Sevoflurane reduces severity of acute lung injury possibly by impairing formation of alveolar oedema

Abstract

Pulmonary oedema is a hallmark of acute lung injury (ALI), consisting of various degrees of water and proteins. Physiologically, sodium enters through apical sodium channels (ENaC) and is extruded basolaterally by a sodium–potassium–adenosine–triphosphatase pump (Na⁺/K⁺-ATPase). Water follows to maintain iso-osmolar conditions and to keep alveoli dry. We postulated that the volatile anaesthetic sevoflurane would impact oedema resolution positively in an in-vitro and in-vivo model of ALI. Alveolar epithelial type II cells (AECII) and mixed alveolar epithelial cells (mAEC) were stimulated with 20 µg/ml lipopolysaccharide (LPS) and co-exposed to sevoflurane for 8 h. In-vitro active sodium transport via ENaC and Na⁺/K⁺-ATPase was determined, assessing ²²sodium and ⁸⁶rubidium influx, respectively. Intratracheally applied LPS (150 µg) was used for the ALI in rats under sevoflurane or propofol anaesthesia (8 h). Oxygenation index (PaO₂/FiO₂) was calculated and lung oedema assessed determining lung wet/dry ratio. In AECII LPS decreased activity of ENaC and Na⁺/K⁺-ATPase by 17·4% ± 13·3% standard deviation and 16·2% ± 13·1%, respectively. These effects were reversible in the presence of sevoflurane. Significant better oxygenation was observed with an increase of PaO₂/FiO₂ from 189 ± 142 mmHg to 454 ± 25 mmHg after 8 h in the sevoflurane/LPS compared to the propofol/LPS group. The wet/dry ratio in sevoflurane/LPS was reduced by 21·6% ± 2·3% in comparison to propofol/LPS-treated animals. Sevoflurane has a stimulating effect on ENaC and Na⁺/K⁺-ATPase in vitro in LPS-injured AECII. In-vivo experiments, however, give strong evidence that sevoflurane does not affect water reabsorption and oedema resolution, but possibly oedema formation.

Introduction

Acute lung injury (ALI) and acute respiratory distress syndrome (ARDS) are a major cause of acute respiratory failure in critically ill patients [1]. The mortality of ARDS has remained high since its first description by Ashbaugh and colleagues [2], although lung protective ventilatory strategies have reduced mortality from 60–70% to 35–40% [3],[4]. While drug treatment is investigated intensively, no pharmacological approach has yet been established [5]–[8].

ALI/ARDS is characterized by capillary leak and reduced fluid reabsorption, resulting in lung oedema. The level of decreased rate of fluid clearance has significant prognostic value for morbidity and mortality [9]. In addition to reduced fluid reabsorption, protein clearance is also impaired. As demonstrated in patients with ARDS, non-survivors have three times higher alveolar protein concentrations than survivors [10],[11].

Several studies have tried to detect the underlying mechanism of impairment of alveolar fluid clearance in ALI/ARDS and various pathways have been suggested [12]–[14]. According to experimental evidence, the active sodium (Na⁺) transport is thereby the most important ion transport mechanism involved in fluid reabsorption out of the alveolar space [15],[16]. The broadly accepted paradigm for Na⁺ transport in the alveoli is a two-step process: Na⁺ enters the cell by epithelial amiloride-sensitive Na⁺-channels (ENaC) located at the apical surface and is extruded by basolaterally located sodium–potassium–adenosine–triphosphatase pumps (Na⁺/K⁺-ATPases) [17],[18]. Research conducted in the last two decades provides evidence that this vectorial transport from apical to basal generates the osmotic force for water to flow out of the alveolar air spaces [19]–[22]. This process can be up- or down-regulated, implying an increased or diminished clearance of alveolar fluid. Studies have demonstrated that net vectorial fluid transport is reduced in human alveolar epithelial cells type II (AEC II) in ALI [23].

Patients suffering from ALI/ARDS most often need to be ventilated mechanically, and therefore remain sedated in intensive care units (ICU) [24]. The overall effect of sedatives and anaesthetics – volatile anaesthetics included – on this disease is unclear. As demonstrated previously, the inflammatory response upon endotoxin stimulation in AEC is partly reversible in the presence of sevoflurane [25]. In an in-vivo model of ALI oxygenation improved in the presence of sevoflurane [26]. However, at the same time volatile anaesthetics are suspected to impair sodium transport [27].

The aim of this work was to investigate the effect of the nowadays commonly used volatile anaesthetic sevoflurane on ENaC and Na⁺/K⁺-ATPase in vitro and in vivo. Based on previous in-vitro and in-vivo results with a positive effect of sevoflurane [26], the hypothesis was raised that in-vitro activity of ENaC and Na⁺/K⁺-ATPase in endotoxin-injured AEC may be increased upon treatment with sevoflurane. Furthermore, an attempt was made to clarify the impact of sevoflurane on oedema in vivo in the endotoxin-induced lung injury model. An improved alveolar fluid clearance upon sevoflurane exposure was postulated.

Materials and methods

In-vitro assay

Alveolar epithelial cells type II (AECII)

The L2 cell line (CCL 149; American Type Culture Collection, Rockville, MD, USA) was derived through cloning of adult female rat lung of AEC type II origin. Cells were cultured in Dulbecco's modified Eagle's medium (DMEM; Invitrogen, Carlsbad, CA, USA), supplemented with 10% fetal bovine serum (FBS; Invitrogen), 1% penicillin–streptomycin and 1% 4-(2-hydroxethyl)-1-piperazineethanesulphonic acid buffer (HEPES; Invitrogen). They were grown for 3 days in uncoated plates (Corning Inc., Corning, NY, USA) to >95% confluence.

Mixed alveolar epithelial cells (mAEC)

Primary AEC were harvested following an established protocol [28],[29]. Briefly, lungs were explanted from male Wistar rats, injected with 10 ml of phosphate-buffered saline (PBS) containing 4 U/ml porcine pancreas elastase (Sigma-Aldrich, Hamburg, Germany) and incubated for 20 min at 37°C. Trachea and large airways were discarded and lungs were minced. Elastase reaction was stopped with 5 ml FBS. After vigorous stirring for 20 min, cells were filtered and incubated for 1 h at 37°C in Petri dishes, coated previously with 1 mg/ml rat immunoglobulin (IgG) (Sigma-Aldrich) in PBS, in order to remove immunocompetent cells. Unattached cells were washed away, and the remaining cells were cultured in DMEM/10% FBS. After a 7-day incubation time, a mixture of type I and type II cells (mAEC) was found (Fig. 1). Type II was detected analysing sodium-dependent phosphate-co-transporter type IIb (NaPi IIb) [30],[31], and type I cells with detection of aquaporin 5 (AQP5) [32],[33].

Fig. 1.

TaqMan real-time polymerase chain reaction (PCR) of alveolar epithelial type II cells (AECII) and mixed alveolar epithelial cells (mAEC). Detection of type I and type II characteristic ion transporters. Values are mean ± standard deviation; n = 6 per group.

Stimulation with LPS and sevoflurane exposure

DMEM/10% FBS of confluent AEC monolayers was replaced by DMEM/1% FBS at least 14 h before starting the experiment. AEC were stimulated with lipopolysaccharide (LPS) from Escherichia coli, serotype 055:B5 (Sigma-Aldrich), in a concentration of 20 µg/ml in DMEM/1% FBS (control group with PBS), according to previous studies [34],[35], and placed in two humidified, preheated (37°C) air-tight chambers (oxid anaerobic jar; Oxoid AG, Basel, Switzerland).

AEC were exposed to 1 minimal alveolar concentration (MAC) = 2·2 vol% sevoflurane (Sevorane®; Abbott AG, Baar, Switzerland) for 8 h, representing a clinically relevant concentration of the volatile anaesthetic as used in previously performed experiments [34]. A mixture of 5% CO₂ and 95% air was directed through a Sevotec 5 Vaporizer (Abbott AG), placed at the entrance of the chamber (for control only CO₂/air mixture). Within 5 min, sevoflurane reached the steady state concentration of 2·2 vol% (monitored by Ohmedia 5330 Agent Monitor; Abbott AG). The chambers were sealed for 8 h and incubated at 37°C. At the end of the experiment sevoflurane concentration was verified again to confirm the value of 2·2 vol%.

²²Na influx studies

Measurement of ²²Na flux through amiloride-sensitive Na⁺ channels was performed as described previously [36]. Culture medium was removed, and cells on six-well plates were rinsed twice and preincubated at 37°C for 20 min in a buffered sodium-free solution containing (in mM): 137 N-methylglucamine, 5·4 KCl, 1·2 MgSO₄, 2·8 CaCl₂ and 15 HEPES (pH 7·4). This solution was replaced by uptake solution composed of (in mM): 14 NaCl, 35 KCl, 96 N-methylglucamine and 20 HEPES (pH 7·4) containing 0·5 µCi/ml of ²²NaCl (37 MBq/mg Na) in the absence or presence of 100 µM amiloride. Amiloride blocks sodium uptake via ENaC and was used as positive control for blocking sodium absorption. After an incubation time of 5 min, cells were washed twice with 1 ml/well of an ice-cold solution containing (in mM): 120 N-methylglucamine and 20 HEPES (pH 7·4). Cells were solubilized in 0·3 ml/well trypsin for 3 min. Tracer activities were determined by liquid scintillation counting (Tri-carb 2900TR, liquid scintillation scanner; Packard, Chicago, IL, USA).

⁸⁶Rubidium influx studies

The measurement of ouabain-sensitive rubidium (⁸⁶Rb) influx was performed as described previously [37],[38]. Assays were performed in a buffered solution A of the following composition (in mM): 120 NaCl, 5 RbCl, 1 MgSO₄, 0·15 Na₂HPO₄, 0·2 NaH₂PO₄, 4 NaHCO₃, 1 CaCl₂, 5 glucose, 2 lactate, 4 essential and non-essential amino acids, 20 HEPES and 0·1% bovine serum albumin (BSA). The osmotic pressure of solution A was adjusted by mannitol to 350 mosM, pH 7·4. After removal of the culture medium half the cells in six-well plates were incubated with 1 ml/well of 4 mM ouabain at 37°C for 30 min. Ouabain blocks Na⁺/K⁺-ATPase and was used as positive control for blocking the transporter. The other half was incubated with solution A. Subsequent plates were washed with 1 ml/well of solution A and incubated for 5 min with 0·6 ml/well of solution A supplemented with 1 µCi/ml ⁸⁶RbCl (370 MBq/mg Rb). Uptake was stopped by washing the cells twice with 1 ml/well of ice-cold rinsing solution containing the following (in mM): 140 N-methylglucamine, 1·2 MgCl₂, 3 NaCl₂, 10 HEPES and 0·1% BSA at pH 7·4. Solubilized cells were traced by liquid scintillation counting.

All chemicals were purchased from Sigma-Aldrich and culture media and their reagents from Invitrogen. Radioactive tracers were supplied by PerkinElmer AG.

Statistical analysis

Each experimental set-up was performed three times, each conducted in sextuplet. Data of the three experiments were taken together and analysed (n = 18). Values are expressed as mean ± standard deviation (s.d.). Optical analysis of box-plots suggested normal distribution of data. Confirmation was performed using a Shapiro–Wilk test. The effects of sevoflurane were compared with the control group (PBS group) for K⁺- and Na⁺-influx and tested by analysis of variances for repeated measurements [one-way analysis of variance (anova)], including a Tukey–Kramer multiple comparison test. Graphpad Prism4® Graphpad Instat3® (GraphPad software, La Jolla, CA, USA) was used for statistical analyses. P-values <0·05 were considered statistically significant.

In-vivo assay

Animal preparation

After approval from the local animal care and use committee (Zürich, Switzerland), experiments were performed with pathogen-free, male Wistar rats (Charles River, Sulzfeld, Germany) (body weight 350–500 g). The rats were kept in standard cages at 22°C (12-h light/12-h dark). Food and water were supplied ad libitum.

Induction of anaesthesia and monitoring was performed as described previously [26]. Rats were tracheotomized. After insertion of a sterile metal cannula, animals were ventilated in parallel (Servo Ventilator 300, Maquet, Solna, Sweden). Pressure-controlled ventilation was set with 30 breaths per minute, pressure was 3/14 cm H₂O, inspiration to expiration ratio 1 : 2 and fractional inspired oxygen concentration (FiO₂) was 100%. Arterial blood was analysed at 0, 2, 4, 6 and 8 h. Using 100% FiO₂ during the whole experiment, the oxygen capability of the lung is represented by the oxygen tension (PaO₂ in mmHg) in arterial blood gas samples (oxygenation index: PaO₂/FiO₂). Body temperature was controlled by rectal temperature measurement and corrected to 37°C by a heating lamp.

Experimental design

Rats were randomized into three different groups, using sealed envelopes: (a) propofol/PBS; (b) propofol/LPS and (c) sevoflurane/LPS (n = 6 in all groups). Rats were instilled intratracheally with 150 µg LPS in 300 µl PBS (control with PBS only) [39], followed immediately by randomization in either propofol or sevoflurane group (co-conditioning). Anaesthesia was performed as described previously [26].

RNA extraction and real-time polymerase chain reaction (PCR) for α-ENaC, γ-ENaC and α₁-Na⁺/K⁺-ATPase

Eight hours after the onset of injury rats were euthanized and lungs were explanted, shock-frozen in liquid nitrogen and stored at −80°C for isolation of mRNA.

Total RNA was isolated form lung tissue using the RNeasy® Mini Kit (Qiagen, Basel, Switzerland), according to the manufacture's protocol. RNA amounts were determined by absorbance at 260 nm.

Reverse transcription and real-time quantitative TaqMan™ PCR were performed as described previously [26]. Specific primers (Microsynth, Balgach, Switzerland) and labelled TaqMan probes (Roche Applied Science, Basel, Switzerland) were designed for α- and γ-subunits of ENaC, for α₁-subunit of Na⁺/K⁺-ATPase and 18S as housekeeping gene. All primers and probes used in the experiments are presented in Table 1. Each experimental PCR run was performed in duplicate with simultaneous negative controls without template.

Table 1.

Primer sequences and probes used for TaqMan real-time polymerase chain reaction (PCR).

Gene	Primer sequence and corresponding probe	Fragment size
α-ENaC	Forward: 5′-TGT GAC TAC CGA AAG CAG AGC-3′	102 bp
	Reverse: 5′-AGG CTT CCG ACA CTT GGA G-3′
	Probe #26
γ-ENaC	Forward: 5′-AGC AAC ACC CCA ACT GGA T-3′	93 bp
	Reverse: 5′-AGG ATT GCT GCA CAC TGA TT-3′
	Probe #26
α1-Na+/K+-ATPase	Forward: 5′-ACT TGG GCA CTG ACA TGG TT-3′	104 bp
	Reverse: 5′-CAC AAG TTT GTC CGT TTT GG-3′
	Probe #26
NaPi IIb	Forward: 5′-CCC AGG AAG AGG AGC AAA A-3′	72 bp
	Reverse: 5′-TCA GGA GCT TTG TGC CAA C-3′
	Probe #64
Aquaporin 5	Forward: 5′-GCT CCG AGC TGT CTT CTA CG-3′	131 bp
	Reverse: 5′-GCG TTG TGT TGT TGT TCA GC-3′
	Probe #20
18S	Forward: 5′-GGA GAG GGA GCC TGA GAA AC-3′	70 bp
	Reverse: 5′-TCG GGA GTG GGT AAT TTG C-3′
	Probe #74
Probe #20	5′-CTG GCT GG-3′
Probe #26	5′-CAG CCC AG-3′
Probe #64	5′-CCA GGC TG-3′
Probe #74	5′-GGC AGC AG-3′

α-ENaC: alpha-subunit of the epithelial amiloride-sensitive Na⁺-channel; γ-ENaC: gamma-subunit of the epithelial sodium channel; α₁-Na⁺/K⁺-ATPase: sodium-potassium-adenosine-triphosphatase (α₁-subunit); AQP5: Aquaporin 5; NaPi IIb: sodium-dependent phosphate co-transporter type IIb; 18S: housekeeping gene (ribosomal origin); bp: base pairs.

For quantitation of gene expression the comparative C_t method was used as described by Livak et al. [40]. The C_t values of samples (propofol/LPS and sevoflurane/LPS) and control (propofol/PBS) were normalized to the housekeeping gene (18S) and calculated as follows: 2^–δδCt, where δδC_t = δC_t,samples – δC_{t, controls.}

Lung wet/dry ratio

Sevoflurane/LPS animals were given 150 µg LPS in 300 µl PBS with or without 100 µM amiloride to block sodium resorption via ENaC [41] (Sigma-Aldrich). After 8 h animals were sacrificed, lungs were explanted and wet weight was measured. Thereafter, lungs were air-dried for 72 h at 65°C and lung dry weight was quantified. Wet/dry ratio (w/d) was calculated as follows [42]: w/d = weight_wet/weight_dry

Statistics

Values are expressed as mean ± s.d., n = 6 per group. Optical analysis of box-plots suggested normal distribution of data. Confirmation was performed with a Shapiro–Wilk test. Vital parameters were tested by analyses of variance for repeated measurements (one-way anova) with a Tukey–Kramer multiple post-hoc test. Real-time PCR and wet/dry ratio data were tested using Student's t-test. Graphpad Prism4® and Graphpad Instat3® (GraphPad Software) were used for statistical analyses. P-values less or equal to 0·05 were considered statistically significant.

Results

Cell survival

As described in previous experiments [25],[34], cell survival was not influenced upon sevoflurane and LPS exposure. This was confirmed with a cytotoxic assay [determination of lactate dehydrogenase (LDH); Promega, Madison, WI, USA, data not shown].

Cell characteristics of mAEC

As seen in Fig. 1, primary culture of mAEC represented both types I and II AEC, detected by real-time PCR (Table 1).

Co-exposure of AECII to LPS and sevoflurane: effects on sodium transport via ENaC and Na⁺/K⁺-ATPase

ENaC activity was assessed in an AECII monolayer measuring ²²sodium (²²Na) influx. As displayed in Fig. 2a, stimulation with LPS impaired ²²Na-influx by 17·4% ± 13·3% s.d. (P < 0·05) compared to the control group. In the presence of sevoflurane, sodium influx into AECII improved in the LPS group, reaching control values (P < 0·05).

Fig. 2.

(a) ²²Na-uptake via epithelial amiloride-sensitive Na⁺-channels (ENaC), relative counts per minute in the different groups. Alveolar epithelial type II cells (AECII) were incubated for 8 h with or without lipopolysaccharide (LPS) and with or without sevoflurane (Sevo). Values are mean ± standard deviation (s.d.); n = 18 per group. *P < 0·05 versus three other groups. (b) Sodium–potassium–adenosine–triphosphatase (Na⁺/K⁺-ATPase) activity measured by ⁸⁶rubidium uptake in alveolar epithelial type II cells (AECII), relative counts per minute. Cells were incubated for 8 h with or without LPS and with or without sevoflurane (Sevo). Values are mean ± s.d.; n = 18 per group. **P < 0·01 versus three other groups.

Activity of Na⁺/K⁺-ATPase, measured by ⁸⁶rubidium (⁸⁶Rb) influx, revealed a 16·2% ± 13·1% (P < 0·01) decrease of ⁸⁶Rb-influx upon LPS stimulation (Fig. 2b). In LPS-stimulated AECII co-exposed to sevoflurane ⁸⁶Rb-influx reached values comparable to the control group (P < 0·01).

Co-exposure of mAEC to LPS and sevoflurane: effects on sodium transport via ENaC and Na⁺/K⁺-ATPase

No difference in ²²Na-influx was observed in all four groups (Fig. 3a).

Fig. 3.

(a) ²²Na-uptake via epithelial amiloride-sensitive Na⁺-channels (ENaC), relative counts per minute in the different groups. A mixture of type I and II alveolar epithelial cells (mAEC) cells were incubated for 8 h with or without lipopolysaccharide (LPS) and with or without sevoflurane (Sevo). Values are mean ± standard deviation (s.d.); n = 18 per group. (b) Sodium–potassium–adenosine–triphosphatase (Na⁺/K⁺-ATPase) activity measured by ⁸⁶rubidium uptake in mAEC, relative counts per minute. mAEC were incubated for 8 h with or without lipopolysaccharide (LPS) and with or without sevoflurane (Sevo). Values are mean ± s.d.; n = 18 per group. *P < 0·05 versus control.

Na⁺/K⁺-ATPase activity in mAEC was increased by 23·7% ± 24·5% in the LPS group, 26·1% ± 38·6% in the sevo/LPS group (both P < 0·05). Sevoflurane did not have a significant impact on LPS-injured mAEC (Fig. 3b).

In-vivo effect of co-stimulation with LPS and sevoflurane on ENaC and Na⁺/K⁺-ATPase mRNA expression

mRNA of α-ENaC was decreased by 58% ± 26·9% in the propofol/LPS compared to the propofol/PBS group (P < 0·05) (Fig. 4a). Sevoflurane co-conditioning did not impact upon the expression of α-ENaC mRNA. γ-ENaC mRNA was down-regulated in both LPS groups compared to propofol/PBS: it decreased by 81·7% ± 12·9% (P < 0·01) in the propofol/LPS and 71·7% ± 17·3% (P < 0·01) in the sevoflurane/LPS group (Fig. 4b), with no intergroup difference.

Fig. 4.

(a) Expression of the α-subunit of epithelial amiloride-sensitive Na⁺-channels (ENaC) mRNA in differently treated animals. Propofol/phosphate-buffered saline (PBS) served as control. LPS: lipopolysaccharide. Values are mean ± standard deviation (s.d.); n = 6 per group. *P < 0·05 versus propofol/PBS. (b) Expression of the γ-subunit of γ-ENaC mRNA in differently treated animals. Propofol/PBS served as control. Values are mean ± s.d.; n = 6 per group. **P < 0·01 versus propofol/PBS. (c) Expression of α₁-subunit of sodium–potassium–adenosine–triphosphatase (Na⁺/K⁺-ATPase) mRNA in differently treated animals. Propofol/PBS served as control. Values are mean ± s.d.; n = 6 per group.

Despite an increased expression of α₁-Na⁺/K⁺-ATPase mRNA in LPS-treated compared to control animals (increase of 46·5% ± 114·6 in the propofol/LPS and 99·4% ± 81·4 in the propofol/LPS group), values between all groups did not differ significantly (Fig. 4c).

In-vivo effect of co-stimulation with LPS and sevoflurane: oxygenation index (pO₂/FiO₂)

While LPS application impaired oxygenation in the propofol group, oxygenation could be maintained in sevoflurane/LPS-treated animals comparable to propofol/PBS (Fig. 5): at 6 h, propofol/LPS animals presented with an oxygenation index of 298 ± 180 mmHg compared to 6 h sevoflurane/LPS animals with 466 ± 50 mmHg (P < 0·05). At 8 h the difference even increased, with 198 ± 142 mmHg in propofol/LPS animals to 454 ± 25 mmHg in LPS animals with sevoflurane application (P < 0·001).

Fig. 5.

Arterial oxygen partial pressure in mmHg of blood gas analysis in differently treated animals. Propofol/phosphate-buffered saline (PBS) served as control. LPS: lipopolysaccharide. Blood gas measurements were taken at the time-points 0, 2, 4, 6 and 8 h. Values are mean ± standard deviation; n = 6 per group. #P < 0·05 versus time-points 0, 2 and 4 h. *P < 0·05 versus 6-h time-point.

Wet/dry ratio of differently treated animals

A 27·7% ± 21·2% higher wet/dry ratio in animals treated with propofol/LPS compared to sevoflurane/LPS was observed (P < 0·05) (Fig. 6a). Sevo/LPS animals treated with amiloride presented similar wet/dry ratios to the group without amiloride application (Fig. 6b).

Fig. 6.

(a) Wet/dry ratio of rat lungs treated with lipopolysaccharide (LPS) and either sevoflurane or propofol for 8 h. Values are mean ± standard deviation (s.d.); n = 6 per group. *P < 0·05. (b) Wet/dry ratio of rat lungs. Animals were treated with LPS and sevoflurane for 8 h. Half the animals received amiloride intratracheally in order to block epithelial amiloride-sensitive Na⁺-channels (ENaC). Values are mean ± s.d.; n = 6 per group.

Discussion

With the current data, two main results can be summarized: first, sevoflurane has a stimulating effect on the pump function of sodium channels in LPS-injured AECII in vitro. However, no such impact was observed in a mixed culture of types I and II AEC (mAEC); rather, this cell composition reflected an in-vivo situation with predominantly type I cells in the lung. In-vivo data underline these findings, demonstrating that the presence of sevoflurane does not influence oedema resolution. Secondly, sevoflurane has a positive impact upon the course of LPS-induced injury in vivo. Animals anaesthetized with sevoflurane presented with better oxygenation.

Transepithelial sodium transport plays an important role in fluid clearance in normal and injured alveoli. α-ENaC thereby seems to be crucial, as α-ENaC-deficient mice died shortly after birth due to lung oedema even without pulmonary inflammation [43]. In addition, previous in-vivo experiments have demonstrated that blocking of epithelial sodium channels in a rat model of hyperoxia-induced lung injury increased extravascular lung fluid volume significantly [44]. With regard to ALI alveolar fluid transport can be up- or down-regulated [45]. Hypoxia inhibits transepithelial sodium transport in ex-vivo lungs [16], while endotoxin A from Pseudomonas aeruginosa stimulates alveolar fluid clearance in rats [46], probably by cytokine-induced stimulation of sodium uptake. Conversely, intratracheal application of endotoxin-impaired alveolar fluid clearance in adult rats at 6 h of injury [26],[47]. Evidence from previous studies indicates that a complex network of inflammatory cytokines and chemokines mediate and modify the inflammatory process in lung injury, including oedema formation [48]–[50].

It is known that inflammation in AEC is mitigated by application of sevoflurane [25]. Our in-vitro investigations in AECII reveal that LPS-induced impairment of both ENaC and Na⁺/K⁺-ATPase is reversed upon co-exposure to sevoflurane. These data suggest that active sodium transport and thus water transport can be increased functionally in injured AECII by administration of sevoflurane. So far, only type II cells were considered as the important regulators for salt and water transport [51]. However, as both types I and II AEC cells express sodium transport channels [52],[53], AECI might also play an important role in water and salt homeostasis in the lung [52]. Therefore, after the positive findings in AECII, in-vitro experiments regarding sodium transport were reassessed in a mixture of types I and II cells, a set-up which more probably reflects the in-vivo situation with only 5% of type II and 95% of type I cells in the lungs. With this mixture of AEC (mAEC), no LPS-induced change or significant influence of sevoflurane was observed for functionality of ENaC. For Na⁺/K⁺-ATPase we could demonstrate increased activity upon LPS exposure, while sevoflurane did not have any significant impact on its function. Therefore, we conclude that AECI are not involved actively in water reabsorption with regard to sodium channels.

A previous study showed evidence that oxygenation improved significantly using sevoflurane in a post-conditioning set-up in an LPS-induced ALI model (intratracheally applied LPS, followed 2 h later by application of sevoflurane compared to propofol anaesthesia) [26]. The present promising in-vitro results from AECII encouraged us to elucidate the question of to what extent sevoflurane may influence either oedema resolution or oedema formation. We were able to demonstrate that wet/dry ratio in the sevoflurane-treated animals was significantly lower compared to the propofol/LPS group, linking better oxygenation to less alveolar oedema. However, when blocking the activity of ENaC using amiloride, the wet/dry ratio remained unchanged. This important result suggests that sevoflurane most probably does not improve alveolar fluid reabsorption, but acts rather on the inflammatory side inhibiting the production of inflammatory mediators, which in consequence decreases vascular permeability. If alveolar water absorption had been more important than oedema formation, one would have expected a clearly increased wet/dry ratio in the case of blocked ENaC [41].

Another interesting observation of the current in-vivo experiments is that co-conditioning with sevoflurane is more effective in amelioration of oxygenation than post-conditioning with the volatile anaesthetic [26]. This finding suggests that early treatment with sevoflurane could inhibit the increase of permeability and attenuate injury-induced vascular leakage.

The present study has several limitations that need to be addressed. Discussion from in-vitro experiments is limited, as the interaction with cells of different character is missing. Another concern lies in the experimental set-up of ALI used. Even if intratracheal application of LPS is defined as a relevant in-vitro and in-vivo animal model for lung injury, it does not fully represent ALI in patients. Therefore, conclusions cannot necessarily be translated to a clinical situation. Furthermore, due to the fact that lungs could not be utilized for both measurement of lung wet/dry ratios and lung RNA analysis, experiments had to be repeated using different animals. This, of course, may create a sample bias, which we tried to minimize by following our strict experimental protocols. Nevertheless, despite these limitations the present study provides new information regarding the protective effect of volatile anaesthetics in ALI.

In conclusion, these data reveal that sevoflurane reverses the inhibitory effect of LPS on the function of ENaC and Na⁺/K⁺-ATPase in AECII in vitro. Sevoflurane exposure can influence positively the course of LPS-induced lung injury with regard to oxygenation. This effect, however, seems not to be mediated by increased fluid clearance, but rather by the anti-inflammatory properties of sevoflurane leading to less oedema formation.

Disclosure

The authors have no conflicts of interest.