INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Lung disease is cystic fibrosis (CF) is characterized by chronic endobronchial infection and inflammation leading to progressive damage to the pulmonary airways. It remains the major cause of morbidity and mortality in patients with CF.

CF is caused by mutations in the cystic fibrosis transmembrane conductance regulator (CFTR) gene (1). CFTR is a cyclic AMP regulated chloride channel found in the apical membrane of cells from a number of different epithelia. In addition to transporting chloride itself, CFTR also appears to be involved in regulating other ion channels including the outward rectifying chloride channel (ORCC) (2) and the epithelial sodium channel (ENaC) (3). Disruption of CFTR results in decreased epithelial chloride permeability and loss of cAMP mediated chloride secretion. In the CF airway epithelium, there is an additional abnormality of sodium transport as a result of which sodium is absorbed at a higher rate than in normal airway epithelium.

The mechanisms by which these known abnormalities of ion transport result in airway inflammation and infection are unclear. A number of different models have been proposed. One suggests that altered ion transport in CF airway results in a reduced depth of airway surface liquid (ASL), which in turn impedes efficient cilial action, leading to mucus retention and subsequent infection. Alternatively, if accelerated salt absorption is not accompanied by water, the depth of the ASL may remain unchanged, but the concentration of sodium and chloride would be lowered, the consequence of which may be reduced neutrophil killing of inhaled bacteria (4). More recently a third model has been proposed in which it is suggested that the airway epithelium, like that of the sweat duct, has a reduced ability to reabsorb salt, resulting in increased concentrations of sodium and chloride which then inhibit the activity of defensin-like molecules at the airway surface (5).

New approaches to the treatment of CF, which attempt to normalize the composition and volume of ASL, will clearly depend on which of these models is correct. In this study, we have analyzed the electrolyte composition of ASL from infants and young children with CF and compared them with that of non-CF controls. For comparison with other published studies, we have also measured ASL from the nasal epithelium in older CF patients and control subjects.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects and Controls

The state of Victoria, Australia (66,000 births/yr) uses a two-tiered newborn screening program for CF based on an estimation of the immunoreactive trypsin from a blood spot Guthrie card, followed by genetic analysis for the F508, G542X, G551D, and R553X mutations. A sweat chloride concentration of greater than 60 Meq/l confirmed the diagnosis in heterozygotes. All patients are admitted to the Royal Children's Hospital, Melbourne, for parental education and baseline assessment. Subsequent management is by the Hospital's CF clinic.

All infants identified by the screening program between November 1995 and November 1996 were eligible for the study (n = 13). Of these 13, 10 families agreed to take part in the study. Samples were also collected from infants and young children born prior to this period undergoing bronchoscopy as part of an on-going infection surveillance study (n = 6) (6) or during admissions for pulmonary exacerbations (n = 3). Using neutrophil percentage and IL-8 level in the BAL fluid markers of pulmonary inflammation, the subjects with CF were divided into three groups. The first group (CF-NI, n = 5) had little evidence of pulmonary inflammation (neutrophil counts less than 15% and IL-8 levels less than 100 pg/ml BAL fluid), the second had obvious pulmonary inflammation (CF-I, n = 7) (neutrophil counts greater than 15% and IL-8 counts over 100 pg/ml), and the third (CF-MI, n = 7) had either a high neutrophil count or a high IL-8 but not both. Seven infants undergoing bronchoscopy for evaluation of stridor formed a disease control group. Plasma concentrations of sodium and chloride were measured in all subjects. Clinical details of the subjects are shown in Table 1.

Nasal samples were collected from 10 subjects with CF (mean age 17 yr) and from 10 control subjects (mean age 29 yr). None of these subjects was using nasal medication and none had current rhinitis or coryzal symptoms.

Bronchoscopy, Collection of Tracheal Samples, and Bronchoalveolar Lavage

Bronchoscopy was performed under halothane general anaesthesia; 0.5 ml of lignocaine was applied above the vocal cords to prevent laryngeal spasm. No lignocaine was used below the vocal cords. In four subjects, lignocaine was omitted and muscle paralysis used. There was no difference in the results using these different methods. All subjects were given 0.01 mg/kg atropine intravenously. A flexible bronchoscope (Olympus model BF3 C20; external diameter, 3.6 mm; suction channel, 1.2 mm; Olympus Optical Co. Ltd., Tokyo, Japan) was introduced into the lower airway through a laryngeal mask, avoiding the use of the suction channel until after the ASL sample had been collected. Once the tip of the bronchoscope was sited just above the carina, a fine polythene protected catheter (Portex Ltd, Kent, UK) containing a thin strip (0.5 mm by 15 mm) of nitrocellulose membrane (Boehringer Mannheim, Mannheim, Germany) was passed through the suction channel. The membrane was then extended out of the catheter and under direct vision touched gently against the tracheal epithelium for 10 s. It was then withdrawn back into the catheter, removed from the bronchoscope and transferred to a microcentrifuge tube on ice, clipping the end of the strip into the top of the tube. Two ASL samples were collected from each patient. The microcentrifuge tubes containing the membrane strips were then spun at 8,000 g for 1 min, and the paper removed. Using a high precision calibrated syringe (SGE International Pty. Ltd., Victoria, Australia) a constant volume aliquot (0.2 µl) of the resulting fluid was then transferred to constant volume (4 µl) of a solution of 25% glycerol containing 30 mM lithium iodide. To control for evaporative losses, all samples were processed in an identical fashion and transferred to the glycerol exactly 15 min after collection. The sample was then stored at 70° C until analysis.

After collection of the ASL sample was complete, the tip of the bronchoscope was wedged in the right middle lobe bronchus and to optimize sampling from endobronchial sites a single, small volume lavage was performed by instilling 1.5 ml/kg of sterile nonbacteriostatic normal saline at room temperature through the suction channel of the bronchoscope over 3-5 s. Using negative suction pressures of 100-150 mm Hg, the saline was immediately aspirated into a suction set over 10-20 s. The bronchoscope was then wedged into the lingula bronchus and using an identical technique, a further single aliquot lavage was performed. The BAL fluid from both lavages was pooled for analysis.

Nasal Samples

Subjects wore a nasal clip for 5 min prior to sampling. Once the clip was released the ASL sample was collected by touching the nitrocellulose membrane strips against the inferior turbinate under direct vision. The membrane was transferred to a microcentrifuge tube on ice and handled and processed as described above for the tracheal samples.

ASL Sample Analysis

The samples were analyzed by Energy Dispersive X-ray Analysis (EDX) using a method adapted from that described by Quinton (7, 8). X-ray analysis was carried out in a Camscan field emission scanning electron microscope, fitted with a Kevex energy dispersive X-ray detector. The electron beam energy was set at 20 keV with a beam current of about 0.5 nA. These instrument settings were maintained for all analyses. The samples were prepared for analysis by contacting drops of sample fluid, expressed from a 1 µl syringe onto double sided carbon tape attached to a microscope stub. The samples were then dried down in high vacuum and carbon coated to prevent charging during analysis. Exposure to vacuum resulted in the formation of circular "cakes" of crystallized material approximately 1-2 µm thick and about 0.5 mm in diameter. The microscope raster was set to be just within the dimensions of each sample, minimizing background signal from the carbon substrate. Five sub-samples were prepared and analyzed in this manner for each ASL sample collected. Also analyzed were the several sets of standards prepared from known sodium chloride solution concentrations.

During analysis, each sample was irradiated with the electron beam and X-ray spectra acquired. Curve fitting of each spectra, including background subtraction, was carried out off-line using Gaussian peak profiles. The intensities of the X-ray lines for the Na K, Cl K were measured and ratioed to the I L intensity for all five sub-samples, and the ratios averaged. Cl/I intensity ratios did not vary by more than 6% about the mean, while Na/I intensity ratios were within 11% of the mean for each ASL sample analyzed. Final concentrations were derived from the calibration curve. When known standards were handled exactly as the patient samples, by absorbing them into the nitrocellulose membrane, leaving them on ice for 15 min and then recovering the sample by centrifugation, the effect was an apparent increase in concentration of 8% to 12%.

BAL Fluid Analysis

BAL fluid from all subjects was analyzed for the presence of infection by quantitative bacterial culture and routine viral culture, and for inflammation by cytology and interleukin-8 (IL-8) analysis as described previously (6). IL-8 concentrations are expressed per milliliter of BAL fluid.

Statistics

The distribution of age, neutrophil count, BAL volume return, and IL-8 concentrations were non-normal, and these parameters are described by medians (interquartile range, IQR). Between group comparisons for these parameters were made using the Mann-Whitney U statistic. The ASL sodium and chloride showed normal distribution within each group and are described using means (SE). Comparisons between groups for these parameters were made using the independent samples t test.

This study was approved by the Human Ethics Committee of the Royal Children's Hospital and written informed consent was obtained from the parents of each child before bronchoscopy.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Tracheal Samples

The BAL data and the ASL sodium and chloride concentrations for the four study groups are shown in Table 2. As expected the BAL fluid neutrophil percentage and IL-8 level were low in the control group. In the CF group without pulmonary inflammation (CF-NI) these parameters were very similar to the control group, although there was a small increase in the IL-8 concentration (30 pg/ml compared with 15 pg/ml in the control group, p = 0.03). In the CF group with mild inflammation (CF-MI) the neutrophil percentage was again similar to controls, and the IL-8 level was significantly higher (226 pg/ml in CF-MI group, p < 0.01). Finally in the CF group with severe inflammation there were marked and highly significant increases in both percentage neutrophils (60% compared with 7% in controls, p < 0.01) and IL-8 levels (1,957 pg/ml in CF-I group, p < 0.01). The number of subjects with positive BAL cultures (> 10⁴ CFUs) correlated with the degree of inflammation. In the control and CF-NI groups, none of the subjects had positive BAL cultures. In the CF-MI group there was one positive culture (Moraxella catarrhalis) and in the CF-I group there were three positive cultures (one Haemophilus influenzae, one Stenotrophomonas maltophilia, and one Staphylococcus aureus).

In the control group, the ASL sodium concentration was significantly lower than the value found in plasma (85 mM in ASL; 137 mM in plasma). The ASL chloride concentration was similar to that found in plasma (108 mM in ASL, 105 mM in plasma). In the CF-NI group, sodium concentration was similar to that in the control group (78 mM compared with 85 mM). In contrast, the ASL chloride concentration was significantly lower (77 mM compared with 108 mM, p < 0.01). In the CF-MI group, the ASL sodium concentration was again similar to the other 2 groups (83 mM) whereas the ASL chloride concentration was an intermediate value (96 mM), which was not significantly different to that of the control group (p = 0.25). Similar results were found in the CF group with the most pulmonary inflammation (CF-I) (ASL sodium concentration 84 mM, ASL chloride concentration 95 mM; these values were not significantly different from those in the control group). Individual data points for the tracheal chloride ASL concentrations are shown in Figure 1. When the CF subjects were considered as a single group, there were no significant differences, compared with the control group, in either the ASL sodium concentration (82 ± 6 mM in the CF group compared with 85 ± 10 mM for controls) or in the ASL chloride concentration (91 ± 5 mM in the CF group compared with 108 ± 5 mM for controls, p = 0.10). There was no effect of genotype on ASL sodium or chloride (data not shown).

View larger version (10K): [in this window] [in a new window]   Figure 1.   Scatter plot of individual tracheal airway surface liquid (ASL) chloride concentrations. CF-NI = CF subjects with no pulmonary inflammation; CF-MI = CF subjects with mild pulmonary inflammation; CF-I = CF subject with marked pulmonary inflammation.

Nasal Samples

There were no significant differences in the nasal ASL sodium or chloride concentration in the CF subjects compared with those found in the controls (CF nasal ASL sodium 116 ± 7 mM versus 106 ± 4 mM for control subjects, p = 0.20; CF nasal chloride 125 ± 4 mM versus 115 ± 4 mM for control subjects, p = 0.08). The individual data points are shown in Figure 2.

View larger version (16K): [in this window] [in a new window]   View larger version (16K): [in this window] [in a new window]   Figure 2.   Scatter plot of individual nasal airway surface liquid (ASL) sodium (A) and chloride (B) concentrations.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

One of the major goals of CF research has been to identify the mechanisms whereby defects in the function of CFTR result in the chronic debilitating lung disease seen in patients with CF. Pulmonary inflammation occurs very early in many CF infants (6, 9), and in older children and adults it is nearly universal. One of the difficulties in research into the pathogenesis of CF lung disease is establishing whether the abnormalities, which are found in the lungs of CF patients are a direct consequence of the CF gene defect or whether they are secondary phenomena.

It is well established that there are defects in ion transport across CF airway epithelium, both in cell culture systems and in vivo (10). These defects of reduced chloride permeability and increased sodium absorption have been explained by a direct effect of the loss of chloride transport through CFTR in the apical membrane and by the influence of CFTR on the ORCC and ENaC. Since water generally follows ion transport across most epithelia, it seems likely that increased sodium absorption and decreased chloride secretion would be associated with either a reduced volume of ASL or a decreased sodium or chloride content. Attempts to measure ASL volume in CF and normal cells in vitro have produced contrasting results (11, 12).

In this study we have measured the sodium and chloride content of ASL from the trachea of normal infants and from those with CF. The neonatal screening program allowed us to identify very young asymptomatic infants with CF, some of whom had little if any evidence of pulmonary inflammation at the time of bronchoscopy. The composition of ASL in these infants is therefore most likely to reflect a direct effect of the gene defect. We have also studied older infants some with clinical exacerbations and some with asymptomatic pulmonary inflammation to separately examine the effect of pulmonary inflammation on ASL composition.

The results of this study show that in infants with CF without pulmonary inflammation tracheal ASL sodium concentration was unchanged, whereas the ASL chloride concentration was significantly lower, when compared with values from non-CF control subjects. This finding is consistent with impaired chloride secretion from the CF airway epithelium. In CF subjects with pulmonary inflammation, the ASL sodium was also unchanged, but the ASL chloride concentration tended to be higher compared with the noninflamed CF group. The increase in ASL chloride may reflect an effect of airway inflammation on the integrity of the airway epithelium, possibly resulting in increased permeability to chloride.

Comparison of our results with other studies measuring tracheal ASL (13, 14) from non-CF subjects shows close agreement for ASL sodium concentration (85 mM in this study compared with 68 mM Wager and coworkers and 82 mM Joris and associates). Like Wager and colleagues, but unlike Joris and associates, we found that ASL chloride was higher than ASL sodium (ASL chloride 127% higher than sodium, compared with 122% Wager and coworkers and 102% Joris and colleagues).

There is only one other published study of tracheal ASL electrolyte composition in CF subjects (14). These workers measured ASL electrolytes in three CF subjects; two adults and one infant aged 6 mo. The results showed an increased concentration of both sodium and chloride (each increased by approximately 150%) compared with normal adult control subjects. The chloride in the CF group was higher than the sodium, and only the increase in the chloride was statistically significant. Although the authors state the infant with CF was not colonized with S. aureus or P. aeruginosa, the ASL sample was collected at bronchoscopy performed because of persistent chest X-ray abnormalities. Thus, it seems likely that this infant, like the other two subjects in the study, had active pulmonary inflammation at the time of sample collection. Our finding of an increase in the ASL chloride in the CF subjects with pulmonary inflammation suggests that the increases in both sodium and chloride concentration found by Joris and coworkers (14) may represent a secondary effect of longer standing and more severe inflammation in their subjects. This suggestion is supported by other results from the study by Joris and coworkers. Patients with airway irritation and acute infection both had increases in ASL sodium and chloride of a similar order of magnitude to that found in the subjects with CF.

Measurement made on nasal ASL taken from young adults with CF showed that the sodium and chloride concentrations were not significantly different to those found in the control subjects. There was a trend for an increase in the chloride concentrations in the CF group (CF nasal chloride 125 ± 4 mM versus 115 ± 4 mM for control subjects, p = 0.08). In a recent study (5), Smith and coworkers measured nasal ASL chloride concentrations in 25 subjects (eight CF and 17 control subjects) and found significantly higher chloride concentrations in subjects with CF when compared with controls. The reason for this apparent qualitative difference of the effect of CF on nasal and tracheal ASL composition, such that there is a decrease in chloride in CF tracheal ASL but either no change or an increase in chloride in CF nasal ASL is not known. It is possible that the atropine used in the collection of the tracheal samples but not the nasal samples is responsible, although this would require the atropine to act differently on normal compared with CF airway epithelium. Alternatively, the observed differences may reflect regional differences in airway epithelial ion transport or possibly an effect of age (the subjects in the nasal studies were young adults), although this seems less likely.

We conclude from our study that the sodium and chloride concentrations in tracheal ASL are not increased in infants with CF who do not have pulmonary inflammation but rather that ASL chloride is decreased. The presence of pulmonary inflammation affects the electrolyte composition of ASL. Our finding of differences in the effect of CF on nasal and tracheal ASL composition suggests that there may be regional differences in the effect of CF on ASL composition within the lung. Thus, the ASL composition in CF small airways, which are the principle site of CF lung disease, may differ from that reported here. Nevertheless our results do not support the hypothesis that increased ASL salt concentrations are important in the initial pathogenesis of lung disease in CF.