Modulation of innate lymphoid cells by enteric bacterial pathogens

Innate lymphoid cells (ILCs) are key regulators of tissue homeostasis, inflammation, and immunity to infections. ILCs rapidly respond to environmental cues such as cytokines, microbiota and invading pathogens which regulate their function and phenotype. Even though ILCs are rare cells, they are enriched at barrier surfaces such as the gastrointestinal (GI) tract, and they are often critical to the host’s immune response to eliminate pathogens. On the other side of host-pathogen interactions, pathogenic bacteria also have the means to modulate these immune responses. Manipulation or evasion of the immune cells is often to the pathogen’s benefit and/or to the detriment of competing microbiota. In some instances, specific bacterial virulence factors or toxins have been implicated in how the pathogen modulates immunity. In this review, we discuss the recent progress made towards understanding the role of non-cytotoxic ILCs during enteric bacterial infections, how these pathogens can modulate the immune response, and the implications these have on developing new therapies to combat infection.


Introduction
Innate lymphoid cells (ILCs) are rare lymphocytes that are primarily tissue-resident and often present in mucosal tissues. Lacking specific antigen receptors, ILCs are mainly activated by cytokines and rapidly respond to infections. ILCs represent innate counterparts of T cells as they share several phenotypic and functional features (1). ILCs can be classified into four distinct subsets: NK cells, ILC1s, ILC2s, and ILC3s based on expression of lineage-specifying transcription factors (TF), surface markers and effector functions (2,3). We will focus on ILCs as NK cells have been well-recognized and studied for decades. ILC1s express T-bet and often protect against intracellular pathogens via production of IFNg but generally are non-cytotoxic (4). ILC2s express the lineage defining TF GATA3 and produce the type 2 cytokines IL-5 and IL-13, and lesser amounts of IL-4 and IL-9. ILC2s are often key to mediating immune responses to helminths and allergens (5). ILC3 development depends on the TF RORgt and these cells produce IL-22 to maintain tissue barrier function and protect against extracellular pathogens (5,6). Further, ILCs can have cellular plasticity and can transdifferentiate to other ILC subtypes depending upon the cytokine and/or environmental milieu (3)(4)(5).
Mucosal surfaces such as those of the GI tract harbor billions of microorganisms as part of the normal microbiota, which is the first line of defense against bacterial pathogens. Nevertheless, many pathogens enter the host through mucosal surfaces causing infection. The immune system has evolved to provide resistance to pathogens while maintaining homeostasis with the microbiota. ILCs are enriched in mucosal surfaces and orchestrate the early defense against invading pathogens. Although ILC3s are cast to be the ILCs that respond to extracellular bacterial infection, ILC1s and ILC2s also can contribute to the immune responses against these pathogens. For every well-honed immune response to combat a pathogen, a host may have a less than perfect response because of many possible reasons, with a primary one being that the pathogen can fight back. Enteric bacteria produce many toxins and other effector molecules that target host cells and can cause apoptosis or interfere in signaling pathways.
Recent studies have highlighted roles of ILCs during infection and modulation of ILC functions by pathogenic bacteria. Although challenging to test the precise role of virulence factors on ILCs in in vivo models due to the requirement for virulence factors for infectivity and the limitations on the tools to specifically target ILCs and not other immune cells, we have discerned much on how ILCs respond to different gastrointestinal (GI) bacterial infections and how the pathogen may affect this response ( Figure 1). Here, we present the current state of the field on modulation of non-cytotoxic ILC functions by enteric bacterial pathogens and how it contributes to pathogenesis or protection against these pathogens.

Modulation of ILC function by bacterial pathogens
Citrobacter rodentium C. rodentium is a murine Gram-negative pathogen that is widely used to model the human pathogen enteropathogenic Escherichia coli (EPEC) (16). Like EPEC, C. rodentium secretes effectors via a type 3 secretion system (T3SS) into intestinal epithelial cells to modulate host cell processes and establish infection (17). The utility of the C. rodentium infection model extends beyond bacterial pathogenesis studies to reveal fundamental aspects of mucosal immunity (18). C. rodentium infection elicits both innate and adaptive immune Pathogenic bacteria modulate ILC function through different mechanisms. Citrobacter rodentium produces the effect EspO which indirectly contributes to activation of ILC3s and their production of secreted IL-22 (7). As an intracellular pathogen Salmonella invades ILC3s which induces their pyroptosis and reduces their numbers in the GI tract (8). Toxin B (TcdB) produced by Clostridioides difficile directly activates ILC3s to produce IL-22 (9). Helicobacter pylori activates epithelial cells to produce IL-33 in a CagA-dependent manner. This IL-33 in turns activates ILC2s in the stomach to produce IL-5 (10,11). Non-gastric Helicobacter species increase frequencies of cytokine producing ILC1s and ILC3s (12). Campylobacter infection induces conversion of ILC3s to ILC1s that can produce IFNg (13). Bacillus anthracis secretes two toxins with opposing effects on ILC3s. Lethal toxin inhibits MAPK signaling which inhibits ILC3 activation (14). In contrast, edema toxin directly activates ILC3s to produce IL-22 (15). Direct interactions are shown with solid arrows, indirect interactions are shown with dashed arrows. Created with BioRender.com. responses that are important for control and clearance of the pathogen in the initial and late phases of infection, respectively. ILC3s are critical immune cells in early control of C. rodentium (19)(20)(21)(22). Following infection, macrophages and dendritic cells (DCs) are activated by bacterial pathogen-associated molecular patterns (PAMPs) and secrete inflammatory cytokines such as IL-23 and IL-1b which activate ILC3s (23)(24)(25)(26)(27). Upon activation, ILC3s rapidly produce IL-22, a critical cytokine in control of C. rodentium infection (5,19,28). ILC3s represent a major early source of IL-22 in infection (21,29), where the cytokine upregulates antimicrobial peptides and mucin production promoting intestinal barrier resistance (19,28).
C. rodentium T3SS effectors modulate innate immune responses (18). It is not known whether C. rodentium directly interacts with ILC3s. However, a recent study showed that the T3SS effector EspO modulates IL-22 signaling (7). While the study reported infection with C. rodentium lacking EspO reduced IL-22 secretion by colonic explants and a subsequent reduction in antimicrobial peptides, no change in the frequency of IL-22 producing ILC3s or T cells was observed (7). It is not clear why a C. rodentium effector would upregulate IL-22. Modulation of IL-22 by other pathogens is known to promote colonization (30) but this is unlikely in the case of C. rodentium as IL-22 is well-described to be protective in C. rodentium infection (19). C. rodentium is also known to modulate metabolism of intestinal epithelial cells (IECs) (18). T3SS effectors, Map and EspF, cause mitochondrial disruption leading to shift in IEC metabolism to aerobic glycolysis causing increased oxygenation of the intestinal mucosa that supports C. rodentium colonization (31-33). Given the role of hypoxia in regulating ILC3 function (34), the question arises whether this increased oxygenation of the GI tract caused by C. rodentium T3SS effectors modulate ILC3. Lastly, C. rodentium infection can cause metabolic rewiring of ILC3s, enhancing their proliferation and cytokine production (35). These "trained ILC3s" show properties of "innate memory" and are better in controlling re-infection than naive ILC3s in mice (35). In summary, C. rodentium infection has been a valuable model to learn much about pathogen-ILC interactions.

Salmonella
Salmonella enterica encompasses several serovars that cause GI infection (36,37). In developing countries, typhoidal serovars, S. Typhi and S. Paratyphi, cause enteric fever. Non-typhoidal serovars such as S. enterica serovar Typhimurium (S. Typhimurium) cause food-borne gastroenteritis that are usually self-limiting but can lead to disseminated infection (36). S. Typhimurium is the most widely studied serovar and findings reviewed here pertain to this serovar. Like many other Gram-negative GI pathogens, S. Typhimurium encodes for T3SS, which in Salmonella is two separate T3SS systems that secrete an arsenal of virulence factors (37).
Interferon-g (IFNg) is a key cytokine in defense against Salmonella infections (38)(39)(40)(41). The main innate source of IFNg during Salmonella GI infection is NKp46+ T-bet+ ILCs that have decreased RORgt levels (41). The IFNg production by these ILCs during infection is driven by IL-12 while IL-23 has no significant effect (41). Although ILCs are mainly tissue-resident cells, Kastele et al. showed that Salmonella infection increased the migratory RORgt+ T-bet+ ILC population in the mesenteric lymph nodes and contributed to IFNg production (42). Further refinement in identifying the IFNg-producing cells, has shown significant IFNg production by ILC1s but not ILC3s (8). IFNg regulates mucin production during Salmonella infection and deficiency of the mucin MUC2 increases susceptibility to infection in a mouse model (39,43). In line with this, depletion of ILCs or genetic ablation of IFNg production by ILCs, results in impaired mucus production (41). Additionally, IFNg-producing ILCs also contribute to intestinal inflammation as interference with IFNg production by ILCs leads to reduced inflammation (41). These studies suggest that IFNg production by ILCs induced by Salmonella infection can be both protective as well as pathologic.
IL-22 is upregulated during Salmonella infection (44). ILC3s via production of IL-22 and lymphotoxin-a mediate in vitro fucosylation of IECs (45). In vivo ILC3s mediate fucosylation during Salmonella infection and mice deficient in intestinal fucosylation are more susceptible to infection (45). Subsequent studies have reported that IL-22 is not protective in Salmonella infection. In fact, increased IL-22 promotes Salmonella infection by improving its ability to compete with microbiota (30,46). This is attributed to its resistance to the antimicrobial peptides induced by IL-22 (30,47,48). In line with these observations, a recent study found that Salmonella induced IL-22 in ILC3s during infection which promoted infection (8). Flagellin activates antigen-presenting cells to produce IL-23, which in turn activated ILC3s. Further, Salmonella directly invade ILC3s leading to capase-1-mediated pyroptosis of ILC3s in a flagellin-independent manner. ILC3 depletion leads to less mortality and reduced disease severity in Salmonella infected mice (8). Thus, Salmonella-induced ILC3 pyroptosis is a possible host defense mechanism against Salmonella by limiting innate IL-22 production after early induction which benefits the bacteria (8). However, the exact signal(s) that induces ILC3 pyroptosis remain unknown. This example of bacterial modulation of ILC3s by regulating cell death is a bacterial defense mechanism that may prove to be relevant to other GI pathogens.

Clostridioides difficile
C. difficile colitis is the most common nosocomial GI infection occurring in patients with perturbed microbiota owing to use of broad-spectrum antibiotics (49). C. difficile strains can encode for three different toxins: toxin A (TcdA), toxin B (TcdB) and C. difficile transferase (CDT) (50). The combined action of these toxins results in disruption of host cytoskeleton in the GI epithelium and ultimately loss of epithelial integrity. Toxin-mediated inflammation can result in symptoms ranging from mild diarrhea to pseudomembranous colitis (50).
Several studies have described the roles of ILCs during C. difficile infection (CDI) (51-53). Nfil3 -/mice, which due to a lack of the transcriptional regulator nuclear factor, interleukin 3 regulated (NFIL3) have reduced numbers of ILCs owing to a developmental defect in ILC maturation, are highly susceptible to CDI (54). Further, Rag2 -/g c -/mice lacking both adaptive immune cells and ILCs have increased mortality following CDI compared to RAG-deficient, lacking only adaptative immunity cells or wild-type mice (51). IFNg is increased in Rag1 -/mice following CDI and selective loss of ILC1s or ILC1-derived IFNg leads to increased disease severity and mortality (51  (34,53,56,57). ILC2s are important players in immunity against helminth infection and allergic response in the GI tract (5). However, these cells have been recently shown to be important in bacterial infections, including C. difficile (52). IL-33, an activator of ILC2s, was found to be upregulated during CDI in cecal tissues of infected mice. IL-33 protected against CDI by activating ILC2s (52). IL-33activated ILC2s increased mucin production by goblet cells, improving epithelial barrier function as well as increasing the number of eosinophils in the colon that are protective during CDI (52,58,59).
Although, the roles of different ILC subsets have been studied in context of CDI, whether C. difficile or its toxins directly interact with ILCs remains poorly defined. Recently, we reported that TcdB directly activates ILC3s in vitro, inducing IL-22 and other effector molecules (9). The TcdB-mediated activation of ILC3s required the toxin's enzymatic activity and was in part mediated by inactivation of the small GTPase CDC42 (9). Gene expression analysis revealed that the toxin-mediated activation of ILC3s is distinct from IL-1bmediated activation (9). Given the protective effects of IL-22 during C. difficile infection, the toxin mediated activation of ILC3s seems surprising. IL-22 may shape the microbiota to favor C. difficile over other species that the pathogen competes for resources. In vivo validation of this observation and function of toxin activated-ILC3s remains to determined.

Helicobacter Spp.
Helicobacter pylori is a highly prevalent Gram-negative pathogen that infects half of the world's population (60). It causes gastritis and is the strongest risk factor for gastric adenocarcinoma (61). The major virulence factors of H. pylori are type four secretion system (T4SS) effectors, which includes cytotoxin-associated gene A product (CagA) (61). ILC2s are the predominant ILC population in the stomach and increase in number during H. pylori infection (62,63). IL-33 is increased in gastric mucosa of patients and H. pylori infected mice with strains encoding CagA. In vitro CagA induces IL-33 production by gastric epithelial cells, leading to ILC2 activation and IL-5 production (10,11). Similarly, increased IL7 mRNA levels are found in the gastric mucosa of H. pylori infected patients (64). In the stomach, ILC2 accumulation and activation is dependent on IL-7 and IL-33, respectively (63). ILC2s coordinate with B cells to produce H. pylori-specific IgA which can coat bacteria in the stomach in a mouse model (63).
Non-gastric Helicobacter species, such as H. apodemus and H. typhlonius, have different effects on ILCs compared to H. pylori (12). In mice lacking T and B cells, colonization of the GI tract with these two Helicobacter species causes activation of ILCs with increased frequencies of IL-22-and/or IL-17-producing ILC3s and IFNgproducing ILC1s, resulting in GI inflammation. However, infection reduces ILC3 number, particularly T-bet-expressing ILC3s (12). How these Helicobacter species cause reduction in ILC3 numbers remains unknown. The reduction in ILC3s was only observed in mice lacking adaptive immune cells and not wild-type mice suggesting that adaptive immunity can sustain ILC3s during Helicobacter colonization.

Campylobacter
Campylobacter species, mainly C. jejuni and C. coli, are the causative agents of one type of food-borne gastroenteritis (65). Although these Gram-negative infections are mostly self-limiting, long-term immune related intestinal dysfunction has been associated with Campylobacter infections (66,67). Several studies have examined the role of ILCs during Campylobacter infection, using a mouse model with a predisposition to intestinal inflammation due to a lack of IL-10 signaling (13,68,69). IL-23, a primary activator of ILC3s, is key driver of inflammation during Campylobacter infection (69). Jing et al. found that IL-23 regulated production of IFNg, IL-17 and IL-22 by ILC1s and ILC3s during Campylobacter infection (69). Consistent with previous studies, IFN-g and IL-17 promoted inflammation while IL-22 was found to be dispensable for inflammation during Campylobacter infection (69). This contrasts with another study in which IL-22 was protective in a mouse model of Campylobacter where IL-10 signaling was intact (70).
Campylobacter infection induces production of IFNg, IL-17 and IL-22 from both ILCs and T cell subsets in mice (68). Similar observations were reported using an ex vivo human gut model of infection (71). Both IFNg and IL-17 were found to promote intestinal inflammation during Campylobacter infection (68). Two recent studies found that ILC-derived IFNg contributes to intestinal pathology during infection (13,69). While ILC1s represent a major source of IFNg, ILC3s can also convert into IFN-g-producing ILC1s by expressing the transcription factor T-bet (5,41). Interestingly, Campylobacter infection induces conversion of ILC3s to ILC1s and these ex-ILC3s in turn produce IFNg that promotes intestinal inflammation (13). IL-12 and IL-23 have been shown to facilitate in vitro ILC3-ILC1 plasticity (72) and both cytokines are elevated during Campylobacter infection (13,68). One may speculate that dysregulation of IL-12 and IL-23 levels during Campylobacter infection promotes ILC3 to ILC1 conversion. However, the mechanism by which Campylobacter regulates ILC3 to ILC1 conversion has not yet been elucidated. Campylobacter spp. produce many toxins that modulate host cells, with the best studied being cytolethal distending toxin (CDT) from C. jejuni, that may modulate ILC function. Together, these studies show that ILCs are a major player in inducing intestinal inflammation and promote pathology during Campylobacter infection.

Bacillus anthracis
B. anthracis is the causative agent of anthrax, a rare but deadly infection of the lungs, GI tract or skin (73). The major virulence factors of the Gram-positive B. anthracis are lethal toxin (LT) and edema toxin (ET) (73). LT or ET can both suppress immune function of both innate and adaptive immune cells (74)(75)(76). In line with the immunosuppressive action of anthrax toxins, LT can suppress ILC3 activation both in vitro and in vivo in a mouse model (14). LT, a zinc metalloprotease, cleaves mitogen activated protein kinases perturbing cell signaling (77). LT reduced IL-22 production by IL-23-activated ILC3s by inactivating MAPK signaling (14). In contrast, ET as an adenyl cyclase, elevates intracellular cAMP levels in ILC3s leading to their in vitro activation (15). How LT, ET and ILC3s interact in vivo remains to be determined. It is likely that LTmediated suppression may overcome ET-mediated activation of ILC3s as a study found suppression of the mRNA encoding IL-22induced antimicrobial peptides, REG3b and REG3g, in the colons of B. anthracis infected mice (75). ET effects can be dose-dependent, and the toxin can suppress T cell proliferation or promote Th17 differentiation (76, 78), which may have similar implications on related ILC3s. Thus, it is also possible that ET may exert suppressive action on ILC3s in vivo. Although not a focus of our review and has yet to be studied, these toxins likely modulate ILCs in the lungs during respiratory disease. Overall, B. anthracis and its two major toxins modulate ILC3 function.

Crosstalk between ILCs and the adaptive immune system
Although ILCs are mainly tissue-resident cells residing at mucosal surfaces, they are also found in the peripheral blood, bone marrow, and primary and secondary lymphoid organs (79-81). Recently, migratory ILCs have been reported in context of infection (42). Their location in lymphoid organs where the adaptive immune response is initiated suggests a role for ILCs in shaping adaptive immunity. Indeed, several studies have underscored the role of ILCs in regulation of adaptive immunity (reviewed in (81, 82)). ILCs can directly or indirectly influence both T and B cell-mediated responses (82, 83). As one of the first responders to pathogens, ILCs influence the cytokine milieu and thus also influence adaptive immune response (82, 83). ILC2s and ILC3s can also directly interact with T cells by expressing MHCII and acting as antigen presenting cells (APCs) (84)(85)(86). ILCs can also express co-stimulatory or auxiliary molecules such as OX40L, ICOS/ICOSL, and PD-L1 (the latter two reported only for ILC2s) which facilitates a direct interaction with T cells thus modulating the adaptive immune response (87-91). Conversely, the adaptive immune system can also regulate ILC function. For example, T cellderived IFNg directly limits ILC2 function (92,93) and regulatory T cells can control ILC2 and ILC3 responses (94)(95)(96). T cells can also enhance ILC2 responses via direct interaction with ILC2s mediated by MHCII or other auxiliary molecules listed above (86,97). Thus, crosstalk between ILCs and adaptive immune cells shape adaptive immunity as well ILC function.
In the healthy GI tract, ILC3s regulate the immune response to the microbiota by regulating T and B cell responses. ILC3s acting as APCs interact with T cells resulting in clonal deletion of microbiota-specific T cells (85). Similarly, ILC3s can act as APCs to interact with T helper follicular cells to regulate the immunoglobulin A (IgA) response to the microbiota (98). ILC3s can also indirectly support IgA production via lymphotoxin signaling in the intestine (99). Ablation of lymphotoxin a signaling in ILC3s results in inhibition of IgA production and composition changes of the GI microbiota (99). Studies on ILC crosstalk during enteric bacterial infection are very limited. As discussed above, ILC3s are important in defense against enteric pathogens such as C. rodentium (21). Several studies have dissected the spatiotemporal interplay of ILCs and the adaptive immune response during C. rodentium infection with ILCs being critical during the early stage of infection and the adaptive response acting later to clear infection (100-102). IL-22, the signature cytokine of ILC3s, regulates the organization and maintenance of colonic lymphoid structures during C. rodentium infection by acting downstream of lymphotoxin signaling (103). Colonic lymphoid structures contain T and B cells and thus are sites of initiation of the adaptive immune response. A recent study showed that Salmonella infection induced migration of intestinal RORgt + T-bet + ILCs to mesenteric lymph nodes which contributed to the protective IFNg response (42). Whether these migratory ILCs contribute to the overall adaptive immune response remains unknown. During colonization by non-gastric Helicobacter species, the adaptive immune system sustains ILC3 number (12). With progress in our understanding of ILC biology, the interplay of ILCs and adaptive immune system has become apparent during healthy, steady-state conditions. However, more studies are needed to uncover the specific nature of ILC and adaptive immune system interactions during enteric infections which will facilitate development of better therapeutics.

Targeting ILCs for treatment of enteric bacterial infections
The standard of care for bacterial GI diseases is antibiotics. However, with the ever-growing rise of antibiotic resistance, this treatment may fail for some patients. Targeting the host immune response, or in combination therapy with antibiotics, has the potential to become an effective treatment to combat GI infections. Although ILCs are rare immune cells, they are critical regulators of tissue homeostasis, inflammation, and immunity against GI infections (5). Increased understanding of ILC biology in recent years has resulted in development of new or co-opting T cell-targeted therapies to therapeutically target ILCs as well (104). Existing ILC targeting strategies for therapeutics include administration of cytokines, adoptive transfer of ILCs, antibodies against ILC-related cytokines, ILC depletion, modulation of ILC plasticity, migration and function, and microbiota manipulation (104). These take advantage of pathways that were developed based on T cell biology. Development of treatments that precisely target ILCs and not T cells is a challenge for the field. Furthermore, manipulation of oxygen levels in the GI tract may be an effective means to counteract a pathogen. Thus, targeting ILCs is increasingly considered for clinical therapy as evident by several preclinical or clinical studies on ILC-targeting therapies (104,105). In fact, several biological therapies targeting ILCs are already approved, and multiple others are in development for Crohn's disease, a chronic inflammatory disease of GI tract (104, 106).
As discussed, enteric bacterial pathogens often modulate ILC function contributing to a pathologic or defensive outcome. Pathogen-induced cytokine production by ILCs can be to the pathogen's advantage or contribute to protection of the host. Further, ILC plasticity and migration can also be modulated by enteric bacterial pathogens. Hence, therapeutics targeting ILC biology are an attractive novel option for the treatment of GI infections. Existing ILC-targeting biologic therapeutics could be adapted to treat enteric bacterial infections. Although studies have started to highlight the roles of ILCs in the context of different enteric bacterial infections, more studies on defining the precise roles of ILCs during infection, especially in humans, are needed before existing ILC-targeting strategies can be adapted to treating enteric infections. Such studies will provide a comprehensive understanding of ILC biology in context of infection leading to adaptation of existing or development of new therapeutic strategies targeting ILCs for enteric bacterial infections.

Conclusions
Despite their relatively recent discovery, ILCs have emerged as key players in barrier resistance to prevent breach by pathogens. ILCs are among the first immune cells that these pathogens encounter. ILCs during enteric bacterial infections can have both protective and pathogenic roles depending on the pathogen and context of infection. Enteric pathogens have evolved strategies to evade host immune responses to help establish their infection in the GI tract. In only a few interactions have we identified the specific bacterial virulence or effector molecules that manipulate specific signaling pathways in ILC3s. Future studies need to examine these different pathogens in the shared and novel ways they can modulate ILC responses. Further, many of these findings are established from animal infection models. There is a need for validation in context of human infection to increase the translation of these basic research studies to better patient outcomes. A better understanding of ILC biology in the context of infection has high potential to lead to ILCtargeted therapies for enteric bacterial infections.

Author contributions
PS contributed to the conceptualization and wrote the first draft. LZ contributed to the conceptualization, reviewed and edited the article, and generated the figure. All authors contributed to the article and approved the submitted version.

Funding
Research in the Zenewicz laboratory is supported in part by the National Institute of General Medical Sciences of the National Institutes of Health (P20GM134973) and in part by the National Cancer Institute Cancer Center Support Grant P30CA225520 awarded to the Stephenson Cancer Center.

Conflict of interest
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Publisher's note
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.