INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Pulmonary surfactant, which covers the large alveolar surface in all mammalian species investigated, is mainly composed of phospholipids (PL) with a predominance of phosphatidylcholine (PC), neutral lipids, and surfactant-specific apoproteins (SP-A, SP-B, SP-C, SP-D) (1). By a far-reaching reduction in alveolar surface tension, pulmonary surfactant stabilizes the alveoli and prevents them from collapse. Alterations of the pulmonary surfactant system have long been implicated in the course of acute respiratory distress syndrome (ARDS), or, more generally, in any lung disease with pronounced alveolar inflammation. Indeed, in clinical studies in ARDS (2) and, more recently, in severe pneumonia (PNEU) (5), a marked impairment of the biophysical surfactant function has been documented. Concerning the composition of surfactant under conditions of ARDS, to date most attention has been paid to the analysis of the PL profile and the apoprotein content. Bronchoalveolar lavage fluid (BALF) concentrations of SP-A and large surfactant aggregate (LSA) concentrations of SP-B were found to be decreased (4), and a marked reduction of PC and phosphatidylglycerol (PG) has been noted throughout.

In marked contrast to PC from other biologic sources, e.g., membranes, surfactant PC is characterized by a high abundance of palmitic acid residues, and this is known to be an essential feature for the high compression of PL during expiration and lowering of the surface tension to values near 0 mN/m. In contrast, the minor PL classes, in particular PG, are assumed to be functionally important because of their high relative contents of unsaturated fatty acids (FA) and their favorable adsorption characteristics (7).

Considering this background, detailed insight into the changes of the FA profile that occur in acute respiratory failure would be highly desirable. Few groups have addressed this issue so far (2, 4, 8), mainly measuring the relative content of disaturated phosphatidylcholine (DSPC) species, with heterogenous results. It is thus largely unclear whether substantial alterations of the FA profile of the different PL classes exist in acute inflammatory lung disease, and whether such alterations may contribute to the decreased surface activity observed in ARDS or PNEU. The current study was performed to assess abnormalities in the FA profiles of all relevant surfactant PL classes isolated from BALF of 39 mechanically ventilated patients with respiratory failure caused by ARDS, PNEU, or the combination of both (ARDS + PNEU).

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Patient Population

This study was conducted at the intensive care unit of the Department of Internal Medicine of the Justus-Liebig University in Giessen, Germany. All 39 patients required mechanical ventilation. The study protocol was approved by the local ethics committee and informed consent was obtained from either the patient or closest relatives.

The patients were classified into the following patient groups:

Acute respiratory distress syndrome (ARDS). This group consisted of eight patients with either severe ARDS resulting from nonpulmonary underlying events (sepsis, n = 5; pancreatitis, n = 1; other reasons, n = 2). Diagnostic criteria were the same as recently proposed by the Consensus Conference on ARDS (9) and included, next to an underlying catastrophic event, bilateral alveolar infiltration, ratios of arterial oxygen tension to fraction of inspired oxygen (Pa_O2/FI_O2) < 200, and the absence of left ventricular failure as documented by pulmonary capillary wedge pressures (Ppcw) within the normal range (Ppcw < 16 mm Hg).

ARDS associated with pneumonia (ARDS + PNEU). This group included nine patients, of whom some were initially diagnosed with PNEU and who showed rapid, diffuse bilateral spreading of the formerly circumscribed infiltrates in sequential chest radiographs, finally developing an ARDS typical pattern (PNEU  ARDS). The other patients were initially classified as having ARDS but acquired microbiologically proven, secondary PNEU within < 72 h (ARDS  PNEU).

Severe pneumonia (PNEU). This group consisted of 22 patients with a clinical history of primary lung infection in the absence of acute or chronic left heart failure (Ppcw < 16 mm Hg). Diagnostic criteria were fever, tachycardia, dyspnea, typical auscultatory findings, characteristic chest radiographs, and microbiologic identification of pathogens in lower respiratory tract specimen (BALF or tracheal suction samples).

Eight nonsmoking healthy volunteers without a history of cardiac or pulmonary disease served as the control group. For demographic data see Table 1.

All bronchoalveolar lavages (BAL) were performed primarily for diagnostic reasons, therefore a strictly time-matched protocol was not applied. Nevertheless, most patients received the BAL within 4 d of onset of mechanical ventilation, and the latest time point of lavage in the current study was 5 d postintubation. Sequential analyses were not performed, and an overlap between the categories did not occur.

BAL Technique

Flexible fiberoptic bronchoscopy was performed in a standardized manner as previously described (5). One segment of the lingula or the right middle lobe was lavaged with a total volume of 200 ml of sterile saline in 10 aliquots, with a fluid recovery of 50 to 70%. The fractions were pooled, filtered through sterile gauze, and centrifuged at 200 × g (10 min, 4° C) to remove cells and membranous debris. The supernatant was immediately divided into aliquots, frozen in liquid nitrogen, and stored at 80° C until further processing. Staining and counting of the pelleted cells were performed according to routine standards.

PL Content and PL Profile

Lipids were extracted from BALF with chloroform/methanol (10), and the PL content was determined by spectrophotometric measurement of phosphorus according to the method of Rouser and coworkers (11). Individual PL classes were separated by high-performance thin-layer chromatography as previously described (5) using silica 60 plates (Merck, Darmstadt, Germany) and chloroform/methanol/acetic acid/water (50/ 37.5/3.5/2, vol/vol) as the developing solvent. This one-dimensional system allows the separation of eight phospholipid classes: lyso-phosphatidylcholine (LPC), sphingomyelin (SPH), PC, phosphatidylserine (PS), phosphatidylethanolamine (PE), PG, phosphatidylinositol (PI), and cardiolipin (CL). Samples (25 µg total PL) and appropriate standards (7 lanes; Sigma, Deisenhofen, Germany) were applied to the plate using a Linomat IV applicator (Camag, Muttenz, Switzerland) and stained with molybdenum blue reagent according to the method of Gustavsson (12). Quantification was accomplished by means of densitometric scanning at 700 nm using a thin-layer chromatography (TLC) scanner II (Camag).

FA Composition

Lipids were extracted from LSA or BALF (data not given in detail) and separated as described previously. After nondestructive visualization of individual PL with primuline (13), the corresponding gel compartments, as well as blanks were eluted after addition of 10 µg pentadecanoic acid as an internal standard with 10 ml of chloroform/methanol (vol/vol). The ester-bound FA were converted to fatty acid methyl ester (FAME) using acid-catalyzed transmethylation with 2 N HCl in methanol. For this purpose, samples (10 to 100 µg PL) were treated with 1 ml of reagent for 12 h at 100° C. Resulting FAME were extracted with 1 ml hexane, dried under a stream of nitrogen, and dissolved in 10 µl of chloroform before further analysis. Gas chromatographic (Carlo Erba Fractovap 2150, Mainz, Germany) separation was performed using a very polar fused silica capillary column (CP-Sil 88, 50 m × 0.25 mm; Chrompack, Frankfurt, Germany) at 198° C, with the flame ionization detector (FID) at 250° C and helium as carrier gas (flow rate 1 ml/min). The injector was set at 250° C and used in the split mode (15:1). The individual FAME were identified by comparison with the retention times of commercial standards (Sigma, Deisenhofen, Germany). The resulting peak areas were corrected against blanks by means of the internal standard and treated with empirically determined mass/response factors.

Isolation of LSA and Surface Tension Measurements

Aliquots of BALF were centrifuged at 48,000 × g (1 h, 4° C) to separate the large from the small surfactant aggregates (4, 14). The pellets were resuspended in a small volume of 0.9% NaCl/3 mM CaCl₂ and assessed for PL content. The pellets were then adjusted to a concentration of 2 mg/ml PL, vortexed for 1 min, and used for FA analysis of LSA or for surface tension measurement, which was assessed by means of a pulsating bubble surfactometer (Electronetics, New York, NY) as previously described (15, 16). The surface tension after 5 min of film oscillation at minimal bubble radius ( min) is given.

Reconstitution Experiments

In nine patients with ARDS or pneumonia, or both, the surface activity of the LSA fraction was additionally characterized upon secondary supplementation of dipalmitoylated phosphatidylcholine (DPPC; Sigman, Deisenhofen, Germany). For this purpose, DPPC was dissolved in glass tubes with chloroform/methanol 2/1 (vol/vol) and dried under a stream of nitrogen. A homogeneous solution of DPPC in distilled water was prepared upon extensive sonication and agitation. Appropriate amounts of this solution were lyophilized in polypropylene tubes and recombined with LSA preparations of single patients to achieve a relative amount of palmitic acid in PC of 80% (thus representing control values). After sonication of untreated and DPPC-enriched LSA, both samples were incubated at 43° C for 1 h with occasional shaking, incubated for an additional 30 min at 37° C, and surface activity was determined as previously described.

Statistics

All results are expressed as mean and standard error with additional calculation of median values where appropriate. Statistical analysis of the differences between the patient categories and the control subjects was performed by testing principal significance diversity first (Kruskal-Wallis H test), followed by a comparison using a nonparametric test (Mann-Whitney U test). Comparison between untreated LSA and DPPC-reconstituted LSA was performed using Wilcoxon's matched-pairs signed ranks test. Values significantly different from control values are indicated with symbols. A Pearson test was performed to demonstrate the correspondence between biophysical and biochemical parameters.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

As could be anticipated from previous studies, patients with ARDS or pneumonia presented markedly elevated neutrophil counts in the BALF (Table 1). Gas exchange was severely impaired throughout all patient groups, with patients with ARDS displaying the lowest Pa_O2/FI_O2 ratios. There was no major difference between the groups in regard to sex, age, or smoking behavior. Control subjects were slightly younger than the patients.

Analysis of PL Classes

A moderate but significant reduction of the relative content of PC was observed in all groups of patients (Table 2). Similarly, the relative amount of PG was markedly reduced in all patients compared with the control group. Accordingly, the relative content of the other minor PL classes was increased, in particular PI and SPH. The changes in the PL profile were most pronounced in the ARDS + PNEU group.

Alterations of FA Profile

In accordance with previous studies, PC obtained from LSA from healthy subjects was highly enriched with palmitic acid, which accounted for approximately 80% of all FA (Figure 1, Table 3). An additional 8% of the FA residues was found to be saturated (14:0, 18:0), thus yielding a relative amount of saturated FA in PC of nearly 90%. Among the unsaturated species, oleic acid was the most important and accounted for approximately 7.5%. All other FA appeared with minor abundance. In contrast, a marked alteration in the PC FA profile was encountered in all patient categories. The relative content of palmitic acid was reduced to 63 to 65%, with the other saturated FA remaining unchanged. The relative percentage of all unsaturated species was markedly increased. This was particularly obvious for oleic acid and linoleic acid but also encountered for palmitoleic acid and arachidonic acid (Table 3).

View larger version (10K): [in this window] [in a new window]   Figure 1.   Relative content of palmitic acid (16:0) in PC isolated from LSA given as percentage of all FA. All single values are given (circles); the mean (triangles) and the median (squares) are depicted. ** p < 0.01; *** p < 0.001 compared with control subjects.

In both healthy subjects and patients, the FA profile of PC isolated from native BALF did not differ markedly from PC isolated from LSA; however, the total amount of saturated species was approximately 5 to 10% lower compared with LSA PC and, more pronounced, the relative concentration of palmitic acid was approximately 6 to 15% lower in BALF PC compared with LSA PC (data not given in detail).

In control subjects, PG was found to be almost equally composed of saturated and unsaturated FA. In contrast to PC, the main FA was oleic acid (~ 42%, Table 4), which accounted for the majority of all unsaturated species (~ 90%). The relative abundance of palmitic acid (~ 26%) only slightly surpassed that of stearic acid (~ 21%). In all patient categories, the relative content of oleic acid and stearic acid decreased markedly, whereas palmitic acid increased. However, the total ratio of unsaturated versus saturated species did not change substantially.

When comparing PI with PG from healthy control subjects, a strikingly similar FA pattern was observed (Table 5). However, changes in the patients contrasted to those in PG, with a reduction of palmitic acid in favor of linoleic acid and arachidonic acid. As a result, the overall content of saturated species was reduced by approximately 10% in the patient groups as compared with control subjects.

On the basis of the PL pattern, it could be assumed that a characterization of the FA profile of PE, PS, and SPH in the control subjects could hardly be performed owing to a low abundance of these PLs. Throughout the control subject and the patient categories, PE was predominantly composed of oleic acid and palmitic acid and a remarkably high content (approximately 10%) of arachidonic acid (Table 6). The FA profiles of PE from the patient categories did not differ significantly from control subjects. In the patient groups, SPH mainly consisted of palmitic acid (45 to 60%), followed by stearic acid (~ 10 to 15%) and small percentages of saturated long chain fatty acids (~ 2% 20:0, ~ 6% 22:0, and ~ 4% 24:0, Table 6). In view of the two control samples that could be analyzed, no comment can be given as to changes under pathophysiologic conditions. The analysis of the FA pattern of PS was possible only in a small number of cases (overall 16). The main component was palmitic acid (~ 50%), followed by oleic acid (~ 20%) and stearic acid (~ 15%). No information with regard to the changes under pathophysiologic conditions could be obtained.

Analysis of Surface Activity

A pronounced impairment of the surface tension-reducing capacity of LSA from all patient groups was encountered. In detail, all eight control samples reached a minimal surface tension ( min) of 0.0 ± 0.0 mN/m, whereas, in contrast, the  min values were markedly elevated to 15.5 ± 2.2 mN/m (ARDS, n = 7), 18.0 ± 1.8 mN/m (ARDS + PNEU, n = 8), and 16.1 ± 1.6 mN/m (PNEU, n = 17). Interestingly, a considerable correlation between  min and the relative content of either 16:0 (Figure 2) or 18:1 in PC of LSA was observed.

View larger version (17K): [in this window] [in a new window]   Figure 2.   Correlation between the relative amount of palmitic acid (16:0; given as percentage of all FA) in PC from LSA and the  min. Open circles represent values from healthy control subjects; solid squares represent values from patients with ARDS or pneumonia, or both.

Reconstitution Experiments

Because of the limited sample amount, only 9 of 39 patient samples could be subjected to the reconstitution experiments addressing the effect of a secondary supplementation with exogenous DPPC. In all cases, a marked reduction of  min values was noted. In detail,  min decreased from 19.1 mN/m to 11.4 mN/m upon reconstitution (p < 0.01; see also Figure 3).

View larger version (11K): [in this window] [in a new window]   Figure 3.   Effect of supplementation of individual LSA samples from patients with ARDS or pneumonia with DPPC on  min. Before DPPC supplementation, the relative amount of palmitic acid in PC was found to range at 64.5 ± 3.2% (mean ± SE). After supplementation, the relative amount of palmitic acid was 80.0% (thus approximating control values). Data represent mean ± SE. ** p < 0.01 compared with untreated LSA.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Human lung surfactant PC, the predominant PL class in unfractionated BALF and in the LSA, was found to be highly enriched in palmitic acid, which accounted for approximately 80% of all FA in the LSA fraction. This is in accordance with previous studies indicating a relative amount of 16.0 of approximately 74% [upon density centrifugation, Shelley and coworkers (17)] and 68% [BALF, Hallman and coworkers (2)]. In PG originating from healthy control subjects, oleic acid was found to be the most abundant FA, accounting for approximately 42% of all FA, followed by palmitic and stearic acid. This FA pattern of PG is similar to that reported by Shelley and coworkers (17), who described 52.2% 18:1, 22.2% 16:0, and 16.8% 18:0. Because of their low abundance under physiologic conditions, data of the FA profile of the minor surfactant phospholipids PI, SPH, PS, and PE are scarce. In the current investigation, PI presented a FA profile nearly equal to that of PG, which is not surprising in view of the common synthesis pathway of these two PLs, whereby both originate from the same precursor, cytidine diphosphodiacylglycerol (CDP-DAG) (18). This observation further supports the assumption that the acidic PI may fully replace PG during the neonatal development period (19), as also suggested by in vitro studies addressing the biophysical activity of PG-enriched versus PI-enriched PL mixtures (20, 21). Finally, the PG and PI FA compositions are in contrast to the composition of membrane-associated PI which possesses a high percentage of arachidonic acid (~ 30%) and low amounts of oleic acid (~ 7%) (22, 23). This suggests that under physiologic conditions the low amounts of PI contained in adult human surfactant represent a surfactant-specific PL pool and are not the result of contamination with plasma or membrane constituents.

SPH from control subjects displayed a FA pattern between that of PC and PG/PI. The most abundant FA was palmitic acid, followed by stearic acid, oleic acid, and linoleic acid. PE consisted of oleic acid > palmitic acid > linoleic acid > stearic acid and contained considerable amounts of arachidonic acid (~ 10%).

Under conditions of acute respiratory failure, whether triggered by primary nonpulmonary (ARDS) or primary pulmonary (PNEU) events or both, a marked and common change in the PL profile and FA composition was encountered. Decreased percentages of PC and PG, with a concomitant increase in SPH, PS, PI, and PE, have already been noted under these conditions (2, 24) and were corroborated in the present study. With regard to the FA profile of single PLs, the most striking observations was the loss of palmitic acid in PC, thus resulting in a 15 to 20% reduction in the overall content of saturated species in PC. Whereas all other saturated FA remained more or less unchanged, this reduction was paralleled by a pronounced increase in unsaturated FA in PC, mostly oleic acid, linoleic acid, palmitoleic acid, and arachidonic acid.

Some investigators have previously analyzed the content of DSPC in ARDS (2, 4, 8, 25) by means of the commonly used osmium tetroxide (OsO₄) method (26), however, with conflicting results. Data on PC FA profiles in ARDS, based on gas chromatography, are scarce. In their early report, von Wichert and Kohl (25) found a decreased ratio of palmitic acid/oleic acid in lung tissue extracts of patients with ARDS (1.64), as compared with normal lungs (2.32). Furthermore, Hallman and colleagues (2) reported a reduction of palmitic acid in PC (BALF) from approximately 68% in control subjects to approximately 47% in patients with ARDS. Both findings thus favorably underscore our data of a markedly reduced percentage of palmitic acid in the PC fraction originating from patients with ARDS, as well as for severe pneumonia.

To our knowledge, this is the first report also addressing the FA profile of minor pulmonary surfactant PL under conditions of respiratory distress. Although the FA pattern of PG and PI was similar in control subjects, it clearly differed in all patient groups. The degree of palmitoylation was reduced in PI, but increased in PG. Concerning the FA profile of SPH, an increase in 16:0 at the expense of 18:1 was encountered in all patients, with no major change of the other FA. Interestingly, the FA profile of PE did not change under conditions of acute respiratory failure.

How relevant are these changes in FA profiles in view of surfactant function? The lowering of surface tension especially at end-expiration is a result of a complex interaction of several surfactant compounds and is still not completely understood. Underlying physicochemical considerations imply that PC may only be substituted by a limited number of other FA with respect to the goal of a  min near 0 mN/m. The underlying reason for this may be the dependency of the critical transition temperature on the FA profile, which increases with chain length and decreases with the number of double bonds. The transition temperature of DPPC is 41° C (27), with this predominant surfactant PL in a gel-crystalline state at physiologic body temperatures, with rigid acyl groups allowing dense packing of the molecule upon lateral compression of the film and thus low surface tension values. Considering this background, it is of interest that the  min of LSA in the present study reached near zero mN/m in control subjects, but dramatically increased in patients with ARDS or severe lung infection, to an extent that must be assumed to result in decreased alveolar stability, thereby contributing to the impairment of gas exchange in these patients. Interestingly, a significant correlation between the relative content of palmitic acid in PC and the  min of the LSA was observed, and supplementation of LSA with exogenous DPPC indeed resulted in a significant improvement, however not a complete normalization, of  min values. Irrespective of the fact that a number of additional pathogenic events appear to contribute to the deterioration of surfactant function in acute respiratory failure, i.e., protein leakage into the alveolar space and changes in the LSA apoprotein content, especially SP-B and SP-C, these findings suggest that the changes in the PC FA composition contribute significantly to the pronounced loss of surface activity.

In conclusion, this study is the first to report the complete FA profiles of different surfactant PL classes in patients suffering from ARDS triggered by nonpulmonary underlying events and severe pneumonia. When compared with control subjects, marked changes were noted, similarly in both entities and in a mixed category (ARDS + PNEU). A highly significant loss in the percentage of palmitic acid, coincident with marked alterations in PG composition, may well contribute to the deterioration of surfactant function encountered under these conditions of acute respiratory distress.