Supercomputing reveals mechanisms behind brain's waste handling

Our simulations show that amyloid beta, the substance underlying plaque aggregates that eventually is killing brain cells in Alzheimer's disease, should be able to diffuse out to the blood vessels. This is in contrast to a recent popular theory that the brain fluid flows like rivers through the tiny clefts around the brain cells, bringing the waste products along. It is, however, not possible to measure the pressure differences within these tiny clefts. We solved this by using electron microscopy images of the clefts and then simulating the flow through these clefts.

This required huge simulations, solving 100 million differential equations, and to our surprise we found that brain fluid will not be able to flow like rivers thorough the extracellular volumes. The cleft are too narrow. However, simulations revealed that amyloid and other waste products is transported out of these clefts by diffusion, the same mechanism bringing molecules of smell into your nose in a room.

The brain waste will end up in volumes surrounding the blood vessels, and we agree with others that along the blood vessels the waste is transported by bulk flow. The big question remaining to understand is the underlying mechanism for bulk flow along the vessels.

We believe a type of water specific channels in the membranes facing the vessels to be important for flow along vessels. These water channels were discovered in the brain by Erlend Nagelhus, Ole Petter Ottersen and coworkers in 1997, so we have a long tradition within this research field at the University of Oslo. We think that a combination of experiments and big simulations is necessary also for revealing the underlying mechanism for flow along vessels.

It is a relatively new approach to use supercomputing within medical research fields. The recent advances in computer hardware, processing power and software tools now make this possible. Such an approach is an important research tool and will be used more extensively in future medical research. We also see that people with background within informatics, physics and mathematics is entering medical research. Karl Erik Holter and Kent-André Mardal are situated at the Department of Mathematics at the University of Oslo and I am a physisist. The project is an interdisciplinary project between medical departments and computational departments at the University of Oslo and at the University of California San Diego.