In this study, we utilized the reverse transcriptase component of telomerase, hTERT, and human papillomavirus type 16 (HPV-16) E6 and E7 genes to transform normal and cystic fibrosis (CF) human airway epithelial (HAE) cells. One cell line, designated NuLi-1 (normal lung, University of Iowa), was derived from HAE of normal genotype; three cell lines, designated CuFi (cystic fibrosis, University of Iowa)-1, CuFi-3, and CuFi-4, were derived from HAE of various CF genotypes. When grown at the air-liquid interface, the cell lines were capable of forming polarized differentiated epithelia that exhibited transepithelial resistance and maintained the ion channel physiology expected for the genotypes. The CF transmembrane conductance regulator defect in the CuFi cell lines could be corrected by infecting from the basolateral surface using adenoviral vectors. Using nuclear factor-κB promoter reporter constructs, we also demonstrated that the NuLi and CuFi cell lines retained nuclear factor-κB responses to lipopolysaccharide. These cell lines should therefore be useful as models for studying ion physiology, therapeutic intervention for CF, and innate immunity.
- cystic fibrosis transmembrane conductance regulator
- human papillomavirus type 16
- human airway epithelial cells
the airway epithelium is not completely understood. It lines the airway ducts that connect the environment with the gas exchange epithelium of the lung. The surface epithelium lining the large human airways is pseudostratified; that is, all cells extend to the basement membrane, but not all cells extend to the luminal surface. The airway epithelium forms a barrier between the external and internal environments, prepares the air for optimal gas exchange at the alveoli, and purifies the air by removing particles and bacteria. The mucus at its surface helps prevent invasion of microorganisms and viruses. Its ciliated surface allows for an efficient mucociliary escalator that allows clearance of particles. The airway epithelium secretes numerous agents into the airway surface liquid, including immunoglobulins and antimicrobial factors; these form part of the defensive shield that protects the airways and lungs from bacterial infection. By active transepithelial transport of electrolytes, it controls the composition and volume of the airway surface liquid covering the epithelium to ensure proper mucociliary clearance and innate immunity. Finally, the airway epithelium plays an important role in the inflammatory response when challenged with environmental factors or infectious agents. It responds to and produces a number of cytokines and other pro- and anti-inflammatory agents that modulate innate immunity.
Several genetic and acquired diseases of the lungs involve the airway epithelia in their pathogenesis and highlight the importance of the airway epithelium (46). These include cystic fibrosis (CF), immotile cilia syndrome, pseudohypoaldosteronims, chronic bronchitis, lung cancer, and viral infections. Therefore, in vitro models of the epithelium would be helpful in understanding the pathogenesis of disease and in developing new therapies.
In earlier studies, methods for culturing and differentiating primary airway epithelial cells have been described. The differentiation was facilitated by air-liquid interface culture (10, 14, 19,49). Compared with in vivo studies, such models have the important advantage of flexibility, control of experimental conditions, and greater opportunities for interventions. However, primary cultures have some disadvantages: limited availability of primary cells and significant variability between donors. Therefore, several cell lines have been developed. These cell lines have been derived from carcinomas or have been transformed using viral genes (for review of airway epithelial cell lines see Refs. 11, 15,51). Although these cell lines have been very valuable, they often have limitations. Primarily, the morphology and some important functions are not always retained.
The present data strongly support the hypothesis that immortalization of human cells requires activation of a mechanism to maintain telomeres (for review see Ref. 40). It has recently been demonstrated that exogenous expression of hTERT, the catalytic component of telomerase, can efficiently immortalize certain human cell types (4) and that these immortal cell lines exhibit few, if any, phenotypic alterations (25). Some cell types, for reasons that are not completely clear, cannot be immortalized by hTERT alone and require abrogation of the retinoblastoma (Rb) and/or p53 pathways, usually by expression of viral oncoproteins (20,32). The goal of the present work was to develop human airway epithelial (HAE) cell lines that maintain phenotypic qualities that would make them useful to study airway epithelial biology, viral pathogenesis, and airway surface liquid composition and to develop models for gene therapy for CF studies.
Primary HAE cell culture model.
Primary epithelial cells were enzymatically isolated from bronchial epithelium of human donor lungs, as previously described (19). During transformation and subsequent passaging, the cells were cultured on collagen-coated plastic dishes (type VI, human placental; catalog no. C-7521, Sigma) in serum-free bronchial epithelial cell growth medium with supplements (catalog no. CC3170, Clonetics/BioWhittaker). For the epithelial cultures, cells were seeded onto collagen-coated, semipermeable membranes (0.6 cm2, Millicell-PCF; Millipore, Bedford, MA), grown at an air-liquid interface as previously described (39, 49, 57), and cultured for the 1st day in 50:50 DMEM-Ham's F-12 medium supplemented with 5% fetal bovine serum. On the day after seeding, the cells were grown and then maintained in 50:50 DMEM-Ham's F-12 medium supplemented with 2% Ultroser G (Biosepra, Cergy-Saint-Christophe, France). Basolateral culture medium was changed every 2–4 days. Viable polarized epithelial cultures are stable for ≥3–6 mo (44). Samples were collected with approval from the University of Iowa Institutional Review Board.
Primary or passage 1 cryopreserved airway epithelial stocks were thawed and expanded for retroviral infections with pLXSN- or pBABE-based retroviral vectors using supernatants generated from stably producing lines (PA317) or from transient transfections of Phoenix amphotrophic packaging lines, as previously described (13,50). Cells were infected with retroviral vector (LXSN) alone or retrovirus expressing hTERT alone or dually infected with hTERT and human papilloma virus (HPV)-16 E6/E7 retroviruses using protocols that have been previously described (13, 20). NuLi (normal lung, University of Iowa)-1 and CuFi (CF, University of Iowa)-1, CuFi-2, and CuFi-3 cells were generated by dual infection with HPV-16 E6/E7-LXSN (13) and hTERT-LXSN (20); NuLi-2 and CuFi-4 cells were generated by dual infection with HPV-16 E6/E7-LXSN and pBabe-neo-hTERT (12). Cells were selected with appropriate antibiotic [G418 (50 μg/ml) and/or hygromycin (8 μg/ml)], split 1:6 with each passage, collected for various assays, and expanded for cryopreservation.
RT-PCR, Western blot, telomere repeat amplification protocol, and cytogenetic analysis.
RT-PCR for HPV-16 E7 was performed using a Retroscript kit (Ambion) according to the manufacturer's protocol and E7 specific primers: 5′-ATG ACA GCT CAG AGG AGG AG-3′ (forward) and 5′-TCA TAG TGT GCC CAT TAA CAG-3′ (reverse). Hot-start PCR was performed with 5 min of denaturing at 95°C followed by 33 cycles of 95°C for 30 s, 52°C for 30 s, and 72°C for 30 s, with a final extension for 5 min at 72°C. This yielded an E7-specific band of 176 bp.
Western analyses were performed as previously described (20) using polyvinylidene difluoride membranes with an antibody for p53 (Oncogene Science) and secondary horseradish peroxidase-conjugated goat anti-mouse IgG (Jackson ImmunoResearch) and visualized using a chemiluminescence system (Renaissance, DuPont NEN).
Telomerase analysis was performed using a telomere repeat amplification protocol ELISA kit (Roche) with 1 μl of cell lysate consisting of 1,000 cells/μl according to the manufacturer's protocol. An HPV-16 E6/E7 immortalized human keratinocyte line was used as a positive control.
Cytogenetic analysis was performed at the University of Iowa Cytogenetics Facility on Giemsa-banded chromosomes on ≥10 metaphase spreads, as previously described (45).
Staining for goblet and ciliated cells.
Primary cultures of human airway epithelium were stained for fluorescence microscopy as follows. Cultures were fixed in 4% paraformaldehyde and then washed three times in phosphate-buffered saline (PBS). Jacalin (JAC) lectin directly conjugated to FITC (1:200; Vector Laboratories, Burlingame, CA) was applied to the apical surface at 4°C for 30 min (29, 41). After the cultures were washed in PBS, they were mounted on glass slides, and coverslips were applied with Vectashield mounting medium (Vector Laboratories). To quantitatively estimate the number of ciliated cells, epithelia were fixed in 4% paraformaldehyde and washed three times in PBS. Epithelia were then permeabilized with 0.2% Triton X-100, rinsed, and incubated with Superblock (Pierce). Epithelia were incubated with mouse antikeratan sulfate monoclonal antibody (Chemicon, Temecula, CA) diluted 1:100 in Superblock and incubated again in a 1:200 dilution of FITC-labeled anti-mouse IgG (Biomeda) (44, 58). Epithelia were analyzed using a scanning laser confocal microscope (model MRC-1024; Bio-Rad, Richmond, CA). Images were obtained en face. To quantitate the percentage of JAC- or keratan sulfate-positive cells per epithelial preparation, epithelial cultures were counterstained with 4′,6-diamidino-2-phenylindole (Vector Laboratories) to label total number of cells.
For scanning electron microscopy, samples were fixed with 2.5% glutaraldehyde and then with 1% osmium tetroxide (43). Samples were dehydrated through a graded series of ethanol and finally with hexamethyldisilizane (Ted Pella, Redding, CA). Samples were mounted onto grids, sputter coated with gold, and imaged on a Hitachi S-4000 scanning electron microscope.
Measurement of transepithelial electrical properties.
For measurement of transepithelial electrical properties, epithelia were mounted in Ussing chambers and studied as previously described (38, 54). Epithelia were bathed in symmetrical solutions containing (in mM) 135 NaCl, 2.4 K2HPO4, 0.6 KH2PO4, 1.2 CaCl2, 1.2 MgCl2, 10 dextrose, and 5 HEPES (pH 7.2) at 37°C and gassed with 100% O2. I sc Amil is the decrease in short-circuit current (I sc) after apical addition of 10 μM amiloride. cAMP-stimulatedI sc (I sc cAMP) is the increase in current after basolateral addition of cAMP agonists (10 μM forskolin + 100 μM IBMX). Bumetanide-sensitiveI sc (I sc Bumet) is the decrease in current after basolateral addition of 100 μM bumetanide to epithelia studied in the presence of apical 10 μM amiloride and basolateral cAMP agonists (10 μM forskolin + 100 μM IBMX);I sc Bumet is a measure of the transepithelial Cl− transport pathway that includes CF transmembrane conductance regulator (CFTR). To investigate the ability to correct the Cl− transport pathway on CuFi cells, we studied the effect of adenovirus (Ad)-mediated expression of CFTR (AdCFTR). Briefly, to disrupt the tight junctions, we pretreated the apical surface of the epithelium for 30 min with 400 μl of 8 mM EGTA in Eagle's modified essential medium (EMEM) containing 50 multiplicities of infection (MOI) of AdCFTR at 37°C. Epithelia were studied in Ussing chambers 3 days later as described above.
Polarity of infection by adenoviruses.
Recombinant adenovirus vectors expressing β-galactosidase, AdβGal, were prepared as described previously (28, 42) by the University of Iowa Gene Transfer Vector Core at titers of ∼1010 infectious units (IU)/ml. At 14 days after the epithelia were seeded, 25 μl containing 50 MOI of the recombinant viruses in PBS were added to the apical or basolateral surface for 30 min (particle-to-IU ratio = 25). After incubation, the viral suspension was removed and the epithelia were rinsed twice with PBS. After infection, the epithelia were incubated at 37°C for an additional 48 h. Total β-galactosidase activity was then measured using a commercially available method (Galacto-Light; Tropix, Bedford, MA). Briefly, after they were rinsed with PBS, the cells were removed from the filters by incubation with 120 μl of lysis buffer (25 mM Tris-phosphate, pH 7.8, 2 mM dithiothreitol, 2 mM 1,2-diaminocyclohexane-N,N,N′,N′-tetraacetic acid, 10% glycerol, and 1% Triton X-100) for 15 min. Light emission was quantified in a luminometer (Analytical Luminescence Laboratory, San Diego, CA).
Adenovirus-mediated gene transfer corrects the CF epithelial Cl− transport defect.
To study the ability to correct the Cl− transport pathway on CuFi cells, we studied the effect of adenovirus-mediated expression of CFTR cDNA on the three CuFi cell lines. Recombinant adenovirus vectors expressing CFTR cDNA under the cytomegaolvirus promoter, AdCFTR, were prepared as described previously (52). Briefly, to disrupt the tight junctions of the epithelia and allow access of the recombinant viruses to the basolateral surface, we pretreated the apical surface of the epithelium for 10 min with 400 μl of 8 mM EGTA in EMEM containing 50 MOI of AdCFTR or adenovirus expressing green fluorescent protein (AdGFP) as control (42). The virus was rinsed, and epithelia were studied to assess Cl− transport in Ussing chambers 3 days later (see above). We also studied epithelia derived from the primary cells that were used to develop the CuFi-1, CuFi-2, CuFi-3, and CuFi-4 cell lines.
Lipopolysaccharide-induced activation of nuclear factor-κB.
The cDNA for luciferase was introduced into primary and NuLi airway epithelial cells by treating the epithelia with 400 μl of 8 mM EGTA in EMEM containing 50 MOI of the adenovirus that expresses nuclear factor-κB (AdNF-κB) (35) luciferase at 37°C. After 1 day, the epithelia were exposed to Pseudomonas aeruginosalipopolysaccharide (LPS) at 0.1 ng/ml, and luciferase activity was analyzed (1). After the cells were rinsed with PBS, they were removed from the filters by incubation with 120 μl of lysis buffer (25 mM Tris-phosphate, pH 7.8, 2 mM dithiothreitol, 2 mM 1,2-diaminocyclohexane-N,N,N′,N′-tetraacetic acid, 10% glycerol, and 1% Triton X-100) for 15 min. Light emission was quantified in a luminometer (Analytical Luminescence Laboratory).
Passaging primary HAE cells.
Primary cells generally are good for electrophysiology studies for only the first couple of passages; then they lose their ability to form tight junctions and transepithelial resistance (R t) when grown at the air-liquid interface. In our experience, the passaging of HAE cells results in increasedR t and decreased I sc bypassage 1 with a significant drop ofR t on further passages. Most of the time, cells passaged two to three times fail to form structures that exhibitR t. In Fig. 1, we show the results of passaging cells from one donor three times.R t increased by 40% on passages 1and 2 and dropped substantially on passage 3(Fig. 1 A). We observed a linear drop inI sc (Fig. 1 B) andI sc Amil (Fig. 1 C) as well as inI sc cAMP (Fig. 1 D). By passage 3, the epithelia exhibited minimal active Na+ and Cl− transport. Data for passages 4 and beyond are not shown, because epithelia that could be studied in Ussing chambers fail. These data suggest that it is important to develop airway epithelial cell lines that retain the ability to differentiate, form tight junctions, and maintain ion channel physiology when grown at the air-liquid interface.
Generation of HAE cell lines.
To develop HAE cell lines, we wanted to first determine whether immortalization could be achieved by infection with retroviruses expressing hTERT alone. Because it has been shown that some cell types require other factors for immortalization, we also decided to perform experiments in parallel in which cells were coinfected with hTERT and HPV-16 E6/E7 retroviral vectors. The latter retrovirus expresses HPV-16 E6 and E7, which are capable of abrogating the p53 and Rb pathways, respectively (20). Two normal HAE and four CF HAE cell strains were used for this study. We found that exogenous expression of hTERT was inefficient at immortalizing HAE cells, and, in general, hTERT-expressing cells grown on submerged cultures senesced atpassages 10–15, which was approximately the same time at which vector control cells senesced (data not shown). Expression of hTERT and HPV-16 E6/E7 together, on the other hand, resulted in efficient extension of lifespan to beyond passage 30 with no apparent crisis or slow down in growth (as evidenced by the time required to achieve confluency after splitting). The hTERT- E6/E7-transduced cell lines derived by this method were designated NuLi-1 and NuLi-2; the cell lines from CF tissue were designated CuFi-1, CuFi-2, CuFi-3, and CuFi-4 (Table1). The CF genotypes of the lines were verified by commercially available multiplex PCR (Genzyme Genetics, Framingham, MA) and are shown in Table 1. All the cell lines were analyzed for telomerase activity, expression of HPV-16 E7, and p53 levels (as a functional assay for expression of HPV-16 E6). RT-PCR analysis for HPV-16 E7 revealed that all the cell lines expressed E7 transcripts (Fig. 2 A). Western analysis demonstrated downregulated levels of p53 protein in all the cell lines compared with normal or vector control cells, as would be expected of cells that expressed HPV-16 E6 (Fig. 2 B). The cell lines were also shown to express high levels of telomerase as measured by an ELISA-based telomere repeat amplification protocol assay (Fig. 2 C). These data demonstrated that the cell lines contained and expressed the expected hTERT and E6/E7 genes.
Cytogenetic analysis was performed on all the cell lines at approximately passage 25. All the cell lines except NuLi-2 were near diploid, but some chromosomal aberrations were present (Table2). Several cell lines were clonal (i.e., NuLi-2, CuFi-3, and CuFi-4); others consisted of mixed populations. Chromosome arm deletions and isochromosomes were rare. For example, NuLi-1 contained one subpopulation with a deletion of the 5q arm, and CuFi-2 contained an isochromosome 7. The most common aberrations were numerical in nature, which may reflect some protection by telomerase from genetic instability and translocations caused by end-to-end fusions. Increase in copy numbers of chromosomes 5 and 20 was the most common alteration (Table 2).
NuLi-1 cells retain the ability to develop epithelia that are active in Na+ and Cl− transport.
Primary lung epithelial cells infected with retroviruses expressing hTERT and HPV-16 E6/E7 (NuLi-1) were cultured on plastic up topassage 30. Every couple of passages the airway epithelial cells were trypsinized and seeded onto semipermeable filters and grown at the air-liquid interface. After 3–4 wk, the epithelia were studied in Ussing chambers, and the I sc, transepithelial voltage (V t), andR t were recorded. R t was slightly higher after passage 7 than in the primary cultures (685 ± 30.7 vs. 532 ± 146.7 Ω), andR t showed a downward trend as the passage number increased (Fig. 3).R t toward passage 29 was only slightly lower at 389 ± 20.7 Ω. To evaluate ion transport, we recorded I sc at baseline (Fig.4 A),I sc Amil (Fig. 4 C),I sc DIDS (Fig. 4 D),I sc cAMP (Fig. 4 E), andI sc Bumet (Fig. 4 F). We also recorded baseline V t under open-circuit conditions (Fig. 4 B). Baseline I scwas preserved throughout passages 7–29 at 30–40 μA/cm2 and was very similar to that of the primary untransformed cells (34.9 ± 7.7 μA/cm2). Most of this I sc could be accounted for by amiloride-sensitive Na+ transport, and, not surprisingly,I sc Amil was also preserved throughout passaging. I sc Amil at passage 29was similar to that of the untransformed passage 1(25.7 ± 9.2 vs. 27.3 ± 1.4). DIDS blocks Cl−transport via non-CFTR Cl− channels.I sc DIDS in primary airway epithelia is <1 μA/cm2, which is much less thanI sc Amil and I sc cAMPand declines as the passage number is increased. CFTR-mediated Cl− transport can be stimulated with the addition of cAMP agonists (I sc cAMP). In primary cultures,I sc cAMP was similar to that in epithelia grown from transformed passage 7 NuLi cells (33.33 ± 2.9 vs. 29.3 ± 2.5). However, we observed a linear decline inI sc cAMP as the passage number increased.Passage 29 exhibited a cAMP-stimulated Cl−current that was only 16.3% of the current in untransformed primary cells. I sc cAMP, a reflection of cAMP-stimulated Cl− current, andI sc Bumet, a reflection of total (cAMP-stimulated + baseline) Cl− current, mirror each other in magnitude and trends with subsequent passages. We also examined the response of NuLi-1 epithelia to 1 mM ATP in the presence of amiloride (I sc ATP).I sc ATP was significantly higher (68%) in NuLi-1 epithelia than in primary cultures of non-CF epithelia (15.0 ± 1.35 vs. 8.9 ± 2.0 μA/cm2,n = 6, P < 0.01). Thus NuLi-1 cells can form electrically tight airway epithelia that preserve Na+ transport compared with primary cells. For reasons unknown, NuLi-2 cells did not exhibit measurableR t when grown in a similar fashion (data not shown). The progressive decline in Cl− transport in the NuLi-1 cells could imply a decrease in CFTR or a decrease in basolateral K+ channels. Although less than ideal, the significant amount of Cl− transport observed in earlier passages would be useful for studies of CFTR function.
NuLi-1 airway epithelia are highly reproducible.
A major advantage of using an immortalized airway epithelial model over primary lung airway epithelial cells would be the ability to decrease the variability from donor to donor. Thus we asked whether NuLi-1 cells independently thawed and grown at the air-liquid interface at different times would exhibit similar electrophysiology. We grew three different cryostocks of NuLi-1 cells at passage 11 at the air-liquid interface and examined their R t. After 3 wk, the epithelia were studied in Ussing chambers, and we recorded the baselineI sc and the change in I scafter sequential administration of 10 μM amiloride, 10 μM DIDS, 10 μM forskolin + 100 μM IBMX, and 100 μM bumetanide. Figure5 shows that three different cryostock-derived epithelia resulted in identicalI sc. This suggests that using NuLi-1-derived epithelia can significantly improve the reproducibility of experimental results.
CuFi cells retain the ability to develop epithelia that are active in Na+ transport and lack Cl− transport.
Primary lung epithelial cells from CF lung transplant recipients were infected with hTERT and HPV-16 E6/E7 to generate four different immortalized CF cells: CuFi-1 (Δ508/Δ508), CuFi-2 (Δ508/Δ508), CuFi-3 (Δ508/R553X), and CuFi-4 (Δ508/G551D). Cells were passaged on plastic up to 30 generations. At passage 18, cryovials were individually thawed and seeded onto semipermeable filters and compared with their primary source. After 3–4 wk, the epithelia were studied in Ussing chambers, and I sc,V t, and R t were recorded. All cell lines except CuFi-2 exhibited good R t(Table 1). The ion transport properties of passaged CuFi-1, CuFi-3, and CuFi-4 epithelia were compared with ion transport properties of epithelia grown from their corresponding primary cells and passaged epithelia in which wild-type CFTR was expressed using a recombinant adenovirus (AdCFTR). We recorded the I sc as we did on NuLi cells. Baseline I sc values for all three cell lines were very similar to values for the primary untransformed cells (Fig. 6,A–C). Most of thisI sc could be accounted for by amiloride-sensitive Na+ transport and, not surprisingly,I sc Amil was also preserved throughout passaging (Fig. 6 D). Addition of cAMP agonists failed to stimulate CFTR-mediated Cl− transport (Fig.6 E). We also examined the response of CuFi-1 epithelia to 1 mM ATP in the presence of amiloride (I sc ATP).I sc ATP in CuFi-1 epithelia was twice as high as in NuLi-1 epithelia and fivefold higher than in primary cultures of CF epithelia (28.8 ± 1.7 vs. 5.5 ± 0.5 μA/cm2, n = 6, P < 0.01). Restoration of I sc cAMP andI sc Bumet in the CuFi epithelia treated with AdCFTR suggests that these epithelia could be used as an intermediate model for therapeutic interventions that restore Cl−current (Fig. 6).
Morphology of NuLi-1-derived airway epithelia.
Growing airway epithelia that can retain electrical properties of the epithelia will be very useful for the understanding and development of therapies directed at restoring Cl− transport. To further characterize the morphology of these epithelia, we compared the number of ciliated and goblet cells as detected by JAC lectin directly conjugated to FITC (goblet cells) and mouse antikeratan sulfate monoclonal antibody (ciliated cells) (58). The number of goblet cells progressively increased as the passage number increased (Fig. 7). Moreover, the number of ciliated cells decreased significantly from >50% on primary cells to ∼5% by passage 20. Thus most of the cells in the epithelia derived from NuLi-1 cells do not appear to be ciliated or to be goblet cell markers. This suggests that although the epithelia retain significant amounts of Cl− and Na+transport, the morphology of the epithelia only slightly resembles that of the human airway in vivo after extensive passaging.
Polarity of infection by adenoviruses.
To further characterize these transformed epithelia, we asked whether receptors are polarized appropriately to the basolateral side. The data from the Ussing chamber experiments clearly suggested that channels were correctly localized; for example, administration of amiloride resulted in blocking of the epithelial Na+ channel. The receptor for adenovirus (CAR) is also localized exclusively to the basolateral side of the epithelial cells. We and others (30) have shown that basolateral administration of adenovirus results in a more efficient fiber-dependent infection of the airway epithelia. Thus we compared the polarity of adenovirus infection on primary and passaged NuLi-1-derived epithelia. Epithelia were incubated with 10 MOI of a recombinant adenovirus that expresses β-galactosidase. After infection, the epithelia were incubated at 37°C for an additional 48 h. Similar to our previous observations, infection of NuLi-1-derived epithelia was 2 logarithmic units more efficient from the basolateral side than from the apical side (Fig. 8). This suggests that NuLi-1-derived epithelia retain the appropriate polarity of expression of receptors such as CAR. Moreover, these results indicate that these cell lines could be a good model for developing vectors targeted to the apical side.
LPS-induced activation of NF-κB.
One of the properties of the airway epithelia is to offer the first-line release of proinflammatory cytokines and recruitment of polymorphonuclear leukocytes to the air space required to maintain a sterile environment (innate immunity). We tested the ability of NuLi-derived airway epithelia to respond to apical administration of LPS. To accomplish this, we utilized a recombinant adenovirus to introduce the cDNA for luciferase driven by an NF-κB response promoter. Thus activation of NF-κB could be easily detected by measuring luciferase activity in a luminometer. Figure9 demonstrates that primary cells expressed little luciferase activity and that exposure to LPS results in a 10-fold increase in the amount of luciferase activity. Passaged cells exhibited increased basal levels of luciferase activity but continue to respond to LPS stimulation. These data suggest that NuLi-1-derived airway epithelia conserve the TLR-4 NF-κB axis and may be utilized to further understand innate immunity.
The development of methods for growing HAE cells that can reliably form electrically tight epithelia has been indispensable for the understanding of the function of epithelial ion transport, composition of airway surface liquid (23, 56), the CF defect (18, 48), transcytosis of immunoglobulins (7,9), pathogenesis of viral and bacterial infections (30,37, 53), and the repair process of the airways and development of novel vectors for gene therapy (17, 24, 27, 33, 44,55). The reliance on primary cultures has limited the access to this resource, and the variability between different donor epithelia complicates the design and interpretation of the data. Here we show a series of airway epithelial cells that retain epithelial properties after multiple passages and will significantly aid in the understanding of epithelial biology. Although these epithelia failed to completely reproduce the morphology and differentiation aspects of distinct airway epithelial cell types, they may become a useful tool to understand the cues required for airway epithelia to assume a ciliated or secretory phenotype. Moreover, these non-CF epithelia exhibited decreased cAMP-stimulated Cl− currents with serial passaging, but these currents were still significantly different from those seen in CF epithelia.
The generation of CuFi cells might also be useful in the development of mutation-specific therapies. Several groups have screened compounds that may affect ΔF508 or G551D CFTR-expressing cells to restore Cl− channel function (3, 8, 22, 34, 36, 47). This work has been done mostly in yeast and heterologous cell lines. Therapeutic candidates could easily be tested in CuFi-4 epithelia, or novel strategies could be developed to perform high-throughput screening in epithelia. Moreover, compounds that induce read through of stop codon mutations have been tested in cell lines and humans. CuFi-3 cells that express R553X (2, 5, 16, 47) could be an excellent candidate for screening those drugs.
Our success at generating HAE cell lines that retain at least some aspects of normal primary cells may be related to exogenous expression of hTERT in conjunction with HPV-16 E6/E7. Most previous attempts to generate HAE cell lines have relied on transfection of the entire HPV-16 or HPV-18 genomes, E6/E7 alone, or infection with the entire genome of simian virus-40 (11, 15). Without maintenance of telomeres by telomerase activation, these strategies have the potential of generating cell lines that are genetically unstable. Cytogenetic analysis of the NuLi and CuFi cell lines that we have developed indicates that they do not contain numerous chromosome rearrangements and are fairly stable in chromosome number. Interestingly, all the cell lines contained extra copies of chromosome 20, and most contained extra copies of chromosome 5. The relevance of these changes is unknown, although 20q amplifications have been observed in uroepithelial cells immortalized by HPV-16 E7 (6). It is possible that exogenous expression of hTERT confers some measure of stability on the cell lines, both genotypically and phenotypically. This needs to be studied in more detail by comparing cells immortalized by E6/E7 alone and cells immortalized by hTERT and E6/E7 together. Even with the addition of hTERT, however, the transduced cell lines still lost some of their phenotypic properties with time in culture, and not all the transduced cell lines retained the ability to form structures that exhibited R t, even at early passage (i.e., NuLi-2 and CuFi-2). The reason for this is unknown. Immortalization of cells by hTERT alone, without the addition of viral oncogenes, might have resulted in more stable cell lines. We were, however, unsuccessful at efficiently and consistently generating immortal cell lines by expression of hTERT alone. In another study (unpublished observations), we were able to obtain one immortal HAE cell line by hTERT expression alone, but this cell line failed to form an electrically tight epithelium when grown at the air-liquid interface. The observation of poor immortalization of cells by hTERT is similar to a study recently reported by others in which immortalization of HAE cells required expression of hTERT and the early region of simian virus-40 (21). It is unknown why HAE cells are resistant to immortalization by hTERT alone, but it may be related to growth conditions. For example, there is evidence that immortalization of human keratinocytes grown on plastic requires telomerase activation and abrogation of the Rb pathway, but when the cells are grown with irradiated feeders, only activation of telomerase is required (31). One way to potentially overcome the phenotypic changes that may be associated with expression of viral oncogenes would be use of a system in which the viral genes could be turned off or removed by a recombinase after initial expansion of the cells. Such a strategy has recently been used to generate reversibly immortal endothelial cells (26).
In summary, we have described a method to efficiently extend the lifespan of CF and non-CF HAE cells while still allowing, in most cases, the retention of normal phenotypic qualities such as the ability to form an electrically tight epithelium. These cell lines should be useful for studies of ion physiology of airway cells, therapeutic intervention for CF, and innate immunity of epithelial cells.
We thank Jan Launspach, Pary Weber, Tamara Nesselhauf, Lisa Einwalter, David Welsh, Daniel Vermeer, Theresa Mayhew, and Rosanna Smith for excellent assistance; Joel Kline and Dwight Look for insightful discussion; Geron for the hTERT cDNA used in construction of pLXSN-hTERT; Robert Weinberg for the hTERT-neo-pBABE retroviral construct; and Denise Galloway for the HPV-16 E6/E7 retrovirus.
This study was supported by the University of Iowa Central Microscopy Research Facility, the Gene Transfer Morphology Core (supported by the National Institute of Diabetes and Digestive and Kidney Diseases), the In Vitro Cell Models Core (supported by the National Institutes of Health and the Cystic Fibrosis Foundation), and Iowa Statewide Organ Procurement and National Institutes of Health Grants HL-61234, DK-60113, and P30 DK-54759 and Cystic Fibrosis Foundation Grant ENGELH98S0.
Address for reprint requests and other correspondence: J. Zabner, Dept. of Internal Medicine, University of Iowa, Iowa City, IA 52242 (E-mail:).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
First published January 10, 2003;10.1152/ajplung.00355.2002
- Copyright © 2003 the American Physiological Society