Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Several recent studies indicate that pulmonary inflammation may occur early in the course of cystic fibrosis (CF) (1), although the existence of a primary inflammatory disorder in CF is controversial. By bronchoalveolar lavage (BAL), most patients with CF can be shown to have increased numbers of airway neutrophils (1). Very young children, aged 3 mo to 7 yr, with CF who have minimal or no clinical evidence of lung disease have elevated numbers of neutrophils and strikingly increased concentrations of the neutrophil chemoattractant interleukin (IL)-8 and tumor necrosis factor (TNF)- in BAL fluid (BALF) (2). Even newborn infants with CF have increased numbers of alveolar neutrophils and increased concentration of IL-8 in BALF (3). Adult and teenage patients with CF with very mild or no detectable lung disease have increased numbers of neutrophils in BALF, usually in conjunction with small numbers of at least one bacterial species known to be pathogenic in CF (4). Investigators have also found that human bronchial epithelial cells in culture expressing mutant CF transmembrane conductance regulator (CFTR) generate more IL-8 in response to several different Pseudomonas gene products than do wild-type bronchial epithelial cells (5, 6). These studies indicate that neutrophilic inflammation in association with increased IL-8 and TNF- concentrations in the airways is a prominent early feature of CF, and suggest that exuberant airway inflammation is a component of the CF phenotype.

Because IL-8 production is increased in CF airways, we initially investigated production of IL-8 by CF bronchial epithelial cells in culture. We used TNF- to stimulate IL-8 production in three cell lines: transformed human bronchial epithelial cells that express mutant CFTR (IB3 cells), the same cell line stably transfected with wild-type CFTR (C38 cells), and transformed normal bronchial epithelial cells (BEAS cells). After finding exaggerated IL-8 production in the IB3 cells in response to TNF-, we hypothesized that increased IL-8 production is related to upregulation of nuclear factor (NF)-B activation in cells expressing mutant CFTR.

NF-B is a protein factor that is required for maximal transcription of many cytokines and other proinflammatory molecules that are thought to be important in the generation of acute inflammatory responses, including intercellular adhesion molecule, inducible nitric oxide synthase, cyclooxygenase-2, IL-6, and IL-8 (7). NF-B attaches to DNA in the promoter region of target genes as a dimer composed of two Rel family proteins, typically p50 and Rel-A (p65). In the NF-B heterodimer, both subunits contact DNA but only Rel-A contains a transactivation domain in the C-terminal end of the protein that interacts directly with the basal transcription apparatus (10). In quiescent cells, NF-B is sequestered in the cytoplasm by its interaction with a member of the inhibitory IB family that includes IB- and IB-. After cell stimulation, IB- is phosphorylated, polyubiquinated, and degraded by the 26S proteasome. IB- degradation unmasks nuclear localization signals that allow NF-B to be transported to the cell nucleus, where NF-B binds DNA containing the sequence 5'-GGGPuNNPyPyCC-3' and activates gene transcription. Activated nuclear NF-B causes an upregulation of IB- messenger RNA levels by binding to NF-B sites in the IB- promoter (11, 12). The newly synthesized IB- helps terminate the NF-B response by resequestering NF-B in the cytoplasm. IB- exists as a basal phosphorylated form that, like IB-, masks the nuclear localization signals on NF-B. Upon cell stimulation, IB- is polyubiquinated and degraded by the proteasome complex, and is resynthesized as an unphosphorylated (or hypophosphorylated) form (13). Unlike IB- and the basally phosphorylated form of IB-, hypophosphorylated IB- is unable to mask the nuclear localization signal and the DNA binding domain of NF-B (13). Therefore, NF-B bound to hypophosphorylated IB- is protected from inactivation by IB- and can enter or remain in the nucleus and mediate persistent transcriptional activation of proinflammatory genes.

In these studies, we found that TNF- stimulation resulted in increased NF-B activation in CF epithelial cells compared with normal and "corrected" CF epithelial cells. We then investigated the mechanism for this upregulation of NF-B activation by assessing processing of inhibitory components (IB- or IB-) in these cells. Although there were no detectable differences in IB- processing or degradation, IB- levels were higher in CF cells and increased hypophosphorylated IB- was found in CF cells after TNF- stimulation, potentially accounting for the upregulation of NF-B in these cells.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Cell Culture

Three different types of transformed human bronchial epithelial cells were studied: one type that expresses mutant CFTR (IB3), one that expresses wild-type CFTR (BEAS), and a "corrected" CF cell line that was derived from IB3 cells stably transfected with wild-type CFTR (C38) (14, 15). C38 cells display normal chloride conductance. IB3 and BEAS cells are human bronchial epithelial cells transformed using an adenovirus-12 simian virus 40 hybrid virus. IB3 cells are F508 compound heterozygotes (F508/W1282X). BEAS and C38 cells were grown in LHC8 media (Biofluids, Rockville, MD) supplemented with penicillin/ streptomycin, and the IB3 cells were grown in Dulbecco's modified Eagle's medium (DMEM)/F12 (GIBCO BRL, Rockville, MD) supplemented with 10% fetal calf serum, glutamine, and penicillin/streptomycin. Both C38 and BEAS were grown on tissue culture plasticware that was coated with LHC8 media containing albumin, fibronectin, and collagen. Before experimentation, IB3 cells were placed in serum-free DMEM/F12 media for 24 h and C38 and BEAS cells were placed in fresh LHC8 media for 24 h. Stimulation of cells with TNF- was carried out in serum-free media.

Preparation of Nuclear and Cytoplasmic Protein Fractions

Nuclear protein extracts were obtained using a modified Dignam protocol (16). A total of 5 to 10 × 10⁶ cells/sample were lysed with Buffer A (10 mM N-2-hydroxyethylpiperazine-N'-ethane sulfonic acid [Hepes], 1.5 mM MgCl₂, 10 mM KCl, 0.5 mM dithiothreitol [DTT], 1 mM phenylmethylsulfonyl [PMSF], 0.01 mM leupeptin, 0.1 nM pepstatin, and 0.5 mM BME) and 1% Nonidet P-40 (NP-40), followed by vortexing to shear the cytoplasmic membrane. Release of nuclei was monitored during this step by visual inspection under light microscopy (×400). Nuclei were pelleted by centrifugation at 6,000 rpm for 6 min in a microcentrifuge and cytoplasmic extracts were saved. Nuclear pellets were washed with low-salt buffer C (20 mM Hepes, 25% glycerol, 1.5 mM MgCl₂, 0.2 mM ethylenediaminetetraacetic acid [EDTA], 0.5 mM DTT, 1 mM PMSF, 0.01 mM leupeptin, 0.1 nM pepstatin, 0.5 mM BME, and 0.01 M KCl), and nuclear proteins were extracted with two 50-µl volumes of high-salt buffer C (20 mM Hepes, 25% glycerol, 1.5 mM MgCl₂, 0.2 mM EDTA, 0.5 mM DTT, 1 mM PMSF, 0.01 mM leupeptin, 0.1 nM pepstatin, 0.5 mM BME, and 0.42 M NaCl). These high-salt extracts were diluted with 200 µl Buffer D (20 mM Hepes, 19% glycerol, 60 mM KCl, 0.2 mM EDTA, 0.5 mM DTT, 1 mM PMSF, 0.01 mM leupeptin, 0.1 nM pepstatin, and 0.5 mM BME) and frozen at 70° C. Total nuclear and cytoplasmic protein concentrations were determined by the Bradford mini- assay (17).

Electrophoretic Mobility Shift Assays

A consensus double-stranded NF-B oligonucleotide (Stratagene, La Jolla, CA) (5'-GATCGAGGGGACTTTCCCTAGC-3') was used for electrophoretic mobility shift assay (EMSA). End labeling was accomplished by treatment with T4 kinase in the presence of ³²P-adenosine triphosphate. Labeled oligonucleotides were purified on a Sephadex G-25 M column (Pharmacia Biotech, Inc., Piscataway, NJ). Nuclear protein, 5 µg, was added to 15 µl incubation buffer (Stratagene) and incubated on ice for 15 min. Next, approximately 100,000 counts per min labeled double-stranded oligonucleotide was added to each sample. This mixture was incubated at room temperature for 20 min and separated by electrophoresis on a 6% polyacrylamide gel in 1× Tris-boric acid-EDTA buffer. Gels were vacuum-dried and subjected to autoradiography. Cold competition was done by adding 50 ng of specific unlabeled double-stranded probe to the reaction mixture. Nonspecific competition was done by adding 50 ng of unlabeled double-stranded oligonucleotide that does not bind NF-B. Supershift assays for NF-B proteins were done with polyclonal antibodies obtained from Santa Cruz Biotechnology, Inc (Santa Cruz, CA). These antibodies were added to the above reaction mixtures at a concentration of 1 µg/15 µl. The samples were then incubated at room temperature for 1 h before gel loading.

Immunoblots

Cell lysates were obtained by removing media, adding 0.5 ml buffer (0.6% NP-40, 150 mM NaCl, 10 mM Hepes [pH 7.9], 1 mM EDTA, and 0.5 mM PMSF), and lifting cells with a cell scraper. This was followed by vortexing and sonication. Proteins were quantitated by the Bradford assay (17) and 75 µg of protein was mixed with an equal volume of 2× sample buffer and boiled for 5 min. Denatured proteins were separated by electrophoresis on sodium dodecyl sulfate (SDS)-polyacrylamide gel along with molecular weight markers and control HeLa cell protein extracts (New England BioLabs, Beverly, MA). Proteins were transferred to an Immobilon-P transfer membrane (Millipore, Bedford, MA) in 20 mM Tris base, 150 mM glycine, and 20% vol/vol methanol overnight at 40 V. Nonspecific binding was blocked by soaking the membrane overnight at room temperature in Tris-buffered saline (TBS), pH 7.6, with 5% nonfat dried milk and 0.1% Tween-20. Immunoreactive proteins were detected by incubating the filter with specific antibodies (IB- and phosphorylated IB- from New England BioLabs; Rel-A/p65 and IB- antibodies from Santa Cruz Biotechnology) overnight at room temperature with constant agitation. Nonspecific binding was washed away by rinsing the filter in TBS containing 0.1% Tween-20. The filters were incubated with horseradish peroxidase-conjugated antirabbit immunoglobulin (Ig) G (New England BioLabs) diluted 1:2,000 in TBS with 5% nonfat dried milk and 0.1% Tween-20 for 1 h. The filter was washed six times for 10 min with TBS containing 0.1% Tween-20. To develop the image, filters were treated with Renaissance Western Blot luminescent reagent (NEN DuPont, Boston, MA) for 1 min and exposed to enhanced chemiluminescence hyperfilm (Amersham, Arlington Heights, IL) for 10 s to 10 min.

Coimmunoprecipitation

Nuclear extracts were prepared as described earlier, followed by immunoprecipitation with polyclonal antibodies to IB- (sc-945; Santa Cruz Biotechnology). Control rabbit IgG (3 µl) and agarose-conjugated control rabbit IgG (20 µl) (both from Santa Cruz Biotechnology) were added to 250 µg of nuclear protein extract. After incubation and centrifugation, 20 µl of agarose-conjugated IB- antibodies were added to the supernatant, followed by overnight incubation at 4° C. The pellet was collected after centrifugation and washed five times with 100 µl 1× phosphate-buffered saline. The pellet was then resuspended in 1× electrophoresis sample buffer, boiled, and separated by electrophoresis on an 8% SDS-polyacrylamide gel along with molecular weight markers. Proteins were transferred to an Immobilon-P transfer membrane and immunoblotting for Rel-A was done using mouse monoclonal Rel-A antibodies (Transduction Laboratories, Lexington, KY) as described earlier.

Measurements of IL-8 by Enzyme-Linked Immunosorbent Assay

IL-8 protein in cell culture supernatant was determined using sensitive enzyme-linked immunosorbent assay kits obtained from R&D Systems (Minneapolis, MN) according to manufacturer's instructions.

Statistical Analysis

For comparison between groups, the unpaired Student's t test for single comparisons was used. Two-tailed P values of  0.05 were considered significant. Differences between IL-8 production by the three cell types were analyzed by Kruskal-Wallas nonparametric analysis of variance with Dunn's post-test analysis using GraphPad InStat version 3.00 for Windows 95/NT (GraphPad Sofware, San Diego, CA).

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Initial studies were done to examine the effect of the CFTR mutation on the production of IL-8 by bronchial epithelial cells after treatment with TNF- (Figure 1). Culture of cells expressing mutant CFTR (IB3), cells expressing wild-type CFTR (BEAS), or "corrected" CF cells (C38) for 48 h in serum-free media resulted in low levels of IL-8 production (solid bars, Figure 1). Treatment with TNF- (30 ng/ml) for 48 h increased production of IL-8 by all three cell types (Figure 1, striped bars); however, the magnitude of this response was significantly greater in IB3 cells compared with BEAS or C38 cells (P < 0.05). Mean TNF--stimulated IL-8 production at 48 h was less than 10 ng per million cells in BEAS and C38 cells, but increased to 68 ng per million cells in IB3 cells.

View larger version (19K): [in this window] [in a new window]   Figure 1.   Production of IL-8 in cell culture supernatants of IB3, BEAS, and C38 cells. Cells were grown in serum-free media without (solid bars) or with (striped bars) the addition of human TNF- (30 ng/ml) for 48 h. Values represent means ± standard error of the mean (SEM) (n = 4 separate experiments). Very little unstimulated production of IL-8 was detected in any of the cell types. TNF--stimulated production of IL-8 by IB3 cells was significantly increased compared with either BEAS or C38 cells (*P < 0.05 compared with BEAS and C38 cells).

Because IL-8 production is regulated at the level of gene transcription by NF-B (18), we next examined whether NF-B activation in CF cells is demonstrably abnormal. Initially, we evaluated the time course for activation of NF-B in nuclear protein extracts from BEAS and IB3 cells after treatment with TNF- (30 ng/ml). By EMSA, NF-B was found to be activated from 1 to 48 h after TNF- treatment in both cell types (not shown). We then attempted to determine whether there were any quantitative differences in TNF--stimulated NF-B activation between normal (BEAS) and CF (IB3) epithelial cells. Figure 2A compares NF-B activation in BEAS and IB3 cells 4 h after treatment with TNF- (30 ng/ml). We chose to evaluate this time point after the addition of TNF- because maximal NF-B activation was apparent at this time. Each lane represents nuclear protein extracts derived from a separate tissue culture plate in the same experiment. Equal amounts of protein were added in each lane. Neither normal (BEAS) nor CF (IB3) epithelial cell lines exhibited substantial basal NF-B activation when cultured for 24 h in serum-free conditions (controls [con], Figure 2A, lanes 1 and 6); however, both cell lines showed activation of NF-B after treatment with TNF- (30 ng/ml) (Figure 2A, lanes 2-5 and 7-10). NF-B activation in IB3 cells was consistently greater than in BEAS cells. The specificity of protein binding in this EMSA is shown by cold and nonspecific competition (Figure 2A, lanes 11 and 12, respectively). Addition of 50 ng of unlabeled oligonucleotide (cold, C) containing an intact NF-B binding site successfully competed for protein binding and eliminated the NF-B band, but addition of 50 ng unlabeled nonspecific (NS) oligonucleotide did not affect binding. Supershifts (Figure 2A, lanes 13 and 14) showed that the detected protein/ DNA binding contained both components of NF-B (p65 [Rel-A] and p50). Addition of antibodies to p65 (Rel-A) (Figure 2A, lane 13) and p50 (Figure 2A, lane 14) resulted in supershift bands with diminution of the primary band (labeled NF-B).

View larger version (42K): [in this window] [in a new window]   View larger version (48K): [in this window] [in a new window]   View larger version (16K): [in this window] [in a new window]   Figure 2.   EMSAs for NF-B in (A) BEAS and IB3 cells and (B) C38 and IB3 cells. (A) EMSA for NF-B using nuclear protein extracts of BEAS and IB3 cells before (control, con) (lanes 1 and 6) and 4 h after treatment with TNF- (30 ng/ml) (lanes 2-5 and 7- 10). Specificity of detected protein binding is shown by cold (C) and nonspecific (NS) competition. Addition of 50 ng of cold, unlabeled NF-B oligonucleotide containing an intact NF-B binding site eliminated the NF-B band (lane 11), but the addition of nonspecific oligonucleotide did not affect protein binding (lane 12). Supershifts (lanes 13 and 14) show that both components of NF-B are present. Addition of antibodies to p65 (Rel-A) (lane 13) and p50 (lane 14) resulted in supershifted bands with diminution of the primary band. (B) EMSA for NF-B using nuclear protein extracts of C38 and IB3 cells before (control, con) (lanes 1 and 5) and 4 h after treatment with TNF- (30 ng/ml) (lanes 2-4 and 6-8). Cold and nonspecific competition and supershifts are also shown (lanes 9-12). (C) Laser densitometry of the NF-B band on EMSAs from three separate experiments involving IB3, C38, and BEAS cells. The mean density of the NF-B band in the TNF--stimulated IB3 cells in each experiment is arbitrarily assigned as 100%. The densities of NF-B bands in the untreated control cells and TNF--stimulated C38 and BEAS cells are reported as percentages of the mean density of the NF-B band from TNF--stimulated IB3 cells in the same experiment (± SEM). NF-B activations in untreated controls from all three cell types were similar so results were pooled. (*P < 0.05 compared with untreated controls, **P < 0.05 compared with untreated controls and TNF--treated IB3 cells).

We performed additional experiments comparing TNF--stimulated NF-B activation in the CF (IB3) cells and the "corrected" CF (C38) cells (Figure 2B). These experiments were done with the same experimental conditions described earlier and showed that IB3 cells had consistently higher NF-B activation than did C38 cells 4 h after TNF- treatment. Figure 2C summarizes laser densitometry readings on EMSAs from three separate experiments involving IB3, BEAS, and C38 cells. The mean density of the NF-B band in the TNF--stimulated IB3 cells was arbitrarily assigned as 100%. The densities of NF-B bands in untreated cells, TNF--stimulated BEAS cells, and TNF--stimulated C38 cells are reported as percentages of the mean NF-B activation in TNF--stimulated IB3 cells in the same experiment. NF-B activations in untreated control cells from all three cell types were similar so these results were pooled. These findings clearly show that NF-B activation is induced in all three cell lines by incubation with TNF- and that NF-B is activated to a greater extent in CF cells than in normal bronchial epithelial cells or corrected CF cells.

Because NF-B activation is dependent on phosphorylation and degradation of inhibitors (IBs), we next examined the effect of the CF mutation on IB- protein levels in whole-cell lysates from the three cell lines. Figure 3 shows basal levels of IB- in whole-cell lysates of BEAS cells (Figure 3, lanes 3-6), C38 cells (Figure 3, lanes 7-10), and IB3 cells (Figure 3, lanes 11-14). In Figure 3, lanes 1 and 2 contain cell lysates from HeLa cells that were unstimulated (lane 2) or stimulated with TNF- (lane 1) to indicate the position of the IB- protein on the immunoblot. No differences in basal IB- levels were detected.

View larger version (25K): [in this window] [in a new window]   Figure 3.   Basal IB- levels detected by immunoblot in whole-cell lysates of unstimulated BEAS, C38, and IB3 cells. HeLa cell control extracts with (+) or without () TNF- stimulation were used as controls (lanes 1 and 2). This immunoblot is representative of three separate experiments. No differences in basal IB- levels were identified.

Figure 4 illustrates IB- phosphorylation, degradation, and resynthesis after treatment with 30 ng/ml of TNF- in whole-cell lysates of C38 and IB3 cells using antibodies specific for the phosphorylated (Figure 4, upper panel) or unphosphorylated (Figure 4, lower panel) form of IB-. C38 cells are represented in lanes 3-8 and IB3 cells in lanes 9-14. In both cell types, phospho-IB- is undetectable at baseline (Figure 4, lanes 3 and 9, upper panel) but is rapidly upregulated by 10 min after the addition of TNF- (Figure 4, lanes 4 and 10). By 30 min, IB- is almost completely degraded in both cell types (Figure 4, lanes 5 and 11 in upper and lower panels). By 60 min, resynthesis of IB- is apparent (Figure 4, lanes 6 and 12), and similar amounts of phosphorylated and unphosphorylated IB- are present in both cell types from 60 to 240 min after TNF- is added. In this figure, slightly more IB- is shown at 240 min in the IB3 compared with the C38 cells; however, this finding was not borne out in other experiments. No differences between the cell lines were identified in TNF--induced processing or resynthesis of IB-.

View larger version (34K): [in this window] [in a new window]   Figure 4.   Immunoblot showing processing of IB- after treatment with TNF- (30 ng/ml) in whole-cell lysates of C38 and IB3 cells. Detection was done using antibodies specific for the phosphorylated form of IB- (upper panel) or the unphosphorylated form of IB- (lower panel). Control HeLa cell protein extracts, both positive (TNF--stimulated) and negative (unstimulated), are shown (lanes 1 and 2, respectively). No phosphorylated IB- is identified in untreated cells (lanes 3 and 9), but increased phospho-IB- is seen at 10 min after TNF- (lanes 4 and 10) in both C38 and IB3 cells. By 30 min, IB- is nearly completely degraded (lanes 5 and 11) and there is recovery of IB- at 60 to 240 min (lanes 6-8 and 12-14) in both cell types.

In addition to identifying the basal levels of IB- and the kinetics of IB- processing after TNF- stimulation in CF and corrected CF bronchial epithelial cells, we wanted to determine whether there were differences in signal-induced phosphorylation or proteasome processing of IB-. To assess these parameters, we utilized a proteasome inhibitor, MG-132 (Sigma, St. Louis, MO), that blocks degradation but not phosphorylation of IB- (19). Figure 5 shows the accumulation of phospho-IB- in whole-cell lysates of C38 cells and IB3 cells after treatment with MG-132 and TNF-. Cells were treated with MG-132 (10 µM) for 2 h before stimulation with TNF- (30 ng/ml). Immunoblots were done for phospho-IB-. Maximal phospho- IB- was identified by 10 min after TNF- treatment, with persistence of the band for up to 4 h. Some diminution of the phospho-IB- band was seen in both cell types between 2 and 4 h. There was no difference in maximal phospho-IB- accumulation at 10 min after TNF-, implying that TNF--induced IB kinase activity is similar in both cell lines. We also assessed whether there were any differences in the concentration of MG-132 that was required to inhibit proteasome function in each cell type. We found that 1 µM of MG-132 almost completely inhibited breakdown of IB- in both cell types (data not shown). No differences were found between the cell lines in signal-induced phosphorylation of IB- or proteasome function that could account for the exaggerated NF-B activation observed in CF cells after treatment with TNF-.

View larger version (15K): [in this window] [in a new window]   Figure 5.   Immunoblot showing time course for accumulation of phospho-IB- in whole-cell lysates of C38 and IB3 cells treated for 2 h with the proteasome inhibitor MG-132 (10 µM) followed by TNF- (30 ng/ml). Detection was done using antibodies specific for the phosphorylated form of IB-. HeLa cell control extracts with (+) or without () TNF- stimulation were used as controls. This immunoblot is representative of three separate experiments. No differences are identified in the level of phospho- IB- after TNF- treatment.

We examined the amount of total cellular Rel-A in IB3 and C38 cells to determine the maximum transactivation potential that could be attributed to NF-B in these cell types. By laser densitometry of Rel-A bands on immunoblots from four separate experiments, 25% more Rel-A antigen was identified in IB3 cells compared with C38 cells, and this concentration of Rel-A in the whole-cell lysates was not affected by treatment with TNF- (data not shown). Therefore, a slightly greater pool of Rel-A is available for activation in IB3 cells compared with C38 cells.

Because more Rel-A is present in IB3 cells than in C38 cells, this component of NF-B must be bound to an inhibitory subunit in the quiescent state. Although IB- levels are similar in the two cell lines, we found that more IB- was present in IB3 compared with C38 cells. Figure 6 shows basal levels of IB- in both cell types as well as the time course for degradation of IB- after treatment of the cells with TNF-. More IB- is present at baseline in the IB3 cells and at each time point up to 240 min after addition of TNF-. IB- levels decreased substantially in C38 cells after 30 min, with a sustained loss to 240 min. In contrast, the loss of basal IB- was blunted in IB3 cells, with substantial antigen detection between 30 and 240 min. More basal IB- was consistently found in lysates of IB3 cells compared with C38 (or BEAS) cells.

View larger version (11K): [in this window] [in a new window]   Figure 6.   Immunoblot showing processing of IB- after treatment with TNF- (30 ng/ml) in whole-cell lysates of IB3 and C38 cells. Detection was done using antibodies specific for the IB-. More IB- is present in IB3 cells at baseline and 10 to 240 min after TNF- compared with C38 cells. This immunoblot is representative of three separate experiments which show that IB- levels at baseline and after TNF- stimulation are higher in IB3 cells compared with C38 cells.

IB- can exist in eukaryotic cells as both a basally phosphorylated form and a hypophosphorylated form. The hypophosphorylated form may be responsible for prolonged activation of NF-B in certain circumstances by protecting activated NF-B from inactivation by IB- (13, 20). Therefore, we thought that the differences we observed in IB- in CF cells could be due to altered production of hypophosphorylated IB-. The presence of increased amounts of hypophosphorylated IB- could help explain the exaggerated NF-B activation and IL-8 production that we observed in these cells. We performed immunoblots for IB- after maximally separating proteins from whole-cell lysates on a 7% polyacrylamide gel and found that we could resolve the detected IB- as two separate bands representing the basally phosphorylated form (Figure 7A, upper band) and the hypophosphorylated form (Figure 7A, lower band). In this figure, which is representative of three separate experiments, unstimulated cells (both C38 and IB3) have very little hypophosphorylated IB-. IB3 cells began to accumulate hypophosphorylated IB- by 30 min after treatment with TNF-, and by 120 to 240 min after TNF- the amount of hypophosphorylated IB- approximated the level of basally phosphorylated IB-. In contrast, C38 cells had much less hypophosphorylated IB- at the 120- and 240-min time points compared with IB3 cells. To show that the hypophosphorylated IB- we have identified after TNF- stimulation represents new synthesis of IB- and not dephosphorylation of basally phosphorylated IB-, we treated IB3 cells with cyclohexamide (CHX) (5 µg/ml) for 30 min before the addition of TNF- and obtained whole-cell lysates 4 h later. Figure 7B shows that minimal hypophosphorylated IB- is present in unstimulated cells, the presence of this band is increased 4 h after TNF-, and pretreatment with CHX largely eliminates the band representing hypophosphorylated IB-. In normal bronchial epithelial cells (BEAS), basal levels of IB- and the appearance of hypophosphorylated IB- after treatment with TNF- were similar to C38 cells (not shown). Differences in hypophosphorylated IB- between the cell lines were present for 48 h after the addition of TNF-. In addition, pretreatment with MG-132 completely blocked degradation of IB- in both IB3 and C38 cells (not shown). Because increased amounts of newly formed, hypophosphorylated IB- were present in CF cells after TNF- treatment, this could explain, at least partially, the exaggerated NF-B activation in these cells.

View larger version (15K): [in this window] [in a new window]   View larger version (23K): [in this window] [in a new window]   Figure 7.   (A) Immunoblot showing processing of IB- and accumulation of hypophosphorylated IB- after treatment with TNF- (30 ng/ml) in whole-cell lysates of IB3 and C38 cells. Proteins were maximally separated on a 7% polyacrylamide gel to identify the phosphorylated form of IB-. The upper unlabeled band is nonspecific. Basally phosphorylated IB- and hypophosphorylated IB- are labeled. Accumulation of hypophosphorylated IB- is identified by 30 min after TNF- stimulation and remains detectable to 240 min. Substantially more hypophosphorylated IB- is present at 120 to 240 min after TNF- in IB3 cells compared with C38 cells. This immunoblot is representative of three separate experiments. (B) Immunoblot showing that treatment with CHX (5 µg/ml) for 30 min before TNF- inhibits formation of hypophosphorylated IB- in IB3 cells. Untreated control cells (con, left two lanes) show no detectable hypophosphorylated IB-. TNF- treatment for 4 h increases formation of hypophosphorylated IB-, which is detected just beneath the basally phosphorylated IB- band (TNF, middle two lanes). Pretreatment with CHX (TNF & CHX, right two lanes) does not block degradation of IB- but inhibits production of the hypophosphorylated form.

If hypophosphorylated IB- functions to allow prolonged, exaggerated NF-B activation in CF cells after treatment with TNF-, then IB- should be detectable in nuclear protein fractions bound to NF-B proteins. Therefore, we performed coimmunoprecipitation experiments to evaluate for the presence of nuclear IB- bound to Rel-A. Using nuclear protein extracts obtained 4 h after TNF- stimulation, immunoprecipitation was done with antibodies to IB-, followed by immunoblotting for Rel-A (Figure 8A). More Rel-A/IB- was identified in IB3 cells in the nuclear compartment compared with C38 and BEAS cells. IB--bound Rel-A was only minimally detectable in nuclear extracts of untreated cells (Figure 8B), consistent with the low levels of NF-B activation detected in this setting. These findings support the hypothesis that TNF- stimulation in IB3 cells results in accentuated NF-B activation through increased production of hypophosphorylated IB-, which binds Rel-A in the nucleus and prevents NF-B from inactivation by newly formed IB-.

View larger version (36K): [in this window] [in a new window]   Figure 8.   Coimmunoprecipitation experiment using nuclear protein extracts from IB3, C38, and BEAS cells. Immunoprecipitation was done with antibodies to IB-, followed by separation of proteins on an 8% polyacrylamide gel and immunoblotting with antibodies to Rel-A. (A) Nuclear extracts were obtained from cells 4 h after TNF- treatment. HeLa cell lysate was run as a control. Rel-A and nonspecific (NS) bands are identified by arrows. (B) Nuclear extracts were obtained from untreated IB3 cells and IB3 cells harvested 4 h after TNF- treatment.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

In these studies, we found that CF bronchial epithelial cells and normal bronchial epithelial cells cultured in serum-free media demonstrated little NF-B activation and IL-8 production in the absence of an inflammatory stimulus. TNF- stimulation resulted in upregulation of NF-B activation and IL-8 generation in all three epithelial cell lines studied, but CF cells showed augmented responses compared with the other cell types. A mechanism for this abnormal NF-B activation in CF appears to be increased generation of hypophosphorylated IB-, which functions to prolong and enhance NF-B activation. We could find no differences in IB- levels, processing, or degradation that could account for our findings. We did find slightly higher Rel-A levels in CF cells compared with the other cell types, which may reflect the increased basal IB- in these cells.

Our finding of exaggerated NF-B activation in CF epithelial cells differs somewhat from a recent report by DiMango and colleagues (21), who found increased basal and stimulated NF-B activation in CF epithelial cells with the F508 mutation compared with normal epithelial cells and C38 cells. In those studies, cell stimulation was done by exposing the cells for 1 h to cytokines (TNF- and IL-1) or Pseudomonas aeruginosa, whereas in our study TNF- was continuously present in the cell culture medium. Differences in duration of stimulation by TNF- may also account for the differences in our findings and those of Black and associates (22), who found no differences in IL-8 production in cultured nasal epithelial cells or transformed airway epithelial cells from CF and normal patients. In addition, the study by DiMango and coworkers (21) was done entirely in the presence of serum-containing media, but our experiments were done in serum-free media. Studying IB3 cells in serum-containing media significantly increases basal IL-8 production (A. Stecenko, unpublished observation), and may account for the differences in NF-B activation found in unstimulated IB3 cells in these studies. Because IB3 cells grow better in serum-containing media and C38 and BEAS cells grow best in LHC8 media without serum, we grew cells under these conditions; however, we changed IB3 cells to serum-free media for our studies to adjust for the effects of serum. Growth of IB3, BEAS, and C38 cells in different serum conditions does not account for the upregulation of IL-8 production in IB3 cells after TNF- treatment. We have found that similar upregulation of IL-8 production occurs in IB3 cells cultured for longer periods in serum-free LHC8 medium (data not shown). Culture of BEAS and C38 cells in LHC8 media with serum did not increase IL-8 production after TNF- stimulation (not shown).

One potential mechanism for alteration in NF-B activation in CF is that processing of mutant, misfolded proteins (including F508 CFTR) has been shown to cause protein accumulation in the endoplasmic reticulum (ER), resulting in an "ER-overload response" (23, 24). Pahl and Baeuerle (24) showed that induction of ER overload in HeLa and 293 cells promoted NF-B activation; however, we found little activation of NF-B in unstimulated IB3 cells, suggesting that ER overload alone is not sufficient to produce NF-B activation in CF epithelial cells under the culture conditions that we used. In fact, we found no differences in IB- processing or proteasome activity in CF cells treated with TNF-. Whether altered regulation of IB- is specific for the CFTR mutation or related to ER overload in general is currently unknown.

Although TNF- treatment has been reported to cause degradation of IB- but not IB- in Jurkat cells (25), Weil and colleagues have shown that E29.1 T-cell hybridomas respond to treatment with TNF- by degradation of IB-, followed by resynthesis in a hypophosphorylated form (26). In 70Z/3 pre-B cells stimulated with lipopolysaccharide, hypophosphorylated IB- binds NF-B but does not prevent entry into the nucleus or binding to DNA (13). In A549 cells, production of hypophosphorylated IB- has been implicated as a cause of persistent NF-B activation after respiratory syncytial virus infection (20). Our findings show that TNF- treatment results in IB- degradation and accumulation of a hypophosphorylated form in all three bronchial epithelial cell lines, which is associated with prolonged activation on NF-B in all these cell lines. However, CF epithelial cells have upregulation of basal IB-, augmented production of hypophosphorylated IB-, and increased nuclear IB- after TNF- treatment compared with normal and corrected epithelial cells. Increased accumulation of hypophosphorylated IB- in CF cells could facilitate increased NF-B activation by protecting a larger fraction of NF-B that is liberated after degradation of IB- and IB- from being deactivated by newly formed IB-. Our studies were done with only a single CF cell line, so it remains to be determined whether these findings are applicable to other types of CFTR mutation or to humans with CF.

CF is a debilitating pulmonary disorder that results in chronic airway infection and inflammation. The exact linkage between expression of mutant CFTR and airway inflammation has not been established, but our studies imply that there may be a primary abnormality in the production of inflammatory mediators in this disease. Our findings are supported by several clinical studies in patients with CF that have shown that inflammation in the alveolar space is a prominent early feature of CF and may precede bacterial colonization of the airways (1). Additional support for the idea that abnormal NF-B activation is present in patients with CF is presented in a recent report by Tabary and associates, who found that cultured bronchial submucosal glands obtained from CF patients undergoing lung transplant exhibited elevated basal NF-B activation and IL-8 production (27).

In summary, CF bronchial epithelial cells have exaggerated activation of NF-B and increased production of IL-8 after treatment with TNF-. These alterations are associated with changes in IB- regulation, which provides a clue to the mechanism of the abnormal inflammatory mediator production by these cells. These findings provide evidence for a primary abnormality in the regulation of inflammatory mediator production in CF epithelial cells.