Despite their obvious similarities all newly discovered ILC populations were given different names based on differential expression of surface markers. Interestingly, this variation is limited to surface markers indicative of the activation state of cells. For example CD25 and Sca-1 are upregulated on T cells upon activation or differentiation. Thus, it is conceivable that the activation state or tissue distribution of the various innate lymphoid populations might influence the expression of surface receptors, as recently observed for nuocytes (Neill and Mckenzie, 2011). Indeed our own unpublished data suggests that ILC, which have expressed the IL-9 gene and are marked by eYFP expression (eYFP⁺ ILC), display variations of surface marker expression before and directly after papain challenge. Differences in the surface phenotype are also observed in ILC induced by intranasal delivery of either IL-25 or IL-33. While challenge with IL-33 induces ILC displaying an activated phenotype, with high expression of CD25 and Sca-1, IL-25 elicited ILC are low for those activation markers (our unpublished data). Interestingly, ILC generated by delivery of IL-33 are also more potent producers of IL-9 upon IL-2 stimulation, suggesting that their activation state could influence the ability to produce IL-9.

Another important but less explored finding is the observation that ILC may depend on the adaptive immune system for maintenance or expression of cytokines. This is best illustrated by the fact that nuocyte numbers (identified by IL-13 via GFP expression) in Rag^−/− mice are greatly reduced at day 6 but not at day 4 after Nippostrongylus brasiliensis infection in contrast to nuocytes induced in wild-type mice (Neill et al., 2010). There are two possible interpretations for this effect: Either nuocytes, while readily induced are not maintained in absence of the adaptive immune system or alternatively their characteristic cytokine expression is lost. Since the classification as nuocytes in this particular setting depended on continuous IL-13 expression monitored by GFP, loss of cytokine expression might erroneously be interpreted as loss of the whole population.

The latter assumption is supported by data from our laboratory in the context of papain mediated lung inflammation. Similar to the report from McKenzie et al. (1998) we observed reduced levels of IL-13 expression in Rag^−/− mice in the absence of the adaptive immune system. Importantly though, we did not observe a significant loss of ILC in Rag^−/− mice based on the classification as lineage⁻ Thy1.2⁺ cells. Since IL-9 promotes expression of IL-5 and IL-13 from ILC and IL-2 induces IL-9, the lack of IL-2 from adaptive immune cells (or NKT cells), and the resulting loss of IL-9 expression is one possible explanation for reduced Th2-type cytokine expression in Rag^−/− mice. In line with this assumption, delivering either exogenous IL-2 or IL-9 to Rag^−/− mice in the context of allergen challenge resulted likewise in increased levels of IL-5 and IL-13 (unpublished data). However, delivery of IL-25 to N. brasiliensis infected Rag^−/− mice results in the rapid expulsion of worms, whereas such mice are normally highly susceptible to helminth infections (Price et al., 2010). Similar results are observed if Rag^−/− IL2rg^−/− (mice which lack ILC in addition to being deficient for T, B, and NK cells) are transferred with ILC and treated with IL-25 (Price et al., 2010). This suggests that continuous exposure to IL-25 and probably IL-33 is able to circumvent the dependency of the adaptive immune system and to substitute for IL-2 presence.

The described dependency on the adaptive immune system either for cytokine expression or survival/maintenance is one of the most intriguing features of ILC. In the past, ablation of adaptive immune cells was believed to only affect adaptive sources of cytokine production. Since expression of cytokines from ILC depends on the adaptive immune system, it is conceivable that studies exploring the effect of T cell mediated cytokine expression unknowingly targeted cytokine expression from ILC simultaneously. In addition, antigen specific cytokine expression has been widely addressed by the restimulation of total splenocytes or lymphocytes with antigen in tissue culture. The presence of T cell (or NKT cell-) derived IL-2 in these cultures most likely will result in the induction of cytokine expression by ILC. This example clearly demonstrates that the cellular sources of cytokines have to be readdressed in the majority of experimental disease models.

It appears that the initiation of ILC happens around day 3–5 in an immune response, thus preceding the adaptive response (Neill et al., 2010; Price et al., 2010). However, whether ILC are needed for the induction of an adaptive type 2 immune response and/or whether they represent the critical and main source of cytokine producers during such immune responses is unknown. These questions are not trivial to address. While the initiation of the ILC response develops independent of the adaptive immune system, their maintenance requires adaptive immune cells. Here the study of Rag^−/− mice can only deliver insufficient answers.

In a model of papain induced lung inflammation the majority of Th2-type cytokine producing cells originated from ILC despite the presence of fully developed Th2 cells (Wilhelm et al., 2011). Here we attempted to address the relative contribution to the overall amount of IL-9 by transferring either CD4⁺ T cells or ILC to naïve Rag^−/−IL2rg^−/− mice, thus limiting IL-9 production to either transferred CD4⁺ T cells or ILC. Despite the shortcomings of this experimental system, it clearly established that ILC are indeed the major producers of cytokines in this particular disease model. However, the question remains if we can relate these findings to other experimental systems and it needs to be addressed if type 2 cytokines produced during helminth infections, influenza or other disease models mainly originate from T cells or ILC. In line with the assumption that ILC are the major source of type 2 cytokines, an elegant set of experiments showed that the transfer of CD4⁺ T cells from il4/il13^−/− mice to Rag^−/− mice, was sufficient to result in expulsion of N. brasiliensis from infected mice (Voehringer et al., 2006). Thus, the need for T cells was emphasized, but these could not have provided the critical source of IL-13, which is absolutely crucial for worm expulsion (Mckenzie et al., 1998; Urban et al., 1998). Further studies are needed to clearly establish the link between the adaptive immune system and the regulation of ILC in different experimental models.

Two recent studies explored the function of lung resident ILC in the context of influenza infection (Chang et al., 2011, p. 103; Monticelli, 2011, p 170). The first study explored the function of innate lymphocytes during the early phase of influenza infection and established that obstruction of the airways during flu infections seems to be dependent on IL-13 by ILC. Interestingly, a second study that focused on the late phase after infection suggested an important role for ILC in the maintenance of airway epithelial integrity, showing that ablation of ILC in Rag^−/− mice during the repair phase after influenza mediated lung infection leads to diminished lung function and impaired airway remodeling in mice. This process appeared to be independent of IL-13. However, administration of amphiregulin was able to restore airway epithelial integrity and tissue homeostasis, thus suggesting a previously unrecognized function of ILC via the expression of amphiregulin. The involvement – if any – of IL-9 in these processes remains to be addressed.

The first evidence for the presence of ILC in humans showed that NHC similar to those of the mouse are present in human fat-associated lymphoid clusters (FALC; Moro et al., 2010). Another, more recent study demonstrated that nasal polyps isolated from patients with atopic disease contain increased numbers of ILC (Mjosberg et al., 2011). These cells were characterized by the expression of the chemoattractant receptor CRTH2 (chemoattractant receptor-homologous molecule expressed on Th2 lymphocytes), were present in lung, gut, and nasal tissues and produced IL-13 in response to stimulation with combinations of IL-2 and IL-25 or IL-33. This strongly suggests that similar mechanism for the induction of ILC in mouse and human exists and highlights their potential in atopic diseases like asthma and allergies.

While the accumulation of ILC has been shown to depend on IL-25 and/or IL-33, it is not entirely clear if IL-25 and IL-33 exposure lead to their recruitment or expansion. In vitro studies suggest that combinations of IL-33 and IL-2 or IL-7 are able to expand the pool of existing ILC. However it remains unknown if ILC are recruited from other sites of the body like liver or bone marrow during an ongoing immune response. Additionally, no data is available with regard to which progenitor cells give rise to ILC.

Finally, the evolutionary origin of ILC remains an interesting as yet unsolved question. Did ILC appear before the development of the adaptive immune system, or did ILC and the adaptive immune system co-evolve as their dependency on the adaptive immune system might suggest? These questions will be hopefully addressed in the future to further help to shape our understanding of this intriguing subset of innate lymphocytes.

It appears that ILC play an important role not only in the initiation and effector phase of an immune response but also in the resolution of inflammation and the tissue repair process. Most of these functions are also attributed to T cells and T cell derived cytokines. However, it seems unlikely that such ILC functions have just evolved as mirror images of those in adaptive T cells. A consideration is the potential of adaptive T cell responses to cause immune pathology or autoimmune syndromes. Although ILC responses have in some cases been shown to have detrimental effects (Buonocore et al., 2010; Chang et al., 2011) these studies by necessity had to be done in relatively contrived experimental models with immunocompromised mice, thus not necessarily reflecting similar pathogenic actions of ILC in wild-type mice. It therefore remains a possibility that immune responses by ILC are a first line of defense coupled with regenerative potential, whereas exacerbation and disbalance toward pathology would only be encountered following the involvement of adaptive T cells. To test any such predictions it would be necessary to develop experimental systems that would allow the inducible deletion of just ILC rather than the current models that can only study ILC function in the absence of an adaptive immune system.