Transcription Factors in the Development and Function of Group 2 Innate Lymphoid Cells

Group 2 innate lymphoid cells (ILC2s) are tissue-resident cells and are a major source of innate TH2 cytokine secretion upon allergen exposure or parasitic-worm infection. Accumulating studies have revealed that transcription factors, including GATA-3, Bcl11b, Gfi1, RORα, and Ets-1, play a role in ILC2 differentiation. Recent reports have further revealed that the characteristics and functions of ILC2 are influenced by the physiological state of the tissues. Specifically, the type of inflammation strongly affects the ILC2 phenotype in tissues. Inhibitory ILC2s, memory-like ILC2s, and ex-ILC2s with ILC1 features acquire their characteristic properties following exposure to their specific inflammatory environment. We have recently reported a new ILC2 population, designated as exhausted-like ILC2s, which emerges after a severe allergic inflammation. Exhausted-like ILC2s are featured with low reactivity and high expression of inhibitory receptors. Therefore, for a more comprehensive understanding of ILC2 function and differentiation, we review the recent knowledge of transcriptional regulation of ILC2 differentiation and discuss the roles of the Runx transcription factor in controlling the emergence of exhausted-like ILC2s. The concept of exhausted-like ILC2s sheds a light on a new aspect of ILC2 biology in allergic diseases.


Introduction
Innate lymphoid cells (ILCs) are lymphocytes responsible for innate cytotoxicity or helper functions [1][2][3]. While ILCs with cytotoxic capability are conventional natural killer (NK) cells which do not have tissue-resident properties, ILCs with helper functions generally reside in the tissue, provoke immune responses against pathogens, and maintain mucosal integrity [4,5]. Helper ILCs are unable to directly sense pathogenic antigens because they do not express antigen-specific receptors such as T cell receptor and B cell receptor. Instead, helper ILCs are generally stimulated by cytokines released from damaged-epithelial cells, phagocytes, and dendritic cells [6][7][8][9][10][11]. Based on cytokine production and requirements of transcription factors, helper ILCs are classified into three distinct subsets; ILC1s, ILC2s, and ILC3s which mirror T H 1, T H 2, and T H 17, respectively [3,12]. ILC1s secrete interferon g (IFNγ) and require T-bet for their differentiation [13]. T-bet is also important for the final maturation of conventional NK cells, which is highly dependent also on eomesodermin (Eomes) [14]. ILC2s are characterized by the production of the T H 2 cytokines interleukin 4 (IL-4), IL-5, IL-13, and granulocyte-macrophage colony-stimulating factor (GM-CSF). All of these expressions are positively regulated by the transcription factor GATA binding protein 3 (GATA-3) [15]. Both RAR-related orphan receptor gamma t (RORγt) and aryl hydrocarbon receptor (AhR) are required for the differentiation of ILC3s as well as for the production of IL-17 or IL-22 by ILC3s [16][17][18]. A substantial number of ILC1s

Bcl11b
Bcl11b is the earliest marker for the ILC2-lineage. The Bcl11b-expressing cells among the CHILP population specifically develop into ILC2s (Figure 1) [41]. Bcl11b ablation in hematopoietic stem cells impairs ILC2 differentiation in the bone marrow and peripheral tissues. Bcl11b expression in ILC subsets is controversial. While Yu et al. suggested that Bcl11b expression among ILC subsets was restricted to ILC2s [43], and Walker et al. showed that ILC1s and ILC3s expressed Bcl11b [42]. Conditional deletion of Bcl11b in ILC2s results in the generation of ILC2s with ILC3 gene signatures [41]. Thus, Bcl11b-deficient ILC2s have a reduction in GATA-3, Gfi-1, and RORα expression, and an increase in sex determining region Y-box 4 (Sox4), RORγt, and AhR expression. Bcl11b is associated with the Gfi1 and Ahr gene loci, but not the Gata3 or Rorc gene loci, suggesting that Gfi1 and AhR expression may be directly regulated by Bcl11b. Enhanced Gfi-1 expression in Bcl11b-deficient ILC2s restores GATA-3 expression and down-regulates RORγt expression. Therefore, Bcl11b positively regulates GATA-3 expression and negatively regulates RORγt expression through Gfi-1 induction. is expressed in early innate lymphoid progenitor cell (EILP), and its expression is maintained in all ILC subsets. Ets-1 up-regulates Id2 in common helper innate lymphoid progenitors (CHILP). Bcl11b-expressing progenitors in the CHILP population give rise to ILC2 (green blocked arrow). Bcl11b positively regulates RORα and Gfi-1 both of which are indispensable for ILC2 differentiation (black blocked arrow from Bcl11b). Gfi-1 induces GATA-3 for ILC2 specification (black blocked arrow from Gfi-1). Runx proteins repress GATA-3 activity by antagonistic binding (red T-bar).

Bcl11b
Bcl11b is the earliest marker for the ILC2-lineage. The Bcl11b-expressing cells among the CHILP population specifically develop into ILC2s (Figure 1) [41]. Bcl11b ablation in hematopoietic stem cells impairs ILC2 differentiation in the bone marrow and peripheral tissues. Bcl11b expression in ILC subsets is controversial. While Yu et al. suggested that Bcl11b expression among ILC subsets was restricted to ILC2s [43], and Walker et al. showed that ILC1s and ILC3s expressed Bcl11b [42]. Conditional deletion of Bcl11b in ILC2s results in the generation of ILC2s with ILC3 gene signatures [41]. Thus, Bcl11b-deficient ILC2s have a reduction in GATA-3, Gfi-1, and RORα expression, and an increase in sex determining region Y-box 4 (Sox4), RORγt, and AhR expression. Bcl11b is associated with the Gfi1 and Ahr gene loci, but not the Gata3 or Rorc gene loci, suggesting that Gfi1 and AhR expression may be directly regulated by Bcl11b. Enhanced Gfi-1 expression in Bcl11b-deficient ILC2s restores GATA-3 expression and down-regulates RORγt expression. Therefore, Bcl11b positively regulates GATA-3 expression and negatively regulates RORγt expression through Gfi-1 induction.

Gfi-1
Gfi-1 expression in ILC2 is relatively higher than other ILCs. Gfi-1 is required for ILC2 differentiation and optimal T H 2 cytokine responses to IL-33 [44]. Gfi-1 positively regulates the expression of Gata3 and Il1rl1 (Figure 1). Deletion of Gfi-1 causes aberrant IL-17a production by ILC2s due to dysregulated expression of RORγt [44]. Increased RORγt expression in Gfi-1-deficient ILC2s may be explained by increased expression of Sox4, which is known to induce RORγt expression. Thus, Gfi-1 suppresses the machinery for the ILC3 lineage and maintains the effector state of ILC2s in allergic diseases.

RORα
Although RORα is expressed in CHILPs and all ILC subsets, it is specifically necessary for ILC2 differentiation among ILC subsets ( Figure 1) [7,45,47]. A RORα-dependent ILC2 differentiation mechanism has not been described yet. Differential interaction of RORα with the gene loci in ILC2s and other ILCs should be examined to understand RORα function in ILC2 differentiation and function. RORα-deficient staggerer mice have been used to observe physiological effects resulting from ILC2 deficiency [7,45]. However, deletion of RORα dampens regulatory T cell (Treg) and ILC3 function, but not their differentiation [48,49]. These data suggest that there might be other unknown immunological defects in the RORα-deficient staggerer mice and the related mice with conditional deletion of RORα.

Ets1
Ets-1 induces Id2 which is critical for differentiation of ILC progenitors, CHILPs [46]. However, for unknown reasons, ILC2s without Ets-1 are reduced only in the bone marrow, but not in the lung. Ets-1 is required for ILC2s to expand and produce IL-5 and IL-13 in response to IL-33. In contrast, IL-6 and IL-9 production are up-regulated in Ets-1-deficient ILC2s cultured with IL-33. It remains to be elucidated how Ets-1 regulates the ILC2 fitness and function.

Global Effects of Runx Proteins in Immune Cells
Runx proteins are a family of transcription factors necessary for many biological processes ranging from differentiation, function, proliferation, tumorigenesis, and cellular identity [50,51]. There are three mammal members in the Runx protein family: Runx1, Runx2, and Runx3. All Runx proteins require heterodimer formation with Subunit b of core binding factor (Cbfβ) to exert their function as transcription factors. All three Runx genes generate two transcripts from distal (P1) and proximal (P2) promoters [52]. These two transcripts generate Runx protein isoforms with distinct N-terminal sequences. Utilization of the promoters differs among cell types. Runx1 and Runx3 genes are expressed in many types of immune cells [50]. Runx1 is required for the emergence of hematopoietic stem cells from hemogenic endothelium. Loss of Runx1 causes embryonic lethality due to bleeding in the brain caused by the absence of platelets [53,54]. Runx1 is also involved in the early thymocyte differentiation and development of invariant natural killer T (iNKT) cells, regulatory T cells and T H 17 cells [51,[55][56][57][58]. Runx3 is a critical transcription factor for the differentiation of CD8 + T cells and T H 1 cells [56,59]. Runx3 is crucial to induce effector genes such as Ifng, Gzmb, Eomes, and to repress Il4 gene in T H 1 cells. Runx3 transcript from the P1 promoter is correlated with protein expression level in T cells because the P2-Runx3 transcript utilizes IRES-mediated translation [52,60]. Runx2 is required for the differentiation of osteoblasts and bone formation [61,62]. Runx2 expression in hematopoietic cells is rather restricted to plasmacytoid dendritic cells [63]. Runx2-deficient plasmacytoid dendritic cells have deficits in migrating to the peripheral tissues.

Runx3 Is Required for the Differentiation of ILC1s and ILC3s
Helper ILCs and conventional NK cells preferentially express Runx3 mRNA among all Runx family genes [64,65]. However, ILC subsets are characterized by the levels of Runx3 P1 transcripts and Runx3 protein expression. High Runx3 protein expression is observed in NK cells, which are innate counterparts of CD8 + T cells, and in ILC1s, innate counterparts of T H 1 cells. ILC3s, innate counterpart for T H 17 cells, express an intermediate level of Runx3, while Runx1 is likely to be a major Runx protein to drive T H 17 program. Runx3 expression is very low in ILC2s, innate lymphoid cells. Ablation of Runx3 in all hematopoietic cells leads to a reduction of NK cells, ILC1s, and ILC3s but not ILC2s. Runx3 is necessary for optimal IFNγ production by ILC1s. Ablation of Cbfβ or a combined loss of both Runx1 and Runx3 result in a dramatic reduction of NK cells and ILC1s, suggesting that Runx1 compensates for the loss of Runx3 in these cells. ILC1s fall into an apoptotic state in the absence of Runx proteins, which is a result of low expression of anti-apoptotic Bcl2. For the development of ILC3s, Runx3 induces expression of an ILC3 master regulator gene, RORγt mainly by direct binding to the intronic enhancer region of the RORγt [66]. AhR is also required for full production of IL-17 and IL-22 together with RORγt [17]. Runx3 deletion impairs expression of Ahr in ILC3s. Upon a half dosage of RORγt expression caused by heterozygous gene knock-out, Ahr expression is reduced by half in the ILC3s. Thus, Runx3 regulates RORγt expression and its downstream target, AhR.
Lymphoid tissue-inducer cells (Lti) cells are indispensable for secondary lymphoid organogenesis and are considered as a subpopulation of ILC3s because Lti cells are characterized by and are dependent on RORγt expression. Lti cells appear in the embryonic intestine at embryonic day 12. An early study clarified the importance of Runx1 for Lti cell differentiation via RORγt induction in the embryos [67]. Although RORγt expression in Lti-cells is reduced by Runx3 deficiency, the number of Lti cells in the embryonic intestine is normal in the absence of Runx3 [64]. These data suggest that there might be distinct functions of Runx proteins for the generation of Lti cells.
Runx proteins also regulate early ILC differentiation. Runx1 and Runx3 expression in ILC progenitors increases as CLPs give rise to CHILPs and ILCPs ( Figure 1) [64]. Expression of Runx1 and Runx3 reaches the highest level at an ILCP stage in the course of ILC differentiation from CLPs to ILC subsets. Runx proteins are essential to induce PLZF expression, which is the marker for ILCPs, by direct interaction with an enhancer of the Zbtb16 locus encoding PLZF [68]. ILC1s and ILC3s down-regulate Runx1 expression but maintain intermediate to high levels of Runx3 expression [64]. In contrast, down-regulation of Runx3 occurs in ILC2 progenitor cells and ILC2s which express more Runx1 than other ILCs. Therefore, the levels of Runx3 expression specifies ILC lineages.

Runx Proteins Prevent Steady-State ILC2s from Overactivation
Runx3 deletion is not enough for obvious phenotypic changes in ILC2s [64,69]. ILC2s lacking Runx3 normally populate the peripheral tissues and are functionally intact because Runx1 is also expressed and compensate Runx3 function in ILC2s. When the function of all Runx proteins is abrogated by ablation of Cbfβ, a binding partner of Runx proteins, in the ILC2s, the ILC2s are normally present in the tissues but exhibit an activated phenotype even in the steady-state condition. Killer cell lectin-like receptor G1 (KLRG1) is a known marker for activated ILC2s, and Thy1 is down-regulated in the ILC2s when they are stimulated by IL-25. Interestingly, loss of Runx function in ILC2s induces high KLRG1 expression and low Thy1 expression with unleashed IL-5 production. As ILC2s are a major source of IL-5, which recruits eosinophils to the lung, eosinophils in the bronchoalveolar space are increased by overproduction of IL-5 from the activated ILC2s lacking Runx function. The infiltration of eosinophils to the bronchoalveolar space seems to occur at subclinical levels because lung epithelial cells are not significantly damaged by the adoptive transfer of the Cbfβ-deficient ILC2s.
The activated ILC2 phenotype caused by Cbfβ deficiency is reminiscent of those observed in the T cells lacking Cbfβ function. T H 2 skewing occurs in CD4 T cells in the absence of Cbfβ, resulting in asthma-like symptoms in the lung [59]. Runx/Cbfβ complexes associate and antagonize GATA-3, a master regulator of T H 2 cells, in T cells [70]. Runx/Cbfβ complexes also suppress the expression of IL-4, which is a critical T H 2 cytokine for type 2 immunity, through binding to a silencer region in the Il4 gene locus [59]. Therefore, the balance between Runx and GATA-3 dictates the functional state of T H 2 cells.
ILC2 activity at steady state is also determined by the balancing action of Runx proteins against GATA-3 ( Figure 2). While GATA-3 overexpression increases IL-5 production in ILC2s, Runx3 cancels such effects in ILC2s [69]. In addition to IL-5, transcriptome analysis demonstrates that many genes regulated by GATA-3 are inversely regulated by Runx proteins. In steady-state ILC2s, IL-5 production is increased by the absence of Cbfβ, presumably via antagonizing GATA-3 function. However, we cannot deny the possibility that Runx/Cbfβ complexes work as a direct repressor for the Il5 gene as is the case for Il4, because Cbfβ associates with several genomic regions around the Il5 gene in ILC2s. Deletion of either Runx1 or Runx3 alone does not induce IL-5 overproduction from ILC2s. However, when both Runx1 and Runx3 are deleted, ILC2s show the activated phenotype, suggesting that the redundant function of Runx proteins restrains the basic activity of ILC2s in steady-state. master regulator of TH2 cells, in T cells [70]. Runx/Cbfβ complexes also suppress the expression of IL-4, which is a critical TH2 cytokine for type 2 immunity, through binding to a silencer region in the Il4 gene locus [59]. Therefore, the balance between Runx and GATA-3 dictates the functional state of TH2 cells. ILC2 activity at steady state is also determined by the balancing action of Runx proteins against GATA-3 ( Figure 2). While GATA-3 overexpression increases IL-5 production in ILC2s, Runx3 cancels such effects in ILC2s [69]. In addition to IL-5, transcriptome analysis demonstrates that many genes regulated by GATA-3 are inversely regulated by Runx proteins. In steady-state ILC2s, IL-5 production is increased by the absence of Cbfβ, presumably via antagonizing GATA-3 function. However, we cannot deny the possibility that Runx/Cbfβ complexes work as a direct repressor for the Il5 gene as is the case for Il4, because Cbfβ associates with several genomic regions around the Il5 gene in ILC2s. Deletion of either Runx1 or Runx3 alone does not induce IL-5 overproduction from ILC2s. However, when both Runx1 and Runx3 are deleted, ILC2s show the activated phenotype, suggesting that the redundant function of Runx proteins restrains the basic activity of ILC2s in steady-state. Figure 2. Runx/Cbfβ complexes differentially regulate ILC2 effector function at steady state and in severe inflammation. In the steady state of the lung, Runx/Cbfβ complexes suppress ILC2 activity through antagonistic binding to GATA-3 which positively regulates ILC2 effector function including constitutive IL-5 production (left red T-bar). During severe allergic airway inflammation, Runx/Cbfβ complexes are essential for ILC2s to respond to IL-33 and maintain their activity. Runx/Cbfβ complexes inhibit the emergence of exhausted-like ILC2s which are normally rare at the site of severe inflammation (right red T-bar). Red dots: IL-5. Blue dots: IL-13.

Runx Proteins Inhibit the Emergence of Exhausted-Like ILC2s during Allergic Inflammation
ILC2s are activated by IL-33, IL-25, TSLP, TNF superfamily ligand TL1A (TL1A), and neuropeptides such as vasoactive intestinal polypeptide (VIP), neuromedin U and, calcitonin generelated peptide [6,11,30,31,[71][72][73]. Furthermore, IL-33 is the most potent ILC2 stimulator in vitro and in vivo and plays an important role in the pathogenesis of allergy. Interestingly, Cbfβ-deficient ILC2s have defects in ILC2 cytokine production and proliferation during in vitro responses to IL-33 [69]. Expression of Il5, Il9, Il13, Csf2 encoding GM-CSF, and Areg encoding amphiregulin upon IL-33 stimulation is all downregulated in Cbfβ-deficient ILC2s. In addition, the expression of many molecules critical for ILC2 activity is dampened in the Cbfβ-deficient ILC2s. ILC2s have machinery for extending immune responses upon their activation. For instance, IL-9 produced by ILC2s binds IL-9 receptor on ILC2s and stimulates their expansion [32]. Inducible costimulatory ligand (ICOSL) induced on the activated ILC2s contributes to their survival and cytokine production by receiving

Runx Proteins Inhibit the Emergence of Exhausted-Like ILC2s during Allergic Inflammation
ILC2s are activated by IL-33, IL-25, TSLP, TNF superfamily ligand TL1A (TL1A), and neuropeptides such as vasoactive intestinal polypeptide (VIP), neuromedin U and, calcitonin gene-related peptide [6,11,30,31,[71][72][73]. Furthermore, IL-33 is the most potent ILC2 stimulator in vitro and in vivo and plays an important role in the pathogenesis of allergy. Interestingly, Cbfβ-deficient ILC2s have defects in ILC2 cytokine production and proliferation during in vitro responses to IL-33 [69]. Expression of Il5, Il9, Il13, Csf2 encoding GM-CSF, and Areg encoding amphiregulin upon IL-33 stimulation is all downregulated in Cbfβ-deficient ILC2s. In addition, the expression of many molecules critical for ILC2 activity is dampened in the Cbfβ-deficient ILC2s. ILC2s have machinery for extending immune responses upon their activation. For instance, IL-9 produced by ILC2s binds IL-9 receptor on ILC2s and stimulates their expansion [32]. Inducible costimulatory ligand (ICOSL) induced on the activated ILC2s contributes to their survival and cytokine production by receiving ICOS signals [74]. Activated ILC2s express a variety of neuropeptide receptors including Vipr2, a VIP receptor, and Nmur1, a neuromedin receptor for expansion and cytokine production of ILC2s during allergy [30,[71][72][73]. Surprisingly, expression of Il9, Il9r, Icos, Vipr2, and Nmur1 in activated ILC2 are all impaired by Cbfβ-deficiency [69]. Furthermore, ILC2s lacking Cbfβ show up-regulated expressions of a series of T-cell exhaustion marker genes, such as Tigit, Il10, Prdm1, Ctla4, and Lag3. It has been documented that T cells acquire a state of dysfunction, called T cell exhaustion, after chronic exposure to cancer and viruses as a result of continuous antigen stimulation [75][76][77]. Exhausted T cells are characterized by impaired ability of proliferation, cytokine production and, effector function after re-stimulation. Expression of a variety of inhibitory receptors such as programmed cell death 1 (PD-1), T cell immunoreceptor with Ig and ITIM domains (Tigit), and lymphocyte Activation Gene-3 (Lag3) are highly induced on exhausted T cells. We, therefore, define hyporesponsive ILC2s with increased expression of T cell exhaustion markers as "exhausted-like" ILC2s.
Since both Tigit and IL-10 are highly expressed in Cbfβ-deficient ILC2s, these two molecules could serve as good markers to define ILC2s with hyporesponsive characteristics. Indeed, a low number of hyporesponsive Tigit + IL-10 + ILC2s emerge at the site of severe allergic inflammation even in wild-type mice [69]. Severe allergy is known to induce IL-10-producing ILC2s, which are associated with decreased eosinophil recruitment in the lung [26]. Tigit expression is restricted to high IL-10-producers in the activated ILC2s [69]. The low reactive Tigit + IL-10 + ILC2s also express high levels of ILC2 activation markers such as PD-1, KLRG1, and glucocorticoid-induced tumor necrosis factor receptor (GITR). When mice are intranasally administrated with papain, a proteinase allergen, the bronchoalveolar space is the most inflamed site in the lung. The Tigit + IL-10 + ILC2s with low reactivity can be found only in the bronchoalveolar space, but not in the lung during severe subacute allergy induced by high doses of papain. The absence of Cbfβ in ILC2s enhances the generation of the Tigit + IL-10 + ILC2s and ameliorates allergic inflammation induced by papain inhalation in the subacute phase. However, when mice are chronically inoculated with papain for a month, increased low reactive Tigit + IL-10 + ILC2s from Cbfβ deficiency do not reduce the degree of chronic inflammation. If mice have the papain treatment interval for recovery after the first course of challenges, the emergence of the low reactive Tigit + IL-10 + ILC2s enhanced by Cbfβ deficiency is associated with attenuation of allergic inflammation induced by second challenges of papain treatment. Therefore, the presence of exhausted-like Tigit + IL-10 + ILC2s is likely to reduce acute deterioration of chronic allergy.
Our data clearly show that pre-activated Cbfβ-deficient ILC2s are hyporesponsive to IL-33 stimulation in vitro and in vivo [69]. If enhanced GATA-3 activation by the absence of Cbfβ is involved in the generation of exhausted-like ILC2s, GATA-3 overexpression alone would lead to the emergence of exhausted-like ILC2s. However, ectopic expression of GATA-3 conversely elicits increased ILC2 reactivity upon IL-33 stimulation in vitro and in vivo. Of note, Runx3 overexpression over GATA-3 expression down-modulates ILC2 responsiveness, suggesting that the reactivity of ILC2s to IL33 stimulation is regulated by the balance between GATA-3 and Runx3 even in the activation phase ( Figure 3). Conversely, transcription factors regulate gene expression through association with regulatory genomic regions such as promoters, enhancers, or repressors. Chromatin immunoprecipitation sequencing data suggest that the Cbfβ binding pattern in ILC2s activated with IL-33 differs from that in ILC2s without IL-33 stimulation. Interestingly, Cbfβ-binding peaks induced specifically upon IL-33 stimulation are enriched in regions nearby the signature genes for exhausted-like ILC2s. For example, the Il5, Il13, Nmur1, Vipr2, Il10, Tigit, Prdm1, Lag3, and Ctla4 gene loci are marked by Cbfβ-bindings specific to IL-33 stimulation. These Cbfβ-binding peaks are not bound by GATA-3, suggesting that regulation of the activity of those possible regulatory regions by Runx/Cbfβ is likely to be independent of GATA-3. Furthermore, Runx/Cbfβ complexes bind the presumed enhancers of ILC2 functional gene loci marked by H3K27 acetylation in the IL-33-stimulated ILC2s (Figure 3). Cbfβ binding peaks overlapping with H3K27 acetylation peaks are found around the Il5, Il9, Il13, Areg, and Nmur1 gene loci. A decrease of H3K27 tri-methylation at the Il10 gene by the absence of Cbfb suggests that Cbfβ could repress the Il10 locus through epigenetic mechanisms ( Figure 3). Therefore, Runx/Cbfβ complexes directly contribute to the phenotypical gene expression of exhausted like ILC2s.

Conclusions
We reviewed how transcription factors regulate ILC2 differentiation and function. A new mechanism by which Runx proteins confer sustained ILC2 reactivity during allergic inflammation was identified. Exhausted-like ILC2s are a small population of over-activated ILC2s expressing inhibitory molecules. PD-1 is a good example of an inhibitory receptor that is induced on activated ILC2s and downmodulate ILC2 activity against worm infection [78]. However, PD-1 on ILC2s could be an activation marker rather than exhaustion marker, because PD-1 + ILC2s generally produce more TH2 cytokines than PD-1 -ILC2s. ILC2s produce IL-10 when the cells are exposed to severe allergic inflammation. However, higher TH2 cytokine production by IL-10-producing ILC2s suggests that IL-10-producing ILC2s are likely to be in more activated than non-IL-10 producing ILC2s [26]. We revealed that Tigit + IL-10 + ILC2s have exhausted-like features at the site of severe inflammation. Exhausted-like ILC2s are very rare in the mouse model of subacute allergy as long as Runx proteins are functional. Whether exhausted-like ILC2s are present in humans remains elusive. However, if they are present, exhausted-like ILC2s might be accumulated in patients with chronic allergy and contribute to constraining allergic responses. Thus, exhausted-like ILC2s may serve as beneficial cells to those patients and have the therapeutic potency to be designated a new target for allergic disorders. Clinical trials are warranted to clarify the physiological effects of exhausted-like ILC2s in chronic allergy.
Author Contributions: T.E and I.T. discuss the concept and wrote this review article.

Conclusions
We reviewed how transcription factors regulate ILC2 differentiation and function. A new mechanism by which Runx proteins confer sustained ILC2 reactivity during allergic inflammation was identified. Exhausted-like ILC2s are a small population of over-activated ILC2s expressing inhibitory molecules. PD-1 is a good example of an inhibitory receptor that is induced on activated ILC2s and downmodulate ILC2 activity against worm infection [78]. However, PD-1 on ILC2s could be an activation marker rather than exhaustion marker, because PD-1 + ILC2s generally produce more T H 2 cytokines than PD-1 -ILC2s. ILC2s produce IL-10 when the cells are exposed to severe allergic inflammation. However, higher T H 2 cytokine production by IL-10-producing ILC2s suggests that IL-10-producing ILC2s are likely to be in more activated than non-IL-10 producing ILC2s [26]. We revealed that Tigit + IL-10 + ILC2s have exhausted-like features at the site of severe inflammation. Exhausted-like ILC2s are very rare in the mouse model of subacute allergy as long as Runx proteins are functional. Whether exhausted-like ILC2s are present in humans remains elusive. However, if they are present, exhausted-like ILC2s might be accumulated in patients with chronic allergy and contribute to constraining allergic responses. Thus, exhausted-like ILC2s may serve as beneficial cells to those patients and have the therapeutic potency to be designated a new target for allergic disorders. Clinical trials are warranted to clarify the physiological effects of exhausted-like ILC2s in chronic allergy.
Author Contributions: T.E and I.T. discuss the concept and wrote this review article.

Conflicts of Interest:
The authors declare no conflict of interest.