LDL retention time in plasma can be -based on causation- estimated by the lipid composition of LDL and other lipoproteins

Abstract

Introduction: Information on LDL's dynamic behaviour of LDL (i.e. production rate and fractional catabolic rate) are of interest if pathologies, lipid-lowering strategies or LDL-metabolism itself are investigated. Determination of these rates is costly and elaborate. Here we studied the interrelationship of LDL mass, its composition and other lipoproteins. Based on this data, we deducted information about LDL's dynamic behaviour.

Methods: Lipoprotein profiles of n = 236 participants are evaluated. Plasma was separated by sequential ultracentrifugation into VLDL, IDL, LDL and HDL. Additionally, LDL and HDL were separated into subfractions. Stepwise multiple linear regressions were used to study LDL's ApoB mass and lipid composition. Relying on these results and on causation, we constructed a mathematical model to estimate LDL's retention time.

Results: The ApoB mass in LDL correlated best among all measured parameters (including corresponding lipid compositions but using no LDL-associated parameters) with the cholesterol ester content in IDL. TG/CE ratios in LDL's subfractions were strongly correlated with the corresponding ratios in IDL and HDL. The constructed mathematical model links the TG/CE ratio of LDL and HDL to LDL's ApoB concentration and enables a good estimate of LDL's retention time in plasma.

Discussion: Relying on our statistic evaluations, we assume that i) the production of nascent LDL via IDL as well as ii) LDL's prolonged retention are mapped by the TG/CE ratio in LDL subfractions. HDL's TG/CE ratio is associated with the change in LDL's TG/CE ratio during its retention in plasma. Our mathematical model uses this information and enables-by relying on causation- a good estimation of LDL's retention time.

Fig 1. Sketch of the model.

Black solid line: exemplary trajectory of the TG/CE ratio of a hypothetical LDL particle with infinite retention time in plasma $\left({\overline{\text{LDL}}}_{\frac{\text{TG}}{\text{CE}}}\left(\text{t}\right)\right)$. The ratio decreases exponentially with rate r. Grey line: The corresponding asymptote $\left({\text{Min}}_{\frac{\text{TG}}{\text{CE}}}\right)$. Dashed line: the observed molar TG/CE ratio in LDL $\left({\text{LDL}}_{\frac{\text{TG}}{\text{CE}}}\right)$, which is the mean of the TG/CE ratio of all present LDL particles.

Fig 2. Lipoprotein characteristics.

Quartiles in NL (black), HTG (green) and HCH (blue). (A) ApoB in the LDL subfractions, (B) CH and TG in total and CH in lipoprotein fractions (C) TG/CE ratios in HDL, LDL and its subfractions. Statistical significance determined by Mann-Whitney-U test (*p < 0.05, **p < 0.01).

Fig 3. Estimating the influence of CETP on LDL’s lipid composition.

The bar represents the coefficient of determination R² of a multiple linear regression model predicting $LDL{1}_{\frac{TG}{CE}}-LDL{6}_{\frac{TG}{CE}}$ using the two explanatory variables $V1:ID{L}_{\frac{TG}{CE}}$ and V2: $HDL2b_{\frac{TG}{CE}}$ in 3 subgroups. Let r²(V1) and r²(V2) be the coefficients of determination, if only V1 or V2 are used as explanatory variable in a corresponding single linear regression, respectively. Three parts contribute to R²: V1 without V2 (white, R²-r²(V2)), V2 without V1 (black, R²-r²(V2)), and V1 and V2 intersecting (dashed, r²(V1)+ r²(V2)-R²).

Fig 4. LDL-FCR estimation.

Quartiles of the FCR estimator 1/LDL_ApoB (grey) and the model parameter $\frac{\mathrm{\mu }}{\mathrm{\mu }+r}$ (black). The model parameter is normalised to the median (dotted line) of 1/LDL_ApoB in ‘Total’).