Eosinophils are bone marrow-derived granulocytes that are typically abundant in inflammatory infiltrates in the defense against helminthic parasites and in allergic diseases. However, various immunoregulatory actions and functions have been also described for eosinophils, such as lymphocyte recruitment, tissue repair, organ development, antigen presentation, antimicrobial and antifungal activities, among other functions (Strandmark et al., 2016). Two of the main eosinophil activation mechanisms are degranulation and, more rarely, phagocytosis (Shamri et al., 2011). Recently, another cell effector mechanism has been described for eosinophils—the release of extracellular DNA traps (Yousefi et al., 2008; Ueki et al., 2013). The release of extracellular DNA traps (ETs) by leukocytes has been considered an important mechanism of the immune response in different inflammatory conditions. Although most ET knowledge is based on neutrophil studies (Brinkmann et al., 2004; Fuchs et al., 2007; Parker et al., 2012), it is well known that other leukocytes are also able to release ETs, including eosinophils (Yousefi et al., 2008), mast cells (von Kockritz-Blickwede et al., 2008), monocytes/macrophages (Chow et al., 2010), dendritic cells (Ramirez-Ortiz et al., 2011; Loures et al., 2015) and basophils (Schorn et al., 2012; Morshed et al., 2014). Various stimuli are capable of inducing the release of ETs, which essentially are structures of disrupted chromatin filaments coated with granular and cytosolic proteins, histones, and proteases (Brinkmann et al., 2004; Fuchs et al., 2007; Guimaraes-Costa et al., 2009; Parker et al., 2012). ET release involves non-lytic (Yousefi et al., 2008, 2009; Yipp et al., 2012) or lytic processes (named ETosis) that in general require cell chromatin decondensation (Brinkmann et al., 2004; Fuchs et al., 2007; Guimaraes-Costa et al., 2009; Parker et al., 2012; Ueki et al., 2013). Some studies indicate that this process is due to histone hypercitrullination by the enzyme PAD4 (peptidylarginine deiminase 4), in which arginine residues in histones are converted to citrulline, allowing DNA fibers to unfold (Wang et al., 2009; Li et al., 2010; Lewis et al., 2015; Van Avondt and Hartl, 2018). However, controversy remains about the relative importance of PAD4 for ETs release (Neeli and Radic, 2013; Kenny et al., 2017; Claushuis et al., 2018; Guiducci et al., 2018; Silva et al., 2019; Thompson-Souza et al., 2020). Interestingly, some studies have indicated that mitochondria are the source of the DNA that composes ETs (Yousefi et al., 2008, 2009). ETs help leukocytes immobilize bacteria, fungi and viruses, creating a microenvironment that favors a more efficient elimination of pathogens (Brinkmann et al., 2004; Fuchs et al., 2007; Papayannopoulos et al., 2010; Parker et al., 2012). However, recent evidence has emerged suggesting that ETs also have a role in non-infectious sterile inflammation (Jorch and Kubes, 2017). In eosinophils, the process by which they produce and release eosinophil extracellular DNA traps (EETs) can result in cell death (named EETosis in eosinophils) (Ueki et al., 2013); or in a non-lytic process by which eosinophils rapidly produce EETs but do not lose their viability (Yousefi et al., 2008, 2009).

Exposure and sensitization to fungal allergens is an important factor in patients with respiratory allergies; in this context, fungi play an important role in the development, severity and persistence of allergic lung diseases (Knutsen et al., 2012; Denning et al., 2014). Allergic bronchopulmonary mycoses (ABPMs) are characterized by fungal colonization and are known to worsen lung function, which is commonly associated with the development of severe asthma (Denning et al., 2013, 2014; Chowdhary et al., 2014). Aspergillus fumigatus is the most common cause of ABPMs; to a lesser extent, C. albicans and Alternaria species are also related to the development of these diseases (Shah and Panjabi, 2002; Chowdhary et al., 2014). ABPA, a form of non-invasive eosinophilic pulmonary aspergillosis, is a multifaceted pulmonary disorder caused by immunological reactions in response to repeated antigen exposure and/or colonization by A. fumigatus (Agarwal et al., 2013; Ueki et al., 2018). Susceptibility is related to the pathophysiology of comorbidities such as asthma, sinusitis, cystic fibrosis and alveolitis (Kousha et al., 2011; Denning et al., 2014). We recently described the presence of EETs in sputum samples from ABPA patients and found that human eosinophils are capable of releasing EETs in vitro in response to A. fumigatus conidia in a lytic non-oxidative process that involves CD11b and the Syk signaling pathway (Muniz et al., 2018). However, the ultrastructural features and signaling events that characterize and drive A. fumigatus-induced EETosis in human eosinophils are not completely understood. Thus, the purpose of this study was to unravel the intracellular mechanisms that direct the process of extracellular DNA trap release by A. fumigatus-stimulated eosinophils.

Eosinophils are the major leukocytes involved in ABPA, which is the most prevalent fungal allergic manifestation among ABPMs (Shah and Panjabi, 2002; Chowdhary et al., 2014). Thus, studies that contribute to a better understanding of ABPA pathology and can elucidate possible therapeutic targets for patient treatment are extremely relevant. Previous studies have shown that EETs are present in the mucus plugs of ABPA patients and that human eosinophils release EETs in response to A. fumigatus conidia in vitro (Muniz et al., 2018; Ueki et al., 2018). Nevertheless, efforts to understand eosinophil-A. fumigatus recognition and the mechanisms that drive this interaction are still in progress. Here, we demonstrated by both TEM and SEM that eosinophils responding to A. fumigatus stimulation present prominent morphological alterations typical of EETosis. Moreover, A. fumigatus-induced EETs release depends on the calcium, Src, PI3K, p38 MAPK and Akt signaling pathways. Interestingly, we determined that human eosinophils release EETs in response to A. fumigatus independently of PAD4 and histone citrullination through a process in which fungus viability is dispensable (Figure 9).

EETosis with a cytolytic profile and extrusion of extracellular traps has increasingly been identified in several eosinophilic diseases (Ueki et al., 2016a, b, 2018). This process of cell death associated to ETs release is morphologically distinct from other classic cell death processes, such as apoptosis and necrosis (Fuchs et al., 2007; Brinkmann and Zychlinsky, 2012), and has been associated with different leukocytes, especially neutrophils (Brinkmann et al., 2004), mast cells (von Kockritz-Blickwede et al., 2008) and eosinophils (Ueki et al., 2013). Here, we used high-resolution TEM to characterize intracellular events associated with EETosis. Our TEM findings revealed that A. fumigatus-stimulated eosinophils elicited nuclear changes typical of EETosis (decondensation, delobulation/rounding, disruption of nuclear envelope, chromatin expansion in the cytoplasm and further release of EETs) (Figure 1). These processes are well characterized in neutrophils (Fuchs et al., 2007), but are poorly understood in human eosinophils. In contrast to the process of ETosis in neutrophils (termed NETosis), which leads to the release of granule contents mixed with released chromatin (Brinkmann et al., 2004; Fuchs et al., 2007), EETosis enables the release of clusters of FEGs, with preserved morphology and contents, together with EETs. Our ultrastructural results showing FEGs by both TEM (Figures 1C, 2D) and SEM (Figures 2A,Ai) are in accordance with previous works demonstrating that EETosis induced by different stimuli, including the lipid mediator lysophosphatidylserine and A. fumigatus conidia, occurs in the presence of punctual MBP labeling (indicative of intact granules) and in the absence of linear immunostaining for MBP or ECP (indicative of disrupted granules) (Ueki et al., 2013; Muniz et al., 2018; Kim et al., 2020). This means that distinct mechanisms are operating in neutrophils and eosinophils in the process of granule product release in association with DNA traps during ETosis. For eosinophils, FEGs are secretory-competent organelles acting as “cluster bombs” that selectively release proteins under specific stimuli (Neves et al., 2008; Ueki et al., 2013). However, the consequences of these secretory entities associated with EETs still require clarification.

The capacity of chromatin to actively participate in NETosis is now recognized (Neubert et al., 2018), in addition to its classical role in regulating gene expression. During NETosis, after certain signaling events and enzymatic reactions (including histone citrullination and phosphorylation events), the chromatin expands, which drives the rupture of the nuclear envelope, determining the point of no return. Here, we noted the same event for human eosinophils undergoing EETosis. Although signaling molecules are usually the focus of studies on cell activation processes, recent works have demonstrated that chromatin changes are crucial in determining cell fate (Neubert et al., 2018). Thus, our ultrastructural observations bring a new appreciation of the importance of linking intracellular signaling pathways to nuclear entities and structural protein modifications to better delineate a stimulus-specific cell activation process.

We previously demonstrated that intracellular signaling involved in the release of EETs in response to A. fumigatus is Syk- and CD11b β2-integrin-dependent (Muniz et al., 2018). Src and Syk kinases have been implicated in the cell signaling of various classes of receptors involved in fungal molecule recognition, including integrins (Mocsai et al., 2002, 2006; Jakus et al., 2007) and C-type lectins (Rogers et al., 2005; Kerrigan and Brown, 2010). Src-family kinases are involved in β₂ integrin-mediated responses in which the Src kinases are implicated in Syk phosphorylation in neutrophils (Mocsai et al., 2006). In line with these findings, our results reveal that A. fumigatus-induced EETs release is blocked in the presence of a Src inhibitor. Additionally, in agreement with our findings, neutrophils respond to A. fumigatus conidia by releasing NETs in a Syk-Src-dependent signaling pathway (Silva et al., 2019).

PI3Ks are enzymes that catalyze the phosphorylation of one or more inositol phospholipids in the 3-position of the inositol ring. PI3Ks can be divided into classes I, II, and III. Class IA PI3Ks, which include PI3K α, β, and δ, are activated by stimulation of tyrosine kinase-based receptors. PI3Kγ, the only member of class IB, is activated by stimulation of GPCR subunits (Vanhaesebroeck et al., 2010). We found that both a non-selective (wortmannin) and a class I-selective (AS605240, 10 μM) PI3K inhibitor blocked A. fumigatus-induced EET release. Because there have been many studies of wortmannin as a non-selective inhibitor of the different PI3K classes and of other possible targets (Makni-Maalej et al., 2013; Germic et al., 2017), we additionally used the compound AS605240. AS605240 at 10 μM has been reported as an inhibitor of class I PIKs but is unable at this concentration to distinguish between class A (PI3K α, β and δ) and B (PI3Kγ) (Sadhu et al., 2003; Silva et al., 2019). In mammals, class I PI3Ks are present in all cell types, including eosinophils, with PI3Kδ and γ highly enriched in leukocytes (Kok et al., 2009). As mentioned, we have previously described the dependence of A. fumigatus-induced EET release on Syk signaling (Muniz et al., 2018). Syk is a tyrosine kinase that is crucial in the signaling pathways mediated by Dectin-1 and the β₂ integrin CD11b/CD18 (Mocsai et al., 2002; Rogers et al., 2005). Taking in account that receptors utilizing protein tyrosine kinase activation are known to be more related to the activation of class IA PI3Ks (PI3K α, β, and δ), we investigated the role of class I PI3K δ in A. fumigatus-induced EET release. We observed that compound IC-87114 at 1 μM (concentration known to be selective for the class I δ isoform) (Sadhu et al., 2003; Silva et al., 2019) completely inhibited EETs release. Indeed, the importance of PI3K δ in eosinophils and in experimental models of allergic inflammation has been previously reported (Lee et al., 2006; Nashed et al., 2007; Kang et al., 2012). Kang and colleagues demonstrated that treatment with ICI87114 reduced murine bone marrow-derived eosinophil adhesion and Mac-1 expression and had inhibitory effects on eotaxin-1-induced chemotaxis and shape change (Kang et al., 2012). However, the authors used 10 μM ICI87114. In our studies, we used 1 μM ICI87114. Some studies suggest that 10 μM ICI87114 might also have prominent effects on PI3Kγ (Sadhu et al., 2003; Silva et al., 2019). Regarding the PI3Kγ isoform, some important effects have also been described in human eosinophils but are related to the activation of GPCRs (Saito et al., 2014). Saito and colleagues described the effects of specific inhibition of the PI3Kγ isoform on human eosinophil chemotaxis, adherence and degranulation induced by eotaxin (Saito et al., 2014). Since PI3Kγ signaling is associated with GPCRs (such as CCR3), whereas PI3K α, β, and δ are activated by receptor tyrosine kinase, it is possible that PI3Kγ and δ isoforms might have distinct roles in eosinophil intracellular signaling and function depending on the stimuli and the cognate receptor involved. In fact, PI3K activation has been implicated as crucial for ROS generation-induced DNA trap release in eosinophils and neutrophils stimulated with GM−CSF and C5a or with low concentrations of PMA (Germic et al., 2017); and in neutrophils stimulated with immobilized immune complexes (Behnen et al., 2014). In contrast, other studies have pointed that ROS and PI3K is dispensable for lysophosphatidylserine-induced EETs extrusion (Kim et al., 2020). The PI3K δ isoform specifically has been described as critically involved in NETs release in response to Leishmania amazonensis (DeSouza-Vieira et al., 2016) and A. fumigatus (Silva et al., 2019) in a signaling pathway upstream of ROS production. In contrast, the PI3Kγ isoform is not involved in A. fumigatus-induced NETs release (Silva et al., 2019).

The activation of Akt has been reported to be the downstream target of PI3K in various cells, including eosinophils (Alessi et al., 1997; Machida et al., 2005; Hawkins et al., 2010). The participation of Akt has been demonstrated in the context of NETs release following immunocomplex recognition and signaling via Mac-1 (Behnen et al., 2014). Moreover, it has been demonstrated that PMA-induced NETs formation is dependent on Akt activation, which suppresses apoptosis via the inhibition of caspases (Douda et al., 2014). In agreement with these findings, we found that the release of EETs in response to A. fumigatus requires the PI3 kinase and Akt signaling pathways. Behnen and colleagues showed that CR3 (CD11b/CD18) activation through the recognition of immunocomplexes by the FcγRIIIB receptor induces NETs release via the Src/Syk, PI3K/Akt, p38 MAPK, and ERK1/2 signaling pathways (Behnen et al., 2014). The involvement of p38 MAPK in NETs release induced by different stimuli, such as bacteria or PMA, was previously reported (Keshari et al., 2013; Behnen et al., 2014; Douda et al., 2015; Ma et al., 2018). In agreement with these studies, we found that the release of EETs was inhibited when eosinophils were pretreated with SB202190, suggesting the involvement of p38-MAPK signaling in A. fumigatus-induced EETs release.

As we observed that A. fumigatus-induced EETs release triggered the Src/Syk, PI3K/Akt and p38 MAPK signaling pathways, we also examined calcium. Changes in leukocyte calcium levels have been consistently related to several leukocyte functions, including cell adhesion, chemotaxis and degranulation, among others (Dixit and Simon, 2012; Hann et al., 2020). The relationship between calcium and ETosis has been supported by different studies in eosinophils, but mostly by studies in neutrophils (Parker et al., 2012; Gupta et al., 2014; Ueki et al., 2016a; Kenny et al., 2017; de Bont et al., 2018). Moreover, the activation of Akt has been found to be regulated by an elevation of calcium during the NETosis process (Douda et al., 2014). Thus, our finding that the calcium chelator BAPTA-AM inhibited EETs release induced by A. fumigatus corroborates these previous observations. Ueki and colleagues found that EDTA inhibited the EETs extrusion induced by A23187, PMA, or immobilized IgG in human eosinophils (Ueki et al., 2016a). In neutrophils, Kenny and colleagues demonstrated the relevance of intracellular calcium for PMA-induced NET production by showing a strong inhibition of NETosis after treating neutrophils with BAPTA-AM (Kenny et al., 2017). In agreement with this observation, another work also provided evidence that external calcium is dispensable for NETs extrusion induced by PMA (Douda et al., 2015). In contrast, a different study showed that IL-8-mediated NET formation requires calcium fluxes from both intracellular and extracellular pools, while only extracellular calcium appeared to be important for PMA-mediated NET generation (Gupta et al., 2014). Accordingly, ionomycin and other calcium ionophores have been widely used as known inducers of NETs and EETs (Douda et al., 2015; Ueki et al., 2016a). In this context, it is clear that calcium plays a crucial role in A. fumigatus-induced ETTs extrusion. However, whether calcium is important for Akt activation during this EETosis process or which is the major source of calcium (intracellular or extracellular or both) are questions still to be answered.

Although several studies have implicated a role for PAD4 in histone citrullination and chromatin decondensation in the process of ETosis (Wang et al., 2009; Li et al., 2010; Lewis et al., 2015; Van Avondt and Hartl, 2018; Kim et al., 2020), others have suggested that PAD4 might not be essential for this process even when H3 citrullination is reduced (Neeli and Radic, 2013; Kenny et al., 2017; Claushuis et al., 2018; Guiducci et al., 2018; Silva et al., 2019; Thompson-Souza et al., 2020). According to most ET studies, the citrullination of histone 3 by the enzyme PAD4, expressed in the nuclei of eosinophils and other granulocytes (Asaga et al., 2001; Nakashima et al., 2002; Kim et al., 2020), results in weakened DNA–histone binding, thereby facilitating the release of DNA from the nucleus and out of the cell (Van Avondt and Hartl, 2018). However, other studies have shown that histone citrullination-independent mechanisms occur in the process of NETs release in response to Candida albicans (Guiducci et al., 2018), bacteria (Streptococcus and Klebsiella pneumoniae) (Kenny et al., 2017; Claushuis et al., 2018), Histoplasma capsulatum (Thompson-Souza et al., 2020) and A. fumigatus (Silva et al., 2019). Under our conditions, we observed that EETs generated in response to A. fumigatus exhibited histone citrullination that was PAD4-dependent; however, PAD4 and H3 citrullination are not essential for EETs extrusion. In fact, the existence of NET release pathways independent of histone citrullination has been described for PMA, one of the most recognized promoters of NET release (Neeli and Radic, 2013; Kenny et al., 2017). In this context, the dependency of ETs extrusion on PAD4 and histone citrullination is questionable and might depend on the stimuli and cell type involved. One question that remains in eosinophils is which molecule is responsible for chromatin decondensation, since it can occur independently of PAD4 and histone citrullination. In neutrophils, elastase and myeloperoxidase have been described as important for this process (Papayannopoulos et al., 2010), but in eosinophils, this is a point for future investigations. Moreover, the consequences of the release of non-citrullinated EETs for the host immune response, for the fungus, and for ABPA development remain unknown. We previously showed that EETs do not contribute to the killing and impairment of A. fumigatus conidia (Muniz et al., 2018). Thus, the impact of non-citrullinated EETs on fungal viability is potentially not critical. However, other consequences of these non-citrullinated EETs regarding ABPA development and the host immune response cannot be discarded. In a study of Klebsiella pneumoniae–induced pneumonia, NETs formed in the absence of PAD4 and histone citrullination did not affect bacterial growth or lung inflammation (Claushuis et al., 2018). Whether the same is valid for eosinophils or eosinophilic lung inflammation in the context of ABPA remains to be addressed. Our findings also demonstrated that A. fumigatus is capable of inducing EETs release independent of fungal viability. In previous works, dead A. fumigatus conidia were also capable of inducing neutrophils to release extracellular DNA traps (Bruns et al., 2010; McCormick et al., 2010). Thus, fungal cell metabolism seems to play no active role in this interaction, both in EETs and NETs formation.

In this study, we identified several components that drive EETs release in response to the fungus A. fumigatus. Based on this and previous studies (Muniz et al., 2018), we believe that the excessive release of EETs may contribute to the formation of sticky mucus, which in turn contributes to airway obstruction and lung function impairment. Furthermore, because EETs lack killing activity against A. fumigatus (Muniz et al., 2018), their association with clusters of functional FEGs may potentiate the pro-inflammatory roles of these structures. Thus, therapeutic interventions to avoid or degrade these DNA traps may represent an interesting approach to minimize inflammatory lung damage without major impacts on fungal clearance. In addition to clarifying mechanisms that underlie the interaction between A. fumigatus and eosinophils, these findings may help improve our understanding of ABPA pathogenesis and treatment.

The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding author/s.

The studies involving human participants were reviewed and approved by Committee on Human Research at Clementino Fraga Filho Hospital (Federal University of Rio de Janeiro). The patients/participants provided their written informed consent to participate in this study.

MVB, IG, MAAD, GAT-S, VSM, and CRIV conducted the experiments and acquired and analyzed the data. MVB, RCNM, RTF, and JSN designed the research studies, analyzed the data and provided reagents. MVB and JSN wrote the manuscript. RCNM and RTF critically revised the manuscript for important intellectual content. All authors contributed to the article and approved the submitted version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.