Acute lung inflammation and injury were induced by intranasal instillation of lipopolysaccharide (LPS) in normal and type 2 nitric oxide synthase (NOS2)-deficient (NOS2-/-) C57BL/6 mice. LPS-induced increases in extravasated airway neutrophils and in lung lavage fluid of TNF-α and macrophage inflammatory protein-2 were markedly lower in NOS2-/- than in wild-type mice, indicating that NOS2-derived nitric oxide (NO·) participates in inflammatory cytokine production and neutrophil recruitment. Instillation of LPS also increased total lung lavage protein and induced matrix metalloproteinase-9 and mucin 5AC, as indexes of lung epithelial injury and/or mucus hyperplasia, and increased tyrosine nitration of lung lavage proteins, a marker of oxidative injury. All these responses were less pronounced in NOS2-/- than in wild-type mice. Inhibition of NOS activity also suppressed production of TNF-α and macrophage inflammatory protein-2 by LPS-stimulated mouse alveolar MH-S macrophages, and this was restored by NO· donors, illustrating involvement of NO· in macrophage cytokine signaling. Oligonucleotide microarray (GeneChip) analysis of global lung gene expression revealed that LPS inhalation induced a range of transcripts encoding proinflammatory cytokines and chemokines, stress-inducible factors, and other extracellular factors and suppressed mRNAs encoding certain cytoskeletal proteins and signaling proteins, responses that were generally attenuated in NOS2-/- mice. Comparison of both mouse strains revealed altered expression of several cytoskeletal proteins, cell surface proteins, and signaling proteins in NOS2-/- mice, changes that may partly explain the reduced responsiveness to LPS. Collectively, our results suggest that NOS2 participates in the acute inflammatory response to LPS by multiple mechanisms: involvement in proinflammatory cytokine signaling and alteration of the expression of various genes that affect inflammatory-immune responses to LPS.
- nitric oxide
- oligonucleotide microarray
nitric oxide (NO·) is an important mediator that is overproduced on infection and in most inflammatory conditions as an additional host defense mechanism and as an immunomodulator. Although NO· is produced by several isoforms of NO synthase (NOS), the increased production of NO· during infection and inflammatory-immune activation is primarily due to induction of type 2 NOS (NOS2) in various cell types (51). In the respiratory tract, the major cells expressing NOS2 are bronchial and alveolar epithelial cells, alveolar macrophages, and infiltrated leukocytes (30). Although the role of NOS2 induction and NO· generation in airway inflammation has been studied actively, there is considerable controversy, inasmuch as pro- and anti-inflammatory properties of NO· have been identified in various models of respiratory infection or of acute or chronic lung inflammation using NOS2-deficient mice or NOS inhibitors. Most previous studies have demonstrated that NOS2 is often critical in host defense against various infections (10, 37, 55) and that NOS2 deficiency often results in an attenuated inflammatory response and reduced injury in various models of airway inflammation (1, 12, 26, 31, 46, 50, 56). However, NOS2-deficient mice have, in some cases, been reported to be more susceptible to lung injury, e.g., by ozone or hyperoxia (27, 29), illustrating anti-inflammatory properties of NO·.
The contribution of NOS2 to airway inflammation and/or injury has often been attributed to the formation of reactive nitrogen species (RNS) when NO· is catabolized by oxidative mechanisms that are activated during inflammation (1, 12, 31, 46). Although RNS can induce oxidative injury to host cells and could, thereby, promote inflammation, NOS2 activation also appears to more specifically regulate cytokine signaling in various mononuclear cells, thereby controlling type I and II immune responses (10, 11, 19, 21). For instance, NOS2 activity has been found to be an essential component of IL-12 signaling in innate immunity (11) and appears to be involved in allergic airway inflammation by selectively downregulating IFN-γ activity (56). In addition, NOS2 has been found to participate in proinflammatory cytokine signaling by activating the transcription factor nuclear factor (NF)-κB in hemorrhagic shock (19). Because the activation of NF-κB is a key event in the initiation of inflammation (3), these observations would implicate NO· in early stages of the inflammatory response, preceding the formation of RNS. However, studies of the regulation of NF-κB by NO· have also uncovered stimulatory and inhibitory effects of NO· (7, 33, 42, 53, 57), effects that are generally attributed to (in)activation of several signaling proteins through S-nitrosation, e.g., p21ras (32) or NF-κB itself (7, 22). It is presumed that the variable effects of NO· on NF-κB signaling depend on the cell type, the relative NO· concentration, and/or the stage of the inflammatory response (22).
Animal exposure to lipopolysaccharide (LPS) is a commonly used model to study acute lung inflammation and injury, and several previous studies have addressed the role of NOS2 in LPS-induced lung injury or mortality (16, 17, 31, 34, 37, 50, 54, 55). However, results of such studies have been somewhat inconsistent, possibly because LPS was usually given systemically in these studies, resulting in systemic multiple organ dysfunction and NOS2 induction, which might obscure local effects of NOS2 on inflammatory processes in the lung. To investigate the involvement of NOS2 in a more confined model of acute airway inflammation, we chose to induce acute lung inflammation and injury by intranasal administration of LPS in wild-type and NOS2-deficient mice. Such LPS instillation is known to cause an acute inflammatory response with transient extravasation of primarily neutrophils in the airways (6, 43). We determined various markers of lung inflammation and injury in lung tissues or lung lavage fluids and performed microarray analysis of global lung gene expression to more fully evaluate the diversity of the transcriptional effects of LPS and the involvement of NOS2 in such acute airway inflammatory responses. Overall, our results indicate that NOS2 defi-ciency suppresses LPS-induced airway inflammation as a result of several changes in basal gene expression and reduced proinflammatory cytokine signaling in response to LPS.
MATERIALS AND METHODS
Animal exposure to LPS. The experiments were approved by the University of California, Davis, Animal Care Committee and conform to National Institutes of Health guidelines regarding animal experimentation. Studies were done using C57BL/6 mice (Charles River Laboratories) and NOS2-deficient mice, which were initially purchased from Taconic Farms (Germantown, NY) and rederived by embryo transfer on a C57BL/6 background to establish a breeding colony designated C57BL/6Ai-(KO)NOS2-N5 (27). Male and female wild-type and NOS2-/- mice (25-30 g) were subjected to brief anesthesia with diethyl ether, and 40-50 μl of an LPS solution in PBS (250 μg/ml, from Escherichia coli serotype 055:B5; Sigma) were instilled directly into their nostrils to reach a dose of 300 μg/kg LPS. Control mice received a similar volume of sterile PBS. Previous studies have demonstrated that a significant fraction of intranasally administered LPS will reach the lungs and that such instillation evokes an acute transient inflammatory response (6, 45).
Collection of lung tissues and lavage fluids. At various times after LPS instillation, mice were killed by administration of pentobarbital sodium (150 mg/kg ip). The tracheae were cannulated, and the lungs were lavaged with four consecutive washes with 0.5 ml of PBS, which were pooled to a total recovered lung lavage fluid of 1.4-1.6 ml. Total cell counts in the collected lavage sample were determined with a hemocytometer, and cell differential counts were determined after Cytospin centrifugation and Diff-Quick staining. Cell-free lung lavage fluids were obtained after centrifugation (4 min at 4,000 rpm) and stored at -20°C until further analysis. The right lung lobes were excised and fixed with 4% paraformaldehyde for paraffin embedding, serial sectioning, and hematoxylin-eosin staining. The left lung lobes were stored in 1 ml of RNAlater (Ambion, Austin, TX) and frozen at -80°C for subsequent RNA extraction.
Inflammatory cytokine and total protein analysis of lung lavage fluids. To quantitate pro- and anti-inflammatory cytokine production in the lung, antigenic concentrations of TNF-α, macrophage inflammatory protein-2 (MIP-2), IFN-γ, and IL-10 in the lung lavage fluids were measured by ELISA (Quantikine M Mouse cytokine assay kit, R & D Systems, Minneapolis, MN).
As a measure of epithelial injury and lung permeability, total protein concentration within lung lavage fluids was measured with a protein assay kit (Bio-Rad, Hercules, CA) based on the method of Bradford, with bovine serum albumin as a standard.
Analysis of NO· production and protein nitration. To quantitate the production of NO· in the lung, its metabolites, and , were measured in lung lavage fluids by chemical reduction to NO· in 120 mM vanadium (III) in 2 M HCl at 95°C and analysis by ozone-enhanced chemiluminescence (model 7020, ANTEK Instruments, Houston, TX) (5).
To evaluate the formation of RNS within the airways, protein 3-nitrotyrosine (3-NT) levels were quantitated using HPLC with electrochemical detection (ECD). Lung lavage fluids were pooled from three mice and concentrated 5-10 times in a SpeedVac, and proteins were digested by pronase (Calbiochem, La Jolla, CA) at a protein concentration ratio of 1:10 at 50°C for 12 h. Samples were filtered on Microcon filters (10,000 mol wt cutoff; Millipore, Bedford, MA) to remove undigested proteins, and the filtrate was injected into the HPLC-ECD 3-NT detection system (Eicom, Kyoto, Japan). Samples were separated on a reverse-phase C18 column (3.0 × 150 mm; Eicom) by elution with 0.1 M sodium phosphate buffer containing 5% methanol and 5 μg/ml EDTA (pH 5.6) at a flow rate of 0.5 ml/min. The elutant was passed through an ECD detector system consisting of two ECD cells; the upstream (reduction) cell was used for reduction of 3-NT at -900 mV, and the downstream (detector) cell was used for detection of the reduced form of 3-NT at an oxidation potential of +300 mV. To determine specificity for 3-NT, similar analysis was performed using a reduction potential of -600 mV, instead of -900 V, which diminishes the response to 3-NT. The amount of 3-NT in each sample was expressed relative to the amount of tyrosine, which was determined by in-line UV detection (274 nm). Control samples that contained only pronase were analyzed similarly to determine the liberation of tyrosine resulting from pronase autodigestion, and results were corrected accordingly.
Gene expression analysis of NOS2, matrix metalloproteinase-9, and mucin 5AC. Lung expression of NOS2, matrix metalloproteinase-9 (MMP-9), or mucin 5AC (MUC5AC) was evaluated by semi-quantitative RT-PCR after lung homogenization and extraction of total RNA with TRIzol reagent (Invitrogen) according to the manufacturer's instructions. First-strand cDNA was synthesized from 5 μg of total RNA using Moloney murine leukemia virus reverse transcriptase (Invitrogen) with an oligo(dT)15 primer (Promega, Madison, WI) at 37°C for 50 min. The PCR was performed using 1 U of platinum Taq DNA polymerase (Invitrogen), and each cycle (of 30 cycles) included 1 min at 94°C for denaturation, 2 min at 60°C for annealing, and 2 min at 72°C for extension. As an internal control, glyceralde-hyde-3-phosphate dehydrogenase (GAPDH) was amplified to normalize the amount of cDNA for each sample. The following primers were designed on the basis of the cDNA sequences: 5′-GTG AGG ATC AAA AAC TGG GG-3′ (sense) and 5′-ACC TGC AGG TTG GAC CAC-3′ (antisense) for NOS2, 5′-CTG TCC AGA CCA AGG GTA CAG CCT-3′ (sense) and 5′-GTG GTA TAG TGG GAC ACA TAG TGG-3′ (antisense) for MMP-9, 5′-CAG CCG AGA GGA GGG TTT GAT CT-3′ (sense) and 5′-AGT CTC TCT CCG CTC CTC TCA AT-3′ (antisense) for MUC5AC, and 5′-TTC ATT GAC CTC AAC TAC AT-3′ (sense) and 5′-GAG GGG CCA TCC ACA GTC TT-3′ (antisense) for GAPDH (Operon Technologies, Alameda, CA). The expected sizes of the PCR products were 380 bp for NOS2, 263 bp for MMP-9, 399 bp for MUC5AC, and 467 bp for GAPDH. Amplification was stopped in the linear phase of the PCR to quantitate the amount of mRNA in each sample. cDNA was diluted 10-fold for NOS2, MMP-9, and MUC5AC and 100-fold for GAPDH, and 30 cycles were used. PCR products were resolved by 3% agarose gel electrophoresis and visualized by ethidium bromide staining, and band densities were quantified by NIH Image version 1.61 software.
LPS stimulation of cultured MH-S alveolar macrophages. Simian virus 40-transformed mouse alveolar macrophage MH-S cells (American Type Culture Collection) (38) were grown in RPMI 1640 medium (Invitrogen) containing 10 mM HEPES, 1 mM sodium pyruvate (Sigma), 4.5 g/l glucose, 1.5 g/l sodium bicarbonate, 0.05 mM 2-mercaptoethanol, and 10% FBS. For experiments, cells were seeded in 24-well plates at 1 × 105 cells/well and cultured until they reached ∼90% confluence. The medium was changed to DMEM (Invitrogen) containing 10 mM HEPES and 0.05 mM 2-mercaptoethanol, and cells were treated with LPS (100 ng/ml) in the absence or presence of 1 mM NG-monomethyl-l-arginine (l-NMMA; Sigma) for 24 h. In some cases, cells were simultaneously exposed to various concentrations of the NO·-releasing compound diethylenetriamine NONOate (Cayman Chemical, Ann Arbor, MI). After incubation, conditioned media were collected and centrifuged at 6,000 rpm for 10 min to remove floating cells, and cytokines and levels were analyzed as described above.
Statistical analysis. Values are means ± SE. Correlation analysis was assessed using a nonparametric Spearman's correlation test. ANOVA was used to test for differences in measured variables between groups. When significant differences were found, specific differences were identified using a sequentially rejective Bonferroni-Dunn post hoc test procedure. Differences were considered significant at P < 0.05.
Oligonucleotide expression microarray (GeneChip) analysis. To obtain further insight into the roles of NOS2 in LPS-induced acute inflammatory response, we used a high-density oligonucleotide expression microarray system (GeneChip, Affymetrix, Santa Clara, CA) to analyze the global gene expression profile in lung tissues. Equal amounts of total lung RNA were pooled from four mice from each treatment group, and the first-strand cDNA was synthesized from 10 μg of pooled RNA using Superscript II reverse transcriptase (Invitrogen) with a T7 oligo(dT)24 primer (Operon) at 42°C for 1 h. Second-strand cDNA was synthesized using DNA polymerase I, DNA ligase, RNase-H, and T4 DNA polymerase (Invitrogen) at 16°C for 2 h. The phenol-extracted double-strand cDNA was used for in vitro transcription by T7 RNA polymerase in the presence of biotin-labeled NTP as described by the manufacturer, and the synthesized biotin-labeled cRNA (50 μg) was purified by an RNeasy kit (Qiagen, Valencia, CA) and fragmented at 95°C for 35 min in 40 mM Tris-acetate, pH 8.1, containing 100 mM potassium acetate and 30 mM magnesium acetate.
The fragmented cRNAs from each treatment group were hybridized to murine genome U74Av2 GeneChip array (Affymetrix) with external controls, stained with streptavidin-phycoerythrin (Molecular Probes, Eugene, OR), and scanned using an Agilent GeneArray Scanner (Affymetrix). Chip hybridization and scanning were performed at the Microarray Core Facility at the University of California, Davis. The output files were analyzed with Microarray Suite 5.0 software (Affymetrix). The expression value (average difference) for each target gene was determined by calculating the average of the difference of perfect match from mismatch intensity (PM-MM difference) between its probe pairs. Genes with a detection P < 0.04 determined by the statistical program are considered to be present call (P), those with 0.04 ≤ P ≤ 0.06 are considered to be marginal call (M), and those with P > 0.06 are considered to be absent call (A). Fold changes were determined by dividing the expression value of the LPS group by that of the control group. D-Chip version 1.1 programs were applied for cluster analysis of detected genes (35, 36). Genes with at least two expression values >20 and with a variation across samples (SD/mean) >0.5 were included in the cluster analysis. Functional category classification was based on the National Center for Biotechnology Information LocusLink (http://www.ncbi.nlm.nih.gov/LocusLink) database, which classifies a gene according to molecular function, and biological process and cellular component using GeneOntology (http://www.geneontology.org) terms.
Acute airway inflammation by intranasal instillation of LPS. Intranasal instillation of 300 μg/kg LPS in C57BL/6 mice induced acute airway inflammation, characterized by increases in total lung lavage cell numbers and protein concentrations, which were maximal 24 h after LPS instillation and declined thereafter (Fig. 1A). The increase in lung lavage cells was almost exclusively accounted for by neutrophils (Fig. 1B, Table 1). Histological analysis of lung tissues with hematoxylin-eosin staining showed patchy areas of neutrophil infiltration in alveolar wall and alveolar space as well as mononuclear cell infiltration in the alveolar wall with edematous thickening (not shown). Lung lavage fluid levels of the proinflammatory cytokines TNF-α and MIP-2 were also increased 24 h after LPS instillation and decreased over the following days. On the other hand, lung lavage fluid levels of IFN-γ and IL-10 increased slightly after 24 h and remained elevated thereafter. On the basis of these initial findings, additional measurements in the next set of experiments were performed 24 h after LPS instillation.
LPS-induced airway neutrophilia is reduced in NOS2-/- mice. Although lavaged lung cells of untreated mice almost exclusively represent alveolar macrophages (97%, Table 1), LPS administration markedly increased total lung lavage cell numbers in wild-type and NOS2-/- mice. These increases were almost exclusively due to neutrophils, and alveolar macrophage numbers did not change significantly (Table 1). The LPS-induced increase in total lavageable cells and neutrophils was ∼50% lower in NOS2-/- than in wild-type mice (P < 0.01; Table 1, Fig. 1B). Similarly, LPS-induced increases in lung lavage protein levels were significantly (22%) lower in NOS2-/- than in wild-type mice (P < 0.05; Table 1).
Administration of LPS to wild-type mice induced NOS2 mRNA within lung tissues, which was accompanied by increased lung lavage levels of (Fig. 2). As expected, NOS2 mRNA was undetectable in LPS-treated NOS2-/- mice, and lung lavage levels of did not change after LPS stimulation. However, basal lung lavage levels from untreated NOS2-/- or wild-type mice were quite similar and may largely originate from other NOS isoforms (NOS1 or NOS3) (47). Quantitative RT-PCR analysis of lung tissue mRNA showed expression of NOS3 and (to a lesser degree) NOS1, and expression of these NOS isozymes was similar in NOS2-/- and wild-type mice (data not shown).
LPS-induced protein nitration in lung lavage fluids. One proposed mechanism by which NOS2-derived NO· might contribute to inflammation is through the production of reactive nitrogen intermediates, which is reflected by increased protein nitration (31). Indeed, analysis of lung lavage proteins revealed increased 3-NT content after LPS instillation in wild-type and NOS2-/- mice (Fig. 3), indicating that formation of 3-NT is not critically dependent on the presence or induction of NOS2. Several previous studies have established that formation of 3-NT results from oxidative events associated with inflammation and involves the activation of granulocyte peroxidases such as neutrophil myeloperoxidase (2, 15, 52). When expressed relative to unmodified protein tyrosine residues, protein 3-NT levels were ∼20% lower in lung lavage fluids from LPS-stimulated NOS2-deficient mice than from wild-type mice (Fig. 3). Because total protein levels in lung lavage fluids were also ∼25% lower in NOS2-/- mice (Table 1), the total extent of protein nitration was actually reduced by ∼40% in NOS2-/- mice, which corresponds with the relative changes in neutrophil extravasation (Table 1). Hence, the results imply that 3-NT formation is associated with the extent of LPS-induced neutrophil influx and occurs as a consequence of oxidative events following neutrophil activation.
LPS-induced production of pro- and anti-inflammatory cytokines. Corresponding with the influx of neutrophils, lung lavage levels of the proinflammatory cytokines TNF-α and MIP-2 were significantly increased 24 h after LPS instillation. This response was significantly suppressed (∼50%) in NOS2-/- mice (Table 2), comparable to the differences in LPS-induced neutrophil influx in the two mouse genotypes. Lung lavage levels of TNF-α and MIP-2 significantly correlated with observed neutrophil numbers (r = 0.916 and 0.821, respectively), and there was strong correlation between TNF-α and MIP-2 levels in lung lavage fluids (r = 0.892). LPS instillation also resulted in increased levels of the anti-inflammatory cytokines IFN-γ and IL-10, but these increases were similar in both mouse genotypes (Table 2). Overall, these findings suggest that NOS2-derived NO· participates in the production of proinflammatory cytokines by LPS.
NOS2 deficiency reduces LPS-induced expression of MUC5AC and MMP-9. In addition to measuring lung lavage protein levels, we also investigated lung gene expression of MUC5AC and MMP-9 in response to LPS, as additional indexes of lung injury and/or mucus secretion (8, 58). We could detect MUC5AC mRNA in lung tissues from untreated mice and found this to be increased after LPS instillation, and this increase was significantly less pronounced (49%, P < 0.05) in NOS2-/- mice (Fig. 4A). Instillation of LPS also markedly increased MMP-9 mRNA levels in the lungs of wild-type mice. Again, the LPS-induced MMP-9 expression was lower in NOS2-/- mice, although basal MMP-9 mRNA levels also appeared to be reduced in NOS2-/- mice compared with wild-type mice (Fig. 4B). The observed differences in MMP-9 or MUC5AC expression were qualitatively comparable to changes in proinflammatory cytokine production and/or neutrophil numbers (in each case reduced by ∼50%) and suggest that NOS2 deficiency may indirectly affect expression of these genes by reducing LPS-induced proinflammatory cytokine production and neutrophil extravasation.
NO· is involved in LPS-stimulated cytokine production by MH-S alveolar macrophages. The effects of NOS2 deficiency on LPS-induced proinflammatory cytokine production suggest involvement of NO· in cytokine formation by LPS-stimulated cells, such as alveolar macrophages. To address the potential involvement of NO· in macrophage cytokine production, cultured mouse alveolar MH-S cells were stimulated with LPS, and production of TNF-α and MIP-2 was followed by ELISA. As shown in Fig. 5, LPS induced the production of TNF-α and MIP-2 in MH-S cells, and this was partially inhibited by the nonselective NOS inhibitor l-NMMA. Inhibition of NOS activity by l-NMMA was demonstrated by almost complete prevention of LPS-induced accumulation of in the culture media (not shown). Addition of the NO·-releasing compound diethylenetriamine NONOate dose dependently restored the inhibitory effects of l-NMMA, providing further evidence for involvement of NO· in LPS-induced production of these cytokines. These results with MH-S cells are consistent with our in vivo results and indicate that NOS2 within alveolar macrophages (and perhaps also in other cell types) participates in proinflammatory cytokine production in response to LPS and contributes to airway neutrophilia.
Global gene expression analysis by high-density oligonucleotide arrays. In an attempt to reveal potential additional transcriptional changes that cause altered LPS responses in NOS2-/- mice, we analyzed global lung expression of 12,451 genes included on the U74Av2 GeneChip. Results of this analysis are illustrated in Fig. 6. According to the detection P value, A call (P > 0.06) was considered to be undetectable/unexpressed genes and M (0.04 ≤ P ≤ 0.06) and P (P < 0.04) calls were considered to be detectable/expressed genes. In lung tissues from wild-type mice, 8,738 genes (70.2%) were not significantly expressed or detected in untreated and LPS-treated groups. Among 3,713 of the expressed/detected genes in the lungs of wild-type mice (29.8%), the overall expression of 3,188 genes (85.9%) was not affected by LPS (0.5- to 2-fold changes), but 298 genes (8.0%) were induced by LPS (>2-fold change), and 227 genes (6.1%) were suppressed by LPS (<0.5-fold change). “Highly induced” genes (>8-fold change) or “highly suppressed” (<0.125-fold change) genes in wild-type mice are indicated in Fig. 6 and listed in Table 3. Highly induced genes include inflammatory cytokines (e.g., TNF-α and IL-1β) or chemokines (e.g., MIPs or monocyte chemoattractant proteins), stress inducible factors (e.g., serum amyloid A3), and other extracellular factors. Highly suppressed genes include cytoskeletal proteins (e.g., actin, myosin, and keratin) and some other cytoplasmic factors.
In NOS2-/- mice, 139 of the 3,748 detected genes (3.7%) were induced by LPS, and 233 genes (6.2%) were suppressed by LPS. Thus markedly fewer genes were induced by LPS in NOS2-/- than in wild-type mice. In general, spots of highly induced or “suppressed genes” in the wild-type group were shifted to the midline in the NOS2-/- group (Fig. 6B). Fold changes in gene expression by LPS in wild-type and NOS2-/- mice are plotted in Fig. 6D. Among the 3,467 expressed/detected genes in both mouse types, only 66 genes (1.9%) were upregulated more than twofold by LPS in both groups (“LPS-inducible genes”). The average expression value of LPS-inducible genes was significantly lower (63%, P < 0.05) in NOS2-/- than in wild-type mice, suggesting that the presence of NOS2 generally contributes to LPS-induced inflammatory gene expression. Comparison of both mouse genotypes showed that basal lung expression of several genes, including some cytoskeletal proteins (e.g., actin, myosin, and keratin) and cell surface proteins (e.g., CD3, CD8, and T cell receptor), was suppressed in NOS2-/- mice compared with wild-type mice (Fig. 6C, Table 3). Such changes may partly explain the reduced ability of NOS2-/- mice to respond to LPS.
Cluster analysis of lung gene expression. To examine the relation between the gene expression pattern and potential functional consequences, we performed cluster analyses using the D-Chip program (35). According to the selection criteria described in materials and methods, 161 genes were selected for cluster analysis, and 7 clusters were identified (Fig. 7A). Gene names and accession numbers are listed in Table 3 (35 uncharacterized expressed sequence tags were excluded). Genes in cluster A were upregulated in wild-type mice, but not in NOS2-/- mice (wild-type-specific upregulated genes). Genes in cluster B1 were upregulated in both groups but were less upregulated in NOS2-/- than in wild-type mice (NOS2-regulated LPS-inducible genes). Genes in cluster B2 were similarly upregulated in both groups (NOS2-independent LPS-inducible genes). Genes in cluster C were similarly downregulated in both groups (NOS2-independent LPS-suppressed genes). Genes in clusters D and E were downregulated in wild-type mice but were not changed or were upregulated in NOS2-/- mice (NOS2-regulated LPS-suppressed genes). Finally, genes in cluster F were upregulated in NOS2-/- mice but were not affected in wild-type mice (NOS2-/--specific upregulated genes).
Many genes of interest are in cluster B1 (a high magnification of this cluster is illustrated in Fig. 7B), which includes all the “highly induced genes” in Fig. 6A and comprises many immune, inflammatory, and stress-responsive genes. Interestingly, this cluster also includes CD14, an LPS-induced gene that is of major importance in LPS binding and recognition. Clusters D and E include all the highly suppressed genes in Fig. 6A and comprise cytoskeletal and mitochondrial genes that are important for cell motility, cell homeostasis, and energy metabolism. The relation between gene clusters and their general functions are illustrated in Fig. 7C.
The main overall conclusion from our present studies is that NOS2-derived NO· is involved in the development of LPS-induced pneumonitis. In this regard, our findings are comparable with those of several previous studies that have indicated a contributing role of NOS2 in other models of airway inflammation induced by intraperitoneal LPS (31), allergen challenge after immunization (56), or exposure to ozone or silica (12, 46). In addition to its contribution to LPS-induced neutrophilia, NOS2 also appears to contribute to lung injury in response to LPS challenge, indicated by protein influx into the alveolar space, which is not strictly related to neutrophil influx (6), and the induction of genes that are associated with epithelial injury, such as MMP-9 and MUC5AC. One limitation of the present studies is that effects of NOS2 deficiency were examined at only one time point after LPS challenge (24 h), and the time course of inflammation in NOS2-/- mice might have been different from that in wild-type mice. Similarly, our results also do not allow us to comment on potential effects of NOS2 deficiency on lung pathology at later time points after LPS challenge, although differences might be expected on the basis of changes in expression of MMP-9 or MUC5AC. Additional studies are required to more specifically address these issues.
One proposed factor in the proinflammatory effects of NO· is the formation of RNS, which can injure cells by oxidative reactions, including the nitration of protein tyrosine residues (12, 31, 46). The presence of elevated amounts of 3-NT in extracellular proteins indeed indicates local formation of RNS in response to LPS instillation in wild-type and NOS2-/- mice. The fact that increased nitration was also observed in NOS2-deficient mice indicates that formation of 3-NT is not critically dependent on increased production of NO· by NOS2 induction but, rather, results from oxidative reactions such as those initiated by activated neutrophils (2, 15, 52). As illustrated in Fig. 2, levels of in lung lavage fluids from NOS2-/- mice did not differ significantly from those of control wild-type mice, suggesting involvement of other NOS isoforms in airway NO· production, which appears sufficient to support the formation of RNS that cause protein tyrosine nitration during the activation of an inflammatory response. Although we observed significant increases in 3-NT in lung lavage proteins after LPS instillation, we could not detect significant changes in protein nitration in lung tissues after intranasal LPS administration by HPLC-ECD analysis (in which case lung tissue 3-NT was largely undetectable; <1 μmol/mol Tyr) or by immunohistochemical analysis with antinitrotyrosine antibodies (not shown). The more abundant nitration observed in extracellular proteins obtained by lung lavage is consistent with the extensive neutrophil extravasation into the air spaces in this more compartmentalized model of acute lung injury, whereas fewer neutrophils appeared to be present within the lung tissue. The close association of 3-NT with the presence and activation of neutrophils (2, 15, 52) strongly suggests that tyrosine nitration and, hence, the formation of RNS that it indicates are a result, rather than a cause, of the inflammatory response.
Our results furthermore indicate that NOS2 is involved in airway production of the proinflammatory cytokines TNF-α and MIP-2 on LPS challenge, implicating NO· in processes that precede neutrophil extravasation and activation and consequent formation of RNS. Previous studies have revealed immune regulatory roles of NO· by developing enhanced Th1 responses after infection and antigen stimulation by promoting IFN-γ and reducing IL-4 or IL-12 (4, 39, 44, 56). Recently, NOS2-/- mice were found to produce more IL-12 than wild-type mice in a similar model of LPS-induced inflammation (44). However, we did not observe significant changes in LPS-induced IFN-γ production between these two mouse strains (Table 2). Our global lung gene expression analysis was in general agreement with these findings and revealed a large group of LPS-inducible genes in wild-type mice, including many genes encoding proinflammatory cytokines, chemokines, and stress factors, that were generally less responsive in NOS2-/- mice. A complicating factor in such global gene expression analysis at a relatively late time point after LPS challenge (24 h) is the fact that results most likely represent a cascade of initial and secondary gene expression, as well as counterregulation. Interestingly, one gene that was found to be induced by LPS in wild-type mice and, to a lesser extent, in NOS2-/- mice is CD14, encoding a major protein involved in the binding and recognition of LPS (48, 49). Because altered CD14 expression in NOS2-/- mice could be a contributing factor in a reduced inflammatory response to LPS, we performed quantitative real-time (TaqMan) PCR analysis of CD14 expression compared with the housekeeping gene hprt using an ABI PRISM 7700 sequence detection system (Perkin-Elmer, Foster City, CA) to confirm these findings. Relative lung expression of CD14 was 1.1 ± 0.1 and 1.1 ± 0.1 for untreated wild-type and NOS2-/- mice, respectively, and 6.9 ± 3.8 and 2.2 ± 1.2 for LPS-challenged wild-type and NOS2-/- mice (n = 3). Thus these results suggest that the differences in LPS-induced inflammation between these mouse strains are not due to differences in CD14 expression. However, induction of CD14 after LPS challenge was markedly reduced in NOS2-/- mice, which may be related to the fact that LPS-induced TNF-α production was also suppressed in these mice (Table 2), because TNF-α appears to mediate LPS-induced CD14 expression (13).
Alveolar macrophages are considered to be major effector cells involved in the inflammatory response to inhaled LPS (28), although recent findings also illustrate the importance of the respiratory epithelium in such innate immune responses (9, 41). In studies with cultured alveolar MH-S macrophages, we confirmed the proinflammatory properties of NO· and illustrated a role for NO· (presumably from NOS2) in the production of the macrophage-derived inflammatory cytokines TNF-α and MIP-2 in response to LPS. Recent studies with chimeric mice, generated using bone marrow transfer between wild-type and NOS2-/- mice, have indicated that leukocyte-derived NOS2, rather than parenchymal cell-derived NOS2, is the primary source of NO· in response to LPS challenge (18). Moreover, the expression or activation of NOS2 within leukocytes, rather than within parenchymal cells, appears to be responsible for the observed pulmonary microvascular leak in sepsis (54). Collectively, these various findings suggest that NO· mediates the acute inflammatory response to LPS, at least in part by participating in proinflammatory cytokine production by macrophages, which results in stimulation/injury of epithelial cells and neutrophil extravasation.
A central event in the inflammatory response to LPS is the activation of the transcription factor NF-κB (9, 41), and NO· has previously been shown to participate in NF-κB activation in alveolar macrophages and in other cell types (7, 33). The precise mechanism by which NO· mediates NF-κB activation is unclear but is presumed to involve intermediate formation of RNS, such as peroxynitrite, by interaction with macrophage-derived oxidants. The observed protein 3-NT in lung lavage fluids could be interpreted as evidence for such peroxynitrite formation, although this protein modification most likely resulted from oxidative reactions after the extravasation and activation of neutrophils (2, 15). Nevertheless, our results cannot exclude the possibility that peroxynitrite is involved in stimulation of NF-κB and proinflammatory cytokine production. An alternatively proposed mechanism for NO·-dependent NF-κB activation involves the S-nitrosation and activation of p21ras, thus initiating signaling cascades that lead to activation of NF-κB (32). In addition to its ability to activate NF-κB, many studies have demonstrated that NO· is also capable of suppressing NF-κB activation by several mechanisms (22, 40, 42). This dichotomous behavior appears to be related to the relative concentrations of NO· and the stage of NF-κB activation and proinflammatory gene expression. Indeed, inhibition of NO· synthesis was found to suppress cytokine production (this study) and NF-κB activation (7, 24) by LPS, but when NO· synthesis was inhibited at later time points after LPS stimulation, NF-κB activity and gene expression were actually found to be enhanced, indicating biphasic effects of NO· with stimulatory and inhibitory properties (7). Consistent with these findings are various reports showing that inhaled NO· actually inhibits NF-κB activation and inflammatory injury by LPS (20, 25). These dual effects of NO· are one major reason for the continued controversy regarding NO· in inflammation and injury. Nevertheless, the present study (in accordance with various previous studies) clearly shows an important role for NOS2-derived NO· in initiating the inflammatory response to LPS. Although such a proinflammatory effect of NOS2 could be interpreted as detrimental, because it is associated with epithelial permeability and injury, it also allows an appropriate host defense against several infectious agents (10, 21, 37, 55) and may, therefore, be beneficial and perhaps essential in some cases.
In addition to the participation of NOS2-derived NO· in LPS-mediated proinflammatory cytokine production, global gene expression analysis of lung tissues from untreated wild-type and NOS2-/- mice revealed several potential additional reasons that might explain the suppressed LPS-induced inflammatory response in NOS2-deficient mice. For example, expression of several genes involved in inflammatory-immune signaling, such as IL-1 receptor-associated kinase and lymphocyte-specific protein tyrosine kinase (which is homologous to p56Lck), was found to be suppressed in NOS2-/- mice compared with wild-type mice (Fig. 7, Table 3) (14, 23). Also, expression of some cytoskeletal proteins was markedly lower in NOS2-/- mice, which might decrease the ability of lung cells to respond to inflammatory stimuli. Future studies, designed to more specifically investigate the cellular origin of these genes and their relation to NOS2 expression or activity, are needed to assess their contribution to LPS-induced inflammation.
In summary, our results confirm that the presence of NOS2 contributes to the initiation of acute airway inflammation in response to LPS, which possibly constitutes a critical role of NOS2 in innate host defense. Moreover, our analysis of global gene expression revealed multifactorial regulation of inflammatory-immune activation by NOS2, with effects on basal expression of several transcription factors and signaling proteins, as well as effects on LPS-mediated induction of many inflammatory genes. Although global gene expression analysis defines the spectrum of genes that are affected by NOS2 deficiency or by LPS stimulation, it does not provide information on the contribution of the various cell types within the lung to the observed changes in gene expression. In addition, several genes that are known to be induced by LPS (including NOS2) were not detected using the GeneChip analysis, indicating some limitations to the application of oligonucleotide arrays, which may be attributed to poor design and selection of the antisense probes and to lack of sensitivity of the detection system. However, our results do indicate that phenotypical changes in LPS-induced inflammation in unconditional NOS2 knockout mice are most likely multifactorial and should not be solely attributed to participation of NO· in LPS-activated signaling pathways or to formation of RNS. Studies with conditional and/or tissue-specific NOS2 deficiency may shed additional light on the regulation of gene expression by NOS2 and its diverse roles in inflammatory-immune responses.
We thank Drs. H. Nishino and S. Azuma (Eicom) for the use of their HPLC-ECD 3-NT detection system and the Vermont Cancer Center for assistance in quantitative RT-PCR analysis.
This work was supported by research grants from the National Heart, Lung, and Blood Institute (HL-60812), the Cystic Fibrosis Foundation, and the University of California Tobacco-Related Disease Research Program (7RT-0167) and by a grant from the Uehara Memorial Foundation (Japan).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2004 the American Physiological Society