INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The adult respiratory distress syndrome (ARDS) is still a major problem in critical care medicine (1). Among a variety of initial events that may lead to ARDS, sepsis is one of the most common causes with the poorest prognosis (2). In their initial study, Ashbaugh and coworkers (3) noticed reduced lung compliance in the early course of ARDS, and postulated that surfactant alterations may be involved in the development of ARDS. Subsequently, this hypothesis has been supported by several studies investigating bronchoalveolar lavage (BAL) material obtained from patients suffering from ARDS (4). Moreover, surfactant replacement therapy yielded promising results both in experimental and clinical studies (5, 7).

Electron microscopic studies of the pathogenetic processes involved in the early development of sepsis-associated ARDS focused on alterations of the endothelium, the role of leukocytes, the occurrence of alveolar edema, and the role of type II pneumocytes in alveolar repair during the chronic phase of ARDS. Little is known, however, about ARDS-related alterations of the pulmonary surfactant system at the ultrastructural level (8). In endotoxin-treated rats, ultrastructural analysis of BAL material demonstrated changes in the relative amounts of surfactant subtypes after 8 and 12 h (9). Recently, we showed that in rats Salmonella minnesota lipopolysaccharide (LPS) resulted in the formation of giant surfactant storing lamellar bodies in type II pneumocytes within 2 h after application, both in vitro and ex vivo (10). However, the question remained whether the formation of giant lamellar bodies was associated with alterations of ultrastructure, composition, and functional activity of secreted alveolar surfactant. Further, it was still unclear which mechanism was responsible for giant lamellar body formation.

Therefore, the present study aimed at a detailed morphometric analysis of the LPS-induced ultrastructural alterations of type II cells, in particular of intracellular surfactant, and of their secretory product, the alveolar surfactant. Due to perfusion fixation, and the particular tissue processing procedure applied (11), surfactant was stabilized and retained in the alveoli so that it could be analyzed in situ. Furthermore, BAL material was analyzed for phospholipids, proteins, and surfactant protein A (SP-A), and its activity was studied by means of a pulsating bubble surfactometer. The main aim of the study was to investigate the effects of endotoxin on pulmonary surfactant. However, we were also interested in studying the contribution of blood-derived leukocytes to these alterations. Therefore, to examine the role of leukocytes LPS treatment was carried out in vivo as well as in the isolated and blood-free perfused rat lung (10, 12, 13). Parts of this work have been published in abstract form (14).

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Experimental Design

All lungs were isolated, perfused, and ventilated for either 10 or 150 min. Then, the lungs were either lavaged or prepared for electron microscopy as described below. Since the viability of an isolated lung preparation is limited to a few hours (12), the present studies focused on the early alterations of surfactant ultrastructure and function, seen within 2 h after treatment. In order to study the effects of LPS in vitro and ex vivo, the following experimental groups were analyzed:

Group 1 (control, in vitro): lungs were prepared from untreated rats and were perfused for 150 min. After 40 min, 500 µl phosphate-buffered saline (PBS)/0.005% hydroxylamine were infused into the tubing leading to the pulmonary artery.

Group 2 (LPS, in vitro): lungs were prepared from untreated rats and were perfused for 150 min. After 40 min, the lungs were perfused with 50 µg/ml LPS (dissolved in 500 µl PBS/0.005% hydroxylamine).

Group 3 (control, ex vivo): rats were injected with 500 µl PBS/ 0.005% hydroxylamine about 110 min prior to start of the perfusion. Subsequently, lungs were isolated, perfused, and ventilated for 10 min.

Group 4 (LPS, ex vivo): rats received an intraperitoneal injection of 20 mg/kg LPS dissolved in 500 µl PBS/0.005% hydroxylamine about 110 min prior to start of the perfusion. Subsequently, lungs were isolated, perfused, and ventilated for 10 min.

Materials

Female Wistar rats (220 to 250 g) were obtained from the Zentralinstitut für Versuchstierzucht in Hannover, Germany. Pentobarbital sodium (Nembutal) was purchased from the Wirtschaftsgenossenschaft Deutscher Tierärzte (Hannover, Germany). Lipopolysaccharide Salmonella minnesota, and heparin was from Sigma (Deisenhofen, Germany); HEPES and glucose from Boehringer Mannheim (Mannheim, Germany). The materials used for preparation of the fixation and staining solutions were: 25% glutardialdehyde, EM grade (Merck, Darmstadt, Germany), paraformaldehyde, extra pure (Merck), dimethylarsinic acid sodium salt trihydrate, for synthesis-grade (Merck-Schuchardt, Hohenbrunn, Germany), osmium tetroxide, crystalline (Paesel+Lorei, Frankfurt, Germany), uranyl acetate dihydrate, for analysis-grade (Merck), and Araldite (Serva, Heidelberg, Germany).

Isolated Perfused Rat Lung Preparation

The rat lungs were prepared and perfused as described recently (12). Briefly, excised lungs were perfused at constant hydrostatic pressure through the pulmonary artery with Krebs-Henseleit buffer that contained 2% albumin, 0.1% glucose, and 0.3% HEPES. The total recirculating amount of buffer was 100 ml. The lungs were suspended by the trachea and were ventilated with humidified and warmed air by negative pressure ventilation with 80 breaths/min and a tidal volume between 1.6 and 2 ml. Every 5 min a deep breath (hyperinflation at 16 cm H₂O) was initiated. A number of physiological parameters such as tidal volume, pulmonary resistance, dynamic pulmonary compliance, and vascular resistance were monitored continuously (10, 12, 13). Static pulmonary compliance was determined during exhalation at 80% (about 8 ml) of the total lung capacity.

Transmission Electron Microscopy (TEM)

The lungs to be examined by electron microscopy were fixed immediately after each experiment as described in our previous study (10). Briefly, at the end of expiration the lungs were fixed via the pulmonary artery by recirculating perfusion at a hydrostatic pressure of 13 cm H₂O for 10 to 15 min. The fixative used was a mixture of 1.5% glutaraldehyde and 1.5% freshly prepared paraformaldehyde in 0.1 M cacodylate buffer (buffer osmolality: approximately 300 mOsm/kg, pH 7.35). After crossclamping of the hilus, the lungs were stored in cold fixative for 24 h. Lung volume was determined by fluid displacement immediately before systematic random sampling of tissue blocks. Briefly, each lung was embedded in 2% agar-agar, and then cut into 3-mm-thick slices using a self-made tissue slicer. Then, a square lattice test system (256 points, d = 5 mm) was superimposed at random over the lung slices. From each site a 3 × 3 × 3 mm³ tissue block was collected.

In order to achieve adequate stabilization of intracellular and alveolar surfactant material, en bloc staining of tissue samples with uranyl acetate was performed before dehydration. This method has previously been shown by means of electron energy loss spectroscopy to improve phospholipid retention in surfactant storing lamellar bodies (11). Tissue processing was performed using an automated tissue processor (Histomat; Bio-med, Theres, Germany). Briefly, samples were rinsed several times, then postfixed for 2 h in 1% OsO₄ in 0.1 M cacodylate buffer, washed again in buffer followed by distilled water, and then transferred to cold, half-saturated aqueous uranyl acetate solution (12 h at 8° C). After dehydration through a graded series of ethanols, tissue samples were transferred to Araldite via propylene oxide, and polymerized for 3 d at 60° C. Tissue blocks were allowed to acquire random orientation in the embedding capsules. For TEM examination with an EM 10 A (Zeiss, Oberkochen, Germany), ultrathin sections were counterstained with lead citrate using an Ultrostainer (Leica GmbH, Hamburg, Germany).

Morphometry

Quantitative analysis of intracellular surfactant. Ultrathin sections of four tissue blocks per lung were cut, and one qualitatively good section per block was analyzed. Following the rules of systematic quadrate (SQ) subsampling, the whole section was examined moving the microscope stage at given uniform intervals along a square lattice test system determined by the stage coordinates. Micrographs were taken of all profiles of type II pneumocytes that hit the center of the fluorescence screen of the microscope whenever a test position was reached. Micrographs were recorded on thin sheet film (Kodak EM film 4489) at a primary magnification of ×5,000 to ensure that the whole cell profile was recorded. The final magnification after photographic reproduction was determined by means of a calibration grid to be ×12,800. A total of 465 profiles of type II pneumocytes were analyzed (mean ± SD, 29 ± 9 cell profiles per lung). A multipurpose test point system (M168) with 168 points, 84 test lines, and a test line length of d = 1.09 µm (at ×12,800) was used for determination of morphometric parameters by means of point and intersection counting (15). Thus, a total of 14,993 points falling on type II pneumocytes were counted (mean ± SD, 937 ± 338 points per lung). To quantify ultrastructural alterations of type II pneumocytes the following parameters were determined: the volume densities (Vv) of lamellar bodies (lb), mitochondria (mi), nucleus (nu), and the remaining cytoplasm (cy) with reference to the whole type II cell (PII) (abbreviations follow Vv_{component,PII}). To exclude that alterations in the volume densities simply reflect changes in the reference space, i.e., the cell, Vv_PII was determined with reference to the constant test system (Vv_PII,TS) (16, 17). In order to further characterize the intracellular surfactant stores, the volume-to-surface-ratio of the lamellar bodies (v/s-ratio_lb), which is proportional to the mean caliper diameter, was calculated according to

(1)

where L_T = test line length; P_lb = number of points hitting lamellar bodies; and I_lb = number of intersections with lamellar bodies. In addition, the numerical density of lamellar bodies per type II pneumocyte (Nv_lb,PII) was determined according to the equation

(2)

where N_a = number of particle profiles per cell profile; A_PII = cell profile area; and = mean caliper diameter (15). The product of the number of test points (P_PII) times squared test line length (L_T2) gives an estimate of A_PII. Because 3-D reconstruction has shown that the lamellar bodies of the rat are approximate spheres (18), the mean caliper diameter could be approximated according to the equation = 6.0 × v/s-ratio_lb (15).

Quantitative analysis of alveolar surfactant. In a second independent study, the same ultrathin sections were examined once more to determine alveolar surfactant and type II pneumocytes relative to parenchyma. According to the rules of SQ subsampling, the whole section was examined moving the microscope stage at given uniform intervals along a square lattice test system determined by the stage coordinates. At each position, point counting was performed online to determine all the points hitting type II pneumocytes and all the points falling on alveolar surfactant. For online morphometric analysis we used a CCD camera (CF8; Kappa Meßtechnik, Germany) connected to a TV monitor (Bosch, Germany). Point counting was performed at a calibrated magnification of ×50,000 using an 88-point test system, which was printed on a transparent sheet and superimposed on the TV screen. From each SQ-sampled area all points hitting the different subtypes of the alveolar surfactant were counted (mean ± SD, 543 ± 198 points per lung) as well as all the points falling on type II pneumocytes (mean ± SD, 1,084 ± 321 points per lung). According to Balis and coworkers (19), four surfactant subtypes were distinguished: (1) lamellar body-like surfactant forms (lbs), (2) tubular myelin (tbm), (3) unilamellar structures (uls), and (4) multilmellar structures (mls). The volume densities of the different surfactant forms were determined with reference to type II pneumocytes (e.g., as Vv_tbm,Pll), which gives an indication of the alveolar surfactant volume per unit cell volume. Furthermore, the volume density of each subtype was calculated with respect to the total alveolar surfactant (e.g., as Vv_tbm,AS), which gives an indication of the relative composition of alveolar surfactant (19). Morphometric quantification could not include the surface lining surfactant monolayer, however, owing to its two- dimensional appearance.

Analysis of BAL

In vivo, loss of pulmonary surfactant function may occur because of decreased amounts or altered structure of surfactant. Therefore, we did not normalize our lavage samples with respect to surfactant concentration, but instead in order to approach the in vivo condition, in our measurements we used the unprocessed first lavage. It was important to ascertain that the recovery was similar under all conditions (Table 1). Lungs were lavaged once with 8 ml of cold (+4° C) PBS and the lavage was centrifuged at 150 g (+4° C) for 10 min. The supernatant was removed and analyzed for biophysical activity as well as for phospholipid and protein content.

Alternatively, to study alterations of surfactant independent of the phospholipid concentration, the phospholipids were extracted to obtain the same concentration of 6.8 mM phosphorus in all samples. Briefly, in a separate set of experiments lungs were lavaged five times, the recovered lavage fluid was centrifuged for 5 min at 150 g (4° C) and subsequently the supernatant was centrifuged for 15 min at 40,000 g (4° C). The pellet was resuspended in PBS containing 1.5 mM CaCl₂, the phospholipid concentration was adjusted to 6.8 mM, and finally these samples were analyzed for their biophysical properties. These conditions were reported to yield mainly large surfactant aggregates (20).

The surface activity of the samples was assessed by means of a pulsating bubble surfactometer. The bubbles were cycled with 0.05 Hz at +37° C, and values were obtained after 1 min of cycling. Given are the maximum and the minimum surface tension, as well as the hysteresis area, which was calculated as the product of surface area and surface tension.

From the supernatant of the first lavage, phospholipids were extracted by methanol/chloroform. Subsequently, the extracts were analyzed for phosphorus (21). The phosphorus concentrations obtained were multiplied by 25 in order to estimate the phospholipid concentrations. Proteins were determined by the method of Bradford and SP-A by an ELISA performed with antibodies that were kindly provided by Dr. D. S. Phelps (Hershey, Pennsylvania).

Statistics

Due to the experimental design, the four groups studied differed not only in LPS being present or absent, but also in the duration of perfusion (10 or 150 min), which resulted in a 2 × 2 table. To analyze these data for the effects of time and LPS treatment simultaneously, a two-factor analysis of variance (ANOVA) was performed. Parametric ANOVA was performed only if the hypotheses of normality and equal variance were not rejected at p < 0.05. In case of interactions between time of perfusion and LPS treatment, post hoc multiple comparisons were performed according to the Student-Newman-Keuls method. A value of p < 0.05 was considered significant. All statistical analyses were performed on an IBM personal computer using the software program SigmaStat (Jandel Scientific, Erkrath, Germany). The statistical power 1- of the performed ANOVA, i.e., the chance of finding a difference, was calculated as detailed by Zar (22).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Functional Changes

The functional consequences of treatment with LPS for 2 h either ex vivo or in vitro have previously been described in detail (10). Briefly, in that study we showed that in perfused rat lungs, LPS causes bronchoconstriction, which is mediated by cyclooxygenase-2-(COX-2)-dependent thromboxane formation (13). Under these experimental conditions, LPS induced neither pulmonary hypertension nor edema formation. Our present findings are in accord with those data and we confirmed the increase in pulmonary resistance induced by LPS in perfused rat lungs (Table 2). In this study, we additionally assessed static pulmonary compliance, which was reduced by time of perfusion, but not by endotoxin treatment (Table 2).

Ultrastructural and Morphometric Analysis of Type II Pneumocytes

The lungs of the four experimental groups showed clear differences in the ultrastructure of the type II pneumocytes (Figure 1) and in the alveolar surfactant (Figures 2 and 3). Notably, these changes were obvious in the absence of any sign of alveolar or interstitial edema. Ultrastructural changes were observed with respect to the time of perfusion/ventilation and the presence or absence of LPS in the perfusate, respectively. In isolated lungs subjected to 150 min of perfusion and negative pressure ventilation (Groups 1 and 2), lamellar bodies were smaller than in lungs of the corresponding groups perfused for only 10 min (Groups 3 and 4). In addition, the cisternae of the endoplasmatic reticulum were slightly swollen in the lungs perfused for 150 min (Figure 1b). The type II pneumocytes of lungs treated with LPS (Groups 2 and 4) contained larger lamellar bodies than the type II cells of control lungs (Group 1 and 3). Giant lamellar bodies (Figures 1c and 1d) were seen in lungs treated with LPS in vitro and ex vivo (Groups 2 and 4). Mitochondria of type II pneumocytes showed slightly more intense staining in control lungs perfused for 150 min (Group 1) and in LPS-treated lungs (Groups 2 and 4) than mitochondria of type II cells after only 10 min of perfusion (Group 3). There were no indications of ultrastructural alterations of the nucleus.

View larger version (235K): [in this window] [in a new window]   View larger version (222K): [in this window] [in a new window]   View larger version (200K): [in this window] [in a new window]   View larger version (197K): [in this window] [in a new window]   Figure 1.   TEM micrographs of type II pneumocytes. (a) Normal ultrastructure in untreated lung after 10 min of isolated perfusion and ventilation (Group 3). (b) Slightly swollen endoplasmic reticulum and nuclear cisterna in untreated lung after 150 min of perfusion and ventilation in vitro (Group 1). Giant lamellar body (GLb) in (c) LPS-treated lung in vivo after 10 min of perfusion and ventilation (Group 4), and in (d ) LPS-treated lung in vitro after 150 min of perfusion and ventilation (Group 2) showing an exceptionally gigantic body. Lb = lamellar bodies; M = mitochondria; Nu = nuclei.

View larger version (179K): [in this window] [in a new window]   View larger version (146K): [in this window] [in a new window]   View larger version (177K): [in this window] [in a new window]   View larger version (184K): [in this window] [in a new window]   Figure 2.   TEM micrographs of tubular myelin subtype (TM) of alveolar surfactant. Normal ultrastructure in (a) untreated lungs after 10 min of isolated perfusion and ventilation (Group 3), and (b) after 150 min of perfusion and ventilation in vitro (Group 1). (c) Small vesicles (arrowheads) with moderately dense content are associated with tubular myelin membranes in LPS-treated lung in vivo followed by 10 min of perfusion and ventilation in vitro (Group 4), but are less prominent in (d ) LPS-treated lung in vitro after 150 min of perfusion and ventilation (Group 2). LbS = lamellar body-like surfactant form.

View larger version (168K): [in this window] [in a new window]   View larger version (139K): [in this window] [in a new window]   View larger version (166K): [in this window] [in a new window]   View larger version (150K): [in this window] [in a new window]   Figure 3.   TEM micrographs of multilamellar subtype (MLS) of alveolar surfactant. (a) Normal ultrastructure in untreated lung after 10 min of isolated perfusion and ventilation (Group 3). (b) Multilamellar structures are slightly more prominent in lungs treated with LPS in vitro after 150 min of perfusion and ventilation in vitro (Group 1). (c) Multilamellar structures are much more prominent in lungs treated with LPS in vivo (Group 4), and reach their largest extension in (d ) lungs treated with LPS in vitro (Group 2). Numerous small vesicles are associated with multilamellar surfactant forms in lungs treated with LPS in vivo. Note that final magnification of C, D is about half the magnification of A, B. TM = tubular myelin.

These observations were further substantiated by the morphometric data concerning the type II pneumocytes which are summarized in Table 3. There were no differences between the four groups studied in the volume density of the type II cell referred to the test system (Vv_PII,TS), which means that there were no alterations in mean cell volume due to perfusion/ventilation and/or LPS treatment. Therefore, the volume densities of intracellular components determined with reference to type II cells were comparable between the different experimental groups (16). An interference of changes in the reference space, i.e., type II cell volume, could be excluded.

In lungs perfused for 150 min (Groups 1 and 2) the volume density of the nucleus and the mitochondria was significantly higher than in lungs perfused for only 10 min (Groups 3 and 4). The volume density of the remaining cytoplasmic components (Vv_cy,PII), which include cytoplasmic ground substance and endoplasmic reticulum, was not affected by perfusion and ventilation for 150 min. There was no significant effect of LPS treatment on nucleus, mitochondria, and cytoplasm, although an interaction of LPS and time was observed with respect to Vv_mi,PII.

Looking at the surfactant-storing lamellar bodies of type II pneumocytes, both LPS treatment and time of perfusion/ventilation resulted in significant and profound, but opposite alterations. The increase in time of perfusion from 10 min (Groups 3 and 4) to 150 min (Groups 1 and 2) resulted in a decrease of Vv_lb,PII (by about 55%, i.e., approximately 24% per hour), and of v/s-ratio_lb. In contrast, LPS treatment in vitro (Group 2) and ex vivo (Group 4) resulted in higher values of Vv_lb,PII (by about 27%) as compared with the corresponding control groups (1 and 3, respectively), which means that the time-dependent decrease of Vv_lb,PII was attenuated. Moreover, v/s-ratio_lb was increased, and Nv_lb,PII was decreased in lungs subjected to LPS treatment (Groups 2 and 4) as compared with untreated lungs (Groups 1 and 3). In summary, irrespective of the time of perfusion/ventilation, LPS resulted in lamellar bodies being larger in size and smaller in number, with their amount per cell being elevated.

Ultrastructural and Morphometric Analysis of Alveolar Surfactant

The alveolar surfactant was observed to be present in different ultrastructural subtypes (Figures 2 and 3). Following Balis and coworkers (19), these surfactant subtypes comprised lamellar body-like forms, multilamellar structures, unilamellar structures, and the characteristic tubular myelin. We regularly observed that lamellar body-like forms were associated with tubular myelin (Figure 2a), which in turn was connected to the surfactant monolayer. Densely coiled lamellae were also seen within multilamellar structures (Figure 3). However, we did not see transitions between multilamellar structures and tubular myelin. As concerns the ultrastructural appearance of lamellar body-like surfactant forms and unilamellar structures, the qualitative examination revealed no differences between the four groups studied. Multilamellar structures (Figure 3), however, were less prominent in control lungs perfused for only 10 min (Group 3), while they were frequently observed in control lungs perfused for 150 min (Group 1), and were abundantly present in all LPS-treated lungs (Groups 2 and 4). Multilamellar structures in lungs of Groups 1, 2, and 4 consisted of more lamellae than in lungs of Group 3. In addition, they were occasionally seen to contain small unilamellar vesicles. In lungs treated with LPS ex vivo (Group 4), focal accumulations of small unilamellar vesicles were observed (Figure 3c). Furthermore, very similar small vesicles, some of which possessed a moderately dense core, were present between the membranes of tubular myelin particularly in lungs treated with LPS ex vivo (Group 4) (Figure 2c). The tubular myelin of control lungs (Groups 1 and 3) showed the normal ultrastructural appearance (Figures 2a and 2b).

The morphometric data regarding the alveolar surfactant are summarized in Table 4. Because one has to take into account that the data strongly depend on each other when subtypes are determined with reference to total alveolar surfactant, we decided to additionally determine the volume density of each subtype per unit cell volume of type II pneumocytes. This reference can be regarded to be fairly constant in our study, because the morphometric data have shown that type II cell volume did not differ between experimental groups (Table 3), and proliferation of type II pneumocytes within 2 h can be excluded. Both duration of perfusion/ventilation and treatment with LPS caused alterations of surfactant subtypes. The volume density of all forms together (Vv_AS,PII) showed a significant increase with increased duration of perfusion and ventilation. Although the increase in Vv_AS,PII caused by LPS treatment was not significant (p = 0.123), it cannot be ruled out (power ~ 0.3). With regard to the individual surfactant subtypes, lungs perfused and ventilated for 150 min (Groups 1 and 2) showed a significant increase in the volume densities of lamellar body-like forms (Vv_lbs,PII), multilamellar structures (Vv_mls,PII), and unilamellar structures (Vv_uls,PII), respectively, as compared with lungs perfused for only 10 min (Groups 3 and 4). LPS treatment resulted in an additional increase of Vv_mls,PII and Vv_lbs,PII. The volume density of tubular myelin (Vv_tbm,PII) was decreased, although at a level of significance of p = 0.075 only (power  0.3). Looking at the relative composition of the alveolar surfactant, both duration of perfusion and LPS treatment resulted in an increase in Vv_mls,AS and a concomitant decrease in Vv_tbm,AS, respectively.

BAL Material: Composition and Activity

Table 1 summarizes our data on the composition of BAL material. For these data, we analyzed the composition of the first lavage. The total amount of lavaged material was not different in the four groups. Protein concentration was significantly increased in lungs perfused and ventilated for 150 min (Groups 1 and 2) as compared with lungs perfused for only 10 min (Groups 3 and 4). The phospholipid concentration in the lavage from LPS-treated lungs was decreased. The concentration of SP-A was 49 ± 7 ng/ml (n = 3) and 73 ± 22 ng/ml (n = 3), in Groups 3 and 4, respectively.

The data obtained from measurements of original BAL fluid (first lavage) by the pulsating bubble surfactometer are given in Table 5. Minimal and maximal surface tensions did not show any significant differences with respect to LPS treatment and duration of perfusion, respectively. However, the hysteresis area obtained from plotting surface area versus surface tension showed a significant decrease in BAL material obtained from LPS-treated lungs (Groups 2 and 4) as compared with BAL material from untreated control lungs (Groups 1 and 3). Table 5 also shows the biophysical properties of BAL material that was standardized to the same concentration of phospholipid. We found that there was no effect of LPS, and only a small effect of perfusion time on hysteresis area.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

General Aspects

At present, inactivation of surfactant by alveolar edema fluid is considered to be an important cause of surfactant alterations in ARDS (5, 23). Our study suggests that alterations in surfactant homeostasis may occur before the manifestation of edema. Most of the evidence supporting a role of surfactant in the development of ARDS is derived from studies of BAL material obtained either from ARDS patients (4, 6, 7, 23) or from experimental models of ARDS (5). However, since BAL analysis is prone to experimental errors (5, 24), such as bronchoconstriction elicited by LPS in vitro (10, 13), BAL analysis alone may be insufficient to study LPS-induced surfactant alterations. Therefore, we studied the composition and activity of surfactant material recovered by BAL, and the ultrastructure of alveolar and intracellular surfactant retained in situ by means of morphometry.

In isolated control lungs, perfusion and ventilation resulted in ultrastructural changes of type II pneumocytes corresponding to those reported by others (16). The increase in perfusion time from 10 to 150 min resulted in a slight swelling of mitochondria and nuclei, as indicated by increased mitochondrial and nuclear volume densities, while swelling of the cells can be excluded because the volume fraction of cells referred to the test point system was not affected. The volume density of intracellular surfactant stores (Vv_lb,PII) decreased by about 24% per hour, which may be explained by the regular deep breaths executed every 5 min (16). The decrease in Vv_lb,PII was accompanied by a decrease in the mean size of lamellar bodies as indicated by their v/s-radio, while their numerical density was unaffected.

Concomitant with the decrease in Vv_lb,PII, the volume density of alveolar surfactant per type II pneumocyte (Vv_AS,PII) increased with time. Looking at the different surfactant subtypes, lamellar body-like forms as well as multilamellar and unilamellar structures showed an increase in Vv, while tubular myelin was not affected. There is good evidence that lamellar body-like forms represent freshly released surfactant, which is transformed into tubular myelin under appropriate conditions (25). Tubular myelin plays a key role in the generation of the surface-active monolayer, while unilamellar structures are thought to represent used surfactant ready for reuptake (25, 26). Our findings, that the amount of tubular myelin per cell was not altered while all other subtypes increased with time, are in line with biochemical data reported by Power and coworkers (28) that after 30 min of hyperventilation the phospholipid content of the tubular myelin fraction was unchanged, while a reversible increase was observed in the other subfractions.

Despite the increase in Vv_AS,PII and despite no change in the amount of phospholipids recovered by lavage, static pulmonary compliance decreased by time of perfusion. This decrease in static compliance appears to be the major factor that limits stability of perfused lung preparations, though the reason for this decrease is not quite clear. Subtle alterations of surfactant, as, for example, in its composition of different subtypes, may contribute to this effect. This is supported by our finding that time of perfusion reduced the hysteresis area of both unprocessed lavage fluid and purified surfactant (Table 5). Though protein concentration in the BAL was slightly increased after 150 min of perfusion, we may exclude edema as a potential cause, since even at the ultrastructural level we could not demonstrate any sign of pulmonary edema (except perivascular) (10).

Effects of Salmonella minnesota Endotoxin

Intracellular surfactant. While mitochondria and nuclei were affected by time of perfusion and ventilation alone, lamellar bodies were the only type II pneumocyte component that showed a significant quantitative response to LPS treatment (Table 3). Irrespective of the presence or absence of blood- derived leukocytes, the same alterations were observed. They were manifest in different ultrastructural characteristics: (1) the volume density of lamellar bodies was significantly higher in LPS-treated than in control lungs, which literally means that the volumetric decrease in lamellar bodies with increasing time of perfusion and ventilation was reduced; (2) as indicated by the v/s-ratio, this effect was paralleled by an increase in their mean size; and (3) their numerical density was significantly lower in LPS-treated than in control lungs. This indicates that several lamellar bodies fuse to form giant bodies rather than that individual bodies grow by incorporation of newly synthesized material.

Enlargement of lamellar bodies was reported to occur in response to various effects and in Chediak-Higashi syndrome (see Reference 10 for references). The increase in lamellar body size observed in these studies has been attributed either to a mere fusion of lamellar bodies and/or a decreased rate of surfactant release. Our findings of an increased volume density of lamellar bodies and a diminished phospholipid concentration in the BAL of LPS-treated versus control lungs indicate that LPS induced a decrease in the rate of surfactant secretion. In line with this, type II pneumocytes isolated from lungs of LPS-treated rats showed a reduction in tetradecanoyl-phorbol-acetate (TPA)-stimulated surfactant secretion, whereas basic secretion was not affected (29). Possibly, this effect of LPS is related to changes in the cytoskeleton. Surfactant release is known to be a stimulus-driven process of exocytosis, in which the cytoskeleton and fusion of cell membranes are intimately involved (25, 30). LPS was shown to bind to the cell membrane of isolated type II pneumocytes and alter its properties (31). Endotoxin stimulation has been reported to affect microtubules of alveolar macrophages and blood monocytes (32). Impairment of microtubule assembly was shown to be involved in the pathogenesis of the Chediak-Higashi syndrome (33).

Alveolar Surfactant

The effects of LPS on intracellular surfactant stores were associated with alterations of the structural composition of alveolar surfactant. LPS induced a relative increase in multilamellar structures at the expense of tubular myelin, which was also observed in ozone-treated (19) and in Escherichia coli-endotoxin-treated rats (9). However, one has to take into account that these parameters strongly depend on each other because they were determined with reference to total alveolar surfactant. Therefore, we additionally determined the volume densities per unit type II pneumocyte volume. This reference can be regarded to be fairly constant in our study, because the morphometric data have shown that type II cell volume did not differ between groups (Table 3), and proliferation of type II pneumocytes within 2 h can be excluded. Notably, these parameters showed that LPS resulted in an increase not only in multilamellar structures, but also in freshly secreted lamellar body-like forms. On the contrary, tubular myelin was decreased in LPS-treated lungs, even though only at p = 0.075. It should be noted, however, that the morphometric quantification of surfactant subtypes is not necessarily reflected in the total phospholipid concentration, particularly since the surfactant monolayer form cannot be included in the determination of volume densities.

The ultrastructural alterations of alveolar surfactant subtypes in response to LPS were paralleled by a decrease in BAL phospholipid concentration. BAL surfactant activity was altered as shown by the reduction in hysteresis area, which is indicative of a reduced compressibility of surfactant in LPS-treated lungs. The altered biophysical activity was not reflected by a change in static pulmonary compliance. Therefore, it appears that the LPS-induced surfactant alterations did not cause profound changes in lung function. It should be noted, however, that in the present study only one time point, i.e., 2 h after LPS administration, was studied. This is a relatively short time period and in future studies it will be important to investigate the relationship between surfactant and lung function after prolonged exposure to LPS.

Decreased surfactant phospholipid concentrations have also been observed in ARDS patients, in patients at-risk to develop ARDS (4), and in endotoxin-treated experimental animals (34, 35). Similar effects on pressure-volume curves were obtained by mixing LPS of various bacteria with surfactant preparations (35). Finally, reduced hysteresis area observed in patients with ARDS after multiple trauma was reported to correlate significantly with the severity of respiratory dysfunction (23). In our study, a reduction in hysteresis area was only observed in the largely unprocessed first lavage, whereas it was unchanged in preparations with standardized phospholipid concentrations. This suggests that the decrease in phospholipid concentration may be responsible for the reduction in hysteresis area. However, one has to take into account that the standardized phospholipid preparations do not contain all of the lavaged surfactant, i.e., they mainly consist of large aggregates (20).

Because LPS reduced the release of intracellular surfactant, one might expect that there was less surfactant material present in the alveoli. In particular, the amount of lamellar body-like forms should have been reduced. However, our finding that lamellar body-like forms per type II cell volume were increased in LPS-treated lungs while tubular myelin was slightly decreased, suggests that LPS also affected alveolar surfactant transformation, which is supported by others. SP-A was shown to aggregate with LPS in the presence of 0.5 mM Ca²⁺ (36), while both SP-A and Ca²⁺ are required for the transformation of lamellar body-like forms into tubular myelin (27). Conversion from heavy to light subfractions of BAL material was delayed in radiation pneumonitis, another model of ARDS (37), and the ratio of small-to-large aggregates was increased in septic adult sheep (38) and in patients with ARDS (6). An interference of LPS with surfactant transformation is supported by our observation that small unilamellar vesicles with a dense core were associated with tubular myelin in lungs treated with LPS ex vivo. Similar vesicles have been observed in mixtures of E. coli-LPS and sheep surfactant obtained in BAL (39).

In summary, our data indicate that LPS induced alterations of intracellular and alveolar surfactant irrespective of the presence or absence of blood-derived leukocytes. The alterations observed are indicative of a reduction in surfactant secretion associated with fusion of lamellar bodies to generate giant forms, and an interference of LPS with alveolar surfactant transformation. This resulted in disturbance of alveolar surfactant composition determined in situ, which was paralleled by moderate biochemical and functional alterations of surfactant material obtained by BAL.