Innate lymphoid cells (ILC) are tissue-resident immune cells lacking receptors to recognize specific antigens; hence their activation depends on cytokines present in the microenvironment (1). Type 2 ILCs or ILC2s are mainly found in the lung, intestine and skin, where they are activated by the epithelial cell-derived alarmins IL-25, TSLP, and IL-33 upon tissue damage (2), although they can also respond to cytokines derived from the immune compartment such as IL-2 and IL-7 (3).

ILC2s are mainly tissue resident cells; however, their precursors reside in the bone marrow, where they maturate after going through different stages of differentiation. Migration of ILC2 precursors from the bone marrow to different peripheral tissues such as skin and lung has been shown to occur in response to intravenous IL-33 administration or IL-33-dependent pulmonary fungal allergen challenge (4). In addition, intranasally administered IL-33 can have a direct effect on bone marrow derived ILC2s, inducing their production of IL-5, demonstrating an important role for these cells in the induction of eosinophilia derived from inflammation in the upper airways (5).

Recent studies have shown that multiple molecules besides cytokines can regulate the functions of ILC2s in different contexts both in vitro and in vivo. These stimuli include the lipid mediators leukotrienes and prostaglandins, neuropeptides, hormones and nutrients (6), thereby demonstrating that the regulation of ILC2 function is tightly controlled by a variety of signals within the cellular microenvironment. Several of these stimuli act through the cAMP/PKA pathway.

Cyclic adenosine monophosphate (cAMP) has been established as a universal regulator of metabolism and gene expression in all living organisms (7). The levels of this messenger are regulated by different enzymes such as adenylate cyclase and phosphodiesterases (8). Although cAMP acts in multiple downstream pathways, its most prominent role is the activation of the protein kinase A (PKA). Upon binding of cAMP, PKA dissociates into its regulatory and catalytic subunits. The catalytic subunits then phosphorylate specific Ser and Thr residues on numerous downstream target proteins involved in different signaling pathways. For example, cAMP-activated PKA binds and phosphorylates cAMP-responsive transcription factors, including cAMP-response element binding protein (CREB), activating transcription factor-1 (ATF-1), Nuclear Factor κB (NFκB), and nuclear receptors (8).

Studies exploring the effects of cAMP, PKA or CREB activation on ILCs are limited. However, on helper T cells, ILCs adaptive cellular counterparts, they regulate cytokine expression, proliferation and apoptosis (8–12). Altogether these reports suggest that activation of the cAMP/PKA pathway has a negative effect on T cell survival by inhibiting cell proliferation and increasing apoptosis (10–12). The cAMP pathway was recently shown to suppress ILC2 function through mediators such as PGE2 and PGI2 (13, 14). Contrastingly, cAMP signaling mediated by neuropeptides activates ILC2s (6), demonstrating a dual regulation of these cells that is context dependent.

ILC2s are characterized by the expression of cytokines associated with the type 2 immune response, including IL-4, IL-5, IL-13 and IL-9, all important in the regulation of immunity against helminth parasites and allergies (15). The expression of IL-9 derived from ILCs was first confirmed in 2011 using a fate reporter in models of sterile inflammation of upper airways (16). A couple of years later, the first evidence of IL-9 expression specifically in ILC2s in vivo was reported, when ILC2s were found to be the main source of IL-9 expression in mice during Nippostrongylus brasiliensis (N. brasiliensis) infection (17). ILC2 cells were confirmed to express high levels of the IL-9 receptor (18) and this cytokine was found to act in an autocrine manner, promoting IL-5, IL-13 and amphiregulin expression as well as ILC2 survival (19). Despite these findings, the mechanisms controlling the expression of IL-9 in ILC2s remain unclear.

Most studies on the regulation of ILC2 cytokines by cAMP-associated pathways have been focused on IL-5 and IL-13 in different inflammation models (13, 20). However, studies exploring how the cAMP pathway could potentially regulate IL-9, and cellular processes such as proliferation and survival, are lacking.

Suppressor of Tumorigenicity 2 (ST2), a component of the IL-33 receptor, is commonly used as a marker for the identification of ILC2s. In mice, ST2 is expressed on ILC2s in the lung, and can also be found in bone marrow and adipose tissue, while its basal expression is limited in tissues such as the gut and skin (21, 22). IL-33 signals through ST2 to activate ILC2s, which produce IL-5 and IL-13 in models of upper airway inflammation and helminth infection (23–25).

In this work, we explored the signals that control IL-9 expression in ILC2s and how the regulation of PKA directs the function of these cells. We found that during N. brasiliensis infection, different populations of ILC2s can be distinguished based on their IL-9 and ST2 expression. These populations are distributed in the different anatomical compartments analyzed. In vitro, IL-33 and the cAMP/PKA pathway modulated the proliferation of ILC2s and regulated their expression of IL-9 and other markers, establishing the stimuli that could be regulating the heterogeneity and function of these cells in the different compartments during infection with N. brasiliensis in vivo.

In this work we report the presence of distinct ILC2 populations based on ST2 and IL-9 expression in a helminth infection model. These populations, which include ST2⁺ IL-9^-, ST2⁺ IL-9⁺, ST2^- IL-9⁺ and ST2^- IL-9^- had not been previously reported in vivo. Here, we provide the first evidence that they are found in tissues and to a lesser extent in lymphoid organs. The ST2⁺ IL-9^- cells predominate in the lung, and along with ST2⁺ IL-9⁺ cells, increase in the N. brasiliensis infection model. Infection with the parasite seems to induce the expression of ST2 and IL-9 in the intraepithelial compartment of the small intestine, however this induction is very discrete and did not reach the statistical significance. In the lamina propria, both IL-9⁺ populations are increased, however the ST2^- IL-9⁺ is the most abundant subset induced by the infection. Two different populations of ILC2 cells have been described: natural (nILC2), characterized by the expression of KLRG1 and ST2; and inflammatory (iILC2), which are induced by N. brasilienis infection and express KLRG1 but apparently lack ST2 (36, 37). This absence of ST2 expression on iILC2s might be explained by downregulation of the receptor driven by its ligand IL-33 in the small intestine, as recently demonstrated (38). IL-33 is constitutively expressed in this tissue and, together with IL-25, is further induced by N. brasiliensis infection to promote the generation of iILC2 cells. Hence, the presence of the ST2^- IL-9⁺ ILC2 population in the small intestine described here might be explained by this regulatory loop, where IL-33 and IL-25 are induced upon infection, promoting iILC2s that lack ST2 and express effector cytokines in vivo. Altogether, our results demonstrate a dynamic transformation of the ILC2 subsets dependent on the tissue and time of infection that can be modulated by the inflammatory context.

An important aspect to consider is that isolating cells from N. brasiliensis infected small intestine is extremely challenging and results in low frequencies of live cells, potentially masking real absolute numbers of cells present in the tissue in vivo. However, analyses of cell frequencies are feasible and reproducible using our protocol. To our knowledge, this is the first report showing hematopoietic cell isolation from N. brasiliensis infected small intestine, followed by analysis of ILC2 cells by flow cytometry, and hence represents a useful tool to evaluate these and other immune cells from this particular complicated tissue.

In order to dissect the signals that originate these ILC2 populations, we developed an in vitro culture model from bone marrow cells in which different combinations of cytokines were tested to determine which ones are important for the induction of ST2 and IL-9. We found that IL-33 is not necessary for the induction of its receptor since the combination of IL-2, IL-7, IL-25 and TSLP is sufficient to generate ST2⁺ ILC2s. On the contrary, IL-33 decreased ST2 expression, which could be due to internalization of the receptor upon ligand binding as described above. Additionally, our studies confirmed IL-33 as a key stimulus to increase IL-9 expression in our cells. This data agrees with previous reports by Mohapatra et al., where lung ILC2s expressed IL-9 in response to IL-33 and this expression was increased when the treatment was performed in combination with other cytokines such as IL-2, IL-7 and TSLP (30). Therefore, we can conclude that the ILC2s in the whole bone marrow culture model resemble the behavior of ILC2s derived from other tissues such as small intestine and lung.

In order to determine the effects of IL-33 and PKA activation directly on ILC2s, we developed another culture model termed BM-ILC2, in which we expanded bone marrow ILC2s in the presence of IL-2, IL-7, TSLP and IL-25 for 3 days. ILC2s where then sorted and cultured 4 additional days in the same cocktail with the addition of IL-33 and/or a cAMP analog. We further demonstrated that with our novel experimental strategy, we were able to generate cultures highly enriched on ILC2 cells expressing GATA-3 and CD25, as expected.

In this model, IL-33 did not have a negative effect on the expression of ST2, differing from the results observed in the whole bone marrow culture. We believe that the effect observed in total bone marrow cultures may depend on further signals provided by other cells, in addition to the presence of IL-33. On the other hand, the IL-33-dependent induction of IL-9 on our BM-ILC2 cells was independent of those potential signals included in the total bone marrow culture, since purified cultures of ILC2s exposed to IL-33 still showed a strong induction of IL-9.

The addition of IL-33 to our sorted ILC2 cultures increased the expression of other markers associated with the identification of these cells such as CD25 and KLRG1. In addition, the differential expression of IL-9 generated three different subpopulations of ILC2 cells, all identified as CD45⁺ lineage^- CD25⁺ GATA-3⁺. These subsets are ST2⁺ IL-9^-, ST2⁺ IL-9⁺ and ST2^- IL-9^-. The first 2 populations correspond to those previously found in the lung and in the intraepithelial compartment of the small intestine. The ST2^- IL-9⁺ ILC2 population identified in lamina propria of N. brasiliensis infected mice, could not be induced in our in vitro model; we found the frequencies of these cells to be less than 3 percent of the total population. As previously mentioned this could be due to the requirement of other signals in combination with IL-33 to induce the negative regulation of ST2 while maintaining the expression of IL-9.

The PKA pathway is known to regulate the expression of cytokines in T lymphocytes and ILC2s. Here, we confirmed that 8-Br-cAMP-mediated PKA activation in the presence of IL-33 increased IL-9 expression, which does not modify the distribution of the ILC2 populations but could enhance the function of ST2⁺ IL-9⁺ ILC2 cells. To our knowledge, there are no previous reports associating PKA activation with an increase in IL-9 in innate lymphoid cells, however this effect has been observed in Th9 lymphocytes where the CGRP/cAMP/PKA pathway promotes IL-9 production via NFATc2 activation (39). This establishes a possible common regulatory pathway between ILC2 and T lymphocytes.

The synergic effect of PKA and IL-33 on IL-9 expression described here is contrasting with studies that report PKA-mediated inhibition of type 2 cytokines in ILC2s (14, 20, 34). This suggests that although our in vitro ILC2s have characteristics in common with lung ILC2s, there are also specific differences associated with the microenvironment of these cells. Along with the increase in IL-9, PKA activation induced CD127 and decreased KLRG1 expression. Previous reports have identified KLRG1 not only as a marker of ILC2 maturation but also as a negative regulator of the function of these cells. The interaction of activated ILC2s with the KLRG1 ligand E-cadherin results in a down regulation of GATA-3, IL-5, IL-13, amphiregulin and reduced ILC2 proliferation (29). In other tissues such as the stomach, ILC2s can be activated by IL-7, inducing proliferation and expression of cytokines (40). We propose that PKA activation could prime these cells for a better activated phenotype, perhaps as a mechanism to ensure proliferation and proper function. However, future experiments are necessary to verify this.

A striking finding is that when PKA is activated, ILC2s seem to decrease in numbers except for the IL-9-producing population. Our results suggest that this can be attributed to the inhibition of proliferation since we did not find PKA activation to be associated with apoptosis. However, the induction of IL-9 by PKA could be compensating for these proliferation effects, which is why we do not see differences in absolute numbers of IL-9⁺ ILC2 cells when we activate PKA. Our results are in accordance with previous studies that report IL-9 to be a potent inducer of proliferation and survival for ILC2 cells (19).

The inhibitory effects on proliferation described here for ILC2s have already been reported in T cells (11, 12, 41). In this model, inhibition of proliferation by the cAMP analog is observed independently of IL-33. cAMP has been reported to inhibit proliferation in T cells through different mechanisms. For example, it can generate cell cycle arrest through the inhibition of cyclins D and E (12) and PKA activation can block IL-2 receptor signaling (11). Further research is necessary to determine the regulatory mechanism carried out by PKA to inhibit proliferation of the ILC2s in our model. In addition, although we did not observed an effect on apoptosis induced by IL-33 and the PKA pathway in our cultures, we did see a lower frequency of apoptotic and death cells on the ILC2 subset expressing IL-9, further supporting the role of IL-9 as a positive signal acting in an autocrine way on these cells (18, 19).

Additional data derived from our BM-ILC2 cultures validate this protocol as a means to generate bona fide ILC2 cells, since our ILC2 populations express both IL-5 and IL-13. In agreement with other works, our data indicate that IL-5 induction is dependent on IL-33 as has been observed in lung-derived ILC2s (42). Contrastingly, IL-13 expression does not require IL-33, however it is significantly enhanced by the presence of the alarmin. Our data is in agreement with previous reports of ILC2s in the intestine, where IL-25 produced by Tuft cells further activated these cells to produce IL-13 in a parasitic infection context (43).

Lung ILC2s have been reported to express IL-5 and IL-13, which can be blocked by cAMP analogs and ligands that activate PKA (13). The effects of PKA on these two cytokines are similar in our model. We established that cAMP has regulatory effects on the numbers of ILC2 cells expressing IL-5 and/or IL-13. These effects were observed only in the presence of IL-33 and were restricted to ILC2s that do not express IL-9. These data suggest that while there is a decrease in the induction of IL-5 and IL-13 in ILC2 cells, the increase in IL-9-expressing cells compensates for these defects, resulting in no change in the absolute numbers of ILC2 that co-express IL-5 or IL-13 with IL-9.

Finally, the induction of IL-4 has been reported in ILC2s through activation with other mediators such as leukotriene D4, but not with IL-33 (44, 45); further supporting the observation in our model that IL-33 alone is unable to promote IL-4 expression in vitro. This could be because in addition to the cytokines used here and in previous reports, other signals are required to generate IL-4⁺ ILC2 cells; however, this has yet to be demonstrated.

In general, the ILC2s differentiated from bone marrow precursors obtained with our in vitro model share various characteristics with tissue-derived ILC2s. It is tempting to speculate that the different ILC2 subpopulations characterized in this work are found in different anatomical locations and that their modulation depends on the conditions and time of stimuli presentation. N. brasiliensis infection for example, results in tissue damage and IL-33 release (46). Worm-derived factors can modulate this IL-33 response (25) and consequently generate the different ILC2 populations. Therefore, it would be interesting to determine if these ILC2 populations are induced in response to IL-33 in vivo as well. Here, we found that stimulation with IL-33 and PKA activation results in an enrichment of IL-9⁺ ILC2 cells that are competent regarding the expression of IL-5 and IL-13. Therefore, expansion of this population could contribute to immunity against helminth infections.

Based in our results, the PKA pathway antagonizes IL-33 in several processes such as proliferation and production of IL-5 and IL-13, while inducing a phenotype that favors its ability to respond to other stimuli in the microenvironment (IL-9⁺, CD127⁺, KLRG1^-). Whether changes in PKA activation are generated during helminth infection with N. brasiliensis is a matter of interest and remains to be determined along with the identification of the potential ligands promoting such activation in vivo.

To conclude, the data presented here provide further evidence of the high heterogeneity between ILC2 subsets in a helminth infection context and calls for careful consideration when choosing the markers to identify these cells. In addition, by providing a protocol for differentiation and expansion of bona fide ILC2 cells, we contribute with a useful tool for studies focused on ILC2 biology. Finally, the findings that IL-33 and PKA induce IL-9 and regulate the different ILC2 population dynamic, could help us better understand the physiology of these cells, allowing their manipulation for therapeutic purposes in specific anatomical locations.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The animal study was reviewed and approved by CICUAL Instituto de Fisiología Celular, UNAM.

EO-M conceived, design and performed the experiments, analyze the data and wrote the manuscript, OM-P performed experiments and discussed the data, BER-M revised the manuscript, JLR-B performed experiments, acquired the data and provided reagents, IL-L discussed and thoroughly revised the results and conclusions of the manuscript, PL-L conceived the idea, design the experiments, discussed and revised the manuscript, supervised the project and acquired the funding. All authors contributed to the article and approved the submitted version.

This work was supported by the following grants to PLL from CONACYT (CB 255287 FORDECYT-PRONACE-303027) and DGAPA (IN209919-PAPIIT). EO-M received a fellowship from CONACYT (481437).

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.

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.