Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

The airway epithelium is composed of several cell types including ciliated, basal, and goblet cells, which are exposed to a myriad of external stimuli. Airway epithelial cells have classically been thought of as barrier cells that are involved in homeostasis and respond to a variety of environmental stimuli, resulting in the alteration of their cellular functions, such as ion transport and movement of airway secretions. Recent evidence, however, suggests that airway epithelial cells may also act as effector cells in response to noxious endogenous or exogenous stimuli; several studies have shown that they produce and secrete a variety of inflammatory mediators including lipid mediators (8, 14), oxygen radicals (2, 17), tumor necrosis factor- (TNF-) (10), granulocyte-macrophage colony-stimulating factor (GM-CSF) (9), and interleukin-6 (IL-6) (4, 22), and chemokines such as IL-8 (4, 26), regulated on activation, normal T-cell expressed and secreted (RANTES) (18, 35), and others. The production of such inflammatory mediators implicates epithelial cells in the initiation and/or exacerbation of airway inflammatory diseases such as asthma and cystic fibrosis.

At present, inhaled and intranasal steroids are among the most efficacious treatments for airway inflammatory diseases. However, the mechanisms by which glucocorticoids mediate their anti-inflammatory effects are not well understood. Several studies have shown that glucocorticoids regulate the transcription of numerous genes involved in inflammation. Glucocorticoids may upregulate the transcription of hormone-inducible genes through binding of the activated glucocorticoid receptor (GR) to specific DNA sequences termed glucocorticoid response elements (GREs), located in the promoter region of the target genes (39). Alternatively, the activated GR may bind to negative GREs (nGREs) (31) and/or interact with transcription factors such as nuclear factor-B (NF-B) (29) and activator protein-1 (AP-1) (15, 33, 41) to downregulate gene expression. In addition, glucocorticoids may decrease the stability of specific mRNA molecules, thereby modulating gene transcription (16). Although there is evidence for glucocorticoid-induced anti-inflammatory proteins, including lipocortin (13) and neutral endopeptidase (6), it is now appreciated that the ability of glucocorticoids to suppress the transcription of genes involved in inflammation may be their primary activity (38). Notable among glucocorticoid-suppressed genes are those for collagenase (15, 33, 41), inducible nitric oxide synthase (11, 28), cyclooxygenase-2, (25), and most cytokine and chemokine genes (32).

DNA differential display analysis (DD-PCR), developed by Liang and Pardee (23), is a powerful new technology that permits the identification and cloning of differentially expressed genes on a global scale. This method involves the reverse transcription of mRNAs with oligo-dT primers anchored to the 5' end of the poly(A) tail, followed by amplification in the presence of random primers, thereby generating double-stranded complementary DNA (cDNA) products of varying lengths. By changing primer combinations, up to 15,000 individual mRNA species from a mammalian cell may by visualized (23). DD-PCR has several advantages over existing methods, including subtractive hybridization (23); DD-PCR is very sensitive, requiring small amounts of total RNA for the visualization of expressed mRNA species in a mammalian cell line; and it is highly reproducible, with approximately 95% of the bands reproduced from one run to another. Moreover, this method allows simultaneous analysis of multiple RNA samples per run, thereby revealing genes unique to a given cell type or cellular process.

In the present study, DD-PCR was utilized to identify genes expressed in human airway epithelial cells that are relevant to airway inflammation and glucocorticoid action. In particular, genes induced in airway epithelial cells as a result of cytokine stimulation, and for which such induced expression was inhibited by glucocorticoids, were targeted. Using this approach, two candidate TNF--inducible and dexamethasone (DEX)-sensitive gene products were identified, including candidates with significant homology to the rhoC small monomeric G protein and the UFO receptor (a novel tyrosine kinase receptor). Putative roles for rhoC and the UFO receptor in airway inflammation are discussed.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Cell Culture and RNA Isolation

The following human cell types were employed: the airway epithelial cell lines IB3-1 (a gift from Dr. Pam Zeitlin, Johns Hopkins University, Baltimore, MD) (42) and BEAS2B (30); cultured normal bronchial epithelial cells (NHBE; Clonetics Corp., San Diego, CA) (20); cultured synovial fibroblasts (a gift from Dr. Joan Bathon, Johns Hopkins University) (3); cultured human umbilical vein endothelial cells (HUVEC) (5); the T-lymphocyte cell line Jurkat (40); the monocyte cell line U937 (36); and the basophil leukemia cell line KU812 (12). All airway epithelial cells, including IB3-1, BEAS2B, and NHBE, were grown in Vitrogen 100 (Collagen Corp., Palo Alto, CA)- coated-flasks. IB3-1 and BEAS2B were cultured in LHC-8 (Biofluids, Inc., Rockville, MD) containing 5% fetal calf serum (FCS), 1% penicillin-streptomycin, and 0.2% fungizone. NHBE were cultured in serum-free bronchial epithelial cell growth medium (BEGM) (supplemented [LHC-9]; complete ingredients available from Clonetics Corp.). All other cell types were cultured in RPMI (Paragon, Inc., Baltimore, MD) containing 10% FCS, 1% penicillin-streptomycin, 0.2% fungizone, and 1% glutamine. Each cell type, adherent cells at 80% confluence and nonadherent cells at 10⁶ cells/ml, was cultured in the presence of TNF- (100 ng/ml; R&D Systems, Minneapolis, MN) with or without the glucocorticoid DEX (1 µM; Sigma Cemical Co., St. Louis, MO) or an equivalent amount of the carrier control dimethyl sulfoxide (DMSO) (Sigma Chemical Co.) for 20 h at 37°C. Following culture, cells were lysed and total RNA was isolated via the RNAzol-B method (Tel-Test Inc., Friendswood, TX) (7).

DNA Differential Display Analysis

Total RNA was treated with deoxyribonuclease I (DNase I) (1 U/µg RNA; Gibco BRL, Grand Island, NY) for 15 min at room temperature; the reaction was stopped with addition of 20 mM ethylenodiamine tetraacetic acid (EDTA) and subsequent heating to 65°C for 10 min. DNAse-treated RNA samples (0.2 µg/reaction) were then reverse transcribed in the presence of the oligo-dT primer T₁₂MN (1 µM; GenHunter Corp., Brookline, MA) (23) where M represents a mixture of the base G, A, or C, and N represents the base G, A, T, or C, Moloney murine leukemia virus (MMLV) reverse transcriptase (100 U/reaction; GenHunter, Inc.); and deoxynucleotide triphosphates (dNTPs) (20 mM; GenHunter Corp.). Reactions were incubated at 37°C for 1 h and then at 95°C for 5 min to inactivate the enzyme. Resulting cDNA products (2 µl/reaction) were amplified via PCR (Hybaid automatic thermocycler; Labnet, Woodbridge, NJ) in the presence of two different 10-mer "random" primers (GenHunter Corp.) that bind randomly to DNA. Amplification reactions contained AmpliTaq polymerase (1 U/reaction; Perkin-Elmer Corp., Norwalk, CT), dNTPs (2 µM; GenHunter Corp.), AP primers 1 to 10 (0.2 µM; GenHunter Corp.), and [-³⁵S]deoxyadenosine triphosphate ([-³⁵S]-dATP) (1,200 Ci/mmole; New England Nuclear/Dupont, Inc., Boston, MA). Amplification reactions were cycled at 94°C for 30 s, 40°C for 2 min, and 72°C for 30 s for 40 cycles; reactions then underwent a final extension at 72°C for 5 min. All reactions were performed in duplicate. PCR products were visualized on a 6% DNA sequencing gel, and selected bands were precipitated. Briefly, bands were cut from the gel, soaked in 100 µl of distilled water (dH₂O) for 15 min at room temperature, and then heated to 100°C for 15 min. DNA was precipitated in the presence of sodium acetate (60 µM), glycogen (100 µg/ml; GenHunter Corp.), and 100% ethanol; samples were set on dry ice for 30 min and DNA was then pelleted and washed with 85% ethanol. Each sample was reamplified with the respective primer pair as described earlier.

T-A Cloning

Amplified cDNA products (ESTs) were electrophoresed in a 1% agarose gel and purified with the QIAquick Gel Extraction Kit (Qiagen Corp., Chatsworth, CA). cDNA fragments were then cloned into the pGEM-T vector using the pGEM T-A Cloning Kit (Promega Corp., Madison, WI). Briefly, cDNA fragments were ligated into the pGEM-T vector with T4 DNA ligase (1 U/reaction) at an insert-to-vector ratio of 10:1. JM109 cells (Promega Corp.) were transformed with the plasmid and plated on Luria broth (LB)-agar containing ampicillin (100 µg/ml; Sigma Chemical Co.), X-gal (20 µg/ml; Sigma Chemical Co.), and isopropylthiogalactopyranoside (IPTG) (0.1 mM; Sigma Chemical Co.). Positive EST-pGEM clones were selected via blue-white screening and expanded in ampicillin-containing (100 µg/ml) LB; plasmid DNA was isolated with the Perfect Prep Kit (5 Prime-3 Prime, Inc., Boulder, CO).

False-positive Analysis

Ten positive pGEM clones for each EST were prepared for dideoxy sequencing of one nucleotide base, using the DNA Sequencing Kit with Sequenase (United States Biochemical-Amersham Co., Arlington Heights, IL), and were analyzed on a 6% DNA sequencing gel. Sequences were then compared, and the majority of identical sequences was considered to be the true-positive EST.

Northern Blot Analysis

Cells were cultured as described earlier and poly-A⁺ RNA was isolated with the Micro-Fast Track Kit (Invitrogen, San Diego, CA). Two micrograms of poly-A⁺ RNA were electrophoresed in a 1% agarose gel and transferred to a nylon filter membrane (GeneScreen Plus; New England Nuclear-Dupont, Inc.). The RNA was then hybridized with a [-³²P]dATP-labeled EST fragment, utilized as a cDNA probe (5 × 10⁵ cpm/ml hybridization buffer; Random Label Kit, Gibco BRL Life Tech., Inc.). The hybridization buffer contained 50% formamide, 2X 1,4-piperazinediethanesulfonic acid (PIPES), 0.5% sodium dodecyl sulfate (SDS) and 100 µg/ml of sheared, heat-denatured salmon sperm DNA (Sigma Chemical Co.). Autoradiograms of the filter were analyzed via densitometry.

Automated DNA Sequencing and BLAST Analysis

The cloned EST fragments were sequenced on both strands, using an automated ABI 373A sequencer according to the manufacturer's procedures and chemistry. Alignments were done with computer software designed by the Genetics Computer Group (University of Wisconsin, Madison, WI).

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

DNA Differential Display Analysis

To identify novel genes associated with airway inflammation, the IB3-1 cell line, an airway epithelial cell line derived from a cystic fibrosis patient and immortalized with the adenovirus 12-SV40 viral construct (42), was cultured in the presence of TNF- (100 ng/ml) with or without DEX (1 µM) or an equivalent amount of the carrier control DMSO for 20 h at 37°C; all cultures were done in duplicate. TNF- was chosen as a stimulus for two reasons: (1) TNF- is a cytokine associated with airway inflammation (21); and (2) TNF--induced signals can be almost completely suppressed by DEX (35). These studies were performed with cells treated for 20 h, since a number of TNF--inducible, DEX-sensitive molecules, such as the chemokine RANTES, are expressed maximally at this duration in airway epithelial cells (35). Following culture, total RNA was isolated from the IB3-1 cells, treated with DNase (1 U/reaction) to remove contaminating DNA, and then analyzed with DNA differential display analysis, as described in MATERIALS AND METHODS, for mRNAs that were differentially expressed in treated and untreated cells (Figure 1).

View larger version (146K): [in this window] [in a new window]   Figure 1.   DNA differential display analysis of mRNAs associated with airway inflammation in IB3-1 cells. IB3-1 cells were cultured in the presence of TNF- (100 ng/ml), DEX (1 µM), or an equivalent amount of DMSO for 20 h at 37°C. Total RNA was isolated and prepared for DD-PCR as described in MATERIALS AND METHODS. The arrowhead indicates the chosen EST induced by TNF- and suppressed by DEX; marks around the EST indicate that the band was removed from the gel for elution. Each reaction was run in duplicate. This autoradiogram is representative of the DD-PCR analysis and EST identification performed.

False-positive and Northern Blot Analyses

Expressed sequence tags (ESTs) representing mRNAs that were upregulated in the presence of TNF- alone and eliminated in the presence of TNF- and DEX were eluted from the gel and reamplified via PCR with the respective 10-mer primer pairs. With this approach, 19 ESTs were identified and successfully reamplified. Following reamplification, ESTs were subcloned into the pGEM-T vector, and positive clones were selected and expanded in order to obtain large amounts of plasmid DNA in preparation for false-positive analysis. Briefly, 10 positive pGEM clones for each EST were prepared for dideoxy sequencing of one nucleotide base, using the DNA Sequencing Kit with Sequenase (United States Biochemical-Amersham Co.), and were analyzed on a 6% DNA sequencing gel (data not shown). Sequences were then compared; the majority of identical sequences was considered to be the true-positive EST.

The regulation of the ESTs by cytokines and glucocorticoids was confirmed through Northern blot analysis. This was essential, as it established that a given candidate gene truly fit the requirements of cytokine inducibility and glucocorticoid sensitivity while at the same time indicating the size of the EST mRNA and its relative levels of expression. To this end, IB3-1 cells were cultured in the presence of TNF-, DEX, or an equivalent amount of DMSO for 20 h at 37°C, as described previously. Following culture, poly-A⁺ RNA was isolated from these cells and prepared for Northern blot analysis. Poly-A⁺ RNA comprises 1 to 2% of total RNA; therefore, the Northern blot analysis of microgram quantities of poly-A⁺ RNA, rather than total RNA, increases the sensitivity of this assay by at least 10-fold. Northern blots were probed with ESTs that had been amplified via PCR from true-positive EST-pGEM clones and labeled randomly. Although 19 ESTs were identified originally via DD-PCR, only EST8 and EST19 yielded the expected pattern of expression by Northern blot analysis; of the remaining ESTs, six were expressed in the presence and absence of TNF- and/or DEX and 11 were not inducible by TNF-. TNF- increased the level of mRNA expression of EST8 (34%, P  0.005, n = 5) and EST19 (41%, P  0.01, n = 5), whereas DEX reduced this expression to resting levels or below (Figure 2). To determine the time point of maximal EST8 and EST19 mRNA expression, IB3-1 cells were cultured in the presence of TNF-, DEX, or DMSO, as described previously, for 30 min and 1, 2, 6, 12, and 20 h at 37°C; poly-A⁺ RNA was isolated and analyzed via Northern blot analysis with the respective EST probes. In this assay, 20 h after culture was the time of maximum expression for both EST8 and EST19 in IB3-1 cells (n = 3, data not shown).

View larger version (49K): [in this window] [in a new window]   Figure 2.   Northern blot and densitometry analyses of EST8 and EST19 in IB3-1 cells. IB3-1 cells were cultured in the presence of TNF-, DEX, or DMSO as described in Figure 1. Poly-A+ RNA was isolated and prepared for Northern blot analysis; blots were probed with randomly labeled amplified cDNA fragments specific for either EST8 (A) or EST19 (B). Densitometry data are expressed relative to DMSO controls after normalization to 28S RNA (EST8: *P < 0.005, n = 5; EST19: *P < 0.01, n = 5).

Again using Northern blot analysis, the expression of EST8 and EST19 was examined in various cell types such as endothelial cells, lymphocytes, monocytes, fibroblasts, and basophils, as well as other airway epithelial cells, including BEAS2B and NHBE cells. All cells were treated with TNF- (100 ng/ml), DEX (1 µM), or DMSO for 20 h at 37°C, as described previously, and prepared for poly-A⁺ RNA isolation and Northern blot analysis. NHBE cells yielded the expected pattern of both EST8 (21%, P  0.009, n = 3) and EST19 (11%, P  0.02, n = 3) expression (Figure 3), whereas KU812 gave the expected pattern of EST8 expression (34%, P  0.01, n = 3) only (Figure 4). Although the remaining cell types expressed both EST8 and EST19, TNF- did not induce the expression of these ESTs in the other cell types tested (n = 3 for each; data not shown).

View larger version (14K): [in this window] [in a new window]   Figure 3.   Densitometry analysis of EST8 and EST19 mRNA expression in normal human bronchial epithelial cells. Normal human bronchial epithelial cells (NHBE) cells were cultured in the presence of TNF- (100 ng/ml), DEX (1 µM), or an equivalent amount of DMSO for 20 h at 37°C. Poly-A+ RNA was isolated and prepared for Northern blot analysis; blots were probed as described in Figure 2 for EST8 (A) or EST19 (B). Densitometry data are expressed relative to DMSO controls after normalization to 28S RNA (EST8: *P < 0.009, n = 3; EST19: *P < 0.02, n = 3).

View larger version (15K): [in this window] [in a new window]   Figure 4.   Densitometry analysis of EST8 mRNA expression in KU812 cells. KU812 cells were cultured in the presence of TNF- (100 ng/ml), DEX (1 µM), or an equivalent amount of DMSO for 20 h at 37°C. Poly-A+ RNA was isolated and prepared for Northern blot analysis; blots were probed as described in Figure 2 for EST8. Densitometry data are expressed relative to DMSO controls after normalization to 28S RNA (EST8: *P < 0.01, n = 3).

DNA Sequencing and BLAST Analyses

Following Northern blot analysis, EST8 and EST19 were sequenced on both strands with an automated ABI 373A sequencer. DNA sequences were then compared with public sequence databases, including GenBank and EMBL, using basic local alignment search tool (BLAST) algorithms (1), to determine whether any of the ESTs shared homology at the nucleotide level with previously identified genes. With this approach, it was determined that EST8 has virtual identity to the small, monomeric G protein rhoC (100% homology, P = 1.6 × 10¹⁰⁰, Figure 5a), which appears to be involved in focal adhesion and cytoskeletal reorganization, (27, 37), whereas EST19 exhibited homology with the human UFO tyrosine kinase receptor (86% homology, P = 5.3 × 10²⁸, Figure 5b) (34).

View larger version (55K): [in this window] [in a new window]   Figure 5.   Sequence analysis of EST8 and EST19. EST8 (A) and EST19 (B) were sequenced on both strands with an automated sequencer and then analyzed for homology with known DNA sequences using the BLAST algorithm. Each sequence is aligned with the most homologous sequence as determined with BLASTN.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

In order to uncover new information about the mechanisms of airway inflammation and glucocorticoid action in humans, we utilized DD-PCR to identify genes in human airway epithelial cells that are relevant to inflammatory disease and glucocorticoid action. In particular, genes that are expressed in airway epithelial cells as a result of cytokine stimulation, and for which such induced expression is inhibited by glucocorticoids, were targeted. Through this analysis, two ESTs, EST8 and EST19, were identified, subcloned, and sequenced from the airway epithelial cell line IB3-1. EST8 and EST19 displayed the expected pattern of mRNA expression in IB3-1 cells as detected via Northern blot analysis: TNF- increased the level of mRNA expression for EST8 (34%, P  0.005) and EST19 (41%, P  0.01), whereas DEX reduced this expression to the resting level (Figure 2). This same pattern of EST8 (21%, P  0.009) and EST19 (11%, P  0.02) mRNA expression was observed in TNF-/DEX-treated normal human bronchial epithelial cells (Figure 3), indicating the potential physiologic relevance of these two ESTs. In addition, TNF- increased the expression of EST8 (34%, P  0.01) but not that of EST19 in the basophil leukemia cell line KU812 (Figure 4). This pattern of expression was not observed in other cell types tested, including human endothelial cells, synovial fibroblasts, lymphocytes, and monocytes, although each expressed EST8 and EST19 at the basal level. Moreover, TNF- did not induce EST8 and EST19 expression over basal levels in other airway epithelial cells including BEAS2B. Such differential expression of EST8 and EST19 among cell types may be the result of differences in culture conditions, lineage, or levels of TNF expression on the cell surface. In addition, the IB3-1 and BEAS2B cell lines may differ with respect to stage of differentiation and/or epithelial cell origin; IB3-1 cells do not form tight junctions (E. M. Schwiebert, unpublished observations), a quality indicative of an undifferentiated phenotype, and appear to be basal cell-like in morphology.

At present, it is unclear why the expression of all TNF--inducible, DEX-sensitive ESTs identified via DD-PCR was not confirmed by Northern blot analysis. Although DD-PCR is a powerful technique, there are potential problems with this method. First, false-positive ESTs (those bands that do not represent authentic mRNAs) may be detected (23). In order to circumvent this problem, the following approaches were taken in this study: (1) all samples for DD-PCR were run in duplicate in order to insure that the banding pattern was reproducible; (2) total RNA samples were treated with ribonuclease (RNAse)-free DNAse I in order to remove contaminating DNA species that might have competed for reagents during amplification; and (3) false-positive ESTs that migrated at the same position during electrophoresis as true-positive ESTs, and might therefore have eluted simultaneously, were identified via partial dideoxy sequencing analysis of EST-pGEM clones. Since only one EST fragment may be inserted into one molecule of pGEM vector during a ligation, multiple positive clones from an EST-pGEM cloning event may represent several different EST species. To determine the true-positive EST, 10 EST-pGEM clones were partially sequenced and the sequences compared; the majority of identical sequences was considered to be the true-positive EST. Second, certain combinations of primers may generate ESTs that are located in the 3' end of an expressed gene, and may therefore represent untranslated sequences, making homology analysis difficult. To decrease this possibility, cDNA products were amplified with pairs of "random" 10-mer primers instead of with a single "random" primer and the respective oligo-dT primer for amplification as described in the original Liang and Pardee protocol (23); the "random" 10-mer primer combination permitted the discovery of ESTs that had sizable open reading frames, many with homology to known genes. Third, rare mRNAs may not be visualized. Despite the high sensitivity of DD-PCR, it may not detect mRNAs that exist at less than 50 copies per cell (23). Taken together, these reasons may also explain why the expression of previously identified genes known to be upregulated by TNF- in airway epithelial cells, such as RANTES (18, 35) and IL-8 (22), was not detected with DD-PCR. Interestingly, recent work has shown that the IB3-1 cell line expresses low levels of some TNF--inducible genes, including RANTES, but not others, such as IL-8, as compared with other airway epithelial cell lines (L. M. Schwiebert, unpublished observations). It is conceivable that modifications to DD- PCR, such as utilizing higher annealing temperatures during the initial cycles of the PCR amplification step together with longer random primers (15- to 20-mers), may further decrease the generation of false positives and enhance the detection of true TNF--inducible genes.

With BLAST analysis, it was determined that EST8 and EST19 displayed significant homology with rhoC and the human UFO tyrosine kinase receptor (UFO-R), respectively (Figure 5). Although TNF- treatment of the IB3-1 airway epithelial cell line induced a modest increase in the expression of mRNA for each of these ESTs, it is possible that such an increase results in a significant enhancement of rhoC and/or UFO-R protein levels and/or function. RhoC is a member of the ras superfamily/rho subfamily of small monomeric guanosine triposphate (GTP)-binding proteins, which also includes rhoA, rhoB, rac, and cdc42. The rho subfamily has been shown to regulate assembly of focal adhesion complexes and actin stress fiber formation in fibroblasts (27, 37). Recently, Laudanna and colleagues reported that rhoA is also involved in chemoattractant-mediated adhesion of leukocytes (19). Specifically, their study demonstrated that inactivation of rho by the rho-specific inhibitor Clostridium botulinum C3 exoenzyme blocked IL-8-induced lymphocyte 41 adhesion to vascular cell adhesion molecule-1 (VCAM-1) and formyl-methionylleucylphenylalanine (fMLP)-induced neutrophil ₂-integrin adhesion to fibrinogen (19). Given this, rhoC may play a similar role in airway epithelium (i.e., rhoC may mediate focal adhesion formation between cells and/or signal transduction through integrins or other receptors expressed on the airway epithelial cell surface). Similarly, UFO-R, an orphan tyrosine kinase receptor expressed in a variety of different cell types (34), may also play a role in signaling the airway epithelial cell during an inflammatory response, possibly resulting in the production and secretion of inflammatory mediators. The human UFO-R contains two immunoglobulinlike and two fibronectin type III repeats in the extracellular portion, a single transmembrane region, and putative tyrosine kinase activity in the intracellular domain (34).

The findings discussed herein indicate the potential usefulness of DNA differential display in identifying genes that may be associated with airway inflammation. Ultimately, it is hoped that information gained from this approach will be important in understanding the cellular and molecular mechanisms of airway inflammatory responses and glucocorticoid actions.