PCSK9 Activity Is Potentiated Through HDL Binding

Abstract

Rationale: Proprotein convertase subtilisin/kexin type 9 (PCSK9) circulates in a free and lipoprotein-bound form, yet the functional consequence of the association between PCSK9 and high-density lipoprotein (HDL) remains unexplored.

Objective: This study sought to interrogate the novel relationship between PCSK9 and HDL in humans.

Methods and results: Comparing lipoprotein and apolipoprotein profiles by nuclear magnetic resonance and targeted mass spectrometry measurements with PCSK9 levels in the community-based Bruneck (n=656) study revealed a positive association of plasma PCSK9 with small HDL, alongside a highly significant positive correlation between plasma levels of PCSK9 and apolipoprotein-C3, an inhibitor of lipoprotein lipase. The latter association was replicated in an independent cohort, the SAPHIR study (n=270). Thus, PCSK9-HDL association was determined during the postprandial response in two dietary studies (n=20 participants each, 8 times points). Peak triglyceride levels coincided with an attenuation of the PCSK9-HDL association, a loss of apolipoprotein-C3 from HDL and lower levels of small HDL as measured by nuclear magnetic resonance. Crosslinking mass spectrometry (XLMS) upon isolated HDL identified PCSK9 as a potential HDL-binding partner. PCSK9 association with HDL was confirmed through size-exclusion chromatography and immuno-isolation. Quantitative proteomics upon HDL isolated from patients with coronary artery disease (n=172) returned PCSK9 as a core member of the HDL proteome. Combined interrogation of the HDL proteome and lipidome revealed a distinct cluster of PCSK9, phospholipid transfer protein, clusterin and apolipoprotein-E within the HDL proteome, that was altered by sex and positively correlated with sphingomyelin content. Mechanistically, HDL facilitated PCSK9-mediated low-density lipoprotein receptor degradation and reduced low-density lipoprotein uptake through the modulation of PCSK9 internalisation and multimerisation.

Conclusions: This study reports HDL as a binder of PCSK9 and regulator of its function. The combination of -omic technologies revealed postprandial lipaemia as a driver of PCSK9 and apolipoprotein-C3 release from HDL.

Trial registration: ClinicalTrials.gov NCT03191513.

Keywords: apolipoproteins; cardiovascular diseases; coronary artery disease; lipoproteins; mass spectrometry.

Figures

Figure 1.

Integrated lipoprotein analysis in plasma.A, Nuclear magnetic resonance (NMR) lipoprotein analysis and targeted apolipoprotein profiling was conducted in the Bruneck study (n=656). Plasma PCSK9 (proprotein convertase subtilisin/kexin type 9) levels, as measured by ELISA, were correlated against NMR lipoprotein attributes including; particle number (P), lipid contents (L), phospholipids (PL), total cholesterol (C), cholesterol esters (CE), free cholesterol (FC), triglycerides (TG), and lastly, each lipid class is also represented as a percentage of total lipids (perc). The lipoprotein particles are resolved by size: extremely large (XXL); extra large (XL); large (L); medium (M); small (S); and very small (XS). Only correlations with a P<0.05 are represented. Each Spearman coefficient is displayed. B, Plasma PCSK9 levels were correlated with the apolipoprotein profiles as measured by targeted mass spectrometry with authentic heavy standards in both the Bruneck (n=656) and SAPHIR (Salzburg Atherosclerosis Prevention Program in Subjects at High Individual Risk; n=270) cohorts. Note the strong association of PCSK9 plasma levels with C apolipoproteins. P values were not adjusted for multiple testing. Overall, 26 tests were performed, for 13 apolipoproteins in each of the 2 cohorts. The Bonferroni adjusted threshold of significance (at the 0.05 level) is 0.0019. CIs are bootstrap percentile CIs based on 1000 bootstrap resamples. CLU indicates clusterin (apolipoprotein J); HDL, high-density lipoprotein; LDL, low-density lipoprotein; and VLDL, very low-density lipoprotein.

Figure 2.

HDL (high-density lipoprotein) composition during the postprandial response.A, Postprandial plasma samples were obtained from 20 healthy individuals at 8, hourly time points (INTERMET [acute effects of interesterification of commercially used fats on postprandial fat metabolism] study). Plasma PCSK9 (proprotein convertase subtilisin/kexin type 9) levels were measured, alongside triglycerides (TGs). B, A linear regression analysis between the PCSK9 and TG postprandial response is represented. C, Postprandial plasma samples from a further 20 healthy individuals at 8, hourly time points (validation cohort) were assessed for PCSK9 and HDL-TG content. D, Nuclear magnetic resonance–based lipoprotein analysis was conducted over the postprandial time course, and the particle concentration of S.HDL (small HDL), M.HDL (medium HDL), L.HDL (large HDL), and XL.HDL (extra-large HDL) is shown. E, Label-free proteomics was conducted upon HDL immuno-isolated from postprandial plasma samples (n=8, 3 time points), and significantly changing proteins over 8 h are represented as a heat map and protein clusters are shown graphically. Significance was determined using the nonparametric Friedman test with Dunn correction, *P<0.05, **P<0.005, ***P<0.0005, ****P<0.0001.

Figure 3.

Crosslinking mass spectrometry (XLMS) identifies PCSK9 (proprotein convertase subtilisin/kexin type 9) as a potential HDL (high-density lipoprotein) interaction partner.A, Ten micrograms of HDL protein (n=8) isolated by either ultracentrifugation (uc HDL) or immunocapture (immunoHDL) were directly compared for PCSK9 content by ELISA. B, Titration of the MS-cleavable crosslinker disuccinimidyl sulfoxide (DSSO) was conducted to determine optimal DSSO concentration, resulting in 1 mmol/L to be used for subsequent analyses. C, ImmunoHDL (n=8 healthy subjects, pooled) was analyzed by XLMS and identified interprotein crosslinks are represented as an interaction network. Apolipoproteins are highlighted in orange. PCSK9 is highlighted in yellow. Crosslink density is represented as gray area. D, Recombinant PCSK9 was spiked into ucHDL before XLMS analysis. PCSK9 crosslinks identified are represented for both immunoHDL and spiked ucHDL. XLMS data visualization was conducted using xiview.org. ANTXR1 indicates anthrax toxin receptor 1; C4B, complement factor 4B; FGA, fibrinogen alpha chain; ITIH1, interalpha-trypsin inhibitor heavy chain 1; LBP, lipopolysaccharide-binding protein; LIPG, endothelial lipase; and TAC3, tachykinin-3.

Figure 4.

Confirmation of PCSK9 (proprotein convertase subtilisin/kexin type 9)-HDL (high-density lipoprotein) interaction.A, Pooled human plasma from healthy subjects (n=3) was separated by size-exclusion chromatography (SEC) and PCSK9 concentrations in collected fractions were determined by ELISA. B, Apolipoprotein abundances in SEC fractions as determined by immunoblotting. C, Recombinant HIS (polyhistidine)-tagged PCSK9 (20 µg/mL) was incubated with ucHDL (ultracentrifuge-isolated HDL), LDL (low-density lipoprotein), and VLDL (very low-density lipoprotein) before SEC separation; PCSK9 alone served as control. D, An anti–apoA1 immunoprecipitation method from human plasma was validated by immunoblotting. E, Mass spectrometry analysis of the apolipoproteins in the bound fraction of the anti–apoA1 pull downs are represented as a heat map. F, Plasma samples (10 µL, n=3 healthy volunteers) were depleted of apoA1. G, PCSK9 abundances were measured by ELISA in apoA1 depleted plasma (supernatant, S/N) and in the bound fractions. Significance was determined by a paired-Student t test. A.U indicates arbitrary units; FPLC, fast protein liquid chromatography; and ve ctrl, isotype control.

Figure 5.

Integrated proteomics and lipidomics analysis of HDL (high-density lipoprotein) in patients with coronary artery disease (CAD).A, PCSK9 (proprotein convertase subtilisin/kexin type 9) distribution across lipoprotein fractions within a cohort of 172 patients with varying CAD-related phenotypes was assessed using a modified sandwich ELISA. Significance was determined using the Kruskal-Wallis test across the groups. B, ucHDL (ultracentrifuge-isolated HDL) from these patients with CAD (n=191, including samples from patients with 6 mo follow-up) was analyzed by quantitative proteomics. The coefficients of variation in HDL protein abundances, as measured by label-free mass spectrometry (MS), were calculated across the whole cohort. C, PCSK9 protein correlations against the core HDL proteome are represented as a volcano plot. D, Quantitation by label-free and tandem-mass tag (TMT) proteomics upon ucHDL revealed proteins that were altered by sex (males, n=98. females n=66). Significant proteins in at least one method are labeled; fold changes across methodologies were compared using a linear regression analysis. E, Three hundred sixty-five lipid species were quantified in HDL by targeted MS with reference standards. A Spearman correlation matrix was generated between the sum of each lipid species in a respective class and the HDL apolipoprotein profile, as well as PCSK9, PLTP, and CLU. A hierarchical cluster analysis is represented as a heat map. F, Crosslinking mass spectrometry (XLMS) analysis of immunoHDL (immuno-isolated HDL) and ucHDL revealed a strong protein-protein interaction between apoA2 and apoC1, green lines represent crosslinks (xiview.org). AC indicates acylcarnitines; ANTXR1, anthrax toxin receptor 1; C4B, complement factor 4B; CE, cholesterol esters; CER, ceramides; CLU, clusterin; DG, diacylglycerides; LBP, lipopolysaccharide-binding protein; LPC, lyso-phosphatidylcholine; PC, phosphatidylcholine; PCYOX, prenylcysteine oxidase; PLTP, phospholipid transfer protein; PON, paraoxonase; RLU, relative light units; SAA, serum amyloid A; SFTPB, surfactant protein B; SM, sphingomyelins; and TG, triglycerides.

Figure 6.

HDL (high-density lipoprotein) potentiates PCSK9 (proprotein convertase subtilisin/kexin type 9) uptake and multimerization.A, PCSK9 and rHDL (reconstituted HDL) were coincubated before HDL immuno-isolation to demonstrate the interaction between rHDL and PCSK9. PCSK9 alone was run through the HDL immunoisolation column as a negative control. B, HepG2 cells were treated with HIS (polyhistidine)-tagged PCSK9 (5 µg/mL), rHDL, or ucHDL (ultracentrifuge-isolated HDL; 25 µg/mL) or a combination of rHDL or ucHDL and HIS-tagged PCSK9 for 6 h before immunoblot analysis for PCSK9 cellular uptake. C, Densitometry analysis of 3 independent replicates. Significance was determined using an independent t test. D, HepG2 cells were treated with HIS-tagged PCSK9 (1 µg/mL), in the presence of increasing amounts of ucLDL (ultracentrifuge-isolated low-density lipoprotein; 0–50 µg/mL). The control lane (Ctrl) represents PCSK9 only. E, Recombinant HIS-tagged PCSK9 at a concentration of 1 µg/mL was incubated in the presence of an increasing concentration of rHDL or ucHDL. Immunoblot analysis was then conducted, with equal amounts of PCSK9 loading for each sample. The control lane represent PCSK9 alone incubated at 4°C. Total protein stain is used to visualize apoA1. LLOD indicates lower limit of detection.

Figure 7.

HDL (high-density lipoprotein) facilitates PCSK9 (proprotein convertase subtilisin/kexin type 9)-mediated LDLR (low-density lipoprotein receptor) degradation.A, HepG2 cells were treated with HIS (polyhistidine)-tagged PCSK9 (1 µg/mL), ucHDL (ultracentrifuge-isolated HDL; 50 µg/mL), or a combination of ucHDL and HIS-tagged PCSK9 for 6 h in the presence of actinomycin D (5 µg/mL), before immunoblot analysis. B, Densitometry analysis of 3 independent replicates. C, The same coincubation experiment was repeated by isolating the membrane protein fraction through cell surface biotinylation and NeutrAvidin agarose enrichment. D, HepG2 cells were treated with HIS-tagged PCSK9 (1 µg/mL), ucHDL (50 µg/mL), or a combination of ucHDL and HIS-tagged PCSK9 for 6 h in the presence of actinomycin D, before the addition of bodipy-LDL (12 µg/mL) for 3 h. Representative fluorescent images for each condition are represented, LDL uptake can be seen in red and a nuclear counterstain in blue (NucBlue), scale bar represents 100 µm. Quantitative analysis of 10 cellular areas for each condition was conducted in ImageJ, summed particle intensities per area were taken, normalized to cell nuclei count. Data is representative of 3 independent experiments. E, Cell surface proteins were isolated from HepG2 cells treated with HIS-tagged PCSK9 (1 µg/mL) or ucHDL (50 µg/mL) for 6 h in the presence of actinomycin D before label-free quantification by mass spectrometry (MS). Proteins with cell surface localization were retained and changes across conditions are represented in volcano plots. Significance was determined using a t test with Welch correction. APP indicates amyloid precursor protein; CTRL, control; SDC, syndecan; and TfR, transferrin receptor.