INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Under a variety of acute or chronic pathophysiologic conditions, neutrophils contribute to the pathogenesis of tissue injury, which predominantly affects the pulmonary microvasculature. However, the accumulation of neutrophils in the pulmonary microvascular bed is not unique to the pathologic situation but represents a basic physiologic phenomenon. Under physiologic conditions, a large pool of neutrophils resides within the microvasculature of the lung (1), exceeding the pool of circulating neutrophils by twofold to threefold (2) and exchanging in a dynamic equilibrium with the latter. This physiologic margination of neutrophils is predominantly confined to the alveolar capillary bed (1) and has previously been attributed to mechanical hindrance of neutrophil transit, as a substantial percentage of capillary segments exhibit diameters distinctly smaller than mean neutrophil diameter (3). Thus, neutrophils have to deform into elongated shapes in order to squeeze through the narrow segments of the pulmonary capillary bed. As neutrophil deformability is low compared with that of red blood cells (4), their transit will be delayed as compared with the latter.

After neutrophil activation, neutrophil deformability is further reduced because of stiffening of the neutrophil cytoskeleton (5) by formation of actin microfilaments. Thus, neutrophil passage through the narrow segments of the pulmonary capillary bed is further impeded, resulting in the additional accumulation of neutrophils referred to as neutrophil sequestration. In contrast, in the systemic microcirculation the pathophysiologic accumulation of neutrophils is a multistep process that is predominantly restricted to postcapillary venules and involves rolling and subsequent firm adherence of neutrophils. The different steps of this adhesion cascade are mediated by specific adhesion molecules expressed on the surface of endothelial cells and neutrophils. Whereas L-, P-, and E-selectins mediate neutrophil rolling, interaction of the integrins Mac-1 and LFA-1 with ICAM-1 results in firm adherence of neutrophils to the endothelium.

Recently, we demonstrated that not only mechanical hindrance of neutrophil transit but also leukocyte/endothelial interactions mediated by selectins contribute to the physiologic margination of neutrophils in alveolar capillaries (6). Moreover, the pathophysiologic sequestration of leukocytes and subsequent development of acute neutrophil-mediated lung injury have been demonstrated to depend on selectins (7, 8), and a particular role in the pathogenesis of acute tissue injury has been proposed for L-selectin (9, 10).

We hypothesized that neutrophil sequestration in the pulmonary capillary bed is not solely the result of impeded transit of stiffened neutrophils but involves L-selectin-mediated leukocyte/endothelial interaction. We investigated the effect of the anti-L-selectin antibody HuDreg 200 on leukocyte sequestration in alveolar capillaries in a model of acute systemic endotoxemia. Endotoxin or its active lipopolysaccharide (LPS) moiety have been demonstrated to induce leukocyte sequestration (11) and subsequent leukocyte-dependent lung injury (12, 13). Intravital fluorescence microscopy was applied, as this is the only method allowing for direct visualization and simultaneous quantification of leukocyte kinetics and microhemodynamic parameters in alveolar capillary networks.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Animal Preparation

Experiments were performed in 29 male New Zealand White rabbits weighing 2.87 ± 0.05 kg. All animals received care in accordance with the Guide for the Care and Use of Laboratory Animals (NIH Publication No. 86-23, revised 1985). The study was approved by the local animal care and use committee of the local government authorities (Regierung von Oberbayern). Anesthetic protocol, surgical preparation, and experimental setup have previously been described in detail (1, 6, 14). In brief, anesthesia was induced intravenously by thiopental sodium (50 mg). Animals were tracheotomized and ventilated mechanically (infant ventilator model IV-100; Sechrist Industries, Inc., Anaheim, CA). Inspired O₂ fraction (FI_O2) was set to 0.33, inspiratory airway pressure to 8 mm Hg, and end-expiratory airway pressure to 2 mm Hg in order to prevent atelectasis. Respiratory rate was adjusted to maintain Pa_CO2 at 35 to 40 mm Hg and arterial blood pH at 7.35 to 7.40. Anesthesia was maintained intravenously by -chloralose (50 mg/kg body weight [bw]). Piritramide (1.5 mg/kg bw) and pancuronium bromide (0.3 mg/kg bw) were administered intravenously for analgesia and neuromuscular blockade, respectively. Adequacy of anesthesia was ensured by continuous monitoring of heart rate and arterial blood pressure. Catheters were introduced into the aorta via the right carotid artery and into the superior vena cava via the right jugular vein for continuous measurement of arterial and central venous blood pressure. A thermodilution catheter (REF-1; Baxter, Unterschleissheim, Germany) was implanted into the descending aorta via the right femoral artery for measurement of cardiac output. A separate catheter was implanted into the left femoral artery and exclusively used for blood withdrawals for endotoxin measurements. After left thoracotomy, the pericardium was incised and a catheter was introduced into the pulmonary artery via the right ventricle for monitoring of pulmonary arterial pressure. Visual access to the surface of the right lung was obtained by partial resection of the right fourth and fifth rib and implantation of a transparent window into the right thoracic wall (1, 6, 14). In order to prevent the lung surface from drying or cooling the window was superfused continuously by a warm (37° C) gas-equilibrated Tyrode buffer.

Cell Labeling and Intravital Fluorescence Microscopy

Autologous erythrocytes were labeled in vitro by fluorescein isothiocyanate (FITC) and reinfused. Leukocytes were stained in vivo by a bolus injection of rhodamine 6G (0.06 µmol/kg bw; Merck, Darmstadt, Germany). A modified Leitz Orthoplan microscope (Wetzlar, Germany) fitted with an L3 filter block for FITC and an N2 filter block for rhodamine 6G fluorescence (all Leitz, Munich, Germany) enabled sequential visualization of FITC-labeled erythrocytes and rhodamine 6G-stained leukocytes in alveolar capillary networks by means of intravital fluorescence microscopy. Kinetics of fluorescently labeled red and white blood cells were monitored by a silicon-intensified video camera (C2400-08; Hamamatsu Photonics, Herrsching, Germany) and recorded on videotape (video recorder AG-7350; Panasonic, Munich, Germany). Microhemodynamics and leukocyte kinetics were investigated during single inspiratory plateau periods prolonged to 5 s each in order to avoid respiratory movements. Microscopic investigations were performed on the lower margin of the right middle lung lobe. Video tape recordings were analyzed off-line in a frame-to-frame technique using a digital image analysis system (Optimas; Bioscan, Edmonds, WA).

L-Selectin Antibody and Flow Cytometry

L-selectin function was blocked intravenously by an infusion of the monoclonal antibody HuDreg 200 (Boehringer Mannheim GmbH, Penzberg, Germany). The humanized Dreg 200 is an IgG₄ mAb directed against L-selectin. It is constructed by combining the CDRs of the murine Dreg 200 antibody previously described to block the lectin function of L-selectin (15) with the human framework and constant regions according to the method described by Co and colleagues (16). HuDreg 200 is eliminated from rabbit serum after a single intravenous dose of 2.0 mg/kg bw with a half-life of about 24 h (U. Martin, personal communication, Boehringer Mannheim GmbH). Prior to administration, HuDreg 200 was diluted in 10 mM KPO₄ and 150 mM NaCl, to a final concentration of 9.3 mg/ml and adjusted to a pH of 7.5. HuDreg 55, the humanized IgG₄ of the murine anti-L-selectin mAb Dreg 55 (15), served as isotype-matched control mAb. Because of their different epitope binding sites (17), HuDreg 200, but not HuDreg 55, cross-reacts with rabbit L-selectin (U. Martin, personal communication, Boehringer Mannheim GmbH).

Anti-L-selectin mAb binding to rabbit neutrophils was determined by flow cytometric analysis using a fluorescence-activated cell sorter (FACSort; Becton Dickinson, Heidelberg, Germany). Five milliliters of whole rabbit blood were collected into heparinized syringes and diluted in phosphate-buffered saline (PBS) to a leukocyte density of 1 × 10⁶ cell/ml, and 250 µl of the suspension were incubated for 30 min at 4° C in the dark with 2.75 µg of an r-phycoerythrin-conjugated F(ab')₂ goat antihuman IgG (Biosource International, Camarillo, CA) alone or in combination with 50 µg HuDreg 200 or HuDreg 55, respectively. For evaluation of in vivo binding, whole blood was collected 5 min and 2 h after a single intravenous dose of 2 mg/kg bw HuDreg 200 and incubated with the secondary antihuman IgG. Contaminating erythrocytes were removed and leukocytes fixed by addition of 1 ml of a lysing solution (FACS Lysing Solution; Becton Dickinson) for 10 min. The samples were centrifuged, washed, and resuspended in PBS prior to flowcytometric analysis. Neutrophils were identified on the basis of their cell size and granularity and selectively analyzed for immunofluorescence after appropriate electronic gating of their population. R-phycoerythrin immunofluorescence was excitated at  = 488 nm and assessed using a standard band-pass filter ( = 595/60 nm). Histograms were generated by cell number versus fluorescence intensity of at least 10,000 cells per sample.

Biochemical Parameters

Arterial and venous blood gas analysis (ABL 300; Radiometer, Copenhagen, Denmark), arterial blood cell count (Coulter Counter T540; Coulter Electronics, Inc., Krefeld, Germany), and manual differential blood count were performed. Endotoxin concentration in plasma was determined in triplicate by quantitative chromogenic limulus amebocyte lysate test (Chromogenix AB, Mölndal, Sweden) and expressed in endotoxin units (EU) with 12 EU  1 ng endotoxin.

Quantification of Microhemodynamics and Leukocyte Kinetics

Microhemodynamics and leukocyte kinetics were analyzed in capillary networks of the subpleural wall of single alveoli as described previously (1, 6, 14). The alveolar wall area (A_A) was measured planimetrically. The passage of at least 30 FITC-labeled red and 30 white blood cells through the alveolar capillary network was evaluated. Total length (L_A) of erythrocyte-perfused capillary segments was determined by superimposing all pathways of FITC-labeled erythrocytes passing the capillary network within the observation interval. Capillary perfusion index reflecting functional capillary density within the alveolar wall area (18) was calculated as

(1)

The velocity of each of the FITC-labeled erythrocytes (V^A_RBC) was calculated as total capillary length passaged by the cell within the alveolar area per time. Mean erythrocyte velocity (^A_RBC) was calculated as harmonic mean of single cell velocities. For each passing leukocyte, time required for transit through the capillary network of the alveolar area was measured as its alveolar capillary transit time (TT _WBC). Leukocytes within the alveolar capillary network that did not move during an entire observation period of 5 s were defined as sticking leukocytes and expressed as number of cells normalized to alveolar wall area (N^A_ST).

Myeloperoxidase Activity

Myeloperoxidase (MPO) activity in lung tissue was used as indirect measure of neutrophil accumulation in total lung and determined according to our previously described method (19).

Study Groups

Animals were divided into four study groups. Group I (n = 6) served as the control group and received intravenously 0.2 ml 0.9% NaCl/kg bw. In group II (n = 6), 2 mg/kg bw of HuDreg 200 were administered intravenously. Group III (n = 8), the endotoxin group, received a single intravenous bolus infusion of 20 µg/kg bw LPS from Escherichia coli 0111:B4 (Sigma, St. Louis, MO). In preceding experiments (n = 4, data not shown), this dosage was found to induce a > 50% drop of leukocyte count in arterial blood without influencing pulmonary arterial pressure or cardiac output. Group IV (n = 6) received intravenously a prophylactic dose of 2 mg HuDreg 200/kg bw 15 min prior to an intravenous bolus infusion of 20 µg/kg bw LPS. An additional three animals received 2 mg/kg bw of the isotype-matched control mAb HuDreg 55 intravenously 15 min prior to LPS infusion (20 µg/kg bw).

Experimental Protocol

After completion of the surgical preparation, previously defined inclusion criteria, i.e., mean arterial pressure > 60 mm Hg and lack of macroscopic visible atelectasis, hemorrhage, or perfusion failure on the lung surface were confirmed and animals were randomly assigned to the study groups. Two additional experiments were performed in the endotoxin group (Group III) as compared with the other groups, as in two experiments the alveolar walls under study were no longer perfused after infusion of LPS and thus kinetic parameters such as the alveolar transit time of leukocytes (TT_WBC) could not be analyzed in these animals. Autologous FITC-labeled erythrocytes were reinfused and a time interval of 30 min was permitted to ensure splenic elimination of rheologically altered labeled red blood cells. Rhodamine 6G was administered immediately prior to investigations. Only perfused capillary networks depicted clearly in the focus plane of the microscope that did not exhibit signs of edema formation, i.e., widening of interalveolar septa, were selected for investigation. In each animal, erythrocyte kinetics within one alveolar wall area were video-recorded during three prolonged inspiratory plateau periods using the filter block for FITC fluorescence. Subsequently, leukocyte kinetics within the identical capillary network were recorded during six inspiratory plateau periods using the filter block for rhodamine 6G. After video recordings, arterial and venous blood samples were obtained for analysis of biochemical parameters.

After completion of baseline recordings, the respective stimulus (NaCl, HuDreg 200, LPS, or the combination of HuDreg 200 or HuDreg 55, respectively, 15 min prior to LPS) was administered intravenously according to the design of each study group. Video-recordings of red and white blood cell kinetics within the identical alveolar capillary network and blood withdrawals were repeated at 5, 30, 60, and 120 min after intravenous infusion of the respective stimulus. Macrohemodynamic pressures were monitored continuously, and cardiac output was determined by thermodilution technique under baseline conditions as well as at 60 and 120 min after stimulus application. At the end of experiments, animals were killed by intravenous injection of saturated potassium chloride, and lungs (weight ~ 150 mg) were excised for measurement of myeloperoxidase activity.

Statistics

All data are given as mean ± SEM. Data were tested within each group for significant differences from baseline using repeated-measures analysis of variance according to the Friedman and Dunn test. For differences between study groups, analysis of variance on ranks according to the Kruskal-Wallis and Dunn test were applied. Nonlinear regression analysis was performed using the software program SigmaPlot (Jandel Corporation, San Rafael, CA). Statistical significance was assumed at p < 0.05.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Flow Cytometry

Flow cytometric analysis demonstrated the binding of the mAb HuDreg 200 to rabbit neutrophils (Figure 1). After in vitro incubation with HuDreg 200 and a secondary r-phycoerythrin-conjugated IgG, 92 ± 2% of neutrophils stained positively (Figure 1, right panel ) as compared with 0.8 ± 0.2% of native neutrophils (Figure 1, left panel ), 1.8 ± 0.5% of neutrophils incubated with the secondary IgG alone (Figure 1, middle panel ), and 2.1 ± 0.4% of neutrophils incubated with the control mAb HuDreg 55 and the secondary IgG. Binding of HuDreg 200 to rabbit neutrophils increased mean channel fluorescence from 18 ± 1 (native), 23 ± 1 (secondary IgG alone), and 22 ± 1 (HuDreg 55 and secondary IgG), respectively, to 353 ± 12. Similar results were obtained for in vivo binding of HuDreg 200 to rabbit neutrophils. Five minutes and 2 h, respectively, after intravenous infusion of HuDreg 200, 86 ± 4% and 91 ± 3% of neutrophils stained positively, with a mean channel fluorescence of 338 ± 14 and 262 ± 11, respectively.

View larger version (14K): [in this window] [in a new window]   Figure 1.   Flow cytometric analysis of the binding of mAb HuDreg 200 to rabbit neutrophils. Flow cytometry histograms of log fluorescence intensity for native neutrophils (left panel ), neutrophils incubated with an r-phycoerythrin-conjugated F(ab')2 goat antihuman IgG alone (middle panel ), and neutrophils incubated with the anti-L-selectin mAb HuDreg 200 and the r-phycoerythrin-conjugated IgG (right panel ).

Macrohemodynamics

Significant decreases of arterial pressure were observed in the LPS and the HuDreg 200/LPS groups 60 and 120 min, respectively, after LPS infusion (Table 1). In contrast, arterial pressure remained constant in the NaCl and the HuDreg 200 groups. Pulmonary arterial pressure (Table 1), central venous pressure, and cardiac output (data not shown) did not change significantly in any of the study groups.

Biochemical Parameters

Pa_O2 (Table 1) and Pa_CO2 (data not shown) remained constant throughout experiments in all study groups. However, white blood cell count decreased both in the LPS and in the HuDreg 200/LPS groups as soon as 5 min after LPS infusion (Table 1). This rapid drop was predominantly due to an acute reduction of circulating granulocytes by 93.6 ± 0.3% and 96.0 ± 0.2%, respectively (p < 0.05 each). Throughout the experiments, white blood cell count also decreased in the NaCl group, but not in the HuDreg 200 group. The plasma concentration of endotoxin (Figure 2) remained almost below detection limit in the NaCl and the HuDreg 200 groups. In contrast, both the LPS and the HuDreg 200/LPS groups exhibited high plasma concentrations of endotoxin after LPS bolus infusion. Time course of endotoxin elimination from plasma was comparable within these two groups and exhibited an endotoxin half-life of 52.6 and 53.2 min, respectively (Figure 2). Endotoxin elimination from plasma is described by the equation:

(2)

View larger version (17K): [in this window] [in a new window]   Figure 2.   Endotoxin concentration in plasma. EU = endotoxin units. Dotted lines serve assignment of data points to the different study groups and do not reflect exact time course of parameters. *p < 0.05 versus baseline; #p < 0.05 versus NaCl, p < 005 versus HuDreg 200.

Microhemodynamics

Microhemodynamics and leukocyte kinetics were quantified in 29 alveolar capillary networks, one per each animal, with alveolar wall areas ranging from 2.1 to 10.8 · 10³ µm². Capillary perfusion index remained constant throughout experiments in the NaCl and the HuDreg 200 groups (Figure 3). However, in the LPS group capillary perfusion index decreased continuously after LPS bolus infusion. Two hours after LPS infusion capillaries within the alveolar walls under study were no longer perfused by red blood cells in two of eight experiments. However, when animals were pretreated with HuDreg 200, the LPS-induced reduction of capillary perfusion was completely blocked and capillary perfusion failure was not detected. Hence, HuDreg 200 significantly ameliorated capillary perfusion index after LPS infusion. Mean erythrocyte velocity ^A_RBC decreased continuously after LPS infusion (Figure 4). However, ^A_RBC remained constant in all other study groups, including the HuDreg 200/LPS group. Thus, the LPS-induced reduction of alveolar capillary perfusion was effectively prevented by HuDreg 200. In contrast, pretreatment of animals with HuDreg 55 did not prevent LPS-induced microhemodynamic changes (data not shown).

View larger version (15K): [in this window] [in a new window]   Figure 3.   Capillary perfusion index (CPI). *p < 0.05 versus baseline; p < 0.05 versus LPS.

View larger version (15K): [in this window] [in a new window]   Figure 4.   Mean erythrocyte velocity VARBC in alveolar capillaries. *p < 0.05 versus baseline; #p < 0.05 versus NaCl; p < 0.05 versus HuDreg 200; p < 0.05 versus LPS.

Leukocyte Kinetics

Leukocyte kinetics in alveolar capillaries differ significantly from those in arterioles and venules. Leukocytes enter the capillary network with a relatively high velocity, but a certain percentage of cells stop at distinct sites of the capillary bed for a variable time period before moving again and continuing their passage through the pulmonary microvasculature (1, 6, 14). The median alveolar transit time of leukocytes (TT_WBC) is determined by the velocity of passing leukocytes, by the percentage of temporarily stopped leukocytes, and by the time elapsed during each of these stops and thus provides a summarizing parameter for leukocyte kinetics in alveolar capillaries (6). Alveolar transit time of leukocytes was 0.39 ± 0.07 s at baseline, but it increased by a factor of 24 within 5 min after LPS infusion (Figure 5). However, 30 min after LPS infusion TT_WBC returned to baseline values and increased then again moderately after 2 h. In contrast, TT_WBC exhibited only a slight increase 60 min after LPS infusion, if animals were pretreated with HuDreg 200. Thus, HuDreg 200 did not completely block, but significantly reduced, the LPS-induced retardation of leukocyte transit through alveolar capillaries. Finally, the number of sticking leukocytes (N^A_ST) increased rapidly and continuously after LPS bolus infusion, indicating a massive intracapillary leukocyte accumulation (Figure 6). In the NaCl group, the HuDreg 200 group, and the HuDreg 200/ LPS group the number of sticking leukocytes remained constant. Hence, the LPS-induced accumulation of leukocytes was completely blocked by pretreatment with HuDreg 200. In contrast, HuDreg 55 did not inhibit LPS-induced leukocyte sequestration in alveolar capillaries (data not shown).

View larger version (13K): [in this window] [in a new window]   Figure 5.   Median alveolar transit time of leukocytes (TTWBC). *p < 0.05 versus baseline; #p < 0.05 versus NaCl; p < 0.05 versus HuDreg 200; p < 0.05 versus LPS.

View larger version (13K): [in this window] [in a new window]   Figure 6.   Number of sticking leukocytes per alveolar wall area (NAST). *p < 0.05 versus baseline; #p < 0.05 versus NaCl; p < 0.05 versus HuDreg 200; p < 0.05 versus LPS.

MPO Activity

MPO activity in lung tissue was increased 2-fold (p < 0.05) in the LPS group (148 ± 21 mU/g) as compared with the NaCl group (71 ± 8 mU), indicating increased neutrophil sequestration in the total lung after infusion of LPS. Pretreatment of animals with HuDreg 200 completely abolished this LPS-induced effect (56 ± 7 mU/g; p < 0.05 versus LPS).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

It was the aim of the present study to investigate the effects of the anti-L-selectin antibody HuDreg 200 on the pathophysiologic sequestration of leukocytes in alveolar capillaries. Leukocyte sequestration was induced by acute systemic endotoxemia and analyzed using intravital fluorescence microscopy of the pulmonary microcirculation in a rabbit model.

The main finding of our study was that pathophysiologic effects on alveolar capillary microhemodynamics and leukocyte kinetics induced by intravenous bolus infusion of 20 µg LPS/ kg bw could be completely blocked by pretreatment of the animals with the anti-L-selectin antibody HuDreg 200. Administration of HuDreg 200 did not only inhibit the LPS-induced intravascular accumulation of leukocytes in alveolar capillaries but also prevented the reduction of alveolar functional capillary density and alveolar capillary perfusion. Hence, L-selectin played an important role in leukocyte sequestration and impaired microvascular perfusion in the present model.

Pathophysiologic sequestration of leukocytes in alveolar capillaries was induced by intravenous infusion of the active LPS moiety of endotoxin from Escherichia coli 0111:B4. Intravenous administration of as little as 33 ng/kg bw LPS induces pulmonary sequestration of leukocytes in rabbits (11). Higher LPS doses (> 500 µg/kg bw) are required to induce significant neutrophil-dependent lung injury (12, 13). In the present study, 20 µg/kg bw LPS were administered as a single intravenous bolus infusion. This dosage significantly elevated endotoxin plasma concentration throughout the time course of investigations, but it did not result in significant lung injury, as indicated by minimal changes in pulmonary hemodynamics and blood gas determinations. A potentially protective role of L-selectin in endotoxin-induced neutrophil-dependent lung injury hence remains speculative. In addition, other models of neutrophil-mediated lung injury, e.g., after complement activation involve adhesion molecules other than L-selectin (7). Therefore, further studies are required to evaluate the relevance of L-selectin-mediated leukocyte/endothelial interaction in other experimental or clinical conditions of leukocyte-dependent acute lung injury.

The present model of endotoxemia was not aimed to induce acute lung injury but to allow for evaluation of mechanisms underlying early leukocyte sequestration in alveolar capillaries. The applied dosage of 20 µg LPS/kg bw was chosen since it induced a significant drop of the peripheral leukocyte count without influencing pulmonary macrohemodynamics in preceding experiments. In the present study, pulmonary blood flow remained constant throughout experiments in all groups, but pulmonary artery pressure increased temporarily by 1.7 mm Hg after LPS infusion. This finding did not reach significance and cannot account for the LPS-induced effects on capillary microhemodynamics and leukocyte kinetics since increased pulmonary artery pressure induces capillary recruitment (18) and reduces capillary leukocyte retention (20). Hence, the LPS effects determined in alveolar capillaries cannot be attributed to variations in pulmonary macrohemodynamics.

Experiments were performed using fluorescently labeled erythrocytes and leukocytes. In vivo labeling of leukocytes by rhodamine 6G excludes cell activation by separation procedures and stains all circulating leukocytes (14). However, as the dye labels all leukocytes, this method does not permit differentiation between different leukocyte subcategories (lymphocytes, neutrophils, basophils, or monocytes), and caution must be taken not to treat leukocytes as a single homogenous population.

The binding of HuDreg 200 to rabbit neutrophils was confirmed by flow cytometric analysis. The binding of anti-L- selectin antibodies to neutrophils has been proposed to activate tyrosine phosphorylation (21) in neutrophils and might therefore alter their kinetics. However, none of the parameters assessed in this study exhibited significant changes after intravenous infusion of HuDreg 200 alone, and leukocyte kinetics in the HuDreg 200 group did not differ from that in the control group receiving a single infusion of normal saline. Thus, direct activation of neutrophils by HuDreg 200 appears unlikely in the presented model, and differences between the LPS group and the HuDreg 200/LPS group can be attributed to the blocking ability of the antibody. In addition, unspecific effects of HuDreg 200 can be ruled out since the isotype-matched control mAb HuDreg 55 did not affect LPS-induced changes of capillary microhemodynamics and leukocyte kinetics.

LPS infusion increased the number of sticking leukocytes in alveolar capillaries continuously over a 2-h interval. Because the leukocyte count in arterial blood remained constant after an initial drop within the first 5 min after LPS infusion, the subsequent accumulation of leukocytes in alveolar capillaries must be attributed to redistribution of sequestered leukocytes to alveolar capillaries or leukocyte release from bone marrow. Five minutes after LPS infusion, a considerable amount of leukocytes was withdrawn from the systemic circulation because of a massive prolongation of their passage through the alveolar capillary network without yet being firmly sequestered in the pulmonary microvasculature. However, within the following 25 min these leukocytes, which presumably underwent activation and stiffening (5), were completely sequestered from the blood and contributed to the successive leukocyte accumulation in alveolar capillaries. In addition, after stimulation a large number of neutrophils can be rapidly released from the bone marrow (22).

LPS-induced accumulation of leukocytes in alveolar capillaries was completely blocked by HuDreg 200. This indicates that the LPS-induced leukocyte sequestration in lung capillaries was mediated by L-selectin. A distinct role for L-selectin-mediated cell/cell interaction as necessary prerequisite for subsequent adherence via ₂-integrins has been described by von Andrian and colleagues (23), and protective effects of anti-L-selectin antibodies have been reported in feline myocardial reperfusion injury (10), in reperfusion injury to the rabbit ear (24), or after hemorrhagic shock (25). However, the concept of a multistep adhesion cascade mediating leukocyte rolling, adherence, and emigration is confined to the venular compartment of the systemic microvasculature. In contrast, in the pulmonary microcirculation, the predominant site of leukocyte sequestration (1) and emigration (26) is the capillary bed, and the contribution of L-selectin-mediated cell/cell interaction to leukocyte sequestration in this microvascular compartment is not clear.

In pulmonary capillaries, neutrophil sequestration has previously been attributed to mechanical retention of neutrophils in the narrow segments of the pulmonary capillary network. However, several studies have suggested the contribution of L-selectin-mediated leukocyte/endothelial interaction in the pathogenesis of pulmonary leukostatis and subsequent acute lung injury. After infusion of complement fragments, L-selectin is required for the permanent sequestration of neutrophils in the lung (27, 28). In addition, L-selectin-deficient mice undergoing 24 h of systemic endotoxic shock have a 90% survival rate as compared with 0% in wild-type mice (9). Monoclonal anti-L-selectin F(ab')2 fragments reduce the pulmonary accumulation of neutrophils and lung injury after systemic complement activation and intrapulmonary deposition of IgG immune complexes, respectively (8). Despite these studies suggesting a contribution of L-selectin to the pathogenesis of acute lung injury in a variety of experimental models, microcirculatory events involved in this process are still unclear and discussed controversially. In the present study, we demonstrated that L-selectin can contribute to the pathophysiologic sequestration of leukocytes in alveolar capillaries by prolongation of leukocyte transit through the alveolar capillary network and firm arrest of leukocytes. Both phenomena might be initiated by upregulation of the L-selectin ligand on the pulmonary capillary wall, which remains to be identified on vascular endothelial cells, or by increase of L-selectin affinity after neutrophil activation because of L-selectin tyrosine phosphorylation (21). Moreover, cross-linking of L-selectin because of L-selectin-mediated leukocyte rolling in extracapillary pulmonary and systemic microvessels may induce cytoskeletal reorganization and upregulation of ₂-integrins on the leukocyte cell membrane (29) and may hence indirectly promote leukocyte sequestration in alveolar capillaries. Finally, activation of leukocytes in endotoxemia might also be attributable to a direct binding of LPS to L-selectin, which can act as low-affinity signaling LPS receptor (30).

Despite inhibition of leukocyte sequestration in alveolar capillaries, HuDreg 200 failed to prevent the rapid onset of systemic neutropenia after LPS infusion. Thus, blockade of L-selectin did not compromise leukocyte sequestration but rather redistributed the cells from alveolar capillaries to other microvascular beds. Leukocytes did not redistribute to extracapillary pulmonary microvessels since HuDreg 200 completely blocked the LPS-induced increase of total lung MPO activity and since leukocyte sequestration is absent in pulmonary arterioles and venules after infusion of HuDreg 200 and LPS (31). In contrast, neutrophil emigration into the peritoneum is independent of L-selectin (32), but requires P- or E-selectin (33). Accordingly, neutrophil emigration is absent in the peritoneum, but normal in pulmonary alveoli in P-selectin/ ICAM-1 double mutant mice (34). These findings suggest the existence of an L-selectin-independent, P- or E-selectin-dependent sequestration of leukocytes in extrapulmonary organs. Determination of site and mechanisms involved in leukocyte sequestration after stimulation by LPS and blockade of L-selectin remains subject to future studies. However, redistribution of leukocyte sequestration from the pulmonary microvasculature to extrapulmonary organs after L-selectin blockade has to be considered cautiously when evaluating the protective potential of anti-L-selectin mAbs.

After intravenous infusion of LPS, the CPI indicating the number of perfused capillaries per alveolar wall area decreased significantly. For reasons discussed above, this effect is not attributable to macrohemodynamic changes but rather to direct obstruction of alveolar capillaries by the intravascular accumulation of leukocytes. We recently reported that release of leukocytes from the lung can recruit alveolar capillaries (6). Similarly, the present study indicates that excessive leukocyte accumulation in the alveolar capillary bed results in capillary derecruitment. HuDreg 200 completely blocked this reduction of capillary perfusion index. Therefore, blockade of L-selectin did not only inhibit the LPS-induced sequestration of leukocytes in alveolar capillaries but also increased the capillary gas exchange surface area of the lung.

Finally, mean red blood cell velocity in alveolar capillaries decreased continuously after LPS infusion. Again, this finding was not attributable to macrohemodynamic changes but rather resulted from a redistribution of blood flow. Blood flow is relatively high in basal lung regions and low in apical lung regions (35), where erythrocyte transit time is therefore prolonged (2). Because of the inverse relationship between pulmonary microvascular perfusion and leukocyte accumulation (14), leukocyte sequestration in the lung exhibits a similar vertical gradient, with relatively high sequestration in apical lung regions as compared with hypostatic lung zones (19). Moreover, the relatively small capillary lumina in apical lung regions further enhance leukocyte sequestration in these areas (4). After LPS infusion, blood flow will redistribute to hypostatic lung regions because of excessive leukocyte sequestration predominantly in the apical lung zones. Pretreatment of rabbits with HuDreg 200 completely abolished the LPS- induced reduction of apical lung perfusion, presumably by preventing obstruction of apical alveolar capillaries by sequestered leukocytes.

In summary, administration of the anti-L-selectin mAb HuDreg 200 effectively blocked the sequestration of leukocytes in alveolar capillaries of rabbit lung after infusion of 20 µg LPS/kg bw. Prevention of leukocyte accumulation inhibited the LPS-induced reduction of alveolar functional capillary density and capillary perfusion. Hence, in the present model of acute endotoxemia, leukocyte sequestration in alveolar capillaries was not the mere result of an impeded pulmonary transit of stiffened neutrophils but involved leukocyte/endothelial interactions mediated by L-selectin. Potentially protective effects of anti-L-selectin mAbs in acute neutrophil-mediated lung injury remain to be evaluated in future experimental and clinical studies. Beneficial effects of anti-L-selectin treatment may be counteracted by suppression of host defense mechanisms and redistribution of activated neutrophils into extrapulmonary organs.