Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Cystic fibrosis (CF) is characterized by chronic sputum production, recurrent infections, and lung destruction (1). Although it is not precisely known how mutation of the CF transmembrane conductance regulator (CFTR) gene leads to the clinical manifestations (2), defective Cl secretion and increased Na⁺ absorption (2, 3) are well documented. Further, these changes in ion transport produce alterations in fluid transport across surface and gland epithelia (4). These resultant alterations in water and salt content of airway surface liquid (ASL) purportedly diminish the activity of bactericidal peptides secreted from the epithelial cells (8) and/or impair mucociliary clearance, thereby promoting recurrent lung infection and inflammation.

Maintenance of the mucociliary clearance requires the coordinate regulation of ciliary motion, ASL depth, and mucin content. The quantity and composition of ASL are controlled by both the surface epithelium and submucosal glands. On the basis of cell-volume estimates (9), it appears that the latter may be a more important source of mucous secretions than goblet cells of the surface epithelium. Secretory tubules of the submucosal glands consist of serous cells in the acini and proximal mucous cells (10). Several lines of evidence suggest that submucosal glands contribute to the pathophysiology of CF lung disease: (1) CFTR is predominantly expressed in the serous cells of the submucosal glands (11, 12); (2) tracheal submucosal gland cultures from patients with CF fail to secrete Cl (12); (3) more than 60% of submucosal gland cultures from non-CF subjects show a baseline secretion, whereas cultures from patients with CF exclusively absorbed fluid (5); and (4) obstruction of submucosal gland ducts is the first pulmonary manifestation in patients with CF, and is followed by marked hyperplasia and hypertrophy (16). The evidence implicating submucosal glands in CF pathogenesis suggests that effective gene therapy for CF lung disease should target these organs. However, though numerous attempts have been made to transfer the CFTR gene to surface airway epithelium (2), little attention has been paid to the submucosal glands. One study showed that low levels of -galactosidase (Gal) expression following intratracheal administration of adenovirus vectors was detectable in submucosal glands (17). Pilewski and colleagues (18), using a xenograft model, also showed that adenovirus instilled into the airway lumen mediated transgene delivery into human airway submucosal glands. However, gland transfection levels were lower than for surface epithelium, and declined markedly with distance from the airway lumen.

Though adenoviral-mediated gene transfection is relatively efficient, the host immune response poses a major problem. Specifically, the viral proteins activate cytotoxicity T lymphocytes that destroy the virus-infected cells (19), thereby terminating gene expression in the lungs of in vivo models examined. Another significant problem is diminished gene transfer upon repeated administration of the adenoviral vectors due to the development of neutralizing antiviral antibodies. While these issues are being addressed by modifying both the vectors and the host immune system, nonviral and nonproteinaceous vectors have been gaining attention as alternative approaches. In particular, cationic lipids have been shown to mediate gene transfer into mammalian cells both in vitro and in vivo (20). However, the efficiency of cationic lipid-mediated gene transfer into submucosal glands has not been determined.

Accordingly, in the present study we have systematically evaluated adenovirus-mediated and cationic lipid- mediated CFTR gene transfer into immortalized CF human bronchial submucosal gland cells. Specifically, we addressed four issues regarding CFTR gene transfer: (1) the relative efficacies and duration of transgene expression of a variety of second-generation adenovirus vectors harboring different enhancers/promoters and polyadenylation signals; (2) the optimal conditions for cationic lipid-mediated gene transfer to submucosal glands; (3) the relative efficiency of adenovirus and cationic-lipid vectors at effecting transfection of submucosal gland cells compared with surface airway epithelial cells; and (4) ability of adenovirus-mediated and cationic lipid-mediated gene transfer to restore CFTR Cl channel activity.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Adenoviral Vector Construction and Preparation

Adenoviral vectors were constructed and prepared as described previously (28, 29). All the vectors were based on an E1-deleted Ad2/E4ORF6 backbone (wild-type E2 and E3, deletion of E4 except ORF6). In Ad2/Gal-4, Ad2 nucleotide sequences between 357 and 3328 were replaced with the cytomegalovirus (CMV) enhancer/promoter, the complementary DNA (cDNA) encoding Gal, and the shortened synthetic bovine growth hormone (BGH) polyadenylation sequence. In Ad2/Gal-E1a and Ad2/Gal-PGK, the CMV enhancer/promoter was replaced with the E1a and phosphoglycerokinase (PGK) promoters, respectively. Ad2/CFTR-5 was constructed in a similar fashion as Ad2/Gal-4 except that the cDNA for Gal was replaced by the human CFTR cDNA.

Cationic Lipid and Plasmid Vector Preparations

Cationic lipid GL-67 was developed and characterized at Genzyme (Cambridge, MA) (27). GL-67 harbors a hydrophobic cholesterol anchor in combination with a polar spermine head group in a T-shape configuration. The plasmid vector pCF1-CFTR, comprised of the CMV enhancer/promoter, a hybrid intron, the human CFTR cDNA, the BGH polyadenylation signal sequence, a pUC origin of replication, and the kanamycin gene, was constructed and produced as reported previously (27, 30). The plasmid vector pCF1-Gal was constructed as pCF1-CFTR except that the human CFTR cDNA was replaced with the cDNA for Gal.

Cell Culture

The tracheobronchial submucosal gland cell line (6CFSMEO) used in this study originated from a CF (F508) patient and was transformed and characterized by Cozens and associates (31). It is not known precisely whether the cell line originated from a serous or mucous phenotype. However, the cell line used in this study expresses antigens that are characteristic of secretory cells (31). The tracheobronchial surface epithelial cell line (CFT1) was generated from a CF (F508) patient and was characterized by Yankaskas and coworkers (32). The cells were cultured essentially as described previously (31, 32). Briefly, submucosal gland cells were seeded onto 12-well cell culture plates at a density of 50,000 cells/cm² and cultured with Dulbecco's modified eagle medium (GIBCO BRL, Gaithersburg, MD) supplemented with 10% fetal bovine serum (FBS; GIBCO BRL). CFT1 cells were seeded onto 12-well cell culture plates at a density of 50,000 cells/cm² and cultured with Ham's F12 medium supplemented with 2% FBS, 5 µg/ml insulin, 3.7 µg/ml endothelial cell growth supplement, 25 ng/ml epidermal growth factor, 30 nM triiodothyronine, 1 µM hydrocortisone, 5 µg/ml transferrin, and 10 ng/ml cholera toxin (GIBCO BRL).

Adenovirus Infection

Cells were infected as described previously (29). In brief, the total number of cells in the wells was estimated. Adenovirus vectors at different multiplicities of infection (MOIs) were added to the cells. After incubation at 37°C for 4 h, the adenoviral suspensions were removed and the cells washed once with regular medium. The cells were assayed for Gal expression or CFTR Cl channel activity 48 h after infection.

Cationic Lipid:Plasmid DNA-Mediated Transfection

Cationic lipid:plasmid DNA (pDNA) complexes were prepared and transfection was performed as reported previously (27). Briefly, cationic lipids were rehydrated in water for 10 min, followed by vortexing at full speed for 2 min. The concentrated lipids were then diluted in Opti-MEM (GIBCO BRL). pDNA was also diluted based on the average molecular weight of a nucleotide (M_r = 330) in Opti-MEM. After the cationic lipids were added to the pDNA, the positively charged cationic lipids and negatively charged pDNA were allowed to interact and form cationic lipid:pDNA complexes at room temperature for 15 min. The cells were incubated with the complexes at 37°C for 4 h. The cells were then washed with Opti-MEM and re-fed with regular medium. Gal expression or CFTR Cl channel activity was assayed 48 h after transfection.

Optimization of the ratios and concentrations of GL-67 and pDNA was carried out using a 96-well assay as described previously (27). In brief, cells were seeded onto 96-well cell culture plates at a density of 7,500 cells/well and allowed to grow to confluence for 5 to 7 d before use. Complexes containing cationic lipids at concentrations ranging from 1.9 to 167 µM and pDNA at concentrations ranging from 1.9 to 240 µM were prepared. Cells were exposed to the complexes (100 µl/well) for 4 h before addition of 50 µl of 30% FBS (GIBCO BRL) in Opti-MEM. At 24 h later, a further 100 µl of 10% FBS in Opti-MEM was added. Following a further 24-h incubation, the cells were assayed for Gal expression.

Measurement of Gal Activity

The cells were assayed for Gal activity using a photometric assay as described previously (27). Briefly, medium was carefully aspirated and cells lysed by the addition of 200 µl lysis buffer (250 mM Tris-HCl and 0.15% Triton X-100, pH 8.0) per square centimeter of growth area. The plates were then frozen in a dry ice/ethanol bath for 5 min and thawed at 37°C for 5 min. The freezing and thawing process was repeated three times. The cell lysates (50 µl) from each well were transferred to a fresh 96-well plate. A total of 150 µl of substrate, chlorophenol red galactopyranoside (1 mg/ml), was added to each well and the plate was read at 580 nm in a microtiter plate reader after the color developed. Gal activity was determined by using a standard curve generated with a Gal standard (Sigma, St. Louis, MO). In some experiments, cells were fixed with 1.8% formaldehyde and 2% glutaraldehyde and then incubated for 6 h with 1 mg/ml X-Gal (5-bromo-4-chloro-3-indonyl--D-galactopyranoside) (Boehringer Mannheim, Indianapolis, IN) in phosphate-buffered saline (pH 7.8). Nuclear-localized blue staining of cells was assessed by light microscopy (27). At least 1,000 cells were counted and the percentage of blue cells was determined.

SPQ Fluorescence Analysis

CFTR Cl channel activity was assessed with the halide-sensitive fluorophore 6-methoxy-N-(3-sulfopropyl)-quinolinium (SPQ) (Molecular Probes, Eugene, OR), as reported previously (33). At 48 h after infection, cells were loaded with SPQ by hypotonic shock for 4 min with 10 mM SPQ in water. The SPQ fluorescence was initially quenched by incubating the cells for up to 30 min in a NaI buffer of the following composition (in mM): 135 NaI, 2.4 K₂HPO₄, 0.6 KH₂PO₄, 1 MgSO₄, 1 CaSO₄, and 10 N-2-hydroxyethylpiperazine-N'-2-ethane sulfonic acid, 10 glucose, pH 7.4. After measuring baseline fluorescence (Fo) for 2 min, the NaI solution was replaced with a solution in which NaI was replaced by NaNO₃. Five minutes later, a cocktail of forskolin (20 µM) and 3-isobutylmethylxanthine (IBMX) (100 µM) was added to stimulate the CFTR Cl channel activity. An increase in halide permeability is reflected by a more rapid increase in SPQ fluorescence. It is the rate of change rather than the absolute change in signal that is the important variable in evaluating anion permeability. Differences in absolute levels reflect quantitative differences between groups in SPQ loading, size of cells, or number of cells studied. The data are presented as means ± standard error of the mean (SEM) of fluorescence at time t minus Fo (the average fluorescence measured in the presence of I for 2 min before ion substitution) and are representative of results obtained under each condition. For each experiment, between 50 and 100 cells were examined on a given day and studies under each condition were repeated on at least 2 d. For each experiment, the responses were compared with those obtained with control or untreated cells. Under the conditions specified above, control or untreated CF cells were unresponsive to added cyclic adenosine monophosphate (cAMP) agonists. There was a broad spectrum in the rate of change in SPQ fluorescence observed with responsive cells. The slope of the response curve in most control or untreated CF epithelial cells was < 0.364. Therefore, we scored cells as responsive if the slope of the response curve, which is indicative of the rate of increase in SPQ fluorescence, was  0.364 after stimulation with cAMP agonists. Because the response was heterogeneous, the data shown are for the 10% of cells in each experiment showing the greatest response. All fields were evaluated, but for clarity of presentation only the top 10% of responders are illustrated in the figures.

Statistical Analysis

Data are expressed as means ± SEM. Statistical analysis was performed using the unpaired Student's t test. A probability (P) value of less than 0.05 was considered to be statistically significant.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Efficiency of Adenovirus-Mediated Gene Transfer

CF surface epithelial (CFT1) cells or CF submucosal gland cells (6CFSMEO) were infected with four different adenovirus vectors harboring different enhancers/promoters and polyadenylation signals at different MOIs. No blue cells were detected by X-Gal staining in naive or Ad2/CFTR-5 (1,000 MOI)-infected cells. In contrast, a significant number of stained cells was observed in submucosal gland cells infected with Ad2/Gal-4 at a MOI of 0.1 (Figure 1a). The percentage of stained cells was MOI-dependent. At a MOI of 100, approximately 90% of the cells were stained blue. However, in submucosal gland cells infected with Ad2/ Gal-E1a or Ad2/Gal-PGK vectors, we detected only positively stained blue cells at higher MOI (> 50). The ranking order for Gal expression with vectors utilizing different promoters/enhancers in CF submucosal gland cells was CMV >> E1a  PGK. In CFT1 cells the ranking order was similar to that obtained in submucosal gland cells (Figure 1b). However, for all three vectors, regardless of the MOI used, the percentage of blue cells was considerably higher in submucosal gland cells than in surface airway epithelial cells (Figure 1).

View larger version (16K): [in this window] [in a new window]   Figure 1.   Comparison of efficiency of different adenovirus vectors in CF submucosal gland cells (a) and CF surface epithelial cells (b). Gal expression was determined by X-Gal staining 48 h after cells were infected with adenovirus vectors at a MOI of 100. Squares, triangles, and circles indicate results attained with Ad2/ Gal-4, Ad2/Gal-E1a and Ad2/Gal-PGK, respectively. Data are expressed as percentage of blue cells assessed by light microscopy (means ± SEM, n = 6).

Because we do not know the sensitivity of detection of Gal on a per-cell basis, it was difficult to estimate Gal expression in cells in which the expression level was below the detection using the X-Gal staining assay. We therefore performed biochemical assays to quantify Gal expression and normalized the levels to total cell protein. Quantitative determination of Gal levels in submucosal gland cells and surface airway epithelial cells indicated that vectors harboring the CMV promoter provided the greatest level of expression (Figure 2). However, this assay revealed that adenovirus vectors utilizing the E1a or PGK promoters yielded comparable levels of Gal expression.

View larger version (14K): [in this window] [in a new window]   Figure 2.   Comparison of Gal expression in CF submucosal gland cells (a) and CF surface epithelial cells (b). Cells were infected with Ad2/Gal-4 (squares), Ad2/Gal-E1a (triangles), and Ad2/Gal-PGK (circles), at a MOI of 100. Gal expression was determined by a quantitative assay and normalized to total cell protein 48 h after infection. Data are expressed as means ± SEM (n = 6).

Optimization of Cationic Lipid-Mediated Gene Transfer to Submucosal Gland Cells

The optimal ratio and concentration of GL-67 and pCF1-Gal in submucosal glands was determined using the 96-well assay as described in MATERIALS AND METHODS. Figure 3 shows that for any given GL-67 concentration, dose-dependent increases in transfection were obtained as pDNA levels were increased, with saturation occurring at approximately 120 µM. In contrast, the cationic lipid dose-dependent curves were bell-shaped, possibly due to toxicity at high concentrations. Maximal transfection levels were obtained with between 120 and 240 µM pDNA and between 10.5 and 21 µM cationic lipid. As the pDNA concentration was increased, the optimal level of lipid for transfection also increased. Thus, with 3.7 µM pDNA, peak Gal expression was seen with 5 µM cationic lipid. When the amount of pDNA used was 240 µM, the optimal lipid concentration was 21 µM.

View larger version (50K): [in this window] [in a new window]   Figure 3.   Optimization of cationic lipid-mediated gene transfer to CF submucosal gland cells. Gal expression was determined using a 96-well assay at 48 h after transfection with complexes of GL-67:pCF1-Gal. Each data point represents results averaged from duplicates. The experiment was repeated once, and similar results were obtained.

We also determined the efficiency of GL-67-mediated gene transfer into submucosal glands by X-Gal staining (Figure 4). Cells were transfected with the cationic lipid GL-67:pCF1-Gal complexes at different molar ratios. At a fixed GL-67 concentration of either 10.5 or 21 µM, the lipid:pDNA ratios were varied between 2:1 and 1:8. Maximal transfection (70% of cells transfected) was attained with GL-67:pCF1-Gal at a molar ratio of 1:4 (21:84 µM), which was approximately equivalent to that observed with Ad2/Gal-4 at a MOI of 100. 

View larger version (14K): [in this window] [in a new window]   Figure 4.   Efficiency of cationic lipid-mediated gene transfer to CF submucosal gland cells. Cells were transfected with complexes of GL-67:pCF1-Gal. The concentration of GL-67 was fixed at either 10.5 (circles) or 21 (squares) mM. The lipid:pDNA ratio was varied from 2:1 to 1:8. Data were pooled from two independent experiments.

Longevity of Gal Expression

We previously reported that surface epithelial cells infected with Ad2/CFTR-5, which utilizes a CMV promoter, exhibited a rapid decline in transgene expression although the initial expression level was higher compared with vectors harboring the E1a promoter (29, 34). In contrast, the adenoviral vector containing the E1a promoter showed more prolonged transgene expression (29). In this study, we tested whether similar findings were true of submucosal glands. We infected cells with Ad2/Gal-4 or Ad2/ Gal-E1a (100 MOI) and assessed Gal expression after 1, 2, 5, 7, 9, 12, 15, and 18 d. In cells infected with either vector the expression peaked on Day 2 and then decreased, with a half-life of approximately 2.5 d (Figure 5a). Thus, on Day 9, only 10% of the expression level detected on Day 2 remained. Gal expression was virtually undetectable by Day 18. The profile of transgene expression was not significantly different between the two adenovirus vectors. When submucosal gland cells were transfected with complexes of GL-67 and pCF1-Gal, Gal expression peaked on Day 1 but declined rapidly on Day 2 (P < 0.05, unpaired Student's t test; Figure 5b). About 10% of the expression level detected on Day 1 was retained after 7 d, with no activity above baseline detected after 9 d.

View larger version (15K): [in this window] [in a new window]   Figure 5.   Persistence of Gal expression in CF submucosal gland cells. Cells were (a) infected with adenoviral vectors at a MOI of 100 or (b) transfected with complexes of cationic lipid GL-67:pCF1-Gal at 21:80 µM. In (a), open and solid columns indicate Ad2/Gal-4 and Ad2/Gal-E1a, respectively. Data are expressed as means ± SEM (n = 3).

CFTR Cl Channel Activity Assessed by SPQ

Surface epithelial and submucosal cultures infected with Ad2/CFTR-5 (100, 300, and 500 MOI) or Ad2/Gal-4 (300 MOI) were loaded with SPQ in I medium and then bathed in nitrate medium. After addition of forskolin (10 µM) and IBMX (100 µM), only a small increase in SPQ fluorescence was detected in less than 1% of the cells infected with Ad2/Gal-4 or in uninfected cells. In contrast, 20 to 89% of the cells infected with Ad2/CFTR-5 showed a rapid increase in SPQ fluorescence in response to forskolin and IBMX (Figure 6a), indicating an increased anion permeability. There was a correlation between the maximal rate of increase in SPQ fluorescence and the MOI used. These results suggest that CFTR Cl channel activity could be restored to CF submucosal gland cells by adenovirus-mediated gene transfer. To compare the efficiency of adenovirus-mediated CFTR gene transfer in surface epithelial cells and submucosal glands, we also infected CFT1 cells with Ad2/CFTR-5. Figure 6b shows that CFT1 cells infected with Ad2/Gal-4 showed little response to the addition of forskolin plus IBMX. In contrast, cells infected with Ad2/CFTR-5 showed a rapid increase in SPQ fluorescence. The pattern of the responses was similar to that observed in submucosal gland cells.

View larger version (29K): [in this window] [in a new window]   Figure 6.   Increase in cAMP-mediated Cl channel activity as determined by SPQ analysis after infection with adenovirus vectors. CF submucosal gland cells (a) or CF surface epithelial cells (b) were infected with Ad2/CFTR-5 or Ad2/Gal-4. Open circles, diamonds, and squares indicate infection with Ad2/CFTR-5 at a MOI of 100, 300, and 500, respectively. Triangles and solid circles indicate Ad2/Gal-4 (500 MOI) and naive controls, respectively. The arrows indicate the addition of forskolin (20 µM) and IBMX (100 µM).

In both submucosal gland cells and CFT1 cells transfected with GL-67:pCF1-Gal, there was no detectable response to the addition of forskolin plus IBMX. However, in cells transfected with GL-67:pCF1-CFTR, the addition of forskolin and IBMX induced a rapid increase in SPQ fluorescence (Figure 7). The pattern of the response was similar in CFT1 and submucosal gland cells (Figure 7). In both cell types, the magnitude of the response and the percentage of responsive cells were related to the ratios and concentrations of GL-67:pCF1-CFTR used. The percentages of responsive cells in submucosal gland cells transfected with either 10.5:20 µM or 21:80 µM of GL-67:pCF1-CFTR complexes were 39 ± 9% and 66 ± 13%, respectively. In surface epithelial (CFT1) cells transfected with the same concentrations of GL-67:pCF1-CFTR (10.5:20 and 21:84 µM) the percentages of responsive cells were 33 ± 10% and 61 ± 8%. There was no significant difference between the two cell types (P < 0.05, unpaired Student's t test). In addition, the magnitude of the cAMP-stimulated response in submucosal gland cells transfected with complexes of GL-67:pCF1-CFTR at the optimized concentration (21:84 µM) and ratio (1:4, lipid:pDNA) was equivalent to that observed in cells infected with Ad2/CFTR-5 at a MOI of 100. These results are similar to that observed using the X-Gal staining assay.

View larger version (26K): [in this window] [in a new window]   Figure 7.   Increase in cAMP-mediated Cl channel activity as determined by SPQ analysis after transfection with cationic lipid:pDNA complexes. CF submucosal gland cells (a) or CF surface epithelial cells (b) were transfected with GL-67:pCF1-CFTR or GL-67:pCF1-Gal. Squares and circles indicate GL-67:pCF1-CFTR at concentrations of 21:80 and 10.5:40 µM, respectively. Triangles indicate results of transfection with GL-67:pCF1-Gal (21:80 µM). The arrows indicate the addition of forskolin (20 µM) and IBMX (100 µM).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Adenovirus-Mediated Gene Transfer into Submucosal Gland Cultures

In human airways, CFTR is predominantly expressed in the serous cells of tracheobronchial submucosal glands (11, 12). Although most non-CF surface airway epithelial cultures absorb fluid under basal conditions, 60% of unstimulated tracheobronchial submucosal gland cultures from non-CF subjects secrete fluid, indicating their importance in maintaining the normal hydration and mucociliary clearance of airways (5). Further, tracheobronchial submucosal gland cultures from CF patients fail to secrete Cl (12, 14, 15) and fluid (5). Despite these indications for a role of submucosal glands in the pathophysiology of CF lung disease, CF gene therapy efforts have focused mainly on correcting the surface airway epithelial cells. However, Mastrangeli and colleagues (17) have shown that following intratracheal instillation of adenovirus vectors, low levels of Gal expression were detectable in a variety of cell types, including submucosal gland cells. Another study also demonstrated that adenoviral vectors were capable of transfecting human airway submucosal glands in a xenograft model (18). In the present study, we determined the relative efficiency of adenoviral vectors at transfecting submucosal gland and surface epithelial cells. The efficiency of adenovirus-mediated transfection of submucosal gland cells as determined by X-Gal staining was equivalent to if not higher than that of surface epithelial cells. The ranking order of transfection efficiency in CF submucosal gland cells was CMV >> E1a ~ PGK, similar to that observed in CF surface epithelial cells (29). Thus, in both cell types, the CMV promoter was more efficient at driving transgene expression than were either the E1a or PGK promoters. It should be noted that approximately 85% of the submucosal gland cells infected with Ad2-CMVGal at a MOI of 10 were positively stained with X-Gal. Increasing the virus vector from 10 to 100 MOI increased the percentage of positively stained cells only by another approximately 10% (Figure 1a). By contrast, there was a 10-fold increase in total Gal expression when measured by a quantitative assay (Figure 2a). A possible explanation for the discrepancy is that an increase in transgene expression per cell rather than in the number of cells expressing the transgene was mainly responsible for the 10-fold increase in Gal expression after infection at a high MOI. This discrepancy was less apparent in surface epithelial cells than in submucosal gland cells (Figures 1b and 2b).

Although the mechanisms underlying the internalization of adenoviral vectors into submucosal gland cells, penetration of endosomes, and nuclear translocation may be different between surface epithelial and submucosal gland cells, our data suggest that submucosal glands are as susceptible as surface epithelial cells to adenoviral infection. Given the morphology of intact submucosal glands in vivo, our data suggest that viral particle access in addition to viral tropism is likely to be another major limiting factor for adenovirus-mediated gene transfer into these cells. Further, it should be noted that results in transformed cells in vitro may differ from results in polarized and differentiated submucosal gland cells. It cannot be ruled out that in polarized and differentiated submucosal gland cells in vivo, the susceptibility of the cells to transfection by the vectors may present a significant barrier.

Cationic Lipid-Mediated Gene Transfer into Submucosal Gland Cells

Cationic lipids have received considerable attention as an alternative approach in CF gene therapy. However, in most studies the efficiency of cationic lipid-mediated gene transfer into surface epithelial cells has been relatively low. Recently we developed and characterized several cationic lipids of novel structure types (27). In surface epithelial cells in vitro and in mouse lungs in vivo, cationic lipid GL-67 showed relatively high efficiency at mediating gene transfer. Using GL-67:pDNA complexes, we provide here the first demonstration that cationic lipids are capable of mediating gene transfer into submucosal gland cells. At the optimal concentrations of GL-67:pCF1-Gal, expression of 20 to 80 mU Gal/µg cell protein was achieved. We found that transformed submucosal gland cells were as susceptible as surface airway epithelial cells to cationic lipid-mediated gene transfer. Further, under optimal conditions, the efficiency of cationic lipid-mediated gene transfer in submucosal gland cells was equivalent to that attained with Ad2/Gal-4 at a MOI of 100.

Restoration of Functional Cl Channel Activity

Defective Cl secretion and increased Na⁺ absorption by surface airway epithelial cells are characteristics of CF airway epithelial cells (2). Tracheobronchial submucosal gland cultures from patients with CF fail to secrete Cl (12, 14, 15) and fluid (5). It has been demonstrated that adenoviral-mediated CFTR gene transfer into CF surface epithelial cells can correct the defects in cAMP-mediated Cl (29, 35) and Na⁺ absorption (29, 39, 40). Recently, the first definitive demonstration of adenovirus-mediated delivery of Gal into human airway submucosal glands in a xenograft model established the feasibility of using gene therapy to target submucosal glands (18). However, the ability of CFTR gene transfer to restore Cl secretion in CF submucosal gland cells has not been documented. In this study, we observed that infection with Ad2/CFTR-5 increased the magnitude of the cAMP-stimulated increase in SPQ fluorescence in submucosal gland cells. The percentage of responsive cells was also significantly greater in cells infected with Ad2/CFTR-5 than with Ad2/Gal-4. The increase in the magnitude and percentage of responsive cells was dependent upon the MOI used. Similar results were obtained from cells infected with other Ad2/CFTR vectors containing different transcription cassettes (data not shown). We also transfected submucosal gland cells with complexes of cationic lipid GL-67:pCF1-CFTR. SPQ analysis showed that transfection with GL-67:pCF1-CFTR complexes resulted in a significant increase in Cl permeability. Thus, adenovirus-mediated and cationic lipid-mediated CFTR gene transfer are capable of correcting cAMP-mediated Cl secretion in CF submucosal gland cells.

Longevity of Transgene Expression

Neither adenovirus nor cationic lipid:pDNA vectors facilitate stable integration of the CFTR gene into the host genome, thereby requiring repeated administration. Because both vectors are capable of eliciting an inflammatory response in the lung, reducing the frequency of administration is desirable. Increasing the longevity of gene expression is one approach by which this problem may be lessened. Different longevities of CFTR or reporter gene expression using adenovirus vectors in various surface epithelia models have been reported, ranging from 11 d to 3 wk (37, 41, 42). A number of nonimmunologic and immunologic factors influence the longevity of transgene expression. We previously reported that for surface epithelial cells in vitro, different promoters resulted in different longevity profiles independent of immunologic responses (29). In this study, we followed Gal expression in submucosal gland cells over 18 d. After administration of Ad2/ Gal-4, Gal expression peaked at Day 2 and dropped by 50% at Day 5. At Day 9 only 10% expression was detected. In contrast to our observation in airway surface epithelial cells, both Ad2/Gal-E1a and Ad2/Gal-PGK had different longevity profiles compared with Ad2/Gal-4. The reason for this discrepancy is unknown. The longevity of cationic lipid-mediated gene expression in submucosal glands was similar to that observed with adenoviral vectors. We speculate that loss of transgene expressing cells due to necrosis or apoptosis may be partially responsible for the transient gene expression. Another possible reason for transient gene expression may be that the pDNA was lost from the cells following mitosis. Finally, we cannot rule out the possibility that the promoters used in these studies were inactivated, for example, by methylation. By concentrating on why expression is transient, future studies in polarized and differentiated submucosal gland cells in vitro or tracheobronchial submucosal glands in vivo should allow us to develop strategies that prolong the duration of transgene expression.