Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Vector-mediated gene transfer has the potential to introduce genes into the airway epithelium to correct cystic fibrosis (CF) and other genetic lung diseases (1, 2). Recently, investigators noted that gene transfer with several recombinant viruses, including retrovirus (3), adeno-associated virus (6, 7), and adenoviruses (8, 9) was very inefficient when the vectors were applied to the apical surface of airway epithelia. Gene transfer efficiency increased greatly when the vectors were applied to the basolateral surface, indicating that their receptors are polarized to the basolateral membrane. This distribution of viral receptors presents a major barrier to gene transfer when the vectors are delivered to the apical surface. For gene transfer to progress as a therapy for lung diseases, approaches must be developed to overcome this barrier.

We previously reported that the application of amphotropic murine leukemia virus (MuLV) vectors to the basal surface, but not apical surface, transduced differentiated airway epithelia in vitro (3, 5). Retroviral-mediated gene transfer from the apical surface was greatly enhanced by formulating the vector in a hypotonic buffer with ethyleneglycol-bis-(2-aminoethylether)-N,N,N',N'-tetraacetic acid (EGTA) (3). However, the mechanism by which formulating vectors with Ca²⁺ chelators in a hypotonic buffer facilitates retroviral gene transfer to epithelia was unclear. Hypotonic solutions and Ca²⁺ chelators have several effects on epithelia. Exposure of airway epithelia to hypotonic conditions causes a rapid decrease in transepithelial resistance and increases paracellular and transcellular permeability to macromolecules (10). Extracellular Ca²⁺ is very important in maintaining the integrity of the epithelium and is required for the formation and stabilization of intercellular junctions (11). Therefore, these conditions could increase the permeability of airway epithelia to viral vectors through either the paracellular or the transcellular routes.

Modulating epithelial junction permeability to enhance gene transfer is an approach that has received little attention. In this article, we studied the effects of Ca²⁺ chelators on viral vector-mediated gene transfer. Formulating vectors with Ca²⁺ chelators in a hypotonic buffer enhanced access to the basolateral membrane and greatly improved gene transfer in vitro and in vivo. Identical interventions resulted in disruption of the junctional complex in human nasal epithelia in vivo. These results suggest that vector-formulation approaches enhance gene transfer to differentiated airway epithelia and may have applications for future human gene therapy trials.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Culture of Human Airway Epithelia

Human airway epithelial cells were obtained from surgical polypectomies or from trachea and bronchi of lungs removed for organ donation. Cells were isolated by enzyme digestion as previously described (12, 13). Freshly isolated cells were seeded at a density of 5 × 10⁵ cells/cm² onto collagen-coated, 0.6 cm² Millicell polycarbonate filters (Millipore Corp., Bedford, MA). The cells were maintained at 37°C in a humidified atmosphere of 7% CO₂ air. Twenty-four hours after plating, the mucosal media were removed, and the cells were allowed to grow at the air-liquid interface (12). The culture medium consisted of a 1:1 ratio mix of Dulbecco's modified Eagle's medium (DMEM)/Ham's F12, 5% Ultroser G (Biosepra SA, Cedex, France), 100 U/ ml penicillin, 100 µg/ml streptomycin, 1% nonessential amino acids, and 0.12 U/ml insulin. Epithelia were tested for transepithelial resistance and for morphology by scanning electron microscopy. Studies were performed on differentiated cells ( 2 wk old). This study was approved by the Institutional Review Board at the University of Iowa.

Reagents

Chemicals. All chemicals were of reagent-grade quality. The sources of the calcium chelators used were as follows: EGTA (#S311-500; Fisher Scientific, Fair Lawn, NJ), ethylenediaminetetraacetic acid (EDTA) (#0732-50G; Ambresco, Solon, OH), and 1,2-bis (2-aminophenoxy)-ethane- N,N,N',N'-tetraacetic acid (BAPTA) (#A4926; Sigma, St. Louis, MO).

Growth factors. To stimulate cell proliferation in studies using MuLV, recombinant human keratinocyte growth factor (KGF; Chiron Technologies-Center for Gene Therapy, Inc., San Diego, CA) was applied to the basal medium of differentiated airway cells at a concentration of 100 ng/ml, as described previously (3). After 24 h of KGF stimulation, differentiated airway cells were subjected to different conditions.

Recombinant viral vectors. Two amphotropic-enveloped MuLV vectors were used in these studies. D17 amphotropic (DA)--galactosidase (gal) expressing a cytoplasmic gal reporter was prepared at Chiron Technologies-Center for Gene Therapy, Inc., as described previously (14). The packaging cell line producing amphotropic nuclear-targeted -gal retrovirus (TA-7gal) was provided by Francois-Loic Cosset, and the virus was prepared in the University of Iowa Gene Transfer Vector Core (15). Titers for the TA-7gal vectors were typically 1 to 5 × 10⁸ colony-forming units (cfu)/ml by blue cell counts on 3T3 cells, whereas titers for the DA-gal vector were 5 × 10⁸ cfu/ml. All the reporter genes were driven by the LTR promoter. The retroviral vector was resuspended in 19.5 mM trimethamine (pH 7.4), 37.5 mM NaCl, and 40 mg/ml lactose. The osmolality of the viral buffer was 105 mmol/kg, measured using a vapor pressure osmometer (Model 5500, Wescor, Inc., Logan, UT). Polybrene was included in all in vitro infection mixtures at a final concentration of 8 µg/ ml. Adenoviral vectors expressing nuclear-targeted -gal (Ad5CMVntgal or Ad5RSVntgal) were produced in the Gene Transfer Vector Core.

In Vitro Transduction of Epithelial Cells with Recombinant Retroviruses

Fully differentiated airway epithelia were stimulated to divide using 100 ng/ml KGF. Twenty-four hours later, the retrovirus (multiplicity of infection [MOI] ~ 20) was combined with hypotonic (osmolality, ~ 40 mmol/kg) or isotonic (osmolality, ~ 220 mmol/kg) solutions containing EGTA, EDTA, or BAPTA. N-2-hydroxyethylpiperazine- N'-ethane sulfonic acid (Hepes) was added to all solutions at a concentration of 10 mM to maintain a pH of 7.4. The final concentration of each chelator was specified in individual experiments. Hereafter, we refer to the Hepes-buffered hypotonic solution as hypotonic buffer and the Hepes-buffered isotonic solution as isotonic buffer. The viral solutions were applied to the apical surface of the epithelia and incubated for different times as described in RESULTS. The transduced epithelia were cultured for 3 d. Transgene expression was detected using the 5-bromo-4-chloro-3-indolyl--D-galactopyranoside (X-gal) histochemical method.

Histochemistry

Epithelial cells were fixed with 2% paraformaldehyde/ phosphate-buffered saline (PBS) solution for 20 min and were rinsed with PBS twice for 10 min each. The X-gal solution was applied, and the cells were incubated at 37°C for 4 h to overnight, as previously described (3). Cells grown on filter membranes were examined microscopically en face for -gal expression. The epithelial cells were counterstained with 4,6-diamidino-2-phenylindole (DAPI) to identify nuclei. For each epithelia, over 1,000 total cells were scanned, and the percentage of blue cells was calculated.

Rabbit airway and lung tissues were fixed in 2% paraformaldehyde/PBS overnight. After X-gal staining and en face examination, tissues were embedded in paraffin, and the sections were cut for histologic examination. Sections were counterstained with nuclear fast red.

Measurement of Transepithelial Resistance In Vitro

Differentiated epithelial cells were treated with 100 ng/ml KGF for 24 h. Different concentrations of each calcium chelator in hypotonic or isotonic buffer were mixed with vector buffer and applied to the apical surface for as many times as specified (3). Transepithelial resistance (RT) was measured with an ohmmeter (EVOM; World Precision Instruments, Inc., Sarasota, FL) by adding cell culture media to the apical surface, and the values were compared with untreated controls.

Confocal Microscopy

Ten micromoles of 5-chloromethyl-fluorescein diacetate (CMFDA; Molecular Probes, Eugene, OR) were added to the basal medium overnight to label cells (16, 17). Latex beads, 100 nm in size (1 × 10⁹ particles; Sigma, St. Louis, MO) were suspended in hypotonic buffer with 6 mM EGTA. The suspension was applied to the apical surface of KGF-stimulated epithelia for 1 h. The epithelia were then washed with PBS 4 to 5 times for 20 min and then fixed in 4% paraformaldehyde/PBS for 20 min. The cell filter membranes were removed, mounted on slides, and examined under a confocal microscope (MRC 1024; Bio-Rad, Richmond, CA).

Electron Microscopy

Transmission electron microscopy was used to study cell morphology after treatment with 6 mM EGTA in hypotonic buffer for 1 h. Differentiated human airway epithelia were fixed in 2.5% glutaraldehyde (0.1 M sodium cacodylate buffer, pH 7.4) overnight at 4°C and then postfixed with 1% osmium tetroxide for 1 h. After serial alcohol dehydration, samples were embedded in Eponate 12 (Ted Pella, Inc., Redding, CA). Sectioning and poststaining were performed using routine methods (18). Samples were examined under a Hitachi H-7000 transmission electron microscope. Epithelia from two different human specimens were examined, and grids from each preparation were studied. A minimum of 100 intercellular junctions were examined in each control or EGTA-treated preparation. Occasionally, cells were observed to loosen from the membranes in the EGTA-treated cells, but cell morphology was well preserved.

Disruption of Tight Junctions in the Airway Epithelium In Vivo

Rabbit studies. The transepithelial electrical potential difference across the rabbit tracheal epithelium (VT) was measured using a modification of previously described methods (19). The reference electrode was a subcutaneous needle connected with sterile normal saline solution to a silver/ silver chloride pellet. Adult New Zealand white rabbits (n = 4 per group) were anesthetized with 32 mg/kg ketamine, 5.1 mg/kg xylazine, and 0.8 mg/kg acepromazine intramuscularly and placed with the head in the dependent position. An exploring electrode (pediatric size 8, Teflon-coated latex double lumen Foley catheter, modified; Rüsch, Inc., Duluth, GA) was inserted into the tracheal lumen via a tracheotomy incision. We filled one lumen of the catheter with sterile normal saline solution and connected it to a silver/silver chloride pellet. Voltage was measured with a voltmeter connected to a strip chart recorder. The catheter was left in that position for the entire recording period. The solutions were administered through the second lumen in the following order: (1) normal saline for 5 min; (2) hypotonic buffer with 10 mM EGTA for 10 min; (3) normal saline with 10 mM EGTA for 5 min; and (4) normal saline with 10 mM EGTA plus 100 µM amiloride for 5 min. Measurements of VT were read by two investigators who were blinded to the treatment. No discrepancies were found in readings of greater than 2 mV between the investigators. Disruption of epithelial junctions was assessed by detecting a fall in nasal VT after perfusion with the chelator and by the loss of amiloride-sensitive VT after perfusion with the chelator solution.

Human studies. We measured the VT across the nasal epithelium using previously described methods (19) in six volunteers. The protocol followed was identical to that described previously for the rabbit studies. The rubber catheter was introduced into the nostril under telescopic guidance, and the side holes of the catheter were placed under the inferior nasal turbinate 6 cm from the most caudal aspect of the columella. The catheter was left in that position for the entire recording period. We administered the solutions through the second lumen in the following order: (1) normal saline for 5 min; (2) hypotonic buffer with 10 mM EGTA or 10 mM EDTA for 10 min; (3) normal saline with 10 mM EGTA or 10 mM EDTA for 5 min; and (4) normal saline with 10 mM EGTA or 10 mM EDTA plus 100 µM amiloride for 5 min. Measurements of VT were read by two investigators who were blinded. These measurements were then checked by a third investigator if there was a discrepancy of more than 2 mV. Disruption of epithelial junctions was assessed by detecting a fall in nasal VT after perfusion with the chelator and by the loss of amiloride-sensitive VT after perfusion with the chelator solution. The study was approved by the Institutional Review Board of the University of Iowa.

Gene Transfer to the Rabbit Tracheal Epithelium In Vivo

Adult New Zealand white rabbits were anesthetized with 32 mg/kg ketamine, 5.1 mg/kg xylazine, and 0.8 mg/kg acepromazine intramuscularly. Using the sterile technique, a ventral midline incision was made, exposing the trachea for tracheotomy. An ~ 1.5-cm tracheal segment cephalad to the tracheotomy was isolated and cannulated on each end with PE 330 tubing (Clay Adams, Becton Dickinson). The tracheal segment was first rinsed and then filled with hypotonic buffer mixed with 12 mM EGTA for 45 min. The EGTA solution was removed and replaced with 300 µl of retrovirus (titer 3 × 10⁹ cfu/ml) or adenovirus (titer 1 × 10¹⁰ plaque-forming units [pfu]/ml) vector. In control animals, saline was substituted for the hypotonic-buffered EGTA treatment. Animals that received the MuLV retrovirus were pretreated with KGF prior to gene transfer to stimulate epithelial proliferation (~ 5 mg/kg, intravenously, and 600 µg via nasal instillation for three doses given 36, 24, and 12 h preceding the procedure) (3, 5). The vector solution was left in place for an additional 45 min, and then the cannulae were removed, the incisions were closed, and animals were allowed to recover. Three days later, animals were killed, and the tissues were studied for -gal expression. The study was approved by the Animal Care and Use Committee at the University of Iowa.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Calcium Chelation Opens Epithelial Junctions

In previous reports, we documented that retrovirus-mediated gene transfer (3, 5) and adenoviral-mediated gene transfer (9) to differentiated airway epithelia occurred preferentially from the basolateral surface. We hypothesized that Ca²⁺ chelation enhanced gene transfer from the apical surface by opening epithelial junctions and allowing vector access to the basolateral membrane. To address this hypothesis, we used fluorescent latex beads (100 nm in size) as a marker for the virus to follow particle deposition in the presence of a Ca²⁺ chelator. Fluorescent beads suspended in a hypotonic 6-mM EGTA solution were applied to the apical side of differentiated human airway cells for 1 h. After washing, the epithelial cells were fixed and examined. In control cells without EGTA treatment, no beads were observed in the paracellular space or within the cell cytoplasm, but they were noted on the apical surface (Figures 1A and 1B). As shown in Figures 1C and 1D, EGTA treatment caused a widening of the intercellular spaces and facilitated deposition of beads (red) in the paracellular space. These results show that Ca²⁺ chelation opens epithelial junctions and allows 100-nm-sized particles access to the basolateral surface.

View larger version (112K): [in this window] [in a new window]   View larger version (153K): [in this window] [in a new window]   Figure 1.   Calcium chelation opens epithelial junctions and allows vector access to the paracellular space. Well-differentiated airway cells were pretreated with 100 ng/ml KGF for 24 h and labeled with vital cell dye for 12 h. Fluorescent latex beads, 100 nm in size (109 particles), were combined with 6 mM EGTA in hypotonic buffer as a marker for a viral vector. After application to the apical surface for 1 h, the beads were washed off, and the epithelia were fixed and examined using confocal microscopy. A and B represent images from control cells. C and D show images from EGTA-treated cells. A representative image from the middle of the Z series shows paracellular deposition of beads (C ). For control cells, the fluorescent beads were suspended in PBS (A). Z-series images were collected every 0.5 µm. Vertical section images were reconstructed (B, control; D, EGTA). (E-F ) Transmission electron microscopic observation of the intercellular junctional complex after EGTA treatment. Intercellular gaps between two control cells (E ) and EGTA-treated airway epithelial cells (F ). Arrow, adherens junction; arrowheads, intercellular gap. EGTA-treated cells show markedly widened paracellular spaces.

Trypan blue dye exclusion was used to screen for toxicity of hypotonic conditions and Ca²⁺ chelators on airway epithelia by exposing cells to 10 mM Hepes buffer alone or 6 mM EGTA in 10 mM Hepes buffer for 4 h. Only rare blue cells (< 1%) were observed in the hypotonic buffer- treated group, whereas 3.9% of cells exposed to EGTA in hypotonic buffer took up the dye (mean value, n = 3 epithelia/condition). Thus, these conditions do not appear to be injurious to the cells.

Transmission electron microscopy was used to directly examine the effect of Ca²⁺ chelation and hypotonic conditions on the epithelial junctional complex. As shown in Figure 1F, EGTA-treated airway cells developed a widened intercellular gap. The space, estimated to be ~ 500 nm across, was approximately five times larger than a retroviral or adenoviral particle. This effect was noted in all the junctions studied in the EGTA treatment group. In contrast, control epithelia maintained tight intercellular junctions with no visible gap (Figure 1E). These data suggest that vector formulation with a hypotonic buffer and a Ca²⁺ chelator allows viral particles to percolate through the intercellular space and transduce airway epithelia at the susceptible basolateral membrane.

Vector-Mediated Gene Transfer to Airway Cells In Vitro Is Enhanced by Calcium Chelation

We evaluated the ability of three different chelators to enhance gene transfer to airway cells. EGTA has a higher affinity for Ca²⁺ over Mg²⁺ (> 10⁵-fold), and its Ca²⁺-binding efficiency is pH-sensitive. In contrast, BAPTA binds Ca²⁺ in an acidic environment and has faster Ca²⁺ binding and release kinetics. In contrast to BAPTA and EGTA, EDTA has a broader spectrum of divalent cation binding. The preference of EDTA for Ca²⁺ is 100-fold greater than that for Mg²⁺. Vector solutions were formulated with 6-mM final concentrations of the chelators in a hypotonic buffer and then applied to the apical surface for 4 h. Gene transfer was examined 3 d later. In control cells, the vectors were applied in an isotonic buffer without a Ca²⁺ chelator. In the absence of Ca²⁺ chelation, no gene transfer occurred for either retrovirus (amphotropic MuLV) or adenovirus vectors (Figures 2A and 2B). All three chelators similarly enhanced gene transfer with retroviral vectors from the apical surface (Figures 2C, 2E, and 2G). Also, gene transfer with adenoviral vectors was enhanced by formulation with Ca²⁺ chelators (Figures 2D, 2F, and 2H). For MuLV, the efficiency of gene transfer declined as the chelator concentration was increased above 6 mM (data not shown). Thus, hypotonic formulation and Ca²⁺ chelation greatly enhances gene transfer when viral vectors are applied to the apical surface.

View larger version (118K): [in this window] [in a new window]   Figure 2.   Gene transfer to differentiated airway epithelia with retrovirus (left panels) and adenovirus (right panels) is enhanced by Ca2+ chelation. For retrovirus experiments, well-differentiated human airway epithelial cells were pretreated with 100 ng/ ml KGF 24 h prior to gene transfer. TA-7gal vector (MOI ~ 20) was combined with the specified Ca2+ chelator (6 mM) in hypotonic buffer. The mixture was then applied to the apical surface of airway epithelia for 4 h. Three days later, transgene expression was detected by X-gal histochemistry. Representative en face views of epithelia stained with X-gal are shown for each condition: A and B, no calcium chelator; C and D, EGTA; E and F, EDTA; and G and H, BAPTA. For adenoviral gene transfer, AdCMVgal (MOI ~ 20) was added to the apical surface of the differentiated epithelia using identical conditions as those for retrovirus, except that KGF pretreatment was omitted.

To learn whether the effects of Ca²⁺ chelation and hypotonic buffers on enhancing gene transfer were specific to apical application, we added adenovirus formulated with 6 mM EGTA in hypotonic buffer to the apical or basolateral surfaces of airway epithelia (MOI 20). Vector formulation in EGTA with hypotonic buffer greatly enhanced apical gene transfer over the vector in normotonic buffer without EGTA (15% -gal-positive cells versus 0% positive, respectively). In contrast, for basolateral application the gene transfer rates for adenovirus formulated in normotonic conditions without EGTA were identical to those for vector formulated 6 mM EGTA in hypotonic buffer (8% -gal-positive cells).

Calcium Chelation under Hypotonic Conditions Enhances Gene Transfer More Quickly Than Isotonic Conditions

To understand whether the tonicity of the vector solution affects epithelial permeability and gene transfer efficiency, hypotonic (osmolality, 40 mmol/kg)- or isotonic (osmolality, 220 mmol/kg)-buffered solutions of 6 mM EGTA, EDTA, or BAPTA were added to the MuLV vector buffer and applied to the apical surface of differentiated airway epithelial cells pretreated with KGF. At specific time intervals, RT was measured using an ohmmeter. As shown in Figures 3A and 3B, all three hypotonic solutions reduced RT to  20% of control values within the initial 10 min. Isotonic solutions showed more variable results. Sixty minutes were required for isotonic EGTA or EDTA to decrease RT to ~ 20% of the controls. In contrast, isotonic BAPTA quickly dropped RT, similar to that of the hypotonic solutions. The time course of RT recovery following a 4-h treatment with 0, 6, or 12 mM EGTA in hypotonic buffer was also studied. RT recovered to baseline within 12, 48, or 72 h, respectively (n = 3 epithelia/time point; data not shown).

View larger version (47K): [in this window] [in a new window]   Figure 3.   The effects of EGTA, EDTA, and BAPTA on RT and gene transfer. Well-differentiated airway epithelia were pretreated with 100 ng/ml KGF for 24 h. Hypotonic (A) or isotonic (B) solutions with 6 mM of the indicated Ca2+ chelators were applied to the apical surface. RT was measured with an ohmmeter and normalized to 100% of control values. For gene transfer experiments, 6 mM (final concentration) hypotonic or isotonic EGTA (C ) or BAPTA (D) was formulated with the TA-7gal vector. The incubation times are indicated on the x-axis. The y-axis shows the percentage of blue cells. Values shown are mean ± SE (n = 3 epithelia/time point; *P < 0.05 by t test for the treatment group versus control at each time point).

Next we evaluated the time course of retroviral gene transfer with EGTA and BAPTA. We focused this study on EGTA and BAPTA because of their specificity for Ca²⁺ over other divalent cations. In the presence of either 6 mM EGTA or 6 mM BAPTA, airway cells were transduced for periods of 15 min to 3 h before removing the vector solution. Three days later, transgene expression was assessed by X-gal histochemistry and quantified by blue cell counting. Figures 3C and 3D show that maximal gene transfer with hypotonic preparations occurred by 1 h. This effect was most apparent with BAPTA. In contrast, greater than 1 h was required for maximal gene transfer with isotonic vector solutions. These data show that the prompt fall in RT noted with hypotonic chelation conditions is associated with gene transfer, requiring a shorter vector incubation period. On the basis of these results, we elected to use the combined conditions of a hypotonic solution and a Ca²⁺ chelator in vivo.

EGTA Treatment Disrupts the Transepithelial Voltage and Amiloride-Sensitivity in Rabbit Tracheal Epithelia and Human Nasal Epithelia In Vivo

Our in vitro data show that treatment of epithelia with a Ca²⁺ chelator in a hypotonic buffer opens epithelial junctions and enhances gene transfer to differentiated human airway epithelia. We wondered if such a formulation strategy could also be applied in vivo. It is not technically feasible to measure RT in vivo. However, measurements of transepithelial voltage (VT) can provide an assessment of RT because when RT falls to zero, VT will also decrease to zero. We first measured the bioelectric responses of the tracheal epithelium in anesthetized rabbits to perfusion with the hypotonic-buffered EGTA solution and saline containing EGTA. As shown in Figure 4A, perfusion of the trachea with the hypotonic solution containing 10 mM EGTA and subsequent treatment with saline and 10 mM EGTA caused a significant fall in VT. After treatment with EGTA in the hypotonic buffer, amiloride had no effect on VT. In contrast, control tracheas perfused with saline exhibited stable baseline voltages and the anticipated fall in VT in response to perfusion with amiloride. The simplest interpretation of these data is that treatment with hypotonic-buffered EGTA opened epithelial junctions in vivo, by reducing RT, and thereby VT.

View larger version (34K): [in this window] [in a new window]   Figure 4.   Perfusion of the rabbit tracheal epithelia and human nasal epithelia with hypotonic buffered EGTA solution inhibits baseline VT and amiloride sensitive VT in vivo. (A) The tracheal transepithelial voltage was measured in anesthetized rabbits via tracheostomy as described in MATERIALS AND METHODS. Left panel: in control animals, the baseline voltage remained stable with saline perfusion. Addition of amiloride caused a significant reduction in VT. Sequential treatment with hypotonic buffer perfusion followed by perfusion with 10 mM EGTA in saline ("intervention") caused a reduction in VT. After EGTA treatment, there was no residual amiloride-sensitive VT. Right panel: comparison of the amiloride-sensitive VT in control and EGTA- treated animals (n = 4 rabbits each for EGTA and control conditions; *P < 0.05). (B) Perfusion of the human nasal respiratory epithelium with hypotonic buffered EGTA solution inhibits baseline VT and amiloride-sensitive VT. Left panel: the nasal epithelium of normal volunteers was sequentially perfused with EGTA in hypotonic buffer, EGTA in saline, and then EGTA in saline with amiloride. The nasal transepithelial potential was measured continuously. In control experiments, the same protocol was performed in the same subjects on another day, omitting EGTA from the perfusate. Under EGTA conditions ("intervention"), there was a significant, reversible drop in the VT and the amiloride-sensitive VT. Right panel: net change in VT in response to amiloride for control and EGTA-treated subjects. EGTA treatment abolished the amiloride-sensitive VT (n = 6 for each condition; *P < 0.01).

The in vitro results in cultured airway epithelia and in vivo studies in rabbits suggest that it is possible to increase epithelial permeability by combined hypotonic/EGTA treatment. Such vector formulation strategies may be useful for enhancing gene transfer to epithelia for the treatment of diseases in humans. To learn if Ca²⁺ chelation also opens epithelial junctions in human airway epithelia in vivo, we performed studies using the nasal epithelium as a model. The baseline nasal VT was recorded; then the epithelium was sequentially perfused with EGTA hypotonic buffer, EGTA in saline, followed by EGTA in saline with amiloride. The nasal voltage was measured throughout the study. In control experiments, the same protocol was performed on a different day in the same subjects, omitting the water treatment and EGTA from the perfusate. Subjects did not experience any adverse effects and could not tell the difference between perfusion with the EGTA solution or saline.

As shown in Figure 4B, when subjects received the hypotonic-buffered EGTA treatment followed by EGTA in saline perfusion, there was a significant decrease in the VT. After EGTA treatment, essentially no amiloride-sensitive VT remained. In control experiments, VT remained stable with solution changes, and there was a significant fall in VT with amiloride treatment (Figure 4B, left panel ). In contrast, EGTA treatment caused a significant fall in nasal VT and loss of amiloride-sensitive VT compared with control conditions. Similar results were obtained when EDTA was substituted for EGTA in the perfusion protocol (data not shown). Thus, treatment with Ca²⁺ chelators increased human nasal epithelial permeability, consistent with the in vitro findings.

Vector Formulation with a Hypotonic Buffer and EGTA Enhances Gene Transfer In Vivo

Next we tested whether conditions that appeared to open epithelial junctions in vivo also facilitated gene transfer. After pretreatment with 12 mM EGTA in hypotonic buffer, a retroviral or adenoviral vector was applied to the luminal surface of the trachea in anesthetized normal adult rabbits. Animals receiving MuLV were pretreated with KGF to stimulate epithelial proliferation (see MATERIALS AND METHODS and References 3, 5, 20-22). Three days later, the tissues were removed and studied for -gal expression. As shown in Figure 5, both vectors transduced cells throughout the tracheal epithelium, including nonciliated and ciliated surface cells, intermediate cells, and basal cells. The number of -gal-expressing cells was qualitatively greater in the adenovirus-treated rabbits (Figures 5E and 5F) than for those that received retrovirus (Figures 5B and 5C). This result may reflect, in part, the titer dependence of gene transfer and the requirement for cell division by MuLV-based vectors. Tracheas that received either viral vector without EGTA or hypotonic buffer treatment showed no evidence of gene transfer (data not shown).

View larger version (147K): [in this window] [in a new window]   Figure 5.   In vivo gene transfer is enhanced by treatment with EGTA in a hypotonic buffer. Tissues were pretreated with 12 mM EGTA in hypotonic buffer for 45 min. Gene transfer was then performed with retroviral (left panels) and adenoviral (right panels) vectors to the rabbit tracheal epithelium in vivo as described in MATERIALS AND METHODS. (A) En face view of trachea following X-gal staining from an animal that received retrovirus formulated in hypotonic/EGTA buffer. -gal-positive cells are scattered across the epithelial surface (n = 3 animals). Control animals treated with retrovirus in control buffer show no -gal expression (n = 1 animal). (B and C ) Cross-sections of trachea from conditions shown in A, revealing transduction of cells throughout the epithelium. NC = nonciliated cell. (D) En face view of trachea after X-gal staining from an animal that received adenovirus formulated in hypotonic/EGTA buffer. -Gal- positive cells are frequently seen across the epithelial surface (n = 3 animals). Control animals treated with adenovirus in control buffer show no -gal expression (n = 1 animal). (E and F ) Cross-sections of trachea from conditions shown in D reveal that cells throughout the epithelium were transduced. When retroviral or adenoviral vectors were preformulated in 12 mM EGTA and hypotonic buffer and applied to the tissues without pretreatment, similar gene transfer was noted (data not shown).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

For gene transfer to progress as a therapy for lung diseases (such as CF), the transduction efficiency must be improved. All the vectors currently in use transduce differentiated airway epithelia inefficiently when applied to the apical surface (3, 6, 8, 9, 20, 23). Modification of viral envelopes or capsids to target receptors on the apical surface of airway cells could improve efficiency. We explored an alternative strategy of enhancing vector access to their native receptors on the basolateral membrane. In this study, we showed that vector formulation with a Ca²⁺ chelator in a hypotonic buffer opens epithelial junctions and facilitates gene transfer from the apical surface in vitro and in vivo. This novel technique overcomes the inefficiency of gene transfer from the apical surface for viral vectors and may be applicable for vector or drug delivery to epithelia.

A complex of specialized intercellular junctions enables epithelial cells to form an electrically tight sheet, maintain cell polarity, and separate the apical and basal compartments. Of the four types of junctions recognized (tight junctions, adherens junctions, desmosomes, and gap junctions [26]), the function of three are clearly calcium-dependent. Tight junctions, the outermost components of the junctional complex, form a continuous point of contact with neighboring cells. Three proteins, occludin, claudin-1, and claudin-2, have been identified in the tight-junction complex (27). A large body of evidence shows that Ca²⁺ is important in the synthesis and maintenance of epithelial tight junctions (11, 26, 30). Adherens junctions, located below the tight junction in the junctional complex, also maintain a circumferential contact zone. Interactions between the extracellular domains of cadherins form the adhesive molecular contacts in the adherens junctions (31). The homotypic interactions of E-cadherin, the initial event in junctional complex formation, is Ca²⁺-dependent (32). Desmosomes contain two types of glycoproteins, the desmogleins and desmocollins; both are Ca²⁺-binding, cadherin-like molecules important in cell-cell adhesion (33). Therefore, calcium is important in the formation and maintenance of the entire epithelial junction complex. By depleting extracellular calcium, Ca²⁺ chelators open the junctional complex in human airway epithelia.

Perfusion of the rabbit tracheal epithelium or the human nasal epithelium with a hypotonic buffer and a Ca²⁺ chelator caused a fall in the VT and a loss of the amiloride-sensitive voltage (Figure 4). In control experiments without Ca²⁺ chelators, the basal voltages remained stable, and amiloride caused the expected fall in voltage. These results parallel those found in primary cultures of human airway epithelia (Figure 2 and [3, 9]) and those published by others (34, 35), and are consistent with the reversible disruption of the epithelial junctional complex. Importantly, when a similar vector formulation protocol was applied to gene transfer with retroviral or adenoviral vectors, we successfully transduced airway epithelia in the rabbit trachea. In previous experiments with MuLV-based vectors, it has proven difficult to achieve gene transfer to the airway epithelium, even after stimulating cells to divide with growth factors (5, 20). The findings are exciting because they show that it is possible to translate these in vitro observations to an in vivo application.

The approaches outlined here may potentially be used in the clinic to open epithelial junctions transiently in vivo and facilitate gene transfer to airway epithelia. Calcium chelators such as EDTA are widely used for several indications (36) and could be applied to vector solutions to enhance delivery to epithelia. In one study, EDTA was aerosolized to the human lung for over a three-month period to assess its potential antibacterial effects in CF patients, and no harmful effects were noted (36). As an alternative to aerosol delivery of a Ca²⁺ chelator, lung lavage could be used to administer the formulated vectors (39, 40). Our studies in the rabbit trachea demonstrate the feasibility of using Ca²⁺ chelation and hypotonic formulation to enhance access of amphotropic retrovirus and serotype 5 adenovirus to their receptors on the basolateral membrane in vivo. The effects of calcium chelation and hypotonic solutions on the junctional complex are rapid and reversible (3, 9).

Persistent expression after gene transfer with integrating vectors requires the transduction of cells with progenitor capacity. Although the progenitor cells in airway epithelia have not been fully defined, intermediate and basal cells are considered to be among them (41, 42). The use of Ca²⁺ chelation to open epithelial junctions could allow the targeting of airway cells with progenitor capacity that are located below the surface layer. Reaching such cells may be required to attain persistent gene transfer. The further development of this methodology and its application in vivo will allow testing of this important hypothesis.