Above left: Culture of human bronchial cells (~ 40 µm tall). Center: Spots of uncaged fluoresciene being transported in normal HBE cultures. Upper culture with, lower culture without mucus. Each pair of images collected at 30 sec intervals. Note lack of transport in mucus-free culture. In transported spot , note the lack of smearing, indicating fluoresceine in PCL being transported along with mucus. Right: XZ images at edge of transporting spot. Red is dextran; green is uncaged fluoresciene. Top row is before uncaging, 2nd - 4th rows, row t=0, 15, 30 sec, resp. Note green fluorescence transport, left to right. Vertical uniformity indicates PCL transport with mucus.

Behavior and fate of periciliary liquid during mucociliary clearance: Cilia, slender µm-diameter projections of the plasma membrane with cores containing microtubule-based motors, rise approximately 7µm from the cell surface, and beat in a whip-like fashion in a layer of mucus-free PCL. Previous theoretical work had predicted that the PCL was more or less stationary during the transport of mucus, which is propelled forward by the power stroke of the cilium. In a vertical profile, it was only the outer about 2µm of PCL that was predicted to undergo net movement, a narrow zone which experiences the cilium's power stroke but which lies above the level of the return stroke [Fulford1986].

What we observed in a recent study [Matsui1998a], however, was contrary to this prediction, namely that within the z-axis resolution of the confocal microscope the entire PCL was transported during mucus transport, and it was transported uniformly at the same rate as the overlying mucus gel. We hypothesized that ciliary mixing of the PCL effectively distributes the momentum imparted to the PCL by its frictional interactions with mucus.

We are testing this hypothesis by using the 3DFM to track the movements of single microspheres added at low concentration to the ASL of human airway epithelial cell cultures [Matsui1998a]. We predict that individual particles will suffer turbulent-like movements throughout the height of the PCL. The 3-D magnetic force capabilities of the instrument should enable us to determine the vertical distribution of propulsive forces generated by the cilia on individual cells. By placing the system under position feedback, we should be able to measure the forces needed to maintain the position of the particle, and therefore determine the forces experienced during the cilium beating cycle.

Effects of volume reduction on mucus viscoelasticity: Mucus is a gel whose scaffolding is made of mucin polymers. The airway mucins are linear polymers of mucin subunits with lengths of at least 30µm. [Sheehan1991; Gendler1995] The mucin monomers are polymerized through end-to-end disulfide bonds and each monomer is a glycoconjugate, 80 - 90 % carbohydrate, with molecular weights exceeding 10 MD. The gels are tangled networks with no inter-molecular crossbridges. [Verdugo1990; Bansil1995] In the ASL of normal humans, the concentrations of mucins in the mucus is approximately 1% by weight, but in cystic fibrosis (CF), an inherited disease characterized by a hyperabsorption of Na+ and liquid from the ASL, the mucin concentra-tion increases to as high as 3 - 4% [Matsui1998b]. We hypothesize that the changes in physical properties of the concentrated mucus in CF (increased viscosity, reduced elasticity, increased adhesivity) impede its transport, creating an environment favorable for bacterial colonization and biofilm formation.

Past investigations of mucus viscoelastic properties [Litt1970; Litt1973; King1974] relied primarily upon cone and plate, or vibrating microball viscometers, techniques that shear and/or align mucin molecules in the gel during the measurement. We propose to use the sub-micron single particle tracking and magnetic force capabilities of the new 3DFM to map the viscoelastic properties of the mucus at high resolution. In samples with low viscosity, the tracking of the Brownian motion of the particles can map the frequency-dependent storage and loss moduli [Gittes1997b; Mason1997]. At high viscosities, it will be necessary to drive the system with the magnetic force capabilities.

In parallel experiments, we are presently studying the growth behavior and adherence of bacteria in mucus subjected to the same series of experimental manipulations. In the final analysis, we hope to provide quantitative data on the changes in mucus viscoelastic properties caused by clinically relevant volume depletion, as well as the effects of these changes on the growth properties of the bacteria that infect CF patients.

Implications for 3DFM: These applications are the first being tested with the system. The viscoelastic property measurements require only particle tracking as a start, as the bead position fluctuations are related to the local viscosity through the fluctuation dissipation theorem. The second measurement requires the local measurement of viscoelastic properties and will require some manipulation of the magnetic particle and long-range tracking.

Updates:
Mucus Rheology and Biofilm Formation

Collaborator: Richard Boucher (Cystic Fibrosis Center)

Cystic Fibrosis patients have higher incidences of respiratory infection caused by bacteria. The primary cause of these infections is p. aeruginosa. CF is also characterized by a thickening of the protective mucus layer lining the lungs. We believe there is a correlation between the thickening of CF-mucus and higher incidence of bacterial infection. The goal of this research is to predict under which conditions p. aeruginosa biofilms will form in mucus in its healthy and disease-like state. Biofilms are promoted by locally high concentrations of auto-signaling molecules, homoserine lactone (HSL). If a bacteria’s motion is restricted by either the viscosity of its host fluid or by being caught in a fluid whose mesh size is smaller than the bacteria, the local concentration of self-signaling molecules will rise as the bacteria divide and remain contained within a given area. We hypothesize that the restriction of bacteria will lead to locally high concentrations of HSL, promoting the formation of bacterial biofilms.

Mean Squared Displacement of 1 um microspheres undergoing thermal diffusion in 2.5% and 8% solids w/w mucus. The results show that there is an order of magnitude difference in the diffusion of the two fluids. Biofilm formation in 2.5% (“healthy”, center) and 8% (“cystic fibrosis-like, right) mucus after 72 hours. Biofilm blooming is obvious in the thicker mucus.

To test this hypothesis, the physical properties of mucus were first determined through the use of microbead rheology. We used the technology developed in CISMM to track the thermal motion of carboxylated microspheres (200 nm – 3 um diameter) in 2.5% (w/w solids) and 8% mucus, with 2.5% mucus representing healthy mucus and 8% a CF-like thickened mucus gel. The thermal motion of the embedded microspheres is tracked, and the viscoelastic properties (viscosity, elasticity, and mesh size) of each fluid are calculated using microbead rheology [Mason and Weitz 1995; Mason 2000]. The viscosity of the fluid is estimated by calculating the average mean squared displacement (MSD, ) of an ensemble of beads (> 200 per size / fluid). According to the Stokes-Einstein relationships, the distance beads diffuse varies inversely with their radius. Thus, by multiplying the MSD of the beads by their radius, we can look to see if there are any size-dependent diffusion properties. Breaks in the size dependence of MSD data generally indicate that objects are above or below the mess size of the fluid.
Microbead Rheology shows us that 2.5% mucus (“healthy”) has an effective viscosity of ~ 0.2 Pa•s and a mesh size ~ 500 nm, compared to 6 Pa•s viscosity and < 100nm mesh in CF-like Mucus. Based on previous bacteria motility assays [Greenberg and Canale-Parola 1977], our result predicts that bacteria will be able to swim at velocities ~ 20-30 um/s in healthy mucus and near zero in CF-like Mucus.
Our hypothesis leads us to believe that when bacteria are able to swim, they will not form biofilms. When they are physically restrained by either being in a high viscosity fluid or caught in a tight mesh, they will form biofilms. Therefore, we predict that biofilms will form in 8% CF-like mucus and not in 2.5% healthy mucus. Testing this hypothesis, GFP p. aeruginosa were cultured in both mucus solutions until their populations became stable (72hrs.). Biofilms formed in 8% mucus and not in 2.5% mucus, confirming out hypothesis.

Cilia Propelled Fluid Dynamics

Collaborators: Richard Boucher (Cystic Fibrosis Center), Richard McLaughlin (UNC Applied Mathematics)

The propulsion of fluids is ubiquitous in biology, with examples including the clearance of mucus in the lung, the flow of cerebrospinal fluid and the flow around the nodal plate in the developing embryo. Understanding these flows requires measurements within biological systems, the development of the theoretical tools to calculate the flows, and the development of model experimental systems to test the new computational models. We have used particle imaging velocimetry to measure the flow profiles of mucus within human bronchial epithelial cell cultures to characterize the strains and strain rates present. To facilitate the calculation of these flows, we have used the 3DFM magnetic systems of CISMM to spin magnetic nanorods in fluid chambers. These rods are 250 nanometers in diameter, 10 microns long and are therefore close mimics of cilia. The rods are spun at 5-15 Hz, the beat rate of cilia, in buffer solutions. To characterize the flows, we have laced the buffer with non-magnetic beads and with fluorescently labeled DNA. The former is tracked using video tracking algorithms developed in CISMM, with the detailed trajectory compared with calculations performed in the McLaughlin group. The agreement is excellent, including the detailed “epicycles” that form when the bead is close to the spinning rod. To further understand the flows, we have performed single molecule imaging of DNA in circulation near the spinning rod. While the DNA folds into a tight ball when unstirred, it straightens to about half its contour length when near the spinning rod. This implies that shear and extensional flows near beating cilia are sufficient to open DNA or other proteins, possibly allowing biochemical activity not available in quiescent solution. For the problem of mucus flow in the lung, these studies indicate that the flows may be accessing strong non-linearities in the mucus rheology related to the stretching of the long chain mucins (~15 microns contour length) near the cilia. This may have profound consequences for the rheological properties necessary for the successful propulsion of mucus by cilia.

(Left) Nanorod (250 nm diameter, 10 microns long, inset) is spun in the 3DFM magnetic system to produce a 5-15 HZ beat cycle. A nonmagnetic bead is entrained in the flow created by the rod and is tracked by CISMM software. The size of the epicycle as a function of the distance of the bead from the rod is shown (triangles) and compared with theory. L-DNA (14 um) is fluorescently tagged and tracked in the flow to measure the shear forces created by the swirling rod. The shear forces are strong enough to extend the DNA to over one half its contour length.