comment Trained immunity, tolerance, priming and diferentiation: distinct immunological processes The similarities and diferences between trained immunity and other immune processes are the subject of intense interrogation. Therefore, a consensus on the defnition of trained immunity in both in vitro and in vivo settings, as well as in experimental models and human subjects, is necessary for advancing this feld of research. Here we aim to establish a common framework that describes the experimental standards for defning trained immunity. Maziar Divangahi, Peter Aaby, Shabaana Abdul Khader, Luis B. Barreiro, Siroon Bekkering, Triantafyllos Chavakis, Reinout van Crevel, Nigel Curtis, Andrew R. DiNardo, Jorge Dominguez-Andres, Raphael Duivenvoorden, Stephanie Fanucchi, Zahi Fayad, Elaine Fuchs, Melanie Hamon, Kate L. Jefrey, Nargis Khan, Leo A. B. Joosten, Eva Kaufmann, Eicke Latz, Giuseppe Matarese, Jos W. M. van der Meer, Musa Mhlanga, Simone J. C. F. M. Moorlag, Willem J. M. Mulder, Shruti Naik, Boris Novakovic, Luke O’Neill, Jordi Ochando, Keiko Ozato, Niels P. Riksen, Robert Sauerwein, Edward R. Sherwood, Andreas Schlitzer, Joachim L. Schultze, Michael H. Sieweke, Christine Stabell Benn, Henk Stunnenberg, Joseph Sun, Frank L. van de Veerdonk, Sebastian Weis, David L. Williams, Ramnik Xavier and Mihai G. Netea

rained immunity has been defined bacteria (for example, Bacille Calmette– pathogens. Epidemiological studies have as one form of adaptation of innate Guérin (BCG)), fungi (for example, Candida shown that vaccination with certain live Thost defense mechanisms or a de albicans) or helminth parasites (for example, vaccines provides heterologous protection facto innate immune memory. Following Nippostrongylus brasiliensis) protects against against unrelated pathogens. For example, exposure to particular infectious agents heterologous infections independently of BCG vaccination in newborn children or vaccines, trained immunity can mount adaptive immunity1. provides protection not only against a faster and greater response against a Furthermore, while the rationale tuberculosis but also against respiratory secondary challenge with homologous or underlying the use of adjuvants in vaccine tract infections and neonatal sepsis, even heterologous pathogens1. Trained formulations is to improve the efficacy and it significantly reduces mortality1. immunity has emerged as a focal point in of adaptive immunity, little attention has BCG-induced trained immunity has also immunology research and has added a layer been given to the direct effects of adjuvants been shown to provide protection against of complexity to our previous understanding on innate immunity and early protection experimental models of yellow fever11 of immune memory, that is, a trait limited against infection. For instance, β-glucan and malaria12 infection. Interestingly, the to antigen-specific responses of the adaptive (mainly encountered as a component of anticancer effects of BCG (for example, in immune system. Although more than 95% fungal cell walls that activates dectin-1) bladder cancer) have also been linked to of species (plants and invertebrates) rely enhances resistance to acute infection with trained immunity13. Thus, trained immunity solely on innate immunity for host defense2, Staphylococcus aureus4 or chronic infection is an evolutionary trait that increases immunological memory has been associated with Mycobacterium tuberculosis (Mtb)5. the fitness of plants, invertebrates and mainly with the adaptive arm of the immune Similarly, administration of agonists of vertebrates against pathogenic microbes. response in vertebrates. However, it is highly NOD-like receptors (for example, NOD2) It is also important to emphasize that, unlikely that a critical evolutionary trait or Toll-like receptors (for example, TLR9) although trained immunity improves the like immunological memory is restricted have been shown to provide protection host’s defense against subsequent pathogenic to adaptive immunity and has not evolved against Toxoplasma6 and sepsis caused by threats, it may also be maladaptive in the in the innate arm of immunity in the Escherichia coli7, respectively. Intriguingly, context of chronic inflammatory disease, entire spectrum of living organisms. In the induction of trained immunity is such as atherosclerosis1. Indeed, in addition fact, systemic acquired resistance (SAR) regulated by a unique set of mediators. For to microbial products, trained immunity can is a well-defined state corresponding to instance, BCG-mediated trained immunity also be induced by endogenous atherogenic innate immunological memory in plants3. requires type II interferon (IFN)8, Mtb substances, including oxidized low-density Similarly, the innate immune system of impaired trained immunity via type I IFN9, lipoprotein particles, lipoprotein (a) and invertebrates (for example, mosquitoes, the and the inflammatory cytokines interleukin catecholamines1,14. In animal models of bumble bee Bombus terrestris, snails, and (IL)-1 and GM-CSF were essential for atherosclerosis, a Western-type diet induces so on) has the capacity to generate memory β-glucan-induced trained immunity10. trained immunity, which persists even after a responses to subsequent reinfection with In addition to what has been seen in switch to a healthy chow diet15. Nevertheless, the same or different pathogens1. There is experimental animal models, there is now it is remarkable that trained immunity also compelling evidence in animal models ample evidence that trained immunity is a induced by BCG vaccination has also been that an initial infection or vaccination with component of the human host response to involved in improved induction of immune

2 Nature Immunology | VOL 22 | January 2021 | 2–6 | www.nature.com/natureimmunology comment

a Differentiation b Priming (HSPCs) in the bone marrow (BM) and in circulating immune cells, and peripherally, at the tissue-specific level. It has been demonstrated that BCG, β-glucan, and a Western-type diet reprogram BM-HSPCs toward myelopoiesis and generate trained immunity8,10,15. These studies provide an explanation for why short-lived innate immune cells can acquire memory with

Innate immune response Innate immune response a persistent phenotype in vivo. However, this may also impact the replacement of tissue-resident innate immune cells with Time Time Changed condition 1st insult 2nd insult new and reprogrammed HSPCs. It has been insult shown that, following pulmonary insults, a reduction in yolk-sac-derived alveolar

c Trained immunity d Tolerance macrophages is compensated for through the accumulation of BM-derived macrophages in the lung airways17. For example, infection of mice with gammaherpesviruses provided protection against allergic asthma, as it caused resident alveolar macrophages (AMs) to be replaced with BM-derived AMs18. Thus, a new imprinted AM may have a completely different functional capacity to that of an original fetal-derived AM. It Innate immune response Innate immune response has also been shown that the metabolism of AMs is significantly different from that Time Time of bone-marrow-derived macrophages 1st insult 2nd insult 1st insult 2nd insult (BMDMs), the latter being more glycolytic and bactericidal19. In addition, a murine Fig. 1 | Schematic presentation of the behavior of innate immune responses during the different model of pulmonary adenoviral infection adaptive programs induced in innate immune cells. a, Differentiation. b, Priming. c, Trained immunity induces trained immunity in AMs, which (innate immune memory). d, Tolerance. is dependent on T cells but independent of BMDMs20. Furthermore, the inflammatory site may also alter the functional capacity of the local stromal cells that induce regulation and immunological self-tolerance involved in innate memory responses trained immunity in residential innate in models of autoimmunity such as type depend solely on epigenetic remodeling, immune cells. Strikingly, following skin 1 diabetes and multiple sclerosis16. While and trained immunity appears to be devoid inflammation, epithelial stem cells maintain the precise molecular mechanisms are of specificity. It has been proposed that prolonged chromatin accessibility at key still not fully understood, this evidence immune memory in innate and adaptive inflammatory genes. This feature expedites supports the broad rationale that exposure immunity represents an evolutionary and heightens their response to subsequent to microbes may help to antagonize continuum in which a more robust stressors and potentially influences stem conditions sharing chronic inflammation immune response evolved first, mediated cell cross-talk with trained immune cells21. and tissue damage, such as autoimmunity. by epigenetic mechanisms, while specificity Therefore, central and peripheral factors, The beneficial effect of BCG vaccination, at evolved later in a subgroup of species or a combination of both, can impact least in some autoimmune disorders, is an (vertebrates) through gene recombination2. the duration and maintenance of trained interesting facet of host–microbe interplay While some of the mechanisms for immunity. in the pathophysiology and treatment of epigenetic remodeling and metabolic immune-mediated diseases. reprogramming during trained immunity Diferentiation, priming, tolerance and have been recently reviewed1, the duration training Mechanisms of trained immunity and maintenance of chromatin-driven Adaptations in innate immune Immunological memory in the two arms innate memory responses are still the subject compartments are exceptionally diverse, as of host defense is mediated through of intense investigation. In the context of innate immune cells demonstrate substantial different processes. During the induction of infectious disease, there are three known plasticity and adapt to various insults such adaptive immune memory, two properties factors that can influence the epigenetic as trauma, infections and vaccination, are induced at the same time: (1) the programming of an immune cell: (1) direct and they continue to adapt as they leave specificity of the response, ensured through infection, (2) pathogen-associated molecular the local microenvironment of the bone the rearrangement of immunoglobulin patterns (PAMPs) from microorganisms, marrow and travel to the blood and tissues. family genes and clonal expansion; and and (3) cytokines released during the It is important to note that the magnitude (2) the amplitude and speed of the response, induction of the host response. We envision (low versus high dose) and duration (short mediated by epigenetic reprogramming that these key factors impact the duration versus long) of stimulation induces specific that modulates the kinetics of gene of trained immunity at a central level, in adaptations in innate immune cells that transcription1. By contrast, the mechanisms hematopoietic stem and progenitor cells reflect their requirement to either enhance

Nature Immunology | VOL 22 | January 2021 | 2–6 | www.nature.com/natureimmunology 3 comment

In vitro In vivo that of adaptive immune cells. However, improved nomenclature has been proposed for both in vitro and tissue-resident monocytes and macrophages24. To investigate the central effects of trained immunity via HSPCs in the BM, experiments have been conducted predominantly in vivo or ex vivo8,10. However, many in vitro experiments have been performed to study the peripheral impacts of trained immunity. Thus,

Vaccine (e.g., BCG) experimental standards are necessary for or expanding our knowledge in this exciting Adjuvants (e.g., β-glucan) field of immunology. Defining the adaptive programs of innate immune cells according to their Training period Training period functional state is important because the (5–7 days) (days to months) molecular mechanisms underlying these processes can often overlap. In this respect, High the epigenetic and metabolic rewiring that specifically program cell differentiation, Immune activation trained immunity, immune priming and tolerance can define these processes. Thus, there are unique signatures that define Low different cellular adaptations. For example, long-term changes in DNA methylation and Epigenetic imprinting stable changes in chromatin accessibility can accompany cell differentiation, whereas specific histone marks characterizing Trained immunity ‘latent enhancers,’ such as monomethylated histone H3 K4 (but not solely that), are often 1 Fig. 2 | Experimental models of in vitro and in vivo trained immunity. ‘tagged’ in trained immunity . However, while specific pathways and markers differ between the various adaptive programs in immune responses or prevent immunity and In ‘trained immunity’ (Fig. 1c), in sharp innate immunity, they all use the same basic excessive immunopathology. Several such contrast to priming, while the first mechanisms (epigenetic, transcriptional and adaptive programs have been described, stimulus leads to changes in the functional metabolic), but with different flavors. including cell differentiation, priming, immune status, the immune activation tolerance and trained immunity (Fig. 1). status returns to the basal level following Experimental models of trained Because innate immune cells can undergo removal of the stimulus, while the immunity any of these functional adaptive programs, it epigenetic alterations persist. However, in The models used to study the adaptive is essential to precisely define the similarities response to homologous or heterologous programs in innate immunity, including and differences between these cellular challenges, both gene transcription and trained immunity, should, therefore, reflect adaptations to ensure the field’s focus and cell function are enhanced at much higher the definitions of these processes (Fig. 2). avoid confusion in the literature. levels than those observed during the In vitro and in vivo models have been The main difference between innate primary challenge. The opposite of trained predominantly used to study peripheral immune cells undergoing these different immunity is innate immune ‘tolerance,’ trained immunity. The most common adaptive programs is their functional wherein the cell is unable to activate gene in vitro model system of trained immunity status prior to secondary challenges. Innate transcription and does not perform its is the training of human peripheral immune cell ‘differentiation’ (Fig. 1a) is functions following restimulation (Fig. 1d). monocytes, in which these cells are exposed often the change of an immature cell into its For instance, repeated or persistent to a stimulus (training period) for a short mature counterpart, which is defined by a exposure of macrophages to a high dose period of time (usually 24 h). Subsequently, long-term change in the functional program of lipopolysaccharide epigenetically cells are incubated for 5–7 days in culture of the cell and is often accompanied by enforces tolerance to prevent the medium without any stimulation. During altered morphological characteristics caused expression of inflammatory genes23. this resting period, the functional program by alterations of the tissue environment Therefore, studies aiming to investigate of the cells returns to steady state. If or chronic exposure to stimuli22. During trained immunity need to clearly identify the primary training stimulus results in ‘priming’ (Fig. 1b), the first stimulus changes the activation state during initial stimulation epigenetic encoding of these differentiated the functional state of these cells, and (represented by effector functions such macrophages, they will show a heightened their immune status (as defined by active as cytokine and reactive oxygen species response to homologous or heterologous gene transcription) does not return to production, phagocytosis, killing, and so secondary stimuli. These models are basal levels before the secondary stimulation on) as well as after the removal of the initial fundamentally different from models or infection. Thus, the impact of a second insult. This is challenging when dealing of cell differentiation or priming in which challenge in primed cells is often additive with monocytes and macrophages, as their the stimulus is either maintained for or synergistic with the original stimuli. spectrum of states is far less defined than a long period of time to induce

4 Nature Immunology | VOL 22 | January 2021 | 2–6 | www.nature.com/natureimmunology comment

differentiation/priming or secondary ex vivo–stimulation assay can actually 1Meakins-Christie Laboratories, Department stimulation is performed very quickly be a model for studying innate immune of Medicine, Department of Microbiology after the initial priming. In these cases, in priming. Furthermore, an ex vivo system and Immunology, Department of Pathology, contrast to training, cells are not allowed could also be used to study (1) the bias of McGill University Health Centre and McGill to return to the functional steady state hematopoiesis (for example, myelopoiesis International TB Centre, McGill University, before secondary stimulation. versus lymphopoiesis) in HSPCs using Montreal, QC, Canada. 2Bandim Health Project, Similar considerations are valid for colony-forming unit assays and (2) the Bissau, Guinea-Bissau. 3Department of Molecular models of the induction of trained immunity functional capacity of myeloid immune cells Microbiology, Washington University in St. Louis, in vivo, although different transcriptional (for example, macrophages or neutrophils). St. Louis, MO, USA. 4Department of Medicine, and epigenetic changes can be seen in Genetic Section and, University of Chicago, Chicago, mature innate immune cells in the periphery Perspectives and conclusions IL, USA. 5Department of Internal Medicine and or in bone marrow progenitors of innate Understanding innate immune memory Radboud Institute for Molecular Life Sciences, immune cells. To reprogram BM-HSPCs for is critical for deciphering new approaches Radboud University Medical Center, Nijmegen, central trained immunity, administration to vaccine development. By dissecting the the Netherlands. 6Institute for Clinical Chemistry of training agents (for example, β-glucan) cellular and molecular mechanisms of trained and Laboratory Medicine, University Hospital, or a vaccine (for example, BCG) leads immunity, we hope to develop new vaccine Technische Universität Dresden, Dresden, Germany. initially to proliferation and expansion of strategies with cross-protective efficacy 7Department of Internal Medicine and Radboud HSPCs. Although the kinetics and dynamics against a range of infections. In addition, we Center for Infectious Diseases, Radboud University of this expansion differ between model can envisage more effective vaccines that Medical Center, Nijmegen, the Netherlands. systems, HSPCs phenotypically return to combine the induction of trained immunity 8Department of Paediatrics, Infectious Diseases a basal status before a secondary challenge with adaptive immune memory. While we Research Group, Murdoch Children’s Research experiment (Fig. 2). Murine models of have made enormous progress in our Institute, Te University of Melbourne, Parkville, trained immunity allow us to study the fundamental understanding of trained Victoria, Australia. 9Te Global Tuberculosis direct impact of BM-HSC training on the immunity in health and diseases, accurate Program, Texas Children’s Hospital, Immigrant immune response to a new homologous nomenclature and experimental and Global Health, Department of Pediatrics, or heterologous challenge. One current standardization are important and would Baylor College of Medicine, Houston, TX, USA. possibility to study this in vivo is to generate encourage progress in the field. By doing 10Biomedical Engineering and Imaging Institute and a chimeric or serial engraftment mouse this work, we hope scientists who are new Department of Oncological Sciences, Icahn School model wherein the HSC compartment to trained immunity will establish accurate of Medicine at Mount Sinai, New York, NY, USA. is reconstituted with trained HSCs8–10. experimental models to close the knowledge 11Laboratory of Chemical Biology, Department of As the reconstitution of hematopoiesis gap in this field. Furthermore, as a more Biochemical Engineering, Eindhoven University of in the recipient is mediated by donor precise mechanistic description of trained Technology, Eindhoven, the Netherlands. 12Division hematopoietic progenitors at earlier time immunity is developed, we will need to of Chemical and Systems Biology, Department points and entirely by the donor HSC formulate updated recommendations on of Integrative Biomedical Sciences, Faculty of compartment at later time points (16 weeks) trained immunity. ❐ Health Sciences, Institute of Infectious Disease and post-transplantation, this system is an Molecular Medicine, University of Cape Town, Cape excellent model to study the long-term Maziar Divangahi 1 ✉ , Peter Aaby2, Town, South Africa. 13Te BioMedical Engineering effects of training on HSC-mediated Shabaana Abdul Khader 3, Luis B. Barreiro4, and Imaging Institute, Icahn School of Medicine innate memory responses to subsequent Siroon Bekkering 5, at Mount Sinai, New York, NY, USA. 14Laboratory homologous or heterologous challenges9. Triantafyllos Chavakis 6, of Mammalian Cell Biology and Development, Human models of trained immunity Reinout van Crevel7, Nigel Curtis8, Rockefeller University, New York, NY, USA. have also been established as proof of Andrew R. DiNardo9, 15Chromatine et Infection G5, Institut Pasteur, concept to study innate immune memory. Jorge Dominguez-Andres 7, Paris, France. 16Department of Medicine, Division BCG vaccination in healthy individuals Raphael Duivenvoorden10,11, of Gastroenterology and the Center for the Study of can induce trained immunity in HSCs Stephanie Fanucchi12, Zahi Fayad 13, Infammatory Bowel Disease, Massachusetts General and circulating monocytes, which gain an Elaine Fuchs14, Melanie Hamon 15, Hospital, Harvard Medical School, Boston, MA, enhanced protective capacity against a range Kate L. Jefrey 16, Nargis Khan1, USA. 17Radboud Center for Infectious Diseases (RCI), of infectious agents1. The function of innate Leo A. B. Joosten 7,17, Eva Kaufmann1, Radboud University Medical Center, Nijmegen, immune cells, as well as their epigenetic and Eicke Latz 18, Giuseppe Matarese 19, the Netherlands. 18Institute of Innate Immunity, transcriptional programs, can be studied Jos W. M. van der Meer20, University Hospitals Bonn and German Center for before and after BCG vaccination (or after Musa Mhlanga 12, Neurodegenerative Diseases, Bonn, Germany. 19Treg any other vaccine or experimental infection, Simone J. C. F. M. Moorlag7,18, Cell Lab, Dipartimento di Medicina Molecolare for that matter). However, to ensure that Willem J. M. Mulder 10,11, Shruti Naik21, e Biotecnologie Mediche, Università di Napoli the initial vaccine has been cleared from Boris Novakovic22, Luke O’Neill 23, “Federico II” and Istituto per l’Endocrinologia e the organism, it is important that there is a Jordi Ochando 24, Keiko Ozato25, l’Oncologia Sperimentale, Consiglio Nazionale delle sufficient interval between vaccination and Niels P. Riksen 5, Robert Sauerwein26, Ricerche (IEOS-CNR), Napoli, Italy. 20Radboud the subsequent assessment of the innate Edward R. Sherwood27, Andreas Schlitzer28,29, University Medical Center, Nijmegen, the immune system by ex vivo restimulation Joachim L. Schultze30, Michael H. Sieweke31,32,33, Netherlands. 21Department of Pathology, with a non-specific stimulus. As an Christine Stabell Benn34,35, Department of Medicine and Ronald O. Perelman example, BCG may be present at the site Henk Stunnenberg36, Joseph Sun 37,38, Department of Dermatology, New York University of vaccination for up to one month25, Frank L. van de Veerdonk 39, School of Medicine, New York, NY, USA. 22Murdoch thus trained immunity assessments should Sebastian Weis 40, David L. Williams41, Children’s Research Institute and Department be performed at later time points. During Ramnik Xavier 42,43,44 and of Paediatrics, University of Melbourne, Royal the first month after vaccination, the Mihai G. Netea 7,17,45 ✉ Children’s Hospital, Parkville, Victoria, Australia.

Nature Immunology | VOL 22 | January 2021 | 2–6 | www.nature.com/natureimmunology 5 comment

23School of Biochemistry and Immunology, for Advanced Study, University of Southern Published online: 8 December 2020 Trinity Biomedical Sciences Institute, Trinity Denmark, Odense, Denmark. 36Department of https://doi.org/10.1038/s41590-020-00845-6 College Dublin, Dublin, Ireland. 24Department of Molecular Biology, Faculty of Science, Radboud 37 References Oncological Sciences, Icahn School of Medicine University, Nijmegen, the Netherlands. Immunology 1. Netea, M. G. et al. Nat. Rev. Immunol. 20, 375–388 (2020). 25 at Mount Sinai, New York, NY, USA. Molecular Program, Memorial Sloan Kettering Cancer 2. Netea, M. G., Schlitzer, A., Placek, K., Joosten, L. A. B. & Genetics of Immunity Section, Division of Center, New York, NY, USA. 38Department of Schultze, J. L. Cell Host Microbe 25, 13–26 (2019). 3. Kachroo, A. & Robin, G. P. Curr. Opin. Plant Biol. 16, Developmental Biology, National Institute of Immunology and Microbial Pathogenesis, 527–533 (2013). Child Health and Human Development, Weill Cornell Medical College, New York, NY, 4. Di Luzio, N. R. & Williams, D. L. Infect. Immun. 20, National Institutes of Health, Bethesda, MD, USA. 39Department of Internal Medicine and 804–810 (1978). USA. 26Radboudumc Center for Infectious Radboud Center for Infectious Diseases (RCI), 5. Moorlag, S. J. C. F. M. et al. Cell Rep. 31, 107634 (2020). 6. Krahenbuhl, J. L., Sharma, S. D., Ferraresi, R. W. & Remington, J. S. Diseases, Department of Medical Microbiology, Radboud University Medical Center, Nijmegen, Infect. Immun. 31, 716–722 (1981). Radboud University Medical Center, Nijmegen, the Netherlands. 40Department of Anesthesiology 7. Muñoz, N. et al. Infect. Immun. 78, 4226–4233 (2010). the Netherlands. 27Department of Pathology, and Intensive Care Medicine and Institute for 8. Kaufmann, E. et al. Cell 172, 176–190.e19 (2018). 9. Khan, N. et al. Cell 183, 752–770.e22 (2020). Microbiology and Immunology, Vanderbilt Infectious Disease and Infection Control, 10. Mitroulis, I. et al. Cell 172, 147–161.e12 (2018). University Medical Center, Nashville, TN, USA. Jena University Hospital, Jena, Germany. 11. Arts, R. J. W. et al. Cell Host Microbe 23, 89–100.e5 (2018). 28Quantitative Systems Biology, Life & Medical 41Department of Surgery, Quillen College of 12. Walk, J. et al. Nat. Commun. 10, 874 (2019). 13. Redelman-Sidi, G., Glickman, M. S. & Bochner, B. H. Sciences Institute, University of Bonn, Bonn, Medicine and Center of Excellence in Infammation, Nat. Rev. Urol. 11, 153–162 (2014). Germany. 29Single Cell Genomics and Epigenomics Infectious Disease and Immunity, East Tennessee 14. Bekkering, S. et al. Arterioscler. Tromb. Vasc. Biol. 34, Unit at the German Center for Neurodegenerative State University, Johnson City, TN, USA. 1731–1738 (2014). 42 15. Christ, A. et al. Cell 172, 162–175.e14 (2018). Diseases and the University of Bonn, Bonn, Broad Institute of MIT and Harvard, 16. Ristori, G., Faustman, D., Matarese, G., Romano, S. & Salvetti, M. 30 43 Germany. Genomics & Immunoregulation, Life Cambridge, MA, USA. Center for Microbiome Curr. Opin. Immunol. 55, 89–96 (2018). and Medical Sciences Bonn, University of Bonn, Informatics and Terapeutics, Massachusetts 17. Guilliams, M. et al. J. Exp. Med. 210, 1977–1992 (2013). 31 18. Machiels, B. et al. Nat. Immunol. 18, 1310–1320 (2017). Bonn, Germany. Aix Marseille University, CNRS, Institute of Technology, Cambridge, MA, USA. 19. Russell, D. G., Huang, L. & VanderVen, B. C. Nat. Rev. Immunol. 32 44 INSERM, CIML, Marseille, France. Center Center for Computational and Integrative 19, 291–304 (2019). for Regenerative Terapies Dresden (CRTD), Biology and Department of Molecular Biology, 20. Yao, Y. et al. Cell 175, 1634–1650.e17 (2018). Technische Universität Dresden, Dresden, Massachusetts General Hospital and Harvard 21. Naik, S. et al. Nature 550, 475–480 (2017). 22. Lavin, Y. et al. Cell 159, 1312–1326 (2014). 33 45 Germany. Max-Delbrück-Centrum für Medical School, Boston, MA, USA. Department 23. Foster, S. L., Hargreaves, D. C. & Medzhitov, R. Nature 447, Molekulare Medizin in der Helmholtzgemeinschaf for Genomics & Immunoregulation, Life and 972–978 (2007). (MDC), Berlin, Germany. 34Bandim Health Medical Sciences Institute (LIMES), University 24. Murray, P. J. et al. Immunity 41, 14–20 (2014). 25. Minassian, A. M. et al. J. Infect. Dis. 205, 1035–1042 (2012). Project, OPEN, Institute of Clinical Research, of Bonn, Bonn, Germany. University of Southern Denmark/Odense University ✉e-mail: [email protected]; Competing interests Hospital, Odense, Denmark. 35Danish Institute [email protected] The authors declare no competing interests.

6 Nature Immunology | VOL 22 | January 2021 | 2–6 | www.nature.com/natureimmunology