Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Little is known concerning the relationships between nonallergic airways inflammation and bronchopulmonary functional disturbances. This may be relevant to the pathogenesis of acute respiratory distress syndrome (ARDS), dominated by an increased alveolocapillary permeability, and of chronic obstructive pulmonary disease (COPD), characterized by neutrophilic inflammation and anatomic rearrangements. Glucocorticosteroids have been used in both instances, but only with limited success (1, 2). We showed that the intraperitoneal or intravenous administration of lipopolysaccharide (LPS) to C57Bl/6 mice is followed by firm neutrophil adhesion onto the pulmonary vascular endothelium (3), in the absence of their migration to the bronchoalveolar lavage fluid (BALF). The myeloperoxidase (MPO) titers of the lungs are elevated as a result of neutrophil sequestration. No overt bronchoconstriction is seen, but bronchopulmonary hyperreactivity (BHR) to aerosolized methacholine is induced, which is due neither to endogenous tumor necrosis factor (TNF)- nor to neutrophil products because it persists after passive neutralization and drug or immune granulocyte depletion, respectively.

To continue this work, we have studied the mechanisms of BHR induced by the intranasal or aerosol administration of LPS. In preliminary experiments, we had observed an intense "direct" effect of LPS on airways resistance in C57BL/6 mice accompanied by marked airway and pulmonary inflammation (4, 5). Because of the current interest in nonallergic lung and airway pathology, such as in ARDS, emphysema, or COPD, we have investigated whether LPS-induced neutrophilic inflammation alters lung functions and induces BHR, and because this was the case, what the mechanism may be and to what extent glucocorticosteroids may be efficacious.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Chemicals

LPS (Escherichia coli 055:B5) was from Difco Laboratories (Detroit, MI). Because unexpected results were obtained with dexamethasone phosphate (see RESULTS), most experiments were duplicated with dexamethasone base, both being from Sigma Chemical Corp. (St. Louis, MO). The steroids were dissolved in dimethyl sulfoxide (DMSO) and 10 µl were injected intraperitoneally. DMSO displayed no effect at this dose. Also from Sigma: hexadecyltrimethyl ammonium bromide (HTAB), ethylenediaminetetraacetic acid (EDTA), O-dianisidine dihydrochloride, and hydrogen peroxide (H₂O₂). Hanks' balanced salt solution (HBSS) was from GIBCO Life Technology, Ltd. (Paisley, UK). Methacholine (acetyl--methacholine chloride) was from Aldrich-Chemie (Steinheim, Germany), vinblastine was from Laboratoires Velbé (France), and granulocyte colony-stimulating factor (G-CSF) was from Immugenex (Los Angeles, CA). Murine recombinant TNF- was kindly provided by Dr. G. R. Adolf (Boehringer-Ingelheim Austria, Vienna, Austria).

Animals

Seven- to eight-week-old male C57BL/6 mice provided by the Centre d'Elevage R. Janvier (Le Genest Saint-Isle, France) were treated intranasally with either different concentrations of LPS in 0.9% NaCl (saline) or with an equivalent volume of saline. Later, at specific time intervals, mice were challenged with methacholine for the study of BHR as described in subsequent text. In another series of experiments, mice were placed in a glass container that allowed aerosolization of LPS (0.3 to 3 mg/ml of 0.9 NaCl solution, saline) or saline using an Aldrich apparatus.

Determination of Bronchoconstriction and BHR

Unrestrained conscious mice, pre-exposed to LPS (see subsequent text) were placed in a whole body plethysmographic chamber (Buxco Electronics, Sharon) to analyze the respiratory waveforms. After 4 min of stabilization, mice received methacholine by aerosolization for 20 s (3 × 10²M in the aerosolator's reservoir). The resulting bronchconstriction was expressed as Penh = [Te (expiratory time)/40% of Tr (relaxation time)  1] × Pef (peak expiratory flow)/Pif (peak inspiratory flow) × 0.67 according to the manufacturer's instructions. Results were expressed as Penh, corresponding to differences between the basal and maximal values.

LPS was administered by two different procedures: intranasally with 20 to 40 µl of saline containing LPS, or unanesthetized mice were exposed to 10 min of LPS aerosol at 0.3 to 3 mg/ml. In this case, the animals were restrained in a nasal box. In both cases, after the administration of LPS, the mice were either placed in the plethysmographic box for online evaluation of Penh before being killed at different intervals or were used later.

Analysis of Airway Inflammation

At given time points after the administration of LPS, mice were anesthetized with 50 mg/kg of urethane intraperitoneally, the tracheas were cannulated and the lungs were washed once with 1 ml saline followed by three washings with 0.6 ml saline, to provide 2 ml of BALF. Aliquots were used to evaluate the total and differential cell number after standard centrifugation. BALF supernatants were collected on ice and TNF- levels were assayed within 2 to 3 h. Protein content in BALF was evaluated with the Bio-Rad protein dye-binding assay (Hercules, CA), according to the manufacturer's instructions.

Evaluation of TNF- by Enzyme Immunometric Assay

TNF- in BALF was determined by an enzyme immunometric assay based on the reaction of thiol groups of monoclonal antibody (mAb) Fab' fragments with maleimido groups introduced into acetylcholinesterase (AChE) (6). As in the case of anti-interleukin-5 antibodies (7), anti-TNF- mAbs MP6-XT22 and MP6-XT3 were purified from ascitic fluids (cloned hybridomas kindly provided by Dr. P. Minoprio, Institut Pasteur, Paris, France), using the affinity chromatography method on a protein G column (HiTrap affinity columns; Pharmacia Biotechnology, Uppsala, Sweden) after precipitation by ammonium sulfate. These rat anti-murine TNF- mAbs were described in detail elsewhere (8).

Immunometric assays were performed in 96-well microtiter plates (MaxiSorps; Nunc, Roskilde, Denmark), coated with 10 µg/ml of the rat antimurine TNF- mAb MP6-XT3 as described (9). The one-step procedure used for immunometric assays involved the simultaneous addition of 100 µl of TNF- standards (7.8 to 1,000 pg/ml) or samples, and 100 µl of the second rat antimurine TNF- mAb MP6-XT22-AchE conjugated at the concentration of 10 Ellman units/ml. After incubation for 18 h at 4°C, the plates were extensively washed, and solid-phase bound AchE activity was determined colorimetrically by addition of 200 µl of Ellman's medium. Absorbance was read at 405 nm with an automatic microplate reader (Dinatech MR 5000; Dinatech Laboratories, Saint Cloud, France). The lower limit of detection of this assay is approximately of 15 pg TNF-/ml sample.

Determination of Lung Myeloperoxidase Activity

Lung tissue MPO activity was determined by minor modifications of a described method (10). After bronchoalveolar lavage, lung vessels were flushed of blood, and the lungs were removed from the thorax, blotted with gauze, and frozen at 20°C until assay. Before being removed, the left atrium was opened and 5 ml of saline were gently perfused. Lungs were then homogenized for 30 s (Potter-Elvejhem glass homogenizer) at room temperature in 1 ml phosphate-buffered saline (PBS). The corresponding extracts were centrifuged (10,000 × g for 10 min at 4°C), and supernatants containing hemoglobin were discarded. The pellets were suspended in 1 ml PBS supplemented with HTAB (0.5%) and EDTA (5 mM), and homogenized again. After centrifugation, 100 µl of supernatants were placed in a test tube with 200 µl PBS-HTAB-EDTA, 2 ml HBSS, 100 µl of O-dianisidine dihydrochloride (1.25 mg/ml), and 100 µl H₂O₂ (0.05% = 0.4 mM). After 15 min of incubation at 37°C in an agitator, the reaction was stopped by addition of 100 µl NaN₃ (1%). The MPO activity was determined as change in absorbance at 460 nM.

Statistical Analysis

Each point corresponds to mean ± standard error of the mean of three to six values obtained from individual mice. Statistical differences were determined using the one-way analysis of variance (ANOVA) and P < 0.05 was considered significant. Individual groups were compared using the unpaired Student's t test. Significance was indicated by an asterisk on the figures when P < 0.05.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Effects of LPS on Neutrophil Recruitment, Production of TNF-, Lung Resistance, BHR and Increased Vascular Permeability

The administration of LPS by aerosol induced a dose-dependent recruitment of neutrophils to BALF and their sequestration in the lungs, as indicated by the increased MPO titers after 3 h and, more intensively, after 24 h (Figure 1). TNF- content in BALF was dose-dependently augmented after 3 h. Thus, 3 h after the aerosolization of 0.1, 0.3, 1, and 3 mg/ml of LPS, 104 ± 21, 437 ± 39, 686 ± 87, and 1,235 ± 130 pg/ml of TNF-, respectively, were found in BALF.

View larger version (18K): [in this window] [in a new window]   Figure 1.   The administration of LPS by aerosol for 10 min at the indicated final concentrations in the aerosolator induces a dose-dependent increase of MPO titers in the lungs (left) and of neutrophil recruitment to BALF (right). Values are for 3 h in the upper row and for 24 h in the lower row. *P < 0.05 compared with NaCl-treated mice.

No TNF- was present in BALF 24 h after LPS administration (not shown). As shown in Figure 2A, lung resistance after intranasal LPS started to increase by 60 to 70 min, peaked by 2 h, and persisted for 3 h or more. This was also the case after the inhalation of LPS but was less intense and persistent (Figure 2B). The influence of LPS on the intensity of subsequent responses to aerosolized methacholine was also studied. As seen in Figure 3, bronchoconstriction induced by methacholine administered 3 to 4.5 h after LPS was dose-dependently and transiently enhanced by LPS (0.1 to 3 mg/ml) and returned to basal values after 24 h, even though neutrophils persisted in BALF and lungs, in terms of MPO titers (Figure 1).

View larger version (28K): [in this window] [in a new window]   Figure 2.   Dose- (A) and concentration-dependent (B) effects of the administration of LPS intranasally and by aerosol, respectively. Vertical scale: Penh (see MATERIALS AND METHODS), as an expression of bronchoconstriction. Horizontal scale: time in minutes after LPS administration.

View larger version (13K): [in this window] [in a new window]   Figure 3.   Intense and brief augmentation of the intensity of the response to methacholine after LPS. Time- (left) and dose-dependent (right) augmentation of the intensity of bronchoconstriction by aerosolized methacholine 3 h after the aerosolization of LPS at the indicated concentrations (left) or times (right, for 3 mg/ml in the aerosolator). Delta A.U.C. refers to the difference between bronchoconstriction by methacholine in LPS-treated and in control (NaCl-treated) mice. *P < 0.05 compared with NaCl-treated mice.

The protein content of BALF was augmented after 24 h (Figure 4) and less so after 3 h (not shown), a time point when the other "direct" effects of LPS were at their peak.

View larger version (15K): [in this window] [in a new window]   Figure 4.   Increased vasopermeation 24 h after the aerosolization of LPS. The concentrations of LPS in the aerosolator are indicated below the columns. Vertical scale: proteins found in BALF in O.D. units (see MATERIALS AND METHODS). *P < 0.05 compared with NaCl-treated mice.

Modulation of the Effects of LPS by Neutrophil Depletion

Vinblastine was administered intravenously (5 mg/kg, 24 to 144 h before intranasal LPS) to deplete neutrophils and study their involvement with LPS-induced bronchoconstriction and hyperreactivity. Figure 5 shows that neutrophils were still recruited to the BALF of vinblastine-treated mice 24 h after intranasal LPS, but that after 72 to 120 h they were virtually absent. At 144 h (6 d), the MPO titers were back to control values, but at this time point, very few neutrophils were found in BALF. The full normalization of neutrophil recruitment to BALF was noted by 11 d after the administration of vinblastine (data not shown). As a control, and in agreement with earlier results (3), intravenous LPS increased the MPO titers as much as did intranasal LPS, but no neutrophils were found in BALF because they remained sequestered on the endothelium (Figure 5). Neutrophil depletion with vinblastine failed to modify the increased TNF- titers in the BALF of LPS- injected mice (data not shown).

View larger version (37K): [in this window] [in a new window]   Figure 5.   Suppression by vinblastine of the LPS-induced recruitment of neutrophils. Vinblastine (vinbl.) was injected intravenously at 5 mg/kg and LPS was aerosolized after the indicated time intervals. Negative controls (sal.-sal.) show no increment in either parameter, whereas positive controls (sal.-LPS) show intense recruitment in terms of MPO (A) and neutrophil counts in BALF (B). Last column corresponds to intravenous LPS (3 mg/kg) with a marked MPO increment in the absence of neutrophil recruitment to BALF.

As shown in Figure 6, bronchoconstriction by LPS was suppressed 72 and 120 h after vinblastine administration, and was back at 144 h, a time point when neutrophils returned to the lungs in terms of MPO titers but were still absent from BALF. Hyperreactivity to methacholine was also suppressed by vinblastine (Figure 6, inset).

View larger version (27K): [in this window] [in a new window]   Figure 6.   Time-dependent suppression by vinblastine of LPS- induced bronchoconstriction followed by recovery. Positive control (NaCl/LPS) shows bronchoconstriction, whereas at the indicated time intervals after 2.5 mg/kg of intravenous vinblastine (72 and 120 h), bronchoconstriction is suppressed and recovers after 144 h. The inset shows the area under the delta curve (A.U.C., see Figure 4) for the responses to methacholine as suppressed by vinblastine and recovered thereafter.

These results suggested that neutrophils in the lungs, but not in BALF, are needed for the bronchoconstrictor effects of LPS. Vinblastine-treated mice were thus treated with G-CSF, which was expected to counteract the effects of vinblastine with respect to neutrophil production. As seen in Figure 7, animals injected with G-CSF (four injections of 10 µg/kg each, 72, 78, 96, and 102 h after vinblastine) recruited neutrophils to the lungs as effectively as nonvinblastine-treated control mice, whereas a lower dose of G-CSF was ineffective. In contrast, G-CSF administration to vinblastine-treated mice failed to allow for BALF replenishment in neutrophils, despite the administration of LPS (Figure 7). Under such conditions, the suppression of LPS-induced bronchoconstriction by vinblastine was only partially counteracted by G-CSF (Figure 8).

View larger version (14K): [in this window] [in a new window]   Figure 7.   (A) Recovery of MPO titers (left) and absence of recovery of neutrophil counts in BALF (right) in vinblastine-treated mice injected with G-CSF. G-CSF was injected two or four times, as indicated, and LPS was aerosolized at 3 mg/ml. Animals were killed 96 h later. (B) Partial recovery of bronchoconstriction by LPS after cell replenishing by G-CSF (four times) in vinblastine-treated animals. Vertical scale: bronchoconstriction in terms of Basal Penh.

View larger version (21K): [in this window] [in a new window]   Figure 8.   Partial recovery of LPS-induced bronchoconstriction in mice pretreated with vinblastine and corrected with four injections of GM-CSF (see MATERIALS AND METHODS). Horizontal scale: time (minutes).

Modulation by Dexamethasone of LPS-Induced TNF- Formation, Neutrophil Recruitment, Bronchoconstriction, BHR, and Vasopermeation

Dexamethasone(s) was dissolved in DMSO, and control animals were treated with DMSO intraperitoneally or intravenously, as indicated, at volumes of 10 µl, which had no effect on their own. TNF- formation (data not shown) and neutrophil recruitment to BALF (Figure 9, middle panel) were suppressed at 24 h by the three doses of dexamethasone base and by its phosphate salt. The MPO lung content was reduced by 1 to 25 mg/kg of dexamethasone, but inhibition plateaued at 60 to 75% (Figure 9, upper panel). In contrast, the accompanying Penh increase was slightly reduced by the low dose and unaffected by the higher doses of 25 mg/kg (Figure 10A). Because these results were surprising, the experiment was performed twice with a similar outcome. The same pattern was seen for vascular leakage into BALF because 1 or 5 mg/kg of dexamethasone was inhibitory, whereas 25 mg/kg had no protective effect (Figure 10B).

View larger version (28K): [in this window] [in a new window]   Figure 9.   Failure of dexamethasone phosphate to suppress neutrophil sequestration in lungs (MPO, upper panel) and protein exudation into BALF (lower panel) as compared with its effectiveness against neutrophil enrichment of BALF (middle panel). *P < 0.05 compared with NaCl-treated mice.

View larger version (26K): [in this window] [in a new window]   Figure 10.   Inhibition of hyperreactivity to methacholine (A) and reduction of LPS-induced bronchoconstriction (B) by 1 mg/kg of dexamethasone as compared with its ineffectiveness at 25 mg/kg. Vertical scale: hyperreactivity to methacholine (in terms of delta A.U.C., A) and bronchoconstriction (in terms of A.U.C., B). Inhibition by 1 mg/kg of dexamethasone is not significantly different from saline-treated. ANOVA, nonsignificant differences.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

This study demonstrates that LPS induces bronchoconstriction when delivered into the murine airways, whether by intranasal instillation or by aerosol. Bronchoconstriction is accompanied by neutrophil recruitment to the lungs and to BALF, and by an enhanced production of TNF-. This is followed by increased responses to the bronchoconstrictor effect of aerosolized methacholine and by a late augmentation of vasopermeation. Some, but not all, of these effects are inhibited by the standard glucocorticosteroid dexamethasone. Different issues raised by these results are discussed in subsequent text.

Role of Neutrophils for the LPS-Induced Bronchoconstriction and BHR

Because of the recognized ability of LPS to induce neutrophil recruitment in vivo, it seemed logical to investigate the potential involvement of neutrophils in the accompanying bronchoconstriction. LPS induced early bronchoconstriction, which started 70 to 90 min after its administration and took 3 to 4 h to return to basal values. At this time point, a short-lived but intense hyperreactivity to methacholine was also observed in the presence of submaximal neutrophil numbers in BALF and in the airways (Figure 1). At 24 h or later (data not shown), more neutrophils were present, but the airways resistance was back to normal, no enhancement of the responses to methacholine being seen (Figure 4). Thus, the presence of neutrophils is not sufficient to support bronchoconstriction or hyperreactivity. Yamamoto and coworkers (11) and Delclaux and colleagues (12) have already suggested that polymorphonuclear leukocyte emigration is necessary but not sufficient for the development of LPS-induced lung injury. Presence of neutrophils after intranasal administration of LPS was reported by Szarka and associates (13) who, as we did (3), noted that the intraperitoneal or intravenous administration does not induce pulmonary injury. In these studies, the "direct" bronchoconstrictor effect of LPS was not reported. In a guinea pig model, Vincent and coworkers (14) demonstrated that hyperreactivity after intratracheal instillation of LPS is platelet- rather than neutrophil-dependent, and indeed platelets are found in the BALF of our mice after LPS aerosolization (data not shown).

To confirm the role of neutrophils, the myelosuppressive drug vinblastine was used as a depleting agent, as we had previously done for eosinophils when studying allergic BHR (15) and for neutrophils in hyperreactivity after intraperitoneal LPS (3). Vinblastine suppressed neutrophil numbers in blood (data not shown) and their LPS-induced recruitment to lungs from 96 to at least 120 h (Figure 6), failing to affect mononuclear cell numbers and functions. The LPS-induced increase of MPO titers, which were also suppressed by vinblastine, started to return to control values at 120 h, recovering fully at 144 h. For contrast, the restoration of neutrophil counts in BALF was delayed, only to be completed after 7 d (data not shown), indicating some form of inability of the newly formed neutrophils, recovering from vinblastine, to cross to the alveolar compartment.

Intravenous LPS induces strong neutrophil adhesion to vascular endothelium, with augmented MPO titers (1) (Figure 6), and also induces hyperreactivity, as does the intraperitoneal administration. Nevertheless, this hyperreactivity differs substantially from that induced by LPS delivered to the airways: it is refractory to vinblastine and the doses of dexamethasone required for inhibition are markedly above those that suppress the production of TNF-. Accordingly, the adhesion of neutrophils to the vascular endothelium and the consequent increment in MPO titers accompany, but are not the cause of, BHR (3). Furthermore, bronchoconstriction in response to LPS returned to the same levels of MPO titers after vinblastine. However, the replenishment of the bronchoalveolar compartment was delayed (Figure 7), suggesting that the neutrophil recruitment to lung parenchyma explains bronchoconstriction and hyperreactivity, both of which were not suppressed by glucocorticosteroids (see subsequent text).

The suppressive effects of vinblastine on neutrophil recruitment to the lungs were surmounted by the coadministration of G-CSF, MPO titers in response to LPS being fully recovered. As in the case of spontaneous recovery, neutrophils did not easily reach the alveolar compartment, this control can be explained either by their relative immaturity and inability to cross the barriers between vessels and the alveoli and/or by their inability to express the required adhesion molecules. Whatever the cause, it is clear that neutrophil sequestration in the lungs is not sufficient to support hyperreactivity. Increasing the time of exposure to G-CSF or the number of injections to stimulate additional maturation of neutrophils was not attempted because it would involve repeating the injections of vinblastine in order to prolong myelosuppression. Experiments with additional growth factors are planned, as is the direct evaluation of the degree of maturation of lung neutrophils.

A correlation between impaired lung function and immune-mediated inflammation was reported recently for Pneumocystis carinii infection (16). Byrne and colleagues (17) have also shown that Pseudomonas aeruginosa injected intravenously to swines induces acute ARDS with reduced lung compliance and augmented vascular permeability. They claimed, rightly so in our opinion, that early changes in pulmonary compliance result from the effects of vasoactive mediators as well as from the sequestration of large numbers of circulating neutrophils in the pulmonary microvasculature in response to sepsis, whereas in the delayed phases the decreased compliance may result from a combination of factors, including altered vascular permeability and protein influx into the airspaces.

Interference of Dexamethasone with LPS-Induced Effects

The anti-inflammatory glucocorticosteroids were expected to be effective in our model, but this was only partially the case. Dexamethasone suppressed TNF- production at the lowest dose that was employed and it inhibited the LPS- induced increase in BALF neutrophil counts (Figure 10), as shown by O'Leary and Zuckerman (18) for rats. In contrast, the lung MPO titers were reduced only up to 60 to 75%, even for the high dose of 25 mg/kg of dexamethasone base or phosphate (Figure 10), which were both used to ascertain that differences in drug distribution would not account for inactivity. Bronchoconstriction by LPS was poorly inhibited by 1 mg/kg of dexamethasone(s), higher doses being inactive (Figure 10B). Interestingly, BHR was also suppressed by the low dose only, and again higher doses were inactive. Protein leakage at 24 h was inhibited by 1 to 5 mg/kg of dexamethasone(s), but 25 mg/kg were inactive or even aggravated leakage (Figure 10). It thus appears that relative failure of dexamethasone(s) to inhibit early (1.5 to 3 h) LPS-induced bronchoconstriction and hyperreactivity is associated with protein leakage at 24 h. It appears as if undesirable effects of glucocorticosteroids were uncovered with higher than usual doses in unsuccessful attempts to maximize therapeutic activity against vasopermeation and the effectiveness of glucocorticosteroids against ARDS, which may have increased the death rates (2). Glucocorticoid-refractory effects of LPS may result from the unmodified production of chemokines such as macrophage inflammatory protein (MIP)-2 in rats (18).

In clear contrast to the inability of dexamethasone(s) to suppress bronchoconstriction and increased pulmonary MPO titers, both the short-lived hyperreactivity to aerosolized methacholine and early TNF- generation in BALF were inhibited by low doses. This might indicate that TNF- accounts for early hyperreactivity, which is compatible with its short half-life in blood and BALF (3). TNF- administered intratracheally to the allergy-prone BP2 mouse strain induces a mild bronchoconstriction and neutrophil recruitment but synergizes markedly with a subeffective dose of antigen to induce hyperreactivity to methacholine (19). Nevertheless, failure of high doses of dexamethasone(s) to modify hyperreactivity, in the presence of TNF- suppression, again suggests the presence of as yet undescribed mediators, generated here from (or in cooperation with) incoming activated neutrophils. One interesting hypothesis raised by the work of O'Leary and Zuckerman (18) showed that dexamethasone up to 30 mg/kg does not inhibit LPS-induced MIP-2 production in rats. It could be that its murine equivalent, KC, may account for the effects we presently describe as resistant to glucocorticosteroids. Dexamethasone-refractory BHR, neutrophil recruitment to the lungs, and protein leakage to the alveolar compartment are models for events occurring during ARDS and should be considered as targets for mechanistic studies and therapeutic intervention. Unraveling the mediators that account for the glucocorticosteroid-resistant effects of LPSbronchoconstriction, most of the hyperreactivity, MPO increment in lungs and increased vasopermeation is our next objective.