INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The pulmonary surfactant prevents alveolar atelectasis by reducing the alveolar surface tension in a manner that depends on surface area (1). In addition, surfactant is an active component of the lung host defense mechanisms (2). Pulmonary surfactant consists of about 90% lipids, mainly saturated phosphatidylcholine, and about 10% proteins, including the surfactant apoproteins SP-A, SP-B, SP-C, and SP-D. It is synthesized, stored, secreted, and to a large extent recycled by type II pneumocytes (1). An intracellular surfactant pool, represented by the lamellar bodies of type II pneumocytes, and an intraalveolar surfactant pool can be distinguished. Ultrastructurally, intraalveolar surfactant consists of several subtypes, namely freshly secreted lamellar body-like forms, tubular myelin, the alveolar lining layer, and small unilamellar vesicles (3, 4). In bronchoalveolar lavage (BAL) studies two subfractions of intraalveolar surfactant are distinguished: large aggregates (LA) or heavy forms, largely corresponding to surface-active tubular myelin, and small aggregates (SA) or light forms, largely corresponding to degraded and inactive small unilamellar vesicles (4). Changes in alveolar pool size are expressed as changes in the SA/LA ratio (5).

The crucial role of the surfactant system for the maintainance of the functional integrity of the lungs is clearly demonstrated by the respiratory distress syndrome of premature neonates (nRDS), in which a primary surfactant deficiency leads to impaired pulmonary function (6). Surfactant impairment is also involved in the pathogenesis of the adult or acute respiratory distress syndrome (ARDS) (5, 7, 8). In addition, surfactant dysfunction has been reported to occur after ischemia and reperfusion (I/R) in experimental lung transplantation (9). Increasing SA/LA ratios have been reported for patients with severe ARDS (13) and in experimental I/R injury (9). Experimental evidence of the inhibition of normal surfactant function by plasma proteins (14, 15) led to the hypothesis that the impairment of the surfactant system after I/R is due to inactivation by extravasated plasma proteins and, therefore, a consequence of the development of intraalveolar edema. In contrast, disturbance in surfactant has been proposed to be a cause of edema formation (reviewed in Reference 5). This study sought to test two hypotheses: whether this impairment of the surfactant system is a mere consequence of I/R-induced intraalveolar edema formation or whether surfactant alterations may be a cause of edema formation. Therefore, we investigated the cumulative effects induced by the whole sequence of transplantation-related events on surfactant ultrastructure and pulmonary function, which includes flush perfusion, cold ischemic storage, and subsequent reperfusion of the lung, rather than examining the individual contribution of each event.

Because intraalveolar edema fluid accumulates only in a fraction of alveoli in the early phase of edema formation, the lungs exhibit both edematous regions and regions free of edema. If surfactant alterations were simply an effect of inactivation due to some edema component(s), then surfactant alterations would be expected in edematous regions only, while regions free of edema should contain normal intraalveolar surfactant. In turn, if surfactant alterations were seen irrespective of the presence of edema fluid, then surfactant alterations might well be a cause of I/R-induced edema formation. Because BAL studies, which may be prone to experimental error (7, 16), do not allow one to determine whether the changes in intraalveolar surfactant pool size derive only from edematous regions or also from regions free of edema, the only way to test these hypotheses is to look into the lungs by means of electron microscopy. Therefore, we have chosen an ultrastructural and stereological approach to quantify the different intraalveolar surfactant subtypes fixed and retained in situ as described (17). Surfactant composition was determined in edematous regions and regions free of edema, which was induced by I/R in a rat isolated perfused lung model. In addition, hemodynamic and respiratory data were recorded during reperfusion of the lungs in order to find alterations in lung function.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Rat Isolated Lung Model

An established rat isolated perfused lung model was used (18). In 10 male Sprague-Dawley rats anesthesia was achieved with pentobarbital (Nembutal, 1 mg/kg body weight, administered intraperitoneally), followed by heparinization via the inferior caval vein (100 IU), operation, and excision of the lungs as described (18). The left lungs of the control animals (n = 5) were fixed for ultrastructural analysis immediately after excision. In the Euro-Collins group (EC; n = 5) the lungs were flush perfused with 20 ml of modified Euro-Collins solution. The organs of the EC group were stored for 2 h at 4° C in EC, with the trachea clamped after the lungs had been inflated with 5 ml of room air, and reperfused with Krebs-Henseleit buffer containing washed bovine red blood cells at a hematocrit of 38-40% for 40 min. During EC perfusion and during reperfusion, ventilation with room air at a tidal volume of 4 ml and a rate of 40 breaths per minute was continuously performed. A positive end-expiratory pressure (PEEP) of 3 cm H₂O was maintained.

Functional Parameters

Hemodynamic and respiratory data were recorded during reperfusion in the EC group as described previously (18). Blood gases were determined at intervals of 10 min during the reperfusion period. The partial pressure of oxygen (PO₂) of the deoxygenated perfusate of the preload pool was defined as venous PO₂ (Pv_O2) whereas the PO₂ of the perfusate collected from the left atrium was defined as arterial PO₂ (Pa_O2). The perfusate oxygenation (PO₂), defined as the arteriovenous oxygenation difference (Pa_O2  Pv_O2), was used to assess the lung capability for gas exchange. Peak inspiratory pressure (PI_max) was registered every 10 min by means of the small-animal respirator used for mechanical ventilation. After 40 min of reperfusion, the right lung was used to determine the wet weight-to-dry weight ratio (W/D ratio) while the left lung was fixed for ultrastructural analysis.

Fixation

Fixation of the left lungs was performed by vascular perfusion via the pulmonary artery at a hydrostatic pressure of 15 cm H₂O for about 2 min. During the fixation process, the airway pressure was adjusted to 12 cm H₂O. The primary fixative consisted of a mixture of 1.5% glutaraldehyde and 1.5% paraformaldehyde in 0.1 M cacodylate buffer (pH 7.35; buffer osmolality, 300 mOsm/kg). At the end of perfusion, the left main bronchus and pulmonary artery were tightly clamped and the organ was stored in cold fixative for 18-24 h before further processing.

Sampling

Systematic uniform random sampling performed at each sampling step yields samples that are representative of the whole organ and thus can be analyzed by means of stereological methods (see Reference 19). Therefore, systematic uniform random samples were collected according to standard methods as described (20). Briefly, each lung was embedded in 2% agar dissolved in twice-distilled water. By means of a tissue slicer the organ was then cut into horizontal slices (3 mm thick), starting with an apical position that was chosen at random. The organ slices were placed in a processing tray with the apical section face up. A transparent point grid with 11 × 11 points (distance between points, 10 mm) was then superimposed over the collection of slices. Whenever a point hit the cut surface of a lung slice, a 3-mm³ tissue block was excised, defining the point as the lower left corner of the tissue block. By this method 5-11 blocks were obtained from each single lung.

Tissue Processing

Processing of tissue blocks was performed as described (20). Standardization of the protocol was guaranteed by use of an automatic tissue processor (Histomat; Bio-Med, Theres, Germany). Briefly, after six rinses over a 30-min period in 0.1 M cacodylate buffer, tissue blocks were osmicated in 1% OsO₄ in 0.1 M cacodylate buffer for 2 h, washed again in 0.1 M cacodylate buffer (four rinses over a 20-min period), rinsed in twice-distilled water (two rinses over a 10-min period), and transferred to half-saturated aqueous uranyl acetate for en bloc staining overnight. All steps were performed at 8° C. After washing in twice-distilled water (six changes over a 30-min period), samples were dehydrated through an ascending series of ethanol, transferred to Araldite via propylene oxide and a 1:1 mixture of propylene oxide- Araldite, embedded in Araldite, and polymerized at 60° C for 3 d. Tissue blocks were allowed to acquire random orientation in the embedding capsules. For histological analysis, semithin sections (0.5 µm thick) from three randomly chosen samples per lung were stained with methylene blue and examined by light microscopy (Leitz Laborlux 11; Leitz, Wetzlar, Germany). For ultrastructural studies, five tissue blocks per lung were chosen at random. From each block, ultrathin sections were cut and counterstained with lead citrate, using an Ultrostainer (Leica, Bensheim, Germany). Qualitative and stereological analysis by transmission electron microscopy was performed using an EM 900 (LEO, Oberkochen, Germany).

Stereological Analysis

Semithin and ultrathin sections were analyzed by established stereological methods (19). At the light microscopic as well as at the electron microscopic level a systematic quadrats subsampling scheme was applied to generate test fields over the whole section, distributed in a systematic random fashion. All parameters were determined on-line by means of point and intersection counting. For light microscopy an eyepiece containing an integration plate (100/25 points) was used at a total magnification of ×1,000. For transmission electron microscopy a coherent multipurpose test system with 192 test points and 96 test lines was applied at a primary magnification of ×3,000, using an image analysis software package (AnalySIS 2.1; Soft Imaging System, Münster, Germany).

The following stereological parameters were estimated: (1) the volume density of intraalveolar edema (ed) related to the gas exchange parenchyma (par), V_Ved/par; (2) the fraction of alveolar epithelium covered with intraalveolar surfactant (epias) related to total alveolar epithelial surface (epi), S_Sepias/epi, using the epithelial basal lamina as a reference marker; (3) the volume density of lamellar bodies (lb) related to type II pneumocytes (pII), V_Vlb/pII, which indicates the intracellular surfactant volume per unit cell volume; (4) the volume densities of intraalveolar surfactant (as) and its subtypes (according to Reference 17) tubular myelin (tm), unilamellar forms (ul), multilamellar forms (ml), and lamellar body-like forms (lbl) related to type II pneumocytes (pII), which indicate intraalveolar subtype volumes per unit cell volume. Owing to its two-dimensional nature, the alveolar lining layer could not be included; (5) the volume densities of each intraalveolar surfactant subtype related to total intraalveolar surfactant as an indicator of relative intraalveolar surfactant composition; and (6) the volume densities of each intraalveolar surfactant subtype in edematous and nonedematous regions related to total intraalveolar surfactant in the respective edematous and nonedematous regions in the EC group. These estimates give an indication of differences in the relative surfactant pool composition in these two regions.

A mean of 202 points falling on intraalveolar surfactant subtypes were counted per individual lung, which is considered to be a sufficient number because the variability among measurements contributes only to a small extent to the total observed experimental variability, which is dominated by the biological variability between the individuals under study (19).

Statistics

Mean values are given ± SEM. Repeated-measures analysis of variance (ANOVA) was performed for statistical analysis of functional data recorded after 10 min and after 40 min of reperfusion in the EC group. Stereological data were tested for significant differences between groups by the Mann-Whitney rank-sum test. All statistical analyses were performed using the software program SigmaStat 2.0 (Jandel Scientific, Erkrath, Germany). Values of p < 0.05 were considered to be significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Functional Parameters

Hemodynamic and respiratory data recorded after 10 min and 40 min of reperfusion in the EC group are given in Table 1. The W/D ratio of lungs subjected to I/R ranged between 8.5 and 11.1 (mean ± SEM; 9.8 ± 0.4), whereas nonischemic lungs perfused for 40 min with the same buffer solution had a normal W/D ratio (5.1 ± 0.2) (18). The perfusate oxygenation PO₂ decreased slightly during the reperfusion period, although not at a significant level. However, whereas tidal volume and PEEP were kept constant, peak inspiratory pressure increased significantly (10 min, 23.4 ± 2.3 cm H₂O; 40 min, 27.3 ± 2.2 cm H₂O; p < 0.05).

Qualitative Findings

The lungs of the two experimental groups showed clear differences at the electron microscopic level (Figure 1). In the control group, the cells forming the alveolar septa showed their normal ultrastructure. The alveolar lumena were almost completely free of edema fluid. The intraalveolar surfactant subtypes exhibited their typical ultrastructural appearance. Freshly secreted surfactant material could be found as lamellar body-like forms close to the alveolar epithelium as well as to tubular myelin. Tubular myelin showed its typical lattice-like arrangement and was usually found in close contact with the alveolar lining layer. The alveolar lining layer was closely apposed to the alveolar epithelium (Figures 1a and 1b). Small unilamellar as well as multilamellar forms were usually seen in groups close to the alveolar epithelium.

View larger version (134K): [in this window] [in a new window]   Figure 1.   TEM micrographs of type II pneumocytes and intraalveolar surfactant subtypes in (a and b) control and (c-f  ) EC groups. (a) Normal ultrastructure in untreated lung fixed by vascular perfusion immediately after excision. Intracellular surfactant stored in lamellar bodies (lb) of type II pneumocyte as well as intraalveolar tubular myelin (tm) are shown. Scale bar = 2 µm. (b) Tubular myelin (tm) in close contact with the alveolar lining layer (open arrows) as well as to the alveolar epithelium. lbl = lamellar body-like form. Scale bar = 0.5 µm. (c) After perfusion with modified Euro-Collins solution, 2 h of ischemic storage at 4° C, and subsequent reperfusion for 40 min, focal intraalveolar edema (ed) developed. Partial swelling of type I pneumocyte (pI). Type II pneumocyte with slightly swollen mitochondria and intact lamellar bodies (lb). Intraalveolar surfactant (lamellar body-like, multilamellar, and unilamellar forms) within edema fluid. Scale bar = 2 µm. (d ) Alveolar septum of EC lung with partial swelling (large arrow) and fragmentation (small arrows) of type I pneumocytes. Alveolar lumena filled with edema fluid (ed). Unilamellar and multilamellar surfactant forms. Scale bar = 2 µm. (e) In the EC group, tubular myelin within intraalveolar edema fluid (ed) appears to be disintegrated. pI = swollen type I pneumocyte. Scale bar = 0.5 µm. (f   ) Disintegrated tubular myelin was also found in nonedematous regions of EC lungs. pI = swollen type I pneumocyte; en = capillary endothelium; er = red blood cell. Scale bar = 0.5 µm.

After perfusion with modified Euro-Collins solution, 2 h of ischemic storage at 4° C, and subsequent reperfusion for 40 min, the lungs of the EC group had developed intraalveolar edema to varying degrees. In severely affected areas, intraalveolar edema was accompanied by extensive hemorrhage, and extended swelling or fragmentation of type I pneumocytes or even denuded basal lamina. In comparison with the type I pneumocytes, the type II pneumocytes appeared to be structurally well preserved in the EC group. Slight mitochondrial swelling as well as slight dilatations of the cisternae of the endoplasmic reticulum and the perinuclear space could occasionally be observed. No ultrastructural changes of lamellar bodies or nuclei were seen. In the EC group, intraalveolar edema accumulations were characterized by the presence of flocculent, electron-scattering material. Owing to the occurrence of intraalveolar edema, the alveolar lining layer was widely separated from the alveolar epithelium (Figures 1c and 1d). Tubular myelin structures often had no contact with the alveolar lining layer and appeared to be in the process of disintegration, occasionally forming whorllike arrangements. Disintegrated tubular myelin was present in edematous as well as in nonedematous regions (Figures 1e and 1f). Unilamellar forms were more prominent and sometimes larger than in the control group.

Stereological Findings

Stereological data are summarized in Tables 2345. By means of light microscopic stereology, intraalveolar edema seen after I/R in the EC group occupied 36 ± 6% (mean ± SEM) of the gas exchange region as compared with 1 ± 1% in the control lungs (p = 0.008) (Table 2). The volume density of lamellar bodies related to type II pneumocytes did not change significantly (10 ± 2% in the control group versus 8 ± 2% in the EC group) (Table 2). The relative surface fraction of alveolar epithelium covered with intraalveolar surfactant decreased significantly (control, 13 ± 2%; EC, 4 ± 1%; p = 0.008) (Table 2). However, this does not necessarily indicate a decrease in the total amount of intraalveolar surfactant, because in the EC group the surfactant material was more often found to be clumped together whereas it was spread over the alveolar epithelium in the control group. The volume densities of intraalveolar surfactant and its subtypes related to type II pneumocytes, which indicate subtype volumes per unit cell volume, showed a slight decrease in the mean values of tubular myelin (2 ± 1 versus 4 ± 1%) and an increase in the mean values of unilamellar forms (55 ± 18 versus 20 ± 4%), although these changes were not significant between groups (Table 3). Prominent changes were seen in relative intraalveolar surfactant composition, which showed an overall decrease in tubular myelin, representing surface active forms (3 ± 1 versus 12 ± 0%; p = 0.008) and an increase in unilamellar forms, representing inactive forms (83 ± 2 versus 64 ± 5%; p = 0.008) in the EC group (Table 4). Analysis of relative intraalveolar surfactant composition revealed that the decrease in tubular myelin and the increase in unilamellar forms occurred in edematous (tubular myelin, 3 ± 1%; unilamellar forms, 88 ± 6%) as well as in nonedematous regions (tubular myelin, 4 ± 3%; unilamellar forms, 77 ± 5%) in the EC group (Table 5). Although the shift toward unilamellar forms was slightly more pronounced in edematous regions, these differences were not significant. No significant changes in multilamellar or lamellar body-like forms were found between the two groups.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The importance of alterations in the pulmonary surfactant system in I/R injury associated with lung transplantation has come into focus (5). Surfactant alterations have been reported to be associated with experimental I/R (9). An impaired surfactant activity has also been shown in human lung transplant recipients (21). Because an optimal lung preservation technique requires a procedure that does not alter the surfactant system (5), the importance of the integrity of surfactant-producing type II pneumocytes has been pointed out (22). Successful exogenous surfactant therapy of donors and recipients in experimental lung transplantation (23, 24) further supports the hypothesis that the maintainance of a functional surfactant system before, during, and after transplantation is of crucial importance for the postoperative outcome. To preserve the intraalveolar surfactant as well as intraalveolar edema fluid for transmission electron microscopy investigations, fixation "from behind," i.e., by vascular perfusion, must be performed (25). Together with phospholipid stabilization by prolonged block staining with uranyl acetate (20), this procedure ensures that surfactant is fixed and retained in situ. Because structural alterations seen after I/R are heterogeneously distributed over the whole lung, histological studies based on biopsy samples are considered to be of limited value as compared with other methods for the assessment of donor lung quality (26). However, the application of unbiased and efficient stereological methods based on systematic uniform random samples that are representative of the whole organ is a prerequisite for the reliable quantification of the ultrastructural findings and, when properly used, yields accurate results with low variance due to stereological methods (19).

This study was performed to test whether the impairment of pulmonary surfactant observed after lung transplantation is a consequence or a cause of intraalveolar edema formation. In our experimental setting, the cumulative effects induced by the whole sequence of transplantation-related events on surfactant ultrastructure and pulmonary function were investigated. Therefore, rat lungs fixed either immediately after excision or after flush perfusion with modified Euro-Collins solution, ischemic storage for 2 h at 4° C, and subsequent reperfusion for 40 min were investigated by means of transmission electron microscopy and stereology. The formation of intraalveolar edema associated with altered intraalveolar surfactant pool compositions observed in the EC group can be described as I/R injury (27). These ultrastructural alterations were associated with a significant increase in PI_max during reperfusion, which, a constant tidal volume and a constant PEEP given, indicates an altered lung compliance. Because we did not observe any structural changes at the level of the airways, we may assume that the increase in PI_max largely reflects an alteration in pulmonary micromechanics resulting from the disturbance of intraalveolar surfactant. Thus, the changes in intraalveolar surfactant subtypes observed by means of transmission electron microscopy and stereology were associated with significant changes in lung function.

The decrease in surface active tubular myelin and the increase in inactive unilamellar forms after I/R observed in this study by means of transmission electron microscopy and confirmed by stereology is in line with the results of Veldhuizen and co-workers (9), who reported an increase in the SA/LA ratio of BAL subfractions in I/R injury. In addition to BAL analysis, which does not allow distinction between surfactant material from edematous regions and regions free of edema, our results show that alterations in intraalveolar surfactant subtypes occur in edematous as well as in nonedematous regions within the same lung. A decrease in surface active tubular myelin and an increase in inactive unilamellar forms were found in edematous as well as in nonedematous regions in the EC group, as revealed by analysis of relative surfactant composition in the respective areas. Therefore, we conclude that surfactant alterations seen after I/R are not dependent on the presence of edema fluid in the alveoli. Because it is not possible to investigate the same alveolus during the time course of reperfusion by means of transmission electron microscopy, we cannot conclusively rule out the possibility that edema fluid was previously present in edema-free regions at an earlier time point during reperfusion and was resolved before fixation. However, the close correlation between the volume fraction of intraalveolar edema within the gas exchange region and PI_max that has been demonstrated previously (28) and the increase in PI_max during reperfusion in our study indicate that I/R injury developed and edema fluid did not resolve.

Alterations in the surfactant system seen after I/R injury have been described as secondary effects due to intraalveolar edema formation (12, 27). In contrast, our results support the hypothesis that I/R has a primary effect on the surfactant system. Disturbances in the intraalveolar surfactant system after I/R are, therefore, not only the result of inactivation due to plasma protein leakage but may also be responsible for an increased permeability of the blood-air barrier, contributing to intraalveolar edema formation and resulting in a vicious cycle with progressing surfactant impairment. This hypothesis is further supported by the experimental observation that surfactant alterations in lipopolysaccharide (LPS)-induced lung injury occur before the onset of edema formation (17). In human donor lungs that developed reperfusion injury, alterations in the lamellar bodies in type II pneumocytes were also present before reperfusion (29). Surfactant alterations result in an increased alveolar surface tension and thus in an increased pressure gradient across the alveoli. On the basis of this pressure gradient, fluid fluxes across the blood-air barrier will result in intraalveolar edema formation (5, 8). Intraalveolar surfactant, therefore, prevents not only atelectasis but also intraalveolar edema formation. It seems likely that the administration of exogenous surfactant to the lung donor before I/R, which has been reported to result in improved postoperative organ function (5, 23), also has this dual protective effect.

The question arises as to whether surfactant alterations after I/R are caused by lung preservation and ischemic storage alone or whether surfactant disturbances occur only after reperfusion (5, 30). Although it has been shown to be possible to preserve the surfactant system during pulmonary artery flushing and subsequent storage under experimental conditions (9, 10), surfactant dysfunction has also been reported to be present after preservation of rat lungs with modified EC solution alone (11). When comparing these data, species differences as well as differences in the experimental conditions must be taken into account. Flush perfusion with modified EC solution, as performed by Andrade and coworkers (11) and in our study, might not be as effective for rat lungs as for other species (5, 30, 31). Thus, the experimental I/R injury induced in our study might have been intensified owing to the choice of species and the preservation solution used. Additional factors must be considered in cases of clinical lung transplantation, where the medical history of the donor may also contribute to surfactant alterations seen before reperfusion (29). Therefore, further studies using different animal models and preservation conditions as well as clinical studies to analyze surfactant structure and function after arterial flushing, after ischemic storage, and after reperfusion are desirable.

The decrease in surface active tubular myelin and the increase in unilamellar vesicles reported in this study, which is comparable to an increased SA/LA ratio, may be caused by a decreased synthesis and/or secretion of surfactant by type II pneumocytes, an increased conversion of tubular myelin into unilamellar vesicles, or a decreased reuptake of unilamellar vesicles by type II pneumocytes. The surfactant apoproteins SP-A and SP-B are important for the maintenance of the integrity of tubular myelin and for the reformation of tubular myelin from surface-lining material when the alveolar surface area decreases (5). Decreased levels of SP-A, which have been reported in ARDS (13, 32) as well as after I/R (9, 12), contribute to an increased formation of unilamellar vesicles (33). Correspondingly, high SP-A levels help prevent surfactant inactivation caused by protease activity (34, 35). Therefore, the question concerning whether I/R directly affects type II pneumocytes, resulting in SP-A imbalance, awaits further examination.

In conclusion, the present study has shown, by means of transmission electron microscopy and stereology, that surfactant alterations in I/R injury are not a secondary consequence due to inactivation by intraalveolar edema fluid, but rather a primary effect that contributes to further edema formation, resulting in a vicious cycle with progressing surfactant impairment. Future studies should focus on the direct effects of I/R on type II pneumocytes.