Structural biologist Nicolas Fawzi will direct Brown's role in a multi-institution project to transform the understanding of cellular protein dysfunction in the earliest stages of neurodegenerative disease.
PROVIDENCE, R.I. [Brown University] - Brown University Professor of Molecular Biology, Cell Biology and Biochemistry Nicolas Fawzi will direct Brown's role in a multi-institution project funded by a federal grant of up to $28.6 million focused on better understanding neurodegenerative disease.
By developing new approaches to detect and characterize protein misfolding in cell models, the researchers will aim to transform the understanding of protein dysfunction in the earliest stages of neurodegenerative disease. Brown is one of seven research institutions with researchers participating in the project, which is called NATIVE-ID and is led by scholars at the Innovative Genomics Institute at the University of California, Berkeley.
The idea is to study protein variants in live, disease-relevant human cells at scale. In parallel, the research team will create a comprehensive machine learning model to predict the complete progression toward disease, from healthy states to misfolding and aggregation, which are the earliest stages of protein dysfunction known to lead to neurodegeneration. The model and the data used to train it will unlock new drug targets to preserve healthy proteins and function, the researchers expect.
Brown's role in the project will be funded by a sub-award of approximately $1 million.
"Our team is kind of a tool-development team," Fawzi said. "We're building a model that will ideally be able to tell us where the problem is for many diseases caused by many different proteins. The idea is that it would be available to not just our team, but to other teams to make impacts on disease targets."
Most neurodegenerative diseases begin with misfolded cellular proteins, Fawzi noted. Among the proteins made by the human body, well over 50% have at least one intrinsically disordered region that does not have a fixed three-dimensional structure. These regions can cause proteins to clump together or misfold in ways that damage cells and cause multiple diseases that have proven challenging to treat.
Scientists have made enormous strides in understanding protein folding in recent years thanks to AI tools like the Nobel Prize-winning AlphaFold2. But as powerful as AI models are, they currently have a key limitation: they can only predict the structured portions of proteins. Fawzi's lab uses nuclear magnetic resonance spectroscopy to be able to analyze protein structure, motions and interactions one atom at a time - an approach, he said, that doesn't require proteins to be in a rigid or structured state.
"In my lab, we look at proteins atom by atom, before they form disease-associated, pathological aggregates," Fawzi said.
Fawzi's team will share data from their analyses with the NATIVE-ID researchers who are building the dataset that will enable them to train novel machine learning models to predict the behavior of variants in intrinsically disordered proteins and how they will relate to disease states. The NATIVE-ID team is focusing on frontotemporal lobar degeneration (FTLD), a neurodegenerative condition in which damage to nerve cells associated with protein misfolding and aggregation can lead to early-onset dementia.
The larger goal is to be able to generalize the findings to other diseases, said Brad Ringeisen, executive director of Berkeley's Innovative Genomics Institute and lead primary investigator on the project.
"Diseases like Alzheimer's disease, Huntington's disease, amyotrophic lateral sclerosis and Parkinson's disease involve proteins that misfold or aggregate, often due to intrinsically disordered regions," Ringeisen said. "The technology is at a state now where we can watch, in real time, proteins move and aggregate inside living cells. We will leverage this amazing capability along with automation, robotics and machine learning to understand what causes proteins known to be involved with neurodegeneration to start misbehaving, enabling new precision treatments that could prevent these diseases from occurring in the first place."
The federal grant comes from an Advanced Research Projects Agency for Health funding program focused on developing new tools for predicting the neurodegenerative health impacts of variations in intrinsically disordered proteins. In addition to U.C. Berkeley and Brown, the project includes researchers from Emory University, Johns Hopkins University, Parallel Squared Technology Institute, Stowers Institute for Medical Research and Texas A&M University.
