Single high-fat challenge and trained innate immunity: A randomized controlled cross-over trial

Abstract

Brief exposure of monocytes to atherogenic molecules, such as oxidized lipoproteins, triggers a persistent pro-inflammatory phenotype, named trained immunity. In mice, transient high-fat diet leads to trained immunity, which aggravates atherogenesis. We hypothesized that a single high-fat challenge in humans induces trained immunity. In a randomized controlled cross-over study, 14 healthy individuals received a high-fat or reference shake, and blood was drawn before and after 1, 2, 4, 6, 24, and 72 h. Incubation of donor monocytes with the post-high-fat-shake serum induced trained immunity, regulated via Toll-like receptor 4. This was not mediated via triglyceride-rich lipoproteins, C12, 14, and 16, or metabolic endotoxemia. In vivo, however, the high-fat challenge did not affect monocyte phenotype and function. We conclude that a high-fat challenge leads to alterations in the serum composition that have the potential to induce trained immunity in vitro. However, this does not translate into a (persistent) hyperinflammatory monocyte phenotype in vivo.

Keywords: Health sciences; Human Physiology; Human metabolism.

Graphical abstract

Figure 1

Study design and postprandial metabolite changes (A) Graphical outline of the in vivo study design. A total of 14 participants received a high-fat shake and a reference shake in a cross-over design. Seven participants first received the high-fat shake, and seven participants first received the reference shake. To exclude carryover effects, the participants received the other shake after a washout period of at least 1 month. Blood was drawn before (t = 0 h) and at several time points after the consumption of both shakes (t = 1 h, t = 2 h, t = 4 h, t = 6 h, t = 24 h, t = 72 h). Starting from the evening before the ingestion of the shake until time point t = 72 h, the participants were provided with standardized meal plans. See also Figure S1. Created with BioRender.com. (B) At t = 0 h, t = 1 h, t = 2 h, t = 4 h, and t = 6 h glucose, triglyceride, free fatty acid, and insulin concentrations were measured for each shake (n = 14). Median ± IQR. ∗ indicates two-sided p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001; no sign indicates no significant differences, as calculated with linear mixed model for repeated measures. Due to technical errors, data from 1 participant was missing for the control shake and from 1 participant for the high-fat shake. Missing data were fitted with the restricted maximum likelihood approach.

Figure 2

Exposure to serum obtained after a high-fat shake, but not after a common breakfast shake, induces trained immunity in healthy human monocytes (A) Graphical outline of the design of the in vitro experiments. Adherent healthy human monocytes were exposed to serum obtained before (t = 0 h; CS0/HFS0) and at several time points after (t = 2 h [CS2/HFS2], t = 4 h [CS4/HFS4], and t = 6 h [CS6/HFS6]) consumption of either the reference or high-fat shake. After 24 h incubation time, cells were rested and differentiated into macrophages. On day 6, the cells were restimulated with Toll-like receptor (TLR)-agonists (TLR4; lipopolysaccharide [LPS] and TLR2; Pam3Cys [P3C]) for another 24 h before cytokine production was measured with ELISA. Created with BioRender.com. (B) Serum obtained after the high-fat shake increased TNF-α and IL-6 production of human monocyte-derived macrophages upon secondary stimulation with LPS, with the highest cytokine production after stimulation with serum obtained at t = 6 h (n = 17). (C) Cytokine production capacity was measured after 24 h exposure to reference shake serum (CS) and the high-fat shake serum (HFS). Data are presented as fold of change to the fasting serum obtained before consumption of the control shake (CS0) or the high-fat shake (HFS0) (n = 11). Median ± IQR. ∗ indicates two-sided p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001, Wilcoxon signed-rank test. CS, serum after the reference shake; HFS, serum after the high-fat shake. See also Figure S2.

Figure 3

High-fat-shake-induced trained immunity is regulated via TLR4 but is not mediated via low-dose LPS, triglyceride-rich lipoproteins, or the saturated fatty acids that are particularly increased after the high-fat shake (A) Adherent human monocytes were pre-incubated for 1 h with plain RPMI or Bartonella LPS (B. LPS). After pre-incubation, the cells were exposed for 24 h to high-fat serum obtained at t = 0 h (HFS0) or at t = 6 h (HFS6), with (PB) or without neutralization of LPS. On day 6, cells were restimulated for 24 h and cytokine production was measured (n = 8). Data are presented as fold of change to HFS0 for each separate inhibitor. (B) HFS0 and HFS6 were depleted from apoB-containing lipoproteins before use in the training experiments (n = 6). (C) The fatty acid composition of the pooled serum obtained at t = 6 h after consumption of the high-fat and the reference shake. Lauric acid (FA(12:0))^a, myristic acid (FA(14:0))^a, and stearic acid (FA(18:0))^a were particularly higher in the high-fat serum. Palmitoleic acid (FA(16:1))^a, oleic acid (FA(18:1))^a, and linoleic acid (FA(18:2))^a were lower. (D) Monocytes were stimulated for 24 h with albumin-conjugated C12:0, C14:0, and C18:0 or with the albumin vehicle (Alb) alone. After resting and differentiation, cytokine production upon restimulation was measured (n = 8). (E) The same inhibition experiments as described under (A) were performed for C12:0 and C14:0. Data are presented as fold of change to the vehicle control for each separate inhibitor (B. LPS n = 10, PB n = 9). Median ± IQR. ∗ indicates two-sided p < 0.05, ∗∗p < 0.01, Wilcoxon signed-rank test. See also Figures S3 and S4.

Figure 4

A single high-fat challenge does not induce persistent changes in white blood cell composition and does not induce increased cytokine production capacity in circulating PBMCs in healthy human volunteers (A) Total white blood cell counts, neutrophil counts, lymphocyte counts, and monocyte counts before and at several time points (t = 0 h, t = 4 h, t = 24 h, and t = 72 h) after the consumption of the high-fat and reference shake (n = 13). (B) Ex vivo cytokine production capacity of PBMCs isolated from the study participants at the same time points (n = 13). There was no significant difference in the PBMC TNF-α production upon LPS restimulation at t = 72 h, as calculated with the Wilcoxon signed-rank test (one-sided p = 0.31). Except for a lower PBMC TNF-α production upon LPS stimulation after the high-fat shake compared to the reference shake at t = 4 h, there were no significant changes, as calculated with linear mixed model for repeated measures. Median ± IQR. ∗ indicates p < 0.05, and no sign indicates no significance.