Normal murine respiratory tract has its mucus concentrated in clouds based on the Muc5b mucin

Abstract

The organization of the normal airway mucus system differs in small experimental animals from that in humans and large mammals. To address normal murine airway mucociliary clearance, Alcian blue-stained mucus transport was measured ex vivo on tracheal tissues of naïve C57BL/6, Muc5b^−/−, Muc5ac^−/−, and EGFP-tagged Muc5b reporter mice. Close to the larynx with a few submucosal glands, the mucus appeared as thick bundles. More distally in the trachea and in large bronchi, Alcian blue-stained mucus was organized in cloud-like formations based on the Muc5b mucin. On tilted tissue, the mucus clouds moved upward toward the larynx with an average velocity of 12 µm/s compared with 20 µm/s for beads not associated with clouds. In Muc5ac^−/− mice, Muc5b formed mucus strands attached to the tissue surface, while in Muc5b^−/− mice, Muc5ac had a more variable appearance. The normal mouse lung mucus thus appears as discontinuous clouds, clearly different from the stagnant mucus layer in diseased lungs.

INTRODUCTION

The normal respiratory tract is exposed to numerous particles, bacteria, and viruses during every breath. Maintaining relatively clean lungs requires an efficient mucociliary clearance system, consisting of the airway surface liquid (ASL) comprising the periciliary liquid (PCL) with the cilia and the surface part of the ASL that we name “ASL-mucus” separate from more concentrated mucus. Impairment of mucociliary clearance triggers mucus accumulation in the airways, a common characteristic of chronic lung diseases such as cystic fibrosis (CF) and chronic obstructive pulmonary disease (COPD).

Airway mucus has two major high-molecular mass, highly glycosylated gel-forming mucins: MUC5AC and MUC5B. In newborn piglets, with numerous submucosal glands, the MUC5B mucin is secreted from the glands, and MUC5AC is secreted from the surface epithelium (30, 34). On the other hand, in adult humans, both MUC5AC and MUC5B are secreted from surface epithelial cells (22). Submucosal glands secrete MUC5B as long linear polymers packed in parallel to form thick mucus bundles around 25 µm in diameter with 1,000–5,000 parallel polymers at exit of the gland (9, 35). Inside the mucin granules, the MUC5B mucin is packed in a highly organized way due to low pH and high concentration of calcium ions. For proper unfolding at secretion, pH in the gland lumen is kept high by the chloride- and bicarbonate-rich fluid secreted from the cystic fibrosis transmembrane conductance regulator (CFTR)–expressing serosal cells located at the most distal part of the gland (7, 8, 31). This pH increase allows the secreted MUC5B mucin polymers to be pulled out into linear molecules. The MUC5AC mucin on the other hand is secreted from the surface secretory cells in the upper respiratory tract of newborn piglets, where it appears as wispy threads and sheets that extend onto the MUC5B bundles (8, 9). The normal respiratory tract, as studied in newborn piglets, is cleaned by the efficient sweeping of the epithelial surface by the MUC5B mucus bundles transported cephalically by the cilia. The movement of these bundles is suggested to be controlled by the MUC5AC mucin attachment/detachment to the goblet cells (7). In contrast, the MUC5B mucin is produced by the secretory cells of the surface epithelium in mouse airways and especially in the distal human airways (22).

The organization of the airway mucus system is different in small animals, even though they are often used for experimental purposes. In humans and pigs, submucosal gland presence extends down to around the 10th bronchial generation, whereas in mice there are only a few submucosal glands in the most proximal trachea. Little or no Muc5ac mucin is expressed in healthy mice and Muc5ac-deficient mice were healthy in contrast to the severely diseased Muc5b^−/− mice (28). Furthermore, old mice exhibited decreased mucociliary clearance and decreased Muc5b levels (14), indicating the importance of Muc5b for normal mucociliary clearance. Recent studies suggest that Muc5ac and Muc5b have distinct functions. Overexpression of Muc5b decreased mucociliary transport and enhanced lung fibrosis caused by bleomycin, whereas absence of Muc5b diminished this response (15). Overexpression of the epithelial sodium channel (ENaC, Scnn1b-Tg mice) caused mucus plugging, and deletion of Muc5b significantly reduced this plugging (20). Muc5ac seems to have a role in lung pathology as it is typically increased in disease (13). In most diseased mouse lung models, both Muc5b and Muc5ac are expressed in surface secretory cells (11). Muc5ac^−/− mice showed significantly decreased mucus plugging and did not develop airway hyperreactivity in models of allergic asthma (10). Muc5ac^−/− mice also showed attenuated lung inflammation and neutrophil trafficking during acute lung injury (18). However, overexpression of Muc5ac provided protection against viral infections, but it still did not cause airway mucus obstruction or impaired mucus clearance, suggesting that mucus hypersecretion alone is insufficient to trigger plugging (6, 10).

Mucus clearance, properties, and biology in normal mouse lungs have not been fully understood. The PCL bathing the cilia is filled with the mucin domains of the MUC1, MUC4, MUC16, and MUC20 transmembrane mucins that together generate a low-friction brush network (3). Floating on the cilia is an ASL-mucus of variable depth (34), containing the mucus. The current understanding has been dominated by the idea that there is a continuous mucus layer covering the tracheobronchial surface (2). This idea was questioned by Van As (32) in the 1970s, but his idea was not adopted. He suggested that there was no continuous mucus layer because this should be impossible to transport, as the distal surface area is much larger than the proximal. Further strengthening his argument was the observation that mucus was transported in the form of discontinuous flakes (33). To illustrate how the mucus is organized in the lungs of smaller animals, we have now studied this in mice using similar approaches as for the pig (8). We have studied the morphology and the functional role of the airway mucins in the healthy mouse. We used naïve C57BL/6 (WT), Muc5b^−/−, Muc5ac^−/−, and enhanced green fluorescent protein (EGFP)-tagged Muc5b reporter mice and studied mucus transport by ex vivo imaging. Our conclusion is that in normal healthy mice, mucus is concentrated in discontinuous clouds floating on a viscoelastic surface fluid (ASL-mucus).

MATERIALS AND METHODS

Animals.

C57BL/6 (WT), Muc5b^−/−, and Muc5ac^−/− mice aged between 8 and 12 wk were used (28). Animals were housed in individually ventilated cages in a specific pathogen-free facility where the exposure to dust was minimized by using selected bedding based on wood shavings, in controlled temperature (21 ± 2°C), relative humidity (55 ± 15%), and a 12:12-h light-dark cycle. Standard chow and water were available ad libitum. The animal care and experimental protocols were approved by the Local Animal Ethics committee, Gothenburg, Sweden.

Muc5b-EGFP reporter mouse.

The Muc5b-EGFP reporter mouse was generated by the introduction of EGFP into the third exon and before the first Cys of the VWD1 of the mouse Muc5b using CRISPR/Cas methodology (Cyagen Biosciences, Santa Clara, CA). The approach and targeted vector sequence is given in Supplemental Fig. S1 (see https://doi.org/10.5281/zenodo.3725593). Cas9 and gRNA (TCCACTTACCGCTCATGCTAGGG, reverse strand) were coinjected into fertilized eggs with donor vector for mice production. The pups were genotyped by PCR and sequenced.

Mouse mucus transport in live tracheal tissues.

Mice were euthanized with an overdose of isoflurane anesthesia. Both trachea and the connected lungs were isolated, placed in ice-cold, oxygenated (95% O₂-5% CO₂) Krebs buffer (composition in mM: 116 NaCl, 1.3 CaCl₂, 3.6 KCl, 1.4 KH₂PO₄, 23 NaHCO₃, and 1.2 MgSO₄), pH 7.4, and kept on ice during transportation (<30 min). The trachea and proximal parts of the primary bronchi were opened and mounted in a petri dish coated with Sylgard 184 Silicone Elastomer (Dow Corning, Midland, MI) using 27 gauge needles. The tissue was covered with 1 mL oxygenated Krebs-glucose buffer (supplemented with 10 mM glucose, 5.7 mM pyruvate, and 5.1 mM glutamate, pH 7.4). The buffer contained 0.4 mM Alcian blue 8G× and/or charcoal and/or one of the carboxylate-modified fluorescent FluoSpheres (40 nm, 580/605; 20 nm, 505/515; 1 µm, 505/515; 1 µm, 365/415) from (ThermoFisher Scientific, Waltham, MA) hereinafter called “beads.” For some experiments, the tissues were treated with either carbachol (10 or 100 µM) or bupivacaine (38396-39-3, Cayman Chemical, Ann Arbor, MI). The petri dish was placed tilted 20° with the larynx end of the trachea upward in a custom-made heating chamber gradually heated to 37°C and kept at a constant temperature during the course of the experiments. Tissues were monitored through a stereomicroscope with color or monochrome charge-coupled device cameras (DS-Fi2 or DS-QiMc, Nikon) and NIS elements software (Nikon Instruments, Tokyo, Japan). The mean speed of the moving Alcian blue-stained mucus in each time-lapse was calculated using NIS-Elements (Nikon). The data are represented as the mean of the measurements in each time-lapse (mean of 2–5 measurements/time-lapse).

Fixation of mouse tissues.

The trachea and lungs were isolated, washed with PBS, and directly fixed with methanolic Carnoy solution (60% anhydrous methanol, 30% chloroform, and 10% glacial acetic acid) for at least 24 h. Tissues were embedded in paraffin, and cut into 4-µm-thick sections, followed either by Alcian blue-periodic acid-Schiff (PAS) staining to assess mucus in the submucosal glands or by immunostainings.

Anti-MUC5B antiserum.

The MUC5B polyclonal antibody used in this study was produced in rabbits by Agrisera (Vännäs, Sweden) using recombinant MUC5B-D3 domain (GenBank accession no. NM_002458; residues 893-1304) as antigen. The protein was expressed with the mammalian episomal expression vector pCEP-His (25) in CHO-K1 cells and purified by nickel affinity chromatography using 1 mL HiTrap Chelating HP column (GE Healthcare) followed by size fractionation on Superdex 200 16/600 column (GE Healthcare) eluted in 20 mM Tris, 150 mM NaCl, pH 7.4. MUC5B-D3 was concentrated up to 1 mg/mL with Amicon Ultra-4 10K (Merck) and used for an initial immunization with 200 µg antigen/rabbit followed by four immunizations with 100 µg antigen/rabbit at 4, 8, 12, and 16 wk. The final bleed was performed at 18 wk. The specificity of the antibodies was verified by immunostaining (Supplemental Fig. S2; see https://doi.org/10.5281/zenodo.3725593).

Immunostaining.

Tissue sections were dewaxed and rehydrated, and microwave antigen retrieval was performed (10 mM citrate buffer pH 6). Samples were blocked for 20 min at room temperature with 5% BSA in PBS. Thereafter, sections were incubated overnight at 4°C with a mouse anti-Muc5ac antibody (clone 45M1; M5293; Sigma-Aldrich; 1:1,000) (19) or the rabbit anti-MUC5B-D3 antiserum (1:5,000) in PBS with 1% BSA. The slides were then incubated for 2 h at room temperature with goat anti-mouse IgG or donkey anti-rabbit IgG secondary antibodies conjugated to Alexa Fluor 488 or 555 (Thermo Fisher Scientific; A11029 and A31572, respectively; 1:2,000) in PBS with 1% BSA. For lectin staining, the slides were incubated with 50 µg/mL rhodamine-labeled Ulex europaeus agglutinin I (UEA1) (Thermo Fisher Scientific) for 2 h at room temperature. Nuclei were counterstained for 10 min using 1 μg/mL Hoechst 34580 (H21486, ThermoFisher Scientific, Waltham, MA), and the slides were mounted with prolong gold anti-fade mounting medium (P36930, ThermoFisher Scientific, Waltham, MA).

Elastase model.

As described previously (11), mucus accumulation and airway plugging were induced by administration of porcine pancreatic elastase intranasally to the mice at day 0 and day 7. Analysis of airways was performed 1 wk after the last administration of elastase.

Lectin staining of live mouse trachea analyzed by confocal microscopy.

The tissue was dissected, opened, and mounted in the petri dish. Krebs-glucose buffer containing rhodamine-UEA1 (ThermoFisher Scientific, Waltham, MA) and bupivacaine (to reduce mucus transport on the tissue) were added to the tissue. After approximately 30 min of incubation at ambient temperature, lectin was removed and fresh Krebs-glucose buffer was added. Tissue was imaged with a Plan-Apochromat ×20/1.0DIC water immersion objective and an upright LSM 700 Axio Examiner 2.1 confocal imaging system (Carl Zeiss, Oberkochen, Germany). Mucus morphology and attachment were analyzed using the Imaris software (version 7.6.5, Bitplane, Zurich, Switzerland).

Mass spectrometry and data analysis.

AB-stained surface liquid, including stained mucus, was collected in 1 mL of Krebs-glucose buffer after measuring the transport velocity, and stored at −80°C until analyzed (n = 6, C57BL/6 WT). The sample was solubilized overnight at 37°C in a GuHCl-containing buffer, according to the filter-aided sample preparation previously described (26). Proteins were sequentially digested with 10 ng/µL of LysC and trypsin (Promega), and peptides were cleaned through Stage Tips (24). Samples were analyzed by nanoLC-ESI-MS/MS in a Q-Exactive mass spectrometer coupled to an EASY-nLC 1000 (ThermoFisher Scientific, Waltham, MA) in a Reprosil-Pur C18-AQ 3 μm (Dr. Maisch, Ammerbuch-Entringen, Germany) column (150 mm × 0.075 mm ID; New Objective, Woburn, MA). Peptides were separated for 60 min on a gradient up to 50% B (A: 0.1% formic acid, B: 0.1% formic acid/80% acetonitrile). The spray voltage was set to 2 kV. Full mass spectra were acquired in profile mode from 400 to 2,000 m/z with resolution 70,000 and AGC target 1e6 (max IT 120 ms). The 12 most intense ions were collided with NCE 30, and tandem mass spectra were acquired with resolution 17,500 and AGC target 5e5 (max IT 64 ms).

Raw files were searched with MaxQuant version 1.5.7.4 (5) against the UniProt mouse reference proteome (55,408 proteins; 21,982 genes; version of November 5, 2019). The search parameters were set as follows: enzyme LysC + trypsin, maximum two missed cleavages, tolerance of 4.5 ppm for peptide main search, and 0.5 Da for MS/MS. Oxidation (M), acetyl (Protein N-term), and deamidation (NQ) were selected as variable modifications, and carbamidomethyl (C) as fixed. The minimum peptide length required was six amino acids, and the maximum peptide mass allowed was 4,600 Da. Peptide and protein false discovery rate were set to 0.01. Protein abundances were relatively estimated by label-free quantification (LFQ). Differential expression was assessed with the LFQ-Analyst web platform (29). Protein localization was predicted with SignalP 5.0 (1). The complete data set is available at the ProteomeXchange Consortium (http://proteomecentral.proteomexchange.org) via the PRIDE partner repository (23).

Statistical analysis.

Data are presented as box-plot showing the minimum, the 25th percentile, the median, the 75th percentile, and the maximum. For comparison of two independent groups, we used the Mann-Whitney U test. For comparison of multiple groups, Kruskal-Wallis test was used with Dunn’s post hoc test to correct for multiple comparisons. Significance was defined as P < 0.05.

RESULTS

The mucus of the healthy murine airways appears as clouds and not as a mucus blanket.

Mice were housed in individually ventilated cages where the exposure to dust was minimized by using wood shavings as bedding. To study the mucus in the healthy mouse lungs, we dissected the mouse trachea and primary bronchi and mounted the tissue with needles. The nonwashed tissue was covered with 1 mL Krebs-glucose containing the positively charged Alcian blue dye at such low concentration that the liquid was almost without blue color. Within minutes, the mucus turned blue due to the negatively charged glycans of the mucins. The mounted tissue was placed in a heated chamber tilted 20° with the larynx oriented on top. This allowed demonstration of the cilia-mediated movement of the mucus upward and against gravity. The tilted mounting allowed the excess liquid to drain and left only a minimal volume of buffer on the tissue, likely corresponding to the ASL. Mucus movement was recorded using video microscopy. The recordings were thus not performed submerged, avoiding any interference with the mucus transport and ensuring that all transport was cilia mediated.

The images showed a discontinuous, patchy Alcian blue-stained mucus that was distributed all over the tissue (Fig. 1A). The mucus had a cloud-like (Cumulus-type) morphological appearance. The trachea of a reporter mouse with EGFP-tagged Muc5b showed the presence of many Muc5b-expressing cells (green) at the carina level (Fig. 1B), but further proximal only few goblet cells were observed. Staining of Carnoy-fixed paraffin-embedded lung tissues with a new antiserum raised against the recombinant MUC5B VWD3 domain confirmed that Muc5b is the mucin present in the Alcian blue-stained cells (Fig. 1C) as also shown before (37). The lack of cross reactivity of this antiserum to Muc5ac was shown by the lack of staining of Muc5b^−/− tissue (Supplemental Fig. S2; see https://doi.org/10.5281/zenodo.3725593).

Fig. 1.

Mucus transport in healthy mouse trachea. A: mucus on the mounted live mouse trachea was stained with Alcian blue. B: tracheas of the enhanced green fluorescent protein (EGFP)-tagged Muc5b reporter mice were analyzed with confocal microscopy. C: paraffin-embedded lung sections were stained with anti-Muc5b antibody. D: mucus transport was monitored by video microscopy (Supplemental Video S1; see https://doi.org/10.5281/zenodo.3552625). Mucus moved upward toward the larynx. Images at different time points from the same time-lapse recording at 16 times the original speed. Arrows with the same color point to the same mucus cloud. E: mucus transport velocity of healthy trachea with bupivacaine treatment, n = 4–9 mice/group. Data are presented as a box-plot showing the minimum, the 25th percentile, the median, the 75th percentile, and the maximum. Kruskal-Wallis and Dunn’s multiple comparisons test. **P = 0.0014. ****P < 0.0001.

In an attempt to analyze the protein composition of the ASL, we tried to separate blue mucus from the surface washings. Unfortunately, this was unsuccessful, but we still analyzed the proteome of the combined material, as shown in Supplemental Table S1 (see https://doi.org/10.5281/zenodo.3725604). As the tissue had to be opened, there is a substantial amount of cellular material although some of these proteins are also abundant in bronchoalveolar lavage fluid, suggesting their presence in ASL (11). The Clara cell-secreted uteroglobin (Scgb1a1) was the most abundant protein (Supplemental Table S1). Staining the tissue with the anti-MUC5AC 45M1 antibody did not show any staining (11), and mass spectrometric analysis confirmed that Muc5b is the only mucin. This confirms our previous observation that young, healthy mouse airways do not express Muc5ac and demonstrates that only the Muc5b mucin is expressed in the airways under healthy conditions (11).

Video recordings showed that the mucus clouds were moving upward toward the larynx, illustrated by sequential images taken from the video (Fig. 1D and Supplemental Video S1; see https://doi.org/10.5281/zenodo.3552625). The red or black colored arrows represent the same cloud followed over time. Over time, the individual small clouds were gathered into bigger ones. Some clouds were extended into threads as they moved along the trachea, and sometimes threads folded back into a cloud-like shape (Fig. 1D). The mucus transport velocity was heterogeneous within the same mouse but also between different mice. Some mucus clouds moved fast, while others moved slower. Moreover, an individual mucus cloud could exhibit both fast and slow transport at different time points. That the individual mucus clouds on the same tissue moved with different velocities revealed that there was no mucus blanket that covered the airways. The magnification and staining method used will only reveal mucus formed by the collection of a high number of mucin polymers. As it was not possible to remove all stained mucus from the surface washings, the proteomic results cannot answer how much Muc5b mucin is present in the air-surface liquid.

The moving Alcian blue-stained mucus moved with a mean velocity of 12 µm/s or 0.7 mm/min (Fig. 1E). The mucus movement followed the direction of ciliary beating because the mucus moved upward toward larynx on the tilted platform. This was supported by the observation that the mucus transport velocity was significantly reduced by decreasing ciliary beating using the well-known cilia-beating inhibitor bupivacaine at three different concentrations (Fig. 1E) (21). The tilted airways ensured that transport of Alcian blue-stained mucus clouds was cilia-dependent.

Beads not in mucus clouds are transported faster than beads associated with the clouds.

Many studies use carboxylated fluorescent beads in vivo and ex vivo to study mucociliary clearance (14, 17, 27, 28). However, we have shown previously that the beads and Alcian blue-stained mucus move with different velocities and that the transport of free beads is much faster than the mucus transport in piglets (8). We interpreted this as the fastest flowing negatively charged beads was transported with the ASL mucus flow, whereas the positively charged Alcian blue is bound to concentrated mucus. Thus, we added 20-nm, 40-nm, or 1-µm fluorescent beads to Alcian blue-stained mucus in our mouse system (Fig. 2, A, B, and C, and Supplemental Video S2; see https://doi.org/10.5281/zenodo.3553490). The beads moved upward with highly variable speed, one group fast and another slower. The slower group moved with the Alcian blue-stain, indicating that they were trapped and collected on the mucus (Fig. 2D). The fast-moving beads were moving upward against gravity, independently of the Alcian blue stain, and almost twice as fast (20 µm/s, 1.2 mm/min). Similar behavior was observed for the different bead sizes 40 nm, 20 nm, and 1 µm, as they moved similarly within the slow- or fast-moving groups. It should be noted that some free beads were initially moving with the excess liquid downward by gravity on the tilted tissue (data not shown). This is similar to observations by Rogers et al. (27) when overloading the tissue with beads.

Fig. 2.

Transport velocity for Alcian blue-stained mucus versus fluorescent beads. Mounted mouse trachea stained with Alcian blue and 1-µm (A), 40-nm (B), or 20-nm fluorescent beads (C). Mucus with both beads and Alcian blue is marked with black arrows. Free moving beads are marked with red arrows. D: the velocity of mucus-trapped beads and free beads separately in comparison with the velocity of Alcian blue-stained mucus, n = 5–9 mice/group, Kruskal-Wallis and Dunn’s multiple comparisons test. *P = 0.0121. ****P < 0.0001 (Supplemental Video S2; see https://doi.org/10.5281/zenodo.3553490). E: effect of carbachol (Cch) at 10 or 100 µM on the transport of Alcian blue-stained mucus; n = 4–9 mice/group, F: effect of 100 µM Cch on 40-nm mucus-trapped beads and free beads; n = 4–7 mice/group. Data are presented as a box-plot showing the minimum, the 25th percentile, the median, the 75th percentile, and the maximum. Kruskal-Wallis and Dunn’s multiple comparisons test. *P = 0.0121. **P = 0.0034.

In our previous work on piglet tissues, we revealed that Alcian blue-stained bundles and bead transport were affected in opposite directions by cholinergic stimulation (7). The ASL-mucus transport measured as fluorescent bead transport was enhanced by acetylcholine or carbachol (Cch), whereas the Alcian blue-stained mucus bundles were retarded. To investigate whether the same situation applies to the mouse respiratory system, we treated live trachea with Cch and studied its effect on both Alcian-blue-stained mucus and 40-nm bead transport. Stimulation by Cch did not affect the mucus transport measured by Alcian blue-staining (Fig. 2E). Moreover, the transport velocity of the 40-nm fluorescent beads, whether for trapped or free beads, did not change (Fig. 2F). We conclude that ASL-mucus and mucus clouds are transported at different velocities in mice and neither is affected by cholinergic stimulation.

Murine submucosal glands formed Muc5b mucus bundles.

The mouse has only a few submucosal glands in the upper trachea concentrated at the anterior boundary between the cricoid cartilage of the larynx and the first tracheal cartilage ring and extending to the first few cartilage rings, depending on the mouse genetic background (34). In our WT mice, a few submucosal glands were confirmed topmost in the trachea (Fig. 3, A and B). Mucus was observed within these glands as indicated by Alcian blue-PAS staining of paraffin-embedded sections. Interestingly, thick mucus bundles were detected at the very proximal part of live trachea stained with Alcian blue (Fig. 3C), similar to the bundles observed in pigs. As shown in Fig. 3D and Supplemental Video S3 (see https://doi.org/10.5281/zenodo.3553492), charcoal added halfway down the trachea was collected by the bundles in the upper trachea. The gel-forming mucin present in the submucosal glands is Muc5b as shown by immunostaining with an anti-MUC5B antiserum and by analyzing live trachea from Muc5b-EGFP reporter mice (Fig. 3, E–H).

Fig. 3.

Submucosal glands at the most proximal mouse trachea produced Muc5b that formed bundle-like mucus. A: paraffin-embedded tracheal tissue stained with Alcian blue-periodic acid Schiff (AB-PAS) showing the location of submucosal glands. B: high magnification of the glands stained with AB-PAS. C: live mounted trachea stained with Alcian blue showing the appearance of the mucus. D: charcoal collected by the mucus (Supplemental Video S3; see https://doi.org/10.5281/zenodo.3553492). E: immunostaining of the trachea tissue with anti-Muc5b antibody. F: upper part of live trachea of enhanced green fluorescent protein (EGFP)-tagged Muc5b reporter mice analyzed by stereomicroscopy. G: live trachea of EGFP-tagged Muc5b reporter mice showing the gland filled with Muc5b (green). H: reconstruction image of Muc5b-producing submucosal gland.

Different morphological appearance of the Muc5ac and Muc5b mucins.

To study the morphological appearance of airway mucins Muc5b and Muc5ac, the mucus of Muc5b^−/− and Muc5ac^−/− mice was analyzed. Although the Muc5ac mucin is not expressed in normal mice kept in our vivarium, it was found in the lungs from Muc5b^−/− mice (28) and also demonstrated by Alcian blue staining in the Muc5b^−/− mouse trachea (Fig. 4B). As expected, Muc5b was present in the Muc5ac^−/− mice (Fig. 4A). This allowed us to study the morphology of these mucins separately, important as the structure of the polymeric forms of the Muc5b and Muc5ac mucins has been suggested to differ. As shown in WT mice, the ex vivo proximal trachea showed Alcian blue-stained Muc5b mucus bundles in Muc5ac^−/− mice (Fig. 4C). In the distal trachea, Muc5b in Muc5ac^−/− mice formed strand-like mucus with thicker blobs of collected mucus (Fig. 4E and Supplemental Video S4; see https://doi.org/10.5281/zenodo.3553496). The video shows that the mucus tends to form long strands, suggesting that the Muc5b mucin forms long linear molecules, just as in piglets (8). When Alcian blue-stained Muc5ac mucin was studied, this formed more round and clumplike mucus assemblies in the proximal trachea (Fig. 4D). In the distal trachea, these were smaller and had not gathered to larger clumps as in the proximal part (Fig. 4F and Supplemental Video S5; see https://doi.org/10.5281/zenodo.3553560). The maximum mucus transport velocity was calculated in the two animal models to be significantly lower in both the Muc5b^−/− and Muc5ac^−/− mice as compared with WT (Fig. 4G). The Muc5ac mucus had a tendency to move even slower than the Muc5b. Together the results suggest that Muc5b and Muc5ac have distinct molecular structure similar to what has been suggested from studies in pigs (7, 8).

Fig. 4.

Airway mucus and its transport velocity of Muc5b in Muc5ac^−/− and of Muc5ac in Muc5b^−/− mice. A: immunostaining of airway sections from Muc5ac^−/− mice with anti-Muc5b antibody showing Muc5b expression (green). B: airway epithelium stained with anti-Muc5ac antibody showing the expression of Muc5ac (red) in Muc5b^−/− mice. C: Alcian blue staining of Muc5b mucin in the live proximal trachea of Muc5ac^−/− mice. D: Alcian blue staining of Muc5ac in the live proximal trachea of Muc5b^−/−mice. E: Alcian blue staining of Muc5b mucin in the live distal trachea of Muc5ac^−/− mice. F: Alcian blue staining of Muc5ac in the live distal trachea of Muc5b^−/−mice. G: mean mucus transport velocities of Muc5b (Supplemental Video S4; see https://doi.org/10.5281/zenodo.3553496) and Muc5ac (Supplemental Video S5; see https://doi.org/10.5281/zenodo.3553560) in the trachea of Muc5b^−/− and Muc5ac^−/− mice. Data are presented as a box-plot showing the minimum, the 25th percentile, the median, the 75th percentile, and the maximum (n = 5–9 mice/group). Mann-Whitney U test was used. ****P < 0.0001.

Mucus is attached to the goblet cells in healthy airways.

Fluorescently labeled lectins are small and useful for staining extracellular mucins due to the high number of lectin epitopes. We have previously described that mucus plugs containing both Muc5ac and Muc5b mucin are formed in the airways of mice induced with elastase (11). Such plugs were effectively stained with the UEA1 lectin conjugated to rhodamine (Fig. 5A). When the same mucus plugs were stained with anti-Muc5b or anti-Muc5ac antibodies, both mucins were present in the plugs (Fig. 5B). As the UEA1 lectin stain was uniform, it indicated that both mucins were stained with this lectin. To further prove this, the Muc5b mucin was shown to be stained in the Muc5ac^−/− elastase-induced mice and the Muc5ac was in the Muc5b^−/− elastase-induced mice (Fig. 5, C and D).

Fig. 5.

Different morphological appearance of murine Muc5b and Muc5ac mucins. A: Carnoy-fixed lung paraffin section of a mucus plug in an elastase-exposed mouse stained with rhodamine-UEA1 lectin. B: same tissue stained with anti-MUC5B and anti-MUC5AC antibodies. Both also stained with Hoechst nuclear stain. C: UEA1-stained Carnoy-fixed lung paraffin sections of the Muc5b mucin in elastase-exposed Muc5ac^−/−. D: UEA1-stained Carnoy-fixed lung paraffin sections of the Muc5ac mucin in elastase-exposed Muc5b^−/−, showing that UEA1 stained both Muc5b and Muc5ac mucins. Both A and B also stained with Hoechst nuclear stain. E, G, and I: tracheal tissue was mounted ex vivo and the Muc5b mucin was stained with UEA1 in Muc5ac^−/− mice. F, H, and J: tracheal tissue was mounted ex vivo and the Muc5ac mucin was stained with rhodamine-UEA1 in Muc5b^−/− mice. Muc5b mucus appear as strands and Muc5ac as less organized mucus. I: Muc5b mucus strand on the surface of the epithelium with two attachment sites to surface goblet cells. J: Muc5ac mucus on the surface of the epithelium with one likely attachment point to the surface goblet cells. Images were acquired by confocal microscopy in z-stacks represented by x/y projections. Images were reconstructed using Imaris software.

UEA1 lectin staining can be performed in live trachea. Both the naïve Muc5ac^−/− and Muc5b^−/− mice were treated with the UEA1 lectin and nuclear stain and studied by confocal microscopy. Muc5b was shown as long strand-like mucus in the trachea in the Muc5ac^−/− mice while Muc5ac formed clump-like mucus in the trachea in the Muc5b^−/− mice (Fig. 5, E–H). When studied in the z-plane, a Muc5b mucus strand was found to extend from inside two goblet cells, suggesting that goblet cells acted as attachment points for the mucus to the tissue (Fig. 5I). The Muc5ac mucus was also found to extend from surface goblet cells, although this was less obvious (Fig. 5J). Together, this suggests that both the Muc5ac and Muc5b mucins could be involved in mucus attachment to goblet cells in the elastase-induced mouse model system. However, as the Muc5ac mucin is essentially absent in normal mice, one can speculate that the Muc5b mucin could retain the mucus to the goblet cell under normal conditions as an explanation to why the mucus clouds were moving slower than the ASL-mucus. Muc5ac might also contribute under diseased conditions.

DISCUSSION

Airway mucus provides one of the major defense mechanisms protecting the lungs from environmental exposure. Although the understanding of the importance of mucus clearance is growing today, and mucus clearance is the focus of diverse research, a full description of the structure and function of the mucus system in mice is not available. We used Alcian blue staining to visualize and track airway mucus, and to measure mucus transport on freshly excised mouse tracheas similar to previous studies in piglets (7, 8). We could conclude that the normal mouse trachea is not covered by a mucus layer or blanket, as has been assumed for a long time. Instead, patchy and unevenly distributed mucus clouds are transported upward toward the larynx by the beating cilia. In this way, the tracheobronchial surface is swept and the airways cleaned. This organization is important because the alternative, a mucus blanket such as in the normal colon, would likely be impossible to transport from the larger surface area of the distal airways to the smaller surface area of the larger airways. The organization of the mucus layer in the colon is not random and organized by sheets of secreted MUC2. Both the MUC2 and MUC5B mucins are packed in the goblet cell in a highly organized way and, once secreted, these mucins form differently organized mucus in the colon and airways of piglets (7) and humans (11). The main difference between these larger animals and the mouse is their high number of submucosal glands. In the pig, the submucosal glands produce thick and long mucus that we have called bundles built around the MUC5B mucin (8). The mucus of the upper normal mouse respiratory tract is appearing instead as clouds, also based on the Muc5b mucin. This mucin is produced by secretory (Clara and goblet) cells in the airway epithelium (37). The Muc5b mucin is secreted in the form of strands and becomes part of the mucus clouds. The few submucosal glands at the very proximal mouse trachea secrete mucus bundles. As suggested from the transport of charcoal, these bundles are efficient in transporting debris the last part over the rim at the larynx. We speculate that these bundles may cover a larger surface area than the clouds and, therefore, can be more efficient in large airways.

The positively charged Alcian blue binds the densely negatively charged mucin domain of the mucins. We have previously used this to stain and measure airway porcine mucus transport ex vivo (7, 8). Herein, we have used the Alcian blue-stained tilted tracheal explant mucus-moving assay to analyze murine mucus. The importance of tilted tissue in such experiments has also been observed for air-liquid interface human bronchial culture models, where most of the fluid drained while the upward transport of mucus continued (4). The upward mucus transport on the cultured cells was maintained during the tilt, but lost in the absence of mucus with only buffer present. This suggested that the viscoelastic properties of mucus are required for transport against gravity.

Having mucus clouds with very few bundles in mouse airways is likely to produce an efficient mucociliary clearance of the small airways. These mucus clouds could be sufficient to remove small inhaled particles. Mucus bundles might not be required in the mouse as large particles are filtered out in the narrow nasopharynx of small animals. However, in large animals like humans and pigs, bundles might be necessary for the removal of larger particles that can reach the lungs through the nasopharynx and mouth. That bundles are important for the transport of large particles is supported by studies by Fischer et al. (12), who showed that bundles from the submucosal glands of pigs are required to remove large particles. In addition, the thick bundles with more than 1,000 parallel packed MUC5B polymers might be too big for the small airways of mice. Interestingly, in larger animals, the submucosal glands are found down to about the 10th bronchial bifurcation, and the smaller airways below, thus, lack mucus bundles. We suggest that the mucus of the small airways of distal human and pig airways might be similar to the mouse mucus clouds, something that is supported by the bronchiole surface secretory cells that largely produce MUC5B (22).

Although Muc5ac expression is scarcely detectable in mouse healthy tissues, it is observed by immunohistology in Muc5b-deficient mice. Muc5b and Muc5ac show different morphology when stained with Alcian blue and lectins. The Muc5b and Muc5ac mucins are both covalently linked COOH-terminal to COOH-terminal via disulfide bonds, forming end-to-end dimeric building units. Later in the secretory pathway, these dimers are sorted to specific vesicles, where they form larger polymers by forming disulfide bonds between two NH₂-termini. The Muc5b mucin forms linear polymers (25, 31). The situation for the Muc5ac mucin is still not clear, but preliminary observations indicate that four NH₂-termini are linked together causing the Muc5ac to form more net-like structures. The interpretation that the Muc5b forms linear polymers is consistent with the Alcian blue-stained mucus strands in this study. Lectins are sufficiently small to efficiently stain secreted mucus for confocal microscopy. Both the Muc5b and Muc5ac mucins were stained by the UEA1-lectin showing similar glycosylation and that these are biosynthesized in the same surface secretory cells. The lectin staining of naïve mice showed long linear strands for Muc5b, but a more variable appearance for Muc5ac, both in agreement with the type of polymers they are suggested to form.

The mean mucus transport velocity in WT mice was 0.7 mm/min (12 µm/s). Other studies have measured mucociliary clearance in normal mouse trachea and reported variable values ranging from 0.03 mm/min (0.5 µm/s) (36) to 11 mm/min (180 µm/s) (14). This discrepancy in measurements is caused by differences in the technique, tissue (or cells), regional differences, and maybe most importantly what is really measured. We observed different velocities of Alcian blue-stained mucus clouds and the fast moving negatively charged beads (1.2 mm/min, 20 µm/s) that we concluded moved with the ASL-mucus. We did not detect any difference in bead transport as a function of size, but once the beads were trapped on the mucus, a slower transport velocity was observed. Henning et al. (16) also noted that the mucociliary transport was independent of the particle size, but highly influenced by the material and thus binding properties of the particles. Notably, in their study, mucoadhesive chitosan-PLGA particles were transported at the slow rate (0.7 mm/min, 12 µm/s) which is the same rate as we measured for Alcian blue-mucus and mucus-collected beads. On the other hand, their particle clearance rate was also faster (3 mm/min, 50 µm/s) (16). Mucus transport decreased dramatically in Muc5b^−/− mice as compared with WT, which is consistent with the previous results and confirm that Muc5b is indispensable for normal mucociliary clearance (28). Interestingly, we also found transport to be decreased in Muc5ac^−/− mice, a finding that is not easily understood especially as Muc5ac is not a component of normal mucus. In pigs, we observed that the mucus bundle movement was controlled by acetylcholine, something that was not observed for the mouse mucus clouds (7).

The ciliary beating is driving the flow of the surface ASL-mucus. This flow, measured as bead transport, was faster than the mucus cloud transport. The difference in velocity was twofold, in contrast to the piglet where an almost 10-fold difference could be observed (7, 8). In the pig, it was concluded that the mucus bundles were held back by transient anchoring of the mucus to the surface goblet cells. Even more interesting was the observation that acetylcholine increased the ASL-mucus movement at the same time as the mucus bundles almost stopped moving. However, acetylcholine had no effect on mucus movement in mouse. As can be suggested from observation in the Muc5ac^−/− mice, the Muc5b mucin might be attached to the surface epithelium in healthy mouse airways. To efficiently clean the tracheobronchial surface, mucus must stay close to the surface and allow control of transport separately from the continuous ciliary beating. It is thus very important to further understand attachment/detachment of the mucus in the normal airways and how this is controlled. Such knowledge may aid understanding of the origins of lung disease and suggest new therapeutic strategies. In newborn piglets lacking the CTFR channel, the bundles were almost stagnant (7). Quantification of goblet-to-bundle Muc5ac attachment sites revealed that these were more frequent in the CF piglets, suggesting that mucus detachment was slower or impaired in CF.

The normal lung responds quickly to particles, environmental pollution, or bacteria by increasing mucus production. This mucus will finally generate a mucus layer as observed in COPD or CF (11). As exposure to the environment starts right after birth, there is a problem to define the nature of the mucus of the normal naïve tracheobronchial surface. When analyzing less than 1-day-old piglets, there was no continuous mucus layer and mucus was only observed as around 25-µm-thick mucus bundles (8). In our 6- to 8-wk-old mice, living in a relatively dust-free environment with a bedding of wood shavings, there was also no mucus layer coating and the mucus observed was in the form of discontinuous mucus clouds. This argues for the surface ASL-mucus of the normal tracheobronchial surface being without a continuous mucus. That the normal tracheobronchial surface has an ASL with a discontinuous mucus was previously pointed out by Van As (32, 33). He described mucus “flakes,” reflecting the same phenomenon as our mucus clouds. We prefer, however, the name “cloud” as this better describes their native three-dimensional appearance as flakes were observed after fixation. Thus, we conclude that the normal mouse lungs have surface ASL in which discontinuous mucus clouds are floating.

GRANTS

This work was supported by the European Research Council (ERC 694181), National Institute of Allergy and Infectious Diseases (U01AI095473), The Knut and Alice Wallenberg Foundation (2017.0028), Swedish Research Council (2017-00958), The Swedish Cancer Foundation, IngaBritt and Arne Lundberg Foundation, the Swedish State under the agreement between the Swedish Government and the County Council, The ALF agreement (236501), Wilhelm and Martina Lundgren’s Foundation, The Cystic Fibrosis Foundation (CFF), Swedish CF Foundation, the Swedish Heart and Lung Foundation, Erica Lederhausen’s Foundation, and the Lederhausen’s Center for CF Research at University of Gothenburg.

DISCLAIMERS

This content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

D.F., S.T.-M., A.E., and G.C.H. conceived and designed research; D.F., A.M.R.-P., and S.T.M. performed experiments; D.F., A.M.R.-P., A.E., and G.C.H. analyzed data; D.F., A.M.R.-P., A.E., and G.C.H. interpreted results of experiments; D.F. prepared figures; D.F. and S.T.-M. drafted manuscript; A.M.R.-P., C.M.E., A.E., and G.C.H. edited and revised manuscript; D.F., A.M.R.-P., S.T.-M., C.M.E., A.E., and G.C.H. approved final version of manuscript.