Abstract

Introduction

The last years have witnessed an unprecedented change in our understanding of innate lymphocyte lineages. It was previously believed, that innate lymphocytes were represented by a single lymphoid lineage, namely natural killer (NK) cells, that, in many aspects, resemble cytotoxic T cells. However, it has become apparent that additional innate lymphocyte subsets exist that use transcriptional programs and display functions distinct from conventional NK (cNK) cells. All innate lymphocytes including cNK cells are now referred to as ILC. In addition to cNK cells, three additional groups of ILC are now being discriminated, ILC1, ILC2, and ILC3. Strikingly, the transcriptional and effector programs of the various ILC populations resemble those of T helper subsets, suggesting that the underlying transcriptional circuitry is evolutionarily more ancient than previously appreciated (Tanriver and Diefenbach, 2014). Here, we will discuss our current view of developmental and transcriptional programs common to all ILC lineages and those required for specification of distinct ILC populations. These recent data provide a framework for our current view of two principal ILC lineages, cytotoxic or killer ILC (i.e., cNK cells) and helper-like ILC (i.e., ILC1, ILC2, ILC3) (Figure 1). We will put a focus on recent progress in dissecting the ILC1 lineage and on common transcriptional programs controlling ILC specification.

Open in a separate window

Figure 1

Refined lineage map for the development of ILC lineages

All lymphoid lineages are the progeny of the common lymphoid progenitor (CLP). After the branchpoint with the B and T lineages an ILC-restricted progenitor may exist (CILP). Downstream of the CILP, two main ILC lineages can be discriminated, killer ILC and helper-like ILC. Killer ILC are represented by cNK cells and helper-like ILC are composed of the various cytokine-producing ILC subsets (i.e., ILC1, ILC2, ILC3). While helper-like ILC express IL-7Rα and require GATA-3 for differentiation, killer ILC do not express IL-7Rα and are normally represented in GATA-3-deficient mice. All helper-like ILC (but not killer ILC) differentiate from the Id2⁺ CHILP. A PLZF⁺ CHILP population has been identified that has more restricted differentiation potential. A precursor/progeny relation between PLZF⁻ and PLZF⁺ CHILP needs to be determined.

CLP: common lymphoid progenitor; CILP: common ILC progenitor; CHILP: comon helper-like ILC progenitor; NKP: cNK-restricted progenitor

Common developmental programs for ILC fate

Given that ILC and adaptive immune system lymphocytes all derive from the common lymphoid progenitor (CLP) (Cherrier et al., 2012; Klose et al., 2014; Moro et al., 2010; Possot et al., 2011; Yang et al., 2011), the question arose if an ILC lineage-restricted progenitor for some or all ILC populations exists that is downstream of the CLP. The existence of such common innate lymphoid progenitor (CILP) or ILC progenitor (ILCP) was predicted because mice lacking expression of the transcriptional regulator inhibitor of DNA binding 2 (Id2) lack all ILC lineages whereas B and T cell development was largely unperturbed (Cherrier et al., 2012; Moro et al., 2010; Satoh-Takayama et al., 2010; Yokota et al., 1999). Id2 is a member of the helix-loop-helix (HLH) family of proteins that have important functions in immune cell fate decisions (Kee, 2009; Murre, 2005). Id2 heterodimerizes with E proteins that function as transcriptional activators or repressors. Four mammalian E proteins have been characterized (E2A, HEB, E12 and E47) that bind to specific DNA sequences called E-box sites. While Id2 cannot bind to chromatin itself (due to the lack of a basic HLH domain), it heterodimerizes with E proteins and sequesters them away from chromatin, thereby controlling the pool of E proteins available for DNA binding (Benezra et al., 1990). Id2 is expressed in high amounts in all ILC lineages (Carotta et al., 2011; Hoyler et al., 2012) whereas naïve T cells and B cells either express only very low amounts of Id2 or are Id2-negative. Interestingly, B cell lineage specification requires the transcription factor early B cell factor 1 (EBF1) (Lin and Grosschedl, 1995) which is a potent repressor of Id2 (Treiber et al., 2010). Current concepts indicate that repression of Id2 is required to establish appropriate concentrations of E2A, which are permissive for the development of B cells (O'Riordan and Grosschedl, 1999). Deletion of EBF1 in B cell progenitors (late pro-B cells) led to de-repression of Id2 and Notch and reprogrammed committed B cell progenitors into ILC and T cells (Nechanitzky et al., 2013). These data suggest that high level Id2 expression is required to establish ILC fate. While Id2 is now recognized as an important knot within the transcriptional network establishing ILC fate, the stage-specific Id2 controlled transcriptional programs on ILC specification and/or maintenance are an important goal for future research.

Differentiation of ILC1

The general challenge when dealing with ILC1 subsets was their phenotypic resemblance (e.g., NKp46⁺, NK1.1 ⁺, T-bet⁺) making it difficult to discriminate subsets and to assign them to separate lineages. Only recently, through the use of genetic reporter systems faithfully reporting the expression of lineage-specifying transcription factors, some progress has been made in separating ILC1 populations from cNK cells. Differentiation of cNK cells depends on sequential steps controlled by the transcription factors Id2, NFIL3 and Eomes. The development and maintenance of cNK cells require IL-15 signaling through IL-2Rβ (CD122) and the cytokine receptor common γ chain (γ_c). However, thymic NK cells develop through a distinct developmental pathway that involves the transcription factor GATA-3 and signaling through the IL-7 receptor (CD127) and produce IFN-γ and granulocyte macrophage colony-stimulating factor (GM-CSF) in much higher amounts than cNK cells (Ribeiro et al., 2010; Vosshenrich et al., 2006) (Table 3). IFN-γ-producing ILC in the oral and intestinal epithelium of both humans and mice (intraepithelial ILC1) are equipped with a unique set of integrins and adhesion molecules such as VLA-1, CD160 and CD103 (in humans), manifest strong transforming growth factor beta (TGF-β) imprinting and do not entirely rely on IL-15 signaling (Fuchs et al., 2013) (Tables 1 and ​and22).

The human oral and intestinal mucosa were found to harbor yet another subset of IFN-γ-producing ILC that express the receptor for IL-7 and lack NK cell markers and lytic enzymes (Bernink et al., 2013). These cells were also called ILC1 and were proposed to arise from the conversion of RORγt⁺ ILC3 under the influence of IL-12 during inflammation. Supporting this, human ILC3 can differentiate into RORγt⁻ IFN-γ-producing cells in vitro when cultured in IL-2 (Cella et al., 2010; Hughes et al., 2010; Hughes et al., 2009). Moreover, fate-mapping experiments using Rorc-Cre mice and ILC3 transfer experiments showed that some IFN-γ-producing ILC do, indeed, originate from RORγt⁺ ILC3 (Klose et al., 2014; Vonarbourg et al., 2010). Together, these studies have suggested that a certain percentage of IFN-γ-producing ILC may stem from ILC3.

Because the development of cNK cells required Eomes and T-bet was required for functional maturation and bone marrow egress (Gordon et al., 2012; Jenne et al., 2009; Townsend et al., 2004), attention has focused on these two transcription factors. Interestingly, TRAIL⁺VLA-1(CD49a)⁺VLA-2⁻ liver NK cells are T-bet and IL-15-dependent, but Eomes-independent (Daussy et al., 2014; Sojka et al., 2014), consistent with a distinct developmental pathway. Indeed, TRAIL⁺VLA-2⁻ liver NK cells did not express Eomes but strictly required T-bet for their development (Daussy et al., 2014; Gordon et al., 2012). Uterine NK cells are both T-bet and Eomes independent (Sojka et al., 2014), perhaps reflecting redundancy of these transcription factors. This may also be the case for salivary gland NK cells, as they express high amounts of both T-bet and Eomes (Cortez et al., 2014).

NFIL3 promotes the development of bone marrow-derived NK cells from CLP in the steady state (Gascoyne et al., 2009; Kamizono et al., 2009; Male et al., 2012). Recent studies have shown that Nfil3 is required for the formation of splenic cNK cells as well as thymic NK cells, possibly by directly regulating the expression of downstream transcription factors such as Eomes and Id2 (Male et al., 2014; Seillet et al., 2014a). In contrast, liver VLA-2^–Eomes^– NK cells (Sojka et al., 2014) and salivary gland VLA-2⁺Eomes⁺ NK cells (Cortez et al., 2014) seem to develop independently of NFIL3. It should be noted, however, that the impact of NFIL3 seems to be affected by the environment and may vary due to infection, inflammation, and perhaps housing conditions. For example, while NFIL3 is required for the development of conventional NK cells in the steady-state, NFIL3 is dispensable for NK cells that express the MCMV-specific receptor Ly49H during MCMV infection (Firth et al., 2013). This variability may explain seemingly contradictory results regarding the impact of NFIL3 on the development of thymic NK cells (Crotta et al., 2014; Seillet et al., 2014a) and liver VLA- 1⁺ NK cells (Sojka et al., 2014). Therefore, dependence on NFIL3 is most likely insufficient to unequivocally define a cell lineage (Table 3).

While cNK cells develop from Id2^– CLP, intestinal ILC1 and VLA-1⁺ liver NK cells develop from Id2⁺PLZF⁺ CHILP (Constantinides et al., 2014; Klose et al., 2014). Thus, these studies establish liver VLA-1⁺ NK cells and intestinal ILC1 as a separate ILC1 lineage, distinct from cNK cells. It will be interesting to see whether uterine and salivary gland NK cells also arise from CHILP or are more related to cNK cells. In addition to originating from discrete precursors, cNK cells and ILC1 may also differ in the timing and site of development. For example, liver ILC1 may develop early on during embryogenesis from the fetal liver and may replicate in situ to maintain stable numbers.

Differentiation of ILC2

Beyond Id2, IL-7 and the cytokine receptor common γ chain (Cao et al., 1995; Moro et al., 2010; Satoh-Takayama et al., 2010; Yokota et al., 1999), differentiation of ILC2 requires the transcription factors GATA-3, RORα, and TCF-1 as well as Notch signaling (Furusawa et al., 2013; Gentek et al., 2013; Hoyler et al., 2012; Klein Wolterink et al., 2013; Mielke et al., 2013; Moro et al., 2010; Wong et al., 2012; Yang et al., 2013). GATA-3 has broader roles in ILC development as deletion of Gata3 in all hematopoietic cells using Vav-Cre (Yagi et al., 2014) or in fetal liver chimeras (Serafini et al., 2014) revealed that GATA-3 is critical for the differentiation of all helper-like ILC subsets but not for cNK cells. Deletion of the Gata3 geneat a later time point using Id2^CreERt2/+ mice resulted in the impaired differentiation of ILC2 but not ILC3 (Hoyler et al., 2012). Similarly, when Rosa26^CreERt2/+ × Gata3^flox/flox bone marrow cells were transplanted into lethally irradiated mice and 4-hydroxytamoxifen was administered into the mice 1 day after the bone marrow transplantation, differentiation of ILC2 but not of ILC3 was strongly suppressed (Furusawa et al., 2013; Yagi et al., 2014). Transgenic overexpression of GATA-3 increased ILC2 numbers concomitant with higher expression of IL-33R (Klein Wolterink et al., 2013). These results suggest that GATA-3 plays important roles in multiple steps of ILC differentiation and/or survival but high level GATA-3 expression in lineage-specified ILC is specifically required for ILC2 maintenance (Table 3). GATA-3 is also required for the differentiation of human ILC2 (Mjosberg et al., 2012). Future research needs to provide a molecular understanding of the GATA-3-directed gene expression networks in the various ILC subsets and in progenitor populations.

RORα, a member of retinoic-acid-receptor-related orphan nuclear receptor, is highly expressed in ILC2 (Moro et al., 2010; Wong et al., 2012). ILC2 differentiation was severely impaired in a natural Rora mutant (Rora^sg/sg mouse) without affecting other ILC subsets (Halim et al., 2012b; Wong et al., 2012), indicating that RORα is a transcription factor specifically required for ILC2 differentiation. Small numbers of ILC2 cells are present in the mesentery of Rora^sg/sg mice and those residual cells are able to produce IL-5 and IL-13 in response to IL-33 (Furusawa et al., 2013), indicating that RORα is dispensable for cytokine production induced by IL-33 signals. Unlike RORα, GATA-3 is essential for the expression of IL-5 and IL-13 in mature ILC2 (Furusawa et al., 2013; Hoyler et al., 2012; Klein Wolterink et al., 2013; Liang et al., 2012; Yagi et al., 2014).

In addition to factors involved in the differentiation of ILC2, other factors play roles in differentiation and functional maturation of ILC2. Vitamin A and its metabolites derived from food control the balance between ILC2 and ILC3 in the intestine as discussed later (Spencer et al., 2014; van de Pavert et al., 2014). Functional specification of ILC2 requires Gfi1 because the lack of the Gfi1 gene resulted in the loss of GATA-3expression and coexpression of IL-13 and IL-17 (Spooner et al., 2013). Gfi1 directly activated the expression of Il1rl1 and Il17rb genes encoding receptors for IL-33 and IL-25, respectively. On the other hand, Gfi1 suppressed the characteristics of ILC3 through the suppression of Sox4-Rorc axis required for the expression of Il17a gene (Spooner et al., 2013). Interestingly, work with human ILC3 had indicated that they can become producers of IL-5 and IL-13 when stimulated with TLR2 ligands (Crellin et al., 2010). These results suggest potential plasticity between ILC2 and ILC3 under certain environmental conditions, which remains to be elucidated. Other outstanding questions are whether class II MHC expression is limited to ILC2 (Oliphant et al., 2014) and ILC3 (Hepworth et al., 2013) and how such class II MHC expression fits with our current understanding of transcriptional regulation of ILC and of ILC being part of the lymphoid branch of hematopoietic cells.

Differentiation of ILC3

Group 3 ILC contain various populations of RORγt-expressing ILC that play important roles in lymphoid organogenesis and that are a substantial innate source of “type 17 cytokines” involved in protecting mucosal barriers against extracellular bacterial and fungal infections (Diefenbach, 2013). The common denominator of the various ILC3 populations is their dependency on the transcription factor RORγt which specifies and defines the ILC3 lineage. RORγt expression has not been reported in other ILC lineages. Rorc(γt), an alternative transcript of the Rorc gene, was initially found to be expressed in all CD4⁺CD8⁺ thymocytes and overexpression of RORγt protected such DP thymocytes against activation-induced cell death (He et al., 1998). Surprisingly, genetic deletion of the Rorc gene had a mild T cell phenotype but these mice lacked all peripheral lymph nodes and Peyer's patches (Eberl et al., 2004; Kurebayashi et al., 2000; Sun et al., 2000). Lack of lymphoid organ development in RORγt-deficient animals was linked to the absence of LTi cells. LTi cells were first identified as innate lymphocytes in lymph node and Peyer's patch anlagen of E13.5-15.5 mice (Adachi et al., 1998; Adachi et al., 1997; Mebius et al., 1997; Yoshida et al., 2002). Interestingly, a phenotypically related and RORγt-dependent lymphoid population was found to be required in the small intestine to coordinate formation of dispersed lymphoid clusters found in the lamina propria of the intestine known as cryptopatches and isolated lymphoid follicles (Eberl, 2005; Eberl and Littman, 2004; Hamada et al., 2002; Kanamori et al., 1996). ILC3 can be differentiated from Lin⁻IL-7Rα⁺Flt3⁻γ₄β₇⁺ fetal liver progenitors that sequentially acquired Id2, CXCR6, and RORγt expression (Cherrier et al., 2012; Possot et al., 2011).

While a role of LTi cells as organizers of lymphoid organ development is well documented, only recently it became clear that they are substantially represented in the intestinal lamina propria of adult mice where they produce copious amounts of IL-22 and IL-17A and contribute to immunity to infections and the pathogenesis of inflammatory diseases (Buonocore et al., 2010; Cella et al., 2009; Cupedo et al., 2009; Luci et al., 2009; Powell et al., 2012; Sanos et al., 2009; Satoh-Takayama et al., 2008; Takatori et al., 2009; Vonarbourg et al., 2010). More recently, it was shown that ILC3 also produce GM-CSF, which established important crosstalk with GM-CSF-responsive mononuclear phagocytes that regulate oral tolerance (Mortha et al., 2014). These intestinal “LTi-like cells” expressed surface receptors often found on NK cells such as NKp46, NKp44, and CD56, raising questions about the relationship between NKp46⁺ ILC3 and NK cells. Compelling evidence from fate-map studies has been provided showing that NKp46⁺ ILC3 develop independent of the NK cell lineage and are the progeny of NKp46⁻ ILC3 (Crellin et al., 2010; Klose et al., 2014; Satoh-Takayama et al., 2010; Sawa et al., 2010; Vonarbourg et al., 2010).

The existence of NKp46⁻ and NKp46⁺ ILC3 raised the question if these are independent ILC lineages or rather differentiation or activation products of the same lineage. Two reports have found that ILC3 can be subdivided into CCR6⁺IL-7Rα^highKit^high and CCR6^−/lowIL-7Rα^lowKit^low cells that may constitute separate ILC3 lineages, both of which require RORγt for development (Klose et al., 2013; Sawa et al., 2010). It has now become clear that CCR6⁺ and CCR6^−/low ILC3 follow distinct transcriptional programs. CCR6^high ILC3 seed the intestine early during fetal development and seem to be long-lived and slowly cycling cells (Eberl and Littman, 2004; Hanash et al., 2012; Sawa et al., 2010). In contrast, CCR6^−/low ILC3 account only for a small fraction of ILC3 at birth but vigorously expand during the first four weeks after birth (Klose et al., 2013). Expansion of CCR6^−/low ILC3 was dependent on the expression of the ligand activated transcription factor aryl hydrocarbon receptor (AhR). Mice genetically lacking AhR expression in all cells or in ILC3 had a largely unaltered compartment of CCR6⁺ ILC3 whereas CCR6^−/low ILC3 were diminished (Kiss et al., 2011; Klose et al., 2013; Lee et al., 2012; Qiu et al., 2011). AhR-independent development of CCR6⁺ ILC3 with LTi function was also documented by the fact that fetally forming lymphoid organs (i.e., lymph nodes, Peyer's patches) were normal in AhR-deficient mice. Curiously, cryptopatches and ILF were absent in AhR-deficient mice, suggesting that CCR6^−/low ILC3 have either non-redundant LTi function or that AhR regulates crucial LTi effector molecules in CCR6⁺ ILC3. The reduced maintenance of CCR6^−/low ILC3 in AhR-deficient mice may be explained by a role of AhR in regulating expression of the receptor tyrosine kinase c-Kit because AhR directly bound to the c-Kit promoter and controlled transcription of the Kit gene (Kiss et al., 2011). AhR is a sensor of various small molecules which confer transcriptional activity to the AhR (McIntosh et al., 2010). Interestingly, nutrient-derived AhR ligands such as glucobrassicins contained in vegetables of the Brassicaceae family (e.g., broccoli, Brussel sprouts) were recently identified to have an important role in driving postnatal expansion of CCR6^−/low ILC3 (Kiss et al., 2011) and intraepithelial γδ T cells (Li et al., 2011) demonstrating a broad role of plant-derived phytochemicals in controlling development and maintenance of immune system components at barrier surfaces.

The majority of NKp46⁻CCR6^−/low ILC3 co-expressed T-bet (Klose et al., 2013; Rankin et al., 2013; Sciume et al., 2012). While plastic (‘ex-RORγt⁺’) ILC3 can be found in humans (Bernink et al., 2013), it remains to be seen if T-bet-expressing CCR6^−/low ILC3 exist in humans. Co-expression of T-bet and RORγt was a remarkable finding because it was believed that lineage-specifying transcription factors such as T-bet and RORγt may be mutually exclusive in expression. T-bet was not required for the differentiation or maintenance of NKp46⁻CCR6^−/low ILC3 but controlled differentiation of NKp46⁻CCR6^−/low ILC3 to NKp46⁺CCR6^−/low ILC3 (Rankin et al., 2013). In addition, an increasing T-bet gradient in CCR6^−/low ILC3 directed a transcriptional program that allowed for functional and phenotypic plasticity (Figure 1). For example, T-bet controlled IFN-γ production in NKp46⁺CCR6^−/low ILC3 that contributed to early immunity to Salmonella enterica infection. When inappropriately stimulated, NKp46⁺CCR6^−/low ILC3 contributed to intestinal inflammatory diseases in mice and humans (Bernink et al., 2013; Vonarbourg et al., 2010). T-bet expression in CCR6^−/low ILC3 was independent of IL-12 signaling but was promoted by IL-23 (Klose et al., 2013), revealing an interesting parallel to the signals driving plasticity of Th17 cells in a mouse model of multiple sclerosis (Hirota et al., 2011).

Recent data revealed an important role for maternal vitamin A metabolites in the differentiation and maintenance of fetal CCR6⁺ ILC3 with LTi function. Mice deficient of retinoic acid receptor (RAR)α expression in hematopoietic cells had reduced numbers of fetal CCR6⁺ ILC3 due to inefficient upregulation of the Rorc(γt) transcript of the Rorc gene (van de Pavert et al., 2014). Reduced ILC3 numbers led to the formation of substantially smaller lymphoid organs, ultimately reducing the fitness of adult mice to virus infections (van de Pavert et al., 2014). Similar findings were reported for adult mice under conditions of vitamin A malnutrition that presented with reduced numbers of IL-22-producing ILC3 (Spencer et al., 2014). Vitamin A depravation of adult mice led to impaired immunity to attaching-and-effacing infections with C. rodentium. While ILC3 numbers and function were reduced under conditions of retinoic acid deficiency, ILC2 were increased in a compensatory manner resulting in enhanced immunity to nematode infection (Spencer et al., 2014). These data and the data from Ahr-deficient mice establish an important trajectory by which nutrients instruct differentiation and maintenance of mucosal ILC subsets.

Conclusions and perspectives

The discovery of new ILC lineages has entirely redefined our hematopoietic lineage maps (Figure 1). It is now clear that the diversity of ILC is much higher than previously appreciated. The description of ILC-restricted progenitors (i.e., CHILP) has now opened opportunities to examine the cues that drive differentiation of the various helper-like ILC populations. While the similarities between helper-like ILC and T helper cell subsets are striking, it is already clear that the cytokine signals required for T helper fate decisions are dispensable for the differentiation into the various helper-like ILC populations (Guo et al., 2014; Klose et al., 2013; Liang et al., 2012). Thus, one of the important future goals will be to identify the signals that drive differentiation of the CHILP. While a naïve CD4 T cell can be differentiated into any Th effector state depending on the cytokine environment provided, it should be considered that ILC fate may be determined through a timed release of progenitors during ontogeny into an environment that allows for the preferential differentiation into a certain ILC fate. Precedent for such temporal organization of lymphoid fate decisions are the various waves of distinct γδ T cell subsets leaving the thymus before birth. Such future lines of research may also reveal an evolutionary perspective on how such transcriptional programs controlling lymphoid fate have emerged. In addition, although ILC2 seem a relatively stable subset compared to ILC1 and ILC3, there is still a possibility of plasticity between ILC2 and other ILC subsets (Crellin et al., 2010; Spooner et al., 2013). Given the emerging roles of transcriptional and functional plasticity of lymphocyte subsets for the pathogenesis of inflammatory diseases (Hirota et al., 2011; Vonarbourg et al., 2010), the molecular cues stabilizing ILC fates and those driving plasticity should be identified because they may reveal therapeutic targets for such debilitating diseases. Thus, future research needs to identify the signals and the spatiotemporal organization of helper-like ILC development.

Research into ILC lineages has also revealed that the population of NKp46⁺ cells, previously addressed as “NK cells”, is more complex than appreciated. For example, the population of NKp46⁺ cells in the lamina propria of the intestine is composed of at least three distinct ILC populations (Klose et al., 2014). The unique contributions of these various NKp46⁺ ILC subsets to immune responses remain to be determined. In that context, it is interesting to note that cNK cells may be a lineage distinct from helper-like ILC. It will be informative to position the branching point of the cytotoxic and helper-like ILC lineages on hematopoietic lineage maps. This will be complemented by analyzing the core transcriptional program determining ILC fate. Id2 has already been discovered as one central hub for ILC fate and NFIL3 and TOX may be other candidates. The DNA binding factor TOX is an interesting candidate because Tox^−/− mice lack NK cell and ILC3 development (Aliahmad et al., 2012; Geiger et al., 2014).

Finally, lymphoid cells use immune recognition receptors for development and to interact with infected or damaged cells. It remains unclear if immunoreceptors expressed by ILC are required for their function and/or development. Mice lacking the Ncr1 gene and consequently not expressing NKp46 showed normal development and function of ILC3 documented by their resistance to C. rodentium infection (Satoh-Takayama et al., 2009). Consecutive work documented a role for activating immunoreceptors for ILC function. Human ILC3 from tonsils and colon could be activated for TNF release when triggered by the NKp44 receptor (Glatzer et al., 2013). In mice, liver-resident ILC1 responded to engagement of various activating immunoreceptors (i.e., NKp46, NKG2D, NK1.1) with the production of TNF, IFN-γ and CCL3 (Daussy et al., 2014). Interestingly, liver-resident ILC1 with an “NK cell phenotype” display antigen-specific memory that does not require recombining receptors (O'Leary et al., 2006; Paust et al., 2010). The phenotype of such memory “NK cells” may indicate that they are an ILC1 subset. It will be highly informative to gain a more molecular understanding of how ILC discriminate between “self” and “non-self” and how previous activation may introduce a memory-like program into innate lymphocytes.

Supplementary Material

Acknowledgements

Footnotes

References