Lung eosinophils elicited during allergic and acute aspergillosis express RORγt and IL-23R but do not require IL-23 for IL-17 production

Exposure to the mold, Aspergillus, is ubiquitous and generally has no adverse consequences in immunocompetent persons. However, invasive and allergic aspergillosis can develop in immunocompromised and atopic individuals, respectively. Previously, we demonstrated that mouse lung eosinophils produce IL-17 in response to stimulation by live conidia and antigens of A. fumigatus. Here, we utilized murine models of allergic and acute pulmonary aspergillosis to determine the association of IL-23, IL-23R and RORγt with eosinophil IL-17 expression. Following A. fumigatus stimulation, a population of lung eosinophils expressed RORγt, the master transcription factor for IL-17 regulation. Eosinophil RORγt expression was demonstrated by flow cytometry, confocal microscopy, western blotting and an mCherry reporter mouse. Both nuclear and cytoplasmic localization of RORγt in eosinophils were observed, although the former predominated. A population of lung eosinophils also expressed IL-23R. While expression of IL-23R was positively correlated with expression of RORγt, expression of RORγt and IL-17 was similar when comparing lung eosinophils from A. fumigatus-challenged wild-type and IL-23p19-/- mice. Thus, in allergic and acute models of pulmonary aspergillosis, lung eosinophils express IL-17, RORγt and IL-23R. However, IL-23 is dispensable for production of IL-17 and RORγt.


Introduction
Fungi of the genus Aspergillus are ubiquitous in the environment [1]. It is estimated that individuals typically inhale hundreds of conidia a day. While in the immunocompetent host this is generally without consequence, in individuals with dysregulated immune systems, the conidia may germinate into hyphae and cause a wide spectrum of diseases ranging from invasive aspergillosis in severely immunocompromised hosts to allergic aspergillosis in atopic individuals [2]. Allergic aspergillosis comprises an overlapping set of diseases including severe asthma with fungal sensitization, allergic bronchopulmonary aspergillosis (ABPA) and chronic pulmonary aspergillosis. The burden of invasive and allergic aspergillosis is estimated at over 200,000 and 5,000,000 persons per year, respectively [3,4]. Only a few of the hundreds of described Aspergillus species regularly cause human disease with A. fumigatus being the most common.
Clinical and experimental studies point to critical roles for multiple cell types in orchestrating innate and adaptive host defenses against pulmonary aspergillosis. Consistent with the wide spectrum of manifestations of aspergillosis, the relative importance of each cell type is dependent upon the model of aspergillosis being studied. It is difficult to discern the role of eosinophils in human invasive aspergillosis; while risk factors such as chemotherapy and corticosteroids deplete eosinophils, they also depress other arms of the immune system. However, in chronic cases of human IPA featuring less immunosuppression, eosinophils are frequently found [5]. We and others have examined the contribution of eosinophils in mouse models of aspergillosis. Eosinophils release extracellular traps [6] and have potent antifungal activity in vitro [7]. The susceptibility to acute infection with A. fumigatus has been studied in mice lacking eosinophils; two studies found increased susceptibility [7,8], whereas one study found decreased susceptibility [9]. The reasons for the disparity are speculative but could be related to different fungal strains used in the models. Data from a variety of models, including those featuring eosinophil-deficient mice, provide strong evidence that eosinophilic inflammation in the airways is pathogenic in allergic pulmonary inflammation [7,10,11].
The antimicrobial and immunoregulatory activity of eosinophils is mediated in large part by effector molecules stored in their primary and secondary granules, including eosinophil peroxidase (EPX) and major basic protein (MBP). Importantly, eosinophils pre-synthesize cytokines and store them in their granules, enabling rapid secretion upon stimulation [12]. Over 35 cytokines, chemokines and growth factors have been identified in at least some eosinophil populations [12,13].
The IL-17 cytokine family is comprised of IL-17A, IL-17B, IL-17C, IL-17D, IL-17E (also known as IL-25) and IL-17F [14]. IL-17A and IL-17F have the most similarity and exist as disulfide linked homodimers and heterodimers. IL-17 family members mediate their biological activity by binding to receptors for IL-17 (IL-17R), of which there are at least five members [14]. Notably, IL-17A, IL-17F and IL-17AF bind to a heterodimeric complex comprised of IL-17RA and IL-17RC. In this manuscript, IL-17 is used to refer to IL-17A. A preponderance of experimental mouse data and human clinical observations have established dichotomous roles for IL-17. For example, while IL-17 generally protects against certain fungal and bacterial infections, it can mediate pathology in allergic and autoimmune diseases [15,16]. Case reports have associated deficient IL-17 responses with invasive aspergillosis [17,18]. However, the contribution of IL-17 to host defenses against aspergillosis has not been well studied.
Originally described as a product of Th17 cells, subsequent studies established that other lymphoid subsets also produced IL-17, including CD8 + T cells, γδ T cells, natural Th17 cells, natural killer T cells and innate lymphoid cells [14]. Induction of lineage fate of IL-17-producing lymphoid cells requires the master transcription factor, RORγt (the gene product of Rorc), and Rorc knockout mice generally phenocopy IL-17-deficient mice [19]. However, RORγt also promotes survival of CD4 + , CD8 + double positive thymocytes [20]. The heterodimeric cytokine, IL-23, promotes the maturation and expansion of Th17 cells and other IL-17-producing lymphocytes [21]. RORγt induces IL-23R and mice deficient in IL-23 generally have defects in the function of IL-17-producing lymphocytes [22,23]. In addition, T cells capable of producing IL-17 independent of IL-23/IL-23R have been described [24].
Regarding myeloid-lineage cells, IL-17 production has been described in subsets of human and murine neutrophils [25]. Using murine models of acute and allergic pulmonary aspergillosis, we and others recently demonstrated that lung eosinophils elicited in response to challenge with live A. fumigatus or sensitization to A. fumigatus crude protein extracts (Af cpe) produce IL-17 [8,9]. Eosinophils have also been implicated as a potential source of IL-17 within the airways of humans with asthma [26]. These findings may be particularly relevant because high levels of IL-17 have been correlated with symptom severity in allergic asthma [27], possibly due to promotion of neutrophilic airway inflammation [28]. In addition, in experimental models of asthma, mice lacking IL-17 signaling have decreased airway hyperresponsiveness and mucus production [29]. The goal of this study was to define the contributions of RORγt and IL-23 to eosinophil IL-17 production. We show that eosinophils express RORγt and IL-23R in acute and allergic models of pulmonary aspergillosis. However, using IL-23p19-deficient mice, we demonstrate that while Aspergillus-induced eosinophils express IL-23R, IL-23 is not required for eosinophil RORγt expression and IL-17 production.

IL-17 expression by lung eosinophils in a modified mouse model of allergic pulmonary aspergillosis
In our published allergic pulmonary aspergillosis protocol, eosinophils recruited to the lungs of sensitized C57BL/6 mice challenged with aerosolized A. fumigatus crude protein extract (Af cpe) produced IL-17 and IL-23 [8]. That protocol required a specialized aerosol chamber and large amounts of allergen. To circumvent these drawbacks, we examined the effect of pulmonary challenge via intranasal instillation of Af cpe on lung eosinophil recruitment and phenotype. Mice on the C57BL/6 background were sensitized with two biweekly intraperitoneal doses of 250 μg Af cpe and received three daily intranasal challenges with 25μg Af cpe (Fig 1A). Flow cytometric analysis of single cell preparations of the lungs of mice analyzed two days following the last challenge revealed the mice developed a robust eosinophilia (Fig 1B and  S1 Fig). The gate used to define eosinophils was almost completely vacant when lung cells from eosinophil-deficient MBP-1 -/-EPX -/mice were analyzed.
While high expression of CD11c on Siglec F + cells is characteristic of alveolar macrophages [30], eosinophil populations expressing low to intermediate CD11c levels can emerge during allergic inflammation [31]. To confirm that these cells were eosinophils in our allergic pulmonary aspergillosis model, Siglec F + lung cells were sorted based on level of expression of CD11c into three sub-populations: (a) CD11c -, (b) CD11c low , and (c) CD11c intermediate (Fig 1C). The sorted cells were then subjected to modified Wright Giemsa (Diff-Quik) and immunofluorescence staining for the eosinophilic granule protein, major basic protein (MBP). Cells from all three gates were predominantly eosinophils as determined by their bilobed and donut shaped nuclear morphology, presence of eosinophilic cytoplasmic granules and immunofluorescence staining for MBP. The negative control, cells from the Siglec F -/CD11cgate, did not contain eosinophils as assessed by both Diff-Quik and immunofluorescence staining. Next, eosinophil expression of intracellular IL-17 was determined in this model of allergic pulmonary aspergillosis. As with our previous model [8], a high percentage of eosinophils stained positive for IL-17 by flow cytometry (Fig 1D) and confocal microscopy ( Fig 1E).

RORγt expression in lung eosinophils from mice with allergic and acute pulmonary aspergillosis
To explore the mechanisms underpinning lung eosinophil IL-17 production, we investigated the contribution of RORγt. This transcription factor induces IL-17 gene expression in T H 17, T C 17, γδ T cells, innate lymphoid cells, and neutrophils [25]. Using the allergic pulmonary aspergillosis model described above, nearly 40% of the lung eosinophils stained positive for RORγt by intracellular flow cytometry (Fig 2A). RORγt was observed in eosinophils by imaging flow cytometry. The top two panels of Fig 2B show representative images of sorted eosinophils exhibiting RORγt staining which co-localizes with the nuclear stain, DAPI, on the merge image. The lower panel shows an apparent alveolar macrophage (AMφ) which, while Siglec F positive, can be distinguished from eosinophils based on its larger size and nuclear morphology. This cell stains negative for RORγt. Note that the antibody to RORγt may also react with RORγ. As monensin was not added in these experiments, we could not determine whether eosinophils which expressed RORγt also expressed IL-17.
We next sought to validate RORγt expression in eosinophils by an alternative technique and to further examine the intracellular localization of this transcription factor. Lung eosinophils from mice with allergic pulmonary aspergillosis were sorted by flow cytometry and separated into nuclear and cytoplasmic fractions. While RORγt was detected in both fractions by immunoblotting, it was significantly enriched in the nucleus ( Fig 2C). The veracity of the fractionation procedure was confirmed by immunoblotting for the nuclear and cytoplasmic proteins, tata-binding protein (TBP) and tubulin, respectively. The specificity of the RORγt antibody was confirmed by immunoblotting cytoplasmic and nuclear lysates obtained from thymocytes of RORγt -/and wild-type mice. As expected, a 58 kDa band (corresponding to the molecular size of RORγt [19]) was observed in lysates from nuclei of wild-type but not RORγt -/thymocytes (S2 Fig).
To determine whether eosinophils obtained from acutely infected lungs also express RORγt, we used our previously described acute pulmonary aspergillosis model in which mice receive a pulmonary challenge with live A. fumigatus conidia and then bronchoalveolar lavage (BAL) is performed 48 hours post-infection [8]. BAL cells were immunostained for RORγt, the eosinophil marker, EPX, and the nuclear stain DAPI. By confocal microscopy, the presence of intranuclear and cytoplasmic RORγt was observed in lung eosinophils ( Fig 2D). Similar results were obtained using another eosinophil marker, MBP (S3 Fig

Expression of mCherry in eosinophils from RORγt-mCherry reporter mice
As an independent means of assessing whether lung eosinophils express RORγt in our aspergillosis models, we used bacterial artificial chromosome (BAC) transgenic reporter mice which express mCherry under control of the Rorc(t) promoter. Cells which express the mCherry signal are predicted to also express RORγt. We sensitized and challenged the mCherry RORγt reporter mice with Af cpe and analyzed the lung eosinophils for the presence of mCherry signal by flow cytometry. Nearly 40% of the lung eosinophils expressed the mCherry reporter ( Fig 3A). In contrast, mCherry expression was not seen in lung eosinophils obtained from naïve mice (Fig 3B). mCherry expression was also seen in BAL eosinophils examined by confocal microscopy in allergic and acute ( Fig 3C) pulmonary aspergillosis models. For these

Expression of IL-23R by lung eosinophils obtained from mice with allergic and acute pulmonary aspergillosis
The heterodimeric cytokine IL-23, acting through the IL-23R, contributes to and enhances the function of IL-17-producing lymphoid cells [14]. In our models of acute and allergic aspergillosis, lung eosinophils express IL-17 [8]. Therefore, we examined whether IL-23R was expressed on the surface of lung eosinophils. With the allergic model of aspergillosis, approximately one third of the lung eosinophils stained positive for IL-23R by flow cytometry (Fig 4A). IL-23 positive eosinophils were also observed by imaging flow cytometry ( Fig 4B). In contrast, alveolar

PLOS PATHOGENS
RORγt and IL-23R expression in Aspergillus-stimulated lung eosinophils

PLOS PATHOGENS
RORγt and IL-23R expression in Aspergillus-stimulated lung eosinophils macrophages, which were distinguished from eosinophils based on their larger size, nuclear morphology and higher CD11c staining, were negative for IL-23R staining (Fig 4B).
We next determined IL-23 receptor expression on eosinophils by immunoblotting ( Fig  4C). Lung eosinophils were flow-sorted from mice with allergic aspergillosis as described in Methods. As a positive control, cells were sorted from the Siglec F -CD11cgate); this gate includes cell types known to express IL-23R, including neutrophils, γδ T cells and T h 17 cells [25, 32, 33]. We found IL-23R was expressed in eosinophils as well as in cells from SigF -CD11cgate ( Fig 4C).
As with other Th17-specifying genes such as il17a, il17f, il21 and il22, Rorc also regulates Il23r gene expression [34]. We therefore sought to determine whether there was a correlation between RORγt and IL-23R expression in lung eosinophils. RORγt-mCherry reporter mice were sensitized and challenged with Af cpe following which, lung eosinophil IL-23R expression was determined by flow cytometry. As observed in our previous experiments (Fig 3A), mCherry-positive and mCherry-negative eosinophil populations were present ( Fig 4D). Importantly, most of the mCherry-positive eosinophils also expressed IL-23R whereas the mCherry-negative eosinophils were largely IL-23R negative ( Fig 4D). These results suggest that RORγt may drive IL-23R expression in eosinophils. We also examined the expression of IL-23R in lung eosinophils 2d post challenge with live A. fumigatus conidia. By flow cytometry, we found a distinct population of IL-23R + lung eosinophils comprising approximately 10% of the gated eosinophils ( Fig 4E). IL-23R was also detected on a subpopulation of the eosinophils by confocal microscopy (Fig 4F).

IL-17 and IL-23r gene expression in lung eosinophils from mice with allergic aspergillosis
To determine whether protein expression was accompanied by active gene expression, we sorted eosinophils from the lungs of wild-type mice with allergic aspergillosis and analyzed RORγt, IL-17, and IL-23r mRNA by reverse transcriptase (RT)-PCR. We found robust expression of all three genes in the sorted eosinophils. As expected, based on literature demonstrating that in lungs stimulated by A. fumigatus, eosinophils are not exclusive producers of IL-17 [8,9,25], total lung cells (unsorted) and lung cells in the eosinophil-negative gate also expressed the above-mentioned genes ( Fig 5A).
Next, we sensitized and challenged wild-type and RORc -/mice with Af cpe and analyzed total lung cells for RORγt, IL-17a, and IL-23r expression by RT-PCR. As RORγt is present in CD4 + CD8 + double positive thymocytes [20], we first confirmed the phenotype of RORc -/mice by analyzing RORγt expression in thymocytes of wild-type and RORc -/mice by RT-PCR. As expected, there were undetectable RORγt transcripts in the knockout mice (S4 Fig). RORγt, IL-17a, and IL-23r expression was not detectable in total lung cells from RORc -/mice compared to its wild-type counterpart (Fig 5B). These data were confirmed by quantitative real-time RT-PCR (Fig 5C).

Role of IL-23p19 in inducing expression of RORγt and IL-17 in eosinophils
Along with IL-6 and TGF-β, IL-23 participates in activating Jak2-Stat3 signaling pathways to upregulate the expression of RORγt in lymphoid cells [35]. We investigated the contribution of IL-23 to eosinophils RORγt expression and IL-17 production by using IL-23p19 -/mice [22]. In the allergic pulmonary aspergillosis model, we found similar expression of RORγt when comparing lung eosinophils from IL-23p19 -/and wild-type mice by flow cytometry (Fig 6A). However, as expected [34], RORγt expression was greatly reduced in CD3 + CD4 + T cells from IL-23p19 -/mice compared to the wild-type mice (Fig 6A and S5 Fig) To further validate eosinophil RORγt expression did not require IL-23 signaling, we induced allergic aspergillosis in IL-23p19 -/and wild-type mice. Lung eosinophils were then purified by FACS, subjected to cellular fractionation and immunoblotted for RORγt. We found similar nuclear expression of RORγt comparing eosinophils from wild-type and IL-23p19 -/mice (Fig 6B). By flow cytometry, similar IL-17A production was found in eosinophils from wild type and IL-23p19 -/mice in the allergic pulmonary aspergillosis model (Fig 6C). Finally, by RT-PCR, we determined RORγt and IL-17a gene expression (Fig 6D) in FACSsorted lung eosinophils and CD4 + T cells from wild-type and IL-23p19 -/mice with allergic aspergillosis. We found transcript levels of RORγt and IL-17a were similar in eosinophils from

PLOS PATHOGENS
RORγt and IL-23R expression in Aspergillus-stimulated lung eosinophils wild-type and IL-23p19 -/mice. In CD4 + T cells, we observed that both RORγt and IL-17a gene expression was reduced in IL-23p19 -/mice compared to that of wild-type (Fig 6D). Our observations in CD4 + T cells are in line with previously published reports where authors have shown the reduced expression of RORγt and IL-17a mRNA in CD4 + T cells from IL-23p19 -/mice [36, 37]. Together these data suggest that lung eosinophils express RORγt and IL-17a in mice with allergic aspergillosis. While RORγt is required for the production of IL-17a, IL-23 is dispensable for lung eosinophils to express RORγt and IL-17a.

Discussion
In acute and allergic pulmonary aspergillosis models, we and others have shown that lung eosinophils elicited in response to challenge with live A. fumigatus and sensitization to A. fumigatus crude protein extracts produce IL-17 [8,9]. Moreover, eosinopenic mice with acute

PLOS PATHOGENS
RORγt and IL-23R expression in Aspergillus-stimulated lung eosinophils aspergillosis had significantly reduced levels of IL-23 heterodimer in their bronchoalveolar lavage fluid compared to their wild-type counterparts [8]. These findings suggest eosinophils are key drivers of the IL-23/IL-17 axis during pulmonary aspergillosis. Here, we studied intrinsic factors that promote eosinophil IL-17 production in acute and allergic aspergillosis. We examined the mechanistic basis for eosinophil IL-17 production in pulmonary aspergillosis, focusing on the roles of RORγt, IL-23 and IL-23R.
Strong evidence implicates RORγt as the master transcription factor indispensable for inducing IL-17-producing lymphoid cells [19,38]. In addition, under certain conditions, neutrophils make RORγt and IL-17 [25]. Using multiple approaches including flow cytometry with intracellular staining, imaging flow-cytometry, immunoblotting and confocal microscopy, we identified a subpopulation of lung eosinophils from mice with allergic and acute aspergillosis that express RORγt. We could not use RORγt-GFP reporter mice to further demonstrate eosinophil RORγt expression because eosinophils have autofluorescence which overlaps with GFP [39]. However, we were able to take advantage of newly created RORγt-mCherry reporter mice to confirm eosinophil expression of RORγt. In Th17 and other IL-17-producing lymphoid cells, expression of RORγt is limited to the nuclear compartment, but in neutrophils, RORγt is found in both nuclear and cytoplasmic compartments [25]. Moreover, stimulation of neutrophils with IL-6 and IL-23 results in translocation of cytoplasmic RORγt to the nucleus [25]. Eosinophils also exhibit both nuclear and cytoplasmic RORγt, as evidenced by cell fractionation studies and confocal microscopy. However, most of the eosinophil RORγt resides in the nucleus. Future studies will be needed to determine the factors activating RORγt nuclear translocation and subsequent IL-17 production in eosinophils.
The intrinsic ability of myeloid cells to produce IL-17 cytokines has been controversial. IL-17 gene expression and protein production has been observed in mouse and human neutrophils by some investigators [25,33]. In contrast, Tamassia et al., failed to detect IL-17 in human neutrophils from healthy donors and active psoriatic patients [40]. Moreover, neutrophils stimulated with IL-6 plus IL-23 were devoid of transcriptionally active histone marks at the IL-17a and IL-17f loci which are present in Th17 cells [40]. Furthermore, a population of human primary mast cells that contain IL-17 but do not express RORγt was described; these cells capture extracellular IL-17A secreted by other cells and then store the cytokine in their intracellular granules [41]. However, the data presented in this manuscript and our previous publication [8] strongly suggest that lung eosinophils synthesize IL-17 following stimulation with A. fumigatus and antigens derived from the fungus. Multiple complementary techniques were used to detect RORγt and IL-17, including the use of IL-17 and RORγt reporter mice. Finally, the generalizability of the findings is suggested by the similar results seen using distinct acute and allergic models of aspergillosis.
In Th17 cells, the heterodimeric cytokine IL-23 acts through the IL-23R to drive and sustain IL-17 expression [32]. IL-23 also promotes survival of myeloid cells at sites of infection by a mechanism independent of IL-17 [42]. RORγt is known to positively regulate IL-23R gene expression [23,43]. We observed IL-23R expression on eosinophils too in our models of acute and allergic aspergillosis. Moreover, in experiments with RORγt-mCherry reporter mice with allergic aspergillosis, we found a significant correlation between mCherry and IL-23R expression in lung eosinophils. Despite the induction of IL-23R, no significant differences were observed comparing allergic lung eosinophils from wild-type and IL-23p19 -/mice with regards to eosinophil recruitment, RORγt expression and IL-17 production. If these data can be extended to humans, they suggest that blocking antibodies to IL-23 and IL-23R would not be effective at reducing eosinophil IL-17 production in allergic lung disease.
The finding of IL-23-independent IL-17 production in eosinophils has also been described in other cell types. A population of γδT cells expressing both RORγt and IL-23R yet can produce IL-17 independent of IL-23/IL-23R signaling was observed in colonic lamina propria [24]. Moreover, in a murine model of laser-induced experimental choroidal neovascularization, IL-23-independent IL-17 production by γδT cells and innate lymphoid cells was demonstrated [44]. Finally, production of IL-17 independent of IL-23R gene expression was reported in human enthesis γδT cells obtained during surgery [45].
The mechanisms by which RORγt is induced in eosinophils and the anatomical location where transcription is turned on remain to be determined. RORγt does not appear to be constitutively expressed in eosinophils as in the RORγt-mCherry reporter mice, unstimulated lung eosinophils were negative for mCherry expression. Interestingly, endogenous RORγt ligands which can act as non-canonical activators have been described, including oxysterols and bile acid metabolites [46,47]. Just as there are subsets of lymphocytes, such as Th17 and Tc17 cells, in which RORγt and subsequent IL-17 expression is turned on, we propose there is an IL-17-producing "Eo 17 " subset of eosinophils that is induced under specific conditions in the host. More data are needed though before this term can be added to the immunology lexicon.
Regardless of the mechanism of activation, our data suggest that RORγt could serve as a drug target in persons with allergic aspergillosis. Identification of the endogenous agonist ligand(s) of RORγt in allergic eosinophils and determining how these ligands activate RORγt may provide insights into the design and development of RORγt modulators that block IL-17 production in eosinophils and inhibit pathology in allergic pulmonary aspergillosis. Indeed, antagonists of RORγt which suppresses IL-17-and IL-23-mediated inflammation are under preclinical investigation for the treatment of autoimmune and other inflammatory diseases [43,48]. Finally, while RORγt is recognized as the master regulator of Th17 differentiation, it does so in concert with other transcription factors, including STAT3, IRF4, BATF, and RUNX1 [49,50]. The contribution of these factors to transcription regulation of Il17a and Il17f in eosinophils requires study.
Asthma is a complex and heterogenous disease [51]. Clinical and experimental data demonstrate that eosinophils contribute to the airway hyperreactivity seen in asthmatic individuals that have allergic airway disease [8,10,11,51]. Moreover, bronchoalveolar fluid IL-17 levels are increased in severe asthma [52], and there appears to be a subset of patients for whom IL-17 cytokines play pathogenic role [51]. However, in addition to eosinophils, numerous other IL-17-producing cell types are found in the lungs. Thus, a key question is whether IL-17 specifically derived from eosinophils plays a significant pathological role. Addressing this knowledge gap should help inform whether approaches targeting eosinophil production of RORγt and IL-17 are likely to be successful.
Allergic airway disease is believed to result from abnormal immune responses stimulated by environmental antigens. Previously, we used a model of allergic pulmonary aspergillosis that required a specialized aerosol chamber and relatively large amounts of antigen for the three pulmonary challenges [8]. In the present study, we simplified the method of inducing allergic pulmonary aspergillosis by changing the challenge to direct aspiration of intranasally administered antigens. The extent of lung eosinophilia observed using this easier and more economical allergic aspergillosis model was similar to previously described models from our lab and others [8,53]. Importantly, we again observed expression of IL-17A by the majority of lung eosinophils elicited in the new allergic aspergillosis model. Moreover, consistent with previously published data with ova and house dust mite induced allergies [31,54], the Siglec F + eosinophils could be divided into subpopulations based on relative expression levels of CD11c. Other models of allergic aspergillosis have been described including one featuring repetitive intranasal instillation of live A. fumigatus conidia [55]. In that model, eosinophil-deficient mice had decreased expression of IL-17A.
In conclusion, we elucidated intrinsic molecular factors responsible for IL-17 production by eosinophils. We demonstrated that lung "Eo 17 " eosinophils express RORγt, the master transcription factor which drives IL-17 production in lymphoid cells. Challenge of lungs with live A. fumigatus or its antigens skews lung eosinophils towards the Eo 17 phenotype by a noncanonical pathway independent of IL-23/IL-23R signaling. Elucidating the mechanisms by which Eo 17 cells are stimulated could inform therapeutic approaches for treating allergic lung diseases in which eosinophilia is a major feature.

Ethics statement
Experimental procedures with mice were approved by the University of Massachusetts Medical School (protocol A1802) and the NYU School of Medicine (protocol IA15-01410) Institutional Animal Care and Use Committees in accordance with National Institutes of Health guidelines for the care and use of laboratory animals.

Mice
All mice were on the C57BL/6 background and were bred and housed in specific-pathogen free conditions at the animal facilities in the University of Massachusetts Medical School. Mice of both genders were used and were generally six to ten weeks old at the start of the experiment. Wild-type C57BL/6J and RORγt -/mice [19] were purchased from The Jackson Laboratory (Bar Harbor, ME) and bred in house. Eosinophil-deficient MBP-1 -/-EPX -/mice were obtained from James Lee and Elizabeth Jacobson (Mayo Clinic, Phoenix, AZ [56,57]. IL-23p19 -/mice were a gift from Nico Ghilardi (Genentech, South San Francisco, CA) [22]. The RORγt-mCherry reporter strain was generated by cloning mCherry-IRES-CreERT2-pA in frame with the initiation codon of Rorc(t), which encodes the RORγt isoform of the nuclear receptor. Mouse founder lines were screened for mCherry reporter activity of in vitro polarized Th17 cells (positive control) compared to Th2 cells (negative control) derived from naïve CD4 + T cells sort-purified from blood samples. Wild-type and control mice were bred and housed in the same animal room.

Murine allergic and acute aspergillosis models
The murine allergic aspergillosis model was as in previous studies [8], except the pulmonary challenges with Aspergillus fumigatus crude protein extract (Af cpe, Greer Laboratories, Lenoir, NC) were via the intranasal (IN) rather than aerosolized route. Briefly, mice were sensitized with 250μg of Af cpe administered via intraperitoneal (IP) injections on days 0 and 14. Mice were lightly anesthetized with isoflurane (Patterson Veterinary) and challenged IN with 25μg of Af cpe on days 28, 29, and 30. On day 32, mice were euthanized, and their lungs were harvested ( Fig 1A).
The murine acute pulmonary aspergillosis model was also as described [8,58,59]. Briefly, conidia from A. fumigatus strain Af293 (Fungal Genetics Stock Center, Manhattan, KS) were harvested after growth on Sabouraud dextrose agar (Remel, San Diego, CA) slants and purified by two rounds of filtration through a 40μm nylon mesh. Conidial suspensions were washed three times in PBS containing 0.01% Tween-20. Mice were anesthetized as above and infected orotracheally with 5x10 7 conidia. Mice were euthanized two days post fungal challenge and their lungs were harvested. For experiments measuring intracellular cytokine expression, mice received an IP injection of 500 μg of monensin (Sigma-Aldrich) 48 hours post Af cpe or conidia challenge [8]. Six hours following monensin treatment, mice were

PLOS PATHOGENS
RORγt and IL-23R expression in Aspergillus-stimulated lung eosinophils euthanized, and their lungs were harvested. Monensin was not used in experiments examining RORγt staining.

Flow cytometry and cell sorting
Lung single cell suspensions were prepared using the MACS lung dissociation kit (Miltenyi Biotec) [8]. Leukocytes were sequentially enriched on a Percoll (GE Healthcare) gradient, incubated with Fc block (2.4G2, BD Pharmingen) and stained with a viability dye (Fixable Viability Dye eFluor 780, eBioscience) for 30 minutes at 4˚C. Cells were washed with FACS buffer (1% BSA/PBS), stained at 4˚C for cell surface markers (Table 1), washed twice to remove unbound antibody conjugates and fixed in 4% paraformaldehyde (Electron Microscopy Sciences) for 20 minutes. For intracellular cytokine staining, cells were permeabilized using commercial kits (BD Perm/Wash and BD Cytofix/Cytoperm, BD Pharmingen) according to the manufacturer's instructions, washed twice and incubated with labeled antibodies (Table 1). Similarly, for staining of nuclear markers, we used the eBioscience Foxp3/Transcription Factor Staining Buffer Set (Invitrogen). Cells were stained for RORγt (Table 1) according to the manufacturer's instructions. Fluorescence minus one and/or isotype controls were included in all experiments. Flow cytometry data were acquired with an LSRII cytometer and analyzed using FlowJo v10 Software (Tree Star Inc.). At least 100,000 events were collected per sample.
Cells were sorted on a Becton Dickinson FACS Aria II cell sorter using an 85μm nozzle operated at 45psi. For eosinophil sorting, lung cells were stained with surface markers CD45, Siglec F and CD11c along with the live dead stain (Fixable Viability Dye eFluor 780). Sorted eosinophils were singlet cells (based on FSC-H vs FSC-A), which were live, CD45 + , Siglec F + , and CD11c -/low/intermediate (S1 Fig). For CD4 + T cell sorting, we sorted cells which were singlets, live, CD45 + , CD3 + , and CD4 + .

Cytological staining
Cells were counted and cytospun onto poly-L-Lysine coated slides (Shandon Cytoslides, Thermo Scientific) at 800 rpm for 3 minutes (Shandon Cytospin 2). Eosinophils were then differentiated by cytological staining using a commercial kit (Diff-Quik, Polysciences, Warrington, PA) per the manufacturer's instructions. Finally, slides were mounted with Entellan Mounting Medium (Electron Microscopy Sciences).

Confocal microscopy
Confocal microscopy was performed as described [8], with slight modifications. Briefly, single cell lung suspensions were prepared as for flow cytometry. Cells were counted, fixed with 2%

Imaging flow cytometry
Lung cells, obtained from mice with allergic aspergillosis, were stained with a live-dead stain (Fixable Viability Dye eFluor 780) and the cell surface markers CD45, Siglec F, CD11c, IL-23R. Where indicated, cells were permeabilized and intracellularly stained for RORγt. Nuclei were distinguished by DAPI-staining. Brightfield and fluorescent images of cells were obtained with an Amnis/Flowsight Imaging Cytometer at 20X magnification and analyzed using IDEAS software (Luminex).

Western blot and immunoblotting
Lung eosinophils, obtained from C57BL/6 mice sensitized and challenged with Af cpe, were purified by fluorescence-activated cell sorting, as described above. Eosinophils were separated into cytoplasmic and nuclear fractions using the Nuclear and Cytoplasmic Extraction Kit (Thermo Scientific). Nuclear and cytoplasmic fractions were resolved by SDS-polyacrylamide gel electrophoresis and probed by western blotting using the antibodies listed in Table 3. The total cell extracts were prepared using RIPA buffer (50mM Tris pH 8.0, 150 mM Nacl, 1% NP40, 0.25% sodium deoxycholate, 0.1% SDS, 2mM EDTA) supplemented with 1X protease inhibitor cocktail (Thermo Fisher).

RT-PCR and Quantitative Real-Time RT-PCR
Total RNA was isolated using the TRIzol (Thermo Fisher) extraction method, as per the manufacturer's instructions. RNA was reverse-transcribed using Protoscript II first strand cDNA synthesis kit (New England Biolabs). The synthesized cDNA was used to study RORγt, IL-17A, IL-23R gene expression using described primers [60]. Gapdh was used as a loading control.

Statistical analysis
GraphPad Prism 7 software was used for statistical analysis. The unpaired two-sample Student's t-test was used to compare means of two groups. The Bonferroni correction was applied for multiple comparisons.

PLOS PATHOGENS
RORγt and IL-23R expression in Aspergillus-stimulated lung eosinophils