Diacylglycerols and Lysophosphatidic Acid, Enriched on Lipoprotein(a), Contribute to Monocyte Inflammation

Abstract

Background: Oxidized phospholipids play a key role in the atherogenic potential of lipoprotein(a) (Lp[a]); however, Lp(a) is a complex particle that warrants research into additional proinflammatory mediators. We hypothesized that additional Lp(a)-associated lipids contribute to the atherogenicity of Lp(a).

Methods: Untargeted lipidomics was performed on plasma and isolated lipoprotein fractions. The atherogenicity of the observed Lp(a)-associated lipids was tested ex vivo in primary human monocytes by RNA sequencing, ELISA, Western blot, and transendothelial migratory assays. Using immunofluorescence staining and single-cell RNA sequencing, the phenotype of macrophages was investigated in human atherosclerotic lesions.

Results: Compared with healthy individuals with low/normal Lp(a) levels (median, 7 mg/dL [18 nmol/L]; n=13), individuals with elevated Lp(a) levels (median, 87 mg/dL [218 nmol/L]; n=12) demonstrated an increase in lipid species, particularly diacylglycerols (DGs) and lysophosphatidic acid (LPA). DG and the LPA precursor lysophosphatidylcholine were enriched in the Lp(a) fraction. Ex vivo stimulation with DG(40:6) demonstrated a significant upregulation in proinflammatory pathways related to leukocyte migration, chemotaxis, NF-κB (nuclear factor kappa B) signaling, and cytokine production. Functional assessment showed a dose-dependent increase in the secretion of IL (interleukin)-6, IL-8, and IL-1β after DG(40:6) and DG(38:4) stimulation, which was, in part, mediated via the NLRP3 (NOD [nucleotide-binding oligomerization domain]-like receptor family pyrin domain containing 3) inflammasome. Conversely, LPA-stimulated monocytes did not exhibit an inflammatory phenotype. Furthermore, activation of monocytes by DGs and LPA increased their transendothelial migratory capacity. Human atherosclerotic plaques from patients with high Lp(a) levels demonstrated colocalization of Lp(a) with M1 macrophages, and an enrichment of CD68⁺IL-18⁺TLR4⁺ (toll-like receptor) TREM2⁺ (triggering receptor expressed on myeloid cells) resident macrophages and CD68⁺CASP1⁺ (caspase) IL-1B⁺SELL⁺ (selectin L) inflammatory macrophages compared with patients with low Lp(a). Finally, potent Lp(a)-lowering treatment (pelacarsen) resulted in a reduction in specific circulating DG lipid subspecies in patients with cardiovascular disease with elevated Lp(a) levels (median, 82 mg/dL [205 nmol/L]).

Conclusions: Lp(a)-associated DGs and LPA have a potential role in Lp(a)-induced monocyte inflammation by increasing cytokine secretion and monocyte transendothelial migration. This DG-induced inflammation is, in part, NLRP3 inflammasome dependent.

Keywords: diglycerides; inflammation; lipidomics; lipoprotein(a); monocytes.

Figure 1.

Blocking oxidized phospholipid (OxPL)–lipoprotein(a) (Lp[a]) results in residual IL (interleukin)-8 secretion in monocytes. CD14⁺ monocytes were stimulated 100 mg/dL Lp(a) in the presence or absence of E06 and Lp(a) for 24 hours. A, Secretion of IL-8 in the cell medium (n=8). P<0.0001 for unstimulated (Unstim) vs Lp(a) and P=0.0063 for Lp(a) vs Lp(a)+E06. B, IL-1β secretion in the cell medium (n=9). P=0.0416 for Unstim vs Lp(a) and P>0.9999 for Lp(a) vs Lp(a)+E06. C, Secretion of MCP-1 (monocyte chemoattractant protein-1) in the cell medium (n=6). No statistically significant difference between the conditions. Grubbs outlier tests were performed to identify outliers. This was followed by a Shapiro-Wilk test to test for normality. When data followed a normal distribution, an ordinary 1-way ANOVA was performed followed by Dunnett multiple comparisons test. When the data were not normally distributed, a Kruskal-Wallis test was performed followed by a Dunn multiple comparisons test (IL-1β secretion). *P<0.05, **P<0.01, ****P<0.0001.

Figure 2.

Subjects with elevated lipoprotein(a) (Lp[a]) exhibit a distinct lipidome characterized by increased levels of diacylglycerols (DGs) and lysophosphatidic acid (LPA). Boxplots demonstrating the total value of DG, LPA (A), phosphatidic acid (PA), and lysoPA/PA (B) ratio in healthy individuals with high levels of Lp(a) (orange, n=12) and low levels of Lp(a) (blue, n=13). Total value is the sum of the relative abundance of all identified lipid species. A Student t test with post hoc Bonferroni correction was used for statistical comparison between the 2 groups. Whiskers: Tukey (C) box plots of differentially expressed genes involved in leukocyte migration, T-cell activation, inflammation, and lipid droplet formation in monocytes from healthy individuals with high levels of Lp(a) (orange, n=10) and low levels of Lp(a) (blue, n=13). P-adjusted values <0.1 were considered statistically significant. *P<0.05, **P<0.01.

Figure 3.

Lipoprotein(a) (Lp[a]) fractions are enriched with diacylglycerols (DGs). A, Per major lipid class, the mean percentage in the total lipidome of the Lp(a) fraction. B, The presence of DGs and the lysophosphatic acid precursor lysophosphatidylcholine (LPC) on Lp(a). Total value is the sum of the relative abundance of all lipid species of that lipid class. 1/2-acyl LPC indicates 1/2-lysophosphatidylcholine; 2-acyl LPE, 2-lysophosphatidylethanolamine; CE, cholesterol ester; Cer(d), ceramide; DG(O)/DG’(O), alkyl/alkenyl-alcyl glycerol; HexCer(d), hexosylceramide; LPC(O)/LPC’(O), alkyl/alkenyllysophosphatidylcholine; PC, phosphatidylcholine; PC(O)/PC’(O), alkyl/alkenylphosphatidylcholine; PE, phosphatidylethanolamine; PE(O)/PE’(O), alkyl/alkenylphosphatidylethanolamine; PI, phosphatidylinositol; SM(d), sphingomyelin/ceramide phosphocholine; SPH(d), sphinganine; TG, triacylglycerol; and TG(O), alkyl diacylglycerol.

Figure 4.

Diacylglycerols (DGs) induce a proinflammatory phenotype in monocytes. A, Boxplots demonstrating the relative abundance of DG(40:6), DG(38:4), and lysophosphatidic acid (LPA) in the plasma of healthy individuals with elevated lipoprotein(a) (Lp[a]) levels. A Student t test with post hoc Bonferroni correction was used for statistical comparison between the 2 groups. B, Volcano plot showing significantly differentially expressed genes in DG(40:6)-stimulated human primary monocytes adjusted for multiple testing using the Benjamini and Hochberg approach. Gene ontology (C) and KEGG (Kyoto Encyclopedia of Genes and Genomes) pathway analysis (D). E, Targeted heatmap displaying significantly upregulated genes. BP indicates biological process; DMSO, dimethyl sulfoxide; GO, gene ontology; IL, interleukin; NF, nuclear factor; NOD, nucleotide-binding oligomerization domain; and TNF, tumor necrosis factor. **P<0.01, ***P<0.001.

Figure 5.

Diacylglycerols (DGs) induce a dose-dependent proinflammatory response in monocytes. A, IL (interleukin)-6, IL-8, IL-1β, and MCP-1 (monocyte chemoattractant protein-1) cytokine secretion in cell medium of monocytes stimulated with increasing concentration of DG(40:6; 24-h stimulation; n=5). B, IL-6, IL-8, IL-1β, and MCP-1 cytokine secretion in cell medium of monocytes stimulated with increasing concentration of DG(38:4; 24-h stimulation; n=5). C, IL-6, IL-8, IL-1β, and MCP-1 cytokine secretion in cell medium of monocytes stimulated with increasing concentration of lysophosphatidic acid (24-h stimulation; n=5). All data are mean±SEM. Grubbs outlier tests were performed to identify outliers. This was followed by a Shapiro-Wilk test to test for normality. When data followed a normal distribution, an ordinary 1-way ANOVA was performed followed by Tukey multiple comparisons test. When the data were not normally distributed, a Kruskal-Wallis test was performed followed by a Dunn multiple comparisons test. DMSO indicates dimethyl sulfoxide; and Unst, unstimulated. *P<0.05, **P<0.01, ****P<0.0001.

Figure 6.

Diacylglycerols (DGs) enriched in the lipoprotein(a) (Lp[a]) fraction activate the inflammasome. A, NLRP3 (NOD [nucleotide-binding oligomerization domain]-like receptor family pyrin domain containing 3) secretion in cell medium of monocytes stimulated with increasing concentration of DG(40:6), DG(38:4), or lysophosphatidic acid (LPA; 24-h stimulation; n=3). B, Representative immunoblot demonstrating a dose-dependent increase in Casp1 (caspase-1) expression in monocytes after 24-hour DG(40:6) stimulation. C, Quantification of the immunoblot (n=3). D, IL-1β secretion in cell medium (n=3). All data are mean±SEM. Grubbs outlier tests were performed to identify outliers. This was followed by a Shapiro-Wilk test to test for normality. Then, the data were analyzed using an ordinary 1-way ANOVA followed by Dunnett multiple comparisons test or Tukey multiple comparisons test. DMSO indicates dimethyl sulfoxide; and Unst, unstimulated. *P<0.05, ***P<0.001, ****P<0.0001.

Figure 7.

NLRP3 (NOD [nucleotide-binding oligomerization domain]-like receptor family pyrin domain containing 3) inhibition can partly diminish lipoprotein(a) (Lp[a])-induced inflammation. Monocytes were pretreated with 1 µM CRID3 (cytokine release inhibitory drug 3) or DMSO (dimethyl sulfoxide) control for 1 hour before stimulation with 1 mg/mL Lp(a) (n=7–8). IL (interleukin)-1β, Casp1 (caspase-1; A), IL-6, and IL-8 (B) secretion were measured in the cell medium. All data are mean±SEM. Grubbs outlier tests were performed to identify outliers. This was followed by a Shapiro-Wilk test to test for normality. When data followed a normal distribution, an ordinary 1-way ANOVA was performed followed by Dunnett multiple comparisons test. When the data were not normally distributed, a Kruskal-Wallis test was performed followed by a Dunn multiple comparisons test. *P<0.05.

Figure 8.

Diacylglycerols (DGs) and lysophosphatidic acid (LPA) facilitate monocyte extravasation. A, Representative images of monocyte transendothelial migration (TEM), which were stimulated with either DMSO (dimethyl sulfoxide) as solvent control or 10 μM DG(40:6). Transmigrated monocytes are represented by black cells with a red asterisk and adhered monocytes by white cells. Scale bar, 100 µm. B, Quantification of the number of migrated cells, number of adhered cells, and the percentage of migrated cells after stimulation with 1 μM, 10 μM DG(40:6), or DMSO control. C, Representative images of TEM of monocytes stimulated with DMSO control or 10 μM DG(38:4). Transmigrated monocytes are represented by black cells with a red asterisk and adhered monocytes by white cells. Scale bar, 100 µm. D, Quantification of the number of migrated cells, number of adhered cells, and the percentage of migrated cells after stimulation with 1 μM, 10 μM DG(38:4), or DMSO control. E, Representative images of TEM of monocytes stimulated with DMSO control or 10 μM LPA. Transmigrated monocytes are represented by black cells with a red asterisk and adhered monocytes by white cells. Scale bar, 100 µm. F, Quantification of the number of migrated cells, number of adhered cells, and the percentage of migrated cells after stimulation with 1 μM, 10 μM LPA, or DMSO control. All data are mean±SEM. Data represent 4 independent experiments. Grubbs outlier tests were performed to identify outliers. This was followed by a Shapiro-Wilk test to test for normality. When data followed a normal distribution, an ordinary 1-way ANOVA was performed followed by Dunnett multiple comparisons test. When the data were not normally distributed, a Kruskal-Wallis test was performed followed by a Dunn multiple comparisons test. *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001.

Figure 9.

Lipoprotein(a) (Lp[a]) colocalizes with M1 macrophages in atherosclerotic lesions. A, Images representing human carotid plaques derived from patients with high Lp(a) and low Lp(a) levels. Nuclei were stained with Hoechst (blue). Macrophages were stained with the pan-macrophage marker CD68 (red). M1 macrophages were determined by colocalization of CD68 and MARCO (macrophage receptor with collagenous structure; green). Apo(a) was stained with LPA4 (yellow). Scale bars, 40 and 20 µm for the zoom images. B, tSNE visualization of 3 distinct macrophage populations. C, Bar plot demonstrating the distribution of macrophage clusters in the Lp(a) high and low groups. ABCA1 indicates ATP-binding cassette transporter 1; Casp1, caspase 1; IL, interleukin; OLR1, oxidized low-density lipoprotein receptor 1; SELL, selectin L; TLR4, toll-like receptor 4; TREM2, triggering receptor expressed on myeloid cells 2; and tSNE, t-distributed stochastic neighbor embedding.

Figure 10.

Lipoprotein(a) (Lp[a]) lowering had no effect on diacylglycerols (DGs) as a lipid class, as well as inflammasome markers. A, Boxplots demonstrating the total value of DG after potent Lp(a) lowering by apo(a) antisense (pelacarsen). Total value is the sum of the relative abundance of all identified lipid species. For the lipidomics analysis, a Student t test with post hoc Bonferroni correction was used for statistical comparison between the 2 groups. B, Graphs demonstrating the relative abundance of DG(40:6) and DG(38:4) after potent Lp(a) lowering by apo(a) antisense (pelacarsen). C, Boxplots and graphs demonstrating the relative abundance of DG(32:3), DG(33:2), and DG(39:1) between healthy individuals with high and low levels of Lp(a), as well as after potent Lp(a) lowering by apo(a) antisense (pelacarsen) in patients with cardiovascular disease. D, Graphs demonstrating the relative abundance of DG(38:0), DG(38:1), DG(40:1), DG(40:2), DG(40:3), DG(42:1), DG(42:2), and DG(42:3) after potent Lp(a) lowering by apo(a) antisense (pelacarsen). E, Box plots of differentially expressed chemotaxis- related genes before and after pelacarsen. P-adjusted values <0.1 were considered statistically significant. F, IL (interleukin)-18 and Casp1 (caspase-1) secretion in plasma before and after pelacarsen treatment. Data were analyzed using a 2-tailed paired t test. *P<0.05, **P<0.01, ***P<0.001.