Innate And Adaptive Immune Memory: An Evolutionary Continuum in The Host's Response To Pathogens(Part 2)

To learn more info plz contact david.wan@wecistanche.com

Central versus Tissue Innate Immune Memory

To understand the mechanisms responsible for the induction of trained immunity, we need to define the two levels at which trained immunity acts. The first level is represented by the cell populations undergoing reprogramming (e.g., hematopoietic stem cell progenitors; see below), and the second level is the intracellular processes responsible for the reprogramming of each cell (e.g., epigenetic and metabolic reprogramming of the cell; see Epigenetic Mechanisms Mediate Induction of Innate immune Memory).

Since the vast majority of myeloid cells, such as monocytes, granulocytes, and dendritic cells, are short-lived, the question arises how innate immune cells maintain and propagate the observed innate immune memory phenotype beyond their own life-span over a period of up to 3 months and longer. Several seminal studies have shed light on some of the mechanisms contributing to these processes. Specifically, it was shown that systemic application of the fungal cell wall component β-glucan leads to a modulation of the transcriptomic, metabolomic, and functional properties of the hematopoietic progenitor cascade in the bone marrow, in turn generating more myeloid cells with a faster kinetic at the expense of the lymphoid lineage(Mitroulis et al,2018). Mechanistically, these effects were attributed to IL1β and GM-CSF signaling events, with induction of cholesterol metabolism and enhanced glycolysis leading to more robust production of myeloid effector cells upon LPS re-challenge. Similar studies in mice injected with BCG demonstrated the effect of vaccination on remodeling myelopoiesis, that in turn mediates improved innate host defense against mycobacteria (Kaufmann et al.,2018). Moreover, in a model of high-fat-diet-induced innate immune training, similar effects on the myelopoiesis have been shown, highlighting the potentially deleterious effects of lifestyle on the reactivity of the immune system at least partly explaining the overt immune activation phenotypes observed in obese individuals (Christ et al., 2018).

So far, studies have only elucidated the effects of systemically applied training stimuli. However, physiologically, topical innate immune training seems very likely as the lung, the mucosae, or the skin are regularly exposed to a wide array of microbial constituents. Indeed, innate immune training can be induced in a skin wound healing model utilizing topical administration of the TLR3 ligand Poly I: C, leading to an enhanced ability to regenerate the skin after injury. This process of enhanced wound repair was dependent on sustained signaling of AIM2 within the reservoir of epithelial stem cells within the affected area of the skin, clearly showing that innate immune training can also be elicited locally, independent of immune cells (Naik et al, 2017). Moreover, earlier studies on the lung suggest that innate immune cells in the lung are indeed able to remember their inflammatory history. Studies investigating the effect of two unrelated subsequent viral infections, LCMV, and influenza A virus, clearly showed that a first infection exerts the ability to alter a secondary innate immune response indicating a degree of innate immune training in these models of viral infections (Mehta et al, 2015). Beyond host defense, the consequences of topical innate immune training or inflammatory memory for the development of autoimmune and auto-inflammatory disorders remain to be investigated. Recently, a study highlighted the importance of prior immune activation on the development of asthma in the setting of a latent gammaherpesvirus infection, which in turn protected affected mice from the development of allergic asthma (Machiels et al,2017). Interestingly, this protective phenotype was accompanied by a replacement of the embryonically derived alveolar macrophage pool with monocyte-derived alveolar macrophages displaying an immune-regulatory phenotype, thereby raising the question of what role tissue-resident macrophages play in the induction and/or maintenance of organ-specific innate immune training.

Taken together, these studies provide evidence that innate immune memory is induced in vivo in two main compartments: centrally in the bone marrow influencing the functional program of immune cell progenitors, and peripherally in the tissues. Especially tissues exposed to the outside world possess the capacity to mount an innate immune training response. This raises the question of how this process is balanced to provide enhanced host defense and counteract the development of auto-inflammatory disorders.

Click here to learn more about Cistanche

Epigenetic Mechanisms Mediate Induction of Innate Immune Memory

The central feature of trained innate immune cells is the ability to mount a qualitatively and quantitatively different transcriptional response when challenged with microbes or danger signals. Evidence supports the convergence of multiple regulatory layers for mediating innate immune memory, including changes in chromatin organization, DNA methylation, and probably non-coding RNAs such as microRNAs (miRNAs) and/or long non-coding RNA (ncRNAs). In myeloid cells, many loci encoding inflammatory genes are in a repressed configuration during homeostasis (Ramirez-Carrozzi et al.,2006,2009; Saccani et al., 2001). Upon primary stimulation, there is a strong gain in chromatin accessibility, increased acetylation, and RNA polymerase ll recruitment. These changes are driven by the recruitment of stimulation-responsive transcription factors (e.g., NF-kB, AP-1, and STAT family members) to enhancers and gene promoters, which are usually pre-marked by lineage-determining transcription factors such as PU.1 (Barozzi et al.,2014; Ghisletti et al., 2010; Heinz et al,2010; Smale and Natoli,2014). In turn, transcription factors control the recruitment of coactivators (including histone acetyltransferases and chromatin remodelers)(Ramirez-Carrozziet al.,2009; Ramirez-Carrozziet al.,2006) that locally modify chromatin to make it more accessible to the transcriptional machinery. Maintenance of such enhanced accessibility underlies the more efficient induction of genes upon restimulation(Foster et al.,2007). One interesting paradigm is provided by latent or de novo enhancers (Kaikkonen et al,2013; Ostuni et al.,2013); these are genomic regulatory elements that are un-marked in unstimulated cells but gain histone modifications characteristic of enhancers (such as monomethylation of histone H3 at K4, H3K4me1)only in response to specific stimuli. In vitro, upon removal of the stimulus, a fraction of latent enhancers retain their modified histones and can undergo a stronger activation in response to restimulation (Ostuni et al.,2013).

Recent studies have investigated the changes in epigenomic programs in innate immune cells during induction of trained immunity. One early study proposed that changes in epigenetic status underlie the repression of inflammatory genes during LPS tolerance (Foster et al,2007).In contrast, during LPS tolerance, the genes involved in anti-microbial responses were either not affected, or their expression was increased (Fos-ter et al.,2007).In turn, exposure of monocytes/macrophages to C.albicans orβ-glucan modulated their subsequent response to stimulation with unrelated pathogens or PAMPs, and the changed functional landscape of the trained monocytes was accompanied by epigenetic reprogramming(Quintin et al, 2012; Saeed et al.,2014).BCG vaccination has also been shown to result in an increase in inflammatory mediators in monocytes from healthy volunteers, which correlated with parallel changes in histone modifications associated with gene activation (Arts et al.,2018; Kleinnijenhuis et al.,2012), as well as with changes in the pattern of DNA methylation (Verma et al.,2017).

Similar to monocytes and macrophages, the induction of CMV-induced NK cell memory is accompanied by dynamic chromatin structure changes (Lau et al.,2018) and at least partially relies on epigenetic reprogramming, which is linked to reduced expression of the transcription factor promyelocytic leukemia zinc finger (PLZF) (Schlums et al,2015) and the tyrosine kinase SYK(Lee et al,2015). Interestingly, a recent comparative study of chromatin structure genome-wide level in mCMV-induced memory NKand CD8Tcells revealed that epigenetic signatures of NK and CD8 T cells, even though very different in naive cells, become similar in effector and memory cells. The few genes that share epigenetic and transcriptional programs in memory NK and CD8 T cells (for example Bach2, Tcf7, and Zeb2) are known to drive the differentiation of CD8 T cells to memory phenotype, suggesting common epigenetic mechanisms underlying memory formation in adaptive and innate immune cells(Lau et al.,2018). Human CMV also drives epigenetic imprinting of the IFNG locus in NK cells, which leads to consistent IFNy production in NKG2C(hi) NK cells, providing a molecular basis for the adaptive feature of these cells (Luetke-Eversloh et al.,2014).

Thus, it can be concluded that epigenetic rewiring is the molecular substrate that sits at the basis of the enhanced response of innate immune cells upon a secondary stimulation(Figure 3).

Classical Adaptive Immune Memory

Changes in chromatin structure accompany not only innate immune memory formation but also adaptive immune memory. The adaptive immune system consists of B and T lymphocytes, which express highly antigen-specific receptors, namely B cell receptor(BCR) and T cell receptor (TCR) emerging during somatic gene recombination(Figure 4), a feature unique to these cells. There have been excellent reviews about the role of T and B cells during an immune response (De Silva and Klein,2015; Kurosakiet al.,2015)and the development of memory of T and B cells once an immune reaction has been resolved. More recently, memory T and B cells have been further subdivided by their location and differential functionalities(Jameson and Masopust, 2018; Kumar et al., 2018).

Although neither naive nor memory T and B cells express effector molecules and they possess largely similar transcriptional programs, their response to secondary TCR or BCR stimulation differs qualitatively and quantitatively (Akondy et al, 2017; Klein et al.,2003). Therefore, the question arises: what are the mechanisms responsible for a more effective yet specific response of lymphocytes during the infection with the same pathogen? These two properties of the adaptive immune response are mediated by two fundamentally different types of mechanisms: first, the higher magnitude and speed of the response are mediated by epigenetic programming, while, second, the specificity of the response is ensured by gene recombination of TCR and BCR and clonal expansion of specific cell subpopulations upon antigen recognition.

Epigenetic Programming in Memory Lymphocytes

To achieve the faster and more pronounced reactivity of T and B lymphocytes upon reinfection with the same pathogen, epigenetic regulation is an ideal regulatory system allowing differential functionality of a cell without losing its identity. On the molecular level, epigenetic mechanisms are essential for the regulation of gene expression (Jaenisch and Bird,2003).DNA methyltransferases, chromatin remodeling, and histone-modifying enzymes rear-range chromatin structure at gene regulatory elements and regulate the accessibility of DNA for the transcriptional machinery. miRNAs silence gene expression at the level of transcription or translation and IncRNAs can either foster or inhibit chromatin interactions.(Houri-Zeevi and Rechavi,2017). Notably, once induced, epigenetic changes persist over time and are preserved through cell divisions, reflecting the past of a cell and allowing them to pass these"memories" to a daughter cell. Therefore, epigenetic regulation might not only be a hallmark of developmental and differentiation processes but explains molecularly the cellular hallmark of T and B cell memory, namely increased magnitude and faster onset of response. Importantly, transcriptional and epigenetic regulation also controls cell proliferation and clonal expansion, a key process of T and B cell memory.

Figure 3. The amplitude of Immune Memory: Epigenetic Mechanisms

Epigenetic rewiring underlies both the adaptive characteristics of innate immune cells during trained immunity and amplification of the response in memory adaptive immune cells. The silencing of effector genes in naive immune cells is maintained by suppressive histone marks, such as H3K27me3. Initial activation of gene transcription is accompanied by loss and gain of specific chromatin marks such as H3K27me3 and H3K4me3, respectively, which are only partially maintained after the elimination of the stimulus. The enhanced status of the innate immune cells, mirrored by the persistence of histone marks such as H3K4me3-and H3K4me1-characterizing"latent enhancers," results in non-specific stronger response to secondary stimuli upon re-challenge.

Recently, the German Epigenome Programme (DEEP) generated a genome-wide epigenetic dataset for human peripheral naive, central, and effector memory CD4+ T cells (Durek et al., 2016). This study showed a progressive loss of DNA methylation, the epigenetic mark mainly associated with gene silencing, from naive to central to effector memory CD4t T-cells. Many regulatory elements that showed decreases in DNA methylation during naive to memory CD4+ T cell transition were linked to genes known to be involved in CD4+T cell differentiation, such as T-bet, IL2 receptor subunits, RUNX3.DNA methylation changes also accompany the differentiation of naive CD8+ T cells into memory cells. Interestingly, DNA methylation was shown to be enriched at loci coding for genes characteristic of naive T cells, such as CD62L or CCR7. Vice versa, genes associated with memory CD8+T cells including T-bet and EOMES show enriched DNA-methylation in naive CD8+ T cells, again suggesting that epigenetic regulation allows differential gene expression between naive and memory T cells (Abdelsamed et al.,2017; Youngblood et al.,2017). If loss of DNA methylation is one of the epigenetic mechanisms allowing memory T cells to react faster and with a higher magnitude, one would propose that gene loci for T cell effector molecules should be de-methylated in memory T cells but not in naive T cells. In fact, loci of genes responsible for CD8+Tcell effector function such as IFNg, Per-forin, granzyme B (GZMB), and GZMK are methylated in naive CD8+ T cells and become demethylated in memory CD8+T cells. Moreover, the methylation profile remains largely unchanged during homeostatic proliferation of memory CD8+T cells in the absence of an antigen (Abdelsamed et al.,2017; Youngblood et al.,2017).

A rearrangement of the DNA methylome has also been observed during the differentiation of naive B cells to germinal center (GC) B cells and to memory cells, with more profound changes between naive and GC B cells than between GC cells and memory and plasma cells(Kulis et al.,2015; Lai et al.,2013).In fact. memory B and plasma cells, although relatively distinct transcriptional, possess similar DNA-methylomes, while naive and memory B cells show more differences in DNA-methylomes despite their similar transcriptional programs. Moreover, among differentially methylated regions were enhancers enriched in transcription factor (TF) binding sites, especially those involved in B cell differentiation (Kulis et al,2015; Lai et al,2013). These observations suggest that decreased levels of DNA methylation at regulatory elements of effector molecules reflect the history of antigen encounter and facilitate faster and more pronounced expression of the effector genes upon reencounter of antigen.

DNA-methylation is not the only epigenetic modification that changes during naive T-and B-cell activation and persists in memory cells. In addition, histone modification patterns on regulatory elements allow the prediction of the subsequent gene expression status. Several studies addressed these changes during T-cell differentiation. A genome-wide analysis of histone modifications (H3K4me1, H3K4me3, and H3K27ac) in human naive, central memory, and effector memory CD4+ T cells revealed that loci coding for various cytokines including IFNY, IL4, IL13, IL17, and IL22 are enriched in activating histone marks as compared to naive cells. In addition, they are more rapidly induced in memory than in naive T cells upon stimulation (Barski et al.,2017; Durek et al,2016). Similar patterns were observed at gene loci coding for T-bet and RORC transcription factors known to be important regulators for effector molecules induced in memory CD4*T cells upon stimulation that, for instance, drive IFNγ and IL17 expression， respectively. Furthermore， there is a positive correlation between a gain of H3K4me3 in memory CD4+ T cells at gene loci that are more inducible in memory cells than naive T cells, suggesting their poised status in memory cells (Barski et al.,2017).

Additionally, epigenetic studies have been conducted on CD8*T cell populations (Henning et al.,2018). The loss of activating H3K4me3 and H3K9ac histone marks and gain of suppressive H3K27me3 were observed on genes downregulated in effector CD8+T cells (such as FOXO1, KLF2, LEF, and TCF7), while the loss of H3K27me3 and/or gain of H3K4me3  were found on genes coding for effector molecules(such as Eomes, TNFa, IFNg, GZMB, CD27, BLIMP1, CCR7, or SELL)in memory cells (Araki et al., 2008; Russ et al.,2014). Interestingly, activating H3K9ac modification was increased in memory cells at loci of genes part of signaling pathways downstream of the TCR. This suggests that not only effector molecules but also early signal transduction can be quickly boosted upon secondary antigen experience in memory T cells(Rodriguez et al,2017). Whether this effect is also responsible for the higher magnitude seen in memory T cells requires further investigation.

Furthermore, activating H3K4me1 and H3K27ac histone modifications were shown to be enriched in effector and memory CD8+ T cells at gene loci induced upon activation of naive cells, IL7r and Id2 for example(Yu et al.,2017)(Crompton et al,2016). Similar to CD4+ T cells, genes highly inducible upon stimulation in memory CD8+ T cells were characterized by enriched H3K4me3 and depleted H3K27me3 modifications at their respective gene loci when compared to naive cells (Araki et al.,2009; Russ et al.,2014). Moreover, the genome-wide distribution of H3K4me3 and H3K27me3 marks in memory cells were more similar to effector than to naive CD8+ T cells (Cromp-ton et al.,2016; Russ et al.,2014).

Figure 4. Specificity of Immune Memory; Antigenic Recognition and Clonal Expansion

Classical adaptive immune memory is induced by antigen presentation to specialized lymphocytes in the lymph nodes. After proliferation, elimination of the pathogen, and contraction, a small number of memory lymphocytes persists to insure long-time specific memory of the target pathogen.

An interesting group of cells that might bridge the development from naive to effector T-cells is recently described as stem-cell-like memory CD8+ T(SCM) cells. Tsim possesses a partially naive phenotype (CD44low CD62L) and memory characteristics∶ of high expression of IL2Rßt and CXCR3+， increased proliferation potential, and cytokine release in response to antigen re-stimulation (Gattinoni et al.,2011; Gattinoni et al.,2009; Zhang et al,2005), and dependence on IL-15 and IL-7 for homeostatic turnover (Cieri et al,2013). The development of Tsim cells is also accompanied by epigenetic changes(Abdel-samed et al.,2017; Akondy et al.,2017). Transcriptomic and histone modification analysis in vitro generated human Tscm suggest that this cell population consists of a developmental continuum from naive via Tsim to effector and memory T cells. Crompton et al.showed a progressive upregulation and downregulation of signature genes from naive T cells, Tsim, and effector T cells to T memory cells, accompanied by the progressive acquisition of H3K4me3 and loss of H3K27me3 histone marks (Cromp-ton et al,2016). Chromatin structure changes were further expanded to the DNA methylome analysis which showed a progressive loss of DNA methylation during the development of Tsim cells from naive T cells(Abdelsamed et al.,2017), indicating that Tsim derives from naive T cells. This hypothesis was challenged however by the observation of long-lived CD8+ T cells generated in yellow fever vaccinated individuals, these authors proposed that Tsim is generated from effector T cells (Akondy et al.,2017). Although the origin of Tsim needs to be better investigated, it is clear that the chromatin structure consists of a basis of enhanced cytokine response and proliferative potential of Tsim.

These data suggest that the elevated expression potential of a gene in memory T-cells is encoded in its chromatin structure and molecularly resembles the functional hallmarks of a higher magnitude and faster response onset. It is primarily gained upon antigen-specific cell activation of naive T cells, then pre-served in memory T cells during in vivo homeostasis.

Less is known about histone modifications in naive and memory B cells, but the assessment of H3K4me1, H3K4me3, and H3K36me3. H3K27me3 and H3ac showed that human B cell subpopulations have very distinct and specific epigenetic profiles. The transition from naive to GC B cells was found to be associated with a gain of activating histone marks H3K4me1, H3K4me3, and H3ac on genes induced during GC formation and a loss of these marks on genes that become silenced (Zhang et al,2014b). Despite little knowledge being available about changes in histone modification landscapes during B cell commitment to memory subsets, several studies show the importance of histone-modifying enzymes in memory B cell formation.

New research has also shed light on the enzymes and mechanisms responsible for the epigenetic programming of memory in lymphocytes. Histone acetyltransferase monocytic leukemia zinc finger protein (MOZ) is a histone modifier found to be important for proper GC and memory B cell formation (Good-Jacob-son et al,2014).MOZ is required during B-cell activation and it was suggested that MOZ-induced histone modification during a primary response can alter the dynamics of secondary responses by affecting the memory B cell repertoire. As DNA methylation plays a role in memory B cell formation mutation of DNA, methyltransferase 3 beta (Dnmt3b) is correlated with a lack of plasma and memory B cells in ICF syndrome(immunodeficiency, centromere instability, and facial anomalies syndrome)(Blanco-Betancourt et al.,2004; Hansen et al,1999; Xu et al., 1999).Dnmt3 deletion early during CD8* T effector cell differentiation resulted in decreased DNA methylation levels and re-expression of genes associated with naive cell state, therefore attenuating the formation of memory cells (Youngblood et al, 2017). In CD4+ T cells, several histone-modifying enzymes, including H3K9 methyltransferase SUV39H1 and Jumonji Domain Containing 3(Jmjd3) H3K27 demethylase, control naive T-cell commitment to effector cells by regulating effector cytokines and transcription factor expression. In the absence of SUV39H1 or Jmjd3, once committed Th1, Th2, or Th17cellsstart to re-express cytokines of another lineage, suggesting that histone-modifying enzymes are required in CD4+ T cells to "remember" their original transcriptional programs (Allan et al, 2012; Let al.,2014).

All in all, a large body of evidence demonstrates the important role played by epigenetic programming in mediating the changes in the magnitude and kinetics of T and B lymphocytes during the induction of immune memory (Figure 3).

The second important property of the adaptive immune memory is represented by the specificity of the responses. It is ensured by the expression of highly specific receptors and immunoglobulins (g)by T and B cells. To be effective, a highly specific immune response requires huge diversity of receptors and antibodies, which is achieved by somatic rearrangement of gene segments coding for TCR and LG. In the classical process of V(D)J recombination, hundreds of gene segments, called variable (V), diversity (D), and joining (J), are assembled into one V-D-J exon. This"cut and paste" process is driven by RAG enzymes (encoded by recombination-activation genes) specifically expressed and indispensable in maturating lymphocytes. V(D)J recombination results in millions of TCR and antibody variants able to recognize and neutralize millions of various antigens. After the successful rearrangement of its receptor, mature B or T cells express functional BCR (composed of transmembrane lgs) or TCR, respectively, ready for an antigen encounter. The presence of RAG proteins is strictly associated with the DNA re-arrangement process, and the appearance of RAG genes during evolution has been believed for decades to be a core stone for the development of adaptive immunity. To date, RAG homologs have been described in many jawed vertebrates but not in jaw-less vertebrates such as lampreys or hagfishes(Kumar et al., 2015). Despite the lack of recombination-activation genes, the immune response of jawless vertebrates does exhibit characteristics of adaptive immunity. This is mediated through lymphocytes carrying antigen-specific variable lymphocyte receptors (VLR) that emerge during somatic DNA rearrangement. Some VLRs can be secreted extracellularly and serve as antibodies (Herrin and Cooper, 2010).

The lymphocyte that carries an antigen-specific receptor undergoes clonal expansion upon activation by the antigen enriching the poll of immune cells in those able to recognize the encountered antigen. Clonal expansion of antigen-activated B and T cells not only assures a better defense during primary infection but also makes the immune response more efficient upon secondary infection by the same agent. Elevated numbers of memory cells and antibody-producing plasma cells generated during clonal expansion augment the chances of an encounter with antigen, making the secondary immune response much faster and more efficient (Campos and Godson,2003). However, despite changes in numbers and relocation of memory cells to sites with increased chances of meeting an antigen (Aiba et al., 2010; Sathaliyawala et al.,2013), there might be intrinsic changes that allow memory cells to react more pronounced and faster to secondary immune challenges (Barski et al,2017).

Figure 5. A Proposed Two-Step Model for the Evolution of Immune Memory

The first step is represented by ancient evolutionary epigenetic processes that insure an increased magnitude of the response to a second infection, and this characterizes both innate and adaptive memory. The second step evolved in vertebrates and insures that the memory is highly specific toward a certain pathogen, by involving the development of specific memory cells selected from a large repertoire obtained through gene recombination.

An Evolutionary Model Involving Two Steps for the Development of Immune Memory

In this Perspective, we reappraised the various mechanisms responsible for the induction of the two major forms of immune memory: classical adaptive immune memory and innate immune memory. Both forms are characterized by an improved response of the host after reinfection and have evolved to enhance the chances of host survival in an environment teeming with potentially lethal pathogenic microorganisms.

Based on the data presented above, we would like to propose a comprehensive concept as a basic framework for understanding the properties of immune memory in various groups of multi-cellular organisms. In this model, we propose that immune memory is a general characteristic of the host defense of all living organisms. The evolution of immune memory in various groups of organisms is a continuum that started with the development of epigenetic mechanisms responsible for increasing the magnitude and speed of the immune response upon reinfection and continued thereafter with the build-up of specificity in vertebrates by mechanisms including gene recombination and clonal selection. Magnitude and kinetics amplification by epigenetic re-wiring characterizes thus a more primitive form of innate immune memory, while both higher magnitude/kinetics and specificity characterize the refined adaptive immune memory in vertebrates (Figure 5).

This is not to be seen as a static model, but it can change upon novel discoveries in the years to come. We can envision, for example, that new forms of adaptive immune memory will be described in complex invertebrate animals. Indeed, in line with the assessments that the advantages of building adaptive immune memory are especially obvious in long-lived organisms, it is conceivable and maybe even likely, that forms of specific adaptive immunity will be described also some groups of com-plex invertebrate animals.

From a molecular perspective, we need to better understand how the epigenetic changes following an initial immune activation are translated to achieve more pronounced secondary responses with a faster onset. What are the epigenetic mechanisms that allow such an adaptation in the behavior of immune cells? When is this behavior evolutionary beneficial? Why do only some stimuli lead to such a cellular response? And, as a consequence, can we identify rules that would allow us to predict immune memory responses to a given chemical entity? What are the modern-life situations that trigger immune memory outside our evolutionary understanding? Is immune memory in these incidences always detrimental? Very recent examples suggest that an unexpected induction of innate immune memory by the Western Diet (Christ et al.,2018) or in the context of Alzheimer's disease (Wendeln et al.,2018) shows the flop-side of an evolutionarily conserved immune mechanism. Furthermore, what is the role of locally induced innate immune memory, and how do tissues erase these memories if needed?

Other questions relate to the specificity of the epigenetic mechanisms and changes induced during the induction of immune memory. Why are only some gene loci affected, and are they from certain classes? Mechanistically, what are the similarities and differences between the epigenetic changes observed in adaptive and innate immune memory? An important area of future research will be to identify and describe the memory characteristics of non-immune cells (e.g., epithelial cells, stromal cells, etc. Indeed, very recent studies have shown epigenetic-cally-mediated long-term changes in epithelial precursors(Naik et al.,2017; Ordovas-Montanes et al.,2018), with important roles in tissue defense and regeneration. The use of newly developed technologies, including single-cell omics sequencing, will represent important support to answer many of these questions within the next decade.

Finally, in addition to a better understanding of immune memory at mechanistic and conceptual levels, we hope that the description of both adaptive and immune memory will lead to a more efficient design of vaccines. Indeed, one can envision that vaccines that are capable of inducing both forms of immune memory at the same time would be more effective. A clear understanding of the processes driving immune memory at all levels is crucial to achieving this aim.