Yuanyuan Liu1
Arslan Hamid2Hannah Hardege1Qian Zhang1Helena Körner1Merle Leffers1Noelia A. Gonzalez3Gerhard Liebisch4Marcus Hoering4Hannes Findeisen1
Katarzyna Placek2Mihai G. Netea2,5Holger Reinecke1Dennis Schwarz1*
Yahya Sohrabi1,6,7*- 1Department of Cardiology I - Coronary and Peripheral Vascular Disease, Heart Failure, University Hospital Münster, Münster, Germany
- 2Department of Immunology and Metabolism, Life and Medical Sciences Institute (LIMES)-Institute, University of Bonn, Bonn, Germany
- 3Institute of Immunology, University of Münster, Münster, Germany
- 4Institute of Clinical Chemistry and Laboratory Medicine, University Hospital Regensburg, Regensburg, Germany
- 5Department of Internal Medicine, Radboud University Medical Center, Nijmegen, Netherlands
- 6Department of Medical Genetics, Third Faculty of Medicine, Charles University, Prague, Czechia
- 7Institute of Molecular Genetics, Czech Academy of Sciences, Prague, Czechia
Granulocyte-macrophage colony-stimulating factor (GM-CSF) contributes to the host defense and the pathogenesis of inflammatory diseases at least in part through inducing trained immunity (TI), however, the mechanism remains poorly characterized. In this paper, we systematically investigated the associated metabolic and epigenetic reprogramming, with a particular focus on the role of liver X receptors (LXRs) in this process. We employed a comprehensive experimental approach, including in vitro isolation and purification of human monocytes from healthy donors, cytokine assays, quantitative PCR, Seahorse metabolic analysis, flow cytometry, and chromatin immunoprecipitation (ChIP), shotgun lipidomics, as well as transcriptomic data analysis to investigate GM-CSF–induced trained immunity. Our results demonstrate that GM-CSF induces TI by enhancing cellular metabolism, as evidenced by increased glycolysis, mitochondrial activity, fatty acid oxidation, and pyruvate metabolism. Lipidomics and RNA sequencing analyses revealed upregulation of lipid synthesis, high triglyceride storage, and acetyl-CoA–producing pathways, leading to increased histone acetylation in GM-CSF–trained cells. Furthermore, glycolysis and mitochondrial metabolism are essential for establishing TI in these cells. Notably, pharmacological inhibition of GM-CSF activated LXR signaling, which potentially mediated via PPARγ, attenuated GM-CSF–induced TI via reducing glycolytic flux and histone acetylation while activation of LXR amplified these effects. Together, these results highlight the role of LXR in linking cellular metabolism with epigenetic reprogramming and demonstrate that elevated metabolic activity and active LXR signaling both are essential for GM-CSF–induced trained immunity. Importantly, these pathways may represent therapeutic targets for modulating GM-CSF–driven maladaptive inflammation in chronic inflammatory diseases.
Graphical Abstract.
1 Introduction
Granulocyte-Macrophage Colony-Stimulating Factor (GM-CSF) is a key cytokine that orchestrates inflammatory responses by inducing differentiation and activation of monocytes and macrophages, leading to increased secretion of pro-inflammatory cytokines such as TNF-α and IL-1β, as well as of reactive oxygen species (ROS) – all of which contribute to inflammatory responses (1–3) and if uncontrolled can lead to tissue damage (4, 5). In addition to promoting the recruitment and survival of myeloid cells, including monocytes and neutrophils, GM-CSF supports effective pathogen clearance. However, its activity can also lead to maladaptive effects such as contributing to the pathogenesis of rheumatoid arthritis (RA) or hyperinflammatory responses to an infection such as COVID-19 (6–8). Furthermore, in the context of cardiometabolic diseases, administration of GM-CSF has been demonstrated to promote atherosclerotic plaque development (9, 10), whereas Csf2-/- mice, which lack GM-CSF signaling, exhibit smaller atherosclerotic lesions (11). Pro-inflammatory stimuli such as oxidized low-density lipoprotein (oxLDL) rapidly induce GM-CSF in macrophages, thereby promoting survival and potentially worsening vascular inflammation in atherosclerosis (12). Conversely, the inhibition of glycolysis reduces plaque formation and enhanced plaque stability in animal models, highlighting the pathological role of inducible glycolysis in atherosclerosis (13). Therefore, targeting GM-CSF or its receptor has the potential to reduce inflammatory activation of monocytes, including cytokine production (4, 5, 14).
Beyond its role in acute inflammation, GM-CSF has recently emerged as a potential regulator of trained immunity. Trained immunity refers to the ability of innate immune or non-immune cells to develop an antigen-agnostic heightened, memory-like response to secondary challenges, a process mediated through epigenetic and metabolic reprogramming (15, 16). In mouse models, β-glucan-induced trained immunity is associated with increased myelopoiesis, at least in part driven by the effects of GM-CSF on hematopoietic precursors (17). In addition, GM-CSF enhances the response to β-glucan in in-vitro models of trained immunity (18). However, GM-CSF alone is also able to prime monocytes to produce heightened inflammatory responses to LPS, suggesting the development of innate immune memory (15). Moreover, priming of monocytes with GM-CSF is linked with marked upregulation of genes involved in fatty acid metabolism and the cholesterol synthesis pathway - genes that are regulated by Liver X Receptors (LXRs) (7, 19). LXRs are nuclear receptors that function as metabolic gatekeepers, particularly in cholesterol homeostasis and lipid metabolism (20, 21). Recent data suggest that GM-CSF-induced inflammation is modulated by LXR activity. LXRs have emerged as critical regulators of the interface between cellular metabolism and inflammation (17, 18). Notably, genetic or pharmacological activation or inhibition of LXRs in human cells results in opposing outcomes (7, 19). LXR activity appears to be highly context-dependent: while LXR activation in macrophage colony-stimulating factor (M-CSF)-differentiated cells can shift their anti-inflammatory phenotype toward a more inflammatory one (19), LXR inhibition in GM-CSF-differentiated monocytes alters their cytokine output and inflammatory tone (7). Previously, we have described the induction of trained immunity by LXR activation through increased glycolysis, and epigenetic modification in histone marks (22). We also showed that this process is associated with elevated intracellular acetyl-CoA, HIF1α accumulation and IL-1β signaling (22). Furthermore, it was shown that during GM-CSF driven differentiation or in inflammatory milieus, LXR inhibition drives an anti-inflammatory reprogramming by up-regulating MAFB and thereby counteracts the GM-CSF inflammatory imprint (7). This highlights that LXR is able to modulate macrophage fate during M-CSF/GM-CSF-dependent differentiation.
In this study, we tend to uncover the regulatory role of LXR in GM-CSF induced trained immunity, which was not addressed yet.
2 Materials and methods
2.1 PBMC and monocyte isolation
Human monocytes were isolated from fresh human blood leukocyte reduction chambers obtained from healthy subjects recruited by the blood bank of the University Hospital Münster as previously described (22). The study was approved by the Scientific and Ethics Committee of the University of Münster and conforms to the principles of the Declaration of Helsinki. Written informed consent was obtained from all donors by the blood bank and leukocyte reduction filters were provided anonymously without sharing additional personal and detailed information. Monocyte isolation was performed by two-step differential density centrifugation over Histopaque® 1077 (Sigma-Aldrich, St. Louis, MO, USA, #10771) and a hyper-osmotic Percoll gradient (46% Percoll in RPMI, GE Healthcare, Uppsala, Sweden, #17089101). Cells were purified further with MACS Pan Monocyte Isolation Kit (Miltenyi Biotec, Bergisch Gladbach, Germany, #130-096-537) and washed once with serum-free RPMI-1640 medium before resuspension in complete RPMI culture medium.
2.2 Monocyte training experiments
Monocytes were cultured at a density of 1 x 106 cells/well, 200,000 cells/well or 50,000 cells/well in a 6-, 24- or 96-well plate (Greiner Bio-OneTM) respectively and primed with 1,000 U/ml rhGM-CSF (ImmonoTools, #11343125) or 30 ng/ml rhM-CSF (ImmonoTools, #11343115) in the presence or absence of 1 μM GW3965 (Cayman, #10054) or 1 μM T0901317 (T13) (#HY-10626 MCE), 2.5 μM GSK2033 (LXR antagonist, Cayman, #25443), 10 μM 3PO (PFKFB3 inhibitor, Axonmedchem, #2175), 10 mM 2-DG (#HY-13966, CME), 10 μM (OSI-27) (HY-10423, MCE), 10 μM KC7F2 (HIF1α inhibitor, Medchemexpress, #HY-18777), 50 μM UK5099 (mitochondrial pyruvate carrier (MPC) inhibitor, Medchemexpress, #HY-15475), 40 μM etomoxir (CPT-1a inhibitor, Medchemexpress, #HY-50202), 40 μM BPTES (Glutaminase inhibitor, Medchemexpress, HY-12683) or vehicle for 24 h in RPMI supplemented with 10% fetal bovine serum, 1 mM sodium pyruvate solution, 2 mM L-glutamine solution and 1% penicillin/streptomycin. The cells were pretreated with corresponding inhibitors/activators 1 h prior to the addition of M-CSF or GM-CSF. The plates were incubated in a 5% CO2 incubator maintained at 37 °C. The medium was changed to fresh complete RPMI 24 h after priming and the cells were rested for 5 days. The medium was exchanged once again on the day 3 of resting. On day 6, the cells were left unstimulated or restimulated with 10 ng/mL LPS (Sigma) for mRNA or ELISA analysis. The cells were lysed after 4–6 h, and supernatant was collected 24 h after restimulation and stored at −20 °C until further analysis.
2.3 Cytokine measurements
The levels of proinflammatory cytokines in supernatants were measured using DuoSet ELISA kits for human TNFα (R&D, #DY210) and human IL-6 (R&D, #DY206) following the manufacturer’s instructions. The absorbance was quantified in a Multimode Plate Reader Victor™ X3, (Perkin Elmer, USA) at 450 nm. Concentrations were calculated by four-parameter logistic regression.
2.4 RNA isolation and qPCR
RNA was obtained from the lysed cells using the NucleoSpin RNA isolation kit (Macherey-Nagel, Düren, Germany, #740955.250) and reverse-transcribed using the RevertAid First Strand cDNA Synthesis Kit (Thermo Scientific, Vilnius, Lithuania, #K1632) according to the manufacturer’s instructions. Real-time qPCR was conducted with the iTaqTM Universal SYBR Green Supermix (Bio-Rad, Hercules, CA, USA, #172-51254) to determine the level of expression for cMYC, GLUT1, HK2, G6PD, ALDOC, ABCG1, HMG-CoAs, ACC1, CPT1A, ACLS, ACSS2, IL-6, TNFα, HMGCR, FASN, ACLY and TFIIB as a housekeeping gene (Table 1).
Table 1. List of primers used for the qRT-PCR gene expression analysis and ChIP-PCR.
2.5 Chromatin immunoprecipitation assay
On the day 6 after training, the cells were fixed, cross-linked, and harvested on day 6 for chromatin immunoprecipitation (ChIP). The assay was performed using the MAGnify Chromatin Immunoprecipitation System (Invitrogen, Carlsbad, CA, USA) according to the manufacturer’s instructions. The immunoprecipitation of the chromatin samples was conducted using an H3K27ac (Diagenode, Ougréé, Belgium, #C15410196) antibody. The rabbit polyclonal IgG (Diagenode, Ougréé, Belgium, #C15410206) antibody was utilized as a negative control. Furthermore, input control and histone proteins linked to DNA were isolated using magnetic beads that were included in the kit. The isolation of DNA was followed by its amplification using real-time quantitative polymerase chain reaction (qPCR). Histone enrichment was identified using promoter-specific primers for IL-6 and TNFα (Table 1).
2.6 Seahorse analysis
2.5 x 104 monocytes per well were seeded in a Seahorse XFp cell culture plate and primed as described before. Oxygen consumption rate (OCR) and extracellular acidification rate (ECAR) were evaluated under basal conditions and following sequential injection of 2 μM oligomycin, 1.5 μM FCCP and 100 nM rotenone plus 1 μM antimycin A (all from Sigma-Aldrich). The cells were maintained in XF Seahorse medium (XF Base Medium Minimal DMEM; Agilent Technologies) containing 10 mM glucose, 2 mM L-glutamine, and 1 mM sodium pyruvate (all from Sigma-Aldrich). The OCR and ECAR were determined using an Agilent Seahorse XFp Mini Analyzer, and the results were analyzed using GraphPad Prism.
For the Seahorse Mito Fuel Flex Test, the medium was changed to XF Seahorse medium containing glucose and glutamine, and cells were incubated at 37 °C in a CO2-free incubator. Glucose dependency and capacity were assessed by sequential addition of 4 μM UK5099, 4 μM Etomoxir and 3 μM BPTES. Fatty acid β-oxidation dependency and capacity were determined by sequential injection of 4 μM Etomoxir, followed by 4 μM UK5099 and 3 μM BPTES. Glutamine metabolism dependency and capacity were assessed by sequential injection of 3 μM BPTES followed by 4 μM UK5099 and 4 μM Etomoxir. The assay was performed using an Agilent Seahorse XFp Mini Analyzer. Data were processed using Seahorse Wave Desktop Software (Agilent Technologies), and metabolic parameters were calculated according to the manufacturer’s guidelines.
Fuel dependency represents the reliance of cells on a specific substrate to maintain basal respiration, calculated as: Dependency (%) = [(Baseline OCR − Target 1 inhibitor OCR)/(Baseline OCR – OCR after all three inhibitors)] × 100. Fuel capacity reflects the ability of mitochondria to utilize a fuel to maintain basal respiration when other fuel pathways are inhibited. Capacity is calculated as: Capacity (%) = [(Baseline OCR – OCR after inhibition of the other two pathways)/(Baseline OCR − OCR after all three inhibitors)] × 100). Following the assay, the cells were lysed in RIPA and the seahorse results were normalized to the protein concentration in respective wells.
2.7 Flow cytometry
On the day six following the training of monocyte, they were harvested using cold PBS containing 5 mM ethylenediaminetetraacetic acid (EDTA). The cells were washed with PBS and stained with surface markers for CD80-FITC (Biolegend, #305206), CD86-APC (Miltenyi Biotec, #130-116-161), CD163-PE (Miltenyi Biotec, #130-097-628), CD206-Vioblue (Miltenyi Biotec, #130-127-809) or respective isotype controls for 30 minutes at 4 °C following manufacturer’s instructions. Following the staining process, the cells were washed with FACS buffer (PBS + 0.5% BSA) and Flow Cytometry analysis was performed using Guava easyCyte (Millipore).
2.8 Re-analysis of RNA-seq and microarray data
We performed a comprehensive re-analysis of two publicly available transcriptomic datasets from the Gene Expression Omnibus (GEO): GSE99056 and GSE156696. GSE99056 is a microarray dataset that profiles gene expression in human monocyte-derived macrophages that have been differentiated using either M-CSF or GM-CSF and then stimulated with 10 ng/ml LPS for four hours (23). In contrast, GSE156696 is an RNA-seq dataset characterizing transcriptional changes in GM-CSF–derived macrophages treated with the LXR agonist GW3965, the LXR antagonist GSK2033, or DMSO during a 7-day differentiation period (7).
For GSE99056, raw microarray data were processed in R software using the limma package (24). The data underwent background correction, quantile normalization, and transformation using the voom function to model mean-variance relationships suitable for linear modeling (25). Probes were annotated using the platform-specific annotation file, and lowly expressed probes were filtered. Donor identity was incorporated as a blocking factor to account for inter-individual variation (24). Differential expression analysis was conducted to compare M-CSF and GM-CSF-derived macrophages under LPS stimulation. Volcano plots, MA plots, and heatmaps of top differentially expressed genes were generated for visualization.
For GSE156696, raw RNA-seq FASTQ files were aligned to the GRCh38 reference genome using HISAT2 (26). Quality control was assessed using FastQC and aggregated with MultiQC (27). Gene-level quantification was performed using featureCounts with GENCODE annotation (28, 29). Low-count genes were filtered based on a minimum count threshold across samples. Differential gene expression analysis was performed using DESeq2 (30), including normalization by the median-of-ratios method, dispersion estimation, and modeling of donor-specific effects. Sample-level quality assessment included principal component analysis and hierarchical clustering based on sample-to-sample distance matrices.
For both datasets, differentially expressed genes were identified using an adjusted p-value threshold of 0.05 and an absolute log2 fold change cutoff. Functional enrichment analysis was conducted using the clusterProfiler package (31), encompassing Gene Ontology categories such as Biological Process (BP), Cellular Component (CC), and Molecular Function (MF), as well as KEGG pathway analysis. The gene universe for enrichment was defined as the set of all expressed or detected genes that passed quality control. Enrichment results were visualized using dot plots, barplots, and enrichment maps. All data processing, statistical analysis, and visualization steps were executed using R and Bioconductor packages (32).
2.9 Shotgun lipidomics
Cells were trained as described before and were lysed in 0.2% SDS lysis buffer for lipid extraction on the 6th days after priming. Lipids were extracted from a volume corresponding to 100 µg of total protein using the Bligh and Dyer method (33). For quantitative lipidomics, internal standards were added prior to extraction. Lipid analysis was performed using direct flow injection analysis (FIA) coupled with a triple quadrupole mass spectrometer (FIA-MS/MS) and a high-resolution hybrid quadrupole-Orbitrap mass spectrometer (FIA-FTMS) (34). FIA-MS/MS was conducted in positive ion mode following previously described methods (35). Triglycerides (TG), diglycerides (DG), and cholesterol esters (CE) were recorded in positive ion mode (m/z 500–1000) as [M+NH4]+ at a resolution of 140,000 (at m/z 200). CE measurements were corrected for species-specific response factors (36). Lipidomics data was normalized to the total protein content of the corresponding cell lysates to account for differences in sample loading.
3 Statistical analysis
The differences among experimental groups were evaluated using one-way analysis of variance (ANOVA) and the Wilcoxon matched-pairs signed-rank test was used to compare the samples in two groups using Graph Pad Prism for Windows. Prior to conducting the analysis, an F-test was performed to ascertain the homogeneity of variance. The exact quantity of individual samples and the variety of experiments is specified in the Figure legends. Technical replicates were averaged to limit the impact of experimental variations. A p-value of <0.05 was considered to be statistically significant.
4 Results
4.1 GM-CSF training of monocytes enhances proinflammatory cytokines expression
We compared GM-CSF- and M-CSF-driven macrophages to capture the distinct differentiation programs elicited by these cytokines. M-CSF typically induces a homeostatic, anti-inflammatory macrophage phenotype, whereas GM-CSF drives a more pro-inflammatory and metabolically active state characterized by enhanced glycolysis and increased oxidative metabolism (37, 38). Using M-CSF-derived macrophages as a baseline therefore enables clear identification whether GM-CSF confers additional metabolic or epigenetic features associated with trained immunity. This comparison therefore allows us to distinguish GM-CSF-specific effects from those related to macrophage differentiation per se. Growing evidence suggests that GM-CSF can induce sustained inflammatory responses in monocytes (15, 39). Therefore, we hypothesized that GM-CSF may have the ability to induce innate immune memory in monocytes through metabolic and epigenetic modifications. Training of human monocytes with GM-CSF led to significantly higher production of inflammatory cytokines upon restimulation with lipopolysaccharide (LPS), when compared to both untreated controls and M-CSF–trained cells (Figures 1A, B). Quantification of TNFα and IL-6 protein concentrations, as well as mRNA expression, revealed elevated amount of these cytokines in GM-CSF-trained cells (Figures 1C, D). To further explore this phenotype, we reanalyzed published microarray datasets, which confirmed that GM-CSF-derived macrophages express higher levels of proinflammatory cytokine genes, such as IL-6, IL-12/23, and type I/III interferons, whereas M-CSF-treated cells preferentially express anti-inflammatory mediators including IL-10 and IGF1 in response to LPS (data not shown) (19, 23). In addition, our phenotyping data indicate that GM-CSF-derived macrophages adopt a unique and plastic phenotype, characterized by low CD163, a moderate increase in CD206, while expression of the co-stimulatory molecules CD80 and CD86 remained relatively unchanged between M-CSF- and GM-CSF-primed cells (Supplementary Figures 1A–D) (40–43).
Figure 1. GM-CSF–trained cells exhibit an innate immune memory phenotype. Monocytes were treated with GM-CSF (1,000 U/mL), M-CSF (30 ng/mL), or left untreated for 24 h. Cells were then rested for 5 days in complete medium and restimulated with LPS (10 ng/mL) for 4 h or 24 h for RNA or cytokine analysis, respectively. IL-6 (A, C) and TNFα (B, D) were quantified at the mRNA or protein levels. Oxygen consumption rate (OCR) and extracellular acidification rate (ECAR) were measured on day 6 in Seahorse miniplates (E–J). One-way ANOVA with Tukey’s post hoc test or Wilcoxon matched-pairs signed-rank test was used as appropriate. Data are mean ± SD (n = 6–8 from ≥3 independent experiments). *p < 0.05, **p < 0.01, ***p < 0.001.
4.2 GM-CSF training reprograms cellular metabolism toward glycolysis
Metabolic reprogramming toward glycolysis is a hallmark of trained immunity (22, 44–47). To assess metabolic function, we performed a Seahorse-based extracellular flux analysis, which demonstrated a significant increase in glycolytic activity in GM-CSF–trained cells (Figures 1E–J). Although basal oxygen consumption rates (OCR) were comparable between GM-CSF- and M-CSF-treated cells, maximal mitochondrial respiration was significantly higher in the GM-CSF group (Figures 1E–G), reflecting a shift toward a more energetically active state capable of supporting anaerobic respiration. Specifically, both basal and maximal extracellular acidification rates (ECAR) – indicative of glycolytic flux – were markedly elevated in these cells (Figures 1H–J). Furthermore, calculating ATP production ratio by glycolysis and mitochondrial respiration revealed that GM-CSF-mediated training enhanced ATP production through glycolysis as well as through mitochondrial respiration in comparison to M-CSF treated cells (Supplementary Figure 2A). Consistent with this, pathway enrichment reanalysis on available microarray data revealed a significant upregulation of glycolytic pathways following GM-CSF training on the day 7 after priming (Figure 2A). qPCR analysis confirmed differential expression of key glycolytic genes involved in glucose uptake and metabolism (Figures 2B–E). Further supporting the functional relevance of this metabolic rewiring, inhibition of glycolysis using 3PO (a PFKFB3 inhibitor), KC7F2 (an HIF-1α inhibitor) or 2-DG significantly reduced inflammatory cytokine production in GM-CSF–trained monocytes (Figures 2F–I, Supplementary Figures 2B, C). This demonstrates that glycolysis is required for sustaining the inflammatory program associated with trained immunity in these cells.
Figure 2. Glycolysis is associated with GM-CSF–induced innate immune memory. RNA sequencing data (GSE99056) were analyzed using the limma and clusterProfiler R packages to identify differentially expressed genes (adjusted p < 0.05, |log2FC| > threshold) and glycolysis pathway enrichment (A). M-CSF– and GM-CSF–primed cells were analyzed for gene expression (B–E). Monocytes were pretreated with 3PO (PFKFB3 inhibitor) or KC7 (HIF1α inhibitor) before GM-CSF or M-CSF training and restimulated with 10ng/ml LPS for 24h for cytokine assays (F–I). Graphs represent mean values ± SD of six individuals in three different experiments. Wilcoxon matched-pairs signed-rank test; data are mean ± SD (n ≥ 6). *p < 0.05, **p < 0.01, ***p < 0.001.
4.3 GM-CSF priming shifts mitochondrial metabolism toward fatty acid oxidation
GM-CSF-primed monocytes exhibited increased oxygen consumption rates (OCR) (Figures 1E–G), suggesting an increase in overall cell metabolism and greater reliance on oxidative phosphorylation (OXPHOS) for mitochondrial ATP production while glycolysis is also upregulated (Figures 1H–J). Consistent with this, reanalysis of microarray data revealed enrichment of OXPHOS-related gene pathways in GM-CSF-treated cells compared to M-CSF group (Figures 3A–C). This suggests a significant role of mitochondrial metabolism in the process of GM-CSF induced trained immunity. Therefore, we performed mitochondrial Mito Fuel Flex Test, which evaluates cellular dependency, capacity, and flexibility for metabolizing glucose, glutamine, and long-chain fatty acids. Interestingly, GM-CSF-trained monocytes displayed a significantly increased oxidative capacity for pyruvate, glutamine, and long-chain fatty acids (Figures 3D–I). Notably, fuel dependency assays demonstrated a marked reliance on lipid β-oxidation in these cells, indicating that fatty acids are a major mitochondrial fuel source following GM-CSF induced trained immunity (Figures 3F–I). These results are supported by expression profiling of M-CSF - versus GM-CSF -trained cells, which showed GM-CSF induced upregulation of genes involved in unsaturated lipid biosynthesis and arachidonic acid metabolism—pathways closely associated with lipid peroxidation (Figures 4A, B). Lipidomics analysis further revealed higher levels of triglycerides in GM-CSF trained cells that serve as the primary storage and supply form of fatty acids for mitochondrial β-oxidation, directly linking them to enhanced catabolic capacity (Figures 4C, F) (48). In contrast, cholesteryl esters (CE) storage, which primarily involved in cholesterol storage and transport but not β-oxidation was lower in GM-CSF but much higher in M-CSF group (49) (Figures 4D, G). In addition, TG/CE ratio were significantly higher in GM-CSF trained monocytes (Figure 4E). A higher TG/CE ratio may reflect a greater availability of fatty acids for beta oxidation as TG stores are more readily mobilized for energy production. To determine the immunological relevance of this metabolic rewiring, we tested whether mitochondrial metabolism contributes to inflammatory cytokine production. Interestingly, pharmacological inhibition of fatty acid oxidation significantly reduced IL-6 and TNFα production in GM-CSF but not in M-CSF trained cells, supporting a key role for FAO in sustaining inflammatory responses (Figures 5A, B, Supplementary Figures 3A, B). However, inhibition of glutamine metabolism selectively reduced IL-6 production, but had no significant effect on TNFα (Figures 5C, D). Unexpectedly, blocking pyruvate entry into mitochondria enhanced IL-6 and TNFα production, likely due to a compensatory increase in aerobic glycolysis (Figures 5E, F) (50). These findings indicate that GM-CSF-trained cells exhibit an enhanced mitochondrial metabolic plasticity and are metabolically primed for efficient mitochondrial fatty acid oxidation.
Figure 3. GM-CSF increases mitochondrial dependence on pyruvate and fatty acid oxidation. RNA sequencing (GSE99056) identified enriched pathways related to pyruvate, glutamate, and fatty acid metabolism using the limma and clusterProfiler R packages. Differentially expressed genes were identified using an adjusted p-value threshold of 0.05 and an absolute log2 fold change cutoff (A–C). M-CSF– and GM-CSF–trained monocytes were analyzed in Seahorse assays to determine the contribution of glucose, glutamine, and fatty acid oxidation to mitochondrial respiration (D–I). Wilcoxon matched-pairs signed-rank test; mean ± SD (n = 6, three experiments). *p < 0.05, **p < 0.01, ***p < 0.001.
Figure 4. GM-CSF training enhances unsaturated fatty acid synthesis, arachidonic acid metabolism, and triglyceride levels. RNA-seq (GSE99056) identified enrichment of pathways regulating unsaturated fatty acid and arachidonic acid biosynthesis using the limma and clusterProfiler R packages. Differentially expressed genes were identified using an adjusted p-value threshold of 0.05 and an absolute log2 fold change cutoff. (A, B). Trained cells were analyzed by shotgun lipidomics to quantify triglycerides (TG), cholesteryl esters (CE), %TG and %CE in total stored lipids, and TG/CE ratios (C–G). Wilcoxon matched-pairs signed-rank test; mean ± SD (n = 5–6, ≥2 experiments). *p < 0.05, **p < 0.01.
Figure 5. GM-CSF–induced trained immunity depends on mitochondrial metabolism, acetyl-CoA production, and histone acetylation. Monocytes were pretreated with Etomoxir (CPT-1a inhibitor), BPTES (glutaminase inhibitor), or UK5099 (mitochondrial pyruvate carrier inhibitor) 1 h before GM-CSF stimulation. Cells were incubated for 24 h, rested for 5 days, and restimulated with LPS (10 ng/mL) for 24 h. IL-6 and TNFα levels were measured in the supernatants (A–F). Cells were trained as described previously, rested for 5 days, and subsequently lysed for analysis of gene expression associated with lipid metabolism (G–N). For histone acetylation analysis, monocytes were primed with GM-CSF or M-CSF for 24 h, rested for 5 days, and lysed on day 6 for ChIP assays using an anti-H3K27ac antibody or control IgG. qPCR was performed on IL6 and TNF promoter regions (O–P). Statistical analysis was performed using the Wilcoxon matched-pairs signed-rank test. Data represent mean ± SD of at least six donors from three independent experiments. p < 0.05, p < 0.01, p < 0.001.
4.4 GMCSF training activates acetyl-CoA-producing pathways and upregulates genes controlling cholesterol hemostasis
Acetyl-CoA is a central metabolic intermediate that serves as a critical link between cellular metabolism and epigenetic regulation by fueling histone acetylation and other protein acetylation events (51, 52). Gene expression profiling revealed that GM-CSF-trained monocytes express significantly higher levels of acetyl-CoA biosynthesis genes compared to M-CSF-treated cells (Supplementary Figure 4A). Specifically, genes such as ACC1, ACSL1, ACSS2 and FASN– key enzymes involved in de novo fatty acid synthesis and acetate utilization – as well as genes that regulate FAO were upregulated in the GM-CSF group (Figures 5G–N). These genes are directly or indirectly associated with acetyl-CoA metabolism in the cells. Taken together with the increased glycolysis, pyruvate utilization, and fatty acid oxidation observed above, these changes suggest a profound impact of GM-CSF on robust increase in intracellular acetyl-CoA production (53).
4.5 GMCSF increases histone modifications and promotes chromatin accessibility at inflammatory gene loci
The development of innate immune memory relies on epigenetic modifications at the promoters and enhancers of inflammatory genes associated with transcriptional activation, including histone H3 lysine 27 acetylation (H3K27ac) (54). In line with our observation of enriched acetyl-CoA metabolism pathways, we investigated whether the enhanced inflammatory responses following GM-CSF training were associated with histone acetylation that facilitates chromatin accessibility and promotes gene transcription. GM-CSF training led to a significant increase in H3K27ac at the IL-6 and TNFα promoter regions (Figures 5O, P), consistent with enhanced transcriptional activity at these loci. These findings indicate that GM-CSF induces a memory-like phenotype in monocytes, characterized by epigenetic remodeling at key inflammatory gene loci.
4.6 LXR regulates cell metabolism and inflammation in GM-CSF primed cells
In addition to metabolic rewiring, GM-CSF priming upregulated the expression of LXRα through PPARγ (Supplementary Figure 4B) (55) and several of its canonical target genes, including ABCG1, FASN, and HMG-CoA synthase, which regulate cholesterol homeostasis and lipid metabolism, suggesting convergence of these pathways on lipid-sensing nuclear receptors (Figures 5G–N). LXR agonists have been reported to potentiate immune responses induced by BCG or oxLDL (22, 56). RNA sequencing reanalysis revealed upregulation inflammatory responses upon LXR activation and downregulation of inflammation with pharmacological inhibition of LXR (Figures 6A, B). Inhibition of LXR signaling significantly reduced TNFα production in GM-CSF trained cells, while the decrease in IL-6 levels did not reach statistical significance (Figures 6C, D). In contrast, treatment of GM-CSF–trained cells with the LXR agonists GW3965 (Figures 6E,F) and T0901317 (T13) (Supplementary Figures 5A, B) further amplified proinflammatory cytokine production (7, 19, 22). These results further underscore the critical role of LXR in regulating GM-CSF-training in monocytes.
Figure 6. LXR activation amplifies, whereas inhibition suppresses, GM-CSF–induced trained immunity. RNA-seq data (GSE156696) were aligned to the GRCh38 reference genome using HISAT2. Differentially expressed genes were identified (adjusted p < 0.05, |log2FC| > threshold), and inflammatory pathway enrichment was analyzed using the clusterProfiler package (A, B). Monocytes were pretreated with the LXR agonist GW3965 or the antagonist GSK2033 for 1 h before GM-CSF stimulation, incubated for 24 h, rested for 5 days in fresh medium, and restimulated with LPS (10 ng/mL) for 24 h. IL-6 and TNFα levels were measured in culture supernatants (C–F). Statistical analysis was performed using the Wilcoxon matched-pairs signed-rank test. Data represent mean ± SD from six to eight donors across three independent experiments. p < 0.05, p < 0.01, p < 0.001.
5 Discussion
Innate immune cells – and even some non-immune cells – can undergo long-term functional reprogramming following an initial stimulus, resulting in enhanced responses upon subsequent challenge, a phenomenon termed trained immunity. This process, triggered by microbial components or metabolic cues such as β-glucan, BCG or oxLDL, involves coordinated metabolic and epigenetic rewiring to sustain a heightened inflammatory state (45, 57). A hallmark of trained immunity is a metabolic shift toward aerobic glycolysis, supplying rapid energy and biosynthetic intermediates. In parallel, epigenetic remodeling – particularly histone acetylation, methylation and lactylation – enhances chromatin accessibility at inflammatory gene loci (14, 15, 25, 26, 28). This process depends on acetyl-CoA derived from glycolysis and the TCA cycle, linking metabolism to histone acetyltransferase (HAT) activity (33, 34).
GM-CSF is a key driver of trained immunity, acting on both mature myeloid cells and hematopoietic progenitors (58). Depending on the context GM-CSF exerts disease-promoting and disease-ameliorating effects making it a powerful modulator of inflammation and immunity. For instance, in the context of cardiovascular diseases, GM-CSF is pro-atherogenic: it exacerbates atherosclerosis by promoting infiltration and activation of immune cells, facilitating plaque progression, and possibly increasing vascularization within plaques, thus reducing their stability and increasing the risk of plaque rupture (9, 11, 59). Administration of recombinant human GM-CSF (rhGM-CSF) can benefit patients undergoing bone marrow transplantation by promoting faster hematopoietic reconstitution and enhancing the ability of macrophages to combat infections and tumor cells (60). Autocrine GM-CSF signaling in hematopoietic stem and progenitor cells (HSPCs) promotes metabolic and functional reprogramming, giving rise to monocytes and macrophages with heightened cytokine responses upon re-challenge (15, 58). This memory-like state can be transferred via adoptive HSPC transfer. In macrophages, GM-CSF induces a pro-inflammatory phenotype, enhances phagocytic and cytokine activity and primes cells for augmented secondary responses (39, 61, 62). In the absence of GM-CSF signaling, macrophages show consistent downregulation of glycolytic enzymes (HK2, PFK1, LDHa), reduced glycolytic flux, lower ability to convert glucose to pyruvate and generate energy through glycolysis (53), highlighting GM-CSF’s essential role in sustaining glycolytic metabolism (53, 63).
Our data demonstrate that GM-CSF is a potent inducer of innate immune memory in human monocytes. GM-CSF-primed cells exhibited amplified cytokine responses, associated with increased glycolysis, acetyl-CoA metabolism, and histone modifications at pro-inflammatory loci. Consistent with previous studies, we observed that GM-CSF trains monocytes for elevated glucose uptake and glycolytic flux via upregulation of glucose transporters (GLUT1/3/4) and glycolytic enzymes (PFKFB3) (61, 64, 65). Inhibition of glycolysis or its regulators (e.g., PFKFB3, HIF1α) reduced cytokine output, confirming glycolytic dependency. These changes meet the energetic and biosynthetic demands of activated macrophages and are reflected by elevated glycolytic capacity and flux, lactate production and extracellular acidification rates (ECAR) (64, 65). Conversely, inhibition of MPC intensified inflammatory responses, suggesting that shunting metabolism toward aerobic glycolysis reinforces inflammatory memory in GM-CSF-trained monocytes (50, 66).
Beyond glycolysis, GM-CSF enhanced mitochondrial activity and fatty acid oxidation, evidenced by increased basal respiration and upregulation of FAO-related genes. This apparent dual activation is consistent with the concept of metabolic flexibility, wherein glycolysis and fatty acid oxidation coexist to balance rapid ATP production, biosynthesis, and mitochondrial energy supply (67, 68). Although these processes are often considered metabolically opposed, their concurrent engagement likely reflects a flexible energetic adaptation. Glycolysis provides rapid production of ATP and biosynthetic intermediates, while fatty acid oxidation sustains more efficient mitochondrial ATP production and redox balance. The relative contribution of each pathway may shift depending on cellular energy demand or signaling context, but their coexistence supports an integrated metabolic response rather than mutual exclusivity (67, 68). They represent complementary arms of a flexible metabolic network and cells dynamically adjust their fluxes through these pathways depending on the energy or metabolic substrates requirements (67–69). In addition, previous data suggest that blocking the conversion of pyruvate to acetyl-CoA forces cells to reduce their reliance on glucose-derived carbon entering the TCA cycle and instead shift toward increased fatty acid oxidation (FAO) to sustain mitochondrial energy production (70, 71).
This metabolic rerouting enhances mitochondrial respiration and can create a cellular environment that favors sustained inflammatory signaling (70). In the context of GM-CSF–trained macrophages, such a shift indicates that GM-CSF-induced training may depend more on FAO. This may shed a new light on the metabolic regulation in innate immune memory induction since it was thought that it is mediated solely through increased glycolysis—often considered the hallmark of trained immunity. Our findings suggest that GM-CSF may generate a hybrid metabolic state in which FAO and mitochondrial oxidative phosphorylation provide essential support for amplified cytokine production and long-term functional reprogramming.
This interpretation is consistent with emerging literature showing that metabolic flexibility, particularly the capacity to engage FAO, underlies several forms of innate immune memory. For example, inhibition of pyruvate utilization has been shown to promote memory T cell formation by driving cells toward FAO-dependent energy production (72, 73). By analogy, enhanced FAO in GM-CSF-treated macrophages may support the development of a more persistent, energetically robust inflammatory phenotype. Together, these findings suggest that pyruvate metabolism is a critical regulatory node that balances glycolysis and mitochondrial lipid oxidation, and that tilting this balance toward FAO may potentiate GM-CSF-induced trained immunity.
Further lipidomics analysis provided additional mechanistic insight, revealing that GM-CSF-trained monocytes displayed a marked increase in intracellular triglyceride (TG) levels accompanied by a substantial reduction in cholesteryl ester (CE) storage. This shift in lipid composition suggests a reprogramming of metabolic pathways that preferentially promotes triglyceride accumulation while limiting esterified cholesterol retention, potentially contributing to the inflammatory phenotype observed in GM-CSF–trained cells (74). Reduced CE formation together with increased TG accumulation can promote a more inflammatory macrophage state. This lipid imbalance is associated with altered membrane composition, sustained low-grade inflammation, and may link to pathological phenotypes such as pro-atherogenic macrophage activation and non-alcoholic steatohepatitis (74–77). While this profile indicates a persistent inflammatory state, accumulated triglycerides may also serve as substrates for mitochondrial lipid oxidation (48, 49, 78, 79). This dual metabolic program appears to support both energy production and provision of acetyl-CoA for histone acetylation. Indeed, ChIP-PCR revealed enrichment of H3K27ac at IL-6 and TNFα promoters, linking metabolism to chromatin accessibility. Importantly, inhibition of glycolysis or blockade of acetyl-CoA–related pathways reduced histone acetylation and trained immunity response, underlining the metabolic control of epigenetic remodeling (61, 64, 65).
In addition to its pro-inflammatory effects, GM-CSF upregulated expression of LXR, a nuclear receptor that regulates cholesterol metabolism and plays a significant role in macrophage differentiation and immune function (7). Modulation of LXR signaling can significantly shift immune responses in GM-CSF- or M-CSF-derived macrophages (7, 19). LXR blockade reduced expression of glycolytic and acetyl-CoA-producing enzymes, dampened ECAR/OCR, and reversed epigenetic activation at inflammatory gene loci. These findings align with reports that LXR antagonists reduce glycolysis and trained immunity in human monocytes (22, 56, 80). Previous studies have shown that LXR activation converts M-CSF–induced anti-inflammatory macrophages into a more inflammatory phenotype (19) and induces inflammatory responses in both M-CSF–differentiated macrophages (56) and monocytes cultured in human serum (22). Consistent with these observations, our results showed that LXR activation increased inflammatory cytokine expression in M-CSF–treated macrophages. Notably, in GM-CSF–trained macrophages, LXR stimulation further amplified inflammatory responses, suggesting synergistic effects between GM-CSF signaling and LXR-mediated metabolic reprogramming (Supplementary Figures 4A, B, Figures 6E, F). In addition, pharmacological inhibition of LXR in GM-CSF–primed cells significantly altered cytokine output, highlighting LXR as an important regulatory node in GM-CSF–driven macrophage activation (Figures 6C, D). Furthermore, LXR inhibition lowers inflammatory cytokines in GM-CSF-trained macrophages, partly via upregulation of MAFB, which promotes anti-inflammatory genes (e.g., IL-10, CCL2) and suppresses pro-inflammatory pathways (7, 19). LXR inhibition suppressed glycolysis-driven inflammation in rheumatoid arthritis and trained immunity (22, 56, 80). Specifically, LXR blockade reduced expression of glycolytic enzyme genes (PFKFB3, LDH-A, HK2, G6PD) in macrophages (80, 81), destabilized HIF-1α (80), and lowered GLUT1 expression, thereby limiting glucose uptake. Furthermore, LXR inhibition reduced the expression of acetyl-CoA-producing enzymes, decreasing substrates for biosynthesis and protein acetylation (80, 81).
Activation of the Akt/mTOR signaling pathway is often involved in the development of innate immune memory following vaccination, infection or exposure to stimuli such as oxLDL (82). Interestingly, the GM-CSF-driven training effect seems to be mTOR independently (82). Inhibition of mTOR using a specific inhibitor OSI-27 did not significantly reduce the production of inflammatory mediators in GM-CSF–trained cells (Supplementary Figures 6A, B), consistent with our previous findings on LXR ligand–dependent trained immunity (22). This suggests that although mTOR is a major regulator of HIF1α transcription, GM-CSF may stabilize HIF1α primarily through alternative pathways, such as c-MYC, ERK, or STAT signaling. This is further supported by our transcriptional analysis, which revealed enrichment c-Myc–associated gene signatures, which was upregulated in our dataset and is known to drive glycolysis and support monocyte proliferation and memory formation (83).
Collectively, in the context of host-pathogen interactions, GM-CSF trained immunity may be harnessed to boost immune responses, conversely, in chronic inflammatory diseases to control disease progression, which could be achieved via modulating LXR. Our data position GM-CSF as a key instructive signal in establishing trained immunity via metabolic and epigenetic reprogramming in myeloid cells, which in part facilitated by upregulation of acetyl-CoA producing pathways. LXR signaling emerges as a regulatory checkpoint in this process, linking cholesterol metabolism to innate immune memory. These insights have translational implications: targeting metabolic regulatory molecules particularly LXR pathways may offer novel strategies to modulate maladaptive inflammation in GM-CSF involved chronic inflammatory diseases such as atherosclerosis and rheumatic diseases.
Data availability statement
The original contributions presented in the study are included in the article/Supplementary Material. Further inquiries can be directed to the corresponding authors.
Ethics statement
The study was approved by the Scientific and Ethics Committee of the University of Münster and conforms to the principles of the Declaration of Helsinki. Written informed consent was obtained from all donors by the blood bank and leukocyte reduction filters were provided anonymously without sharing additional personal and detailed information.
Author contributions
YL: Investigation, Writing – original draft. AH: Writing – review & editing, Data curation. HH: Writing – review & editing, Investigation. QZ: Writing – review & editing, Investigation. HK: Writing – review & editing, Investigation. ML: Investigation, Writing – review & editing. NG: Writing – review & editing. GL: Writing – review & editing, Data curation, Investigation. MH: Investigation, Writing – review & editing, Data curation. HF: Writing – review & editing. KP: Writing – review & editing. MN: Writing – review & editing. HR: Writing – review & editing, Funding acquisition. DS: Writing – original draft, Methodology, Conceptualization, Writing – review & editing. YS: Supervision, Writing – review & editing, Investigation, Writing – original draft, Data curation, Conceptualization, Funding acquisition.
Funding
The author(s) declared that financial support was not received for this work and/or its publication.
Acknowledgments
We thank Prof. Ángel Corbí for his valuable comments and insights during the development of this study. YS, HF, NG, and HR are supported by the Faculty of Medicine, University of Münster, YS and HF acknowledge the Dr. Robert Pfleger Foundation (grant ZUW80298). DS is supported by the SFBTRR332 clinical science fellowship program and the fund Innovative Medical Research of the University of Münster Medical School. YL and QZ acknowledges the support of the China Scholarship Council program. KP and MN are supported by Deutsche Forschungsgemeinschaft (DFG) – SFB 1454 – project number 432325352. NG is supported by Deutsche Forschungsgemeinschaft (DFG) – SFBTRR332 – project number 449437943. MN is supported by German Research Foundation (DFG) excellence cluster ImmunoSensation (EXC 2151– project number 390873048). We acknowledge support from the Open Access Publication Fund of the University of Muenster.
Conflict of interest
MN is a scientific founder of Biotrip, Salvina, TTxD and Lemba.
The remaining author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
Generative AI statement
The author(s) declared that generative AI was used in the creation of this manuscript. During the preparation of this manuscript, the authors used ChatGPT (GPT-4o, mini variant; OpenAI, accessed June 2025) to enhance the clarity and readability of the text. All suggestions were critically reviewed, revised where necessary, and approved by the authors, who take full responsibility for the final content.
Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.
Publisher’s note
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.
Supplementary material
The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fimmu.2025.1685796/full#supplementary-material
References
1. Ingelfinger F, De Feo D, and Becher B. GM-CSF: Master regulator of the T cell-phagocyte interface during inflammation. Semin Immunol. (2021) 54:101518. doi: 10.1016/j.smim.2021.101518
2. Hamilton JA. GM-CSF-dependent inflammatory pathways. Front Immunol. (2019) 10:2055. doi: 10.3389/fimmu.2019.02055
3. Petrina M, Martin J, and Basta S. Granulocyte macrophage colony-stimulating factor has come of age: From a vaccine adjuvant to antiviral immunotherapy. Cytokine Growth Factor Rev. (2021) 59:101–10. doi: 10.1016/j.cytogfr.2021.01.001
4. Lotfi N, Zhang GX, Esmaeil N, and Rostami A. Evaluation of the effect of GM-CSF blocking on the phenotype and function of human monocytes. Sci Rep. (2020) 10:1567. doi: 10.1038/s41598-020-58131-2
5. Lee KMC, Achuthan AA, and Hamilton JA. GM-CSF: A promising target in inflammation and autoimmunity. Immunotargets Ther. (2020) 9:225–40. doi: 10.2147/ITT.S262566
6. Avci AB, Feist E, and Burmester GR. Targeting GM-CSF in rheumatoid arthritis. Clin Exp Rheumatol. (2016) 34:39–44.
7. de la Aleja AG, Herrero C, Torres-Torresano M, Schiaffino MT, Del Castillo A, Alonso B, et al. Inhibition of LXR controls the polarization of human inflammatory macrophages through upregulation of MAFB. Cell Mol Life Sci. (2023) 80:96. doi: 10.1007/s00018-023-04745-4
8. De Luca G, Cavalli G, Campochiaro C, Della-Torre E, Angelillo P, Tomelleri A, et al. GM-CSF blockade with mavrilimumab in severe COVID-19 pneumonia and systemic hyperinflammation: a single-centre, prospective cohort study. Lancet Rheumatol. (2020) 2:e465–73. doi: 10.1016/S2665-9913(20)30170-3
9. Haghighat A, Weiss D, Whalin MK, Cowan DP, and Taylor WR. Granulocyte colony-stimulating factor and granulocyte macrophage colony-stimulating factor exacerbate atherosclerosis in apolipoprotein E-deficient mice. Circulation. (2007) 115:2049–54. doi: 10.1161/CIRCULATIONAHA.106.665570
10. Zhu SN, Chen M, Jongstra-Bilen J, and Cybulsky MI. GM-CSF regulates intimal cell proliferation in nascent atherosclerotic lesions. J Exp Med. (2009) 206:2141–9. doi: 10.1084/jem.20090866
11. Shaposhnik Z, Wang X, Weinstein M, Bennett BJ, and Lusis AJ. Granulocyte macrophage colony-stimulating factor regulates dendritic cell content of atherosclerotic lesions. Arterioscler Thromb Vasc Biol. (2007) 27:621–7. doi: 10.1161/01.ATV.0000254673.55431.e6
12. Biwa T, Hakamata H, Sakai M, Miyazaki A, Suzuki H, Kodama T, et al. Induction of murine macrophage growth by oxidized low density lipoprotein is mediated by granulocyte macrophage colony-stimulating factor. J Biol Chem. (1998) 273:28305–13. doi: 10.1074/jbc.273.43.28305
13. Poels K, Schnitzler JG, Waissi F, Levels JHM, Stroes ESG, Daemen M, et al. Inhibition of PFKFB3 hampers the progression of atherosclerosis and promotes plaque stability. Front Cell Dev Biol. (2020) 8:581641. doi: 10.3389/fcell.2020.581641
15. Borriello F, Iannone R, Di Somma S, Loffredo S, Scamardella E, Galdiero MR, et al. GM-CSF and IL-3 modulate human monocyte TNF-alpha production and renewal in in vitro models of trained immunity. Front Immunol. (2016) 7:680. doi: 10.3389/fimmu.2016.00680
16. Sohrabi Y, Godfrey R, and Findeisen HM. Altered cellular metabolism drives trained immunity. Trends Endocrinol Metab. (2018) 29:602–5. doi: 10.1016/j.tem.2018.03.012
17. Mitroulis I, Ruppova K, Wang B, Chen LS, Grzybek M, Grinenko T, et al. Modulation of myelopoiesis progenitors is an integral component of trained immunity. Cell. (2018) 172:147–61.e12. doi: 10.1016/j.cell.2017.11.034
18. Walachowski S, Tabouret G, Fabre M, and Foucras G. Molecular analysis of a short-term model of beta-glucans-trained immunity highlights the accessory contribution of GM-CSF in priming mouse macrophages response. Front Immunol. (2017) 8:1089. doi: 10.3389/fimmu.2017.01089
19. Gonzalez de la Aleja A, Herrero C, Torres-Torresano M, de la Rosa JV, Alonso B, Capa-Sardon E, et al. Activation of LXR nuclear receptors impairs the anti-inflammatory gene and functional profile of M-CSF-dependent human monocyte-derived macrophages. Front Immunol. (2022) 13:835478. doi: 10.3389/fimmu.2022.835478
20. Schulman IG. Liver X receptors link lipid metabolism and inflammation. FEBS Lett. (2017) 591:2978–91. doi: 10.1002/1873-3468.12702
21. Bilotta MT, Petillo S, Santoni A, and Cippitelli M. Liver X receptors: regulators of cholesterol metabolism, inflammation, autoimmunity, and cancer. Front Immunol. (2020) 11:584303. doi: 10.3389/fimmu.2020.584303
22. Sohrabi Y, Sonntag GVH, Braun LC, Lagache SMM, Liebmann M, Klotz L, et al. LXR activation induces a proinflammatory trained innate immunity-phenotype in human monocytes. Front Immunol. (2020) 11:353. doi: 10.3389/fimmu.2020.00353
23. Cuevas VD, Simon-Fuentes M, Orta-Zavalza E, Samaniego R, Sanchez-Mateos P, Escribese M, et al. The gene signature of activated M-CSF-primed human monocyte-derived macrophages is IL-10-dependent. J Innate Immun. (2022) 14:243–56. doi: 10.1159/000519305
24. Ritchie ME, Phipson B, Wu D, Hu Y, Law CW, Shi W, et al. limma powers differential expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Res. (2015) 43:e47. doi: 10.1093/nar/gkv007
25. Law CW, Chen Y, Shi W, and Smyth GK. voom: Precision weights unlock linear model analysis tools for RNA-seq read counts. Genome Biol. (2014) 15:R29. doi: 10.1186/gb-2014-15-2-r29
26. Kim D, Paggi JM, Park C, Bennett C, and Salzberg SL. Graph-based genome alignment and genotyping with HISAT2 and HISAT-genotype. Nat Biotechnol. (2019) 37:907–15. doi: 10.1038/s41587-019-0201-4
27. Ewels P, Magnusson M, Lundin S, and Kaller M. MultiQC: summarize analysis results for multiple tools and samples in a single report. Bioinformatics. (2016) 32:3047–8. doi: 10.1093/bioinformatics/btw354
28. Liao Y, Smyth GK, and Shi W. featureCounts: an efficient general purpose program for assigning sequence reads to genomic features. Bioinformatics. (2014) 30:923–30. doi: 10.1093/bioinformatics/btt656
29. Frankish A, Diekhans M, Ferreira AM, Johnson R, Jungreis I, Loveland J, et al. GENCODE reference annotation for the human and mouse genomes. Nucleic Acids Res. (2019) 47:D766–73. doi: 10.1093/nar/gky955
30. Love MI, Huber W, and Anders S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol. (2014) 15:550. doi: 10.1186/s13059-014-0550-8
31. Yu G, Wang LG, Han Y, and He QY. clusterProfiler: an R package for comparing biological themes among gene clusters. OMICS. (2012) 16:284–7. doi: 10.1089/omi.2011.0118
32. Huber W, Carey VJ, Gentleman R, Anders S, Carlson M, Carvalho BS, et al. Orchestrating high-throughput genomic analysis with Bioconductor. Nat Methods. (2015) 12:115–21. doi: 10.1038/nmeth.3252
33. Bligh EG and Dyer WJ. A rapid method of total lipid extraction and purification. Can J Biochem Physiol. (1959) 37:911–7. doi: 10.1139/y59-099
34. Horing M, Ejsing CS, Krautbauer S, Ertl VM, Burkhardt R, and Liebisch G. Accurate quantification of lipid species affected by isobaric overlap in Fourier-transform mass spectrometry. J Lipid Res. (2021) 62:100050. doi: 10.1016/j.jlr.2021.100050
35. Liebisch G, Lieser B, Rathenberg J, Drobnik W, and Schmitz G. High-throughput quantification of phosphatidylcholine and sphingomyelin by electrospray ionization tandem mass spectrometry coupled with isotope correction algorithm. Biochim Biophys Acta. (2004) 1686:108–17. doi: 10.1016/j.bbalip.2004.09.003
36. Horing M, Ejsing CS, Hermansson M, and Liebisch G. Quantification of cholesterol and cholesteryl ester by direct flow injection high-resolution fourier transform mass spectrometry utilizing species-specific response factors. Anal Chem. (2019) 91:3459–66. doi: 10.1021/acs.analchem.8b05013
37. Lacey DC, Achuthan A, Fleetwood AJ, Dinh H, Roiniotis J, Scholz GM, et al. Defining GM-CSF- and macrophage-CSF-dependent macrophage responses by in vitro models. J Immunol. (2012) 188:5752–65. doi: 10.4049/jimmunol.1103426
38. Fleetwood AJ, Lawrence T, Hamilton JA, and Cook AD. Granulocyte-macrophage colony-stimulating factor (CSF) and macrophage CSF-dependent macrophage phenotypes display differences in cytokine profiles and transcription factor activities: implications for CSF blockade in inflammation. J Immunol. (2007) 178:5245–52. doi: 10.4049/jimmunol.178.8.5245
39. Shi H, Chen L, Ridley A, Zaarour N, Brough I, Caucci C, et al. GM-CSF primes proinflammatory monocyte responses in ankylosing spondylitis. Front Immunol. (2020) 11:1520. doi: 10.3389/fimmu.2020.01520
40. Samaniego R, Palacios BS, Domiguez-Soto A, Vidal C, Salas A, Matsuyama T, et al. Macrophage uptake and accumulation of folates are polarization-dependent in vitro and in vivo and are regulated by activin A. J Leukoc Biol. (2014) 95:797–808. doi: 10.1189/jlb.0613345
41. Rios I, Herrero C, Torres-Torresano M, Lopez-Navarro B, Schiaffino MT, Diaz Crespo F, et al. Reprogramming of GM-CSF-dependent alveolar macrophages through GSK3 activity modulation. Elife. (2025) 14. doi: 10.7554/eLife.102659
42. Jiemy WF, van Sleen Y, van der Geest KS, Ten Berge HA, Abdulahad WH, Sandovici M, et al. Distinct macrophage phenotypes skewed by local granulocyte macrophage colony-stimulating factor (GM-CSF) and macrophage colony-stimulating factor (M-CSF) are associated with tissue destruction and intimal hyperplasia in giant cell arteritis. Clin Transl Immunol. (2020) 9:e1164. doi: 10.1002/cti2.1164
43. Draijer C, Penke LRK, and Peters-Golden M. Distinctive effects of GM-CSF and M-CSF on proliferation and polarization of two major pulmonary macrophage populations. J Immunol. (2019) 202:2700–9. doi: 10.4049/jimmunol.1801387
44. Moorlag S, Khan N, Novakovic B, Kaufmann E, Jansen T, van Crevel R, et al. beta-Glucan Induces Protective Trained Immunity against Mycobacterium tuberculosis Infection: A Key Role for IL-1. Cell Rep. (2020) 31:107634. doi: 10.1016/j.celrep.2020.107634
45. Kleinnijenhuis J, Quintin J, Preijers F, Joosten LA, Ifrim DC, Saeed S, et al. Bacille Calmette-Guerin induces NOD2-dependent nonspecific protection from reinfection via epigenetic reprogramming of monocytes. Proc Natl Acad Sci U.S.A. (2012) 109:17537–42. doi: 10.1073/pnas.1202870109
46. Moorlag S, Folkman L, Ter Horst R, Krausgruber T, Barreca D, Schuster LC, et al. Multi-omics analysis of innate and adaptive responses to BCG vaccination reveals epigenetic cell states that predict trained immunity. Immunity. (2024) 57:171–87.e14. doi: 10.1016/j.immuni.2023.12.005
47. Ziogas A, Novakovic B, Ventriglia L, Galang N, Tran KA, Li W, et al. Long-term histone lactylation connects metabolic and epigenetic rewiring in innate immune memory. Cell. (2025) 188:2992–3012. doi: 10.1016/j.cell.2025.03.048
48. Selen ES, Choi J, and Wolfgang MJ. Discordant hepatic fatty acid oxidation and triglyceride hydrolysis leads to liver disease. JCI Insight. (2021) 6:(2). doi: 10.1172/jci.insight.135626
49. Alves-Bezerra M and Cohen DE. Triglyceride metabolism in the liver. Compr Physiol. (2017) 8:1–8. doi: 10.1002/cphy.c170012
50. Wang D, Wang Q, Yan G, Qiao Y, Zhu B, Liu B, et al. Hypoxia induces lactate secretion and glycolytic efflux by downregulating mitochondrial pyruvate carrier levels in human umbilical vein endothelial cells. Mol Med Rep. (2018) 18:1710–7. doi: 10.3892/mmr.2018.9079
51. Fanucchi S, Dominguez-Andres J, Joosten LAB, Netea MG, and Mhlanga MM. The intersection of epigenetics and metabolism in trained immunity. Immunity. (2021) 54:32–43. doi: 10.1016/j.immuni.2020.10.011
52. Yu X, Wu H, Su J, Liu X, Liang K, Li Q, et al. Acetyl-CoA metabolism maintains histone acetylation for syncytialization of human placental trophoblast stem cells. Cell Stem Cell. (2024) 31:1280–97.e7. doi: 10.1016/j.stem.2024.07.003
53. Wessendarp M, Watanabe-Chailland M, Liu S, Stankiewicz T, Ma Y, Kasam RK, et al. Role of GM-CSF in regulating metabolism and mitochondrial functions critical to macrophage proliferation. Mitochondrion. (2022) 62:85–101. doi: 10.1016/j.mito.2021.10.009
54. Novakovic B, Habibi E, Wang SY, Arts RJW, Davar R, Megchelenbrink W, et al. beta-glucan reverses the epigenetic state of LPS-induced immunological tolerance. Cell. (2016) 167:1354–68.e14. doi: 10.1016/j.cell.2016.09.034
55. Chawla A, Boisvert WA, Lee CH, Laffitte BA, Barak Y, Joseph SB, et al. A PPAR gamma-LXR-ABCA1 pathway in macrophages is involved in cholesterol efflux and atherogenesis. Mol Cell. (2001) 7:161–71. doi: 10.1016/S1097-2765(01)00164-2
56. Findeisen HM, Voges VC, Braun LC, Sonnenberg J, Schwarz D, Korner H, et al. LXRalpha regulates oxLDL-induced trained immunity in macrophages. Int J Mol Sci. (2022) 23:(11). doi: 10.3390/ijms23116166
57. Netea MG, Dominguez-Andres J, Barreiro LB, Chavakis T, Divangahi M, Fuchs E, et al. Defining trained immunity and its role in health and disease. Nat Rev Immunol. (2020) 20:375–88. doi: 10.1038/s41577-020-0285-6
58. Bono C, Guerrero P, Jordan-Pla A, Erades A, Salomonis N, Grimes HL, et al. GM-CSF programs hematopoietic stem and progenitor cells during candida albicans vaccination for protection against reinfection. Front Immunol. (2021) 12:790309. doi: 10.3389/fimmu.2021.790309
59. Singhal A and Subramanian M. Colony stimulating factors (CSFs): Complex roles in atherosclerosis. Cytokine. (2019) 122:154190. doi: 10.1016/j.cyto.2017.10.012
60. Kumar A, Taghi Khani A, Sanchez Ortiz A, and Swaminathan S. GM-CSF: A double-edged sword in cancer immunotherapy. Front Immunol. (2022) 13:901277. doi: 10.3389/fimmu.2022.901277
61. Na YR, Gu GJ, Jung D, Kim YW, Na J, Woo JS, et al. GM-CSF induces inflammatory macrophages by regulating glycolysis and lipid metabolism. J Immunol. (2016) 197:4101–9. doi: 10.4049/jimmunol.1600745
62. Kroner A, Greenhalgh AD, Zarruk JG, Passos Dos Santos R, Gaestel M, and David S. TNF and increased intracellular iron alter macrophage polarization to a detrimental M1 phenotype in the injured spinal cord. Neuron. (2014) 83:1098–116. doi: 10.1016/j.neuron.2014.07.027
63. Menegaut L, Thomas C, Jalil A, Julla JB, Magnani C, Ceroi A, et al. Interplay between liver X receptor and hypoxia inducible factor 1alpha potentiates interleukin-1beta production in human macrophages. Cell Rep. (2020) 31:107665. doi: 10.1016/j.celrep.2020.107665
64. Singh P, Gonzalez-Ramos S, Mojena M, Rosales-Mendoza CE, Emami H, Swanson J, et al. GM-CSF enhances macrophage glycolytic activity in vitro and improves detection of inflammation in vivo. J Nucl Med. (2016) 57:1428–35. doi: 10.2967/jnumed.115.167387
65. Na YR, Hong JH, Lee MY, Jung JH, Jung D, Kim YW, et al. Proteomic analysis reveals distinct metabolic differences between granulocyte-macrophage colony stimulating factor (GM-CSF) and macrophage colony stimulating factor (M-CSF) grown macrophages derived from murine bone marrow cells. Mol Cell Proteomics. (2015) 14:2722–32. doi: 10.1074/mcp.M115.048744
66. Zhong Y, Li X, Yu D, Li X, Li Y, Long Y, et al. Application of mitochondrial pyruvate carrier blocker UK5099 creates metabolic reprogram and greater stem-like properties in LnCap prostate cancer cells in vitro. Oncotarget. (2015) 6:37758–69. doi: 10.18632/oncotarget.5386
67. Muoio DM. Metabolic inflexibility: when mitochondrial indecision leads to metabolic gridlock. Cell. (2014) 159:1253–62. doi: 10.1016/j.cell.2014.11.034
68. Faubert B, Solmonson A, and DeBerardinis RJ. Metabolic reprogramming and cancer progression. Science. (2020) 368:(6487). doi: 10.1126/science.aaw5473
69. O'Neill LA, Kishton RJ, and Rathmell J. A guide to immunometabolism for immunologists. Nat Rev Immunol. (2016) 16:553–65. doi: 10.1038/nri.2016.70
70. Chen Y, Wang J, and Li Y. Mitochondrial pyruvate carrier: a metabolic-epigenetic checkpoint orchestrating T cell differentiation. Signal Transduct Target Ther. (2022) 7:280. doi: 10.1038/s41392-022-01101-z
71. Wenes M, Jaccard A, Wyss T, Maldonado-Perez N, Teoh ST, Lepez A, et al. The mitochondrial pyruvate carrier regulates memory T cell differentiation and antitumor function. Cell Metab. (2022) 34:731–46.e9. doi: 10.1016/j.cmet.2022.03.013
72. Pearce EL, Walsh MC, Cejas PJ, Harms GM, Shen H, Wang LS, et al. Enhancing CD8 T-cell memory by modulating fatty acid metabolism. Nature. (2009) 460:103–7. doi: 10.1038/nature08097
73. van der Windt GJ, Everts B, Chang CH, Curtis JD, Freitas TC, Amiel E, et al. Mitochondrial respiratory capacity is a critical regulator of CD8+ T cell memory development. Immunity. (2012) 36:68–78. doi: 10.1016/j.immuni.2011.12.007
74. Arts RJ, Joosten LA, and Netea MG. Immunometabolic circuits in trained immunity. Semin Immunol. (2016) 28:425–30. doi: 10.1016/j.smim.2016.09.002
75. Li AC and Glass CK. The macrophage foam cell as a target for therapeutic intervention. Nat Med. (2002) 8:1235–42. doi: 10.1038/nm1102-1235
76. Zanotti I, Favari E, and Bernini F. Cellular cholesterol efflux pathways: impact on intracellular lipid trafficking and methodological considerations. Curr Pharm Biotechnol. (2012) 13:292–302. doi: 10.2174/138920112799095383
77. Min HK, Kapoor A, Fuchs M, Mirshahi F, Zhou H, Maher J, et al. Increased hepatic synthesis and dysregulation of cholesterol metabolism is associated with the severity of nonalcoholic fatty liver disease. Cell Metab. (2012) 15:665–74. doi: 10.1016/j.cmet.2012.04.004
78. Welte MA. Expanding roles for lipid droplets. Curr Biol. (2015) 25:R470–81. doi: 10.1016/j.cub.2015.04.004
79. Ipsen DH, Lykkesfeldt J, and Tveden-Nyborg P. Molecular mechanisms of hepatic lipid accumulation in non-alcoholic fatty liver disease. Cell Mol Life Sci. (2018) 75:3313–27. doi: 10.1007/s00018-018-2860-6
80. Zheng DC, Hu JQ, Mai CT, Huang L, Zhou H, Yu LL, et al. Liver X receptor inverse agonist SR9243 attenuates rheumatoid arthritis via modulating glycolytic metabolism of macrophages. Acta Pharmacol Sin. (2024) 45:2354–65. doi: 10.1038/s41401-024-01315-7
81. Dianat-Moghadam H, Khalili M, Keshavarz M, Azizi M, Hamishehkar H, Rahbarghazi R, et al. Modulation of LXR signaling altered the dynamic activity of human colon adenocarcinoma cancer stem cells in vitro. Cancer Cell Int. (2021) 21:100. doi: 10.1186/s12935-021-01803-4
82. Sohrabi Y, Lagache SMM, Schnack L, Godfrey R, Kahles F, Bruemmer D, et al. mTOR-dependent oxidative stress regulates oxLDL-induced trained innate immunity in human monocytes. Front Immunol. (2018) 9:3155. doi: 10.3389/fimmu.2018.03155
Keywords: acetyl-CoA, glycolysis, GM-CSF, histone modification, mitochondrial metabolism, trained immunity
Citation: Liu Y, Hamid A, Hardege H, Zhang Q, Körner H, Leffers M, Gonzalez NA, Liebisch G, Hoering M, Findeisen H, Placek K, Netea MG, Reinecke H, Schwarz D and Sohrabi Y (2026) Active LXR signaling, coupled with elevated mitochondrial and glycolytic metabolism contributes to GM-CSF–induced trained immunity. Front. Immunol. 16:1685796. doi: 10.3389/fimmu.2025.1685796
Received: 14 August 2025; Accepted: 08 December 2025; Revised: 03 December 2025;
Published: 07 January 2026.
Edited by:
Emmanuel Atangana Maze, UMR9198 Institut de Biologie Intégrative de la Cellule (I2BC), FranceReviewed by:
Nidhi Kedia-Mehta, Maynooth University, IrelandSathyabaarathi Ravichandran, Jackson Laboratory for Genomic Medicine, United States
Copyright © 2026 Liu, Hamid, Hardege, Zhang, Körner, Leffers, Gonzalez, Liebisch, Hoering, Findeisen, Placek, Netea, Reinecke, Schwarz and Sohrabi. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
*Correspondence: Dennis Schwarz, RGVubmlzLlNjaHdhcnpAdWttdWVuc3Rlci5kZQ==; Yahya Sohrabi, eWFoeWEuc29ocmFiaUB1a211ZW5zdGVyLmRl