Apolipoprotein E content of VLDL limits LPL-mediated triglyceride hydrolysis

Fig. 1. LPL-mediated TG hydrolysis in VLDL fractions in whole plasma separated by size-exclusion chromatography. Citrated plasma (350 μl) was incubated with increasing amounts of recombinant human LPL in vitro and then separated by fast protein liquid chromatography (FPLC) over a Superose 6 gel filtration column in PBS as described in the Materials and methods section. About 300 μl fractions were collected, and TG was quantified colorimetrically. VLDL eluted within fractions 15–23 and LDL eluted in fractions 25–33 in this system. Orlistat, a lipase inhibitor, was used as a control and to quench the reactions. Samples include whole plasma without exogenous LPL (closed circle), whole plasma with 1.5 μg/ml LPL and 50 μM Orlistat (open circle), plasma with 0.37 μg/ml LPL (closed diamond), plasma with 0.75 μg/ml LPL (open downward triangle), plasma with 1.0 μg/ml LPL (closed upward triangle), and plasma with 1.5 μg/ml LPL (closed square). Experiments were performed in triplicate with representative traces shown.

Fig. 2. Comparison of rapid microplate assay to size-exclusion chromatography assay of LPL activity. Human VLDL isolated by ultracentrifugation from a single donor at 20 mg/dl (protein) in LPDP was incubated with increasing amounts of human LPL for 30 min at 37°C. The samples were then analyzed by size-exclusion chromatography on a Superose 6 column (as for Fig. 1) or put through the 96-well desalting plate protocol described in the Materials and methods section. A: Percent of TG hydrolysis by LPL as analyzed by both methods. B: Correlation plot of TG hydrolysis as measured by the microplate and size-exclusion methods. The line shows a simple linear regression (r2 = 0.9708). All experiments were performed in triplicate, and error bars indicate one sample standard deviation.

Fig. 3. Variation in LPL hydrolysis of ultracentrifugally isolated VLDL from nine donors in cohort 1. A: VLDL was isolated by ultracentrifugation from nine apparently healthy volunteers at T = 0 days, then again at T = 30 days. Compared at equal TG mass (20 mg/dl), the VLDL particles from each donor were evaluated as substrates for a fixed amount (0.262 μg) of human LPL. The percent of TG hydrolysis is plotted for each donor at each time point. B: We also compared the VLDL samples at equal protein concentrations (20 mg/dl). Data are shown from the T = 0 point. A one-way ANOVA indicated significant differences exist among the samples (P < 0.001). C: Correlation between TG/protein ratio of each VLDL particle and LPL hydrolysis. The line shows a simple linear regression through all data shown in panel (A), r2 = 0.6906. For all panels, experiments were performed in triplicate, and error bars show one sample standard deviation. For all panels, filled circle, donor 1; open circle, donor 2; filled down triangle, donor 3; open up triangle, donor 4; filled square, donor 5; open square, donor 6; closed diamond, donor 7; open diamond, donor 8; and filled up triangle, donor 9.

Fig. 4. SDS gel electrophoresis of VLDL particles isolated from cohort 1. Ultracentrifugally isolated VLDL obtained from each donor in Fig. 3 was loaded by equal protein on a 4–15% SDS gradient gel and run at 200 V for 30 min under reducing conditions followed by Coomassie staining. The lane numbers correspond to the donor who provided the particles. Expected protein molecular weights for common VLDL: apoB100 (MW, 480 kDa), APOE (MW, 34 kDa), and the apoCs (MW, between 6.6 and 8.8).

Fig. 5. MS-based proteomic analysis of ultracentrifugally isolated VLDL from each donor in cohort 1. About 50 μg of total VLDL protein from each donor was analyzed by MS after delipidation, reduction, alkylation, and digestion with trypsin as described in the Materials and methods section. Data are expressed as the label-free quantitation intensity as calculated by MaxQuant. The plotted data are an average of three separate runs and normalized to apoB concentration.

Fig. 6. Correlation of VLDL proteome abundance and TG hydrolysis by LPL among donors. A: A Pearson correlation analysis was performed between the percent of TG hydrolysis in the presence of human LPL and the MS spectral counts for each detectable VLDL proteome member. The correlation coefficients are listed for the top nine most abundant proteins detected. B: Correlation between percent of TG hydrolysis and APOC2 spectral count. The line shows a simple linear regression (r2 = 0.157). C: Correlation between percent of TG hydrolysis and APOC1 spectral count. The line shows a simple linear regression (r2 = 0.005). D: Correlation between percent of TG hydrolysis and APOE spectral count. The line shows a simple linear regression (r2 = 0.949). The data are an average of three separate runs and normalized to apoB concentration from the T = 0 time point.

Fig. 7. Validation of the APOE effect on LPL-mediated VLDL TG hydrolysis in a second cohort of subjects. A: A second group of six subjects was recruited (cohort 2). Instead of isolating VLDL by ultracentrifugation as in cohort 1, each donor plasma was separated by size-exclusion chromatography. The traces, as tracked by TG, are shown for each donor (see the legend). In this system, VLDL elutes between fractions 15 and 20, LDL is centered on fraction 25, and HDL elutes after about fraction 32. B: TG hydrolysis of pooled VLDL fractions assayed at equal TG concentration as in Fig. 3 (n = 3, error bars show one sample standard deviation). A one-way ANOVA indicated that significant differences existed between the samples (P < 0.001). C: Correlation of TG hydrolysis versus concentration of APOE measured by ELISA and normalized to APOB content (also by ELISA). Line shows a simple linear regression (r2 = 0.889).

Fig. 8. The effect of adding exogenous APOE on VLDL TG hydrolysis by LPL. A: VLDL (20 mg/dl of total protein) was incubated alone, with 0.05 mg/ml human plasma-purified APOA1, or with 0.05 mg/ml human plasma APOE for 1 h at 37°C prior to performing an LPL TG hydrolysis assay. A one-way ANOVA indicated difference exist among the samples (P < 0.001), and the ∗ shows a difference from VLDL only by Tukey test at P < 0.05. B: A similar experiment to that in panel (A) except that amount of human plasma APOE was systematically varied as shown. The same source of VLDL was used for each experiment, which were performed in triplicate with error bars representing one sample standard deviation. A one-way ANOVA indicated that differences exist among the samples (P < 0.001), and the ∗ shows a difference from VLDL with no APOE addition by Tukey test at P < 0.05.

Fig. 9. Effect of APOE isoforms on VLDL TG hydrolysis by LPL. The experiment was performed as in Fig. 8 except that 25 μg/ml of either human plasma APOE or recombinant APOE2, APOE3, or APOE4 were incubated with the VLDL particles (total protein of 20 mg/dl) for 1 h at 37°C prior to performing an LPL TG hydrolysis assay. The same source of VLDL was used for each experiment, which were performed in triplicate with error bars representing one sample standard deviation. P < 0.001 by one-way ANOVA. ∗ indicates difference from VLDL alone, P < 0.05 by Tukey test.