Abstract
Airway disease in cystic fibrosis (CF) is characterised by impaired mucociliary clearance, persistent bacterial infection and neutrophilic inflammation. Lipoxin A4(LXA4) initiates the active resolution of inflammation and promotes airway surface hydration in CF models. 15-Lipoxygenase (LO) plays a central role in the “class switch” of eicosanoid mediator biosynthesis from leukotrienes to lipoxins, initiating the active resolution of inflammation. We hypothesised that defective eicosanoid mediator class switching contributes to the failure to resolve inflammation in CF lung disease.
Using bronchoalveolar lavage (BAL) samples from 46 children with CF and 19 paediatric controls we demonstrate that the ratio of LXA4to leukotriene B4(LTB4) is depressed in CF BAL (p<0.01), even in the absence of infection (p<0.001).
此外,15-LO2成绩单是显著的less abundant in CF BAL samples (p<0.05). In control BAL, there were positive relationships between 15-LO2 transcript abundance and LXA4/LTB4ratio (p=0.01, r=0.66) and with percentage macrophage composition of the BAL fluid (p<0.001, r=0.82), which were absent in CF.
Impoverished 15-LO2 expression and depression of the LXA4/LTB4ratio are observed in paediatric CF BAL. These observations provide mechanistic insights into the failure to resolve inflammation in the CF lung.
Abstract
Reduced 15-LO2 expression in the lower airways of children with CF, associated with a depressed LXA4/LTB4ratiohttp://ow.ly/tzZWa
Introduction
Cystic fibrosis (CF) is caused by mutation of the gene encoding the cystic fibrosis transmembrane conductance regulator (CFTR) chloride channel and characterised by impaired mucociliary clearance, persistent bacterial infection and neutrophil mediated inflammation.
Sequential biosynthesis of eicosanoid classes in inflammatory exudates (termed “class switching”) directs the temporal progression of acute inflammation from initiation to propagation and finally to active resolution [1]. Prostaglandins initiate the inflammatory response. Leukotrienes follow and are typified by leukotriene B4(LTB4) which amplifies and propagates inflammation [1]. LTB4and interleukin (IL)-8 are both potent neutrophil chemoattractants [2,3] and negatively correlate with pulmonary function in CF [4,5]. Lipoxin A4(LXA4) biosynthesis initiates the active resolution phase of inflammation [6], followed by biosynthesis of the resolvins and protectins.
LXA4inhibits neutrophil effector functions [7] and counter-regulates the effector functions of LTB4, most notably inhibiting LTB4-induced neutrophil transmigration [8–10]. LXA4also suppresses production of IL-8 by leukocytes and bronchial epithelial cells [11–13]. Mice treated with analogues of LXA4and subsequently challenged withPseudomonas aeruginosacontained the bacterial challenge more effectively [14]. LXA4augments airway epithelial innate defence by stimulating tight junction formation [15], enhancing calcium-activated chloride secretion [16], and by restoration of airway surface liquid height in CF airway epithelial cells [17]. LXA4concentration in CF bronchoalveolar lavage (BAL) has been variously reported as significantly suppressed or not significantly different from controls [14,18].
LXA4is biosynthesised in the respiratory tract by transcellular cooperation of neutrophils [19], eosinophils [20], alveolar macrophages [21] or airway epithelial cells [22], each expressing different lipoxygenase (LO) enzymes (fig. 1) [2,23]. Upregulation of 15-LO activity favours LXA4生物合成的白三烯biosynthesis [1,24]. This occurs both as a result of 15-LO product, 15-hydroxyeicosatetraenoic acid (HETE), competing for catalytic sites at the 5-LO enzyme and thereby reducing the formation of 5(S) HETE, and by competition for the common biosynthetic intermediate leukotriene A4[1,22–24].
We hypothesised that “eicosanoid mediator class switching” is defective in CF and may play a mechanistic role in the failure to resolve inflammation in CF lung disease. We measured eicosanoids associated with the “class switch” [1] from propagation (leukotrienes) to resolution of inflammation (lipoxins), and the relative expression of mRNA transcripts for eicosanoid synthetic enzymes (5-LO, 12-LO and 15-LO1/2 and leukotriene A4hydrolase (LTA4H)) in BAL from children with CF and compared this with paediatric controls. In addition, we conducted subgroup analysis based on the presence or absence of microbial pathogens (>104CFU·mL−1) detected in the BAL fluid.
Methods
Study population
Children with CF were recruited through SHIELD CF (Study of Host Immunity and Early Lung Disease in Children with CF), a longitudinal study established around the preschool surveillance bronchoscopy programmes at Our Lady’s Children Hospital, Crumlin (OLCHC), the National Children’s Hospital, Tallaght and University Hospital Limerick, Limerick, Ireland. Children without CF undergoing bronchoscopy for clinical reasons were recruited as paediatric controls (see the online supplementary material for indications and treatment characteristics). Assessments were undertaken when the children were in a stable clinical condition. Children were excluded from the analysis if they were taking leukotriene receptor antagonists or treatment doses of antibiotics in the 2 weeks prior to bronchoscopy. Ethical approval for this study was granted by the research ethics committee at each hospital and written informed consent was obtained from the parents of all participating subjects. Bronchoscopy, BAL collection and processing to obtain a supernatant and cell pellet were performed as previously described (online supplementary material) [25].
Measurements
落下帷幕的整除被送到了微生物学实验室at OLCHC where testing, including conventional bacterial and viral studies, was performed according to Clinical Pathology Accreditation guidelines (online supplementary material). Samples were considered to have positive microbial culture if one or more pathogen was detected on viral studies or cultured at a concentration of >104CFU·mL−1.
LXA4and LTB4concentrations were measured by ELISA (Oxford Biomedical Research, Rochester Hills, MI, USA and Cayman Chemical, Ann Arbor, MI, USA) (for specificities of antisera see online supplementary table E1). IL-8 was measured by ELISA using mouse anti-IL8 capture antibody and biotinylated goat anti-IL8 detection antibody (R&D Systems, Abingdon, UK). Neutrophil elastase (NE) activity was measured as described previously (online supplementary material) [25]. Total and differential cell counts were obtained by the trypan blue exclusion method and light microscopy (for further detail see the online supplementary material).
Total RNA was extracted from the BAL cell pellets using RNAqueous kit (Ambion Inc., Austin, TX, USA). RNA was reverse transcribed using a High Capacity cDNA Reverse Transcription kit (Applied Biosystems, Dun Laoghaire, Ireland). Real-time PCR was performed on a 7900HT Fast Real-Time PCR System (Applied Biosystems) using TaqMan probes to detect genes coding for arachidonate 5-lipoxygenase (ALOX5), arachidonate 12-lipoxygenase (ALOX12), arachidonate 15-lipoxygenase (ALOX15), arachidonate 15-lipoxygenase type B (ALOX15B) and LTA4H (Applied Biosystems) and 18S ribosomal RNA used as internal control. The threshold cycle for each target gene was normalised using the Ct of the endogenous control (18S). Relative fold expression for each target was calculated as (2(−ΔΔCt)) in comparison to a reference patient.
Statistical analysis
Graphpad prism (GraphPad Software, Inc., La Jolla, CA, USA) was used to perform statistical analyses. Results are presented as mean±sdor mean±sem. The groups were compared using Fisher’s exact test for categorical variables and unpaired t-test for continuous variables. Correlations were analysed using the Spearman test. No correction was made for multiple comparisons. Results were reported as statistically significant with p-values <0.05, <0.01 and <0.001.
Results
51 BAL samples from children with CF and 22 from paediatric control patients were assessed for eligibility in this study. Five children with CF and three control children were excluded because of leukotriene receptor antagonist treatment or recent use of treatment dose antibiotics. CF and control children did not differ significantly in age, sex or BAL neutrophil count (table 1). Control children with positive BAL microbial culture were younger than control children with negative BAL microbial culture. Children with CF with positive BAL microbial culture were older than children with CF with negative BAL microbial culture. Children with CF with positive BAL samples had higher mean neutrophil counts and more frequently detectable NE than culture negative CF samples (table 1). The most common bacterial pathogens cultured wereHaemophilus influenzae,Staphylococcus aureusandStreptococcus pneumonia.P. aeruginosa>104CFU·mL−1was identified in two CF and one control BAL sample. Viral studies were negative for all BAL samples (CF and controls).
We found no significant differences between control and CF samples in mean LXA4or LTB4concentration (fig. 2a and c). LXA4and LTB4concentrations adjusted for neutrophil and macrophage count and examined by microbial culture result are presented in online supplementary figure E1. IL-8 was higher in CF samples than in control samples (1198±149 ng·mL−1versus379±110 ng·mL−1, p<0.01) (fig. 2e).
Comparison of culture negative and positive control and CF samples revealed no significant difference in LXA4concentration (fig. 2b). LTB4concentration was significantly higher in infected samples within a group (control with positive microbial culture 155±40 pg·mL−1versuscontrol with negative culture 26±12 pg·mL−1, p<0.05; and CF with positive microbial culture 1383±454 pg·mL−1versusCF with negative culture 156±35 pg·mL−1, p<0.01) (fig. 2d). IL-8 was higher in CF BAL with negative culture than in control BAL with negative culture (817±178 ng·mL−1versus110±28 ng·mL−1, p<0.05) (fig. 2f). CF BAL samples with positive microbial culture had higher IL-8 concentrations than either CF samples with negative culture (1614±215 ng·mL−1versus817±178 ng·mL−1, p<0.01) or control samples with negative culture (1614±215 ng·mL−1versus536±159 ng·mL−1, p<0.01).
To evaluate the relative concentrations of these mediators in BAL we calculated LXA4/LTB4and LXA4/IL-8 ratios. Children with CF had lower LXA4/LTB4ratios in BAL fluid compared with control children (2.9±0.7versus7.1±1.9, p<0.01) (fig. 3a). Among control samples, the LXA4/LTB4ratio in sterile samples was significantly higher (13.1±3.6versus3.1±3.6,p < 0.01),在矿山测量样本from which pathogens were cultured (fig. 3b). In CF BAL, the LXA4/LTB4ratio was uniformly depressed and did not vary with infection status (p=0.98). The LXA4/LTB4ratio measured in CF samples with negative microbial culture was significantly depressed in comparison to control BAL with negative culture (2.9±0.7versus13.1±3.6, p<0.001).
The LXA4/IL-8 ratio was significantly higher in control BAL than in CF BAL (3.5×10−3±1.7×10−3versus0.5×10−3±0.1×10−3, p<0.01) (fig. 3c). The LXA4/IL-8 ratio was significantly higher in control samples with negative microbial culture than in controls with positive BAL microbial culture (7.9×10−3±4.2×10−3versus0.9×10−3±0.2×10−3, p<0.05) (fig. 3d). The LXA4/IL-8 ratio measured in CF BAL with negative microbial culture was significantly lower compared with control BAL with negative microbial culture (0.6×10−3±0.2×10−3versus7.9×10−3±4.2×10−3, p<0.01). In CF BAL, the LXA4/IL8 ratio was uniformly depressed and the correlation with infection seen in control BAL was absent.
14 control (seven with negative and seven with positive microbial culture) and 11 CF (six with negative and five with positive microbial culture) BAL cell pellets with a similar distribution of infection, and neutrophil (p=0.95) and macrophage (p=0.71) compositions were selected from the original sample set to study the expression of genes involved in LXA4and LTB4生物合成在气道腔(fig. 4e and f). The cellular fraction obtained from the airway lumen contained the full repertoire of transcripts required for LXA4biosynthesis (fig. 4). ALOX15B mRNA was significantly less abundant in CF samples compared to control samples (0.44±0.13versus2.63±0.92 fold expression, p<0.05) (fig. 4a). A nonsignificant trend towards lower ALOX15 expression was observed in CF BAL than in control BAL (1.92±0.49versus5.51±1.89 fold expression, p=0.09) (fig. 4b). ALOX5 and LTA4H mRNA abundance did not significantly differ between control and CF samples (fig. 4c and d). ALOX12 mRNA was measurable in only six out of 14 control samples and in four out of 11 CF samples.
In control BAL samples there was a significant relationship between ALOX15B mRNA abundance in the cell pellet and LXA4/LTB4ratio (r=0.66, p=0.01) (fig. 5a). In CF samples this correlation was absent (p=0.31) (fig. 5b). Neutrophils and macrophages are the most abundant cell types in the BAL cell pellet, and in control samples ALOX15B mRNA abundance correlates positively, strongly and significantly with macrophage percentage composition of the pellet (r=0.82, p=0.0003) and negatively with neutrophil percentage composition (r= -0.82, p=0.0003) (fig. 5c and d). In CF samples, despite variability in the macrophage composition of the samples, the correlation of ALOX15B abundance with macrophage/neutrophil composition was absent (p=0.20 and p=0.17, respectively) (fig. 5e and f). Variability in the abundance of ALOX5 mRNA, recognised to be expressed abundantly both by neutrophils and alveolar macrophages [26–28], was not significantly correlated with cell type for control or CF samples.
Discussion
This is the first study to demonstrate defective eicosanoid class switching, a key step in initiating the active resolution of inflammation, in the lower airways of young children with CF despite the absence of recognised pathogens. We presentin vivoevidence of impoverished 15-LO2 gene expression associated with a depressed LXA4/LTB4ratio in the lower airways of children with CF. We found that the correlations between LXA4/LTB4ratio, infection, abundance of 15-LO2 transcripts and macrophages observed in control BAL were lost in CF BAL. This report provides new and important mechanistic insights into the failure to resolve inflammation in CF lung disease.
In order to test the hypothesis that eicosanoid class switching is defective in CF, giving rise to a failure to resolve acute inflammation, we compared eicosanoids temporally associated with the “switch” from propagation to resolution, LTB4and LXA4, respectively. In addition, with the expectation that the phase of inflammation would be affected by the infectious context [29], we conducted subgroup analysis based on the presence or absence of microbial pathogens detected in the BAL fluid. By comparing relative mediator concentrations, expressed as a LXA4/LTB4ratio, we demonstrate that the balance in eicosanoid biosynthesis weighs heavily in favour of LXA4production over LTB4production in control BAL when compared to CF BAL (fig. 3a). This finding indicates an abnormality in the active resolution of inflammation in CF and is consistent with the reduced LXA4/IL-8 ratio in CF BAL compared to control BAL also reported here (fig. 3c).
Consistent with a previous report [18], we did not find significant differences when comparing control and CF BAL, directly or in subgroup analyses (based on microbial culture result), as regards the absolute content of LXA4(fig. 2a and b). One other study suggested there is a failure to resolve acute inflammation in CF and reported a significant suppression in the ratio of LXA4/neutrophils in CF [14]. In our study, we did not find any significant difference in the LXA4/neutrophils ratio between control and CF BAL (online supplementary fig. E1a). This could potentially be explained by differences in the control population between that study and ours, especially with regards to allergic airways disease and infection. Although LTB4concentration did not significantly differ between control and CF BAL (fig. 2c), we found that LTB4concentration was always higher in infected BAL subgroups (fig. 2d) consistent with its role in neutrophil recruitment and host defence. It is noteworthy that a negative correlation has been described between LTB4and pulmonary function outcomes in CF [4].
LXA4and its synthetic analogues have been shown,in vitro, to suppress IL-8 production by leukocytes and bronchial epithelial cells [11,12]. As such, the LXA4/IL-8 ratio represents a compound outcome incorporating both the local concentration of each mediator, and the effectiveness of LXA4in antagonising IL-8 production. In control samples the LXA4/IL-8 ratio was significantly lower in the presence of infection in BAL, consistent with an appropriate pro-inflammatory response to infection. In contrast, the higher LXA4/IL-8 ratio seen in uninfected control BAL, favouring of the resolution of inflammation, was absent in uninfected CF BAL where the ratio remained low independent of infection.
In the absence of airway infection in control children (control samples with negative microbial culture), the balance between LXA4and LTB4biosynthesis was skewed in favour of LXA4production, reflecting ongoing active resolution of inflammation. Furthermore, our study providesin vivoevidence that LXA4/LTB4balance in the lower airways of control children correlates strongly with the abundance of 15-LO2 transcripts in the cellular phase of BAL. This finding is consistent within vitroreports that 15-LO enzyme activity promotes LXA4生物合成的白三烯biosynthesis [1,24]. In the control population, we also find a positive correlation between 15-LO2 transcript abundance and the percentage macrophage composition of the BAL cell pellet, and, reciprocally, a negative correlation with the percentage neutrophils (fig. 4e and f). This is consistent with the recognised expression of 15-LO, and 15(S)-HETE and LXA4generation by human alveolar macrophages [21]. Although neutrophils and macrophages in induced sputum both expressed 15-LO, the majority has been attributed to macrophages [30].
Taken together our observations suggest that macrophages, by expressing 15-LO2 and participating in transcellular cooperation with neutrophils expressing 5-LO, make a significant contribution to LXA4synthesis in the bronchial lumen and regulate eicosanoid balancein vivoin the airway.
In CF, in contrast to our findings in BAL samples from controls, the bias in favour of LXA4production in uninfected BAL was absent. LXA4/LTB4and LXA4/IL-8 ratios were not significantly different between not infected and infected CF BAL subgroups (fig. 3b and d). In order to further our understanding of impaired eicosanoid class switching in CF we considered the roles played by various enzymes involved in eicosanoid biosynthesis (12-LO, LTA4H, 5-LO, 15-LO1 and 15-LO2). The activity of 12-LO in platelets from CF patients has been reported to be reduced [31]; however, in our study, 12-LO transcripts were detected at very low copy numbers in only 10 out of 25 airway samples tested, a finding which could possibly represent blood cell contamination from a friable and inflamed airway wall. Overexpression of LTA4H could result in a diversion phenomenon, routing the common immediate leukotriene A4away from LXA4synthesis towards LTB4synthesis giving rise to the reduced LXA4/LTB4ratio observed; however, expression was not significantly different in CF BAL than in control BAL. We observed nonsignificant trends towards lower abundance of 15-LO1 and 5-LO transcripts in the CF airway (fig. 4b and c).
Importantly, we found reduced 15-LO2 gene expression in CF samples (fig. 4a). Furthermore, in BAL from children with CF there was a breakdown in the relationships between 15-LO2 transcript abundance, LXA4/LTB4ratio and percentage macrophage composition of the BAL. The macrophage and neutrophil content of the CF BAL cell pellets analysed had a wide range of composition and a similar distribution to the control samples, thus, reduced expression of 15-LO2 is not attributable to the neutrophil cell population simply overwhelming the macrophage population numerically. In fact, reduced 15-LO activity in the face of preserved 5-LO activity provides a reasonable explanation for the excess of LTB4over LXA4biosynthesis seen in CF subjects free from infection in this study.
These findings, describing impoverished ability to restore normal eicosanoid mediator balance provide mechanistic insights into the persistence of neutrophil mediated inflammation in lieu of its active resolution in CF lung disease. In light of reports that LXA4/LTB4balance is disturbed across the spectrum of asthma severity [32–35] and in scleroderma lung disease [36], and given that the D-series resolvins [37], protectin D1 [38] and 15(S) HETE (an agonist of the anti-inflammatory nuclear receptor peroxisome proliferator-activated receptor-γ) [39] are also 15-LO products, these mechanistic insights may have wider relevance in the pathophysiology of inflammatory respiratory disease.
15-LO expression is a hallmark feature of the alternative activation phenotype (M2) in macrophages [40], induced by the action of IL-4 and IL-13 (produced by T-helper 2 cells) [41,42]. While recent studies have identified a bronchial mucosal lymphocytic infiltrate in the CF airway their role in CF airway disease is poorly understood [43]. Deficient expression of IL-4/IL-13, delayed or impaired macrophage differentiation (such as that seen in bronchial epithelial cells in CF [44),或者分化走向经典激活(M1) over alternative activation could all contribute to the observed reduction in the expression of 15-LO reported here. In an example of feed-forward amplification, the 15-LO product resolvin D1 promotes M2 macrophage differentiation and, thus, 15-LO deficiency could lead to reduced resolvin D1 biosynthesis and consequently to reduced M2 macrophage differentiation, further compounding the deficiency in 15-LO expression [45]. Alveolar macrophage defects are widely reported in CF, although M1/M2 balance in CF alveolar macrophages remains a subject of contention (reviewed in [46]). Hostile environmental factors could play a role in the failure to actively resolve inflammation observed here. While we examined the impact of conventional microbial infection upon eicosanoid balance we cannot exclude the possibility that sterile inflammatory stimuli (e.g.pulmonary aspiration [47] or adherent mucus plaques) or non-culturable microbes [48] affected eicosanoid expression and synthesis. Finally, CFTR expression has been demonstrated in non-CF macrophages and CFTR inhibition disrupted macrophage function; thus, it is plausible that CFTR deficiency could directly affect the regulation of ALOX15B expression, for exampleviaendoplasmic reticulum stress and the unfolded protein response [49].
Strengths and limitations
Although there were differences between the mean ages of children with infection in their BAL within the control and CF subgroups, we found no significant correlation between LXA4/LTB4ratio and age among control children (p=0.31) or children with CF (p=0.49). The measured neutrophil count in our CF cohort is consistently lower than that recently reported for another large study of early lung disease in CF, the AREST CF (Australian Respiratory Early Surveillance Team for Cystic Fibrosis) cohort [50]. There are differences in lavage sampling protocol between the two studies. In our study lavage fluid is obtained from two instillations into the right middle lobe and two into the lingula, with the entire lavage then being pooled and analysed as a homogenous fluid. In the AREST CF study the right middle lobe is lavaged three times, followed by either the lingula or the “most affected lobe on computed tomography (CT)” and different lavage fractions are used for microbiology and inflammation studies [50]. These protocol differences may give rise both to dilutional variation and differences arising from the selection bias introduced by the selection of lavage zone by CT in the AREST CF study. In our study the proportion of children, stratified by age, with infection cultured from BAL was lower than that reported for the AREST CF cohort at similar ages, 14.2% and 33.3% of children in the second and third years of lifeversus21.1% and 40.2% at the same ages in the AREST CF cohort [50]. The definition of infection we have employed differs from that used by AREST CF, with a more stringent criterion of 104CFU·mL−1versus105CFU·mL−1for positive culture of pathogenic bacteria, whereas the AREST study classified “any”P. aeruginosacultured as infection. The overall prevalence ofP. aeruginosawas low in this cohort with two children with CF culturing >104CFU·mL−1and only one further child with CF culturing “any”P. aeruginosaat 900 CFU·mL−1. Repeat analysis was performed classifying “any”P. aeruginosaculture as “infection”, but did not significantly affect any outcomes reported. Whilst the ELISA method has good sensitivity and specificity, and is widely accepted as a method to measure lipid mediators, it is acknowledged that the mass spectrometry method yields better specificity.
Conclusions
This study providesin vivoevidence of impoverished 15-LO2 gene expression associated with a depressed LXA4/LTB4ratio in the lower airways of children with CF. We report new mechanistic insights into the failure to resolve inflammation in early CF lung disease, even in the absence of infection.
Footnotes
This article has supplementary material available fromwww.qdcxjkg.com
Support statement: This work was supported by grants from the National Children’s Research Centre, Ireland (E10/08), the Health Research Board of Ireland (1403 HRA_POR/2011/21) and Institut National de la Santé et de la Recherche Médicale, France.
Conflict of interest: Disclosures can be found alongside the online version of this article atwww.qdcxjkg.com
- ReceivedJune 21, 2013.
- AcceptedFebruary 3, 2014.
- ©ERS 2014