Trained Immunity

Trained immunity (also termed innate immune memory) is the biological process by which cells of the innate immune system—specifically monocytes, macrophages, dendritic cells, natural killer (NK) cells, and bone-marrow hematopoietic stem and progenitor cells (HSPCs)—undergo long-lasting epigenetic and metabolic reprogramming after a primary encounter with a pathogen-associated molecular pattern (PAMP), endogenous danger signal, or live-attenuated vaccine, enabling them to mount a faster and higher-amplitude transcriptional and antimicrobial response upon a subsequent encounter with either the same or a completely unrelated (heterologous) pathogen (Netea et al., 2011; Netea et al., 2020). Formulated and named in 2011 by Romanian-Dutch immunologist Mihai G. Netea and colleagues at Radboud University Medical Center, trained immunity overturned the century-old textbook dogma that immunological memory is the exclusive property of the adaptive immune system (T and B lymphocytes), explaining both how Jawless vertebrates, plants (systemic acquired resistance), and invertebrates survive reinfection without Rag-dependent V(D)J recombination and why live vaccines such as Bacillus Calmette–Guérin (BCG) reduce all-cause infant mortality from unrelated respiratory and diarrheal infections.

Innate immune adaptation spans two opposing functional poles depending on the nature and dose of the initial stimulus:

• Endotoxin Tolerance (Innate Immunoparalysis): High doses of bacterial lipopolysaccharide (LPS) engaging Toll-like receptor 4 (TLR4) epigenetically silence pro-inflammatory cytokine promoters (TNF, IL6, IL1B) via histone deacetylation and H3K9me2/3 methylation to prevent lethal septic shock during Gram-negative bacteremia, while preserving antimicrobial peptide expression (Foster et al., 2007).
• Trained Immunity (Innate Potentiation): Exposure to fungal cell-wall β-1,3-glucan (from Candida albicans via the C-type lectin receptor Dectin-1), mycobacterial muramyl dipeptide (MDP) or BCG vaccine (via the cytosolic receptor NOD2), oxidized low-density lipoprotein (oxLDL), or Plasmodium falciparum hemozoin stably poises inflammatory and antimicrobial enhancers so that secondary stimulation weeks to months later triggers a 2- to 10-fold higher surge of TNF-α, IL-1β, IL-6, and reactive oxygen species (Quintin et al., 2012; Kleinnijenhuis et al., 2012):

Trained immunity is sustained by a tight reciprocal feedback loop between chromatin-modifying enzymes and central cellular metabolism—in which intermediate metabolites of the tricarboxylic acid (TCA) cycle and cholesterol synthesis act directly as obligate substrates, cofactors, or competitive inhibitors of histone methyltransferases and demethylases (Cheng et al., 2014; Arts et al., 2016):

1. Epigenetic Chromatin Poising (H3K4me1, H3K4me3, H3K27ac, and UMLILOs)

In a resting naive monocyte, the promoters and enhancers of pro-inflammatory genes (TNF, IL6, IL1B, PTGS2) are tightly compacted around nucleosomes and transcriptionally silent.

• Primary Stimulation: Engagement of Dectin-1 (via the Raf-1 or Syk–CARD9 pathway) or NOD2 activates NF-κB and AP-1 transcription factors, recruiting MLL/COMPASS histone methyltransferases and CBP/p300 histone acetyltransferases—often guided in cis across topologically associating domains (TADs) by immune-gene-priming long non-coding RNAs (IPLs such as UMLILO, Upstream Master LncRNA of the Inflammatory Chemokine Locus) (Fanucchi et al., 2019).
• Rest Period (The Poised "Latent Enhancer" State): After the primary pathogen is cleared and gene transcription returns to baseline, H3K27ac is partially removed, but monomethylation of histone H3 lysine 4 (H3K4me1) at distal latent enhancers and residual trimethylation (H3K4me3) at gene promoters are retained along with open chromatin accessibility (ATAC-seq peaks), keeping the chromatin bookmarked for rapid RNA polymerase II recruitment upon secondary challenge (Saeed et al., 2014).

2. Metabolic Rewiring: Warburg Glycolysis, Fumarate, and the Mevalonate Pathway

To supply the acetyl-CoA and block the removal of H3K4me3 during the resting phase, trained monocytes remodel three core metabolic pathways:

• Akt–mTOR–HIF-1α Aerobic Glycolysis (The Warburg Shift): Dectin-1 or BCG stimulation activates the PI3K–Akt–mTOR–HIF-1α signaling axis, shifting the monocyte from oxidative phosphorylation toward high-flux aerobic glycolysis (Cheng et al., 2014). Elevated glycolysis generates high cytosolic NAD⁺/NADH ratios and feeds pyruvate into the mitochondria to synthesize citrate, which is exported to the nucleus and cleaved by ATP-citrate lyase (ACLY) into acetyl-CoA to fuel H3K27ac histone acetylation. Blocking mTOR with rapamycin or HIF-1α genetically completely abolishes trained immunity.
• Glutaminolysis and Fumarate Inhibition of KDM5 Demethylases: Trained monocytes upregulate glutaminolysis, converting glutamine into glutamate and α-ketoglutarate to replenish the TCA cycle and accumulate high intracellular concentrations of fumarate (Arts et al., 2016). Because fumarate is a structural analogue of α-ketoglutarate, accumulated fumarate competitively inhibits the α-ketoglutarate-dependent KDM5 family of histone demethylases (which normally erase H3K4me3)—thereby locking H3K4me3 marks onto the TNF and IL6 promoters!
• Mevalonate Accumulation and IGF1R Autocrine Loop: Simultaneously, activation of the cholesterol / mevalonate synthesis pathway leads to the intracellular accumulation of mevalonate (rather than cholesterol itself). Mevalonate binds and activates the insulin-like growth factor 1 receptor (IGF1R) in an autocrine loop, sustaining mTOR activation; consequently, HMG-CoA reductase inhibitors (statins) block mevalonate synthesis and prevent the induction of trained immunity (Bekkering et al., 2018).

Because circulating human classical monocytes (CD14⁺⁺CD16⁻) have a bloodstream half-life of only 1 to 7 days, early critics questioned how BCG vaccination could maintain trained blood monocytes for 3 to 12 months in humans. In 2018, Kaufmann et al. and Mitroulis et al. resolved this paradox by discovering that trained immunity operates at two anatomical levels:

1. Central Trained Immunity (Bone-Marrow HSPCs): Intravenous or intradermal BCG vaccination ( via type II interferon IFN-γ) or systemic β-glucan administration (via IL-1β and GM-CSF signaling plus up-regulated cholesterol biosynthesis) reaches the bone-marrow niche and epigenetically reprograms long-lived hematopoietic stem cells (LT-HSCs) and multipotent progenitors (MPP3) toward myelopoiesis. These centrally trained stem cells continuously produce epigenetically poised daughter monocytes and granulocytes for months to years (Kaufmann et al., 2018; Mitroulis et al., 2018).
2. Peripheral (Tissue-Resident) Trained Immunity: Long-lived self-renewing tissue macrophages—such as alveolar macrophages in the lungs trained by prior adenovirus or influenza infection with help from CD8⁺ T-cell IFN-γ—maintain local autonomous trained immunity inside mucosal barriers (Yao et al., 2018).

Trained immunity is a double-edged sword in human medicine:

• Beneficial Heterologous Vaccine Protection and Cancer Immunotherapy: Epidemiological trials in West Africa (Aaby & Benn) and controlled human challenge studies (Arts et al., 2018) demonstrated that BCG vaccination reduces neonatal mortality by ~30%–45% and reduces experimental yellow fever vaccine viremia in proportion to IL-1β epigenetic training. Furthermore, intravesical instillation of live BCG has served for over 40 years as the gold-standard immunotherapy for non-muscle-invasive bladder cancer, eradicating urothelial tumors via local trained macrophage and NK-cell cytotoxicity.
• Maladaptive Trained Immunity in Atherosclerosis and Autoinflammation: Conversely, sterile endogenous danger signals—such as oxidized LDL (oxLDL), lipoprotein(a), uric acid crystals, and Western high-fat/high-sugar diets (acting via the NLRP3 inflammasome)—induce maladaptive trained immunity in bone-marrow myeloid progenitors, driving persistent arterial wall plaque inflammation in atherosclerosis, gout, hyper-IgD syndrome (where mevalonate kinase deficiency causes constitutive mevalonate accumulation), and neurodegenerative neuroinflammation (Christ et al., 2018).

1. Netea, M. G., Quintin, J., & van der Meer, J. W. M. (2011). "Trained immunity: A memory for innate host defense." Cell Host & Microbe, 9(5), 355–361. https://doi.org/10.1016/j.chom.2011.04.006
2. Saeed, S., Quintin, J., Kerstens, H. H. D., Rao, N. A., Aghajanirefah, A., Matarese, F., et al. (2014). "Epigenetic programming of monocyte-to-macrophage differentiation and trained innate immunity." Science, 345(6204), 1251086. https://doi.org/10.1126/science.1251086
3. Cheng, S.-C., Quintin, J., Cramer, R. A., Shepardson, K. M., Saeed, S., Kumar, V., et al. (2014). "mTOR- and HIF-1α-mediated aerobic glycolysis as metabolic basis for trained immunity." Science, 345(6204), 1250684. https://doi.org/10.1126/science.1250684
4. Kaufmann, E., Sanz, J., Dunn, J. L., Khan, N., Mendonça, L. E., Pacis, A., et al. (2018). "BCG educates hematopoietic stem cells to generate protective innate immunity against tuberculosis." Cell, 172(1–2), 176–190.e19. https://doi.org/10.1016/j.cell.2017.12.031
5. Netea, M. G., Domínguez-Andrés, J., Barreiro, L. B., Chavakis, T., Divangahi, M., Fuchs, E., et al. (2020). "Defining trained immunity and its role in health and disease." Nature Reviews Immunology, 20(6), 375–388. https://doi.org/10.1038/s41577-020-0285-6