Isoflurane Regulates Atypical Type-A γ-Aminobutyric Acid Receptors in Alveolar Type II Epithelial Cells

γ-AMINOBUTYRIC acid (GABA) is a major inhibitory neurotransmitter in the central nervous system. GABA generates fast inhibition in mature neurons via activation of type-A GABA (GABA_A) receptors, a class of pentameric ion channels that are highly permeable to chloride anions (Cl⁻). GABA_A receptors are also widely expressed in nonneuronal cells in organs outside the central nervous system, including the lung, pancreas, and ovaries.¹  Recent studies,^2,3  including one of ours,⁴  showed that rodent alveolar type II (ATII) epithelial cells express GABA_A receptors. ATII cells also produce GABA,³  and stimulation of this autocrine GABA–GABA_A receptor signaling system causes an outward chloride flux and membrane depolarization in these cells.⁴ 

A variety of inhaled and injectable general anesthetics, including the volatile anesthetic isoflurane, increases the activity of GABA_A receptors in neurons.^5,6  The increase in GABA_A receptor activity contributes to the profound neurodepressive properties of these drugs.⁷  When administered to patients, volatile anesthetics rapidly diffuse across the alveolar epithelium en route to the brain. We postulated that inhaled anesthetics alter the function of ATII cells, at least in part, by upregulating the activity of GABA_A receptors. ATII cells regulate innate pulmonary immune responses by producing the pulmonary collectin surfactant protein-A,⁸  as well as proinflammatory factors such as cyclooxygenase-2 (COX-2)⁹  and cytokines.¹⁰  Stimulating GABA_A receptor⁵  decreases the secretion of cytokines from lung epithelial cells.¹¹  Isoflurane also reduces the release of cytokines from ATII cells¹⁰  and decreases cytokine-augmented expression of surfactant proteins in these cells.¹²  Animal models have indicated that these actions of isoflurane may protect the lung from infection-related lung injuries.^13,14 

The current study was undertaken to determine whether isoflurane increases the function of GABA_A receptors in human ATII cells and, if so, whether isoflurane alters the expression of COX-2 via regulation of GABA_A receptors. The prototypic inhaled anesthetic isoflurane was selected for most of the studies reported here because it is widely used in clinical practice and because the effects of this inhaled anesthetic in the lungs have been previously characterized.

Human lung sections were obtained from the University of British Columbia James Hogg Research Centre tissue bank. Use of these lung sections was approved by the iCAPTURE Centre of St. Paul’s Hospital, Vancouver, Canada. The procedures used to obtain human lung tissues and to prepare lung sections for studies such as this one have been described.^15,16  Briefly, the lung tissues were obtained from patients who underwent lung resection for treatment of tumors at St. Paul’s Hospital. Each resected lung was inflated with a solution containing 50% cryomatrix solution (Shandon, Pittsburg, PA), rapidly frozen solid in liquid nitrogen vapor, cut into 2-cm-thick transverse slices, and sampled with a power-driven hole saw to obtain tissue cores 1.5 cm in diameter and 2 cm in length, which were subsequently stored at −80°C. Sections of lung without evidence of cancer were selected for this study, cut into 10-μm slices, and mounted on glass slides.

A549 cells, a line of cells derived from human ATII cells,¹⁷  were propagated as previously described.¹⁸  Briefly, stock cultures of A549 cells were maintained at 4-day passage intervals, and individual cells were used for no more than 15 passages. For experimental tests, cells were removed from monolayer stock cultures with a trypsin–EDTA solution (0.05% trypsin, 0.481 mm Na-EDTA, Life Technologies, Carlsbad, CA), counted with a hemocytometer, and plated at a density of about 1 × 10⁵ cells/ml in 35-mm tissue culture dishes or on glass coverslips in Dulbecco’s modified Eagle medium (Life Technologies) supplemented with 10% fetal calf serum. The A549 cells were incubated at 37°C in a humidified atmosphere of 5% CO₂ and were used for immunocytochemical and/or electrophysiologic studies 24–36 h after plating.

The expression of GABA_A receptor subunits in lung tissues and A549 cells was studied by immunofluorescence staining. GABA_A receptors are hetero-pentamers made up of combinations of 19 different subunits.¹⁹  In the current study, we examined the expression of five specific subunits (α5, β2, β3, δ, and π), for two reasons: first, our polymerase chain reaction assays detected abundant messenger RNAs for the α5, β3, δ, and π subunits in A549 cells⁴  and second, the functional GABA_A receptor pentamers are often composed of two α and two β subunits and one of the other subunits.²⁰  As previously described,⁴  paraffin sections of mouse lung tissue were deparaffinized with xylene and then hydrated sequentially in 100, 95, and 70% ethanol. Epitopes were unmasked by heating the tissue sections in citrate buffer at pH 6 in a microwave. The tissues were permeabilized with 0.1% Triton X-100 and blocked with 10% normal serum for 1 h. The lung tissue sections were incubated overnight with primary antibodies for diverse GABA_A receptor subunits, including antibody for the α2 subunit (Alomone Labs, Jerusalem, Israel, 1:100 dilution) and antibody for the π subunit (Abcam, Cambridge, MA, 4 μg/ml). These subunits were specifically studied because the minimal requirement to produce a GABA-gated ion channel is the inclusion of both α and β subunits²¹  and because a high level of messenger RNA encoding the π subunit has been identified in rat ATII cells.^2,3  The ATII cells in lung sections were identified by immunostaining presurfactant protein C using an antipresurfactant protein C antibody (Santa Cruz Biotechnology, Santa Cruz, CA, 1:200 dilution). The lung sections were subsequently incubated with Cy3-conjugated or fluorescein isothiocyanate-conjugated secondary antibodies (Jackson ImmunoResearch Laboratories, West Grove, PA). As a negative control, to verify stain specificity, immunoglobulin G (Santa Cruz Biotechnology) was used to replace primary antibodies. Immunocytochemistry of cultured cells was performed as previously described,^22–24  using antibody against GABA_A receptor β2 and β3 subunits (Millipore, Temecula, CA, 1:100 dilution) and an antibody against COX-2 (Santa Cruz Biotechnology, 1:200 dilution). A fluorescein isothiocyanate-conjugated antibody for pan-cytokeratin (Sigma, Oakville, Ontario, Canada, 1:100 dilution) was used to visualize the A549 cells.

Immunostained lung tissue and A549 cells were studied through an inverted Zeiss microscope using the Zeiss LSM program (Carl Zeiss Canada Ltd., Toronto, Ontario, Canada). The detection threshold for immunofluorescence of the studied protein was set just below the negative control. The microscope image fields were randomly selected by arbitrarily moving the sample, and multiple confocal microscopy images (12–25 images per group) were obtained and saved for imaging analysis, as previously described.^22–24  Briefly, the fluorescence signal was adjusted with Image-J software (National Institutes of Health, Bethesda, MD) such that the fluorescence-stained cellular structures appeared black on a white background. The total area of staining per image field was collected for statistical analysis.

To demonstrate the expression of GABA_A receptor subunit protein in human-derived ATII cells, lysates of A549 cells were prepared and then used for immunoblotting assays with primary antibodies against specific subunits of GABA_A receptors. The general procedures of Western blotting were the same as previously described.²²  Briefly, A549 cells were lysed in ice-cold phosphate-buffered saline with 1% Triton X-100 and 0.5% sodium deoxycholate supplemented with protease inhibitors. The primary antibodies were purchased from the following companies: anti-GABA_A receptor α5 and δ subunits from Millipore (dilution of 1:1000 and 1:1500), respectively; anti-GABA_A receptor β2 and β3 subunits from ABR Affinity BioReagents (Golden, CO, 1:1000 dilution); anti-GABA_A receptor π subunit from Abcam (1:2000 dilution); anti-COX-2 from Santa Cruz Biotechnology (1:1000 dilution); and anti-ß-actin from Sigma (1:5000 dilution). Prepared solutions were incubated overnight at 4°C. Lysate (containing 40 µg protein) of mouse cerebral cortex was used as the positive control for blotting assays of the α5, β2, β3, and δ subunits, and the lysate (containing 50 µg protein) of Jurkat cells (an immortalized line of T lymphocytes) was used as the positive control for blotting assays of the π subunit.

A fresh solution of isoflurane (Abbott Laboratories Ltd., Chicago, IL) was prepared on each experimental day as previously described.²⁵  Specifically, liquid isoflurane was added to extracellular solution (ECS), and the solution was stored in a tightly sealed glass container at room temperature (22–24°C) for at least 2 h. Under these conditions, the ECS, saturated with isoflurane, formed an upper layer, while liquid isoflurane stayed below the ECS. The isoflurane-saturated ECS was sampled and quickly diluted in regular ECS at the following dilutions: 1:1000, 1:333, 1:100, 1:33, and 1:10. We previously measured the concentration of isoflurane in the perfusion barrels and found that the final concentrations of isoflurane in test solutions containing 0.10, 0.33, 1.0, 3.33, and 10.0% of isoflurane-saturated ECS were about 2.5, 8.3, 25, 83, and 250 µm, respectively.²⁵  All test solutions containing isoflurane were stored in sealed glass containers until used. Solutions of sevoflurane (Abbott Laboratories Ltd.) were prepared according to the procedure described earlier for isoflurane solutions.

After removal of the culture media, A549 cells were rinsed with ECS containing (in mm) 155 NaCl, 1.3 CaCl₂, 5.4 KCl, 25 HEPES, and 33 glucose, at pH 7.4 and osmolarity about 315 mOsm. An Axopatch-1D amplifier (Axon Instruments Inc., Foster City, CA) was used to make patch recordings at room temperature (22–24°C). The patch electrodes were filled with intracellular solution containing (in mm) 155 KCl, 15 KOH, 10 HEPES, 2 MgCl₂, 1 CaCl₂, and 2 tetraethylammonium, at pH 7.35 and osmolarity about 315 mOsm. The A549 cells were continuously bathed in regular ECS delivered through a multibarrel perfusion system. All solutions flowed at a speed of about 5 ml/min through silicone tubing that connected the glass solution reservoirs and the perfusion barrels, which were directed to the test cells. To reduce the evaporative loss of volatile anesthetics, any solution reservoir containing a volatile anesthetic was capped and the silicone tube clamped before the solution was applied. Solutions were rapidly applied to the test cell for a short duration by switching the perfusion barrels via a computer-controlled device (SF-77B, Warner Instruments LLC, Hamden, CT). To evaluate the potency of GABA for GABA_A receptors in A549 cells, GABA at concentrations of 0.1–100 μm was applied to the cells. In some experiments, the GABA_A receptor antagonist bicuculline and/or picrotoxin was also applied to the cells. To study the effect of isoflurane on the activity of GABA_A receptors, this anesthetic was co-applied during application of GABA. To study the effect of sevoflurane on the activity of GABA_A receptors, GABA was co-applied during application of sevoflurane.

With the system in voltage clamp mode (with the holding potential set at –60 mV), whole cell patch clamp recordings were performed. Cells showing “leaky current” (>20 pA) were excluded from testing. GABA-induced currents were acquired online using pCLAMP software (Axon Instruments) and a Digidata 1322A or 1200 digitizer (Axon Instruments). The electrical signals were filtered (1 kHz) and saved in a computerized database. The digitized data were analyzed offline using Clampfit software (Axon Instruments), as previously described.^4,24,26,27  To construct the GABA dose–response plots for currents recorded in A549 cells, the peak of GABA-evoked current was measured. For analyzing the effect of isoflurane on GABA-induced currents, the averaged amplitude of GABA-evoked current elicited during application of isoflurane was normalized to that recorded 20 ms before application of isoflurane.

At 24 h after plating, A549 cells grown in culture dishes or on glass coverslips were treated with muscimol (30 μm), isoflurane, or isoflurane plus bicuculline (50 μm) for 4 h. For isoflurane treatment, isoflurane-saturated Dulbecco’s modified Eagle medium was prepared on the day of use at room temperature (22–24°C) and stored in a tightly sealed glass container. Dulbecco’s modified Eagle medium (2 ml) containing 250 μm isoflurane was added to each cell culture dish every 30 min over a period of 3.5 h (i.e., eight times). All A549 cells were used 4 h after the treatment.

Trypan blue staining was used to examine cell viability. Briefly, A549 cells cultured in 35-mm dishes were incubated, for 4 h, in control culture medium, in medium containing 250 μm of isoflurane, or in isoflurane-saturated medium. The cells were then incubated with control medium containing 0.2% trypan blue (Sigma) for 10 min. After two washings with phosphate-buffered saline, the cells were fixed with 4% paraformaldehyde for 10 min at room temperature. Multiple pictures of randomly selected trypan blue-treated cells in each dish were obtained under a Zeiss inverse microscope with 40× lens. The number of cells stained with trypan blue was counted and expressed as a percentage of all cells in the image field (i.e., 150–400 cells).

Statistical analyses were performed with SigmaPlot software (Systat Software Inc., Chicago, IL). For analyzing GABA_A receptor activity, the peak amplitude of GABA-induced currents was normalized to either the maximum response or the control current. To determine the half-maximal effective concentration (EC₅₀) of GABA-induced currents in A549 cells, the normalized amplitude of GABA-induced current was plotted against the concentration of GABA, and the data were fitted with the equation f = (a/1) + (x/x₀)^b. Group data (two tailed) were examined with paired or unpaired Student t tests, as appropriate. All data are expressed as mean ± SD. P values less than 0.05 were considered significant.

The immunohistochemical assays showed that ATII cells stained positively for presurfactant protein C were located in the alveolar area of mouse lungs (fig. 1A) and human lungs (fig. 1B), as described previously.²⁸  The ATII cells were immunoreactive to antibodies to the α2 and π subunits of the GABA_A receptor (fig. 1, A and B), as well as to an antibody selective for both the β2 and β3 subunits (data not shown). Next, immunocytochemistry was performed on the human ATII cell line A549, under nonpermeabilized conditions, to confirm the expression of GABA_A receptor subunits by these cells and to determine whether the receptors are expressed on the plasma membrane. Specifically, A549 cells were immunostained with the antibody directed against both the β2 and β3 subunits, and immunofluorescent clusters of β2/β3 subunits were identified on the cell surface (fig. 1C). Immunoblotting of lysates of A549 cells revealed protein bands that corresponded to α5, β2, β3, δ, and π subunits of the GABA_A receptor (fig. 1D).

To study the functional properties of GABA_A receptors in ATII cells, whole cell voltage clamp recordings were performed in A549 cells. The application of GABA (0.1–100 μm) evoked an inward current that increased in amplitude with increasing concentrations of GABA (fig. 2A-1). The threshold GABA concentration that evoked a detectable current was approximately 0.1– 0.3 µm (n = 6). The peak amplitude of the current generated by 100 μm GABA was 117 ± 38 pA (n = 9). The EC₅₀ and the maximal effective concentration (EC_Max) of GABA, estimated from curves fitted to the concentration–response plot for GABA-evoked currents in A549 cells (n = 6), were 2.45 ± 0.28 and 32.66 ± 6.32 μm, respectively (fig. 2A-2). Currents evoked by GABA concentrations above 3.0 µm peaked rapidly and then decayed or “desensitized” to a steady state. The peak current activated by 3 µm GABA was 84.5 ± 3.6% of that evoked by 100 µm GABA (n = 6). The ratio of peak to steady state of the current evoked by 3 µm GABA was 0.62 ± 0.02 (n = 8), similar to that of the current evoked by 100 µm GABA (0.61 ± 0.04, n = 9).

Surprisingly, the current evoked by 30 μm GABA was only partially inhibited by the GABA_A receptor antagonist bicuculline. Specifically, the amplitude of current evoked by 30 μm GABA in the absence and presence of a typically saturating concentration of bicuculline (100 µm) was 80 ± 7.3 pA (n = 7) and 27 ± 4.1 pA (n = 7), respectively (fig. 2, B-1 and B-2). Higher concentrations of bicuculline did not further reduce the current. In contrast, the current evoked by 30 µm GABA was strongly inhibited by another GABA_A receptor antagonist, picrotoxin (50 µm; fig. 2B-2). This result suggests that the pharmacologic properties of GABA_A receptors in ATII cells differ significantly from those of GABA_A receptors in most neurons,²⁹  as the receptors in neurons are highly sensitive to both bicuculline and picrotoxin when studied under identical recording conditions.²⁹ 

Next, we investigated whether isoflurane modulates the activity of GABA_A receptors in ATII cells. Specifically, we tested the effects of isoflurane (2.5–250 μm) on the currents evoked by 0.5, 10, and 100 µm GABA (I_GABA0.5, I_GABA10, and I_GABA100, respectively). As depicted in figure 3, A and B, isoflurane enhanced the amplitude of I_GABA0.5 and I_GABA10 in a dose-dependent manner, with similar EC₅₀ values (EC₅₀ of isoflurane on I_GABA0.5: 16.2 ± 2.9 μm, n = 7 cells; EC₅₀ of isoflurane on I_GABA10: 19.66 ± 5.4 μm, n = 6 cells; P = 0.11; fig. 3B). However, the maximum effect of isoflurane on I_GABA0.5 was significantly greater than its maximum effect on I_GABA10 (I_GABA0.5: 2.4 ± 0.6 times of the amplitude of control current, n = 7 cells; I_GABA10: 1.6 ± 0.2 times of the amplitude of control current, n = 6 cells; P < 0.001; fig. 3B). Notably, a remarkable “after response” of the GABA-evoked current (I_GABA0.5 or I_GABA10) was observed when a high concentration (83 or 250 μm) of isoflurane was applied to the cell (fig. 3A, lowest row, and C-1). Moreover, high concentrations of isoflurane actually decreased the amplitude of I_GABA100, and terminating the application of isoflurane elicited a large and long-lasting rebound current in some cells (fig. 3C-2). Notably, isoflurane applied alone did not evoke any current (data not shown).

To examine whether other volatile anesthetics also regulate the function of GABA_A receptors in ATII cells, we tested the effect of sevoflurane on GABA-evoked currents in A549 cells. Sevoflurane displayed a comparable effect on the activity of GABA_A receptors in A549 cells. Specifically, at a low concentration (1:1000 dilution), sevoflurane enhanced the current induced by 10 μm GABA (fig. 3D-1), but at a high concentration (1:10 dilution), it inhibited the current (fig. 3D-2). Taken together, these results indicate that volatile anesthetics modify the activity of atypical GABA_A receptors in ATII cells in a complex fashion, as they both potentiate and inhibit the current, depending on the activity levels of the receptors.

Given that ATII cells express GABA_A receptors that are highly sensitive to isoflurane and sevoflurane, we next sought to determine whether an anesthetic, through upregulation of GABA_A receptor activity, would modify a key inflammation-related function in these cells. Immunohistochemistry combined with confocal microscopy revealed the immunoreactivity of COX-2 in surfactant-expressing alveolar cells of control mouse lungs (fig. 4A), which indicates that the gene encoding for COX-2 is constitutively expressed in ATII cells. Similarly, immunoreactivity of COX-2 was detected in ATII cells of human lungs (fig. 4B) and in control A549 cells, with high-magnification microscopic images displaying immunoreactive clusters of COX-2 in the cytosol of these cultured cells (fig. 5, A-1 to A-4, with comparison to negative control in fig. 5A-5). Analysis of the number of COX-2 clusters via Image-J software (fig. 5B) revealed significantly fewer clusters of COX-2 in A549 cells treated with the GABA_A receptor agonist muscimol (30 μm; fig.5, A-2 and C) relative to control. This finding indicates that stimulation of GABA_A receptors in ATII cells inhibits COX-2 expression.

Autocrine GABA_A receptors in ATII cells, including A549 cells, are constantly activated by endogenously secreted GABA.^2,3  We treated A549 cells with isoflurane (initial concentration 250 μm) to examine whether this anesthetic affects COX-2 expression by increasing autocrine GABA_A receptor activity. Immunocytochemistry showed that isoflurane treatment decreased the immunoreactivity of COX-2 in A549 cells (fig. 5, A-3 and C). The isoflurane-induced reduction in COX-2 expression was partially reversed by co-treating the A549 cells with bicuculline (50 μm; fig. 5, A-4 and C). The inhibitory effects of both muscimol and isoflurane on COX-2 expression in A549 cells were confirmed with immunoblotting assays (fig. 5D). Exposure of neuronal cells to a high concentration of isoflurane causes apoptosis.³⁰  Indeed, almost all of the A549 cells treated with a high concentration of isoflurane (initial concentration 2500 μm) stained for trypan blue (95.1 ± 3.9%; fig. 5E, Iso-S). However, it is unlikely that the inhibition of COX-2 expression by a lower concentration of isoflurane was the result of a cytotoxic effect on the cells, given that very few of the A549 cells treated with the test concentration of isoflurane (initial concentration 250 μm) stained for trypan blue (1.76 ± 0.21%; fig. 5E, Iso-250), a result that was not significantly different from that for control cells (0.12 ± 0.2%; fig. 5E, Ctrl).

To the best of our knowledge, this is the first study to show that the activity of GABA_A receptors in ATII cells is upregulated by an inhaled anesthetic. Specifically, the study shows that (1) ATII cells in human lungs express GABA_A receptors; (2) GABA_A receptors in A549 cells exhibit a high affinity for GABA and relative resistance to inhibition by bicuculline (compared with GABA_A receptors in neurons); (3) isoflurane and sevoflurane at clinically relevant concentrations increase the activity of GABA_A receptors in ATII cells in a complex manner; and (4) isoflurane decreases expression of COX-2 in ATII cells, an effect that is mediated, at least in part, via GABA_A receptors.

ATII cells in the alveolar areas of human and rodent lungs expressed subunits of the GABA_A receptor (fig. 1, A and B). The expression of various GABA_A receptor subunits in A549 cells was confirmed by immunoblots (fig. 1, C and D). More importantly, these GABA_A receptor subunits formed functional channels in A549 cells, as evidenced by the picrotoxin-sensitive GABA-evoked currents in the cells (fig. 2).

Our results suggest that the pharmacologic properties of GABA_A receptors in ATII cells differ fundamentally from those expressed in neurons in several important respects. First, the EC₅₀ and EC_Max of GABA in A549 cells in the current study were approximately 2.5 and 32 µm, respectively (fig. 2A-2), whereas the EC₅₀ and EC_Max of GABA in central neurons are much higher: 19 and 600 μm, respectively.²⁹  These differences indicate that the GABA receptors in ATII cells have a high affinity for their endogenous ligands. The high affinity of GABA_A receptors in A549 cells likely results from the unique subunit composition of these receptors. In particular, π subunits are abundant in ATII cells³  but are rarely expressed in neurons.³¹  Receptors containing the π subunit typically display a higher affinity for GABA than receptors without this subunit.³² 

Second, ATII cells express glutamic acid decarboxylase, a key enzyme involved in the synthesis of GABA.^2,3  It has been postulated that these cells are the primary source of GABA in the alveoli and that GABA_A receptors localized to the apical membrane of alveolar epithelia are activated by GABA in an autocrine fashion.³  ATII cells do not generate action potentials; and are devoid of any mechanisms that would facilitate the synchronized release of large amounts of GABA into the alveolus, as occurs at synapses in the central nervous system. The high affinity for GABA of the GABA_A receptors in ATII cells would allow these receptors to sense the low ambient concentration of endogenous GABA. A hypothetical scheme by which GABA_A receptors are activated in ATII cells is shown in figure 6. Here, we propose that GABA molecules released from ATII cells disperse into the alveolar liquid and spread over the vast surface area of the alveolus. It is conceivable that a low concentration of GABA is continually present in the alveolar surface liquid, causing persistent activation of GABA_A receptors in the ATII cells. The physiological role of this GABA_A receptor-mediated tonic current is now being studied, but it has already been reported that GABA–GABA_A receptor signaling in ATII cells regulates alveolar water–electrolyte homeostasis.³ 

In addition, our results revealed that GABA_A receptors in ATII cells were partially inhibited by high concentrations of bicuculline but were effectively blocked by picrotoxin (fig. 2B). Bicuculline-insensitive GABA_C receptors were previously described as pentameric constructs containing only ρ subunits. These receptors could be effectively blocked by picrotoxin but not bicuculline.³³  The ρ subunits may also be expressed heterogeneously in cells, forming hetero-oligomers with other GABA_A receptor subunits.^34,35  As such, the ρ1–ρ3 subunits have been classified as GABA_A receptor subunits.^36,37  The relative insensitivity to bicuculline of GABA_A receptors in A549 cells raises the possibility that these receptors may include ρ subunits. Consistent with this suggestion, high levels of messenger RNA for the ρ1–ρ3 subunits have been identified in rat ATII cells.³  Confirmation of the expression of ρ subunits in primary ATII cells of human lungs awaits further immunohistochemical assays with specific antibodies.

Isoflurane had complex effects on the activity of GABA_A receptors in ATII cells, as it both potentiated and inhibited the currents evoked by different concentrations of GABA. Such multifaceted effects of isoflurane exhibited several interesting features. First, the degree of enhancement by isoflurane of GABA-evoked currents decreased with increasing concentrations of GABA (fig. 3, A and B). Furthermore, high concentrations (≥250 μm) of isoflurane inhibited the currents evoked by a high concentration of GABA (100 μm; fig. 3C-2). Terminating the administration of high-concentration isoflurane caused a marked rebound of the GABA-evoked currents (fig. 3C-2). The molecular mechanism underlying these complex effects of isoflurane on the activity of GABA_A receptors in ATII cells awaits further investigation. In this regard, previous studies in neurons showed that high concentrations of GABA desensitize GABA_A receptors.³⁸  Isoflurane inhibition of GABA-evoked currents may be secondary to an increase in desensitization of the receptors, as has been observed in neurons.³⁹ 

The current study revealed that isoflurane at clinically relevant concentrations (8.3–83 μm) significantly facilitated GABA_A receptor activity evoked by low concentrations of GABA (0.5 and 10 μm; fig. 3, A and B). As both the total alveolar surface and the corresponding total volume of alveolar liquid are very large, the concentration of endogenous GABA in the alveoli must be low. As illustrated in figure 6, we postulate that under in vivo conditions, GABA_A receptors in ATII cells are constantly activated by the low concentration of GABA and that inhaled isoflurane promptly diffuses into the alveolar liquid, thus upregulating the activity of GABA_A receptors in ATII cells. In the current study, the reported isoflurane concentrations were calculated on the basis of our previously reported measurement of isoflurane solutions.²⁵  As some isoflurane could evaporate from solutions during tests, the effect of isoflurane on GABA_A receptor activity in A549 cells might be underestimated.

The expression of COX-2 increases rapidly in response to numerous inflammatory stimulants in a variety of cells. COX-2 is constitutively expressed in ATII cells, as evidenced by immunostaining of ATII cells isolated from human lungs (fig. 4) and A549 cells (fig. 5, A and D). COX-2 is a key enzyme involved in the production of prostaglandin E2,^40,41  a bioactive substance that is critical for various pulmonary functions. For example, it increases the production of surfactant in ATII cells.⁴²  Prostaglandin E2 is also a potent mediator of lung inflammation.^43,44  Given that isoflurane decreases prostaglandin E2 levels in the hypothalamus⁴⁵  and also suppresses the biosynthesis of surfactant in the lung,^12,46  we postulate that isoflurane may inhibit COX-2 expression in ATII cells by modulating GABA_A receptors. Indeed, our results from both the immunocytochemical and immunoblotting assays confirmed that treating A549 cells with either muscimol or isoflurane significantly reduced the expression of COX-2. The effect of isoflurane on COX-2 expression was largely prevented by bicuculline (fig. 5), which confirms that the effect of isoflurane was realized through GABA_A receptors.

Available data have shown that the expression level of COX-2 in ATII cells changes rapidly in the presence of different concentrations of inflammatory cytokines.⁴⁷  In microglia, a decrease in COX-2 expression happens after diminishing calcium (Ca²⁺) entry through store-operated channels⁴⁸  (also known as Ca²⁺ release-activated Ca²⁺ [CRAC] channels). The opening of CRAC channels occurs in response to Ca²⁺ store depletion following various stimulations. The activated CRAC channels allow Ca²⁺ to enter down its concentration and electrical gradients, and the CRAC channel-mediated Ca²⁺ current is characterized by a pattern of inwardly rectifying.⁴⁹  It is proposed that activating GABA_A receptors in T-lymphocytes results in membrane depolarization that in turn decreases Ca²⁺ entry through CRAC channels⁵⁰  by reducing the electrical driving force for Ca²⁺. CRAC channels are expressed in ATII cells.⁵¹  Therefore, we hypothesize that isoflurane increases the activity of GABA_A receptors in ATII cells, causing more Cl⁻ efflux and greater membrane depolarization, which in turn leads to a decrease in the entry of Ca²⁺ through CRAC channel and hence a decline in COX-2 expression (fig. 6). Collectively, data from this study suggest that inhaled anesthetics, including isoflurane, produce antiinflammatory action in the lung, at least in part, by enhancing the activity of GABA_A receptors in ATII cells.

The authors thank Ella Czerwinska, M.Sc., Research Technician, Departments of Anesthesia and of Physiology, University of Toronto, Toronto, Ontario, Canada, for culturing cells for some studies; and Michael Jackson, Ph.D., Adjunct Professor, Department of Physiology and Pharmacology and Robarts Research Institute, University of Western Ontario, London, Ontario, Canada, for critically commenting on the manuscript.