We recently reported that Na+/H+ exchanger isoform 1 (NHE1) activity in astrocytes is stimulated and leads to intracellular Na+ loading after oxygen and glucose deprivation (OGD). However, the underlying mechanisms for this stimulation of NHE1 activity and its impact on astrocyte function are unknown. In the present study, we investigated the role of the ERK1/2 pathway in NHE1 activation. NHE1 activity was elevated by ∼75% in NHE1+/+ astrocytes after 2-h OGD and 1-h reoxygenation (REOX). The OGD/REOX-mediated stimulation of NHE1 was partially blocked by 30 μM PD-98059. Increased expression of phosphorylated ERK1/2 was detected in NHE1+/+ astrocytes after OGD/REOX. Moreover, stimulation of NHE1 activity disrupted not only Na+ but also Ca2+ homeostasis via reverse-mode operation of Na+/Ca2+ exchange. OGD/REOX led to a 103% increase in intracellular Ca2+ concentration ([Ca2+]i) in NHE1+/+ astrocytes in the presence of thapsigargin. Inhibition of NHE1 activity with the NHE1 inhibitor HOE-642 decreased OGD/REOX-induced elevation of [Ca2+]i by 73%. To further investigate changes of Ca2+ signaling, bradykinin-mediated Ca2+ release was evaluated. Bradykinin-mediated intracellular Ca2+ transient in NHE1+/+ astrocytes was increased by ∼84% after OGD/REOX. However, in NHE1−/− astrocytes or NHE1+/+ astrocytes treated with HOE-642, the bradykinin-induced Ca2+ release was increased by only ∼34%. Inhibition of the reverse mode of Na+/Ca2+ exchange abolished OGD/REOX-mediated Ca2+ rise. Together, our data suggest that ERK1/2 is involved in activation of NHE1 in astrocytes after in vitro ischemia. NHE1-mediated Na+ accumulation subsequently alters Ca2+ homeostasis via Na+/Ca2+ exchange.
- intracellular pH
- cortical astrocytes
- sodium/calcium exchange
- intracellular sodium ion
na+/h+ exchanger isoform 1 (NHE1) mediates H+ efflux and Na+ influx through the cell membrane and plays a significant role in intracellular pH (pHi) regulation and recovery from acidosis (28). Inhibition of NHE1 reduces the resting pHi of cortical astrocytes of rats (23, 30) and mice (13). We recently reported (13) that NHE1 activity in cortical astrocytes is stimulated by ∼75% during reoxygenation (REOX) after 2-h oxygen and glucose deprivation (OGD). Stimulation of NHE1 activity leads to an ∼5.6-fold increase in intracellular Na+ concentration ([Na+]i). However, the mechanisms underlying this stimulation of NHE1 activity and its impact on ion homeostasis of astrocytes are unknown.
NHE1 can be directly stimulated by acidosis via a H+-mediated allosteric effect (36). In addition, NHE1 is regulated by phosphorylation, which alters the affinity of the transmembrane internal H+ transport site of NHE1 (9, 24, 28). Activation of the ERK pathway and one of its downstream effectors, p90 ribosomal S6 kinase (p90RSK), has been reported to stimulate NHE1 activity during myocardial ischemia and reperfusion (24). There is growing evidence implicating ERK1/2 activation in promotion of cell death during oxidative neuronal injury (7). However, whether the ERK signaling cascade plays a role in stimulation of NHE1 activity in astrocytes after in vitro ischemia remains unknown.
Overstimulation of NHE1 activity may cause ischemic cell injury. Inhibition of NHE1 attenuates myocardial damage after ischemia and reperfusion (2). The NHE-specific inhibitor N-aminoiminomethyl-1-methyl-1-indole-2-carboxamide methanesulfonate (SM-20220) significantly reduces Na+ accumulation and edema in brain after transient cerebral ischemia in rats (14). Elevation of [Na+]i can trigger reverse-mode operation of Na+/Ca2+ exchanger and causes intracellular Ca2+ overload, which in turn may lead to Ca2+-dependent irreversible cell damage during anoxia and ischemia (6, 18). In the present study, we investigated whether NHE1-mediated accumulation of intracellular Na+ affects intracellular Ca2+ signaling in astrocytes after in vitro ischemia. We report here that stimulation of NHE1 activity in astrocytes after in vitro ischemia depends in part on activation of the ERK1/2 signaling pathway. NHE1 activation not only causes Na+ accumulation but also alters Ca2+ signaling and loading of intracellular Ca2+ stores.
MATERIALS AND METHODS
Eagle's modified essential medium (EMEM) and HBSS were from Mediatech Cellgro (Herndon, VA). FBS was obtained from Hyclone Laboratories (Logan, UT). Collagen type I was from Collaborative Biomedical Products (Bedford, MA). The acetoxymethyl esters of BCECF (BCECF-AM), fura-2 (fura-2 AM), and sodium-binding benzofuran isophthalate (SBFI-AM) were obtained from Molecular Probes (Eugene, OR). Pluronic acid was purchased from BASF (Ludwigshafen, Germany). Nigericin, dibutyryl cAMP (DBcAMP), monensin, FCCP, oligomycin, bradykinin, thapsigargin, and gramicidin were purchased from Sigma (St. Louis, MO). HOE-642 was a kind gift from Aventis Pharma (Frankfurt, Germany). 2-[4-[2,5-(Difluorophenyl)methoxy]phenoxy]-5-ethoxyaniline (SEA-0400) was from EMD Biosciences (San Diego, CA). 2-(2-Amino-3-methoxyphenyl)-4H-1-benzopyran-4-one (PD-98059) was from Tocris Cookson (Ellisville, MO). Antibody for phosphorylated ERK1/2 (pERK1/2) [Tyr(P)204/Thr(P)202] was from Cell Signaling Technologies (Beverly, MA), and antibody for total ERK (K-23) was from Santa Cruz Biotechnology (Santa Cruz, CA).
Animals and Genotype Analysis
The NHE1-null mutant (NHE1−/−) mice used in the present study were generated as previously described (3). We obtained NHE1 homozygous mutant and wild-type (NHE1+/+) mice by breeding gene-targeted NHE1 heterozygous mutant mice. Tail biopsies were obtained from 1-day-old mice. Genotypes were determined using PCR analysis of DNA from tail biopsies as described in our previous study (13).
Primary Culture of Mouse Cortical Astrocytes
Dissociated cortical astrocyte cultures were established as described previously (31). In brief, cerebral cortices were removed from 1-day-old NHE1+/+ or NHE1−/− mice. The cortices were incubated in a trypsin solution (0.25 mg/ml in HBSS) for 25 min at 37°C, and the tissue was mechanically triturated. The dissociated cells were rinsed and resuspended in EMEM containing 10% FBS. Viable cells (1 × 104 cells/well) were plated in six-well plates or on glass coverslips coated with collagen type 1. Cultures were maintained in a 5% CO2 atmosphere at 37°C and refed every 3 days throughout the study. To obtain morphologically differentiated astrocytes, confluent cultures (10 days in culture, DIV10) were treated with EMEM containing 0.25 mM DBcAMP to induce differentiation. DBcAMP has been widely used to mimic neuronal influences on astrocyte differentiation (10, 33). Experiments were routinely performed in DIV15–DIV28 astrocytes. This culture age was chosen because immature astrocyte cultures are resistant to damage under ischemic conditions (12).
NHE1+/+ or NHE1−/− cultures grown on coverslips were rinsed twice with an isotonic OGD solution containing (in mM, pH 7.4) 0 glucose, 20 NaHCO3, 120 NaCl, 5.36 KCl, 0.33 Na2HPO4, 0.44 KH2PO4, 1.27 CaCl2, and 0.81 MgSO4. Cells were incubated in 1.0 ml of OGD solution in a hypoxic incubator at 37°C (Forma model 3130, Thermo Forma, Marietta, OH) equilibrated with 94% N2-1% O2-5% CO2. The oxygen level in the medium of cultured cells was monitored with an oxygen probe (model M1-730, Microelectrodes, Bedford, NH) and decreased to ∼2–3% after 60 min in the hypoxic incubator. Normoxic control cells were incubated in 5% CO2 and atmospheric air in isotonic control buffer containing 5.5 mM glucose, with the rest of the components in the buffer identical to those in the isotonic OGD solution.
Measurement of pHi
The HEPES-buffered solution (pH 7.4) and pH calibration solutions were described in our previous study (13). Cells grown on coverslips were incubated with 2.5 μM BCECF-AM in HEPES-buffered solution for 10 min at ambient temperature as described previously (13). The coverslips were washed and placed in an open-bath imaging chamber (model RC24, Warner Instruments, Hamden, CT) containing HEPES-buffered solution. The chamber was mounted on the stage of the TE 300 inverted epifluorescence microscope, and the astrocytes were visualized with a ×40 Super Fluor oil immersion objective lens (13). The cells were excited every 10–30 s at 440 and 490 nm, and the ratio of the emission fluorescence at 535 nm was recorded. Images were collected with a Princeton Instruments MicroMax charge-coupled device (CCD) camera and analyzed with MetaFluor image-processing software (Universal Imaging, Downingtown, PA). The ratio of the fluorescence emissions (F490/F440) was calibrated using the high-K+/nigericin technique (5). The background-corrected data were fit with a variant of the Michaelis-Menten equation (5).
Cells were subjected to an acid load by the transient application (1–2 min) of a NH4+/NH3 solution as described previously (13). NH4+/NH3 solution was prepared by replacing 10 mM of NaCl in the HEPES-buffered solution with an equimolar concentration of NH4Cl. pHi rose as NH4+ accumulated during the prepulse. Cells were subsequently returned to a HEPES-buffered solution. Acidification of the cytoplasm occurred when NH3 quickly diffused out of the cell.
Determination of buffer capacity.
Buffer capacity (βi) was determined over a range of pHi values by subjecting the cells to progressively decreasing concentrations of NH4+, based on βi = ΔNH4+/ΔpHi (22). These experiments were done in HCO3−- and Na+-free buffers to block acid extrusion mechanisms. The change in intracellular NH4+ concentration ([NH4+]i) was determined from the intracellular NH3 concentration ([NH3]i), pK, and pH [pHi = 9.3 − log([NH4+]i/[NH3]i)], with an assumption of [NH3]i = extracellular NH3 concentration ([NH3]o) and Δ[NH+4]i = change in intracellular H+ concentration (Δ[H+]i) as reported previously (22). H+ flux rates (J, mM H+/min) were determined by multiplying βi by ΔpHi/Δt for each interval during pHi recovery (1–2 min) after prepulse treatment.
Intracellular Ca2+ Measurement
Astrocytes grown on coverslips were incubated with 5 μM fura-2 AM for 45 min (16). The cells were washed, and the coverslips were placed in the open-bath imaging chamber containing HEPES-MEM at ambient temperature. With a Nikon TE 300 inverted epifluorescence microscope and a ×40 Super Fluor oil immersion objective lens, astrocytes were excited every 10 s at 345 and 385 nm, and the emission fluorescence at 510 nm was recorded. Images were collected and analyzed with MetaFluor image-processing software. At the end of each experiment, the cells were exposed to 1 mM MnCl2 in Ca2+-free HEPES-MEM. The Ca2+-insensitive fluorescence was subtracted from each wavelength before calculations (16). The MnCl2-corrected 345-to-385-nm emission ratios were converted to concentration with the Grynkiewicz equation (8) as described previously (16).
Bradykinin-induced Ca2+ release from intracellular stores was assessed by exposing astrocytes to 1 μM bradykinin in Ca2+-free HEPES-MEM (1 mM EGTA) to prevent entry of extracellular Ca2+. Peak Ca2+ release was measured after application of bradykinin. After washout of bradykinin, cells were exposed to 1 μM thapsigargin to assess residual Ca2+ release from bradykinin-insensitive Ca2+ stores. Thapsigargin-induced peak Ca2+ release from intracellular stores was measured in the presence of 1 μM thapsigargin in HEPES-MEM (1.2 mM Ca2+).
Reverse mode of Na+/Ca2+ exchange (NCX) was induced in astrocytes as described by Hoyt et al. (11). Astrocytes were loaded with Na+ by applying Ca2+-free HEPES buffer (2 μg/ml gramicidin) for 1 min. Reverse mode of NCX was then initiated by exposing cells to a Na+-free buffer (1.2 mM Ca2+) for 30–40 s, which triggered a rise in [Ca2+]i. Cells quickly regulated [Ca2+]i to baseline values when control HEPES-MEM was reintroduced.
Gel Electrophoresis and Western Blotting
Cultured cells were washed with ice-cold PBS (pH 7.4) that contained 2 mM EDTA and protease inhibitors (32). Cells were collected by scraping from the six-well plates in an antiphosphatase buffer containing protease inhibitors (37). The protein content was measured by the bicinchoninic acid method (Pierce, Rockford, IL). Samples and prestained molecular mass markers (Bio-Rad, Hercules, CA) were denatured in SDS reducing buffer (1:2 by volume, Bio-Rad) and heated at 37°C for 15 min before gel electrophoresis. The protein samples (10 μg lysate protein) were loaded and separated by SDS-PAGE (12%). After transfer to a polyvinylidene difluoride membrane, the blots were incubated in 7.5% nonfat dry milk in Tris-buffered saline (TBS) and then incubated with a primary antibody, anti-pERK1/2 polyclonal antibody. For a protein loading control, the blot was restripped and reprobed with the anti-nonphospho-ERK1/2 polyclonal antibody, which recognized nonphosphorylated ERK 42-/44-kDa proteins. Both antibodies were used at a concentration of 1:1,000 in a solution of 7.5% milk-TBS buffer. After being rinsed, the blots were incubated with horseradish peroxidase-conjugated secondary IgG. Bound antibody was visualized by enhanced chemiluminescence assay (ECL, Amersham).
To elucidate a quantitative increase in pERK1/2, densitometric measurements of each protein band were performed. Protein bands on the blot were selected, and average pixel density was recorded with Un-Scan-It gel (Silk Scientific, Orem, UT). A ratio of pERK1/2 and nonphospho-ERK1/2 was calculated in normoxic control or OGD or REOX samples. The percent increase in pERK1/2 expression was then evaluated by comparing values in normoxic control vs. OGD or REOX samples.
Cultured cells grown on coverslips were rinsed with PBS (pH 7.4) and fixed with 4% paraformaldehyde in PBS for 20 min at room temperature. After being rinsed, cells were incubated with blocking solution (10% normal goat serum, 0.4% Triton X-100, and 1% bovine serum albumin in PBS) for 1 h. Cells were then incubated with anti-glial fibrillary acidic protein monoclonal antibody (1:100) or anti-pERK1/2 antibodies (1:100) in blocking solution for 1 h. Cells were rinsed with PBS and incubated with goat anti-mouse Alexa 488- or anti-rabbit Alexa 595-conjugated antibodies (1:200) for 1 h. Cell images were captured by a Nikon TE 300 inverted epifluorescence microscope (×20), a Princeton Instruments MicroMax CCD camera and MetaMorph image-processing software (Universal Imaging). Identical digital imaging acquisition parameters were used in negative control and experimental images.
Measurement of Cell Swelling
Cell volume change in single cells was measured by using calcein as a marker of intracellular water volume as described previously (16). Briefly, astrocytes grown on coverslips were placed in a closed-bath imaging chamber and incubated with 2.5 μM calcein AM. Calcein fluorescence in cells was monitored with a ×40 Super Fluor oil immersion objective and a FITC filter set (excitation 480 nm, emission 535 nm; Chroma Technology, Rockingham, VT) until the fluorescence plateaued (∼30 min). Images were collected every 60 s for 10 min to determine baseline data. The chamber was then switched to OGD buffer for 60 min. The fluorescence signals were corrected for baseline drift, and calibration with the calibration standard buffers was performed at the end of each experiment as described previously (16).
SigmaStat (Systat Software, Point Richmond, CA) software was used for statistical analysis. The data are reported as means ± SD; n values refer to the number of cells tested in each experimental condition. Significance between groups was tested with either a t-test or a Mann-Whitney test. An F-test was used for comparison of linear and nonlinear regression fits.
βi in NHE+/+ Astrocytes Does Not Change After OGD/REOX
To investigate changes in proton flux to reflect NHE1 activity, we first examined whether the intrinsic buffer capacity βi of NHE1+/+ astrocytes was altered by either OGD/REOX or genetic ablation of NHE1. βi was determined in NHE1+/+ and NHE1−/− astrocytes under control conditions and in NHE1+/+ astrocytes under OGD/REOX conditions. A representative buffer capacity experiment from a single cell is shown in Fig. 1A. Changes in pHi were induced by applying progressively decreasing concentrations of NH3/NH4+ (Fig. 1A). In NHE1+/+ astrocytes βi increased linearly from a value of ∼7.8 mM/pH unit at a pHi of 7.0 to ∼18 mM/pH unit at a pHi of 6.2 (Fig. 1B). The fitted regression was βi = 98 + (−12.9 * pHi), which is consistent with the reported values in rat astrocytes (4, 22, 30). βi was also determined in NHE1+/+ astrocytes after 2-h OGD and 1-h REOX (Fig. 1B). OGD/REOX did not cause a significant change in the fitted line of βi vs. pHi [βi = 75 + (−9.6 * pHi); P > 0.05]. Likewise, the line fitted to βi vs. pHi in NHE1−/− astrocytes was not different from that in NHE1+/+ astrocytes under control conditions [βi = 103 + (−13.6 * pHi); P > 0.05].
Stimulation of NHE1 Activity After OGD/REOX Partially Depends on Activation of MEK-Mediated Pathways
Figure 2A shows representative pHi tracings from NHE1+/+ astrocytes subjected to an acid load. After 2-h OGD and 15-min REOX there is an increase in pHi recovery rate in NHE1+/+ astrocytes (Fig. 2A). This increase in pHi recovery reflects an increase in NHE1 activity (13). pHi recovery in the presence of the MEK inhibitor PD-98059 was not different from the normoxic control (data not shown). In contrast, inhibition of MEK activity with 30 μM PD-98059 significantly attenuated the pHi recovery rate after OGD and REOX. To further investigate the kinetics of NHE1 activity under these conditions, we calculated J during the prepulse-induced pHi recovery and plotted it against the corresponding [H+]i (Fig. 2, B–D). In control NHE1+/+ astrocytes, the velocity of H+ efflux from cells increased as [H+]i increased and was best described by a sigmoid fit (Km of 0.264 ± 0.005 μM, Vmax of 7.21 ± 0.23 mM/min; Fig. 2B) with a Hill coefficient of 2.14 ± 0.11. The latter is consistent with other reports and suggests that there might be more than one binding site for H+ on the intracellular face of the protein (1). The values for Km and Vmax are comparable to those obtained for NHE1 expressed in fibroblasts as well as in C6 rat glioma cells (17, 29). When NHE1+/+ astrocytes were exposed to 2-h OGD plus 15-min REOX, the kinetics of NHE1 activity were dramatically changed (Fig. 2C). Vmax increased about three times (28.6 ± 2.2 mM/min), whereas Km shifted to a higher concentration of protons (0.71 ± 0.02 μM). The sigmoidal curve fit of the OGD/REOX data was significantly different from the curve fit to the normoxic control data in NHE1+/+ astrocytes (P < 0.05). The Hill coefficient for the OGD/REOX data was 2.03 ± 0.23. Inhibition of MEK1/2 activity with 30 μM PD-98059 affected OGD/REOX-mediated changes on NHE1 kinetics (Fig. 2D). Vmax was reduced by ∼41% in the presence of 30 μM PD-98059 (16.7 ± 0.4 mM/min) but still remained significantly higher than the normoxic control values. No further inhibition of Vmax was found with 60 μM PD-98059 (data not shown). In addition, in the presence of PD-98059, Km shifted back to a lower concentration of protons and was not significantly different from that of the normoxic control (0.377 ± 0.013 μM; P > 0.05). This suggests that the MEK1/2 and ERK1/2 pathways are partially responsible for activation of NHE1 activity in astrocytes after OGD/REOX.
We further investigated the OGD/REOX-mediated increase in ERK1/2 activity by measuring the phosphorylation status of ERK1/2 in astrocytes. As shown in Fig. 2E, top, when lysates of NHE1+/+ astrocytes were analyzed by immunoblotting, there were no significant differences in expression of total ERK1/2 proteins (∼42–44 kDa) in response to either OGD or OGD/REOX conditions. In contrast, either 1-h OGD or 2-h OGD + 1-h REOX led to an increase in expression of phosphorylated ERK1/2 proteins (pERK1/2, ∼42–44 kDa) in NHE1+/+ astrocytes (Fig. 2E, bottom). pERK1/2 expression was increased by 24% after 1-h OGD and by 16 ± 1% after 2-h OGD and 1-h REOX. The increase in pERK1/2 protein expression was further demonstrated in NHE1+/+ astrocytes after 1-h OGD by immunofluorescence staining. As shown in Fig. 2F, a and c, no significant morphological changes were found in NHE1+/+ astrocytes after 1-h OGD. However, immunoreactive signals of pERK1/2 protein was significantly increased in NHE1+/+ astrocytes at the end of 1-h OGD (Fig. 2F, b and d). This signal was absent when the anti-pERK1/2 primary antibody was omitted from the procedures (Fig. 2Fb, inset), which indicates that the immunofluorescent signal is specific to pERK1/2 protein.
To confirm that phosphorylation of ERKs correlates with NHE1 activation, we also determined NHE1 activity in astrocytes immediately after 1-h OGD. The NHE1-mediated H+ extrusion rate in NHE1+/+ astrocytes was significantly increased to 0.48 ± 0.022 pH U/min at 6.6 pHi after 1-h OGD, compared with a control of 0.33 ± 0.08 pH units/min at pHi 6.4 (P < 0.05, n = 3). Thus the rise in ERK phosphorylation correlates with NHE1 activation after either 1-h OGD or 2-h OGD/REOX.
Prolonged Acidosis Is Not Responsible for NHE1 Activation After OGD/REOX
It has been reported that prolonged acidosis of pHi = 6.55 ± 0.05 for 3 min in cultured rat ventricular cardiomyocytes led to ERK1/2-dependent stimulation of NHE1 activity (9). We previously found (13) that 2-h OGD did reduce pHi to 6.80 ± 0.02 from a resting level of 7.09 ± 0.07 in NHE1+/+ astrocytes. Therefore, we investigated whether prolonged acidosis plays a role in activation of NHE1 activity after OGD. NHE1+/+ astrocytes were exposed to 30 mM NH3/NH4+ for 1 min and were returned to the control HEPES buffer containing Na+. J was determined during pHi recovery (Fig. 3A). The prolonged acidosis was induced by exposing cells to the second NH3/NH4+ treatment and then returning them to a Na+-free HEPES buffer to maintain cellular acidosis (Fig. 3A). pHi was reduced from 6.99 ± 0.03 to 6.25 ± 0.07. After 3 min, extracellular Na+ was reintroduced and the change of J as pHi recovered was determined (Fig. 3B). The control kinetics in these experiments were Vmax = 7.93 ± 0.79 mM/min and Km = 0.30 ± 0.01 μM. After prolonged acidosis at pHi of 6.25, Vmax was 7.77 ± 0.51 mM/min and Km was 0.28 ± 0.01 μM. The line fit to the control data was not significantly different from the line fit to the prolonged acidosis data (P > 0.05). This suggests that the response of NHE1 activity to prolonged acidosis in cortical astrocytes is different from that of cardiomyocytes (9). Because NHE1 activity is not altered by the prolonged acidosis in NHE1+/+ astrocytes, it is unlikely that the ERK1/2-dependent stimulation of NHE1 activity after OGD/REOX is initially triggered by an increase in [H+]i.
Inhibition of ERK1/2 Pathways Attenuates NHE1-Mediated Increase in [Na+]i After OGD/REOX
We previously observed (13) that stimulation of NHE1 activity causes a substantial increase in [Na+]i in astrocytes at the end of 60-min REOX following 2-h OGD. In the present study, we further monitored NHE1-mediated intracellular Na+ accumulation every 5 min during 1-h REOX (Fig. 4). As shown in Fig. 4A, [Na+]i in NHE1+/+ astrocytes was 11.7 ± 3.4 mM at the end of 2-h OGD. It was not significantly different from normoxic control values (7.9 ± 1.5 mM). However, on REOX, [Na+]i rose sharply at a rate of 3.2 mM/min after 5 min of REOX, and it reached a plateau (71.4 ± 19.8 mM) at ∼25 min of REOX. In the presence of the NHE1 inhibitor HOE-642 (Fig. 4A), NHE1+/+ astrocytes exhibited a delayed rise in [Na+]i. A significant elevation of [Na+]i only occurred at 20 min of REOX, and its rate was much slower (0.8 mM/min). Moreover, the peak value of [Na+]i in the presence of HOE-642 was 37.7 ± 11 mM. Similar results were found in NHE1−/− astrocytes ([Na+]i =10.3 ± 3.4 mM at 0 min of REOX and 39.3 ± 11.3 mM at 60 min of REOX). These findings further confirm that NHE1 activation leads to accumulation of Na+ in astrocytes after OGD.
Because data in Fig. 2A suggest that activation of MEK1/2 and ERK1/2 pathways is involved in OGD/REOX-mediated stimulation of NHE1 activity, we next examined whether inhibition of MEK1/2 would block NHE1-mediated Na+ accumulation after OGD/REOX. As shown in Fig. 4B, 30 μM PD-98059 did not affect the basal levels of [Na+]i. In the presence of PD-98059, [Na+]i after 2-h OGD did not change (11.3 ± 1.3 vs. 11.8 ± 3.3 mM). In contrast to a sharp rise in [Na+]i at 10 min of REOX in the absence of PD-98059, [Na+]i increased at a markedly reduced rate (0.8 mM/min) in the treated astrocytes (P < 0.01). In the presence of PD-98059, although [Na+]i continued to increase during REOX over the next 50 min, it reached a value 59% less than in untreated astrocytes (33.4 ± 9.0 vs. 79.2 ± 17.2 mM; Fig. 4B). Interestingly, the effects of PD-98059 on the rate of Na+ increase and its peak value are similar to those observed in response to the NHE1 inhibitor HOE-642. The data imply that activation of MEK1/2 pathways may stimulate NHE1 activity and lead to a rise in Na+.
The OGD/REOX-mediated intracellular Na+ overload suggests that Na+ extrusion mediated by Na+-K+-ATPase is either reduced or remains unchanged and thus fails to maintain Na+ homeostasis. OGD (8 h) reduced Na+-K+-ATPase activity by ∼44% (measured by ouabain-sensitive Rb influx; Ref. 16). To further address this issue, we measured the Na+ accumulation rate in astrocytes during REOX in the absence or presence of the Na+-K+-ATPase inhibitor ouabain. The basal Na+ accumulation rate was 0.08 ± 0.07 mM/min under normoxic control conditions. Inhibition of Na+-K+-ATPase activity with 1 mM ouabain increased the Na+ accumulation rate to 2.9 ± 0.5 mM/min in normoxic astrocytes. However, OGD/REOX alone significantly elevated the Na+ accumulation rate (3.95 ± 0.40 mM/min; n = 3, P < 0.05) in the absence of ouabain. Blocking of Na+-K+-ATPase activity with ouabain did not significantly change the Na+ accumulation rate (3.16 ± 0.86 mM/min). This implies that Na+-K+-ATPase function is reduced after OGD/REOX and fails to maintain the physiological level of intracellular Na+. Thus the loss of Na+ homeostasis after OGD/REOX results from accelerated Na+ influx and decreased Na+ extrusion by Na+-K+-ATPase.
We previously demonstrated (13) that NHE1 activation leads to intracellular Na+ overload and swelling in astrocytes. We then examined whether blocking of NHE1 activation with PD-98059 affects cell swelling. As shown in Fig. 4C, inhibition of ERK pathways with PD-98059 significantly attenuated the OGD/REOX-induced swelling, and this effect is similar to that mediated by NHE1 inhibition by either HOE-642 or genetic ablation (13). This further supports our view that OGD/REOX-triggered stimulation of NHE1 activation depends in part on ERK activation.
Stimulation of NHE1 Activity Contributes to Increase in [Ca2+]i in NHE1+/+ Astrocytes After OGD/REOX
Figure 4 shows that activation of NHE1 activity after OGD/REOX led to an approximately fivefold increase in [Na+]i in NHE1+/+ astrocytes. We hypothesized that the NHE1-mediated Na+ overload could cause an influx of Ca2+ via a decrease and/or reversed operation of NCX. In this experiment, we measured changes in bulk [Ca2+]i after OGD/REOX. Basal [Ca2+]i under normoxic control conditions was 60 ± 5 nM. [Ca2+]i remained at 59 ± 12 nM in NHE1+/+ astrocytes after 2-h OGD and 1-h REOX. We speculated that a lack of increase in bulk [Ca2+]i after OGD/REOX might be due to sequestration of Ca2+ into the endoplasmic reticulum (ER) (16). To investigate this possibility, we monitored changes of [Ca2+]i when the ER Ca2+-ATPase was inhibited. As shown in Fig. 5A, application of 1 μM thapsigargin to NHE1+/+ astrocytes under normoxic conditions caused a transient increase in [Ca2+]i. The Ca2+ transient peaked at ∼1 min (382 ± 233 nM; Fig. 5A, inset) and declined to baseline values over the next 6 min. After OGD/REOX in NHE1+/+ astrocytes, the thapsigargin-induced Ca2+ transient increased to 777 ± 567 nM (Fig. 5A, inset), which suggests an increase in ER Ca2+ loading during OGD/REOX. In contrast, when NHE1+/+ astrocytes were treated with 1 μM HOE 642 during OGD/REOX, the thapsigargin-induced Ca2+ transient was only 431 ± 194 nM (Fig. 5A, inset), which was significantly lower than values in OGD/REOX (P < 0.001). We then examined whether a reduction in Ca2+ transient can be detected in NHE1−/− astrocytes (Fig. 5B). The thapsigargin-induced peak Ca2+ transient in NHE1−/− astrocytes was 316 ± 204 nM under normoxia. After OGD/REOX, it reached 421 ± 226 nM, which was not significantly different from normoxic controls in NHE1+/+ astrocytes (P > 0.05). These results indicate that stimulation of NHE1 activity affects ER Ca2+ loading after OGD/REOX.
We performed some experiments to examine whether inhibition of ERK pathways via PD-98059 would reduce Ca2+ overload after OGD/REOX. PD-98059 at 30 μM significantly reduced the thapsigargin-mediated rise in Ca2+ in NHE1+/+ astrocytes under either control or OGD/REOX conditions (133 ± 16 and 111 ± 36 nM, respectively; Fig. 5C). Thus ERK pathways are involved in Ca2+ loading under both normoxic and OGD/REOX conditions.
To further investigate a role for NHE1 in ER Ca2+ signaling after OGD/REOX, we measured releasable Ca2+ from the ER. As shown in Fig. 6, A and B, 1.0 μM bradykinin triggered Ca2+ release from intracellular Ca2+ stores and raised [Ca2+]i from a baseline of 69 ± 19 nM to 356 ± 151 nM in NHE1+/+ astrocytes under normoxic conditions. However, OGD/REOX caused an increase of bradykinin-induced Ca2+ release (654 ± 215 nM, P < 0.001) in NHE1+/+ astrocytes (Fig. 6, A and B). In NHE1−/− astrocytes, bradykinin evoked a 336 ± 79 nM Ca2+ release under normoxic control conditions, which was similar to that observed in NHE1+/+ cells (Fig. 6, C and D). In contrast, bradykinin triggered significantly less Ca2+ release in NHE1−/− astrocytes after OGD/REOX (450 ± 94 nM, P < 0.001; Fig. 6, C and D). This implies that OGD/REOX caused an increase in Ca2+ uptake by the intracellular Ca2+ stores in NHE1+/+ astrocytes but not in NHE1−/− astrocytes.
We then confirmed this finding by analyzing the effect of the potent NHE1 inhibitor HOE-642 on Ca2+ signaling. Treatment of NHE1+/+ astrocytes with 1 μM HOE-642 did not change either baseline [Ca2+]i (84 ± 14 nM) or bradykinin-induced Ca2+ release (342 ± 96 nM) under normoxic conditions (Fig. 7). In contrast, treatment of NHE1+/+ cells with 1 μM HOE-642 during OGD/REOX significantly decreased bradykinin-induced Ca2+ release (423 ± 120 vs. 654 ± 215 nM, P < 0.05; Fig. 7). This is consistent with the observations in NHE1−/− astrocytes.
As shown in Figs. 6 and 7, a small fraction of Ca2+ release was insensitive to bradykinin but triggered by thapsigargin. These results imply that astrocytes have both bradykinin-sensitive and bradykinin-insensitive Ca2+ stores. OGD/REOX led to an increase in thapsigargin-triggered Ca2+ release in NHE1+/+ but not in NHE1−/− astrocytes.
NHE1 Activity Contributes to Mitochondrial Ca2+ Accumulation in NHE1+/+ Astrocytes
Our data indicate that bradykinin-sensitive Ca2+ stores contain higher levels of Ca2+ after OGD/REOX. However, mitochondria are also known to buffer increases in cytosolic Ca2+ after pathological stimuli. We therefore investigated whether OGD/REOX affects Ca2+ loading in mitochondria of NHE1+/+ astrocytes. NHE1+/+ astrocytes were exposed to 10 μM FCCP and 2.5 μg/ml oligomycin under Ca2+-free conditions, after either 3 h of normoxia (Fig. 8A) or 2 h of OGD + 1 h of REOX (Fig. 8B). After a 10- to 20-s delay, [Ca2+]i increased in response to FCCP + oligomycin in both normoxic astrocytes (166 ± 32 nM) and OGD/REOX-treated astrocytes (261 ± 43 nM) (Fig. 8). Cytosolic Ca2+ was subsequently cleared (Fig. 8, A and B). When NHE1+/+ astrocytes were treated with 1 μM HOE-642 during OGD/REOX, the peak increase in [Ca2+]i was only 222 ± 47 nM and significantly smaller than it was in the absence of HOE 642 (P < 0.001; Fig. 8, B and C). These results suggest that mitochondria do buffer increases in cytosolic Ca2+ after OGD/REOX, but to a smaller degree than ER Ca2+ stores. In addition, stimulation of NHE1 activity also increases Ca2+ loading in mitochondria.
Increase in [Ca2+]i Is in Part a Result of Reverse-Mode Operation of NCX
In our recent study (16), we demonstrated that overloading of intracellular Na+ via Na+-K+-Cl− cotransporter-mediated Na+ influx resulted in reverse-mode operation of NCX and a rise in cytosolic Ca2+. This increase in NCX activity is blocked by the NCX blocker KB-R7943 (16). In the present study, we investigated whether reverse operation of NCX contributes to the increase in [Ca2+]i following stimulation of NHE1 activity as shown in Figs. 7 and 8. First, we established the effects of the potent and selective inhibitor SEA-0400 on the reverse mode of NCX in NHE1+/+ astrocytes (Fig. 9A). The reverse operation of NCX was triggered by removing extracellular Na+, which switches the NCX from the forward (Ca2+ exit) to the reverse (Ca2+ entry) mode. The reverse mode of NCX in NHE1+/+ astrocytes led to a large and repeatable increase in [Ca2+]i (Fig. 9A). However, in the presence of 100 nM SEA-0400, peak Ca2+ entry through NCX was reduced by ∼90% (Fig. 9A). This confirmed that SEA-0400 is a potent inhibitor of the reverse operation of NCX in astrocytes, consistent with the report in other cell types (20).
We investigated whether the NHE1-mediated effect on the loading of Ca2+ stores during OGD/REOX depends on the reverse-mode operation of NCX. As shown in Fig. 9B, both KB-R7943 (3 μM) and SEA-0400 (100 nM) significantly reduced the OGD/REOX-induced increase in the thapsigargin-induced Ca2+ transient. The untreated cells exhibited a rise in Ca2+ of 777 ± 567 nM after OGD/REOX (Fig. 9, B and C). In the presence of KB-R7943 or SEA-0400, Ca2+ transient values were reduced to 247 ± 107 and 322 ± 129 nM, respectively (Fig. 9C). This result further suggests that ER Ca2+ loading after OGD/REOX is attributable to Ca2+ influx via the reverse mode of NCX.
To further support this finding, we performed theoretical thermodynamic analysis of NCX in astrocytes. The dependence of NCX reversal potential on [Na+]i was modeled with LabHEART (version 4.9.5 simulation software; Ref. 27). The key variables, [Na+]o, [Ca2+]o, [Na+]i, and [Ca2+]i, were modified with our experimental data under normoxic control or OGD/REOX conditions. An NCX current-voltage plot was generated as [Na+]i was varied from 10 to 80 mM. The membrane voltage at which each simulation generated a current of zero was then plotted vs. [Na+]i (Fig. 10). With this simulation, a reversal potential of −43 mV was generated for control ionic conditions at [Ca2+]i of 100 nM. As [Na+]i increased, the reversal potential for NCX decreased. Resting plasma membrane potential in astrocytes has been reported to be about −60 mV (21); thus, under these conditions, NCX is in the forward mode and may help clear Ca2+ from the cell. The actual membrane potential during OGD/REOX is not measured in NHE1+/+ astrocytes and may remain either constant or depolarized. However, in either situation, the simulation predicts that an increase in [Na+]i during OGD/REOX will strongly favor an inwardly directed Ca2+ current (Fig. 10). Under ischemic conditions, [Ca2+]i will rise and may reduce the reverse-mode operation of NCX. To take this factor into consideration, a simulation was performed in which [Ca2+]i was elevated to 777 nM, a level that was reached after OGD/REOX in the presence of thapsigargin. If plasma membrane potential remains constant during ischemia, then NCX is predicted to function in the reverse mode when [Na+]i increases above 25 mM with [Ca2+]i at 777 nM. In fact, [Na+]i was elevated to ∼45 mM at 10 min of REOX and reached 80 mM at 60 min of REOX. However, when NHE1 activity was inhibited by HOE-642, NHE1+/+ astrocytes exhibited ∼50% less intracellular Na+ accumulation, which would tend to drive NCX in the forward mode. This thermodynamic analysis further supports the view that NHE1-mediated Na+ overload triggers the reverse operation of NCX after OGD/REOX.
Role for NHE1 in pHi Regulation in Cortical Astrocytes
We reported recently (13) that NHE1 plays a role in maintenance of resting pHi in mouse cortical astrocytes. In the absence of HCO3−, pHi recovery rate after NH4+/NH3-mediated acid loading was increased by ∼75% in NHE1+/+ astrocytes after 2-h OGD. Either inhibition of NHE1 activity with HOE-642 or genetic ablation of NHE1 abolished this elevated pHi regulation (13). We concluded that NHE1 activity is stimulated in astrocytes on REOX in response to in vitro transient ischemia. In this study, we further analyzed the kinetics of NHE1 in astrocytes after in vitro ischemia. First, we determined that the intrinsic buffering power was unaltered in NHE1+/+ astrocytes after OGD/REOX or in NHE1−/− astrocytes. At pHi 7.0, βi was 7.8 mM/pH unit in NHE1+/+ astrocytes under normoxic conditions. This βi value was unchanged in NHE1+/+ astrocytes after OGD/REOX. In addition, a similar βi value was found in NHE1−/− astrocytes under normoxic conditions. There were no significant changes in the slopes of the lines fitted to βi vs. pHi under these conditions. This finding suggests not only that the cytosolic intrinsic buffering capacity is unchanged in astrocytes after 2-h OGD but also that ablation of NHE1 does not trigger a compensatory effect on intrinsic buffering capacity. Therefore, the elevation of pHi recovery rate after OGD, which is sensitive to both [Na+]o and the NHE1 inhibitor HOE-642, is largely the result of an increase in NHE1 activity in NHE1+/+ astrocytes.
Our kinetic analysis revealed that when NHE1+/+ astrocytes were reoxygenated after 2-h OGD, the Vmax of Na+/H+ exchange was increased threefold and the Km for H+ was doubled. This implies that Na+/H+ exchange is more efficient in astrocytes after OGD. The causes of the significant elevation in Vmax are unknown. It could be due in part to ERK1/2-mediated phosphorylation of NHE1 or to increased expression of NHE1 in the surface of astrocytes via protein trafficking after OGD. A significant increase in NHE1 synthesis is unlikely considering the short term of OGD treatment.
Cell shrinkage results in stimulation of NHE1 activity in C6 glioma cells as a result of an alkaline shift in the pHi dependence of the exchanger without changing the Vmax (29). Serum-mediated growth signals also shift NHE1 toward the high-affinity form (15). In contrast, in the current study, the NHE1 was less sensitive to acidification after OGD. This may imply that altering Vmax is a more efficient way for astrocytes to accelerate H+ extrusion than an alkaline shift of Km. Two-hour OGD induced an intracellular acidosis in astrocytes (pHi of 6.8 vs. 7.10 of normoxic control; Ref. 13). Therefore, an increase in Vmax and a decrease in H+ affinity would ensure that the NHE1 system would not be saturated under acidic conditions after OGD.
Activation of ERK1/2 Signaling Pathways Is in Part Responsible for OGD/REOX-Mediated NHE1 Stimulation in Astrocytes
In addition to an NH2-terminal transmembrane ion translocation domain, NHE1 contains a COOH-terminal cytoplasmic regulatory domain that modulates transport activity most likely by altering the affinity of a H+ transport site in the transmembrane domain (28). A phosphorylation domain at the distal COOH terminus of NHE1 contains a number of serine residues that are phosphorylated by serine/threonine kinases acting downstream of distinct signaling pathways.
It has been reported that the serine/threonine kinase p90RSK, a downstream substrate of ERK1/2, directly phosphorylates NHE1 and activates exchanger activity in response to growth factors (24, 34). MEK-dependent pathways, including p90RSK and ERK1/2, stimulate phosphorylation and activation of NHE1 in ischemic myocardium (24). The ERK signaling cascade is emerging as an important regulator of neuronal and glial responses to oxidative stress-related injury (7, 38). However, it is unknown whether ERK activation triggers NHE1 stimulation in astrocytes after OGD. In the present study, we found that 1-h OGD led to activation of ERK1/2 phosphorylation and that this stimulation was sustained in astrocytes after 2-h OGD and 1-h REOX. In addition, inhibition of MEK activity with PD-98059 not only significantly reduced NHE1-mediated H+ efflux but also attenuated NHE1-mediated Na+ accumulation by ∼60%. Therefore, we concluded that OGD/REOX-induced stimulation of NHE1 activity is in part dependent on MEK-ERK phosphorylation in astrocytes. It remains to be established whether ERK1/2 activation leads to phosphorylation of p90RSK and whether this in turn directly regulates NHE1 activity in astrocytes after OGD/REOX.
A sustained intracellular acidosis (3 min, pHi 6.5) leads to stimulation of NHE1 activity in rat ventricular myocytes, which requires activation of ERK1/2 and p90RSK (9). We found that 2-h OGD caused a reduction in pHi from 7.09 to 6.80 (13). In the current study, we examined whether OGD/REOX-induced stimulation of NHE1 activity could be the result of OGD-mediated moderate and sustained acidosis. In cortical NHE1+/+ astrocytes, neither the Vmax nor the Km of NHE1 was altered after 3 min of acidosis at a pHi of 6.25. Thus sustained acidosis failed to trigger NHE1 activation in mouse cortical NHE1+/+ astrocytes, suggesting that it might be a cell type-specific phenomenon. Activation of MEK-ERK phosphorylation in astrocytes is unlikely to be the result of a sustained intracellular acidosis after OGD/REOX.
It is known that many signals can trigger activation of ERK pathways in response to oxidative stress (7). A 4-h period of glucose deprivation in mouse cortical astrocytes caused an approximately twofold increase in ERK1/2 phosphorylation (38). In the present study, 2-h OGD in astrocytes led to an ∼25% reduction in cellular ATP level (16). Thus an increase in intracellular Ca2+, reduction of ATP, glucose deprivation, and the generation of reactive oxygen species may subsequently stimulate MEK-ERK phosphorylation in astrocytes after OGD/REOX.
On the other hand, it was recently observed in vascular smooth muscle cells that NHE1 activity regulates ERK phosphorylation (25). We have performed additional experiments in NHE1−/− astrocytes to investigate whether ERK phosphorylation was altered in the absence of NHE1 function. One-hour OGD moderately stimulated ERK phosphorylation in NHE1−/− astrocytes (∼10%).
Stimulation of NHE1 Activity in Astrocytes Overloads Ca2+ Stores in ER and Mitochondria After OGD/REOX
In the present study, there was a twofold increase in releasable Ca2+ from ER Ca2+ stores when NHE1+/+ astrocytes underwent 2-h OGD and 1-h REOX. In addition, either inhibition of NHE1 activity or genetic ablation of NHE1 abolished the OGD/REOX-mediated Ca2+ loading of the ER. This finding was confirmed by two different measurements: thapsigargin-induced ER Ca2+ release in the presence of extracellular Ca2+ and bradykinin-mediated ER Ca2+ release in the absence of extracellular Ca2+. Moreover, we also evaluated the mitochondrial response to the OGD/REOX-mediated Ca2+ loading. Mitochondria exhibited a moderate level of increase in Ca2+ loading after OGD/REOX. The ER plays a central role in the control of cellular Ca2+ homeostasis. ER dysfunction (depletion of ER Ca2+ stores, inhibition of ER protein folding and processing) is linked to ischemic cell damage (26). The ER has close physical and functional contacts to mitochondria, and the coordination between Ca2+ signaling of the ER and mitochondria is facilitated by the strategic location of mitochondria at sites of ER Ca2+ release (6, 19). It has been reported that Ca2+ released during the ER stress response may promote mitochondrial Ca2+ overload, fragmentation, and apoptosis (6). Therefore, during severe ischemic episodes in astrocytes, the NHE1-mediated ER stress response may have a significant impact on mitochondrial function. This view is further supported by a recent finding that inhibition of NHE1 activity with HOE-642 significantly suppressed mitochondrial Ca2+ overload and loss of mitochondrial membrane potential induced by H2O2-mediated oxidative stress (35).
Stimulation of NHE1 Activity Triggers a Reverse-Mode Operation of NCX
The present study indicates that an increase in Ca2+ entry during OGD/REOX, which is buffered by ER and mitochondrial Ca2+ stores, is linked to NHE1 activity. We reported recently (16) that intracellular Na+ loading mediated by Na+-dependent Cl− transport triggers Ca2+ influx via a reverse-mode operation of NCX after OGD. In the present study, 3 μM KB-R7943, which blocked ∼90% of the NCX activity (16), abolished the OGD/REOX-induced enhancement in ER Ca2+ stores. A more selective inhibitor of NCX1, SEA-0400 (IC50 of 5 nM in astrocytes; Ref. 20), also blocked the OGD/REOX-mediated increase in Ca2+ transient. The data support our view that OGD/REOX initiates the reverse-mode operation of NCX in astrocytes and causes Ca2+ transient influx. We believe that this process is triggered by NHE1-mediated Na+ overload. Our thermodynamic analysis revealed that the reverse-mode operation of NCX would occur when [Na+]i increased above 25 mM. At the end of 1-h REOX following 2-h OGD, [Na+]i in NHE1+/+ astrocytes was increased to ∼80 mM as a result of accelerated Na+ influx via NHE1 and decreased Na+ extrusion via Na+-K+-ATPase. In contrast, NHE1−/− astrocytes exhibited a 50% reduction in Na+ overload. These data led us to conclude that stimulation of NHE1 activity overloads astrocytes with intracellular Na+ and subsequently triggers Ca2+ influx via the reverse mode of NCX.
In summary, we found that NHE1 activity was significantly stimulated in astrocytes in response to in vitro ischemia. This stimulation was in part dependent on activation of ERK1/2 signaling pathways. NHE1 activity led to intracellular Na+ overload and affected Ca2+ stores in ER and mitochondria. Therefore, overstimulation of NHE1 in astrocytes may contribute to ER stress and mitochondrial dysfunction under ischemic conditions.
This work was supported in part by National Institutes of Health Grants R01-NS-38118 and R01-NS-048216 (to D. Sun) and R01-HL-61974 (to G. E. Shull) and American Heart Association Established Investigator award grant N0540154 (to D. Sun).
The authors thank Dr. Tom Cook for consultation on the statistical analysis in this study.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2005 the American Physiological Society