Ubiquitously expressed volume-regulated anion channels (VRACs) are activated in response to cell swelling but may also show limited activity in nonswollen cells. VRACs are permeable to inorganic anions and small organic osmolytes, including the amino acids aspartate, glutamate, and taurine. Several recent reports have demonstrated that neurotransmitters or hormones, such as ATP and vasopressin, induce or strongly potentiate astrocytic whole cell Cl− currents and amino acid release, which are inhibited by VRAC blockers. In the present study, we explored the intracellular signaling mechanisms mediating the effects of ATP on d-[3H]aspartate release via the putative VRAC pathway in rat primary astrocyte cultures. Cells were exposed to moderate (5%) or substantial (30%) reductions in medium osmolarity. ATP strongly potentiated d-[3H]aspartate release in both moderately swollen and substantially swollen cells. These ATP effects were blocked (≥80% inhibition) by intracellular Ca2+ chelation with BAPTA-AM, calmodulin inhibitors, or a combination of the inhibitors of protein kinase C (PKC) and calmodulin-dependent kinase II (CaMK II). In contrast, control d-[3H]aspartate release activated by the substantial hyposmotic swelling showed little (≤25% inhibition) sensitivity to the same pharmacological agents. These data indicate that ATP regulates VRAC activity via two separate Ca2+-sensitive signaling cascades involving PKC and CaMK II and that cell swelling per se activates VRACs via a separate Ca2+/calmodulin-independent signaling mechanism. Ca2+-dependent organic osmolyte release via VRACs may contribute to the physiological functions of these channels in the brain, including astrocyte-to-neuron intercellular communication.
- volume-regulated anion channels
- protein kinase C
- calcium/calmodulin-dependent kinase II
- glutamate release
- neuron-glia communication
the overwhelming majority of mammalian cells respond to swelling by the activation of volume-regulated potassium and anion channels, resulting in a net loss of osmolytes, causing regulatory volume decrease (32, 46). Ubiquitously expressed volume-regulated anion channels (VRACs) are permeable to a variety of inorganic anions, small organic anions, and uncharged molecules, including the amino acids taurine, glutamate, and aspartate (30, 49, 51, 71). Although the molecular identity of VRACs has not been established, there is good evidence that, at least in some cell types, more than one type of anion channel may contribute to swelling-activated Cl− and organic osmolyte fluxes (2, 26, 31, 78). In addition to their role in cell volume homeostasis, VRACs are also thought to par-ticipate in a variety of other processes including cell proliferation, apoptosis, and mechanosensitivity in endothelial and muscle cells (13, 52).
In the brain, VRACs contribute to physiological and pathological amino acid release. In ischemia and other brain pathologies, uncontrolled cell swelling, primarily seen in astrocytes, causes massive efflux of excitatory amino acids that is sensitive to VRAC inhibitors (45, 57, 69), and such release has been implicated in ischemic brain damage (27, 29). Under physiological conditions, VRACs are functional in the supraoptic and paraventricular nuclei of the hypothalamus. In these brain areas, small changes in extracellular osmolarity tonically regulate taurine release via a VRAC permeability pathway in specialized subpopulations of astrocytes (9, 20). Extracellular taurine via glycine receptors modulates the electric activity of magnocellular neurons, which secrete key hormones of body water homeostasis, vasopressin and oxytocin (21, 22).
A role for VRACs in nonpathological glutamate and aspartate release has not yet been demonstrated. Although astrocytes show changes in their volume in response to neuronal stimulation in situ (1), the small degree of astrocytic swelling alone seems insufficient to activate VRACs to the level of functional significance. However, the findings of a few recent in vitro studies (42, 44, 62) suggest that neurotransmitters and neuromodulators that evoke intracellular Ca2+ increases may induce amino acid release via a putative VRAC pathway in nonswollen or moderately swollen cells. Such VRAC-mediated organic osmolyte release may contribute to Ca2+-dependent astrocyte-to-neuron signaling, which is currently thought to be mediated by glutamate (6, 18). In the present study, we used d-[3H]aspartate release as a measure of VRAC activity to explore the intracellular signaling mechanisms mediating VRAC activation and/or modulation by extracellular ATP in cells exposed to moderate (a 5% reduction in medium osmolarity) and substantial (a 30% reduction in medium osmolarity) hyposmotic gradients. We employed moderate cell swelling to test for VRAC activity under conditions resembling those to which astrocytes are exposed upon physiological stimulation. Substantial cell swelling, on the other hand, allowed us to examine the mechanisms of full VRAC activation as well as the effects of ATP on fully activated VRACs. Several reports (35, 36) have suggested that depending on the degree of swelling, cells may utilize different ion channels or signaling mechanisms of ion channel activation to regulate their volume. Portions of the data reported in this article were presented in preliminary form (43).
MATERIALS AND METHODS
Confluent primary astrocyte cultures were prepared from cortices of newborn Sprague-Dawley rat pups as previously described (15), with minor modifications. Pups were killed by decapitation according to the procedure conforming to the Public Health Service Policy on Humane Care and Use of Laboratory Animals and approved by the Albany Medical College Institutional Animal Care and Use Committee. The cerebral cortices were separated from meninges and basal ganglia, and tissue was dissociated using the neutral protease dispase and DNase I. Dissociated cells were seeded on poly-d-lysine-coated 18 × 18-mm glass coverslips (Carolina Biological Supply, Burlington, NC) and grown for 2–4 wk in minimal essential medium supplemented with 10% heat-inactivated horse serum, 50 U/ml penicillin, and 50 μg/ml streptomycin at 37°C in a humidified 5% CO2-95% air atmosphere. Culture medium was replaced twice weekly. After 10 days of cultivation, penicillin and streptomycin were removed from the culture medium. Immunocytochemistry showed that ≥95% of the cells stained positively for the astrocytic marker, glial fibrillary acid protein.
Excitatory amino acid release.
Release of excitatory amino acids was measured as previously described (42) using d-[3H]aspartate, a nonmetabolized analog of l-glutamate and l-aspartate, which is taken up by glutamate transporters in the same manner as l-glutamate. Astrocytes were loaded overnight with 4 μCi/ml d-[3H]aspartate (final concentration 270 nM) in 2.5 ml of serum-containing minimum essential medium in an atmosphere of 5% CO2-95% room air at 37°C. In double-labeling experiments, cells were loaded with 4 μCi/ml d-[3H]aspartate and 1 μCi/ml [14C]taurine (final concentration 9.3 μM). Before the start of the efflux measurements, the cells were washed free of extracellular isotope and serum-containing medium in HEPES-buffered solution. The basal HEPES-buffered medium contained (in mM) 122 NaCl, 3.3 KCl, 0.4 MgSO4, 1.3 CaCl2, 1.2 KH2PO4, 10 d-glucose, and 25 HEPES. pH was adjusted to 7.4 with NaOH (∼15 mM). The coverslips were inserted into a Lucite perfusion chamber that had a depression precisely cut in the bottom to accommodate the coverslip and a Teflon screw top, leaving a space above the cells of ∼100–150 μm in height. The cells were superfused at a constant flow rate of 1.2 ml/min in an incubator set at 37°C with HEPES-buffered media. Hyposmotic media were prepared by 5% dilution with H2O (a 5% decrease in medium osmolarity) or by 50 mM reduction of [NaCl] (a 30% decrease in medium osmolarity). The osmolarities of all buffers were checked using a freezing point osmometer (μOsmomette; Precision Systems, Natick, MA) and were measured to be 288 ± 2, 273.5 ± 2, and 197 ± 3 mosM for isosmotic, −5% hyposmotic, and −30% hyposmotic media, respectively. Superfusate fractions were collected at 1-min intervals. At the end of each experiment, the isotope remaining in the cells was extracted with a solution containing 2% sodium dodecyl sulfate plus 8 mM EDTA. Ecoscint scintillation cocktail (4 ml; National Diagnostics, Atlanta, GA) was added, and each fraction was counted for [3H] or [3H]/[14C] in a Packard Tri-Carb 1900TR liquid scintillation analyzer (Packard Instrument, Meriden, CT). Percent fractional isotope release for each time point was calculated by dividing the radioactivity released in each 1-min interval by the radioactivity left in the cells (the sum of all the radioactive counts in the remaining fractions up to the beginning of the fraction being measured, plus the radioactivity left in the cell digest) using a custom computer program, as previously described (41).
Data are presented as the means ± SE of 4–10 experiments performed on at least two different astrocyte preparations. Effects of all agonists and inhibitors of intracellular signaling were always compared with the controls performed on the same day and on the same culture preparation. In all cases we compared maximal amino acid release values measured during the second to third minute of exposure to hyposmotic medium. The data were analyzed by one-way ANOVA followed by the post hoc Newman-Keuls test when multiple comparisons were made. Significance levels of P < 0.05 were accepted as statistically different. Origin 7.5 (OriginLab, Northampton, MA) and Statistica 6.1 (StatSoft, Tulsa, OK) were used for statistical analysis.
d-[3H]aspartate (specific activity 16.2 Ci/mmol) and [14C]taurine (specific activity 108.5 mCi/mmol) were purchased from PerkinElmer Life Sciences (Boston, MA). Dispase (neutral protease dispase grade II) was obtained from Roche Applied Science (Indianapolis, IN). All cell culture reagents were obtained from GIBCO-Invitrogen (Grand Island, NY). Bisindolylmaleimide I (Gö-6850), ML-7, KN-62, KN-93, and Ro-32-0432 were obtained from Calbiochem (San Diego, CA). BAPTA-AM, chelerythrine chloride, trifluoperazine, and other chemicals, unless otherwise specified, were from Sigma (St. Louis, MO).
Characterization of ATP-stimulated d-[3H]aspartate release pathway.
In the present study, moderate cell swelling, induced by a 5% reduction in medium osmolarity, produced little effect on d-[3H]aspartate release in cultured astrocytes (∼20–40% increase above the basal levels; Fig. 1A). When 10 μM ATP was applied in combination with moderate cell swelling, it caused a transient 250–450% (variation between different cell cultures) increase in d-[3H]aspartate efflux rate compared with basal release levels (Fig. 1A). In our previous study (44), we found that the potentiation of amino acid release by 10 μM ATP was close to maximal and not statistically different from the effects of 100 μM and 1 mM ATP. Even in those cultures showing the highest ATP sensitivity, the maximal rate of the ATP-induced d-[3H]aspartate release in moderately swollen cells was much smaller than the release induced by the substantial hyposmotic swelling (compare the release values in Fig. 1B). In substantially swollen cells, 10 μM ATP additionally potentiated d-[3H]aspartate release approximately threefold (Fig. 1B; P < 0.001). In our previous studies (42, 44), we found that both swelling-activated d-[3H]aspartate release and ATP-induced d-[3H]aspartate release were sensitive to the nonselective VRAC blockers DIDS (200 μM) and 5-nitro-2-(3-phenylpropylamino)benzoic acid (NPPB; 100 μM) and to the open-pore VRAC blocker extracellular ATP (10 mM). In addition, 100 μM phloretin, which at this concentration largely discriminates VRACs vs. Ca2+-sensitive and CFTR Cl− channels (14), potently blocked the swelling-activated and the ATP-induced d-[3H]aspartate release (42). We have repeated our previous experiments with NPPB (42) in substantially swollen cells (30% reduction in medium osmolarity), in moderately swollen cells (5% reduction in medium osmolarity) in the presence of 10 μM ATP, and in substantially swollen cells in the presence of 10 μM ATP, and we found 69, 82, and 78% inhibition, respectively (n = 3, all P < 0.01 vs. respective controls; data not shown). To test the amino acid selectivity of the transport pathway involved in the ATP-induced amino acid release in moderately swollen cells, we measured simultaneously [14C]taurine and d-[3H]aspartate release (Fig. 1C). ATP increased the release of both isotope-labeled amino acids, with [14C]taurine showing larger stimulation compared with the basal release levels.
Ca2+ and calmodulin dependence of ATP-stimulated d-[3H]aspartate release.
P2Y receptor-mediated cellular responses to ATP are primarily due to the activation of phospholipase C (PLC) and subsequent intracellular Ca2+ release from inositol 1,4,5-trisphosphate (IP3)-sensitive intracellular Ca2+ stores (58). Therefore, we tested the Ca2+ sensitivity of the ATP-mediated d-[3H]aspartate release in both moderately and substantially swollen cells. Preloading astrocytes for 20 min with the intracellular Ca2+ chelator 10 μM BAPTA-AM strongly suppressed basal d-[3H]aspartate release and inhibited the ATP-induced amino acid release in moderately swollen astrocytes by ≥75% (Fig. 2A; P < 0.001). In substantially swollen cells, the ATP-induced increase in d-[3H]aspartate release was suppressed by ≥90% (Fig. 2B). In contrast, the hyposmotic medium-stimulated release itself showed a low sensitivity (∼25% inhibition; P = 0.052) to intracellular Ca2+ chelation (Fig. 2B), which was similar to our previously published data for [3H]taurine release (41).
We further checked whether the Ca2+-dependent activation and/or modulation of VRACs in cultured astrocytes involves a calmodulin-dependent step. The calmodulin antagonists 20 μM trifluoperazine (IC50 = 5.8 μM) and 50 μM chlorpromazine (IC50 = 17 μM) essentially eliminated the ATP-induced d-[3H]aspartate release in moderately swollen cells (Fig. 3A). Trifluoperazine (20 μM) also completely suppressed the ATP-induced increase in amino acid release in substantially swollen cells (Fig. 3B). As in the case of BAPTA-AM, trifluoperazine was ineffective in reducing control, swelling-induced d-[3H]aspartate release (Fig. 3B, P = 0.427). These data are very similar to our previously published findings for swelling-activated [3H]taurine release (41).
Effects of tyrosine kinase inhibitors on ATP-dependent d-[3H]aspartate release.
In cultured astrocytes, hyposmotic medium-induced cell swelling causes Ca2+-dependent activation of tyrosine kinase cascades (67), and tyrosine kinase signaling has been implicated in astrocytic VRAC activation or modulation (7, 10, 17). Therefore, we tested for tyrosine kinase involvement in the ATP-induced potentiation of amino acid release. In moderately swollen cells, both the receptor tyrosine kinase inhibitor tyrphostin A51 (20 μM; Fig. 4A) (P = 0.005) and an inhibitor of nonreceptor tyrosine kinases of the Src family, 4-amino-5-(4-chlorophenyl)-7-(t-butyl)pyrazolo[3,4-d]pyrimidine (PP2; 10 μM) (n = 5, P = 0.023; data not shown), reduced the ATP-induced d-[3H]aspartate efflux by 50–60%. In substantially swollen cells, tyrphostin A51 did not affect the ATP-induced increase in d-[3H]aspartate release (Fig. 4B, P = 0.551) or the release under control hyposmotic conditions (Fig. 4B, P = 0.144). Therefore, the ATP-induced organic osmolyte release seems differentially sensitive to tyrosine kinase inhibition depending on the degree of cell swelling.
MLCK does not contribute to the ATP effect on organic osmolyte release.
The myosin light chain kinase (MLCK) is a calmodulin-dependent enzyme that contributes to VRAC activation in vascular endothelial cells (50). This enzyme is expressed in astrocytes both in primary culture and in situ (12). Because we had found that the regulation of d-[3H]aspartate release by ATP is calmodulin dependent, we tested for the contribution of MLCK in mediating the ATP effects. Hyposmotic medium-stimulated d-[3H]aspartate efflux was surprisingly potentiated by the MLCK inhibitor ML-7 (10 μM) (2-fold potentiation; Fig. 5, P < 0.001). Furthermore, ML-7 was completely ineffective when cells were exposed to a combination of hyposmotic shock and ATP (Fig. 5, P = 0.565). Thus these data do not support a critical role for MLCK in ATP signaling.
Effects of CaMK II inhibitors on ATP- and swelling-induced organic osmolyte release.
Another major Ca2+/calmodulin-sensitive enzyme, one that is expressed in astrocytes and has been reported to regulate VRAC activity, is Ca2+/calmodulin-sensitive kinase II (CaMK II) (4, 5). To test for the CaMK II involvement in the ATP-induced d-[3H]aspartate release, we used the selective inhibitors of this enzyme, KN-62 (IC50 = 900 nM) and KN-93 (IC50 = 370 nM), at 5 and 10 μM. In substantially swollen cells, both inhibitors, when used at the higher 10 μM concentration, suppressed d-[3H]aspartate release by ∼40–50% either under control hyposmotic conditions or when hyposmotic stress was coapplied with ATP (Fig. 6, A and B). Although the absolute values of the ATP-induced d-[3H]aspartate release were reduced by the KN-93 and KN-62 compounds, relative values of VRAC activation by ATP remained similar in cells treated or not treated with CaMK II inhibitors. Therefore, CaMK II activation seems insufficient to explain the full ATP-induced potentiation of VRAC activity. We were unable to use higher concentrations of the KN compounds because of their limited solubility. In moderately swollen cells, 10 μM KN-93 inhibited the ATP-induced d-[3H]aspartate release by ∼50% (n = 5, P = 0.008; data not shown), similar to the data obtained in substantially swollen cells.
Effects of PKC inhibitors on ATP-dependent modulation of amino acid release.
Astrocytes express a variety of protein kinase C isoforms, several of which are activated by elevations in intracellular [Ca2+] ([Ca2+]i) (70). PKC was recently implicated in the positive modulation of volume-dependent organic osmolyte release by muscarinic receptors in neuroblastoma cells (34). We therefore tested for PKC involvement in the ATP-induced activation of d-[3H]aspartate release by using several specific PKC inhibitors. In moderately swollen cells, both 1 μM bisindolylmaleimide I and 10 μM chelerythrine potently inhibited the ATP-induced VRAC activation by 70–80% (Figs. 7A and 8A, respectively). In substantially swollen cells, bisindolylmaleimide or chelerythrine strongly reduced d-[3H]aspartate release in the ATP-treated cells but showed no effect (chelerythrine; P = 0.762) or little effect (bisindolylmaleimide I, 27% inhibition; P = 0.037) on the release induced by substantial hyposmotic swelling (Figs. 7B and 8B). Overall, these inhibitors strongly reduced (≥70% inhibition) the ATP-induced increment in d-[3H]aspartate release, suggesting that PKC is an important element of the ATP-induced VRAC modulation. In contrast, PKC is not obligatory or plays only a minor role in VRAC activation by hyposmotic swelling alone. Two other potent and selective PKC inhibitors, Ro-32-0432 (1 μM) and Gö-6983 (1 μM), were less effective against the ATP-induced d-[3H]aspartate release in substantially swollen cells (30–35% inhibition, n = 4, P < 0.05 for both inhibitors; data not shown) and showed no inhibition under control hyposmotic conditions (n = 3–4, P > 0.4 for both inhibitors; data not shown).
Additive action of PKC and CaMK II inhibitors on ATP-induced amino acid release.
Because the pharmacological inhibition of CaMK II or PKC produced strong but incomplete inhibition of the ATP effects on d-[3H]aspartate release, we tested for the additive actions of these two signaling enzymes. A combination of the most effective PKC inhibitor, bisindolylmaleimide I (1 μM), and the CaMK II inhibitor KN-93 (10 μM) completely blocked the ATP-induced increase in d-[3H]aspartate release from moderately swollen (Fig. 9A; P < 0.001) and substantially swollen (Fig. 9B; P < 0.001) astrocytes. In sharp contrast to the ATP-regulated component, control hyposmotic medium-induced amino acid release was only insignificantly affected by the combination of PKC and CaMK II blockers (Fig. 9B; 30% inhibition, P = 0.335).
In the present study, we explored intracellular signaling mechanisms involved in the regulation of cell volume-dependent astrocytic organic osmolyte release by extracellular ATP. It was recently found that in cultured astrocytes and several other cell types, receptor agonists stimulating the PLC/[Ca2+]i signaling pathway, such as ATP, bradykinin, and vasopressin, activate VRAC-like Cl− currents and organic osmolyte release even in the absence of cell swelling, but they do so to a much larger extent in swollen cells (8, 42, 62, 75, 76). Exocrine release of endogenous ATP was initially postulated to be a necessary step of VRAC activation (76). However, subsequent studies have shown that in the majority of cell types, endogenous ATP release is not necessary or sufficient for VRAC activation but, instead, positively modulates already active VRACs (19, 25, 42, 60, 75). We demonstrated in the present study that in cultured astrocytes, the ATP-induced, Ca2+-dependent regulation of organic osmolyte release via a VRAC-like pathway involves at least two Ca2+-dependent intracellular signaling cascades that incorporate PKC and CaMK II. This ATP-dependent mechanism of VRAC regulation is different from VRAC activation by hyposmotic cell swelling, which is largely independent of intracellular Ca2+ increases and calmodulin.
What transport pathway mediates the ATP-stimulated amino acid release?
This question of what specific transport pathways are responsible for ATP-induced amino acid release was initially addressed in our previous study (42), in which we found that ATP-induced d-[3H]aspartate release is inhibited by several structurally unrelated VRAC blockers and by small degrees of cell shrinkage. However, our present findings of substantial differences in the intracellular signaling mechanisms contributing to the regulation of d-[3H]aspartate release by hyposmotic cell swelling and by ATP naturally return us to the possibility that hyposmotic cell swelling and ATP activate different or multiple transport systems. Besides VRACs, which are activated under hyposmotic conditions and are permeable to excitatory amino acids (3, 23, 28, 63), cultured astrocytes express several other transport pathways potentially contributing to excitatory amino acid release. These pathways include excitatory amino acid transporters working in a reverse mode, the P2X7 ATP receptor channels, connexin hemichannels, and Ca2+-dependent glutamate release through an exocytotic mechanism (11, 54, 55, 65, 77).
Both ATP-induced aspartate release and swelling-activated aspartate release are potently inhibited by several VRAC blockers that we have tested, including 100 μM NPPB, 200 μM DIDS, 10 mM extracellular ATP, and 100 μM phloretin (42, 44). One previous study found that reversal of amino acid transporters is insensitive to NPPB and extracellular ATP (64). P2X7 and connexin hemichannels are fully active only at low extracellular Ca2+ and Mg2+ concentrations and show little or no sensitivity to typical VRAC blockers (11, 77). Therefore, they are also unlikely to be major contributors to hyposmotic and ATP-regulated d-aspartate release. The Ca2+-dependent vesicular glutamate release is less characterized in terms of its sensitivity to VRAC inhibitors. However, as shown by our data, the ATP-activated release pathway in moderately swollen cells is permeable to both d-[3H]aspartate and [14C]taurine. Because in glial cells taurine is known to be localized in the cytoplasm and released through a VRAC route (9, 24), this finding seems to exclude a substantial contribution of the vesicular release pathway. Furthermore, the ATP-induced d-aspartate release is strongly potentiated by cell swelling and completely suppressed by cell shrinkage (42).
On the basis of the foregoing studies, we concluded that under our experimental conditions, ATP-induced excitatory amino acid release in cultured astrocytes occurs predominantly via a VRAC-like channel, and we therefore use the term VRAC throughout this article. We cannot exclude the possibility, however, that other transport systems may provide a minor contribution to d-[3H]aspartate fluxes measured in our experiments or that more than one VRAC-like permeability pathway may exist in cultured astrocytes (26, 47).
Different roles for [Ca2+]i/calmodulin signaling in VRAC regulation by ATP and cell swelling.
Our data suggest that in cultured astrocytes, VRAC regulation by ATP requires increases in [Ca2+]i and calmodulin. This is in contrast to VRAC activation by hyposmotic cell swelling, which is essentially Ca2+ and calmodulin independent. In hyposmotically swollen astrocytes, d-[3H]aspartate release was only modestly (≤25%) inhibited by the calmodulin inhibitor trifluoperazine or by chelation of intracellular Ca2+ with BAPTA-AM. In contrast, the same pharmacological agents nearly completely suppressed the ATP-induced increases in d-[3H]aspartate release in both moderately and substantially swollen cells (see Figs. 2 and 3). These data are in line with previous findings of van der Wijk et al. (75) on the Ca2+-dependent modulation of swelling-activated 125I− fluxes in human intestine 407 cells. In cultured cerebellar astrocytes, calmodulin antagonists and BAPTA-AM have shown little effect on hyposmotic medium-induced [3H]taurine release but completely inhibited the positive modulation of such release by the calcium ionophore ionomycin (5, 48). A small component of swelling-activated organic osmolyte release, which is sensitive to Ca2+ chelators and calmodulin inhibitors, observed by us and others (8, 42, 75), may be due to autocrine ATP release and subsequent VRAC modulation.
In contrast to the Ca2+-independent VRAC activation by hyposmotic cell swelling observed in this study, Li et al. (33) found a complete dependency of swelling-induced Cl− and taurine currents on calmodulin and intracellular Ca2+ (33) and strong inhibition of Cl− currents by an intracellular application of anti-calmodulin antibodies (53). However, they used high concentrations of calmodulin inhibitors (100 μM trifluoperazine and 300 μM W-7), which we could not test in our experiments because of their pronounced nonspecific effects, and a combination of 1 mM extracellular EGTA and 20 mM intracellular BAPTA, which likely reduces [Ca2+]i below the permissive levels required for VRAC functioning (66, 73).
ATP-dependent VRAC regulation involves several Ca2+-dependent protein kinase cascades.
Several Ca2+ and calmodulin-sensitive signaling enzymes have been reported to contribute to VRAC activation and/or modulation, three of which (MLCK, CaMK II, and PKC) are highly expressed in astrocytes (4, 12, 70). MLCK contributes to the swelling-induced VRAC activation in endothelial cells (50). However, in cultured astrocytes (data from the present work) and in NIH/3T3 mouse fibroblasts (56), the MLCK inhibitor ML-7 potentiated (rather than inhibited) organic osmolyte release in hyposmotically swollen cells and had no effect on the release in substantially swollen cells treated with ATP. Therefore, a positive modulation of VRAC by the MLCK appears to be cell-type specific and not relevant to the ATP-stimulated organic osmolyte release in astrocytes.
The second candidate enzyme, CaMK II, likely modulates VRAC activity in astrocytes. In our experiments, the CaMK II inhibitors KN-62 (10 μM) and KN-93 (10 μM) attenuated the VRAC-mediated d-[3H]aspartate release by ∼40–50% under all conditions tested. Although the absolute values of the ATP-induced d-[3H]aspartate release were reduced, the relative degree of VRAC stimulation by ATP remained similar in the cells treated or untreated with the CaMK II inhibitors. Therefore, CaMK II does not appear to be crucial for the ATP-dependent VRAC modulation. In contrast to our data, Cardin et al. (5) found that in cultured cerebellar astrocytes, 10 μM KN-93 completely inhibits the Ca2+-dependent upregulation of [3H]taurine release by ionomycin but does not affect control hyposmotic medium-stimulated amino acid release (5), pointing to a pivotal role for the CaMK II in VRAC modulation in ionomycin-treated cells.
Two PKC inhibitors, chelerythrine and bisindolylmaleimide I, potently suppressed the ATP-induced d-[3H]aspartate release in our experiments while minimally affecting d-[3H]aspartate release induced by substantial hyposmotic swelling. Thus PKC appears to be a critical element of the ATP-induced VRAC regulation in astrocytes but does not contribute to VRAC activation by hyposmotic swelling. However, PKC involvement is not consistent with our data presented in Fig. 3, showing strong calmodulin dependence of the ATP effect, because Ca2+-sensitive members of the PKC family are directly regulated by [Ca2+]i and do not require calmodulin activation. This contradiction may be explained by a calmodulin-dependent regulation of PLC, an upstream enzyme of the PKC signaling cascade (37, 68). Similar Ca2+- and PKC-dependent modulation of anion channel-mediated organic osmolyte efflux has been found in neuroblastoma cells (34). PKC contributes to VRAC activation or positive modulation in some (34, 39, 59, 61), but not all (40, 79), cell types.
Although PKC inhibitors strongly attenuated the ATP-induced d-[3H]aspartate release, inhibition was incomplete. We therefore tested for the additive action of PKC and CaMK II. A combination of the inhibitors of both enzymes caused complete suppression of the ATP-induced VRAC regulation in moderately and substantially swollen cells. This finding implies that both PKC and CaMK II contribute to the ATP effects, with PKC playing the dominant role. Cooperative action of PKC and CaMK II has been found to be critical for the regulation of delayed rectifier potassium channels by angiotensin II in CATH.a cells (72), activation of phospholipase D by muscarinic acetylcholine receptors in a heterologous expression system (38), and ATP-dependent activation of ERK-1/2 in smooth muscle cells (16).
Tyrosine kinases differently regulate VRAC activity depending on degree of cell swelling.
Both the PKC and the CaMK II inhibitors, when applied in combination or alone, blocked the ATP-induced VRAC activation in astrocytes independently of the degree of cell swelling. In contrast, tyrosine kinase inhibitors showed strong inhibition of the ATP-induced d-[3H]aspartate release in moderately swollen cells but were ineffective in substantially swollen cells with or without ATP. Thus tyrosine kinase actions are limited to smaller degrees of cell swelling and are not essential for VRAC activation in substantially swollen cells. A similar trend for stronger inhibition of the volume-dependent organic osmolyte fluxes by tyrosine kinase blockers in “less swollen” compared with “more swollen” cells has been reported in epithelial cells, isolated brain supraoptic nuclei, and cultured cerebellar astrocytes (5, 10, 74). One suggested explanation of this phenomenon is that tyrosine kinases modulate volume sensitivity of the hypothetical volume sensor rather than directly regulate the VRAC function (10). “Saturation” of the volume signal in substantially swollen cells may override the modulatory contribution of tyrosine kinases (74). Because combined inhibition of PKC and CaMK II essentially abolishes the ATP-induced amino acid release in moderately swollen cells, tyrosine kinase signaling may work in the same signaling cascade, upstream or downstream of PKC/CaMK II.
In summary, on the basis of the present data and our previous findings (42), we suggest that extracellular ATP regulates organic osmolyte release via a VRAC-like transport pathway through activation of P2Y receptors and at least two Ca2+/calmodulin-dependent intracellular signaling cascades that incorporate PKC and Ca2+/calmodulin-dependent kinase II (Fig. 10). In contrast, substantial hyposmotic cell swelling activates VRACs through a separate Ca2+/calmodulin-independent signaling cascade that was not identified in this study. The ATP-dependent positive modulation of VRACs likely accelerates the regulatory volume decrease process in substantially swollen cells and may induce functionally significant VRAC activity in nonswollen and moderately swollen cells. In the brain, the ATP-induced release of excitatory amino acids via a VRAC-like permeability pathway from nonswollen or moderately swollen astrocytes may contribute to intercellular glutamate signaling.
This work was supported in part by National Institutes of Health Grants R01 NS-35205 (to H. K. Kimelberg) and F05 TW-05329 (to A. A. Mongin).
We thank Renee E. Haskew-Layton for critical reading of and helpful comments on the manuscript.
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