Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Nitric oxide (NO), a short-lived, free radical gas, is a multifunctional effector in many physiologic processes (1, 2). Its biologic effects in the lung include modifying airway tone, regulating pulmonary vascular tone, stimulating mucin secretion, and modulating mucociliary clearance through effects on ciliary beat frequency, as well as microbial and tumor cell killing (1). In addition to the beneficial functions, a high level of NO production has been associated with tissue damage, septic shock, and airway inflammation of asthma (4). NO is synthesized by NO synthase (NOS) which converts L-arginine to L-citrulline and NO. Three genes encode isoforms of NOS in different tissues: inducible NOS (NOS2) that produces high levels of NO, and two constitutive NOSs that produce low levels of NO (7). The normal human airway epithelium expresses NOS2 in vivo, and asthmatic airways express even higher levels of NOS2 (8). Given the wide spectrum of positive and negative effects of NO in the lung, study of NOS2 gene expression in respiratory epithelial cells is of particular interest.

In general, cytokines induce NOS2 in murine and human cell lines through transcriptional regulation of the gene (9). For example, interferon (IFN)- induces a high level of NOS2 gene expression in primary cultured human airway epithelial cells, whereas the combination of interleukin (IL)-1, tumor necrosis factor (TNF)-, and IFN- leads to NOS2 induction in the lung adenocarcinoma cell line A549 (10). In contrast, transformed human bronchial epithelial cell lines BET-1A and BEAS-2B do not express NOS2 in response to cytokine combinations by Northern or Western analysis (10, 11). The lack of NOS2 expression in an immortal respiratory epithelial cell line has hampered the investigation of the mechanisms regulating NOS2 gene expression in the human lung. Moreover, elucidation of conditions that lead to NOS2 expression in BEAS-2B and BET-1A cells may reveal pathways necessary for high-level NOS2 expression in the human airway, and provide a convenient model system for in-depth investigation of NOS2 regulation. Mechanisms regulating NOS2 expression in human respiratory epithelial cells are complex, with the requirements of activating multiple signal transduction pathways and of new protein synthesis (6, 12). Recently, we have shown that NOS2 gene expression in human airway epithelial cells occurs in response to double-stranded RNA (dsRNA), through rapid activation of dsRNA-activated protein kinase (PKR), and subsequent activation of signaling proteins including nuclear factor (NF)-B and IFN regulatory factor (IRF)-1 (13). Loss of NOS2 induction in response to a wide variety of cytokines and mediators in PKR-null cells verified the central role of dsRNA-activated pathways in the generalized signaling to NOS2 (13). In this context, we hypothesized that BEAS-2B and BET-1A cells may express NOS2 in response to dsRNA. Here, we show that although cytokines (IFN-, TNF-, and IL-1) do not induce NOS2 expression in BEAS-2B or BET-1A cells, addition of dsRNA to this cytokine mix enables BEAS-2B cells to express NOS2 messenger RNA (mRNA). Similarly, BET-1A cells express NOS2 in response to dsRNA in combination with IFN- and serum. Importantly, dsRNA strongly activates NF-B and IRF-1 in BEAS-2B cells, transcription factors essential for NOS2 gene expression in other cell lines. Collectively, these results support the concept that dsRNA activation of intracellular signal transduction pathways are required for NOS2 gene expression in human bronchial epithelial cells.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Cell Culture and Treatments

BEAS-2B and BET-1A cells, human bronchial epithelial cell lines transformed by adenovirus12-Simian virus 40 (SV-40) virus and SV-40 T antigen, respectively (14), were cultured in LHC9 serum-free Lechner and LaVeck (LHC) medium (LHC8; Biofluids, Inc., Rockville, MD) which contains additives 0.33 nM retinoic acid and 2.75 µM epinephrine, on plates precoated with coating medium containing 29 µg/ml collagen (vitrogen; Collagen Corp., Palo Alto, CA), 10 µg/ml bovine serum albumin (BSA) (Biofluids), and 10 µg/ml fibronectin (Calbiochem, La Jolla, CA) for 5 min. The cells were passaged at 60 to 80% confluence by dissociation from plates with 0.02% trypsin (E-PET; Biofluids) that was neutralized with soybean trypsin inhibitor (Biofluids). A549 cells, an epithelial cell line derived from lung adenocarcinoma (American Type Culture Collection, Rockville, MD), were cultured in Eagle's minimum essential medium (MEM) (GIBCO Laboratories, Grand Island, NY) with 10% fetal calf serum (FCS), 2 mM L-glutamine, 50 U/ml penicillin, and 50 µg/ml streptomycin (10). In some experiments, FCS from GIBCO was used as a conditioning agent for the indicated time before detecting NOS2 mRNA induction.

Human IFN- was a gift from Genentech, Inc. (South San Francisco, CA), or purchased from R&D Systems, Inc. (Minneapolis, MN). Recombinant human IL-1 and TNF- were purchased from Genzyme Corp. (Cambridge, MA). Polyinosinic-polycytidylic acid (poly IC) was from Sigma Chemicals (St. Louis, MO). Reagents used in cell culture were tested for endotoxin contamination using a limulus endotoxin detection assay kit (Biowhitaker, Walkersville MD). Only endotoxin-free reagents were used in studies.

RNA Extraction and Northern Analysis

Total RNA was extracted and evaluated by Northern analysis as previously described using a ³²P-labeled, 1.9-kb NOS2 complementary DNA (cDNA) probe (pCCF21) or, as a control, 2-kb -actin cDNA probe (pHFA-1) (11), and then subjected to autoradiography. Expression of NOS2 mRNA relative to -actin was accomplished using a Phosphorimager (Molecular Dynamics, Inc., Sunnyvale, CA). Otherwise, the images of ethidium bromide staining of 28S ribosomal RNA were electronically digitalized by scanning and the RNA integrity was quantitated by the software, ImageQuant v. 1.2 (Molecular Dynamics).

Western Analysis

Cell lysate was prepared by freeze/thaw of BEAS-2B, BET-1A, or A549 cells cultured for the indicated times with IFN-, TNF-, and/or poly IC, and isolated in lysate buffer (3 mM dithiothreitol [DTT], 5 µg/ml aprotinin, 1 µg/ml leupeptin and pepstatin A, 0.1 mM phenylmethylsulfonyl fluoride [PMSF], 1% Nonidet P-40, and 40 mM N-2-hydroxyethylpiperazine-N'-ethane sulfonic acid [Hepes], pH 7.5). Total protein was measured by bicinchoninic protein assay (BCA) (Pierce, Rockford, IL). Primary antibodies used for Western analyses included a rabbit polyclonal primary antibody directed against the C-terminal 10 amino acids of human NOS2 (NO53; Merck, Rahway, NJ), and anti-IRF-1 antibody (Santa Cruz Biotechnology, Santa Cruz, CA) (17). Total proteins were separated by 6% and 10% sodium dodecyl sulfate polyacrylamide gel electrophoresis under denaturing and reducing conditions for NOS2, and IRF-1, respectively. Signal detection was accomplished using a peroxidase-linked, species-specific, donkey antirabbit secondary antibody (Amersham, Arlington Heights, IL) and enhanced chemiluminescence (Amersham).

Electrophoretic Mobility Shift Assay

Whole-cell extracts were prepared by a modification of a previously described method (15). In brief, adherent cells were harvested by a cell lifter, and the cell suspensions were centrifuged and washed with phosphate-buffered saline and resuspended in ice-cold low-salt buffer (10 mM Hepes [pH 7.9], 1.5 mM MgCl₂, 10 mM KCl, 0.5 mM PMSF, and 0.5 mM DTT). After a 5-min incubation on ice, cells were washed in the same buffer, then pelleted. A volume of the high-salt extraction buffer equal to the volume of the cell pellet was added (20 mM Hepes [pH 7.9], 25% glycerol, 0.42 M NaCl, 1.5 mM MgCl₂, 0.2 mM ethylenediaminetetraacetic acid [EDTA], 0.5 mM PMSF, 0.5 mM DTT, 5 µg/ml leupeptin, 2 µg/ml aprotinin, and 1 µg/ml pepstatin) and the mixture was placed on ice for 30 min. Whole-cell extracts were clarified by centrifugation at 12,000 × g for 20 min at 4°C. The protein concentration was measured by BCA (Pierce).

Oligonucleotides used for electrophoretic mobility shift assay (EMSA) included the IFN- activation site (GAS) (5'-tcgaGCCTGATTTCCCCGAAATGACGGC-3') corresponding to human IRF-1 promoter from base pairs (bp) 130 to 106 sequence relative to transcription start point (10), B sequence motif (5'-gatctACTCCGGGAATTTCC CTGGCC-3') corresponding to human GRO  gene promoter from bp 82 to 60 sequence relative to transcription start point (16), and the multimerized hexamer probe of the consensus sequence, (AAGTGA)₄, originally used to clone IRF-1 (17). These synthetic oligonucleotides were either end-labeled with [-³²P]adenosine triphosphate by polynucleotide kinase or fill-in labeled with [-³²P]deoxycytidine triphosphate. For binding reactions, whole-cell extracts (5 µg total protein) were incubated in 24 µl of total reaction volume containing 20 mM Hepes (pH 7.9), 10% glycerol, 60 mM NaCl, 5 mM MgCl₂, 4 mM Tris-HCl, 1 mM DTT, 0.6 mM EDTA, 200 µg/ml BSA, and 2 µg polydeoxyinosinic:polydeoxycytidylic acid (Amersham) for 15 min at 4°C. The ³²P-labeled oligonucleotide (0.2 ng, 2 × 10⁵ counts/min) was added to the reaction mixture and incubated for 20 min at room temperature. To specifically identify signal transducer and activator of transcription (STAT)-1, NF-B and IRF-1 proteins in binding complexes, 2 to 4 µg of rabbit anti-STAT-1 polyclonal antibody, rabbit anti-P65 polyclonal antibody, and rabbit anti-IRF-1 or anti-IRF-2 antibody (Santa Cruz Biotechnology) were added to the binding reaction mix and incubated for 30 min at 4°C before adding the ³²P-labeled oligonucleotide. The reaction products were analyzed by electrophoresis on a 4% polyacrylamide gel with 0.25 × TBE buffer (22.3 mM Tris, 22.2 mM borate, and 0.5 mM EDTA) for NF-B or a 4% polyacrylamide gel containing 50 mM Tris-HCl (pH 7.5), 0.38 M glycine, and 2 mM EDTA for IRF-1. The gels were dried and analyzed by autoradiography.

Statistical Analysis

Data in the text and figures are expressed as means ± standard error of the mean (SEM).

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

dsRNA in Combination with Cytokine Mix Enables BEAS-2B Cells to Express NOS2 mRNA

The combination of proinflammatory cytokines such as IFN-, TNF-, and IL-1 has been shown to be a potent inducer for NOS2 in human respiratory epithelial cells (10, 11). Thus, BEAS-2B cells cultured in serum-free LHC9 medium were stimulated with IFN-, TNF-, IL-1, or synthetic dsRNA (poly IC) for 24 h, either alone or in combination, and harvested to detect NOS2 mRNA by Northern or Western analyses. The combination of IFN-, TNF-, and IL-1 induced barely detectable NOS2 (%NOS2 mRNA/ -actin: 0.78 ± 0.03), whereas addition of dsRNA enabled BEAS-2B cells to express abundant levels of NOS2 in combination with IFN-, TNF-, and IL-1 (%NOS2 mRNA/-actin: 19 ± 4). Individually, neither IFN- nor dsRNA induced NOS2 (Figures 1A and 1B, lanes 2 and 5). These findings provide the first evidence that dsRNA, in combination with cytokines, stimulates high levels of NOS2 induction in BEAS-2B cells.

View larger version (53K): [in this window] [in a new window]   View larger version (35K): [in this window] [in a new window]   Figure 1.   (A) NOS2 mRNA expression in BEAS-2B cells. BEAS-2B cells cultured in serum-free medium were stimulated with IFN- (10,000 or 1,000 U/ml [lane 6]), TNF- (10 ng/ml), IL-1 (10 U/ml), or poly IC (100 µg/ml) for 24 h and harvested to detect NOS2 mRNA by Northern analysis (5 µg total RNA/lane). -actin is shown as a control. Similar results were obtained in two separate experiments. (B) NOS2 protein expression in BEAS-2B cells. BEAS-2B cells cultured in serum-free medium were stimulated with IFN-, TNF-, IL-1, or poly IC for 24 h and harvested to detect NOS2 protein by Western analysis (60 µg total protein/ lane). A549 cells cultured in MEM containing 10% serum and stimulated with cytokines are shown as a positive control (lane 6).

Decreased NOS2 mRNA Induction in A549 Cells Cultured in Serum-Free Medium

BEAS-2B and BET-1A cells were cultured in serum-free conditions for prolonged times. Serum contains a multitude of factors that affect cell growth and inflammatory responses (18). The effect of serum on NOS2 inducibility was assessed by incubating A549 cells in serum-free LHC9 medium. A549 cells that had originally been cultured in 10% serum-containing medium were washed and recultured for the indicated period in serum-free LHC9 medium. The ability of A549 cells to express NOS2 mRNA upon stimulation with IFN-, TNF-, and IL-1 was assessed by Northern analysis. Control culture of A549 cells in MEM with 10% FCS showed high levels of NOS2 mRNA, whereas A549 cells transferred from serum-containing medium to serum-free medium showed a progressive decrease in NOS2 mRNA induction (%NOS2/28S: 10% serum-containing media 49 ± 5, 1-d serum-free 16 ± 2, 2-d serum-free 11.7 ± 0.6). The 5-d culture of A549 cells in serum-free medium no longer expressed NOS2 mRNA in response to cytokines (Figure 2A, lane 6), although cells were viable as assessed visually and extracted RNA and protein of good integrity. Western analyses of NOS2 in A549 cells stimulated with IFN- for 24 h in the presence or absence of serum showed similar results, i.e., A549 expressed NOS2 protein in the presence of serum but did not in the absence of serum for 5 d (Figure 2B). Thus, serum contains factors that endow A549 cells with the ability to express NOS2 mRNA upon stimulation and, conversely, the specialized serum-free LHC9 medium used for culture of transformed bronchial epithelial cells lacks factors necessary for cytokine induction of NOS2.

View larger version (61K): [in this window] [in a new window]   View larger version (46K): [in this window] [in a new window]   View larger version (35K): [in this window] [in a new window]   Figure 2.   (A) Loss of NOS2 mRNA induction in A549 cells cultured in serum-free medium. A549 cells were cultured in MEM containing 10% FCS or serum-free LHC9 medium for the indicated times. Cells were then stimulated with IFN- (100 U/ml), TNF- (10 ng/ml), and IL-1 (10 U/ ml), for 8 h and harvested to detect NOS2 mRNA by Northern analysis (5 µg total RNA/ lane). Ethidium bromide stains show equal loading of RNA. Similar results were obtained in two separate experiments. (B) Loss of NOS2 protein induction in A549 cells cultured in serum-free medium. A549 cells were cultured in MEM containing 10% FCS or serum-free LHC9 medium for 5 d. Cells were then stimulated with IFN- (100, 1,000, or 10,000 U/ml) for 24 h and harvested to detect NOS2 protein by Western analysis (60 µg total protein/lane). (C) Serum requirement for NOS2 expression in BET-1A cells. BEAS-2B and BET-1A cells were cultured in serum-free LHC9 medium or 10% FCS-containing LHC9 medium for 4 d. Cells were then stimulated with IFN- (10,000 U/ml) and poly IC (100 µg/ml) for 24 h and harvested to detect NOS2 mRNA by Northern analysis (5 µg total RNA/lane). Ethidium bromide staining of the gel shows the amount and integrity of RNA loaded. Similar results were obtained in two separate experiments.

Serum Effect on BEAS-2B and BET-1A Cell Expression of NOS2 mRNA

On the basis of the loss of NOS2 expression in A549 in serum-free LHC9 medium, we investigated the potential effect of serum on BEAS-2B and BET-1A cell expression of NOS2. BEAS-2B and BET-1A cells were cultured in serum-free LHC9 medium or 10% FCS-containing LHC9 medium for 4 d. Cells were then stimulated with IFN- and synthetic dsRNA (poly IC) for 24 h and harvested to detect NOS2 mRNA by Northern analysis. Serum-exposed BET-1A cells expressed NOS2 mRNA in response to IFN- and dsRNA (Figure 2C). Serum did not lead to NOS2 mRNA induction in BEAS-2B cells (Figure 2C).

Serum Conditions BET-1A Cells for Subsequent NOS2 mRNA Induction upon Stimulation with IFN- and dsRNA

BET-1A cells were cultured in LHC9 medium with or without 10% FCS for 6 d. Cells were then stimulated with IFN-, TNF-, IL-1, or synthetic dsRNA (poly IC) and harvested in 24 h to detect NOS2 mRNA by Northern analysis or protein by Western analysis. BET-1A cells expressed NOS2 in response to IFN- and poly IC in serum-containing LHC9 medium (%NOS2 mRNA/28S: 28 ± 4) but not in serum-free LHC9 medium (Figures 2C and 3A). The combination of IFN-, TNF-, and IL-1 without poly IC led to only very low-level induction of NOS2 in serum-free LHC9 or in serum-containing media (%NOS2 mRNA/28S: serum-free 0.8 ± 0.7, 10% serum-containing media 2.5 ± 2.5) (Figures 3A and 3B). Thus, serum is necessary to condition BET-1A cells for NOS2 induction but dsRNA is still essential for optimal induction of NOS2 expression.

View larger version (59K): [in this window] [in a new window]   View larger version (39K): [in this window] [in a new window]   View larger version (12K): [in this window] [in a new window]   View larger version (11K): [in this window] [in a new window]   Figure 3.   Serum-dependence of NOS2 expression in BET-1A cells. BET-1A cells were cultured in serum-free LHC9 medium or 10% FCS-containing LHC9 media for 6 d. Cells were then stimulated with IFN- (10,000 U/ml), TNF- (10 ng/ml), IL-1 (10 U/ml), or poly IC (100 µg/ml) for 24 h and harvested to detect (A) NOS2 mRNA by Northern analysis (5 µg total RNA/lane) and (B) NOS2 protein by Western analysis (60 µg total protein/lane). Ethidium bromide stains show equal loading of RNA. Similar results were obtained in two separate experiments. (C) Serum concentration-dependent NOS2 mRNA induction. BET-1A cells were cultured in 0.1%, 1%, or 10% FCS-containing LHC9 medium for 4 d. Cells were then stimulated with IFN- (10,000 U/ml) and poly IC (100 µg/ml) for 24 h and harvested to detect NOS2 mRNA by Northern analysis (5 µg total RNA/lane). (D) Time required for NOS2 mRNA induction by 10% FCS. BET-1A cells were cultured in 10% FCS-containing LHC9 medium for 1 to 4 d. Cells were then stimulated with IFN- (10,000 U/ ml) and poly IC (100 µg/ml) for 24 h and harvested to detect NOS2 mRNA by Northern analysis (5 µg total RNA/lane). Abundance of NOS2 mRNA is expressed relative to 28S, with the peak intensity designated as 100. The graph displays means ± SEM relative to the peak intensity from two separate experiments.

Mode of Serum Conditioning of BET-1A Cells for NOS2 mRNA Expression

BET-1A cells were cultured in LHC9 medium with 0.1%, 1%, or 10% FCS for 4 d. Cells were then stimulated with IFN- and poly IC for 24 h and harvested to detect NOS2 mRNA by Northern analysis. The effectiveness of serum conditioning was dependent upon the concentration of serum used, with maximal NOS2 mRNA induction at a FCS concentration of 10% (Figure 3C). We examined the time required to condition BET-1A cells for NOS2 mRNA induction, using a fixed concentration of FCS (10%). No significant NOS2 mRNA induction was observed until the cells had been in serum-containing medium for at least 3 d (Figure 3D).

STAT-1 DNA-Binding Activity in BEAS-2B Cells

Although numerous intracellular signal transduction pathways have been implicated in NOS2 gene expression (22, 23), IFN- activation of STAT-1 and IRF-1 is absolutely required for NOS2 expression in murine cells (24) and likely also essential for NOS2 induction in human cells (8, 13). BEAS-2B cells are human bronchial epithelial cells transformed by infection with adenovirus 12-SV-40 hybrid virus (14). Adenovirus early region may impair STAT-1 activation and lead to lack of NOS2 induction by IFN- (25). To assess STAT-1 activation, BEAS-2B cells were incubated for 15 min with IFN-, whole-cell extracts were isolated, and EMSA was performed to detect STAT-1 DNA-binding activity. IFN- induced STAT-1 activation, demonstrated by the supershift with STAT-1 antibody (Figure 4). Taken together with the knowledge that BEAS-2B cells express chemokines in response to IFN- (26), we conclude that IFN- signaling pathway and subsequent IFN--inducible gene transcription are intact in BEAS-2B cells.

View larger version (51K): [in this window] [in a new window]   Figure 4.   STAT-1 activation by IFN- in BEAS-2B cells. BEAS-2B cells were incubated for 15 min with IFN- (10,000 U/ml). Whole-cell extracts were isolated, and EMSA was performed (5 µg/lane) to detect STAT-1 DNA-binding activity. Where indicated, whole-cell extracts were preincubated with STAT-1 antibody. Similar results were obtained in two separate experiments. NS, no stimulation.

NF-B Activation and IRF-1 Expression/Activation by dsRNA in BEAS-2B Cells

A transcription factor of particular importance in immune and inflammatory responses including NOS2 induction in A549 cells is NF-B (23, 27). In addition, IRF-1 is essential for NOS2 activation in murine macrophages, and mediates transcriptional activation via specific cis-acting elements resident in the promotors of IFN-stimulated genes (24). To determine the involvement of NF-B and IRF-1 in NOS2 induction in BEAS-2B cells, whole-cell extracts for EMSA were prepared from BEAS-2B cells stimulated for 1 h with IFN-, TNF-, IL-1, or poly IC. Poly IC and TNF- strongly induced NF-B activation (Figure 5). IL-1 also stimulated, to a lesser degree, NF-B activation. IFN- had little effect on basal NF-B activity. To detect IRF-1 induction and activation, BEAS-2B cells were stimulated with IFN-, TNF-, poly IC, or IL-1 for 4 h and evaluated by Western analysis and EMSA. IRF-1 protein was induced by IFN- and poly IC alone, and to a lesser degree by TNF- (Figure 6A). Further, poly IC potentiated IFN- induction of IRF-1 protein expression. Stimulation with IFN- or poly IC resulted in the activation of IRF-1 DNA binding (Figure 6B).

View larger version (106K): [in this window] [in a new window]   Figure 5.   NF-B activation in BEAS-2B cells. BEAS-2B cells were incubated for 1 h with 10,000 U/ml IFN-, 10 ng/ml TNF-, 10 U/ml IL-1, or 100 µg/ml poly IC. Whole-cell extracts were isolated and EMSA was performed (5 µg protein/lane). Where indicated, whole-cell extracts were preincubated with either p65 or c-Rel antibody. Similar results were obtained in two separate experiments. NS, no stimulation.

View larger version (46K): [in this window] [in a new window]   View larger version (61K): [in this window] [in a new window]   Figure 6.   IRF-1 expression and activation in BEAS-2B cells. BEAS-2B cells were incubated for 4 h with 1,000 U/ml IFN-, 10 ng/ml TNF-, 10 U/ml IL-1, or 100 µg/ml poly IC. IRF-1 protein induction (A) and DNA-binding activity (B) were determined by Western analysis (30 µg/lane) and EMSA (5 µg/lane), respectively. Where indicated, whole-cell extracts were preincubated with either IRF-1 or IRF-2 antibody. Similar results were obtained in two separate experiments. NS, no stimulation.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Because airway epithelia are a major cellular source of NOS2 in the lung (6, 28), a detailed understanding of intracellular events regulating the expression of NOS2 in human respiratory epithelial cells is prerequisite to ascertaining the role(s) of NOS2 in lung biology and diseases. Although researchers have noted high-level NOS2 expression in primary cultured human airway epithelial cells and the lung cancer cell line A549 in response to a variety of cytokines, such as IFN-, IL-1, IL-6, or TNF-, it has been difficult to convincingly demonstrate NOS2 expression by transformed bronchial epithelial cell lines BEAS-2B or BET-1A (10, 11). In this study, we show that NOS2 expression in transformed human bronchial epithelial cells is critically dependent upon exposure to dsRNA. Addition of dsRNA to the cytokines IFN-, TNF-, and IL-1 enables BEAS-2B cells to induce NOS2 mRNA. Thus, coordinate activation of multiple signaling pathways is likely necessary for BEAS-2B cells to express NOS2. In particular, regulation of NOS2 gene expression appears to require effector proteins responsive to dsRNA. PKR activates several signaling pathways leading to NOS2 expression in response to dsRNA (29). Recently, we have demonstrated that PKR protein expression in human airway epithelial cells is increased and activated by dsRNA, indicating a functional PKR pathway in these cells (13). Thus, dsRNA is a physiologically relevant activator of gene expression in the airway. The airway is frequently the first site of contact for viral infections, and most viruses induce the synthesis of dsRNA at some time during their replication cycles (13, 30, 31). Viral infection also leads to NOS2 expression in the lung. Our findings support that dsRNA is important for NOS2 induction in virus-infected cells, and subsequent generation of NO by the airway epithelium for antiviral host defense. In addition, dsRNA-like molecules are also present in some uninfected cells, and are capable of activating PKR (31). In this context, we and others have demonstrated a central role of PKR in the generalized signaling pathway to NOS2, using genetically PKR-deficient murine cells (13) or cells stably expressing dominant negative mutants of PKR (30). Taken together, dsRNA induction of NOS2 mRNA in BEAS-2B or BET-1A cells may be mediated by PKR activation of intracellular signal transduction pathways, including but not limited to NF-B and/or IRF-1.

Although the regulation of NOS2 gene expression by transcription factors is complex (9), the expression of NOS2 is driven primarily by Janus kinase/STAT-1 signaling pathway and NF-B (22, 23). In addition, IRF-1 can mediate transcriptional activity via specific cis-acting elements in the NOS2 promoter (24). Viral transformation of cells can lead to impaired ability to activate transcription factors, e.g., STAT-1 (25). Thus, we investigated the competence of signaling pathways in BEAS-2B cells, which are important for NOS2 expression. IFN- led to STAT-1 activation and IRF-1 induction and activation in BEAS-2B cells, whereas TNF- and IL-1 both led to NF-B activation. Similarly, dsRNA activated NF-B and IRF-1 in BEAS-2B cells. IRF-1 protein induction by IFN- was potentiated by dsRNA, but no other discernible effects were noted in activation of signaling pathways. Although these results certainly support the finding that the expression of NOS2 in transformed human bronchial epithelial cells may depend on induction/activation of STAT-1, IRF-1, and NF-B, other pathways must also be involved on the basis of the requirement of dsRNA for NOS2 expression in BEAS-2B cells.

In this context, culture of A549 cells in serum-free medium led to time-dependent loss of NOS2 inducibility by cytokines. Further, exposure of BET-1A cells to serum resulted in expression of NOS2 mRNA in response to IFN- and dsRNA, supporting the importance of serum-activated pathways for NOS2 expression. Serum contains numerous factors that affect many cellular events such as cell growth and death, cell proliferation and differentiation, and inflammatory responses (18). Although the precise mechanism of how serum conditions BET-1A cells for NOS2 expression is unknown, serum activates mitogen- activated protein kinase via protein kinase C-dependent and -independent (activated Ras) pathways (32). Activation of these pathways results in transcription of immediate early genes (e.g., c-fos) and/or phosphorylation of c-Jun protein or other kinases, and ultimately activates transcription factors such as activator protein (AP)-1 or NF-B, which may induce NOS2 gene expression. However, serum effects on AP-1 and NF-B are fairly rapid and unlikely to be a primary mechanism contributing to NOS2 induction, inasmuch as the serum effect required a minimum of 3 d exposure. The delayed effect more likely suggests a requirement for ongoing new protein synthesis. Interestingly, a similar conditioning action by serum has been noted for macrophage function. For example, rabbit alveolar macrophages fail to undergo a significant respiratory burst unless they are conditioned in serum for 24 to 48 h before triggering with phorbol myristate acetate or concanavalin A (33). Similarly, mouse peritoneal macrophages show fungistatic capacity against Cryptococcus neoformans only if they are cultured in 10% FCS-containing medium (34).

Although BEAS-2B and BET-1A cells are derived from normal human bronchial epithelial cells (14), the conditions for NOS2 expression are different in the two cell lines. Although dsRNA is necessary for optimal NOS2 expression in both, BET-1A cells also require exposure to serum for optimal NOS2 expression. Transformation of these cell lines is different, with BEAS-2B cells transformed by adenovirus12-SV-40 hybrid virus whereas BET-1A cells were transformed by SV-40 T antigen (14). Differences in transformation and/or subsequent differences in genetic mutations may be responsible for the variation in NOS2 activation in these cell lines (14). Regardless of these differences, clearly dsRNA-signaling pathways are important for NOS2 expression in human respiratory epithelial cells. Whereas primary cells are more applicable to in vivo situations, mechanistic studies frequently require the use of model systems that include transformed cells. With conditions for NOS2 expression characterized, these cell lines may be used as a model system to further investigate the complex mechanisms leading to NOS2 gene expression in human respiratory epithelial cells.