Fatty acid desaturation and lipoxygenase pathways support trained immunity

Abstract

Infections and vaccines can induce enhanced long-term responses in innate immune cells, establishing an innate immunological memory termed trained immunity. Here, we show that monocytes with a trained immunity phenotype, due to exposure to the Bacillus Calmette-Guérin (BCG) vaccine, are characterized by an increased biosynthesis of different lipid mediators (LM) derived from long-chain polyunsaturated fatty acids (PUFA). Pharmacological and genetic approaches show that long-chain PUFA synthesis and lipoxygenase-derived LM are essential for the BCG-induced trained immunity responses of human monocytes. Furthermore, products of 12-lipoxygenase activity increase in monocytes of healthy individuals after BCG vaccination. Grasping the underscoring lipid metabolic pathways contributes to our understanding of trained immunity and may help to identify therapeutic tools and targets for the modulation of innate immune responses.

Fig. 1. Increased PUFA biosynthesis in BCG-trained monocytes is associated with eicosanoid and docosanoid biosynthesis after 24 hours.

Levels of PUFA and oxylipids in stimulated monocytes at 24 h with BG, BCG, and LPS (n = 4-8 biologic replicates, pooled from 4 independent experiments), a Arachidonic acid pathway (AA), b Eicosapentaenoic acid (EPA), and c Docosahexaenoic acid (DHA) pathway. Violin plots showing median and the interquartile region, Friedman’s test followed by Dunn’s multiple comparisons test. Area ratios (A.R), corrected for internal standards and protein content are shown. (Ctrl control, BG β-glucan, BCG Bacillus Calmette-Guérin, LPS lipopolysaccharide, LOX lipoxygenase, COX cyclooxygenase, PG prostaglandin, Tx thromboxane, LT leukotriene, HETE hydroxyeicosatetraenoic acid, HEPE hydroxyeicosapentaenoic acid, HDHA hydroxy docosahexaenoic acid).

Fig. 2. FADS2 activity contributes to the induction of trained immunity.

a Monocyte FADS1, FADS2, SCD gene expression 4 h and 24 h after BCG exposure (n = 8, except for FADS2 at 4 h which is n = 10 biologic replicates). b Schematic representation of the in vitro model used for the pharmacological inhibition of the induction of BCG-trained immunity. c Production of TNFα and IL-6 by BCG-trained macrophages incubated with (5 µM, 15 µM) SC-26196 for the first 24 h of culture and restimulated at day 6 with LPS for 24 h. (n = 6 biologic replicates pooled from 2 independent experiments). d Production of TNFα and IL-6 by BCG-trained macrophages incubated with (5 µM) GSK2033 for the first 24 h and restimulated with LPS at day 6 of culture for further 24 h. (n = 6 biologic replicates pooled from 2 independent experiments). Mean + SD. c, d two-way ANOVA, Sidak’s multiple comparisons test. (Ctrl: control; BCG: Bacillus Calmette-Guérin; LPS: lipopolysaccharide; FADS: fatty acid desaturase SCD: stearoyl-CoA desaturase).

Fig. 3. Inhibition of the COX pathway does not decrease BCG-induced cytokine production.

a TNFα and IL-6 secreted by macrophages exposed to BCG and (0.5 mM) Aspirin, (30, 300 nM) FR122047, (1, 50 µM) Celecoxib or (5, 10 µM) Itraconazole for the first 24 h of culture and 5 days later stimulated with LPS for 24 h (n = 9 (Aspirin and Itraconzole), n = 12 (FR122047), n = 6 (Celecoxib) biologic replicates, pooled from 3 (Aspirin and Itraconazole), 4 (FR122047) or 2 (Celecoxib) independent experiments). b Relative COX1 and COX2 expressions in monocytes incubated with BCG for 4 h (n = 6 biologic replicates pooled from 2 independent experiments). c Schematic representation of in vitro LM supplementation experiments. d TNFα and IL-6 secreted by macrophages exposed to (10, 100 pM) PGE₂ for the first 24 h of culture and 5 days later stimulated with LPS for 24 h (n = 6 biologic replicates, pooled from 2 independent experiments). a Mean, 10–90 percentile and whiskers extend to the most extreme points, b, d Mean + SD. a two-way ANOVA, Sidak’s multiple comparisons test, b two-tailed Wilcoxon matched-pairs signed rank test, d Friedman’s test followed by Dunn’s multiple comparisons test. (Ctrl control, BCG Bacillus Calmette-Guérin, LPS lipopolysaccharide, COX cyclooxygenase, CP cytochrome P, PG prostaglandin).

Fig. 4. LOX pathways play a role in BCG-induced trained immunity.

a Production of TNFα and IL-6 by BCG-trained macrophages incubated with (10, 100 µM) Zileuton, (2, 20 µM) ML355 or 10 µM PD146176 for the first 24 h and restimulated with LPS for another 24 h at day 6 of culture. (n = 9 biologic replicates, pooled from 3 independent experiments). b–d Production of TNFα and IL-6 by macrophages incubated with b (3, 30, 300 nM) 5-HETE, (0.1, 1, 5 µM) 7-HDHA, (1, 10 nM) LTB₄, (c) (0.03, 0.3, 3 µM) 15-HETE, (0.1, 1, 5 µM) 17-HDHA, d (0.1, 10 nM) 12-HETE (0.1, 10 nM) 12-HEPE or (0.1, 10 nM) 14-HDHA for 24 h and restimulated at day 6 of culture with LPS for 24 h. (n = 6 (b, c) or n = 9 (d) biologic replicates, pooled from 2 (b, c) or 3 (d) independent experiments). Mean + SD, a two-way ANOVA, Sidak’s multiple comparisons test, b–d Friedman test followed by Dunn’s multiple comparisons correction. (Ctrl: control; BCG: Bacillus Calmette-Guérin; LPS: lipopolysaccharide; LOX: lipoxygenase).

Fig. 5. Lipid desaturation and the LOX pathway are involved in the in vivo induction of trained immunity.

a, b Heatmaps of the p values of the association between SNPs around FASD1, FADS2, SCD, LOX5, LOX12 and LOX15 genes (+/− 250 kb) and the fold change of the production of (a) TNFα and IL-6 after in vitro exposure to β-glucan or BCG and LPS restimulation, (n = 238 healthy volunteers for TNFα and n = 251 healthy volunteers for IL-6) and b TNFα, IL-6 and IL-1β upon BCG vaccination and S. aureus ex vivo restimulation (n = 278 healthy volunteers). Boxplots show selected SNPs with the lowest p-values stratified according to genotype, shown in the heatmap in blue. Boxplots show median, upper, and lower quartiles, and whiskers extend to the most extreme point less than 1.5 times the interquartile range from the box. c Fold change of intracellular monocyte levels of 12-LOX products of AA (12-HETE), EPA (12-HEPE), and DHA (14-HDHA) one month after BCG vaccination (visit 2) relative to levels prior to vaccination (visit 1) (n = 6 healthy volunteers per group, one outlier per group was removed (ROUT test (Q = 5%)), two-way ANOVA, Sidak’s multiple comparisons test. (Ctrl control, BCG Bacillus Calmette-Guérin, LPS lipopolysaccharide, PBMC peripheral blood mononuclear cells, FC fold change, A adenine, T thymine, C cytosine, G guanine; V visit).