IL-17A both initiates (via IFNγ suppression) and limits the pulmonary type 2 immune response to nematode infection

Nippostrongylus brasiliensis is a well-defined model of type 2 immunity but the early lung-migrating phase is dominated by innate IL-17A production and neutrophilia. While the importance of IL-17A is well established during microbial infections, its contribution to type 2 immunity is poorly understood. Using N. brasiliensis infection we show that Il17a-KO mice exhibited an impaired type 2 immune response correlating with increased worm burden. Neutrophil depletion and reconstitution studies demonstrated that neutrophils contributed to the subsequent eosinophilia but were not responsible for the ability of IL-17A to promote type 2 cytokine responses. Transcriptional profiling of the lung on day 2 of N. brasiliensis infection revealed an increased IFNγ signature in the Il17a-KO mice confirmed by enhanced IFNγ protein production. Depletion of early IFNγ restored type 2 immune responses in the Il17a-KO mice supporting the hypothesis that the suppression of IFNγ by innate IL-17A sources promotes the protective type 2 immune response. Notably, when IL-17A was blocked later in infection, there was an increased type 2 response. Combined data from both N. brasiliensis and Trichuris muris, a gut-specific pathogen, demonstrated that IL-17A regulation of type 2 immunity was lung-specific. Together our data suggest that IL-17A is a major regulator of type 2 immunity in the lung which supports the development of a protective type 2 immune response but subsequently limits the magnitude of that response.


Introduction
Innate and adaptive sources of interleukin-17A (IL-17A) are responsible for a range of neutrophil-associated inflammatory conditions as well as protection from many bacterial and fungal pathogens [1] [2]. In contrast, type 2 immunity is required for effective control of most helminth infections [3] and is characterized by eosinophilic inflammation and the cytokines 60 IL-4, IL-5 and IL-13. When both type 2 and IL-17 responses are present during helminth infection enhanced pathology is observed, as shown for human schistosomiasis [4], [5] and onchocerciasis [6]. The detrimental relationship between IL-17A and type 2 associated diseases has also been extensively documented in allergic asthma in which the most severe symptoms occur in patients with both high Th2 and Th17 cell responses [7]. Critically, type 2 cytokines 65 can actively suppress IL-17A production which may be an important feedback mechanism to avoid extreme IL-17A-driven pathology [8]- [10]. Despite the evidence for an important relationship between IL-17A and type 2 immune responses during chronic disease, how these responses are connected remains poorly understood.

70
We and others have demonstrated a prominent role for IL-17A during infection with the lungmigrating nematode Nippostronglyus brasiliensis [8] [11], a well-defined pulmonary model of type 2 immunity. After entering the host via the skin, N. brasiliensis larvae migrate through the blood vessels into the lung, causing tissue damage and haemorrhage. IL-17RA-dependent neutrophil recruitment is largely responsible for the lung damage in this model [12]. We 75 previously found that the chitinase-like protein Ym1 induces expansion of IL-17A-producing γδ T cells and Ym1 blockade or IL-17A-deficiency protects mice from peak lung damage [8].
More surprising was our finding that Ym1 neutralisation or IL-17A-deficiency prevents the development of a full type 2 response during N. brasiliensis infection [8], [13].

80
The notion that IL-17A is required for development of a type 2 response appears counter to the evidence that type 2 cytokines suppress IL-17A production [9], [10]. However, previous studies using murine models of allergic inflammation also show impaired type 2 immunity in the face of IL-17A deficiency [14], [15] or blockade [16]. In an infection or injury context, the specific tissue as well as timing might all play decisive roles in whether IL-17A augments or suppresses 85 type 2 responses. We therefore used N. brasiliensis infection to address the contribution of γδ T cell-derived IL-17A and neutrophils to the development of a subsequent type 2 immune response in the lung. We found that IL-17A suppressed early IFNγ production and that this suppression was essential for the optimal development of a type 2 response. Although neutrophils acted as a major driver of subsequent lung eosinophilia, neutrophils were not 90 responsible for IL-17A-mediated enhancement of type 2 immunity. Once the type 2 response was established, IL-17A acted as a negative regulator, revealing distinct roles during innate and adaptive stages of the response. Notably, Trichuris muris, a nematode restricted to the gastro-intestinal tract also induced a lung type 2 response that was IL-17A-dependent.
However, we found no evidence that IL-17A regulated the intestinal type 2 response. Thus, IL-95 17A serves as a lung-specific regulator of the type 2 immune response.

Mice and ethics statement
For experiments using only WT mice, C57BL/6 J mice were obtained from Charles River. C57BL/6 Il17a Cre Rosa26 eYFP mice were originally provided by Dr Brigitta Stockinger [17], 105 [18]. For Il17a-KO experiments C57BL/6 WT mice and C57BL/6 Il17a Cre Rosa26 eYFP homozygote mice were bred at the University of Manchester. Mice were age-and sex-matched and all mice were housed in individually ventilated cages. Both males and females were used. Mice were not randomized in cages, but each cage was randomly assigned to a treatment group. Mice were culled by asphyxiation in a rising concentration of CO2. 110 Experiments were performed in accordance with the United Kingdom Animals (Scientific Procedures) Act of 1986.
At various time points mice were euthanised, bronchoalveolar lavage (BAL) was performed with PBS containing 1% BSA and lungs were taken for further analysis. For worm counts, the small intestines of infected mice were collected in PBS. Small intestines were then cut longitudinally along the entire length, placed in a 50 ml Falcon and incubated at 37°C for 120 4h. Settled worms were then counted with the aid of a dissecting microscope.

Flow cytometry.
Single-cell suspensions of the lung were prepared by digesting minced lung lobes for 30 min at 37 °C with 0.2 U/ml Liberase TL (Roche) and 80 U/ml DNase (Life Tech) in Hank's balanced-salt solution before forcing tissue suspensions through a cell strainer (70 µm, 125 Greiner). Red blood cells were lysed using Red Blood Cell Lysing Buffer Hybri Max (Sigma) for 3 min at RT and reaction was stopped by diluting samples in PBS. Total live cells were counted with AO/PI dye on an automated cell counter (Auto2000, Nexcelom).

Quantification of cytokines.
Single-cell suspensions of splenocytes, lung-draining lymph nodes or whole lung were stimulated ex vivo with N. brasiliensis excretory secretory product (E/S) antigen [20] (1 µg/ 150 ml) or anti-CD3 (1 µg/ml). Cell supernatants were harvested 72 h later and were stored at −20 °C until further analysis. Mouse IL-13 DuoSet ELISA kit (R&D Systems) was used for measurement of IL-13 levels. Mesenteric lymph node (MLN) cells from T. muris infected or uninfected mice were collected, cultured and restimulated ex vivo for 36 h with E/S as previously described [21]. The concentrations of IL-5, IL-6, IL-9, IL-10, IL-13, IL-17A, TNFα 155 and IFNg in the MLN culture supernatant were measured by cytokine bead array (CBA, BD Biosciences, UK) as per the manufacturer's protocol.

Antibody depletion and cell transfer experiments
Neutrophils were depleted via intraperitoneal injection of anti-Ly6G antibody (clone 1A8, BioXcell, West Lebanon, NH, USA) (500 µg/mouse/day) on days -1 and 1 of infection 160 with N. brasiliensis. Control mice were injected with an equal volume of IgG2a isotype control (BioXcell). Neutrophil depletion was confirmed by anti-Gr1 (RB6-8C5) staining for flow cytometry. For neutrophil transfer, neutrophils were purified from bone marrow using Histopaque-based density gradient centrifugation as described before [22]. Collected neutrophils were washed twice with RPMI 1640 and counts and viability was determined 165 (average purity 85% ± 4.42% and 90% ± 1.5% viability). Neutrophils were resuspended in PBS and 0.5 x 10 6 cells (40µl) transferred intranasally into mice on day 1 post infection with N. brasiliensis. IFNg was depleted using an anti-IFNg monoclonal antibody (clone XMG1.2) and injected intraperitoneally (500 µg/mouse/day) on days -1 and 1 of infection with N.

Histology
For histology, the left lobe of the lung was isolated at different time points following N. 175 brasiliensis infection, inflated and fixed in 10% neutral-buffered formalin (Sigma). Whole left lung lobes were processed using a tissue processor (Leica ASP300S) and were then embedded in paraffin. Paraffin blocks were then sectioned to 5 µm using a microtome (Leica RM2235) and routinely stained with Mayer's haematoxylin (Merck Millipore HX86014349) and eosin (Sigma) for histological analysis. Slides were imaged using an Olympus slide scanner 180 and high-resolution image files were exported using Pannoramic Viewer software (3DHISTECH). The images were then processed in a KNIME software workflow to obtain 50 random regions of interests (ROIs) across the whole lung section. ROIs that contained lobe boundaries or extensive artefacts were excluded from the analysis. The ROIs were then converted to binary images and lacunarity (L) was quantified using the FracLac plugin for 185 ImageJ. The L values of all the ROIs were averaged to obtain estimates for the entire lobe.

Extraction of RNA and quantitative real-time PCR
A fragment of the right lung lobe was stored in RNAlater (Ambion) before homogenization of tissue in Qiazol reagent with a TissueLyser (Qiagen). RNA was prepared according to 190 manufacturer's instructions. RNA was quantified using a ND-1000 Spectrophotometer (NanoDrop Technologies). Reverse transcription of 1 µg of total RNA was performed using Tetro reverse transcriptase (Bioline). For reverse transcription, total RNA was treated with 50 U Tetro reverse transcriptase (Bioline), 40 mM dNTPs (Promega), 0.5 µg PolyT primer for cDNA synthesis (Roche) and RNasin inhibitor (Promega). The abundance of transcripts 195 from the genes of interest was measured by quantitative real-time PCR with the Light Cycler 480 II system (Roche) with a Brilliant III SYBR master mix (Agilent) and specific primer pairs. PCR amplification was analysed by the second-derivative maximum algorithm (Light Cycler 480 Sw 1.5; Roche), and expression of the gene of interest was normalized to that of the housekeeping gene Actb (beta-actin). A list of primer sequences used are shown in Table  200 2.

Trichuris muris infection and E/S products 205
T. muris eggs were prepared from chronically infected stock mice as described previously [23]. Individual worms were then counted by eye under a binocular dissecting microscope. T. muris 210 adult excretory secretory product antigen (E/S) was prepared as described by [23]. In brief, adult T. muris were cultured ex vivo at 37 o C, the culture supernatant was collected and centrifuged to remove eggs and worms. The resultant supernatant was then filter sterilised and stored at -20 o C until use for in vitro re-stimulation of MLN cells.

Nanostring RNA Profiling 215
Samples were run on an Agilent 2200 Tape Station system to ensure high quality lung RNA; samples with a RIN value of < 6.5 were excluded. RNA was diluted to 20ng/μL in RNase free H2O, measured using Qubit™ RNA HS Assay Kit (Thermofisher) and run on a Nanostring nCounter® FLEX system using the Myeloid Innate Immunity v2 panel (XT-CSO-MMII2-12) as per manufacturer's instructions. Raw data was loaded into nSolver version 4.0 using default 220 settings. non-normalised counts were then exported and subsequent analyses were performed in R (version 3.6) [24]  Heatmaps were then generated from normalized counts of differentially expressed genes using the ComplexHeatmaps R package [26]. The networks and functional analyses of differentially expressed genes were generated with Ingenuity pathway analyser (QIAGEN Inc., 230 https://www.qiagenbio-informatics.com/products/ingenuity-pathway-analysis). Data were then imported into R for visualisation [27].

Statistical analysis.
Prism 7.0 (version 7.0c, GraphPad Software) was used for statistical analysis. Differences 235 between experimental groups were assessed by ANOVA (for normally distributed data, tested using Shapiro-Wilk normality test) followed by Sidak's multiple comparisons test. For gene expression data, values were log2 transformed to achieve normal distribution. Comparisons with a P value of less than 0.05 were considered to be statistically significant.

IL-17A-deficient mice mount a diminished type 2 response
In keeping with the known ability of N. brasiliensis to induce a strong pulmonary type 2 immune response on day 6 post infection (d6pi), we found the BAL and lungs of C57Bl/6 mice to be dominated by eosinophils (Suppl. Fig. 1A, B). This response was accompanied by 245 elevated numbers of CD4 + T cells as well as induction of Group 2 Innate lymphoid cells (ILC2, Suppl. Fig. 1C, D). The establishment of a type 2 response was further confirmed by increased  Fig.1A), which, together with N. brasiliensis larvae migration, is known to cause acute lung injury responses [11].
To investigate the role of IL-17A during the development of type 2 immune responses, we infected Il17a-KO mice and corresponding WT C57BL/6J controls with N. brasiliensis L3's. 255 Larvae leave the lung within 48 h and are expelled from the gut within 6-8 days. Consistent with our previous findings [8] Il17a-KO mice were significantly more susceptible to infection exhibiting an intestinal worm burden almost twice as high as in the WT controls on d4pi and d6pi Fig. 1B). As expected, the early d2 neutrophilia in response to N. brasiliensis infection was muted in the Il17a-KO mice relative to the WT controls (Fig. 1C). Between day 2 and 6 260 post infection, there was a switch from neutrophilic to eosinophilic responses in the lungs (Suppl. Fig 1A). Whilst increased eosinophil numbers were observed in all infected d6pi mice relative to naïve animals, this increase was less evident for Il17a-KO mice (Fig. 1D). ILC2 displayed a similar pattern, as cell numbers significantly increased with infection in WT but not in Il17a-KO mice (Fig. 1E). Additionally, WT mice exhibited an increased type 2 cytokine 265 expression signature at d6pi, whilst Il17a-KO mice had reduced Il4 expression and a significantly delayed increase in Il5 expression (Fig. 1F). IL-13 protein levels in whole lung samples restimulated with a-CD3 on d6pi were significantly higher in WT mice compared to naïve controls whilst Il17a-KO mice completely failed to upregulate IL-13 levels ( Fig. 1G), further demonstrating an impairment of protective type 2 responses. We also measured 270 expression levels of the major mucins in the lung because host mucin production is another feature of anti-helminth protective type 2 responses [28]. N. brasiliensis infection drove an early increase in mucins Muc5ac and Muc5b expression in the lungs of WT mice at day 2 postinfection corresponding to a time when the larvae are transitioning from the lungs. Expression of the mucin Muc5ac was significantly increased in WT mice compared to naïve controls, but 275 reduced in Il17a-KO mice, although not significant (Fig. 1H). A similar pattern was observed for Muc5b expression, which was upregulated in infected WT mice by day 2, but failed to increase in Il17a-KO infected mice (Fig. 1H). By d6pi, when the worms had already migrated to the small intestines, increased mucin expression was no longer observed in infected animals ( Fig. 1H). Together these data demonstrate an impairment of type 2 immune responses during 280 helminth infection in the absence of IL-17A.

IL-17A regulates T cell activation and polarization during N. brasiliensis infection
Next we aimed to determine whether the impact of IL-17A-deficiency on type 2 immunity was due to changes in T cell activation or polarisation during N. brasiliensis infection. Using flow 285 cytometry, we observed a reduction in total numbers of CD4 + T cells on d7pi in the lungs of Il17a-KO mice compared to WT controls ( Fig. 2A). In contrast, there were no significant differences in CD4 + T cell numbers in the lung-draining lymph nodes ( Fig. 2A).

Not only were there fewer CD4 + T cells in the lungs of Il17a-KO mice, the CD4 + T cells of
Il17a-KO mice were producing significantly less IL-13 and IL-5 (Fig. 2B). We also found 290 significantly fewer IL-5 + and IL-13 + CD4 + T cells in the Il17a-KO mice compared to WT controls ( Fig. 2C). At the same time point post-infection, we found that expression of the activation marker CD69 was upregulated in the lungs of WT but not Il17a-KO mice (Fig. 2D).
However, CD69 did not differ between all tested groups in the lung-draining lymph nodes (Fig.   2E). Recently, Minutti et al. described a role for epidermal growth factor receptor (EGFR) on 295 T cells in protection against N. brasiliensis. EGFR forms a complex with ST2 to allow for IL-33-induced IL-13 expression at the site of infection [29]. We therefore analysed surface expression of EGFR and ST2 on lung and lung-draining lymph node T cells. N. brasiliensis infection increased the number of CD4 + T cells expressing these markers in the lung, but this increase was significantly reduced in Il17a-KO mice. (Fig. 2D). Changes to ST2 and EGFR 300 expression between WT and KO mice were not observed in the lung-draining-lymph nodes ( Fig. 2E). We also measured PD-1 expression, an important regulator of T cell function during helminth infection [30], [31]. N. brasiliensis infection in WT mice led to increased numbers of CD4 + T cells expressing PD-1 in the lung (Fig. 2D) with significantly fewer of these cells in Il17a-KO mice (Fig. 2D). Whilst there were also increased numbers of PD-1 + CD4 + T cells in 305 the lymph node following infection of WT mice, these were not altered by the absence of IL-17A (Fig. 2D). Overall, these data demonstrate the lung-specific impact of IL-17A-deficiency on CD4 + T cell numbers, type 2 cytokine production and activation status.

Neutrophils regulate eosinophil recruitment but not type 2 responses 310
IL-17A is a major driver for neutrophil recruitment and activation [32], raising the possibility that impaired neutrophilia in Il17a-KO mice is responsible for diminished type 2 immune responses. This hypothesis was supported by data indicating that neutrophils from N. brasiliensis infected mice can express type 2 cytokines [11]. Therefore, we assessed whether neutrophils contributed to the type 2 immune response by depleting neutrophils during N. 315 brasiliensis infection (Fig. 3A). Injection of anti-Ly6G effectively prevented neutrophil accumulation in the BAL at d2pi compared to isotype control and depletion was still evident in the blood at d6pi (Fig. 3B, Suppl. Fig. 2A). Neutrophilia at day 2 post N. brasiliensis infection is a major driver of the lung injury [11] and here we confirmed these findings (Fig.   3C). Histological sections demonstrated infection-induced injury at d2pi in isotype-treated WT 320 mice, an effect that is almost absent in infected mice depleted of neutrophils. Using lacunarity [33] as an indicator of acute lung injury [34], we show that neutrophil depleted mice show no significant signs of damage (Fig. 3C). Whilst intestinal worm burdens at d2pi and d4pi were not affected by neutrophil depletion, a small but significant increase in parasite numbers at d6pi was observed in mice treated with neutrophil-depleting antibody (Fig. 3D). As expected, 325 eosinophils were prominent in the lung and BAL of isotype-treated mice at d6pi (Fig. 3E).
However, depletion of neutrophils in infected mice significantly reduced eosinophils in the lung and BAL compared to the isotype-treated animals. Notably, blood and bone marrow eosinophils were not affected, suggesting a defect in eosinophil recruitment (Supplement Fig.   2B). We therefore assessed chemokines with known eosinophil chemotactic effects (Ccl5, 330 Ccl11, Ccl22), but neutrophil-depletion did not alter mRNA expression of these chemokines in the lung. In addition, we measured Ccl8 as Chen et al. demonstrated that N. brasiliensisprimed neutrophils from d2pi exhibit a notable upregulation of Ccl8 [11]. Expression of Ccl8 was increased in infection, but significantly lower in neutrophil-depleted mice (Suppl. Fig. 2C).
As expected, N. brasiliensis infection significantly increased CD4 + T cells in the lung, but the 335 number of CD4 + T cells were significantly reduced following neutrophil depletion (Fig. 3F).
Despite effects on both CD4 + T cells and eosinophilia, there were limited changes in type 2 cytokine expression following neutrophil depletion (Fig, 3G, H). Whilst IL-4 levels significantly increased in infected isotype treated mice, there was no significant upregulation of IL-4 expression in whole lung or by CD4 + T cells in neutrophil-depleted mice (Fig. 3G-H). 340 IL-13 and IL-5 were not also altered (Fig. 3G, H, Suppl. Fig. 2D). Thus, neutrophil depletion prevented eosinophil recruitment to the lungs of infected mice, but had a limited effect on type 2 cytokine expression.
To further address the role of neutrophils, we tested whether transfer of WT neutrophils into Il17a-KO mice could rescue the defect in the development of type 2 immunity. Neutrophils 345 were isolated from the bone marrow of CD45.1 + mice using the histopaque-gradient method [22] and intranasally transferred into Il17a-KO mice (Fig. 3I). Neutrophil transfer restored airway neutrophil frequency in Il17a-KO mice (Fig. 3J). Interestingly, this neutrophilic response was due to increased recipient-derived cells and not the transferred CD45.1 + neutrophils (Suppl. Fig. 2E). One day post-neutrophil transfer (on d2pi), eosinophil numbers 350 were increased in Il17a-KO mice, reaching similar numbers to those observed in infected WT mice (Fig. 3K). Ccl8 levels were also increased following neutrophil transfer (Suppl. Fig. 2F), consistent with the possibility that this chemokine plays a role during eosinophil recruitment to site of infection. At a later timepoint, when eosinophilia was much more dominant and pronounced, the Il17a-KO mice that received neutrophils also displayed an increased lung 355 eosinophilia (Fig. 3L). Despite changes to eosinophil numbers, transfer of neutrophils did not rescue the deficit in type 2 cytokines in Il17a-KO mice (Fig. 3M). Together, this data provides evidence that neutrophils support eosinophil recruitment at the site of infection, but suggests they do not directly contribute to type 2 cytokine responses.

IL-17A leads to a downregulation of early IFNg during Nippostronglyus infection
Rapid early IL-17A production is known to be critical for protective immune responses in different settings of lung immunity [1], [35]. To better understand the early events unfolding in the lung during N. brasiliensis infection, we performed a Nanostring gene expression array WT mice (Fig. 4B). This led us to hypothesize that IL-17A may suppress IFNg, which would facilitate Th2 cell development and explain why mice deficient in IL-17A cannot induce a full type 2 immune response. This hypothesis was also consistent with our unpublished and published [36] finding that Ym1, which induces IL-17A, strongly suppresses IFNg. To test this possibility, we assessed IFNg responses in Il17a-KO mice after N. brasiliensis infection. While 375 WT mice exhibited significant suppression of Ifng in whole lung at 2dpi compared to uninfected controls, mice deficient in IL-17A did not show this phenotype (Fig. 4C). By intracellularly staining for IFNg, we observed that Il17a-KO mice infected with N. brasiliensis failed to exhibit the early downregulation of IFNg expression seen in infected WT mice (Fig.   4D). Importantly, this failure of suppression was observed across different types of IFNg-380 producing cells. Lung CD4 + T cells, CD8 + T cells as well gd T cells from Il17a-KO mice produced IFNg at the same proportion as the naïve controls, while WT mice exhibited a downregulation in IFNg production (Fig. 4D). We further assessed whether early IFNg was produced by gd T cell subsets that differ in their expression of CD27 and CD44 [37]. At 16 h post N. brasiliensis infection, IFNg frequencies were also significantly increased in gd T cells 385 of Il17a-KO mice compared to WT controls (Fig. 4E). Consistent with expectations [37], the CD27 + gd T cells were the main producers of IFNg after infection (Fig. 4F). Overall, our data demonstrated that IL-17A-deficiency enhanced IFNg production during infection, supporting the hypothesis that IL-17A plays an important role in downregulating IFNg at the site of infection during the lung migratory phase of N. brasiliensis infection. 390

IFNg neutralization in Il17a-KO mice rescues the impaired type 2 immune response
To directly test whether active suppression of IFNg by IL-17A is required for the full development of a type 2 immune response during N. brasiliensis infection, IFNg was neutralised at day -1 and 1 of infection in Il17a-KO mice and responses examined at day 8 395 post-infection. (Fig. 5A). The significant defect in eosinophilic responses in Il17a-KO mice compared to WT mice was still evident at day 8 post-infection. However, blocking IFNg in Il17a-KO mice enhanced eosinophil numbers (Fig 5B). The same pattern was observed for numbers of CD4 + T cells in the lungs (Fig. 5C). To determine whether IFNg neutralisation altered the type 2 response, we firstly assessed expression of key type 2 cytokines Il4 and Il13 400 in the lungs of mice. Whilst expression of these genes was significantly reduced in infected Il17a-KO mice compared to WT mice, IFNg depletion completely recovered expression of these cytokines in Il17a-KO mice compared to isotype-treated animals, and Il4 surpassed the levels seen in WT mice (Fig. 5D). Additionally, we assessed the type 2 marker, Chil3, which also demonstrated enhanced expression in Il17a-KO mice treated with anti-IFNg compared to 405 similarly treated WT mice (Fig. 5E). Similarly, analysis of numbers of IL-5 and IL-13 producing CD4 + T cells at d8 post infection showed restoration of the type 2 response in Il17a-KO mice that received the neutralising IFNg antibody (Fig. 5F). Consistent with the ability of IFNg to regulate type 2 cytokines, IFNg depletion also restored the activation status of CD4 + T cells in the Il17a-KO mice (Fig. 5G) and increased the numbers of CD4 + T cells expressing 410 type 2 markers EGFR, PD1 and ST2 (Fig. 5H). Again, no effect was observed in these parameters in IFNg-depleted WT mice. Here, we show that an initial reduction in IFNg levels during N. brasiliensis infection mediated by IL-17A, allows the subsequent development of type 2 immunity in the lung.

IL-17A suppresses an established type 2 response in the lung
IFNg depletion in Il17a-KO mice not only restored the type 2 response, but in some cases exceeded WT levels. Therefore, we hypothesized that IL-17A may suppress the type 2 immune response during later stages of N. brasiliensis infection. To test this hypothesis, we neutralized IL-17A at d4pi, d5pi and d6pi in WT mice and assessed immune responses at d7pi (Fig. 6A). 420 Blocking of IL-17A led to a significant increase in both ILC2 numbers and frequencies in the lung (Fig. 6B), as well as the numbers of ILC2s producing IL-5 and IL-13 (Fig. 6C). However, CD4 + T cell numbers in the lung were comparable between isotype-treated and anti-IL-17Atreated WT mice (Fig. 6D). However, the ability of CD4+ T cells to produce type 2 cytokines may partly rely on IL-17A, as mice administered anti-IL-17A showed a slight increase IL-5 425 and IL-13 (Fig. 6E). This data demonstrates that IL-17A can have differential effects depending on the time and status of infection. While early IL-17A promotes the type 2 response, later in infection IL-17A acts to suppress and limit excessive the type 2 immunity, particularly in ILC2s.  [39]. Here, worm counts in the caecum of Il17a-KO mice and WT mice were comparable at d19pi and d32pi (Fig. 7A), indicating IL-17A does not alter parasite expulsion rate. Cell numbers in the caecum were analysed and no differences in eosinophil and neutrophil frequency were observed between Il17a-KO mice and WT controls on d19pi and d32pi 445 (Supplement Fig. 3A, B), suggesting an IL-17A-independent recruitment mechanism for both these cell types. CD4 + T cell numbers in the MLN were also comparable on d19pi and d32pi between Il17a-KO mice and WT controls (Suppl. Fig 3C). Although there was an induction of type 2 cytokines in infected mice as measured by intracellular cytokine staining, the numbers of IL-5 or IL-13-producing CD4 + T cells in the MLNs did not significantly differ between the 450 groups (Suppl. Fig 3D). Secreted levels of IL-5, IL-9 and IL-13 in MLN cells also did not show any significant differences between Il17a-KO mice and WT controls (Suppl. Fig 3E). Together, these data failed to provide any evidence that IL-17A was an important regulator of type 2 immunity in the intestine during T. muris infection.

455
Previous studies have shown that despite the restriction of the T. muris lifecycle to the gastrointestinal tract of the mammalian host, evidence of a type 2 immune response can be observed at distant sites, such as the lung [40] . Therefore, the immune response in the lung of T. muris infected WT vs Il17a-KO mice at d19pi and d35pi was assessed. Neutrophils numbers were increased in infected WT animals at d19pi and d35pi but this was significantly reduced in 460 Il17a-KO mice on d19pi (Fig. 7B). No significant changes were observed for eosinophils (Fig.   7C). Whilst lung CD4 + T cell numbers in infected animals did not change compared to naïve controls (Fig. 7D), Il17a-KO mice had significantly fewer IL-5 + CD4 + T cells at d19 and d32pi compared to WT controls (Fig. 7E). Although the effect on IL-13 + CD4 + T cells was less evident, infected Il17a-KO mice failed to significantly increase numbers of Il-13 + CD4 + T cells 465 compared to uninfected controls (Fig. 7E). Supporting the intracellular cytokine staining, qRT-PCR analysis in whole lung tissue showed an impairment of type 2 cytokines in the Il17a-KO mice, with significantly decreased expression of Il4 (d32pi) and Il5 (d19pi) (Fig. 7F). Similar to infection with N. brasiliensis, we also observed an upregulation of IFNg in the lung during T. muris infection in Il17a-KO mice. Both the number and the frequency of IFNg + CD4 + 470 T cells in the lung were significantly increased in Il17a-KO compared to WT infected mice on d19pi (Fig. 7G). This data utilising T. muris infection models suggests that IL-17A-dependent suppression of IFNg allows promotion of the type 2 immune response specifically in the lungs but not the intestine.

475
Discussion IL-17A, the key cytokine of the IL-17 family, is central to barrier immunity, combating fungal infections and inducing antimicrobial proteins as well as neutrophil activating and recruiting chemokines [2]. Here, we explored the relationship between IL-17A and type 2 immunity 480 during N. brasiliensis infection and show that IL-17A, by impinging on IFNg expression, allows type 2 immunity to fully develop within the lung.
As expected, IL-17A was important for innate neutrophilic responses during N. brasiliensis infection. However, neutrophil depletion and transfer experiments demonstrated that 485 neutrophils themselves only partly contributed to the development of type 2 immunity, mainly through enhancement of the subsequent eosinophilia. Our data highlighted a potential role for the chemokine Ccl8 in neutrophil mediated recruitment of eosinophils. A correlation between CCL8 and eosinophil numbers in allergic airway inflammation has been described previously [41] and CCL8 was associated with type 2 inflammatory responses in the lung during the 490 protective response to Klebsiella pneumonia [42]. However, neither study demonstrated a direct eosinophil chemotactic function of CCL8. More recently, Puttur et al. revealed that CCL8 is important for the appropriate mobilisation of ILC2s and specifically the localised production of IL-5 by ILCs, possibly explaining why we observed reduced eosinophilia in neutrophil depleted mice [43]  confirming early neutrophilia as a main driver of acute lung injury following N. brasiliensis infection [11].
The ability of type 2 cytokines to suppress IL-17 has been well documented [47]- [49], but our study suggests that the development of the type 2 response in the first place can require IL-505 17A. Our data is supported by a model of murine atopic dermatitis, in which Nakajima et al.
show that IL-4 production is reduced by IL-17A deficiency and that in vitro Th2 differentiation from naïve T cells is enhanced with the addition of IL-17A [14]. Notably, the IL-17A-induced IL-4 [14] may subsequently suppress IL-17A production as a feedback mechanism, consistent with the previous observations that type 2 cytokines negatively regulate IL-17A [50]. In 510 allergic airway inflammation models, deficiency of IL-17A also leads to decreased type 2 cytokine production [15] and Th17 cells contribute to Th2-cell-mediated eosinophilic airway inflammation [51]. A combination of type 2 cytokines and IL-17A can be also found as signature for severe disease pathology. For example in a model for severe airway hyperresponsiveness, IL-17A contributes to asthma pathology by enhancing IL-13 activity 515 [52]. Furthermore, during murine and human schistosomiasis and onchocerciasis, a dysregulated balance between IL-17A and type 2 responses can exacerbate the pathology [4]- [6], [53].
In our effort to understand how IL-17A might be required for full type 2 immunity, we 520 identified a role for IL-17A mediated suppression of IFNg. Evidence that IL-17A can regulate IFNg exists in the literature, but whether IFNg is upregulated or downregulated is controversial and likely depends on the setting, timing and location [54]- [56]. During experimental visceral leishmaniasis, Il17a-KO mice have enhanced production of a protective IFNg response [54], while in Toxoplasma gondii infection, excessive IFNg in Il17a-KO mice has severe 525 pathological consequences [55]. Evidence also exists in the context of helminth infection, where a lack of IL-17A drives elevated IFNg during infection with the filarial nematode Litomosoides sigmondontis [57] or Schistosoma japonicum [56] and Schistosoma mansoni [5].
Together these findings support a potentially widespread role for IL-17A in regulating IFNg.

530
Our study shows that early in infection, IL-17A suppression of IFNg in the lung has consequences for the subsequent development of type 2 immune responses. Notably, in Il17a-KO mice the protective type 2 immune response in the small intestine was not impaired as mice were still able to expel N. brasiliensis. To examine whether IL-17A-dependent regulation of IFNg is site specific, we used high dose infection with the colon dwelling parasite T. muris. 535 The protective type 2 immune response against T. muris in the intestine was intact in mice lacking IL-17A. The surprising finding however was that even though T. muris does not have a lung stage, the concurrent type 2 response in the lung was impaired in Il17a-KO mice. CD4 + T cells in the lung produced less type 2 cytokines and both eosinophils and neutrophils were significantly downregulated in the Il17a-KO mice. Consistent with our findings in N. 540 brasiliensis, CD4 + T cells in Il17a-KO mice produced significantly higher amounts of IFNg than in their WT counterparts. Together, our data suggest that the impact of IL-17A on type 2 development may be lung restricted. Nonetheless, there may still be a fundamental requirement for IFNg suppression for type 2 immunity to progress. Artis et al. demonstrated that the type 2 immune response during T. muris requires TSLP, and in very similar experiments to those 545 described here, demonstrated that TSLP functions to suppress IFNg [58]. Thus, early suppression of IFNg may be a general pre-requisite for the development of a type 2 environment with a requirement for IL-17A in the lung and TSLP (or other factors) in the gut.
Our data also reveal a potential time dependence on the regulation of type 2 immunity by IL-550 17A. While the early gd T cell-derived IL-17A supported the type 2 response, late IL-17A, derived from both Th17 cells and gd T cells, appeared to be a negative regulator. We observed that depletion of early IFNg in Il17a-KO mice not only rescued their impaired type 2 immune responses on d8pi, but the type 2 mRNA signature significantly exceeded the WT levels. Thus, removal of IFNg in IL-17A deficient mice, which allowed normal type 2 progression, revealed 555 a potential role for IL-17A in suppressing type 2 cytokines. To our knowledge, IL-17A suppression of type 2 cytokines has not previously been described in vivo.
We have not yet addressed the full mechanism behind IL-17A-mediated suppression of IFNg during N. brasiliensis infection but it is notable that IL-17A not only impairs type 2 cytokine 560 production, but also alters the cellular activation status and expression of type 2 markers.
Interestingly, in our model, we only observe impairment of type 2 immune responses in the lung itself and not in the Th2 cells from lung-draining lymph nodes. Expression of EGFR and ST2, two markers closely associated with type 2 settings [29], were reduced on the CD4 + T cells of Il17a-KO mice in the lung. EGFR expression on Th2 cells is critical for resistance 565 during GI helminth infection and a signalling complex between EGFR and ST2 can activate Th2 cells to secrete IL-13 in an antigen-dependent manner upon IL-33 exposure. Our data would suggest that this "licensing" of Th2 cells does not occur in the Il17a-KO mice during N.
brasiliensis infection, indicating that IL-17A is needed for a proper induction of the adaptive Th2 response in the lung. 570 Together we demonstrate that the early events in the lung shape the protective type 2 immune response, with IL-17A as a critical regulator through the downregulation of IFNg. In combination with previous data [58], suppression of IFNg at barrier sites may be a central paradigm for type 2 immunity. The ability of IL-17A to subsequently suppress type 2 responses reveal an important feedback loop that must go awry during severe asthma and other 575 type 2 conditions, in which IL-17 plays a damaging and pathogenic role.

Disclosures 585
The authors have no financial conflicts of interest. Data was tested for normality using Shapiro-Wilk test and analysed using One-Way Anova followed by Sidak's multiple comparisons test for selected groups. NS not significant, *P<0.05, **P<0.01. parameters were investigated at d19 and d32pi. Worms counts in the caecum (A). Frequency of neutrophils (B), eosinophils (C) and CD4 + T cells (D) in the lung at d19pi and d32pi. Absolute counts of IL-5 + and IL-13 + CD4+ T cells per gram of lung tissue (E). Relative expression of cytokines Il4 and Il5 from whole lung RNA (log2 expression relative to actb (βactin)) (F). Absolute cell counts and frequency of IFNγ by CD4 + T cells per gram of lung tissue 985 (G). Data are expressed as mean ± s.e.m.) and are representative of 3 individual experiments with at least 4 mice per infected group and one mouse per control group. Data was tested for normality using Shapiro-Wilk test and analysed using One-Way Anova followed by Sidak's multiple comparisons test for selected groups or student's t-test. *P<0.05, **P<0.01