How does the immune system recognize a wide variety of microorganisms but avoid attacking the body’s own tissues (which are referred to as ‘self ’)? A century of research into this question led to two central principles of immunology. The first principle of immunology is self–non-self discrimination1,2, or the recognition of danger3, altered self and discontinuity4. This is achieved through several mechanisms, such as the expression of innate immune receptors that recognize microorganism-associated molecular patterns (MAMPs) or danger- associated molecular patterns (DAMPs), the elimination of developing T cells that react with self (a process that is known as thymic selection) and the generation of regulatory T (Treg) cells. The second principle is the generation and selection of diverse immune repertoires5–7. To recognize the large diversity of viral, bacterial, fungal and eukaryotic microorganisms, as well as altered self, the process of somatic V(D)J recombination leads to the generation of a large repertoire of receptors on B cells and T cells8,9, which are then selected for by antigens from microorganisms and altered self.

However, these two principles are not sufficient to understand several intriguing aspects of the immune system. For example,

On the basis of immune equilibrium, it can be assumed that the immune system is able to distinguish between ‘good’ and ‘bad’ — for example, between mutualistic and pathogenic microorganisms — and develop either anti-inflammatory or pro- inflammatory immune responses11,12. However, how it would be able to distinguish good from bad is still an open question16. In fact, there are three, possibly four, types of immune response that are mutually inhibitory. These include type 1 responses against intracellular threats (such as viruses, intracellular bacteria and tumours), type 2 responses against large extracellular threats (such as helminths) and type 3 responses against extracellular microorganisms (such as extracellular bacteria and fungi). Activation of one type of response inhibits another type, and the immune equilibrium is maintained by the competing immune responses. This principle forms the basis of the equilibrium model of immunity. In contrast to earlier immunological principles, the equilibrium model of immunity does not predict how an immune response is triggered (as all types of immune response are already active during the steady state) but instead predicts how this active immune system behaves when facing new threats, which are almost always present in combination. I believe that the equilibrium model of immunity provides a broad but simple and testable framework to explain complex immune phenomena such as tolerance, autoimmunity, allergy and resistance or susceptibility to secondary infections. It might also open up new approaches for immunotherapy.

The equilibrium model of immunityThe foundations of immunology are rooted in the work of Metchnikoff (1845–1916) and Paul Ehrlich (1854–1915), who drew on the discovery by Louis Pasteur (1822–1895) and Robert Koch (1843–1910) of the pathogenic potential of microorganisms (BOX 1). Therefore, the immune system is primarily viewed as a system that opposes pathogens through elaborate mechanisms of recognition and destruction. However, the realization that the immune system is in a constant state of activation, even in the absence of pathogens, indicates that the immune system does not react only to

the very existence of autoimmunity and Treg cells suggests a failure to avoid the recognition of self, and thus a need for tolerance mechanisms. Furthermore, these principles cannot explain why the immune system develops self-threatening allergic reactions, and why allergies and autoimmune pathologies are increasing in industrialized nations in association with greater hygiene and less exposure to infectious diseases10.

On the basis of the principle of self–non-self discrimination, it was assumed that the immune system is at rest when not exposed to pathogens, danger signals or altered self. However, it is now clear that the immune system is never at rest, in particular at mucosal sites11,12 but also systemically13 and in germ-free conditions14. Indeed, self is constantly altered and danger is continually present as cells die, tissues are injured, and microorganisms proliferate within and outside the body. More than a century ago, Élie Metchnikoff stated that “physiological inflammation” is required to maintain harmony in animals15. These observations prompted a third principle of immunology — immune equilibrium — in which anti- inflammatory forces of the immune system regulate pro-inflammatory forces to maintain homeostasis. When the anti-inflammatory forces fail, severe pathology occurs.

E S S AY

Immunity by equilibriumGérard Eberl

Abstract | The classical model of immunity posits that the immune system reacts to pathogens and injury and restores homeostasis. Indeed, a century of research has uncovered the means and mechanisms by which the immune system recognizes danger and regulates its own activity. However, this classical model does not fully explain complex phenomena, such as tolerance, allergy, the increased prevalence of inflammatory pathologies in industrialized nations and immunity to multiple infections. In this Essay, I propose a model of immunity that is based on equilibrium, in which the healthy immune system is always active and in a state of dynamic equilibrium between antagonistic types of response. This equilibrium is regulated both by the internal milieu and by the microbial environment. As a result, alteration of the internal milieu or microbial environment leads to immune disequilibrium, which determines tolerance, protective immunity and inflammatory pathology.

524 | AUGUST 2016 | VOLUME 16 www.nature.com/nri

PERSPECTIVES

© 2016

Macmillan

Publishers

Limited.

All

rights

reserved. ©

2016

Macmillan

Publishers

Limited.

All

rights

reserved.

pathogens. At mucosal sites in particular, symbiotic microorganisms induce diverse immune effector cells that establish a vital equilibrium between the host and its microbiota11,12. This is also the situation systemically, where symbiotic viruses chronically activate the immune system13.

Negative regulation has a crucial role in the immune system, a concept that was first proposed in the 1970s by Niels Jerne, Heinz Kohler and Geoffrey Hoffmann in the context of the immune network theory17–19, and by Richard Gershon in the context of the suppressor T cell theory20,21. Negative regulation has now become an immunological paradigm through the characterization of Treg cells, the absence of which leads to dramatic inflammatory pathology22–25. It is generally agreed that a healthy immune system, as well as a healthy immune response, is based on an equilibrium between effector T cells and Treg cells, and that it involves regulation through a variety of molecules, such as interleukin-10 (IL-10), transforming growth factor-β (TGFβ), programmed cell death 1 (PD1; also known as PDCD1), cytotoxic T lymphocyte antigen 4 (CTLA4) and CD25 (also known as high-affinity IL-2 receptor (IL-2RA))26,27.

However, Treg cells are not the only cells to regulate immune responses. Other types of immune cells have regulatory functions, such as regulatory B cells28 and myeloid-derived suppressor cells29. Furthermore, it is well known that T helper 1 (TH1) cells, TH2 cells30–32 and TH17 cells negatively regulate each other in in vitro differentiation assays, and that one type of T helper cell characteristically dominates a particular immune response in vivo33. More generally, immune responses can be characterized as antagonistic type 1, type 2 or type 3 responses (with type 3 responses defined as those inducing TH17 cells) (BOX 2), which involve an array of lymphoid, myeloid and stromal cells that are adapted to the type of microorganism, pathogen or injury that is affecting the organism.

The equilibrium model of immunity proposes that the immune system relies on an equilibrium between these different types of immune response: an equilibrium that defines homeostasis. As a consequence, a microorganism or injury that triggers one type of response induces the suppression of the other types of response. Conversely, an absence of stimulation of one type of response leads to the exacerbation of the other types of response, with potentially pathological consequences (FIG. 1).

of strong inflammation and extensive tissue injury, during which both intracellular and extracellular threats may be recognized, lymphoid cells produce both IFNγ and IL-17 and therefore have a mixed type 1 and type 3 phenotype38,39.

Type 2 responses that develop against large organisms, such as helminths, seem to have a different nature and purpose. The destruction of these large organisms by immune cells is difficult, and so the immune system constructs a barrier using tissue- repair mechanisms to provide protection40. For example, at mucosal surfaces, type 2 responses induce mucus production and collagen deposition. These responses are shaped by the production of IL-25, IL-33 and thymic stromal lymphopoietin (TSLP) by non-haematopoietic cells41, leading to the activation of ILC2s35 and eosinophils, resulting in the development of TH2 cells and IgG1+ or IgE+ B cells and the production of IL-4, IL-5 and IL-13. Alternatively-activated macrophages (AAMs)42 have an important

Type 1 and type 3 responses are induced by microorganisms, and they fit the original defensive role that was assigned to the immune system by Metchnikoff and Ehrlich. Type 1 responses are shaped by the production of IL-12 by dendritic cells (DCs) and macrophages34 and lead to the activation of natural killer (NK) cells and group 1 innate lymphoid cells (ILC1s)35, followed by the activation of TH1 cells, cytotoxic CD8+ T cells and, in mice, IgG2+ B cells. The main molecular effectors of type 1 responses are interferon-γ (IFNγ) that is produced by lymphoid cells and cytotoxic molecules, such as perforin and oxygen radicals. Type 3 responses are shaped by the production of IL-1β and IL-23 by DCs and macrophages36,37, which leads to the activation of ILC3s35 and TH17 cells. The type 3 effector phase is characterized by the production of IL-17 and IL-22 by lymphoid cells, antimicrobial peptides (AMPs) by epithelial cells and the recruitment of neutrophils. In the context

Box 1 | A brief history of pathology and immunology

The Greek physician Hippocrates (460–370 BCE) is credited as being among the first to propose that diseases are not caused by the Gods but by environmental factors, diet and living habits. His medical system was based on humourism, which states that health is determined by the equilibrium, within an individual, of four fundamental body fluids, the humours. This notion of internal equilibrium is central to the more recent concept of ‘milieu intérieur’, defined by Claude Bernard (1813–1878). In his view, an organism has to ensure internal stability, or homeostasis, as defined later by Walter Cannon (1871–1945)63, to maintain health in the face of external variations.

Directly contradicting humourism, Louis Pasteur (1822–1895) and Robert Koch (1843–1910) demonstrated the germ theory of disease, in which the environment is a source of diseases through infectious microorganisms. The proponents of this theory had to postulate the existence of an immune system that protects the body from pathogenic microorganisms. Élie Metchnikoff (1845–1916), who was recruited by Pasteur in 1888, demonstrated the existence of phagocytes that destroy microorganisms by ingestion. Paul Ehrlich (1854–1915), a close friend of Koch, showed the role of humoral immunity, conveyed by antibodies, in the defence against microorganisms and bacterial toxins.

A century of immunology research followed to try to understand how the immune system recognizes the enormous diversity of microorganisms, and how it destroys pathogens without destroying the organism that it means to protect. Paradigms emerged, such as the self–non-self discrimination principle proposed by Frank Macfarlane Burnet (1899–1985)1 and Niels Jerne (1911–1994)2. More recently, Polly Matzinger formulated the danger model3, which proposes that the immune system is triggered not only by the recognition of invading microorganisms (non-self), as discussed by Charles Janeway (1943–2003)131, but also by the danger associated with these microorganisms and, more generally, by injured tissue. Finally, Thomas Pradeu, Sébastien Jaeger and Eric Vivier proposed that the immune system is fundamentally activated by a change in normality or discontinuity4.

An alternative view of immune reactivity, known as the immune network theory, was proposed in the 1970s by Jerne, Heinz Kohler and Geoffrey Hoffmann17–19. According to this theory, the immune system is composed of a network of cells and antibodies, recognizing both non-self antigens and new self-antigens (such as hyper-variable regions of antibodies). This network of recognition includes both activators and suppressors that maintain the system in a state of equilibrium during homeostasis. Disturbance of that equilibrium by a new antigen induces a response, which is followed by a gradual return to homeostasis. Although this theory has been generally abandoned by the immunological community, the idea of an equilibrium within the immune system that is maintained by a network of activators and suppressors is a concept that is supported by more recent studies, and encapsulated by the view that an equilibrium between pro-inflammatory and anti-inflammatory immune responses is required to maintain health.

P E R S P E C T I V E S

NATURE REVIEWS | IMMUNOLOGY VOLUME 16 | AUGUST 2016 | 525

© 2016

Macmillan

Publishers

Limited.

All

rights

reserved. ©

2016

Macmillan

Publishers

Limited.

All

rights

reserved.

role in type 2 responses during tissue repair43,44 and were originally described by Metchnikoff as the phagocytes that eliminate abnormal cells during fetal development45. Mixed type 2 and type 3 responses may occur during allergic responses, possibly induced by concomitant tissue damage and the presence of extracellular microscopic particles46.

A type 4 immune response has also been proposed33,47. This response does not develop against microorganisms or parasites that infect or injure tissues, but rather aims to block microorganisms and parasites before they reach sensitive tissues. For example, in the eye, inflammation is poorly tolerated and causes irreversible damage48, and in the intestine, the extensive microbiota must be kept away from the epithelium to avoid the induction of tissue-damaging chronic inflammation49,50. Type 4 immunity requires secretory IgA, which is released in large quantities into the intestinal lumen, and is also released in tears, saliva, sweat and secretions from the genitourinary tract and the respiratory epithelium. IgA has a broad ‘anti-inflammatory’ effect51, partly because it neutralizes microorganisms before they reach the host tissues where they would induce type 1 or type 3 immune responses52. Other elements of the proposed type 4 responses include the production of mucus49 and AMPs, which are also secreted in substantial amounts by epithelial cells50.

So, do Treg cells constitute a separate type of response that is dedicated to suppressing all other responses? Treg cells have been described that express the signature transcription factors of TH1 cells (T-bet; also known as TBX21)53, TH2 cells (GATA binding protein 3

that represses the others until the infection is cleared. By contrast, following a decrease in one type of trigger, for example, during antibiotic therapy to eliminate extracellular bacteria that induce type 3 responses, the other types of response are dysregulated and exacerbated to levels that may lead to pathological inflammation. I propose that the original and intuitive concept of equilibrium — the basis of Claude Bernard’s (1813–1878) principle of ‘milieu intérieur’, and of Walter Cannon’s (1871–1945) concept of homeostasis63 — is a core principle of immunity (BOX 1).

Evidence for the equilibrium modelInflammatory pathologies. Type 2 responses drive allergic and pro-fibrotic pathologies through the production of IL-4, IL-5 and IL-13 (REF. 64). By contrast, type 3 responses are involved in autoimmune inflammatory diseases, such as inflammatory bowel disease, rheumatoid arthritis and multiple sclerosis39,65,66, through the production of IL-17, granulocyte–monocyte colony- stimulating factor36 and lymphotoxin67, whereas type 1 responses are involved in systemic lupus erythematosus (SLE) and type 1 diabetes (T1D) through the production of IFNs68. In industrialized nations, the decreased incidence of infectious diseases thanks to better hygiene, vaccines and antibiotics, is associated with an increase in the incidence of allergic and autoimmune inflammatory diseases — an association that is described by the hygiene hypothesis10. I propose that inflammatory pathologies may not only be a consequence of exacerbated type 1, type 2 or type 3 responses, but may also be a consequence of diminished type 1, type 2 or type 3 responses.

Epidemiological and experimental data show that a loss of exposure to microorganisms (not necessarily pathogens) leads to an increased susceptibility to allergic responses69. For example, children raised on farms are less susceptible to developing allergies than those not raised on farms70. Conversely, children who are treated at an early age with multiple doses of antibiotics are more susceptible to allergies than those who are not exposed to repeated antibiotic therapy71,72. The same effect is found in mice treated early in life with antibiotics73 or maintained in germ-free conditions until weaning74, and germ-free mice also develop high levels of serum IgE14. Several mechanisms have recently been reported to explain how the microbiota inducing type 3 responses represses pro-allergic type 2

(GATA3))54,55 and TH17 cells (retinoic acid receptor-related orphan receptor-γt (RORγt))27,56, as well as chemokine receptors that are associated with these T helper cell subsets. It has been suggested that the cytokine environments that induce type 1, type 2 or type 3 responses also promote a partial differentiation of Treg cells into these types, allowing them to migrate to effector sites and to efficiently regulate the associated T helper cell functions57. In addition, the generation of Treg cells, in particular, type 3 Treg cells, is induced at the expense of the associated effector cells56,58. Nevertheless, this does not occur during inflammation27, and type 3 Treg cells do not regulate TH17 cells, but instead regulate TH2 cells, in accordance with the equilibrium model of immunity56. The inverse situation is also true: type 2 Treg cells regulate TH17 cells54.

Finally, and importantly, although one type of immune response is dominant at one site at a particular moment, different responses can simultaneously develop at different sites or sequentially develop at the same site. For example, a type 3 response that is induced by symbiotic microbiota in the intestine does not preclude the presence of a type 2 response in adipose tissues59,60, at least in the steady state61. Furthermore, destructive type 1 and type 3 responses must be followed by repair-associated type 2 responses to restore homeostasis in the affected tissue40,62.

In summary, the four arms of the immune system are balanced in the healthy state. The normal microbiota induces all four types of response and thereby establishes a healthy balance. During infection or injury, one arm of the immune system is stimulated

Box 2 | Molecular regulation of type 1, type 2 and type 3 immunity

The molecular regulation of type 1, type 2 and type 3 immunity has been extensively studied in T helper (TH) cells and their innate counterparts, the innate lymphoid cells (ILCs). Each type of TH cell and ILC is induced by one of the master regulator transcription factors: T‑bet induces TH1 cells and ILC1s, GATA binding protein 3 (GATA3) induces TH2 cells and ILC2s, and retinoic acid receptor-related orphan receptor-γt (RORγt) induces TH17 cells and ILC3s132,133. During inflammation, the cytokine environment activates specific signal transducers and activators of transcription (STATs) that favour the expression of a particular master regulator in naive or precursor cells. For example, interleukin-23 (IL-23) induces the phosphorylation and activation of STAT3 that drives the expression of RORγt, which in turn induces the expression of the effector cytokines IL-17 and IL-22 (REF. 134). These master regulators function within a network of transcription factors that support their lineage-determination function135 and negatively regulate the expression and function of the ‘competing’ master regulators. T-bet inhibits GATA3 (REF. 136) and the differentiation of TH17 cells137, whereas GATA3 inhibits both type 1 and type 3 responses132. In type 3 cells, STAT3 suppresses type 2 responses56,138,139. The master regulator of regulatory T (Treg) cells is forkhead box P3 (FOXP3), which binds GATA3 (REF. 140), RORγt27,141 and STAT3 (REF. 140) and which dominates to induce its regulatory programme. It should be noted, however, that these master regulators can have additional roles during the development of lymphoid cells and in mature effector cells142. For example, GATA3 expression is also required for the development of the common ILC precursor143 and for the further differentiation of ILC3s into a type 1‑like NKp46+ subset that co‑expresses T‑bet, possibly through the regulation of RORγt144.

P E R S P E C T I V E S

526 | AUGUST 2016 | VOLUME 16 www.nature.com/nri

© 2016

Macmillan

Publishers

Limited.

All

rights

reserved. ©

2016

Macmillan

Publishers

Limited.

All

rights

reserved.

responses. First, exposure of pre-weaned mice to the microbiota leads to long-term epigenetic modification of the gene encoding CXC-chemokine ligand 16 (CXCL16), which recruits invariant NKT cells that produce IL-13 and increase susceptibility to type 2 inflammation74. Second, we showed that the microbiota induces TH17 cells and RORγt+ Treg cells that inhibit the generation of type 2 T cell responses56. Third, it has been shown that B cells respond directly to bacterial signals to limit serum IgE levels and basophil numbers75. In addition, type I

responses are enhanced if the adipose tissue is inflamed — for example, after a diet- and inflammation-induced increase in intestinal permeability that causes microorganisms and microbial products to access the circulation and to reach fat tissues61,83,84. As a consequence, type 2 responses are inhibited and blood glucose levels rise, increasing the risk of type 2 diabetes.

Type 3 immunity is regulated by viral and bacterial infections that induce type 1 responses85. In particular, IL-17 production that is required against bacterial

IFNs that are induced by murine norovirus (MNV) block the elevated type 2 response that is found in germ-free mice76, and IFNs directly block ILC2s77,78.

Another case of pathological interference between type 2 and type 3 responses involves fat metabolism, and the regulation by AAMs of blood glucose consumption to generate heat79. Monocytes that are recruited to fat are converted to AAMs by IL-4 that is produced by eosinophils80, which are themselves recruited by IL-5 that is produced by ILC2s81,82. However, type 3

Helminths

Nature Reviews | Immunology

IL-33

IL-12

IFNγ

IL-23

Type 1 response

IL-4 IL-17

TGFβ

sIgA

Health

a

b Autoimmunity Allergy

Type of response1 2 3 4

Type of response1 2 3 4

Type of response1 2 3 4

Type 2 response

Type 4 response

Type 3 response

Extracellularmicroorganisms

Intracellular viruses and bacteria, and tumours

Exclusion of microorganisms

Figure 1 | The equilibrium model of immunity. a | The equilibrium model of immunity is based on the idea that the immune system is never at rest but instead relies on a dynamic equilibrium between four types of competing and mutually inhibitory immune responses. Type 1 responses are directed against intracellular threats, such as viruses, some bacteria and tumours, whereas type 3 responses are directed against extracellular micro organisms, such as most bacteria and fungi. Type 2 responses are directed against large parasites, such as helminths, which cannot be destroyed by type 3 responses but that can be kept at bay by the construction of barriers through the production of mucus and collagen. Finally, type 4 responses develop to exclude

microorganisms and to protect sensitive tissues from potentially destructive inflammation, through the secretion of, for example, large amounts of IgA and antimicrobial peptides into the gut lumen or eye secretions. b | Health is determined by an equilibrium between these four types of response. Following an infection that induces one type of response, the other types of response are inhibited. Conversely, the absence of one type of response leads to the exacerbation (red) of the other types, potentially leading to inflammatory pathologies, such as autoimmunity and allergy. IFNγ, interferon-γ; IL, interleukin; sIgA, secretory IgA; TGFβ, transforming growth factor-β.

P E R S P E C T I V E S

NATURE REVIEWS | IMMUNOLOGY VOLUME 16 | AUGUST 2016 | 527

© 2016

Macmillan

Publishers

Limited.

All

rights

reserved. ©

2016

Macmillan

Publishers

Limited.

All

rights

reserved.

infections is inhibited by type 1 and type 2 IFNs86–88. Furthermore, helminth-induced type 2 responses limit autoimmune inflammation, at least in rodent models89,90. These mechanisms might explain the increased incidence of arthritis, multiple sclerosis and inflammatory bowel disease in industrialized countries, where there is a decreased incidence of infections by helminths, viruses and intracellular bacteria such as Mycobacterium tuberculosis10. Conversely, a loss in the diversity of bacterial symbionts, as a consequence of increased antibiotic use, may lead to a loss of type 3 responses and, therefore, to an increase in autoimmunity that is driven by type 1 responses, such as SLE and T1D68. In support of this mechanism, segmented filamentous bacteria, intestinal symbionts and potent inducers of type 3 responses91,92, are associated with protection from T1D in mice93,94.

Infectious pathologies. Recent data have provided evidence to explain how a pre-existing infection can influence the outcome of a second infection (or superinfection). For example, mice that are latently infected with a γ-herpesvirus develop increased resistance to superinfection by the bacteria Listeria monocytogenes95. Both microorganisms elicit a type 1 response by the host, and the increased level of IFNγ that is induced by the latent (symbiotic) virus confers protection against L. monocytogenes. However, the presence of a virus can also increase susceptibility to superinfection by a microorganism that is controlled by a different type of response. For example, a previous infection with lymphocytic choriomeningitis virus (LCMV) increases the susceptibility of mice to a subsequent infection with L. monocytogenes or Staphylococcus aureus through a mechanism that involves the type 1 IFN-mediated apoptosis of granulocytes96. In this case, the type 1 response that is induced by the virus blocks the type 3 response that is mediated by the granulocytes to limit the early spread of bacteria. Similarly, a previous infection with influenza A virus or respiratory syncytial virus renders mice highly susceptible to superinfection with pneumonia-inducing bacteria, such as Neisseria meningitides97, Streptococcus pneumoniae, Haemophilus influenzae, Streptococcus pyogenes and S. aureus98. These viruses may also induce type 2 repair responses during later stages of infection that further interfere with the antibacterial type 3 response.

Predictions from the equilibrium modelThymus-derived Treg cells and repair responses. A key principle of adaptive immunity is the clonal selection of immature T cells in the thymus107–109, during which T cells that recognize self-antigens are eliminated (through negative selection). An apparent violation of this rule is the generation in the thymus of Treg cells that react against self 110,111. It is generally thought that thymus-derived Treg (tTreg) cells are generated to suppress responses by effector T cells that weakly react to self-antigen; these cells react with self-antigen too weakly to be eliminated during negative selection, but do react strongly enough to have the potential to become autoreactive disease-causing cells during inflammation112. So, tTreg cells are selected in the thymus owing to their intermediate reactivity to self.

The equilibrium model of immunity suggests an alternative explanation for the generation of tTreg cells, as well as a new classification for thymus-derived and peripherally derived Treg (pTreg) cells (FIG. 2). We have recently reported that, in the intestine, microbiota-induced pTreg cells express RORγt, which is the marker for type 3 lymphoid cells, and initially differentiate along a pathway that is common to TH17 cells and then through a distinct pathway in the presence of retinoic acid56. By contrast, Treg cells that express GATA3, the marker for type 2 lymphoid cells, can develop in the absence of the microbiota56 and constitute another major population of Treg cells in the intestine54,55 and adipose tissue113,114. On the basis of these observations, we proposed that Treg cells differentiate according to the associated effector T cells into type 1, type 2 or type 3 Treg cells, and thus determine the level of the local immune response56. This suggests that recognition of self by developing tTreg cells, which occurs in the absence of microbiota-derived antigens, should lead to the generation of type 2 Treg cells, which may contribute to tissue repair when activated in the context of sterile tissue injury. During the late stages of an infection that require tissue repair, type 2 Treg cells may also contribute to the inhibition of type 1 and type 3 responses54,115. This idea could be tested by assessing the expression of molecules that are associated with type 2 immunity and tissue repair by tTreg cells.

Neonatal and oral tolerance. Immune tolerance is considered to be a general mechanism of immune control that is required to avoid inflammatory pathology.

Conversely, helminths that induce type 2 responses impair antiviral type 1 responses. Mice that are infected with MNV develop virus-specific CD8+ T cell and TH1 cell responses. However, pre-infection with the helminth Trichinella spiralis blocks the generation of virus-specific T cells and leads to increased viral loads99. Microbiota-induced type 3 responses also block antiviral type 1 responses. Mice treated with antibiotics resist MNV infection through the production of IFNλ by epithelial cells and the activation of the transcription factors signal transducer and activator of transcription 1 (STAT1)  and interferon- regulatory factor 3 (IRF3), which are mediators of type 1 immunity. However, in untreated mice, the microbiota induces type 3 immunity and blocks this type 1 response100. Finally, infectious microorganisms or helminths can manipulate the immune system for their own benefit. For example, the fungus Aspergillus fumigatus induces a potent pro-allergic type 2 response in the lungs101, but neutrophils and IL-17-mediated type 3 responses are required for its efficient elimination102.

Tumours. Type 1 cytotoxic responses, which involve CD8+ T cells and NK cells, are most effective for fighting tumours. However, tumours induce several responses that inhibit antitumour type 1 responses. Tumour-induced type 3 responses favour tumour growth through the activation of the anti-apoptotic transcription factor STAT3. In a mouse model of spontaneous breast cancer, tumour-infiltrating macrophages produce IL-1β, which induces the production of IL-17 by γδ T cells and the recruitment of neutrophils103. The neutrophils inhibit the activation of tumour-specific CD8+ T cells and thereby facilitate tumour metastasis.

Tumours also induce TGFβ-driven ‘anti-inflammatory’ responses, which inhibit tumour-specific CD8+ T cells by promoting the generation of Treg cells (which type of Treg cell remains to be determined)104. TGFβ also promotes the generation of IgA+ B cells, which are the proposed main effectors of type 4 responses105. Recently, IgA+ B cell responses were found in the tumour microenvironment of mice treated with the chemotherapeutic drug oxaliplatin. The generation of such B cells was dependent on TGFβ receptor signalling and led to the expression of IL-10 and PD1 ligand 1 (also known as CD274), which inhibited the cytotoxic T cell-mediated response against the tumour106.

P E R S P E C T I V E S

528 | AUGUST 2016 | VOLUME 16 www.nature.com/nri

© 2016

Macmillan

Publishers

Limited.

All

rights

reserved. ©

2016

Macmillan

Publishers

Limited.

All

rights

reserved.

However, as in the case of Treg cells, tolerance may instead reflect an inhibition of one type of immunity by another type33 (FIG. 2).

For example, neonatal tolerance occurs on the exposure of neonates to antigens116. It is commonly accepted that the immature status of the neonatal adaptive immune system leads to the elimination rather than to the activation of T cells specific for such antigens. This view is an extension of the principle, formulated by Frank Macfarlane Burnet (1899–1985), that embryonic antigens are defined as self and should therefore induce tolerance1. However, as discussed above for the generation of tTreg cells, neonatal antigens, as well as embryonic antigens, may induce type 2 cells (TH2 cells, type 2 Treg cells, ILC2s and AAMs) and thereby induce the repair responses that are required to control developing tissues. Neonatal type 2 responses are predicted to inhibit the type 1 and type 3 responses induced by most experimental challenges33. It may also be predicted that an absence of type 2 responses during the neonatal period

Resistance to infection. An important implication of the equilibrium model of immunity is that resistance to infection is partly determined by the state of the immune system before infection (FIG. 3). For example, a primary infection confers resistance or susceptibility to a subsequent infection, depending on the types of response that are engaged by the two infectious agents. This idea may be generalized to the modulation of the immune system by the symbiotic microbiota, which includes type 1-inducing viruses, type 2-inducing helminths and allergens, and type 3-inducing bacteria and fungi. Therefore, analysis of the immune state of an organism before infection, in terms of the balance between the different types of immune response, may allow researchers to predict the outcome of an infection or the efficiency of immunotherapy and vaccination.

Preventive and therapeutic avenuesOn the basis of the equilibrium model of immunity, novel types of preventive and therapeutic strategies may be developed.

increases susceptibility to pathological tissue damage and life-threatening organ dysfunction.

Oral tolerance is a mechanism by which antigen delivered through the gastrointestinal tract suppresses effector responses against that antigen throughout the organism117. However, although tolerance suggests an absence of response, intestinal responses to orally-delivered antigens are not eliminated but instead comprise type 3 Treg cells56 and the production of IgA47,118. Consistent with the equilibrium model of immunity, the induced type 3 response inhibits type 1 and type 2 responses against the orally-delivered antigen. Accordingly, RORγt-deficient mice, which lack type 3 immunity, develop a pathological form of type 1 immunity against intestinal antigens119. Similarly, the type 4 response comprising IgA inhibits the other types of response. Oral tolerance can be breached by the administration of antigen together with mucosal adjuvants, such as cholera toxin, which shifts the type 4 IgA response to a local type 3 response120 or to a systemic type 2 response121.

Nature Reviews | Immunology

T-bet

RORγ t

pTreg

cell

pTreg

cell

B cellsIgA

Intracellularmicroorganismsand tumours

Extracellularmicroorganisms

Exclusion of microorganisms

Thymic and neonatal tolerance

Virus- and tumour-induced tolerance

Bacteria- and fungi-induced tolerance

Oraltolerance

Tissuerepair

Type 2 response Type 3 response

Type 1 response

GATA3

tTreg

cell

Type 4 response

Figure 2 | Tolerance in the equilibrium model of immunity. The four types of immune response are mutually inhibitory. Therefore, tolerance, as measured by the elimination of one type of response, may in many cases reflect inhibition by another type of response rather than a total absence of immune responses. For example, measures of the type 1 response against a virus (interferon-γ (IFNγ) levels and antigen-specific cytotoxic T lymphocyte responses) can be significantly decreased by the presence of symbiotic

bacteria that induce type 3 responses and helminths that induce type 2 responses. Regulatory T (Treg) cells are an important component of mutual regulation, but they do not lack effector functions: thymus-derived Treg (tTreg) cells react to self-antigens, not only to inhibit type 1 and type 3 responses, but also to promote tissue repair115, which is a trait of type 2 responses. GATA3, GATA binding protein 3; pTreg, peripherally derived Treg; sIgA, secretory IgA.

P E R S P E C T I V E S

NATURE REVIEWS | IMMUNOLOGY VOLUME 16 | AUGUST 2016 | 529

© 2016

Macmillan

Publishers

Limited.

All

rights

reserved. ©

2016

Macmillan

Publishers

Limited.

All

rights

reserved.

For example, to protect against viruses or tumours type 1 responses need to be induced. This can be achieved either by directly enhancing type 1 responses or by blocking type 2 and type 3 responses. William Coley famously used S. pyogenes to treat patients with cancer in the 1890s122, a strategy that was later shown to involve either tumour necrosis factor (TNF) or IL-12 (REF. 123). In addition, one of the most effective treatments against non-invasive bladder cancer is intravesical delivery of Mycobacterium bovis bacillus Calmette-Guérin (BCG), which is a strong inducer of type 1 responses124. To increase the safety and feasibility of this approach, MAMPs could be used or synthesized to promote type 1 responses. Antibiotic treatment has been shown to prevent persistent infection by MNV, as the type 3

using non-pathogenic bacteria, fungi or viruses, or MAMPs derived from these microorganisms127, that inhibit pro-allergic type 2 responses. Furthermore, the induction of type 1 or type 2 responses using viruses, helminths, allergens or related MAMPs, could be used to inhibit autoimmune type 3 inflammation, an approach that has been investigated using helminths128,129 and helminth-derived proteins89,130.

Approaches that are based on the positive manipulation of the immune equilibrium have an important benefit compared with the use of drugs that target microorganisms or inflammation. These approaches re-equilibrate the immune system and strengthen the equilibrium through the presence of ‘regulatory’ microorganisms, rather than weaken it through drugs or losses to the host microbiota.

Gérard Eberl is at the Institut Pasteur, Microenvironment and Immunity Unit, 75724 Paris, France, and the Institut National de la Santé et de la

Recherche Médicale (INSERM) U1224, 75724 Paris, France.

Correspondence to G.E. gerard.eberl@pasteur.fr

doi:10.1038/nri.2016.75 Published online 11 Jul 2016

1. Burnet, F. M. & Fenner, F. The Production of Antibodies 2nd edn (Macmillan and Co., 1949).

2. Jerne, N. K. The somatic generation of immune recognition. Eur. J. Immunol. 1, 1–9 (1971).

3. Matzinger, P. Tolerance, danger, and the extended family. Annu. Rev. Immunol. 12, 991–1045 (1994).

4. Pradeu, T., Jaeger, S. & Vivier, E. The speed of change: towards a discontinuity theory of immunity? Nat. Rev. Immunol. 13, 764–769 (2013).

5. Jerne, N. K. The natural-selection theory of antibody formation. Proc. Natl Acad. Sci. USA 41, 849–857 (1955).

6. Ehrlich, P. in Nobel Lecture, December 11, 1908 (Elsevier Publishing Company, 1967).

7. Burnet, F. M. The Clonal Selection Theory of Acquired Immunity (Vanderbilt Univ. Press, 1959).

8. Weigert, M. G., Cesari, I. M., Yonkovich, S. J. & Cohn, M. Variability in the lambda light chain sequences of mouse antibody. Nature 228, 1045–1047 (1970).

9. Hozumi, N. & Tonegawa, S. Evidence for somatic rearrangement of immunoglobulin genes coding for variable and constant regions. Proc. Natl Acad. Sci. USA 73, 3628–3632 (1976).

10. Bach, J. F. The effect of infections on susceptibility to autoimmune and allergic diseases. N. Engl. J. Med. 347, 911–920 (2002).

11. Sansonetti, P. J. War and peace at mucosal surfaces. Nat. Rev. Immunol. 4, 953–964 (2004).

12. Sansonetti, P. J. & Di Santo, J. P. Debugging how bacteria manipulate the immune response. Immunity 26, 149–161 (2007).

13. Virgin, H. W., Wherry, E. J. & Ahmed, R. Redefining chronic viral infection. Cell 138, 30–50 (2009).

14. Cahenzli, J., Koller, Y., Wyss, M., Geuking, M. B. & McCoy, K. D. Intestinal microbial diversity during early-life colonization shapes long-term IgE levels. Cell Host Microbe 14, 559–570 (2013).

15. Metchnikoff, E. Lectures on the Comparative Pathology of Inflammation Delivered at Pasteur Institute in 1891 (Dover Press, 1989).

16. Eberl, G. A new vision of immunity: homeostasis of the superorganism. Mucosal Immunol. 3, 450–460 (2010).

response that is induced by the bacterial microbiota (which is inhibited by the antibiotics) inhibits the production of antiviral IFNλ by epithelial cells100. Type 2 or type 3 responses can be further targeted using neutralizing antibodies against key cytokines, such IL-33 or IL-23, or using antagonists against key transcription factors such as RORγt125,126. Similar strategies may be developed to enhance antihelminth type 2 responses or antibacterial and antifungal type 3 responses.

Conversely, the immune equilibrium could be manipulated to dampen allergic and autoimmune inflammation. Current strategies for treating allergy and inflammation rely on targeting effectors, such as histamine and TNF, or the use of broad anti-inflammatory drugs. Instead, it might be possible to induce type 1 or type 3 responses

Nature Reviews | Immunology

• Symbionts (viruses, bacteria and helminths)

• Homeostatic tissue repair• Allergens and oral antigens

• Increased resistance to viruses, intracellular bacteria and tumours

• Decreased susceptibility to allergy

• Decreased resistance to helminths, extracellular bacteria and fungi

• Decreased tissue repair

• Increased resistance to helminths

• Increased tissue repair• Decreased resistance to

viruses, tumours, bacteria and fungi

• Increased susceptibility to allergy

• Increased resistance to extracellular bacteria and fungi

• Decreased susceptibility to allergy

• Decreased resistance to viruses, tumours and helminths

• Decreased tissue repair

1 2 3 4

1 2 3 4

1 2 3 4

Type of response

Type of responseType of response

Type of response

1 2 3 4

Figure 3 | Microorganisms in the equilibrium model of immunity. The induction of one type of response by microorganisms or helminths inhibits the other types of response. Thus, viruses are predicted to decrease susceptibility to allergy, but at the same time to decrease tissue repair (a property of type 2 responses) and increase susceptibility to infection by helminths, bacteria and fungi. By contrast, helminths are predicted to increase tissue repair, but at the same time decrease resistance to viruses and bacteria and increase susceptibility to allergy. Following the same logic, most bacteria and fungi are predicted to decrease susceptibility to allergies, but also to increase susceptibility to viruses, tumours and helminths, and to inhibit tissue repair.

P E R S P E C T I V E S

530 | AUGUST 2016 | VOLUME 16 www.nature.com/nri

© 2016

Macmillan

Publishers

Limited.

All

rights

reserved. ©

2016

Macmillan

Publishers

Limited.

All

rights

reserved.

17. Jerne, N. K. Towards a network theory of the immune system. Ann. Immunol. (Paris) 125C, 373–389 (1974).

18. Hoffmann, G. W. A theory of regulation and self-nonself discrimination in an immune network. Eur. J. Immunol. 5, 638–647 (1975).

19. Rowley, D. A., Kohler, H. & Cowan, J. D. An immunologic network. Contemp. Top. Immunobiol. 9, 205–230 (1980).

20. Gershon, R. K. & Kondo, K. Cell interactions in the induction of tolerance: the role of thymic lymphocytes. Immunology 18, 723–737 (1970).

21. Gershon, R. K. A disquisition on suppressor T cells. Transplant Rev. 26, 170–185 (1975).

22. Germain, R. N. Maintaining system homeostasis: the third law of Newtonian immunology. Nat. Immunol. 13, 902–906 (2012).

23. Bennett, C. L. et al. The immune dysregulation, polyendocrinopathy, enteropathy, X-linked syndrome (IPEX) is caused by mutations of FOXP3. Nat. Genet. 27, 20–21 (2001).

24. Brunkow, M. E. et al. Disruption of a new forkhead/winged-helix protein, scurfin, results in the fatal lymphoproliferative disorder of the scurfy mouse. Nat. Genet. 27, 68–73 (2001).

25. Wildin, R. S. et al. X-linked neonatal diabetes mellitus, enteropathy and endocrinopathy syndrome is the human equivalent of mouse scurfy. Nat. Genet. 27, 18–20 (2001).

26. Josefowicz, S. Z., Lu, L. F. & Rudensky, A. Y. Regulatory T cells: mechanisms of differentiation and function. Annu. Rev. Immunol. 30, 531–564 (2012).

27. Lochner, M. et al. In vivo equilibrium of proinflammatory IL-17+ and regulatory IL-10+ Foxp3+ RORγt+ T cells. J. Exp. Med. 205, 1381–1393 (2008).

28. Mauri, C. & Bosma, A. Immune regulatory function of B cells. Annu. Rev. Immunol. 30, 221–241 (2012).

29. Serafini, P., Borrello, I. & Bronte, V. Myeloid suppressor cells in cancer: recruitment, phenotype, properties, and mechanisms of immune suppression. Semin. Cancer Biol. 16, 53–65 (2006).

30. Mosmann, T. R. & Coffman, R. L. TH1 and TH2 cells: different patterns of lymphokine secretion lead to different functional properties. Annu. Rev. Immunol. 7, 145–173 (1989).

31. Bettelli, E. et al. Reciprocal developmental pathways for the generation of pathogenic effector TH17 and regulatory T cells. Nature 441, 235–238 (2006).

32. Harrington, L. E. et al. Interleukin 17-producing CD4+ effector T cells develop via a lineage distinct from the T helper type 1 and 2 lineages. Nat. Immunol. 6, 1123–1132 (2005).

33. Matzinger, P. Friendly and dangerous signals: is the tissue in control? Nat. Immunol. 8, 11–13 (2007).

34. Trinchieri, G. & Gerosa, F. Immunoregulation by interleukin-12. J. Leukoc. Biol. 59, 505–511 (1996).

35. Eberl, G., Colonna, M., Di Santo, J. P. & McKenzie, A. N. Innate lymphoid cells: a new paradigm in immunology. Science 348, aaa6566 (2015).

36. Cua, D. J. et al. Interleukin-23 rather than interleukin-12 is the critical cytokine for autoimmune inflammation of the brain. Nature 421, 744–748 (2003).

37. Sutton, C., Brereton, C., Keogh, B., Mills, K. H. & Lavelle, E. C. A crucial role for interleukin (IL)-1 in the induction of IL-17-producing T cells that mediate autoimmune encephalomyelitis. J. Exp. Med. 203, 1685–1691 (2006).

38. Klose, C. S. et al. A T-bet gradient controls the fate and function of CCR6–RORγt+ innate lymphoid cells. Nature 494, 261–265 (2013).

39. Buonocore, S. et al. Innate lymphoid cells drive interleukin-23-dependent innate intestinal pathology. Nature 464, 1371–1375 (2010).

40. Allen, J. E. & Sutherland, T. E. Host protective roles of type 2 immunity: parasite killing and tissue repair, flip sides of the same coin. Semin. Immunol. 26, 329–340 (2014).

41. Saenz, S. A., Taylor, B. C. & Artis, D. Welcome to the neighborhood: epithelial cell-derived cytokines license innate and adaptive immune responses at mucosal sites. Immunol. Rev. 226, 172–190 (2008).

42. Gordon, S. & Martinez, F. O. Alternative activation of macrophages: mechanism and functions. Immunity 32, 593–604 (2010).

43. Knipper, J. A. et al. Interleukin-4 receptor α signaling in myeloid cells controls collagen fibril assembly in skin repair. Immunity 43, 803–816 (2015).

44. Sadtler, K. et al. Developing a pro-regenerative biomaterial scaffold microenvironment requires T helper 2 cells. Science 352, 366–370 (2016).

73. Russell, S. L. et al. Early life antibiotic-driven changes in microbiota enhance susceptibility to allergic asthma. EMBO Rep. 13, 440–447 (2012).

74. Olszak, T. et al. Microbial exposure during early life has persistent effects on natural killer T cell function. Science 336, 489–493 (2012).

75. Hill, D. A. et al. Commensal bacteria-derived signals regulate basophil hematopoiesis and allergic inflammation. Nat. Med. 18, 538–546 (2012).

76. Kernbauer, E., Ding, Y. & Cadwell, K. An enteric virus can replace the beneficial function of commensal bacteria. Nature 516, 94–98 (2014).

77. Moro, K. et al. Interferon and IL-27 antagonize the function of group 2 innate lymphoid cells and type 2 innate immune responses. Nat. Immunol. 17, 76–86 (2016).

78. Duerr, C. U. et al. Type I interferon restricts type 2 immunopathology through the regulation of group 2 innate lymphoid cells. Nat. Immunol. 17, 65–75 (2016).

79. Thomas, S. A. & Palmiter, R. D. Thermoregulatory and metabolic phenotypes of mice lacking noradrenaline and adrenaline. Nature 387, 94–97 (1997).

80. Qiu, Y. et al. Eosinophils and type 2 cytokine signaling in macrophages orchestrate development of functional beige fat. Cell 157, 1292–1308 (2014).

81. Lee, M. et al. Activated type 2 innate lymphoid cells regulate beige fat biogenesis. Cell 160, 74–87 (2015).

82. Brestoff, J. R. et al. Group 2 innate lymphoid cells promote beiging of white adipose tissue and limit obesity. Nature 519, 242–246 (2015).

83. Burcelin, R., Garidou, L. & Pomie, C. Immuno-microbiota cross and talk: the new paradigm of metabolic diseases. Semin. Immunol. 24, 67–74 (2012).

84. Kim, K. A., Gu, W., Lee, I. A., Joh, E. H. & Kim, D. H. High fat diet-induced gut microbiota exacerbates inflammation and obesity in mice via the TLR4 signaling pathway. PLoS ONE 7, e47713 (2012).

85. Yang, J. Y. et al. Enteric viruses ameliorate gut inflammation via toll-like receptor 3 and toll-like receptor 7-mediated interferon-β production. Immunity 44, 889–900 (2016).

86. Henry, T. et al. Type I IFN signaling constrains IL-17A/F secretion by γδ T cells during bacterial infections. J. Immunol. 184, 3755–3767 (2010).

87. Wu, V. et al. Plasmacytoid dendritic cell-derived IFNα modulates Th17 differentiation during early Bordetella pertussis infection in mice. Mucosal Immunol. 9, 777–786 (2016).

88. Chong, W. P. et al. NK-DC crosstalk controls the autopathogenic Th17 response through an innate IFN-γ-IL-27 axis. J. Exp. Med. 212, 1739–1752 (2015).

89. Finlay, C. M. et al. Helminth products protect against autoimmunity via innate type 2 cytokines IL-5 and IL-33, which promote eosinophilia. J. Immunol. 196, 703–714 (2016).

90. Reddy, A. & Fried, B. An update on the use of helminths to treat Crohn’s and other autoimmunune diseases. Parasitol. Res. 104, 217–221 (2009).

91. Gaboriau-Routhiau, V. et al. The key role of segmented filamentous bacteria in the coordinated maturation of gut helper T cell responses. Immunity 31, 677–689 (2009).

92. Ivanov, I. I. et al. Induction of intestinal Th17 cells by segmented filamentous bacteria. Cell 139, 485–498 (2009).

93. Kriegel, M. A. et al. Naturally transmitted segmented filamentous bacteria segregate with diabetes protection in nonobese diabetic mice. Proc. Natl Acad. Sci. USA 108, 11548–11553 (2011).

94. Sofi, M. H. et al. pH of drinking water influences the composition of gut microbiome and type 1 diabetes incidence. Diabetes 63, 632–644 (2014).

95. Barton, E. S. et al. Herpesvirus latency confers symbiotic protection from bacterial infection. Nature 447, 326–329 (2007).

96. Navarini, A. A. et al. Increased susceptibility to bacterial superinfection as a consequence of innate antiviral responses. Proc. Natl Acad. Sci. USA 103, 15535–15539 (2006).

97. Alonso, J. M. et al. A model of meningococcal bacteremia after respiratory superinfection in influenza A virus-infected mice. FEMS Microbiol. Lett. 222, 99–106 (2003).

98. McCullers, J. A. The co-pathogenesis of influenza viruses with bacteria in the lung. Nat. Rev. Microbiol. 12, 252–262 (2014).

99. Osborne, L. C. et al. Virus-helminth coinfection reveals a microbiota-independent mechanism of immunomodulation. Science 345, 578–582 (2014).

45. Tauber, A. I. & Chernyak, L. Metchnikoff and the Origins of Immunology: From Metaphor to Theory (Oxford Univ. Press, 1991).

46. Besnard, A. G. et al. Dual role of IL-22 in allergic airway inflammation and its cross-talk with IL-17A. Am. J. Respir. Crit. Care Med. 183, 1153–1163 (2011).

47. Matzinger, P. & Kamala, T. Tissue-based class control: the other side of tolerance. Nat. Rev. Immunol. 11, 221–230 (2011).

48. Forrester, J. V., Xu, H., Kuffova, L., Dick, A. D. & McMenamin, P. G. Dendritic cell physiology and function in the eye. Immunol. Rev. 234, 282–304 (2010).

49. Van der Sluis, M. et al. Muc2-deficient mice spontaneously develop colitis, indicating that MUC2 is critical for colonic protection. Gastroenterology 131, 117–129 (2006).

50. Vaishnava, S. et al. The antibacterial lectin RegIIIγ promotes the spatial segregation of microbiota and host in the intestine. Science 334, 255–258 (2011).

51. Mkaddem, S. B. et al. IgA, IgA receptors, and their anti-inflammatory properties. Curr. Top. Microbiol. Immunol. 382, 221–235 (2014).

52. Macpherson, A. J. & Uhr, T. Induction of protective IgA by intestinal dendritic cells carrying commensal bacteria. Science 303, 1662–1665 (2004).

53. Koch, M. A. et al. The transcription factor T-bet controls regulatory T cell homeostasis and function during type 1 inflammation. Nat. Immunol. 10, 595–602 (2009).

54. Wohlfert, E. A. et al. GATA3 controls Foxp3+ regulatory T cell fate during inflammation in mice. J. Clin. Invest. 121, 4503–4515 (2011).

55. Schiering, C. et al. The alarmin IL-33 promotes regulatory T-cell function in the intestine. Nature 513, 564–568 (2014).

56. Ohnmacht, C. et al. The microbiota regulates type 2 immunity through RORγt+ T cells. Science 349, 989–993 (2015).

57. Wohlfert, E. & Belkaid, Y. Plasticity of Treg at infected sites. Mucosal Immunol. 3, 213–215 (2010).

58. Sefik, E. et al. Individual intestinal symbionts induce a distinct population of RORγ+ regulatory T cells. Science 349, 993–997 (2015).

59. Hams, E., Locksley, R. M., McKenzie, A. N. & Fallon, P. G. Cutting edge: IL-25 elicits innate lymphoid type 2 and type II NKT cells that regulate obesity in mice. J. Immunol. 191, 5349–5353 (2013).

60. Molofsky, A. B. et al. Innate lymphoid type 2 cells sustain visceral adipose tissue eosinophils and alternatively activated macrophages. J. Exp. Med. 210, 535–549 (2013).

61. Kim, H. Y. et al. Interleukin-17-producing innate lymphoid cells and the NLRP3 inflammasome facilitate obesity-associated airway hyperreactivity. Nat. Med. 20, 54–61 (2014).

62. Bleriot, C. et al. Liver-resident macrophage necroptosis orchestrates type 1 microbicidal inflammation and type-2-mediated tissue repair during bacterial infection. Immunity 42, 145–158 (2015).

63. Cannon, W. B. Organization for physiological homeostasis. Phys. Rev. IX, 399–431 (1929).

64. Wick, G. et al. The immunology of fibrosis. Annu. Rev. Immunol. 31, 107–135 (2013).

65. Geremia, A. et al. IL-23-responsive innate lymphoid cells are increased in inflammatory bowel disease. J. Exp. Med. 208, 1127–1133 (2011).

66. Wu, H. J. et al. Gut-residing segmented filamentous bacteria drive autoimmune arthritis via T helper 17 cells. Immunity 32, 815–827 (2010).

67. Van Praet, J. T. et al. Commensal microbiota influence systemic autoimmune responses. EMBO J. 34, 466–474 (2015).

68. Theofilopoulos, A. N., Baccala, R., Beutler, B. & Kono, D. H. Type I interferons (α/β) in immunity and autoimmunity. Annu. Rev. Immunol. 23, 307–336 (2005).

69. Prioult, G. & Nagler-Anderson, C. Mucosal immunity and allergic responses: lack of regulation and/or lack of microbial stimulation? Immunol. Rev. 206, 204–218 (2005).

70. von Mutius, E. & Vercelli, D. Farm living: effects on childhood asthma and allergy. Nat. Rev. Immunol. 10, 861–868 (2010).

71. Foliaki, S. et al. Antibiotic use in infancy and symptoms of asthma, rhinoconjunctivitis, and eczema in children 6 and 7 years old: International Study of Asthma and Allergies in Childhood Phase III. J. Allergy Clin. Immunol. 124, 982–989 (2009).

72. Droste, J. H. et al. Does the use of antibiotics in early childhood increase the risk of asthma and allergic disease? Clin. Exp. Allergy 30, 1547–1553 (2000).

P E R S P E C T I V E S

NATURE REVIEWS | IMMUNOLOGY VOLUME 16 | AUGUST 2016 | 531

© 2016

Macmillan

Publishers

Limited.

All

rights

reserved. ©

2016

Macmillan

Publishers

Limited.

All

rights

reserved.

100. Baldridge, M. T. et al. Commensal microbes and interferon-λ determine persistence of enteric murine norovirus infection. Science 347, 266–269 (2015).

101. Skov, M., Poulsen, L. K. & Koch, C. Increased antigen-specific Th-2 response in allergic bronchopulmonary aspergillosis (ABPA) in patients with cystic fibrosis. Pediatr. Pulmonol 27, 74–79 (1999).

102. Margalit, A. & Kavanagh, K. The innate immune response to Aspergillus fumigatus at the alveolar surface. FEMS Microbiol. Rev. 39, 670–687 (2015).

103. Coffelt, S. B. et al. IL-17-producing γδ T cells and neutrophils conspire to promote breast cancer metastasis. Nature 522, 345–348 (2015).

104. Yu, P. et al. Intratumor depletion of CD4+ cells unmasks tumor immunogenicity leading to the rejection of late-stage tumors. J. Exp. Med. 201, 779–791 (2005).

105. Coffman, R. L., Lebman, D. A. & Shrader, B. Transforming growth factor β specifically enhances IgA production by lipopolysaccharide-stimulated murine B lymphocytes. J. Exp. Med. 170, 1039–1044 (1989).

106. Shalapour, S. et al. Immunosuppressive plasma cells impede T-cell-dependent immunogenic chemotherapy. Nature 521, 94–98 (2015).

107. Sprent, J., Lo, D., Gao, E. K. & Ron, Y. T cell selection in the thymus. Immunol. Rev. 101, 173–190 (1988).

108. Kisielow, P., Bluthmann, H., Staerz, U. D., Steinmetz, M. & von Boehmer, H. Tolerance in T-cell-receptor transgenic mice involves deletion of nonmature CD4+8+ thymocytes. Nature 333, 742–746 (1988).

109. Vukmanovic, S., Bevan, M. J. & Hogquist, K. A. The specificity of positive selection: MHC and peptides. Immunol. Rev. 135, 51–66 (1993).

110. Jordan, M. S. et al. Thymic selection of CD4+CD25+ regulatory T cells induced by an agonist self-peptide. Nat. Immunol. 2, 301–306 (2001).

111. Annacker, O., Pimenta-Araujo, R., Burlen-Defranoux, O. & Bandeira, A. On the ontogeny and physiology of regulatory T cells. Immunol. Rev. 182, 5–17 (2001).

112. Sakaguchi, S. et al. Immunologic tolerance maintained by CD25+ CD4+ regulatory T cells: their common role in controlling autoimmunity, tumor immunity, and transplantation tolerance. Immunol. Rev. 182, 18–32 (2001).

113. Feuerer, M. et al. Lean, but not obese, fat is enriched for a unique population of regulatory T cells that affect metabolic parameters. Nat. Med. 15, 930–939 (2009).

114. Kolodin, D. et al. Antigen- and cytokine-driven accumulation of regulatory T cells in visceral adipose tissue of lean mice. Cell. Metab. 21, 543–557 (2015).

115. Arpaia, N. et al. A distinct function of regulatory T cells in tissue protection. Cell 162, 1078–1089 (2015).

132. Zhu, J. & Paul, W. E. Peripheral CD4+ T-cell differentiation regulated by networks of cytokines and transcription factors. Immunol. Rev. 238, 247–262 (2010).

133. Serafini, N., Vosshenrich, C. A. & Di Santo, J. P. Transcriptional regulation of innate lymphoid cell fate. Nat. Rev. Immunol. 15, 415–428 (2015).

134. Ivanov, I. I. et al. The orphan nuclear receptor RORγt directs the differentiation program of proinflammatory IL-17+ T helper cells. Cell 126, 1121–1133 (2006).

135. Ciofani, M. et al. A validated regulatory network for Th17 cell specification. Cell 151, 289–303 (2012).

136. Usui, T. et al. T-bet regulates Th1 responses through essential effects on GATA-3 function rather than on IFNG gene acetylation and transcription. J. Exp. Med. 203, 755–766 (2006).

137. Mukasa, R. et al. Epigenetic instability of cytokine and transcription factor gene loci underlies plasticity of the T helper 17 cell lineage. Immunity 32, 616–627 (2010).

138. Milner, J. D. et al. Impaired TH17 cell differentiation in subjects with autosomal dominant hyper-IgE syndrome. Nature 452, 773–776 (2008).

139. Minegishi, Y. et al. Dominant-negative mutations in the DNA-binding domain of STAT3 cause hyper-IgE syndrome. Nature 448, 1058–1062 (2007).

140. Rudra, D. et al. Transcription factor Foxp3 and its protein partners form a complex regulatory network. Nat. Immunol. 13, 1010–1019 (2012).

141. Zhou, L. et al. TGF-β-induced Foxp3 inhibits TH17 cell differentiation by antagonizing RORγt function. Nature 453, 236–240 (2008).

142. Tindemans, I., Serafini, N., Di Santo, J. P. & Hendriks, R. W. GATA-3 function in innate and adaptive immunity. Immunity 41, 191–206 (2014).

143. Yagi, R. et al. The transcription factor GATA3 is critical for the development of all IL-7Rα-expressing innate lymphoid cells. Immunity 40, 378–388 (2014).

144. Zhong, C. et al. Group 3 innate lymphoid cells continuously require the transcription factor GATA-3 after commitment. Nat. Immunol. 17, 169–178 (2016).

AcknowledgementsThe author thanks M. Daëron, T. Pradeu, N. Cerf-Bensussan, A. Freitas and B. Marsland for the many discussions, ideas, corrections and critical reading of the manuscript. The author also thanks members of the Microenvironment and Immunity unit of the Institut Pasteur for discussions and experiments over the years that led to the ideas presented in this Essay, as well as the students and postdoctoral researchers of the Department of Immunology of the Institut Pasteur for the Forest seminar series and critical reading of the manuscript.

Competing interests statementThe author declares no competing interests.

116. Streilein, J. W. Neonatal tolerance: towards an immunogenetic definition of self. Immunol. Rev. 46, 123–146 (1979).

117. Wu, H. Y. & Weiner, H. L. Oral tolerance. Immunol. Res. 28, 265–284 (2003).

118. Cerutti, A., Chen, K. & Chorny, A. Immunoglobulin responses at the mucosal interface. Annu. Rev. Immunol. 29, 273–293 (2011).

119. Lochner, M. et al. Microbiota-induced tertiary lymphoid tissues aggravate inflammatory disease in the absence of RORγt and LTi cells. J. Exp. Med. 208, 125–134 (2011).

120. Datta, S. K. et al. Mucosal adjuvant activity of cholera toxin requires Th17 cells and protects against inhalation anthrax. Proc. Natl Acad. Sci. USA 107, 10638–10643 (2010).

121. Snider, D. P., Marshall, J. S., Perdue, M. H. & Liang, H. Production of IgE antibody and allergic sensitization of intestinal and peripheral tissues after oral immunization with protein Ag and cholera toxin. J. Immunol. 153, 647–657 (1994).

122. Coley, W. B. The treatment of malignant tumors by repeated inoculations of erysipelas. With a report of ten original cases. 1893. Clin Orthop Relat Res, 3–11 (1991).

123. Tsung, K. & Norton, J. A. Lessons from Coley’s Toxin. Surg. Oncol. 15, 25–28 (2006).

124. Silverstein, M. J., DeKernion, J. & Morton, D. L. Malignant melanoma metastatic to the bladder. Regression following intratumor injection of BCG vaccine. JAMA 229, 688 (1974).

125. Solt, L. A. et al. Suppression of TH17 differentiation and autoimmunity by a synthetic ROR ligand. Nature 472, 491–494 (2011).

126. Huh, J. R. et al. Digoxin and its derivatives suppress TH17 cell differentiation by antagonizing RORγt activity. Nature 472, 486–490 (2011).

127. Hammad, H. et al. House dust mite allergen induces asthma via Toll-like receptor 4 triggering of airway structural cells. Nat. Med. 15, 410–416 (2009).

128. Finlay, C. M., Walsh, K. P. & Mills, K. H. Induction of regulatory cells by helminth parasites: exploitation for the treatment of inflammatory diseases. Immunol. Rev. 259, 206–230 (2014).

129. Weinstock, J. V. & Elliott, D. E. Helminth infections decrease host susceptibility to immune-mediated diseases. J. Immunol. 193, 3239–3247 (2014).

130. Driss, V. et al. The schistosome glutathione S-transferase P28GST, a unique helminth protein, prevents intestinal inflammation in experimental colitis through a Th2-type response with mucosal eosinophils. Mucosal Immunol. 9, 322–335 (2015).

131. Janeway, C. A. Jr. Approaching the asymptote? Evolution and revolution in immunology. Cold Spring Harb. Symp. Quant. Biol. 1989. 54: 1–13 J. Immunol. 191, 4475–4487 (2013).

P E R S P E C T I V E S

532 | AUGUST 2016 | VOLUME 16 www.nature.com/nri

© 2016

Macmillan

Publishers

Limited.

All

rights

reserved.