Michael Mitchell Receives NIH Director’s New Innovator Award

Michael Mitchell Receives NIH Director’s New Innovator Award

“High-Risk, High-Reward” Research Will Apply Machine Learning to Engineer Delivery Technologies for Gene Editing and Immunotherapy

In the era of personalized medicine, the differences between individual cases of a cancer do not just vary from patient to patient, but also from tumor to tumor, and even from cell to cell.

Michael Mitchell, Skirkanich Assistant Professor of Innovation in Penn Engineering’s Department of Bioengineering, is drawing on a variety of fields — biomaterials engineering, data science, gene therapy and machine learning — to tailor the next generation of drug delivery vehicles with this level of precision.

His work in this field has earned him a $2.4 million NIH Director’s New Innovator Award, which is part of the NIH Common Fund’s High-Risk, High-Reward Research program. The High-Risk, High-Reward Research program supports innovative research proposals that might not prove successful in the conventional peer-review process despite their potential to advance medicine.

“This program supports exceptionally innovative researchers who have the potential to transform the biomedical field,” said NIH Director Francis S. Collins, in a statement. “I am confident this new cohort will revolutionize our approaches to biomedical research through their groundbreaking work.”

“We’re proposing a ‘4D’ platform: data-driven drug delivery,” says Mitchell. “This combines what we commonly use in drug delivery — engineering or designing new biomaterials and nanotechnologies that encapsulate drugs and getting them where they need to go — with machine learning and data science so a computer can help us identify which molecular properties of a drug delivery technology allow us to target diseased cells in the body. The hope is that one day, instead of testing these delivery systems one at a time, we can test many of them at once to define the physical and chemical parameters that make them more or less effective. The idea is that we will be able to predict, before we administer these vehicles to a patient, which delivery technology is best suited for a particular tissue, and even a specific subset of diseased cells.”

Mitchell and his colleagues are particularly interested in designing delivery technologies for gene editing and immunotherapy to treat bone marrow cancers, such as leukemia and multiple myeloma. Myeloma is particularly hard to treat; patients with chemotherapy-resistant tumors can have very poor prognoses, measured in months. Finding the cancerous mutation in a myeloma tumor may prove fruitless, as different parts may have different mutations. Killing the slow-growing part of a tumor may even enable the faster-growing part to spread more aggressively.

“What makes our research high-risk and a potential paradigm shift in the field is that we’re not developing therapeutics that target the tumors themselves; we’re targeting the microenvironment,” Mitchell says. “We’re targeting the vessels that surround the myeloma tumor, the stromal cells, the immune cells and the adhesion receptors that those cancer cells are exposed to. Rather than targeting a tumor’s specific mutation, we’re targeting how it interacts with its surroundings. We are manipulating these interactions using nucleic acid therapeutics, including small RNAs, messenger RNA, and CRISPR-Cas9 genome editing approaches.”

“If we can change how myeloma moves and crawls and feels its environment,” he says, “can we use this as a new way to overcome chemoresistance and make clinical drugs work better? By controlling adhesion and migration, we can influence how myeloma and potentially other cancers get into vessels and travel. If we disrupt those interactions with our delivery systems, via nucleic acid technologies, for example, that in itself could be therapeutic by slowing down the spread of a range of cancer types throughout the body.”

The key to Mitchell’s approach is in both the high-throughput and the computational approaches he and his colleagues propose to develop delivery vehicles that are tailored to these multifaceted criteria. Starting in mouse models of multiple myeloma, the researchers will launch a fleet of drug delivery vehicles with different combinations of shapes, sizes and chemical decorations on their exteriors.

“Instead of administering them one at a time, seeing where they go, then trying again, we’re using techniques such as molecular barcoding to sequence where many different delivery vehicles go,” Mitchell says. “That allows us to accelerate the screening process and identify key parameters — is there something unique about the chemistry that makes it go to the bone marrow where a tumor is, versus getting filtered out by the liver or the kidney?”

By screening those parameters and using machine learning algorithms to figure out which sets of traits are most effective at clearing the various hurdles drug delivery vehicles encounter, Mitchell and colleagues aim to tailor those vehicles for each treatment.

“Looking forward, our ultimate goal is to utilize this drug delivery technology platform for personalized medicine in collaboration with clinicians at Penn’s Perelman School of Medicine,” Mitchell says. “If you are going to perform a bone marrow biopsy on a patient who potentially has multiple myeloma, can we — one day — introduce many different FDA-approved biomaterials at a low dose, identify which one best gets into a patient’s tumor, and then go back and administer a drug delivery system and a therapeutic that are now both uniquely suited for that patient’s needs?”

Mitchell is one of seven Penn researchers to receive NIH Director’s Awards from the NIH Common Fund’s High-Risk, High Reward Research program this year, among 89 awardees nationally. He is the sole recipient of an NIH Director’s New Innovator Award from Penn Engineering. The $2.4 million award is disbursed over a five-year period.