Though she calls herself a mathematician, Franziska Michor's work on mathematical models of cancer doesn't fit neatly in that field or in the field of cancer biology. Instead, Michor is working in uncharted scientific territory, building bridges among math, computer science, biology, and medicine to answer questions about the origins of cancer, relationships among cancer types, and the emergence of drug-resistant tumors.
"I'm less interested in puzzle solving or very basic things that are not applicable to real-life situations," says Michor, who is currently based at Memorial-Sloan Kettering Cancer Center (MSKCC) in New York City. Though her research skills involve equations and computers rather than a pipette or a scalpel, her goal is the same as any other researcher in the oncology field: to eliminate cancer.
This unique approach to translational research earned her, in 2008, an R01 grant from the National Institutes of Health to model the biology of cancer stem cells. And in 2009, Michor became the principal investigator of one of the National Cancer Institute's 12 new Physical Sciences-Oncology Centers, a program that supports collaborations between natural scientists and clinical researchers to study cancer using new approaches. As part of that center, Michor and Eric Holland, an MSKCC physician-scientist, are working to predict the cell of origin for brain cancers and certain types of leukemia. If researchers better understood when and in what type of cell mutations arise, they'd have a better idea of how to choose the right treatment or develop new treatments, says Michor, who is just 27 years old.
Michor "has a skill for communicating with medical people, and probably that is the most important aspect of her success," says theoretical biologist Yoh Iwasa of Kyushu University in Fukuoka, Japan, one of Michor's longtime collaborators. "She's not just a translator," he adds: She captures the essence of a medical question and reframes it as a problem she can study using mathematics.
The family business
Blending mathematics and medicine is a novel twist on the family business: Michor's father Peter and sister Johanna are both mathematicians in her native Austria. "Math has always been a part of my life," Michor says. The abstract rigor of the subject appeals to her. "You solve something and it's true forever," she says. Her mother Elli, a nurse, provided the medical influence, with stories about her work with patients in the hospital.
Michor wanted to combine her interests, but the university system in Austria pushes students to specialize early, and an undergraduate degree includes only courses essential to understanding that specialty. So in 2000, when Michor started at the University of Vienna, she pursued simultaneous courses of study in mathematics and in molecular biology. She studied at her own blistering pace, taking examinations in 27 courses during her first year at the university. After completing her degree in 2002, she looked for graduate programs in which she could hone her research skills to focus on cancer. At 19, she moved to the United States for a Ph.D. program in evolutionary biology at Harvard University, where she developed evolutionary models of cancer.
One part of that work involved looking at models of chronic myeloid leukemia and resistance to imatinib (Gleevec), a drug that inhibits a specific protein that's characteristic of CML. Using mathematical models and a 169-patient data set, she and her colleagues examined the speed at which cancer cells were depleted when patients were treated. Among this group of patients, Michor concluded, imatinib did not eliminate the cancer stem cells: Instead, it targeted more differentiated cells, such as progenitor cells. The research offered an explanation for how cancer could recur in these patients.
Proximity to the bench and bedside
After completing her Ph.D. in 2005, Michor stayed on at Harvard as a postdoc. During the second year of her postdoc she was based at the Dana-Farber Cancer Institute to be closer to the patients and data that are vital to her research. She sought out a similar arrangement when looking for a faculty member position in 2007; she chose MSKCC, where computational biologists occupy a single floor, sandwiched above and below by several floors of experimental laboratories.
Close proximity to experimental colleagues has offered many opportunities to collaborate and see biology research in action. To better understand the raw data and biological systems she studies, Michor spends time in collaborators' labs looking at petri dishes, picking up a pipette occasionally, and checking in on mice in the animal facility. She has observed lung cancer surgeries performed by one of her clinical collaborators, MSKCC surgeon Robert Downey.
Franziska Michor, Robert Downey, and colleagues are working to create a phylogenetic tree of cancer. The figure above shows a tree for liposarcoma subtypes, with human mesenchymal stem cells on the left and normal fat cells -- which are fully differentiated -- on the right, and liposarcoma subtypes in between. The goal, the authors hope, is to identify genes related to differentiation that can serve as novel therapeutic targets.
Downey approached Michor as she arrived at MSKCC about working on a project to organize a phylogenetic tree of cancer based on clinical data, including gene expression. So far they've included samples of lung cancer, leukemia, breast cancer, and liposarcomas. It's miles from complete, Downey says, but it's beginning to reveal relationships between tumors. And the model is holding up to laboratory experiments intended to validate it.
In another project, Michor is examining dosing strategies that might help avoid drug resistance among patients receiving targeted drugs such as imatinib. By using mathematical models to simulate the effects of different treatment scenarios, such as a low daily dose of a drug compared with a high dose given less often, her work can narrow the number of strategies that should be tested in the lab and eventually in the clinic. Otherwise, she says, "what you have to do is implement every possible strategy in enough mice or people long enough so that you can see statistically significant differences."
Her work with MSKCC physician-scientist Holland for the Physical Science-Oncology Center is looking for the cell of origin for brain cancers and certain types of leukemia. It's a question that can't be addressed with experiments, Holland notes. If you manipulate the genes of a living organism to see what happens, it doesn't actually simulate the evolution of cancer as it occurs naturally in an unperturbed cell. So Michor and her group are using biological information to build mathematical models to predict the cell of origin; then researchers in Holland's group will use mouse models to verify the mathematical predictions. "That's an example of where math tells you something that biology simply can't tell you," Holland says.
While her work has attracted the attention of collaborators and funders, publishing her research has proved less straightforward, she says. Even though computational work can be done relatively quickly, Michor's work relies on experiments done by her collaborators to verify that models are biologically or medically relevant. Because of these cross-disciplinary methods, a paper can include sequencing data, gene expression data, and growth data alongside the mathematical models. As with some other interdisciplinary fields, it's often difficult to find the right journal or reviewers who can evaluate the combination of mathematics and biology as a cohesive whole rather than as individual components.
But even with the challenges of her work, the opportunity to approach familiar problems with different questions is gratifying, Michor says. Her collaborators appreciate that perspective: "She's looking at [a research question] from the outside in and then applying tools and a wisdom vantage point that really makes you think about it differently," Holland says.
But it's a complex role. "With math, if you do it once -- unless you made a mistake -- it's going to be the same every time you do it," Michor says. "However, if I put on my biology hat, it's very hard to come up with a mathematical model that abstracts at the right level because [the biology is] very complex." A good model must be simple enough to provide conceptual insight but not overlook important components, she says. "It's almost an art, really."
A long journey ahead
Carrying out her work in a medical center and next to a hospital provides an added motivation for her work. Just walking to the hospital to attend meetings and encountering cancer patients serves as a powerful reminder, she says. "I think, 'Oh my gosh, I've got to get back to work. This is really pressing.' " Downey also finds that the collaboration with Michor helps him make sense of the apparent chaos of cancer in the clinic -- a disease that, at times, seems wildly out of control. "She's able to show that cancer is to some degree a coherent process," he says.
In September, Michor is moving her lab and its eight researchers to Boston, to the Dana-Farber Cancer Institute and Harvard University's School of Public Health -- familiar stomping grounds, and an institution that, like MSKCC, offers opportunities to work closely with clinicians. The move will allow her to work in the same city as her husband, Roland Fryer Jr., a Harvard economist who models educational systems. Her collaborations with many of her MSKCC colleagues will continue, and the Physical Sciences-Oncology Center will move with her.
Just like other cancer researchers, she hopes her work will eventually save lives, but, like a true mathematician, she calls the probability of that "something very, very tiny with lots of zeroes after the decimal point." Answering her fundamental questions about cancer could take decades, but fortunately she's young enough to have time on her side: "It's very risky, but I think it's worth it."
Sarah Webb writes about science, health, and technology from Brooklyn, New York.