broader outside the two of you, versus ... As I was listening to both of you at the beginning, it felt not coincidental that this how you're spending this part of your professional lives. I mean, Dr. Gilkes you've studies and worked in a range of different sciences, chemistry, the oncology. Dr. Macklin, you talked about how you could of been piano, you studied German. You've also ended up in applied mathematics. so, it seems like the two of you, individually, have this as part of your DNA's. Sorry to get all scientific on there with the DNA.

So, do you think it's you guys? Do you think it's that this mash up is the right model to extrapolate the future or science? Is it both? Do you need both the right model and the right people? Enough of my question, Dr. Gilkes ... What's the answer?

Dr. Gilkes: You make a good point. I hadn't really thought of the actually people and their willingness to reach out and be interdisciplinary. I think that's important, but I think also bringing these people together to, sort of, find each other is important too because from my end, and even academics you really see departments. The departments are generally very focused on one topic. So, being so focused you kind of forget that you're apart of a whole university that has many many departments. Sometimes you don't reach out to those other department just because of your low cal.

So, I think you need a spark, sort of the mashing up concept, and also you need people that are willing to listen to one another. Also, to be able to learn their language. There is a little bit upfront cost to the whole thing, where you need to spend the time to really learn what each other is doing. You can't be completely closed.

Chris Riback: Yeah, cannot be too parallel processes, I imagine.

Dr. Gilkes: Yes.

Chris Riback: Dr. Macklin, any view on your side.

Dr. Macklin: Yeah, I think that's right. I mean, this is kind of the whole soil and seed thing. You need an environment that's conducive to people working together and you need the people who are ready to do it. The Amigos Program literally has the word germinating somewhere in the acronym. Germinating solutions or something like that. So, I guess it goes with the soil and seed. I guess I hadn't thought about it in that way that we were kind of pre wired in our DNA to do this based on our backgrounds. With that said, I think there is this trend in both our fields to start working together on a more regular basis.

When I enter mathematical modeling in cancer more than a decade ago, it was a very rare thing to find a clinician or a biologist at a math meeting. We have to beg people to talk to us. So, you really end up with a lot of the toy models that had very limited impact, but now ten years later I think do the wonderful and insightful programs like this ... I think the cultures of both fields is changing for the better. You find a lot more biologist showing up at conferences and clinicians. You find that even a lot of biology programs are teaching their biologist coding and computational biologist is part of the core curriculum in way that didn't happen ten years ago. So, I'd like to think that this is the trend for the future, that you'll see more of this and we'll take a little bit less of this ... Maybe it won't take lightening anymore. Maybe it will only take, you know, some kind of smaller sparks to make this happen.

Chris Riback: Yes, I certainly hope and speaking for others, hope that this is just the beginning. To take your soil and seed metaphor, speaking for other, I know we're all looking forward to seeing the fruits of your labor. So, thank you both for taking the time and thank you both for your work. I really appreciate, and many of us appreciate it on both fronts.

Dr. Macklin: Well thanks very much for today, for the change to discuss with you and come back to us anytime if you ever want to learn more or have other questions.

Chris Riback:  I have no shortage of questions. They keep on coming. Thank you and thank you Dr. Gilkes as well.

Dr. Gilkes: Thank you, Chris.

Dr. Macklin: Thanks.


To learn more about breast cancer research, or to subscribe to our podcast, go to



Unite and Donate

Other Ways To Give

Donate By Check

If you prefer sending a donation by check, you can mail it to:

The Breast Cancer Research Foundation

60 East 56th St., 8th Floor

New York, NY 10022

Other Options

From paying tribute to a loved one to workplace giving, there are a variety of ways to support research.

Learn More