A single tumor often features multiple varieties of wildly proliferating cells, with each variety undergoing genetic mutations. Chasing these moving targets, researchers tailor therapies that arrest—but only rarely vanquish—the disease. However, momentum may finally be shifting in the quest for effective cancer treatments.
“For so long, therapy has been reactive,” says Michael Hemann, who has joined forces with fellow MIT faculty member Douglas Lauffenburger to address this challenge. “What if we could instead steer tumors toward an outcome we know how to manage, or toward becoming better behaving tumors?”
This is no wishful thinking. Hemann is professor of biology at MIT’s David H. Koch Institute for Integrative Cancer Research, which has just celebrated its fifth anniversary. He and Koch Institute extramural faculty member Lauffenburger—the Ford Professor of Biological Engineering, Chemical Engineering, and Biology, and head of MIT’s Department of Biological Engineering—have been collaborating on studies that suggest it is possible to predict and shape a cancer’s unique trajectory, and to determine points along that journey when it may be especially susceptible to treatment.
Their work, in a field of research called systems biology, creates a finely detailed portrait of the complex evolution and drug responsiveness of certain kinds of cancer.
“Only by embracing and comprehending the complexity can you hope to come up with effective treatment,” says Lauffenburger. “And we have the experimental tools to be as comprehensive about studying complexity as we want.”
Hemann arrived at MIT in 2006 skilled in the latest methods for manipulating genes in living organisms. As a graduate student at Johns Hopkins University, and then as a postdoctoral fellow at Cold Spring Harbor Laboratory, he had trained to use viruses and RNA interference to grow specific cancers in mice. He more recently added CRISPR to his repertoire, an even faster technique for modifying the genome of living cells.
“We have an entirely new toolset for manipulating genetic systems in vivo,” says Hemann. “We can now perform big genetic screens with mouse models, looking at lots of phenotypes [expression of genetic traits], introducing many changes at once to see how a tumor emerges or resists a cancer therapy.”
These kinds of experiments generate reams of genetic data that require sophisticated analysis—an area that falls directly in Lauffenburger’s wheelhouse. On the MIT faculty since 1995, Lauffenburger calls himself “half cell biologist, half engineer,” and was prepared when “biology hit the omics era.” (“Omics” refers collectively to the study of genes and proteins that comprise living organisms.)
With research interests in cancer and biomedical engineering, Lauffenburger devises computational strategies for capturing changes within complex biological systems from the molecular level up. By “creating conceptual frameworks,” Lauffenburger says, he aims “to get the most power out of omics experimental methods.”
“The unexpected can be transformational”
With their common interests and complementary skillsets, Hemann and Lauffenburger seem like an obvious research match. But they needed a well-placed nudge to forge a union. This was delivered in 2008 by Justin Pritchard PhD ’12, then a graduate student of Lauffenburger’s, who was intrigued by the cancer data flowing out of Hemann’s lab. “This situation is the epitome of MIT,” says Lauffenburger. “It’s the brilliant, fearless, creative graduate students who find connections between labs.” They “facilitate our interaction in a deep way,” adds Hemann. “Students are the glue that holds us together.”
In the course of investigating how combinations of drugs worked on B-cell lymphoma, a type of blood cancer, Hemann had generated a very large data set. “In Mike’s world, you can perturb hundreds to thousands of gene products,” says Lauffenburger. “The issue is figuring out what’s important to tumor biology.”
Based on decades of experience treating patients, clinicians have discerned that some drugs in combination can achieve a kind of one-two punch against B-cell lymphoma. But the biological mechanisms behind their efficacy remained unknown. Pritchard realized that by using Lauffenburger’s computational models, he could mine the giant data set for patterns of drug impacts, gleaning likely pathways of tumor susceptibility, and identifying which drugs worked best, and in what combination.
Pritchard’s research “gave a multivariate genetic foundation” to a common clinical practice, says Lauffenburger, helping provide “a biological rationale for this kind of drug treatment.”
This study, the basis for Pritchard’s thesis and multiple journal articles, was the launching point for a series of collaborative ventures between Lauffenburger and Hemann—all facilitated by graduate students. It is a partnership that is taking both labs into new scientific territory, and breaking new ground in cancer research. As Hemann puts it: “In biology, the unexpected can be transformational.”
The two laboratories began to zero in on what Hemann calls “one of the essential problems in cancer biology”: tumor heterogeneity. While tumor cells start as single cells, they begin growing uncontrollably, and then differentiate into diverse subpopulations. There can be heterogeneity of tumors across patients with the same cancer, as well as heterogeneity within the same tumor.
Given such wild variation in a given type of cancer, how do researchers identify effective treatments, especially when the treatments themselves promote mutations and further drug resistance?
“Now we can stay ahead of the game”
Another joint graduate student, Boyang Zhao PhD ’16, began to crack this puzzle. In Hemann’s lab, he created heterogeneous lymphoma tumors in mice, and then tested these tumors with single and combination drug therapies. Zhao used Lauffenburger’s computational tools to analyze data comprised of 10,000 heterogeneous tumor compositions and their response to six drugs.
“The computational work allowed Bo to simulate an evolving mix of tumor cells, so he could predict the response of these heterogeneous cells to treatment,” says Hemann. “This really moved us forward.”
The team’s focus on heterogeneity began paying off rapidly. Using mouse models of acute lymphoblastic leukemia, Zhao discovered that at an early stage of the evolution of the cancer, it developed an acute sensitivity to drugs that had demonstrated no previous efficacy in the treatment of this disease. “This is the awkward phase, the teen years, for the tumor, when it’s hypersensitive to drugs,” says Hemann.
This research suggests not only that modeling can predict optimal times for treating this leukemia, but that it might also be possible to dynamically modify cancers in order to sensitize them to therapy. “Now we can stay ahead of the game,” says Lauffenburger. “We know at what point to hit the cancer with a new drug.”
These findings, published in the March 24, 2016, issue of the journal Cell, have the potential to improve treatment for a range of blood cancers, including acute myeloid leukemia, with its diversity of genetic subpopulations, and others such as chronic myelogenous leukemia, where only patients with a specific genetic mutation find relief through a highly targeted drug regimen.
But the team’s hybrid approach, combining genetic manipulation in vivo and powerful computational frameworks, has even broader potential. “We believe it could apply to any type of heterogeneous cancer, which really means any type of cancer,” says Lauffenburger. “It leads us to a new world of drug screening,” adds Hemann. “We can now determine a tumor’s unexpected sensitivities, and find new compounds that have efficacy during the evolution of the disease.”
While their work raises the possibility of rapid testing of new and more targeted cancer drugs, it also points to better application of current drugs. “If we could block the protective signals in some tumors that make them drug resistant, and find the best time to administer the drug, then current frontline chemotherapy could work better at lower doses,” says Hemann. “Our big mission is to make therapies more effective and less toxic.”
To that end, the researchers will be partnering with clinicians at local teaching hospitals. “One of our next proving grounds will be drug resistance in lung cancer,” says Lauffenburger. This means expanding their collaboration: “We will be adding more students to our labs,” notes Hemann.
It’s a prospect both scientists relish. “Our interactions are multifaceted, with the science and personal dimensions all intertwined,” says Lauffenburger.
“We know how to defer to each other’s expertise, and this has allowed us to move in directions we never would have before,” says Hemann. “As with any good relationship, you find a situation that endures because it’s both productive and exciting.”