Discovering adaptive features of innate immune memory

Conventionally, it was thought that innate immunity operated through a simple system of nonspecific responses to an insult. However, this perspective now seems overly simplistic. It has become evident that intricate cooperation and networking among various cells, receptors, signaling pathways, and protein complexes are essential for regulating and defining the overall activation status of the immune response, where the distinction between innate and adaptive immunity becomes ambiguous. Given the evolutionary timeline of vertebrates and the success of plants and invertebrates which depend solely on innate immunity, immune memory cannot be considered an innovation of only the lymphoid lineage. Indeed, the evolutionary innate immune memory program is a conserved mechanism whereby innate immune cells can induce a heightened response to a secondary stimulus due to metabolic and epigenetic reprogramming. Importantly, the longevity of this memory phenotype can be attributed to the reprogramming of self‐renewing hematopoietic stem cells (HSCs) in the bone marrow, which is subsequently transmitted to lineage‐committed innate immune cells. HSCs reside within a complex regulated network of immune and stromal cells that govern their two primary functions: self‐renewal and differentiation. In this review, we delve into the emerging cellular and molecular mechanisms as well as metabolic pathways of innate memory in HSCs, which harbor substantial therapeutic promise.


| E VOLUTIONARY SUCCE SS OF INNATE AND ADAP TIVE IMMUNE SYS TEMS
To struggle for survival, all living organisms must evolve with a defense mechanism, and the outcome of this tug of war depends on the immunity of both host and microbe.While there are several hierarchies, nomenclatures, and classifications to describe the complexities of immunity, the fundamental role of the immune system is to discriminate self from nonself.Indeed, multicellularity followed shortly after the ability to eliminate non-self via phagocytosis, which was evident even in the deepest branches of Archaea and Bacteria phylogeny. 1,2From an evolutionary perspective, the origin of the most ancient immune system is thought to be derived from prokaryote and bacterial ancestors.In fact, the origin of the first eukaryotes is described by the endosymbiotic theory, interpreted from the Greek "endo," "syn," and "bios," to mean the close association of cells, one inside the other. 3This theory describes a process whereby bacterial genes entered the eukaryotic lineage via organelle ancestors 4 ; the most cited instances being the formation of the mitochondrion as a product of an archaeal cell and an aerobic proteobacterium, 4,5 and the chloroplast deriving its origin from the integration of a cyanobacterium into a heterotrophic eukaryote. 6This has been demonstrated by the conserved prokaryotic biochemistry of both these organelles, and has been described as the first two endosymbiotic events afforded by gene transfer from prokaryotes. 7e implications of symbiosis have been extended to the innate immune system-by definition, survival demands that immune genes be adaptable and malleable.In fact, innate immune defenses exhibit a fossil record of shared genetic factors across several species.
The conservation of innate immune signaling modalities between prokaryotes and eukaryotes points toward common ancestry over convergent evolution.For example, the cyclic GMP-AMP synthase (cGAS)-stimulator of interferon genes (STING) pathway is an antiviral innate immune strategy; it acts as a sensor for double-stranded DNA which leads to the activation of STING and results in the transcription of interferon and other antiviral factors. 8This pathway is found in several invertebrates, humans, and even bacteria.Bacterial cGAS protects against phage infection and has conserved protein architecture to its human counterpart.Wein et al. proposed that these cell-autonomous innate proteins first evolved in prokaryotes as anti-phage strategies, and through an endosymbiotic event these defense strategies were inherited by eukaryotes. 5e origin for the study of innate immunity was marked by Elie Metchnikoff, who identified myeloid-like cells in starfish larvae phagocytosing a foreign body (a rose thorn), in 1882.The innate immune system is characterized by its rapid responses via a range of pathogen recognition receptors (PRRs), which are germline-encoded proteins allowing for the recognition of pathogen extracellular architectures (e.g., the bacterial cell wall) and/or their intracellular nucleic acids (e.g., viral DNA/RNA).Toll-like receptors (TLRs) were the first family of PRRs that Janeway speculated to recognize conserved microbial products. 80][11] This quality is contrasted to the somatically assembled antigen receptors of the adaptive arm of the immune system, which allow for an unlimited molecular reservoir of pathogen recognition and present as a key evolutionary novelty of vertebrates.A notable feature of the adaptive immune response is the ability to respond with greater speed to a secondary encounter of the same antigen, a process termed immune memory-however this does not appear to be an innovation of the lymphoid lineage.
While the endosymbiotic theory has been postulated to describe the development of the innate immune system, it has been proposed that the adaptive immune system is a unique merger of cellular innovation (the lymphocyte) and the repurposing of an ancient, autonomous toxin-antitoxin program. 12,13Lymphocytes recognize highly specific antigen configurations of pathogens and subsequently trigger clonal amplification, cellular differentiation, and the production of antigen-specific antibodies. 14However, the distinction between innate and adaptive immunity cannot simply be drawn by "nonspecific" versus "specific", by the diversity of somatically generated antigen receptors, or even by the capacity to generate memory responses.

F I G U R E 1 Somatic diversification in the vertebrate subphylum. (A)
The branching of jawless (agnathans) and jawed (gnathosomes) vertebrates is considered the evolutionary birth of somatic diversification.While gnathosomes were the first organisms found to possess T and B cells with TCR, BCR, and MHC genes, agnathans were also determined to have lymphocytes but with recombinatorial assembly of a different type of modular genetic units, to generate variable lymphocyte receptors (VLRs). 14In experimental models, agnathans have been reported to generate specific agglutinins after a primary and secondary stimulation; however, these were not similar to the gnathosome immunoglobulin-like structure. 22(B) The specificity and memory capacity of T and B cells are conferred by the variable (V) domains of the Ig and TCR.A seminal study 18 uncovered the first antigen receptor chain that employs two V domains, the NAR-TCRδ.This molecule observed in sharks has a specialized TCR δV domain which bears similarity to the TCR subtype IgNAR (new antigen receptor) of a cartilaginous fish ancestor, dating back 200 million years.Direct antigen recognition by the NAR-TCR was observed using the typical γδ TCR as a scaffold, and given the lack of MHC-restriction of these T cells, these findings support the notion that γδ TCRs have afforded evolutionary freedom to vertebrates from a very early gnathostome ancestor.
The concept of somatic diversification was thought to have evolved at the branching of jawless and jawed vertebrates (Figure 1).
The central elements of our adaptive immune defenses are shared by all jawed vertebrates. 14It was conventionally considered that the adaptive immune system in jawed vertebrates was evolved as a result of the horizontal transfer of the recombination activating gene (RAG) transposon from an invading organism into an immunoglobulin superfamily gene member; this led to the initiation of antigen receptor gene rearrangement and the optimization of preexisting effector mechanisms of innate host defense. 15,16Jawless vertebrates (agnatha) were once thought to be devoid of adaptive immunity; however, the discovery of variable lymphocyte receptors (VLRs), which are the foundation of the second convergent adaptive immune system in jawless fish, challenged this conventional dogma.
Jawless vertebrates indeed possess an arm of defense with similar cellular selection features, but with receptors based on TLR genes rather than immunoglobulin genes and T-and B-cell systems. 13,14is finding presents an interesting crossroad: while the cellular principles were in place for jawless vertebrates, the evolving system opted not to implement this cellular adaptive immune system. 13terestingly, the comparison of the lymphocyte lineages in jawed and jawless vertebrates via VLRs shows similarities with γδT cells, which recognize microbe antigens in an MHC-independent manner and can perform both innate and adaptive functions. 17,18wless vertebrates are not the only or even the oldest eukaryotes to lack cellular adaptive immunity-plants and invertebrates also lack a somatic adaptive immune system and mobile immune cells 19,20 and according to fossil records the first land plants were established approximately 480 million years ago. 21This deficiency of the kingdom plantae and subphyla of invertebrates, informs a redundancy or repetition of biological features in the adaptive immune system.Instead of adaptive immunity, plants depend on two branches of their innate immune system: pattern-triggered immunity (PTI) and effector-triggered immunity (ETI).In PTI, cell surface pattern recognition receptors (PRRs) recognize microbe-associated or damageassociated molecular patterns and through a series of signaling cascades and kinase activation steps, lead to the transcriptional induction of pathogen-responsive genes, the production of reactive oxygen species, and the deposition of callose to reinforce the plant cell wall, etc. 20,21 In response, phytopathogens have acquired resistance mechanisms, for instance by interfering with recognition, and thus plants in turn established a more specialized mechanism of detection.ETI describes the production of disease resistance proteins that respond to effector molecules shed by pathogens to establish infection.A common outcome in ETI is programmed cell death induction at the site of infection with the aim of limiting pathogen spread.ETI can also lead to systemic acquired resistance (SAR): the production of mobile immune signals such as methyl salicylic acid (MeSA), which are transported systemically to uninfected tissues for transcriptional programming and the production of antimicrobial proteins to allow for protection from a subsequent exposure.SAR has been designated as canonical evidence for the capacity of innate immune system to establish memory to a previous insult. 23vertebrates, which represent 99% of animal diversity and which singularly possess innate immune systems, have also been evidenced to possess immune memory. 24Despite being an exceptionally diverse taxa, memory phenotypes following challenge have been evidenced in several subphyla including, but not limited to, arthropods, 25 mollusks, 26 and ctenophores. 27Very recent work using the invertebrate model of copepods 28 has uncovered through infection and secondary challenge with species of their natural parasite, tapeworms, that this protective recall ability can be attributed specifically to "innate immune memory."Using metabolic and transcriptomic analyses, it was determined that this innate memory could respond in a specific manner (homologous secondary challenge) by regulating histone modification on gene splicing or in an unspecific manner (heterologous secondary challenge) which involved metabolic rewiring. 28e importance of the innate immune system is reflected by its existence and remarkable diversity at almost every level of the evolutionary tree of life.Considering the evolutionary success of innate immunity, which is the only mode of host defense in about 99% of all species on Earth, 29 it is imperative to consider that a critical evolutionary trait like immunological memory must have been evolved in the innate arm of immune system.In this review we will discuss the cellular and molecular mechanisms involved in "innate immune memory," which has been termed "trained immunity" in vertebrates.

| TR AINED IMMUNIT Y
Trained immunity is underscored by epigenetic and metabolic rewiring of innate immune cells and has been observed in many different cell types including but not limited to monocytes/macrophages, 30,31 natural killer (NK) cells, 32 neutrophils, 33 and innate lymphoid cells (ILCs). 34Induction of trained immunity has been documented to occur by both sterile factors (i.e., oxidized low-density lipoprotein, oxLDL 35 ) or microbial agents and factors (i.e., the BCG vaccine strain of Mycobacterium bovis, 31,36,37 the fungal cell wall component β-glucan 36,38,39 ).To describe the changes involved during training, there must be a delineation of chronological stages: baseline, primary stimulation, return to baseline, and secondary stimulation. 40At baseline, due to low-energy demand, chromatin remains condensed and there is reduced basal metabolic activity.Upon primary stimulation, a cellular response is initiated by PRRs or cytokines, which leads to a large energy expenditure.Metabolic activity is enhanced often involving an augmentation of glycolysis in innate immune cells (e.g., macrophages) or lymphocytes (e.g., T cells). 41Despite being an energetically inefficient source of ATP, this shift from mitochondrial oxidative phosphorylation (OXPHOS) to glycolysis (known as the "Warburg effect") leads to increased ATP production due to the speed of reaction, mediated by increased expression of glycolytic enzymes which is attributed to an epigenetic modification.
Following the removal of the initial stimulus, the host enters a resting phase, wherein histone modifications are incorporated at sites known as "latent enhancers." 23The epigenetic reprogramming of trained immunity has been extensively reviewed. 42Briefly, these genomic regulatory elements are normally unmarked; they acquire histone modifications during the return to baseline stage of trained immunity to allow the chromatin surrounding the promoters of proinflammatory and metabolic genes to remain accessible and enable a faster response upon second stimulation. 42This return to baseline status differentiates trained immunity from immune "priming," in which the secondary challenge occurs at the metabolically active and pro-inflammatory state of the initial stimulus. 43Finally, during the second stimulation of trained immune cells, the same or a different stimulus induces a metabolic and pro-inflammatory response at a superior degree than the primary stimulation. 42e protective effects of trained immunity have been observed to last for months to years after primary stimulation, 36,44 but the effectors of trained innate immune cells (e.g., monocytes and neutrophils) possess a relatively short lifespan, on the scale of days. 45This has led to the discovery that trained immunity can occur at the level of the long-lived hematopoietic stem cell (HSC) in the bone marrow, where their epigenetic reprogramming have been transmitted to progenitors and subsequently circulating lineage-committed innate immune cells. 30,36,46With these findings, the definition has been extended to incorporate that training can occur both at the level of the bone marrow HSCs (central trained immunity) as well as in mature, tissue-resident, or circulating immune cells in peripheral sites, such as the lungs 47 (peripheral trained immunity).However, many questions remain as to the migratory cues, signaling cascades, and cell-cell interactions within the melting pot of the bone marrow, and the regulatory mechanisms of how these elements must harmonize to lead to a "protective" training phenotype.

| BONE MARROW ARCHITEC TURE AND HEMATOP OIE TIC ARCHITEC TS
The bone marrow (BM) is a highly organized tissue containing a mosaic of microenvironments to support different cellular niches.The bone marrow contains various hematopoietic and stromal cells, the latter of which include sympathetic nerves, adipocytes, osteoblasts, osteocytes, and various fibroblasts. 48The most well-studied niche is that of the HSCs: nonmotile cells that reside close to sinusoids, endosteum, and perivascular cells. 48,49During hematopoiesis, HSCs give rise to various subsets of multipotent progenitors (MPPs) which become lineage-committed and segregate broadly into erythroid, myeloid, or lymphoid progenitor pools. 50HSCs are maintained with the help of growth factors and signals produced by mesenchymal and perivascular stromal cells as well as endothelial cells, including CXCL12 for BM retention, 49 stem cell factor (SCF) for HSC regeneration and angiopoietin for regeneration of the sinusoids. 50,51Cs require many growth factors, membrane bound ligands, and secreted gene products for which the cellular sources remain uncertain. 50Concurrently, there are many distinct clusters of leukocytes that reside in the BM and their roles to influence the HSC niche are only recently becoming apparent. 52For instance, neutrophils are fast responders to injury, abundant in cell number, and turnover at a rapid rate (approximately every 12 h); upon their senescence, they must return to the BM for recycling and clearance by macrophages. 53This process triggers signaling that reduces the size of the HSC cellular niche, mediated by a decrease in CXCL12 and egress of progenitors into the circulation and periphery. 53The magnitude of neutrophil activity, as determined by their rate of senescence in the BM, has been hypothesized to inform the demand for macrophage recruitment to peripheral tissues. 53BM-resident macrophages themselves have also been reported to directly impact the HSC niche, by promoting the retention of HSCs.In one study, the depletion of BM phagocytic macrophages in two different in vivo models led to the inhibition of HSC-supportive cytokine expression (including CXCL12, angiopoietin, and kit-ligand), suppression of endosteal osteoblasts, and the mobilization of HSC and progenitors into the circulation. 54In another independent study, it was demonstrated that reductions in BM mononuclear phagocytes (both monocytes and macrophages) via clodronate administration led to reduced CXCL12 and egress of HSCs and progenitors into the bloodstream.
Upon further characterization, it was demonstrated that specific deletion of CD169 + macrophages (and not CD169 − monocytes) was sufficient to induce HSC egress, and that this effect was mediated by crosstalk between macrophages and Nestin + mesenchymal stem cells (MSCs). 55Finally, another branch of leukocytes in the BM that has garnered much interest is the niche of mature T cells.[58] In striking resemblance to HSCs, 59 memory T cells must balance two essential tasks: the maintenance of self-renewal over differentiation and the control of cell migration into the periphery. 60To do so the "two-niche hypothesis" was proposed, 60 in which memory T cells in the BM are localized in either a "proliferative niche" or a "quiescence niche" in which the BM microenvironment would allow for either maintenance or mobilization upon stimulation.However, many questions remain as to the purpose of such a large memory T-cell reservoir in the BM and its influence on HSCs.In a seminal study, Belkaid and colleagues 61 uncovered that dietary restriction induced an accumulation of memory T cells in the BM which led to BM remodeling, involving increased adipocytes and T-cell trophic factors such as CXCL12.The influx of T cells was shown to protect the memory pool from unfavorable conditions and improve their function upon a secondary challenge, demonstrated using a model of acute oral bacterial infection. 61In addition to the memory pool of T cells, FoxP3 regulatory T cells (Tregs) have also been observed to modulate HSC quiescence.The BM contains a high frequency of FoxP3 Tregs relative to total CD4 + T cells as compared to both the spleen and lymph nodes. 62Using intravital microscopy techniques, it has been shown that Tregs were colocalized with the HSPC niche at the endosteal surface.It was proposed that this localization allowed for discrete sites of immune privilege to prevent allogenic rejection after transplantation and shielded HSPCs from autoimmunity or excess inflammation. 62Building on this observation, another group show that these HSC niche Tregs are marked by their expression of HSC marker CD150, and that the selective depletion of these cells expanded the numbers of HSCs in the BM. 63Importantly, they found that adenosine produced by Tregs protected HSCs from oxidative stress and maintained their quiescence. 63In addition to Treg support for HSC function and protection against immune attack, it has also been suggested that Tregs are entrapped by HSCs to construct a pro-survival niche against premature aging. 64llectively, the fluctuations of neutrophils, macrophages, and memory T cell pools highlight a crucial host strategy to adapt and remodel BM microenvironments and cellular niches during nonhomeostatic conditions to preserve immune function and defense.
Given the critical role of HSCs to maintain all hematopoietic lineages, their functions must be tightly regulated preventing dysregulated immune signaling.One such tightly regulated feedback loop within HSCs is their cellular metabolism.

| H SC ME TABOLIS M AND FUN C TI ON
The hallmark of HSCs is their dual capacity for self-renewal and multipotent differentiation. 65Considering fully differentiated cells have a limited lifespan, the ability of HSCs to both replenish aged mature cells as well as perpetuate themselves via self-renewal is critical throughout their lifetime.While the maintenance of the HSC population depends on the balance between self-renewal and differentiation, stress factors (e.g., infection) can become catastrophic, resulting in the decline of the HSC population and subsequent exhaustion.This can also explain why HSCs lack phagocytic machinery, to prevent them from directly being infected by pathogens. 30,36,66,67Therefore, understanding the regulatory mechanisms involved in HSC decisions during homeostasis or stress conditions is essential to demarcate their contribution to trained immunity.
The equilibrium between HSC self-renewal and differentiation is bioenergetically and biosynthetically is regulated via cellular metabolism (Figure 2).Under steady state, HSCs utilize glycolysis, glutamine, and fatty acid metabolism to support self-renewal and survival.However, when HSCs undergo differentiation, a burst of robust mitochondrial metabolism (OXPHOS) supplies the energy for this rapid transition. 68,69For instance, it has been shown that the inhibition of mitochondrial respiration in HSCs via conditional inactivation of protein tyrosine phosphatase mitochondrial 1 (Ptpmt1) blocked HSC differentiation and resulted in rapid hematopoietic failure. 70Despite the relative hypervascularity of the bone marrow cavity, oxygen levels are lower than in the venous circulation due to higher levels of oxygen consumption by proliferating hematopoietic cells. 71Hypoxic conditions in the bone marrow cavity are therefore required for the maintenance of HSC self-renewal capacity via glucose and fatty acid metabolism.Thus, it is not surprising that the hypoxic environment of the bone marrow informs the utilization of anaerobic glycolysis rather than mitochondrial OXPHOS for ATP production.In support of this, it has been determined that one of the key transcription factors that is essential for the HSC response to hypoxia is hypoxia inducible factor 1 (HIF1), which promotes glycolysis. 72,73Depletion of HIF1α in HSCs resulted in decreased glycolysis, increased OXPHOS, and subsequent HSC exhaustion. 74Consequently, a key feature of selfrenewing HSCs is their ability to limit mitochondrial respiration in order to remain in a quiescent state. 74,75terestingly, cellular metabolism also plays a crucial role in orchestrating the epigenetic reprogramming of innate immune cells, driven by trained immunity.It is widely recognized that metabolites have the ability to influence the activity of the enzymes responsible for modifying chromatin structure. 76Therefore, the metabolic restructuring of innate immune cells is pivotal in regulating their adaptability and epigenetic alterations during trained immunity.For instance, β-glucan-trained monocytes exhibit heightened aerobic glycolysis, a characteristic process facilitated by the AKT-mTOR-HIF1α pathway.Blocking this pathway effectively diminished trained immunity. 41Moreover, the induction of trained immunity by BCG vaccination necessitates a metabolic shift in monocytes toward aerobic glycolysis, enabling them to mount a more robust response upon subsequent stimulation. 77However, our understanding of how trained immunity alters HSC metabolism is incompletely understood.Collectively, these studies indicate a two-way dialogue between signals from the BM microenvironment and HSC-intrinsic networks, which regulate HSC metabolism and subsequently the balance between self-renewal and differentiation.

| INTERPL AY B E T WEEN C Y TOK INE S AND H SC FUN C TI ON
HSCs exist in a quiescent state and are maintained over the lifespan of an organism to replenish hematopoietic populations with age and use; however, they have also been implicated as firstresponders during immune responses to infection. 78HSCs express pathogen PRRs such as TLRs and are known to be intricately regulated by Type I and Type II interferons (IFNs) and TLR agonists.The Type I interferon IFNα, has been demonstrated to induce HSCs to exit G 0 and enter the active cell cycle via STAT1 and AKT1 signaling, resulting in the upregulation of stem cell antigen 1 (Sca-1). 79terestingly, this stimulatory effect on HSCs was not observed in vitro, 79 implicating cellular crosstalk in the BM microenvironment.It has been determined that VEGF signaling mediated at least in part by HSCs induced stimulation of endothelial cells, resulting in increased BM vascularity and vascular leakage. 80While acute IFNα stimulation in vivo induced HSC proliferation, 79,81 chronic exposure has been observed to impair HSC function and lead to exhaustion. 79Similarly, Type II interferon (IFNγ) signaling has been observed to mediate activation of quiescent long-term HSCs during bacterial infection with Mycobacterium avium. 66Again, a short-term IFNγ exposure during infection accelerated HSC differentiation to enhance myeloid development as an innate immune response. 82Conversely, the effect of chronic IFNγ exposure during M. avium infection induced a defect in HSC self-renewal and increased susceptibility to stress-induced apoptosis, leading to HSC exhaustion. 83This was also seen in sterile chronic IFNγ exposure, where it was mechanistically demonstrated that IFNγ disrupted the interaction between HSCs and CXCL12-abundant reticular cells, leading to HSC migration dependent on the cell surface protein BST2. 84Indeed, it has been shown that IFNγ modulates the BM microenvironment; IFNγ-treated BM mesenchymal stromal cells displayed impaired support for the HSC niche by the increased expression of interleukin-6 (IL-6) and SCF, which ultimately led to impaired HSC differentiation. 85ide from interferons, the pro-inflammatory cytokine interleukin-1 (IL-1) has also been shown to influence the balance between differentiation and self-renewal in HSCs.In isolated HSC culture and in vivo, it has been shown that IL-1 directly regulates HSC fate, leading to accelerated cell division and differentiation into the myeloid lineage via activation of the PU.1 transcriptional circuit.Not unexpectedly, while acute exposure to IL-1 allowed for myeloid recovery, long-term exposure reversibly ablated HSC self-renewal, transplantation capacity, and lineage output. 86Excessive IL-1 production has also been linked to a model of hematopoietic aging, in which the production of IL-1β by endosteal mesenchymal cells was responsible for creating a pro-inflammatory BM milieu, HSC expansion, and increased myelopoiesis. 87Importantly, this phenotype was reversible by acute IL-1 blockade which induced delayed niche ageing. 87 has become increasingly apparent that demand-driven hematopoiesis, during non-homeostatic conditions and in a prolonged state, can induce harmful effects to the balance of classical stem cell functions.Recent breakthroughs on the signaling cascades underlying central trained immunity 30,36,38 depend on the signaling cascades that are known to modulate HSCs: interferons (Figure 3).F I G U R E 2 HSC function pivots on cellular metabolism.During homeostatic conditions, HSCs undergo anaerobic glycolysis, fatty acid and glutamine metabolism for energy production.This mechanism is regulated by the hypoxic conditions in the HSC niche, which induce hypoxia-inducible factor (HIF)-1α to drive the expression of pyruvate dehydrogenase kinase (PDK) 2 and 4, to prevent pyruvate from entering the TCA cycle and triggering downstream mitochondrial oxidative phosphorylation (OXPHOS).HSCs are reported to engage mitochondrial OXPHOS for functions involving high-energy demand, such as differentiation.These mechanisms highlight the importance of cellular metabolism to modulate the dichotomy of HSC function: self-renewal, which is dependent upon anaerobic glycolysis to maintain quiescence, and differentiation, which requires rapid and heightened levels of ATP production available singularly by mitochondrial OXPHOS.

| Type II IFN
One of the earliest works to pioneer the field of central trained immunity, was mechanistically established by signaling via Type II interferon, or IFNγ. 30In this study we showed that when the BCG vaccine gained direct access to the BM by intravenous (iv) injection, this led The protective effect of BCG-iv was also sustainable in the absence of continuous BCG stimulus, as shown through antibiotic-treated BM adoptive transfer experiments, with protection maintained despite no detectable bacterial burden in the BM.
In both central and peripheral trained immunity, IFNγ plays an important role for protection against diverse models of infection.
In a study of Leishmania donovani infection, BM IFNγ production by CD4 + T cells was found to be indispensable to mediate the expansion of activated Ly6C hi monocytes, displaying higher proportions of monocyte expression of MHCII and iNOS and greater control of BM parasite load. 88Looking outside the BM, training by respiratory adenoviral infection has been identified to lead to IFNγ production by effector CD8 + T cells.This allowed for the development of memory alveolar macrophages with rapid antimicrobial responses. 47Type II IFN has also been linked to the generation   89 Despite these advances, the source of IFNγ and timing of its production in this model of central trained immunity has yet to be elucidated.

| Type I IFN
Given the well-established role of Type I IFNs to modulate the BM-HSC niche, its function to generate modulate immunity would also appear highly likely.Indeed, the central trained immunity induced by fungal-derived β-glucan has been credited with the transcriptomic and epigenetic rewiring of granulopoiesis. 33In an unprecedented work, Chavakis and colleagues demonstrated a protective antitumor phenotype for reprogrammed neutrophils at the level of the BM, in a process necessitating Type I interferon signaling (abrogated in IFNAR1 −/− mice) but not requiring the adaptive arm of immunity (demonstrated using Rag1 −/− mouse model). 33In contrast, in a model of virulent bacterial infection, the protective effects of Type I IFN in the BM were not corroborated.During Mtb-iv infection and unlike BCG-iv vaccination, our group has revealed a Type I IFN-dependent reprogramming of HSCs that led to suppressed myelopoiesis and defective trained immunity to a subsequent Mtb challenge.The detrimental role of type I IFN was tied to dysregulated iron metabolism profile in the BM, as well a necroptotic cell death program observable specifically in the myeloid compartment. 36Interestingly, we have also shown that treatment of mice with two doses of β-glucan induces myelopoiesis via IL-1 and provides a remarkable protection against subsequent Mtb infection via monocytes/macrophages. 38However, unpublished data from our group showed that treatment of mice with one dose of β-glucan promotes granulopoiesis and provides protection against influenza A virus (IAV) infection in a type I IFN dependent, but IL-1 independent manner.This phenotype was associated with increased mitochondrial mass and oxidative respiration evidenced by increased oxidative phosphorylation (OXPHOS) in neutrophils.
These trained neutrophils were essential to reduced morbidity and mortality without affecting IAV load, implicating their functions in promoting disease tolerance.Taken together, the role of Type I IFNs to induce trained immunity depends on the magnitude of signaling involved in training agents as well as context dependent.
Thus, understanding the early mechanisms by which signaling is initiated and sustained presents important future directions for the field.

| CON CLUS IONS
The traditional differentiation between innate and adaptive immunity, which was once solely based on criteria like nonspecific versus specific responses, as well as the vast array of somatically generated antigen receptors and the ability to generate memory responses, is now becoming blurred.A growing body of evidence indicates that innate immunity can display characteristics typically associated with adaptive immunity, and vice versa.For instance, we have recently shown that while the BCG vaccine generates antigen-specific effector memory T cells, these memory T cells can imprint on alveolar macrophages in an IFNγ-dependent manner, providing long-term cross-protection against influenza viral infection. 90A similar study also demonstrated that the BCG-mediated T cell/IFNγ axis provided cross-protection against influenza and SARS-CoV-2 infection via enhanced antiviral immunity in lung epithelial cells. 91This brings another layer of complexity to immunological memory: the dialogue between adaptive and innate memory responses, a phenomenon which is currently in its infancy.Nonetheless, we envision that novel discoveries will emerge from this exciting field of the research.Additionally, how HSC metabolism switches from quiescence to an actively proliferating state presents another gap in the field.
While cytokine signaling is central to trained immunity, several questions remain unanswered surrounding the induction of training.
First, how does central trained immunity ensure protection and not exhaustion in HSCs, and how is the duration of the signal controlled?Second, where are the signaling sentinels (such as IFNs) produced in the BM in response to inducers of trained immunity, and how are BM-resident leukocytes involved in this crosstalk?And finally, is there a unique cytokine profile, which is linked to a training agent for inducing central trained immunity and is this is associated with HSC metabolism?Although several epigenetic mechanisms have been linked to trained immunity [92][93][94] whether the epigenetic reprogramming of HSCs is maintained following self-renewal, or only transmitted to progenitors and subsequent differentiated cells, and how this occurs is still undefined.We speculate that the activation of these to transcriptional reprogramming of HSCs and MPPs, and an expansion of the myeloid lineage.BM-derived macrophages (BMDMs) from BCG-iv vaccinated mice displayed enhanced protection against virulent Mycobacterium tuberculosis (Mtb) infection, as well as epigenetic reprogramming.Using a murine parabiosis model with CCR2 −/− mice (defective macrophage recruitment) and an adoptive transfer model with RAG1 −/− mice (deficient in adaptive immune cells), these BCG-iv-trained macrophages were shown to confer in vivo protection against pulmonary Mtb infection.Using mice deficient in IFNγ receptor (IFNγR −/− ) to derive BMDMs, BCG-iv-mediated protection was reversed against Mtb infection, implicating a key role for IFNγ signaling in macrophage training to enable a protective phenotype.

F I G U R E 3
Effects of vaccine, adjuvant, or pathogenic stimuli on HSC reprogramming.There are few well-described stimuli of trained immunity, among which are the BCG vaccine and the fungal cell wall component β-glucan.While BCG-mediated trained immunity is afforded by Type II IFN signals in the LKS + HSC compartment, the effects of β-glucan are dependent upon the IL-1 and Type I IFN signaling pathways.These key cytokines induce metabolic and epigenetic rewiring of HSCs, leading to the induction downstream of myelopoiesis.Interestingly, in a dose-dependent fashion, β-glucan can bias hematopoiesis either toward granulopoiesis (one dose) or myelopoiesis (two doses).In direct contrast, the presence of Mycobacterium tuberculosis in the HSC niche impairs the training program via the Type I IFN signaling cascade.This leads to mitochondrial stress in HSCs and progenitor cells, mediated by Type I IFN-dependent disruption of mitochondrial membrane potential (ΔΨm).
of peripheral trained immunity to counter Candida albicans infection.In peripheral blood mononuclear cells (PBMCs) isolated from patients with STAT1-dependent chronic mucocutaneous candidiasis (CMC) and STAT3-dependent hyper-immunoglobulinemia E syndrome (HIES), training induced by C. albicans was defective and IFNγ production by NK cells was diminished only in CMC patient samples.When NK cells were depleted from PBMC samples of HIES patients, these cells displayed defective training measured by TNFα and IL-6 production, as compared to healthy patient controls.
Another unexplored area of research in trained immunity is the link between stem cell metabolism and trained immunity.Although many studies have demonstrated that cellular metabolism is central to the memory response in fully differentiated immune cells mainly by switching from mitochondrial OXPHOS metabolism to the glycolytic pathway, how the reprogramming of HSCs via a training agent (e.g., β-glucan or BCG) affects the cellular metabolism of HSCs and progenitor cells remains unknown.This is particularly important as the dual function of HSCs in self-renewal and differentiation requires a unique cellular metabolism, which is atypical in comparison to trained fully differentiated immune cells.Distinct metabolic programming of HSC, progenitor, and differentiated cells indicates that their regulation is controlled epigenetically and/or via external factors, such as the level of oxygen that changes significantly from the bone marrow (hypoxia) to the blood circulation (normoxia).