anton 2 sims suggest how
fish oil health benefits
start at cell membrane
Molecular simulations coupled with AI analysis reveal how omega-3 polyunsaturated fatty acids may regulate the size of membrane “rafts”
The omega-3 polyunsaturated fatty acids (n-3 PUFAs) found in fish oil have healthy effects on heart risk, blood pressure and even rheumatoid arthritis. But doctors don’t know how they work. A team of scientists from Indiana University – Purdue University Indianapolis (IUPUI) used Anton 2, a special-purpose supercomputer designed and developed by D. E. Shaw Research (DESRES) and hosted at PSC, combined with artificial intelligence (AI) analysis of the results, to show how these oils affect the structure of the cell membrane. Their results hint at a fundamental mechanism that may underlie the n-3 PUFAs’ health benefits.
Above: Because some components of the cell membrane, like N-palmitoyl-sphingomyelin (PSM, red in bottom row) or cholesterol (white at bottom) tend to be solid at body temperature and others, like phospholipids (blue at bottom) tend to be liquid, the former tend to form solid “rafts” floating in the two-dimensional liquid of the membrane (left). By making the phospholipids contain omega-3 polyunsaturated fatty acids (PDPC, blue) and simulating the membrane on Anton for six microseconds (right), scientists showed that the PDPC formed a “not-solid, not-liquid” buffer zone (“Interface,” dark blue in the middle row) around the rafts that would allow them to combine and grow larger. Reprinted with permission from Canner SW, Feller SE, Wassall SR, Polyunsaturated Phospholipid Bilayer: A Supervised Machine Learning Analysis of Molecular Dynamics Simulations. J. Phys. Chem. B 125(48) 13158-13167. Copyright 2021, American Chemical Society.
Why It’s Important
In a world of dietary supplements with dubious health claims, the n-3 PUFAs found in fish oil seem to be the real deal. There’s good evidence they decrease the risk of dying from heart disease; reduce high blood pressure; lower high triglycerides and cholesterol; and relieve pain and other symptoms of rheumatoid arthritis. Problem is, doctors are largely in the dark about how the docosahexaenoic acid (DHA) and eicosapentaenoic acid (EPA) found in fish oil produce so many health benefits.
“My lab is interested in how polyunsaturated fatty acids modify the architecture of membranes. Polyunsaturated fatty acids … incorporate into membrane phospholipids, and one idea is that they modify the architecture of ‘lipid rafts.’ A lipid raft is a region that hosts signaling proteins … When they come together, the signaling proteins they contain trigger. And there’s been quite a lot of debate about how polyunsaturated fatty acids would affect that.”—Stephen Wassall, IUPUI
That’s why Samuel Canner, then an undergraduate student working in the laboratory of Stephen Wassall of IUPUI, decided to find out how n-3 PUFAs affect the cell membrane at a molecular scale. Canner is currently a graduate student at Johns Hopkins University.
Earlier work by the lab had shown that DHA and EPA, when incorporated by the cell into phospholipids that make up much of the membrane, allow “rafts” of fatty molecules in the membrane to combine and grow larger. Scientists suspected that when the rafts combine, it brings together receptor proteins in the rafts that then trigger healthy changes in the body’s metabolism. Canner used a novel combination of simulations on Anton with artificial intelligence to test how the n-3 PUFAs cause the rafts to grow.
How PSC Helped
Much of the cell membrane is a kind of two-dimensional sea of the fatty chemicals that make it up. Embedded in the membrane are the receptor proteins that sense the cell’s surroundings and change its behavior in response to signals from elsewhere in the body. But these proteins don’t float in a simple fluid — some chemicals found naturally in the membrane, such as cholesterol, tend to be solid at body temperature. Others, like the phospholipids, are liquid. Because of this, the more solid components form compact rafts that float in the surrounding liquid and contain the receptors.
Scientists knew that these receptor proteins play an important role in moderating the body’s metabolism. Many of the receptors activate when they group together. By helping the rafts to grow larger, as the lab results suggested, the n-3 PUFAs might be causing the receptors to group and activate. But that’s as far as the evidence went.
The IUPUI team turned to Anton at PSC because its specialized design allows it to more quickly carry out molecular dynamics simulations of large chemical systems at longer timescales than are possible with general-purpose supercomputers. In other words, Anton could perform the needed simulations in far less real time than general-purpose supercomputers. Anton was designed and made available without cost by DESRES and hosted at PSC thanks to operational funding from the National Institutes of Health.
“We used Anton 2 to pre-form a raft, and put it in the center of a disordered, liquid patch containing DHA. Then we just let it run for six microseconds and study how that time evolution works and see how the mixing of these lipids goes about … You can look at the picture and you’ve got an idea … but where do you draw the line that defines the edge of a raft? It’s very arbitrary, and the machine learning gives you a very distinct and easy way to define it …”—Samuel Canner, IUPUI
There was a hitch. Anton could simulate the massive system as the fatty molecules interacted and show how n-3 PUFAs affected the mixing. But the challenge of interpreting the results seemed to call for some additional tools. So Canner performed a second computational step. Using the Big Red II supercomputer at Indiana University, he trained an AI program using a technique called machine learning to classify the regions in the Anton simulation as solid or liquid, taking human bias out of the loop.
The simulations and AI classification revealed that phospholipids containing the n-3 PUFAs didn’t mix into the rafts. Instead, they helped form a region around the raft that wasn’t quite solid and wasn’t quite liquid. By forming a buffer state between raft and liquid, the fish oils stabilized the rafts, allowing them to grow larger. The scientists reported their results in The Journal of Physical Chemistry in November 2021.
Canner’s results were the first molecule-by-molecule demonstration of how the n-3 PUFAs cause rafts to grow. But it’s just an initial step in the group’s plan to understand the dietary effects of these fatty acids. For one thing, previous research methods couldn’t distinguish between the effects of the two n-3 PUFAs, DHA and EPA. The Anton simulations likely can, offering a window into whether each has distinct health effects. The scientists would also like to demonstrate that the larger rafts really do activate the receptor proteins. Finally, the many cell, tissue and bodily responses to the receptor signals must be traced to figure out how they generate the health benefits.