See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/261730491
Transcompartmental Inflammatory Responses in Humans: IV Versus Endobronchial Administration of Endotoxin ARTICLE in CRITICAL CARE MEDICINE · APRIL 2014 Impact Factor: 6.31
READS
14
7 AUTHORS, INCLUDING: Ronan Martin Griffin Berg
Kevin A Evans
Frederiksberg Hospital
University of South Wales
67 PUBLICATIONS 512 CITATIONS
15 PUBLICATIONS 194 CITATIONS
SEE PROFILE
SEE PROFILE
Lars Konge
Kirsten Møller
University of Copenhagen
Rigshospitalet
109 PUBLICATIONS 294 CITATIONS
159 PUBLICATIONS 3,548 CITATIONS
SEE PROFILE
SEE PROFILE
Available from: Ronni Th. R. Plovsing Retrieved on: 03 February 2016
Transcompartmental Inflammatory Responses in Humans: IV Versus Endobronchial Administration of Endotoxin Ronni R. Plovsing, MD1; Ronan M. G. Berg, MD2; Kevin A. Evans, PhD3; Lars Konge, MD, PhD4; Martin Iversen, MD, DMSc5; Peter Garred, MD, DMSc6; Kirsten Møller, MD, PhD, DMSc2,7
Objectives: Transcompartmental signaling during early inflammation may lead to propagation of disease to other organs. The time course and the mechanisms involved are still poorly understood. We aimed at comparing acute transcompartmental inflammatory responses in humans following lipopolysaccharide-induced pulmonary and systemic inflammation. Design: Randomized, double-blind, placebo-controlled, crossover study.
Department of Intensive Care, University Hospital Rigshospitalet, Copenhagen Ø, Denmark. 2 Centre of Inflammation and Metabolism, Department of Infectious Diseases M7641, University Hospital Rigshospitalet, Copenhagen Ø, Denmark. 3 Neurovascular Research Laboratory, Faculty of Health, Science and Sport, University of Glamorgan, South Wales, United Kingdom. 4 Centre for Clinical Education, University of Copenhagen and the Capital Region of Denmark, Copenhagen, Denmark. 5 The Heart Centre, Department of Lung Transplantation, University Hospital Rigshospitalet, Copenhagen Ø, Denmark. 6 Laboratory of Molecular Medicine, Department of Clinical Immunology M7631, University Hospital Rigshospitalet, Copenhagen Ø, Denmark. 7 Neurointensive Care Unit 2093, Department of Neuroanesthesiology, University Hospital Rigshospitalet, Copenhagen Ø, Denmark. Drs. Plovsing and Møller contributed to study design; Drs. Plovsing, Berg, Evans, Konge, and Iversen contributed to data collection; Drs. Plovsing, Berg, and Møller contributed to data analysis; Drs. Plovsing, Berg, Konge, Iversen, Garred, and Møller contributed to data interpretation; and Dr. Plovsing drafted the article. All authors made critical revisions and read and approved the final article. Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal’s website (http://journals.lww.com/ ccmjournal). Dr. Plovsing’s institution received grant support from The National Research Council (09-064930/FFS), University Hospital Rigshospitalet, The Foundation of Merchant Jakob Ehrenreich and Grete Ehrenreich, The Beckett Foundation, The Gangsted Foundation, and The Hørslev Foundation (all grants in support for preparation and conduction of the present study). The remaining authors have disclosed that they do not have any potential conflicts of interest. For information regarding this article, E-mail:
[email protected] Copyright © 2014 by the Society of Critical Care Medicine and Lippincott Williams & Wilkins DOI: 10.1097/CCM.0000000000000320 1
Critical Care Medicine
Setting: ICU. Subjects: Healthy male volunteers. Interventions: Fifteen volunteers (mean age, 23; sd, 2 yr) received Escherichia coli endotoxin (lipopolysaccharide, 4 ng/kg) IV or endobronchially on two different study days. Groups were evaluated by bronchoalveolar lavage at baseline (0 hr) and 2, 4, 6, 8, or 24 hours postchallenge. Cardiorespiratory variables were continuously recorded throughout the study day, and plasma and bronchoalveolar lavage fluid markers of inflammation were measured. Measurements and Main Results: IV endotoxin elicited a systemic inflammatory response with a time-dependent increase and peak in tumor necrosis factor-α, interleukin-6, and leukocyte counts (all p < 0.001). Furthermore, a delayed (6–8 hr) increase in bronchoalveolar lavage fluid interleukin-6 concentration (p < 0.001) and alveolar leukocyte count (p = 0.03) and a minor increase in bronchoalveolar lavage fluid tumor necrosis factor-α were observed (p = 0.06). Endobronchial endotoxin was followed by progressive alveolar neutrocytosis and increased bronchoalveolar lavage fluid tumor necrosis factor-α, interleukin-6, and albumin (all p < 0.001); a systemic inflammatory response was observed after 2–4 hours, with no change in plasma tumor necrosis factor-α. Conclusions: Acute lung or systemic inflammation in humans is followed by a transcompartmental proinflammatory response, the degree and differential kinetics of which suggests that the propagation of inflammation may depend on the primary site of injury. (Crit Care Med 2014; XX:00–00) Key Words: acute lung injury; bronchoalveolar lavage; human experimentation; inflammation; lipopolysaccharide; sepsis; systemic inflammatory response syndrome
A
cute respiratory distress syndrome (ARDS) and sepsis are heterogeneous clinical conditions with a high mortality (1–3). Both diseases are associated with a complex interplay of different inflammatory modulators and cell types and are accompanied by a markedly imbalanced cytokine response (4–11). This, in turn, leads to an increase in epithelial and vascular permeability that may favor the www.ccmjournal.org
1
Copyright (c) Society of Critical Care Medicine and Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited
Plovsing et al
transfer of inflammatory mediators from one compartment to another (5, 9, 12). It has been suggested that locally released inflammatory mediators pass from the lungs into the bloodstream and subsequently trigger systemic inflammation (8, 13). This phenomenon may contribute to the development of severe systemic inflammation and the extrapulmonary organ dysfunction that occur in a large number of patients (5, 12). Conversely, severe systemic inflammation may possibly lead to ARDS through the transfer of inflammatory mediators from the bloodstream to the lungs. The time course from the initial insult to propagation of inflammation and the mechanisms involved in this putative transmission of disease between compartments have yet to be established. The human endotoxin model may be used to study early inflammatory mechanisms (14, 15). Administration of lipopolysaccharide (LPS), a well-characterized p athogen-associated molecular pattern (PAMP) (16), leads to toll-like receptor 4 (TLR4)-mediated activation of nuclear factor-κB-dependent proinflammatory gene expression. This may be analogous to that which occurs during some cases of acute inflammation in the clinical setting (17–21). Several studies have been published on the compartmental inflammatory effects of LPS administration in healthy volunteers and animals (18, 22–30). However, the goal of our study was to provide a direct comparison of the acute inflammatory responses induced by IV and endobronchial administration of equal endotoxin doses, which was given to volunteers in a randomized, double-blind, placebo-controlled, crossover design. We hypothesized that systemic injection of endotoxin would
elicit a secondary (transcompartmental) pulmonary inflammatory response and that the degree and kinetics of inflammation would be similar to the systemic inflammatory response induced by lung inflammation.
MATERIALS AND METHODS Subjects This randomized, double-blind, placebo-controlled, crossover study was approved by the Research Ethical Committee of Copenhagen and Frederiksberg Municipalities, Denmark (protocol no. H-2-2009-131), and performed in accordance with the Declaration of Helsinki. Oral and written informed consent was obtained from all volunteers prior to participation. Fifteen healthy nonsmoking male volunteers were recruited by advertising. All had an unremarkable medical history. A thorough physical examination, 12-lead electrocardiogram, spirometry, and routine blood samples did not reveal any abnormality and no signs of infection occurred within 4 weeks ahead of the study day. Study Design All volunteers participated on two study days, which were separated by at least 24 days and included the following interventions (Fig. 1): ●●
●●
Study day A: Escherichia coli LPS 4 ng/kg IV + saline endobronchially. Study day B: LPS 4 ng/kg endobronchially + saline IV. Seven volunteers were randomized to the order AB, and another eight were randomized to BA; volumes of saline were identical to those of LPS (referred to as “endotoxin” hereafter).
Figure 1. Study flowchart. Volunteers were challenged with Escherichia coli lipopolysaccharide (LPS, 4 ng/ kg) as an IV or endobronchial bolus on the first study day, whereas the opposite intervention was given on the second study day (n = 15 in each study group). Bilateral bronchoalveolar lavage (BAL) was performed at 2, 4, 6, 8, or 24 hr postchallenge (n = 3 at each time point).
2
www.ccmjournal.org
Study Day Volunteers (mean age, 23 ± 2 yr; mean height, 186 ± 5 cm; mean weight, 80 ± 9 kg) reported to the ICU at 6:45 am following an overnight fast. They were catheterized with an arterial catheter in the left radial artery following local anesthesia (lidocaine, 20 mg/mL), and a peripheral venous catheter was inserted in the antecubital region (for injection of endotoxin or placebo). After placement of intravascular catheters, a bronchoalveolar lavage (BAL) was done, which was immediately followed by the interventions described above; endobronchial and IV injections were performed simultaneously. XXX 2014 • Volume 42 • Number XXX
Copyright (c) Society of Critical Care Medicine and Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited
Clinical Investigation
Volunteers were randomized in an unblinded fashion to undergo a bilateral BAL at 2, 4, 6, 8, or 24 hours postintervention, respectively, on both study days (n = 3 in all groups). Thus, all volunteers underwent BAL twice on each study day. Invasive arterial blood pressure, heart rate, respiratory rate, and arterial pulse oximetry were continuously recorded at 1 kHz throughout the first 0–8 hr using an analog-to-digital converter (PowerLab 4/30, ADInstruments, Oxford, United Kingdom) interfaced with a personal computer and afterward recalculated as the mean during 30-minute epochs. Rectal temperature and arterial blood samples were obtained hourly (0–8 hr), whereas venous blood samples were drawn at 16 and 24 hours, respectively. Clinical signs and subjective symptoms of distress were quantified throughout the study day. Volunteers were allowed to drink tap water ad libitum, but were otherwise fasting until removal of the arterial catheter after 8 hours. They were discharged after 24–26 hours upon complete alleviation of all symptoms and normalization of vital variables. Volunteers received financial compensation as approved by the Research Ethical Committee.
Statistical Analysis A linear mixed model was used for evaluation of r epeated-measures data; baseline values were included as a covariate for comparison of between intervention group effects (endotoxin-time interaction), while baseline was compared with subsequent time points for within intervention group effects (endotoxin effect). If an overall intervention group effect was identified, post hoc pairwise comparisons for the estimated differences of the least-squares means were evaluated using the Tukey-Kramer adjustment for multiple comparisons. Model assumptions were assessed by residual diagnostics, that is, normality and variance homogeneity of conditional studentized residuals. Data were logarithmically transformed to attain normal distribution and variance homogeneity if needed and log-transformed variables were plotted on the log-scale in order to achieve symmetric CIs. All analyses were performed using SAS statistical software version 9.2 for Windows (SAS Institute Inc., Cary, NC). Data are expressed as median (interquartile range), unless otherwise stated. A p value of less than 0.05 was considered to represent a statistically significant difference.
Bronchoalveolar Lavage A detailed description of the BAL procedure is provided in the online supplement (Supplemental Digital Content 1, http:// links.lww.com/CCM/A946). In brief, on both study days a BAL was performed in the lingula (control segment) at baseline (t = 0 hr), immediately followed by the interventions. Endobronchial instillation of either endotoxin or saline was delivered in a subsegment of the right middle lobe (challenged segment). All BAL procedures were performed according to guidelines from the American Thoracic Society (31).
RESULTS
Assays Total peripheral leukocyte counts and differentials were determined by an automated analyzer (Sysmex XE-2100, Sysmex Europe GmbH, Hamburg, Germany), and plasma C-reactive protein (CRP) was measured by particle-enhanced immunoturbidimetry (Roche/Hitachi Modular Pre-Analytics Plus System, Roche Diagnostics GmbH, Mannheim, Germany). Bronchoalveolar lavage fluid (BALF) was kept at –80°C until measurement of both plasma and BALF albumin and total protein which was done by the colorimetric method (Roche/Hitachi Modular Pre-Analytics Plus System, Roche Diagnostics GmbH). Plasma for cytokine measurements was obtained by centrifuging whole blood in EDTA-containing tubes, whereas unfiltered BALF was centrifuged in tubes containing a carrier protein (10% bovine serum albumin in phosphate-buffered saline), both at 3,500 rpm at 4°C for 15 minutes. BALF supernatants and plasma samples were kept at –80°C until analysis. Tumor necrosis factor (TNF)-α and interleukin (IL)-6 were determined by means of electrochemiluminescent detection using a multiarray ultrasensitive immunoassay and a SECTOR Imager 2400 (Meso Scale Diagnostics, Gaithersburg, MD). Samples were analyzed in duplicate, and mean concentrations were calculated. The lower limit of detection was 0.12 and 0.26 pg/mL for TNF-α and IL-6, respectively. Critical Care Medicine
Clinical Signs and Symptoms of Systemic and Pulmonary Inflammation IV challenge with endotoxin was associated with fever, decreased blood pressure, tachycardia, and increased respiratory rate (Fig. 2A–D, endotoxin-time interaction; p < 0.001 for all variables), together with increases in pulmonary and systemic symptom scores (Table E1, Supplemental Digital Content 1, http://links.lww.com/CCM/A946). By contrast, endobronchial instillation of endotoxin elicited only a very transient increase in mean arterial blood pressure, heart rate (Fig. 2, B and C), and symptom scores (Table E1, Supplemental Digital Content 1, http://links.lww.com/CCM/A946), with no change in core temperature or respiratory rate (Fig. 2, A and D). Biochemical Markers of Systemic Inflammation As expected, IV challenge with endotoxin elicited a systemic inflammatory response with a time-dependent increase and peak in plasma cytokines (TNF-α, 2 hr; IL-6, 2–4 hr) and a transient leukopenia that was later followed by leuko- and neutrocytosis (Fig. 3A–D, endotoxin-time interaction; p < 0.001 for all variables). TNF-α, total peripheral leukocytes, and neutrophils were still significantly elevated 24 hours postchallenge (Fig. 3A, C, and D), with levels of CRP peaking at 70 mg/L (48–89 mg/L) (endotoxin-time interaction; p < 0.001). Similarly, endobronchial instillation of endotoxin was associated with increased plasma levels of IL-6, leukocytes (Fig. 3B–D), and CRP (3; 1–5 mg/L at 24 hr; endotoxin effect; p < 0.001), albeit to a lesser extent than after IV challenge. IL-6 peaked after approximately 4 hours, whereas no change in plasma TNF-α levels was observed (Fig. 3, A and B). Of note, the levels of cytokines and peripheral leukocyte counts did not differ according to the order of randomization (online supplement, Supplemental Digital Content 1, http:// links.lww.com/CCM/A946). www.ccmjournal.org
3
Copyright (c) Society of Critical Care Medicine and Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited
Plovsing et al
Figure 2. Temperature, invasive hemodynamics, and respiratory responses to IV or endobronchial administration of endotoxin in healthy volunteers. Data presented as median (black line) and interquartile range (error bars). Closed circles: IV endotoxin (n = 15); open circles: endobronchial endotoxin (n = 15). Rectal temperature (A) was obtained hourly (0–8 hr) and at 16 and 24 hr, whereas mean arterial blood pressure (MAP) (B), heart rate (C), and respiratory rate (D) were continuously sampled at 1 kHz throughout the first 0–8 hr and afterward recalculated as the mean during 30-minute epochs. Longitudinal data were analyzed by use of a linear mixed model with post hoc adjustments (Tukey-Kramer) for multiple comparisons. Between intervention group effects (endotoxin-time interaction), p < 0.001 for all variables. Different from baseline within endobronchial endotoxin group, *p < 0.01; different from baseline within IV endotoxin group, †p < 0.05, ‡p < 0.01; different from corresponding value between interventions, #p < 0.05, ##p < 0.01.
Biochemical Markers of Pulmonary Inflammation IV administration was associated with an increased BALF leukocyte count (Fig. 4B, endotoxin effect; p = 0.03) and a tendency toward an increased neutrophil count (endotoxin effect; p = 0.05). Furthermore, IV challenge was associated with an increase in BALF IL-6 levels (Fig. 4F, endotoxin effect; p < 0.001), and TNF-α likewise tended to increase (Fig. 4D, endotoxin effect; p = 0.06). The increase in IL-6 levels following IV administration occurred bilaterally (Fig. 4F, control vs challenged segment; p = 0.66) and was at its maximum after 6–8 hours. Endobronchial instillation of endotoxin resulted in a progressive increase in BALF leukocyte count in the challenged segment (Fig. 4A, endotoxin effect; p < 0.001). The observed increase in cellularity was mainly due to alveolar neutrocytosis (Table E2, Supplemental Digital Content 1, http://links.lww.com/CCM/ 4
www.ccmjournal.org
A946; endotoxin effect; p < 0.001); furthermore, instillation was associated with an increase in the number of BALF macrophages and lymphocytes (Table E2, Supplemental Digital Content 1, http://links.lww.com/CCM/A946; endotoxin effect; p < 0.01). Instillation of endotoxin did also result in a profound increase in BALF cytokines from the challenged segment different from that observed following IV challenge (Fig. 4C–F, endotoxin-time interaction; p = 0.003 and 0.02 for TNF-α and IL-6, respectively). TNF-α and IL-6 peaked 4 hours after endobronchial instillation. The observed cellular response after endobronchial instillation did not differ from that after IV administration, although there was a trend toward higher BALF leukocyte and neutrophil counts (Fig. 4, A and B; and Table E2, Supplemental Digital Content 1, http://links.lww.com/CCM/A946; endotoxin-time interaction; p = 0.05 and 0.07, respectively). Total BALF XXX 2014 • Volume 42 • Number XXX
Copyright (c) Society of Critical Care Medicine and Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited
Clinical Investigation
Figure 3. Plasma cytokine and peripheral leukocyte responses to IV or endobronchial administration of endotoxin in healthy volunteers. Data presented as median (black line) and interquartile range (error bars). Closed circles: IV endotoxin (n = 15); open circles: endobronchial endotoxin (n = 15). Tumor necrosis factor (TNF)-α (A), interleukin (IL)-6 (B), total peripheral leukocyte count (C), and peripheral neutrophil count (D) were obtained hourly from a radial artery catheter (0–8 hr) and again at 16 and 24 hr (venous samples). Longitudinal data were analyzed by use of a linear mixed model with post hoc adjustments (Tukey-Kramer) for multiple comparisons. Between intervention group effects (endotoxin-time interaction), p < 0.001 for all variables. Different from baseline within endobronchial endotoxin group, **p < 0.05, *p < 0.01; different from baseline within IV endotoxin group, ‡p < 0.01; different from corresponding value between interventions, ##p < 0.01.
protein and albumin concentrations changed markedly after endobronchial installation (endotoxin-time interaction; p < 0.01), with a three- to 25-fold increase after 24 hours (Table E2, Supplemental Digital Content 1, http://links.lww.com/ CCM/A946). By contrast, no changes were observed after IV administration (Table E3, Supplemental Digital Content 1, http://links.lww.com/CCM/A946). Although the predicted effects model of inflammation corresponded well with the observed data (Fig. 4), a diverse individual pulmonary response to systemic inflammation was observed. Hence, the yield of cytokines from the lungs was found to be augmented in approximately one third of volunteers following IV challenge (Fig. 4, D and F). Rerunning the analysis without any potential influential observation did not statistically affect between (TNF-α, p = 0.04; IL-6, p = 0.04) or within intervention group effects (TNF-α, p = 0.54; IL-6, p < 0.001). Critical Care Medicine
DISCUSSION In the present human-experimental study, administration of endotoxin elicited both primary (compartmental) and secondary (transcompartmental) inflammatory responses. Our findings indicate that, within hours after the onset of systemic inflammation, local alveolar synthesis or leakage of IL-6 from the circulation to the lungs is followed by transmigration of leukocytes. The mechanisms responsible for this transcompartmental signaling may differ from that caused by a pulmonary insult, where the local emergence of proinflammatory mediators is much more closely related to the ensuing systemic response. Thus, the degree and differential kinetics of inflammation suggest that propagation of disease may indeed depend on the primary site of injury. In accordance with previous findings in humans, IV challenge with endotoxin was associated with physiological symptoms, high www.ccmjournal.org
5
Copyright (c) Society of Critical Care Medicine and Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited
Plovsing et al
Figure 4. Bronchoalveolar lavage fluid (BALF) cytokine and BALF leukocyte responses to IV or endobronchial administration of endotoxin in healthy volunteers. Data presented as predicted means according to the mixed effects model (large dots with connecting line) with corresponding individual values shown for each lung segment and time point (small dots). Total alveolar leukocyte count (number × 103/ mL BALF) following endobronchial (A) and IV endotoxin (B); levels of tumor necrosis factor (TNF)-α (pg/ mL BALF) following endobronchial (C) and IV endotoxin (D); levels of interleukin (IL)-6 (pg/mL BALF) following endobronchial (E) and IV endotoxin (F). Data were obtained from the control segment (lingula, left side) at baseline (n = 15) and from both control segment and the challenged segment (right middle lobe, right side) 2, 4, 6, 8, or 24 hr after endotoxin administration (n = 2 for Figure A, left side, 24 hr and Figure B, right side, 6 hr, otherwise n = 3 [A–F]). Longitudinal data were analyzed by use of a linear mixed model for comparison of between intervention group effects (endotoxintime interaction). Endobronchial administration was associated with significantly higher pulmonary levels of both TNF-α (p = 0.003) and IL-6 (p = 0.02), whereas a trend toward higher total leukocyte counts was observed (p = 0.05). A within intervention group (endotoxin) effect was observed for all variables after endobronchial administration (A, C, and E; p < 0.01), with levels of cytokines peaking after 4 hr (C, E). IV administration was associated with an increase in total leukocyte count (B) and levels of IL-6 (F) (endotoxin effect; p = 0.03 and p
