Mapping the Human Platelet Lipidome Reveals Cytosolic Phospholipase A2 as a Regulator of Mitochondrial Bioenergetics during Activation

Abstract

Human platelets acutely increase mitochondrial energy generation following stimulation. Herein, a lipidomic circuit was uncovered whereby the substrates for this are exclusively provided by cPLA2, including multiple fatty acids and oxidized species that support energy generation via β-oxidation. This indicates that acute lipid membrane remodeling is required to support energetic demands during platelet activation. Phospholipase activity is linked to energy metabolism, revealing cPLA2 as a central regulator of both lipidomics and energy flux. Using a lipidomic approach (LipidArrays), we also estimated the total number of lipids in resting, thrombin-activated, and aspirinized platelets. Significant diversity between genetically unrelated individuals and a wealth of species was revealed. Resting platelets demonstrated ∼5,600 unique species, with only ∼50% being putatively identified. Thrombin elevated ∼900 lipids >2-fold with 86% newly appearing and 45% inhibited by aspirin supplementation, indicating COX-1 is required for major activation-dependent lipidomic fluxes. Many lipids were structurally identified. With ∼50% of the lipids being absent from databases, a major opportunity for mining lipids relevant to human health and disease is presented.

Graphical abstract

Figure 1

The Resting Human Platelet Lipidome and How It Changes upon Thrombin Activation Platelet lipid extracts from three genetically unrelated donors (D1–D3) were analyzed by LC-FTMS on the Orbitrap Elite, at 60,000 resolution, then processed using SIEVE 2.0 followed by in-house-generated software as described in Supplemental Information. (A) Venn diagram showing the number of ions in each donor lipid isolate. Common and donor specific ions are demonstrated. (B) Table showing the number of ions in each donor and mean, total and lipids common to two or more donors. Calculated from (A). (C) Predominant lipid classes in platelets. Ions were grouped according to the most predominant classification family based on classifications in databases. (D) Scatter diagrams showing elution of lipids from polar or non-polar (NP) columns, in either negative- or positive-ion mode. Lipids are color-coded according to classification from databases. To view in interactive mode, download Data S2, which allows zooming and viewing of individual accurate m/z values for all ions. Putative identifications for all are in Data S1. (E–H) The thrombin-stimulated lipidome. Platelet lipid extracts from three genetically unrelated donors (D1–D3), before and after activation for 30 min using 0.2 U/mL thrombin, were analyzed as above, showing ions that are elevated at least 2-fold. (E) Venn diagram thrombin-upregulated lipids in two or more donors. (F) Total number of lipids in each donor, shown individually (G) The effect of thrombin on basal platelet lipids. Lipids that are newly detected or elevate at least 2-fold in two or more donors are shown. Out of 5,245 lipids, only 107 were upregulated, with an additional 646 appearing only in the thrombin-activated samples. (H) Scatter diagrams showing elution of lipids from polar or NP columns, in either negative- or positive-ion mode. Putative identifications for all are in Data S1.

Figure 2

Time Course of Selected OxPL Generation by Human Platelets in Response to Thrombin Washed platelets were activated using 0.2 U/min thrombin for varying times, then lipids were extracted and analyzed using a 4000 Q-Trap as described in Experimental Procedures, and with parent to daughter transitions as listed on the chromatograms. Left: time courses of integrated areas compared to internal standard (n = 3, mean ± SEM). Middle: sample chromatograms showing almost absence of lipids at 0 min followed by upregulation from 10 to 120 min. Right: MS/MS spectra for all lipids shown. More detailed MS/MS data are available in Data S4.

Figure 3

Radar Plot Showing Relative Levels of OxPLs Generated by Platelets, Identification of Four Oxidized FAs Generated via COX-1/TXS, and Effect of Aspirin on the Platelet Lipidome (A) Radar plot showing time course of the 52 most upregulated oxPLs. The lipids are color-coded based on sn2 fatty acid, and show that monohydroxylipids of AA, 22:4, and 22:5 predominate (red, brown, and burgundy labels are on right-hand side of plot), while poly-hydroxylated FAs are less abundant. (B) MS/MS spectra of four lipids identified as 14-HNTE (n6), 14-HNTE (n3), 14-HNTrE (n6), and 14-HNDE (n6). Note a major ion, identified in (C). NR, not related (an ion from an unrelated isobaric lipid). (C) Mechanism of formation of 14-HNTE (n6) (left) and fragmentation of all four isomers (right). (D) Aspirin blocks generation of 45% of platelet lipids upregulated by thrombin. Platelet lipid extracts from three genetically unrelated donors with or without 7 days on 75 mg/day aspirin were analyzed after activation for 30 min using 0.2 U/mL thrombin, as in Figure 1. Ions noted in Figure 1G to be upregulated by at least 2-fold in two or more donors are shown. (E) Predominant lipid classes upregulated by thrombin and aspirin inhibited in platelets. Putative ions were grouped according to the most predominant classification family based on classifications in LipidMaps, LipidHome, or HMDB, as for Figure 1C.

Figure 4

Heatmaps and Network Analysis Showing the Diversity of Thrombin and Aspirin Responses for Free and Oxidized FAs Generated by Three Genetically Unrelated Donors Heatmaps were generated using full-scan FTMS data for each donor using the pheatmap package in R, and lipids are color-coded according to lipid group. Levels of treatment response are represented by a color gradient ranging from blue (decrease in response) to white (no change) to red (increase in response). Lipids are color-coded by group and clustered by similarity in overall response to the treatments. Network analysis was done using Cytoscape 3.2.1. For each donor, pairwise correlations between lipids were calculated in R. Nodes represent individual lipids, and degree (size) is the number of links. Edge thickness represents strength of correlation between nodes. Full data are available in Data S6 for all individual lipids.

Figure 5

Thrombin-Stimulated Generation of FAs, Eicosanoids, OxPLs, and Mitochondrial Oxygen Consumption Is Sensitive to cPLA₂ Inhibition, while sPLA₂ Inhibition Suppresses OxPLs Only Heatmaps were generated using the pheatmap package in R and lipids color-coded according to lipid group. Levels of treatment response are represented by a color gradient ranging from blue to red (increase in response). Lipids are color-coded by group and clustered by similarity in overall response to the treatments (A–C) Effect of PLA₂ inhibitors on generation of FAs and oxPLs by platelets. Washed platelets were incubated with inhibitors or vehicle for 15 min, then activated using 0.2 U/mL thrombin for 30 min and generation of lipids determined using LC/MS/MS as described in Supplemental Information. Inhibitors were as follows: (A) cPLA₂, 50–100 nM cPLA₂i; (B) iPLA₂, 50 nM BEL; and sPLA₂, 2 μM OOEPC. (C) Inhibition of oxPLs by PLA₂ inhibitors. oxPLs are grouped depending on the sn2 fatty acid to aid data visualization. Results are shown as heatmaps, with full data in Figures S3–S5. (D–G) cPLA₂ inhibition suppresses mitochondrial oxygen consumption, but not glycolysis. Platelets were plated on Cell-Tak-coated XF96 plates, and pre-treated with cPLA₂i (<200 nM) for 15 min prior to bioenergetic measurements. (D) Basal OCR of platelets was measured prior to injection of thrombin (0.5 U/mL), followed by 1 μg/mL oligomycin (O), 0.6 μM FCCP (F), and 10 μM antimycin A (Anti). (E) cPLA₂i (0–200 nM) dose response of thrombin-linked OCR presented as a percentage of control. (F) Indices of mitochondrial function, basal, thrombin responsive, ATP-linked, proton leak, maximal, reserve capacity, and non-mitochondrial OCR were calculated. (G) Basal and thrombin-responsive ECARs were calculated from parallel ECAR measurements. Data expressed as mean ± SEM from one representative donor, n = 3–5 replicates per sample. ^∗p < 0.05, ^∗∗p < 0.01, and ^∗∗∗p < 0.005 different from control (ANOVA and Bonferroni post hoc test).

Figure 6

Eicosanoids and Fatty Acids Acutely Feed into β-Oxidation during Platelet Activation and Support Maintenance of Membrane Phospholipid Asymmetry Heatmaps were generated using the pheatmap package in R and lipids color-coded according to lipid group. Levels of treatment response are represented by a color gradient ranging from blue (decrease in response) to white (no change) to red (increase in response). Lipids are color-coded by group and clustered by similarity in overall response to the treatments. (A) Several eicosanoids are substrates for platelet β-oxidation on thrombin generation. Washed platelets were incubated with inhibitors or vehicle at room temperature (RT), then activated using 0.2 U/mL thrombin at 37°C for 30 min and generation of lipids determined using LC-MS/MS as described in Supplemental Information (n = 3). Inhibitors were as follows: etomoxir (Eto), 25 μM; cPLA₂i, 100 nM. (B) Inhibition of glycolysis leads to a failure to generate free fatty acids and eicosanoids on thrombin activation of platelets. Washed platelets were incubated with inhibitors or vehicle at RT, then activated using 0.2 U/mL thrombin at 37°C for 30 min, with generation of lipids determined using LC-MS/MS as described in Supplemental Information (n = 3). Cells were incubated in the presence or absence of 5 mM glucose (Glc). Inhibitors were as follows: Eto, 25 μM; 2-deoxy-D-glucose (2DG), 120 mM. (C) Blocking β-oxidation leads to a failure to maintain phospholipid asymmetry in platelets. Washed platelets were incubated with inhibitors or vehicle for up to 240 min at RT, then stained for PS using Annexin V-FITC, for 15 min at RT in the dark, and analyzed by flow cytometry. Inhibitors were as follows: Eto, 25 μM; antimycin A (Anti), 10 μM; 2DG, 120 mM. Data were analyzed using one-way ANOVA with Bonferroni post hoc test (n = 3, mean ± SEM), comparing Eto with vehicle and Eto+2DG with 2DG. (D) Scheme for proposed role of cPLA₂ in acutely feeding substrates into mitochondrial oxidative phosphorylation and generation of oxPLs.