Expansion of the Innate Lymphocyte Family: Discovery of IL-22–Producing ILC3s

Innate lymphoid cells (ILCs) compose a heterogeneous population in the immune system that lacks Ag-specific receptors (1–6). It has been less than two decades since ILCs were first discovered. Previously, NK and lymphoid tissue inducer (LTi) cells were considered the only innate lymphocytes (7, 8). However, since the identification of ILCs by several different groups, our appreciation of these subsets has deepened. The investigation of ILCs has impacted our knowledge of immunity in tissue homeostasis and in many different diseases. In this Pillars of Immunology commentary, we underline the fundamental discovery made by Cella et al. (9) in identifying a novel lymphoid cell type, IL-22–producing NK cells—so called NK-22 cells—in MALT, which were later renamed as group 3 ILCs (ILC3s).

During the early days of their discovery, numerous names were being used to describe the same subsets of ILCs, which created significant confusion. For example, IL-22–producing ILCs were called NK-22 cells, NKR-LTi cells, NCR22 cells, and ILC22s (9–12). The development of ILCs shares several features with NK and LTi cells, including their dependency on the common cytokine receptor γ-chain and the transcriptional repressor inhibitor of DNA binding 2 (ID2) (13, 14), as well as IL-7 signaling (1, 11, 15). However, cytokine production patterns were distinct in the ILC subsets, similar to what was shown in different Th cell subsets. Therefore, a nomenclature to categorize ILCs into three ILC subsets was introduced in an effort to minimize confusion (1). Current ILC nomenclature mirrors that of T cells: ILC1s produce IFN-γ, ILC2s produce IL-5 and IL-13, and ILC3s produce IL-17 and/or IL-22. With an improved understanding of ILC development during the last two decades, ILCs are now classified into five subsets, that is, NK cells, ILC1s, ILC2s, ILC3s, and LTi cells (6).

Before 2009, the field recognized two subsets of innate lymphocytes, NK and LTi cells, which are now considered subsets of ILCs. In the classical view, NK cells were further divided into two subsets: one that is specialized in lysis of infected cells through perforin and granzyme, and another that is capable of producing IFN-γ and other cytokines to prime immune cells of various types, thus promoting inflammation (16). Given these functions, most studies focused on NK cells in peripheral blood and secondary lymphoid tissues in humans. It was not until publication of this highlighted Pillars of Immunology article, together with a series of other reports, that attention shifted to innate lymphocytes in mucosal tissues, such as gut (9, 17, 18). Through analysis of human tonsils and mouse Peyer’s patches, Cella et al. (9) discovered that ILC3s localized at mucosal sites that express one of the natural cytotoxicity receptors, NKp44 (or NKp46 in mice), which is not expressed on blood NK cells. Cella et al. (9) went on to show that the NKp44⁺ ILC3s in MALT produce IL-22, IL-26, and LIF. IL-23 and TLR-activated monocytes triggered the secretion of IL-22, but not IL-17, by NKp44⁺ ILC3s. They also showed that ILC3-derived cytokines stimulate epithelial cells to proliferate and secrete IL-10. In mice, this subset was identified not only in Peyer’s patches, but also in the small intestine lamina propria during Citrobacter rodentium infection.

The presence of IL-22– and IL-17–producing ILC3s was originally described in tonsils and the gut (9, 19). A critical early finding was that this subset of ILCs can mediate colitis in the absence of T cells (19). Although ILC3s share transcription factors and many functions with Th17 cells, their responses to external stimuli are different. Both ILC3s and Th17 cells are critical sources of effector molecules that control microbial invasion at mucosal tissues. However, microbial stimuli are dispensable for ILC3s, as they can develop even in germ-free mice (20, 21). In contrast, Th17 cell differentiation requires stimulation by microbiota (22, 23). Moreover, IL-22 production by ILC3s is strictly dependent on IL-23 (9), whereas exposure to IL-6 is required for Th17 polarization to IL-22 secretion (24). Mice that lack ILC3s but have T cells have only a transient defect in controlling C. rodentium infection (25), whereas mice with IL-22–deficient Th cells cannot clear this gut pathogen (24). Although Th17 cells can override ILC3 functions as adaptive immunity kicks in, the latter cells are still an essential frontline responder to combat bacterial infection at mucosal barrier sites. Furthermore, ILC3s play an even more important role in regulating mucosal barrier homeostasis.

The featured study by Cella et al. (9) paved the way for further investigation into ILC plasticity based on how ILCs react to changes in the microenvironment (10, 26). In subsequent studies, human RORγt⁺ ILC3s were shown to convert to ILC1-like cells that are responsive to IL-12 and produce IFN-γ under in vitro culture conditions with IL-2 or IL-15, which induces expression of T-bet and IL-12Rβ2 (26, 27). The differentiation of ILC3s to ILC1s was also observed in patients with Crohn’s disease (27, 28). In mice, ILC3 plasticity was demonstrated in vivo using fate mapping techniques, which revealed that NKp46⁺ ILC3s have the ability to convert to IFN-γ⁺NK1.1⁺ ILC1s (known as ex-ILC3s) (10, 26, 27, 29). The mechanisms of ILC3 plasticity are now known to involve downregulation of RORγt, as well as upregulation of T-bet and Notch signaling, for converting ILC3s to ILC1s (10). These findings were supported by other studies showing that the conversion is not observed in Tbx21^−/− mice (30–32) and mice deficient in Notch2 or RBP-Jκ, the Notch downstream signaling molecule (33). IFN-γ–producing ex-ILC3s responded to IL-12 and IL-23, exacerbated colitis, and controlled Salmonella infection-caused inflammation in mice (10, 19, 25, 26, 31, 34). Whether ILC3 plasticity is required for changing their primary functions from homeostatic to pathogenic, similar to what has been observed with Th17 cells, requires further investigation.

In summary, the work by Cella et al. (9) was the first to describe what is now known as ILC3s, and it laid the foundation for our understanding of ILC3s in the context of their function, development, and plasticity. Since its initial publication in 2009, the article has influenced diverse research areas in ILC biology and immunology. For example, we now better understand ILC regulation by dietary nutrients, such as the role of retinoic acid in ILC3 functions. Recent studies showed that ILC3s express receptors for neurotrophic factor RET (35) and neuropeptide VIP, suggesting that neuroimmune interactions through ILC3s regulate inflammation and infection (36, 37). The advance of sequencing techniques also has contributed to our expanding knowledge of ILC heterogeneity. Although it is still challenging to translate these findings into therapeutic applications, further investigation of ILC3s might provide essential tools for understanding the basic mechanisms of tissue homeostasis and could provide strategies for treating diverse diseases, especially at the mucosal sites.