Roles of Innate lymphoid Cells (ILCs) in Allergic Diseases: The 10-year anniversary for ILC2s

Abstract

In the twelve years since the discovery of innate lymphoid cells (ILCs), our knowledge of their immunobiology has expanded rapidly. Group 2 ILCs (ILC2s) respond rapidly to allergen exposure and environmental insults in mucosal organs, producing type 2 cytokines. Early studies showed epithelium-derived cytokines activate ILC2s, resulting in eosinophilia, mucus hypersecretion and remodeling of mucosal tissues. We now know that ILC2s are regulated by other cytokines, eicosanoids, and neuropeptides as well, and interact with both immune and stromal cells. Furthermore, ILC2s exhibit plasticity by adjusting their functions depending on their tissue environment and may consist of several heterogenous subpopulations. Clinical studies show that ILC2s are involved in asthma, allergic rhinitis, chronic rhinosinusitis, food allergy and eosinophilic esophagitis. However, much remains unknown about the immunological mechanisms involved. Beneficial functions of ILCs in maintenance or restoration of tissue well-being and human health also need to be clarified. As our understanding of the crucial functions ILCs play in both homeostasis and disease pathology expands, we are poised to make tremendous strides in diagnostic and therapeutic options for patients with allergic diseases. This review summarizes discoveries in immunobiology of ILCs and their roles in allergic diseases in the past 5 years, discusses controversies and gaps in our knowledge and suggests future research directions.

Introduction

Allergic diseases result when the immune system inappropriately responds to a nonpathogenic environmental cue. Immunologically, allergic diseases generally follow type 2 immune programs (1). For many years, the CD4⁺ Th2 cell has been recognized as the driving force behind these type 2 immune responses and clinical symptoms, releasing cytokines IL-4, IL-5, IL-9 and IL-13, which promote class switch to IgE, mucous production, smooth muscle hyperreactivity and increased vascular permeability (1). Th2 cells also recruit and activate inflammatory cells, including eosinophils, basophils and mast cells, which mediate the effector phases of the immune responses via release of toxic proteins, cytokines, radical oxygen species (ROS), histamine and leukotrienes (1). Studies also suggested that the type 2 immune response is initiated by production and/or extracellular release of cytokines, such IL-33, IL-25, and TSLP, from mucosal epithelial cells, sentinel macrophages and potentially other cell types, which results in activation of dendritic cells (DCs) and proliferation and differentiation of antigen-specific CD4⁺ T cells (2).

However, in the early 2000s, it became clear that initial stages of type 2 immune responses also involve activation of an unknown cell type, which does not express T cell receptors (3, 4). In 2010, three publications released almost simultaneously identified this novel cell as a lymphoid cell with innate type 2 immunologic activities (5–7). Since then, extensive research has uncovered the basic biology of this cell type, now known as the group 2 innate lymphoid cell (ILC2), and its integral roles in health and diseases (8, 9). The field of ILC2s and other innate lymphoid cells (ILCs) has expanded quickly with advanced knowledge of basic immunology and availability of cutting-edge technologies, such as single cell RNA sequencing (scRNAseq) and multiplex flow and mass cytometry. In this review, we will briefly discuss the basic biology of ILC2s and their closely related counterparts, group 1 innate lymphoid cells (ILC1s) and group 3 innate lymphoid cells (ILC3s). We will summarize the recent advances in our understanding of the immunobiology of ILC2s. We will then discuss the roles of ILC2s and other ILCs in allergic diseases with emphasis on human studies and papers published during the past 5 years; we will build upon the excellent review article published in the JOURNAL (10). Please note, as ILCs are relatively new, there are several controversies and conflicting reports in the field. While we attempt to summarize the information in an unbiased manner, some details may not be fully reconciled among the papers covered in this review.

Immunobiology of ILCs

Innate lymphoid cells (ILCs) reside mainly in mucosal tissues, where they are poised to respond quickly to environmental pathogens and insults. ILCs lack conventional lineage markers and do not express rearranged antigen-specific receptors but instead are activated by cytokines and other mediators released by epithelial cells, macrophages and DCs (8, 9). Once triggered, ILCs secrete a large quantity of cytokines that recruit other immune and inflammatory cells, activate adaptive immune cells, and mediate physiologic and pathologic responses (9). ILCs are described as innate counterparts to CD4⁺ T helper cells, mimicking cytokine profiles released by Th1, Th2 and Th17 cells but without the need for activation by specific antigens (8).

The current knowledge regarding the subsets and origin of human ILCs is summarized in Figure 1. ILC1s are regulated by the transcription factor T-bet and release IFN-γ and TNF-α upon activation by IL-12, IL-15 and IL-1. ILC1s provide protection against intracellular viruses and bacteria (8, 9). ILC2s are under the control of GATA-binding protein 3 (GATA3) (11, 12) and release IL-5, IL-13 and other type 2 cytokines, similar to Th2 cells (5–7), and play crucial roles in tissue homeostasis, helminth clearance and pathophysiology of allergic diseases (5–8, 13, 14). ILC3s, like Th17 cells, are dependent on the retinoic acid-related orphan receptor γt (RORγt) and aryl hydrocarbon receptor (AHR), and release IL-17 and IL-22 upon activation by IL-1β or IL-23 (15). ILC3s provide protection against extracellular bacteria and promote autoimmune diseases (15). Similar to ILC3s, lymphoid tissue inducer cells (LTi), that are involved in the formation of secondary lymphoid organs during embryogenesis, are also dependent on RORγt (16).

Figure 1. Development of ILCs.

ILCs begin as common lymphoid progenitors (CLP) which differentiate into common innate lymphoid progenitors (CILP) under the control of Id2. CILPs can differentiate into common helper innate lymphoid progenitors (CHILP) or into NK cell precursors (NKP). NKPs further differentiate into NK cells while CHILPs can follow one of three programs. First, lymphoid tissue inducer progenitors (LTiP) arise from CHILPs and can give rise to lymphoid tissue inducer cells (LTi). Second, CHILPs under the influence of Id3 differentiate into regulatory innate lymphoid cells (ILCreg). Third, CHILPs give rise to innate lymphoid cell precursors (ILCP). ILCPs differentiate into ILC1, ILC2 or ILC3 subsets regulated by the lineage-specific transcription factors T-bet, GATA3/RORα, or RORγt/AHR, respectively. Note that differentiated ILC subsets maintain some degree of plasticity as represented in Figure 3 for ILC2s. AHR, aryl hydrocarbon receptor; Areg, amphiregulin; EOMES, Eomesodermin; GATA3, GATA Binding Protein 3; GM-CSF, granulocyte macrophage colony-stimulating factor; Id2, inhibitor of DNA binding 2; Id3, inhibitor of DNA binding 3; IFN-γ, interferon-γ; IL, interleukin; ILC, innate lymphoid cell; NK, natural killer; RORα, RAR-related orphan receptor a; RORγt, RAR-related orphan receptor γt; T-bet, T-box transcription factor; TGF-β, transforming growth factor–β; TNF-α, tumor necrosis factor-a; TSLP, thymic stromal lymphopoietin

Recently, a subpopulation of ILCs, named regulatory ILCs (ILCregs), that secretes IL-10 and suppresses activation of other ILCs and intestinal inflammation, was identified in mouse gut (17). ILCregs expand from common helper innate lymphoid progenitors (CHILPs), depend on the transcriptional regulator Id3 and autocrine TGF-β1, and express a set of genes that are distinct from other ILCs or CD4⁺ regulatory T (Treg) cells. In addition, in vitro culture of human or mouse ILC2s with a cocktail of cytokines (IL-33, IL-2 and IL-7) and retinoic acid (RA) induced production of IL-10 and promoted their capacity to suppress effector functions of CD4⁺ T cells and ILC2s (18). ILCs and ILC2s that produce IL-10 were also identified in nasal polyp tissues from patients with chronic rhinosinusitis (CRS) and in the lungs of mice exposed to allergens or IL-33 (17–19). These findings suggest the existence of ILCs with a regulatory function, similar to Treg cells. Finally, NK cells, the innate counterpart to cytotoxic CD8⁺ T cells, were originally grouped with ILC1s (9). However, fate-mapping studies showed that NK cells branch off the development pathway earlier than do the other ILC groups and should be considered a unique subset (9).

ILC2s are likely most relevant to allergic diseases although other ILCs may also be involved. Early studies using animal models implicate ILC2s in the pathogenesis of allergic airway diseases (20–24), atopic dermatitis (25, 26) and food allergy (27–29) as described more in detail below (Table 1). ILC2s are activated indirectly when allergens and other environmental cues stimulate tissue cells and immune cells to release several types of immunologic and biological molecules, including epithelium-derived cytokines (IL-33, IL-25, TSLP) (5–7, 22–24), eicosanoids (LTD₄, PGD₂) (21, 30) and neuropeptides and hormones (neuromedin U (NMU), vasoactive intestinal protein (VIP)) (31–34). ILC2-derived type 2 cytokines and amphiregulin mediate multiple homeostatic, immunologic and pathologic processes, including helminth clearance, epithelial repair from injury and beiging of white adipose tissue (5–7, 13, 35, 36), as well as eosinophil recruitment, fibrogenesis and goblet cell hyperplasia and epithelial mucus production (22–24, 37, 38). Some processes, such as helminth clearance and epithelial repair, may be beneficial for the host. For example, stomach ILC2s protect against Helicobacter pylori infection in mice by promoting IgA antibody production (39). Others drive the immunopathology of type 2 responses. Altogether, ILC2s likely have both beneficial and pathologic effects to the host, and it will be critical to understand their immunobiology better and to dissect fully the context-dependent roles for ILC2s in human diseases.

Table 1.

Roles of ILC2s in allergic diseases: Insights from animal models^*

Model	Activating signals	ILC2-derived mediators	Pathophysiology	Ref
Alternaria alternata; intranasal (i.n.)	IL-33	IL-5, IL-13 (putative)	Eosinophilia, mucus overproduction, goblet cell hyperplasia, airway inflammation	23
Papain; i.n.	IL-33, TSLP	IL-5, IL-13 (putative)	Eosinophilia, mucus overproduction	20
OVA; intraperitoneal (i.p.) sensitization, aerosolized challenge	IL-33, IL-25 (putative)	IL-5, IL-13	NA	22
House dust mite (HDM); intratracheal (i.t.) sensitization, i.n. challenge	NA	IL-5, IL-13	Eosinophilia	24
Alternaria alternata; i.n.	LTD4	IL-5, IL-13, IL-4	NA	21
Calcipotriol (MC903); topical (C57BL/6)	TSLP	IL-5, IL-13 (putative)	Stratum corneum thickening, epidermal hyperplasia, dermal inflammatory infiltration	26
Skin-specific overexpression of IL-33	IL-33	IL-5	Eosinophilia, skin lesions, itch, eosinophilia, mast cell activation	25
OVA; i.p. sensitization, gavage challenge	IL-25	IL-13	goblet cell hyperplasia, intestinal permeability, dysregulated gut barrier function	27
Peanut; gavage sensitization and challenge	IL-33	IL-4	Decreased body temperature, peanut-specific IgE	28
Alternaria alternata; i.n.	IL-33	IL-5	Eosinophilopoiesis	184
Alternaria alternata	LTC4	IL-5, IL-13	Airway eosinophilia	122
IL-33; i.n.	IL-33	IL-13	Disruption of epithelial tight junctions and epithelial barrier	118

^*Mouse models of allergic diseases that examined the roles of ILC2s are summarized by focusing on earlier studies. Note that models with helminths are not included to provide more rooms for the models using natural and model allergens.

Recent advances in immunobiology of ILC2s

During the past several years, our knowledge of the immunobiology of ILC2s has advanced rapidly. New pathways and molecules involved in activation and regulation of ILC2s as well as their unique roles within a network of type 2 immunity have been described. We have learned that previous phenotypic and functional definitions of ILC2s may be too simplistic, and that ILC2s demonstrate plasticity and heterogeneity.

Figure 2 summarizes our current knowledge regarding factors that activate or inhibit effector functions of ILC2s and their potential cellular sources. One of the new observations during the past 5 years involves neuropeptides. In mouse models of allergic airway disease, both VIP and NMU amplified allergen- or cytokine-induced IL-5 production from lung ILC2s (31, 34). The NMU receptor, Nmur1, was expressed by human blood and intestinal ILC2s, and NMU alone was sufficient to stimulate type 2 cytokine production by mouse ILC2s (32, 33).

Figure 2. Factors that activate and inhibit ILC2s and their potential cellular sources with focus on allergic immune responses. Activation.

Activation of ILC2s by cytokines, eicosanoids and neuropeptides generated by epithelial cells, keratinocytes, macrophages, dendritic cells, Tuft cells, brush cells, and neurons in allergic diseases has been demonstrated in both mice and humans. Inhibition. Epithelial cells, NK cells, DCs, Tregs, ILCregs, neurons and neutrophils inhibit ILC2s both by direct interaction and by generation of cytokines, eicosanoids and neuropeptides. CD62P, P-selectin; CGRP, calcitonin gene-related peptide; cysLT, cysteinyl leukotriene; DC, dendritic cell; DR3, death receptor 3; FFA, free fatty acid; G-CSF, granulocyte colony-stimulating factor; IFN-α, interferon-α IFN-β, interferon-β; IFN-γ, interferon-γ; IL, interleukin; LTC₄, leukotriene C₄; LTD₄, leukotriene D₄; NK, natural killer; NMU, neuromedin U; PGD₂, prostaglandin D₂; PGE₂, prostaglandin E2; PGI₂, prostaglandin I₂; PMN MDSC, polymorphonuclear myeloid-derived suppressor cell; PSGL-1, P-selectin glycoprotein ligand-1; SCF, stem cell factor; TGF-β, transforming growth factor–β; TL1A, tumor necrosis factor-like cytokine 1A; Treg, regulatory T; TSLP, thymic stromal lymphopoietin; VIP, vasoactive intestinal peptide

Calcitonin gene-related peptide (CGRP) is also reported to activate or regulate mouse ILC2s. CGRP is produced by pulmonary neuroendocrine cells (PNECs) and augments IL-33-mediated activation of lung ILC2s (40). Mice lacking CGRP showed reduced airway inflammation after airborne allergen exposure (40). Furthermore, PNECs are increased in the lungs of patients with asthma (40), suggesting pathological roles. Conflicting observations have been reported using mice deficient in the CGRP receptor, which showed that CGRP inhibits IL-33-mediated ILC2 activity and proliferation (41). Differences in mouse strains and experimental models may explain the controversy. Norepinephrine may also regulate murine ILC2 function through the β₂-adrenergic receptor (β₂AR) (42). Indeed, β₂ agonists inhibited ILC2 proliferation and effector function and reduced allergen-induced airway inflammation (42). Further studies will be necessary to clarify the neuronal regulation of ILC2s and elucidate whether and how the novel information from mouse studies applies to human diseases.

Another layer of neuronal regulation of ILC2s is provided by the epithelium. A subset of epithelial cells, known as chemosensory epithelial cells (EpCs), are particularly adapted to respond to neuronal signals and are found in all mucosal tissues (43). They are referred to as Tuft cells in the intestine, brush cells (BrCs) in the trachea, and solitary chemosensory cells (SCCs) in the nasal respiratory mucosa. In mouse intestine, Tuft cells are activated by microbial metabolites to release IL-25, which subsequently activates gastrointestinal (GI) ILC2s to promote a type 2 response against helminth infection (44, 45). In mouse lungs, BrC generation of IL-25 promotes eosinophilic inflammation in response to aeroallergens (46). Recently, two populations of choline acetyltransferase-expressing EpCs, comparable to tracheal BrCs, were identified in the respiratory and olfactory mucosa in mice (47). These mouse nasal BrCs express high levels of transcripts for cysteinyl leukotriene (cysLT) pathway enzymes and IL-25, are activated by ATP via the P2Y2 receptor to produce cysLTs, and mediate innate and adaptive type 2 immune responses in response to aeroallergens (47). As cysLTs are among the many activators of ILC2s (48) (Figure 2), nasal BrCs and other chemosensory epithelial cells may play a major role in ILC2 activation and function in allergic diseases. Exploration of how chemosensory epithelial cells and ILC2s work together to promote and regulate allergic immune responses will provide future insights.

New information has emerged regarding transcriptional regulation of ILC2s. Along with the transcription factor GATA3, other transcription factors and non-coding RNA are also involved in regulation of ILC2s. In mice, BCL11B, previously thought to be a T cell-specific transcription factor, was necessary to maintain mature ILC2s but was dispensable for differentiation of ILC2s (49). Another transcription factor, interferon regulatory factor 7 (IRF7), promotes transcription of Bcl11b, and ILC2 functions are impaired in the absence of Irf7 (50). The microRNA miR-155 is required for airway ILC2 expansion, eosinophilia and lung pathology in allergen sensitization and challenge models in mice (51). miR-155 increased in ILC2s after IL-33 stimulation (51) and may protect ILC2s from apoptosis without affecting their ability to produce cytokines (52). The transcriptional controller core binding factor β (CBFβ) also regulates gene transcription in murine ILC2s (53, 54). CBFβ is dispensable for ILC2 homeostatic maintenance but is necessary for effector functions of ILC2s. Finally, short-chain fatty acids (SCFA) may provide the link between the microbiome and immune tolerance (55). One SCFA, butyrate, inhibited histone deacetylase (HDAC) activity, leading to downregulation of GATA3 and reduced proliferation of murine ILC2s (56). Butyrate also inhibited production of type 2 cytokines by mouse and human ILC2s and reduced airway inflammation and pathology after allergen sensitization and challenge in mice (56). Thus, studies related to molecular regulation of ILC2s have provided considerable new insight, and further investigation into the regulatory mechanisms for ILC2 transcription factors and their repressors will provide valuable information to better understand the biology of ILC2s as well as how to harness them to develop novel strategies to treat allergic diseases.

Role of ILC2s in the network of type 2 immunity

While previous studies highlighted the direct effects of ILC2’s products on immunopathology of diseases and tissue homeostasis, we have started to recognize that ILC2s are a major player in a network of type 2 immunity. Several reports have highlighted the contribution of ILC2s to orchestrate the T cell- or B cell-mediated adaptive type 2 immune response. For example, ILC2s are required to amplify CD4⁺ Th2 cell responses upon infection with Nippostrongylus (N.) brasiliensis or in airways exposed to the cysteine protease papain (57–59). ILC2s from mediastinal lymph nodes (LNs), spleen, and small intestines of naïve mice express MHC-II, and peptide-pulsed ILC2s directly induced proliferation and IL-2 production by T cells in an MHC-II-dependent manner (58, 59). A previous study showed that in vitro stimulation of human DCs with TSLP can lead to upregulation of OX40L and Th2 cell polarization of naïve CD4⁺ T cells (60). Recently, the OX40-OX40L-dependent cellular interaction between ILC2s and naïve CD4⁺ T cells was found to enhance expression of type 2 cytokines by both ILC2s and T cells (61, 62). Upon treatment with IL-33 or papain, expression of OX40L was selectively increased in ILC2s in the lungs (62). Interestingly, OX40L costimulation by ILC2s simultaneously expanded not only effector Th2 cells but also GATA3⁺ Treg cells that express OX40 (62), suggesting ILC2s may enhance or suppress type 2 immunity depending on the context. In addition, ILC2-derived IL-13 promoted Th2 cell responses indirectly by promoting migration of CD40⁺ DCs to the draining LNs and by enhancing priming and differentiation of naive T cells (57). Thus, the information so far suggests that ILC2s may promote or modulate type 2 immunity by multiple mechanisms. Further investigation will be required to better understand the roles that ILC2s play as a member of the type 2 immune network and molecular and cellular mechanisms involved.

How can ILC2s be inhibited?

While the molecules that activate ILC2s have been studied extensively, less is known regarding the mechanisms that inhibit ILC2s. Previous studies show that the cytokines involved in type 1 immunity, such as type 1 IFNs, IFN-γ, and IL-27, and certain lipid mediators, such as prostaglandin I₂ (PGI₂), suppress ILC2 function (10) (Figure 2). In addition, several molecules expressed by ILC2s may also serve in regulation of their own functions. For example, while PD-1 is associated with exhausted T cells, mouse lung ILC2s or human peripheral blood ILC2s also express PD-1 when they are stimulated with IL-33 (63). Engagement of PD-1 suppressed viability of mouse ILC2s and downregulated their effector functions by metabolically reprogramming ILC2s (63). Interactions between ILC2s and other immune cells may also suppress ILC2s. For example, exposure to microbes in the lungs plays important roles in regulating lung ILC2s. Administration of TLR agonists, such as TLR7/8 agonist R848 and TLR9 agonist CpG, inhibited IL-33- or allergen-induced innate type 2 responses in mouse lungs (64, 65). pDCs likely play a role in mediating the suppressive effects of TLR agonists by producing IFN-α quickly, which causes apoptosis of ILC2s (64). pDCs also inhibit ILC2s indirectly by promoting IFN-γ production by NK cells (65). Polymorphonuclear (PMN) myeloid-derived suppressor cells (MDSC) effectively suppressed type 2 cytokine production by ILC2s in vitro and in vivo, via a cyclooxygenase-1 (COX-1) pathway-dependent mechanism (66). In addition, neutrophils have a capacity to inhibit ILC2s indirectly by decreasing the circulating levels of G-CSF, thus inhibiting G-CSF promotion of type 2 cytokine production by lung ILC2s (67).

Sex hormones affect various cell types involved in allergic diseases (68), and they likely play a major role in regulating ILC2s as well. In mice, ILC2s are more prevalent in the lungs of females than those of males (69). Testosterone inhibits ILC2 function (70) and estrogen mediates expansion of ILC2s in female reproductive organs (69). ILC2s from lungs of female mice produce more type 2 cytokines in response to IL-33 than do those from male mice (71). Indeed, lung ILC2s from female mice were more metabolically active and were similar transcriptionally to memory T cells (71). Thus, sex differences in allergic diseases may be explained in part by differences in the numbers and activities of ILC2s. Altogether, ILC2s are regulated by products of other immune cells in the local tissues, and the tissue environment likely plays a major role in defining the identity and function of ILC2s.

Plasticity of ILC2s

Generally, ILCs are tissue-resident and are maintained and expanded locally under specific physiologic and pathologic conditions. Recent studies suggest that ILC2s may consist of several subpopulations and that they bear some plasticity that allows them to vary expression of inflammatory mediators and cell surface molecules depending on their environment. In other words, ILC2s are able to adapt to local tissue environments. Evidence also suggests that ILC2s can migrate between different organs. The plasticity of ILC2s (summarized in Figure 3) likely plays important roles in clinical phenotypes and endotypes of allergic diseases.

Figure 3. Plasticity of ILC2s.

ILC2s can be directed by cytokines to change their abilities to produce certain sets of cytokines. While ILC2s regularly produce type 2 cytokines and amphiregulin, they can be re-programmed to produce IFN-γ (i.e., ILC1-like) when stimulated with IL-12 and IL-1β. Importantly, ILC1-like ILC2s retain the ability to produce IL-13, and this transition can be reversed with IL-4. ILC2s can be induced to produce IL-10 (i.e., IL-10⁺ ILC2s) when cultured with IL-4 or RA in the presence of a cocktail of IL-33, IL-2 and IL-17. ILC2s can produce IL-17 when stimulated with IL-1β, IL-6, IL-23 and TGF-β (i.e., ILC3-like). Note that our knowledge of ILC2 plasticity is in an early stage, and further studies are necessary to identify the optimal conditions to drive transition of ILC2s from one subset to another in vivo and in vitro and to uncover the mechanisms involved. Areg, amphiregulin; IFN-γ, interferon-γ; IL, interleukin; RA, retinoic acid; TGF-β, transforming growth factor–β

In 2015, Huang et al. described a subset of ILC2s, named inflammatory ILC2s (iILC2s), with increased expression of the IL-25 receptor subunit IL-17RB and decreased expression of the IL-33 receptor subunit ST2; conventional ILC2s express high levels of ST2 and low levels of IL-17RB (72). iILC2s responded vigorously to IL-25 and developed into IL-17-producing ILC3-like cells or into nILC2-like cells, depending upon the local cytokine milieu, and contributed to expulsion of the helminth N. brasiliensis and to immunity against Candida albicans (72, 73). This data suggests iILC2s may be a transient progenitor of ILC2s in mice. To date, similar subsets have not been reported in human ILC2s.

More recent studies show that ILC2s are able to migrate to diverse tissues in response to cytokines and the presence of microbes (74). For example, a portion of lung iILC2s were derived from resting ILC2s residing in intestinal lamina propria and were recruited to the lungs in response to systemic administration with IL-25 or to N. brasiliensis infection. Importantly, the migration of ILC2s was dependent on sphingosine 1-phosphate (S1P)-mediated chemotaxis (74), which promotes their lymphatic entry, blood circulation and accumulation in peripheral tissues, similarly to the dynamics of T cells. Bone marrow may also provide another source of ILC2 progenitors for peripheral tissues. By using a scRNAseq approach, Zeis et al. found that at the steady state Il18r1⁺ ILC progenitors (ILCPs) in the bone marrow can traffic into the adult lung and become resident there in mice (75). Furthermore, infection with the parasite N. brasiliensis led to the differentiation of Il18r1⁺ ILCP into three types of ILC2s, including the Il18r1⁺ST2⁻, Il18r1⁻ST2⁺, and Il18r⁺ST2⁺ populations, which can contribute to ILC2 renewal during adulthood. Importantly, a comparable immature population of ILC2s (IL7R⁺IL18R⁺TCF7⁺), which lack lineage markers but express GATA3, were identified in human peripheral blood and lungs (75).

ILC2s generally produce type 2 cytokines such as IL-5 and IL-13 in response to IL-33, IL-25 or TSLP. However, ILC2s can be re-programmed into IFN-γ-producing ILC1-like cells or IL-17-producing ILC3-like cells (76–80) (Figure 3). For example, mouse and human ILC2s can convert into IFN-γ-producing ILC1s when cultured with IL-1β and IL-12 in vitro (76, 78, 81). In vivo conversion of ILC2s to ILC1s has been reported in mice infected with influenza virus (78). ILC2s did not respond directly to IL-12 but required IL-1β that induced expression of IL-12 receptor (IL-12R) as well as ST2 and IL-17RB (77); thus, IL-1β appears to function as a signal to prime ILC2s to respond to the local environment and to promote their plasticity. Interestingly, ILC1-like ILC2s retained IL-13 production and lost their potential to produce IL-5 (76, 81, 82), suggesting that Il5 and Il13 genes are regulated differently. The ILC2 to ILC1 conversion can be reversed by IL-4, derived from basophils, eosinophils or Th2 cells, through activation of STAT6 and induction of GATA3 (12, 76).

ILC2s can also be induced to produce IL-17 (Figure 3). IL-25-responding mature mouse ILC2s produced IL-17 in vitro when cultured with a cocktail of IL-1β, IL-6, IL-23, and TGF-β (80), similar culture conditions are used to induce Th17-type CD4⁺ T cells. In addition, mouse IL-17RB⁺ iILC2s that express both RORγt and GATA3 have the capacity to produce both IL-13 and IL-17 in vitro and in vivo (72). Intestinal ILC2s have accessible regulatory elements for Rorc and Il17 (83). Furthermore, ST2⁺ ILC2s in the lungs produced IL-17 as well as IL-5 and IL-13, dependent on AHR, when mice were treated systemically with IL-33 or intranasally with papain (79). In humans, two subsets of CRTH2⁺ ILC2s were identified in peripheral blood, distinguished by their expression of CD117 (84, 85). The CRTH2⁺CD117⁺ ILC2 population expressed features of ILC3s, such as RORγt and CCR6, in addition to a ILC2 signature, and produced IL-17 in response to IL-1β and IL-23 (85–87). While the molecular mechanisms involved in the ILC2 to ILC3 transition are not fully understood, the transcription factor GFI1 and its upstream regulator BCL11B may be involved as they promote expansion of Th2 cells and block the generation of Th17 and Treg cells (49, 88). Indeed, GFI1 deficiency resulted in decreased expression of ST2, GATA3, and IL-5 and increased production of IL-17 by murine ILC2s (88).

As described above, in vitro culture of human or mouse ILC2s with cytokines such as IL-33, IL-2 and IL-7 together with the vitamin A metabolite retinoic acid (RA) converted them to IL-10-producing ILC2s that downregulate the functions of conventional ILC2s or Th2 cells (18, 19); RA serves an agonist for a transcription factor retinoic acid receptor (RAR) in ILC2s (89). Repeated intranasal exposure to allergens also induced IL-10-producing lung ILC2s in mice, which may emerge during chronic allergic airway inflammation in humans (54, 90). Generation of IL-10-producing ILC2s was dependent on IL-4 that led to increased signaling by the transcription factors cMaf and Blimp-1 (90). Generally, the transcription factors Runx1 and Runx3 negatively regulate IL-10 production in CD4⁺ T cells (54, 91, 92). These transcription factors may also play a pivotal role in regulation of IL-10-producing ILC2s as ILC2s lacking them produced a larger quantity of IL-10 and expressed higher levels of PD-1 and Tigit (54).

Immunological memory is defined as the ability of immune cells to specifically remember the first encounter with a given antigen and to produce a secondary response that is faster and greater in magnitude than the primary response (93). The plasticity of ILC2s may also promote this “memory” response in an innate setting. Martinez-Gonzalez et al reported that some IL-33-stimulated mouse ILC2s can persist in the lungs and LNs for up to 3–4 months and respond vigorously to suboptimal doses of IL-33, TSLP or IL-25 (94). This transition of activation-experienced or “memory” ILC2s to a more aggressive phenotype was accompanied by a conversion from IL-33-responsiveness to IL-25-responsiveness, but not expression of RORγt as observed in iILC2s. The expression levels of ST2 and TSLPR were not upregulated in these “memory” ILC2s, suggesting cell-intrinsic pathways, such as changes in chromatin accessibility and epigenetic modification, may explain functional activation of these ILC2s (95). Together, growing evidence suggests mouse and human ILC2s have considerable functional plasticity to allow them to produce different sets of cytokines and to respond vigorously to exogenous cytokines.

Heterogeneity of ILC2s

In addition to plasticity, tissue environment likely provides an important cue for heterogeneity of ILC2s. Recent analyses of murine and human ILCs have identified transcriptional heterogeneity among ILC2s in the tissues, such as bone marrow, fat, lungs, gut and skin (96–99). For example, in mice deficient in all key tissue-derived cytokines for ILC2s, including IL-33, IL-25 and TSLP, the population of IL-5-expressing ILC2s was significantly reduced in lungs, fat and gut, but they were sustained in skin (99). Skin ILC2s, but not those in other organs, expressed IL-18R and produced IL-5 in response to IL-18. Alternatively, in the lungs of mice exposed to the fungal allergen Alternaria, IL-5-producing ILC2s that do not express ST2 or CD127 have been identified; they were less frequent than the conventional ST2⁺CD127⁺ double-positive population (100). These unconventional ILC2s expressed transcripts for IL-10 and ILC1- and ILC3-related genes as well as those for type 2 cytokines. Similarly, the phenotypes of lung ILC2s were affected by airway exposure to allergens or cytokines, and the changes in these phenotypes were influenced by mouse strain (101). Together, these findings suggest ILC2s express different surface molecules, depending on the organs or inflammatory conditions, which likely allows them to adjust their responses to the environment.

The mechanisms behind this tissue specification and heterogeneity of ILC2s can be explained in part by dynamic regulation of ILC2s throughout their life. A sophisticated fate-mapping study in mice showed that ILC2 development is characterized by three distinct waves of dispersal, expansion and activation coordinated by tissue-specific transcriptional programs (102). ILC2s appear during late gestation; however, a majority of peripheral ILC2 pools are generated by de novo expansion during the neonatal period, leading to expression of tissue-specific genes. Thus, the neonatal period provides a critical time window for expansion and differentiation of ILC2s. It is conceivable that the tissue environment during this period may influence the risk for developing allergic diseases later in life; this hypothesis needs to be examined. Furthermore, IL-33, IL-25 and IL-18 likely provide important signals for tissue specification of ILC2s in the lungs, gut and skin, respectively, during this period.

The heterogeneity of ILC2s has been recognized in humans. Earlier, human ILC2s were identified by their expression of prostaglandin D2 receptor, CRTH2, and a lectin-type receptor, CD161 (103). Subsequently, other molecules, such as CD127, ST2 and CD25, have been used to identify ILC2s in human peripheral blood and tissues. Rather than describing the different methods employed by the investigators in detail, we have summarized staining protocols in Table 2. Recently, in peripheral blood of healthy individuals, two distinct subsets of ILC2s, distinguishable by c-Kit (CD117) expression (85), were identified. Functional analyses showed that c-Kit^lo ILC2s produce more type 2 cytokines as compared to c-Kit^hi ILC2s (85), suggesting that c-Kit^lo ILC2s represent fully mature and committed ILC2s. In contrast, c-Kit^hi ILC2s produced more IL-17 and IFN-γ under ILC3- and ILC1-promoting conditions, respectively, suggesting that they represent a more immature and plastic phenotype (85). Another study showed that CRTH2 may not always be a reliable marker for human ILC2s. Indeed, lung ILC2s expressed lower levels of CRTH2 than did blood ILC2s, suggesting that CRTH2 may be downregulated during migration to the lungs (104).

Table 2.

Variation in human ILC2 characterization^*

CRTH2	CD127	CD161	CD45	GATA3	IL-5	ST2	CD25	IL-17RB	References
X	X	X	X						18, 32, 76, 85, 87, 144, 145, 166
X	X	X							81, 116, 120, 124
X	X		X						30, 62, 63, 65, 66, 106, 110, 114, 118, 128, 142, 162, 171
X	X								84, 108, 113, 119, 121, 136, 137, 161, 163, 164, 165
X		X	X						154 (nasal tissue)
X			X						67, 82, 146, 154 (blood), 155, 160, 180
	X		X			X			112
	X					X			109, 115
	X		X						170
	X								143
						X	X		26
								X	135
				X					165, 172 (both IF)
					X				107

^*ILC2s were defined as lineage-negative cells expressing the receptors and/or transcription factors as designated in this table. Although not included here, lineage-negative panels also varied greatly among the studies described.

Heterogenous populations of ILC2s are also reported in peripheral blood of patients with asthma. For example, the chemokine receptor CCR10 and its ligands CCL17 and CCL28 are known to be involved in epithelial immunity (105). Subjects with severe asthma showed increased frequency of CCR10⁺ ILC2s, compared to non-asthmatic or mild asthma subjects (106). Functionally, CCR10⁻ ILC2s showed robust production of type 2 cytokines while CCR10⁺ ILC2s displayed an ILC1-like phenotype with robust IFN-γ production and little to no IL-5 or IL-13 (106).

Liu et al. used IL-5 expression as a functional surrogate for ILC2s and fully characterized ILC2s for their expression of CD127 (IL-7R) and CRTH2 and effector functions in subjects with asthma (107). While CRTH2⁺CD127⁺ ILC2s produced type 2 cytokines alone, CRTH2⁻CD127⁺ ILC2s and CRTH2⁻CD127⁻ ILC2s produced IL-17 and IFN-γ, respectively, in addition to type 2 cytokines. Importantly, a new set of markers, including CD30 and TNFR2, was proposed to be more suitable to encompass the entire population of IL-5-producing ILC2s (107). Together, ILC2s likely include several subpopulations that may distinguish themselves from the canonical ILC2s by their expression of cell surface molecules and ability to produce non-type 2 cytokines. Furthermore, conventional markers, such as CRTH2 and CD127, may failed to capture certain ILC2 subpopulations. Further investigation into the plasticity and heterogeneity of ILC2s will provide important information to understand the roles of ILC2s in endotypes and clinical manifestation of allergic diseases. New technologies, such as high dimensional flow cytometry and mass cytometry, likely provide robust tools to accomplish the task.

ILCs in Allergic Diseases

Asthma

Asthma and chronic rhinosinusitis (CRS) are chronic inflammation of upper and lower airways, in which the roles of ILCs have been studied extensively in humans. Increased frequencies of ILC2s have been reported in blood and bronchoalveolar lavage (BAL) fluids from patients with allergic asthma as compared to healthy control subjects (108–110) and in blood and sputum specimens of patients with severe asthma compared to mild asthma (111, 112). The numbers of circulating ILC2s correlated with eosinophil counts in sputum and peripheral blood (111, 113). However, some conflicting observations are also reported regarding blood ILC2s. For example, circulating ILC2 frequencies did not differ among well-controlled, partially controlled or uncontrolled asthma patients (110). ILC2 frequencies were increased in sputum, but not in peripheral blood, of pediatric patients with severe asthma (114). Finally, no correlation was found between circulating ILC2s and eosinophils in non-allergic asthma (113). Altogether, while increased numbers of ILC2s in airway specimens from asthma patients are consistently observed, some controversies remain regarding ILC2s in peripheral blood, which may be explained partially by heterogeneity in ILC2 populations and differences in severity of the disease.

Investigation of the activation status of circulating ILC2s has provided consistent outcomes among asthma patients. For example, at steady-state, subjects with severe asthma had more IL-5⁺ ILC2s in peripheral blood and sputum than subjects with mild asthma or disease controls (107, 112). Similarly, increased numbers of IL-13⁺ ILC2s were found in peripheral blood of patients with uncontrolled asthma (110). In vitro stimulation of peripheral blood mononuclear cells (PBMCs) with IL-33 or IL-25 showed increased production of type 2 cytokines in subjects with asthma as compared to healthy controls or those with seasonal allergic rhinitis, suggesting enhanced activity of innate type 2 responses in asthma (108).

Other observations suggest that ILC2s in the airways increase in response to airborne exposure to allergens. For example, allergen challenge in subjects with mild allergic asthma increased the number of activated ILC2s in sputum, but decreased blood ILC2s (115). Similarly, segmental allergen challenge in subjects with allergic asthma resulted in increased BAL fluid ILC2s and decreased blood ILC2s (116), suggesting recruitment of ILC2s from circulation to the airways in response to allergen exposure. Human ILC2s express α₄β₁ and α_Lβ₂ integrins (117), which may play a key role for their recruitment to the airway. Production of PGD₂ and CXCL12 (SDF1) in the airway during allergen exposure may also promote recruitment of ILC2s (116). Thus, the immune responses to environmental allergens may in part explain the increased numbers of ILC2s in the airways of subjects with asthma.

While the mechanisms of ILC2 activation in human asthma are not fully understood, animal models of allergen-induced airway inflammation suggest that ILC2-derived IL-5 and IL-13, which are induced by epithelium-derived cytokines, such as IL-33, IL-25 and TSLP, play major roles in airway eosinophilia, mucus production and airway remodeling (Table 1). In particular, ILC2-derived IL-13 may explain the chronicity of asthma. In mice exposed to a cocktail of allergens for a prolonged period to mimic human asthma, a clinical phenotype of asthma persisted even after discontinuation of allergen exposure (109). A feedback loop involving epithelial cell-derived IL-33 and ILC2-derived IL-13 likely plays a major role in the pathological process as IL-13 promoted IL-33 release and upregulated ST2 expression in epithelial cells where IL-33 worked as an autocrine molecule. Furthermore, IL-13 derived from human ILC2s disrupted tight junctions between human bronchial epithelial cells and induced barrier dysfunction in vitro by suppressing expression of tight junction proteins zona occludens-1 and occludin (118).

Besides cytokines, eicosanoids likely play a critical role in promoting migration and activation of ILC2s in asthma. PGD₂ induced type 2 cytokine production by human blood ILC2s in vitro (30, 119, 120). Furthermore, PGD₂ mobilized ILC2s from peripheral blood to the lungs (104, 121). ILC2s also constitutively express hematopoietic prostaglandin D₂ synthase (HPGDS) and produce PGD₂ upon activation with IL-25, IL-33 and TSLP (120), suggesting that PGD₂ may serve as an autocrine mediator. Leukotrienes such as LTD₄ and LTC₄ also activate ILC2s via their receptor cysLTR1 (21, 122). Indeed, mouse models of asthma demonstrated that LTD₄, LTC₄ and PGD₂ exacerbate the disease (21, 122, 123). On the other hand, certain lipid mediators inhibited ILC2s. For example, PGE₂ suppressed GATA3 and ST2 expression and inhibited activation of ILC2s in vitro (124, 125). Lipoxin A₄ (LXA₄), PGE₂ and PGI₂ ameliorated pathology in mouse models of asthma (119, 125, 126). Furthermore, inhibition of the COX pathway enhanced ILC2 responses and promoted airway pathology after aeroallergen exposure, likely via the loss of PGE₂- and PGI₂-mediated inhibition of IL-33 release from airway epithelial cells (126). Thus, interaction between ILC2s and lipid mediators likely plays a major role in regulating pathophysiology of asthma.

Involvement of ILC2s was also observed with other triggers of asthma besides allergens. For example, in mice, respiratory syncytial virus (RSV) infection increased both lung ILC2 numbers and IL-13 production by ILC2s (127). In experimental rhinovirus (RV) infection in humans, patients with asthma showed increased levels of IL-33 and type 2 cytokines in nasal fluids (128). Supernatants from RV-infected bronchial epithelial cells induced type 2 cytokine release by isolated ILC2s (128), suggesting that epithelium-derived IL-33 may activate ILC2s during respiratory virus infection. Similarly, in a mouse model, exposure to cigarette smoke extract (129) and high concentrations of oxygen (130) activated ILC2s and promoted asthma-like pathology in the airways. Importantly, these ILC2 responses were dependent on IL-33 and treatment with the Nrf2 agonist sulforaphane alleviated both IL-33 release and airway inflammation (130), suggesting critical roles for oxidative stress in airway epithelium. As oxidative stress also plays a role in allergen-induced IL-33 release from airway epithelium (131), it may provide a fundamental mechanism involved in innate type 2 responses to a variety of agents that the respiratory tract encounters.

The gastrointestinal microbiome also likely plays a role in regulating ILC2s. Crosstalk among intestinal microbiota, macrophages and ILC3 has been known to maintain immune homeostasis in the gut (132). Chua et al. discovered the link between colonization with spore forming Lachnospiraceae bacteria Ruminococcus (R.) gnavus and asthma development in infants (133). Feeding with R. gnavus promoted allergic sensitization to ovalbumin (OVA) antigen in mice, which was associated with increased numbers of ILC2s in the intestines (133). The authors suggested a mechanism in which production of IL-33, IL-25 and TSLP by colonic epithelium in response to R. gnavus colonization led to activation of ILC2s and increased numbers of eosinophils and Th2 cells in the intestine, which in turn migrate and induce immunopathology in the lungs (133).

As we understand better the link between ILC2s and human asthma, studies have been initiated to elucidate the effects of asthma treatment on ILC2s and to modulate ILC2s as a potential treatment option for asthma. Conventional treatment for asthma, including subcutaneous immunotherapy (SCIT) and glucocorticoids, can decrease ILC2 numbers and suppressed their effector functions. For example, in mice sensitized to birch pollen, SCIT resulted in decreased ILC2 numbers and function in the airways (134). In patients with asthma, blood ILC2 frequency and IL-13 levels were reduced with SCIT and pharmacotherapy compared to pharmacotherapy alone (135). Significant decreases in circulating ILC2 numbers were observed in newly diagnosed patients with asthma after budesonide treatment (113). Budesonide also inhibited phosphorylation of STAT proteins and JAK3 and greatly reduced type 2 cytokine production from ILC2s in vitro (136).

Conversely, lack of ILC2 response to glucocorticoids may explain steroid resistance in asthma. Indeed, BAL fluid ILC2s from asthmatic subjects displayed steroid resistance, and they continued to produce type 2 cytokines even in the presence of dexamethasone (137). Importantly, TSLP, but not IL-25 or IL-33, promoted this ILC2 resistance to glucocorticoids (137, 138), and TSLP levels in BAL fluids correlated with dexamethasone resistance in BAL ILC2s (137). Furthermore, treatment with inhibitors that block the downstream signaling of TSLPR reversed steroid resistance in BAL fluid ILC2s (137). In addition, when mice lacking the IL-33 receptor ST2 were exposed to a cocktail of allergens, ILC2s that produce IL-9 and IL-13 mediated persistent mucus production and airway hyperreactivity (139); blocking TSLP reversed the immunologic and pathologic changes in this model. Taken together, ILC2s may contribute to steroid resistance and persistent airway pathology, and TSLP may play a key role.

Another potential treatment for asthma as related to ILC2s includes targeting the glucagon-like peptide 1 receptor (GLP-1R) on airway epithelial cells. Treatment with a GLP-1R agonist reduced ILC2 activation and subsequent eosinophilia and pathology in mice exposed to Alternaria (140). The activity of the GLP-1R agonist to inhibit IL-33 release from epithelial cells likely explains the mechanism of its efficacy (140). Cyclosporin A also suppressed ILC2-mediated airway inflammation induced by protease papain (141). Further studies using pharmacologic or biologic agents likely will provide useful information to dissect the roles of ILC2s in pathophysiology of asthma and, at the same time, to develop novel treatment strategies for asthma by targeting ILC2s.

Chronic rhinosinusitis

Chronic rhinosinusitis (CRS) is marked by chronic inflammation of nasal mucosa and paranasal sinuses lasting greater than 12 weeks. Because ILCs are resident in mucosa tissues, it is not surprising that human ILC2s were first reported in nasal polyp tissue from patients with CRS (103). Since then, a number of studies showed that ILC2s are increased and demonstrate an activation phenotype in nasal polyps in patients with CRS with nasal polyps (CRSwNP) (76, 103, 142). Inflamed ethmoid sinonasal tissue from patients with CRSwNP showed higher frequencies of ILC2s than comparable tissue from patients with CRS without nasal polyps (CRSsNP) (143–145). Eosinophilic nasal polyps or ethmoid tissues also showed increased ILC2 frequencies compared to non-eosinophilic nasal polyps (145–147). Among patients with CRS, ILC2s were more abundant in those with asthma compared to those without asthma, and ILC2s in sinus tissue correlated with nasal symptom scores (145). These findings suggest roles for ILC2s in eosinophilic inflammation and polypoid changes in nasal and sinus mucosa in CRS and their association with clinical symptoms. Importantly, the frequency of peripheral blood ILC2s was similar between patients with eosinophilic nasal polyps compared to non-eosinophilic nasal polyps (147) and did not differ between CRSwNP and healthy controls (142), suggesting compartmentalized regulation of ILC2s in respiratory mucosa.

An extensive study of all major ILC subsets in CRS tissues in the United States found that ILC2s are the predominant ILCs in nasal polyps and are 100-fold more frequent in nasal polyps as compared with the sinus mucosa of control subjects (142). A small but significant increase in ILC2s and ILC3s was also observed in ethmoid tissues from subjects with CRSsNP, likely reflecting heterogeneic nature of CRSsNP(148). ILC2 frequencies in nasal polyps were not altered by glucocorticoid treatment, asthma status or aspirin sensitivity (142) while another study showed that systemic corticosteroid treatment reduces the frequency of ILC2s in eosinophilic nasal polyps (146).

The heterogeneity of ILC2s has been recognized in CRS. For example, ILC2s in upper airways of subjects with CRSwNP were CD117⁺IL-1R1⁺ and their levels correlated with the levels of Th2 cytokines and tissue eosinophilia as well as allergic and asthmatic status (149). In contrast, ILC2s from those with CRSsNP showed a CD117⁻IL-1R1⁻ phenotype (149). Functional plasticity of nasal polyp ILC2s has also been reported. As expected, nasal polyp ILC2s produced IL-5 and IL-13 after stimulation with IL-33 (12, 143). IL-12 added to nasal polyp ILC2s induced a shift from IL-5-producing ILC2s to IFNγ-producing ILC1-like cells; this transformation was reversed by IL-4, suggesting that IL-4, likely produced by eosinophils in nasal polyps, might maintain the type 2 characteristics of ILC2s in vivo (76). IL-10-producing ILC2s were also increased in nasal polyps from patients with CRSwNP (18). Clearly, further studies are necessary to elucidate the heterogeneity of ILC2s and other ILC populations in CRS and to understand their roles in clinical phenotypes and endotypes of the disease.

While the mechanisms involved in proliferation and activation of ILC2s in CRS are not fully understood, mediators known to activate ILC2s (Figure 2), including TSLP, IL-4, IL-13 and a variety of lipid mediators, are elevated in nasal polyps compared with healthy sinonasal tissues (150). The levels of other ILC2 activators, such as IL-25 and IL-33, are rather controversial (151). Recently, receptor activator of NF-κB (RANK) ligand (RANK-L, TNFSF11) was found to be elevated in nasal polyps (151). ILC2s expressed RANK-L and RANK. Further, an agonistic antibody to RANK induces type 2 cytokines in human ILC2s in vitro (151), suggesting that the RANK/RANK-L pathway may provide another mechanism for ILC2 activation in CRS. Importantly, membrane-bound RANK-L is detected mainly on the surface of Th2 cells (151), suggesting potential physical interaction between ILC2 and Th2 cells in nasal polyps.

An association between increased production of cysLTs and tissue eosinophilia in nasal polyps also needs to be highlighted (152). Increased expression of cysLTR1 was observed in leukocytes in nasal biopsy specimens from patients with aspirin-sensitive CRS and aspirin exacerbated respiratory disease (AERD) (153). In subjects with AERD, ILC2s in nasal scrapings increased while blood ILC2s decreased after aspirin challenge (154). Furthermore, mouse lung ILC2s attached to platelets through the P-selectin ligand, and depletion of platelets resulted in rapid loss of ILC2s from lungs and reduced production of type 2 cytokines (155). Platelet-leukocyte aggregates are known to facilitate cysLT production in AERD (156), suggesting pathologic interaction between platelets and ILC2s during aspirin challenge. In contrast, mice lacking the ptges gene, which codes for PGE₂ synthase, that had been sensitized and challenged with house dust mite (HDM) extract developed an AERD-like disease upon challenge with aspirin (157), suggesting inhibitory roles for PGE₂ in AERD. Furthermore, cysLTs directly inhibited ILC2 proliferation and cytokine release through the CysLT1R while cysLTs promoted IL-33 release from alveolar type 2 cells that express CysLT2R (158). An in vitro study using human ILC2s showed that LTE₄-mediated chemotaxis and type 2 cytokine production was inhibited in the presence of the CysLT₁ antagonist montelukast (159); however, little else has been reported regarding the effects of leukotriene inhibitors on ILC2 number and function. Altogether, these findings suggest important roles for the interaction between a variety of lipid mediators and ILC2s in pathophysiology of AERD. Further studies in this area will likely provide valuable information to better understand the mechanisms of CRS.

Allergic Rhinitis

Allergic rhinitis (AR) is generally considered an IgE-mediated nasal, ocular and inflammation response to airborne allergens. Association between AR and ILC2s has been recognized in allergen challenge models and natural exposure to airborne allergens. For example, ILC2s increased in peripheral blood of subjects with cat allergy after allergen challenge (160). Numbers of blood ILC2s and ILC3s increased during the peak season in subjects with grass allergy (161, 162). Subjects with both allergy and asthma had more blood ILC2s than subjects with allergy alone; and the number of ILC2s correlated with disease severity (162). In addition, AR subjects who are sensitive to a perennial allergen, HDM, showed higher blood ILC2 frequencies than non-allergic controls (163, 164). In local tissues, ILC2s (Lin⁻GATA3⁺ by immunofluorescence staining) were detected in the nasal mucosa of HDM-sensitive AR patients (165). The number of circulating ILC2s in patients with AR also correlated with markers of disease severity, such as symptom scores (164). Altogether, similar to patients with asthma, exposure to allergens likely increases the number of circulating ILC2s in AR.

A study elucidated the mechanisms for activation of ILC2s in AR. ILC2s, myeloid dendritic cells (mDCs), and plasmacytoid dendric cells (pDCs) were detected in the nasal mucosa of patients with AR (165). When cultured together in vitro, mDCs activated ILC2s and promoted type 2 cytokine production. Importantly, the decoy receptor soluble ST2 (sST2) inhibited mDC-induced cytokine production by ILC2s and circulating mDCs expressed IL-33, suggesting that mDCs directly activate ILC2s via the IL-33/ST2 pathway (165). In contrast, co-culture with plasmacytoid DCs (pDCs) inhibited cytokine release from ILC2s in an IL-6-dependent manner (165).

Immunotherapy has been reported to affect the frequency or activity of ILC2s in AR. For example, SCIT inhibited seasonal increases in ILC2 frequency (161). In HDM-sensitive AR subjects who underwent sublingual allergen immunotherapy (SLIT), the numbers of ILC2s in peripheral blood correlated with the clinical efficacy of the immunotherapy. Indeed, ILC2s decreased in AR subjects who responded successfully to the treatment but not in non-responders (166). Further analysis of other ILC populations revealed that IL-10⁺CTLA4⁺ ILCs increased in responders but not in non-responders while the number of IL-4⁺CRTH2⁺ ILC2s were not significantly affected (166), suggesting that ILCs with a regulatory phenotype might be generated by immunotherapy. Circulating IL-10⁺ ILC2s increased in patients with grass pollen allergy after 12 months of grass pollen SLIT. No changes in IL-10⁺ ILC2s were seen in patients treated with placebo SLIT (167). Thus, the current evidence suggests that ILCs, including ILC2s and those with a regulatory phenotype, are modulated upon allergen exposure and immunotherapy in patients with AR. Further studies are necessary to investigate the immunological mechanisms that explain these dynamic changes in ILCs in response to allergens and to elucidate their roles in the pathophysiology and prevention of AR.

Atopic Dermatitis

Atopic dermatitis (AD) is a chronic inflammation of skin marked by severe pruritus and eczematous lesions and often mediated by impaired skin barrier functions. Studies have been performed both in mouse models and humans to investigate the roles of ILCs in AD. In mice, repeated topical administration of the vitamin D analog calcipotriol induced AD-like lesions characterized by eczematous dermatitis, thickening of the stratum corneum, epidermal hyperplasia, and dermal infiltration of inflammatory cells (26). ILC2s increased in the skin after calcipotriol treatment and depletion of ILC2s ameliorated the AD-like skin pathology (26). Importantly, in contrast to ILC2s in the intestine and lung, the ILC2 responses in the skin were independent of IL-33 or IL-25 but were dependent on TSLP. Similarly, activation of dermal ILC2s with an IL-2-anti-IL-2 immune complex induced AD-like lesions with an increased number of ILC2s even in T and B cell-deficient Rag1^−/− mice (168). Mice engineered to overexpress IL-33 in the skin developed AD-like lesions (25, 169), and IL-5-producing ILC2s played a pivotal role in development of skin lesions as examined by antibody-mediated depletion and transfer of bone marrow from ILC2-deficient Rora^sg/sg mice (169). Finally, filaggrin-deficient mice showed increases in IL-5-producing ILC2s and spontaneously developed dermatitis (170). Thus, several distinct mouse models suggest pathologic roles for skin ILC2s in promoting development of AD-like lesions even in the absence of adaptive immunity.

ILC2s are increased in the skin of patients with AD. In humans, lesional skin biopsies from patients with AD showed increased CD25⁺ST2⁺ ILC2s (26, 171). Increased numbers of CD127⁺CRTH2⁺ ILC2s were also identified in AD patients with FLG mutation as compared to those without the mutation (170). ILC2s from subjects with AD expressed higher levels of ST2, IL17RB and TSLPR, the receptors for IL-33, IL-25 and TSLP, respectively, as compared to skin ILC2s from healthy individuals (171). Furthermore, by immunohistochemistry, not only ILC2s but also a prominent AHR⁺ ILC3 population was detected in AD skin near the epidermis and in close proximity to T lymphocytes. In contrast, more prominent populations of ILC1s and RORC⁺ ILC3s were observed in skin biopsy specimens from subjects with psoriasis (172), reflecting different immunological mechanisms for AD and psoriasis.

ILC2s may contribute to the pathophysiology of AD through several mechanisms. As expected from conventional immunobiology of ILC2s, ILC2-derived IL-13 was responsible for inflammation and barrier dysfunction in lesional skin in a mouse model (173). In a model of poison ivy-induced allergic contact dermatitis (ACD), urushiol promoted production of IL-33 by skin keratinocytes, leading to activation of ST2-expressing dorsal root-ganglion neurons and subsequent scratching behavior (174). Given the ILC2’s response to neuropeptides, ILC2s may promote itch and skin inflammation related to poison ivy-induced ACD. In contrast, in 2,4,6-trinitrochlorobenzene-induced contact hypersensitivity, hapten challenge increased natural killer cells in the skin (175). Interestingly, in this model, ILC2-deficient Rora^sg/floxIl7r^Cre/+ mice showed an enhanced ear swelling response, suggesting that ILC2s can protect from type 1-dominant contact dermatitis.

Identification of factors that contribute to expansion and activation of skin ILC2s is an active area of research. As described above in the calcipotriol model, the ILC2 responses in the skin were critically dependent on TSLP (26). In contrast, IL-33, but not IL-25 or TSLP, induced both type 2 cytokine production and migration of skin ILC2s in vitro (171). A recent clinical trial further supports potential involvement of IL-33 in the pathophysiology of AD in humans (176). In contrast, after tape stripping and sensitization with OVA antigen, mouse keratinocytes produced IL-25, which promoted IL-13 production by ILC2s and induced skin inflammation and barrier dysfunction (173). Furthermore, PGD₂ robustly triggered transmigration of healthy skin ILC2s and promoted production of type 2 cytokines through the receptor CRTH2 (30). In mice, resting skin IL-5⁺ ILC2s express CD218 (IL-18R1), and IL-18 likely plays roles in allergic responses of skin ILC2s (99). Thus, the mechanisms involved in activation of skin ILC2s appear to be multiple and may reflect complex etiology and pathophysiology of AD involving genetic factors, skin microbiome and environment. In addition, potential interaction with other immune and inflammatory cells also needs to be recognized. For example, IL-4 production by basophils was required for ILC2 expansion and cytokine production in an IL-33-driven mouse model of AD (169). In another AD model, RORa-expressing skin Tregs dampened ILC2 activation and IL-5 production, resulting in diminished skin eosinophilia and AD-like pathology through TNF ligand–related molecule 1 (TL1A) and its receptor death receptor 3 (DR3) (177).

Food Allergy and Eosinophilic Esophagitis

As compared to other allergic diseases, fewer studies have been performed to investigate the roles for ILC2s in food allergy and eosinophilic esophagitis (EoE). A majority of information is derived from mouse models of these diseases. For example, in mice sensitized by intragastric administration of peanut extract and cholera toxin as an adjuvant, IL-33 played a major role in development of systemic anaphylaxis as well as expansion of intestinal ST2⁺ ILC2s (178). Subsequently, in mice with gain-of-function mutation of IL-4Rα, intragastric administration with peanut butter expanded ILC2s and ILC3s in mesenteric LNs and small intestine (28). In the model, IL-4 derived from gastrointestinal ILC2s and mast cells in response to epithelial IL-33 was necessary for IgE antibody production and development of systemic anaphylaxis. Interestingly, in a phase 2 clinical trial, anti-IL-33 antibody treatment desensitized patients with peanut allergy (179), suggesting potential involvement of IL-33 and downstream cell types, perhaps CD4⁺ Th2 cells and ILC2s, in food allergy.

Besides IL-4 and IL-33, ILC2-derived IL-13 may also be involved in food allergy. Mice that overexpress IL-25 in the intestine and were sensitized by intraperitoneal injection with OVA developed acute anaphylaxis when challenged repeatedly with gastric administration of OVA antigen (27). In this model, ILC2s isolated from the lamina propria produced type 2 cytokines in response to IL-25, and IL-13-deficient ILC2s lost their ability to promote IgE antibody production (27). Importantly, intestinal CD4⁺ Th2 cells supported IL-13 production by ILC2s, suggesting an interaction between ILC2s and Th2 cells. More recently, mice exposed to peanut flour particles by inhalation produced IgE antibody to peanut and developed systemic anaphylaxis (29). During the sensitization process, IL-13 produced by lung ILC2s induced activation and migrations of DCs, which promoted the development of T follicular helper cells that are necessary for IgE class switch in germinal center B cells (29). Thus, animal models describe several pathways that ILC2s are involved in development of food allergy. Nonetheless, a majority of these mouse models use genetic manipulation or adjuvants to promote sensitization to food antigens. Therefore, the roles for ILC2s described in these models need to be verified in humans.

EoE is a chronic hypersensitivity response to food antigens (10). A few studies suggest the potential roles for ILC2s in EoE. For example, ILC2s were increased in esophageal biopsies from children with active EoE compared to those with inactive EoE and healthy controls; the number of esophageal ILC2s correlated with tissue eosinophilia in biopsy specimens (180). Biopsy specimens from EoE patients contain increased levels of TSLP and IL-33 (181, 182), which may promote expansion and activation of ILC2s in esophageal mucosa. Indeed, oral antigen challenge of mice that had been sensitized percutaneously with MC903 and OVA resulted in EoE-like tissue pathology, which was dependent on type 2 cytokines, TSLP and basophils (181). Apparently, our knowledge regarding the roles of ILC2s in EoE is limited and further studies are necessary.

Summary and future directions

As summarized above, increasing evidence in the field suggests strong association between ILC2s and allergic disease. A number of conceptual advances have also been made regarding immunobiology of ILCs by identifying biological molecules beyond cytokines that can activate ILCs (e.g. neuropeptides) and by recognizing the plasticity and heterogeneity of ILCs. Furthermore, while ILC2s were initially identified as a cell type with innate effector functions, new information suggests that ILC2s are an important part of a network of type 2 immunity and that they may promote tissue pathology and maintain homeostasis in concert with other cell types. Nonetheless, three major areas of future research can be identified.

First, the heterogeneity of ILC2s causes a challenge in clinical studies to investigate the roles of ILC2s in allergic diseases. Human ILC2s have been conventionally identified by the lack a set of markers (i.e. lineage markers) and by expression of a combination of another set of markers (e.g. ST2, ICOS, CD127, CD25, CRTH2, intracellular IL-5, see Table 2). However, there has not yet been complete agreement on which markers exactly define an ILC2. A major question remains whether any sets of markers can identify ILC2s comprehensively given the heterogeneity of their surface molecules. Indeed, investigators have sought to parse out subsets of ILC2s with unique functions or expression of non-conventional markers. Production of type 2 cytokines, such as IL-5 and IL-13, and expression of the transcription factor GATA3 likely provides fundamental characteristics of ILC2s. Further analysis of ILC2 populations in peripheral blood and tissues specimens from patients with allergic diseases based on this fundamental principle will help to characterize heterogeneity of ILC2 populations more precisely and their association with disease phenotypes. Cutting-edge technologies, such as scRNAseq and high dimensional mass cytometry and flow cytometry, will likely provide a useful tool to accomplish this task.

Second, the mechanisms to explain how ILC2s are involved allergic diseases need to be defined further. As summarized in this review, ILC2s are increased and activated in various allergic diseases in humans. The cause-effect relationship between ILC2s and diseases and the immunologic mechanisms to explain how ILC2s are involved in the pathophysiology of allergic diseases have been elucidated by a variety of animal models. However, many of these new observations need to be verified in humans, which is particularly important as there are some species differences in ILCs between humans and mice. For example, whereas ILC2 comprise a substantial proportion of ILCs in mouse intestine and human fetal intestine, ILC2s are largely absent throughout the healthy human GI tract (103, 183). As ILC2s are a potent innate effector cell as well as a player in the immune network, it will be highly important in the future to elucidate how ILC2s are activated and regulated by tissue cells and other immune cells. Candidates are many, including several subsets of epithelial cells, T cells, DCs, eosinophils, mast cells, basophils and potentially others. Reciprocal interaction between ILC2s and these cell types also needs to be investigated. Clinical trials with biologics that are targeted to the molecules involved critically in immunobiology of ILC2s, such as IL-33, will likely provide important insight.

Finally, ILC2-driven type 2 responses are now recognized in diverse immune processes, different anatomical locations, and homeostatic or pathological settings. ILC2s are likely governed by their microenvironment, tissues and organs. Therefore, it would be highly important to elucidate how ILC plasticity fits into pathophysiology of allergic diseases and their endotypes and phenotypes. Our understanding of how plasticity of ILC2s is initiated, maintained and reversed is still very much in its infancy. Low prevalence of ILCs makes investigating the cellular and molecular mechanisms involved in regulation of ILCs, such as epigenetics and post-transcriptional modification, technically challenging, but that will be overcome by new technologies. With all of the knowledge we’ve gained regarding ILC2s in the past ten years, we now recognize that the ILC2 forms an important nexus of the immune system and may present an attractive target for immune modulation in diseases. How can we leverage what we know about ILC2s and other ILCs to develop treatments that can prevent, alleviate, or potentially even cure allergic diseases? That has always been, and remains, the goal of allergic disease research.

What do we know?

• ILCs are innate lymphoid cells that lack rearranged receptors and are present in small numbers, but quickly produce large quantities of cytokines in response to cytokines, eicosanoids, neuropeptides and other signals from sentinel immune and stromal cells.

• ILC2s are strongly associated with various allergic diseases, such as asthma, AR, CRS, and AD; we are beginning to recognize roles played by other ILCs as well.

• ILC2s are heterogeneous and plastic. Their heterogeneity and plasticity provide insight into how they promote tissue pathology and homeostasis but also complicate our ability to define what they are and how all their subsets intersect.

• ILC2’s roles in allergic diseases extend beyond their previously-recognized innate effector functions. ILC2s are actively involved in promoting adaptive immune responses and interact reciprocally with stromal and immune cells in a network of type 2 immunity.

What is still unknown?

• What roles do ILCs, particularly ILC2s, play in food allergy and EoE?

• How integral are ILCs other than conventional ILC2s, including ILC1s, ILC3s and regulatory ILC2s, both in promoting and protecting from allergic diseases? How does ILC2 plasticity serve to ameliorate or exacerbate disease?

• In light of some variability in how ILC2s are identified in previous studies, particularly in humans, are there any phenotypes that can be used universally to define ILC2s? How does the heterogeneity in ILC2 populations affect disease endotypes and phenotypes?

• How do new findings in mouse models translate into human diseases?

• As our understanding of how various mediators and cell types interact with ILC2s increases, how can we leverage our knowledge to provide more effective, affordable and safe treatments of allergic diseases?

Abbreviations

ACD

allergic contact dermatitis

AD

atopic dermatitis

AERD

aspirin exacerbated respiratory disease

AHR

aryl hydrocarbon receptor

AR

Allergic rhinitis

β₂AR

β₂-adrenergic receptor

BrCs

brush cells

CBFβ

core binding factor β

CGRP

calcitonin gene-related peptide

CHILP

common helper innate lymphoid progenitors

CILP

common innate lymphoid progenitors

CLP

common lymphoid progenitors

COX-1

cyclooxygenase-1

CRS

chronic rhinosinusitis

cysLT

cysteinyl leukotriene

DCs

dendritic cells

EoE

eosinophilic esophagitis

EpCs

chemosensory epithelial cells

GATA3

GATA-binding protein 3

GI

gastrointestinal

GLP-1R

glucagon-like peptide 1 receptor

HDAC

histone deacetylase

HDM

house dust mite

iILC2s

inflammatory ILC2s

ILC2s

group 2 innate lymphoid cells

ILCs

innate lymphoid cells

ILCregs

regulatory ILCs

ILCPs

ILC progenitors

IRF7

interferon regulatory factor 7

LNs

lymph nodes

LTi

lymphoid tissue inducer cells

LXA₄

lipoxin A₄

mDCs

myeloid dendritic cells

MDSC

myeloid-derived suppressor cells

NKP

NK cell precursors

NMU

neuromedin U

OVA

ovalbumin

PBMCs

peripheral blood mononuclear cells

pDCs

plasmacytoid dendric cells

PGI₂

prostaglandin I₂

PMN

polymorphonuclear

PNECs

pulmonary neuroendocrine cells

RA

retinoic acid

RAR

retinoic acid receptor

RORγt

retinoic acid-related orphan receptor γt

ROS

reactive oxygen species

RSV

respiratory syncytial virus

RV

rhinovirus

S1P

sphingosine 1-phosphate

SCCs

solitary chemosensory cells

SCFA

short-chain fatty acids

SCIT

subcutaneous immunotherapy

scRNAseq

single cell RNA sequencing

SLIT

sublingual allergen immunotherapy

Treg

T regulatory