Infiltration of activated neutrophils into the lung appears to be a key element in the severe lung injury that develops in animal models of acute lung injury. Partial liquid ventilation with perflubron has been shown to ameliorate tissue damage compared with conventional mechanical ventilation in acute lung injury models. Pilot experiments indicated that indirect exposure to perflubron could modulate the degree to which subsequent neutrophil binding to endothelial cell monolayers was upregulated after lipopolysaccharide activation. Endothelial cell monolayers preexposed to perflubron showed >40% reductions in the surface steady-state levels of E-selectin and intercellular adhesion molecule-1 achieved after proinflammatory activation (P < 0.05), which correlated with a reduction in the real-time association constants measured by biosensor techniques. These results indicate that direct contact with the perflubron liquid phase is not necessary to attenuate inflammatory responses. Rather, diffusion of perflubron from the alveolar space into the adjacent pulmonary vascular endothelial layer may modulate neutrophil adhesion and thereby reduce the rate of infiltration of activated neutrophils into the injured lung.
- acute lung injury
- partial liquid ventilation
- neutrophil infiltration
- intercellular adhesion molecule-1
acute lung injury (ALI) and acute respiratory distress syndrome (ARDS) are manifestations of acute lung inflammation resulting from a variety of direct and indirect insults to the lung (33, 34). Marked elevations in neutrophil counts in bronchoalveolar lavage fluid have been noted in ARDS patients compared with those in control ventilated or nonventilated patients (2, 13, 33-35, 42). Studies of animal models of ALI have confirmed that the marked increase in the number of activated neutrophils migrating into the lung plays a key role in the lung injury that develops after the initiating lung insult (5, 25).
Partial liquid ventilation (PLV) with perflubron has been shown to improve gas exchange to a greater extent than conventional mechanical ventilation (CMV) in a variety of animal models of ALI (15, 16, 27,39). Histological analysis has demonstrated that this benefit in gas exchange correlates with a reduction in alveolar hemorrhage, vascular leakage, and an influx of activated leukocytes (5, 15, 16). Rats with cobra venom-induced lung injury had a marked reduction in vascular edema accumulation (measured by 125I-albumin influx) as well as an infiltration of neutrophils when ventilated with PLV with perflubron compared with that with CMV treatment (5). In vitro studies with isolated alveolar macrophages have demonstrated a decrease in reactive oxygen species production (36) and cytokine production (38) after exposure to perflubron. Likewise, isolated human neutrophils exhibited a decreased attachment to and lysis of target epithelial cells in vitro in the presence of perflubron (41). These data are consistent with the possibility that perflubron may exert an anti-inflammatory effect. However, whether this is due to direct contact of cells with a hydrophobic liquid perflubron phase that, in effect, separates effectors from targets or whether perflubron has a specific effect on cellular function has not been well defined.
Emigration of neutrophils from the circulation into areas of inflammation entails a sequence of rolling, tethering, and adhesion events to the activated endothelium of the vasculature before diapedesis across the endothelial layer (4, 19, 37). Leukocyte rolling to the vascular endothelium under conditions of flow is mediated by selectin binding on activated endothelial cells (ECs) (4, 19, 20, 37). Subsequent firm adhesion is mediated by upregulation in the expression of vascular cell adhesion molecule (VCAM) and intercellular adhesion molecules (ICAMs) on the EC surface that bind to their cognate integrin receptors expressed on the leukocyte plasma membrane (4, 19, 37). In vitro studies with human umbilical vein ECs (HUVECs) have indicated that E-selectin expression is upregulated 4–6 h after activation by inflammatory cytokines or lipopolysaccharide (LPS); thereafter, selectins are internalized or shed into the medium, and expression returns to basal values (3, 31, 32). VCAM and ICAMs take longer to reach maximal surface levels (6 and 24 h after activation, respectively), which, in the case of ICAM-1, has been shown to remain fairly constant through 72 h in the continued presence of inflammatory stimuli (23, 31, 32). Neutralizing anti-P-selectin and anti-ICAM-1 antibodies have been shown to reduce the extent of lung injury in a variety of animal models (8, 24, 25), indicating the importance of these adhesion molecules in the migration of neutrophils into the inflamed lung tissue as well as the importance of neutrophil infiltration in the development of ALI or ARDS.
During PLV with perflubron, the endothelial lining of the pulmonary capillary bed is not in direct contact with the hydrophobic liquid perflubron phase. Rather, it is separated from the liquid perflubron phase in the lung by a thin (<1-μm) layer of interstitial extracellular matrix and epithelial cytoplasm. Therefore, we used the two-chambered BIOCOAT system to mimic this spatial separation of endothelial cells from liquid perflubron to address the question of whether indirect exposure to perflubron could modify the rate of neutrophil infiltration into the acutely injured lung by modifying the prerequisite step of neutrophil adhesion to activated endothelium. In this model system, low levels of perflubron diffused into the HUVEC membranes over time. EC monolayers exposed to perflubron before activation with LPS or inflammatory cytokines showed an attenuation in the number of neutrophils binding to ECs, which, in turn, correlated with a reduction in the degree to which E-selectin and ICAM-1 surface protein levels were upregulated. Biosensor measurements of the initial binding rate of neutrophils to activated ECs confirmed that the rate of binding was slower to the cells that had been preexposed to perflubron than to the cells exposed to medium alone.
MATERIALS AND METHODS
The perflubron [medical grade, 100% (vol/vol) perfluorooctyl bromide; LiquiVent] used in this study was supplied by Alliance Pharmaceutical (San Diego, CA). All other chemicals were obtained from Sigma (St. Louis, MO) unless stated otherwise.
HUVECs (American Type Culture Collection, Manassas, VA) were cultured at 37°C in a humidified, 5% CO2-95% air atmosphere in modified MCDB 131 medium (Becton Dickinson Labware, Bedford, MA) containing 10% fetal bovine serum, 1 mg/ml of hydrocortisone, 10 μg/ml of epidermal growth factor, 3 mg/ml of bovine brain extract, and 50 μg/ml of amphotericin B. A stock of cells was maintained by passaging every 4–6 days for a maximum of five passages to prevent cells from becoming confluent. For the experiments, 1 × 105 HUVECs in 0.5 ml of medium were seeded in type I collagen-coated inserts of 24-well BIOCOAT plates (Becton Dickinson Labware). The inserts were placed in the lower compartments containing medium, and the cells were allowed to attach and reach confluency (usually 24 h) before being used for experiments.
Human neutrophils were isolated fresh on the day of the experiment. Blood was drawn into EDTA-anticoagulated Vacutainer tubes. Six milliliters of whole blood were layered over 3 ml of Polymorphprep (Nycomed-GIBCO, Oslo, Norway) in a 15-ml centrifuge tube and centrifuged at 700 g for 40 min at room temperature. This yielded an upper leukocyte band consisting of monocytes and lymphocytes and a lower leukocyte band consisting of neutrophils. This lower band was withdrawn into another 15-ml centrifuge tube, underlaid with 1 ml of Polymorphprep, and centrifuged at 1,200 g for 15 min at room temperature. The entire upper cellular fraction was removed (leaving the pellet that contained some red blood cells), placed into a 50-ml centrifuge tube, and washed twice with 25 ml of calcium- and magnesium-free HEPES-buffered salt solution. The neutrophils were centrifuged at 300 g for 10 min and 150 g for 10 min during the first and second washes, respectively. The neutrophils were resuspended gently in medium, a cell count was obtained with an automated Baker 9000 cell counter (Serono Diagnostics, Allentown, PA), and the cell suspension was adjusted to a density of ∼5 × 106 cells/ml.
Measurement of Perflubron Diffusion
HUVECs were seeded at a density of 1×105 cells in 0.5 ml medium/insert, with three sets of three wells containing medium alone. After the cells had formed a confluent monolayer, the culture medium in the lower compartment was replaced with either 1 ml of culture medium (control) or perflubron. As shown in Fig.1 A, the cells were attached to the upper surface of the collagen-coated, 3-μm porous membrane of the insert that separated the cells from direct contact with the bulk liquid perflubron layer in the lower compartment well. The inserts were immersed in culture medium before the contents of the lower well were replaced with perflubron. Because perflubron is almost twice as dense as water (density of 1.9), perflubron in the lower well is unlikely to replace the aqueous medium trapped in the membrane pores, further minimizing the chance for direct cell contact with the liquid perflubron. Finally, liquid perflubron can be readily distinguished from the culture medium due to phase separation, but none was ever observed to pass through the membrane to the upper chamber even if the level of perflubron was raised above that of the medium in the insert. Duplicate sets of the medium (medium-only wells) or cells were removed from the wells at 0, 4, 8, 24, and 48 h of incubation, and the cell contents from the wells were resuspended in 0.5 ml PBS/sample. Triplicate 0.1-ml fractions of the sample were extracted with isooctane solution containing perfluorodecalin as an internal standard. These were then extracted twice with water, and the perflubron content in the upper isooctane phase was analyzed by gas chromatography with a DB-1 capillary column and an electron capture detector.
Neutrophil Binding Assay
HUVECs were seeded onto collagen-coated inserts at a density of 1 × 105 cells/insert and were cultured to confluency, forming a coherent monolayer with tight cell contacts and a clear basolateral polarity (typically this occurred within 24 h). The lower compartments of the BIOCOAT plates were then completely filled with 1 ml of either cell medium (control) or perflubron and left overnight at 37°C. Fresh cell medium was added to the upper chamber (0.5 ml), with a subset of HUVECs being activated by adding 100 ng LPS/well for 3–4 h at 37°C. Control ECs (with and without perflubron) were incubated without LPS treatment to determine the basal level of neutrophil binding to unactivated ECs. The ECs were gently washed twice with cell medium, taking care not to disturb the cell monolayer or puncture the membrane barrier. Neutrophils were added to each upper compartment insert (5 × 105 cells/insert), incubated for 45 min in the cold (4°C) to maximize cell binding without transmigration through the membrane, and washed with cell medium. Mixed alkyl trimethylammonium bromide (MATAB; 0.5%) in 50 mM potassium phosphate buffer, pH 6.1 (150 μl), was added to each upper compartment insert, and the cell lysates (50 μl) were transferred to a 96-well microtiter plate and spectrophotometrically measured for myeloperoxidase (MPO) content with the assay described inNeutrophil MPO Assay.
Neutrophil MPO Assay
Substrate. o-Dianisidine, a colorimetric substrate for MPO, was dissolved by warming in distilled water at a concentration of 18 mg/ml. The solution was kept from the light and made up fresh on the day of the assay. The substrate solution was made by adding 400 μl ofo-dianisdine solution to 20 ml of potassium phosphate buffer. Hydrogen peroxide (30% wt/vol) was diluted 100-fold in water, and then 400 μl of this were added to the substrate solution just before use.
Neutrophil lysates. MATAB (0.5 g) was dissolved in 100 ml of phosphate buffer (50 mM, pH 6.1) and used to lyse the neutrophil membranes to release the MPO contained in the cytoplasmic granules. Freshly isolated human neutrophils were used to generate a standard curve to compare the linearity of MPO content as a function of cell count. Neutrophils in phosphate buffer were added to microtiter centrifuge tubes to give the following cell numbers per tube: 50 × 103, 40 × 103, 30 × 103, 20 × 103, 10 × 103, 5 × 103, and 2.5 × 103. The cells were centrifuged at 1,200 g for 15 min to pellet the cells, and the supernatants were discarded. MATAB buffer (150 μl) was added to each tube, and the contents were mixed and then centrifuged at 12,000 rpm for 2 min to pellet the cellular debris, leaving the MPO in the supernatant fraction.
MPO assay. Either MPO standard (50 μl) or an aliquot (50 μl) of each neutrophil lysate was added in duplicate to the designated wells of a 96-well microtiter plate, with one set of wells containing 50 μl of MATAB buffer alone to serve as a blank control. Then, 200 μl of the MPO substrate solution were added to each well, and the samples were incubated for 5 min before the absorbance was measured at 450 nm with a microtiter plate reader (Thermomax, Molecular Devices, Sunnyvale, CA). Absorbance units were converted to total neutrophils bound per well based on a linear neutrophil standard curve (r 2 = 0.99).
Surface Expression of Cell Adhesion Molecules
HUVECs were seeded onto collagen-coated inserts, cultured to confluency, and then exposed to perflubron or plain medium overnight as described in Measurement of Perflubron Diffusion. Fresh cell medium was added to the upper chamber (0.5 ml), and a subset of HUVECs was activated by adding 10 ng/ml of tumor necrosis factor-α (TNF-α) and 1 ng/ml of interleukin-1β (IL-1β) per insert. The maximal expression of E-selectin and ICAM-1 at the plasma membrane was observed 4 and 24 h later, respectively. Control HUVECs (with and without perflubron) were incubated without cytokine treatment to determine the surface protein levels in the absence of activation, i.e., of resting cells. Although these primary cells could be passaged a limited number of times, we noted that the degree to which the cell could be maximally stimulated decreased after the fifth or sixth passage (data not shown). Therefore, for all experiments, particular note of the passage number was made, and cells were used betweenpassages 2 and 5. The cells were washed once with cell medium and incubated at 4°C for 45 min with PBS containing a monoclonal mouse antibody directed against either E-selectin or ICAM-1 (PharMingen, San Diego, CA). The anti-ICAM-1 antibody was directly coupled to phycoerythrin (PE). In the case of HUVECs incubated with anti-E-selectin antibodies, the cells were washed again with cell medium and incubated with a secondary FITC-labeled goat anti-mouse IgG antibody for an additional 45 min. The cells were washed with PBS and fixed with 1% Formalin, and the fluorescence intensity was measured in situ with a Jen Sedival epifluorescence microscope fitted with a Hamamatsu charged-coupled device C5810 camera that was linked to an ImagePro Plus (Media Cybernetics) image-analysis program. For each treatment, a total of three fields per well and two wells were measured for a given passage. The experiment was repeated a total of three times. For each fluorescent tag, the background level was set at the value generated by the isotypic control antibody. The FITC signal was measured by maximizing the green channel relative to the red and blue channels, and the orange PE signal was measured by minimizing the blue channel. Thereafter, all measurements for each fluor were run with the same macro program (identical exposure time and gating channel settings). In parallel, the cell count per field was also obtained by scoring the number of Hoechst-stained nuclei. The overall fluorescence intensity per field was then normalized to cell count.
Kinetics of Cell Adhesion
Confluent HUVEC monolayers were exposed to either perflubron or medium for 40–48 h at 37°C in a 5% CO2-95% air atmosphere as described in Measurement of Perflubron Diffusion. TNF-α (10 ng/ml) and IL-1β (1 ng/ml) were added to the upper chamber medium for 4 or 24 h to induce maximal surface expression of E-selectin or ICAM-1, respectively. The cells were washed and detached from the upper compartment inserts with dispase, a Bacillus-derived neutral protease specifically for type I collagen and, therefore, with minimal effect on adhesion molecules during the dissociation of the tight EC monolayer to a single-cell suspension. The cell contents of duplicate inserts per sample were combined, washed twice with PBS, and analyzed for the rate of neutrophil adhesion with an IAsys biosensor equipped with a manual single cuvette (Affinity Sensors, Paramus, NJ; for a detailed description of this method for real-time kinetic analysis, see Ref. 11). Briefly, HUVECs from each treatment group were added to a cuvette with a hydrophobic surface (aminosilane). After sufficient time for cell adsorption on the cuvette surface, excess cells were removed by washing with PBS and BSA was added to block nonspecific binding. The new baseline established for the refractive index (evanescent field) after cell adsorption onto the cuvette surface and the washing away of excess unwashed cells was similar for both control and perflubron-exposed cells. This indicated that there was little difference in the overall cell binding to the cuvette surface between the two treatment groups. There was little difference in the refractive index on addition of the blocking agent BSA. Neutrophils (1 × 105) were then added to the cuvette, the cells were kept in suspension with gentle stirring, and the initial rate of neutrophil binding to the HUVECs was calculated with the IAsys system Fast-fit software to curve fit the change in the relative refractive index over the 10-min period after addition of the neutrophils. The apparent association constant (K a) is defined as the initial slope of the curve and is expressed in Arc Secs (refractive index change per second; see Fig. 6 A for a representative trace). After the addition of neutrophils and initiation of the PBS wash step, the resultant small decrease represents both the decrease due to the signal from unbound neutrophils and the dissociation rate of loosely bound as well as specifically bound neutrophils. The contribution of excess neutrophils in suspension to the overall signal was determined from the trace of neutrophil binding to BSA-coated cuvettes and found to be minimal (refractive index change per sec ∼10 Arc Secs). Quadruplicate measurements were made for each treatment group.
Data are presented as means ± SE. Where possible, the differences between perflubron-exposed and control (nonexposed) cells were analyzed by two-way analysis of variance with Systat software (SPSS, Chicago, IL). In all cases, the null hypothesis of no difference was accepted at P ≥ 0.05.
Perflubron Diffuses Into Cells
The 24-well BIOCOAT plate was used to approximate the in vivo conditions during PLV with perflubron when the pulmonary vascular endothelium is closely juxtaposed (<1 μm) but not in direct contact with the liquid perflubron phase in the alveoli. Preliminary experiments confirmed that liquid perflubron did not pass through the membrane onto the lower collagen-coated surface of the insert (perflubron is readily distinguished from culture medium due to the phase separation of aqueous and perfluorocarbon phases). Because the basal surface of the cell monolayer was never in direct contact with the liquid perflubron layer, perflubron could not be taken up by nonspecific mechanisms such as pinocytosis. Rather, any perflubron content will depend on diffusion and the relative solubility of perflubron for a given compartment. As shown in Fig. 1 B, no perflubron could be detected in the culture medium even after 18 h of incubation at 37°C as would be expected given the extremely low solubility of perflubron in water. In comparison, small but detectable perflubron levels (1.8 μg/105 cells) were present in the cellular fractions after 4 h of incubation, increasing to 5.5 μg/105 cells at 8 h and plateauing between 7 and 8 μg/105 cells after 24–48 h of indirect exposure to perflubron. Perflubron is highly hydrophobic [partition coefficient between octanol and water phases (log P) ∼ 7] but has a solubility in olive oil, which is considered to be a good surrogate for biological membrane lipids, of 37 mM (21, 30). Consequently, the cell-associated perflubron most likely reflects perflubron that has diffused across the insert membrane and partitioned into the cellular membranes (1, 29).
Neutrophil Binding to Activated ECs Is Attenuated After Preexposure to Perflubron
To determine whether the low levels of perflubron that partitioned into the cells could impact neutrophil binding to activated endothelial cells, HUVECs were grown to confluency on collagen-coated insert membranes and then cultured in the presence and absence of perflubron in the lower chamber for 18 h. Subsets of the control and perflubron-exposed wells were then incubated in the presence of LPS for 3–4 h before neutrophils were added to assess binding under static conditions. After a 45-min incubation with control, nonactivated HUVECs, only 1.31 × 103 ± 0.20 × 103 neutrophils bound. There was no detectable difference in the number of neutrophils binding to resting control cells compared with resting perflubron-exposed cells (1.30 × 103 ± 0.29 × 103). After HUVEC activation with LPS, the number of neutrophils binding to HUVEC monolayers that had not been exposed to perflubron rose 6.3-fold (to 8.31 × 103± 0.71 × 103/well) compared with a 4.1-fold increase (to 5.40 × 103 ± 0.77 × 103/well) to HUVECs that had been preexposed to perflubron. This represented a 35% reduction in the number of neutrophils binding to activated HUVECs under steady-state conditions (P < 0.05; Fig. 2). A similar effect was observed in pilot studies with cultured human pulmonary venous ECs (Cell Systems, Kirkland, WA; data not shown), but because of the difficulty in obtaining a steady supply of these primary cells, HUVECS were chosen for the remaining experiments.
Effect of Perflubron on Cell Adhesion Molecule Expression on HUVECs
To assess whether this attenuation in the maximal number of neutrophils bound to activated EC monolayers reflected a reduction in the degree to which surface levels of cell adhesion molecules (CAMs) are upregulated by proinflammatory cytokines after exposure to perflubron, the levels of E-selectin and ICAM-1 were determined by immunolabeling. The fluorescence intensity of the cells was measured in situ by ImagePro analysis of the FITC (E-selectin) and PE (ICAM-1) fluorescent signals. There was a marked difference in the degree to which both E-selectin (Fig. 3) and ICAM-1 (Fig.4) steady-state levels were upregulated at the cell surface after activation. Consistent with the literature, unactivated HUVECs showed a negligible signal with anti-E-selectin or anti-ICAM-1 immunofluorescence (37). There was no significant difference between the background fluorescence of resting control cells and resting perflubron-exposed cells. In three independent experiments, computerized quantitation of the fluorescence signal indicated a significant reduction in the degree to which E-selectin and ICAM-1 were upregulated after prolonged exposure to perflubron (Fig. 5, A andB, respectively). E-selectin-specific fluorescence was reduced overall by 46% (P = 0.003), and ICAM-1-specific fluorescence was reduced by 49% (P = 0.016 vs. activated control cells). No differences in cell density based on the number of nuclei per field (Hoechst staining) or viability as determined by trypan blue staining and coherence of the endothelial cell monolayer under phase optics were evident between cells grown with medium in the lower well and those cells exposed to perflubron in the lower well for 48 h. Therefore, we saw no evidence that these changes in adhesion molecule surface levels were due to reduced cell viability after exposure to perflubron. Similar results of the attenuating effects of perflubron preexposure on expression of these adhesion molecules were obtained in a pilot study with fluorescence-activated cell-sorting analysis to measure the fluorescence intensity (20 and 36% reduction in E-selectin- and ICAM-specific fluorescence, respectively, of perflubron-exposed cells; data not shown). However, the in situ method does not require trypsinization of the cell monolayer and, therefore, was the preferred approach to measure the fluorescent signal.
Neutrophils Bind at a Slower Rate to Activated HUVECs That Have Been Preexposed to Perflubron
Selectins and CAMs exhibit relatively low affinity for their cognate leukocyte receptors under resting conditions. After activation, there was an upregulation in their density at the cell surface after exposure to proinflammatory stimuli, which provides a mechanism to increase the avidity of leukocyte binding to activated endothelium. Given that surface levels were upregulated under both treatment conditions, it was important to determine whether the relative differences in the extent of upregulation translated into a substantive difference in the initial rate of neutrophil binding to activated HUVECs. In one set of wells, HUVECs were incubated with perflubron in the lower chamber for 40–48 h at 37°C. Another set of HUVECs was incubated with cell medium under the same conditions (control). Subsets of wells from each treatment group were activated with TNF-α and IL-1β for 4 or 24 h. The cells were readily detached from the inserts by the collagenase-specific protease dispase, which had no detectable effect on the surface adhesion molecule proteins (data not shown). The HUVECs were allowed to adsorb onto the cuvette surface, and the initial rate of neutrophil binding was analyzed by the IAsys biosensor as the relative change in refractive index (Fig.6 A shows a representative trace). The initial K a was estimated with the Fastfit software of the IAsys system to define the slope of the change in refractive index after neutrophil addition. After activation for 4 h with TNF-α plus IL-1β (the time course for maximal upregulation of surface levels of E-selectin), control cells exhibited a 3.5-fold increase in the initial K a of neutrophils relative to that of resting control ECs, whereas there was only a 1.4-fold increase in perflubron-exposed cells (P < 0.05; Fig. 6 B). After 24 h of activation (the time course for maximal upregulation of ICAM-1 levels), the K a of neutrophils to control cells increased 1.5-fold. In contrast, perflubron-exposed cells showed no increase in the initial rate of neutrophil binding on inflammatory activation (Fig. 6 C).
PLV with perflubron has been proposed to have an anti-inflammatory effect on the development of ALI in vivo (5, 15, 16) as well as on the activity of isolated neutrophils and alveolar macrophages in vitro (36,38, 41). However, it has been unclear whether these are nonspecific effects due to the juxtaposition of the hydrophobic, non-lipid-miscible perflubron liquid with cellular surfaces having an impact on cell-cell interactions and extracellular protein conformation (a barrier function) or whether perflubron exerts a more specific effect on cellular function. The study presented by Virani et al. (41) suggested that the physical presence of perflubron was required to prevent attachment of neutrophils to target epithelial cells and their subsequent lysis. In other words, perflubron acted simply as a physical barrier.
We have evidence that some of these effects can be induced by indirect exposure to perflubon correlating with partitioning of perflubron into cell membranes. Given that the pulmonary microvasculature is in close proximity but separated from direct contact with perflubron during PLV therapy, it was of interest to determine whether indirect exposure of endothelial cells to perflubron could modulate their subsequent interactions with neutrophils. In the studies presented in this report, the BIOCOAT system was used to mimic the spatial separation of the pulmonary vascular endothelium from the bulk perflubron liquid phase in the airways that occurs in vivo. We were able to demonstrate that perflubron accumulates slowly (low levels after 4 h of exposure, plateauing around 24–48 h of exposure) under these conditions. Because the cells were separated from direct physical contact with the liquid perflubron, any perflubron associated with the cellular fraction represents perflubron that has diffused and partitioned into lipid compartments and not liquid perflubron that has been taken up from the liquid perflubron phase by nonspecific mechanisms such as pinocytosis. The maximal levels that accumulated within the cellular fraction were reached after 24 h of indirect exposure and were on the order of ∼10 μg/105 cells. Perflubron has a solubility in olive oil, a lipid with similar chain-length characteristics to membrane lipids forming bilayers, of 37 mM (21, 30). Perfluorochemicals with moderate chain lengths, such as perflubron, have been hypothesized to intercalate and form an interface between the leaflets of the bilayer (1, 29). Unpublished studies from our laboratory (G. Neslund and V. Obraztov) have confirmed that under conditions of indirect exposure, all the perflubron content is associated with the membranes and that perflubron intercalates into lipid bilayers. Therefore, it seems reasonable to hypothesize that during PLV therapy (which typically lasts for 2–5 days), perflubron could diffuse into adjacent pulmonary capillary EC membranes and exert an effect on membrane-mediated endothelial responses to proinflammatory stimuli. The results presented here confirmed that perflubron diffusing from the lower chamber into the ECs growing in the upper chamber did not prevent, but could modulate, their subsequent response to proinflammatory stimulation and affect the degree to which neutrophil binding was upregulated.
After an inflammatory stimulus, cultured ECs as well as vascular endothelium in vivo have been shown to upregulate the surface expression of selectins and to trigger neutrophil rolling and adhesion (19, 20, 40). E-selectin expression has been shown to peak around 4 h after inflammatory stimulation; protein surface levels then exhibit a slow steady decline over the next 20 h (3, 31, 32). Subsequent firm adhesion and extravasation into sites of inflammation appear to be mediated by integrin-mediated binding to ICAM-1 and platelet EC adhesion molecule (PECAM)-1 (4, 18, 28, 37). ICAM-1 expression reaches peak values over 6–24 h poststimulation in vitro (31, 32) and is upregulated in the pulmonary vasculature in response to IgG immune complexes within 4 h (26). Therefore, the impact of perflubron preexposure on the relative upregulation of surface E-selectin and ICAM-1 expression by proinflammatory stimuli was evaluated. The results demonstrate that preexposure to perflubron resulted in a marked, significant reduction in the degree to which their surface expression was upregulated on HUVECs.
It is important to note that preexposure to perflubron did not abolish the upregulation of E-selectin or ICAM-1. Neither CAM is expressed in nonactivated, resting ECs; rather, TNF-α or IL-1β signaling pathways, operating via the nuclear factor-κB pathway, activate de novo transcription and protein synthesis (3, 4, 13, 31, 32, 37, 40). However, whether preexposure to perflubron exerts its effects by modifying the initial signaling pathway through the cytokine receptors, the extent of CAM synthesis at the translational level and/or the trafficking of newly synthesized protein to the plasma membrane will require further study.
The reduction in the maximal levels of the individual adhesion molecules attained after proinflammatory stimulation of cells preexposed to perflubron compared with that in control cells was 46 and 49% with E-selectin and ICAM-1, respectively. The avidity of neutrophil binding depends on the overall concentration of these molecules at the membrane surface. Because these adhesion molecules were upregulated under both conditions (control and perflubron exposed) compared with the resting condition (nonactivated), it was important to assess whether these differences in steady-state levels would correlate with a difference in the rate of neutrophil binding to ECs. As shown in Fig. 6, the initial rate at which neutrophils bound to activated ECs that had been preexposed to perflubron was slower than the rate of binding to control activated ECs. The transit time of blood through the lung capillary bed has been estimated to be on the order of 1–3 s (6, 17), although the less deformable, larger neutrophils have been estimated to have a much slower transit time, with median transit times ranging from 60 (dogs) to 120 (humans) s (17, 18, 22). If the modulation in adhesion molecule expression and its impact on the rate and extent of neutrophil binding to endothelial monolayers after perflubron exposure that we observed in vitro also occur in vivo, prolonged perflubron exposure may impact the binding interactions that regulate neutrophil extravasation into the lung. This, in turn, may exert an effect on the migration rate of activated neutrophils into the injured lung. Our data would predict that any such effect of perflubron would take some time to manifest during PLV therapy because of the relatively slow rate of perflubron diffusion into endothelial membranes.
Although a variety of blocking antibody studies (8-10, 12, 14,24-26) have indicated a critical role for P-selectin, L-selectin, and ICAM-1 in mediating neutrophil infiltration and development of lung injury in at least some animal models of ALI, studies (8, 10, 24) with knockout mice have indicated that although these molecules appear to play a key role in the emigration of neutrophils from the systemic circulation, they do not appear to be essential for neutrophil infiltration from the pulmonary vasculature in response to a variety of proinflammatory insults. Interestingly, L-selectin has been shown to play a key role in the prolonged sequestration of neutrophils within the lung in complement- and bacterial pneumonia-induced lung inflammation but not in the emigration of neutrophils out of the vasculature into the lung space (10). Similarly, a study (12) with blocking antibodies and neutrophil inhibitory factor indicated that neutrophil-induced lung injury results from both CD18-dependent and -independent mechanisms depending on the stimulus. In the systemic circulation, most neutrophil adhesion and margination occurs in the postcapillary venules (4, 37), whereas in the lung, neutrophil sequestration has been shown to occur in the capillaries (7,22, 43). Neutrophil margination occurs under normal conditions in the lung due to the longer transit times of neutrophils through the lung, with the wide range (3 s to 20 min) presumed to reflect differences in regional flow through the larger vessels compared with that in the pulmonary microcapillary vasculature (17, 18, 22). Taken together, this has caused some debate over the role of selectin-mediated rolling phenomena in neutrophil margination within the lung. However, it is clear that there is increased neutrophil sequestration within the lung in response to inflammatory insults. Some (43) have postulated that this may simply be due to increased resistance to flow resulting from increased neutrophil rigidity in response to chemoattractants, whereas others (9, 10) have shown that this increased sequestration shows L-selectin dependence. Furthermore, the elevated levels of marginated neutrophils in the normal pulmonary vasculature compared with those in the systemic circulation do not equate with increased extravasation during normal conditions (7, 10, 18). Rather, increased migration occurs only in response to proinflammatory insults of diverse types such as pneumonia, LPS, and cigarette smoke (7, 18). Clearly, this transition must be associated with changes in cell-cell adhesion interactions that facilitate migration and diapedesis. Recent studies (7, 28) have suggested that PECAM-1- rather than ICAM-1-mediated interactions may be more important because migration across the endothelium occurs at the lateral cell borders between neighboring ECs, correlating with PECAM-1 localization.
Irrespective of the exact mechanism regulating neutrophil infiltration within the injured lung, the principle that perflubron preexposure can modulate neutrophil EC interactions by inducing a reduction in the levels of adhesion molecules at the plasma membrane, at least in vitro, is of interest in the pathophysiology of acute lung diseases such as ARDS. ARDS represents a condition of maximal, uncontrolled influx of inflammatory cells into the lung. Should a similar downregulation of key adhesion molecules occur in vivo after prolonged PLV therapy, then such a mechanism might help contribute to a reduced neutrophil content, thereby aiding in the resolution of the neutrophil inflammatory component of ARDS.
We are grateful to Dr. Helen Ranney for helpful discussions of the manuscript and to Rich Jones for analysis of perflubron levels in cell samples.
Address for reprint requests and other correspondence: C. M. Woods, Dept. of Biological Research, Alliance Pharmaceutical Corp., 3040 Science Park Rd., San Diego, CA 92121 (E-mail:).
Present address of S. F. Flaim: Galileo Laboratories Inc., 5301 Patrick Henry Dr., Santa Clara, CA 95054.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.
- Copyright © 2000 the American Physiological Society