Naloxone and Ouabain in Ultralow Concentrations Restore Na+/K+-ATPase and Cytoskeleton in Lipopolysaccharide-treated Astrocytes*

Astrocytes respond to inflammatory stimuli and may be important modulators of the inflammatory response in the nervous system. This study aimed first to assess how astrocytes in primary culture behave in response to inflammatory stimuli concerning intracellular Ca2+ responses, expression of Toll-like receptor 4 (TLR4), Na+/K+-ATPase, actin filament organization, and expression of cytokines. In a cell culture model with lipopolysaccharide (LPS), astrocyte response was assessed first in the acute phase and then after incubation with LPS for 1–48 h. The concentration curve for LPS-stimulated Ca2+ responses was bell-shaped, and the astrocytes expressed TLR4, which detects LPS and evokes intracellular Ca2+ transients. After a long incubation with LPS, TLR4 was up-regulated, LPS-evoked Ca2+ transients were expressed as oscillations, Na+/K+-ATPase was down-regulated, and the actin filaments were disorganized. Interleukin-1β (IL-1β) release was increased after 24 h in LPS. A second aim was to try to restore the LPS-induced changes in astrocytes with substances that may have dose-dependent anti-inflammatory properties. Naloxone and ouabain were tested separately in ultralow or high concentrations. Both substances evoked intracellular Ca2+ transients for all of the concentrations from 10−15 up to 10−4 m. Neither substance blocked the TLR4-evoked Ca2+ responses. Naloxone and ouabain prevented the LPS-induced down-regulation of Na+/K+-ATPase and restored the actin filaments. Ouabain, in addition, reduced the IL-1β release from reactive astrocytes. Notably, ultralow concentrations (10−12 m) of naloxone and ouabain showed these qualities. Ouabain seems to be more potent in these effects of the two tested substances.

Infection, tissue damage, or other conditions can lead to a neuroinflammatory state. Lipopolysaccharide (LPS) exposure can be used experimentally as a model to generate an inflammatory response both in vitro and in vivo. Endotoxin, or LPS, is a potent inflammatory activator (1) in astrocytes and microglia, with stimulation of Toll-like receptor 4 (TLR4) (2). The expres-sion of TLR4 is up-regulated in astrocytes under neuroinflammatory conditions (3). Activation of TLR4 leads to activation of nuclear factor-B, a factor that regulates expression of genes involved in immune responses, including release of the proinflammatory cytokines tumor necrosis factor-␣ (TNF-␣) and interleukin-1␤ (IL-1␤) (2).
Astrocytes in networks, positioned between the vasculature and synapses, monitor neuronal signaling, including rebuilding of synapses (4,5). Astrocytes express almost the same repertoire of receptors and ion channels as neurons. They regulate synaptic transmission via a bidirectional communication with neurons, and they release gliotransmitters as well as factors, including cytokines, fatty acid metabolites, and free radicals (6 -8).
The Ca 2ϩ signaling in astrocytes over long distances is analogous to, but much slower than, the propagation of action potentials in neurons (9). These astrocytic Ca 2ϩ waves (10,11) can be evoked by transmitters released from neurons and glial cells, followed by activation of especially G protein-coupled receptors. Cytosolic Ca 2ϩ plays a key role as a second messenger, and the control of Ca 2ϩ signals is therefore critical. This involves coordination of Ca 2ϩ entry across the plasma membrane (PM), 2 Ca 2ϩ release from the endoplasmatic reticulum (ER), refilling of the ER stores, and extrusion across the PM (12). The Na ϩ -Ca 2ϩ exchanger, a Ca 2ϩ transporter that controls the intracellular Ca 2ϩ concentrations, is driven by the Na ϩ electrochemical gradient across the PM. This Na ϩ pump, Na ϩ /K ϩ -ATPase, indirectly modulates Ca 2ϩ signaling (13), and inflammatory stimuli influence Ca 2ϩ homeostasis in the astrocyte networks (5, 14 -16).
The cytoskeleton seems to be important in this system for controlling of PM microdomains and the ER complex. The adaptor protein ankyrin B is associated with the Na ϩ pump and also with ER proteins, such as the inositol 1,4,5-trisphosphate (IP 3 ) receptor. The main cytoplasmic matrix of proteins, spectrin and actin, are attached to ankyrin B. An intact cytoskeleton is required for the propagation of astrocytic Ca 2ϩ waves (17), and disruption of the cytoskeleton abolishes the Ca 2ϩ oscillations by changing the balance between the Ca 2ϩ -regulating processes (18).
Parameters associated with neuroinflammation are downregulation of Na ϩ transporters, changing of Ca 2ϩ signaling in the astrocyte networks, and release of cytokines (16, 19 -25). With LPS as an inflammatory inducer, we wanted to evaluate how astrocytes behave concerning Ca 2ϩ signaling and cytokine release after inflammatory stimulation. We also wanted to evaluate how the astrocytes behave over time during incubation with LPS. We hypothesized that the astrocytes are influenced by the inflammation inducer LPS and that it is possible to restore some aspects of the cells to a homeostatic state using exposure to a substance that may have anti-inflammatory properties. We aimed to test this using naloxone, particularly in very low concentrations (26,27). We further aimed to test ouabain in low and high concentrations and its effects of Na ϩ /K ϩ -ATPase activity (28) and evoked Ca 2ϩ transients in astrocytes (28 -30).

EXPERIMENTAL PROCEDURES
In this study, the in vitro model involved astrocytes co-cultured with brain endothelial cells (16,24). The biological rationale for co-culturing astrocytes with endothelial cells is that the astrocytes are affected by substances released from the capillary endothelial cells of the BBB (31). These interactions are essential for a functional neurovascular unit (4,32). The endothelial cells are not directly in physical contact with the astrocytes, and interaction in the model is brought about through the shared medium. The co-cultured astrocytes are morphologically differentiated with long, slender processes, and they exhibit greater Ca 2ϩ responses and cytokine release than monocultured astrocytes (16,24).
Chemicals-All chemicals were obtained from Sigma-Aldrich if not stated otherwise.
Microvascular Endothelial Primary Cultures-Brain capillary fragments were isolated and endothelial cells cultured according to Hansson and co-workers (16, 24) using a modified version of the method used by Abbott and co-workers (34).
Astrocytes Co-cultured with Adult Rat Brain Microvascular Primary Cultures-The experimental astrocytes were obtained after co-cultivation with primary brain microvascular endothelial cultures and primary astrocytic cultures. Astrocytic cultures at 6 days in vitro were co-cultured with newly prepared microvascular cultures. The endothelial cells were grown in inserts above the astrocytic cultures. The cells from the two different cultures were never in contact. The cells were grown together for 9 -11 days. At the time of the experiments, the astrocyte cultures were 15-17 days old, including the 9 -11 days of cocultivation. The endothelial cells were removed before the experimental procedure (16,24).
Calcium Imaging-Astrocytes co-cultured with brain microvascular endothelial cells were incubated at room temperature with the Ca 2ϩ -sensitive fluorophore probe Fura-2/AM (Invitrogen Molecular Probes, Inc., Eugene, OR) for 30 min (8 l in 990 l of Hanks' HEPES-buffered saline solution, containing 137 mM NaCl, 5.4 mM KCl, 0.4 mM MgSO 4 , 0.4 mM MgCl 2 , 1.26 mM CaCl 2 , 0.64 mM KH 2 PO 4 , 3.0 mM NaHCO 3 , 5.5 mM glucose, and 20 mM HEPES, dissolved in distilled water, pH 7.4). The fluorophore probe was dissolved with 40 l of DMSO and 10 l of pluronic acid (Molecular Probes, Leiden, The Netherlands). After incubation the cells were rinsed three times with Hanks' HEPES-buffered saline solution before exposure to a substance. The antagonists were applied 3.5 min before the agonist. LPS, ouabain, or naloxone, were tested to verify if they induced Ca 2ϩ transients on their own. To determine the underlying Ca 2ϩ source of the LPS-, ouabain-, or naloxone-evoked Ca 2ϩ transients, internal stores were depleted by preincubation with a sarcoendoplasmic reticulum Ca 2ϩ -ATPase inhibitor, thapsigargin (1 M) and caffeine (20 mM) (35), or incubated with a Ca 2ϩ -free buffer (exchange of CaCl 2 with MgCl 2 , and 1 mM EGTA). The experiments were performed at room temperature using a calcium imaging system and Simple PCI software (Compix Inc. Imaging Systems, Hamamatsu Photonics Management Corp., Cranberry Twp, PA) and an inverted epifluorescence microscope (Nikon ECLIPSE TE2000-E) with a ϫ20 (numerical aperture 0.45) fluorescence dry lens and a Polychrome V, monochromator-based illumination system (TILL Photonics GmbH). The various substances were infused into the cell baths using a peristaltic pump (Instech Laboratories, Plymouth Meeting, PA) at an approximate rate of 600 l/min. One minute before the start of the recording sequence, the stimulating substance was injected into the pump tubes for 30 s. The substance took ϳ80 s to reach the cells through the tubes. Hanks' HEPES-buffered saline solution continued to flow through the pump tubes and onto the cells throughout the experiment. The images were captured with an ORCA-12AG (C4742-80-12AG), high resolution digital cooled CCD camera (Hamamatsu Photonics Corp.).
The total areas under the transients (i.e. amounts of Ca 2ϩ released) (36) were analyzed to provide measures of the vigor of the Ca 2ϩ responses. The amplitude was expressed as the maximum increase of the 340/380 nm ratio. The area under the Ca 2ϩ peaks was calculated in Origin (Microcal Software Inc., Northampton, MA).
The cells were washed with PBS-BSA-Sap. for 3 ϫ 5 min and then incubated with a mixture of FITC-conjugated F(abЈ) 2 donkey anti-rabbit IgG and Texas Red-conjugated F(abЈ) 2 donkey anti-mouse IgG secondary antibodies (Jackson Immunoresearch, Westgrove, PA), both diluted 1:150, and the nuclei marker Hoescht 33258 diluted 1:1000 in PBS-BSA-Sap. The cells were washed with PBS-BSA-Sap three times for 5 min each and finally rinsed with PBS. The coverslips were mounted on microscope slides with a fluorescent mounting medium (DAKO, Glostrup, Denmark) and viewed in a Nikon Eclipse 80i microscope. Pictures were taken with a Hamamatsu C5810 color-intensified 3CCD camera.
Viability Assay-A LIVE/DEAD viability assay kit (Invitrogen Molecular Probes) for mammalian cells was used, and assays were performed as follows.
The cells were gently washed twice with Dulbecco's PBS (Invitrogen). A mixture of 4 M ethidium homodimer-1 (EthD-1) and 2 M calcein AM diluted in Dulbecco's PBS was added, and the cells were incubated for 30 -45 min at room temperature. Following incubation, 10 l of fresh LIVE/DEAD reagent solution or Dulbecco's PBS was added to a clean microscope slide. Cells were then mounted upside down and immediately viewed in a Nikon Eclipse 80i microscope. Pictures were taken with a Hamamatsu C5810 color-intensified 3CCD camera.
SDS-PAGE and Western Blotting-Cells were rinsed twice in PBS and immediately lysed for 20 min on ice in cold radioimmune precipitation assay lysis buffer containing 50 mM Tris-HCl, 150 mM NaCl, 1 mM phenylmethylsulfonyl fluoride, 1 mM EDTA, 1% Triton X-100, 1% sodium deoxycholate, 0,1% SDS, 5 g/ml aprotinin, and 5 g/ml leupeptin, pH 7.4. The procedure was done according to Persson et al. (37). Separate aliquots were taken for protein concentration determination. All samples were correlated for total protein contents, and an equal loading of 20 g total protein of each sample was applied in each lane of the gel. ␤-Actin was used as control for equal loading.
SDS-PAGE was conducted using the Novex precast gel system (Invitrogen) according to the manufacturer's recommendations using 4 -12% BisTris gels (Invitrogen) at 200 V for 50 min. The separated proteins were then transferred at 30 V for 60 min to a nitrocellulose membrane (Invitrogen) using NuPAGE transfer buffer (Invitrogen) supplemented with methanol and NuPage antioxidant. The membranes were rinsed twice with distilled water, and the proteins were visualized with Ponceau S solution (Sigma). Proteins were blocked with 5% fat-free skim milk (Semper AB) in TBST (50 mM Tris-HCl, 150 mM NaCl, and 0.05% Tween) for 60 min at room temperature. The membranes were then probed with primary antibodies, washed four times for 2 min each with TBST, subsequently probed with secondary horseradish peroxidase (HRP)-conjugated secondary antibodies, and finally washed several times in TBST. The primary antibodies used were TLR4 (rabbit polyclonal) (Santa Cruz Biotechnology, Inc.) diluted 1:500 and Na ϩ /K ϩ ATPase (␣-subunit) (mouse monoclonal) (Sigma) diluted 1:250. The secondary antibodies used were HRP-conjugated donkey anti-mouse or -rabbit F(abЈ) 2 fragment (both from Jackson Immunoresearch) diluted 1:10000. All primary and secondary antibodies were diluted in 5% fat-free skim milk in TBST. Protein was then detected with an enhanced chemiluminescence kit (PerkinElmer Life Sciences) and visualized with a Fuji Film LAS-3000 (Tokyo, Japan).
Cytokine Release-Co-cultured astrocytes were stimulated for 24 h with 10 ng/ml LPS diluted in unsupplemented minimum Eagle's medium to measure IL-1␤ release. The inserts with endothelial cells were removed just before incubation. When anti-inflammatory substances were used, the cells were incubated for 30 min with the antagonists before LPS treatment and remained in the medium throughout the experiment. Control (i.e. untreated) cells were kept in unsupplemented minimum Eagle's medium. Supernatants were collected for enzymelinked immunosorbent assay (ELISA), and the cells underwent protein measurements.
Rat IL-1␤ (Nordic Biosite, Täby, Sweden) were used according to the manufacturer's instructions to measure the amounts of cytokine released. Between every incubation step, several washings were performed. The protein concentration was  100%). B, LPS in a Ca 2ϩ -free medium evoked Ca 2ϩ transients in half of the responding cells. C, to determine the underlying Ca 2ϩ source, internal stores were depleted with thapsigargin, followed by caffeine, which was used to complete the store depletion. All cells were blocked. D, all cells were blocked with thapsigargin, caffeine, and Ca 2ϩ -free buffer. The amplitudes were expressed as the maximum increase of the 340/380 nm ratio. The cells were from four different coverslips, from two different seeding times. Results shown are from a typical experiment. FIGURE 3. Astrocytes express TLR4. A, astrocytes express TLR4 visualized with Western blot, the protein band expressed at 89 kDa. B, astrocytes were stained for the TLR4 receptor (green), revealing that these cells do express TLR4 receptors. The astrocytes were stained for glial fibrillary acidic protein (GFAP) (red), and the nuclei were visualized with HO33258 (blue). C, astrocytes stimulated with LPS were blocked with the TLR4 antagonist LPS-RS. These results show that astrocytes express TLR4 and that LPS-induced Ca 2ϩ transients were blocked by the selective TLR4 antagonist LPS-RS.
determined as described below. The amount of IL-1␤ release was correlated to the protein amount for each well.
Protein Determination-The protein determination assay was performed in accordance with the manufacturer's instructions using a DC Protein Assay (Bio-Rad), based with some modifications on the method used by Lowry et al. (38). Both a standard (0 -4 mg/ml BSA) and samples were mixed with the reagents, incubated for 15 min at room temperature, read at 750 nm with a Versa-max microplate reader, and analyzed using SoftMax Pro 4.8, both from Molecular Devices (Sunnyvale, CA).
Actin Visualization-The astrocytic cytoskeleton was evaluated by staining of actin filaments after incubation with 10 ng/ml LPS for 0, 1, 4, 8, 24, or 48 h. Cultures were fixed with 4% paraformaldehyde and made permeable with PBS (Invitrogen) containing 1% BSA and 0.05% saponin followed by an Alexa TM 568-conjugated phalloidin probe (Invitrogen) diluted 1:40 in PBS supplemented with 1% BSA. The coverslips were rinsed three times in PBS and then mounted on microscope slides with Dako's fluorescent mounting medium (Dako Co., Glostrup, Denmark) before viewing with a fluorescence dry objective lens attached to an inverted Nikon Optiphot-2 microscope.
Statistics-The level of significance was analyzed using oneway analysis of variance followed by Dunnet's multiple comparisons test. Error bars show S.E.

RESULTS
Stimulation with LPS-Astrocytes were stimulated with LPS (0.01-100 ng/ml) and Ca 2ϩ imaging experiments were performed on 15-17-day-old co-cultured astrocytes. The endothelial cell inserts were removed just before the experiments. In 0.01 ng/ml LPS, 47.5% of the astrocytes responded (19 of 40 cells); in 0.10 ng/ml LPS, 57.5% responded (23 of 40 cells); in 1.0 ng/ml LPS, 67.5% responded (27 of 40 cells); in 10 ng/ml LPS, 73.0% responded (29 of 40 cells); and in 100 ng/ml LPS, 25.0% responded (10 of 40 cells). The amplitudes and areas under the Ca 2ϩ peaks were calculated, and the number of peaks was counted for each cell. The concentration curve showed a bellshaped appearance visualized both with maximum ratio increase and amount of Ca 2ϩ released (Fig. 1). The astrocytes were most sensitive to LPS at 0.10, 1.0, and 10 ng/ml. In all further experiments, the cells were stimulated with 10 ng/ml LPS.
To investigate if the Ca 2ϩ elevation was cross-membrane or intracellular-dependent, LPS in a Ca 2ϩ -free medium evoked Ca 2ϩ transients in half of the responding cells. All cells were blocked when the cells were incubated with thapsigargin, a depletor of intracellular Ca 2ϩ stores (33), followed by caffeine, which was used to complete the store depletion (33). All cells were blocked with thapsigargin and Ca 2ϩ -free buffer (Fig. 2).
Astrocytes Express TLR4-Astrocytes express TLR4 visualized with Western blot (Fig. 3A). The astrocytes were stained for glial fibrillary acidic protein and for the TLR4, revealing that these cells do express TLR4 receptors. The cell nuclei are visualized with HO33258 (Fig. 3B). Astrocytes stimulated with LPS were blocked with the TLR4 antagonist LPS-RS (39) (Fig. 3C).
Time-dependent Stimulation with LPS-For astrocytes incubated with LPS for 30 min or 1, 4, 6, or 24 h, the LPS-evoked Ca 2ϩ transients were captured. At 6 and 24 h of exposure of LPS, the LPS-induced Ca 2ϩ transients increased as well as the areas under curves; p Ͻ 0.001 at 6 h, p Ͻ 0.01 at 24 h (Fig. 4, A  and B). The amplitudes in the control are the same as in Fig. 2A. The viability of the cells visualized with a live-dead kit shows very few dead cells with red nuclei in the figure. The viability did not change over time (Fig. 4C). TLR4 was up-regulated when the cells were incubated with LPS for 4 h and visualized with Western blot (Fig. 4D).
Na ϩ /K ϩ -ATPase Expression Over Time-Expression of Na ϩ /K ϩ -ATPase was studied with Western blot for cultures incubated with LPS for 1, 4, 8, 24, or 48 h. The expression of Na ϩ /K ϩ -ATPase initially increased and decreased after 8 h of incubation (Fig. 5).
Actin Filaments-Untreated astrocytes stained with an Alexa TM 488-conjugated phalloidin probe were dominated by F-actin organized in stress fibers. After a 1-or 4-h incubation with LPS, respectively, small amounts of ring-formed actin filaments were observed. At 8 h, these ring structures became more pronounced, and the actin filaments were organized more diffusely. These rearrangements of the actin filaments became further apparent after the cells were incubated with LPS for 24 or 48 h (Fig. 6).
Stimulation with Naloxone-Astrocytes were stimulated with naloxone (10 Ϫ15 to 10 Ϫ4 M), and Ca 2ϩ imaging was recorded where all astrocytes responded (40 of 40 cells in the respective concentrations). One peak was observed, and the areas under the curves were calculated. Naloxone-evoked Ca 2ϩ transients were obtained in both low and high concentrations (Fig. 7A).
To investigate if the Ca 2ϩ elevation was cross-membrane or intracellular-dependent, thapsigargin and/or Ca 2ϩ -free buffer was used. Naloxone (10 Ϫ12 or 10 Ϫ5 M) in a Ca 2ϩ -free medium evoked the same Ca 2ϩ transients as control, suggesting that Ca 2ϩ elevation occurs due to efflux from the storage in ER. All cells were blocked when the cells were incubated with thapsigargin followed by caffeine. All cells were blocked with thapsigargin, caffeine, and Ca 2ϩ -free buffer (Fig. 7B).
Naloxone did not block LPS-induced Ca 2ϩ transients. This was visualized by first stimulation with LPS, after which the same cells were treated with naloxone (10 Ϫ5 M) 30 s before LPS was added to the medium quality (Fig. 7C). Naloxone did not block TLR4, which detects LPS-evoked Ca 2ϩ transients.
The main results here were that naloxone evoked Ca 2ϩ transients in both low and high concentrations. Ca 2ϩ was released from intracellular stores, and naloxone did not block the TLR4 Ca 2ϩ -evoked responses.
Stimulation with Ouabain-Astrocytes were stimulated with ouabain (10 Ϫ15 to 10 Ϫ3 M), and Ca 2ϩ imaging experiments were performed. All astrocytes responded, 100% (40 of 40 cells in the respective concentrations). The Ca 2ϩ transients were evoked partly as oscillations, and the areas under the curves were calculated. The concentration curve showed that ouabain evoked Ca 2ϩ transients in both low and high concentrations (Fig. 8A).  To investigate if the Ca 2ϩ elevation was cross-membrane or intracellular-dependent, thapsigargin and/or Ca 2ϩ -free buffer was used. Ouabain (10 Ϫ12 or 10 Ϫ5 M) in a Ca 2ϩ -free medium evoked the same Ca 2ϩ transients as control, an indicator that Ca 2ϩ elevation occurred due to efflux from the storage in ER. All cells were inhibited when the cells were incubated with thapsigargin followed by caffeine. No Ca 2ϩ transient amplitudes were seen when the cells were blocked with thapsigargin, caffeine, and Ca 2ϩ -free buffer (Fig. 8B).
Ouabain did not block LPS-induced-Ca 2ϩ transients. This was visualized by first stimulation with LPS; thereafter, the same cells were treated with ouabain (10 Ϫ5 M) 30 s before LPS was added to the medium (Fig. 8C).
The main results here were that ouabain evoked Ca 2ϩ transients in both low and high concentrations. Ca 2ϩ was released from intracellular stores, and ouabain did not block the TLR4 Ca 2ϩ -evoked responses.
Naloxone and Ouabain in Ultralow Concentrations Prevent LPS Down-regulation of Na ϩ /K ϩ -ATPase-Astrocytes were incubated with naloxone or ouabain (10 Ϫ12 and 10 Ϫ5 M) 30 min before the cells were treated with naloxone or ouabain, respectively, and LPS for 1, 4, 8, 24, or 48 h. The expression of Na ϩ / K ϩ -ATPase visualized with Western blot was unaltered during the time curves for both naloxone and ouabain (Fig. 9). Naloxone as well as ouabain prevented LPS-induced down-regulation of Na ϩ /K ϩ -ATPase.
Naloxone and Ouabain Restore Actin Filaments-When astrocytes were incubated with LPS for 24 h, the actin filaments were reorganized. In cells treated with naloxone or ouabain (10 Ϫ12 or 10 Ϫ5 M), 30 min before the cells were incubated with both LPS and naloxone, respectively, the disrupted actin filaments were to a great extent restored. At 10 Ϫ12 M, the actin filaments were better restored for both naloxone and ouabain compared with 10 Ϫ5 M (Fig. 10).
Ouabain Reduces IL-1␤ Release-Astrocytes incubated with LPS for 24 h released IL-1␤. Cells were treated with naloxone or ouabain (10 Ϫ12 or 10 Ϫ5 M) 30 min before the cells were incubated with both LPS and naloxone or ouabain, respectively. Ouabain, at low or high concentrations, reduced the LPS-induced IL-1␤ increased release (Fig. 11).

DISCUSSION
Regulation of Ca 2ϩ dynamics by transmitters and soluble factors is a possible mechanism by which the astrocyte network detects changes in the CNS microenvironment and regulates brain activities, including inflammatory processes. The complexity of Ca 2ϩ signaling has made it difficult to determine the physiological role of these phenomena, although this probably is an important function of astrocytes. During inflammation, TLR4 is activated, and astrocytes express TLR4, as demonstrated in the present study and by others (39,40). LPS has been shown to be a TLR4 agonist (41), and this receptor is thought to be largely responsible for LPS-related signaling (2). It has been FIGURE 7. Stimulation with naloxone. A, astrocytes were loaded with the Ca 2ϩ probe, Fura-2/AM, for 20 min and stimulated with naloxone (10 Ϫ15 to 10 Ϫ4 M). All astrocytes responded, 100% (40 of 40 cells in the respective concentrations). One peak was observed, and the areas under the curves (AUC) were calculated. The concentration curve showed that naloxone evoked Ca 2ϩ transients in low and high concentrations. n ϭ 40 from four different coverslips, from two different seeding times. Data are mean Ϯ S.E. (error bars). Naloxone-evoked Ca 2ϩ transients were dependent on intracellular Ca 2ϩ stores. B (top left) 10 Ϫ12 or 10 Ϫ5 M naloxone elicited Ca 2ϩ responses in astrocytes, in all cells tested (40 of 40; 100%). Top right, Ca 2ϩ -free medium evoked similar Ca 2ϩ transients as controls. Bottom left, To determine the underlying Ca 2ϩ source, internal stores were depleted with thapsigargin, followed by caffeine, which was used to complete the store depletion (40 of 40 cells; 100%). Bottom right, similar results were seen with thapsigargin, caffeine, and Ca 2ϩ -free medium. The amplitudes were expressed as the maximum increase of the 340/380 ratio. The cells were from four different coverslips, from two different seeding times. Results shown are from a typical experiment (10 Ϫ12 M).
Naloxone did not block LPS-evoked Ca 2ϩ transients. C, naloxone was not able to block LPS-induced Ca 2ϩ transients visualized by stimulation with LPS. The same cells were treated with naloxone (10 Ϫ5 M) 30 s before LPS was added to the medium.
proposed that TLRs drive inflammation that gives rise to symptoms (42,43). This study was focused on demonstrating that astrocytes react to inflammatory activation. LPS is frequently used as a model for the induction of inflammation both in vitro and in vivo. In this study, astrocyte-enriched cultures from newborn rat brain were co-cultured with endothelial cultures from adult rat brain. These astrocytes express TLR4, a feature that is further induced after incubation with LPS. This is in contrast to findings in astrocyte-enriched cultures with no cocultivation, where TLR4 is expressed rather sparsely and is more difficult to induce. Although LPS normally does not readily pass the intact BBB, LPS itself can impair the integrity of the BBB and thereby affect the flux system between blood and brain (44). Therefore, although our findings in the present study must primarily be conceived as purely in vitro model data, they may have some bearing also for the understanding of in vivo systemic actions following the release of LPS (e.g. following brain trauma or sepsis), when a disruption of the BBB is prone to occur.
In this study, we found that acute stimulation of astrocytes with LPS affects the Ca 2ϩ signaling; LPS-elicited Ca 2ϩ transients were observed in a concentration-dependent bell-shaped distribution, where also the induced Ca 2ϩ peak was blocked by the TLR4 antagonist LPS-RS. The Ca 2ϩ for these transients originated from intracellular stores. Next, we observed that astrocyte behavior changed over time when they were incubated with LPS for longer time periods; the one peak LPSevoked Ca 2ϩ transient changed to Ca 2ϩ oscillations after 1 h, and the Ca 2ϩ oscillations increased in intensity from 6 h on and were still dominant at 24 h. It has been proposed that the size of the intracellular Ca 2ϩ store in astrocytes is important when the Ca 2ϩ response goes from transient to oscillatory and that this could be a property of reactive astrocytes (19). This has also been shown when astrocytes were stimulated by growth factors (19). A similar mechanism for converting this response pattern is shown in the present study when astrocytes are exposed to the inflammatory stimulus LPS. This can be one mechanism by which astrocytes detect changes and regulate brain activities such as processes of inflammation.
In renal inflammation, LPS induces down-regulation of sodium transporters, such as Na ϩ /K ϩ -ATPase, Na ϩ -K ϩ -2Cl Ϫ co-transporter, Na ϩ /H ϩ exchanger, and epithelial sodium channel (20). LPS also induces down-regulation of sodium transporters in liver cells (21), kidney cells (22), and intestinal epithelial cells (23). In vitro in cortical duct cells, proinflammatory cytokines, such as TNF-␣ and IL-1␤, decreased the expression of Na ϩ /K ϩ -ATPase (20). From observations in different cellular systems, it can be surmised that inflammation (not only cytokine release) is important as well as down-regulation of Na ϩ transporters. Down-regulation of Na ϩ transporters means dysfunction of the Na ϩ /K ϩ -ATPase activity (20), which can . Bottom left, to determine the underlying Ca 2ϩ source, internal stores were depleted with thapsigargin, followed by caffeine, which was used to complete the store depletion. Bottom right, similar results were seen with thapsigargin, caffeine, and Ca 2ϩ -free medium. The amplitudes were expressed as the maximum increase of the 340/380 ratio. The cells were from four different coverslips, from two different seeding times. Results shown are from a typical experiment (10 Ϫ12 M). Ouabain did not block LPS-evoked Ca 2ϩ transients. C, ouabain was not able to block LPS-induced Ca 2ϩ transients visualized by stimulation with LPS. The same cells were treated with ouabain (10 Ϫ5 M) 30 s before LPS was added to the medium. lead to increased intracellular Na ϩ concentration. Astrocytes have strong resistance to Na ϩ influx only when Na ϩ /K ϩ -ATPase activity is maintained (45). This is why we evaluated protein expression for Na ϩ /K ϩ -ATPase in astrocytes after the cells were incubated with LPS over time. We found that the protein expression of Na ϩ /K ϩ -ATPase was reduced after 8 h, as an indication that the Na ϩ transporters were down-regulated.
The cytoskeleton controlling the PM microdomains and the ER complex is important with regard to calcium transients. The adaptor protein ankyrin B is associated with Na ϩ /K ϩ -ATPase and also with ER proteins, such as IP 3 . The main cytoplasmic matrix proteins, spectrin and actin, are attached to ankyrin B. An intact cytoskeleton is required for the propagation of astrocytic Ca 2ϩ waves (17), and disruption abolishes the Ca 2ϩ oscillations by changing the balance between the Ca 2ϩ -regulating processes (18). We found that the actin filaments were disorganized after many h of LPS exposure. The actin filaments, which are normally organized in stress fibers, changed into ring structures after 1 h of treatment and were disrupted and disorganized at 24 h. We have observed similar patterns previously when astrocytes were exposed to ethanol, ammonium chloride, or lactate (46,47). Others have seen morphological changes induced by LPS in cultured astrocytes (48,49). In line with earlier published data, the present findings confirm that there is a connection between rearrangement of actin filaments and changes in astrocytic morphology and cell volume (46,47).
Another notable finding in our study concerned substances that may have properties with anti-inflammatory qualities and up-regulation or restoration of markers related to inflammation: Na ϩ /K ϩ -ATPase protein expression, actin filament organization, Ca 2ϩ oscillations, and IL-1␤ release. We have reasoned that when Na ϩ /K ϩ -ATPase is down-regulated, TLR4 is activated, and when Na ϩ /K ϩ -ATPase has a normal pump activity, TLR4 is not overexpressed, although it is still unclear if a direct causal connection exists between these processes. Our findings showed that TLR4 protein expression was up-regulated when the astrocytes were incubated with LPS in a timedependent manner. LPS activated TLR4, leading to more intracellular Ca 2ϩ time-dependent release. In a previous study (16), increased IL-1␤ production was noted, and this seems to be initiated through TLR4 activation (50). The Na ϩ /K ϩ -ATPase is an energy-transducing pump, and expression of these proteins was down-regulated with time. Inhibiting or modulating this pump has an influence on the intracellular Na ϩ concentration, which in turn results in an increases in intracellular Ca 2ϩ concentration via the Na ϩ -Ca 2ϩ exchange (28). We chose two substances to study in LPS-activated astrocytes: naloxone and ouabain. We opted to use them instead of more traditional anti-inflammatory agents because naloxone in ultralow concentrations has been suggested to have anti-inflammatory properties (27) in addition to its powerful effect as a -opioid receptor antagonist. Also, ultralow dose naloxone is effective in restoration of the antinociceptive effect of morphine in rats and in addition increases anti-inflammatory cytokine IL-10 expression (26).
Ouabain, a digitalis-derived glycoside is a well recognized Na ϩ /K ϩ -ATPase inhibitor, especially pronounced at high concentrations, and is also known to regulate intracellular Ca 2ϩ release in cardiac myocytes, peritoneal macrophages, and astrocytes in high concentrations (13,29,51). Ouabain leads to increases in intracellular Ca 2ϩ release in hippocampal astrocytes, which is IP 3 receptor-dependent (29). Ouabain also enhances LPS-down-regulated inducible nitric-oxide synthase activity in peritoneal macrophages (51). Interestingly, at low concentrations (nanomolar and picomolar), ouabain stimulates Na ϩ /K ϩ -ATPase activity (28) and activates complex signaling cascades in kidney cells (30). In subnanomolar concentrations, ouabain stimulates proliferation and differentiation of cardiac and smooth muscle cells (52).
In the present study, we observed that stimulation with naloxone evoked intracellular Ca 2ϩ transients that were IP 3 receptor-dependent. The intracellular Ca 2ϩ release was not concentration-dependent, and this was unexpected. Both at ultralow and high concentrations, Ca 2ϩ transients had the same appearance. Stimulation with ouabain also evoked intracellular Ca 2ϩ transients that were IP 3 receptor-dependent, and the intracellular Ca 2ϩ release was not concentration-dependent. Ouabain and naloxone behaved in a similar way regarding Ca 2ϩ transients in both ultralow and high concentrations. At high ouabain concentrations, Na ϩ /K ϩ -ATPase activity is inhibited, resulting in an elevation of intracellular Na ϩ concentration, and this in turn raises intracellular Ca 2ϩ concentration via diminished Na ϩ -Ca 2ϩ exchange (28). At low concentrations, Na ϩ /K ϩ -ATPase is instead stimulated via activation of the Scr/ epidermal growth factor receptor intracellular way, and this raises intracellular Ca 2ϩ concentration via the phospholipase C/IP 3 receptor pathway (28). This means that intracellular FIGURE 10. Naloxone and ouabain restored actin filaments. Astrocytes were stained with an Alexa TM 488-conjugated phalloidin probe. The untreated culture was dominated by F-actin organized in stress fibers. Cultures were preincubated with naloxone (Nalox.; 10 Ϫ12 or 10 Ϫ5 M) or ouabain (Ouab.; 10 Ϫ12 or 10 Ϫ5 M) for 30 min before the cultures were treated with LPS (10 ng/ml) for 1, 4, 8, 24, or 48 h. Both naloxone and ouabain restored actin filaments. FIGURE 11. IL-1␤ release. Cultures were preincubated with naloxone (10 Ϫ12 or 10 Ϫ5 M) or ouabain (10 Ϫ12 or 10 Ϫ5 M) for 30 min before the cultures were treated with LPS (10 ng/ml) for 24 h. Ouabain in both low and high concentrations reduced the LPS-increased IL-1␤. n ϭ 4. The level of significance was analyzed using one-way analysis of variance followed by Dunnet's multiple comparisons test. Data are mean Ϯ S.E. ***, p Ͻ 0.001; n.s., nonsignificant.
Ca 2ϩ release comes from intracellular Ca 2ϩ stores but via two different mechanisms.
Because the LPS triggered astrocytes exert a lowered Na ϩ / K ϩ -ATPase activity, we found it relevant to investigate whether this down-regulation could be counteracted by naloxone and/or ouabain in ultralow concentrations. These two substances also had the ability to restore the actin filaments, which was even more pronounced at ultralow concentration. This shows the importance of a well working Na ϩ pump activity, which is coupled to the cellular cytoskeleton. Most notable here is that ultralow concentrations seem enough to induce the effects. In addition, ultralow concentrations of ouabain led to decreased IL-1␤ release. The mechanisms behind the reduced IL-1␤ release need to be further evaluated. The biosynthesis of IL-1␤ is complex and is regulated at multiple levels. In microglia, it has been shown that LPS stimulates TLR4, which induced IL-1␤ production depending on NF-B, and that the purinergic receptor P2X7 is involved (53). However, ouabain did not inhibit the TLR4 Ca 2ϩ -evoked responses in our astrocytes. Furthermore, naloxone is known to reduce IL-1␤ synthesis in rats (27), which we were unable to reproduce in our astrocytes.
In conclusion, using a co-cultured model of astrocytes, we have shown that after a long incubation with LPS, TLR4 is upregulated, LPS-evoked Ca 2ϩ transients are expressed as oscillations, Na ϩ /K ϩ -ATPase is down-regulated, and the actin filaments are disorganized. Two substances with proposed anti-inflammatory properties in a low dose range, naloxone and ouabain, demonstrate the ability to limit these astrocytic LPSinduced alterations. Naloxone evokes intracellular Ca 2ϩ transients but not in a concentration-dependent manner, and there is a similar pattern with ouabain, although neither blocks TLR4. Ultralow concentrations of naloxone prevent the LPS-induced down-regulation of Na ϩ /K ϩ -ATPase and restore the actin filaments. Ouabain demonstrates the same effect, although ouabain also reduces IL-1␤ release. Further study is warranted concerning modulation of the inflammatory response in astrocytes.