A Finer Fat Model
Models of healthy and diseased lipid profiles could prove valuable diagnostically.
When it comes to heart disease risk, “bad” and “good” cholesterol—also known as low density lipoproteins [LDL] and high density lipoproteins [HDL]—do not tell the whole story. These particles that carry fat through the blood can be broadly classified based on their density, but they actually vary widely in their composition and clinical risk. A new computational model, described in the May issue of PLoS Computational Biology, allows scientists to see this diversity for the first time, providing additional information to aid in diagnoses and treatment planning.
“We look at lipoprotein profiles in greater detail in order to find possibly relevant abnormalities in the lipid values that would remain undetected by looking only at LDL or HDL,” says lead author Katrin Hübner, PhD, a postdoctoral research fellow at the University of Heidelberg who completed much of the work while a PhD student at the Charité University hospital in Berlin. The model has several potential clinical applications.
Unlike previous models of blood lipid metabolism, which considered just four lipoprotein density classes (very low, low, intermediate, and high), Hübner and colleagues modeled the whole spectrum of individual lipoproteins—by combining any of three proteins (apoB, apoA, and other) and three fat molecules (cholesterol, triglycerides, and phospolipids) in varying amounts. The particles undergo 20 reactions, including particle birth from the liver, particle death from cell uptake, and transfer of fats between particles.
In initial simulations, Hübner and colleagues generated virtual blood lipoprotein profiles that closely matched experimental values from healthy individuals. Then they tweaked the parameters in their model to mimic three known lipid disorders. For example, to simulate familial hypercholesterolemia, which involves a malfunctioning LDL receptor, they decreased the rate of cellular uptake of apoB-containing particles (which are recognized by the receptor) by 75 percent. The simulations accurately reproduced the characteristic lipid profiles of the three diseases.
The model could help pinpoint the underlying molecular defect in patients with abnormal lipid profiles of unknown origin, Hübner says. It could also be used to predict the impact of specific treatments, such as drugs or lifestyle changes, on a patient’s lipid profile.
“This work addresses an important issue in modeling lipoprotein metabolism, which is the heterogeneity of lipoproteins,” says Brendan O’Malley, PhD, Project Leader of Systems Biology of Lipid Metabolism at Unilever Corporate Research in the United Kingdom, who also works on lipoprotein modeling (using a different approach).
“This is one of the first works in this area, so there’s still quite a lot of work to be done,” he says. For example, the model needs to be further validated with high quality patient data. But, in the future, it could lead to improved diagnostics and personalized treatments for cardiovascular disease, he adds.
“It’s not ready for the clinic yet,” Hübner agrees. “But we’ve made a promising first step.”