Published ahead of print on February 4, 2005, doi:10.1165/rcmb.2004-0410OC
© 2005 American Thoracic Society DOI: 10.1165/rcmb.2004-0410OC Viscoelastic Gel Formulations Enhance Airway Epithelial Gene Transfer with Viral VectorsProgram in Gene Therapy, Department of Pediatrics, Carver College of Medicine, The University of Iowa; and Division of Pharmaceutics, College of Pharmacy, The University of Iowa, Iowa City, Iowa Correspondence and requests for reprints should be addressed to Paul B. McCray, Jr., Department of Pediatrics, 240G EMRB, The University of Iowa, Iowa City, IA 52242. E-mail: paul-mccray{at}uiowa.edu
Advances in gene transfer to the conducting airways for the treatment of pulmonary diseases such as cystic fibrosis have identified several vector classes that transduce airway epithelia in vitro and in animal models. One barrier to epithelial gene transfer is the rapid removal of materials from the airway surface via mucociliary clearance. This host defense mechanism limits gene transfer efficiency to airway epithelial cells. Here we show that formulation of gene transfer vectors with viscoelastic gels provides longer epithelial residence time and increases vector-mediated gene transfer efficiency. Gene transfer with adenoviral, adeno-associated, and lentiviral vectors all significantly improved after formulation with viscoelastic gels designed to slow mucociliary clearance. Importantly, viscoelastic gel formulations enhanced vector transduction to the conducting airways, the desired treatment target for diseases such as cystic fibrosis.
Key Words: gene transfer • viral vectors • viscoelastic gel
Gene transfer vectors can be readily delivered to the airways by aerosol or instillation, but once at the mucosal surface they must overcome numerous barriers for efficient uptake and expression in epithelia. Cellular barriers to gene transfer include epithelial tight junctions that restrict paracellular uptake and access to some receptors, low endogenous rate of endocytosis at the apical membrane, extracellular glycocalyx, and rapid mucociliary clearance of foreign material entrapped in secreted mucins (reviewed in Ref. 1). Early attempts to bypass cellular barriers included the use of agents to transiently disrupt tight junctions, such as EGTA, to increase paracellular permeability (2–5), or the targeting of specific cellular receptors to enhance vector uptake (1, 3, 6–8). Because of safety considerations associated with modifying tight junctions in vivo, efforts to directly transduce cells from the luminal surface have increased. As a result, several different viral vectors are being investigated for improved gene transfer (1, 9). Progress has been reported with lentiviral vectors, adenoviral vectors, and adeno-associated viral vectors (AAV) (9–13). Early success in the translation of gene-based therapies with adenoviral viral vectors was seen in many cell culture and animal models. However, the location of the adenovirus receptor (CAR) on the basolateral surface of the airway epithelium diminished the transduction efficiency unless tight junctions were disrupted (7). In addition, adenovirus is a nonintegrating vector; thus chronic, repeated administration of the gene therapy would likely be required to sustain a population of normal respiratory epithelial cells. Other viral vectors, notably AAV5 (13), AAV6 (11), AAV9 (14), and pseudotyped lentiviruses (9), can target receptors on the luminal airway surface. These vector systems are promising tools for gene therapy. Mucociliary clearance is a barrier to efficient gene transfer. The luminal surfaces of the large airways are populated with ciliated epithelia that act to remove inhaled particles trapped in mucus. Efficient mucociliary transport is accomplished by the coordinated beating of the cilia through the periciliary liquid layer with intermittent ciliary contact with the overlying mucus layer. Mucociliary clearance rates from the large airways of humans range between 0.25 and 0.5 cm/min (15). Consequently, the contact time of a gene transfer vector with the epithelial surface may be too short to achieve significant uptake of a gene-based therapeutic agent, regardless of the uptake mechanism. Seiler and coworkers (16) previously demonstrated that formulating recombinant adenovirus in a thixotropic solution containing 1% Avicel RC-591, a mixture of microcrystalline cellulose and carboxymethylcellulose, decreased virus clearance and enhanced gene transfer to airway epithelia. Results in this study indicated that such gel formulated adenoviral vectors enhanced gene transfer in vitro and in vivo by inhibiting mucociliary clearance. We hypothesized that optimized viscoelastic gel formulations might increase the transduction efficacy for multiple classes of gene transfer vectors. We tested three viscoelastic gels, carboxymethylcellulose (CMC), methylcellulose (MC), and poloxamer 407 (P407), for their ability to enhance gene transfer of adenoviral, AAV5, and lentiviral vectors to the airways of mice.
Vector Production Recombinant adenoviral vectors expressing ß-galactosidase (Ad-ßGal) or luciferase (Ad-Luc) were prepared as previously described (17) by the University of Iowa Gene Transfer Vector Core at titers of  1010 plaque-forming units (pfu) per ml. Recombinant AAV5 vector expressing eGFP (AAV-eGFP) at titer of 4 x 1012 viral genomes (vg) per ml was a kind gift from Jay Chiorini (13, 18). The FIV vector used in this study was reported previously (6, 19). The FIV vector construct expressed a nuclear targeted ß-galactosidase cDNA directed by the CMV promoter (FIV-ßGal) or a firefly luciferase cDNA directed by the RSV promoter (FIV-Luc). The Autographa californica GP64 envelope has efficiently pseudotyped an HIV-based vector (20), and will efficiently pseudotype the FIV-based vector (P.L.S. and P.B.M., unpublished observations).
Vehicle Preparation
In Vitro Viral Vector Administration The NuLi cells were transduced by diluting vector preparations in serum-free media to achieve an MOI of 10, and 100 µl of the solution was applied to the apical cell surface for a 2-h period at 37°C. The cells were rinsed with media and cultured for 4 d. To transduce airway epithelia from the basolateral side, the Millicell culture insert containing the epithelium was turned over and virus applied to the bottom surface of the filter for 2 h in 100 µl of media. After the 2-h incubation, the vector was removed, cells were rinsed in media, and the culture insert was turned upright and cultured at 37°C, 5% CO2, for 4 d. After the 4-d incubation, cells were harvested and ß-galactosidase activity quantified. The Galacto-light chemiluminescent reporter assay (Tropix, Bedford, MA) was used to quantify ß-galactosidase activity according to the manufacturer's protocol. The relative light units (RLUs) were quantified using a luminometer (Monolight 3010; Pharmingen, San Jose, CA) and standardized to total protein as determined by modified Lowry assay (23240; Pierce Biotechnology, Rockford, IL) using the manufacturer's protocol.
In Vivo Viral Vector Pulmonary Administration
In Vivo Viral Vector Nasal Administration
Bioluminescence Imaging
Mucociliary Transport Rates
Statistics
Formulating Adenoviral Vectors in Viscoelastic Gels Markedly Enhances Gene Transfer to the Conducting Airways Gene transfer to mouse airway epithelia with Ad-ßGal formulated with saline vehicle (neat) is inefficient (Figures 1A and 1B). Consistent with previous observations (4, 5), the efficiency was enhanced by pretreating the airways with 400 mM EGTA (Figures 1C and 1D). We hypothesized that transient inhibition of mucociliary clearance by viscoelastic gel formulations would improve gene transfer efficiency. To test this hypothesis, we formulated Ad-ßGal with final concentrations of 1% MC, 1% CMC, or 7.5% P407 and delivered each preparation to the lungs of mice via orotracheal intubation. Four days after vector delivery, the lungs were excised and X-gal stained. The appearance of ß-galactosidase–positive cells was readily apparent to the naked eye in lungs receiving gel formulated vector (Figures 2A–2C), but not in naïve animals (Figure 2D) or control animals receiving the adenovirus without gel formulation (data not shown). Upon microscopic examination, the transduction of the conducting airway epithelium with Ad-ßGal formulated with 1% MC or 1% CMC was striking (Figures 2E, 2F, 2I, and 2J), whereas the conducting airway cell transduction achieved with the P407 was considerably less (Figures 2G and 2K). No ß-galactosidase–positive cells were observed in the naïve control (Figures 2H and 2L).
The overall survival rates after orotracheal delivery of 50 µl of formulated vector was as follows: neat 31/35 (89%), MC 29/32 (90%), CMC 36/40 (90%), and P407 12/28 (43%). Mice tolerated vector formulated with MC and CMC well; however, less than half of the mice receiving vector formulated with P407 survived. The 1 in 10 deaths that occurred after delivery of MC or CMC formulated vector were not significantly different than unformulated vector and were likely a combined result of anesthesia and fluid volume. Due to the low survival rate and the lower gene transfer efficiency, P407 was not included as a tested vehicle in subsequent in vivo experiments. To quantify relative gene transfer efficiencies, serotype 5 adenoviral vector expressing firefly luciferase (Ad-Luc) was delivered to mouse airways via orotracheal intubation. Bioluminescent signal was assessed using a CCD camera 5 min after intraperitoneal luciferin delivery (Figure 3). As shown, the maximum gene transfer was observed with Ad-Luc formulated with 1% MC. Neat and EGTA-pretreated animals displayed approximately equal levels of gene transfer in the lung (Figure 3). However, EGTA-pretreated animals had much greater expression in the nasal epithelia, presumably due to enhanced vector access to CAR.
Effects of Viscoelastic Gels on Mucociliary Clearance and Transepithelial Resistance The effect of gel formulation on mucociliary transport rate (MTR) was measured using a bovine tracheal explant model (23) (Table 1). The transport rate of 1% CMC, 1% MC, or 7.5% P407 was measured and compared with the transport rate of a control charcoal:mucin suspension. CMC was most effective at decreasing MTR (40% decrease). MC and P407 also decreased MTR (14% and 10% decrease, respectively). These data suggest that one method of action for polymeric gels is the slowing of mucociliary clearance. The resulting increase in mucosal contact time likely facilitates significantly improved uptake of the Ad-ßgal vector via receptor-independent mechanisms (24).
Increasing vector residence times on cells will lead to increased receptor-independent transduction with adenovirus applied to the apical surface (24). To test the effectiveness of our viscoelastic gel formulations at increasing receptor-independent cellular transduction, we used NuLi cells grown at an air–liquid interface with previously demonstrated basolateral CAR expression (21). Ad vector was formulated with gels or EGTA and applied to the apical surface of NuLi cells (Figure 4A). Control cells received vector formulated in serum-free media (MOI = 10) applied at the apical or basolateral surface. As expected, in the absence of formulation (neat), the Ad vector transduced the basolateral surface with greater efficacy than the apical surface. Consistent with previous observations, EGTA was effective at increasing transduction at the apical surface (7). In addition, each gel formulation enhanced transduction from the apical surface.
EGTA is known to disrupt tight junctions and allow vector access to basolateral localized receptors (25, 26). To determine if the viscoelastic gels also disrupt tight junctions, formulations identical to those used in the in vivo experiment were applied to the apical surface of NuLi cells (Figure 4B). The viscoelastic gels did not significantly affect the transepithelial resistance. In contrast, vector formulated with EGTA transiently lowered the resistance. Taken together, these data suggest that the gel formulations enhance adenoviral gene transfer by increasing the vector residence time at the cell surface leading to receptor-independent uptake.
Formulating an AAV in Viscoelastic Gels Enhances Gene Transfer to the Conducting Airways
Formulation of Lentiviral Vectors with Viscoelastic Gels Enhances Gene Transfer to Nasal Epithelia Based on our in vivo and in vitro data, we conclude that viscoelastic gel formulations have no significant deleterious effect on the titer of encapsidated vectors such as adenovirus and AAV. However, stability of enveloped viruses in such formulations has not been assessed. To test if such gels inactivated enveloped viruses, we first incubated each gel with feline immunodefiency virus (FIV)-based vector for 30 min at 37°C, then titered the vector by limiting dilution on HT1080 cells. Importantly, there was no significant drop in vector titer after incubation with any of the gels evaluated (Table 2). This result suggests that these, and perhaps other, gel formulations might be developed to enhance delivery of enveloped viral vectors to airway epithelia.
FIV-based vector expressing luciferase (FIV-Luc) or an Ad5-Luc control were formulated as described in MATERIALS AND METHODS and delivered directly to the nasal epithelia of mice (Figure 6A). The nasal epithelium was selected as a relevant target because cystic fibrosis (CF) mouse models are available that manifest measurable ion transport and other defects at this site (27). Gene transfer efficacy was evaluated 19 d after vector delivery by quantifying bioluminescence as described in MATERIALS AND METHODS. As shown in Figure 6, the MC formulation greatly enhanced FIV gene transfer to the nose of mice in contrast to the neat or EGTA controls. Unexpectedly, pretreating the nasal epithelia with EGTA before delivering FIV-Luc formulated with MC resulted in less transduction than the MC formulation alone. The reasons for this discrepancy are currently unclear and are under further investigation.
A disadvantage of luciferase as a reporter is the difficulty in determining cell types expressing the transgene. It is necessary to determine the transduced cell types in nasal airways (28); therefore, in parallel experiments, FIV expressing nuclear targeted ß-galactosidase (FIV-ßGal) was formulated with MC and delivered to the nasal epithelia of mice (Figure 6B). Three weeks after vector delivery, paraffin-embedded sections of nasal tissues were prepared as described in MATERIALS AND METHODS. Morphologic analysis confirmed ß-galactosidase expression in both respiratory and olfactory epithelia (Figure 6B).
Progress in gene transfer to the airways has resulted from focused studies of both epithelial biology and vector development. Combining improved vectors with modified formulations appears to be a promising strategy for enhancing the efficiency of airway epithelial gene transfer for CF and other airway diseases (2, 3, 5, 6, 8, 16). These studies demonstrate that two simple gel formulations (MC and CMC) dramatically enhance the efficiency of gene transfer to the airways. Importantly, by formulating Ad5 and AAV5 vectors with viscoelastic gels, gene transfer is concentrated in the conducting airways, with limited gene transfer to the alveoli. This is a novel result with important implications for future CF applications, as the cells of the conducting airway epithelium are a key target population for correction. In contrast, it is believed that gene transfer to the alveolar epithelium may not be required for prevention or treatment of CF lung disease. By restricting gene transfer to the airway epithelium and avoiding delivery to the large alveolar surface area, we also effectively increase the vector MOI delivered to the target cells compared with simple water-based vehicles. Rat lungs have been reported to have 5 x 107 total epithelial cells from trachea to respiratory bronchioles and 9 x 108 alveolar epithelial cells (29). Estimating 109 total cells/gram (30), 3.57% of a 1.4-g rat lung is composed of airway epithelial cells and 64.3% is alveolar epithelia. Assuming such ratios hold true for mice, a 230-mg mouse lung contains 8.2 x 106 epithelial cells from the trachea to the respiratory bronchioles and 1.5 x 108 alveolar epithelial cells. We delivered 2.5 x 108 pfu of Ad5 vector to the lungs. Assuming an even distribution of vector across the entire airway and alveolar surface area, we achieved an MOI of 1.6. Restriction of the vector to the conducting airways with viscoelastic gel increases the effective MOI to 30.5. This 20-fold increase in MOI is consistent with the significant increase in gene transfer of the conducing airways observed with viscoelastic gel formulated vector. We presumed that gel formulations of adenoviral vectors enhanced gene transfer via CAR-independent mechanisms (24). The receptor for AAV5, PDGFR (31), appears to be functionally present on the apical surface of airway epithelia (13). Likewise, GP64 pseudotyped FIV preferentially transduces airway epithelia from the apical surface (P.L.S. and P.B.M., unpublished observations). The finding that CMC and MC gel formulations substantially enhanced AAV5 and GP64 pseudotyped FIV gene transfer is consistent with the idea that improving the vector residence time by inhibiting mucociliary clearance increases the probability for receptor binding and uptake.
Based on the cow tracheal explant model, the concentrations of viscoelastic gels used in this study transiently decreased mucociliary transport rate by < 50% (Table 1). Higher gel concentrations may be more effective at enhancing gene transfer to conducting airways. Indeed, 2% methylcellulose gels can decrease mucociliary transport rate by Importantly, all of the gel-forming polymers tested in these studies have been shown to be safe for use in humans (32–34). MC, CMC, and P407 are present in numerous FDA-approved commercial drug products that are administered topically for cutaneous, ophthalmic, oral, or nasal applications. Although further testing of the safety of these and other gels in the lungs is important, preliminary examinations (35) of the respiratory tissues treated with MC or CMC do not show any evidence of local toxicity, and the safety history of these gel-forming polymers suggests that these are promising formulations for clinical use. Interestingly, MC formulations increased the transduction efficiency for both enveloped and nonenveloped viral vectors. Although not a subject of our studies, gel formulations might also increase the transduction efficiency of nonviral vectors such as liposome/DNA complexes or naked DNA. These studies indicate that viscoelastic gel formulations are promising vehicles for several classes of gene transfer vectors.
The authors are grateful for the contributions of Micheal Henry, Haley Sinn, Beverly Davidson, Melissa Hickey, and Christine Rowley. They acknowledge the assistance of the DNA Sequencing Core, Cell Morphology Core, Cell Culture Core, and the Gene Transfer Vector Core partially supported by the Cystic Fibrosis Foundation, NHLBI (PPG HL-51670), and the Center for Gene Therapy for Cystic Fibrosis (NIH P30 DK-54759).
This work was supported by the Cystic Fibrosis Foundation SINN04G0 (P.L.S.), NIH RO1 HL-61460 (P.B.M), and PPG HL-51670 (P.B.M.). Conflict of Interest Statement: P.L.S has no declared conflicts of interest; A.J.S. has no declared conflicts of interest; M.D.D. has no declared conflicts of interest; and P.B.M. has no declared conflicts of interest. Received in original form December 22, 2004
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Top
Abstract
Introduction
x cm2 confirmed by transepithelial resistance measurements using a volt-ohm meter (World Precision Instruments, Sarasota, FL). The volt-ohm meter was also used to measure transepithelial resistance at indicated time points after addition of viscoelastic gels to NuLi cells. 
