Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Recent data support the idea that apical surface fluid (ASF), the thin fluid layer that covers the mucosal surface of airway epithelia, provides an important mucosal defense against bacterial colonization of the airway (1, 2). In studies with primary cell cultures of human airway cells, it was demonstrated that ASF contained antibacterial activity to a wide spectrum of bacteria, including Pseudomonas aeruginosa, Staphylococcus aureus, and Escherichia coli (3). This study demonstrated decreased antibacterial activity in ASF from primary cultures of cystic fibrosis (CF) airway cells. This was attributed to a difference in the salt concentration of normal and CF ASF because when antibacterial activity of ASF washings from normal and CF cells was assayed in vitro with a controlled ionic composition of the medium, no differences were observed. In vitro assays also demonstrated that ASF washings were salt sensitive and that the sensitivity was a function of ASF protein concentration, with greater levels of salt required to inhibit antibacterial activity at higher levels of ASF protein (4). Although genotype-dependent differences in the salt concentration of ASF from primary cultures (5, 6) and human airways (7, 8) have proven to be controversial, with some groups obtaining conflicting results, it is clear that the antibacterial activity of CF lung is less than that of normal lung. Therefore, therapeutic approaches that increase ASF antibacterial activity may be of benefit to patients with CF, irrespective of in vivo ASF salt concentration.

Potential ways to increase antibacterial activity in the CF lung include increasing the secretion of antibacterial proteins and promoting the secretion of additional antibacterial factors that may not be present under basal conditions. Because bacterial killing activity is dependent on the protein concentration of antibacterial factors (4), this approach could be of benefit to patients with CF, irrespective of the issues about differences in the salt content of ASF. Although there are a large number of identified antibacterial proteins and peptides in ASF, and in other organs as well, a complete elucidation of these factors has not been achieved. However, the following are known to be present in human ASF: -defensin-1 (9), -defensin-2 (10), cathelicidin LL-37 (11), lactoferrin, and lysozyme (12). These factors appear to provide an innate broad spectrum, first line of defense against infection that is active even in the absence of inflammation.

Little is known about the regulation of synthesis and secretion of these proteins in the airway. Most antibacterial proteins appear to be constitutively synthesized and secreted, and in general, protein-containing vesicles are not seen in cells that are thought to secrete antibacterial proteins. The best-characterized regulatory mechanism is that of interleukin-1-dependent enhancement of -defensin-2 expression; this appears to involve a transcription factor nuclear factor (NF)-B-dependent mechanism for activation (13). Increases in -defensin-2 expression have also been observed with lipopolysaccharide and tumor necrosis factor  (14, 15).

Studies employing primary cultures of airway cells have a number of difficulties. These include sample procurement and the maintenance of cell cultures in sufficient number and for sufficient periods of time to perform well-controlled pharmacologic experiments. To avoid these problems, we have undertaken to study antibacterial activity in an airway cell line. We have chosen Calu-3 cells, a well-characterized airway cell line that forms high resistance monolayers when grown on permeable supports (16, 17). Calu-3 cells express high levels of cystic fibrosis transmembrane conductance regulator (CFTR), as well as surface antigens like those of serous cells from the glandular regions of the large airway (18). As antibacterial factors are synthesized by serous cells (19), we postulated that Calu-3 cells should be an excellent model system for the study of ASF antibacterial activity. We also hypothesized that clinically relevant drugs could potentiate or attenuate the innate antibacterial properties of airway cells.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Reagents

Cell culture reagents were purchased from GIBCO BRL (Grand Island, NY), with the exception of human placental collagen (Cohesion Technologies, Inc., Palo Alto, CA). Beclomethasone-dipropionate, budesonide, triamcinolone-acetonide, terbutaline-sulfate, salbutamol, and ibuprofen were obtained from Sigma (St. Louis, MO). Cromolyn sodium was obtained from the hospital pharmacy. E. coli M15 were obtained from Stratagene (La Jolla, CA) and P. aeruginosa were obtained from Dr. Terry Machen (University of California, Berkeley, CA). Calu-3 cells were obtained from the American Type Culture Collection (Rockville, MD).

Cell Culture

Calu-3 cells were cultured at 37°C and 5% CO₂ in a 50%/50% mixture of Dulbecco's modified Eagle's medium/Ham's F-12 (DMEM/F12) that was supplemented with 15% fetal calf serum (FCS), 500 U/ml penicillin, and 50 µg/ml streptomycin. Cells were maintained in 75-cm² tissue culture flasks and split when 80 to 90% confluent. For all studies, cells were plated at 2 × 10⁶ cells/cm² onto Costar Transwell inserts (0.4-µm pore size, 12 mm diameter, clear polyester membrane; Cambridge, MA) that had been coated with human placental collagen. Culture medium within the insert was removed on the second day after plating, and cells were grown at an air-liquid interface, in the absence of antibiotics, by feeding from the basolateral side with DMEM/F12 with 15% FCS. Transepithelial resistance (TER) was assayed before medium changes by adding medium to the apical side and measuring TER with chopstick electrodes and an epithelial volt-ohm meter (World Precision Instruments, Sarasota, FL). Medium was then removed from both sides and replaced only on the basolateral side. Confluent monolayers with a TER greater than 300 ohm · cm² were obtained in approximately 7 d.

Isolation and Assay of ASF

Only after a TER greater than 300 ohm · cm² was obtained were cells cultured in the presence or absence (sham-treated monolayers) of pharmacologic agents. For control experiments, monolayers were sham-treated with sterile water. All agents were added to the basolateral medium and cells were cultured in their presence for 48 h. ASF was collected by washing the apical surface three times with 60 µl of sterile, nonbacteriostatic water and combining the washes. ASF in 180 µl of water is referred to as the ASF washing. After a recovery period of at least 2 d, cultures were again incubated with agents for 48 h. ASF was collected from cells for up to 25 d after plating on filters. ASF washings were stored at 4°C, or assayed at the time of collection. Standard inocula of E. coli M15 or P. aeruginosa were incubated overnight with shaking at 37°C in 2 ml of Luria broth (LB). Bacteria were washed three times with sterile, nonbacteriostatic water and resuspended to 10⁴ or 10⁶ cells/µl. Serial dilutions were used to obtain samples with the desired number of colony-forming units (CFU) per microliter. In this fashion, known amounts of bacteria, assayed by plating on LB plates, were added to 30 µl of ASF washings and incubated overnight at 37°C. The bacteria-ASF mixture was then plated onto LB plates and CFU were counted after 18 h.

Reverse Transcriptase/Polymerase Chain Reaction of Messenger RNA from Calu-3 Cells

Total cell RNA was isolated from Calu-3 cells with a Qiagen RNA isolation kit (Chatsworth, CA). Primers for human -defensin-1, human -defensin-2, cathelecidin LL-37, lactoferrin, lysozyme, and CFTR were designed according to the sequences obtained from GeneBank (Genome Database, Johns Hopkins University, Baltimore, MD) and purchased from Integrated DNA Technologies (Coralville, IA). The primers amplified fragments that could be distinguished from genomic DNA or did not amplify genomic DNA, and in all cases polymerase chain reaction products were sequenced to confirm identity. The primer sequences were: -defensin-1, 5' AATCCTGAGTGTTGCCTGCCAGTC 3' and 5' ACTTCT GCGTCATTTCTTCTGGTC 3'; -defensin-2, 5' GCCATGAG GGTCTTGTATCTC 3' and 5' TCTGAATCCGCATCAGC CAC 3'; LL-37, ATAGATGGCATCAACCAGCGGTCC 3' and 5' GACTCTGTCCTGGTACAAGATTCC 3'; lactoferrin, 5' GGA TAGACCTGTGGAAGGATATCTT 3' and 5' ACAGCAGCG CAAAGTCTGCCAGCTT 3'; and lysozyme, 5' GCTACAGG GGAATCAGCCTAG 3' and 5' CCACAACCTTGAACATAC TGACGG 3'. Reverse transcriptase/polymerase chain reaction (RT-PCR) was performed using avian myeloblastosis virus reverse transcriptase (Boehringer Mannheim, Indianapolis, IN) and the same amount of cellular RNA was used in all assays. Reverse transcription was performed at 50°C for 30 min. For complementary DNA (cDNA), amplification samples were held at 94°C for 120 s; then 10 cycles of 94°C for 30 s, 55°C to 45°C (decreasing at 1°C per cycle) for 30 s, and 50°C for 30 s; an additional 25 cycles at 94°C for 30 s, 50°C for 30 s, and 68°C for 30 s, and then 7 min at 68°C. Expected RT-PCR products for -defensin-1, -defensin-2, LL-37, lactoferrin, and lysozyme were 267, 235, 383, 440, and 331 bp, respectively.

Gel Electrophoresis

ASF washings (40 µl) were loaded onto 10% gels and proteins were separated by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE). Tricine SDS-PAGE with 16.5% gels was used to analyze the low-molecular-weight proteins in ASF washing.

Statistical Analysis

All the data are expressed as mean ± standard error of the mean (SEM). One-way and repeated analysis of variance with post hoc analysis by the Student-Newman-Keuls method was used with P  0.05 considered significant.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

To determine if Calu-3 ASF has antibacterial activity, Calu-3 cells were cultured on filters in air-liquid interface in the absence of antibiotics, and ASF washings were collected 2 d after a resistance of at least 300 ohm · cm² was achieved. Antibacterial activity was assayed by incubating ASF washings with bacteria overnight at 37°C and counting recovered CFU. The antibacterial activity is summarized in Figure 1. As shown in Figure 1, when E. coli or P. aeruginosa were incubated in sterile water there was no significant change in CFU number after 18 h. In contrast, when bacteria were incubated with ASF washings at a concentration of 100 µg protein/ml, recovered CFU were not statistically different from zero provided no more than 10⁴ CFU were added. Only when 10⁶ CFU were added was the antibacterial activity unable to reduce the number of recovered CFU. However, it should be pointed out that even under these conditions, there was considerable bacterial killing because the effective doubling time of bacteria was approximately 18 h. In contrast, when bacteria were incubated with 100 µg/ml bovine serum albumin (BSA) for 18 h, bacterial growth was seen at all levels of added CFU; even at 10 and 100 CFU more than 10⁴ CFU were recovered after 18 h (data not shown). This observation suggests that in the presence of ASF washings bacteria can multiply and that it is only when the antibacterial activity of the ASF washings exceeds the initial rate of bacterial growth are recovered CFU less than added CFU.

View larger version (33K): [in this window] [in a new window]   Figure 1.   Antibacterial activity of ASF from Calu-3 cells. The indicated number of E. coli (open bars) or P. aeruginosa (shaded bars) CFU were incubated for 18 h at 37°C with a 30 µl sample of an ASF washing (crosshatched bars) or sterile water (non-crosshatched bars). Recovered CFU are presented as a percentage (mean ± SEM) of added CFU. For E. coli, n = 30 and for P. aeruginosa, n = 10.

Because studies with primary cultures of airway cells have demonstrated that ASF antibacterial activity is salt sensitive, we tested the salt sensitivity of ASF washings from Calu-3 cells. In Figure 2, ASF washings, collected in sterile water, were assayed in the presence of NaCl. In the absence of NaCl, antibacterial activity like that in Figure 1 was observed, but concentrations of NaCl greater than 50 mM inhibited antibacterial activity and bacterial growth was greater than bacterial killing. The data suggest that NaCl-dependent inhibition does not vary with the number of added CFU. To determine if NaCl-dependent inhibition is dependent on the level of antibacterial activity, the salt dependence of antibacterial activity was assayed at two different protein concentrations. At the higher concentration, 300 mM NaCl was required to inhibit antibacterial activity. In contrast, at the lower ASF concentration, only 50 mM NaCl was needed to inhibit this activity.

View larger version (20K): [in this window] [in a new window]   Figure 2.   Inhibition of ASF antibacterial activity by NaCl. In the main figure, CFU recovered is plotted as a function of NaCl concentration in the assay for 102 (open circles), 103 (open squares), and 104 (open triangles) CFU of E. coli. All assays contained 30 µl of an ASF washing and were incubated with bacteria for 18 h at 37°C. In the inset, 102 CFU of E. coli were assayed with 0.4 mg/ml (open circles) or 0.2 mg/ml (open squares) of ASF protein. CFU recovered is plotted as a function of NaCl concentration. For both figures, data are representative of three experiments.

The bimodal nature of the data in Figure 2 is most likely a reflection of two aspects of this assay. In the absence of antibacterial activity, bacterial number will increase exponentially, whereas bacterial killing is at best a linear function of time and ASF protein level. For a system like this, two possible outcomes can occur: (1) the initial rate of bacterial growth exceeds the rate of bacterial killing. In this case, there will ultimately be a large increase in the number of recovered CFU. (2) The initial rate of bacterial growth is less than the rate of bacterial killing. In this case, provided a limited number of killing units is not exceeded, there will ultimately be zero bacteria in the assay. Because we routinely assay CFU by incubating ASF washings and bacteria for 18 h, these considerations provide a reasonable explanation for the bimodal nature of our results. To test this, CFU were assayed at 30-min intervals in ASF washings with different levels of antibacterial activity (protein concentration). As shown in Figure 3, at low levels of antibacterial activity bacterial number increased in an exponential fashion, whereas at higher levels of antibacterial activity bacterial number declined. Note that greater amounts of ASF washings, and therefore higher levels of antibacterial activity, were required to inhibit bacterial growth when the initial number of CFU was increased. This suggests that antibacterial killing occurs at a limited rate and not by a mechanism where the rate of killing is proportional to the number of bacteria in the assay.

View larger version (16K): [in this window] [in a new window]   Figure 3.   Time course for bacterial proliferation in the presence of ASF. ASF washings were mixed with BSA at an equal protein concentration 100% ASF (open squares), 66% ASF (open triangles), 33% ASF (inverted triangles), and 0% ASF (open circles), and 30 µl was incubated with 2000 (A) or 200 (B) CFU of E. coli. At the indicated times, aliquots were removed and CFU determined. Data are representative of two experiments.

To further characterize the antibacterial activity of ASF, ASF was collected from Calu-3 epithelia that had been seeded onto filters between 4 and 42 d before ASF collection. As shown in Figure 4, Calu-3 cells require approximately 7 d to develop a TER of 300 ohm · cm² or greater. ASF washings collected from cells before 7 d in culture had reduced antibacterial activity. At the time a TER of 300 ohm · cm² is achieved, a high titer of antibacterial activity is observed and this is maintained for over 2 wk. However, the antibacterial titer drops to nearly zero after 30 d in culture, surprisingly there is no decrease in TER at this time. Based on these results, all samples of ASF used in this work were collected between 7 and 25 d after cells were plated on filters. We also investigated the rate at which ASF antibacterial activity was regenerated after apical washings. The recovery of appreciable antibacterial activity required between 24 and 30 h (data not shown). This observation led us to collect ASF no more often than every other day. Last of all, it was of interest to determine the stability of the antibacterial activity in our ASF washing samples. Once collected, ASF washings were stored at 4°C or assayed immediately. When the activity of stored samples of ASF washings were compared with that of samples that were assayed at the time of collection, there was no statistically significant decrease in antibacterial activity for up to 10 mo (data not shown).

View larger version (17K): [in this window] [in a new window]   Figure 4.   Kinetics of antibacterial factor production and the generation of TER. CFU recovered (open circles) is plotted as a function of days in culture. In all cases, 103 CFU of E. coli were assayed with 30 µl of an ASF washing for 18 h at 37°C. For comparison, TER (closed circles) is plotted as a function of days in culture. Values are mean ± SEM.

To define the antibacterial factors that are expressed by Calu-3 cells, total cell RNA was isolated from Calu-3 cell monolayers and assayed for -defensin-1, -defensin-2, cathelicidin LL-37, lactoferrin, and lysozyme by RT-PCR. As shown in Figure 5, for all five proteins cDNA of the expected size was obtained and in each case cDNA was sequenced to confirm its identity. Calu-3 monolayers expressed low levels of messenger RNA (mRNA) for -defensin-2 and lactoferrin, but mRNA for these proteins was increased by incubation with 2 µg/ml lipopolysaccharide (data not shown). It is interesting that mRNA for -defensin-1, -defensin-2, and lactoferrin could only be detected when cells were cultured on filters and not when grown on plastic. Because cell culture in the presence of topical corticosteroids and ₂-agonists altered ASF antibacterial activity (see subsequent text), we assayed mRNA for -defensin-1, -defensin-2, cathelicidin LL-37, lactoferrin, and lysozyme in cells that had been cultured with 10 µg/ml budesonide or 10 µg/ml salbutamol. In neither case was there a detectable change in message expression when assayed by RT-PCR (data not shown). However, when ASF washings were characterized by SDS-PAGE, one prominent protein (molecular weight [mol wt] ~ 62 kD) was greater in ASF washings from budesonide-treated cells and less in washings from salbutamol-treated cells (Figure 6). No differences were observed in proteins of mol wt < 10 kD (data not shown).

View larger version (82K): [in this window] [in a new window]   Figure 5.   RT-PCR of Calu-3 antibacterial proteins. Calu-3 cell mRNA was isolated from monolayers that had been cultured at an air- liquid interface for 10 d. RT-PCR products for -defensin-1 (lane 2), -defensin-2 (lane 3), cathelicidin LL-37 (lane 4), lactoferrin (lane 5), and lysozyme (lane 6) are shown. Lane 1 contains molecular weight standards.

View larger version (94K): [in this window] [in a new window]   Figure 6.   Separation by SDS-PAGE of proteins from Calu-3 cell ASF. Equal volumes, 40 µl, of ASF washings from Calu-3 monolayers treated with 100 µg/ml of budesonide (lane 1), vehicle (lane 2), or 100 µg/ml of salbutamol (lane 3) are shown. Molecular weight standards (kD) are indicated on the left.

To determine if ASF antibacterial activity can be increased, we tested the effects of culturing Calu-3 cells under a number of conditions on ASF antibacterial activity. In our initial studies, we have examined the effects of culturing Calu-3 cells in the presence of clinically approved pharmaceuticals that are currently used in the treatment of CF and other airway diseases. In these studies, agent or vehicle was added to basolateral medium for 48 h before ASF washings were collected. Our major focus was on topical corticosteroids and ₂-agonists. As shown in Figure 7, when ASF washings from Calu-3 cells that had been cultured in the presence of 100 µg/ml budesonide were compared with ASF washings from control cells, ASF from budesonide-treated cells had greater antibacterial activity. In contrast, ASF washings from cells that had been cultured with 100 µg/ml salbutamol had less antibacterial activity than did ASF washings from control cells. To extend these findings, dose-response studies were carried out. Calu-3 cells were cultured for 48 h in the presence of budesonide at concentrations between 0.01 µg/ml and 100 µg/ml, and the antibacterial activity of ASF washings assayed with 10⁴ CFU of E. coli. As shown in Figure 8, a dose-dependent inhibition of bacterial growth was observed. Similar dose-dependent effects were also observed with the topical corticosteroids beclomethasone and triamcinolone. In all cases, the addition of drug directly to ASF washings from control cultures was without effect on antibacterial activity. In contrast, as shown in Figure 9, dose-response curves for the ₂-agonist salbutamol displayed the opposite effect, with increased cell growth in ASF washings from cells treated with salbutamol. A decrease in the antibacterial activity of ASF washings was also observed for cells that had been cultured with terbutaline. As with the studies of corticosteroids, the addition of salbutamol and terbutaline (10 µg/ml) directly to ASF washings from control cells had no effect on antibacterial activity (data not shown). We also cultured Calu-3 cells with the nonsteroidal anti-inflammatory agents ibuprofen and cromolyn sodium. Cromolyn sodium at concentrations up to 100 µg/ml and ibuprofen at concentrations up to 100 µg/ml, the plasma concentration required for anti-inflammatory effects in patients with CF and the concentration at which ibuprofen can inhibit CFTR function, had no effect on bacterial killing (data not shown) (20, 21).

View larger version (14K): [in this window] [in a new window]   Figure 7.   Effects of culture conditions on antibacterial activity of Calu-3 cell ASF. Calu-3 cell monolayers were cultured in the presence of 100 µg/ml salbutamol (closed triangles), 100 µg/ml budesonide (closed squares) or vehicle (closed circles) for 48 h before the collection of ASF. ASF washings (30 µl) were assayed with indicated number of E. coli CFU at 37°C for 18 h and recovered CFU plotted as a function of CFU added. Data are representative of three experiments for each condition.

View larger version (35K): [in this window] [in a new window]   Figure 8.   Dose-response relationship for antibacterial activity of Calu-3 cell ASF after culturing with topical corticosteroids. Calu-3 cells were cultured with the indicated concentrations of budesonide (open bars), beclomethasone (light gray bars), or triamcinolone (dark gray bars) for 48 h before the collection of ASF. ASF washings (30 µl) were assayed with 104 E. coli at 37°C for 18 h and recovered CFU plotted as a function of corticosteroid concentration. In the absence of steroid, 1,930 ± 70 CFU were recovered. Data are presented as mean ± SEM for 3 monolayers at each condition. *P  0.05 compared with control.

View larger version (28K): [in this window] [in a new window]   Figure 9.   Dose-response relationship for antibacterial activity of Calu-3 cell ASF after culturing with 2-agonists. Calu-3 cells were cultured with the indicated concentrations of salbutamol (open bars) and terbutaline (shaded bars) for 48 h before the collection of ASF. ASF washings (30 µl) were assayed with 103 E. coli at 37°C for 18 h. Recovered CFU in the absence of 2-agonist was 108 ± 56. Data are presented as mean ± SEM for three monolayers at each condition. *P  0.05 compared with control.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

These studies demonstrate that the ASF of Calu-3 cells has antibacterial activity and that this activity can be altered by drugs that are used clinically to treat airway diseases. We observed that antibacterial activity can be increased by treating Calu-3 cells with the topical corticosteroids budesonide, triamcinolone, and beclomethasone. The effect was most pronounced with budesonide, a newer topical corticosteroid with higher potency. In contrast, treating Calu-3 cells with the ₂-agonists salbutamol or terbutaline leads to a reduction in antibacterial activity. In our assays, the nonsteroidal anti-inflammatory agents ibuprofen and cromolyn sodium had no effect on bacterial killing activity. Because we measured antibacterial activity in an in vitro assay system where ionic composition was controlled, our results suggest that these agents influence ASF antibacterial activity by altering the secretion of antibacterial proteins.

Calu-3 cells, an immortalized airway cell line, were used to facilitate the experimental design and allow a sufficient number of experiments to be performed. Calu-3 cells were selected for study for several reasons. They form high resistance monolayers when grown on permeable supports, which allows for culture at an air-liquid interface (17). Under these conditions Calu-3 cells form a well-differentiated monolayer with epithelial characteristics (16). The cell line has been characterized extensively and shown to express high levels of functional CFTR (18). Calu-3 cells are derived from a human carcinoma and express surface antigens like those of serous cells from the glandular regions of the large airway (16). The submucosal gland is thought to be the major source of airway surface fluid, mucins, and immunologically active substances (22). Both serous cells and Calu-3 cells express and secrete antibacterial peptides. In addition, serous cells in the submucosal glands are the major site of CFTR in the human airway (25), whereas native airway epithelial cells generally express far lower levels of CFTR. For these reasons, Calu-3 cells appear to represent a relevant model system for the study of airway antibacterial defenses.

The antibacterial activity of ASF washings from Calu-3 cells has a number of similarities to previous characterizations of ASF antibacterial activity from primary cultures of airway cells. These similarities include the fact that (1) the antibacterial activity of Calu-3 cell ASF washings is inhibited by elevated concentrations of NaCl and NaCl dependence is a function of protein concentration, (2) ASF antibacterial activity accumulates with time, and (3) antibacterial protein expression is similar to that observed in other airway cell systems. These similarities are significant because the use of immortalized cell lines may make comparisons to the in vivo condition difficult.

Two points need to be emphasized about this data. First, when cultured at an air-liquid interface Calu-3 cells express and secrete antibacterial products into the ASF. Although we have identified mRNA for -defensin-1, -defensin-2, LL-37, lactoferrin, and lysozyme in Calu-3 cells, our studies do not establish the identity of the antibacterial agent(s) in isolated ASF washings. Second, the antibacterial activity of Calu-3 cell ASF can be altered by changes in culture conditions. Culturing cells in the presence of pharmacologically relevant concentrations of topical corticosteroids increased the antibacterial activity of ASF washings, whereas culture with ₂-agonists decreased this activity. Although antibacterial activity has been described for primary cultures of airway epithelial cells, our studies are the first to demonstrate the antibacterial activity of an airway cell line. By demonstrating agonist-dependent regulation of ASF antibacterial activity, it is now possible to investigate this aspect of innate antibacterial activity. Our studies suggest that agonist-dependent regulation is the result of alterations in protein secretion and not necessarily changes in the ionic composition of ASF. However, additional agonist-dependent changes in the ionic composition of ASF cannot be ruled out by these studies.

There are currently two disparate views on the mechanism by which ASF antibacterial activity is altered in CF airway (2). Several studies have suggested that the ASF of CF airway cells has an increased concentration of NaCl and that this inhibits the antibacterial activity of -defensins and proteins such as lysozyme (4, 10, 26). In vitro assays have confirmed that the antibacterial activity of these agents can be inhibited by elevated concentrations of NaCl. However, despite earlier reports of elevated salt in the ASF of the CF airway (8), more recent studies have failed to observe significant differences between the ASF Na and Cl concentrations of CF and normal subjects (7). The second mechanism does not discount a role for antibacterial peptides and proteins, but based on the failure to observe differences in ASF NaCl levels, postulates that mucous dehydration and reduced rates of mucociliary clearance are the primary reasons for increased infection in the CF lung (22). However, both hypotheses predict that an increase in antibacterial protein secretion could be of clinical benefit to patients with CF.

It is of note that topical corticosteroids increased ASF antibacterial activity. This observation is perhaps surprising because -defensin expression has been shown to be regulated by an NF-B-dependent pathway and corticosteroids are known to inhibit NF-B-dependent gene expression (27). However, it should be noted that although small changes in antibacterial peptide message or protein levels cannot be excluded from our experiments, our data failed to observe major changes in -defensin-1 or -defensin-2 mRNA or protein levels when Calu-3 cells were cultured in the presence of budesonide. This would suggest that the topical corticosteroids we studied increase ASF antibacterial activity by the induction of other antibacterial proteins and that the expression of these proteins is activated by steroid-dependent mechanisms that are NF-B independent. This seems to be supported by the changes in protein composition we noted that were limited to proteins with molecular weights of greater than 10 kD. The identity and function of these proteins, however, remains undetermined by these experiments. An NF-B-independent mechanism is also suggested by the observation that ₂-agonists inhibit antibacterial activity, as they are not known to affect the NF-B- dependent pathways (28). Our data suggest that antibacterial ASF proteins are altered by culturing Calu-3 cells in the presence of corticosteroids or ₂-agonists. Studies are presently under way to identify differences in ASF proteins.

Our results suggest that clinical agents can influence the antibacterial activity of ASF in vitro. We have described for the first time the facilitation and inhibition of antibacterial activity by drugs that are commonly used for a variety of airway diseases, including CF. Despite clinical benefits in terms of lung function and inflammatory markers (29, 30), long-term oral corticosteroid therapy in CF can be associated with serious side effects (29, 31). This has lead many physicians to consider the use of inhaled corticosteroids in airway care for patients with CF, and in fact, this is fairly common. It is estimated that at least 12% of patients with CF use these agents to suppress pulmonary inflammation (32). There have been a number of clinical studies evaluating the efficacy of inhaled corticosteroids in CF with mixed results (33). The overall number of patients in these trials has been small, the agents, while including budesonide and beclomethasone, have varied among the studies, and the effect on lung function has not been consistent. Whereas several of these studies examined specific markers of inflammation in CF sputum, usually with encouraging results, specific antibacterial activity has yet to be assessed in a clinical trial. Whether the beneficial effect of inhaled corticosteroids seen in these studies is due to an anti-inflammatory effect alone or due to enhanced innate antibacterial activity is unknown. Similarly, inhaled ₂-agonists are frequently prescribed to patients with CF and no adverse effect on airway infection has been measured. Ibuprofen is used less frequently in CF than steroids (32), despite promising results in a clinical trial (21). We found that ibuprofen had no effect on the antibacterial properties of ASF, either at clinically important (21) or CFTR inhibitory concentrations (20). Assessing the in vivo antibacterial effect of any of these drugs will be complex, but our results suggest that this sort of manipulation may be possible and that the use of such an outcome measure may be clinically relevant.

In summary, we have developed a model to study the antibacterial properties of Calu-3 ASF and have found that our model shares many of the features of primary cultures and airway explants. We have found that pharmaceutical agents used clinically in CF may enhance or attenuate the antibacterial activity of ASF. These results suggest the therapeutic potential for modifying the antibacterial activity of ASF in vivo. In addition, an in vivo method for assaying ASF antibacterial activity may be useful for assessing therapeutic efficacy in CF clinical trials.