Abstract

Introduction

Type 2 inflammation, including tissue eosinophilia, epithelial mucus metaplasia, and IgE production, is characteristic of allergic diseases, such as asthma [1]. Classically, conventional CD4⁺ Th2 cells that secrete IL-4, IL-5, and IL-13 have been considered the primary orchestrators of the allergic response in tissues. Although the study of conventional Th2 cells has provided significant insight into the pathogenesis of asthma and allergy, many phenotypes of human disease are not adequately explained by the Th2 cell paradigm. For example, viruses are thought to be a primary trigger of asthma exacerbations, and evidence is accumulating that viruses may play a role in the development of asthma [2]. Furthermore, environmental insults, such as tobacco smoke and ozone exposure, stimulate innate-immune pathways and may contribute to asthma pathogenesis [3, 4]. This suggests that additional innate-immune pathways are active, and recent studies support ILC2s as important contributors in models of type 2 diseases. Here, we will review cellular and molecular mechanisms of ILC2s in several animal models and stress their potential contributions to diseases of the respiratory tract and skin. The accompanying review by Peebles [5] will focus on the roles of ILC2s in clinical disease and highlight the need for strategies to target ILC2s in humans.

DISCOVERY OF ILC2s: A NEW ERA IN TYPE 2 IMMUNITY

The presence of a non-B, non-T cell population that produces Th2 cytokines has been suggested for over 1 decade. Beginning in 2001, investigations of IL-25-induced responses in mice led to the discovery of novel non-B, non-T cells that produced Th2 cytokines [6–8]. Three pivotal studies reported in 2010 confirmed the presence of innate lymphocytes that robustly produce Th2 cytokines and are now termed ILC2s by consensus [9–12]. The first report showed that the IL-1 family cytokine IL-33 induced high levels of IL-5 and IL-13 from lineage-negative (lack of expression for CD3, CD4, CD8, TCR-β, TCR-δ, CD5, CD19, B220, NK1.1, Ter-119, GR-1, Mac-1, CD11c, and FcεR1) and c-kit⁺ Sca-1⁺ lymphocytes, present in the mesenteric fat of mice [11]. The cells were termed NHCs, and the authors showed that IL-33 stimulation of 5000 purified NHCs induced microgram levels of IL-5 and IL-13, demonstrating a robust Th2 cytokine production capacity. Furthermore, adoptive transfer of NHCs into RAG2/γc knockout mice that lack NHCs resulted in intestinal goblet cell hyperplasia after Nippostrongylus brasiliensis infection. A subsequent report showed that IL-25 and IL-33 given to IL-13 reporter mice led to the accumulation of an IL-13⁺ non-T cell population in the MLNs and small intestines [9]. The cells were named “nuocytes,” and further work showed that nuocyte expansion after helminth infection required IL-25 and IL-33. Importantly, nuocyte IL-13 production was also required for helminth expulsion. The 3rd report described a similar population of lineage-negative IL-13⁺ cells, termed Ih2s, which accumulated in the MLNs, spleens, livers, and lungs of mice after N. brasiliensis infection. Ih2s were also sufficient to induce worm expulsion after stimulation with IL-25 in the absence of adaptive immunity [10]. Collectively, these landmark studies demonstrated that NHCs, nuocytes, and Ih2s (now all collectively referred to as ILC2s) were novel Th2 cytokine-producing cells critical for early helminth immunity (see Table 1).

ILC2 DEVELOPMENT

Similar to other ILCs, ILC2s originate from CLPs that require Id2, Notch, and IL-7R that signals through the common γc for development [11, 13–15] (see Fig. 1). A complete absence of mesenteric fat-associated c-kit⁺ Sca-1+ NHCs was shown initially in Id2 knockout mice and is consistent with the requirement of Id2 for innate lymphoid development [11]. A follow-up report demonstrated that bone marrow NHCs are marked by a history of RAG expression and developed from common lymphoid, but not myeloid or erythroid, progenitors after studies were performed of progenitors transplanted into irradiated congenic mice [13]. IL-7 or IL-7R is required for development of ILC2s, and IL-7R^−/− mice lack innate type 2 lung inflammation after exposure to the fungal allergen Alternaria alternata compared with a robust response induced in RAG^−/− mice that have ILC2s but lack T or B cells [11, 16, 17]. Although several ILC2 studies have used IL-7R^−/− mice, interpretations of ILC2 function in these mice can be challenging, given the potential role of IL-7R in eosinophil, T cell, and macrophage function that could contribute to type 2 inflammation [18, 19].

Open in a separate window

Figure 1.

ILC2 development. ILC2s derive from a CLP through expression of Id2, Notch, IL-7R, GATA-3, TCF-1, PLZF, and possibly, Ets-1 Pre-ILC2s require GATA-3 and RORα for maturation into ILC2s. GATA-3 and possibly Ets-1 are required for activated ILC2 function.

The master Th2 cytokine transcription factor GATA3 has been shown to play a prominent role in ILC2 development and function and represents shared transcriptional programming of ILC2s and Th2 cells [20–22]. Aside from a clear role in the development of ILC2s, GATA3 is also critical to the development of ILC3s and all IL-7R-expressing ILCs [23, 24]. This suggests that GATA3 has a broad, early role in ILC development, apart from a later role in ILC2 cytokine production. In addition to GATA-3 expression, our group has reported that human peripheral blood ILC2s and mouse lung and bone marrow ILC2s highly express the transcription factor Ets-1 [17], which has been shown to be critical for NK and NKT cell development (25, 26). Further, Ets-1^−/− T cells cultured in Th2-polarizing conditions are impaired in Th2 cytokine production, suggesting that Ets-1 may be required for ILC2 development and/or Th2 cytokine production similar to GATA-3 [27].

Since the discovery of ILC2s, investigations have identified novel transcription factors that regulate ILC2 development. The nuclear receptor RORα was shown to be required for ILC2 development from CLPs into nuocytes in vitro, and IL-25 administration to RORα knockout mice resulted in severely impaired nuocyte expansion, as well as reduced goblet cell hyperplasia compared with WT mice [28]. A subsequent report confirmed that RORα contributed to NHC development and importantly, showed that responses from ILCs distinct from ILC2s, as well as conventional Th2 cells, were not impaired [29]. In contrast to GATA-3 regulation of ILC2s and Th2 cells, RORα may specifically regulate ILC2s without a significant role in CD4⁺ Th2 cell differentiation or cytokine production. Interestingly, GATA3^−/− CLPs do not express RORα or Id2 mRNAs, suggesting that ILC2s undergo complex, sequential regulation by several molecules in their development (Fig. 1) [21]. From a translational aspect, RORα has been identified previously as an asthma-susceptibility gene in genome-wide association studies and possibly links ILC2s to human asthma [30].

In addition to RORα, Notch signaling member TCF-1 has been shown to be required for the development of functional ILC2s, as mice that lack TCF-1 displayed impaired ILC2 cytokine responses [31]. Additionally, the transcription factor PLZF has been detected at high levels in all ILC precursors, although bone marrow chimera studies revealed that PLZF appears to be required more specifically for development of ILC2s, in addition to a known role for NK T cell development [32]. Overall, the development of ILC2s from lymphocyte progenitors depends on GATA3-induced Id2 and RORα, in addition to Notch, TCF-1, PLZF, and IL-7R signaling.

The varied surface-marker expression and tissue distribution initially reported with mouse Ih2s, nuocytes, and NHCs have led to some confusion regarding whether these populations are indeed distinct from each other. The common surface-marker expression shared by these populations includes CD45, Thy1.2, CD44, CD69, and c-kit, but differences exist in surface expression of Sca-1, T1/ST2 (IL-33R), and IL-7Rα (Table 1) [9–11]. However, the commonality that Ih2s, nuocytes, and NHCs are lineage-negative lymphocytes that produce Th2 cytokines has led to a consensus to include these populations as ILC2s [12]. Differences detected among these cell populations may relate to the level of activation, kinetics, and tissue distribution, as well as methods of detection or stimulation. Human ILC2s have also been detected in multiple tissues, including the skin, respiratory, and gastrointestinal tract, and are defined as lineage-negative lymphocytes that express the PGD2R (or DP2; also known as CRTH2) and CD161 [33]. Ongoing work will, no doubt, identify novel ILC2 surface markers, and the hope is that a more ILC2-specific marker is discovered to allow for improved detection and possibly targeting.

PRODUCTS OF ILC2s IN HEALTH AND DISEASE

Studies with animal models have shown that the products of ILC2s contribute to normal tissue homeostasis, pathologic changes in disease, and repair responses after mucosal insult. The initial studies of ILC2s demonstrated an early role in helminth immunity, as ILC2s were shown to produce high levels of IL-5 and IL-13, leading to tissue eosinophilia, mucus production, and worm expulsion [9–11]. Interestingly, the quintessential Th2 cytokine IL-4 was not detected after ILC2 stimulation with IL-33 in these early reports [9, 11]. Subsequent studies have shown that TSLP and LTD4 can induce IL-4 production by ILC2s, suggesting that under specific conditions, ILC2s are a source of IL-4 [20, 34]. Notably, IL-33 stimulation of conventional T cells is reported to induce a Th2 cell population that produces IL-5, but not IL-4, and thus, lack of IL-4 production by ILC2s in the early reports may be related to a specific effect of IL-33R signaling [35].

The first study to demonstrate a role for lung ILC2s in a mouse model of disease showed that ILC2 IL-13 expression was required for AHR in mice challenged with influenza virus [36]. A subsequent report showed that adoptively transferred WT ILC2s were sufficient to restore AHR in IL-13 knockout mice after IL-25 exposure, thus further highlighting the importance of ILC2s as a source of IL-13 [37]. Lung ILC2s have since been shown to produce IL-5 constitutively, which controls eosinophil homeostasis, although IL-13 production requires stimulation with IL-25 or IL-33 [38].

Additional studies have demonstrated that ILC2s also produce IL-6 and IL-9 [11, 39]. IL-9 is also produced by Th2 or Th9 cells and contributes to airway eosinophilia, mast cell accumulation, and mucus production in animal asthma models [40]. The first report, showing that ILC2s produce IL-9, demonstrated that mouse lung ILC2s were a dominant source of IL-9 in WT mice challenged with the protease papain [39]. ILC2 IL-9 production was transient after papain challenge, peaking at 12 h and minimally present by 24 h, and promoted production of ILC2 IL-5, IL-6, and IL-13. Interestingly, ILC2 IL-9 production was dependent on IL-2 and intact adaptive immunity, thus suggesting adaptive immune cells can control ILC2 function.

Early studies provided important insight into the contribution of ILC2s to pathologic changes in type 2 immune disease, including tissue eosinophilia, mucus production, and AHR. However, the finding that ILC2s also promote tissue-repair responses through production of the EGFR ligand amphiregulin has changed our perspective of these novel cells [41]. Monticelli et al. [41] showed that depletion of ILC2s in RAG knockout mice infected with influenza virus led to hypoxemia, hypothermia, and loss of epithelial barrier integrity. ILC2s were found to produce high levels of amphiregulin and adoptive transfer of ILC2s, or administration of amphiregulin, into ILC2-depleted mice restored airway integrity after influenza infection. Our group has also found that lung ILC2s rapidly produce amphiregulin after innate challenge with the fungal allergen Alternaria [17]. Whether ILC2 amphiregulin can also contribute to pathologic remodeling, as is found in asthma, is unknown. Overall, the context of disease may dictate whether ILC2s contribute to pathogenic responses or play a role in normal repair processes.

Recently, the role of ILC2s in conditions of homeostasis has been investigated. With the use of IL-5 reporter and deleter mice, a study showed that ILC2s are the dominant source of IL-5 in many tissues at rest, including brain, heart, lung, kidney, skin, intestine, and uterus [38]. In the naïve mouse lung, IL-5 production by ILC2s was constitutive, but IL-13 required further stimulation. In contrast, ILC2s in the small intestine constitutively produced IL-13, which was increased further after food intake. Interestingly, vasoactive peptide, a hormone activated by eating, induced further ILC2 IL-5 production, suggesting that homeostatic control of ILC2s may occur through neuropeptide stimulation. Another report by the same group demonstrated that ILC2s were the major source of IL-5 and IL-13 in mouse visceral adipose tissue and contributed to eosinophil and alternatively activated macrophage accumulation [42]. The study also showed that IL-5 and eosinophil-deficient mice displayed impaired glucose control after a high-fat diet. A similar investigation found that depletion of ILCs in RAG^–/– animals fed a high-fat diet led to increased obesity and impaired glucose tolerance [43]. Adoptive transfer of ILC2s or NKT cells into obese mice induced weight loss and improved glucose control. Together, these reports suggest that ILC2s are critical in maintaining tissue eosinophil homeostasis, as well as visceral fat abundance and glucose control. Overall, the functions of ILC2s under homeostatic conditions and in disease should be greatly explored before targeting these cells for therapy in human type 2 inflammatory diseases.

REGULATION OF ILC2 FUNCTION

The list of mediators that can modulate ILC2 function is continually expanding and is summarized in Fig. 2. The cytokines IL-33 and IL-25 were first described to induce ILC2 cytokine production and proliferation [6–11]. IL-33 exists in a chromatin-bound, biologically active proform, present in epithelial cells, endothelial cells, and macrophages, and can be processed into a more potent cleaved protein by neutrophil elastase and cathepsin G [44, 45]. IL-33 binds to a heterodimer of ST2 and the IL-1R accessory protein (T1/ST2). IL-25 (IL-17E) is a member of the IL-17 family and binds specifically to the heterodimer of IL-17RB and IL-17RA. IL-25 is secreted by many cell types, including Th2 cells, eosinophils, as well as epithelial cells [46]. Ligation of T1/ST2 by IL-33 or IL-25R by IL-25 results in activation of MAPK and NF-κB signaling [36, 46, 47]. IL-25 and IL-33 are induced in type 2 inflammatory responses in mice and are thus available for potent ILC2 Th2 cytokine production [17, 48, 51]. Notably, a report has shown that IL-33 is more potent than IL-25 in the induction of ILC2 IL-13 production responsible for AHR in mice [52]. Shortly after the characterization of mouse ILC2 responses, human ILC2s from fetal gut and peripheral blood were shown to produce IL-13 in response to IL-25 and IL-33 [33]. Thus, IL-33 and IL-25 induce activation of human and mouse ILC2s, and the availability of these cytokines from different cell types may dictate the dominant pathway of ILC2 activation. Aside from ILC2 activation, IL-25 and IL-33 activate many cells, including DCs, eosinophils, macrophages, basophils, T cells, and NKT cells that could contribute amplification of the inflammatory response and indirectly activate ILC2s [47, 53].

Open in a separate window

Figure 2.

Tissue ILC2 responses. Epithelial cytokines TSLP, IL-33, and IL-25 can activate ILC2s to produce Th2 cytokines, including IL-4, IL-5, IL-9, and IL-13. PGD2 and LTD4, produced by mast cells, eosinophils (Eos), and macrophages (Mac), can also activate ILC2s, whereas LXA4 may have inhibitory effects on ILC2 function. TNF member TL1A, produced by DCs and other cells, might further promote ILC2 activation through binding of DR3. ILC2s may also respond to basophil IL-4 production to promote tissue eosinophilia (not shown). ILC2 and Th2 cells work in parallel to promote tissue type 2 responses. IL-5 production stimulates recruitment, activation, and survival of eosinophils. IL-13 induces AHR, DC migration, and along with IL-9, promotes mucus production. ILC2 IL-4 secretion may contribute to Th2 cell differentiation and IgE production by B cells. Amphiregulin (AREG) secreted by ILC2s binds to EGFR in the airway epithelium to induce repair responses and possibly pathologic remodeling. MC, Mast cell.

Recent work has identified mediators, apart from IL-25 and IL-33, that regulate ILC2 responses including the epithelial cytokine TSLP, which has been detected at increased levels in the airway epithelium of asthmatics and in the skin of atopic dermatitis patients and was initially was found to induce expression of the OX40L on DCs leading to Th2 cell differentiation [54, 55]. More recently, a role for TSLP in ILC2 responses was demonstrated as human peripheral blood and nasal polyp ILC2s cultured with TSLP and IL-33 displayed enhanced Th2 cytokine production (including IL-4) above IL-33 alone [20]. TSLP has also been shown to activate mouse ILC2s and independent of IL-33 in the skin [56, 57].

GATA3 is considered a master Th2 cytokine transcription factor that controls ILC2 development, as well as function [17, 20–22]. Interestingly, TSLP stimulation of human ILC2s was shown to increase the level of GATA3 expression [20]. Forced GATA3 expression in DP2-negative (CRHT2-negative) human ILCs also up-regulated DP2 (CRTH2), T1/ST2, and TSLPR levels and led to Th2 cytokine production compared with control vector-transduced ILCs. Furthermore, knockdown of GATA3 in human ILC2s led to reduced IL-13 levels. Overall, studies with human and mouse ILC2s have convincingly shown that GATA3 regulates much of the ILC2 differentiation and activation program. Although ILC2s and Th2 cells share high GATA3 expression, ILC2s express GATA3 in the bone marrow, suggesting that they are primed for Th2 cytokine production, in contrast to conventional Th2 cells that differentiate in the periphery [17, 22].

Cytokines, including IL-25, IL-33, and TSLP, were the first mediators discovered to activate ILC2s, but more recent reports have demonstrated that lipid mediators present in type 2 inflammatory diseases also modulate ILC2 activity. Eicosanoids, including PGs and LTs, are arachidonic acid-derived lipid mediators produced by many cells, including eosinophils, mast cells, DCs, and macrophages that have been shown to modulate ILC2 function [34, 58, 59]. PGD2 binds to DP2 (CRTH2), expressed on human ILC2s, as well as Th2 cells and eosinophils, and promotes chemotaxis, cytokine production, and survival [58, 60–64]. In human peripheral blood ILC2 cultures, PGD2 potentiated IL-13 production above IL-2, IL-25, and IL-33 stimulation that was also mimicked by a DP2 (CRTH2) agonist [58]. The same study also demonstrated that human ILC2s express LXA4R/N-formyl peptide receptor 2, the receptor for LXA4, and LXA4 inhibited ILC2 IL-13 production in the presence of IL-2, IL-25, IL-33, and PGD2. In addition to a role in cytokine production, DP2 (CRTH2) ligation by PGD2 functions in recruitment of many inflammatory cells into tissues, including basophils, eosinophils, and Th2 cells [60]. Our group [65] and others [61] have recently shown that PGD2 induces chemotaxis of human ILC2s via DP2 (CRTH2). PGD2 has also been tied to severe asthma, and PGD2 binding to DP2 (CRTH2) on ILC2s may have multiple functional consequences, including chemotaxis and cytokine production in asthma [62].

CysLTs are another group of eicosanoids that are elevated in asthma and CRS [59]. LTD4 binds with high affinity to the CysLT1R that is highly expressed on the surface of mouse lung ILC2s, as we have reported [34]. In vitro, LTD4 potently induced ILC2 calcium influx and cytokine production, including release of IL-4. Furthermore, intranasal LTD4 potentiated allergen-induced ILC2 proliferation and lung eosinophilia in RAG2 knockout mice (that have ILC2s but lack T and B cells). As eicosanoid molecules, including PGs, LTs, and lipoxins are largely produced by nonepithelial cells, including macrophages, mast cells, eosinophils, and DCs, these reports suggest novel ways in which ILC2s can be modulated (Fig. 2) [59].

The TNF/TNFR family members control a wide array of innate and adaptive immune responses [66, 67]. The TNF member TL1A is produced by activated DCs and T cells and binds to DR3, resulting in signaling through TNFR-associated death domain protein and intracytoplasmic death domain [68, 69]. A recent report demonstrated that DR3 is expressed on human and mouse ILC2s, and TL1A directly induces ILC2 cytokine production in vitro and promotes ILC2 cytokine production and expansion in vivo [70]. Importantly, DR3^−/− mice were impaired in clearance of N. brasiliensis infection and had reduced lung inflammation and numbers of ILC2s compared with controls. A role for TL1A/DR3 was shown previously to be critical for the development of allergic lung inflammation in an OVA model of asthma, and DR3⁺ ILC2s may have contributed to these findings [68]. Overall, the expanding numbers of mediators, including cytokines and lipid mediators that modulate ILC2 function, support the idea that the inflammatory milieu created by varying neighboring cell types likely dictates ILC2 activity. Thus, the potential for redundancy and context specificity in ILC2 activation will require further exploration if therapy to target upstream mediators of ILC2 activation in humans is undertaken.

ILC2 CROSS-TALK WITH OTHER IMMUNE CELLS

The relationship between ILC2s and other immune cells is summarized in Fig. 3. ILC2s have been shown recently to participate in cross-talk with DCs, CD4⁺ T cells, B cells, mast cells, and basophils. Early studies revealed that γc cytokines IL-2 and IL-7 support ILC2 survival and cytokine production in vitro [11, 56]. As conventional T cells are a source of IL-2, this suggested that adaptive immune cells could be involved in the maintenance of ILC2 responses. A recent study investigated such interactions between mouse ILC2s and naïve CD4 cells and found that anti-CD3/CD28-stimulated naïve CD4⁺ cells induced ILC2 proliferation and IL-4 and IL-5 production that was dependent on IL-2 [71]. Furthermore, ILC2s induced contact-dependent CD4⁺ T cell-Th2 cytokine production, and ILC2s could act as MHCII-dependent APCs in vitro to stimulate T cell responses. In an attempt to translate these findings in vivo, the authors adoptively transferred OVA-specific CD4⁺ cells and ILC2s into ST2^−/− mice challenged with IL-33 plus OVA. Only cotransfer of ILC2s with CD4 cells resulted in MLN OVA-specific IL-13 production. A subsequent report showed that ILC2s promoted CD4 T cell responses in vitro that were partially dependent on OX40/OX40L interactions and IL-4 [72]. Although these reports suggest that ILC2s induce CD4 cell proliferation and Th2 cytokine production in a contact-dependent manner in vitro, whether ILC2 and CD4 cells require direct contact in vivo for promoting Th2 responses is unclear, especially given the low numbers of ILC2s in tissues and lymph nodes. Additionally, these studies have used naïve and primary activated T cells and not memory T cells that likely propagate chronic type 2 inflammatory diseases and trigger exacerbations.

Open in a separate window

Figure 3.

ILC2 cross-talk. (A) ILC2s and CD4⁺ T cells may functionally interact directly through MHC2/TCR signaling and possibly OX40/OX40L. IL-2, secreted by T cells, supports ILC2 survival and proliferation, and IL-4, secreted by ILC2s or Th2 cells, may lead to mutual stimulation. (B) Mast cell release of lipid mediators, including CysLTs and PGD2, can lead to chemotaxis (PGD2) and activation of ILC2s. In turn, activated ILC2s produce IL-9 that leads to mast cell accumulation in tissues. (C) Basophils are an innate source of IL-4 required for ILC2 activation after airway exposure to protease. (D) ILC2s produce IL-4 and IL-6 that can induce B cell production of IgE and IgA, respectively. ILC2s express ICOS, and B cells express the ICOS ligand (ICOSL), although an interaction between these cells via ICOS/ICOS ligand is unknown.

The production of IL-4 from ILC2s could provide an important signal leading to Th2 cell polarization in vivo [20, 34, 71, 72]. However, a recent study showed that ILC2 IL-13, but not IL-4, was required for papain-induced Th2 cell responses in RORα^−/− bone marrow transplant mice that lack ILC2s but have normal Th2 cell responses [73]. In another report, the same group showed that RORα^−/−bone marrow transplant mice were impaired in generating Th2 cell responses in a house dust-mite model, but not an OVA/alum model, suggesting that the type of antigen and/or route of sensitization (mucosal vs. systemic) may dictate whether ILC2s contribute to Th2 differentiation [74]. Although the precise role of ILC2s in initiating adaptive Th2 responses may depend on the context or model, the idea that ILC2s could be critical for development of adaptive type 2 inflammatory conditions has broad implications.

Basophils have also been implicated in promoting Th2 responses, including those induced by papain [75, 76]. Recently, a potential link between basophil IL-4 production and ILC2 activation in papain-induced lung inflammation was reported [77]. The authors generated IL-4 enhancer mutant mice (IL-4 3′UTR^−/− mice) that selectively lacked basophil IL-4 production, whereas mast cells and CD4⁺ T cells were capable of IL-4 production. IL-4 3′UTR^−/− mice were impaired in developing eosinophilic lung inflammation, mucus production, and AHR after 3 papain challenges compared with WT mice. In addition, lung ILC2 activation and numbers were reduced in the IL-4 3′UTR^−/− mice. Lung ILC2s express IL-4Rα, as we have previously reported [17], and this study showed that IL-4 stimulation of ILC2s resulted in production of CCL11, CCL3, CCL5, and IL-9, as well as potentiation of IL-33-induced IL-5 and IL-13 secretion. Importantly, adoptive transfer of IL-4^−/− basophils in basophil-depleted mice led to reduced numbers of ILC2s, eosinophilic lung inflammation, and mucus production compared with mice receiving WT basophils. Thus, basophil/ILC2 cross-talk may initiate early type 2 responses to protease allergens. Whether these findings might translate to ongoing allergic inflammation is not yet clear.

In addition to basophil/ILC2 cross-talk, mast cells may communicate with ILC2s. Mast cells and ILC2s were found to be colocalized in the dermis of mice, and interactions between these cell types were visualized for up to 30 min [78]. Furthermore, mast cell release of IL-6 and TNF-α was reduced by IL-13, suggesting that ILC2 IL-13 production may have a suppressive role in mast cell function. Mast cells and ILC2s have also been detected in close proximity in human lung [58]. One possible reason for being such close neighbors could be a result of mast cell production of PGD2 that is chemotactic for ILC2s, as our group [65] and others [61] have shown. In turn, ILC2 IL-9 production could induce mast cell accumulation in tissues [78].

Finally, B cells and ILC2s may have meaningful interactions through direct contact, as well as cytokine production. ILC2s express the costimulatory molecule ICOS and could be stimulated by ICOS ligand-expressing B cells, but any functional significance of this possible interaction has yet to be shown [9]. The initial description of ILC2s in mesenteric fat demonstrated that production of IL-5 and IL-6 led to IgA antibody secretion by B1 B cells [11]. Furthermore, an ILC2-like population that is Sca-1⁺ Thy1⁺ IL-18R⁺ but does not express IL-7R or T1/ST2 was shown to promote B cell IgE production in vitro in the presence of ILC2 stimulated with IL-18 [79].

In addition to interactions with several immune cell types, ILC2s may have an important influence on epithelium, smooth muscle cells, and fibroblasts, as is suggested by the report that ILC2 amphiregulin production promotes lung epithelial tissue repair [41]. Overall, ILC2s have been shown to interact with many immune cell types known to contribute to type 2 diseases, and the uncovering of the functional significance of such interactions will likely improve our understanding of the role of tissue ILC2s.

ILC2s IN EXPERIMENTAL ASTHMA AND LUNG DISEASE MODELS

Human asthma is characterized by peribronchial inflammation, AHR, epithelial mucus production, and airway remodeling with principle triggers that include respiratory viruses and allergens. Mouse model protocols use airway challenges with viruses or allergens to induce airway inflammation and have been able to replicate features of human asthma to varying extents. Lung ILC2s from mice were first shown to contribute to AHR after influenza virus exposure [36]. Adoptive transfer of WT ILC2, but not IL-13 knockout ILC2, restored AHR in IL-13 knockout mice, suggesting that ILC2 IL-13 was critical for influenza-induced AHR. Since this initial report, several groups have shown that ILC2s contribute to type 2 lung inflammatory responses and AHR in mice exposed to multiple allergens, including Alternaria, papain, house dust mite, and OVA, and findings are summarized in Table 2 [16, 17, 36, 37, 56, 80].

TABLE 2.

ILC2s in selected lung and skin disease models

Allergen/virus/compound	ILC2 involvement in model	Ref.
Lung inflammation models
Influenza	AHR dependent on IL-33 and ILC2 IL-13 (RAG2−/− mice) Epithelial tissue repair dependent on ILC2 amphiregulin (RAG1−/− mice) ILC2 depletion by CD90.1 antibody in RAG−/− mice	[36, 41]
Fungal allergen (Alternaria)	Initiation of airway eosinophilia (RAG2−/− mice) dependent on IL-33 ILC2 transfer to IL-7R−/− sufficient for early eosinophilic airway inflammation	[16, 17]
OVA	Th2 cytokine producing ILC2s comparable with number of CD4+ Th2 cells ILC IL-13-restored AHR restored in IL-13−/− mice Partial ILC2 steroid resistance in OVA + IL-33 model	[37, 80, 81]
House dust mite	ILC2 promotion of Th2 cell priming in RORα−/− BMT mice	[74, 80]
Papain (cysteine protease)	Initiation of lung eosinophilia (RAG1−/− mice) ILC2 IL-9 production ILC2 promotion of Th2 cell priming in RORα−/− BMT mice ILC2 activation by basophil IL-4	[39, 56, 77]

Open in a separate window

Dermatitis models
MC903 (vitamin D analog)	ILC2 responses and skin inflammation dependent on TSLP, not IL-33 signaling Ear swelling reduced in RORαsg/sg and CD90.1-depleted RAG1−/− mice Ear swelling and skin ILC2 number dependent on IL-25, IL-33, and TSLP ILC2 activation by basophil IL-4	[57, 82, 83]

Open in a separate window

BMT, Bone marrow transplantation; ^sg/sg, staggerer.

Most of the lung disease models have used RAG knockout mice (that lack adaptive immunity) or have evaluated ILC2 responses before generation of adaptive immunity. Thus, the contribution of ILC2s compared with other cell types during an ongoing adaptive response is unclear. However, a few studies have investigated the relative levels of Th2 cytokine-producing cell types during early and later antigen challenges [37, 80]. One study demonstrated that the numbers of lung IL-5⁺ ILC2s were about ½ of IL-5⁺ T cells after 1 and 3 challenges but similar after 10 challenges with house dust mite [80]. Lung IL-13⁺ ILC2s remained less than ½ of IL-13⁺ T cells after 1, 3, and 10 house dust-mite challenges. However, ILC2 and T cell IL-5⁺ and IL-13⁺ cells were similar in BAL regardless of the number of challenges. With the use of an OVA model, the same study showed that ILC2s were a major source of IL-5, but neither T cells nor ILC2s were the dominant source of IL-4 and IL-13. Another study used IL-4 and IL-13 reporter mice and compared numbers of IL-4⁺ and IL-13⁺ cells within the CD4⁺ and CD4⁻ compartments in the BAL after a 12 or 25 day OVA model [37]. The IL-13⁺ CD4⁺ cell levels were 3× that of CD4⁻ (which included ILC2 and other CD4⁻ cells) in the 12 day model and nearly 10× greater in the 25 day model. IL-4⁺ cells in the BAL appeared to be neither CD4⁺ cells nor ILC2s. Together, these reports suggest that the numbers of cytokine-positive ILC2s may be less than CD4⁺ and other cell types during later antigen challenges, but ILC2s continue to produce Th2 cytokines at later stages of inflammatory responses.

One limitation of these studies is that the methods do not account for the level of cytokine production/cell, and previous reports suggest that ILC2s may have a larger capacity for Th2 cytokine production in vitro compared with other cell types. On a per-cell basis, ILC2 have been reported to produce significantly more IL-5 and IL-13 (microgram amounts/5000 cells) compared with mast cells, basophils, NKT cells, and Th2 cells when stimulated with IL-33 [11, 84–87]. Strategies to deplete ILC2s, beginning at later stages of allergic inflammation, as has been done with CD4 cells [88], will be required to address their role in chronic antigen-driven responses.

Mice exposed to the fungal allergen Alternaria develop rapid increases in IL-33, activation of lung ILC2s, and airway eosinophilia, dependent on IL-33 [16, 48]. Alternaria is a protease allergen, and activation of ILC2s through IL-33 may be a common pathway shared with other proteases, including papain [56]. We have reported further that single intranasal administration of Alternaria induces rapid CysLT production that could activate ILC2s through CysLT1R, suggesting that IL-33-independent pathways of ILC2 activation are present early after allergen challenge [34]. Our work has also shown that Alternaria-induced ILC2 activation in mice occurs during a conventional adaptive response to a completely different allergen, suggesting that high Alternaria spore exposure could worsen lung inflammation caused by another allergen [89]. From a translational aspect, Alternaria exposure is associated with severe human asthma, including fatal attacks, although whether rapid IL-33/ILC2 activation occurs in humans is unknown [90].

In addition to promoting pathogenic type 2 lung responses, there is evidence that ILC2s induce tissue repair responses after viral exposure. Depletion of ILC2s in RAG knockout mice infected with influenza virus led to hypoxemia, hypothermia, and loss of epithelial barrier integrity [41]. ILC2s were found to produce high levels of amphiregulin and adoptive transfer of ILC2s, or administration of amphiregulin to ILC2-depleted mice restored airway integrity after influenza infection. We have also reported that ILC2 production of amphiregulin and epithelial expression of the amphiregulin receptor EGFR are rapidly up-regulated after one Alternaria challenge, thus supporting a role for ILC2s in tissue repair postallergen challenge as well [17].

Rhinovirus exposure is a major cause of morbidity for human asthmatics and is the primary cause of asthma exacerbation [2]. In addition to asthma flares, viruses are implicated in the development of asthma [2, 91]. A very recent study reported that rhinovirus infection in neonatal mice, but not adult mice, expanded IL-13⁺ lung ILC2s [92]. ILC2 expansion as well as mucus metaplasia and AHR were dependent on IL-25. Interestingly, the expanded ILC2s persisted for 3 weeks after infection and could provide a robust type 2 response to additional stimuli, resulting in the development of asthma features.

Corticoisteroid resistance can be associated with severe asthma and whether ILC2s are steroid resistant has important therapeutic implications [93]. A recent study suggested that mouse lung ILC2s may be corticosteroid resistant in the presence of TSLP [81]. The authors performed OVA/alum i.p. sensitizations, followed by IL-33 and OVA challenges, in mice and determined that dexamethasone administration did not reduce lung ILC2s. Furthermore, an absence or blockade of TSLP led to a partial increase in steroid sensitivity of lung ILC2s in vivo, and TSLP stimulation of naïve lung ILC2s in vitro led to partial dexamethasone resistance. Our recent work suggests that systemic corticosteroids induce apoptosis of mouse lung ILC2s during repetitive intranasal Alternaria challenges [94]. This suggests that ILC2 corticosteroid resistance may occur under specific conditions that need to be explored further.

In addition to a role for ILC2s in asthma models, ILC2s have been implicated in pulmonary fibrosis. A recent report showed that lung ILC2s expanded 10-fold by Day 7 after bleomycin treatment, and ST2^−/− (IL-33R) mice were resistant to lung fibrosis and lacked ILC2 expansion [95]. Furthermore, adoptive transfer of ILC2s into WT mice receiving bleomycin increased lung inflammation and collagen production as well as lung IL-13 and TGF-β1 mRNA levels. IL-33 also induced M2 polarization of lung macrophages that were required for lung fibrosis. Although the precise role of ILC2s during bleomycin-induced fibrosis is not clear, the suggestion is that IL-33 induces ILC2 IL-13 production, which may have multiple inflammatory and profibrotic effects on several cell types, including macrophages, alveolar epithelium, and fibroblasts.

ILC2s IN HUMAN ASTHMA

A report in 2009 demonstrated the presence of a CD34⁺ non-B/non-T lymphocyte population that produced IL-5 and IL-13 in asthmatic sputum after airway challenge with allergen [96]. Whether these cells are the later-described CRTH2-expressing ILC2s is not known, but the work does suggest that a non-T cell Th2 cytokine-producing cell type may be active in human asthma. Subsequently, ILC2s were shown to be present in human lung and BAL [41, 58]. Barnig et al. [58] reported that peripheral blood ILC2s (lineage-negative CRTH2+ IL-7R⁺) cells were not different in numbers among severe asthmatics, mild asthmatics, and controls. The group went on to show that IL-13 production by peripheral blood ILC2s was enhanced by PGD2 and inhibited by LXA4. A more recent report demonstrated that peripheral blood ILC2s (lineage-negative CRTH2⁺ IL-7R⁺) were increased in patients with allergic asthma compared with those with allergic rhinitis and healthy controls [97]. Furthermore, PBMC production of IL-5 and IL-13, in response to IL-25 and IL-33, was increased in those with allergic asthma. Although both studies differ in correlation of peripheral blood ILC2 numbers with asthma, the patient populations are likely heterogeneous, and the allergic status of individuals in the former study is not clear. Although the function of ILC2s in the lungs of asthmatic patients is not known, upstream cytokines and mediators of ILC2 activation, including IL-25, IL-33, TSLP, LTs, and PGD2, are elevated in human asthma [62, 98–101].

ILC2s IN HUMAN CRS

CRS affects up to 12–13% of the adult general population and is characterized by inflammation of the nasal and paranasal mucosal surfaces, often associated with asthma, allergic disease, aspirin sensitivity, and cystic fibrosis [102, 103]. A subset of patients with CRSwNPs that cause significant morbidity and has led to classification of “endotypes” within CRS [103]. ILC2s were first reported as enriched in nasal polyps in 2011 [33], and a subsequent study identified an increased percentage of ILC2 in inflamed sinonasal mucosa from patients with CRSwNP compared with CRS without nasal polyps [104]. Another report showed a correlation with increased nasal polyp ILC2s and allergic sensitization, as well as numbers of polyp Th2 cells [105]. Our group recently reported that ILC2s were increased in eosinophilic compared with a noneosinophilic polyp endotype and were reduced in patients that had received systemic corticosteroids [94].

Further investigation into mechanisms of ILC2s in nasal polyps revealed that nasal polyp ILC2s produced IL-13 in a GATA3-dependent manner after stimulation with IL-2, IL-33, and TSLP [20]. Furthermore, TSLP enhanced the level of ILC2 GATA3 expression, possibly accounting for increased ILC2 cytokine production by TSLP. Similar to asthma, mediators that activate ILC2s, including epithelial cytokines TSLP and IL-33, as well as LTs, have been detected at higher levels in patients with CRS and are thus available for potent ILC2 stimulation [104, 106, 107].

CRS also includes patients with chronic allergic rhinitis. Our group investigated changes in peripheral blood ILC2s immediately after cat-allergen challenge in allergic rhinitis subjects with positive cat challenges [108]. We found that the percent of CRTH2+ ILC2s approximately doubled after cat-allergen challenge compared with diluent control challenges in the same patients performed at a separate visit. In support of this, a subsequent report showed that peripheral blood ILC2s are increased in grass pollen allergic rhinitis subjects during the pollen season and are reduced by s.c. immunotherapy [109]. Taken together, the available literature suggests that ILC2s may have a role in CRS, including nasal polyposis and allergic rhinitis, and further studies are needed to characterize the relative functional role of ILC2s compared with mast cells, basophils, T cells, and eosinophils.

ILC2 IN SKIN-DISEASE MODELS

Atopic dermatitis is a common chronic skin condition characterized by itchy, scaly rashes, along with skin barrier disruption, eosinophilic infiltration, and high serum IgE. Mouse models of atopic dermatitis are limited, although transgenic expression or instillation of cytokines as well as use of a vitamin D analog have recapitulated some of these clinical features. ILC2s are found in the skin of mice, and the majority of IL-13-positive cells under homeostatic conditions in the skin of IL-13 reporter mice are CD3 negative [78]. This is consistent with other reports showing that ILC2s are the primary Th2 cytokine-producing cell type in different organs in the absence of disease [38, 42].

In a mouse model of atopic dermatitis induced by the vitamin D analog calcipotriol, ILC2s contributed to dermal thickening and inflammation that was dependent on TSLPR signaling but independent of IL-33 [57]. This model has been shown previously to be TSLP dependent, and whether IL-33 is available and active is unclear [110]. Moreover, another recent report showed that skin-specific overexpression of IL-33 induced atopic dermatitis-like pathology with increased dermal eosinophils, skin Th2 cytokines, and expansion of ILC2s [111]. Thus, the availability of 1 or more upstream mediators of ILC2 activation, such as TSLP and/or IL-33, likely drives ILC2 cytokine responses and accumulation that could lead to dermatitis. Interestingly, IL-2 administration to RAG1^−/− mice also resulted in skin ILC2 expansion, Th2 cytokine production, and pathologic features of dermatitis, including epidermal hyperplasia, dermal thickening, and eosinophilic infiltration, suggesting that multiple cytokines under specific conditions can promote ILC2-mediated dermatitis [78]. Similar to lung-disease models, basophils have also been shown recently to support ILC2 accumulation in the skin through IL-4 production [82]. Overall, ILC2s appear to contribute to dermatitis in these studies, but models of atopic dermatitis are limited, and relevance to human disease is largely unclear.

ILC2s IN HUMAN ATOPIC DERMATITIS

Elevated levels of TSLP and IL-33 have been detected in the skin of atopic dermatitis patients, suggesting that ILC2s may also be active in the disease [112, 113]. ILC2s have also been found at higher levels in the skin of atopic dermatitis patients compared with healthy controls [57, 83]. A recent study of ILC2s in human atopic dermatitis demonstrated that human skin ILC2s from diseased patients produce Th2 cytokines in response to IL-33 but not TSLP or IL-25 [83]. Human skin ILC2s express killer cell lectin-like receptor subfamily G member 1, which binds to the epithelial adhesion molecule E-cadherin that is down-regulated in atopic dermatitis. The authors found that E-cadherin inhibited ILC2 proliferation and IL-5 and IL-13 production [83]. Furthermore, ILC2s were recruited to skin of house dust mite-allergic subjects administered intradermal house dust mite. Thus, there may be a role for ILC2s in atopic dermatitis, in part, as a result of reduced inhibition by keratinocyte E-cadherin and increased recruitment upon allergen exposure.

TARGETING ILC2s IN HUMAN DISEASE

Many challenges remain in specifically targeting ILC2s in human disease. First, there is no specific marker or target currently known to be present on ILC2s. Second, the targeting of upstream cytokines, including IL-33, TSLP, and PGD2, will likely have effects on multiple inflammatory and structural cell types known to be regulated by these mediators, and the targeting of 1 mediator may not be sufficient. Finally, some caution going forward is prudent in the targeting of tissue ILC2s that may have a role in metabolic homeostasis and tissue repair. As information regarding the role of ILC2s in human disease accumulates, there is reason to be hopeful that such therapeutic targeting in some patients may be promising.

SUMMARY

The recent discovery of ILC2s has generated significant excitement and has led to an improved understanding of allergic diseases and asthma. ILC2s produce large amounts of Th2 cytokines and are found in human tissues, including respiratory tract and skin. Mouse models have shown that ILC2s can contribute to eosinophilic infiltration, mucus production, AHR, and skin disease. Furthermore, ILC2s are implicated in Th2 cell sensitization, thus linking ILC2s with conventional adaptive Th2 cell paradigms of allergic disease that include specific IgE production. Importantly, ILC2s appear to have roles in metabolic homeostasis and tissue repair, suggesting multifaceted effector functions. Studies correlating ILC2 levels or function with human diseases are emerging, and the targeting of ILC2s may represent a potential therapeutic strategy. However, several important questions regarding the role of ILC2s in allergic diseases remain, including how ILC2s can be specifically targeted and to what extent ILC2s contribute during chronic allergic inflammation (Table 3). The hope is that the discovery of ILC2s will eventually translate into therapy for human disease.

ACKNOWLEDGMENTS

Glossary

Footnotes

REFERENCES