As the COVID-19 pandemic and its consequences grew rapidly, many MBL affiliates and alumni thought, “What can I do to help?” This is the first in a series of articles about MBL community members who in some way pivoted their energies to address the coronavirus crisis. Suggestions for future profiles may be sent here.
By Diana Kenney
David Odde has spent years tracking how cancer cells migrate and metastasize. His group at the University of Minnesota is especially focused on glioblastoma, the aggressive form of brain cancer that took the lives of Senators Edward Kennedy and John McCain.
Using live tumor tissue from patients at his university’s medical center, Odde captures video of the cancer cells as they move and exit the tissue. He then develops computer models that reveal how the cells navigate: in what direction, with what force and speed. These models, or “flight simulators,” can be used to predict which drug regimens may most effectively disrupt the cancer’s progress.
Last March, as the coronavirus pandemic spread, “we began wondering why we weren’t applying our modeling expertise to COVID-19,” Odde says. “The more I thought about it, the more sense it made.”
Learning to Pivot at the MBL
To pivot his modelling from cancer to the novel coronavirus, Odde’s experience on the MBL Physiology course faculty from 2009 to 2013 offered a lift. “MBL was the place where we learned to rapidly prototype models,” he says.
Each year, Odde and his teaching assistant, Brian Castle, spent two weeks helping the Physiology students develop computer models for “whatever they were interested in,” he says. “We learned how to rapidly assimilate the [scientific] literature on a new problem, and in two weeks the students would produce a model,” he says. “It taught us to work with a sense of urgency. And this prepared us to respond to COVID-19.”
Over the past two months, with advice from experts in immunology, virology, and infectious disease, Odde and Castle rapidly read up on the first SARS epidemic in 2002-2003. “There is a lot of good cell biology literature around that virus, and we can leverage it because of its similarity to the 2019 coronavirus (SARS-CoV2),” Odde says. “Together with emerging biological and clinical information on the current coronavirus outbreak, we have pretty rich information that we can use to build and constrain our computational models.”
Odde’s group first developed a model for the life cycle of the novel coronavirus: how it infects the host cell, commandeers the cell’s machinery to produce more viral particles, and moves on. A second simulator will show how the virus spreads in respiratory tissue. And a third will model the immune response to the infection. They just launched a website for their COVID-19 modeling efforts.
Rationalizing the Myriad of COVID-19 Clinical Trials
Their overarching goal is to use the models to predict which drugs will most effectively combat the coronavirus, and thus should be prioritized for testing in clinical trials. “A big problem right now is there are so many different drugs and therapies that people want to try,” Odde says. “How do we decide which are the most promising? We think our models can help make more rational decisions about where to direct precious clinical resources, by leveraging our understanding of the biophysics [of the infection] and existing domain knowledge. We’ve been trying to do this with cancer therapies, so it’s not a big shift. “
Odde’s group is working closely with the principal investigators of three clinical trials at the University of Minnesota Medical Center. These trials, which are among a “cacophony” of more than 100 trials open nationwide to discover COVID-19 treatments, are testing the drugs hydroxychloroquine, remdesivir, and losartan. Odde’s model for the coronavirus life cycle can help them identify which parts of the life cycle may be more sensitive to which drugs, and at what dose.
“Ideally, you want [the drug] to target the weak points in the virus’s life cycle, the places where it is most vulnerable. We want to find where those vulnerabilities lie and align that with the current slate of clinical trials,” he says. “Our modeling may also help predict what combinations of therapies may be most effective. I’m excited about that because we may be in for some surprises. Two drugs that work reasonably well apart may not work together, or the reverse – in combination they might be very strong.”
Odde notes that the National Institutes of Health has launched a public-private partnership to prioritize the most promising vaccine and drug candidates for COVID-19. “We won’t be able to conduct all of these different trials,” Odde says. “So how do you narrow the options down, and do it rationally and fairly? We think modeling can play an important role in that process.”