© 2002 American Thoracic Society DOI: 10.1165/rcmb.4746 Altered Phospholipid Composition and Aggregate Structure of Lung Surfactant Is Associated with Impaired Lung Function in Young Children with Respiratory InfectionsChild Health, Infection Inflammation and Repair Division, School of Medicine, Southampton General Hospital, Southampton; Department of Respiratory Medicine, Bristol Royal Hospital for Children, Bristol, United Kingdom; and Departments of Paediatric Pulmonology and Neonatology, Hannover Medical School, Hannover, Germany Address correspondence to: A. D. Postle, Ph.D., Child Health, Level G (803), Centre Block, Southampton General Hospital, Tremona Road, Southampton SO16 6YD, UK. E-mail: adp{at}soton.ac.uk
Alterations to pulmonary surfactant structure, composition, and function contribute to the severity of respiratory infections. Analysis of bronchoalveolar lavage fluid (BALF) from children undergoing diagnostic bronchoscopy for structural abnormalities (control group, n = 24), asthma (n = 18), lung infection (n = 30), and cystic fibrosis (CF, n = 15) showed that BALF phospholipid concentration decreased with age for the control group and was elevated in all disease groups. The fractional concentration of the major surface active component, dipalmitoyl phosphatidylcholine (PC16:0/16:0), correlated (r2 = 0.608, P < 0.01) with airway resistance (FEV1% predicted), and decreased PC16:0/16:0 was accompanied by increased concentrations of phospholipid components characteristic of cell membranes (PC16:0/18:1 and PI18:0/20:4). Median minimal surface tension, measured by pulsating bubble surfactometer, was elevated (P < 0.01) in both infection (17.5 mN/m) and CF (17.1 mN/m) compared with the control group (1.5 mN/m). Centrifugation (60,000 x g, 40 min) of BALF indicated that infection was accompanied by accumulation of large aggregate forms of surfactant, in contrast to previous reports of increased conversion to inactive small aggregate surfactant particles in ventilated patients with respiratory failure. This accumulation of surface-inactive, large aggregate forms of surfactant, possibly due to mixing with membrane material from inflammatory cells, may contribute to severity of lung disease in children with respiratory infections.
Abbreviations: acute respiratory distress syndrome, ARDS • bronchoaveolar lavage fluid, BALF • cystic fibrosis, CF • dipalmitoylphosphatidylcholine, DPPC • electrospray ionization mass spectrometry, ESI-MS • large aggregate, LA • phosphatidylcholine, PC • phosphatidylglycerol, PG • phosphatidylinositol, PI • phospholipid, PL • small aggregate, SA •
Although primary surfactant deficiency is the major factor underlying the pathogenesis of neonatal respiratory distress syndrome in preterm infants (1), secondary alterations to surfactant composition and structure may also contribute to the etiology of other lung diseases. For instance, increased concentrations of plasma protein in airway fluid of patients with acute respiratory distress syndrome (ARDS) (2) or asthma (3) can inhibit the surface tension–lowering abilities of lung surfactant. As the high surface pressure generated by the surfactant film in the lungs is a major force keeping the air spaces relatively dry, such inhibition of surfactant may contribute to the severity of airway edema and mucus plugging characteristic of these conditions. Similarly, hydrophilic surfactant proteins A and D (SP-A, SP-D) are integral components of the innate immune defenses of the lungs against microbial lung infection (4), and decreased concentrations of these collectin proteins may make the lungs more susceptible to infection. As antioxidant properties have recently been described for SP-A and, especially, SP-D (5), any decrease in their concentrations may promote oxidative damage of surfactant and cell membrane phospholipid (PL). Surfactant PL itself is essentially immunosuppressive, inhibiting superoxide generation by granulocytes (6), cytokine production by alveolar macrophages (7), and proliferation of peripheral blood mononuclear cells (8). Altered surfactant composition of bronchoalveolar lavage fluid (BALF) or tracheal aspirates has been reported for a number of respiratory infections in children. We have recently shown that disease progression of severe respiratory syncytial virus bronchiolitis in children was accompanied by a progressive deterioration of surfactant phospholipid composition (9), shown by a decreased fractional concentration of dipalmitoyl phosphatidylcholine (PC16:0/16:0), the major surface tension–reducing component of surfactant. Similarly, children requiring mechanical ventilation for respiratory failure, due to pneumonia, viral pneumonitis, and acute RDS, exhibited lower ratios in tracheal aspirate samples of lecithin:sphingomyelin and of SP-A:total protein (10). The decreased concentrations of SP-A (11, 12) and SP-D (12) reported in BALF from children with cystic fibrosis (CF) have been suggested to contribute to the increased susceptibility to the chronic respiratory colonization with opportunistic bacteria that is characteristic of this disease. We recently reported that, compared with children with structural abnormalities to the lungs but no history of lung infection (12), the phospholipid composition of BALF from children with CF and other respiratory infections was not significantly altered. However, potential confounding factors identified included the lower age distribution of the control group, widespread variation in the severity of infection, and relatively small numbers of children in each group. Consequently, we here detail the phospholipid composition, aggregate structure, and surface tension function of BALF surfactant from a larger group of children with a wide range of respiratory conditions, which has enabled the separate effects of age and disease to be distinguished. These analyses have tested the hypothesis that disturbance of phospholipid composition and aggregate structure of lung surfactant are associated with, and may contribute to, the impaired lung function of young children with respiratory infections.
Study Population BALF was collected from children undergoing bronchoscopy for diagnostic purposes. This experimental design, although the only realistic protocol for the ethical collection of such samples, imposed certain restrictions on the study. First, the age and disease distribution of the children involved was entirely dependent on the spectrum of clinical presentation. Second, physiologic measurements such as lung function testing could only be performed when clinically indicated, and consequently such data could not be collected from children in the control group. Twenty-four children with no detectable pulmonary inflammation or infection were categorized as control subjects for the purposes of this study. The majority of these presented with stridor or a persistent cough, but clinical indication for bronchoscopy included investigations following removal of polypoid (1) or carcinoid tumor (1), chronic iron-deficiency anemia (1), suspected hemoptysis (1), and to confirm the resolution of previous chest infections (2). A separate clinical category comprised 18 children with diagnosed asthma. At the time of the bronchoscopy, they had no evidence of lung infection with clear chest radiographs. Bronchoscopies were performed because of poor responsiveness to treatment. All had recurrent wheezing, in the older children substantiated by lung function evidence of variable airflow limitation. All had significant improvements in symptoms and/or function after administration of bronchodilator. Thirty children suffered from a variety of lung infections identified by shadowing on chest radiographs and for all but seven of the subjects, positive growth of organisms from culture of the BALF. The seven without positive cultures were producing purulent secretions in the presence of persistent radiographic changes. A fourth group of 15 children had CF, all diagnosed unequivocally by positive sweat iontophoresis measuring both sodium and chloride levels. Ethical permission for this study was obtained from the Southampton and South West Hampshire Local Research Ethics Committee. Ranges of ages and disease severity within each of the four subject groups are summarized in Table 1. In common with our previous study (12), age distributions within subject groups were all skewed toward lower values, and median values were significantly lower for children in the control group (2.0 yr) compared with those with more established lung disease (asthma, 5.5 yr, P < 0.002; non-CF infection, 4.0 yr, P < 0.02; CF, 4.3 yr, P < 0.02). The range of disease severity within subject groups was assessed from measurements of lung function, microbiologic growth on culture of BALF, and the presence of neutrophils in the BALF cell pellet. Lung function was measured in children above 5 or 6 yr of age as the percentage of predicted FEV1. Distributions of the type and severity of pulmonary bacterial infections were similar in the non-CF infection and CF patient groups. Both the total cell concentrations and the proportions of neutrophils recovered from the BALF were significantly greater for patients with asthma (P < 0.02), non-CF infection (P < 0.001), and CF (P < 0.001) compared with control children.
Sample Processing Fiberoptic bronchoscopy was performed for clinical reasons as previously described (12). Parents gave informed written consent for the procedure, which routinely included BAL and the retention of a BALF aliquot for research purposes. Lungs of infants < 3 yr of age were lavaged with 3 x 10 ml aliquots of sterile saline, whereas children > 3 yr of age received 3 x 20 ml. BALF samples were retained on ice throughout processing, filtered through a 100-µm mesh strainer to remove mucus, and then centrifuged at 400 x g for 10 min to recover cells for differential leukocyte analysis. Cells were resuspended in Hanks' Buffered Saline Solution and total nucleated cells counted in a Neubauer manual hemocytometer (Weber, UK), assessing viability by identifying dead cells with Trypan Blue stain. Cytospin preparations were fixed and stained using a Haema-Gurr staining kit (BDH Laboratories, Poole, Dorset, UK) and the proportions of macrophages, lymphocytes, neutrophils, and eosinophils were calculated from counts of greater than 500 leukocytes. Aliquots of cell-free BALF were retained for determination of total phospholipid and protein concentrations, and the remainder centrifuged at 60,000 x g for 40 min to separate the large aggregate (LA) and small aggregate (SA) surfactant fractions. The LA surfactant pellet was suspended in 100–500 µl 0.9% (wt/vol) NaCl, and both LA and SA fractions were stored at -80°C for compositional analysis of phospholipid and for surface tension measurements.
Extraction and Determination of Total Phospholipid
An aliquot of LA surfactant containing
Electrospray Ionization Mass Spectrometric Analysis of Phospholipid
Protein Analysis
Surface Tension Analysis
Statistical Analyses
Effect of Development on BALF Phospholipid and Protein Concentrations in Children Given the significantly lower ages of the control group of children, we first determined the effect of age on surfactant concentration in this group to assess the additional impact of various lung diseases. Total phospholipid concentration of post cell BALF supernatant from children in the control group was inversely related to age (Figure 1). It was elevated and highly variable for children younger than 2 yr, but was relatively constant for children over 3 yr. The increased phospholipid concentrations in BALF from younger children was not secondary to the smaller instilled lavage volumes used, as there was no relationship between measured phospholipid concentration and instilled saline volume (results not shown). When each subject group was divided into children younger and older than 3 yr (Table 2), BALF phospholipid concentration (median and ranges) in older children was significantly higher in the infection (78, 10–265 µM, P < 0.01) and CF (143, 85–343 µM, P < 0.001) groups compared with the control group (46, 30–78 µM). Furthermore, BALF phospholipid concentrations did not correlate with either the volume of lavage recovered or the efficiency of lavage recovery in any of the subject groups (results not shown). This comparison suggested that the elevated BALF phospholipid concentration in children with CF was not secondary to inefficient recovery of lavage fluid.
By comparison with total phospholipid, total protein concentration in BALF was not related to children's ages in the control group (r2 = 0.1879; P > 0.05, results not shown). The higher total phospholipid concentration in BALF from the younger control children was therefore consistent with an increased surfactant content, rather than just reflecting a general increase in the total amount of material recovered. Compared with control subjects, protein concentrations (Table 2) were significantly higher in BALF from children with non-CF infections (P < 0.01) and in children with CF (P < 0.01).
Surfactant Phospholipid Composition in Disease Proportional concentrations of the major phospholipid classes PC, PG, and PI in the large aggregate surfactant fraction, expressed as a percentage of their sum, were not significantly different for children from the control, asthma, and non-CF infection groups. However, the proportions of PI were significantly higher for children with CF (5.0 ± 2.8 mol %) compared with control subjects (2.7 ± 1.4 mol %, P < 0.002). Although the PC molecular species composition of the large aggregate material was not significantly different between control subjects and subjects with asthma, there were subtle differences in PC composition for children from the infection and CF groups. Dipalmitoylphosphatidylcholine (PC16:0/16:0), the major surfactant phospholipid species thought to be primarily responsible for generating low surface tension values in the lungs, represented 52 mol % of total PC in control subjects, but was decreased in the infection (47.1 mol %) and CF (49.2 mol %) groups (Figure 2). There was a significant association (r2 = 0.608, P < 0.01) between proportional PC16:0/16:0 concentration in the large aggregate surfactant and measurement of percentage predicted FEV1, which were made on some of the older children immediately before bronchoscopy (Figure 3). These results suggested that although PC16:0/16:0 ranged from 32–56 mol % of total PC across the whole study population, very low values (< 40%) were only observed in the children with most severe lung disease. This association was principally dependent on infection status; the correlation between %PC16:0/16:0 and % predicted FEV1 was greater for children in the infection and CF groups (Figure 3, closed symbols; r2 = 0.784, P < 0.01) compared with children with no apparent infection (Figure 3, open symbols; r2 = 0.129, n.s.).
This molecular species analysis showed clearly that infiltration of plasma phospholipid was not a predominant cause of altered PC composition of large aggregate material from infection and CF groups. The dominant plasma PC species is palmitoyllinoleoyl phosphatidylcholine (PC16:0/18:2) (17), and infiltration of plasma PC would be expected to increase the fractional content of this molecular species. This was not the case, and the concentration of PC16:0/18:2 was not raised in either infection group. Instead, there appeared to be an accumulation of membrane phospholipid in the large aggregate surfactant fraction, possibly derived from inflammatory cells such as the neutrophil. Palmitoyloleoyl phosphatidylcholine (PC16:0/18:1) is the major species of neutrophil PC, whereas stearoylarachidonoyl PI (PI18:0/20:4) comprises  70 mol % of neutrophil PI but only 10 mol % of surfactant PI (18). Neutrophil-mediated inflammation was associated with increased concentration in BALF of PC16:0/18:1 in patients with CF (18.6 ± 3.1 mol % total PC) compared with control subjects (15.5 ± 2.1 mol %, P < 0.05) and of PI18:0/20:4 in the infection (15.1 ± 7.7 mol %) and CF (16.1 ± 9.5 mol %) groups compared with the control group (10.0 ± 7.6 mol %, P < 0.05 in both cases).
Surface Tension–Reducing Function of LA Surfactant
Analysis of Aggregate Structure of BALF Surfactant There was a wide variation in the ratio of small to large aggregate fractions (SA:LA) of BALF phospholipid for all patient groups, which was apparently related to the pathology of the underlying disease. This was especially notable for the inverse association between the SA:LA ratio and numbers of neutrophils in BALF (Figure 4). Using the concentration of neutrophils in BALF as an indicator of inflammation associated with lung infection, this result suggests that, in contrast to an increased ratio in ventilated patients with ARDS (19), the SA:LA ratio did not increase with severity of inflammatory lung disease in these children. Irrespective of diagnosis, the SA:LA ratio was not increased above the range of the control group for any child who exhibited elevated numbers of neutrophils in their BALF (Figure 4). When stratified by diagnosis, the SA:LA ratio was unchanged for children with asthma and non-CF infection compared with the control group, but was significantly lower for children with CF (P < 0.01) (Figure 5). This conclusion was supported by the comparisons between SA:LA ratio and both  min of the LA fraction (Figure 6) % predicted FEV1 (Figure 7). Both comparisons clearly show that the SA:LA ratio, contrary to expectations, did not increase either with impaired surface tension–reducing properties or with decreased airway function. Together, these observations imply an accumulation in the lungs of children with lung infection of LA surfactant, possibly complexed with inflammatory cell phospholipid, with inferior surface tension–reducing properties.
This study demonstrates clearly the extent and nature of the alterations to lung surfactant composition, structure, and function in children with a variety of respiratory diseases. Importantly, all these BALF samples were obtained at diagnostic bronchoscopy, and none of these children required ventilatory support. Consequently, these children exhibited a heterogeneous pattern of disease that was generally less severe than in previous reports of surfactant changes in ventilated children (10). This heterogeneity enabled the changes to surfactant composition to be correlated with clinical parameters of lung function, and the detailed analysis of the changes provided evidence for the mechanisms involved in modifying BALF phospholipid composition. The observation that total phospholipid concentration was much higher in younger children and decreased progressively with age (Figure 1) agrees with previous reports (20). Studying a smaller population of children, we previously highlighted that the median age of children was lower in the control compared with the disease groups, and suggested that this discrepancy might have masked any significant difference to BALF phospholipid concentration between the various groups (12). The results presented in Table 2 demonstrate that phospholipid concentration in BALF from children over 3 yr of age was indeed significantly elevated for all disease groups. Previous analysis of phospholipids in BALF from patients with respiratory infection have generally reported decreased contents of total PC and PG, and of the fractional content of disaturated PC, with correspondingly increased contents of sphingomyelin, PS, PE, and PI (10, 11). Such changes are typical of the accumulation of cell membrane–derived phospholipids within the surfactant phospholipid fraction. However, these analyses were performed on patients with severe infection, often requiring ventilatory support, and we previously found only relatively minor changes to the molecular species compositions of surfactant phospholipids in BALF obtained from children with a variety of respiratory conditions who underwent diagnostic bronchoscopy (12). The detailed analyses presented here demonstrate alterations of BALF phospholipid compositions in those children over 3 yr of age with more severe lung disease. In these children, the decreased content of PC16:0/16:0 associated with lung infection (Figure 2) was accompanied by increased fractional concentrations of PC16:0/18:1 and PI18:0/20:4, which are both major components of cell membranes. Consequently, the altered phospholipid composition of BALF LA material in lung infection is most likely due to mixing with membrane fragments from dead and dying inflammatory cells, principally granulocytes (18). Contamination of surfactant with membrane fragments from resident alveolar macrophages would not be apparent by this technique, as the phospholipid composition of such cells is dominated by ingested surfactant material (unpublished observation). We could not determine in this study the nature of the complexes formed between this membrane-derived phospholipid and native surfactant, as this would have required further purification of surfactant by density gradient centrifugation. Although there was too little material in the small volume aliquots of BALF available for analysis from these routine bronchoscopies to undertake this procedure, ESI-MS phospholipid analysis demonstrated that SA and LA fraction of BALF had essentially identical compositions (results not shown). This result strongly implies that the altered phospholipid composition of the LA fraction from BALF was not simply due to co-precipitation of surfactant with membrane fractions. In that case, the phospholipid composition of the SA fraction remaining after removal of membrane material would not have altered with disease status. The increased BALF phospholipid concentration in younger children is unlikely to have been due to preferential sampling of alveolar rather than airway material. Protein concentration in BALF was equivalent in younger and older children in the control group (Table 2), and comparable postnatal developmental patterns of surfactant phospholipid concentration have been shown in material obtained by exhaustive lung lavage of experimental animals (21). It is possible, however, that bronchoscopy of damaged lungs resulted in preferential sampling of airway material from children with more severe lung disease. One alternative potential mechanism for altering surfactant phospholipid compositions is infiltration of plasma lipoprotein-derived phospholipid into the airways. We have recently demonstrated that such a mechanism contributes to the severity of the asthmatic response, and that the extent of such infiltration can be monitored by any increased concentration in BALF of the principle component of plasma phospholipid, PC16:0/18:2 (17). The lack of any increased concentration of PC16:0/18:2 in BALF from either infection group indicates that for these children the epithelial:endothelial barrier was relatively intact and retained its ability to prevent infiltration of lipoprotein material into the airways. This conclusion is supported by the relatively modest alteration to BALF protein concentration in the various disease groups (Table 2) in contrast to the much larger protein infiltration characteristic of the asthmatic response (3).
Measurement of surface tension in the large aggregate pellet demonstrated the expected increased Initially, this conclusion appeared to be at variance with a wealth of evidence demonstrating that surfactant dysfunction in conditions such as ARDS is characterized by conversion of large, surface-active aggregate forms into small, surface-inactive aggregates (19, 23). However, these studies were conducted on severely ill patients supported by positive pressure ventilation. By contrast, sepsis in spontaneously breathing rats is characterized by accumulation of inactive large aggregate surfactant (24) that was essentially comparable to our findings in children with more severe lung disease. A similar increased fractional concentration of LA surfactant coupled with impaired activity was observed in rats after short-term exposure to ozone (25). In ARDS, the conversion of surfactant into SA fractions may be potentiated by mechanical distension of the alveolus secondary to mechanical ventilation (26). In contrast, lung infection is characterized by focal areas of infection, consolidation, and minimal ventilation, together with regions of relatively normal function. Under such conditions, with overall decreased tidal volumes due to the extent of lung disease, surface area changes in the lungs would be considerably reduced, and may in turn explain the absence of conversion of large to small aggregates of surfactant in spontaneously breathing children with lung infection. Conversely, accelerated conversion from LA to SA forms of surfactant may be one mechanism that contributes to the extent of the reported lung dysfunction in patients ventilated with high tidal volumes (27). One prediction from this observation is that formation of SA forms of surfactant should be decreased in patients with ARDS maintained by high frequency oscillatory ventilation. It is important to remember, however, that the LA surfactant that accumulated in children with more severe lung disease was itself essentially inactive (Figure 6). These conclusions highlight the complexity of the changes to lung surfactant composition and function in children with respiratory infection, and emphasize the potential problems with any future attempt to administer therapeutic surfactant preparations to lungs which already contain elevated concentrations of inactive surfactant.
This work was supported by a British Lung Foundation/John Ellerman Fellowship in Paediatric Lung Disease (A.M.) and by Cystic Fibrosis Trust project grant PJ412. The authors are grateful to the Wellcome Trust for an equipment grant to purchase the mass spectrometer. The authors are grateful to Christa Acevedo for assistance with the pulsating bubble surfactometer analyses.
The nomenclature for phospholipid structures denotes first the class of phospholipid – phosphatidylcholine (PC) or phosphatidylinositol(PI) – followed by the fatty acid substitutes at the sn-1 and sn-2 positions of the glycerol backbone. Each fatty acid is described by numbers of carbon atoms and unsaturated double bonds. For instance, PC16:0/16:0 is dipalmitoyl PC, PC16:0/18:1 is sn-1-palmitoyl sn-2-oleoyl PC and PI18:0/20:4 is sn-1-stearoyl sn-2-arachidonoyl PI. Received in original form October 15, 2001 Received in final form June 20, 2002
This article has been cited by other articles:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Top
Abstract
Introduction
P < 0.05.
P < 0.01.
