Nematode-Infected Mice Acquire Resistance to Subsequent Infection With Unrelated Nematode by Inducing Highly Responsive Group 2 Innate Lymphoid Cells in the Lung

The immune responses against helminths have been investigated individually, and it is well-established that infected hosts develop an immunological memory to resist reinfection by the same pathogen. In contrast, it is poorly understood how the host immune system responds to subsequent infection by unrelated parasites after elimination of the first infection. We previously reported that infection of mice with Strongyloides venezuelensis induces the accumulation of group 2 innate lymphoid cells (ILC2s) in the lung. Here, we demonstrated that S. venezuelensis-experienced (Sv-exp) mice became significantly resistant against infection by Nippostrongylus brasiliensis. N. brasiliensis infection induced enhanced accumulation of ILC2s and eosinophils with increased expressions of mRNA for Th2 cytokines in the lungs of Sv-exp mice. The resistance was dependent on ILC2s, and eosinophils but not on CD4+ T cells. Furthermore, pulmonary ILC2s in Sv-exp mice acquired a highly responsive “trained” phenotype; in response to N. brasiliensis infection, they rapidly increased and produced IL-5 and IL-13, which in turn induced the early accumulation of eosinophils in the lungs. IL-33 was required for the accumulation of ILC2s and the resistance of mice against N. brasiliensis infection but insufficient for the induction of trained ILC2s. In conclusion, animals infected with one type of lung-migratory nematodes acquire a specific-antigen-independent resistance to another type of lung-migrating nematodes, providing animals with the capacity to protect against sequential infections with various lung-migratory nematodes.

The immune responses against helminths have been investigated individually, and it is well-established that infected hosts develop an immunological memory to resist reinfection by the same pathogen. In contrast, it is poorly understood how the host immune system responds to subsequent infection by unrelated parasites after elimination of the first infection. We previously reported that infection of mice with Strongyloides venezuelensis induces the accumulation of group 2 innate lymphoid cells (ILC2s) in the lung. Here, we demonstrated that S. venezuelensis-experienced (Sv-exp) mice became significantly resistant against infection by Nippostrongylus brasiliensis. N. brasiliensis infection induced enhanced accumulation of ILC2s and eosinophils with increased expressions of mRNA for Th2 cytokines in the lungs of Sv-exp mice. The resistance was dependent on ILC2s, and eosinophils but not on CD4 + T cells. Furthermore, pulmonary ILC2s in Sv-exp mice acquired a highly responsive "trained" phenotype; in response to N. brasiliensis infection, they rapidly increased and produced IL-5 and IL-13, which in turn induced the early accumulation of eosinophils in the lungs. IL-33 was required for the accumulation of ILC2s and the resistance of mice against N. brasiliensis infection but insufficient for the induction of trained ILC2s. In conclusion, animals infected with one type of lung-migratory nematodes acquire a specific-antigen-independent resistance to another type of lung-migrating nematodes, providing animals with the capacity to protect against sequential infections with various lung-migratory nematodes.

INTRODUCTION
Soil-transmitted intestinal parasites cause one of the most prevalent neglected tropical diseases (NTD) (1). Soil-transmitted helminths infect approximately 1.7 billion people in developing countries, creating both health-and economics-related problems (1-4). Particularly, in tropical and subtropical regions, such as sub-Saharan Africa, the Americas, China, and East Asia, the distribution of many infectious diseases caused by parasitic nematodes, such as Ascaris lumbricoides, hookworm (Ancylostoma duodenale and Necator americanus), Trichuris trichiura, and Strongyloides stercoralis, overlap one another (1, 2,5). People in these areas can be infected by multiple parasites concurrently or sequentially.
These soil-transmitted helminths are gastrointestinal nematodes, and they are divided into two groups based on the infection routes. A. lumbricoides and T. trichiura parasitize directly to the intestinal tract by oral infection, hookworm and S. stercoralis transcutaneously invade and once go to the lungs and then migrate to the intestinal tract to mature. For this reason, the latter is known to cause eosinophilic inflammation in the lungs (Loeffler syndrome).
To resist such intestinal nematodes infection, hosts generally develop Th2 immune responses, which induce IgG1 and IgE production, intestinal mastocytosis, systemic eosinophilia, goblet cell hyperplasia, and intestinal smooth muscle contraction (6)(7)(8)(9). Among these various reactions, the response required for appropriate protection depends on the type of infecting parasite.
In an experimental animal model, the induction of goblet cells, epithelial cell turnover, and smooth muscle contraction, all of which are induced by the action of Th2 cell-derived cytokines (IL-4 and IL-13), are indispensable for the rapid expulsion of N. brasiliensis, a gut-dwelling rodent nematode (10)(11)(12).
Strongyloides venezuelensis, another rodent nematode, is a counterpart of the human pathogen S. stercoralis (13). In contrast to the required responses for N. brasiliensis expulsion, cytokineinduced mastocytosis and antibody (Ab)/FcRγ-dependent mucosal mast cell (MMC) activation are important for the rapid expulsion of S. venezuelensis from the intestine (14)(15)(16)(17)(18). Since the host memorizes this particular Th2 type immune response, when the same species of nematode invades again, a stronger immune response occurs to reject the infection (19,20).
To date, the immune responses against helminths have been investigated individually, and it is well-established that helminth-infected hosts develop an immunological memory to resist a reinfection by the same pathogen. In contrast, it is poorly understood how the host immune system responds to subsequent infection by an unrelated intestinal nematode after elimination of the first infection. The aim of this study is to Abbreviations: Ab, Antibody; BALF, Bronchoalveolar lavage fluid; GATA3, GATA-binding protein 3; ILC2, Group 2 innate lymphoid cell; L3, Third stage larva; NK, Natural killer; PBL, Peripheral blood leukocyte; Q-PCR, Quantitative PCR; RORα, Retinoic acid receptor-related orphan receptor alpha; Sv-exp, Strongyloides venezuelensis experienced; TSLP, Thymic stromal lymphopoietin. investigate the effect of primary infection on secondary infection with a different species of intestinal nematode and analyze S. venezuelensis-infected hosts acquisition of specific-antigenindependent resistance to infection by subsequent unrelated lung-migrating nematode by the induction of trained ILC2s. Animals C57BL/6 mice were purchased from Charles River Laboratories Japan. Wistar rats, and Sprague-Dawley (SD) rats were purchased from SLC Japan. Il33 −/− mice were produced and reported previously (31). RORα sg/j mice were obtained from Jackson Laboratory. dblGATA mice were purchased from Jackson Laboratory and backcrossed for 11 generations into the C57BL/6 background. All mice were used at 6-11 weeks of age except for the donor mice for bone marrow transplantation experiment. To prepare Rora sg/sg bone marrow chimera mice, recipient 6week-old C57BL/6 wild-type (WT) mice were lethally irradiated (5.5 Gy × 2, 4 h interval). Donor bone marrow cells (3-10 × 10 6 cells per body) prepared from 3-to 4-week-old RORα sg/+ or RORα sg/sg mice were transferred intravenously into recipients. Chimera mice were used for experiments from 12 weeks after transplantation. The numbers of mice used in each experiment were indicated in each figure legend.

Helminth Infections
S. venezuelensis was provided by Dr. H. Maruyama (University of Miyazaki) and maintained in male Wistar rats in our laboratory (15). Mice were subcutaneously inoculated in the back with 4,000-5,000 S. venezuelensis L3. N. brasiliensis was provided by Dr. M. Yamada (Kyoto Prefectural University of Medicine) and maintained in male SD rats (33). Mice were subcutaneously inoculated in the inguinal region with 500 N. brasiliensis L3. The degrees of infection of individual mice were monitored by counting the number of eggs excreted in feces daily (eggs/g of feces). For S. venezuelensis eggs, we counted eggs by direct smear method. In brief, feces fleshly passed from each mouse were weighed and suspended in a known volume of water in centrifuge tubes. The eggs were counted using a microscope (13). For N. brasiliensis eggs, we counted eggs after concentration with formalin-ether sedimentation method with minor modification (34). In brief, feces were weighed and dissolved in water, filtered through one layer of gauze, and centrifuged at 700 g for 3 min, the sediment was suspended with 10% formalin, after 20 min, ether was added, and shaken for 30 s, and the tube was again centrifuged and the pellet was recovered. Lung cells (prepared as described below), BALF cells, and PBLs were collected at 0, 7, 14, 21, and 28 days after S. venezuelensis infection.
To quantify lung stage larvae, mice were euthanized by cervical dislocation 2 days after infection, the whole lung from each mouse was harvested, minced, and incubated on a steel mesh put on a beaker containing 100 ml of PBS for 2 h at 37 • C. Worms that emerged from the lung tissue were counted using a microscope (40x). To assess intestinal worm burdens, the whole small intestine from each infected mouse at the indicated day was removed and incubated on a steel mesh as above. The number of worms that emerged from the intestinal tissue was counted.

Intraduodenal Implantation of Adult Worms
Adult worms were prepared from the intestines of SD rats inoculated with 4,000 infective larvae 7 days previously. The adult worms were suspended in 200 µl of PBS and injected into the duodenum of recipient mice (175 worms/mouse) under anesthesia using a 1-ml syringe with an 18-gauge angiocath (BD Biosciences). Sixteen hours later, the mouse intestines were removed, cut open longitudinally, and incubated at 37 • C in PBS for 2 h. The number of worms that emerged from the intestines was counted. For histological analysis, intestine specimens were sampled at about 6-12 cm distal from pylorus, fixed with 4% (wt/vol) paraformaldehyde (PFA) and embedded in paraffin. Those tissues were sectioned at 8-µm thickness and deparaffinized sections were stained with hematoxylin and eosin.

Bronchoalveolar Lavage (BAL)
BAL was performed on the left lung with three aliquots of 0.5 ml of PBS per mouse, with a ligature of the right main bronchus by forceps, after perfusing the mouse via the right ventricle with PBS under anesthesia. The resulting BALF was examined for cytokine concentrations by ELISA using an ELISA Duo set for mouse IL-5 and IL-13 (R&D systems), and BALF cells were analyzed using flow cytometry.

Flow Cytometry
BALF cells, lung cells, and PBLs were stained with antibodies for surface antigens and analyzed using a FACSCalibur, LSRFortessa X-20 (BD Biosciences), or SP6800 (SONY) flow cytometer. Cells were classified as indicated in each figure legend. For intracellular cytokine staining, lung cells were stimulated with 50 nM PMA and 1 µM ionomycin for 4 h in the presence of brefeldin A or with 10 ng/ml IL-7, 100 ng/ml IL-25, and 100 ng/ml IL-33 for 24 h (brefeldin A was added for the last 6 h). Cells were stained with eBioscience Fixable viability dye eFlour506 (Thermo Fisher) and with antibodies against surface antigens. Fixation, permeabilization, and intracellular cytokine staining were performed using an eBioscience Foxp3/Transcription factor staining buffer kit and antibodies for IL-5 and IL-13 and were analyzed using a SP6800 flow cytometer.

Quantitative RT-PCR (Q-PCR)
Total RNA was extracted with a RNeasy Mini Kit (Qiagen), and cDNA was synthesized using RevertraAce (TOYOBO). The expression levels of genes were quantified with TaqMan Gene Expression Assays (Applied Biosystems). The results are shown as the relative expression standardized against the expression of the gene encoding β-actin. Specific primers and probes used for quantitative RT-PCR were Il33 (Mm00505403_m1), Il5 (Mm00439646_m1), Il13 (Mm00434204_m1), Epx (Mm00514768_m1), and actin-b (4352933E) (Applied Biosystems).

Statistics
All data are shown as the mean ± S.D. The numerical data were analyzed using either a Student's t-test, one-way ANOVA with Bonferroni's post-hoc test, or two-way ANOVA with Bonferroni's post-hoc test using Prism (GraphPad Software), and p-values of less than 0.05 were considered significant. In all figures, * p < 0.05, * * p < 0.01, * * * p < 0.001, and * * * * p < 0.0001.

Infection-Experienced Mice Develop an Enhanced Lung Inflammation and Show a Substantial Resistance to Nippostrongylus brasiliensis Infection
In the life cycles of S. venezuelensis and N. brasiliensis, the third stage larvae (L3) infect the skin and migrate to the lung before they mature in the intestine (36,37). Within the lung, they induce severe inflammatory changes (i.e., mouse Loeffler syndrome), characterized by severe eosinophilic infiltration and goblet cell hyperplasia via the action of Th2 cytokines mainly from IL-33activated ILC2s (31). However, mice that have been previously infected with these nematodes inhibit re-infection with the same pathogen by preventing larval migration to the lung via acquired immunity (38,39).
To examine the effect of prior nematode infection on subsequent infection with a different nematode, we subcutaneously infected 6-week-old naive WT C57BL/6 mice with L3 of S. venezuelensis. The infected WT mice normally expel this parasite after 2 weeks due to the action of mast cells (16,17). At 4 weeks after the primary infection, S. venezuelensis-experienced (Sv-exp) or age-matched uninfected control mice were infected with L3 of N. brasiliensis, which is another species of lung migrating intestinal nematode, or were re-infected with L3 of S. venezuelensis ( Figure 1A). First, we confirmed that comparable numbers of N. brasiliensis larvae migrated to the lungs in control and Sv-exp mice (Figure 1B), indicating the occurrence of normal larval migration from the skin to the lungs in Sv-exp mice. We found that, at days 5 and 7 after N. brasiliensis infection (day 33 and day 35 after S. venezuelensis infection), Sv-exp mice had significantly lower worm burdens compared with S. venezuelensis-uninfected control mice ( Figure 1C). Further, we were unable to detect eggs in the feces from Sv-exp mice (Figure 1D), indicating that Sv-exp mice acquired resistance to infection by N. brasiliensis.
Because host animals expel nematodes via the effects of Th2 cytokines, we analyzed the cytokine expression levels in the intestine. Sv-exp mice did not show an enhancement in cytokine expression compared with control mice (Figure S1A), indicating that the resistance to infection by N. brasiliensis is not caused by a severe inflammatory response in the intestine. Furthermore, we detected comparable worm burdens in both groups when we inoculated N. brasiliensis adult worms directly into the intestine (Figure S1B). They migrated into the intervillous space of jejunum (Figure S1C), suggesting that the intestines in Sv-exp animals were conducive to parasitic infection.
To investigate the immune responses in the lung, we analyzed the accumulation of inflammatory cells in the bronchoalveolar lavage fluid (BALF). Following N. brasiliensis infection, Sv-exp mice showed larger accumulations of ILC2s and eosinophils in the lung alveolus compared with control (no primary infection) mice ( Figure 1E and Figure S2). As ILC2s are known to produce large amounts of the Th2 cytokines IL-5 and IL-13 (40,41), it is not surprising that we detected higher levels of IL-5 and IL-13 in the BALF from N. brasiliensis-infected Sv-exp mice compared with control mice (Figure 1F). Accordingly, preinfection with S. venezuelensis also enhanced the expression of mRNA for IL-5 and IL-13 in the lungs ( Figure 1G). In contrast, reinfection with S. venezuelensis did not induce such hyperresponses (Figures 1E,G), probably due to the prevention of larval migration to the lung via acquired immunity provoked by the primary infection (39).
Since S. venezuelensis infection induces an activation of ILC2s that is IL-33-dependent (31), we analyzed the level of Il33 mRNA expression in Sv-exp and S. venezuelensis infection-naive mice. The experience of S. venezuelensis infection did not influence the expression of Il33 mRNA (Figure 1G), either before or after N. brasiliensis infection, indicating that the increase in the number of ILC2s was not caused via an enhancement of IL-33 production.
To determine if these results are specific responses to S. venezuelensis infection, we conducted the reverse experiment, beginning with N. brasiliensis as the primary infection. Notably, the experience of N. brasiliensis infection similarly reduced the number of eggs per gram of feces in S. venezuelensis-infected mice and increased the number of eosinophils in the BALF (Figure S3).
We were able to detect a protective effect of prior S. venezuelensis infection even 3 months later (Figures S4A,B). The numbers of ILC2s and eosinophils tended to be enhanced by N. brasiliensis infection 3 months after S. venezuelensis infection, though these differences failed to reach statistical significance ( Figure S4C). Overall, the protective effect induced by experiencing an S. venezuelensis infection lasts at least 3 months after the first infection.
The above data suggest that the lung is the important site for protection against a subsequent infection by other pulmonary migratory nematode. To investigate the changes in the lungs after S. venezuelensis infection, we analyzed cells in the BALF at different time points after S. venezuelensis infection. In agreement with our previous report, S. venezuelensis infection resulted in a rapid increase in the number of inflammatory cells in the BALF (Figure 1H) (31). After the initial expansion, the numbers of ILC2s, eosinophils, and lymphocytes all peaked at day 14. Around day 12, adult worms were expelled from the intestine in WT mice (16,17). Eosinophil and macrophage populations declined to their basal levels at day 21 and day 14, respectively, indicating the resolution of inflammation. In contrast, many ILC2s persisted in the BALF, even at day 28. These data suggest that the ILC2s remaining in the lungs after infection contribute to protection against a subsequent infection by other pulmonary migratory nematode. FIGURE 1 | of ILC2s, eosinophils, CD3 + /B220 + cells, and monocytes among the BALF cells at the indicated days were analyzed by flow cytometry (FACScalibur). Cell populations were defined as follows: ILC2s, FSC lo SSC lo Lin(CD3, CD4, CD8, CD19, NK1.1, IgE, Gr-1, siglecF) − Sca-1 + ST2 + ; Eosinophils, CD45 + CD3 − B220 − CCR3 + , CD3/B220: CD45 + CCR3 − CD3 + /B220 + ; and Monocytes, CD45 + Autofluorescence high . Data were compared with day 0 data by one-way ANOVA. Data are representative of two independent experiments.

ILC2s Contribute to the Resistance Against Nippostrongylus brasiliensis Infection in Sv-Exp Mice
Since Sv-exp mice were found to have an increased number of ILC2s in the lungs, we next examined whether ILC2s are required for the hyper-responses in the lung and the lower N. brasiliensis worm burden that occur in these mice. To investigate the role of ILC2s, we made ILC2-deficient Rora sg/sg bone marrow chimera mice and infected them with nematodes (Figure 2A). First, we confirmed the absence of ILC2s in Rora sg/sg chimera mice ( Figure 2B). Notably, the eosinophil accumulation was reduced in ILC2-deficient mice compared with ILC2-sufficient Rora sg/+ control chimera mice ( Figure 2B). The expression levels of Th2 cytokines, which are major products of ILC2s, were also decreased in the lungs of Rora sg/sg chimera mice (Figure 2C), though low levels of cytokine expression remained. Probably other Th2 cytokine-producing cells, such as T cells, mast cells and basophils, cause the remaining cytokine expressions. Furthermore, S. venezuelensis-infection-induced resistance against N. brasiliensis was impaired in ILC2-deficient mice (Figure 2D), suggesting that the protective phenotype of Sv-exp mice requires RORα-dependent innate lymphoid cells.
In addition to ILC2s, S. venezuelensis infection induces strong Th2 differentiation of CD4 + helper T cells, as a general result of nematode infection (7,42,43). To examine the contribution of Th2 cells to the enhanced eosinophil accumulation and protection against subsequent N. brasiliensis infection, we depleted CD4 + T cells by treating Sv-exp mice with an anti-CD4 Ab before N. brasiliensis infection ( Figure S5A). The resistance against N. brasiliensis infection and the alveolar accumulation of ILC2s and eosinophils in Sv-exp mice were not affected by the absence of CD4 + T cells (Figure 2E and Figure S5B), indicating that Th2 cells are dispensable for these responses. These data demonstrate that Sv-exp mice acquired resistance to N. brasiliensis via ILC2-dependent innate immune responses.

Eosinophils Are Essential for the Protection of Sv-Exp Mice From Nippostrongylus brasiliensis Infection
Since ILC2s are a major source of IL-5 production, we investigated the requirement for IL-5 in the protective phenotype of Sv-exp mice. The blockade of IL-5 completely abolished the protective effect of S. venezuelensis pre-infection on N. brasiliensis infection ( Figure 3A). As expected, the IL-5 depletion strikingly inhibited the accumulation of eosinophils in the BALF from N. brasiliensis-infected mice as well as the expression of mRNA for eosinophil peroxidase (Epx) in the lungs without affecting the numbers of ILC2s and CD4 + T cells or the mRNA expression levels of Il33 and Il13 (Figures 3B,C).
To confirm the role of eosinophils, we infected eosinophildeficient dblGATA mice with S. venezuelensis, followed later by infection with N. brasiliensis. In agreement with the results from the IL-5 neutralization experiment, eosinophil-deficient Sv-exp mice failed to reject N. brasiliensis infection, without displaying reductions in the accumulation of ILC2s and CD4 + T cells or in the mRNA expression levels of Il33, Il5, and Il13 (Figures 3D-F). These results clearly demonstrate that the enhanced IL-5 production by ILC2s in Sv-exp mice increases the number of eosinophils in the lungs, and these eosinophils contribute to the protection against N. brasiliensis infection.

IL-33 Is Critical for the Resistance of Sv-Exp Mice to Infection by Nippostrongylus brasiliensis
We previously reported that IL-33-deficient mice had an impairment in the early accumulation of eosinophils and ILC2s in response to S. venezuelensis infection (31). To determine if the reduction of ILC2s in Il33 −/− mice was restored by other mechanisms to the level of WT mice after the first 2 weeks post-infection, we infected WT and IL-33-deficient mice with S. venezuelensis and analyzed the cell population in the BALF over the following 4 weeks. Il33 −/− mice were unable to accumulate ILC2s during these 4 weeks (Figure 4A). Although there were significant differences on day(s) 7 and/or 14 in the number of eosinophils and mRNA expression levels of Il5 and Il13 between WT and Il33 −/− mice, by day 21, these differences between WT and Il33 −/− mice became negligible (Figures 4A,B).
To examine the role of IL-33 in the enhanced lung inflammation induced by subsequent helminth infection, we infected Sv-exp WT and Il33 −/− mice with N. brasiliensis. A lack of IL-33 severely impaired the accumulation of ILC2s and eosinophils along with the induction of cytokine mRNA expression in Sv-exp mice (Figures 4C,D), indicating that IL-33 is required for the accumulation of ILC2s and the severe lung inflammation in sequential nematode infection. Furthermore, S. venezuelensis infection-induced resistance to infection by N. brasiliensis was abrogated in Il33 −/− mice ( Figure 4E). These data clearly demonstrate the critical role of IL-33 in the enhanced lung inflammation and host defense of Sv-exp mice.

Strongyloides venezuelensis Infection Induces Trained ILC2s
N. brasiliensis larvae enter the lung around day 2 after they are subcutaneously injected into a host, and they migrate to  Figure S5. The numbers of worms in the intestine were counted at day 5 after N. brasiliensis infection (n = 10-11). Pooled data from two independent experiments are shown (Mean ± SD). the intestine around day 5 (38); therefore, early accumulation and activation of ILC2s and eosinophils in the lungs may be important for the nematode infection resistance of Sv-exp mice. To examine the timing of the inflammatory cell increase in the lungs after N. brasiliensis infection, we analyzed the lung cells of Sv-exp mice and control mice daily. We found that the numbers of pulmonary eosinophils, ILC2s, and CD4 + T cells increased in Sv-exp mice on day 2 post-infection, whereas these cell populations did not increase significantly in control mice (Figure 5A and (Figure S6). Because the early eosinophil accumulation could potentially be due to an increased responsiveness of ILC2s rather than only an increase in the number of these cells, next we compared the reactivity of ILC2s from the lungs of Sv-exp and control mice. The experience of S. venezuelensis infection significantly increased the proportion of IL-5/IL-13 single-and double-positive lung ILC2s in response to PMA and ionomycin stimulation (Figures 5B,C), indicating that ILC2s from Sv-exp mice have more cytokine-producing capacity compared with those from control mice.
We also analyzed the cytokine expressions in sorted ILC2s from the lungs of control mice or Sv-exp mice 2 days after N. brasiliensis infection. N. brasiliensis infection clearly increased the Il13 mRNA expression in ILC2s from Sv-exp mice compared with that in ILC2s from control mice (Figure 5D), indicating that the ILC2s in Sv-exp mouse lungs responded promptly and strongly to N. brasiliensis infection in vivo. To investigate the differences between ILC2s from Sv-exp and control mice, we analyzed the expression of various transcription factors and surface markers ( Figure 5E) and (Figure S7). We found that the expression level of Irf4 mRNA was significantly higher in ILC2s from Sv-exp mice compared with control mice, whereas the mRNA levels of Rora and Gata3, which are critical transcription factors for ILC2 development and function (23)(24)(25), were not different. The expression level of mRNA for TSLPR (Crlf2) was also enhanced in Sv-exp ILC2s compared with control ILC2s.
We did not detect any enhanced surface expression of IL-25R or ST2, suggesting that the enhanced ILC2 response was not caused by a change in the expression of receptors for IL-25 and IL-33. Together, these results demonstrate that the ILC2s in Svexp mice not only increased their number in the lungs but also changed their phenotype to that of highly responsive ILC2s. Since these ILC2s have been trained by S. venezuelensis infection to acquire a highly responsive phenotype, we call these cells "trained ILC2s."

IL-33 Is Not Sufficient to Induce Trained ILC2s or Protective Immunity Against Nippostrongylus brasiliensis Infection
We and other groups previously reported that IL-33 has a strong capacity to induce the proliferation and activation of ILC2s (31,40,41,44). To determine if IL-33 is sufficient for the induction and retention of trained ILC2s in the lung and for protection against N. brasiliensis infection, we administered IL-33 intranasally into WT C57BL/6 mice and analyzed various populations in the BALF cells weekly. An accumulation of ILC2s was induced promptly and intensively in IL-33-treated mice at day 7. At day 21, when the number of nematode infectioninduced ILC2s reached their population peak, the number of IL-33-induced ILC2s decreased. Finally, the number of ILC2s in IL-33-treated mice returned to the level of phosphate-buffered saline (PBS)-treated control mice at day 28 ( Figure 6A). To examine the capacity of ILC2s to produce cytokines, lung cells from PBS-treated, IL-33-treated, or Sv-exp mice were stimulated with PMA + ionomycin. S. venezuelensis infection induced hyperreactive ILC2s, whereas IL-33 treatment failed to enhance the responsiveness of ILC2s (Figures 5B,C, 6B,C). Similar results were also obtained following cytokine (IL-7 + IL-25 + IL-33) stimulation ( Figure S8). Furthermore, IL-33treated mice failed to be protected against N. brasiliensis infection ( Figure 6D) and failed to show accumulations of ILC2s and eosinophils in the lungs (Figure 6E). These differences following S. venezuelensis infection and IL-33 treatment suggest that the training of ILC2s requires some other stimuli in addition to IL-33.

DISCUSSION
Here, we clarified the immune modification effect of S. venezuelensis infection on a subsequent N. brasiliensis infection and determined its mechanism. Our data demonstrate that a host suffering from S. venezuelensis exposure is resistant to a subsequent N. brasiliensis infection. Our results showed that IL-33 as well as lung ILC2s and eosinophils are necessary for this resistance. Furthermore, we found that the lung ILC2s of Svexp mice gained a "trained" phenotype and had a rapid and high reactivity against N. brasiliensis infection. Together, our findings suggest that a host may be resistant to subsequent helminth infection within the weeks following a primary infection with other pulmonary migratory nematode due to the production of high cytokine levels by trained ILC2s waiting in the lung, which induce a rapid accumulation of eosinophils in response to IL-33.
Since mice that have been infected with S. venezuelensis develop immune memory, most S. venezuelensis larvae cannot migrate to the lungs during reinfection attempts (39). In agreement with previous findings, reinfection of mice by S. venezuelensis induced only weak Il33 mRNA expression compared with primary infection and was unable to induce Th2 cytokine expression despite the relatively high numbers of ILC2s remaining in the lungs of Sv-exp mice.
It is well-known that the main components of immune memory are antigen-specific T cells and B cells, but, in some cases, natural killer (NK) cells also can be antigen-specific memory cells (45). For example, antigen-specific memory NK cells have the capacity to protect DAP12-deficient neonatal mice from mouse cytomegalovirus infection (46). Furthermore, stimulation with the cytokines IL-12, IL-15, and IL-18 can induce long-lasting non-antigen-specific NK cells (47). In addition to NK cells, it has recently been clarified that myeloid cells, such as macrophages, have a mechanism to remediate pathogen exposure (48,49). These phenomena are described as "trained immunity" or "innate immune memory." We found that Sv-exp mice acquired specific-antigenindependent immunity; specifically, mice that experienced a S. venezuelensis infection gained significantly higher resistance to infection by N. brasiliensis. Similarly, mice that experienced a N. brasiliensis infection become resistant to infection by S. venezuelensis; our findings are consistent with a previous report by Baek et al. which suggested the contribution of eosinophils to this process (50). In the present study, we demonstrated the critical role of eosinophils in the resistance to subsequent infection by a different species of parasitic worm, and we further showed that this protection depends on IL-33 and ILC2s, but not on CD4 + cells, suggesting that mice acquired CD4independent "innate immune memory" or "trained immunity" following infection with S. venezuelensis. As the cytotoxic activity of eosinophils against N. brasiliensis larvae has already been demonstrated (51)(52)(53), it seems likely that the reduction in the number of worms in the small intestine is due to larval injury by eosinophils. However, we cannot completely exclude the contribution of acquired immunity to this protection because antibodies produced in response to the first infection may possibly bind to the larvae of N. brasiliensis via antigen crossreaction.
Our data suggest that the lung is an important site for the resistance to subsequent nematodes that use this organ in their life cycle. Previous work showed that macrophages activated by CD4 + T cells and ILC2s injure larvae in the lungs upon reinfection by N. brasiliensis (54). Additionally, neutrophils produce IL-13 and induce macrophages to develop into M2 macrophages, which are long-lived macrophages with parasite killing activity (55). Long-lived macrophages may remain in the lungs of S. venezuelensis-infected mice. Since eosinophils also have the ability to produce IL-4 and IL-13 (56,57), it is possible that the final effector cells are macrophages. However, at least in our experimental system, ILC2s and eosinophils were found to be essential for host defense against a second infection.
Guo et al. reported that Ascaris suum-infected mice were resistant to N. brasiliensis infection due to the activation of antigen-independent CD4 + Th2 cells by IL-33 stimulation (58). In their experiments, the mice were re-infected early after the first infection (day 11), and accumulated Th2 cells from the initial infection were likely to respond to the next infection. In contrast, our model examined infection after a longer period post-initial infection (day 28), and it seems that the contribution of ILC2s becomes higher than that of CD4 + T cells when the second infection occurs later.
During the initial infection with S. venezuelensis, the numbers of ILC2s and eosinophils in the lung increase after day 7 of infection. However, larvae pass through the lungs from day 1 to day 4 of infection (36), so eosinophils would need to accumulate in the lungs within 4 days for these cells to attack lung larvae. Here, we demonstrated that the number of eosinophils increased in Sv-exp mice beginning from day 2 of N. brasiliensis infection. This rapid increase in eosinophils is probably due to the increased number of ILC2s resulting from prior S. venezuelensis infection along with the ability of ILC2s to rapidly and strongly respond to N. brasiliensis infection.
IL-33, a critical cytokine for activation of ILC2s, is expressed in many types of cells including epithelial and endothelial cells and released in response to necrotic damages (59). In the lung, type II alveolar epithelial cells possess IL-33 even in the normal state and increase its expression by nematodes infection (31). Here we demonstrated IL-33 is actually essential for the activation of ILC2s to accumulate eosinophils. However, as the IL-33 expression before and after N. brasiliensis infection is comparable between control and Sv-exp mice, the increase of IL-33 expression is transient and the previous responses do not appear to be memorized.
Martinez-Gonzalez and colleagues reported that IL-33 induces memory ILC2s (60). They showed that intranasal treatment with IL-33 or papain induces memory ILC2s in mediastinal lymph nodes in mice, and these cells respond strongly to subsequent challenge with papain or IL-33, respectively. In our study, IL-33 was essential for the induction of highly reactive ILC2s; however, following IL-33 administration, the ILC2s in the lungs disappeared promptly after a transient proliferation. This result is consistent with their report that memory ILC2s are not in the lungs. However, highly reactive ILC2s are present in the lungs of S. venezuelensis-infected mice, and they respond strongly to N. brasiliensis infection. Nevertheless, when we infected IL-33treated mice with N. brasiliensis, we did not observe the resistance to subsequent nematode infection and increases in pulmonary ILC2s and eosinophils that occurred in Sv-exp mice. There are two differences in the experimental design between these two studies. First, we used a dose of IL-33 capable of inducing the same ILC2 level as that induced by nematode infection, which is lower than the dose used in the prior work. Second, different materials were used as the challenge. We employed a nematode as the challenge instead of papain. Our results suggest that the activation of ILC2s requires IL-33 but the induction of trained ILC2s likely requires some factors other than IL-33.
We found that ILC2s in Sv-exp mice showed enhanced expression of mRNA for TSLPR and IRF4. Mohapatra et al. reported that N. brasiliensis infection induced TSLP release in the BALF at 25 h and that TSLP/IL-33 stimulation of ILC2s induced the production of Th2 cytokines in an IRF4-dependent manner (61), suggesting the enhanced function of ILC2s in Sv-exp mice is due to their increased expression of these molecules. They also demonstrated that stimulation of TSLP with IL-33 induced IL-9 production by ILC2s (61). IL-9 has been shown to act autocrinally to maintain lung ILC2s and to play a role in tissue repair after infection with N. brasiliensis (29). Additional work is needed to determine why ILC2s remain for a long time after S. venezuelensis infection and why they show enhanced reactivity.
In conclusion, we have shown that S. venezuelensis-infected hosts acquire specific-antigen-independent "trained immunity" that contributes to a substantial resistance to pulmonary migratory nematodes. This is beneficial for host animals that live in areas where there is more than one parasitic infestation.