Local Inflammatory Responses following Bronchial ... - ATS Journals

Local Inflammatory Responses following Bronchial Endotoxin Instillation in Humans NAOMI P. O’GRADY, HUGH L. PREAS II, JÉRÔME PUGIN, CARMEN FIUZA, MARGARET TROPEA, DEBRA REDA, STEVEN M. BANKS, and ANTHONY F. SUFFREDINI Critical Care Medicine Department, Warren G. Magnuson Clinical Center, National Institutes of Health, Bethesda, Maryland; and the Division of Medical Critical Care, University Hospital of Geneva, Geneva, Switzerland

To study local lung inflammation, 34 subjects had endotoxin (1–4 ng/kg) instilled into a lung segment and saline instilled into a contralateral segment followed by bronchoalveolar lavage (BAL) at 2 h, 6 h, 24 h, or 48 h. Endotoxin instillation resulted in a focal inflammatory response with a distinct time course. An early phase (2 h to 6 h) revealed an increase in neutrophils (p 0.0001) with elevated cytokines (tumor necrosis factor [TNF]-, TNF receptors [TNFR], interleukin [IL]-1, IL-1 receptor antagonist, IL-6, granulocyte–colony-stimulating factor [G-CSF], all p 0.002, but no change in IL-10) and chemokines (IL-8, epithelial neutrophil activating protein-78, monocyte chemotactic protein-1, macrophage inflammatory protein [MIP]-1, MIP-1, all p 0.001, but no change in growth-regulated peptide-). A later phase (24 h to 48 h) showed increased neutrophils, macrophages, monocytes, and lymphocytes (all p 0.02), and a return to basal levels of most mediators. Elevated levels of inflammatory markers (TNFR1, TNFR2, L-selectin, lactoferrin, and myeloperoxidase) persisted in the BAL at 48 h (p 0.001). Increased permeability to albumin occurred throughout both phases (p 0.001). Blood C-reactive protein, serum amyloid A, IL-6, IL-1ra, G-CSF, but not TNF- increased by 8 h (all p 0.008). The local pulmonary inflammatory response to endotoxin has a unique qualitative and temporal profile of inflammation compared with previous reports of intravenous endotoxin challenges. This model provides a means to investigate factors that initiate, amplify, and resolve local lung inflammation.

The rapid response of the innate immune system to infection occurs through pattern-recognition molecules (1). These germline-encoded proteins recognize microbial macromolecules and are either secreted (i.e., C-reactive protein) or reside on cell surfaces (i.e., toll-like receptor, CD14). Cell receptor activation may result in endocytosis of the offending pathogen or lead to cell signaling with release of inflammatory mediators. These mediators control adaptive immune responses by regulating T cell costimulatory molecule expression and by eliciting diverse cytokine and chemokine patterns that activate and recruit cells to a site of infection (2, 3). The bacterial cell wall lipopolysaccharide, endotoxin, is a prototypic microbial trigger that stimulates innate immunity (2, 3). The resultant inflammatory responses are essential in early host defense but may also contribute to later organ injury.

Endotoxin is a major mediator of septic shock and gram-negative bacterial pneumonia (4, 5). To investigate the acute inflammatory responses associated with this bacterial product, humans have been given endotoxin intravenously, which rapidly results in a systemic inflammatory response including fever, leukocytosis, and increased blood levels of activated cells and inflammatory mediators (reviewed in [6]). Most of these acute-phase mediators (i.e., cytokines, chemokines, leukocyte granule contents, coagulation–fibrinolytic proteins) are maximum in blood after 1 h to 3 h and approach or return to baseline levels by 6 h to 8 h (6, 7). The lungs are relatively protected during this systemic inflammation. Lung permeability increases to small ( 500 Da) but not large molecules and no changes occur in lavage cellularity, protein content, or cytokine mRNA or protein (8, 9). The contribution of endotoxin to dust-related occupational lung disease has been studied in normal subjects and subjects with asthma using inhalation challenges with either endotoxin, killed whole gram-negative bacteria, or organic dusts that contain endotoxin as well as other bacterial and fungal elements (10–12). These exposures result in whole lung inflammation with increases in neutrophils and cytokines in the lung lavage (12, 13). Limited information is available regarding the mediators and the temporal sequence of pulmonary inflammation due to endotoxin in this model (12, 14). Important variables in the inhalation models include the amount of inhaled material delivered to the lung and the composition of the organic dusts. To assess whether the same dose of endotoxin as given with intravenous challenges would elicit a qualitatively similar inflammatory response in the lung, we developed a model of local pulmonary inflammation based on direct segmental instillation of endotoxin. We sought to determine the temporal sequence and focal nature of any inflammatory response. The dose of endotoxin used in the current study was identical to that used in human studies of intravenous endotoxin administration (6). Using bronchoalveolar lavage to sample the challenged segments, we show that the initiation and resolution of inflammation in a localized pulmonary focus have a unique qualitative and temporal profile compared with previous reports of intravenous endotoxin-induced systemic inflammation.

METHODS (see Online Data Supplement) Subjects

(Received in original form October 1, 2000 and in revised form January 23, 2001) Funding by NIH intramural funds. Correspondence and requests for reprints should be addressed to Anthony F. Suffredini, M.D., Bldg 10, Room 7D-43, CCMD/CC/NIH, 10 Center Drive, Bethesda, MD 20892-1662. E-mail [email protected] This article has an online data supplement, which is accessible from this issue’s table of contents online at www.atsjournals.org Am J Respir Crit Care Med Vol 163. pp 1591–1598, 2001 Internet address: www.atsjournals.org

Our institutional review board approved the study and written informed consent was obtained from 34 healthy subjects (20 men, 14 women, 28.5 1.2 yr).

Endotoxin Dosing To establish tolerability and a time course, six subjects were challenged at baseline (0 h) with endotoxin (E. coli O:113) (1 ng/kg, n 2, 2 ng/kg, n 2, 4 ng/kg, n 2). Bronchoalveolar lavage (BAL) was then performed at 6 h and 24 h. After this initial study, 28 subjects were studied using 4 ng/kg at 0 h followed by BAL at either 2 h (n 3), 6 h (n 14), 24 h (n 6), or 48 h (n 5).

1592

AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE

TABLE 1. SYSTEMIC RESPONSES IN 28 SUBJECTS FOLLOWING ENDOBRONCHIAL ENDOTOXIN INSTILLATION*

Symptoms† Temperature, C Heart rate, beats/min MAP, mm Hg PaO2, mm Hg PCO2, mm Hg pH

Baseline

Maximum Change (time)

0.46 0.11 36.74 0.07 81.5 2.9 98.71 2.01 103.6 7.0 37.6 0.9 7.42 0.01

0.64 0.23 (6 h) 37.29 0.09 (8 h)‡ 67.7 3.2 (3 h)‡ 86.92 1.43 (4.5 h)‡ 86.5 2.6 (8 h)‡ 38.3 0.65 (2 h)‡ 7.41 0.01 (7 h)

Definition of abbreviation: MAP mean arterial pressure. * Endotoxin dose instilled 4 ng/kg. † The presence and severity of symptoms were quantified by an hourly questionnaire over the initial 8-h study period (see METHODS). The number of subjects experiencing symptoms and their level of severity were: malaise—1 subject (severity of 2), headache —3 (1), myalgias—1 (2), secretions—6 (1), chest pain—0, dyspnea—2 (1) and 1 (2), wheeze—4 (1), and cough—13 (1) and 4 (2). ‡ p 0.01 for change over 8 h by ANOVA. All endotoxin-related effects occurred independent of the time of the second BAL (endotoxin–time interaction, p NS).

VOL 163

2001

180 ml of saline into the endotoxin-challenged segment. Total BAL leukocyte number and cell differential were obtained.

Assays Blood and BAL inflammatory markers and mediators were measured by enzyme-linked immunosorbent assay (ELISA). Measurements of tumor necrosis factor (TNF)- bioactivity (7), proinflammatory activity (e.g., upregulation of intercellular adhesion molecule [ICAM]-1 on cultured human alveolar type II-like A549 cells) (15), and gelatinase-B zymography were performed as previously described (16). Serum samples were assayed for C-reactive protein and serum amyloid A, total and differential blood leukocyte counts, and TNF-, interleukin (IL)-6, granulocyte–colony-stimulating factor (G-CSF), and IL-1 receptor antagonist (IL-1ra) as previously described (7).

Symptom Assessment Symptoms were quantified using an hourly questionnaire (0 h–8 h) regarding the presence of systemic (malaise, headache, myalgias) or pulmonary symptoms (dyspnea, cough, wheezing, chest pain, and secretions) and their severity (0—none, 1—minimal, 2—moderate, and 3—severe).

Statistical Methods Bronchoscopy and BAL Bronchoscopy was performed after applying topical lidocaine. A balloon-tipped catheter was placed through the bronchoscope to occlude the challenged segment. Saline (0.9%, 10 ml) followed by 10 ml of air was instilled into the control segment (either lingula or right middle lobe). Endotoxin (approximate volume 1–2 ml), followed by saline and air, was then instilled in the contralateral lung segment. The subsequent BAL instilled 180 ml of saline into the control segment and

Data were analyzed using analysis of variance (ANOVA) (17). For lavage data, analysis included main effects for time point of lavage, patient (nested within time point), and an indicator of endotoxin exposure (e.g., whether the lavage came from the site that received endotoxin or not). In addition, an interaction between time and endotoxin was included in the model to determine if a significantly different time course existed for those segments challenged with endotoxin or saline (e.g., endotoxin–time interaction). All remaining interactions were pooled to form the error term in the ANOVA. Remaining systemic

Figure 1. Changes in BAL cellularity at 2, 6, 24, and 48 h after saline (open bar) and endotoxin (shaded bar) challenge of lung segments. Endotoxin was associated with increases in (A) alveolar macrophages (endotoxin effect, p 0.024; endotoxin–time interaction, p NS), (B) neutrophils (p 0.0001 and p 0.0001), (C) lymphocytes (p 0.0001 and p 0.04), and (D) monocytes (p 0.0001 and p 0.003). BAL (E) total protein (p 0.0002 and p NS) and (F) albumin (p 0.0011 and p NS) increased after endotoxin compared with saline challenged segments. Data presented as mean and SEM and analyzed by ANOVA. The 2 h time point in A (n 3) includes one outlier. Removing this subject results in mean values on the saline control side of 20.9 (8.8 and 33) and endotoxin-challenged segment of 19.7 (16.6 and 22.8) 104 cells/ml, respectively.

O’Grady, Preas, Pugin, et al.: Endotoxin Effects in Human Lung data could not isolate local effects of endotoxin. These ANOVA models include main effects for timing of lavage, patient (nested within timing of lavage), and time postendotoxin administration. An interaction between timing of lavage and time postendotoxin administration was also included in the model (e.g. group–time interaction), with remaining interactions pooled to form the error term. For all ANOVA models, normality of residuals was assessed using a Shapiro–Wilk test, and, if necessary, data were transformed to improve the distribution of the residuals. Data are presented as mean SEM.

RESULTS (see Online Data Supplement) In the pilot study, sequential lavage at 6 h and 24 h in subjects challenged with 1, 2, or 4 ng/kg of endotoxin revealed a localized inflammatory focus in the lung. All six subjects developed increases in cellularity and neutrophils in the endotoxin-challenged segments compared with their respective control sides. Total BAL cell counts increased 6- to 7-fold compared with the control segment (p 0.002) (see Table E1 in the online daa supplement). The percentage of neutrophils and monocytes increased and the percentage of macrophages decreased in the BAL (all p 0.002). A more detailed examination of the inflammatory response was performed in 28 subjects challenged with 4 ng/kg of endotoxin. The most common symptom was cough (17 subjects)

1593

occurring around the time of the bronchoscopy (Table 1). Temperature increased by 8 h (p 0.0001) and heart rate and mean arterial pressure fell and reached nadir values at 3–4.5 h (p 0.0001). PaO2 fell at 8 h (p 0.001) and pCO2 rose slightly from baseline at 2 h (p 0.01) (Table 1). These changes occurred independent of the time of the BAL (group–time interaction, p NS). The return BAL volume was greater from the endotoxinchallenged segments compared with the control (endotoxin 90.2 27.3 versus control 78.8 5.2 ml, p 0.027). BAL cellularity increased after endotoxin compared with control (endotoxin effect, p 0.0001 and endotoxin–time interaction, p NS) with absolute increases in alveolar macrophages (p 0.024 and p NS), neutrophils (p 0.0001 and p 0.0001), lymphocytes (p 0.0001 and p 0.024), and monocytes (p 0.0001 and p 0.003) (total and differential counts—see Table E2 in the online data supplement, cells per ml of BAL—Figure 1A–1D). Neutrophils predominated in the BAL beginning at 6 h and remained elevated at 24 h and lymphocytes and monocytes rose at 24 h and 48 h. Eosinophils were less than 1% of the control and endotoxin lavage (data not shown). Total protein and albumin increased 2- to 3-fold in the endotoxin BAL compared with control (endotoxin effect, p 0.0011 and endotoxin–time interaction, both p NS) (Figure 1E and 1F).

Figure 2. Changes in acute-phase cytokines and proinflammatory activity at 2, 6, 24, and 48 h after saline (open bars) and endotoxin (shaded bar) challenge of lung segments. Endotoxin was associated with increases in (A) TNF- bioactivity (endotoxin effect, p 0.004; endotoxin–time interaction, p NS), (B) proinflammatory activity (p 0.0025 and p 0.029), (C) TNFR1 (p55) (p 0.0001 and p NS), (D) TNFR2 (p75) (p 0.0001 and p NS), (E) IL-1 (p 0.0001 and p 0.01), (F) IL-1ra (p 0.0001 and p 0.006), (G) IL-6 (p 0.0001 and p 0.0001), and (H) G-CSF (p 0.0001 and p 0.02). Data presented as mean and SEM and analyzed by ANOVA.

1594


TNF- bioactivity (Figure 2A) was detected as early as 2 h, was greatest at 6 h, and then was absent at 24 h and 48 h (endotoxin effect, p 0.004 and endotoxin–time interaction, p NS). TNF- antigen levels paralleled TNF- bioactivity (control versus endotoxin: 2 h, 2 1 versus 215 185; 6 h, 0.33 0.2 versus 322 77; 24 h, 1 1 versus 9 4; 48 h, 0.03 0.1 versus 1.5 0.5, pg/ml, p 0.0003 and p 0.02). Proinflammatory activity (e.g. measured as upregulation of intercellular adhesion molecule-1 on human alveolar type II-like A549 cells) increased at 2 h and 6 h after endotoxin and then returned to baseline values at 24 h and 48 h (p 0.025 and p 0.029) (Figure 2B). TNF receptor type 1 (TNFR)1 (p55) rose 3-fold and type 2 TNFR2 (p75) rose 10-fold at 6 h and both remained elevated at 24 h and 48 h (endotoxin effect: both p 0.0001 and endotoxin–time interaction, p NS) (Figure 2C and 2D). A small rise in interleukin (IL)-1 occurred at 2 h and 6 h (p 0.0001 and p 0.01) accompanied by a rise in IL-1 receptor antagonist (IL-1ra) (p 0.0001 and p 0.006), which was approximately 100-fold in excess of IL-1. Both IL-1 and IL-1ra returned to baseline values by 24 h and 48 h (Figure 2E and 2F). IL-6 rose significantly at 2 h and 6 h after endotoxin (endotoxin effect, p 0.0001 and endotoxin–time interaction, p 0.0001) and then returned to baseline values at 24 h (Figure 2G). No IL-10 was detected at 2, 6, 24, or 48 h. G-CSF was significantly elevated at 6 h and 24 h and returned to basal levels at 48 h (p 0.001 and p 0.02) (Figure 2H). Two C-X-C chemokines, IL-8, epithelial neutrophil activating protein (ENA)-78, but not growth regulated peptide (GRO)- and three C–C

VOL 163

2001

chemokines, monocyte chemotactic protein (MCP)-1, macrophage inflammatory protein (MIP)-1 and MIP-1, rose acutely at 6 h and then all returned to baseline values at 24 h (all chemokines, endotoxin effect, p 0.0007 and endotoxin–time interaction, p 0.02) (Figure 3A and 3F). Markers of inflammatory cell activation were detected in the BAL (Figure 4A and 4D). L-selectin was elevated at 6 h and remained greater than control values at 24 h and 48 h (p 0.0001 and p NS). Gelatinase-B was elevated at 6 h (p 0.009 and p 0.01) and returned to baseline levels at 24 h and 48 h. Both myeloperoxidase and lactoferrin levels were significantly elevated at 6 h and remained greater than control levels at 24 h and 48 h (endotoxin effect, both p 0.0001 and endotoxin–time interaction, both p NS). Systemic inflammatory responses following local endotoxin instillation included the development of a mild leukocytosis with a predominance of neutrophils (p 0.0001) (Table 2). During this period, blood levels of G-CSF, IL-1ra, and IL-6 increased 3- to 6-fold from baseline (all p 0.0001) whereas levels of TNF- antigen did not change (p NS) (Table 2). At 24 h serum C- reactive protein rose 8-fold (p 0.0003) and serum amyloid A had risen almost 30-fold (p 0.0008) (Table 2).

DISCUSSION The study of human innate immunity in vivo is limited to three models. Acute systemic inflammation to endotoxin has been studied extensively using intravenous challenges (6). Whole lung exposure to inhaled endotoxin has been used to study its

Figure 3. Changes in BAL C–C and C–X–C chemokines at 2, 6, 24, and 48 h after saline (open bar) and endotoxin (shaded bar) challenge of lung segments. Endotoxin was associated with increased levels of (A) IL-8 (endotoxin effect, p 0.0007; endotoxin–time interaction, p 0.02), (B) MCP-1 (p 0.0001 and p 0.0002), (C) GRO- (p NS), (D) MIP-1 (p 0.0001 and 0.002), (E) ENA-78 (p 0.0007 and p 0.02), and MIP-1 (p 0.0001 and p 0.0005). Data presented as mean and SEM and analyzed by ANOVA.

1595

O’Grady, Preas, Pugin, et al.: Endotoxin Effects in Human Lung

Figure 4. Changes in markers of cell activation in BAL at 2, 6, 24, and 48 h after saline (open bar) and endotoxin (shaded bar) challenge of lung segments. Endotoxin was associated with increases in (A) L-selectin (endotoxin effect, p 0.0001; endotoxin–time interaction, p NS), (B) gelatinase-B (p 0.0009 and p 0.01), (C) myeloperoxidase (p 0.0001 and p NS), and (D) lactoferrin (p 0.0001 and p NS). Data presented as mean and SEM and analyzed by ANOVA.

contribution to dust-related occupational lung disease (10– 12). A third model, skin blister windows, allows assessment of neutrophil exudation and phagocytic activity (18). Each of these models has features that address specific correlates of human pathophysiology. We have developed a model of localized inflammation to endotoxin to understand the requirements necessary to initiate, amplify, and resolve endotoxin-induced inflammation in the lung. The dose of endotoxin used in the current model is 100 times smaller than doses used in inhalation models of endotoxin-induced lung inflammation (e.g., 0.30 g instilled over 2–3 s versus 30–60 g inhaled over 60 min) and a dose of 1 ng/kg (10 endotoxin units/kg) is sufficient to initiate this local inflammatory response (10–12). We show that endotoxin instillation into a pulmonary segment results in localized inflammation that occurs in two phases. The early phase from 2 h to 6 h is composed of a neutrophil influx with elevated levels of cytokines and chemokines. The later phase, from 24 h to 48 h, is manifested by a persistence of elicited neutrophils, and increased numbers of macrophages, monocytes, and lymphocytes. During this latter period most of the cytokines and chemokines measured return to basal levels. Throughout both periods, permeability to albumin is increased. Inflammatory mediator profiles and clearance differ from previously described systemic inflammatory responses after intravenous endotoxin (6, 7). Following intravenous endotoxin, most acute phase mediators (e.g., TNF-, TNFR, IL-1, IL-1ra, IL-6, IL-8, MIP-1, MCP-1, GRO-, G-CSF, and IL-10) are maximum in blood after 1 h to 3 h and returned to basal levels by 6 h to 8 h (6, 7). In contrast, some of these mediators (e.g., TNF-, IL-1, IL-1ra, IL-6, IL-8, MIP-1, and MCP-1) rise in BAL at 2 h, are greatest at 6 h, and return to control levels by 24 h whereas others remain elevated at 24 h (e.g., G-CSF) and 48 h (e.g., TNFR, lactoferrin, and myeloperoxidase). Two mediators of the systemic response to intravenous endotoxin (e.g., IL-10 and GRO-) do not increase during the local lung inflammatory response to endotoxin. These data show that endotoxin-induced inflammation in the lung has a unique qualitative and temporal pattern of inflammatory mediator and cell marker expression compared with systemic endotoxin responses.

The BAL following endotoxin challenge is highly biologically active. TNF-, and to a lesser extent IL-1, propagated the inflammatory response initiated by endotoxin. Despite the dilution factor of the BAL and elevated levels of soluble TNFR, TNF- bioactivity was detectable during the first 6 h suggesting very high local lung concentrations of TNF-. This increase in TNF bioactivity paralleled the increase in proinflammatory activity, as measured by the upregulation of ICAM-1 on human type II-like A549 epithelial cells. The contribution of IL-1 to the proinflammatory activity was likely diminished by the 100-fold excess of IL-1ra to IL-1. In contrast to acute respiratory distress syndrome (ARDS), where IL-1 is the predominant proinflammatory molecule, the proinflammatory activity following instilled endotoxin is similar to conditioned supernatants from endotoxin-treated alveolar macrophages (15).

TABLE 2. BLOOD INFLAMMATORY RESPONSES IN 28 SUBJECTS FOLLOWING BRONCHIAL ENDOTOXIN INSTILLATION*

Total leukocyte, 109 cells/L Neutrophil, % Bands, % Lymphocytes, % Monocytes, % G-CSF, pg/ml IL-6, pg/ml IL-1ra, pg/ml TNF-, pg/ml C-reactive protein (mg/L) Serum amyloid A ( g/ml)

Baseline

Maximum Change (time)

6.11 0.23 52.0 2.2 1.79 0.41 35.0 2.0 7.9 0.49 11.0 2.9 0.4 0.1 189.9 16.3 1.6 0.3 1.72 0.55 10.45 2.32

10.56 0.051 (8 h)† 70.7 2.10 (8 h)† 4.5 0.90 (8 h)† 18.47 1.90 (8 h)† 5.2 0.50 (8 h)† 62.6 9.2 (8 h)† 12.8 2.8 (6 h)† 509.9 73.7 (8 h)† 1.37 0.3 (6 h) 14.79 2.90 (24 h)† 380.00 142.00 (24 h)†

Definition of abbreviations: G-CSF granulocyte–colony-stimulating factor; IL-6 interleukin 6; IL-1ra interleukin 1 receptor antagonist; TNF- tumor necrosis factor-. * Endotoxin dose instilled 4 ng/kg. † p 0.008 and reflects change over study period by ANOVA. These endotoxin-related effects occurred independent of the time of the second BAL (endotoxin-time interaction, p NS). Serum G-CSF, IL-6, IL-1r, and TNF- were measured at 0, 2, 5, 6, 8, and 24 h. C-reactive protein and serum amyloid A were measured at 0 and 24 h.

1596


Several factors may have contributed to attenuation of the inflammatory response in the lung. Both TNFR1 and TNFR2 were elevated simultaneously with TNF- bioactivity at 6 h and remained elevated at 24 h and 48 h, after TNF- bioactivity had abated and TNF- antigen was no longer detected. However, the effects of soluble TNFR on local inflammatory responses depend on the relative proportion of receptor to ligand. High concentrations of TNFR can interfere with TNF binding to cells whereas low concentrations of TNFR can stabilize the TNF trimer and enhance TNF- presentation to cells. This may prolong TNF effects in the lung compartment where clearance rates are slow (19). Further, transgenic mice producing human TNFR2 (p75) develop severe systemic inflammation including the lung, which evolves independent of TNFR ligands suggesting that under some conditions, soluble TNFR2 (p75) may serve as an inflammatory mediator and contribute to tissue inflammation (20). Three additional measured factors may have contributed to the abatement of the local inflammatory response. IL-6 suppresses transcription of IL-1 and TNF- in mononuclear cells and induces the synthesis of IL-1ra by macrophages (21, 22). Coadministration of IL-6 with intratracheal endotoxin to rats decreases lung TNF levels and neutrophil exudation (23). G-CSF may have contributed to the antiinflammatory response of the lung as well. In addition to mobilizing and activating neutrophils, G-CSF-treated subjects release less TNF- and more IL-1ra and TNFR in endotoxin-stimulated whole blood assays (24). MCP-1, which was also elevated at 6 h, has antiinflammatory effects in endotoxin-treated mice by suppressing TNF release and enhancing antiinflammatory cytokine expression (25). IL-10, a potent antiinflammatory cytokine, did not contribute to the attenuation of the pulmonary inflammatory response to endotoxin. We found no IL-10 in the BAL of the saline or endotoxin-challenged segments at any time point. Recovery of recombinant IL-10 incubated with BAL was 90% (supplementary methods) suggesting that interfering molecules did not account for the absence of IL-10. In contrast, IL-10 levels rise rapidly in humans after intravenous endotoxin challenge (7). Conflicting data exist regarding BAL levels of IL-10 in normal subjects. High BAL levels of airway epitheliumderived IL-10 have been described in healthy controls whereas IL-10 is absent in the BAL of ventilated patients and in mild asthma (26–28). Further, IL-10 has been noted to be either absent or present in supernatants of human alveolar macrophages stimulated with endotoxin (26, 28–32). However, mice challenged with intranasal endotoxin develop lung inflammation but have no IL-10 in whole lung preparations or BAL and when their alveolar macrophages are stimulated with endotoxin, no IL-10 protein or mRNA is expressed (33). These investigators suggest that the lung microenvironment may suppress IL-10 induction by alveolar macrophages or that these cells acquire a unique phenotype compared with other types of tissue macrophages (33). The recruitment of neutrophils to the endotoxin-challenged segment coincided with increased BAL levels of C–X–C chemokines, IL-8 and ENA-78, at 2 h and 6 h. However, increased neutrophils persisted in the 24-h BAL after these chemokines had returned to basal levels, suggesting the presence of other unmeasured neutrophil chemoattractants and the effects of trophic factors such as G-CSF, which remained elevated in the BAL at 24 h. The fall in neutrophil numbers at 48 h suggests an end to the neutrophil influx with clearance of apoptotic cells by alveolar macrophages. GRO-, a neutrophil chemoattractant found in the BAL of patients with ARDS, pulmonary infection, or ozone exposure, did not change after

VOL 163

2001

pulmonary endotoxin challenge (34, 35). Yet levels of GRO- as well as other chemokines (e.g., IL-8 and MCP, MIP-1) rise in the blood after intravenous endotoxin suggesting that unique expression patterns of chemokines may occur in tissue and blood in response to endotoxin (7, 36). In contrast to the temporal relation of neutrophils and C–X–C chemokines, the greatest influx of mononuclear cells occurred at 24 h–48 h. The C–C chemokines, MCP-1, MIP-1, and MIP-1, attract distinct populations of monocytes and lymphocytes and were elevated at 6 h with the initiation of the mononuclear cell influx into the lung (37). However, BAL monocytes and lymphocytes continue to increase at 24 h and 48 h after these C–C chemokines returned to basal levels, suggesting the presence of other mononuclear attractants at these later time points. One class of molecules that could explain this persistent elevation of mononuclear cells is derived from neutrophil granules. Neutrophils stimulated by IL-8 or other stimulants release chemoattractants (i.e., azurocidin/CAP37, -defensins) that mediate T cell and monocyte accumulation at sites of inflammation (38). These mediators in conjunction with C–C chemokines recruit adaptive immune cells to sites of inflammation and form an important link between innate and acquired immune responses (2, 3). Other markers of inflammatory cell receptor shedding or degranulation were detected in BAL after endotoxin challenge. Soluble L-selectin, an adhesion molecule shed from neutrophils and mononuclear cell, may function as a local buffer system inhibiting leukocyte–cell interactions at sites of inflammation (39). After contact with inflammatory mediators or chemokines, gelatinase-B is released within minutes by neutrophils and within hours by mononuclear cells (16). This metalloproteinase facilitates cell transmigration by degrading the extracellular matrix. BAL levels in the current study were maximum at 6 h whereas after intravenous endotoxin serum levels are greatest at 1 h to 3 h (16). Both myeloperoxidase, a primary neutrophil granule constituent that catalyzes the production of hypochlorous acid, and lactoferrin, an iron binding molecule found in the secondary specific granules of neutrophils, were elevated as late as 48 h postendotoxin challenge (40). These increases paralleled the increase in neutrophils in the BAL and represent indices of neutrophil activation and possibly delayed clearance of the granule constituents from the air space. The integrity of the alveolar epithelial–endothelial interface is a major variable in limiting the systemic effects of local lung inflammation. Humans challenged with intrabronchial leukotriene B4 develop a neutrophil influx into the lung without a change in protein permeability (41). Animals given intratracheal endotoxin develop high BAL TNF levels without increases in blood levels, but when the alveolar–capillary interface is injured, TNF placed or produced in the lung can be detected in the circulation or lung perfusate (42–44). However, pulmonary endotoxin exposure to humans is associated with systemic inflammatory responses. Following endotoxin inhalation, subjects with mild asthma or normal subjects develop a leukocytosis, small increases in blood TNF-, and increases in CRP (14, 45). In the current study, the changes in local permeability to albumin and a high concentration gradient would favor the lung as the source for the increases in blood G-CSF, IL-1ra, and IL-6, which may have contributed to the rise in acute phase proteins and the leukocytosis. However, blood TNF- levels did not change despite high BAL levels at 6 h, suggesting that it may be cleared rapidly from the systemic circulation by acute phase reactants or that inflammatory mediators do not leak uniformly into the systemic circulation.

O’Grady, Preas, Pugin, et al.: Endotoxin Effects in Human Lung

Exposure of lung segments to small amounts of endotoxin rapidly initiates a sequence of inflammatory responses over 48 h characterized by an early persistent neutrophil influx followed by a later influx of mononuclear cells. Although most inflammatory mediators are released within 6 h and return to basal levels by 24 h, elevation of some inflammatory molecules and markers persists (e.g., G-CSF, TNFR, myeloperoxidase, and lactoferrin) at 24 h and 48 h. Our findings highlight important differences in the temporal sequence and character of the inflammatory response to endotoxin in the lung compared with the systemic circulation. This model of local inflammatory responses provides a means to delineate early events in the initiation and resolution of innate immune responses in the lung. References 1. Medzhitov R, Janeway CJ. Innate immunity. N Engl J Med 2000;343: 338–344. 2. Fearon DT, Locksley RM. The instructive role of innate immunity in the acquired immune response. Science 1996;272:50–53. 3. Medzhitov R, Janeway CA, Jr. Innate immunity: impact on the adaptive immune response. Curr Opin Immunol 1997;9:4–9. 4. Hurley J, Levin J. The relevance of endotoxin detection in sepsis. In: Brade H, Opal SM, Vogel SN, Morrison DC, editors. Endotoxin in health and disease, 1st ed. New York: Marcel Dekker, Inc.; 1999. p. 841–854. 5. Pugin J, Auckenthaler R, Delaspre O, van Gessel E, Suter PM. Rapid diagnosis of gram negative pneumonia by assay of endotoxin in bronchoalveolar lavage fluid. Thorax 1992;47:547–549. 6. Suffredini AF, O’Grady NP. Pathophysiologic responses to endotoxin in humans. In: Braude H, Opal SM, Vogel SN, Morrison DC, editors. Endotoxin in health and disease, 1st ed. New York: Marcel Dekker, Inc.; 1999. p. 817–830. 7. Suffredini AF, Reda D, Banks SM, Tropea M, Agosti JM, Miller R. Effects of recombinant dimeric TNF receptor on human inflammatory responses following intravenous endotoxin administration. J Immunol 1995;155:5038–5045. 8. Suffredini AF, Shelhamer JH, Neumann RD, Brenner M, Baltaro RJ, Parrillo JE. Pulmonary and oxygen transport effects of intravenously administered endotoxin in normal humans. Am Rev Respir Dis 1992; 145:1398–1403. 9. Boujoukos AJ, Martich GD, Supinski E, Suffredini AF. Compartmentalization of the acute cytokine response in humans after intravenous endotoxin administration. J Appl Physiol 1993;74:3027–3033. 10. Rylander R, Bake B, Fischer JJ, Helander IM. Pulmonary function and symptoms after inhalation of endotoxin. Am Rev Respir Dis 1989;140: 981–986. 11. Von Essen S. The role of endotoxin in grain dust exposure and airway obstruction. Curr Opin Pulmon Med 1997;3:198–202. 12. Jagielo PJ, Thorne PS, Watt JL, Frees KL, Quinn TJ, Schwartz DA. Grain dust and endotoxin inhalation challenges produce similar inflammatory responses in normal subjects. Chest 1996;110:263–270. 13. Deetz DC, Jagielo PJ, Quinn TJ, Thorne PS, Bleuer SA, Schwartz DA. The kinetics of grain dust-induced inflammation of the lower respiratory tract. Am J Respir Crit Care Med 1997;155:254–259. 14. Wesselius LJ, Nelson ME, Bailey K, O’Brien-Ladner AR. Rapid lung cytokine accumulation and neutrophil recruitment after lipopolysaccharide inhalation by cigarette smokers and nonsmokers. J Lab Clin Med 1997;129:106–114. 15. Pugin J, Ricou B, Steinberg KP, Suter PM, Martin TR. Proinflammatory activity in bronchoalveolar lavage fluids from patients with ARDS, a prominent role for interleukin-1. Am J Respir Crit Care Med 1996; 153:1850–1856. 16. Pugin J, Widmer MC, Kossodo S, Liang CM, Preas HL, Suffredini AF. Human neutrophils secrete gelatinase B in vitro and in vivo in response to endotoxin and proinflammatory mediators. Am J Respir Cell Mol Biol 1999;20:458–464. 17. Scheffe H. The analysis of variance. New York: John Wiley & Sons; 1959. 18. Fiuza C, Salcedo M, Clemente G, Tellado JM. In vivo neutrophil dysfunction in cirrhotic patients with advanced liver disease. J Infect Dis 2000;182:526–533. 19. Aderka D, Engelmann H, Maor Y, Brakebusch C, Wallach D. Stabilization of the bioactivity of tumor necrosis factor by its soluble receptors. J Exp Med 1992;175:323–329.

1597 20. Douni E, Kollias G. A critical role of the p75 tumor necrosis factor receptor (p75TNF-R) in organ inflammation independent of TNF, lymphotoxin alpha, or the p55TNF-R. J Exp Med 1998;188:1343–1352. 21. Schindler R, Mancilla J, Endres S, Ghorbani R, Clark SC, Dinarello CA. Correlations and interactions in the production of interleukin-6 (IL-6), IL-1, and tumor necrosis factor (TNF) in human blood mononuclear cells: IL-6 suppresses IL-1 and TNF. Blood 1990;75:40–47. 22. Tilg H, Trehu E, Atkins MB, Dinarello CA, Mier JW. Interleukin-6 (IL-6) as an anti-inflammatory cytokine: induction of circulating IL-1 receptor antagonist and soluble tumor necrosis factor receptor p55. Blood 1994;83:113–118. 23. Ulich TR, Yin S, Guo K, Yi ES, Remick D, del Castillo J. Intratracheal injection of endotoxin and cytokines: II. Interleukin-6 and transforming growth factor beta inhibit acute inflammation. Am J Pathol 1991;138:1097–1101. 24. Hartung T, Docke WD, Gantner F, Krieger G, Sauer A, Stevens P, Volk HD, Wendel A. Effect of granulocyte colony-stimulating factor treatment on ex vivo blood cytokine response in human volunteers. Blood 1995;85:2482–2489. 25. Zisman DA, Kunkel SL, Strieter RM, Tsai WC, Bucknell K, Wilkowski J, Standiford TJ. MCP-1 protects mice in lethal endotoxemia. J Clin Invest 1997;99:2832–2836. 26. Bonfield TL, Konstan MW, Burfeind P, Panuska JR, Hilliard JB, Berger M. Normal bronchial epithelial cells constitutively produce the antiinflammatory cytokine interleukin-10, which is downregulated in cystic fibrosis. Am J Respir Cell Mol Biol 1995;13:257–261. 27. Donnelly SC, Strieter RM, Reid PT, Kunkel SL, Burdick MD, Armstrong I, Mackenzie A, Haslett C. The association between mortality rates and decreased concentrations of interleukin-10 and interleukin-1 receptor antagonist in the lung fluids of patients with the adult respiratory distress syndrome. Ann Intern Med 1996;125:191–196. 28. Magnan A, van Pee D, Bongrand P, Vervloet D. Alveolar macrophage interleukin (IL)-10 and IL-12 production in atopic asthma. Allergy 1998;53:1092–1095. 29. Thomassen MJ, Divis LT, Fisher CJ. Regulation of human alveolar macrophage inflammatory cytokine production by interleukin-10. Clin Immunol Immunopathol 1996;80:321–324. 30. Berkman N, John M, Roesems G, Jose PJ, Barnes PJ, Chung KF. Inhibition of macrophage inflammatory protein-1 alpha expression by IL-10. Differential sensitivities in human blood monocytes and alveolar macrophages. J Immunol 1995;155:4412–4418. 31. John M, Lim S, Seybold J, Jose P, Robichaud A, O’Connor B, Barnes PJ, Chung KF. Inhaled corticosteroids increase interleukin-10 but reduce macrophage inflammatory protein-1alpha, granulocyte-macrophage colony-stimulating factor, and interferon-gamma release from alveolar macrophages in asthma. Am J Respir Crit Care Med 1998; 157:256–262. 32. Boehringer N, Hagens G, Songeon F, Isler P, Nicod LP. Differential regulation of tumor necrosing factor-alpha (TNF-alpha) and interleukin-10 (IL-10) secretion by protein kinase and phosphatase inhibitors in human alveolar macrophages. Eur Cytokine Netw 1999;10:211–218. 33. Salez L, Singer M, Balloy V, Creminon C, Chignard M. Lack of IL-10 synthesis by murine alveolar macrophages upon lipopolysaccharide exposure. Comparison with peritoneal macrophages. J Leukocyte Biol 2000;67:545–552. 34. Villard J, Dayer-Pastore F, Hamacher J, Aubert JD, Schlegel-Haueter S, Nicod LP. GRO alpha and interleukin-8 in Pneumocystis carinii or bacterial pneumonia and adult respiratory distress syndrome. Am J Respir Crit Care Med 1995;152:1549–1554. 35. Krishna MT, Madden J, Teran LM, Biscione GL, Lau LC, Withers NJ, Sandstrom T, Mudway I, Kelly FJ, Walls A, et al. Effects of 0.2 ppm ozone on biomarkers of inflammation in bronchoalveolar lavage fluid and bronchial mucosa of healthy subjects. Eur Respir J 1998;11:1294– 1300. 36. Kopydlowski KM, Salkowski CA, Cody MJ, van Rooijen N, Major J, Hamilton TA, Vogel SN. Regulation of macrophage chemokine expression by lipopolysaccharide in vitro and in vivo. J Immunol 1999; 163:1537–1544. 37. Taub DD, Proost P, Murphy WJ, Anver M, Longo DL, van Damme J, Oppenheim JJ. Monocyte chemotactic protein-1 (MCP-1), -2, and -3 are chemotactic for human T lymphocytes. J Clin Invest 1995;95:1370– 1376. 38. Chertov O, Michiel DF, Xu L, Wang JM, Tani K, Murphy WJ, Longo DL, Taub DD, Oppenheim JJ. Identification of defensin-1, defensin-2, and CAP37/azurocidin as T-cell chemoattractant proteins released from interleukin-8-stimulated neutrophils. J Biol Chem 1996;271:2935–2940. 39. Schleiffenbaum B, Spertini O, Tedder TF. Soluble L-selectin is present

1598


in human plasma at high levels and retains functional activity. J Cell Biol 1992;119:229–238. 40. Pittet JF, Mackersie RC, Martin TR, Matthay MA. Biological markers of acute lung injury: prognostic and pathogenetic significance. Am J Respir Crit Care Med 1997;155:1187–1205. 41. Martin TR, Pistorese BP, Chi EY, Goodman RB, Matthay MA. Effects of leukotriene B4 in the human lung. Recruitment of neutrophils into the alveolar spaces without a change in protein permeability. J Clin Invest 1989;84:1609–1619. 42. Nelson S, Bagby GJ, Bainton BG, Wilson LA, Thompson JJ, Summer WR. Compartmentalization of intraalveolar and systemic lipopolysac-

VOL 163

2001

charide-induced tumor necrosis factor and the pulmonary inflammatory response. J Infect Dis 1989;159:189–194. 43. Tutor JD, Mason CM, Dobard E, Beckerman RC, Summer WR, Nelson S. Loss of compartmentalization of alveolar tumor necrosis factor after lung injury. Am J Respir Crit Care Med 1994;149:1107–1111. 44. Kurahashi K, Kajikawa O, Sawa T, Ohara M, Gropper MA, Frank DW, Martin TR, Wiener-Kronish JP. Pathogenesis of septic shock in Pseudomonas aeruginosa pneumonia. J Clin Invest 1999;104:743–750. 45. Michel O, Ginanni R, Le Bon B, Content J, Duchateau J, Sergysels R. Inflammatory response to acute inhalation of endotoxin in asthmatic patients. Am Rev Respir Dis 1992;146:352–357.