Abstract: We present a dynamical model of lipoprotein metabolism derived by combining a cascading process in the blood stream and cellular level regulatory dynamics. We analyse the existence and stability of equilibria and show that this low-dimensional, nonlinear model exhibits bistability between a low and a high cholesterol state. A sensitivity analysis indicates that the intracellular concentration of cholesterol is robust to parametric variations while the plasma cholesterol can vary widely. We show how the dynamical response to time-dependent inputs can be used to diagnose the state of the system. We also establish the connection between parameters in the system and medical and genetic conditions.
Introduction
High levels of low-density lipoprotein (LDL)—the so-called "bad cholesterol"—are a primary driver of atherosclerosis and cardiovascular disease. Traditional understanding often relies on simple linear logic: increased dietary fat leads to higher LDL. However, this fails to explain the dramatic variability in cholesterol levels between individuals or the difficulty some patients face when trying to lower their levels through diet alone.
This research bridges the gap between highly specific fluid dynamics and abstract linear compartmental models. By developing a low-dimensional nonlinear model, the authors link specific genetic mutations and clinical conditions, such as Familial Hypercholesterolemia, to the underlying biochemical "wiring" of the human body.
Model Formulation
The researchers constructed a system of five nonlinear ordinary differential equations (ODEs) to track the "lipoprotein cascade." This process begins with the liver secreting Very Low-Density Lipoprotein (VLDL), which degrades into Intermediate-Density Lipoprotein (IDL) and eventually into LDL within the bloodstream.
Key features of the model include:
- Receptor-Mediated Uptake: Cells utilize LDL receptors (LR) to extract cholesterol from the blood.
- Regulatory Feedback: A critical nonlinear term where the concentration of intracellular cholesterol ([IC]) regulates the synthesis of new receptors.
- Bifurcation Analysis: The system's behavior was analyzed to find "fixed points" (steady states) and determine how these states shift as parameters like diet or medication (e.g., statins) vary.
Parameterization and Data
As a theoretical modeling study, the authors synthesized "nominal values" and parameter ranges from decades of established physiological literature, including work by Goldstein, Brown, and Packard. These values represent the metabolic rates of an average healthy human and the known ranges for various clinical conditions.
Key Insights and Results
The Bistability Trap and Hysteresis
The most significant finding is that for certain ranges of VLDL secretion and cholesterol metabolism, the system can exist in two different stable states simultaneously: a "low" healthy branch and a "high" diseased branch. This leads to hysteresis, or a "sticky" disease state. If a poor diet pushes a patient onto the high-cholesterol branch, simply returning to a healthy diet may not be sufficient to return the system to the low branch; the system "remembers" its diseased state.
Cellular Robustness vs. Plasma Volatility
Sensitivity analysis revealed that cells are highly effective at protecting their internal environment. Intracellular cholesterol ([IC]) remains remarkably constant even when external parameters change. However, the cell achieves this stability by "dumping" the variance into the plasma [LDL]. This explains why blood tests are so volatile while internal cellular functions remain stable. Mathematically, [LDL] was found to be an order of magnitude more sensitive to parameter variations than [IC].
Genetic and Diagnostic Mapping
The model successfully maps the five classes of Familial Hypercholesterolemia (FH) mutations to specific parameters. For instance, Class 1 and 2 mutations affect the synthesis parameter $c$, while Class 3 and 4 affect binding rates $d_I$ and $d_L$. Furthermore, the authors suggest that the relative amplitude of LDL oscillations in response to periodic inputs (like daily meals) could serve as a superior diagnostic marker compared to static blood tests.
Limitations and Future Directions
While the model provides a robust framework, it predicts unphysiologically high LDL levels in the extreme "high cholesterol" branch, suggesting the existence of secondary elimination mechanisms not yet captured. Future iterations aim to include High-Density Lipoprotein (HDL) as a dynamic variable rather than a constant rate and to distinguish between the regulatory roles of liver cells versus peripheral tissue cells.
Summary updated