Computer simulations indicate new drugs
Few diseases are as emotionally painful, both for the victims and for their families, as Alzheimer's. The anguish of watching the gradual deterioration of a loved one's mental faculties for decades is accompanied by the harsh realization that the process is irreversible. Against this backdrop of despair, recent simulations from one of the world's most powerful supercomputers provide cause for hope.
For the majority of people diagnosed with Alzheimer's, the degenerative brain ailment is a deadly two-protein disease. Amyloid protein lurking outside neurons, the fundamental units of the body's nervous system, forms plaques. Tau protein, loitering inside neurons, forms neurofibrillary tangles. At present, drugs delay symptoms but ultimately do not halt formation of plaques and tangles. Tomorrow's drugs, however, may turn into a medical myth the prevailing view that Alzheimer's disease is unstoppable.
In 2007 ORNL researchers Edward Uberbacher and Phil LoCascio used 100,000 processor hours on the Laboratory's Cray XT4 Jaguar supercomputer to investigate the mechanisms by which a new class of drugs acts. The drugs, called caprospinols, may stop the growth of Alzheimer's fibrils and even disassemble the threadlike fibers.
"This is the first time that molecular dynamics have been used to simulate the mechanisms these drugs use to interact with and reduce the growth of Alzheimer's fibrils," says Uberbacher, who leads a joint ORNL-University of Tennessee team. "We learned that these drugs work several different ways, and the findings gave us new ideas about how to improve the drugs."
Uberbacher, LoCascio and colleagues used a software code called LAMMPS to develop a computational simulation of the interactions of different drugs with Alzheimer's fibrils. "This simulation is very much like an experiment," Uberbacher says. "The simulation shows us lots of different possible drug interactions with the fibril at once."
The researchers used the information to explore mechanisms by which drugs attach to and reconfigure small proteins called peptides bound in fibrils, which aggregate in the Alzheimer's brain as plaques. The buildup of proteins may cause loss of neurons and vascular damage, leading to degeneration of the brain.
Prior to running the simulation, the scientists mathematically represented the chemical bonds within the drugs and fibrils, which set the parameters for possible types of molecular interactions.
"Because we can perform quantum-mechanical, ab initio calculations on one thousand or so atoms, we can generate this knowledge in a way that is more accurate and useful than what pharmaceutical companies usually produce," explains LoCascio. "Hopefully this method will become more widespread in industry and lead to better drug design."
During the simulation, the ORNL researchers used the super-computer to perform molecular mechanics calculations to predict each drug's activity. Drug molecules interacted with the protein molecules of the fibrils, but they also interacted with each other.
Results show promise. Some drug molecules were found to bind to the growing ends of Alzheimer's fibrils, impeding further growth. A drug developed by researchers at Georgetown University and licensed by Samaritan Pharmaceuticals prevented an Alzheimer's peptide from changing to a conformation that would allow addition of peptides to a growing fibril. Another drug unraveled tangled fibrils by causing their peptides to dissociate.
Collaborators at the University of Tennessee are conducting laboratory experiments to evaluate promising compounds in mouse brains. UT researchers have developed specialized micro-CT and MRI technologies for imaging Alzheimer's plaques in the brains of small animals. In addition, UT hosts a transgenic colony of mice engineered to harbor a gene associated with human Alzheimer's disease.
The specialized brain imaging and genetically unique animals at UT and supercomputer simulations at Department of Energy facilities have improved understanding of how drugs act on fibrils. The insight paves the way for rational design of new drugs, Uberbacher says.
"The simulations performed on Jaguar are an important demonstration of a new paradigm for dynamic modeling of drug-protein interactions," Uberbacher says. "As a bonus, the collaboration is a model for how DOE computing facilities can interact with medical universities."
Researchers believe the awesome power of the world's largest computers may be what is needed to break the hold that Alzheimer's disease has on elderly populations. —Dawn Levy
Web site provided by Oak Ridge National Laboratory's Communications and External Relations