The interplay between innate lymphoid cells and microbiota

ABSTRACT Innate lymphoid cells (ILCs) are mainly resident in mucosal tissues such as gastrointestinal tract and respiratory tract, so they are closely linked to the microbiota. ILCs can protect commensals to maintain homeostasis and increase resistance to pathogens. Moreover, ILCs also play an early role in defense against a variety of pathogenic microorganisms including pathogenic bacteria, viruses, fungi and parasites, before the intervention of adaptive immune system. Due to the lack of adaptive antigen receptors expressed on T cells and B cells, ILCs need to use other means to sense the signals of microbiota and play a role in corresponding regulation. In this review, we focus on and summarize three major mechanisms used in the interaction between ILCs and microbiota: the mediation of accessory cells represented by dendritic cells; the metabolic pathways of microbiota or diet; the participation of adaptive immune cells.

Then, we focus on arranging and concluding mechanisms of the interactions between ILCs and microbiota.

Nomenclature and classification
NK cells were first identified and regarded as the only innate lymphoid cells (9). As more ILC subsets were discovered, they were divided into three groups based on the producing of characteristic cytokines: NK cells and ILC1s constitute Group 1 ILCs, ILC2s alone represent Group 2 ILCs and NCR + ILC3s, NCR − ILC3 and lymphoid tissue-inducer cells constitute Group 3 ILCs (10). According to similar expression of cytokines, these three groups of ILCs successively serve as innate immune counterparts for CD4 + T helper (Th)1, Th2, and Th17/22 cells (11). In recent years, according to the developmental trajectories and functions of the ILC subsets, ILCs have been re-divided into five subsets: NK cells, ILC1s, ILC2s, ILC3s, and LTi cells (7). In addition, several new ILC subsets were discovered. For example, regulatory ILCs (ILCregs) have an unique genetic profile and can alleviate innate intestinal inflammation (12). ILCX is independent of the three typical ILC groups, with its functions to be studied (13). A list of classification and characteristic of ILCs is presented in Table 1.

NK and ILC1s
NK cells are mainly distributed in peripheral blood, secondary lymphoid tissues, and peripheral immune organs, and are recruited to sites of inflammation upon pathogen invasion (14). ILC1s were first identified in human tonsil and ileum, and mainly resident in human mucosal tissues (15). Both NK cells' and ILC1s' development and functions depend on the expression of T-box transcription factor T-bet (encoded by Tbx21), with IFN-γ as the main effector cytokine (7,13). NK cells mostly express high level of EOMES compared to lower expression in ILC1s, thus serving as a marker to distinguish between NK cells and ILC1s (16). IFN-γ, TNF-α, perforin, and granzyme produced by NK cells and ILC1s exert functions of antitumor and antiintracellular pathogenic microbial infection, as well as cytotoxic effect including induction of cell lysis and apoptosis (17)(18)(19). IL-2, IL-12, IL-15, and IL-18 all have different activation effect on Group 1 ILCs (16). Differently, the cytotoxic effect of ILC1s is relatively weak or even absent than that of NK cells (14).

ILC3s
ILC3s are mainly located in mucosal tissues such as lamina propria of small intestine, with minor distribution in lymphoid tissues such as tonsil (24). ILC3s are characterized by the expression of transcription factor retinoic acid-related orphan receptor gamma subtype t (RORγt) and the production of cytokines such as IL-17A, IL-17F, IL-22, and granulocytemacrophage colony-stimulating factor (GM-CSF) (25)(26)(27). According to the expression of NKp46 (in mouse) or NKp44 (in human) on the cell surface, ILC3s are divided into two subsets: NCR + and NCR − ILC3s (28). Moreover, the main homeostasis cytokine produced by ILC3s is IL-22, through which it maintains the intestinal barrier, promotes differentia tion of mucus-producing goblet cells, prevents bacterial translocation, fight against the pathogen invasion, and other regulatory effect to maintain homeostasis (7). Differently, IL-17 supports the release of chemokines, which recruits proinflammatory neutrophils and may lead to intestinal inflammation (29).

LTi cells
LTi cells are mainly located in secondary and tertiary lymphoid structures and are critical for the formation of secondary lymph nodes and Peyer's patch in fetal stage through lymphotoxin (30). Like ILC3s, LTi cells also strictly depend on RORγt expression, with similar expression of characteristic transcription factors and production of cytokines. Differently, LTi cells express c-Kit and CCR6, but do not express NCR (17,19).

DCs together with macrophage and monocyte
Dendritic cells (DCs), macrophages, and monocytes first need to be activated by the microbiota, and then act as antigen-presenting cells to transmit signals to ILCs. Bacterial stimulation to DCs may be achieved by pattern recognition receptor TLR on DCs, such as TLR2 can recognize peptidoglycan and lipoteichoic acid in the cell wall of Lacto bacillus acidophilus (L. acidophilus), which induces the maturation of mouse DCs and the production of IL-12 (14). Specific to NK cells, Lactobacillus pentosus (L. pentossus) strain S-PT84 can induce IFN-γ production by activating CD11c + DCs in TLR2 and/or TLR4 dependent manner (31). Although NK cells themselves also express TLR, bacteria cannot directly stimulate IFN-γ production by them (16,32). It is different in term of virus that DCs produce IFN-Ⅰ (type Ⅰ interferon) stimulated by Mouse Cytomegalo Virus (MCMV) and Lymphocytic Choriomeningitis Virus (LCMV) infection, and then DCs are upregulated to trans-present IL-15 to NK cells after receiving IFN-Ⅰ stimulation through own IFN-receptors ( Fig. 1) (33,34). Among many effector cytokines produced by NK cells, it has been well studied that IL-2 can activate and enhance NK cell functions (35). Moreover, IL-15 forms a complex with IFN-Ⅰ-dependent IL-15Rα on DCs to activate NK cells in a trans-presenting manner by cell-cell contact (36). It has also been proved that IL-12 (p70), IL-18, and IL-1β can help NK cells' activation and functions, which produced by CD11c + myeloid dendritic cells (CD11c + mDCs) in Salmonella typhimurium (S. typhimurium) infection model in vitro ( Fig. 1) (16). Notably, a research indicates that IL-12 is redundant in human response to mostly microorganisms (37).
In addition to NK cells, Group 3 ILCs have also been proved to require the help of DCs, macrophages, monocytes to establish connections with microbiota. Citrobacter roden tium (C. rodentium) can promote IL-22 production by ILC3s through stimulating CD11c + mDCs to produce IL-23 and IL-1β ( Fig. 1). Notably, IL-23 and IL-1β are redundant in stimulating IL-22 production by colonic ILC3s (38). Moreover, the critical role of adaptor protein MyD88 on DCs in TLR-mediated innate immune activation through recognition of microbe-associated molecular patterns (MAMPs) has been demonstrated. S. typhimu rium selectively enhances IL-22 production of ILC3 by secreting flagellin to activate TLR5-MyD88-IL-23 signaling pathway in antigen presenting cells (APCs), such as DCs (39). Similarly, in C. rodentium infection, MyD88 signaling pathway in DCs is sufficient to stimulate the expression of proinflammatory cytokines including IL-6, IL-1β, IL-23, and then induce the early response of Group 3 ILCs ( Fig. 1) (40). GM-CSF can stimulate the release of IL-10 and other cytokines from DCs and monocytes to promote regulatory T cell (Treg) differentiation, thus maintaining intestinal tolerance. At steady state, GM-CSF is mainly produced by RORγt + ILC3s, and this process depends on IL-1β released by macrophages following microbiota stimulation ( Fig. 1) (41). In addition to the positive regulatory pathways described above, negative ones have also been reported. In inflammatory bowel disease (IBD), GPR109a signaling pathway may inhibit the microbialinduced production of multiple inflammatory cytokines by DCs, including IL-23 in colon, thereby inhibiting ILC3s to suppress inflammation ( Fig. 1) (29). Interestingly, ILC3s sometimes exhibit the characteristics of Group 1 ILCs. After bacterial stimulation, colonic mDCs express high level of IL-15Rα and trans-present IL-15 to ILC3s. Moreover, cispresentation of IL-15 to ILC3s is also allowed because of the co-expression of all three Minireview mBio IL-15R subunits (IL15Rα/β/γ). Ultimately, granzyme B expression of ILC3s is increased, with some ILC3 subsets even co-expressing perforin to exert a cytotoxic effect similar to NK cells ( Fig. 1) (42).

Intestinal epithelial cells
IECs directly contact with microbiota in the gut, playing an integral role in mediating microbiota-ILCs interactions. Previous studies have focused on ILC3-dependent-IL-22 induced by microbiota (Fig. 1). IL-22 exerts functions by binding to IECs through its receptor. Mechanistically, IL-22 phosphorylates transcription factor STAT3 and modulate gene expression, after binding to its receptor (26,43). Specifically, IL-22 promotes differentiation of mucus-producing goblet cells and maintains differentiation of crypt stem cells into IECs, which serve to protect the intestinal barrier (44,45). Moreover, upon stimulation with IL-22, IECs and special Paneth cells produce high level of antibacte rial peptides including S100A8/A9 and RegⅢγ to suppress pathogenic bacteria and maintain intestinal homeostasis (46). Besides being passively regulated by IL-22, several studies in recent years have identified that IECs can take the initiative. Depending on IFN-γ-producing NK1.1 + cells and IL-12, IECs can produce IL-7 induced by C. rodentium infection. IL-7 plays an important protective role during the early phase of infection and acts through its receptor on APCs and CD4 + LTi cells ( Fig. 1) (47). In the same infection of C. rodentium, IECs' intrinsic expression of IKKα is shown to be critical for IL-22 production by Group 3 ILCs in the gut (48).

Ahr
The aryl hydrocarbon receptor (Ahr) is an environmental sensor that works in both innate and adaptive immune systems (49). Ligands for Ahr can be derived from diet (e.g., cruciferous plants), microbial flora (e.g., Lactobacillus tryptophan metabolite indole), and/or host cells (50,51). Both bacteria-derived and diet-derived ligands contribute to the maintenance of RORγt + ILCs (50). It has been shown that certain bacteria in the gut can produce metabolites from tryptophan that activate Ahr in vitro (52). Notably, ligands derived from host cells may make a significant contribution to the development of RORγt + ILCs, because deprivation of Ahr ligands from mouse diet has little effect in the accumulation of RORγt + ILCs in the gut (53). Studies have demonstrated that the co-expression of Ahr and RORγt in RORγt + ILCs can significantly enhance the recruitment of Ahr to the IL-22, synergistically promoting IL-22 expression and contributing to the maintenance and functions of RORγt + ILCs, which play important roles in resistance to C. rodentium infection (Fig. 2) (54). Similarly, absent of Ahr has been proved to cause the loss of IL-22 + ILC3s, resulting in the decline of resistance to C. rodentium infection (55,56). In addition, in a cell-intrinsic manner, Ahr signaling pathway suppresses the expression of IL-33 receptor ST2 on ILC2s mediated by Gfi1 transcription factor, and the expression of ILC2 effector molecules IL-5, IL-13, and amphiregulin. Depletion of Ahr enhances protective immunity in the gut, for instance antihelminth as well as resistence to adult Heligmosomoides polygyrus bakeri (H. polygyrus bakeri) infection (Fig. 2) (57). However, Ahr signaling pathway can also contribute to type 2-associated gut inflammation, such as ulcerative colitis and food allergy (58,59).

SCFAs
SCFAs can be produced by bacterial fermentation of dietary fiber in the colon (60). The intestinal pathosymbiont Trichomonas can degrade diet fiber into SCFAs, and succinate therein can be recognized by GPR91 in tuft cells. Then tuft cells secrete IL-25 to activate ILC2s, and IL-13 produced by ILC2s can in turn promote proliferation of tuft cells. It is important that the loop above can help the host fight against parasitic infections (Fig. 2) (61). Moreover, Group 3 ILCs can also sense environmental signals through SCFA receptor Ffar2 (GPR43). In murine, not only C. rodentium infection but also DSS-induced inflamma tion, Ffar2 agonists or propionate differentially activate AKT or ERK signaling pathways, and then increase ILC3-derived IL-22 through the AKT and STAT3 axis to confer protec tion against infection (Fig. 2) (62). In mice with hepatocellular carcinoma (HCC), gut microbiota-derived SCFAs, especially acetate, reduce SOX13 expression by inhibiting the activity of histone deacetylases. Thereby, it decreases IL-17A production by hepatic ILC3s, inhibiting tumor growth and improving prognosis (Fig. 2) (63).

Retinoic acid
Retonic acid (RA), a metabolite of vitamin A, is an important dietary and nutritional signal in humans. RA signaling pathway prevents and protects against colitis induced by DSS or C. rodentium infection through promoting IL-22 production by ILC3s to form an early immune response (Fig. 2)  Under a vitamin A deficient environment, severely reduced IL-22-producing ILC3s lead to impaired immunity to acute bacterial infections. However, the dramatic expansion of IL-13-producing ILC2s result in increased resistance to nematode infections (Fig. 2) (66). In addition to diet-derived RA, gut commensal bacteria such as segmented filamen tous bacteria (SFB) can produce RA in the gut by expressing aldehyde dehydrogenase Minireview mBio enzymes. This bacteria-derived RA can initiate innate immunity possibly mediated by ILCs, and promote early protection against enteric C. rodentium infection (67).

Ketogenic diet
Ketogenic diet (KD) first appeared to simulate fasting and is metabolically characterized by a low carbohydrate ratio and a high fat ratio (68). Recently, it has been shown that KD can reduce RORγt + CD3 − Group 3 ILCs and the production of related inflamma tory cytokines including IL-17α, IL-18, IL-22, and Ccl4 in DSS induced colitis, protecting intestinal barrier function (Fig. 2). Notably, it has been demonstrated by fecal transplant that KD can mediate the regulation of ILC3s and the protection from colitis, depending on the modification of gut microbiota (69). Moreover, in ILC2-driven type 2 airway inflammation induced by Alternaria alternata (A. alternata), β-hydroxybutyrate (BHB) reduces the proliferation of ILC2s, type 2 cytokine responses, and immunopathology by suppressing IL-2-producing mast cells. BHB inhibits mast cell functions partly through activation of GPR109A, with similar effect found in KD and 1,3-butanediol (Fig. 2) (70).

B cells secreting IgA
Although ILC2s have not been shown to directly targeted regulate commensals, previous studies have demonstrated that ILC2s can regulate the composition of microbiota indirectly by influencing IgA secretion by B cells (20,71). In mouse model of Helicobacter pylori (H. pylori) infection, H. pylori enhances the induction of IL-7 and IL-33 production in the stomach, and then triggers the activation and proliferation of ILC2s. IL-5 producted by ILC2s leads to the differentiation of B cells into IgA-secreting plasma cells, and then IgA envelops the pathogen H. pylori and plays a role in resistence to infection (Fig. 3) (72).

Minireview mBio
Similarly, in the contrast of SPF and GF mice, it can be found that commensal bacteria such as Bacteroidaceae family S24-7 can active ILC2s by increasing induction of gastric IL-7 production, and then upregulate IgA secretion by B cells. This process is important for gastric immune maturation, and prevents oral pathogen infection to maintain gastric homeostasis (73). Once loss the induction of commensals, such as vancomycin-mediated eradication of Actinobacteria and Bacteroidetes, it can result in a significant reduction of IgA, demonstrating that commensals are indispensable for the maintenance of gastric homeostasis (74). Notably, the upregulation of B cells secreting IgA by ILC2s is independ ent on the engagement of T cells, either in gastric H. pylori infection or in commensals (72,73). Moreover, accumulating evidence implicates that Group 3 ILCs are key regulators of B cells secreting IgA in the gut (46). The secretion of IgA enhances physical separation of commensals from intestinal barrier, controls the colonization balance of commensals in gut microenvironment, neutralizes potentially harmful bacterial toxins and dietary metabolites (75). In Peyer's patches, tissue-resident B cells are activated by microbiotaderived antigens, and then upregulate CCR6 and attracted to sup-epithelial dome (SED) by CCL20. In SED, B cells interact extensively with CD11b + DCs and class switch to IgA. Importantly, CD11b + DCs expressing LTβR is maintained by local supply of LTα1β2 from ILC3s (Fig. 3). Unlike ILC2s, this process depends on T cells (76).

CD4 + T cells
RORγt + ILCs express major histocompatibility complex class Ⅱ (MHCⅡ) and can process and present antigens. Under steady state, ILC3s restrict commensals' specific CD4 + T cell pathological responses in a MHCII-dependent manner to maintain intestinal homeostasis (77). Microbiota-induced IL-23, depending on phosphorylation of mTORC1 and STAT3 in NCR − ILC3s, reversibly inhibits key signals of MHCII in ILC3s. As a result, the ability that ILC3s present antigens to intestinal mucosal T cells is reduced, associating with immune tolerance (Fig. 3) (78). Moreover, MHCII + ILC3s can directly induce apoptosis of activated commensal bacteriaspecific CD4 + T cells (79,80). However, it is different in spleen that IFN-γ can induce CD4 + T cells by MHCII expression on NCR − ILC3s (78). Under the stimulation of IL-1β, peripheral ILC3s upregulate surface MHCII and express costimula tory molecules (Fig. 3) (81). In IBD, IBD-associated microbiota induce the release of TL1A from CX3CR1 + mononuclear phagocytes (MNPs), and then enhance the expression of TL1A dependent costimulatory molecule OX40L in MHCII + ILC3s. This process leads to antigenspecific T cell proliferation and pathogenic Th1 cell expansion in a model of chronic colitis (Fig. 3) (82). In a model of Helicobacter hepaticus infection, ILC3s located in mesenteric lymph nodes are able to screen and promote the differentiation of gut microbiotaspecific RORγt + Tregs, and suppress Th17 cells. Thereby, it plays a critical role in establishing immune tolerance to the gut microbiota (83). In colorectal cancer (CRC) mouse model, the interplay between ILC3s and T cells through MHCII supports the colonization of microbiota, which subsequently induces type 1 immunity. Once lacking ILC3specific MHCII, mice developed invasive CRC and resistance to anti-PD-1 immunotherapy (Fig. 3) (84). In a mouse model of house-dust-mite-induced allergic airway inflammation, antigen-presenting MHCII + ILC3s significantly limit expansion of allergenspecific CD4 + T cells and mite-associated microbes induced Th17 cells (Fig. 3) (85).

Conclusion and perspectives
It has been extensively studied that the role and regulatory mechanisms of ILCs in homeostasis, IBD, and cancer (18,86,87). However, when it comes to ILCs-microbiota interactions, few articles have made systematic summary. In this review, we cite the latest research advances to dissect the mechanisms of ILCs-microbiota interactions as far as possible.
In the first major mechanism of accessory cells, DCs are considered as the key role because most of the interactions between ILCs and microbiota rely on mediators secreted by DCs. In addition to DCs, macrophages and monocytes of both APCs and IECs of non-APCs are also involved. Moreover, some studies have found that glial cells can regulate ILC3s production of IL-22 in an MYD88-dependent manner to protect the gut against pathogenic bacteria (88,89). Of course, it remains to be investigated that the role of glial cells and other accessory cells in the interaction of ILCs with microbiota.
In the second major mechanism of metabolic pathways, it is worth thinking about whether dietary therapy can be used as an effective therapeutic manner to improve the interactions between ILCs and microbiota. In addition to KD which playing a protection and mitigation role in mouse IBD, it is worthwhile to investigate whether other metabolic pathways can affect bacterial susceptibility or disease progression through changes in diet.
In the third major mechanism of adaptive immune cells, not only IgA-secreting B cells can directly provide immune effect, but also MHCII-presented CD4 + T cells can provide immune tolerance. Furthermore, the latest study found that the ratio of Treg to Th17 in the intervening CD4 + T cells was closely related to ILCs and microbiota (83). Therefore, the selective role of ILCs on gut microbiotaspecific Tregs and other CD4 + T cell subsets may be focus of the following study.
Finally, in addition to the three major mechanisms of ILC-microbiota interactions above, it remains to be explored and refined. We look forward to establishing a complete ILC-microbiota interaction framework that will contribute to human health.

ACKNOWLEDGMENTS
This research was supported by the National Natural Science Foundation of China (82102408), Project funded by China Postdoctoral Science Foundation (2022M712681) and Xuzhou Medical University Excellent Talent Introduction Project (D2019030).
The authors declare no competing interests.