Full Knock out mice (ADAM28^-/-) for ADAM28 BALB/c mice and Wild type (ADAM28^+/+) counterparts were previously described (20) and were enrolled in our studies according to “Principles of Laboratory Animal Care” (National Society for Medical Research). Experimental protocols described were approved by the animal ethical committee of the University of Liège (under the references #1597 and #2146). All experiments were performed on male and female mice. Animals were included in the asthma protocol aged 8 weeks and their weight was between 20 and 25 grams. Three protocols of asthma induction were used and referred to as a “short-term” (ST), “intermediate-term” (IT) and a “long-term” (LT) exposure protocol ( Figure 1 ). Eight weeks-old BALB/c mice were sensitized on days 1 and 8 (ST model) or on days 1 and 12 (IT and LT models) by intraperitoneal injection of 10 μg of ovalbumin (OVA Grade V; ref#A5503 Sigma-Aldrich, Schnelldorf, Germany) emulsified in 2 mg aluminium hydroxide (AlumInject; ref#77161 Perbio, Erembodegem, Belgium). Mice were subsequently divided into 4 groups for daily nebulization in standard Plexiglas boxes (30 × 20 × 15 cm): 2 groups of mice (ADAM28^-/- and ADAM28^+/+ mice) were only exposed to PBS (ref#17-516Q, Lonza, Verviers, Belgium) aerosol (control cohorts), and the 2 other groups (ADAM28^-/- mice and ADAM28^+/+) were subjected to ovalbumin 1% aerosol for 30 min (OVA Grade III; ref#A5378 Sigma-Aldrich, Schnelldorf, Germany). Aerosols were generated daily by ultrasonic nebulizer from days 22 to 26 for the ST exposure model and from days 22 to 56 for the IT model (aerosols were continued until day 90 for the LT model) 5 days/week (every odd week).

Mice were exposed to the 90 days model of exposition to OVA. At the day of sacrifice, they were anesthetized by intraperitoneal injection of pentobarbital (70 mg/kg, Nembutal^®, Sanofi Animal Health, Belgium) and ketamin (75 mg/kg, Ketalar^®, Pfizer, Brussels, Belgium). A surgical tracheotomy was performed and was followed by insertion of an 18-gauge polyethylene catheter into the trachea. Mice were ventilated by using a FlexiVent small animal ventilator (SCIREQ, Montréal, Québec, Canada). Baseline lung-function was evaluated during FlexiVent manoeuvres measuring lung compliance, tissue elastance, and tissue hysteresivity in basal conditions. Lung hysteresis, the area between the ascending and descending portions of the pressure-volume curve is a reflect of the property of the lung to dissipate the received energy (the energy applied to the lung in inspiration is not completely recovered in expiration). Hysteresivity (η) is the ratio of tissue damping (G) over tissue elastance (H). After basal evaluation, mice were exposed by inhalation to increasing doses of nebulized methacholine (1.25, 2.5, 5, 10, 20 mg/ml; ref#190231 ICN, Asse-Relegem, Belgium, PBS was used as a diluting solvent), and a dose-response (airway resistance) curve was obtained for each animal.

Right main bronchus was clamped and right lung was excised and immediately frozen in liquid nitrogen. The remaining left lung was inflated with 4% paraformaldehyde (PFA, Ref#8.18715.1000, Sigma-Aldrich, Schnelldorf, Germany) and resected. After a night in PFA bath, lungs were embedded in paraffin. Five-µm sections were stained with haematoxylin–eosin. A peribronchial inflammation score was determined, related to cell infiltration around the bronchi, by quantification of peribronchial inflammatory cells (eosinophils, lymphocytes, macrophages, etc.), as previously described (21). When no inflammatory cells were detectable around the bronchi a value of 0 was given. A value of 1 was given when there were occasionally inflammatory cells, a value of 2 when most bronchi were surrounded by a thin layer (one to five cells) of inflammatory cells, and a value of 3 when most bronchi were surrounded by a thick layer (more than five cells) of inflammatory cells. The score was measured on seven randomly selected bronchial sections per mouse and peribronchial inflammation scores are expressed as a mean value per animal.

Masson’s Trichrome staining was used to measure collagen deposition around the bronchi. As previously described, a collagen deposition score was applied (allowing to give a score from 0 to 3 to each observed bronchi) (18). A score of 0 was recorded when no collagen was stained by Trichrome Masson around the bronchi, a score of 1 for a thin layer of collagen, 2, for a cluster of collagens and 3, for a thick layer of collagen. Immunohistochemistry for alpha-smooth muscle actin (α-SMA) was performed using mouse primary antibody anti-α-SMA-FITC (ref#F3777 Sigma–Aldrich, Schnelldorf, Germany). Digitalized slides corresponding to 7 bronchi per mice were analysed. Smooth-muscle-cell layer was measured and reported to epithelial basement membrane perimeter. Glandular hyperplasia was evaluated by measuring the percentage of periodic acid-Schiff (PAS)-stained goblet cells per total epithelial cells (percentage of 300 randomly counted cells) (PAS ref#1.09034.1000, Merck, Hoeilaart, Belgium).

Fibroblasts were isolated from the lung of ADAM28^+/+ mice exposed or non-exposed to allergens in LT model of asthma. They were cultured according to already validated protocols (22). Immediately after sacrifice, chests of mice were disinfected with 70% ethanol and lungs were excised using sterile tools. They were cut into small pieces by cutting with razor blades. Tissue digestion was performed by collagenase (Collagenase from Clostridium histolyticum, ref#C9891 Sigma-Aldrich). Fibroblasts were cultured in hypoxia (incubator at 37°C, 5% of CO2 and 3% of O2). Confluence of the fibroblasts was followed using microscope. When they covered 60% of the plate medium was changed to remove debris and unattached cells. Between days 7 and 14, according to fibroblasts confluence, cells were harvested and divided in several plates for analyse. One of them was reserved for further RNA extraction. A CyQUANT cell proliferation assay was performed following manufacturer’s instructions. To evaluate fibroblasts growth rate, fluorescence was analysed 48 hours after first passage.

Mice were sacrificed, and a bronchoalveolar lavage was immediately performed using 4 x 1 ml PBS–EDTA 0.05 mM (ref#324503 Calbiochem, Darmstadt, Germany) as previously described (13, 14). Cells were recovered by gentle manual aspiration. After centrifugation (1,200 rpm for 10 min, at 4°C), supernatant was collected and frozen at –80°C. BAL was collected from ADAM28^+/+ and ADAM28^-/- mice exposed and non-exposed to OVA in a LT model of asthma. Cultured lung fibroblasts were isolated from the lungs of naïve ADAM28^+/+ mice. Enrichment of fibroblasts culture medium was performed by adding 10% of previously collected BAL (ADAM28^+/+ and ADAM28^-/- exposed to OVA or to PBS) to culture medium and cell proliferation was measured by a CyQUANT assay (Ref#C7026, Invitrogen, Merelbeke, Belgium).

The right lobe of the lung previously frozen was disrupted with a Mikro-Dismembrator (Sartorius Stedim Biotech, Vilvoorde, Belgium). Total lung RNA was extracted and purified using High Pure RNA Tissue Kit (Roche ref#12033674001, Mannheim, Germany) according to the manufacturer’s instructions. Total ADAM28, membrane-bound (Variant 1, Var1) and soluble form (Variant 4, Var4) expressions were assessed by semiquantitative RT-PCR and normalized to the 28s rRNA. As 28S rRNA is the product of the precursor 45S rRNA, some portions of 45S rRNA are also amplified by the primers chosen to target 28S.

The following primers targeting respectively 28S, total ADAM28, ADAM28 Var1 and ADAM28 Var4 were used: (F) 5’- GTTCACCCACTAATAGGGAACGTGA -3’, (R) 5’-GGATTCTGACTTAGAGGCGTTCAGT-3’, (F) 5’-CTACTTGAGCTGCAAGTGTCCATC-3’ and (R) 5’-CAGGTCTTGCTCACAGCATTTG-3’, (F) 5’-AGCCTCCACCTGATGTCCTAATCA-3’, (R) 5’-TAACCCACTTTCCAGGGGTCAGTT-3’, (F) 5’-AGCCTCCACCTGATGTCCTAATCA-3’, and (R) 5’- cctgagggttaagagcgctagtaa-3’. These latter primers amplifying ADAM28 variant 4 also amplify variant 5 that is a non-coding sequence.

For real-time RT-PCR, 100ng of cDNA were used, produced using RNA extracted, from whole lungs of OVA-exposed mice and control counterparts (PBS).

To assess and quantify inflammatory cells infiltration, lungs of ADAM28^+/+ and ADAM28^-/- were harvested and digested in collagenase C (1mg/ml; Gibco) prior to red blood cell lysis (Red Blood Cell Lysis Buffer, Sigma Aldrich, Saint-Louis, Missouri). Cells were stained with fluorochrome-conjugated surface antibodies during 30 minutes, fixed and permeabilized using Cytofix/Cytoperm (ref#554714 BD Biosciences, Erembodegem, Belgium) before intracellular antigen staining. Antibodies used for flow cytometry analysis were: CD5-BV421 (ref#53-7.3, BioLegend), CD4-PERCP CY 5.5 (ref#RM4-5, BD Biosciences), CD8a-BB515 (ref#53-6.7, BD Biosciences), IFN-γ-PE CY7 (ref#XMG 1.2, BD Biosciences), IL-4-APC (ref#11B11, eBioscience, Thermo Fisher Scientfic), CD45R-APC-eFluor780 (B220) (ref#RA3-6B2, eBioscience). Data were acquired on FACS CANTO II flow cytometer (BD Biosciences) and analyzed using BD FACSDiva software (BD Biosciences).

Results are expressed as columns and scatter plots with means ± SE. Statistic comparison between groups was performed, using GraphPad InStat (GraphPad; http://www.graphpad.com). The D’Agostino Pearson and Bartlett’s tests were performed to assess normality of the values and difference of variances. Non-normal law variables were compared by nonparametric Mann Whitney test, while normal-law variables were compared using the parametric t-test. When multiple comparisons were performed a one-way ANOVA was used. Values of P<0.05 were considered as significant.

In order to assess ADAM28 expression in lung in mouse models mimicking acute to chronic features of asthma, mRNAs corresponding to total ADAM28, soluble form (referred to as Variant 4: Var4) and membrane-bound form (or Variant 1: Var1) were measured by RT-PCR after ST, IT and LT allergen exposure ( Figures 2A–C ).

In the ST exposure model, total ADAM28 expression was lower in lungs of allergen-exposed mice as compared to PBS-exposed animals and ADAM28 expression correlated with the duration of exposure, i.e., animals challenged with OVA in the LT protocol had higher ADAM28 expression as compared to animals exposed to OVA in the ST protocol or in the IT protocol ( Figure 2A ). These results were confirmed by q-PCR (data not shown). Expression of the membrane-bound variant (var1) of ADAM28 was not significantly modulated at any stage of OVA exposure as compared to PBS ( Figure 2B ). On the other hand, expression of the soluble ADAM28 form (var4) was about twice higher when animals were exposed to OVA in the IT or LT protocol while not modulated in the ST protocol after OVA exposure ( Figure 2C ) and ADAM28 expression increased progressively according to the duration of allergen exposure, as it was observed for the total ADAM28.

In order to assess a potential role for ADAM28 in airway remodelling, ADAM28^+/+ or ADAM28^-/- mice were exposed to OVA in the LT protocol. OVA-exposed animals displayed increased levels of peribronchial inflammation as compared to PBS-exposed animals regardless of the ADAM28 expression ( Figure 3A ). Nevertheless, bronchial inflammation score was significantly lower in OVA-exposed ADAM28^-/- as compared to OVA-exposed ADAM28^+/+ mice. In PBS-exposed animals, levels of bronchial inflammation were low and similar in ADAM28^-/- and ADAM28^+/+ ( Figures 3A–E ). CD 4 + T lymphocytes were measured in lungs of PBS- or OVA-exposed animals by flow cytometry. ADAM28^-/- displayed a slightly higher recruitment of CD 4 + T cells in the lungs upon allergen exposure as compared to ADAM28^+/+ ( Figures 3F, G ). In the meantime, levels of CD 8 + T lymphocytes were lower in non-exposed ADAM28^-/- mice as compared to the corresponding non-exposed ADAM28^+/+( Figure 3G ). Nevertheless, Th₂ lymphocytes were significantly less recruited after OVA exposure in ADAM28^-/- animals as compared to wild-type counterparts ( Figures 3H, I ). B lymphocytes failed to increase in the lungs of ADAM28^-/- after OVA exposure in contrast with what was measured in ADAM28^+/+ mice ( Figures 3J, K ).

At baseline, i.e., in PBS-exposed mice, there were no significant differences between ADAM28^-/- and ADAM28^+/+ animals regarding lung function parameters measured by Flexivent^® (compliance (C), hysteresis and airway resistances) ( Figures 4A–C ). In the LT protocol, OVA-exposed ADAM28^+/+ mice displayed changes in lung function (significant decrease of compliance, and increase of hysteresis and airway resistances) when compared to PBS-exposed mice. In contrast, lung function measurements in the LT protocol (evaluated by compliance, hysteresis and airway resistances measurements) in ADAM28^-/- were not modified after OVA exposure and remained similar to the values obtained from sham-exposed mice.

Bronchial responsiveness following exposure to increasing doses of methacholine was measured using the FlexiVent^® system in ADAM28^+/+ or ADAM28^-/- mice exposed for 90 days to OVA or PBS. During methacholine challenge, OVA-exposed ADAM28^+/+ animals displayed a classical increase in airway resistances when compared with PBS exposed counterparts while, in sharp contrast, OVA-exposed ADAM28^-/- animals did not reach similar levels of airway resistance after being exposed to the highest doses of methacholine and displayed values similar to PBS-exposed animals ( Figure 4D ).

Airway remodelling was evaluated in the LT protocol after allergen (OVA) or PBS exposure by histology and immuno-histological analysis. Mucous hyperplasia was measured by Periodic Acid-Schiff (PAS) staining. Percentages of mucous cells in the airway epithelium were higher in OVA-exposed mice as compared to PBS-exposed counterparts ( Figures 5A–C ). Percentages of mucous cells after OVA exposure were significantly lower in ADAM28^-/- as compared to ADAM28^+/+ animals ( Figures 5A–C ). Airway smooth muscle hyperplasia was quantified by the measurement of alpha-smooth muscle actin (α-sma) in immunohistochemistry ( Figures 5D–F ). In OVA-exposed ADAM28^+/+ mice, the area occupied by α-sma was significantly increased in airway walls as compared to PBS counterparts while this area did not increase in ADAM28^-/- exposed to allergens ( Figures 5D–F ). ADAM28^+/+ mice exposed to allergens displayed a significantly increased collagen deposition in the airway walls as measured by Masson’s trichrome staining. In contrast, the extent of collagen deposition in the airway walls was significantly lower in ADAM28^-/- exposed to allergens ( Figures 5G–I ).

In order to study the potential role of ADAM28 in the modulation of airway remodelling, fibroblasts were isolated from lungs of OVA- or PBS-exposed animals (LT protocol) and cultured in order to evaluate their ability to produce ADAM28 mRNA and to proliferate. Proliferation rate of freshly isolated fibroblasts was evaluated after the first passage by measuring DNA levels in cell culture during 48h by Cyquant analysis. Fibroblast proliferation was significantly increased after OVA exposure, regardless of their extraction from ADAM28^+/+ or ADAM28^-/- mouse lungs ( Figure 6A ). However, the proliferation of fibroblasts extracted from lungs of OVA-exposed ADAM28^-/- animals was significantly lower as compared with OVA-exposed ADAM28^+/+ ( Figure 6A ). No significant difference was found regarding ADAM28 mRNA expression measured by RT-PCR in fibroblasts extracted from OVA- or PBS-exposed wild-type mice ( Figure 6B ).

Lung fibroblasts were isolated from naïve ADAM28^+/+ mice and cultured with a medium containing 10% BAL. Cell proliferation was evaluated by measuring DNA levels using CyQUANT during 48 hours. Fibroblast proliferation was significantly increased when cells were cultured with BAL from OVA-exposed ADAM28^+/+ (LT model) in contrast to what was observed when fibroblasts were incubated with BAL collected from OVA-exposed ADAM28^-/- ( Figure 6C ).