Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Mortality from the acute respiratory distress syndrome is reported to be ~ 50% and has not changed substantially since the syndrome was first described in 1967 (1, 2). This mortality rate has led to the development of numerous innovative treatment and support modalities throughout the last two decades for patients suffering from various types of acute respiratory failure (3). A recent development in the treatment of respiratory failure has been the use of liquid-assisted ventilation techniques (6). These techniques are based on the observation that diseased alveoli are more easily inflated with liquid than air because of a lower surface tension at the liquid-epithelial interface compared with the gas-epithelial interface (9, 10). Perfluorocarbons have been extensively studied for this purpose because of their high spreading coefficient and low surface tension as well as their high oxygen and carbon dioxide solubility (10, 11). Two types of liquid ventilation have been studied, total liquid ventilation and partial liquid ventilation (PLV), in which the lung is filled with perfluorocarbons approximating functional residual capacity and ventilated using a conventional ventilator with gas tidal volumes (5, 10, 11).

PLV has been studied in healthy animals as well as in experimental animals with lung injury, and recently the first clinical studies were reported in premature infants and adults (7, 8, 12). These studies have focused on demonstrating the effectiveness of PLV in improving the gas exchange and lung mechanics in animals and humans (13, 14). Animal studies have also shown qualitative improvement of lung histology (15), and recently semiquantitative histologic studies have confirmed these observations (16). Several in vitro studies have suggested that perfluorocarbons also influence the inflammatory response, which could contribute to the improvement of gas exchange and histologic changes observed in these studies (17, 18). These observations were supported by studies showing reduced polymorphonuclear leukocytes (PMN) accumulation in the injured lung of animals treated with perfluorocarbons during mechanical ventilation (19, 20).

This study was designed to quantitate lung injury and the inflammatory response in the lung in animals treated with either conventional mechanical ventilation (CMV) or PLV following induced acute lung injury at a light microscopic level using morphometric techniques. Fluorescence microscopy was used to quantitate alveolar epithelial cell damage, and transmission electron microscopy was used to examine ultrastructural damage to alveolar walls.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Animals

Adult New Zealand white rabbits (n = 17, weighing 2.5 to 3.5 kg) were used in this study. All experiments were approved by the Animal Experimentation Committee of the University of British Columbia.

Experimental Protocol

We studied three groups of animals after the induction of acute injury: (1) CMV with gas (n = 6); (2) CMV combined with PLV (n = 5); and (3) control group with no ventilation (control, n = 3). For the fluorescence and electron microscopy studies, the lungs of a group of normal rabbits (normal, n = 3) with no intervention (no acute lung injury or ventilation) were evaluated. Rabbits were anesthetized with ketamine (35 mg/kg intramuscularly) and xylazine (1.5 mg/kg intramuscularly). The spontaneously breathing animals were pre-oxygenated using a nose cone with 100% oxygen, and anesthesia was deepened using halothane 1 to 3%. A venous catheter was placed in the marginal ear vein to facilitate analgesia and sedation as a continuous infusion of ketamine (20 µg/kg/min) and xylazine (1 mg/kg/h) was supplied. The anterior neck was dissected and a tracheostomy was performed. A cuffed endotracheal tube (3.5 mm internal diameter; Sheridan Catheter Corp., Argyle, NY) was placed in the trachea and secured. A vascular catheter was inserted into the left carotid artery and the right internal jugular vein, and central venous and arterial pressures were continuously displayed on a monitor bracket (Hewlett-Packard Omni-Care, Andover, MA). Normothermia was maintained with overhead heaters and electric warming pads. Maintenance fluids were provided by continuous infusion of 0.9% saline solution containing 5% dextrose. A dopamine infusion (5 µg/kg/min) was commenced to minimize the hemodynamic instability that we have experienced during the lung lavage. Sodium bicarbonate was given as required to maintain arterial pH > 7.25.

  Mechanical ventilation was initiated (Servo 900C; Siemens- Elema, Solna, Sweden) after cannulation. Muscle paralysis was induced by intravenous administration of pancuronium bromide (0.2 mg/kg) with intermittent boluses of 0.1 mg/kg as required to control animal movement. The initial ventilator settings, using the volume control mode, were tidal volumes of 10 to 12 ml/kg as measured by a pneumotachygraph (NVM-1; Bear Medical Systems, Riverside, CA) placed between the endotracheal tube and the ventilator tubing (to exclude compliance of tubing as a variable in this measurement). Positive end-expiratory pressure (PEEP) was set at 5 cm H₂O, with a fraction of inspired oxygen (FI_O2) of 0.5, and a respiratory rate to achieve a Pa_CO2 value within the normal range (35 to 45 mm Hg [4.7 to 6.0 KPA]).

Acute Lung Injury

Lung injury was induced with a normal saline whole lung lavage procedure as previously described (21). A syringe was used to instill warmed (37°C) saline (20 ml/kg) into the lung via the endotracheal tube. Saline was allowed to remain in the lung for 10 to 15 min, then removed, and a new bolus instilled until adequate lung injury was evident. This was reached when the P/F ratio (Pa_O2/FI_O2) was < 100 on FI_O2 of 1.0 and PEEP of 6 cm H₂O. This procedure took 30 to 45 min and the animals were then randomized to one of the three treatment groups.

Ventilation of Animals

Ventilation management was based on current clinical practice and recommendations. After the lung injury was established, the animals assigned to CMV were ventilated with gas in the pressure control mode with PEEP set at 6 cm H₂O and peak pressures adjusted to maintain an effective tidal volume of 10 ml/kg. The respiratory rate was set at 35 breaths/min and adjusted to maintain Pa_CO2 in the normal range (35 to 45 mm Hg). An increase in the FI_O2 was used to achieve and maintain Sa_O2 > 85%. PEEP was maintained at 6 cm H₂O throughout the experiment for all groups. For the animals randomized to PLV, perfluorocarbon (Liquivent; Alliance Pharmaceutical Corp., San Diego, CA) was instilled into the lungs through a side port on the endotracheal tube connector in an amount that approximated their functional residual capacity (20 to 25 ml/kg). Instillation was complete when the meniscus of the fluid column in the endotracheal tube was level with the anterior chest wall at an end-expiratory pressure of zero. During the process, the animals were ventilated with 100% oxygen. After the instillation of the perfluorcarbon, mechanical ventilation was applied to the liquid-filled lung using the pressure control mode, PEEP of 6 cm H₂O, an effective tidal volume of 10 ml/kg, a respiratory rate of 35 breaths/min, and inspiratory to expiratory ratio of 1:1. The respiratory rate was adjusted to maintain normocarbia and the FI_O2 was adjusted to maintain Sa_O2  85%. The meniscus of fluid was checked intermittently and losses replaced as required. Animals that were randomized to the control group were killed after the saline lavage.

Measurements

Cardiorespiratory physiologic data were measured continuously and recorded at baseline, 30 min after the injury was established, and hourly thereafter. Ventilator settings at these points were recorded using the digital displays on the ventilator. Arterial blood samples for blood gas analysis (ABL 330; Radiometer, Copenhagen, Denmark) were collected at baseline, after the lung injury, and then hourly until animals were killed. Alveolar-arterial oxygen gradient [(P_b P_H2O × FI_O2)  (Pa_CO2/0.8)  Pa_O2] and the oxygenation index [(100 × mean airway pressure [] × FI_O2)/Pa_O2] were calculated and mean airway pressure recorded using the digital display on the ventilator.

Processing of Lung Tissue

Light microscopy and morphometry. Animals were killed with an overdose of pentobarbital, the chest was opened, and the heart was ligated immediately to trap the blood in the lung vessels. The lungs were excised and the left lung was fixed for 4 h by inflating it with 10% buffered formalin at 25 cm H₂O. Lungs were weighed and volumes were determined by water displacement. Fixed lungs were cut into 2-cm thick slices perpendicular (90°) to the plane of gravity, and eight random samples were obtained from these slices in the gravitational dependent, middle, and nondependent regions of the lung. Tissue was embedded in paraffin and cut in 3-µm thick tissue sections, placed on slides coated with 3-aminopropyl-triethoxysilane (Sigma Chemical Co., St. Louis, MO), and baked for 16 h at 37°C. The paraffin was removed in two 10-min washes with xylene, sections were rehydrated in graded ethanol from 100 to 70%, rinsed twice with distilled water, and stained with hematoxylin and eosin.

Morphometric analysis. The inflammatory response in the lung was quantitated by evaluating  100 alveoli in each section in computer-generated random fields of view at ×800 magnification on a Nikon Microphot-fx light microscope (Nikon, Tokyo, Japan). Random fields that consisted largely of lung parenchyma (alveoli and small vessels) were evaluated. Three sections in the gravitational dependent region and three in the gravitational nondependent region were evaluated. To evaluate  100 alveoli, five to 10 random fields of view were evaluated per section. The number of PMN, red blood cells (RBC), and alveolar macrophages in alveoli were counted, as well as the number of alveoli with edema or hyaline membranes in the CMV (n = 6), PLV (n = 5), and control (n = 3) groups. All values were expressed as the number/alveoli of either the dependent or independent region (mean of 10 fields of view in each region).

  To enumerate the total number of PMN in the lung, we used a sequential level stereologic analysis described by Cruz-Orive and Weibel (22) in three animals of each group. A point counting grid was placed over the cut surface of the lung slices that were examined at ×4 magnification. The number of points falling on lung parenchyma, large blood vessels, and airways (> 1 mm diameter) were determined. The volume fraction (Vv) of each component was estimated as follows: Vv (object) = sum of points on object  Sum of total points.

  Paraffin-embedded sections were then point counted at ×800 magnification using a Nikon Microphot-fx light microscope (Nikon) and using an image analysis system (Bioview; Infrascan Inc., Richmond, BC, Canada). To determine the Vv of the different lung components, computer-generated random fields of view were evaluated (20/animal). Because acute lung injury is largely a lung parenchymal disease, mostly peripheral lung tissue was sampled by skipping random fields of view that contain mostly larger central airways or blood vessels. This ensures that the 20 random fields evaluated per animal are therefore mostly peripheral lung tissue. Twenty fields of view were selected because our calculated coefficient of error (CE) showed that evaluating more than this number of fields per animal does not decrease the CE or add to the data. The number of points falling on airspace, tissue, and PMN in the airspace or tissue were counted and the Vv of each component was estimated. Tissue in this context consisted of alveolar walls and included capillaries and interstitial spaces. The identity of slides was masked and evaluated with the observer unaware of the group from which it came.

  The number of PMN in the airspace or alveolar walls was calculated. The calculation for the number of PMN in the airspace is shown as an example:

where 143 fl is the assumed volume of a rabbit PMN (23). Results are expressed as the total number of PMN/ml of airspace or tissue in the different lung regions.

Fluorescence Microscopy

The right lung was inflated with OCT (diluted 50% in distilled water), an optimal cutting temperature embedding medium (Miles, Elkhart, IN), and stored at 70°C. To examine the alveolar epithelial cells, the lung was divided in three different gravitational regions. For each region, serial cryosections (~ 5,000 nm) were obtained on a cryostat (Frigocut 2800 N; Leica; Nussloch, Germany) and mounted on 3-aminopropyltriethoxysilane-coated slides. After fixing (15 min) with 1% paraformaldehyde, specimens were washed (10 min) with phosphate-buffered saline (pH 7.3) and then incubated (30 min at room temperature) with the fluorescein isothiocyanate (FITC)-conjugated lectin (150 µg/ml) extracted from Maclura pomifera (Sigma), PBS, or fluoroscein alone (150 µg/ml) as a control. The lectin extracted from M. pomifera recognizes carbohydrate moieties of the alveolar epithelium. Previous immunocytochemical studies demonstrated the specificity of this lectin toward a membrane glycoprotein on the surface of type II cells with little to no binding to type I cells (24, 25). On the basis of these studies, this lectin binds to a single membrane glycoprotein of the pulmonary alveolar epithelium and is useful to identify and characterize type II cells. Pilot studies were done to establish the optimal conditions for this labeling procedure. In these studies we have shown that low concentration of paraformaldehyde (1%) and a relative short period of fixation are crucial to minimize the background fluorescence induced by the fixative alone. Furthermore, FITC alone does not differentially bind to type II or type I pneumocytes but gives a dimmer diffuse background binding. Cell morphology and topographic location were used to assist in the identification of type II cells, and all cells present in the alveoli space were excluded from the analysis. Furthermore, the pilot study also established the lectin concentration (150 µg/ml). This concentration gives little to no background labeling while staining type II cells and not flattened elongated type I cells. After rinsing, specimens were mounted in 1,4-diazabicyclo (2.2.2) octane (DABCO) antifading solution. Photographs were taken on a Zeiss fluorescence microscope (Oberkochen, Germany) at ×800 magnification.

Quantitative Cytochemical Analysis

We used fluorescence microscopy to examine FITC-M. pomifera-labeled type II cells and enumerated these cells in lung sections from each animal group. Cells dispersed in the alveolar lumen were excluded from the analysis. Type II-labeled cells were counted on 30 randomly selected fields per section (one section per animal) using ×125 magnification. To determine the perimeter and the labeling intensity of type II cells, fluorescent images (10 images per section and three sections per animal) were captured in computer-generated random fields of view using image analysis software provided by the National Institutes of Health (NIH Image). Results were expressed as the number of labeled type II cells/field, the average perimeter of type II cells (µm), and the relative mean density of these cells. Images were evaluated with the observer unaware of the group being analyzed.

Transmission Electron Microscopy

In two animals per group, tissues were collected and processed from transmission electron microscopy (TEM) studies. A lobe from the left lung was inflated (pressure, ~ 25 cm H₂O) and immersion-fixed (1 h) with 2.5% glutaraldehyde using 0.1 M Na cacodylate buffer, pH 7.3. Fixed lung was divided into five slices perpendicular to the gravitational field. Lung tissue samples (about 2 mm³) were taken from three different locations of both the dependent and nondependent regions of the lung, further fixed for 1 h with 2.5% glutaraldehyde, washed in a buffer (1 h), and postfixed (1 h) with 1% osmium tetroxide (OsO₄). Fixed samples were rinsed, dehydrated in a graded ethanol series, and embedded in LR white. Sections used for TEM were obtained as follows: 1 section per grid × 5 grids per block × 3 blocks per region × 2 regions per animal × 2 animals per group × 4 groups of animals. Sections were cut on a CR 2000 RMC microtome (Wein, Austria). Ultrathin sections (~ 100 nm) were mounted on 200-mesh copper grids, stained with uranyl and lead citrate, and examined on a Philips transmission electron microscope 400 (Eindhoven, The Netherlands). Grids were evaluated with the observer unaware of the group being analyzed.

Statistical Analysis

Results are expressed as mean ± standard error (SE). To analyze the morphometric data, a randomized block analysis of variance (ANOVA) was used to compare the number of PMN, RBC, alveolar macrophages, and edema/alveoli between groups. This analysis takes in account the differences within an animal (20 fields of view) and between animals and groups. As reference value, historic control animals (normal animals of the same weight and age with no intervention) were used (26) (Table 1). Oxygenation and lung mechanics over time were compared between groups using a two-way ANOVA for repeated measures. Bonferroni's correction was used for multiple comparisons. The square root of the data was used to normalize the data for type I cell mean fluorescence intensity. P < 0.05 was accepted as statistically significant.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Oxygenation and Lung Mechanics

Blood gas, lung mechanics, and hemodynamic status did not differ between the groups before the lung injury. The animals needed 6 ± 4 saline washouts before the target respiratory failure abnormalities were reached. The amount of perfluorocarbon initially needed to fill the lungs was 20 ± 5 ml, and the amount needed for replacement periodically throughout the experiment was 2 ± 0.5 ml/kg/h. Peak airway pressures ranged from 15 to 25 cm H₂O and was similar in both groups. Table 2 shows the arterial-alveolar oxygen gradient (Aa_DO2), the oxygen index (OI), and the in the different groups before and after the injury was induced as well as over the study period. Oxygenation improves significantly after PLV with little or no change in the CMV-gas group. FI_O2, respiratory rate, peak inspiratory pressures, and did not differ between the groups.

Lung Injury

Because of the presence of perfluorocarbon in the PLV group, the wet/dry ratio of lung tissue could not be used as a parameter of acute lung injury. Figure 1 shows a representative view of tissue taken from the dependent lung region of the control (Figure 1A), CMV- (Figures 1B and 1C), and PLV- (Figure 1D) treated animals. At light microscopic level, acute lung injury was quantitated as alveolar hemorrhage and edema/hyaline membrane formation. The number of RBC per alveoli was higher in the CMV (n = 6) and PLV (n = 5) groups compared with the control group (n = 3) (Figure 2A). The percentage of alveoli with edema or hyaline membranes (Figure 2B) was higher in the CMV (n = 6) and PLV (n = 5) groups compared with the control (n = 3) group. In the CMV group, there was more alveolar hemorrhage and edema compared with the PLV group (P < 0.05, Figures 2A and 2B, respectively). There was no difference in alveolar hemorrhage (5.1 ± 1 versus 5.4 ± 0.9 RBC/alveolus) or edema (25.2 ± 3.2 versus 22.7 ± 4.1% alveoli with edema) in the gravitational dependent versus nondependent lung regions in the CMV group. Similarly, there was no difference in alveolar hemorrhage (2.6 ± 0.3 versus 2.1 ± 0.6 RBC/alveolus) or edema (12 ± 2 versus 10.9 ± 3.1% alveoli with edema) in the gravitational dependent versus nondependent lung regions in the PLV group.

View larger version (80K): [in this window] [in a new window]   Figure 1.   Photomicrographs of light microscopic appearance of peripheral lung tissue in the control (A), CMV (B and C), and PLV (D) groups at ×400 magnification. Sections from formalin-fixed, paraffin-embedded lung tissues were stained with hematoxylin and eosin. In the control group (A), there are congested alveolar walls with leukocyte sequestration (arrowhead) but no intra-alveolar PMN or red cells. Note the alveolar edema (open arrow) and hyaline membranes (closed arrow) with alveolar wall disruption (B and C) in the CMV group. Numerous PMN (closed arrowheads) and RBC (open arrowheads) are present in the alveolar spaces. In the PLV group (D), PMN and RBC are present in airspaces, but the integrity of alveolar walls is better preserved.

View larger version (7K): [in this window] [in a new window]   View larger version (8K): [in this window] [in a new window]   Figure 2.   Alveolar damage as measured by alveolar hemorrhage (A) and alveolar edema/hyaline membrane formation (B). Alveolar hemorrhage is expressed as the number of RBC/alveolus and was significantly higher in the PLV (n = 5) and CMV (n = 6) groups compared with the control group (n = 3) (* P < 0.01), and the CMV group was higher than the PLV group (# P < 0.05). Alveolar edema/hyaline membrane formation is expressed as the percentage of alveoli that contains either edema fluid or hyaline membranes. This percentage was higher in the PLV and CMV groups compared with the control group (* P < 0.01), and the CMV group was higher than the PLV group (# P < 0.05).

Leukocytes in the Lung

In the nondependent regions of the lung, the number of PMN in the alveolar walls (sequestration) and alveolar space (migration) was similar between the CMV (n = 5) and PLV (n = 5) groups (Table 1). However, in the dependent lung regions, there were more PMN in the alveolar walls and alveolar space in the CMV group than in the PLV or control groups (P < 0.05). Sequestration of PMN in alveolar walls in all groups was approximately five to 10 times higher than in normal, noninjured animals, using historic data (26). Interestingly, there were fewer alveolar macrophages/alveoli in the PLV group than in the CMV or control groups (P < 0.05) (Figure 3).

View larger version (9K): [in this window] [in a new window]   Figure 3.   Alveolar macrophages in the airspaces expressed as the number of macrophages per alveolus. Note the reduced number of macrophages in the PLV (n = 5) group (*P < 0.05) compared with either control (n = 3) or CMV (n = 6) groups.

Fluorescence Microscopy

Cytochemical studies show paraformaldehyde (PFA) and fluorescein alone cause diffuse dim background fluorescence and the lectin extracted from M. pomifera labels type II and not type I cells. Figure 4A represents a lung section from a normal, nontreated rabbit and shows that type II cells (arrowheads) display a cuboidal shape and cover a small portion of the alveolar surface, whereas type I cells are flattened and cover most of the alveolar surface. Figure 4B represents a lung section from a control (saline-treated) rabbit and shows that type II-labeled cells are smaller than those from the nontreated lungs. As illustrated in Figure 4C, type II-labeled cells from the CMV group are fewer and appear less intensely labeled than type II cells from the PLV group (Figure 4D).

View larger version (21K): [in this window] [in a new window]   Figure 4.   Fluorescence microscopy of peripheral lung tissue of normal (A), control rabbits after repeated saline lavage (B), CMV (C), and PLV (D) animals. The arrowheads point to fluorescent-labeled type II alveolar epithelial cells. Note the one to two labeled type II cells per field of view in the normal lung tissue (A) that appeared smaller after the saline lavage (B), the reduced labeling density of type II cells in the lungs of mechanically ventilated animals (C), and the increased labeling density of type II cells in the lungs of animals ventilated with perfluorocarbons (D). Microphotographs represent a ×800 magnification.

Table 3 summarizes data from the quantitative analysis of type II-labeled cells. Alterations in type II cells include reduced cell perimeter after saline injury, a decrease in the number of M. pomifera-positive cells in the CMV group, and an increase in the mean density (brightness) of type II cells in the PLV group (P < 0.05).

Transmission Electron Microscopy

Figure 5 illustrates structural features of alveolar epithelial cells in normal (nontreated) (Figure 5A), control (saline-treated) (Figures 5B and 5C), CMV (Figures 5D and 5E), and PLV (Figure 5F) lungs of rabbits. Figure 5A shows a type II cell in tight interaction with a type I cell and underlying mesenchymal cells (arrowheads). The type II cells display apical microvilli and large cytoplasmic lamellar bodies (L) (Figure 5A). The latter contain profiles of myelin figures, suggesting the presence of surfactant phospholipids and surfactant-related apoproteins.

View larger version (163K): [in this window] [in a new window]   Figure 5.   TEM of lung tissue sections from normal (A), control (B and C), mechanically ventilated (D and E), and PLV (F) rabbits. (A) Type II cells with close interaction with type I cells (arrowheads) and large cytoplasmic lamellar bodies (L) are shown. (B) In saline-treated lungs, type II cells show loose interaction with mesenchymal (open arrowhead) and type I cells (closed arrowhead), reduced microvilli (arrow) with lamellar bodies in the process of exocytosis (L, inset). (C) Type I cells with increased electron density, membrane blebbing, and cytoplasmic vesicles (arrowheads) are shown. (D) A type II cell from a gas-ventilated animal with reduced microvilli (arrow), dilated mitochondrial cristae (open arrowhead), and few lamellar bodies (L) is shown. (E) A type II cell dislodged from its basement membrane (arrow) in the CMV group is shown. (F) Type II cells in liquid ventilated animals with close juxtaposition to mesenchymal and type I cells, distinct microvilli, prominent mitochondria (M), and smaller lamellar bodies (L) than normal cells are shown. Magnification of photomicrographs of A and B are ×8,700, C is ×5,200, D is ×7,400, E is ×2,000, F is ×6,600, and the inset in B is ×10,000.

In saline-treated lungs (Figures 5B and 5C), type II cells show loose interaction with mesenchymal (open arrowhead) and type I cells (closed arrowheads). They show reduced microvilli in the central cap (arrow) with lamellar bodies in the process of exocytosis (L) (Figure 5B, inset). As illustrated in Figure 5C, type I cells display structural features typical of cell damage with increased electron density, membrane blebs, and cytoplasmic vesicles (arrowheads) (Figure 5C).

Morphologic changes in type II cells from gas-ventilated lungs (Figures 5D and 5E) include reduced cell surface microvilli (arrows), dilated mitochondrial cristae (open arrowhead), and few lamellar bodies (L). As shown in Figure 5E, some type II cells were dislodged from their basement membrane (arrow), whereas others appear to be dedifferentiating into type I cells (picture not shown).

In liquid-ventilated lungs (Figure 5F), type II cells are cuboid and in close juxtaposition to mesenchymal and type I cells. Type II cells often appear in groups of two, possibly owing to cell division. Morphologic characteristics of type II cells include distinct microvilli, prominent mitochondria (M), and smaller lamellar bodies (L) than in type II cells of normal lungs. Alveolar epithelial cells in the PLV group show fewer morphologic characteristics of damage than those in the CMV group. These ultrastructural features reported apply to both the dependent and the nondependent regions of the lung as there was no apparent difference between lung regions.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

In this study we used quantitative and qualitative histologic methods to characterize lung damage and the inflammatory response in a model of acute lung injury treated by either CMV or PLV. The light microscopy analysis clearly showed that the use of CMV resulted in more structural damage to alveolar walls with alveolar edema formation and hemorrhage. Quantitative fluorescence microscopy analysis showed damage and a reduced number of type II alveolar epithelial cells in animals treated with CMV. These observations were supported by qualitative ultrastructural studies using electron microscopy. Extensive damage to alveolar epithelium in gas-ventilated animals was observed, whereas less damage was seen in animals treated with PLV. Light microscopic analysis further showed that the structural lung damage results in an inflammatory response characterized by more sequestration and migration of PMN in the dependent regions of gas-ventilated animals. Together these studies show that perfluorocarbon has a protective effect on the structural damage to alveolar walls induced by CMV.

Perflubron, the perfluorocarbon used in this study, has a high oxygen and carbon dioxide solubility, low surface tension, and good spreadability, all qualities that make it an excellent respiratory medium (6, 10, 11, 27). This compound is also considered to be biochemically inert, nontoxic, and compatible with biologic systems. In this study, the use of perflubron combined with CMV caused a sustained improvement in oxygenation over the 4-h study period (Table 2). This finding is consistent with results from similar studies (13, 15, 16, 28, 29). Liquid ventilation improves oxygenation in severe respiratory failure by lowering the intra-alveolar surface tension, thus improving lung compliance, ventilation-perfusion mismatching, and shunt, presumably by alveolar recruitment and lowering of the pulmonary vascular resistance (6, 9). It also allows the use of lower peak pressures because of the lower intra- alveolar surface tension (30, 31) with lower stretch and stress on lung tissues during positive pressure ventilation resulting in less alveolar wall disruption, hemorrhage, and edema formation. In this study we tested this hypothesis by performing quantitative light and fluorescence microscopic studies as well as qualitative electron microscopic studies. The quantitative light microscopy studies support previous qualitative and semiquantitative (15, 16) analyses, demonstrating reduced alveolar hemorrhage, alveolar edema, and hyaline membrane formation (Figures 2A and 2B) with better preservation of alveolar wall integrity (Figure 1).

The inflammatory response induced by the acute lung injury and subsequent mechanical ventilation was evaluated by the influx of PMN into the lung. After acute lung injury, PMN are rapidly recruited into the injured lung, and these PMN have been implicated in the pathogenesis of the acute respiratory distress syndrome (32, 33). PMN-mediated tissue injury is attributed to the ability of these cells to generate oxygen radicals and release various hydrolytic enzymes (34). Colton and colleagues (38) have shown reduced PMN accumulation in the lung with PLV treatment of acute lung injury. Interestingly, in our study, the total number of sequestered and migrated PMN in the lungs was not different between the CMV and PLV groups. However, the number of sequestered and migrated PMN in the dependent lung regions of the CMV group was higher than that of the PLV group (Table 1), suggesting that perflubron preferentially influences the inflammatory response in the dependent regions of the lung. The relatively high density of perfluorocarbons (specific gravity, ~ 1.9 g/ml) tends to distribute the liquid to the dependent lung regions, which may preferentially improve the mechanical properties of these regions, resulting in less tissue damage. Perfluorocarbons have been shown to inhibit alveolar macrophage (17) and neutrophil activation in vitro (18) and reduce the proinflammatory cytokines interleukin (IL)-1 and IL-6 in bronchoalveolar lavage fluids of patients treated with PLV (39), suggesting that perflubron also has a direct anti-inflammatory effect. Moreover, the repeated instillation of perfluorocarbons into the lung could lavage the alveoli and have a dilutional effect on the inflammatory cells and mediators in the alveoli. It is not possible to establish from our data whether the beneficial effect of PLV in our model was due to the anti-inflammatory qualities of perfluorocarbons or due to the stabilization and recruitment of alveoli.

Recent work in the area of prevention of lung injury and improving oxygenation has focused on the concepts of minimizing alveolar stretch and lung recruitment strategies (40, 41). Lung recruitment strategies have included the use of higher levels of PEEP, high frequency oscillation, and PLV. Each of these strategies has been demonstrated to improve oxygenation and reduce lung injury. In our experiment, PEEP was maintained at a fixed level of 6 cm H₂O in order to minimize confounding variables. Clearly, other protective modes of ventilation also reduce lung injury, particularly when used in combination with PLV (42, 43).

The pulmonary alveolar surface is formed by type II and type I cells, and these two cells are structurally and functionally different from each other. Type I cells cover most of the alveolar surface, are terminally differentiated, and provide the surface for gas exchange between air and capillary blood. In contrast, type II cells cover only 7% of the alveolar surface area, can differentiate into type I cells, and play a vital role in the synthesis and secretion of surfactant (9), the synthesis of basement membrane components, and the transport of solutes (22, 23, 44, 45). In this study, a fluorescein-labeled lectin was used to examine the integrity of type II cells. This lectin, M. pomifera agglutin, recognizes complex carbohydrate moieties of alveolar epithelium, and immunocytochemical studies using ferritin- and gold-labeled conjugates of M. pomifera agglutin have demonstrated the specificity of this lectin for a single (230 kD) membrane glycoprotein on the surface of type II alveolar epithelial cells (24, 25, 46). This hydrophilic integral membrane protein has terminal N-acetylgalactosamine residues, and studies suggest that this protein is important in the regulation of surfactant secretion and recycling (24). Our fluorescence microscopy studies show fewer lectin- labeled type II cells in the CMV group (Table 3), suggesting that these cells were either damaged or lost. Electron microscopic studies support the notion that these cells are damaged (Figure 5E) with loose attachments to both type I cells and the underlying basement membrane. Detachment of type II cells at the basal surface severs their connection with interstitial cells, which is thought to be important in promoting surfactant production (9). The electron microscopic studies show fewer surfactant-containing lamellar bodies in type II cells of the CMV group and support the importance of type II cell detachment and damage in their ability to produce surfactant.

Damage to the type II cell surface coat could affect cell function and responsiveness to endogenous and exogenous mediators (47). This damage could result from direct physical or chemical trauma or indirectly from the action of proteinases released by inflammatory cells. Damage to this surface coat may facilitate microbial infection either by changing the charge of the cell surface or by exposing receptors important for microbial adhesion. Varsano and colleagues (48) showed that the gram-negative bacteria Pseudomonas aeruginosae only adhere to injured alveolar epithelial surfaces.

In the PLV group, the type II cells were similar in number and size and showed more lectin binding compared with the normal lungs (Table 3). This increase in lectin binding could result from conformational changes of the molecule or from increased production or redistribution of surface glycoproteins. The electron microscopic studies showed type II cells in the PLV group with prominent mitochondria and Golgi complexes, suggesting active metabolic activity. On the basis of these observations, we speculate that the renewal of alveolar epithelial cell surface glycoproteins during liquid ventilation could protect the injured alveolar epithelial cells from bacterial attachment and subsequent infections.

Type I cells show structural features suggestive of damage with an increase in electron density and membrane blebbing (Figure 5C). Epithelial type I cells are very susceptible to injury and their renewal depends on the survival of type II cells. The latter switch from one differentiated phenotype to another without involving cell division (25). This morphologic transformation of type II into type I cells has been reported with exposure to NO₂ (45). The larger number and less damage to type II cells in the PLV group could accelerate this transformation of type II into type I cells and lung repair. The electron microscopic studies are observational in nature and should be interpreted with caution. However, they provide a basis for further quantitative studies to test the hypothesis that PLV protect the type I and type II pneumocytes from ventilator- induced injury and accelerate the repair of these cells after injury. Furthermore, studies to delineate the mechanisms by which perfluorocarbons influence the inflammatory response and enhance lung repair are clearly necessary.

In this study, we documented the structural damage to alveolar walls that occurred with CMV. The inflammatory response to this damage was mostly neutrophilic in nature and more intense in the dependent regions of the lung. PLV using perfluorocarbons protects the lung from this damage and attenuates the inflammatory response. Interestingly, quantitative fluorescent microscopic and qualitative electron microscopic studies show that type II alveolar epithelial cells, numbers, and functional integrity were preserved in animals treated with perfluorocarbons. Because type II cells are crucial for surfactant production and repair of damaged type I cells, we speculate that a mechanism by which perfluorocarbons improve acute lung injury is by preventing type II cell damage and enhancing their function.