Abstract

   Asthma is phenotypically heterogeneous with several distinctive
   pathological mechanistic pathways. Previous studies indicate that
   neutrophilic asthma has a poor response to standard asthma treatments
   comprising inhaled corticosteroids. Therefore, it is important to
   identify critical factors that contribute to increased numbers of
   neutrophils in asthma patients whose symptoms are poorly controlled by
   conventional therapy. Leukocytes release chromatin fibers, referred to
   as extracellular traps (ETs) consisting of double-stranded (ds) DNA,
   histones, and granule contents. Excessive components of ETs contribute
   to the pathophysiology of asthma; however, it is unclear how ETs drive
   asthma phenotypes and whether they could be a potential therapeutic
   target. We employed a mouse model of severe asthma that recapitulates
   the intricate immune responses of neutrophilic and eosinophilic airway
   inflammation identified in patients with severe asthma. We used both a
   pharmacologic approach using miR-155 inhibitor-laden exosomes and
   genetic approaches using miR-155 knockout mice. Our data show that ETs
   are present in the bronchoalveolar lavage fluid of patients with mild
   asthma subjected to experimental subsegmental bronchoprovocation to an
   allergen and a severe asthma mouse model, which resembles the complex
   immune responses identified in severe human asthma. Furthermore, we
   show that miR-155 contributes to the extracellular release of dsDNA,
   which exacerbates allergic lung inflammation, and the inhibition of
   miR-155 results in therapeutic benefit in severe asthma mice. Our
   findings show that targeting dsDNA release represents an attractive
   therapeutic target for mitigating neutrophilic asthma phenotype, which
   is clinically refractory to standard care.

   Keywords: asthmatic inflammation, neutrophilic asthma phenotype,
   extracellular trap, miR-155, exosome delivery

Introduction

   Over the last two decades, asthma incidence in the United States has
   increased from 5.5% to 7.8% of the general population and is a major
   health concern for patients and medical professionals (CDC, Most Recent
   National Asthma Data, 2019). Asthma is the most common chronic
   inflammatory disease with several characterized endotypes and
   phenotypes ([45]1–[46]3). Pathophysiologic mechanisms in asthma and
   treatment strategies are determined by inflammatory phenotypes ([47]4).
   Unlike the relatively well-defined mechanisms that result in type 2
   high eosinophilic asthma inflammation, those leading to non-type 2 high
   mixed neutrophilic/eosinophilic asthma are poorly understood, and the
   specific treatments based on molecular pathogenesis are yet to be
   developed. Although only 10% of asthmatics are categorized as mixed
   cellular ([48]5), they account for nearly 200,000 hospitalizations and
   3,500 deaths per year in the United States ([49]6). Non-type 2 severe
   asthma is poorly responsive to corticosteroid therapy, compared to the
   highly steroid-responsive type 2 eosinophilic asthma ([50]7). Recent
   evidence indicates that asthma exacerbations resulting from excessive
   activation of the innate immune responses are manifested as airway
   neutrophilia, although details of the mechanism are not fully
   elucidated ([51]8–[52]10). Given the frequency, severity, and
   disability associated with the neutrophilic or mixed cellular asthma
   phenotypes, it is crucial to identify molecular mechanisms that
   contribute to mixed cellular asthma that is poorly responsive to
   conventional therapy.

   Although several mediators have been implicated in neutrophil
   recruitment to the airways ([53]9, [54]11, [55]12), these have not
   proven to be pivotal in the pathogenesis of neutrophilic and/or mixed
   neutrophilic/eosinophilic asthma, indicating that there are gaps in our
   knowledge of the immune response in the airways of those with severe
   asthma. Therefore, it is important to identify critical factors
   contributing to increased neutrophils in asthmatic airways of these
   refractory phenotypes. To understand the complex immune responses of
   severe human asthma, it is imperative to develop a mouse model of
   severe asthma that displays mixed granulocytic infiltrates in the
   airways ([56]7). Although no ideal mouse model provides a comprehensive
   nature of severe human asthma ([57]13), our model is characterized by
   mixed eosinophilic and neutrophilic inflammation, which is ideal for
   determining mechanisms and potential therapeutic targets in
   corticosteroid-refractory severe asthma.

   Neutrophils have an important role in innate defense mechanisms.
   However, in asthma, airway neutrophils have been associated with
   disease severity and acute exacerbations ([58]14). Elevated sputum
   eosinophils and neutrophils are associated with the lowest lung
   function in asthma ([59]15–[60]17). Specifically, the interaction of
   eosinophils and neutrophils is associated with a greater decline in
   lung function compared to any single inflammatory pathway in severe
   asthma ([61]17). Th1/IFN-γ and Th17/IL-17 pathways play critical roles
   in neutrophil recruitment to the lungs in both neutrophilic and mixed
   neutrophilic/eosinophilic severe asthma ([62]18–[63]20). The bacterial
   second messenger cyclic (c)-di-GMP has been shown to be a mucosal
   adjuvant that induces a Th1–Th17 response ([64]21, [65]22). Upon
   activation, neutrophils release chromatin fibers, referred to as
   extracellular traps (ETs), consisting of host double-stranded (ds)DNA,
   histones, and granule contents ([66]23). Web-like structure ETs have
   antimicrobial defense activity by ensnaring microbes in a sticky matrix
   of extracellular chromatin and bactericidal proteins ([67]24) but also
   have physiologic and pathologic roles in asthmatic airways
   ([68]25–[69]27). Although multiple lines of evidence showed that ETs
   released by activated neutrophils and eosinophils increase asthma
   severity ([70]26, [71]28, [72]29), the mechanisms by which ETs
   influence the progression of mixed neutrophilic/eosinophilic severe
   asthma have not yet been defined.

   Our study shows that ETs were released in the bronchoalveolar lavage
   fluid (BALF) of patients with asthma who underwent subsegmental
   bronchoprovocation with allergen (SBP-AG) challenge. In a murine asthma
   model that resembles the complex immune responses identified in severe
   human asthma, we observed that microRNA (miR)-155 regulates the release
   of ETs, which is associated with the severity of allergic immune
   responses. Using a serum-derived exosome-based oligonucleotide delivery
   system, we blocked the release of ETs in vivo and attenuated the
   severity of allergic lung inflammation in the mouse model. Our results
   provide strong pre-clinical evidence that targeting host dsDNA release
   is an effective therapeutic approach for mitigating mixed
   neutrophilic/eosinophilic asthma lung inflammation and alleviating
   severe airway hyperresponsiveness.

Materials and methods

   Detailed methods are described in the [73]Supplementary Material.

Subsegmental bronchoprovocation with allergen bronchoscopy protocol

   This protocol was approved by the Institutional Review Board (IRB) of
   the University of Illinois (Chicago, IL, USA), and an investigational
   new drug (IND) was obtained from the Food and Drug Administration (FDA)
   for bronchoscopic administration of allergens to volunteers. The
   details of the protocol were described in our previous publication
   ([74]30, [75]31). In brief, we recruited study subjects who are on step
   1 asthma therapy according to the National Asthma Education and
   Prevention Program (NAEPP) Asthma Guidelines. After informed consent
   was obtained, subjects underwent screening for inclusion and exclusion
   criteria, which included skin prick testing for dust mite, short
   ragweed, and cockroach allergens. Once the screening was complete, BAL
   was performed per standard research guidelines approved by the IRB.
   Pre-challenge BAL samples were obtained right before SBP-AG. At 48 h
   after the SBP-AG, post-challenge BAL samples were obtained from the
   allergen exposed, at the adjacent and contralateral subsegment. The
   maximum challenge dose for SBP-AG was 5 ml of 100 binding antibody
   units (BAU)/ml or 1:2,000 wt/vol concentration of allergen. The mean
   age of the UIC was 24 (range 18–36), and the male-to-female ratios were
   2:4.

Mice

   C57BL/6, miR-155KO (007745), and PAD4KO mice (030315) were purchased
   from the Jackson Laboratories (Bar Harbor, ME, USA). Myeloid miR-155
   knockout mice and miR-155 ^fl/fl mice (026700) were crossed with
   LysM-cre mice (004781, Jackson Laboratories, Bar Harbor, ME, USA) to
   homozygosity (miR-155 ^fl/fl LysMcre). Littermates were used as
   controls. DNA extraction and genotyping were performed as described
   previously ([76]32). Unless otherwise stated, mice at age-matched 8 to
   12 weeks were used in this study.

Induction of murine asthma model

   We used the published triple-allergen (DRA)-induced mouse asthma model
   with minor modifications ([77]30, [78]33, [79]34). DRA mixture includes
   extracts of dust mite (Dermatophagoides farina, 830 µg/ml), ragweed
   (Ambrosia artemisiifolia, 8.34 mg/ml), and Aspergillus fumigatus (830
   µg/ml; Stallergenes Greer, Lenoir, NC, USA). Mice were sensitized with
   the DRA allergen mixture on days 0 and 5 by intranasal (i.n.)
   injection. Seven days later (day 12), mice were challenged with DRA
   mixture (200 µg) at the same concentration used for sensitization in 20
   µl of saline once daily for 3 days/week for 4 weeks. Twenty-four hours
   following the final challenge, mice were sacrificed, and BALF and lung
   tissues were collected for further analysis. All sensitized animals had
   elevated levels of total IgE.

   In the severe asthma model, mice were sensitized with DRA (100 µg) and
   cyclic-di-GMP (c-di-GMP) (5 µg; Tocris, Minneapolis, MN, USA)
   intranasally on days 0 and 5. Seven days later (day 12), mice were
   subjected to three challenge sets involving three consecutive
   challenges with DRA (100 µg) and c-di-GMP (0.5 µg) with a rest of 4
   days in between challenge sets ([80]7). Mice were then euthanized, and
   features of allergic airway inflammation were assessed, 24 h following
   the final challenge. The schematic of this model is shown in
   [81]Figure 1A.

Figure 1.

   [82]Figure 1
   [83]Open in a new tab

   Leukocytes released ET-derived host dsDNA in asthma. (A) PicoGreen
   assays of extracellular dsDNA in the BALF samples obtained pre- and
   post-allergen challenge in the subsegmental bronchoprovocation with
   allergen (SBP-AG) protocol (N = 3). (B) Western immunoblot for citH3 in
   BALF samples obtained pre- and post–allergen challenge in the SBP-AG
   protocol. (C) NE from BALF in the SBP-AG protocol. (D) Mice were
   sensitized and challenged by daily intranasal (i.n.) administration of
   DRA or saline, 3 days a week for 4 weeks. Cell staining for markers of
   neutrophils (CD45^+Ly6G^+) and eosinophils
   (CD45^+Ly6G^−SiglecF^+CD11c^−). (E) Airway hyperresponsiveness (AHR) in
   mice subjected to the 4 weeks triple allergen asthma model. Shown are
   central airway resistance (Newtonian resistance, Rn) values. (F)
   Concentration of extracellular dsDNA and NE in the BALF of mice
   subjected to triple allergen asthma model (N = 6). (G) The level of
   neutrophil chemokine KC and MPO in BALF from DRA-induced asthmatic mice
   (N = 6). (H) Representative confocal microscopy images of lung
   macrophages from allergen-exposed mice producing ETs with co-expression
   of citH3 (arrow). Blue, staining for DNA; red, citH3; green,
   macrophages. Graphs were plotted as mean ± SE. p-Values were obtained
   using a two-tailed Student’s t-test. BALF, bronchoalveolar lavage
   fluid; ET, extracellular trap; NE, neutrophil elastase; MPO,
   myeloperoxidase.

Delivery of small RNA-loaded exosomes in vivo

   Serum exosomes were isolated according to the manufacturer’s protocol
   (Thermo Fisher Scientific, Waltham, MA, USA). Nanoparticle tracking
   analysis was performed to determine the size and concentration of serum
   exosomes (NanoSight NS300). Modified calcium-mediated transfection was
   used to introduce small RNAs, including the miR-155 inhibitor (catalog
   no. MSTUD0080, Sigma, St. Louis, MO, USA) and inhibitor control
   (catalog no. NCSTUD001, Sigma) into serum exosomes ([84]35). To deliver
   small RNAs into serum exosomes, 100 pmol of small RNA was loaded into
   100 µg of serum EXOs quantified by protein content ([85]36). For in
   vivo experiments, small RNA-loaded exosomes in 30 µl of saline were
   i.n. instilled into the mouse lung 30 min before the challenge at 3
   weeks. One day after the last instillation, mice were sacrificed, and
   the cellular profile in the BALF was evaluated.

Measurements of airway hyperresponsiveness

   Mechanical properties of the mouse lung were assessed in mice
   anesthetized with ketamine/xylazine using the forced oscillation
   technique as previously described ([86]34). Anesthetized mice were
   mechanically ventilated on a FlexiVent computer-controlled ventilator
   (SCIREQ, Montreal, QC, Canada) with 8 ml/kg tidal volume at a frequency
   of 150 breaths/min and 2–3 cmH[2]O positive end-expiratory pressure.
   Continuous EKG and pulse oximetry monitoring (Starr Life Sciences,
   Oakmont, PA, USA) was performed. Pressure and flow data (reflective of
   the airway and tissue dynamics) were used to calculate lung resistance,
   static lung compliance, and dynamic lung compliance at baseline with
   the use of the single-compartment and constant phase model.
   Standardized recruitment maneuvers were performed prior to each
   physiologic measurement to prevent atelectasis and standardize volume
   history. Maximal airway responsiveness was measured following exposure
   to increasing doses of nebulized methacholine (MCh) (0–50 mg/ml).

DsDNA measurement in bronchoalveolar lavage fluid

   DsDNA was measured in the acellular fraction of the BALF, which was
   obtained after a double centrifugation and supernatant collection.
   Levels of dsDNA were determined with Quant-iT PicoGreen dsDNA reagent
   (Invitrogen, Carlsbad, CA, USA) according to the manufacturer’s
   protocol. For immunoblot analysis for citH3 to detect ETs, anti-citH3
   (ab5103, Abcam, Cambridge, UK) was used.

Single-cell RNA sequencing

   Single cells were obtained from asthma lung tissues by mechanical and
   enzymatic dissociation ([87]33). After enzymatic digestion, the cells
   were placed in an inactivation buffer (Roswell Park Memorial Institute
   (RPMI) 1640 + 10% fetal bovine serum (FBS)) and filtered to remove any
   remaining tissue, the filtrate was centrifuged at 300g, the cell pellet
   was suspended in red blood cell (RBC) lysis buffer and resuspended in
   inactivation buffer, and viability was determined to be 85%–95% before
   submitting for approximately 0.75 to 1 × 10^5 cells for single-cell RNA
   sequencing (scRNA-seq) analysis. Library preparation was carried out
   using 10x Genomics Chromium controller (10x Genomics). Chromium 10x
   Genomics single-cell library was sequenced as 150 base pair single-end
   reads on an Illumina NovaSeq 6000 (Illumina, Inc., San Diego, CA, USA),
   with cells from each of the two microchips sequenced across three flow
   cell lanes.

Bioinformatics analysis for next-generation sequencing data

   Single-cell data in feature-barcode matrices were processed using Cell
   Ranger 6.0.0 ([88]37) using the GRCm38 (mm10) genome (downloaded March
   2021 refdata-gex-mm10-2020-A build from10x) to identify unsupervised
   cell clusters. The data were natural log transformed and normalized for
   scaling the sequencing depth to a total of 1e4 molecules per cell,
   followed by regressing out the number of UMI using Seurat 4.0.4
   ([89]38). Briefly, cells were selected based on RNA content thresholds:
   greater than 300 RNA features, less than 2,000 unique expressed genes
   (nFeature_RNA), and mitochondrial gene content less than 25% using the
   “subset” function. Preprocessed data from the six samples were
   integrated using the Seurat function, the FindIntegrationAnchors
   function, and IntegrateData using 20 principal components as default
   parameters normalized with NormalizeData function. The variable
   features were found with FindVariableFeatures function using 2,000
   features. After scaling and principal component analysis (PCA)
   normalization with default parameters, clustering was performed using
   Seurat functions FindNeighbors and FindClusters functions using 20 PCs
   at resolution of 0.8 with the Louvain algorithm. The identified
   clusters were then visualized using Uniform Manifold Approximation and
   Projection (UMAP) embedding using 20 PCs. The feature plots were
   generated using Seurat function FeaturePlot using the RNA assay. Module
   score was calculated from inputting genes using the Seurat-implemented
   function AddModuleScore and generated using the FeaturePlot function as
   described ([90]38). The average expression levels of each cluster at
   the single-cell level are subtracted by the aggregated expression of
   random control feature sets (200 by default). For dot plot figures,
   normalized expression values were generated using NormalizeData
   function from the RNA assay after splitting clusters by condition. This
   would create a matrix of natural log-normalized gene expression with a
   scaling factor of 10,000. The size of the circle represents the number
   of cells in the cluster when the gene was detected. z-Score dot plots
   were generated using raw expression data using the scale function in
   Seurat dot plot function. Pathway enrichment analysis for lymphocyte
   clusters was performed using Ingenuity Pathway Analysis (IPA) ([91]39).
   Raw data for IPA analysis was found using Seurat implemented function
   FindAllMarkers using parameters: only.pos = FALSE, min.pct = 0.2,
   thresh.use = 0.2. IPA pathway z-score of greater than 2 is considered
   activated, while a score of less than −2 is considered an inhibited
   pathway. Canonical Pathway figures were generated using ggplot2
   [[92]https://cran.r-project.org/web/packages/ggplot2/citation.html] in
   R. Functional enrichment analysis by cluster was performed using
   enrichR [[93]https://maayanlab.cloud/Enrichr/] DEenrichRPlot function
   using the Gene ontology (GO) 2021 pathways and Kyoto Encyclopedia of
   Genes and Genomes (KEGG) 2019 mouse databases
   [[94]https://www.genome.jp/kegg/kegg1.html].

Statistical analysis

   Data were represented as box-and-whisker plots unless otherwise noted.
   Results were expressed as means ± SE. A comparison between two groups
   was performed with Student’s t-test for unpaired variables. Statistical
   significance is indicated in figure legends. A p < 0.05 was considered
   statistically significant.

Results

Leukocytes release extracellular trap-derived dsDNA in asthma

   ETs have been shown to be present in the airways of asthmatics ([95]25,
   [96]26). High expression of ETs-derived dsDNA correlates with airway
   neutrophilia and epithelial cell damage and is related to mucus
   production ([97]5, [98]9). Consistent with these findings, our data
   show that leukocyte-released ET-derived dsDNA presents in BALF of
   volunteers with asthma who were experimentally challenged with
   allergens ([99]Figure 2A). As illustrated, there was a dramatic
   increase in extracellular dsDNA in the allergen-challenged site
   compared with the pre-challenge or contralateral post-challenge sites,
   with a significantly lower level in the immediate subsegment adjacent
   to the allergen-challenged site. To confirm that the extracellular
   dsDNA is derived from ETosis, we measured hyper-citrullinated histone
   H3 (citH3), a characteristic of ETs. CitH3 was present in the
   post-challenged sites, the highest at the experimental site
   ([100]Figure 2B). Consistent with this finding, neutrophil elastase
   (NE) was also higher in the allergen-challenged site ([101]Figure 2C).

Figure 2.

   [102]Figure 2
   [103]Open in a new tab

   Host dsDNA release was greatly increased in the airways of severe
   asthma (SA) mice. (A) Schematics of the mouse model of SA used in the
   study. Mice were sensitized and challenged with c-di-GMP and DRA for 3
   weeks, and BALF was collected on day 29. Representative gating strategy
   outlining airway neutrophils and eosinophils in SA mice. (B) PicoGreen
   assay and Western immunoblot for citH3 showing ET-associated DNA
   present in BALF. (C) NE in the BALF of mice in the SA model. (D) We
   observed amorphous, unshaped, extracellular DNA structures that
   colocalized with citH3 (red) and MPO (green) in severe asthma mouse
   lungs, as identified by immunofluorescence. (E) Representative confocal
   microscopy images of BAL cells from SA mice producing MPO with
   co-expression of PAD4 (arrow). Blue, staining for DNA; purple, MPO;
   green, macrophages; cyan, PAD4. (F) Expression of PAD4 as determined by
   Western immunoblot. (G) The level of neutrophil chemokine chemokines
   and cytokines in BALF from SA mice (N = 6). Graphs were plotted as mean
   ± SE. p-Values were obtained using a two-tailed Student’s t-test. ET,
   extracellular trap; BALF, bronchoalveolar lavage fluid; NE, neutrophil
   elastase.

   We also observed an increase in BALF dsDNA and NE levels in a murine
   model of chronic asthma ([104]Figure 2F), in which mice were challenged
   with the common triple allergen DRA (dust mite, ragweed, and
   Aspergillus) for 4 weeks as our group previously reported
   ([105]Figure 2D) ([106]33). As shown, neutrophils (Neu, CD45^+Ly6G^+)
   and eosinophils (Eos, CD45^+Ly6G^−SiglecF^+CD11c^−) were detected in
   the airways of allergen-challenged mice ([107]Figure 2D), with was
   associated with accentuated MCh-induced airway responsiveness
   ([108]Figure 2E). Notably, the levels of neutrophil chemokine CXCL1
   (also known as Gro-α in humans and KC in mice) and myeloperoxidase
   (MPO), a surrogate marker of increased neutrophils, were significantly
   increased in BALF from the allergen-challenged mice compared to
   saline-challenged control ([109]Figure 2G). Finally, we detected
   extracellular DNA structures that prominently colocalized with citH3 in
   pulmonary macrophages and granulocytes from the mouse model of asthma
   ([110]Figure 1H, arrows). These data indicate that recruited
   leukocytes, neutrophils and eosinophils, are the major cellular source
   of dsDNA in the asthmatic airways in humans and mice.

Host dsDNA release is significantly increased in the airways of severe asthma
mice

   Airway neutrophilia is one of the hallmarks of acute exacerbation or
   severe subgroups of asthma ([111]14). In order to address this
   important clinically relevant issue, we employed a mouse model of
   severe asthma that recapitulates the intricate immune responses of
   neutrophilic and eosinophilic airway inflammation identified in
   patients with severe asthma ([112]16, [113]17, [114]40). In this severe
   asthma model (hereafter, SA model), mice are sensitized to the three
   allergens (DRA) along with c-di-GMP followed by 1 week of rest and then
   3 weeks of intranasal challenge ([115]Figure 1A) ([116]7). A key
   feature of this model is increased numbers of eosinophils and
   neutrophils in BAL samples that were characterized by flow cytometry.
   In this model, the recruited neutrophils underwent ETosis in the lung,
   as evidenced by increased BALF dsDNA, citH3, and NE ([117]Figures 1B,
   [118]C). Using immunofluorescence microscopy, we observed amorphous and
   unshaped structures colocalized with citH3 and MPO in severe asthma
   mouse lungs ([119]Figure 1D). Peptidyl arginine deiminase 4 (PAD4) has
   been shown to be involved in ETosis by hyper-citrullination of histones
   ([120]5). Notably, high-resolution confocal microscopy indicated the
   presence of MPO/PAD4-positive neutrophils in the BAL cells of the
   severe asthma mice ([121]Figure 1E), indicating that recruited
   neutrophils are a cellular source of dsDNA and citH3 in severe
   asthmatic airways. The level of PAD4 protein was much higher in the
   whole lung tissue of the severe asthma mice than in the saline-treated
   control ([122]Figure 1F) along with high levels of BAL-KC and BAL-MPO
   in the severe asthma mice ([123]Figure 1G).

   Neutrophils and eosinophils were increased in BAL fluids of the mice
   subjected to the SA model, compared with saline-challenged control
   ([124]Supplementary Figures 1A, [125]B), but the SA model had fewer BAL
   eosinophils compared to the triple allergen (DRA)-only model
   ([126]Supplementary Figure 2). The DRA/c-di-GMP challenge significantly
   induced the expression of genes associated with Th2 and Th17 immune
   responses (including Il13, Il17, Ifng, Kc, and Muc5a/c) compared to the
   saline-treated control ([127]Supplementary Figure 1C). Periodic
   acid-Schiff (PAS) staining detected that mucus glycol conjugates of
   goblet cells were also increased in the severe asthma mice
   ([128]Supplementary Figure 1D). Of note, the severe asthma mice had
   highly increased bronchial hyper-reactivity to MCh ([129]Supplementary
   Figure 1E), compared to that of the triple allergen (DRA)-only model
   ([130]Figure 2E). Total serum IgE and the levels of BALF Th2
   cytokines/chemokines such as CCL17, CCL22, IL-13, IL-17, and periostin
   were elevated in the severe asthma mice compared to the saline-treated
   control ([131]Figure 1G).

MiR-155 is involved in extracellular trap formation in severe asthma

   Recent studies indicate that miR-155 plays an indispensable role in
   Th1/Th2/Th17 immune responses ([132]41, [133]42). In this study, we
   hypothesized that miR-155 is involved in ETs formation and related
   cellular functions in macrophages. We observed that lungs and sorted
   alveolar macrophages (AMs) from the severe asthma mice contained higher
   levels of miR-155 than those derived from the saline-treated control
   ([134]Figure 3A and [135]Supplementary Figure 2B). Of interest, the
   myeloid cell-specific miR-155-deficient mice (miR-155^fl/flLysM-cre)
   showed a reduction in extracellular dsDNA level in BALF as compared to
   control littermates (miR-155^fl/fl) at a steady state
   ([136]Supplementary Figures 3A–[137]C). However, there was no
   difference in BAL cell numbers ([138]Supplementary Figure 3D).

Figure 3.

   [139]Figure 3
   [140]Open in a new tab

   MiR-155 is involved in ET formation in severe asthma (SA). (A)
   Expression of miR-155 in lung from SA mice (N = 3–5). (B) Total cells
   and (C) leukocyte differentials were counted in the BALF of the mouse
   model of SA used in the study (N = 6-8). (D) PicoGreen assay and
   Western immunoblot for citH3 showing ET-associated DNA present in BALF.
   (E) NE in the BALF from WT and miR-155KO mice (N = 5). (F) PAD4 and
   citH3 proteins as determined by Western immunoblotting from SA mouse
   lung. (G) Representative confocal microscopy images of BAL cells from
   SA mice producing MPO with co-expression of PAD4 (arrow). Blue,
   staining for DNA; cyan, MPO; green, macrophages; red, PAD4. (H)
   PAS-stained sections from WT (top) and miR-155KO (bottom) mouse lungs
   subjected to the SA model. Boxed regions are shown enlarged at right.
   (I) Airway hyperresponsiveness (AHR) in WT and miR-155KO mice subjected
   to the SA model. Graphs were plotted as mean ± SE. p-Values were
   obtained using a two-tailed Student’s t-test. ET, extracellular trap;
   BALF, bronchoalveolar lavage fluid; NE, neutrophil elastase; MPO,
   myeloperoxidase; PAS, periodic acid-Schiff; WT, wild type.

   When mice were subjected to the SA model, miR-155 whole-body deficiency
   (miR-155KO mice) significantly decreased the total number of BAL cells
   ([141]Figure 3B) with a marked reduction in BAL eosinophil and
   neutrophil counts compared with that in wild-type (WT) mice
   ([142]Figure 3C). Furthermore, the miR-155KO mice had markedly reduced
   cell-free dsDNA ([143]Figure 3D) and NE ([144]Figure 3E) in BAL fluids,
   which are characteristics of ETosis. Consistent with these findings,
   miR-155KO mouse lungs subjected to the SA model had decreased PAD4 and
   citH3 expressions ([145]Figure 3F). This lower propensity of the
   miR-155KO mice to ETosis in severe asthma seems to be associated with a
   decreased number of PAD4-positive neutrophils in BALF ([146]Figure 3G).
   Together, these findings indicate that miR-155 regulates ETosis and is
   pivotal to neutrophilic lung inflammation such as severe asthma.

   The miR-155KO mice subjected to the SA model had significantly
   decreased mRNA levels of characteristic genes involved in eosinophilic
   (Il4, Il13, Ccl17, and Muc5a/c) and neutrophilic (Il17, Mmp12, and Kc)
   inflammation compared to the WT counterparts ([147]Supplementary
   Figure 4A). Similarly, the miR-155KO mice subjected to the SA model had
   decreased PAS staining ([148]Figure 3H) and blunted MCh-induced airway
   responsiveness ([149]Figure 3I) as well as marked deficiency in Th2 and
   Th17 cytokine productions ([150]Supplementary Figure 4B). Consistent
   with these findings, BAL Th2 and Th17 cytokines were also significantly
   decreased in the miR-155KO mice ([151]Supplementary Figure 4C).
   Collectively, these data indicate that miR-155 regulates Th2 and Th17
   allergic lung inflammation via ETosis process.

Inhibition of miR-155 alters eosinophilic and neutrophilic inflammation in
the mouse model of severe asthma

   Next, we examined whether miR-155 could be targeted to decrease
   allergic lung inflammation by inducing defective ETosis. Emerging
   evidence indicates that exosomes can serve as vehicles for therapeutic
   drug delivery ([152]43, [153]44). We established a protocol using
   serum-derived exosomes to deliver the specific inhibitor against
   miR-155 into the lung in vivo, as previously described ([154]35). We
   purified exosomes from mouse serum and determined the size of
   serum-derived exosomes and inflammatory lung response in vivo.
   Nanoparticle tracking analysis measurement (NanoSight) showed that the
   purified serum-derived exosomes were ~78 nm in diameter, as expected
   from the previous study ([155]45) ([156]Supplementary Figure 5A).
   Incubation of serum-derived exosomes with primary alveolar macrophages
   showed no cytotoxic effect ([157]Supplementary Figure 5B).
   Serum-derived exosome treatment on mouse lungs did not alter the number
   of BAL inflammatory cells ([158]Supplementary Figure 5C). To evaluate
   the efficiency of miR-155 inhibition in vivo, a synthetic miR-155
   inhibitor (mmu-miR-155-5p, Sigma-Aldrich) was loaded into exosomes
   using modified calcium-mediated transfection and delivered to the lung
   via intranasal instillation ([159]Supplementary Figure 5D). As shown,
   exosome-mediated delivery of the specific miR-155 inhibitor decreased
   the miR-155 level in the lung, compared with the group treated with a
   control vector ([160]Supplementary Figure 5E).

   In the SA model, the neutrophil influx into BAL was significantly
   decreased in miR-155 inhibitor/exosome-treated mice compared with
   control/exosome-treated mice ([161]Figure 4A, 19.3% vs. 34%). The
   exosome-mediated delivery of miR-155 inhibitor significantly decreased
   the total number of BAL cells with a marked reduction in BAL eosinophil
   and neutrophil numbers, compared with control/exosome-treated mice
   ([162]Figures 4B, [163]C), although both groups were equally sensitized
   to the allergens because total serum IgE levels were identical in the
   SA model ([164]Figure 4D). Moreover, BAL-dsDNA and citH3 were markedly
   decreased in BALF of the miR-155 inhibitor/exosome-treated mice
   subjected to the SA model as compared with their counterpart control
   ([165]Figure 4E), indicating the critical roles of miR-155 in ETosis.
   Concordant with the decrease of ETosis, the miR-155
   inhibitor/exosome-treated mice had attenuated levels of BAL-NE
   ([166]Figure 4F).

Figure 4.

   [167]Figure 4
   [168]Open in a new tab

   Delivery of miR-155 inhibitor has broad immunosuppressive potential in
   severe asthma (SA) mice. (A) Delivery of miRNA inhibitor via
   serum-derived exosomes into the lung macrophages in vivo. Serum
   exosomes measuring 100 μg transfected with 100 pmol of inhibitor
   control or miR-155 inhibitor (miR-155-5p) were given. Mice were
   sensitized and challenged with c-di-GMP and DRA and treated with
   miR-155 inhibitor or control on days 25, 26, and 27. Neutrophil influx
   in BAL from SA mice is analyzed by flow cytometry (N = 5-7). (B) Total
   cells and (C) leukocyte differentials were counted in the BALF. (D) All
   SA animals had elevated levels of total IgE. (E) PicoGreen assay and
   Western immunoblot for citH3 showing ET-associated DNA present in BALF.
   (F) NE in the BALF of mice to the SA model. Graphs were plotted as mean
   ± SE. p-Values were obtained using a two-tailed Student’s t-test. BAL,
   bronchoalveolar lavage; BALF, bronchoalveolar lavage fluid; ET,
   extracellular trap; NE, neutrophil elastase.

   Next, we examined the asthma features of the SA model with the
   exosome-mediated delivery of a miR-155 inhibitor. The treatment with a
   miR-155 inhibitor in the lung in the SA model induced a marked
   reduction in the expression of Il13, Il17, Ifng, Kc, and Ccl17 genes in
   the lung ([169]Figure 5A); decreased goblet cell hyperplasia
   ([170]Figure 5B); and lowered PAD4 protein expression in the lung as
   compared to the control group ([171]Figure 5C). Airway
   hyperresponsiveness (AHR) was also markedly lower in the miR-155
   inhibitor/exosome-treated mice ([172]Figure 5D). Similarly, there was a
   significant deficiency in Th2 and Th17 cytokine production
   ([173]Supplementary Figure 5F) and the secretion of BAL-cytokines such
   as KC, IL-17, IL-13, IL-5, and CCL17 in the miR-155 inhibitor-treated
   group ([174]Figure 5E). Together, these data indicate that miR-155 has
   a pivotal role in promoting ETosis, and miR-155 could be a target to
   decrease mixed Th1/Th2/Th17 immune responses.

Figure 5.

   [175]Figure 5
   [176]Open in a new tab

   MiR-155 inhibition alters eosinophilic and neutrophilic inflammation in
   the airways of severe asthma (SA) mice. (A) The mRNA expression in lung
   tissues from Con/Exo- and 155-5p/Exo-treated mouse lungs subjected to
   the SA model (N = 5–7). (B) Representative histologic lung sections
   stained with PAS. Boxed regions are shown enlarged at right. (C)
   Expression of PAD4 as determined by Western immunoblot. (D) Airway
   hyperresponsiveness (AHR) in Con/Exo- and 155-5p/Exo-treated mice
   subjected to the SA model. (E) The level of cytokines in BALF (N =
   5–7). Graphs were plotted as mean ± SE. p-Values were obtained using a
   two-tailed Student’s t-test. PAS, periodic acid-Schiff.

MiR-155 deficiency alters heterogeneity of lung immune cells in a mouse model
of severe asthma

   Next, we performed scRNA-seq to investigate the transcriptional
   profiles of lungs affected by miR-155. Single-cell suspensions of whole
   lungs were generated from the WT and miR-155KO mice subjected to the SA
   model (N = 3 each). Consistent with our previous data ([177]Figure 3C),
   the miR-155KO mice had decreased inflammatory cell influx
   ([178]Supplementary Figures 6A–[179]C). Cells were immediately
   encapsulated and barcoded for library preparation using a 10x Genomics
   microfluidics system, followed by sequencing. We recovered
   approximately 10,000 individual cells, which were annotated based on
   established cell markers ([180]38, [181]46, [182]47). Unsupervised
   clustering analysis revealed 27 different clusters corresponding to
   distinct cell types of the lung, including epithelial, stromal,
   endothelial, and leukocyte lineages with specific molecular markers
   ([183]Figure 6A). The miR-155KO mice subjected to the SA model showed a
   decrease in lung immune cells, whereas the numbers of structural cells
   such as epithelial and endothelial cells were slightly increased
   compared to their WT counterparts ([184]Figure 6B and
   [185]Supplementary Figure 6D). We evaluated the expression of genes
   associated with asthma-related responses and airway remodeling per cell
   clusters ([186]10, [187]47). Several genes were among the most highly
   expressed factors, including Egr1, Il1b, Stat1, Rel, Tcf4, and Nfkb1
   ([188]Supplementary Figure 6E).

Figure 6.

   [189]Figure 6
   [190]Open in a new tab

   Single-cell RNA-seq analysis of the lung cells in severe asthma (SA)
   mice. (A) Single-cell RNA-seq was performed on single-cell suspensions
   from whole lungs of mice. Cells were clustered using a graph-based
   shared nearest-neighbor clustering approach plotted by a UMAP plot.
   UMAP plot of all scRNA-seq data showing a total of 27 distinct cellular
   populations that were identified. (B) UMAP plots of lung cells from the
   miR-155KO and WT SA groups (N = 3 per group). All cells were analyzed
   using canonical correlation analysis with the Seurat R package. (C) The
   expression levels of ET formation-related Ccl7, Lpc1, Tkt, Ak2, Actn4,
   Atf3, Atf4, Bax, Ptgs2, Arg2, Il1b, Rps3, Prdx6, Jun, Cd74, C3, and
   Hmgb1 genes are shown. Each dot size corresponds to the percentage of
   cells in the cluster expressing a gene, and dot color corresponds to
   the mean expression level for the gene in the cluster. (D) UMAP plots
   compare gene module scores between miR-155KO and WT SA mouse lungs
   using ET formation-related genes (Atf3, Atf4, Bax, Ptgs2, Arg2, Il1b,
   Rps3, Prdx6, Jun, Cd74, C3, and Hmgb1). (E) UMAP plots show the
   expression of the indicated ET-related genes in each cluster. UMAP,
   Uniform Manifold Approximation and Projection; WT, wild type; ET,
   extracellular trap.

   Next, we examined ETs formation-related genes in the identified lung
   cell clusters. By using dot plots, we showed that the transcripts of ET
   formation-related genes were substantially downregulated in the
   miR-155KO mice subjected to the SA model ([191]Supplementary
   Figure 6F). We evaluated the expression of genes associated with ETosis
   ([192]48–[193]52) including Atf3, Atf4, Bax, Ptgs2, Arg2, Il1b, Rps3,
   Prdx6, Jun, Cd74, C3, and Hmgb1 in leukocyte clusters of the SA model
   ([194]Figure 6C). We also detected low-level expression of several
   genes in the miR-155KO mice subjected to the SA model. Gene module
   scores for these signature genes were compared between the miR-155KO
   and WT mice. The miR-155KO mice subjected to the SA model have less
   enriched gene module scores of these signature genes compared to their
   counterpart control ([195]Figure 6D). Among these genes, Il1b and Ptgs2
   were differentially expressed in granulocyte clusters between the WT
   and miR-155KO mice. Rps3 was predominantly expressed in T-cell
   clusters, whereas Arg2 was expressed in macrophage clusters
   ([196]Figure 6E). Gene set enrichment analysis using IPA presented the
   lung leukocyte clusters, which showed decreased activation of pathways
   of RHOGDI (Rho GDP-dissociation inhibitor) signaling, apoptosis
   signaling, TWEAK (TNF-related weak inducer of apoptosis) signaling,
   IL-17 signaling, and PAK signaling, which regulate cellular function
   including stress-related response and allergic airway diseases [197]10,
   [198]18, [199]41), during severe asthma in the miR-155KO mice. In
   contrast, PPAR signaling, PPARα/RXRα signaling, and PTEN signaling
   pathways in miR-155KO leukocyte clusters were higher than those of the
   WT counterparts ([200]Supplementary Figure 6G).

PAD4 inhibition results in decreased dsDNA with no effect on granulocytes
influx in severe asthma

   To further address the functional impact of ETosis on asthma, we used
   PAD4 knockout mice (PAD4KO) that have defective ETosis ([201]53).
   PAD4KO mice had a markedly lower level of BALF dsDNA, although
   inflammatory cells infiltrating the lungs were higher than those of
   their WT counterpart ([202]Supplementary Figures 7A–C). Similarly, the
   specific inhibitor for PAD4, GSK484 (10 mg/kg, i.p., three times),
   decreased PAD4 activity in the lung and reduced dsDNA release and NE
   productions in BALF ([203]Supplementary Figures 7D, [204]E). However,
   PAD4 inhibition had no change in the number of total cells or
   granulocyte counts in the lung ([205]Supplementary Figures 7F, [206]G)
   and did not alter miR-155 expression in the SA model
   ([207]Supplementary Figure 7H). These data indicate that PAD4 is
   downstream of miR-155 with a limited function in dsDNA releases, and
   miR-155 has a broader impact on asthma phenotypes along with regulating
   ETosis.

Intranasal instillation of deoxyribonuclease reduces airway eosinophils but
has no effect on airway inflammation in a mouse model of severe asthma

   We also tested the potential impact of naked dsDNA on asthma phenotypes
   with deoxyribonuclease (DNase) I treatment (200 U, i.n., three times)
   ([208]Supplementary Figure 8A). DNase I treatment modestly decreased
   BAL eosinophils in the SA model, whereas BAL neutrophils remained
   unchanged ([209]Supplementary Figures 8A, [210]B). DNase I had a
   similar partial decrease in airway mucus cell metaplasia and decreased
   IL-13 and muc5a/c in the lung, suggesting that DNase I did not affect
   neutrophil response to allergen challenge ([211]Supplementary
   Figures 8C, [212]D).

Discussion

   Asthma is a consequence of complex gene–environment interactions, with
   heterogeneity in clinical presentation ([213]54). Based on the
   development of relevant biological therapies, type 2-high is the most
   well-defined endotype of asthma ([214]55). The more severe form of
   asthma, the mixed granulocytic phenotype, has few endotype-specific
   therapeutic options, and therefore, further study of its underlying
   mechanisms is needed. The severe asthma endotype with airway
   eosinophils and neutrophils is associated with asthma exacerbations and
   is resistant to inhaled and systemic corticosteroids. To address this
   important clinically relevant issue, we developed a severe asthma mouse
   model that recapitulates the immune response that has been identified
   in the airways of patients with mixed granulocytic asthma. Abundant
   neutrophils infiltrate the lung, and mixed neutrophilic/eosinophilic
   inflammation was greatly increased in our severe asthma mouse model,
   which is consistent with the literature ([215]5, [216]7, [217]21,
   [218]22). Our data indicate that ETosis is involved directly and
   indirectly in regulating neutrophil and eosinophil accumulation, which
   have a pathophysiological role in promoting excessive inflammation and
   tissue damage. We identified a marked increase in lung
   eosinophils/neutrophils and extracellular dsDNA in animals exposed to
   allergens with c-di-GMP. We also discovered that short non-coding
   miR-155 acts as an early regulator of the release of host dsDNA and
   ETosis in severe asthma. The severity of the disease is reduced in
   miR-155-deficient mice compared to their WT counterparts. The
   neutrophil influx in BAL was decreased in miR-155
   inhibitor/exosome-treated severe asthma mice. There was a significantly
   reduced BALF dsDNA in PAD4 deficient; however, the infiltration of lung
   immune cells was not decreased. DNase I instillation decreased the
   intensity of airway inflammation but did not impact miR-155-derived
   neutrophilia. Together, these findings indicate that miR-155 promotes
   the release of host dsDNA that contributes to airway damage and mixed
   granulocytic inflammation.

   Outside of the original observation of ETs antimicrobial activity,
   recent studies have described an assortment of ET-associated
   immunologic processes and pathologies. ETs contain histones and
   granular proteins such as NE and MPO. Soluble ET components prime other
   immune cells to induce sterile inflammation or immune cell recruitment.
   The released extracellular DNA acts as a danger signal and triggers
   cytokine production for additional immune response ([219]24). A novel
   perspective has emerged in light of recent reports, which showed that
   ETs are generated by both eosinophils and neutrophils in human atopic
   asthmatic airways ([220]25, [221]26). Patients with mixed neutrophils
   and eosinophils in their lungs have more severe airflow obstruction and
   a higher frequency of exacerbations and are resistant to treatment with
   corticosteroids ([222]15). Both neutrophils and eosinophils undergo
   ETosis to promote inflammation and tissue damage in response to many
   pathogens and pro-inflammatory stimuli. Therefore, key aspects of ETs
   in the immune response are sufficient to recapitulate many features of
   asthma pathophysiology. In this study, we focused on the role of ETs in
   regulating asthmatic airway inflammation directly or indirectly by
   modulating other immune cell recruitments. Our severe asthma mouse
   results unequivocally indicate that there is a substantial level of
   biologically active ETs present in the airways of asthmatics with
   neutrophil/eosinophil-enriched inflammatory response. It is important
   to note that increased ETosis associated with mild asthma as well,
   suggesting that an imbalance of ET formation in the lung may also
   contribute to enhanced Th2 response in asthma pathogenesis.

   MicroRNAs are dynamic post-transcriptional regulators of gene networks.
   Profiling studies of miRNAs in human biopsy specimens and mouse models
   of asthma showed differential expression in ~10%–20% of miRNAs
   ([223]56). MiR-155 is one of the first identified and most studied
   miRNAs up to date. It is thought to play a role in the functioning of
   the human immune system ([224]57, [225]58). MiR-155 is essential for
   Th2-mediated eosinophilic inflammation through the transcription factor
   PU.1 ([226]59). Additionally, ILC2 and IL-33 signaling are regulated by
   miR-155 in Th2 eosinophilic allergic airway inflammation ([227]42).
   Thus, miR-155 is of considerable interest for understanding immune
   regulation in allergic asthma conditions. Herein, we showed that host
   dsDNA release from ETosis in mixed neutrophilic and eosinophilic asthma
   phenotypes is dependent on miR-155 expression levels. Many neutrophils
   infiltrated into the lungs, and increased NE productions were
   significantly reduced in the miR-155KO mice compared to the WT mice. We
   made the novel observation that the miR-155KO mice were unable to
   induce both neutrophil and eosinophil influx in the lung in response to
   severe asthma conditions, which suggests another role for miR-155 in
   neutrophilic inflammation in severe asthma. MiR-155-deficient mice also
   showed decreased PAD4 expression levels as compared to the WT mice,
   which is consistent with the results of in vitro study ([228]60). Our
   data confirm previous in vitro data indicating that after stimulation,
   activation of PAD4 occurs but is regulated by miR-155 in exaggerated
   ETs generation. These observations predict that miR-155 functions as a
   positive regulator of ETosis via regulation of PAD4 gene expression. We
   first asked if myeloid-specific miR-155 deficiency is sufficient to
   recapitulate the phenotype observed in global miR-155 knockout mice, as
   well as to gauge the effectiveness of inhibitor delivery to rescue the
   phenotype. However, we show that myeloid-specific deletion of miR-155
   in miR-155^fl/flLysMCre mice leads to a reduction of BALF dsDNA without
   affecting macrophage, neutrophil, or eosinophil numbers in a model of
   severe asthma. This might explain the myeloid lineage in
   miR-155^fl/flLysMCre mice with residual miR-155 in the lungs since in a
   model of severe asthma, inflammatory cell influx is completely blunted
   in whole-body miR-155KO mice. Future studies should identify the role
   of myeloid miR-155 and their associated phenotypes in neutrophilic
   severe asthma in the lung.

   Various clinical trials using miRNA mimetics or anti-miRNAs as
   therapeutic targets are currently underway and show promising results
   ([229]61). Our therapeutic approach involving a miR-155 inhibitor has
   informed the concept to deliver small RNA molecules using serum-derived
   exosomes into the lung in vivo. An exosome-based drug delivery system
   has recently attracted increasing attention ([230]36). Notably,
   exosome-mediated drug delivery has the following advantages in
   comparison with other approaches, such as nanoparticles, liposomes, and
   viruses. Nanoparticles and liposomes delivered into the lung via the
   i.t. route usually forms aggregate ([231]35). Exosomes are potentially
   less toxic and less immunogenic compared with exogenous delivery
   vehicles ([232]62). They also have been shown to be a mode of transport
   across the blood–brain barrier ([233]63). However, the delivery of
   exosome-based therapeutics to the lungs via i.n. or i.t. remains
   unexplored. Our results indicate that the exosome-loaded miR-155
   inhibitor is successfully delivered via i.n. instillation and exerted
   functions in lungs in vivo. Nonetheless, the correct timing of miR-155
   inhibition is critical to be functional in the lungs while limiting
   inflammatory cell influx and tissue damages. A partial targeting
   miR-155 after severe asthma onset did not produce salutary effects on
   reducing BALF dsDNA and lung injury.

   In the model tested herein, we used DNase I and PAD4 inhibitors to
   assess the functional importance of ETs in severe allergic asthma. None
   of these approaches are fully specified for suppression of the
   heightened Th2/Th17 response. It is important to note that the
   biological activity of the inhibitors may not be optimal in the lung
   tissue and could be influenced by factors that are inherent to the
   experimental models used ([234]64). Treatment with DNase I and PAD4
   inhibitor reduced dsDNA release in BALF; however, it had the same
   global outcome on the development of severe asthma compared to vehicle
   severe asthma mice, suggesting that ETs could mediate the onset of
   severe asthmatic airway inflammation in the models tested here.

   Single-cell profiling of lung immune cells and identification of
   molecular pathways in asthma pathogenesis are critical fields of
   interest in asthma research. This also provides a vital information
   resource for the immune cellular network in the lung during an asthma
   exacerbation. Here, our results provide data-driven different
   transcriptional profiles present in the immune cells of the lung at the
   single-cell level that contribute to the pathogenesis of asthma and
   decode the network by using gene set enrichment analysis. In our
   established severe asthma mouse model, we identified 27 major cell
   clusters by using SingleR. While numerous immune cells are present in
   the lungs of severe asthma mice, eosinophils, T lymphocytes,
   neutrophils, and macrophages are the most abundant cells, whose levels
   are consistent across different phenotypes of asthma. The profile of
   the identified cell clusters in miR-155-deficient and WT mice in the
   presence of c-di-GMP with DRA challenge was not altered, although
   miR-155KO suppressed c-di-GMP/DRA-induced increased levels of
   eosinophils and neutrophils. Detailed analysis revealed distinct
   transcriptional modulation in heightened Th2/Th17 response when
   compared in mice lacking miR-155. Inflammatory factors and
   innate/adaptive immune responses are essential in driving the
   development and exacerbation of asthma. We were able to find the
   expression of multiple key factors of persistent inflammation, adaptive
   immune response, and ETosis pathway in our experimental model. These
   factors critically regulate exaggerated inflammatory responses,
   inducing asthma exacerbation ([235]10). The pathway analysis supports
   that miR-155 influences ETosis, at least in part, through significant
   modulation of neutrophil, eosinophil, and macrophage responses.
   ET-derived DNA and granule proteins from these myeloid linage cells are
   key components in driving the pathogenesis of asthma exacerbation in
   this model. It has become clear that uncontrolled ET formation or their
   insufficient removal can have serious consequences; however, ET
   formation pathway is still under investigation.

   There are several limitations of this study, which might serve as
   directions for future studies. Excessive release of a serine protease
   NE can damage surrounding tissues and contribute to lung dysfunction
   ([236]65). Neutrophils are the dominant cellular source of NE, but it
   is also produced by macrophages ([237]66–[238]68). Our group has
   reported that a majority of the resident cells in the human BALF from
   our SBP-AG protocol were CD163-positive AMs ([239]30). In the
   allergen-challenged experimental site, the newly recruited eosinophils
   appeared to have co-existed with recruited monocyte-derived AMs (but
   not neutrophils) ([240]69). We reported that the newly recruited
   macrophage population in the airway is a very different phenotype from
   these resident macrophages. Remarkably, elevated NE was observed in
   BALF obtained from human subjects with asthma after SBP-AG protocol
   without any neutrophil influx. This phenomenon was speculated to occur
   due to increased innate immune cells (i.e., monocyte-derived AMs) in a
   mild intermittent allergic asthma model. Further work is required to
   address the precise role of macrophage-derived NE in allergic lung
   inflammation. Second, our experiments using mice treated with PAD4
   inhibitor and DNase I did not fully reveal miR-155 regulation pathways.
   Independent evaluation of miR-155 regulation pathways including the
   PAD4 axis in the global outcome of the development of severe asthma is
   further needed. Next, although we found that miR-155^fl/flLysMCre mice
   had reduction of BALF dsDNA, the role of myeloid miR-155 and their
   associated phenotypes in neutrophilic severe asthma in the lung needs
   to be further studied. Besides, our next study should address how
   miR-155 modulates cellular function in specific cell subsets in both
   innate and adaptive immune responses by scRNA-seq in the future.

   In summary, ETosis and associated host dsDNA are cardinal features of
   mixed granulocytic inflammation in severe asthma. Our data indicate
   that short non-coding miR-155 contributes to the release of
   extracellular dsDNA, which exacerbates many features of allergic
   inflammation. We developed a novel therapeutic tool using serum-derived
   exosomes as oligonucleotide delivery vehicles, leading to the control
   of the release of ETs in vivo. Our single-cell transcriptional analysis
   of mouse lung immune cells provides an insightful framework for
   understanding the distinct expression patterns of ET-related genes at
   the individual cell level. These findings provide pre-clinical evidence
   that targeting host dsDNA release represents an attractive therapeutic
   approach for mitigating the inflammatory asthma features for mixed
   granulocytic severe asthma patients.

Data availability statement

   The datasets presented in this study can be found in online
   repositories. The name of the repository and accession number can be
   found below: NCBI Gene Expression Omnibus; [241]GSE203656.

Ethics statement

   The studies involving human participants were reviewed and approved by
   University of Illinois Center for Clinical and Translational Science.
   The patients/participants provided their written informed consent to
   participate in this study.

Author contributions

   JK and SC performed the experiments, analyzed data, and drafted the
   manuscript. MK, HL, JE, and TL assisted in the mouse and cellular
   experiments. PS, MP, and PY obtained and analyzed the scRNA-seq data.
   NP, GP, and JC served as consultants for this project. GP and JC
   supported the SBP‐AG protocol and assisted with data analyses. SC
   supervised the project. All authors contributed to the writing of the
   final manuscript.

Acknowledgments