Abstract
Background Severe asthma is associated with multiple comorbidities, including gastro-oesophageal reflux disease (GORD), which can contribute to exacerbation frequency and poor quality of life. Since epithelial dysfunction is an important feature in asthma, we hypothesised that in severe asthma the bronchial epithelium is more susceptible to the effects of acid reflux.
Methods We developed an in vitro model of GORD using differentiated bronchial epithelial cells (BECs) from normal or severe asthmatic donors exposed to a combination of pepsin, acid pH and bile acids using a multiple challenge protocol (MCP-PAB). In addition, we analysed bronchial biopsies and undertook RNA sequencing of bronchial brushings from controls and severe asthmatics without or with GORD.
Results Exposure of BECs to the MCP-PAB caused structural disruption, increased permeability, interleukin (IL)-33 expression, inflammatory mediator release and changes in gene expression for multiple biological processes. Cultures from severe asthmatics were significantly more affected than those from healthy donors. Analysis of bronchial biopsies confirmed increased IL-33 expression in severe asthmatics with GORD. RNA sequencing of bronchial brushings from this group identified 15 of the top 37 dysregulated genes found in MCP-PAB treated BECs, including genes involved in oxidative stress responses.
Conclusions and clinical implication By affecting epithelial permeability, GORD may increase exposure of the airway submucosa to allergens and pathogens, resulting in increased risk of inflammation and exacerbations. These results suggest the need for research into alternative therapeutic management of GORD in severe asthma.
Abstract
Using a combination of in vitro and ex vivo approaches, this study identified reflux causing significant effects on the bronchial epithelial structure and function, which were greater in patients with severe asthma https://bit.ly/31XV9tq
Introduction
Severe asthma is associated with multiple comorbidities, including gastro-oesophageal reflux disease (GORD), which is particularly common and is associated with exacerbation frequency and poor quality of life [1]. Until recently, this association was explained by three mechanisms: vagal reflex [2], neuroinflammation [3] and microaspiration directly triggering airway inflammation [4–7]. While studies of reflux in animal models [4–6, 8] and cultures of bronchial epithelial cells [9, 10] have shown varying impact of gastro-oesophageal refluxate on mediators of inflammation and airway remodelling, direct in vivo evidence for these mechanisms in patients with asthma has been limited. We recently undertook an in-depth analysis of sputum proteomics in severe asthmatics and identified 11 proteins differentially abundant in patients with GORD, including elevated levels of antimicrobial proteins and reduced levels of proteins involved in systemic inflammatory responses and epithelial integrity [11], providing the first direct evidence that reflux is associated with changes in the microenvironment on the epithelial surface of the airways. Recognising that defective epithelial barrier function, dysregulated repair mechanisms and modified epithelial immune responses to pathogens and allergens are important features in asthma [12], we further hypothesised that the presence of GORD in severe asthma would significantly influence global epithelial gene expression. Applying unbiased topological data analysis of microarray data derived from bronchial brushings, we identified a subset of severe asthmatics with a clinical phenotype defined by obesity, presence of GORD and treatment with proton pump inhibitors (PPIs) [13], characterised by upregulated airway remodelling signalling and downregulated mechanisms of immune cell recruitment, possibly linked to both bile acid exposure and PPI treatment [13].
In the current study, we sought to elucidate further the underlying mechanisms of GORD-associated dysregulation of the airway epithelium in severe asthma using a combination of in vitro and ex vivo approaches. We developed an in vitro model of GORD in which fully differentiated air–liquid interface (ALI) cultures of primary bronchial epithelial cells were exposed to a multiple challenge protocol using pepsin, acid pH and bile acids (MCP-PAB). Consistent with our previous in vivo observations [13], we observed that ex vivo exposure of epithelial cells to refluxate results in significant structural and functional changes. We then extended our studies using bronchial biopsies and bronchial brushings from severe asthmatics with GORD and confirmed the effects on interleukin (IL)-33 and changes in expression of a selection of genes identified from the in vitro study.
Methods
Study participants and sample collection
Severe asthmatics (step 4/5 of British Thoracic Society/Scottish Intercollegiate Guidelines Network guidelines) and healthy control participants were recruited prospectively and assessed for GORD using 24-h pH/impedance studies. The severe asthmatics were further stratified into those with documented GORD but no PPI treatment; those with documented GORD and PPI treatment; and those without GORD. Epithelial cells were harvested by bronchoscopic brushings and processed into RNA for subsequent RNA sequencing (RNA-seq) analysis or used in primary bronchial epithelial cell culture [14]. In addition, bronchial biopsies were taken and fixed in paraffin for immunohistochemical analyses.
The study was approved by the South-Central Hampshire A research ethics committee (reference numbers: 13/SC/0182 and 14/WM/1226) and all participants gave their informed consent.
Analysis of the ex vivo effect of MCP-PAB
Initial dose and time-course studies were conducted with 16HBE cells exposed to MCP-PAB (pepsin and chenodeoxycholic acid (CDC) at acidic pH); optimised conditions were subsequently confirmed to be appropriate for fully differentiated ALI cultures (see supplementary material and supplementary figure E1 for full details). MCP-PAB conditions were applied to the apical epithelial surface for 30 min before washing twice. After 4 h recovery, epithelial permeability was measured using transepithelial electrical resistance (TEER) and fluorescein isothiocyanate (FITC) dextran 4 kDa [14]. Apical supernatants were collected for cytokine measurements. Cells were then lysed with TRIzol lysis reagent (Life Technologies, Paisley, UK) and frozen at −80°C until analysis or fixed for immunofluorescence staining or electron microscopy analysis.
Immunostaining and electron microscopy
ALI cultures were fixed with 4% paraformaldehyde and processed for immunofluorescence staining, as described previously [14]. In addition, the cultures were processed for transmission electron microscopy (TEM) and analysed for epithelial integrity.
Bronchial biopsies were processed as described previously [15], embedded into paraffin; sections were stained using goat polyclonal anti-human IL-33 (R&D Systems, Abingdon, UK). Results were expressed as positive nuclei per total number of epithelial cells.
Cytokine analyses
IL-8 concentrations in conditioned media were measured using an IL-8 DuoSet ELISA (R&D, Abingdon, UK) while IL-6, tumour necrosis factor (TNF)-α and IL-1α were measured using VPlex (MSD, MD, USA).
Analysis of gene expression in epithelial brushings and differentiated cells
Total RNA was extracted from epithelial brushings or cultured cell lysates using miRNeasy Mini Kit and RNase-Free DNase Set (Quiagen, Manchester, UK). cDNA libraries were prepared using NEBNext Ultra (nonstranded) mRNA library prep kit with polyA pulldown for mRNA enrichment. Paired-end 150 bp sequencing to a depth of 20 million reads (epithelial brushings) or 50 million reads (differentiated cells) was performed on an Illumina HiSeq2500 by Novogene (Cambridge, UK). FASTQ files were aligned to human genome build GRCh38 using STARv2.6.0; reads were counted with HTSeq; and differential gene expression analysis conducted with edgeR. Details are given in the supplementary material. Data are deposited in the Gene Expression Omnibus repository.
Statistical analyses
Paired t-tests were applied to transcriptomic data, while clinical and experimental data were analysed using Kruskal–Wallis, Mann–Whitney U-test or t-tests depending on data distribution; p<0.05 was considered significant. False discovery rate correction was applied to the transcriptomic data.
Results
MCP-PAB causes epithelial damage and alters barrier and secretory function
To analyse the impact of GORD on the airway epithelium of severe asthmatics, ALI cultures derived from bronchial brushings of participants with severe asthma, GORD and PPI treatment (n=8) and healthy controls (n=5) (table 1) were exposed for 30 min to MCP-PAB conditions consisting of 50 µg·mL−1 of pepsin and 250 µM CDC at pH 5.
Characteristics of participants
MCP-PAB-induced epithelial damage was characterised by TEM as enlargement of intercellular spaces and beginning of cell detachment (supplementary figure E2). In addition, MCP-PAB-exposed ALI cultures had markedly increased ionic and macromolecular permeability, as shown by decreased TEER (figure 1a) and increased FITC dextran 4 kDa passage, respectively (figure 1b), with a significantly greater impact of MCP-PAB on permeability of cultures from severely asthmatic donors compared with healthy controls. Analysis of epithelial tight junctions and adherens junctions in ALI cultures from severely asthmatic donors showed a marked disruption of the junctions in MCP-PAB-exposed cultures (supplementary figure E3).
Effects of multiple challenge protocol using pepsin, acid pH and bile acid on epithelial permeability. Bronchial epithelial air–liquid interface cultures from healthy controls (HC) (n=5) and severely asthmatic (SA) (n=8) donors were untreated or exposed to multiple challenge protocol using pepsin, acid pH and bile acid (MCP-PAB) for 30 mins, washed and allowed to recover for 4 h before a) ionic and b) macromolecular permeability were measured.
In addition to having a marked impact on epithelial structure, MCP-PAB caused an increase in the secretion of CXCL8, IL-1α, and TNF-α (figure 2). These results were supported by analysis of epithelial gene expression (supplementary figure E4).
Stimulation of epithelial cytokine release by multiple challenge protocol using pepsin, acid pH and bile acid. Bronchial epithelial air–liquid interface cultures from healthy controls (HC) (n=4) and severe asthmatic (SA) (n=6) donors were untreated or exposed to multiple challenge protocol using pepsin, acid pH and bile acid (MCP-PAB) for 30 mins, washed and allowed to recover for 4 h before a) CXCL8, b) interleukin (IL)-1α, c) tumour necrosis factor (TNF)-α and d) IL-6 protein release was measured in apical secretions.
Artificial refluxate upregulates unfolded protein responses, damage responses and epithelial remodelling mechanisms
To further analyse the mechanisms involved in MCP-PAB-induced epithelial dysregulation in ALI cultures, we analysed mRNA transcriptomes obtained by RNA-seq.
Comparison of gene expression in unstimulated ALI cultures showed 147 genes upregulated and 266 downregulated in cultures from severe asthmatics when compared with healthy controls (figure 3, supplementary table E1). Application of MCP-PAB resulted in a profound effect on gene expression, especially in ALI cultures from severely asthmatic donors, which had a significantly higher number of differentially expressed genes (DEGs) (n=599) compared to cultures from healthy donors (n=87 DEGs). Among the most prominent modulated genes were IL1RL1 (IL-1 receptor like 1, the receptor for IL-33), CHAC1 (cation transport regulator-like protein 1), involved in oxidative balance and unfolded protein response (UPR) and SERPINB9, a serine protease inhibitor.
Changes in epithelial gene expression caused by multiple challenge protocol using pepsin, acid pH and bile acid in vitro. Bronchial epithelial air–liquid interface (ALI) cultures from healthy controls (n=5) and severely asthmatic (n=6) donors were untreated or exposed to multiple challenge protocol using pepsin, acid pH and bile acid (MCP-PAB) for 30 mins, washed and allowed to recover for 4 h before harvesting for RNA sequencing. Heatmap of the top dysregulated epithelial genes from low (blue) to high (red) levels of expression.
Gene ontology analysis (AmiGO) of all MCP-PAB-induced DEGs identified a number of differentially controlled biological processes (p<0.05). Taking a cut-off of two-fold increase in gene expression, we found 57 processes upregulated in cultures from severe asthmatic donors and 25 in those from healthy donors. In order to identify the processes with the greatest impact, we undertook a further selection of gene expression with a cut-off of five-fold; this showed 16 processes upregulated in cultures from severe asthmatic donors and 11 in cultures from healthy controls (table 2).
Top upregulated biological processes in multiple challenge protocol exposed cultures when compared with control cultures (details provided if greater than five-fold enrichment, p<0.05)
The most significant enrichment due to MCP-PAB exposure was in the protein kinase R-like endoplasmic reticulum kinase (PERK)-mediated UPR; this was significant in cultures from both asthmatics and healthy donors, but was three times greater in healthy participants. Cultures from healthy individuals exposed to MCP-PAB were also enriched in other stress-response processes (table 2). In contrast, MCP-PAB-exposed epithelial cultures derived from severe asthmatic donors were enriched in epidermal growth factor receptor (EGFR) signalling, cell migration and vasculature development, suggesting upregulation of tissue repair and remodelling responses.
Having established MCP-PAB-induced epithelial damage, we next explored the impact of MCP-PAB on damage signalling. We analysed the damage-associated cytokine IL-33 and found that artificial refluxate caused increased nuclear IL-33 staining cultures from severe asthmatic donors (figure 4). We confirmed IL-33 expression in ALI cultures using quantitative PCR (healthy controls n=5; severely asthmatic patients n=6), and showed that MCP-PAB was associated with a 67% increase in IL-33 expression in severe asthma and a −11% change in healthy controls (p=0.01 for between-group comparison).
Regulation of epithelial expression of interleukin (IL)-33 by multiple challenge protocol using pepsin, acid pH and bile acid (MCP-PAB). Bronchial epithelial air–liquid interface (ALI) cultures from severely asthmatic donors were untreated (control) or exposed to MCP-PAB for 30 mins, washed and allowed to recover for 4 h before fixing and immunofluorescence staining. Images show IL-33 nuclear staining (green) and 4′,6-diamidino-2-phenylindole (DAPI) (red). Images are representative of experiments using ALI cultures from six donors. Scale bars=25 µm.
Comparison of in vitro findings with in vivo epithelial changes in severe asthma with GORD
To determine the relevance of our in vitro findings with epithelial changes in severe asthma with GORD, we first performed immunohistochemistry for IL-33 using bronchial biopsies from severely asthmatic (n=9) or healthy control subjects without GORD (n=4); the severe asthmatics were subgrouped as follows: 1) severe asthma with no GORD (SA-no GORD; n=5), and 2) severe asthma with documented GORD, but who had abstained from their regular PPI treatment for 2 weeks to avoid potential bias of systemic or local (through micro-inhalation) impact of PPI on epithelial gene expression (SA-GORD; n=4). As shown in figure 5, there was a significantly higher number of IL-33-positive nuclei in SA-no GORD compared to healthy controls, with a further significant increase in SA-GORD.
Epithelial interleukin (IL)-33 expression is increased severe asthma with gastro-oesophageal reflux disease (GORD). Typical patterns of immunohistochemical staining for IL-33 in bronchial biopsies from a) healthy control participants without GORD (HC); b) severe asthmatics with no documented GORD (SA-no GORD; n=5); and c) severe asthmatics with documented GORD who had abstained from their regular proton pump inhibitor (PPI) treatment for 2 weeks (SA-GORD; n=4). d) Quantitation of positive nuclei expressed as percentage of total epithelial cells.
Finally, we analysed mRNA transcriptomes of bronchial brushings from SA-GORD (n=6), SA-no GORD (n=4) and healthy control subjects (n=12). RNA-seq analysis identified that, of the top 37 genes whose expression was modified in ALI cultures in response to MCP-PAB, 15 were similarly modified ex vivo in patients with GORD (table 3). Of note, the expression of CHAC1, the top upregulated gene involved in the UPR process which was identified as the main mechanism induced by MCP-PAB in ALI cultures (supplementary table E1), was also increased in bronchial brushings obtained from SA-GORD when compared to SA-no GORD, confirming a similar impact of refluxate on epithelial responses to endoplasmic reticulum stress in vivo.
Differential epithelial expression of 37 genes in air–liquid interface (ALI) cultures and expression of these genes in bronchial brushings
Discussion
Using a combination of in vitro and ex vivo approaches, we have obtained compelling evidence in support of reflux having a significant impact on bronchial epithelial structure and function, with a profound effect on the epithelium of severe asthma patients. Application of MCP-PAB conditions to epithelial ALI cultures caused marked acute structural damage, including disruption of adherens and tight junctions, increased permeability and induction of stress responses, as shown by enrichment of the UPR genes and modulation of the alarmin IL-33 and its receptor IL1RL1. These in vitro findings were supported by observations in bronchial biopsies and by global gene expression analysis of epithelial brushings from severe asthmatics without or with GORD and healthy controls.
GORD is a chronic disorder caused by abnormal reflux of acid, pepsin and bile acids, defined as time of acid exposure >6% during 24-h monitoring of oesophageal pH [16]. Combined with impedance measurement, pH monitoring allows the detection of acidic (pH <4), weakly acidic (pH 4–7) and non-acid reflux (pH >7), the latter occasionally persisting despite treatment with PPIs [17]. Whether and to what extent gastric juice contents penetrate the lungs in subjects with GORD has been uncertain [18], although our own studies have provided evidence that clinical GORD is associated with changes in several biomarkers [11, 13]. When deciding on the composition of the ingredients in the MCP-PAB for in vitro testing, we took into account physiological concentrations in gastric secretions of acid [19], pepsin [20] and total bile acids [21] and previous reports of effects of pepsin [9] and bile acids [22] on epithelial cells. Based on dose-ranging experiments, we chose 50 µg·mL−1 of pepsin and 250 µM CDC at pH 5 for 30 min, because this resulted in measurable damage without causing extreme cytotoxicity. Thus, we observed enlargement of intercellular spaces, disruption of intercellular junctions and increased permeability, effects similar to observations in vivo in oesophageal and laryngeal epithelium exposed to chronic refluxate [23, 24]. This, coupled with previous studies showing that ALI cultures derived from severe asthmatic donors exhibit phenotypic features similar to those found in vivo [14, 25], led us to conclude that exposure of ALI cultures to MCP-PAB conditions is a reliable model to analyse the effect of reflux on the bronchial epithelium in severe asthma.
Our study revealed a marked reflux-induced increase in the nuclear expression of the alarmin IL-33, as well as upregulation of IL1RL1, the gene encoding the IL-33 receptor. IL-33 is a member of the IL-1 cytokine family localised in the nucleus of airway epithelial cells and its release can be triggered by damage caused, for example, by allergens or viruses [26, 27]. It is a known asthma susceptibility gene [28, 29] and plays a crucial role in type-2 innate immunity through activation of group 2 innate lymphoid cells to trigger production of IL-4, IL-5 and IL-13 [30]. Our findings of IL-33 upregulation in bronchial biopsies from severe asthma with GORD are in concordance with the observed upregulation of IL-33 nuclear expression in the oesophageal mucosa in patients with reflux oesophagitis [31] and symptoms of heartburn [32].
Refluxate-induced damage also included a response to oxidative stress through PERK-mediated UPR [33]. A recent study, using an oesophageal squamous epithelial cell line, identified bile acid-mediated activation of the PERK-mediated UPR [34]. Our study provides the first evidence of refluxate-triggered UPR activation in the airway epithelium. PERK is a type I endoplasmic reticulum (ER) transmembrane protein activated by misfolded proteins inside the ER. Its stimulation induces transcription of UPR-related genes, leading to autophagy, apoptosis and redox homeostasis [35]. UPR is considered a master regulator in inflammatory diseases and its role in asthma development has been suggested [35]. UPR can be activated by various asthma triggers, including allergens, cigarette smoke and viruses, and regulates oxidative stress in asthma [35]. UPR regulates NF-κB activity and NF-κB-mediated inflammation and can induce apoptosis in case of prolonged activation of ER stress. The relatively limited enrichment in the PERK-mediated UPR process that we observed in exposed cells from severe asthmatics may reflect an ineffective response to multiple types of damage and so explains the vulnerability of the bronchial epithelium in severe asthma to a range of environmental challenges.
We observed MCP-PAB-induced epithelial changes in cultures from both normal and severe asthma donors. However, MCP-PAB had a more pronounced impact on the structural and functional properties of the asthmatic epithelium, including a greater increase in epithelial permeability and a higher number of MCP-PAB-associated DEGs. While our ex vivo transcriptome analysis of bronchial brushings did not completely match our in vitro results from ALI cultures, this may be because the in vitro model represents a single acute stress event caused by exposure to MCP-PAB conditions, which do not fully reflect the complexity of gastric juice, whereas the clinical condition of GORD is characterised by repeated exposures to various components of gastric juice. In addition, severe asthma patients with GORD treated with PPIs may present a dysregulated aerodigestive microbiome, with a potential role in severe asthma [36, 37]. Nonetheless, we were able to identify 15 dysregulated genes in brushings from severe asthmatics with GORD among the 37 top dysregulated genes in MCP-PAB-exposed ALI cultures, with the extent of dysregulation being higher in cultures from severe asthmatics than in those from healthy controls. Among the DEGs were genes involved in oxidative stress responses (CHAC1, BACH2), cell adhesion (CEACAM1), cytoskeleton organisation (CDC42EP1) and cilia formation (GMNC), pointing to impact on epithelial structure regulation. Exposure of severe asthma cultures to refluxate also caused enrichment in EGFR- and cell migration-related processes that were not changed in cultures from healthy individuals. Our results suggest that refluxate might contribute to defective epithelial barrier function and EGFR-mediated remodelling, key features of asthma [14, 38].
In summary, our study has identified a direct impact of refluxate on the airway epithelial structure, barrier permeability and modulation of gene expression, including UPR, responses to oxidative stress and wound healing processes. This suggests a possible role for GORD in increasing exposure of the subepithelial airway mucosa to allergens and infectious pathogens, resulting in increased risk of inflammation and exacerbations, as well as a possible role in airway remodelling, a key feature of severe asthma. These results suggest the need for research into alternative therapeutic management of GORD in severe asthma.
Supplementary material
Supplementary Material
Please note: supplementary material is not edited by the Editorial Office, and is uploaded as it has been supplied by the author.
Supplementary methods and table. ERJ-01634-2021.Supplement
Supplementary figure E1. Development of the refluxate mix. Exposure of 16HBE cell cultures to hydrochloric acid (a, b), chenodeoxycholic acid (c, d) and/or artificial refluxate (e, f) induced cytotoxicity (a, c, e) and increase in epithelial ionic permeability (b, d, f). CDC: chenodeoxycholic acid (μM); MCP-PAB: multiple challenge protocol using pepsin, acid pH and bile acid. *: p<0.05; **: p<0.005; ***: p<0.0005. ERJ-01634-2021.Figure_E1
Supplementary figure E2. Enlargement of intercellular spaces by artificial refluxate. Bronchial epithelial ALI cultures from severe asthmatic donors were untreated (a, c) or exposed to multiple challenge protocol using pepsin, acid pH and bile acid (MCP-PAB) (b, d) for 30 mins, washed and allowed to recover for 4 hours before fixing. TEM photographs of intercellular spaces (white arrows) are representative of experiments using 4 donors. ERJ-01634-2021.Figure_E2
Supplementary figure E3. Disruption of epithelial tight junctions by artificial refluxate. Bronchial epithelial ALI cultures from severe asthmatic donors were untreated or exposed to multiple challenge protocol using pepsin, acid pH and bile acid (MCP-PAB) for 30 mins, washed and allowed to recover for 4 hours before fixing and immunofluorescence staining. a) ZO-1 (green) and 4′,6-diamidino-2-phenylindole (DAPI) (red). b) E-cadherin (green) and 4′,6-diamidino-2-phenylindole (DAPI) (red). Images are representative of experiments using 6 donors. Scale bar 25 μm. ERJ-01634-2021.Figure_E3
Supplementary figure E4. Stimulation of epithelial cytokine expression by artificial refluxate. Bronchial epithelial ALI cultures from healthy controls (HC) (N=5) and severe asthmatic (SA) (N=6) donors were untreated (CTL) or exposed to multiple challenge protocol using pepsin, acid pH and bile acid (MCP-PAB) for 30 mins, washed and allowed to recover for 4 hours before RNA extraction for CXCL8 (a) and IL-1ɑ (b) gene expression analysis. *: p<0.05 versus control. **: p<0.005 versus control. ERJ-01634-2021.Figure_E4
Shareable PDF
Supplementary Material
This one-page PDF can be shared freely online.
Shareable PDF ERJ-01634-2021.Shareable
Footnotes
Author contributions: Substantial contributions to the conception or design of the work: J-M. Perotin, G. Wheway, J.P.R. Schofield, P. Howarth, D.E. Davies and R. Djukanovic. Acquisition, analysis or interpretation of the data: J-M. Perotin, G. Wheway, K. Tariq, A. Azim, R.A. Ridley, J.A. Ward, C. Barber, D.E. Davies and R. Djukanovic. Drafting the manuscript or revising it critically for important intellectual content: J-M. Perotin, D.E. Davies and R. Djukanovic. Final approval of the manuscript version to be published: all authors
Conflict of interest: J-M. Perotin, G. Wheway, K. Tariq, A. Azim, R.A. Ridley, J.A. Ward, C. Barber and D.E. Davies have nothing to disclose. J.P.R. Schofield reports being director and shareholder in TopMD Precision Medicine Ltd. P. Howarth reports personal fees from GSK outside the submitted work. R. Djukanovic reports receiving fees for lectures at symposia organised by Novartis, AstraZeneca and TEVA, consultation for TEVA and Novartis as member of advisory boards, and participation in a scientific discussion about asthma organised by GlaxoSmithKline; and is a co-founder and current consultant, and has shares in Synairgen, a University of Southampton spin-out company.
Support statement: J-M. Perotin acknowledges the support of the European Respiratory Society (fellowship LTRF 2017) and of the Asthma, Allergy and Inflammation Research Charity. Funding information for this article has been deposited with the Crossref Funder Registry.
- Received June 9, 2021.
- Accepted December 10, 2021.
- Copyright ©The authors 2022. For reproduction rights and permissions contact permissions{at}ersnet.org
References
