A Department of Energy program that opens some of the world’s most powerful computers to researchers around the globe has generated a promising lead for a Parkinson’s disease treatment.
A team led by researcher Igor Tsigelny of the San Diego Supercomputer Center (SDSC) at the University of California-San Diego (UCSD) used the Blue Gene/P supercomputer at DOE’s Argonne National Laboratory to simulate how proteins called alpha-synucleins (aS) damage neurons. The simulation showed in detail how aS molecules join into ring-like structures that penetrate the cell membranes and create pores. In the case of Parkinson’s disease, the pores can lead to death in the brain’s dopamine-producing cells, causing loss of mobility and other symptoms that worsen over time.
Tsigelny is collaborating with UCSD neuroscientist Eliezer Masliah to verify the modeling and simulation findings. The computational work also generated leads for a drug to slow Parkinson’s progression. Tsigelny and Masliah already have experimental results that support their hypothesis.
The SDSC team did its simulation with a grant of 1.2 million processor hours on the Blue Gene/P, ranked as the world’s fastest computer dedicated to unclassified research, through DOE’s Innovative and Novel Computational Impact on Theory and Experiment (INCITE) program. INCITE opens DOE’s most powerful computers to scientists and engineers for simulation and modeling projects with the potential to make major advances.
Tsigelny’s team has received INCITE allocations on Blue Gene for three straight years, starting with 16,000 processor hours in 2006 and 75,000 in 2007.
“Because this study needs a lot of computation, a supercomputer is absolutely necessary,” Tsigelny said. “Without supercomputing, I wouldn’t be able to make these discoveries."
At least half a million people in the U.S. are believed to suffer from Parkinson’s and about 50,000 new cases are diagnosed each year, the National Institute of Neurological Disorders and Stroke reports.
Clumps of aS in the brain have long been associated with Parkinson’s and other degenerative diseases, but Tsigelny said by the time clumps appear, the damage has already been done. The disease starts when aS aggregation on cell membranes creates pores – long before clumps appear.
Proteins are the cell’s workhorses, carrying out vital maintenance and metabolic functions. Every organism’s cells has thousands of different proteins made of amino acids chains that in most cases fold into precise shapes called conformations. Some proteins, including aS, can have a significant number of different conformations, some of which can cause problems, including disease.
The SDSC researchers developed software called membrane-associated-proteins assessment (MAPAS) that predicts which sections of amino acid chains are most likely to contact and bind a protein to the lipids or fats on the cell membrane. MAPAS helped identify the aS conformations that would lead to the ring-like aggregates on cell membranes.
The team’s molecular-scale simulations, which ran on the Blue Gene/P at Argonne’s Leadership Computing Facility, helped to identify not only the aS amino acid sections that bind to the cell membrane, but also the contact sections where aS molecules join to form the damaging rings. “We found the vulnerable positions in these rings, where we can now talk about drug design” to prevent them – and the damaging pores – from forming, Tsigelny said.
Masliah’s experiments with cultured neurons and electron microscopy confirmed the simulation’s findings. Tsigelny and Masliah also are working to develop compounds that can stop aS aggregation in cell cultures – a first step toward a drug to treat Parkinson’s disease. Preliminary cell culture experiments show the approach is effective, and the team is conducting experiments in mice.
“It’s a very, very good collaboration,” Tsigelny says, with simulation results suggesting approaches for Masliah’s experiments and the experiments supporting the molecular models.
Because the set of programs and methods developed by SDSC scientists can simulate a number of protein-cell membrane interactions, Tsigelny also is using it to study other conditions, including Alzheimer’s disease, kidney diseases and some cancers. The same molecular modeling concepts also can be used to simulate bacteria designed to absorb metals. Such bacteria could have a range of uses, including environmental decontamination.