Th2 cells promote eosinophil‐independent pathology in a murine model of allergic bronchopulmonary aspergillosis

Repeated inhalation of airborne conidia derived from the fungus Aspergillus fumigatus (Af) can lead to a severe eosinophil‐dominated inflammatory condition of the lung termed allergic bronchopulmonary aspergillosis (ABPA). ABPA affects about 5 million individuals worldwide and the mechanisms regulating lung pathology in ABPA are poorly understood. Here, we used a mouse model of ABPA to investigate the role of eosinophils and T cell‐derived IL‐4/IL‐13 for induction of allergic lung inflammation. Selective deletion of IL‐4/IL‐13 in T cells blunted the Af‐induced lung eosinophilia and further resulted in lower expression of STAT6‐regulated chemokines and effector proteins such as Arginase 1, Relm‐α, Relm‐β, and Muc5a/c. Eosinophil‐deficient ΔdblGata mice showed lower IL‐4 expression in the lung and the number of Th2 cells in the lung parenchyma was reduced. However, expression of the goblet cell markers Clca1 and Muc5a/c, abundance of mucin‐positive cells, as well as weight gain of lungs were comparable between Af‐challenged ΔdblGata and WT mice. Based on these results, we conclude that T cell‐derived IL‐4/IL‐13 is essential for Af‐induced lung eosinophilia and inflammation while eosinophils may play a more subtle immunomodulatory role and should not simply be regarded as pro‐inflammatory effector cells in ABPA.


Introduction
Allergic bronchopulmonary aspergillosis (ABPA) is a chronic inflammatory condition affecting about 5 million people worldwide with a high incidence in asthma and cystic fibrosis patients [1,2]. ABPA is elicited by the omnipresent fungal opportunistic pathogen Aspergillus fumigatus (Af), resembles the most common manifestation of allergic bronchopulmonary mycosis and is also considered as one of the severest forms of allergic eosinophilic asthma [3,4]. Treatment do date is largely limited to symptomsuppressive azole antimycotica and glucocorticoid therapy to reduce fungal burden and immunopathology, respectively [5].
Dependent on the host immune status Af can cause various diseases in humans. In contrast to invasive aspergillosis, which primarily affects immuno-compromised patients, ABPA is characterized by a type 2 immune response directed against inhaled fungal conidia [6][7][8][9]. This includes increased production of the cytokines IL-4, IL-5, and IL-13, elevated serum IgE levels, pronounced goblet cell hyperplasia with mucus deposition in airways and accumulation of Th2 cells and eosinophils in the lung [10][11][12][13].
IL-4 and IL-13 activate STAT6-regulated genes in various cell types and play a central role in mediating the effector phase of type 2 immune responses, including ABPA [14][15][16][17][18]. Both cytokines can be produced by many different cell types of the innate and adaptive immune system including Th2 cells, basophils, mast cells, type 2 innate lymphoid cells (ILC2s), and eosinophils [19][20][21]. It is therefore challenging to identify the predominant cellular source of IL-4/IL-13 in type 2 immune responses. In a model of OVA sensitization and challenge, we could previously show that T cell-derived IL-4/IL-13 is required for induction of allergic inflammation of the lung [22]. This might be different in ABPA since the abundantly recruited eosinophils could be a more important source of these cytokines as compared to Th2 cells. Furthermore, Af infection of the lung is also known to cause a Th17 response that can be crossreactive to Candida albicans and may counteract a protective Th1 response [23,24]. In addition, repeated weekly intranasal exposure of mice to live Af conidia resulted in coevolution of a mixed Th1, Th2, and Th17 response where arterial remodeling in the lung was promoted by IL-4 and IL-10 from non-T cells [9].
Eosinophils are classically regarded as terminal effector cells in various allergic diseases [25][26][27]. This view is largely based on the fact that eosinophils store several cytotoxic proteins like eosinophil peroxidase and major basic protein in cytoplasmic granules that are released after activation and thereby inflict damage on surrounding tissue [28]. However, eosinophils may also have immunomodulatory functions and eosinophil-derived major basic protein has been shown to be dispensable for allergic airway inflammation [29][30][31].
In this study, we used a mouse model of ABPA with repetitive intranasal application of live Af conidia to immunocompetent mice to address the contribution of Th2 cells and eosinophils for allergic inflammation of the lung. We demonstrate that Af conidia induced massive recruitment of eosinophils to the lung while the number of neutrophils did not increase. T cell-derived IL-4/IL-13 appeared to be essential for lung eosinophilia, induction of chemokines associated with eosinophil recruitment, and goblet cell hyperplasia. Eosinophil-deficient dblGata mice showed lower levels of IL-4, IL-17, and IL-23, and contained fewer Th2 cells in the lung parenchyma as compared to WT mice. Despite these effects, eosinophils appeared to be largely dispensable for goblet cell hyperplasia, mucus production, and lung pathology.

Aspergillus fumigatus conidia elicit lung eosinophilia without neutrophil recruitment
Eosinophils and neutrophils are both known as potent innate effector cells associated with antifungal immune responses. However, the kinetics and mechanisms of tissue accumulation of both cell types can vary depending on the organ, fungal burden, and the quality of the adaptive immune response. Here, we established a mouse model for ABPA by repetitive intranasal administration of 2 × 10 6 live conidia of Af and analyzed the lungs at different time points for eosinophil and neutrophil accumulation (Fig. 1A). We observed a selective increase of eosinophils but not neutrophils over time. On day 17, 3 days after the fifth Af conidia administration, eosinophils reached about 30% of total cells and a cell count of more than 5 × 10 6 in the lung while the number of neutrophils remained as low as in untreated mice ( Fig. 1B-D, Supporting Information Figs. S1 and S2). Basophils and ILC2s showed only a small increase in total numbers (Supporting Information Fig. 3). We therefore conclude that this repetitive low-dose administration of Af conidia elicits a strong type 2 immune response with massive lung eosinophilia as it is also seen in human ABPA patients.

Af-induced lung eosinophilia requires T cell-derived IL-4/IL-13
The role of T cells and the Th2 cell-associated cytokines IL-4 and IL-13 in the context of allergy and more specifically in ABPA have been the subject of numerous studies [16][17][18]. Given the fact that IL-4 and IL-13 can be expressed by many different cell types including a variety of innate effector cells such as eosinophils, basophils, and ILC2s, it was unclear whether T cell-derived IL-4/IL-13 plays a crucial or rather redundant function in this ABPA model. To address this point, we compared WT mice with T cell-specific IL-4/IL-13-deficient (4-13Tko) mice, where IL-4 and IL-13 is specifically deleted only in the T cell lineage [32]. Surprisingly, without IL-4 and IL-13 from T cells, we observed completely abrogated eosinophilia ( Fig. 2A-C). The Af-induced increase of mRNAs for IL-4 and IL-13 in the lung of WT mice was blunted in 4-13Tko mice (Fig. 2D). This indicates that Th2 cells are either the main source of these cytokines or play a critical role to recruit other IL-4/IL-13 producers to the lung. The impaired IL-4 and IL-13 expression was further accompanied by a lack of induction of IL-5, Ccl11 (eotaxin-1), and Ccl24 (eotaxin-2), critical molecules for eosinophil recruitment, and survival (Fig. 2D). In contrast to 4-13Tko mice, Af-challenged WT mice further upregulated the IL-4/IL-13-induced chemokine and ABPA marker Ccl17 (TARC) [33,34] as well as Cxcl1 and Ccl2 associated with neutrophil and monocyte recruitment, respectively (Fig. 2D). IFN-γ as an indicator for type 1 immunity was not upregulated upon Af administration in WT or 4-13Tko mice. Intriguingly, IL-17A was strongly induced by Af but comparable between both strains of mice, while IL-23A expression implicated in Th17 and Treg induction was generally very low and only induced in Af-challenged lungs of WT mice (Fig. 2D). Hence, repetitive intranasal administration of Af conidia in the absence of IL-4/IL-13-expressing Th2 cells appeared not to shift the immune response to a Th1-or Th17-dominated phenotype.  Fig. S1). Frequencies (C) and total numbers (D) of eosinophils and neutrophils in the lung at indicated days after Af conidia administration. Displayed are the mean + SEM from pooled data of two independent experiments except for d1 (only 1 experiment) with at least five mice per group and time point. Statistical significance was determined by Kruskal-Wallis test with Dunn's post hoc method against the control group (d0). ***p < 0.001.

4-13Tko mice show normal Af-induced eosinophil development but reduced blood eosinophilia
Given the impaired Af-induced lung eosinophilia in 4-13Tko mice, we next addressed the question, whether eosinophil development or maturation in the BM is also affected in 4-13Tko mice. Therefore, we analyzed BM and blood from naïve and Af-challenged WT and 4-13Tko mice by flow cytometry. The frequency and total number of eosinophils (Siglec-F + SSC hi ) in the BM slightly increased after Af exposure in both WT and 4-13Tko mice ( Fig. 3A and B). The upregulation of the eotaxin receptor CCR3 indicates an important maturation step for eosinophils in the BM [35]. Using this marker, we observed no significant differences in the number of mature eosinophils between WT and 4-13Tko mice ( Fig. 3A and C). However, we found that the frequency of peripheral blood eosinophils was increased in Af-challenged WT mice but not in 4-13Tko mice ( Fig. 3D and E). These results suggest that 4-13Tko mice are impaired in mobilizing eosinophils from the BM into the blood circulation and/or promoting survival of blood eosinophils while generation and maturation of eosinophils in the BM appear normal.

T cell-derived IL-4/IL-13 is required for Af-induced allergic lung inflammation
Having shown that T cell-derived IL-4/IL-13 is critical for Afinduced lung and blood eosinophilia, we further investigated whether this correlates with histopathological changes and expression of effector molecules associated with allergic lung inflammation. Hematoxylin-and eosin-stained lung sections of Afchallenged WT mice revealed massive eosinophil-dominated infiltrates that were not observed in 4-13Tko mice (Fig. 3F). Furthermore, periodic acid-Schiff (PAS) staining showed that only Afchallenged WT mice developed goblet cell hyperplasia in the lung (Fig. 3G). Quantitative RT-PCR analyses revealed that markers of IL-4/IL-13-activated goblet cells like Calcium-activated chloride channel regulator 1 (Clca1, also named Gob5) and Resistin-like-β (Retnlb) were only induced in Af-challenged WT mice (Fig. 3G). Similarly to mucus production, we could also only observe an increase in IgE levels in the serum of WT, but not 4-13Tko mice (Fig. 3H).
We further determined the frequency and total numbers of AAMs in the lung by flow cytometry using PD-L2 as a surface marker [36]. AAMs only increased in the lung of Af-challenged WT mice ( Fig. 4A-C). The AAM-associated effector molecules Resistinlike-α (encoded by the Retnla gene) and Arginase 1 (Arg1) [37,38] were also strongly induced in Af-challenged WT mice, but not in 4-13Tko mice (Fig. 4D).
These findings demonstrate that the major effector pathways of Af-induced type 2 immunity-associated lung inflammation including eosinophilia, goblet cell hyperplasia, IgE production, and AAM accumulation are all critically dependent on T cell-derived IL-4/IL-13.

Eosinophils promote the IgE response and Th2 cell accumulation in the lung of Af-challenged mice
So far, we described the critical role of Th2 cells for lung eosinophilia and type 2 immunity in a model of ABPA. In addition to a variety of pro-inflammatory effector molecules, eosinophils can also produce significant amounts of IL-4 and IL-13 and may therefore promote Th2 accumulation and down-stream effector pathways in the lung. To determine the role of eosinophils for accumulation of Th2 cells in the lung of Af-challenged mice, we compared IL-4eGFP reporter mice (4get mice) with eosinophildeficient 4get mice on day 17 after the first administration of Af conidia. Flow cytometric analysis demonstrated that the frequency of Th2 cells among all CD4 T cells in the lung increased from 2.7% in naïve to 25% in Af-challenged 4get mice and from 2.3% to 22% in 4get dblGata mice (Fig. 5A). Yet in total numbers, we observed about 1 × 10 5 more Th2 cells in the lung of Af-challenged 4get mice as compared to 4get dblGata mice (Fig. 5B). This difference was not apparent in the BAL where Th2 cell numbers were about fivefold lower as compared to total lung tissue in both strains of mice (Fig. 5B). The spleen did not show an increase of Th2 cells after Af challenge indicating that intranasal administration of Af conidia does not result in a systemic Th2 response ( Fig 5B). Further analysis revealed comparable AAM accumulation in the lung ( Fig. 5C and D) but lower serum IgE levels in Af-challenged 4get dblGata as compared to 4get mice (Fig. 5E). These results demonstrate that eosinophils can enhance accumulation of Th2 cells in the lung and promote the IgE response in this model of ABPA.

Eosinophils promote expression of IL-4, IL-17, and IL-23 in the lung of Af-challenged mice
To further investigate eosinophil-dependent changes of cytokines and chemokines in the lung of Af-challenged mice, we performed quantitative RT-PCR analysis from whole lung tissue of 4get and 4get dblGata mice. IL-5 and IL-13 were strongly and IFN-γ was only moderately induced to similar levels in both strains of mice (Fig. 6A). In contrast, IL-4 upregulation was significantly stronger in Af-challenged 4get mice (Fig. 6A). This result could be confirmed at the protein level by ELISA (Fig. 6B). We also noticed increased expression of IL-17A and IL-23A in Afchallenged 4get mice that is consistent with a previous report demonstrating expression of both cytokines by eosinophils after acute high-dose Af conidia administration to the lung [39]. With regard to chemokines, we observed strong and comparable induction of Ccl11 and Ccl24 in both strains of mice, while Ccl17 and Cxcl2 were expressed at higher levels and Cxcl1 at lower levels in eosinophil-deficient 4get dblGata mice (Fig. 6A). This indicates that although eosinophils can promote IL-4 expression in the Af-challenged lung, this does not result in increased expression of IL-4/IL-13-induced chemokines. Furthermore, eosinophils seem to promote a mixed Th2/Th17 response in this ABPA model.  Quantitative RT-PCR data of AAM-associated markers from whole lung tissue. Expression was normalized to the Hprt gene. Bars show the mean + SEM of pooled data from three experiments with a total of -seven to nine (flow cytometry) or five to eight (qRT-PCR) mice per group. Statistical significance was determined by two-way ANOVA with Holm-Sidak post hoc testing. ***p < 0.001.

Eosinophils are dispensable for ABPA-associated mucus production and lung edema
Eosinophils are generally believed to promote tissue damage and lung pathology in ABPA patients because they release a variety of pro-inflammatory proteins including major basic protein, eosinophil cationic protein, eosinophil peroxidase, and eosinophilderived neurotoxin. Therefore, we further investigated whether absence of eosinophils in 4get dblGata mice was associated with reduced goblet cell hyperplasia and less severe lung pathology. Unexpectedly, we observed strong and comparable induction of the goblet cell markers Clca1 (Gob5) and Muc5ac in the lung of Af-challenged mice of both mouse strains (Fig. 7A). PAS staining further indicated similar percentages of PAS + airways in 4get and 4get dblGata mice ( Fig. 7B and C). As a further parameter of pathology, we determined the weight of the lung from naïve and Af-challenged mice. The weight of the lung increased from 0.2 g in naïve mice to 0.4 g in Af-challenged mice of both mouse strains (Fig. 7D). Based on these results, we conclude that eosinophils are largely dispensable for Af-induced lung edema in this established model of ABPA.

Discussion
In this study, we demonstrate that T cell-derived IL-4/IL-13 is an important signal to induce lung eosinophilia and ABPAassociated pathology in mice, when repetitively confronted with low numbers of Af conidia to mimic chronic exposure to this fungal pathogen. Despite their well-known pro-inflammatory effector function, eosinophils appeared to be dispensable for lung pathology that developed with comparable severity in WT and eosinophil-deficient mice.
The lack of IL-4/IL-13 expression in T cells blunted the lung eosinophilia while eosinophil development in the BM was not affected. This supports the view that Th2 cells get locally activated in the lung and secrete IL-4/IL-13 to induce expression of the STAT6-regulated and eosinophil-recruiting chemokines Ccl11, Ccl17, and Ccl24 in lung endothelial cells [40]. In addition, Th2derived IL-5 may enhance the survival of recruited eosinophils in the lung.
We also observed that IL-4 expression and the number of Th2 cells in the lung parenchyma were reduced in the absence of eosinophils. This suggests that eosinophils can promote further recruitment or survival of Th2 cells in the Af conidia challenged lung consistent with previous reports using other models of allergic lung inflammation [41][42][43]. However, neither the frequency of Th2 cells among all CD4 T cells in the lung nor the total number of Th2 cells in the BAL was diminished in 4get dblGata mice. Interestingly, the expression of the Th2 recruiting chemokine Ccl17, which can be used as diagnostic marker in ABPA patients [34], was even higher in 4get dblGata as compared to control mice. This suggests that eosinophils promote the accumulation of CD4 T cells in the lung parenchyma rather than stimulating differentiation or selective recruitment of Th2 cells in this ABPA model.  Eosinophils can also express IL-13 and it has been demonstrated that IL-13-driven pathology can occur independently of IL-4 in other models of allergic airway inflammation [44][45][46]. Indeed, eosinophil-derived IL-13 can be sufficient for allergic lung inflammation as demonstrated in an adoptive transfer system [47]. However, we observed no reduction of total lung IL-13 in 4get dblGata mice indicating that eosinophils are not a major source of this cytokine in the ABPA model with repetitive low-dose exposure to Af conidia. Interestingly, the induction of IL-17A and IL-23A expression was affected in 4get dblGata mice indicating that eosinophils can also promote a type 3 immune response in the lung that is characterized by increased levels of these cytokines. In fact, lung eosinophils have been demonstrated to express IL-17A and IL-23A in fungal respiratory disease [39,48].
The only minor difference of the type 2 inflammatory response in 4get dblGata mice was unexpected since several other reports C 2020 The Authors. European Journal of Immunology published by WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim.
www.eji-journal.eu  had indicated that eosinophils are important effector cells that promote allergic inflammation in the lung. For example, eosinophils had been found to induce airway damage and bronchial hyperreactivity in OVA-induced allergic lung inflammation in studies with IL-5-deficient mice [49] or anti-IL-5 treated mice [50]. Moreover, the Af cell wall component chitin elicits eosinophil recruitment and lung pathology [51]. Others have shown that allergic lung inflammation induced by repeated intranasal administration of Af extracts is attenuated in dblGata and CCR3-deficient mice on BALB/c background [27]. This result appears discrepant to our findings but might be explained by the fact that we used dblGata mice on C57BL/6 background [42] and, more importantly, live Af conidia instead of Af extracts, which better resembles the natural situation.
Recently, human ABPA cases have been described, where application of the IL-5 receptor targeting antibody mepolizumab acted C 2020 The Authors. European Journal of Immunology published by WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim.
www.eji-journal.eu beneficially in ameliorating pathology, implying a somehow proinflammatory role for eosinophils in this allergic disease [52][53][54]. However, although this treatment led to reduced eosinophilia, it should be taken into account that also other allergy-related cells, like basophils or antibody-producing B cells, are responsive to IL-5 and might be affected as well [55,56].
In conclusion, we found a dominant role for T cell-derived IL-4/IL-13 in establishing eosinophilia and allergic inflammation of the lung in a murine model of ABPA using live Af conidia. Eosinophils appeared to be largely dispensable for goblet cell hyperplasia, mucus production, and lung pathology. It remains to be investigated in further detail whether eosinophils are involved in fungal control and resolution of inflammation rather than exacerbation of ABPA.

Mice
T cell-specific IL-4/IL-13-deficient (4-13Tko) [32] and eosinophildeficient dblGata [57] mice have been described. For this study, dblGata mice had been crossed to 4get [58] mice in which expression of IL-4 mRNA is marked by the eGFP fluorescent protein. All mice were on C57BL/6 background, kept under SPF conditions and used at the age of 7-12 weeks. Animal experiments were performed with approval from the Local Government of Lower Franconia and according to the German animal protection law and guidelines of the European Union (Directive 2010/63/EU).

Intranasal administration of Af conidia
Instillation of freshly harvested fungal spores was performed as published before [59]. Briefly, mice were anaesthetized and received 2 × 10 6 live conidia of A. fumigatus ATCC 46645 (ref. PMID: 6160400) in 50 μL PBS intranasally every 3-4 days for up to five times. Mice were analyzed on day 17 of the experiments unless indicated otherwise.

Organ sampling and generation of single cell suspensions
Spleens were carefully mashed through a 70 μm cell strainer with a 5 mL syringe plunger. Femur and tibia bones were flushed out with FACS buffer (2% FCS, PBS) and single cell suspensions were generated by repetitively passing through a 20G syringe (BD Microlance, Franklin Lakes, NJ). Lungs were perfused with 8 mL of PBS through the right ventricle of the heart, placed into RPMI1640 (PAN-Biotech Inc., Aidenbach, Germany), cut into small pieces and digested with DNase DN25 (Sigma-Aldrich) and Liberase TM (Roche, Basel, Switzerland) at 100 μg/mL for 30 min at 37°C under continuous rotation. Debris was removed by gently mashing through a 100 μm cell strainer. Blood was collected in FACS blood buffer (2% FCS, 2000 U/mL Heparin, PBS) from the heart. Erythrocytes were lysed with ACK buffer (155 mM NH 4 Cl, 9.99 mM KHCO 3 , 0.10 mM EDTA) and cell suspensions filtered through a 70 μm cell strainer afterward. BAL was flushed out with 2 mL PBS 0.1 mM EDTA intratracheally. To retrieve serum samples, blood was also collected from the heart of mice, transferred into SST TM tubes (BD Microtainer, Franklin Lakes, NJ), allowed to clot for 30 min and centrifuged afterward according to manufacturer's protocol. Resulting supernatants were used for immunoglobulin ELISA measurements.

Flow cytometry
At first Fc-block (α-CD16/32, clone 2.4G2, BioXcell, West Lebanon, NH) was added to single cell suspensions to prevent unspecific staining. The staining was undertaken at 4°C in the dark for 20 min. For this, fluorophore-conjugated antibodies against the following targets were used: SIGLEC-F BV421 ( [60].

Quantitative RT-PCR
RNA from whole lung tissue was isolated with TriSure reagent (Bioline, Meridian Life Science Inc., Memphis, TN) and chloroform extraction. cDNA was generated with a High Capacity Reverse Transcription Kit from Applied Biosystems/Thermo Fisher Scientific (Waltham, MA). For the actual qRT-PCR, the SYBR Select Master Mix from Thermo Fisher Scientific (Waltham, MA) was used. Measurements were performed on a CFX-Connect instrument (Bio-Rad Laboratories, Inc., Hercules, CA) or a ViiA 7 Real-Time PCR System (Applied Biosystems/Thermo Fisher Scientific). Ct value results were normalized to Hprt as a reference gene. A list of primers used in this study is provided in Supporting Information Table S1. Primers are listed in Supporting Information

Enzyme-linked immunosorbent assay
BAL supernatants and serum samples were analyzed via sandwich ELISA. For IL-4, an antibody pair of 11B11 (BioXcell, West Lebanon, NH) and BVD6-24G2-biotin (Biolegend), for IL-13 the murine IL-13 Standard ABTS ELISA Development Kit from Peprotech (Rocky Hill, NJ) were used. In both cases, alkaline phosphatase-conjugated Streptavidin (Southern Biotech Birmingham, AL) was applied to detect bound antibodies. IgE levels were assayed by pairing rat-anti-mouse IgE capture antibody (R35-72, BD Biosciences) and alkaline phosphatase-conjugated polyclonal goat-anti-mouse IgE detection antibody (Southern Biotech). Para-Nitrophenylphosphate (Southern Biotech) was used as chromogenic substrate and absorption was measured at a wavelength of 405 nm on a Multiskan FC 3.0 instrument (Thermo Fisher Scientific).

Histology
Lungs were fixed overnight in 4% para-formaldehyde (PFA) at 4°C, dehydrated, and embedded in paraffin. 3 μm sections were cut on a microtome and stained with H&E or Periodic-Acid-Schiff's reagent (PAS). Stained tissue sections were analyzed using a 5× objective for overview pictures or 20× and 40× objectives for selected representative bronchi and cellular infiltrates on a Zeiss Axio Vert.A1 microscope.

Statistical analysis
Statistical analysis of acquired data was performed with Sigmaplot (Version 12.3, Systat Software Inc., San Jose, CA). For the kinetic analysis ( Fig. 1), criteria for a one-way ANOVA were not met and thus a Kruskal-Wallis test with the nominal variable of time was performed with Dunn's post hoc test against the control group of time point d0. In all experiments, comparing two different nominal variables (treatment and genotype), statistical significance was determined with two-way ANOVA and a Holm-Sidak post hoc test.