Inhibition of pulmonary nuclear factor kappa-B decreases the severity of acute Escherichia coli pneumonia but worsens prolonged pneumonia

Introduction Nuclear factor (NF)-κB is central to the pathogenesis of inflammation in acute lung injury, but also to inflammation resolution and repair. We wished to determine whether overexpression of the NF-κB inhibitor IκBα could modulate the severity of acute and prolonged pneumonia-induced lung injury in a series of prospective randomized animal studies. Methods Adult male Sprague-Dawley rats were randomized to undergo intratracheal instillation of (a) 5 × 109 adenoassociated virus (AAV) vectors encoding the IκBα transgene (5 × 109 AAV-IκBα); (b) 1 × 1010 AAV-IκBα; (c) 5 × 1010 AAV-IκBα; or (d) vehicle alone. After intratracheal inoculation with Escherichia coli, the severity of the lung injury was measured in one series over a 4-hour period (acute pneumonia), and in a second series after 72 hours (prolonged pneumonia). Additional experiments examined the effects of IκBα and null-gene overexpression on E. coli-induced and sham pneumonia. Results In acute pneumonia, IκBα dose-dependently decreased lung injury, improving arterial oxygenation and lung static compliance, reducing alveolar protein leak and histologic injury, and decreasing alveolar IL-1β concentrations. Benefit was maximal at the intermediate (1 × 1010) IκBα vector dose; however, efficacy was diminished at the higher (5 × 1010) IκBα vector dose. In contrast, IκBα worsened prolonged pneumonia-induced lung injury, increased lung bacterial load, decreased lung compliance, and delayed resolution of the acute inflammatory response. Conclusions Inhibition of pulmonary NF-κB activity reduces early pneumonia-induced injury, but worsens injury and bacterial load during prolonged pneumonia.


Introduction
Acute lung injury (ALI) and acute respiratory distress syndrome (ARDS) are life-threatening disorders, for which no specific therapy is known. When ARDS occurs in the setting of multisystem organ failure, mortality rates more than 60% have been reported, with significant morbidity in 50% of survivors [1]. ALI and ARDS develop most commonly in the context of severe sepsis [2], particularly infection with gram-negative bacilli such as Escherichia coli (E. coli) [3], and sepsis-induced ARDS has the worst outcome [4].
Nuclear factor kappa B (NF-B) is a key transcriptional regulator in the setting of inflammation and injury and plays a role in diverse inflammatory disorders, including acute lung injury [5]. Activation of NF-B occurs in response to diverse stimuli, such as endotoxin, which bind to cell-surface receptors that in turn activate the canonic and/or noncanonic signaling pathway. This signaling cascade ultimately results in the phosphorylation and inactivation of the cytosolic inhibitor IB complex, which then dissociates, allowing NF-B to translocate to the nucleus to initiate gene transcription [5]. Inhibition of NF-B reduces injury in preclinical models of ALI, including ischemia-reperfusion [6], endotoxemia [7], and cecal ligation and puncture-induced sepsis [8]. However, NF-B also exerts important cytoprotective effects, promoting cell survival, resolution of inflammation, and wound repair [9]. Of interest, NF-B signaling plays a central role in the host response to lung bacterial infection [10]. Consequently, inhibition of NF-B may constitute a doubleedged sword, particularly in pneumonia-induced ALI/ ARDS, in which immune competence is essential to eradication of the infectious agent [11].
We wished to determine the potential for inhibition of pulmonary NF-B activity to modulate the severity of pneumonia-induced lung injury. We used a gene-based therapy approach, via intrapulmonary delivery of three different doses of adenoassociated viral vector encoding the NF-B inhibitor IBa gene (AAV-IBα), to modulate the NF-B signaling pathway in the lung. We hypothesized that pulmonary overexpression of the NF-B inhibitor IBα would (a) attenuate the severity of the lung injury induced by acute E. coli pneumonia; but would (b) worsen the severity of prolonged E. coli pneumonia-induced lung injury; and (c) a dose-response relation would exist, with higher AAV-IBα doses having the greatest effect.

Materials and methods
Specific-pathogen-free adult male Sprague-Dawley rats (350 to 450 g) were used in all studies. The experimental model was based on those previously reported [12][13][14]. All work was approved by the National University of Ireland Galway Research Ethics Committee and conducted under license from the Department of Health, Ireland.

Preparation of AAV vectors
AAV-vector production was carried out as previously described, with several modifications [15]. The IBa-SuperRepressor (IBa-SR) gene (1,566 bp) was ligated into the pAAV-MCS vector (Agilent Technologies Inc., Santa Clara, CA, USA), and plasmid size confirmed by gel electrophoresis and validated by sequencing (Eurofins MWG Operon, Ebersberg, Germany). The IBa-FLAG plasmid DNA and AAV serotype 6 envelope were generated and sent to Virapur for AAV production (VIRAPUR, 6160 Lusk Blvd., Suite C-101, San Diego, CA, USA). AAV serotype 6 was chosen because AAV has reduced immunogenicity, the virus plasmid size is sufficient for the IBa-SuperRepressor (IBa-SR) gene, and serotype 6 has demonstrated tropism for lung epithelial cells.
Viral vector particle titers were determined with quantitative real-time polymerase chain reaction (qRT-PCR) and aliquoted and stored at -80 o C. As required, an aliquot was thawed and added to 75 μl of the porcine surfactant Curosurf (120 mg/ml) (Trinity-Chiesi Pharmaceuticals Limited, Cheadle, UK), and a final instillate volume of 300 μl was made up with PBS. For those animals receiving vehicle only, the instillate was 75 μl of Curosurf mixed with 225 µl of phosphate-buffered saline (PBS). Curosurf surfactant was added to each instillate, as it was demonstrated in prior studies to enhance spread and improve transgene expression [16].

Vector instillation
Animals were anesthetized by inhalational induction with isoflurane and an intraperitoneal injection of 40 mg/kg ketamine (Pfizer, Kent, UK). After confirmation of depth of anesthesia, laryngoscopy was performed (Welch Allyn Otoscope; Buckinghamshire, UK), and the trachea intubated with a size 16 intravenous catheter (BD Insyte; Becton Dickinson Ltd., Oxford, UK). After instillation of vector or vehicle, depending on the specific series, animals were extubated and allowed to recover from anesthesia.

E. coli/vehicle instillation
The E. coli used in these experiments is E5162 (serotype: O9 K30 H10) and was supplied by the National Collection of Type Cultures, Central Public Health Laboratory, London, England. The E. coli were stored on preservative beads (Protect, Lancashire, England) at -80°C. Beads were placed in 3-ml vials of peptone water (Cruinn Diagnostics, Dublin, Ireland) and incubated at 37°C for 18 to 24 hours to allow bacterial concentrations to reach a plateau. The bacterial suspension was then centrifuged and washed in phosphate-buffered saline to produce the inoculum. The bacterial load in each inoculum was determined by plating serial dilutions on agar plates. Preliminary experiments were performed to determine the bacterial load of intratracheal E. coli required to produce a lung injury over a 4-hour and over a 72-hour period.

Acute pneumonia protocol
Ninety-six hours after virus instillation, animals were anesthetized with intraperitoneal 80 mg/kg ketamine and 8 mg/kg xylazine, and anesthesia maintained with Alfaxalone (Alfaxadone 0.9% and alfadadolone acetate 0.3%). A tracheostomy tube was inserted, and intraarterial access was sited in the carotid artery. Muscle relaxation was induced with cisatracurium besylate, and the lungs were mechanically ventilated with 30% O 2 in 70% N 2 . After 20 minutes, arterial blood gas measurement was performed and 1 × 10 11 E. coli in a 300-μl PBS suspension (or vehicle alone) instilled via the tracheostomy [12,13]. Animals were ventilated for 4 hours, with systemic arterial blood pressure, peak airway pressure, and body temperature continually measured. Lung compliance and arterial blood gas analysis was measured hourly, and body temperature was maintained at 36°C to 37.5°C.

Prolonged pneumonia protocol
Animals were anesthetized by inhalational induction with isoflurane and intraperitoneal 40-mg/kg ketamine (Pfizer, Kent, UK). After confirmation of anesthesia depth, 5 × 10 9 E. coli in a 300-μl PBS suspension was instilled into the trachea under direct vision, and the animals allowed to recover [17]. Animals were monitored closely for 72 hours after E. coli instillation, and then reanesthetized, tracheostomized, and mechanical ventilation was instituted, and injury severity assessed, as described [18].

Postmortem analyses
At the end of the protocols, the animals were killed by exsanguination under anesthesia, and the plasma snapfrozen for later analysis. The heart-lung block was dissected from the thorax, bronchoalveolar lavage (BAL) was performed, and BAL fluid differential leukocyte counts and lung bacterial colony counts were completed. BAL fluid was centrifuged, and the supernatant was snap-frozen and stored at -80°C. BAL concentrations of IL-1β, TNF-a, IL-6, CINC-1, IL-10, and KGF were determined by using ELISA (R&D Systems, Abingdon, UK), and BAL protein concentrations were measured (Micro BCA Protein assay kit; Pierce, Rockford, IL, USA) [16]. The left lung was isolated and fixed, and the extent of histologic lung damage determined by using quantitative stereologic techniques [18].

Assessment of transgene expression and efficacy
IBa-FLAG transgene expression was determined in lung homogenates with real-time PCR and Western blotting, as previously described [16,19]. In brief, RNA was extracted from lung tissue, cDNA was synthesised, and quantitative PCR was performed for IBa-FLAG, normalized against a GAPDH control product. For Western blot IBα-FLAG analysis, total cell protein was extracted, protein concentration was determined, and samples were electrophoresed on an SDS-PAGE gel and transferred to nitrocellulose [20]. Primary anti-human IBα-FLAG monoclonal antibody (Sigma-Aldrich, St. Louis, MO, USA) was used, with secondary antibody conjugated to horseradish peroxidase (Cell Signaling Technology, Danvers, MA, USA), and the membrane incubated with a chemiluminescent substrate (Super-Signal West Pico; Pierce).
The effect of IBa overexpression on the activation of the NF-B pathway was assessed by measurement of nuclear accumulation of the activated P65 subunit of NF-B [19]. Nuclear extracts were performed on homogenized rat lung tissue by using a NE-PER Nuclear and Cytoplasmic Extraction Kit (Fisher Scientific Ireland, Dublin, Ireland), and NF-B (p65) measured by using an NF-B Transcription Factor Assay Kit (Cayman Chemical Company, Ann Arbor, MI, USA).

Data presentation and analysis
Continuous responsive variables are summarized by using mean (SD) and median (interquartile range, IQR) as necessary. The proportion of animals surviving was analyzed by using the χ 2 test. All other data were analyzed with one-way ANOVA, followed by the Dunnett test or by Kruskal-Wallis, followed by the Dunn test, with the vehicle group used as the reference group for all comparisons. The assumptions underlying all models were checked by using suitable residual plots. A P value of <0.05 was considered statistically significant.
IBα expression and function Pulmonary instillation of AAV-IBa produced a dose-dependent increase in lung IBa gene transcription ( Figure 1A) and protein production ( Figure 1B, C). Densitometry of IBα-FLAG Western blot (n = 3 per group) confirmed this dosedependent increase in IBα protein ( Figure 1C). IBa overexpression decreased E. coli-induced NF-B activation, as measured by nuclear accumulation of the activated P65 subunit of NF-B ( Figure 1D).
Survival and bacterial load IBα overexpression enhanced animal survival, with survival rates of 33% in the Vehicle group, 58% with 5 × 10 9 IBa, and 92% with 1 × 10 10 IBa Of concern, the highest IBa (5 × 10 10 ) dose abolished the survival benefit (Table 1). IBa overexpression also enhanced the duration of animal survival, which followed this same pattern, with the effect greatest with the1 × 10 10 IBa dose (Table 1). Importantly, IBα did not alter lung bacterial loads (Table 1).
E. coli-induced decrement in arterial oxygenation, with benefit greatest at intermediate (1 × 10 10 ) IBa vector dose, but abolished at the higher (5 × 10 10 ) IBa dose ( Figure 2A and Table 1). A similar pattern was seen with regard to the decrement in static compliance ( Figure 2B), and pulmonary permeability, as assessed by protein leak into the BAL fluid ( Figure 2C), but no effect was seen in regard to BAL neutrophil or mononuclear cell counts ( Figure 2D and Table 1). IBa reduced E. coli-induced histologic evidence of lung injury, with the intermediate (1 × 10 10 ) vector dose of IBa again most effective ( Figure 3A through E).

Animal Survival and Bacterial Load
IBα did not alter the number or duration of animal survival (Table 4). However, IBα substantially increased lung E. coli bacterial loads at all three IBα doses (Table 4).
sham (vehicle) instillation. The null or IBα transgenes did not alter lung function in sham infected animals (Table 5).

Discussion
Pulmonary overexpression of IBα attenuated acute pneumonia-induced lung injury, decreasing the severity of the decrement in lung function, whereas also decreasing the inflammatory response. In contrast, IBα worsened lung injury and inflammation induced by prolonged pneumonia, increasing lung bacterial load and delaying resolution of the inflammatory response. These findings provide novel insights into the effects of NF-B in pneumonia-induced lung injury, and raise concerns regarding therapeutic potential of inhibiting NF-B, particularly in prolonged untreated pneumonia.

E. coli acute and prolonged pneumonia
Infection with gram-negative bacilli such as E. coli is the commonest cause of ARDS [21,22] and is also a very common complication of ARDS due to other causes [23]. We studied E. coli-induced pneumonia, a wellcharacterized animal model that mimics the clinical development of ARDS very closely [23][24][25][26][27][28]. Intratracheal instillation of 1 × 10 11 E. coli resulted in physiological and pathologic changes consistent with a severe lung injury over a 6-hour period, similar to that previously reported [12,13]. Instillation of a lower dose of 5 × 10 9 E. coli produced a more gradually evolving injury over a 72-hour period, as previously described [14,17].

NF-B: role in lung inflammation
The role of the NF-B signaling in the host immune response to lung injury is increasingly well understood [19]. NF-B activation pathway gene polymorphisms alter the susceptibility to [29] and severity [30] of clinical ARDS. NF-B is a dimer of a number of related proteins, including RelA (also known as P65), p50, p52, RelB, and cRel, with the RelA and p50 heterodimer, the most common form. On cell activation, by stimuli such as gram-negative bacterial endotoxin, diverse signaling pathways are activated, which converge to phosphorylate and activate the IB kinase complex proteins (IKK), which then phosphorylate and inactivate the IB proteins, which dissociates from NF-B. The active NF-B translocates to the nucleus and binds to specific cognatebinding sequences in the promoter or enhancer regions of different target genes to initiate transcription [5].
The therapeutic potential of strategies to inhibit NF-B is evident from the demonstration that NF-B inhibition decreases injury in nonseptic ALI models, including pulmonary [6] and systemic reperfusion, and endotoxemia [7]. Pulmonary overexpression of the RelB member of the NF-B family decreases cigarette smoke-induced lung injury [31].

ARDS, pneumonia, and NF-B
The effects of modulation of the NF-B in the setting of lung bacterial infection are less well understood [19]. NF-B decoy oligodeoxynucleotides reduce ALI in mice in the early phases of cecal ligation and puncture-induced sepsis [8]. Selective inhibition of vascular endothelial NF-B activity in endotoxemic transgenic mice reduced lung inflammation and increased survival [32]. Of interest, this approach improved survival and systemic organ function, but did not alter bacterial clearance, in septic mice [32]. However, others have found inhibition of NF-B signaling to exert detrimental effects in the setting of infection. Inhibition of hepatocyte NF-B activity reduced Listeria monocytogenes clearance, decreasing murine survival [11]. In contrast, clearance of pseudomonas bacteria from the mouse lung was enhanced by pulmonary overexpression of the RelA subunit of NF-B [33], suggesting that the NF-B pathway plays a pivotal role in maintaining immune competence and is essential to eradication of the infectious agent [11].

NF-B: role in lung inflammation, injury, and repair
NF-B also promotes cell survival, resolution of inflammation, and repair after injury. Inhibition of NF-B signaling retards pulmonary [20] and intestinal [34] epithelial wound healing. Maturational differences in lung NF-B activation profiles exist, with NF-B activation protecting the lung against hyperoxia [35] and endotoxemia [36] in neonatal rats. Consequently strategies to inhibit NF-B, particularly if not spatially or temporally targeted, may have detrimental effects.

Targeting pulmonary NF-B
We wished to determine whether inhibition of pulmonary NF-B activity could modulate the severity of pneumoniainduced lung injury. We found that pulmonary overexpression of the IBa gene did reduce acute pneumonia-induced injury. IBα decreased the decrement in arterial oxygenation and lung static compliance, decreased alveolar protein leak, and decreased histologic injury, compared with vehicle. IBα modulated the cytokine response to E. coli instillation, decreasing alveolar IL-1β but increasing CINC-1 concentrations. Of importance, these effects were dose dependent, with benefit maximal at the intermediate (1 × 10 10 ) IBa dose, and a loss of efficacy at the higher (1 × 10 10 ) IBa concentration. In contrast, in prolonged untreated E. coli pneumonia, pulmonary IBa gene overexpression worsened the lung injury, and increased lung bacterial load. IBα delayed the resolution of the acute inflammatory response, increasing the proportion of alveolar neutrophils while decreasing alveolar mononuclear cells, which comprise lymphocytes and macrophages and are considered important to repair. IBα increased alveolar concentrations of TNF-α and IL-1β, cytokines implicated in the early phase of the response to septic insult. IBα increased alveolar IL-10 concentrations, which may partly explain the ineffective response to the bacterial insult.

NF-B and sepsis: an integrated paradigm
The contrasting effects of NF-B inhibition in acute versus prolonged pneumonia generate a number of important insights. First, the finding that NF-B inhibition reduced the severity of acute pneumonia and decreased the host response induced ALI suggests that any benefit in this setting is mediated via a decreased host response to the instillation of E. coli [37]. This is consistent with the concept that in the early phases of pneumonia, secreted toxins, and the host immune response may be the predominant source of injury [37,38]. This is supported by the fact that NF-B inhibition is protective in nonseptic inflammatory ALI models. Second, this effect was dose dependent, with lower and intermediate IBα doses protective and the beneficial effects ablated at higher doses. This suggests a U-shaped dose-effect curve, reducing the therapeutic utility of approaches to inhibit NF-B in the setting of acute pneumonia. Third, IBα worsened the severity of prolonged pneumonia, substantially increasing lung bacterial loads, and resulting in a persisting pulmonary inflammatory response. IBα, by inhibiting the early host response to E. coli instillation, may have led to a persistence of infection, ultimately worsening lung injury. Finally, the beneficial effects of NF-B inhibition in acute pneumonia were lost at the highest dose. The reasons for this is unclear, but may include NF-B inhibition mediated reduction in lung epithelial repair [20], apoptosis of airway epithelial cells [16] and effects on lung macrophage function [39].

Conclusions
IBα overexpression reduces injury severity and inflammation in early pneumonia, but slows resolution of inflammation and bacterial clearance and worsens injury during prolonged pneumonia. These findings raise serious questions regarding the utility of strategies that inhibit the innate immune response, such as NF-B, in live bacterial pneumonia.
Key Message: Inhibition of pulmonary NF-B activity decreases the severity of early pneumonia-induced lung injury, but worsens injury severity and bacterial load during prolonged pneumonia.
Abbreviations AAV: adenoassociated virus; ALI: acute lung injury; ANOVA: analysis of variance; ARDS: acute respiratory distress syndrome; BAL: bronchoalveolar lavage; CINC-1: cytokine-induced neutrophil chemoattractant-1; DNA: deoxyribonucleic acid; E. coli: Escherichia coli; IκBα: inhibitory factor kappa B alpha; IL: interleukin; IQR: interquartile range; KGF: keratinocyte growth factor; qRT-PCR: quantitative real-time polymerase chain reaction; NF-κB: nuclear factor kappa B; PBS: phosphate-buffered saline; RNA: ribonucleic acid; TNF-α: tumor necrosis factor-α Authors' contributions JD performed the animal experiments, assays, and histologic analyses, analyzed the data, drafted the manuscript, and agreed to the final submitted version. GFC and MH assisted with animal experiments, contributed to drafting the manuscript, and agreed to the final submitted version. CM, BA, and DOT performed assays and histologic analyses, contributed to drafting the manuscript, and agreed to the final submitted version. TOB conceived and designed the experiments, provided viral vector expertise, drafted the manuscript, and agreed to the final submitted version. JL conceived and designed the experiments, analyzed the data, drafted the manuscript, and agreed to the final submitted version.

Competing interests
The authors declare that they have no competing interests.