Induction of ciliary orientation by matrix patterning and characterization of mucociliary transport

Abstract

Impaired mucociliary clearance (MCC) is a key feature of many airway diseases, including asthma, bronchiectasis, chronic obstructive pulmonary disease, cystic fibrosis, and primary ciliary dyskinesia. To improve MCC and develop new treatments for these diseases requires a thorough understanding of how mucus concentration, mucus composition, and ciliary activity affect MCC, and how different therapeutics impact this process. Although differentiated cultures of human airway epithelial cells are useful for investigations of MCC, the extent of ciliary coordination in these cultures varies, and the mechanisms controlling ciliary orientation are not completely understood. By introducing a pattern of ridges and grooves into the underlying collagen substrate, we demonstrate for the first time, to our knowledge, that changes in the extracellular matrix can induce ciliary alignment. Remarkably, 90% of human airway epithelial cultures achieved continuous directional mucociliary transport (MCT) when grown on the patterned substrate. These cultures maintain transport for months, allowing carefully controlled investigations of MCC over a wide range of normal and pathological conditions. To characterize the system, we measured the transport of bovine submaxillary gland mucin (BSM) under several conditions. Transport of 5% BSM was significantly reduced compared with that of 2% BSM, and treatment of 5% BSM with the reducing agent tris(2-carboxyethyl)phosphine (TCEP) reduced viscosity and increased the rate of MCT by approximately twofold. Addition of a small amount of high-molecular-weight DNA increased mucus viscosity and reduced MCT by ∼75%, demonstrating that the composition of mucus, as well as the concentration, can have significant effects on MCT. Our results demonstrate that a simple patterning of the collagen substrate results in highly coordinated ciliated cultures that develop directional MCT, and can be used to investigate the mechanisms controlling the regulation of ciliary orientation. Furthermore, the results demonstrate that this method provides an improved system for studying the effects of mucus composition and therapeutic agents on MCC.

Significance

Mucociliary clearance is the process whereby the coordinated activity of thousands of cilia interact with and propel mucus from the airways, removing inhaled pathogens and toxins from the pulmonary system. Our understanding of how cilia align and how they interact with the viscoelastic mucus to efficiently clear the airways is incomplete. We demonstrate that a pattern of ridges or grooves introduced into a collagen substrate can induce human airway epithelial cells to orient their cilia and transport mucus in a predetermined path. Furthermore, we examine how changes in mucus rheology and composition affect transport. These studies show that ciliary alignment can be regulated by the extracellular matrix and enables investigations of new treatments for improving mucociliary clearance in muco-obstructive lung diseases.

Introduction

Mucociliary clearance (MCC) is a critical innate defense mechanism that is essential to maintain the health of the pulmonary system (1,2). Through the integrated processes of fluid secretion and absorption, the secretion of large mucin biopolymers and other proteins that compose the viscoelastic mucus, and the proper orientation, coordination, and regulation of ciliary activity, the MCC system continuously captures and removes pathogens and other harmful materials from the airways. The importance of MCC is clearly illustrated by diseases in which impaired MCC is central to disease pathogenesis (3). For example, in cystic fibrosis (CF), the absence or reduced function of the CF transmembrane conductance regulator protein results in dehydration of the airway surface liquid and the accumulation of a thick, viscous mucus that is difficult to clear by MCC or cough (4). The accumulated mucus becomes a nidus for infection, inflammation, and increased mucin production, setting in motion a vicious cycle of lung infection and damage. Primary ciliary dyskinesia is a rare disease caused by genetic defects that impair the function of the motile cilia that line the airways, resulting in reduced or absent MCC, leading to recurrent airway infections (5). Furthermore, impaired MCC also plays a role in the pathogenesis of other chronic diseases, including chronic obstructive pulmonary disease and asthma (3,6). A more complete understanding of the MCC system and how it fails in different disease states will allow the development of specific and efficient therapies to improve MCC and reduce morbidity and mortality.

Currently, the most direct method of studying MCC is to measure the clearance of radiolabeled particles from human subjects over time (7). In a typical study, a subject inhales nonabsorbable technetium-99m (^99mTc)-labeled sulfur-colloid particles and images of the lungs are obtained with a γ-camera over a series of time points after the inhalation. The disappearance of the ^99mTc signal over time is taken to represent the rate of MCC. Although this method provides a direct measure of MCC in humans, the technique requires a high level of expertise and is not widely available. Multiple different methods have been utilized to measure MCC in animal models, including measuring the disappearance of radioactive particles as described above (8,9) and tracking the transport of fluorescent or other particles, either in live animals or in ex vivo (excised) tissues (e.g., (10,11). Although these and other methods provide valuable information, they each have limitations, including the use of highly specialized equipment or techniques (e.g., synchrotron (12), optical coherence tomography (13)), the requirement to add large volumes of fluid to the airway surface, and/or the inability to perform repeat studies on the same animal.

As an alternative to animal models, many investigators have utilized well-differentiated air/liquid interface (ALI) cultures of human airway epithelial (HAE) cells to study aspects of MCC in vitro. Occasionally, these cultures have been shown to spontaneously coordinate their ciliary activity to produce areas of circular mucociliary transport (MCT), colloquially referred to as “mucus hurricanes” (14). ALI cultures have been used to investigate MCT in a wide range of studies, particularly in relation to CF (e.g., (15, 16, 17, 18)) and toxic exposures (e.g., (19,20)). In addition, several investigators have used ALI cultures to measure flow by tracking fluorescent particles over short distances (micrometers or millimeters (21)). We recently reported that by including an insert in the center of the culture membrane to form a circular track (i.e., an MCT device (MCTD)), we could increase the percentage of ALI cultures that developed complete circular transport (CCT) by approximately fivefold (22). In addition to increasing the percentage of cultures demonstrating transport, the MCTD (or “racetrack” cultures), allow for the directional measurement of MCT under controlled conditions, measurement of MCT over longer distances (centimeters), and repeat measurements of the same culture under different conditions over extended time periods. However, because the percentage of cultures that developed CCT was still relatively low under these conditions (∼25–45%), we explored other modifications to our protocol to attempt to increase the number of cultures developing CCT. Here, we report that by introducing a pattern of ridges and grooves into the collagen matrix used to prepare the MCTD, more than 90% of cultures studied demonstrated coordination of their cilia and CCT. We further demonstrate the utility of this system by exploring the effects of different modifications of the mucus layer, including mucin concentration, treatment with a reducing agent, and addition of genomic DNA, on the rate of transport. Importantly, we find that the inclusion of a small amount of DNA in the mucus causes a disproportionately large reduction in transport rate.

Materials and Methods

A Supporting materials and methods section is provided in the online supplemental materials.

Culture of HAE cells

Primary HAE were obtained from the UNC Cell and Tissue Culture Core Facility under protocols approved by the UNC Institutional Review Board, and cultured as previously described (23,24). Briefly, first, passage HAE cells (250,000 cells/cm²) were plated on 30 mm Millicell Culture Inserts (PICM03050; MilliporeSigma, Burlington, MA) coated with collagen as described below. Cells were fed both apically and basolaterally with ALI media until confluent, and then fed only from the basolateral surface. Cultures were fed three times per week for the first 2 weeks of culture and two times weekly for the remainder of the experiments.

Preparation of collagen coated inserts

MCTDs were prepared as previously described by attaching a central ring to the center of Millicell Culture Inserts using silicone sealer, creating a circular track (22). Inserts were coated with human type IV by one of two methods. In the standard procedure, a solution of collagen was added to the apical compartment of the Millicell and allowed to air dry. In the “brushed” procedure, a nylon brush was used to coat the track of the MCTD with a collagen solution. After drying, inserts were irradiated with ultraviolet light in a culture hood for 30 min and stored at 4°C until use. Additional details and videos demonstrating these procedures are described in the Supporting materials and methods section and Videos S4 and S5.

Analysis of ciliary orientation

Immunofluorescence was performed using the whole-mount staining protocol previously described (25). In brief, cultures were immunostained with rabbit anti-centriolar protein homolog B (POC1b) and mouse anti-centriolin (CNTRL). Membranes were imaged using a Nikon N-SIM microscope capturing four to six images per MCTD. Each image was taken collecting 25–35 optical sections of 0.1 μm, including the entire basal bodies-basal feet complexes of all the cells in the field of view at a resolution of 33.333 pixels/μm. Images were processed and analyzed using FIJI (26).

Rheological measurement of mucus solutions

Biophysical properties of the prepared mucus solutions were assessed at 23°C using particle tracking microrheology according to previously published methods (27). In brief, the thermally driven motion of 1-μm-diameter carboxylated Fluospheres added to prepared samples was tracked to determine their viscoelastic properties. Bead motion was recorded for 30 s at a rate of 60 frames/s using a 40× air objective on a Nikon Eclipse TE2000U microscope. Individual bead trajectories were measured automatically using a custom Python program that uses TrackPy (https://doi.org/10.5281/zenodo.34028) for bead localization and tracking. Statistical significance for particle tracking microrheology were performed using the Kurskal-Wallis test. For this analysis, each bead was considered an independent observation to observe the heterogeneity of the specimen.

Measurement of MCT rate

Cultures were visualized with a Nikon Eclipse TE2000 inverted microscope and videos were captured at 37°C with a Redlake ES-310T camera driven by SAVA software (Ammons Engineering, Clio, MI), as previously described (22). Fluorescent beads were added to the mucus solutions to more easily visualize MCT. Recorded videos were analyzed using TrackPy to determine the average speed of MCT.

Results

Collagen ridges induce ciliated cell alignment

In previous studies, we found that providing HAE cells with a directional cue (i.e., a central cylinder (22)) improved the percentage of ALI cultures that developed coordinated transport. To provide an additional directional cue during differentiation of HAE cells, we used a small nylon brush to apply a thin collagen coating to preformed MCTDs. The collagen was applied in thin layers around the circular path of the MCTD and allowed to dry briefly between layers. This resulted in a series of concentric ridges/grooves in the collagen layer (Fig. 1 B) that followed the circular pattern of the MCTD. In contrast, although inserts coated with a solution of collagen showed some faint ridges/grooves, these crossed the inserts in a straight-line pattern from one edge to the other (Fig. 1 A), and we speculate that these ridges are formed by the pattern of fibers inherent in the Millicell membrane. HAEs from multiple donors were plated on MCTDs that had collagen applied either by the addition of a bulk collagen solution (“standard MCTD”), or by the brushing technique described above (“brushed MCTD”), and cultured at an ALI until well differentiated (6–8 weeks) (22,28). In these studies, only a small number of the standard MCTD cultures initiated developed CCT (<10%; 13 donor samples), although many had areas of coordinated cilia that demonstrated local transport. In contrast, 85 out of 92 (∼92%; 22 donor samples) consecutive cultures initiated on brushed MCTDs developed CCT (see Video S1). Once differentiated, cultures maintained CCT for months, with some cultures demonstrating transport >1 year after initiation. Between twice-weekly feedings, the apical surface of most cultures would become dry, causing transport to cease, but this was easily restored by addition of fluid and/or washing the surface to remove accumulated mucus. The transport speed of endogenous mucus varied depending on the cell donor, the amount of mucus present, and the hydration state of the culture. However, when endogenous mucus was removed by washing and replaced with a standard solution of 2% bovine submaxillary mucin (BSM), MCT was typically between 50 and 100 μm/s. This is in good agreement with previous estimates of in vivo MCT rates (29) and other studies of HAE cultures.

Figure 1.

Introduction of ridges into the collagen coating. (A) Diagram (left) illustrating the location of a representative scanning electron micrograph (right) of an MCTD culture insert coated with collagen applied by bulk solution (standard). Note the straight-line pattern of ridges that extends into the side of the MCTD. (B) Diagram (left) illustrating the location of a representative scanning electron micrograph (right) of an MCTD culture insert coated with collagen by repeated application with a small brush. Note the curved lines of the ridges follow the circular track of the MCTD.

Video S1. Video showing continuous mucociliary transport around an entire MCTD

The culture surface was overloaded with 1 μm fluorescent beads to label the mucus. Doing this causes the mucus to clump and appear as strands, but allowed the mucus to be visualized at a low enough resolution to see the entire width of the track in a single field of view (compare to Videos S2 and S3). The video was captured at 10 fps using a 2× objective and is played back at 10 fps. The field of view was moved around the track to show mucus flowing continuously around the entire track. At eight evenly spaced positions, the field of view was paused for a few seconds to show the mucus flowing independently of the motion of the field of view.

The proper alignment of cilia, both within and between cells, is essential for effective MCC (25,30). The direction of the ciliary beat is indicated by the alignment of the basal body-basal foot within individual cells (31). To evaluate the alignment of the basal bodies-basal feet, standard ALI cultures and brushed MCTDs that demonstrated CCT were examined by fluorescent immunocytochemistry. Basal bodies were labeled with a POC1b antibody and basal feet were labeled with CNTRL (Fig. 2, A and B; (32)). A custom-derived script was used to detect basal body-basal foot pairs (Fig. 2, C, C′, D, and D′) and to determine the mean vector, the length of the mean vector of each individual cell (Fig. 2 E), and the overall length of the mean vector “r” as a measurement of the overall orientation of and degree of coordination of all the cilia in the image (Fig. 2 F). Multiple images from several standard and brushed MCTD cultures were analyzed. The brushed MCTD cultures displayed a significantly higher degree of ciliary coordination compared with the standard cultures, both within (0.54 ± 0.01 vs. 0.22 ± 0.01, Fig. 2 E) and between (0.46 ± 0.03 vs. 0.13 ± 0.013, Fig. 2 F) ciliated cells. In cells grown on standard inserts, some areas displayed high degrees of intracellular coordination, but the overall orientation was lower, with length of the mean vector between 0.05 < r < 0.2 values (Fig. 2 F). These results confirm that the brushed collagen matrix substantially improved the intra- and intercellular coordination between ciliated cells.

Figure 2.

Ciliary coordination analysis. Basal bodies (BB) were labeled with an anti-POC1 centriolar protein B antibody (green) and basal feet (BF) were labeled with CNTRL antibody (magenta) to determine ciliary orientation in standard (A) or brushed (B) cultures. The boxed area in the left panel is enlarged in the right panel. (C, C′, D, D′) Representative images generated by analysis of (A) and (B) using a custom program to detect BB-BF pairs. The starting point of each vector is determined by calculating the centroid of each cell from its cell border. The green circle shows the magnitude = 1 distance for the mean vector. The unit vectors were used to quantify the degree of ciliary coordination. (E) Length of the mean vector (r) in individual cells from standard (n = 115 cells) or brushed (n = 165 cells) cultures. 0 < r < 1, where 0 means no coordination and 1 means full coordination of cilia orientation. Error bars indicate SD. (F) Length of the mean vector (r) indicating the overall orientation of the cilia in the scanned fields from standard (n = 16 fields) or brushed (n = 21 fields) cultures. Images were collected from three cultures; each from a different donor. The quadruple asterisks indicate p < 0.0001 using an unpaired t-test with Welch correction. Error bars indicate SD.

Measurement of MCT

To establish a standard protocol for measuring MCT, cultures exhibiting CCT were washed to remove endogenous mucus and then a solution of 2% BSM containing fluorescent microbeads was added to the apical surface and allowed to equilibrate. The movement of the fluorescent beads was recorded at eight equally spaced fields around the culture, analyzed automatically using TrackPy software, and averaged to obtain the rate of MCT. Initial experiments compared the rate of transport of 2% BSM, representing a “healthy” mucus, to that of 5% BSM, representing the concentration of mucin that is typically observed in CF sputum (33). Cultures were washed, and the rate of transport of 2% BSM was measured. In this set of cultures, the transport rate was 74.7 ± 15.7 μm/sec, slightly faster than our earlier study in which transport of 2% BSM was 56.1 ± 5.0 μm/sec in standard cultures (22). The same culture was washed, and the rate of transport of 5% BSM was measured followed by washing and repeating the measurement of transport of 2% BSM again. As shown in Fig. 3, the rate of transport of 5% BSM was substantially reduced compared with the initial baseline transport rate of 2% BSM (see Videos S2 and S3). Importantly, the results also show that after replacing the 5% BSM solution with 2% BSM, the rate of MCT was restored to near baseline, demonstrating the reproducibility and robustness of the system.

Figure 3.

Effect of mucin concentration on MCT. The speed of MCT in 2 or 5% BSM was measured in brushed collagen MCTD cultures. MCT was measured sequentially in the same cultures (n = 3 from two different donors) using 2%, 5%, and then 2% (labeled 2%′ in the chart) BSM solutions to demonstrate the reproducibility of the system. Individual data points for three cultures are shown (diamond, square, circle). The columns show the mean and error bars indicate SEM. A Kruskal-Wallis one-way analysis of variance indicated the groups did not belong to the same distribution (p = 0.03) apparently because of reduced transport of the 5% BSM.

Video S2. Representative video of a brushed MCTD showing mucociliary transport of fluorescent beads in a 2% solution of BSM

Eight equally spaced fields around the insert are recorded (as illustrated in the diagram on the lower left) and analyzed to determine the average speed of MCT. (original recording speed 30 fps; playback 30 fps).

Video S3. Representative video of a brushed MCTD showing mucociliary transport of fluorescent beads in a 5% solution of BSM

Eight equally spaced fields around the insert are recorded (as illustrated in the diagram on the lower left) and analyzed to determine the average speed of MCT. The rate of transport is obviously reduced compared to Video S1. (original recording speed 30 fps; playback 30 fps).

Video S4. Assembling the MCTD

Apply glue to the central cylinder. Wipe away glue from the outer edge of the cylinder. Place it on the jig. Place the insert over the cylinder and press the insert membrane onto the cylinder. Remove the MCTD from the jig and use an applicator to spread the glue over the membrane inside the central cylinder. Set aside to dry.

Video S5. Preparing the collagen coating

Put the volume of collagen to apply in a separate tube. Use 50 ul of collagen solution per centimeter squared of membrane. For 30 mm Millicells (23 mm membrane outer diameter) and a central cylinder with an outside diameter of 15 mm, we use 122 μl per MCTD). Dip the brush into the collagen solution. Brush the collagen onto the MCTD membrane in a circular motion. Set the MCTD aside to dry. Repeat the dip and brush for each MCTD. When all MCTDs have dried, brush again, always keeping to the same brushing direction, counterclockwise in the example shown. Continue until all the collagen has been applied. After drying, inserts are irradiated with UV light in a culture hood for 30 min. and stored at 4°C until use.

Effect of a reducing agent on MCT

Recent reports have suggested that reducing the disulfide cross-links in mucin polymers may reduce the viscosity of airway mucus in diseases, including CF, asthma, and chronic obstructive pulmonary disease, thereby increasing MCC and improving clinical outcomes (17,34). We therefore measured the effect of the reducing agent tris(2-carboxyethyl)phosphine (TCEP) on the rheology and MCT of 2 and 5% BSM. Microrheological studies demonstrated that 10 mM TCEP significantly reduced the viscosity of both 2 and 5% BSM solutions (Fig. 4 A). Treatment with TCEP also increased the MCT rate of 5% BSM to approximately that of 2% BSM, although having no significant effect on the MCT rate of 2% BSM (Fig. 4 B). Interestingly, we find that when comparing MCT with complex viscosity (Fig. 4 C), there is a minimal viscosity that the mucus solution must surpass before MCC begins to be reduced. These results provide further support for the development of reducing agents as mucolytic therapies for obstructive lung diseases.

Figure 4.

Effect of the reducing agent TCEP on rheology and MCT of BSM. (A) Solutions of 2% and 5% BSM were treated with or without 10 mM TCEP and rheological properties were determined (^∗∗∗p < 0.001). Error bars indicate SD. (B) MCT of 2% and 5% BSM treated with or without TCEP. n = 6 cultures from three donors for 2% BSM, and n = 8 cultures from three donors for 5% BSM. ^∗∗p < 0.01 was found by using Student t-test. Error bars indicate SEM (C) MCT versus mucus viscosity showing that there is a threshold viscosity that mucus must exceed to limit transport. Error bars indicate SD.

Effect of DNA

In CF airways, neutrophils infiltrating in response to infection release DNA, adding another high-molecular-weight component to the mucus. The relative amount of DNA to mucin varies among CF subjects, ranging from 1:100 in children with little disease to 1:20 in adults with worse disease (35,36). Because of the higher rigidity of DNA compared with mucins, we hypothesized that small amounts of DNA could have significant effects on mucus rheology and transport. To test this, we measured the viscosity of 2, 10, and 50 mg/ml BSM solutions with and without the inclusion of small amounts of DNA. The results show that even at concentrations 20- to 100-fold less than the concentration of the mucin, the presence of DNA significantly increased mucus viscosity (Fig. 5 A). Furthermore, the rate of transport of 1% BSM was reduced ∼75% by the presence of 0.04% DNA (Fig. 5 B). These results clearly show that other components of mucus, even when present at low relative amounts, can have disproportionate effects on the overall rheology and transport properties of the mucus solution.

Figure 5.

Effect of DNA on mucus rheology and transport. (A) Increase in rheological properties of standard solutions of BSM with increasing amounts of DNA. Error bars indicate SD. (B) Reduction of MCT after addition of high-molecular-weight DNA. For each test condition, the baseline rate of MCT was determined in 1% BSM followed by measuring the rate of MCT after the addition of control (PBS) or DNA solution. Figure shows three individual experiments from two donors and the average data. Error bars indicate SEM.

Discussion

Previous studies have shown that when cultured at an ALI, HAE cells tend to spontaneously develop areas of coordinated ciliary activity (14). However, in a typical culture insert, these areas develop in random fashion and usually do not coordinate with each other. The mechanisms controlling the coordination of ciliary beating are still poorly understood, although recent studies have highlighted the involvement of the planar cell polarity (PCP) pathway and hydrodynamic forces (37,38). Thus, Guirao et al. (37) reported that subjecting differentiating ependymal cells to an external fluid flow could orient ciliary beating in the direction of the flow. Based on our previous studies that demonstrated growing HAE cells in a circular track improved ciliary alignment, coordination, and MCT (22), we reasoned that providing the differentiating cells with an additional directional cue might further enhance their ability to coordinate. Attempts to use fluid flow to improve ciliary alignment in our experience were largely unsuccessful, likely in part because of the requirement for an ALI to obtain optimal differentiation. However, our results show that plating the cells on an MCTD with a preformed pattern introduced into the collagen substrate greatly enhanced the number of cultures that demonstrated CCT.

How the presence of ridges/groves is sensed by the dissociated airway epithelial cells and what signaling pathways are used by the cells to align their cilia are unknown. One likely possibility is that the ridges/grooves provide an axis for the cells to align on. For example, the orientation of keratocytes has been shown to be strongly influenced by the presence of aligned collagen fibrils or nanoscale groves and ridges in the culture substrate (39,40). Binding to the collagen matrix through cellular receptors may orient the cells in the direction of the ridges, establishing a cell polarity, similar to the proximal-distal orientation of developing airway. By mechanisms that are incompletely understood, PCP is established in the developing cells, resulting in a polarized microtubule network and asymmetric localization of PCP proteins, followed by basal body docking and ciliogenesis (38). Because flow has been shown in some cases to orient or refine ciliary beating (41), it is then reasonable to speculate that as ciliary differentiation proceeds and ciliary activity increases, the direction of ciliary beating is further refined. Although the alignment of cilia, both within and between cells, was improved in the brushed cultures, the alignment of cilia may not be as high as in vivo. For example, Vladar et al. reported a mean vector per cell of 0.83 in P5 mouse tracheas that increased with age (38). However, in these studies, orientation was determined by a different method (TEM), and we did not investigate the age of culture as a variable. Future studies to compare the development of PCP and ciliary coordination in vitro with the in vivo situation are warranted. Interestingly, in some cultures, we have observed transport moving in opposite directions on approximately one-half of the insert and forming a circular “hurricane” at their juncture. This suggests that the determination of transport direction can occur independently at different areas around the culture, and once established, may not be easily altered. This agrees with previous animal studies, in which cilia in sections of trachea that were reversed in vivo continued to beat in their original direction (42). The precise mechanisms controlling the development, refinement, and maintenance of ciliary alignment or how these processes are altered in different diseases are not known. Although the establishment of directed MCC in vivo during development occurs in a different manner than the development of CCT in an ALI culture, this system, in which the directionality of ciliary beat can be directed by a simple modification of the collagen substrate, will provide a valuable model to investigate the mechanisms involved in detail.

The method we have developed to reproducibly obtain directionally transporting cultures can also be exploited to study all aspects of MCC in a relatively simple in vitro system using HAE cells. For example, by washing off the endogenous mucus produced by the HAE cells and replacing it with a solution of known composition, we investigated the effect of mucin concentration on transport. Not surprisingly, increasing the concentration of BSM from 2 to 5% substantially reduced the transport rate, as previously reported (22). In agreement with recent studies (13,17,33), reducing disulfide bonds in the 5% mucin solution reduced the viscosity of the solution and restored transport rates to that of the 2% mucin solution. Importantly, although treating 2% BSM with the reducing agent also reduced viscosity, this did not significantly affect MCT. When comparing MCT rate with mucus viscosity, we see that MCC is decreased as the viscosity of mucus exceeds 0.01 Pa.s, or 10 times the viscosity of water, with the most pronounced result occurring as η approaches 0.1 Pa.s. We therefore speculate that the viscosity of mucus must exceed a threshold on the order of 10–100 times the viscosity of water (i.e., 0.01–0.1 Pa. s) to slow MCC. This speculation is consistent with the previous reports of Puchelle that found CBF was slowed by viscosities greater than 0.1 Pa.s (43) and by studies of the force generation of cilia (44) that found cilia were tuned to beat against a drag force of ∼0.1 Pa. s. Therefore, we would not expect the chemical reduction of 2% BSM to affect MCC, whereas reducing 5% mucus would be predicted to have a more significant effect. Casting this result into a clinically relevant context, we would not expect the use of a mucolytic to negatively impact MCC in a healthy individual, but would expect a mucolytic to have a positive effect on patients with pathologically concentrated mucus.

To begin to examine the effect of mucus composition in disease states on MCC, we added genomic DNA to the BSM solution to simulate purulent mucus from a CF airway. Surprisingly, we found that that adding a low concentration of DNA (1% of the mucus concentration) had a significant effect on the rheology of the mucus solution, increasing viscosity significantly. This effect was further amplified by adding DNA at a concentration of 5% of the mucin concentration. Furthermore, the addition of 0.04% DNA to 1% mucin reduced MCT by ∼75%. We hypothesize that because DNA molecules are more rigid than mucin molecules, the presence of DNA has a disproportionately larger effect on mucus rheology, and subsequently, MCC. Thus, the composition of mucus, as well as its concentration, can have profound effects on the rate of transport. These results are particularly important for studies of muco-obstructive lung diseases such as CF (35) and non-CF bronchiectasis (36), both of which have elevated DNA concentrations.

The model system presented here has many advantages for studying MCC. First, the ability to induce ciliary coordination and directed mucus transport in a reproducible fashion greatly increases the number of studies that can be performed. Second, because once the cultures are entrained, they maintain transport for long periods of time (>1 year), multiple measurements can be obtained from the same culture. Third, the cultures can be easily manipulated, so different protocols can be developed to study all aspects of MCC. For example, to measure the effect of DNA on MCT, we first measured a baseline rate of transport in cultures that were already transporting 1% BSM. The test solution (buffer, DNA) was then added to the mucus layer and allowed to mix before measuring the experimental rate of transport. Importantly, all of these can be studied under controlled, reproducible conditions in cultures of normal or diseased HAE cells. We anticipate that other improvements may further increase the utility of the system described here. For example, a previous study utilized inserts of reduced dimensions to increase throughput (19). Improvements in the composition of the media used to culture HAE cells may improve cellular differentiation to more closely resemble the in vivo airway. Moreover, additional modifications of the collagen coating, including varying the height and number of ridges and the use of a three-dimensional printer or other technology to generate the circular pattern of ridges in the collagen may further improve reproducibility.

Together, the results presented here demonstrate that this model can be adapted to study many different aspects of MCC, including the development of ciliary coordination, the effects of the composition of mucus and ciliary beat on MCC, and the regulation of ion and fluid transport. Finally, these results demonstrate that this system can be used to test and compare the effect of various therapeutics, including different combinations of agents, to improve MCC in different diseases with different mucus compositions.

Author contributions

L.E.O., D.B.H., P.R.S., and R.S. designed the research. P.R.S., X.M.B.-M., H.G., and M.R.M. performed research. P.R.S., X.M.B.-M., H.G., M.R.M., L.E.O., and D.B.H. analyzed data. L.E.O., D.B.H., P.R.S., M.R.M., and X.M.B.-M. wrote the manuscript.

Editor: Takayuki Nishizaka.