Airway Elastin is increased in severe asthma and relates to proximal wall area: histological and computed tomography findings from the U‐BIOPRED severe asthma study

Airway remodelling, which may include goblet cell hyperplasia / hypertrophy, changes in epithelial integrity, accumulation of extracellular matrix components, smooth muscle hypertrophy and thickening of the lamina reticularis, is a feature of severe asthma and contributes to the clinical phenotype.

epithelial area and total collagen were not different between the four groups. Due to small numbers of suitable CT scans, it was not feasible to compare airway morphometry between the asthma groups.

| INTRODUC TI ON
Severe asthma has a heterogeneous clinical phenotype and pathology. This pathology includes airway remodelling, features of which may be goblet cell hyperplasia/hypertrophy, changes in epithelial integrity, accumulation of extracellular matrix components, smooth muscle hypertrophy and thickening of the lamina reticularis. [1][2][3][4] In this current study, we focus on the evaluation of goblet cell hyperplasia / hypertrophy and the presence of the extracellular matrix components, collagen and elastin, and how these relate to central airway wall and lumen measurements by computed tomography (CT). This study was undertaken as data regarding changes in these parameters, in relation to disease, are conflicting.
Goblet cell hyperplasia has been observed in the lungs of patients who have died from asthma 5,6 and is also seen in bronchial biopsy tissue in milder asthma, 7 although not consistently so. 8 To date, there are no reported studies in living severe asthmatics. Collagens 1, 3 and 5 form part of the extracellular matrix in the bronchial wall, contributing to the structural support of the airway. 1,2 Whilst there is a report of increased submucosal collagen 9 in bronchial biopsies from mild and moderate asthma compared with healthy controls, other studies have not consistently reproduced this finding. [10][11][12][13] In severe asthmatics, Benayoun et al 12 and Chakir et al 13 have reported an increase in submucosal collagen compared with both milder asthmatics and healthy controls but again this finding in biopsy tissue is inconsistent. 11,14 Furthermore, post-mortem studies have not reported any differences in total collagen content in the lungs of cases of fatal asthma as compared to that in non-asthmatic deaths. 15,16 Elastin, which contributes to airway patency and elastic recoil, has been reported to be increased in the central airways 17 and in longitudinal bundles, 15 but decreased in the distal airways 18 in fatal asthma compared with non-asthma controls, whereas no disease-related differences were observed by Godfrey. 19 An increase in the proportion of elastin within the airway smooth muscle has been observed in cases of fatal asthma compared with non-fatal asthma. 20 In bronchial biopsies from mild and moderate asthma, no difference in the proportion of elastic fibres, as compared to healthy controls, is reported and this is not affected by corticosteroid treatment. Due to these conflicting findings and the need to better understand the remodelling pathology in severe asthma and its relationship to the clinical assessment by CT, the U-BIOPRED study of severe asthma has explored the relationship between CT central wall parameters and proximal airway wall biopsy changes in remodelling.

| Study design
As previously described in detail, 29 the U-BIOPRED (Unbiased Biomarkers for the Predictions of Respiratory Disease Outcomes) multi-centre pan-European severe asthma study included three adult asthma groups: severe non-smokers (SAn), severe current/ex-smokers (SAs/ex) and those with mild-moderate disease Conclusion: These findings identify a link between extent of elastin deposition and airway wall thickening in severe asthma.

| Bronchoscopy procedure, biopsy collection and analysis
The bronchoscopy cross-sectional sub-study, for the collection of airway samples, has been described in detail by Wilson et al 30  The observers (HMP and JAW) undertaking the image analysis were blinded to the participant grouping and ID.

| Computed tomography
The clinical characterization of the participants also included CT.
Volumetric whole lung scans were obtained at full inspiration (total lung capacity) and full expiration (residual volume) using a standardized protocol for each scanner manufacturer and model. All participants were coached in the breath-holding techniques and practiced breath-holding immediately prior to scanning. Participants were scanned 10-60 minutes after receiving 400 μg salbutamol. Postprocessing was performed using the VIDA Apollo software (VIDA Diagnostics, Iowa, USA) as described previously. 32,33 Quantitative CT parameters included large airway morphometry (measured in mm 2 ): lumen area (LA), wall area (WA) and percentage wall area

| Statistical analysis
Data were initially analysed by ANOVA to test. for differences between groups and then where relevant, either non-parametric or parametric analysis applied to evaluate the significance of group differences using SPSS (version 19). Spearman's rank test was applied to test for pair-wise correlations. We also included our previously published data 30 for lamina reticularis thickness and airway smooth muscle area fraction when testing for these relationships.

Severe asthma nonsmoker (SAn)
Severe asthma current or ex-smoker (SAs/ex)

| RE SULTS
The demographics for the participants included in this study are summarized in Table 1. The participants in both severe asthma groups were older and had a higher body mass index (BMI) and lower forced expiratory volume in one second (FEV1), forced expiratory flow (FEF25-75) than both the MMA and healthy controls. Peak expiratory flow (PEF) and specific airway conductance (sGaw) were also lower in both severe groups compared with healthy controls, and in the SAs/ex, but not the SAn, these measures were both lower than the MMA. The MMA had a lower FEV 1 , FEF25-75, PEF and SGAW than the HC.

| Biopsy remodelling features
Representative images of Movat's pentachrome staining are shown in Figure 1. This dye-based technique enables the identification of mucin, elastic and collagen in one section. Summary data are shown in  Figure 2a). In most cases, these elastic fibres had a mixed appearance (wispy and lamella). However, a lamellar elastic appearance was more frequent in the severe asthmatics than the HC, where a wispy phenotype was more predominant (Figure 2b).

| Computed tomography
For the bronchoscopy cohort included in this study, suitable CT images for analysis were only available for 18 expiratory scans and 20 inspiratory scans the majority of these being for the SAn group (n = 12) (Table 3). Therefore, it was not reasonable to make betweengroups comparisons of these data. CT scans had to be excluded for the following reasons: i) deviations in the CT acquisition protocol, ii) technical errors in data capture or transfer, iii) error in CT procedure identified by lung density being greater in the inspiratory compared to expiratory scans or iv) CT and body plethysmography lung volumes, when compared, being discrepant by more than 3 SD from the mean difference.

| Relationship between biopsy remodelling, CT measures and lung function data
Due to the low number of participants with CT data, we tested for relationships between biopsy remodelling data and all participants with CT data (n = 17-18 asthma, n = 1-2 HC). We observed a positive relationship between the percentage elastin in the submucosa and the per cent wall area (expiratory) (Figure 2c), which also related to We also did not observe any relationship between remodelling features and measures of lung function.

| DISCUSS ION
In this study, which is the largest to date to assess submucosal airway remodelling in severe asthma, we have observed more elastin in the airways of severe asthmatics, irrespective of smoking status, compared to healthy controls. The elastic fibres in the asthmatics were thickened and more lamellar in appearance than in the HC.
The amount of elastin in the airways and its appearance had a relationship to airway wall area, assessed by CT. There was no relationship between the amount of elastin and the demographic or clinical measures reported here.
The proportion of elastin we observed in our severe asthma groups is similar to that reported by Godfrey et al 19 in the airway wall in asthma deaths and Araujo 20 in airway smooth muscle. Neither of these studies report differences between fatal asthma and nonasthma controls. Increases in elastin have been reported in the central airways in fatal asthma compared with non-asthma controls. 15,17 These elastic fibres were observed to be similar in appearance to those observed in our study, being thickened and more lamellar in appearance, as has also been reported by Bousquet et al. 21 The differences in findings with respect to elastin content in re- Increased bronchoconstriction has, in mild asthmatics, been shown to drive a remodelling response, 38 and this could also be a possible mechanism initiating the remodelling changes we have observed. The presence of increased bronchoconstriction could also account for the increased lamellar appearance of the elastin, as reported by Mauad et al. 17 The positive relationship we observed between elastin and per cent wall area is novel. We also noted that this was associated with an altered appearance of the elastic fibres, with the higher percentage of elastin and greater wall area having elastic fibres with thickened lamellar appearance as shown in Figure 1. This suggests that as TA B L E 2 Summary data-biopsy remodelling.

Severe asthma non-smoker (SAn)
Severe asthma current or ex-smoker (SAs/ex) the amount of elastin in the airway increases and thickens so does the airway wall area.
We did not observe any relationship between airway remodelling measured by CT with mucin, collagen, ASM fraction or lamina reticularis assessed histologically. This result for lamina reticularis concurs with that of Aysola et al 25  In this study, we did not observe a relationship between these structural changes and measures of lung function. In mild asthma, an inverse correlation between PC 20 methacholine and airway elastin content in paraffin-embedded bronchial biopsies is observed 39 implying that increasing severity of airway hyper-reactivity is linked with increased elastin. Howarth et al 40 report that airway hyper-reactivity remains abnormal in severe asthma despite therapy. However, in the U-BIOPRED study PC 20 methacholine was not assessed, so we are unable to explore relationships between this aspect of abnormal physiology and airway tissue morphology.
Our data showing that total submucosal collagen did not vary with disease concur with the observations of others in mild, moderate and severe asthmatics [10][11][12]14 and in asthma deaths, 15,16 but does differ from the reports of Wilson et al, 9 Benayoun et al 12 and Chakir et al 13 In these latter studies, the methodological approaches employed differ from the present study, which may explain the The study of Kaminska et al 27 is the only previously reported study that has examined the relationship between CT measures of remodelling and submucosal collagen. Like our current study, they did not observe any relationships.
Our study, which is larger than previously reported studies assessing goblet cell hyperplasia in asthma, [5][6][7][8]  As previously reported, 30 we did not observe any disease-related differences in the thickness of the lamina reticularis or the proportion of airway smooth muscle as a fraction of the submucosa in these steroid-treated asthmatics as compared to healthy controls. We also did not observe, in this current study, any relationship with CT measures of remodelling and these morphological measures. This lack of relationship concurs previous studies. [25][26][27][28] Although this is one of the largest studies assessing submucosal remodelling in bronchial biopsies from living severe asthmatics that yielded some novel findings, this study does have some limitations.
The low number of suitable CT images that could be analysed meant we were unable to compare remodelling measured by CT across the different asthma groups or with HC. This could have also lead to a type II error.
In conclusion, this study has revealed increased elastin in severe asthma that relates to CT scan measures of airway wall thickening.
It also supports previous work showing that submucosal collagen deposition and goblet cell number do not differ with asthma disease severity and that this, and airway smooth muscle and laminar reticularis thickness do not relate to CT measures of remodelling.

ACK N OWLED G EM ENTS
The The authors also recognize that without the voluntary help from the asthmatic participants and the healthy control volunteers, this study would not have been possible and are indebted to their generosity.
Members of the U-BIOPRED study group are detailed in the supplementary acknowledgements. The data that support the findings of this study are available from the corresponding author upon reasonable request.

AUTH O R S' CO NTR I B UTI O N S
SJW lead the immunopathological analysis for the U-BIOPRED study including the analysis presented here. JAW and HMP undertook the staining and analysis described. SB and CB were responsible for the CT part of this study. KFC, DS, BD, BB, NK, TS and PHH were leads at the clinical centres undertaking the bronchoscopy aspects of this work, and SRS and PHH were the bronchoscopy study leads. DS and ARS were responsible for the clinical data analysis for the main study. KFC, RD and PJS designed and were the leads for the U-BIOPRED study.