Common gene therapy viral vectors do not efficiently penetrate sputum from cystic fibrosis patients

Abstract

Norwalk virus and human papilloma virus, two viruses that infect humans at mucosal surfaces, have been found capable of rapidly penetrating human mucus secretions. Viral vectors for gene therapy of Cystic Fibrosis (CF) must similarly penetrate purulent lung airway mucus (sputum) to deliver DNA to airway epithelial cells. However, surprisingly little is known about the rates at which gene delivery vehicles penetrate sputum, including viral vectors used in clinical trials for CF gene therapy. We find that sputum spontaneously expectorated by CF patients efficiently traps two viral vectors commonly used in CF gene therapy trials, adenovirus (d∼80 nm) and adeno-associated virus (AAV serotype 5; d∼20 nm), leading to average effective diffusivities that are ∼3,000-fold and 12,000-fold slower than their theoretical speeds in water, respectively. Both viral vectors are slowed by adhesion, as engineered muco-inert nanoparticles with diameters as large as 200 nm penetrate the same sputum samples at rates only ∼40-fold reduced compared to in pure water. A limited fraction of AAV exhibit sufficiently fast mobility to penetrate physiologically thick sputum layers, likely because of the lower viscous drag and smaller surface area for adhesion to sputum constituents. Nevertheless, poor penetration of CF sputum is likely a major contributor to the ineffectiveness of viral vector based gene therapy in the lungs of CF patients observed to date.

Figure 1. Sample 20 s trajectories.

Representative trajectories of (A) Adenovirus (Adv), (B) AAV5, (C) muco-adhesive PS control nanoparticles and (D) muco-inert control PS-PEG nanoparticles. All trajectories have MSD values within one standard deviation of the ensemble average.

Figure 2. Averaged ensemble mean squared displacements of viral vectors and synthetic particles with respect to time scale.

Data represent n = 5 independent experiments with ∼100 particles tracked per experimental condition. Error bars represents standard error. *Statistically significant difference when compared with AAV5, AdV or PS (p<0.05).

Figure 3. Distribution of logarithms of individual particle effective diffusivities (D_eff).

Distributions of D_eff of AdV, AAV5, PS and PS-PEG particles represented as a percentage of particles in CF sputum at a time scale of 1 s.

Figure 4. Transport mode distributions of AdV, AAV5, PS and PS-PEG particles.

Particles were classified into either (i) immobile or hindered and (ii) diffusive , –. *Statistically significant difference when compared with AAV5, AdV or PS within the same transport mode classification (p<0.05).

Figure 5. Theoretical model of particle penetration across a sputum layer.

(A) Schematic of the model, where particles are deposited in airway lumen on top of the CF sputum layer and must penetrate a 10 µm sputum layer to reach the epithelial cells. The pericilliary layer is much smaller in CF patients due to the collapsed cilia from the accumulated sputum . (B) Estimated fraction of viral and synthetic particles that are capable of penetrating a 10 µm thick layer of CF sputum over time using Fick's second law. *Statistically significant difference when compared with AAV5, AdV or PS (p<0.05).