LDL subclass lipidomics in atherogenic dyslipidemia: effect of statin therapy on bioactive lipids and dense LDL

Abstract

Atherogenic LDL particles are physicochemically and metabolically heterogeneous. Can bioactive lipid cargo differentiate LDL subclasses, and thus potential atherogenicity? What is the effect of statin treatment? Obese hypertriglyceridemic hypercholesterolemic males [n = 12; lipoprotein (a) <10 mg/dl] received pitavastatin calcium (4 mg/day) for 180 days in a single-phase unblinded study. The lipidomic profiles (23 lipid classes) of five LDL subclasses fractionated from baseline and post-statin plasmas were determined by LC-MS. At baseline and on statin treatment, very small dense LDL (LDL5) was preferentially enriched (up to 3-fold) in specific lysophospholipids {LPC, lysophosphatidylinositol (LPI), lysoalkylphosphatidylcholine [LPC(O)]; 9, 0.2, and 0.14 mol per mole of apoB, respectively; all P < 0.001 vs. LDL1-4}, suggesting elevated inflammatory potential per particle. In contrast, lysophosphatidylethanolamine was uniformly distributed among LDL subclasses. Statin treatment markedly reduced absolute plasma concentrations of all LDL subclasses (up to 33.5%), including LPC, LPI, and LPC(O) contents (up to -52%), consistent with reduction in cardiovascular risk. Despite such reductions, lipotoxic ceramide load per particle in LDL1-5 (1.5-3 mol per mole of apoB; 3-7 mmol per mole of PC) was either conserved or elevated. Bioactive lipids may constitute biomarkers for the cardiometabolic risk associated with specific LDL subclasses in atherogenic dyslipidemia at baseline, and with residual risk on statin therapy.

Trial registration: ClinicalTrials.gov NCT01595828.

Keywords: ceramides; isopycnic density gradient ultracentrifugation; lipoprotein-associated phospholipase A2; liquid chromatography electrospray ionization-tandem mass spectrometry; low density lipoprotein; low density lipoprotein subclass heterogeneity; lysophosphatidylcholine; metabolic syndrome; pitavastatin calcium.

Conflict of interest statement

M.J.C. has received research funding and/or modest honoraria for consultancy from Amarin, Amgen, AstraZeneca, Kowa, MSD, Sanofi, Regeneron, and Pfizer. All other authors declare that they have no conflicts of interest with the contents of this article

Copyright © 2020 Chapman et al.

Figures

Fig. 1.

Plasma concentrations of individual lipid classes across the apoB-containing lipoprotein subfractions (VLDL+IDL, LDL1-5) at baseline (D0), and the effect of pitavastatin calcium treatment (4 mg/day) for 180 days (D180) in obese mixed dyslipidemic male MetS subjects. Values are expressed as mean ± SEM (n = 12) in picomoles of each lipid class per milliliter of plasma; absolute values are presented in supplemental Table S1. Density ranges: VLDL+IDL P < 0.001; **0.001 < P < 0.01; and *0.01 < P < 0.05 versus D0. A–V: COH (A); CE (B); TAG (C); DAG (D); SM (E); PC (F); PC(P) (plasmalogen) (G); PC(O) (H); LPC (I); LPC(O) (J); PE (K); PE(P) (plasmalogen) (L); PE(O) (M); LPE (N); PI (O); LPI (P); Cer (Q); dhCer (R); MHC (S); DHC (T); THC (U); GM3 (V).

Fig. 1.

Plasma concentrations of individual lipid classes across the apoB-containing lipoprotein subfractions (VLDL+IDL, LDL1-5) at baseline (D0), and the effect of pitavastatin calcium treatment (4 mg/day) for 180 days (D180) in obese mixed dyslipidemic male MetS subjects. Values are expressed as mean ± SEM (n = 12) in picomoles of each lipid class per milliliter of plasma; absolute values are presented in supplemental Table S1. Density ranges: VLDL+IDL P < 0.001; **0.001 < P < 0.01; and *0.01 < P < 0.05 versus D0. A–V: COH (A); CE (B); TAG (C); DAG (D); SM (E); PC (F); PC(P) (plasmalogen) (G); PC(O) (H); LPC (I); LPC(O) (J); PE (K); PE(P) (plasmalogen) (L); PE(O) (M); LPE (N); PI (O); LPI (P); Cer (Q); dhCer (R); MHC (S); DHC (T); THC (U); GM3 (V).

Fig. 1.

Plasma concentrations of individual lipid classes across the apoB-containing lipoprotein subfractions (VLDL+IDL, LDL1-5) at baseline (D0), and the effect of pitavastatin calcium treatment (4 mg/day) for 180 days (D180) in obese mixed dyslipidemic male MetS subjects. Values are expressed as mean ± SEM (n = 12) in picomoles of each lipid class per milliliter of plasma; absolute values are presented in supplemental Table S1. Density ranges: VLDL+IDL P < 0.001; **0.001 < P < 0.01; and *0.01 < P < 0.05 versus D0. A–V: COH (A); CE (B); TAG (C); DAG (D); SM (E); PC (F); PC(P) (plasmalogen) (G); PC(O) (H); LPC (I); LPC(O) (J); PE (K); PE(P) (plasmalogen) (L); PE(O) (M); LPE (N); PI (O); LPI (P); Cer (Q); dhCer (R); MHC (S); DHC (T); THC (U); GM3 (V).

Fig. 1.

Plasma concentrations of individual lipid classes across the apoB-containing lipoprotein subfractions (VLDL+IDL, LDL1-5) at baseline (D0), and the effect of pitavastatin calcium treatment (4 mg/day) for 180 days (D180) in obese mixed dyslipidemic male MetS subjects. Values are expressed as mean ± SEM (n = 12) in picomoles of each lipid class per milliliter of plasma; absolute values are presented in supplemental Table S1. Density ranges: VLDL+IDL P < 0.001; **0.001 < P < 0.01; and *0.01 < P < 0.05 versus D0. A–V: COH (A); CE (B); TAG (C); DAG (D); SM (E); PC (F); PC(P) (plasmalogen) (G); PC(O) (H); LPC (I); LPC(O) (J); PE (K); PE(P) (plasmalogen) (L); PE(O) (M); LPE (N); PI (O); LPI (P); Cer (Q); dhCer (R); MHC (S); DHC (T); THC (U); GM3 (V).

Fig. 2.

Molar ratios of individual lipid classes normalized to moles of apoB in plasma VLDL+IDL, LDL1, LDL2, LDL3, LDL4, and LDL5 subfractions from obese mixed dyslipidemic male MetS subjects at baseline (D0) and after pitavastatin calcium treatment [4 mg/day for 180 days (D180)]. Values are expressed as mean ± SEM (n = 12) in picomoles of each lipid class per picomole of apoB. Percent change (%) was calculated relative to baseline values. ***P < 0.001; **0.001 < P < 0.01; and *0.01 < P < 0.05 versus D0. Density ranges: VLDL+IDL <1.019 g/ml, LDL1 = 1.019–1.023 g/ml, LDL2 = 1.023–1.029 g/ml, LDL3 = 1.029–1.039 g/ml, LDL4 = 1.039–1.050 g/ml, and LDL5 = 1.050–1.063 g/ml. Lines and square symbols in blue represent data at baseline, D0; red lines and triangular symbols represent data following 180 days of pitavastatin treatment (D180). A–V: COH (A); CE (B); TAG (C); DAG (D); SM (E); PC (F); PC(P) (plasmalogen) (G); PC(O) (H); LPC (I); LPC(O) (J); PE (K); PE(P) (plasmalogen) (L); PE(O) (M); LPE (N); PI (O); LPI (P); Cer (Q); dhCer (R); MHC (S); DHC (T); THC (U); GM3 (V).

Fig. 2.

Molar ratios of individual lipid classes normalized to moles of apoB in plasma VLDL+IDL, LDL1, LDL2, LDL3, LDL4, and LDL5 subfractions from obese mixed dyslipidemic male MetS subjects at baseline (D0) and after pitavastatin calcium treatment [4 mg/day for 180 days (D180)]. Values are expressed as mean ± SEM (n = 12) in picomoles of each lipid class per picomole of apoB. Percent change (%) was calculated relative to baseline values. ***P < 0.001; **0.001 < P < 0.01; and *0.01 < P < 0.05 versus D0. Density ranges: VLDL+IDL <1.019 g/ml, LDL1 = 1.019–1.023 g/ml, LDL2 = 1.023–1.029 g/ml, LDL3 = 1.029–1.039 g/ml, LDL4 = 1.039–1.050 g/ml, and LDL5 = 1.050–1.063 g/ml. Lines and square symbols in blue represent data at baseline, D0; red lines and triangular symbols represent data following 180 days of pitavastatin treatment (D180). A–V: COH (A); CE (B); TAG (C); DAG (D); SM (E); PC (F); PC(P) (plasmalogen) (G); PC(O) (H); LPC (I); LPC(O) (J); PE (K); PE(P) (plasmalogen) (L); PE(O) (M); LPE (N); PI (O); LPI (P); Cer (Q); dhCer (R); MHC (S); DHC (T); THC (U); GM3 (V).

Fig. 2.

Molar ratios of individual lipid classes normalized to moles of apoB in plasma VLDL+IDL, LDL1, LDL2, LDL3, LDL4, and LDL5 subfractions from obese mixed dyslipidemic male MetS subjects at baseline (D0) and after pitavastatin calcium treatment [4 mg/day for 180 days (D180)]. Values are expressed as mean ± SEM (n = 12) in picomoles of each lipid class per picomole of apoB. Percent change (%) was calculated relative to baseline values. ***P < 0.001; **0.001 < P < 0.01; and *0.01 < P < 0.05 versus D0. Density ranges: VLDL+IDL <1.019 g/ml, LDL1 = 1.019–1.023 g/ml, LDL2 = 1.023–1.029 g/ml, LDL3 = 1.029–1.039 g/ml, LDL4 = 1.039–1.050 g/ml, and LDL5 = 1.050–1.063 g/ml. Lines and square symbols in blue represent data at baseline, D0; red lines and triangular symbols represent data following 180 days of pitavastatin treatment (D180). A–V: COH (A); CE (B); TAG (C); DAG (D); SM (E); PC (F); PC(P) (plasmalogen) (G); PC(O) (H); LPC (I); LPC(O) (J); PE (K); PE(P) (plasmalogen) (L); PE(O) (M); LPE (N); PI (O); LPI (P); Cer (Q); dhCer (R); MHC (S); DHC (T); THC (U); GM3 (V).

Fig. 2.

Molar ratios of individual lipid classes normalized to moles of apoB in plasma VLDL+IDL, LDL1, LDL2, LDL3, LDL4, and LDL5 subfractions from obese mixed dyslipidemic male MetS subjects at baseline (D0) and after pitavastatin calcium treatment [4 mg/day for 180 days (D180)]. Values are expressed as mean ± SEM (n = 12) in picomoles of each lipid class per picomole of apoB. Percent change (%) was calculated relative to baseline values. ***P < 0.001; **0.001 < P < 0.01; and *0.01 < P < 0.05 versus D0. Density ranges: VLDL+IDL <1.019 g/ml, LDL1 = 1.019–1.023 g/ml, LDL2 = 1.023–1.029 g/ml, LDL3 = 1.029–1.039 g/ml, LDL4 = 1.039–1.050 g/ml, and LDL5 = 1.050–1.063 g/ml. Lines and square symbols in blue represent data at baseline, D0; red lines and triangular symbols represent data following 180 days of pitavastatin treatment (D180). A–V: COH (A); CE (B); TAG (C); DAG (D); SM (E); PC (F); PC(P) (plasmalogen) (G); PC(O) (H); LPC (I); LPC(O) (J); PE (K); PE(P) (plasmalogen) (L); PE(O) (M); LPE (N); PI (O); LPI (P); Cer (Q); dhCer (R); MHC (S); DHC (T); THC (U); GM3 (V).

Fig. 3.

Molar ratios of individual lipid classes normalized to moles of PC in VLDL+IDL, LDL1, LDL2, LDL3, LDL4, and LDL5 subfractions from obese mixed dyslipidemic male MetS subjects at baseline (D0) and after pitavastatin calcium treatment [4 mg/day for 180 days (D180)]. Values are expressed as mean ± SEM (n = 12) in femtomoles of lipid class per picomole of PC. Percent change (%) was calculated relative to baseline values (D0). ***P < 0.001; **0.001 < P < 0.01; and *0.01;<;P < 0.05 versus D0. Density ranges: VLDL+IDL <1.019 g/ml, LDL1 = 1.019–1.023 g/ml, LDL2 = 1.023–1.029 g/ml, LDL3 = 1.029–1.039 g/ml, LDL4 = 1.039–1.050 g/ml, and LDL5 = 1.050–1.063 g/ml. A–U: COH (A); CE (B); TAG (C); DAG (D); SM (E); PC(P) (plasmalogen) (F); PC(O) (G); LPC (H); LPC(O), lysoalkylphosphtidylcholine (I); PE (J); PE(P) (plasmalogen) (K); PE(O) (L); LPE (M); PI (N); LPI (O); Cer (P); dhCer (Q); MHC (R); DHC (S); THC (T); GM3 (U).

Fig. 3.

Molar ratios of individual lipid classes normalized to moles of PC in VLDL+IDL, LDL1, LDL2, LDL3, LDL4, and LDL5 subfractions from obese mixed dyslipidemic male MetS subjects at baseline (D0) and after pitavastatin calcium treatment [4 mg/day for 180 days (D180)]. Values are expressed as mean ± SEM (n = 12) in femtomoles of lipid class per picomole of PC. Percent change (%) was calculated relative to baseline values (D0). ***P < 0.001; **0.001 < P < 0.01; and *0.01;<;P < 0.05 versus D0. Density ranges: VLDL+IDL <1.019 g/ml, LDL1 = 1.019–1.023 g/ml, LDL2 = 1.023–1.029 g/ml, LDL3 = 1.029–1.039 g/ml, LDL4 = 1.039–1.050 g/ml, and LDL5 = 1.050–1.063 g/ml. A–U: COH (A); CE (B); TAG (C); DAG (D); SM (E); PC(P) (plasmalogen) (F); PC(O) (G); LPC (H); LPC(O), lysoalkylphosphtidylcholine (I); PE (J); PE(P) (plasmalogen) (K); PE(O) (L); LPE (M); PI (N); LPI (O); Cer (P); dhCer (Q); MHC (R); DHC (S); THC (T); GM3 (U).

Fig. 3.

Molar ratios of individual lipid classes normalized to moles of PC in VLDL+IDL, LDL1, LDL2, LDL3, LDL4, and LDL5 subfractions from obese mixed dyslipidemic male MetS subjects at baseline (D0) and after pitavastatin calcium treatment [4 mg/day for 180 days (D180)]. Values are expressed as mean ± SEM (n = 12) in femtomoles of lipid class per picomole of PC. Percent change (%) was calculated relative to baseline values (D0). ***P < 0.001; **0.001 < P < 0.01; and *0.01;<;P < 0.05 versus D0. Density ranges: VLDL+IDL <1.019 g/ml, LDL1 = 1.019–1.023 g/ml, LDL2 = 1.023–1.029 g/ml, LDL3 = 1.029–1.039 g/ml, LDL4 = 1.039–1.050 g/ml, and LDL5 = 1.050–1.063 g/ml. A–U: COH (A); CE (B); TAG (C); DAG (D); SM (E); PC(P) (plasmalogen) (F); PC(O) (G); LPC (H); LPC(O), lysoalkylphosphtidylcholine (I); PE (J); PE(P) (plasmalogen) (K); PE(O) (L); LPE (M); PI (N); LPI (O); Cer (P); dhCer (Q); MHC (R); DHC (S); THC (T); GM3 (U).

Fig. 3.

Molar ratios of individual lipid classes normalized to moles of PC in VLDL+IDL, LDL1, LDL2, LDL3, LDL4, and LDL5 subfractions from obese mixed dyslipidemic male MetS subjects at baseline (D0) and after pitavastatin calcium treatment [4 mg/day for 180 days (D180)]. Values are expressed as mean ± SEM (n = 12) in femtomoles of lipid class per picomole of PC. Percent change (%) was calculated relative to baseline values (D0). ***P < 0.001; **0.001 < P < 0.01; and *0.01;<;P < 0.05 versus D0. Density ranges: VLDL+IDL <1.019 g/ml, LDL1 = 1.019–1.023 g/ml, LDL2 = 1.023–1.029 g/ml, LDL3 = 1.029–1.039 g/ml, LDL4 = 1.039–1.050 g/ml, and LDL5 = 1.050–1.063 g/ml. A–U: COH (A); CE (B); TAG (C); DAG (D); SM (E); PC(P) (plasmalogen) (F); PC(O) (G); LPC (H); LPC(O), lysoalkylphosphtidylcholine (I); PE (J); PE(P) (plasmalogen) (K); PE(O) (L); LPE (M); PI (N); LPI (O); Cer (P); dhCer (Q); MHC (R); DHC (S); THC (T); GM3 (U).

Fig. 4.

Lp-PLA2 activity and mass across VLDL+IDL and LDL1-5 subfractions as a function of density at baseline and effect of pitavastatin calcium treatment [4 mg/day for 180 days (D180)]. Lipoprotein subfractions were isolated from plasma as indicated in the Materials and Methods section, and LpPLA2 mass (expressed as nanograms per milliliter; n = 3; for all data points for apoB-containing lipoproteins at D0 and D180, the mean value for SEM was 6.1 IU/l) and activity (expressed as IU/l; n = 4; for all data points for apoB-containing lipoproteins at D0 and D180, the mean value for SEM was 2.8 IU/l; see the Materials and Methods). Blue line and blue diamond symbols, Lp-PLA2 mass at D0; red line and red square symbols, Lp-PLA2 mass at D180; green line and green triangular symbols, Lp-PLA2 activity at D0, and violet line and violet cross symbols, Lp-PLA2 activity at D180.

Fig. 5.

Arterial entry and retention of vsdLDL5: relevance to atherogenesis. The small particle size of vsdLDL particles (LDL5) favors enhanced arterial wall entry by endothelial transcytosis (13, 21, 51). These particles are preferentially enriched in three bioactive lysophospholipids: LPC, 9 mol per LDL5 particle; LPC(O), i.e., lyso-PAF, 1 mol per seven LDL5 particles; and LPI, 1 mol per six LDL5 particles, and equally in Lp-PLA2. sdLDLs display elevated binding affinity for proteoglycan components of the extracellular matrix, favoring enhanced intimal retention (21). Complexes of dense LDL5 particles with proteoglycans (PG) and glycosaminoglycans (GAG) are avidly taken up by human monocyte-derived macrophages (HMDMs) (52). Systemic oxidative stress is a characteristic of dyslipidemic MetS subjects (53); LDL5 lipids are highly susceptible to oxidative modification (21). Oxidative modification leads to scavenger receptor uptake in HMDMs with conversion to pro-inflammatory prothrombotic foam cells, key components of atherosclerotic plaques (1, 21). Hydrophilic lysophospholipids [LPC, LPC(O), LPI] may diffuse out from LDL5 particles into the aqueous phase in arterial tissue. LPC can exert a spectrum of biological effects via two pathways: the indirect pathway involving autotaxin-mediated hydrolysis of LPC with formation of lysophosphatidic acid (LPA), a potent cell signaling mediator acting through multiple G-coupled receptors (32, 54); and the direct pathway involving i) stimulation of biglycan expression and PG-chain elongation in vascular smooth muscle cells (VSMCs) with induction of an osteogenic phenotype in these same cells (55, 56), ii) stimulation of monocyte chemoattractant protein-1 (MCP-1) expression in both VSMCs and vascular endothelial cells (ECs) (32), iii) stimulation of VSMC migration (32), iv) induction of adhesion protein expression in ECs and VSMCs (32), v) stimulation of release of arachidonic acid from ECs via the PKC pathway (32), and vi) induction of the production of IL-1β in HMDMs (57). Together these effects of LDL5-associated LPC indicate a key role for these particles in the acute and chronic inflammatory dimensions of atherogenesis (21, 32, 55, 58). By contrast, LPC suppresses tissue factor expression in human monocytes (59). The degree to which LDL5-associated lyso-PAF may be transformed to active pro-inflammatory PAF depends largely on the potential for its acylation by HMDMs in situ (60). Finally, the LPI/GPR55 axis may initiate a wide range of cellular responses, including cytokine and chemokine secretion, cell proliferation and migration, and platelet aggregation (61, 62). Considered together, these findings indicate that the enhanced intimal entry, retention, and oxidation of lysophospholipid-laden LDL5 particles may exert a spectrum of biological actions in arterial tissue, which together favor accelerated atherogenesis in dyslipidemic individuals.