We examined the effects of dissolved nitric oxide (NO) gas on cytoplasmic calcium levels ([Ca2+]i) in C6 glioma cells under anoxic conditions. The maximum elevation (27 ± 3 nM) of [Ca2+]i was reached at 10 μM NO. A second application of NO was ineffective if the first was >0.5 μM. The NO donor diethylamine/NO mimicked the effects of NO. Acute exposure of the cells to low calcium levels was without effect on the NO-evoked response. Thapsigargin (TG) increased [Ca2+]iand was less effective if cells were pretreated with NO. Hemoglobin inhibited the effects of NO at a molar ratio of 10:1. 8-Bromo-cGMP was without effect on the NO-evoked response. If cells were pretreated with TG or exposed chronically to nominal amounts of calcium, NO decreased [Ca2+]i. The results suggest that C6 glioma cells have two receptors for NO. One receptor (NOA) elevates [Ca2+]i and resides on the endoplasmic reticulum (ER). The other receptor (NOB) decreases [Ca2+]i and resides on the plasmalemma or the ER. The latter receptor dominates when the level of calcium within intracellular stores is diminished.
nitric oxide (NO) modulates an extraordinary array of cellular characteristics and interactions such as endothelial cell tension, astrocyte neurotransmitter uptake, platelet aggregation, and neuronal communication via neurotransmitters. NO commonly exerts effects through cGMP: the reduction of glutamate uptake in glial cells (39), the reduction of the effectiveness of glutamate in retinal horizontal cell receptors (20), inhibition of platelet aggregation (22), vasorelaxation (36), and generation of calcium waves in astrocytes (38) all involve cGMP. NO also elevates cGMP levels in cerebellar astrocytes (26).
Recently, NO has been found to have two actions on cytoplasmic calcium levels. NO releases the cation from internal stores, elevating cytoplasmic calcium levels in human platelets (14, 33), endothelial cells (35), and glia (38). Paradoxically, NO also promotes the refilling of these stores (and lowers cytoplasmic calcium levels) following calcium store depletion in human platelets (33). By contrast, these actions of NO on internal calcium stores appear to be independent of cGMP.
Our studies of NO began during an investigation of a (sporadic) effect of NO on cell volume of C6 glioma cells. Noting that cell volume, cytosolic calcium levels, and NO may be linked, we decided to characterize the effect of NO on cytoplasmic calcium levels in glioma cells and to examine the role of cGMP. Although there are a plethora of studies of calcium homeostasis in glial cells (34), there appear to be only two reports of NO elevating glial intracellular calcium levels in astrocytes derived from mouse (29) and rat (38). Using C6 glioma cells in suspension, we extend their findings and those of Trepakova et al. (33) and suggest that two receptors for NO exist and possibly act in yin-yang fashion: one receptor functions to elevate cytoplasmic calcium, while the other decreases cytosolic calcium levels. The latter NO receptor becomes active following Ca2+ store depletion.
C6 glioma cells were grown as previously reported (4,5). Confluent cells were dissociated with trypsin (0.5 mg/ml, 25°C) in PBS, washed three times, and resuspended at final cell density of 0.75 × 106 cells/ml in a saline solution containing (in mM) 120 NaCl, 5 KCl, 1 CaCl2, 1 MgCl2, 10 glucose, and 20 HEPES, pH 7.4. All solutions were made in column-distilled water. Cells were loaded with fura 2-AM (Sigma Chemical, St. Louis, MO, or Molecular Probes, Eugene, OR) with Pluronic F-127 acid (0.05%; Molecular Probes) for 60 min in the dark at 25°C. After loading, cells were washed twice with saline solution and resuspended in oxygen-free saline. The stock cell suspension was kept at room temperature (∼23°C) in a N2 atmosphere.
Fluorescence measurements were made using an SLM 8000C spectrofluorometer (SLM Instruments, Urbana, IL). Two milliliters of the stock cell suspension were placed in a quartz cuvette maintained at 37°C with constant stirring. The chamber enclosing the quartz cuvette was continuously gassed with N2. After a 5-min equilibration, the cell suspension was repeatedly illuminated at wavelengths of 340, 380, and 363 nm. Four emission readings at 510 nm were averaged at each excitation wavelength. One cycle of excitation wavelengths (340 → 380 → 363 → 340 nm), including averaging, took ∼2.5 s. Slit widths were 4 nm for excitation and emission. The gain and high voltage were set to 1 and 1,000 V, respectively. A stock solution of NO was made by adding oxygen-free column-distilled water (i.e., previously bubbled with oxygen-free N2 gas) to a tonometer (Tech Glass Instruments, Marion, NY) filled with NO gas (a generous gift from Dr. Robert W. Noble). The gas-filled tonometer was stored at 23°C and replenished with NO every 8–12 wk. Stock NO solutions were stored in a gas-tight syringe with the needle inserted in a rubber stopper. The syringe solution was replenished from the tonometer every 2 days. A gas-tight syringe was used to draw aliquots from the stock syringe for use in the experiments. The following equation was used to calculate the concentration of NO in stock solutions Equation 1where Q = T + 273.15, A = −62.8086,B = 82.3420, and C = 22.8155. T is the temperature (in °C) (16). The calculated concentration of stock NO was 2.0 mM at 23°C. All additions of NO to the cell suspension (total volume 2.0 ml) were made with 20 μl of oxygen-free water as the vehicle. Less concentrated stock solutions of NO were made by diluting the concentrated stock solution in oxygen-free column-distilled water.
The concentration of NO in solutions decreases with time in solution (9). To determine the lifetime of NO, we measured directly the concentration of NO under the conditions of our experiment (37°C, N2 atmosphere, saline solution, 2.0-ml volume, quartz cuvette). We used an ISO-NO mark II NO meter (World Precision Instruments, Sarasota, FL).
To determine whether NO affected the dissociation constant (K d) of fura 2, we made calcium solutions buffered with EGTA containing 1 μM fura 2 (the salt form), 100 mM KCl, and 10 mM HEPES, pH 7.4 at 37°C. In each experiment, we made seven solutions with calcium levels ranging from 1 to 0.300 μM.K d values were computed from the slope of the line (solving Eq. 2 for K d). Calcium concentrations in the EGTA solutions were computed from the absolute stability constants (log10) (20°C, 0.1 M):K 1 = 9.47 ± 0.02,K 2 = 8.85 ± 0.01,K 3 = 2.66 ± 0.02,K 4 = 2, K Ca = 10.97 ± 0.10, K CaHEGTA = 5.30 (12, 19) after correcting for pH, temperature, and ionic strength (18). The errors in the stability constants were carried through the calculations to estimate the maximum uncertainty in the calculated calcium concentration. The uncertainty in pH was assumed to be 0.05 pH units. The dielectric constant was computed using a method described by Bers et al. (1). An in-house computer program, written in C (a computer language), facilitated the computations. The maximum uncertainty in computed calcium concentrations was ∼30%.
Cell calcium concentrations were calculated by using aK d of 0.285 μM (37°C) (10) of fura 2 fluorescence following standard methods (11) Equation 2R is the ratio of two fluorescence emission intensities excited at 340 and 380 nm. EGTA380 and Dig380 are the fluorescence emission intensities of the fura 2-loaded cell suspension excited at 380 nm in the presence of EGTA and digitonin, respectively. Rmax is the ratio of emission intensities measured at excitation wavelengths of 340 and 380 nm in the presence of 2 μM digitonin and was computed from an average of 20 points (see Fig.1 A, points between the 3rd and 4th arrows). After digitonin was added, 10 mM EGTA was added to the cell suspension and Rmin was similarly computed. Values for Rmax, Rmin, Dig380, and EGTA380 were 5.37 ± 1.10, 0.882 ± 0.116, 0.0342 ± 0.0144, and 0.081 ± 0.037, respectively (means ± SD, n= 116 in 37 experiments). There were no significant differences in these parameters between cells loaded in the presence (1 mM, 29 experiments) or nominal absence (∼1 μM, 8 experiments) of calcium. NO (20 μM) slightly reduced the K d of fura 2 by 0.005 ± 0.004 μM (mean ± SD, 5 experiments; not shown). The baseline calcium levels between aliquots from the same cell suspension varied by ∼0.020 μM. Therefore, we did not correct the data for the small effect of NO on the K d of fura 2.
A nonlinear curve-fitting procedure (2) was used to estimate parameters in the Hill equation: EC50, the change in maximum calcium level (Δ[Ca2+]max), and the apparent number of NO binding sites (n app). Observed means and variances (not standard errors) were used for input. Briefly, parameter space was searched to find reasonable starting values using a Monte Carlo method, after parameters were limited to reasonable boundaries: EC50, 0.1 to 100 μM; [Ca2+]max, 5 to 50 nM;n app, 1 to 5. Χ2 was minimized with a gradient-search method, and the search was stopped when χ2 stopped changing by ∼1%. To estimate errors in the parameters, we generated many synthetic data sets by randomly choosing new mean values for each data point, within the corresponding experimental variance. The Χ2 fitting method was repeated for each data set. The estimated errors in the parameters, which are slightly larger than those corresponding errors for the individual fits, are reported. A description of the synthetic data method of error estimation has been published (2). The fitting procedure was written in Gauss (Gauss 386i VM; Aptech Systems, Maple Valley, WA).
Hemoglobin (Hb; a gift from Dr. Robert W. Noble) was prepared from the erythrocytes of a human donor, as described previously (7,25). This Hb was dissolved in distilled H2O (dH2O) and stored in liquid N2. Before being used, an aliquot was diluted in oxygen-free dH2O. Lyophilized hemoglobin was obtained from the Sigma Chemical (catalog no. H-7379) and stored at 0–5°C. The lyophilized Hb was dissolved in oxygen-free dH2O. The air space over the Hb solutions was gassed exhaustively (hours) with N2. The concentration of Hb given refers to the number of heme groups (i.e., high-affinity NO binding sites), not Hb molecules per se.
Figure 1 A shows fluorescence emission intensities (measured at 510 nm) from a fura 2-loaded cell suspension excited at three different wavelengths: 340, 363, and 380 nm. Addition of 20 μM NO evoked an increase in the emission excited at 340 nm and a concomitant decrease at 380 nm. The fluorescence excited at 363 nm was unaffected by NO (Fig.1 B), suggesting that NO does not affect the fura 2 dye directly. The K d of fura 2 was not affected significantly by NO (see methods). Figure 1 Cshows that the concentration of NO decays with time in saline solution (without cells) under conditions otherwise identical to those in Fig.1 A. The decay in NO concentration was described well by a single exponential plus a constant (solid line) and reached one-half of its initial value (20 μM) in 2.47 min (Fig. 1 C, open arrow). Figure 1 D shows the computed cell calcium levels from the traces shown in Fig. 1 B. [Ca2+]i just before the addition of NO was 113 ± 77 nM (mean ± SD, n = 98 determinations in 37 experiments), in agreement with previous findings (34). Cell calcium levels peaked within 14 s after addition of NO, during which the NO concentration decreased, at most, by 6% (Fig. 1 D, open circles). The decay in calcium signal (Fig. 1 D) parallels the decay of NO concentration in solution (Fig. 1 C). This finding suggests that at least some of signal loss reflects the decaying concentration of NO in solution. In addition, the sustained nature of the cell's response is a reflection, at least in part, of the remaining NO in solution. Addition of the vehicle (H2O) used to deliver NO to the cell suspension was without effect in internal calcium levels (5 determinations in 4 experiments; not shown). A second addition of 20 μM NO was similarly without effect (15 determinations in 10 experiments). The ineffectiveness of the second application of NO may reflect a diminished (or empty) pool of calcium or the inactive (desensitized) state of a NO receptor.
NO was without significant effect on cell viability, as tested by exclusion of trypan blue (%viability: controls, 96 ± 1%; +20 μM NO, 93 ± 2%; means ± SD, n = 7 determinations in 3 experiments).
We examined the ability of three classes of NO donors to effect a response. The nitric oxide donor diethylamine/NO (DEA/NO), a NONOate that spontaneously releases NO (9), evoked increases in cell calcium levels (Fig. 2, open circles) identical to equimolar gaseous NO (Fig. 2, closed circles). Subsequent addition of DEA/NO or NO was without effect, confirming the phenomenon shown in Fig. 1. By contrast, other NO donors tested were without effect on calcium levels in cells otherwise responsive to gaseous NO: sodium nitroprusside (SNP, an iron nitrosyl; 0.1 and 1.0 mM, 8 and 6 determinations, respectively), isosorbide dinitrate (ISDN, an organic nitrate; 0.1 mM, 2 determinations),S-nitroso-N-acetylpenicillamine (SNAP, anS-nitrosylthiol; 0.1 mM, 3 determinations), or 3-morpholinosydnonimine (SIN-1, an oxime; 10 μM, 1 experiment). The ineffectiveness of SNP confirms a previous study of this donor in C6 glioma cells (29). These results may suggest that the requirements (9) for release of NO from some donors may not be met under our experimental conditions (seediscussion).
We examined the possibility that inducible nitric oxide synthase (iNOS) levels in our cells may be elevated, perhaps due to making the cell suspension. N ω-nitro-l-arginine (l-NNA; 0.1 mM), a reversible blocker of iNOS, was ineffective in either the presence or absence of l-arginine (0.1 mM), the substrate for iNOS (12 determinations, 3 experiments; not shown). The ineffectiveness of l-NNA andl-arginine suggests that the trypsin dissociation procedure does not elevate iNOS activity, and that iNOS activity is low, under our experimental conditions.
Although NO elevates cytoplasmic calcium levels in endothelial cells via a cGMP-independent pathway (35), NO evokes increases in cGMP levels in cerebellar astrocytes in primary cell culture (26). We therefore examined whether the NO-evoked elevation of [Ca2+]i was mediated by cGMP. 8-Bromo-cGMP (8-Br-cGMP) did not affect internal calcium levels or the NO-evoked response (2 experiments, 6 determinations; not shown). These results suggest that NO does not elevate internal calcium levels through a cGMP-dependent pathway.
In the presence of oxygen, NO forms many compounds (3). Although oxygen levels should be minimal under the anoxic conditions used in our experiments, we decided to determine whether by-products formed in the presence of atmospheric oxygen cause an elevation in [Ca2+]i. Adding NO to cell suspensions equilibrated with atmospheric oxygen (∼200 μM dissolved O2) or preexposing the NO stock solution to the atmosphere was without effect on internal calcium levels (not shown). Furthermore, the ability to exclude trypan blue was not affected by adding 20 μM NO in cells exposed to atmospheric oxygen (1 experiment, 6 determinations). These results suggest that the evoked elevation in internal calcium levels is not due to species generated by NO and superoxide and that NO-evoked responses are unrelated to acute cell toxicity.
To confirm that NO is required for the calcium response, we examined the ability of Hb to block the response. Hb avidly binds NO (21). However, Hb solutions (0.5–10 μM) made from lyophilized human Hb evoked a substantial increase in [Ca2+]i (7 experiments; not shown), precluding its use in this experiment. A similar result has been noted with lyophilized bovine Hb (see Ref. 15,methods). By contrast, human Hb from another source (not subjected to lyophilization) was without effect on cell calcium levels (6 experiments). We used this Hb in the following experiment. See our methods and discussion for further details.
Addition of 1 μM NO evoked an increase and a concomitant decrease in the fluorescence emission excited at 340 and 380 nm, respectively (Fig.3, A and C). Subsequent addition of 20 μM NO was without effect (Fig. 3,A, C, and E). Addition of 10 μM Hb evoked a decrease in fluorescence emission excited at the three wavelengths (Fig. 3, B and D) but did not affect calcium concentration (Fig. 3 F). The solution in the cuvette had a brown tint, typical of Hb solutions. In the presence of 10 μM Hb, addition of 1 μM NO was without effect, but subsequent addition of twofold excess NO (relative to Hb) evoked a substantial increase in cell calcium (Fig. 3, B, D, and F). This result suggests that Hb inhibition is not attributable to interference with the fura 2 dye or block of the NO receptor by Hb. It may be argued the brown-tinted solution reduced the sensitivity (compare Fig. 3, A and B) and that a successful response to 1 μM NO was missed in Fig. 3, B, D, and F. If this were the case, then subsequent addition of 20 μM NO should have been without effect (see Fig. 3, A,C, and E; also see Fig. 7). The results suggest that 10 μM Hb rapidly bound all the 1 μM NO, preventing its use at the NO receptor, and that NO is responsible for the elevation of cytoplasmic calcium.
Addition of 10 μM Hb was without effect on the fluorescence emission excited at the calcium-insensitive wavelength (363 nm) of fura 2 (Fig.3 D). By contrast, a subsequent addition of 20 μM NO evoked a decrease in fluorescence emission excited at 363 nm (most noticeable in Fig. 3 D, circles). This phenomenon, never observed in the absence of Hb (Fig. 1), suggests quenching of the fura 2 dye (seediscussion).
The minimum effective concentration of NO was 0.2 μM (Fig.4). The effect of NO reached maximum at ∼10 μM. NO (100 μM) caused a less-than-maximal elevation in [Ca2+]i, perhaps due to desensitization (see Figs. 1 C, 2, and 3). A fit of the Hill equation to the NO dose-response to all data points results in an EC50 of 0.7 ± 0.3 μM, an apparent Hill coefficient of 2 ± 0.4 (1.4, rounded up; see Ref. 27), and a [Ca2+]max of 27 ± 3 nM. The results suggest that two NO molecules are involved in the elevation of [Ca2+]i. The minimum effective concentration of NO (0.2 μM) and the EC50 are similar to the corresponding parameters of the gas's relaxation of rings of cerebral arteries (8).
To determine whether NO causes calcium entry from the external solution, we added 2 or 10 mM EGTA to the bath containing 1 mM CaCl2 ([Ca2+]o = 50 ± 20 nM or 5 ± 2 nM, respectively). Adding 10 mM EGTA evoked a decrease in cytoplasmic free calcium concentration (Fig.5 A), as expected (6,17). Subsequent addition of NO evoked an increase in [Ca2+]i of ∼25 ± 8 nM (Fig.5 A), quantitatively indistinguishable from the size of the response in 1 mM external calcium (27 ± 6 nM) (Fig. 4). Identical results were obtained when 2 mM EGTA was added to the bath (not shown). These results suggest that external calcium plays no role in the NO response.
To approach the question of the source of calcium from a different direction, we loaded cells with fura 2 in the nominal absence ([Ca2+]o∼ 1 μM) of calcium (cells exposed to saline without added CaCl2 ∼1.2 h before study). From the Nernst equation, we reasoned that the driving force for calcium entry from the external medium would be reduced by approximately fourfold (assuming the membrane potential remains constant). If calcium was entering the cell from the medium, there should be a corresponding diminution of the response. By contrast to the EGTA results, NO evoked a significant decrease in [Ca2+]i (−13 ± 5 nM, mean ± SE, 10 determinations in 4 experiments) (Fig.5 B). Subsequent addition of 1 mM CaCl2 to the bath caused a substantial increase in internal Ca2+(60 ± 7 nM, 10 determinations in 4 experiments), as expected (17). A further addition of 20 μM NO evoked a very slight transient decrease in internal calcium levels (−1 ± 3 nM, 5 determinations in 3 experiments) (Fig. 5 B). Together, these results suggest that the bulk of the calcium is from internal stores of calcium. These stores likely become depleted after chronic exposure to low calcium concentrations (∼1 μM for ∼1 h) but not when acutely exposed to low (∼5–50 nM for ∼3 min) external calcium levels. Internal NO-sensitive stores may not be filled readily (within the time frame of the experiment) after addition of CaCl2 to the medium. The results suggest that a second NO receptor dominates when the calcium level within the stores is reduced or depleted. Under these conditions, this second NO receptor removes calcium from the cytoplasm.
To investigate internal sources of calcium further, we studied the effect of thapsigargin (TG), an inhibitor of microsomal Ca2+-ATPase (6, 32). This pump translocates calcium from the cytoplasm into the endoplasmic reticulum (ER). TG depletes this store as calcium ultimately leaks into the cytoplasm following inhibition of this calcium pump (6). If NO and TG released calcium from the same internal store, prior depletion of this store should reduce the effectiveness of subsequent addition of TG or NO. Addition of 20 μM NO evoked an increase in [Ca2+]i, and a subsequent addition of 100 nM TG evoked a further increase (Fig.6 A). By contrast, addition of TG evoked a much larger increase in [Ca2+]i, which eventually reached a steady-state level. Subsequent addition of NO evoked a decrease in calcium concentration (Fig. 6 B). The effectiveness of TG is reduced by approximately one-half if NO precedes TG. These results suggest that NO and TG release calcium from a common pool in the ER and confirms that a second NO receptor exists. The effects of this second NO receptor surface when the ER is depleted of calcium (by TG or by chronic exposure of the cells to low calcium).
We examined the ineffectiveness of the second application of NO by changing the concentration of NO in the first application and testing the response to a subsequent addition of a fixed concentration of NO (20 μM; Fig. 7 A, test application). Diminution of the test application was minimally detectable at 0.10 μM and maximum at >0.50 μM NO (Fig.7 B). These results suggest that the receptor maximally desensitizes to NO at concentrations ∼20-fold lower than those required for maximum cytoplasmic calcium elevation (10 μM NO; Fig. 4) and rules out the possibility of diminished pools of internal calcium as a cause for the ineffectiveness. To confirm that internal pools of calcium are still capable of releasing calcium when the response to NO is eliminated, we examined the effect of prior application of 1.0 or 20 μM NO on the response evoked by 0.1 μM TG. After treatment with 1.0 μM NO (which eliminates the response to a subsequent addition of 20 μM NO; see Fig. 7 B), TG still evokes a substantial response (Fig. 7 C, closed circles). Increasing the concentration of NO to 20 μM reduces the effectiveness of TG by ∼40% (Fig. 7 C, compare open and closed circles). These results confirm that TG-sensitive pools of calcium still exist when a second application of NO becomes ineffective and that NO releases calcium from a TG-sensitive pool (Fig. 6). These results also confirm that the less-than-maximal response to 100 μM NO in Fig. 4 is due to desensitization of the receptor and support the idea that desensitization of the first receptor may be helpful (but probably not required) for operation of the second NO receptor.
The experiments so far were carried out in anaerobic conditions. To determine whether the first NO receptor can be activated under aerobic conditions, we compared the effects of gaseous NO and the NO donor DEA/NO on cells suspended in medium equilibrated with air (∼200 μM oxygen). Under these conditions, 20 μM NO and 20 μM DEA/NO both evoked increases in cytosolic calcium levels, although the effect of gaseous NO was less then that of DEA/NO (Fig.8 A). We suggest that gaseous NO reacts with oxygen, limiting its effectiveness. Although reactive species can form under these conditions (3), these species, if formed, were without effect on cell calcium levels. As noted earlier, stock solutions of NO exposed to the atmosphere were also without effect. When bound to DEA, NO is probably protected from reacting with oxygen, and DEA/NO releases NO close to the first NO receptor, providing a continuous source of agonist, hence the larger effect of DEA/NO under aerobic conditions.
We examined the ability of DEA/NO to elicit a response from the second NO receptor under aerobic conditions. DEA/NO (20 μM) evokes a decrease in cytosolic calcium levels (Fig. 8 B). The response evoked by DEA/NO under aerobic conditions is identical to that evoked by gaseous NO under anaerobic conditions (Fig. 8 B, compare closed and open circles). Application of an identical dose of DEA/NO or NO is without effect, suggesting either desensitization of the second NO receptor or depletion of the source of calcium. The restorative effects of bath-applied 1 mM CaCl2 is similarly independent of the oxygen tension (Fig. 8 B, 3rd arrow). These results demonstrate that the functional expression of the first and second NO receptor is independent of the oxygen tension in the suspending medium.
Model of NO regulatory effects.
A mechanism has been proposed for NO regulation of calcium levels in human platelets (33), and our model is similar. Figure9 shows a model of the effects of NO on cell calcium. The solid and dashed arrows indicate mechanisms by which NO may increase or reduce cytoplasmic calcium levels, respectively.
NO acts on a receptor (NOA) located on the ER, releasing calcium from this store (fig. 9, solid arrow). This receptor desensitizes (Figs. 1 C, 2, 3 E, and 7), because the source of calcium is still available when the receptor cannot be activated by a repeated dose of NO (Fig. 8).
Two lines of evidence suggest that NO activates a second receptor, NOB, which moves calcium to a space not occupied by fura 2:1) the NO-evoked decrease in [Ca2+]i for cells chronically exposed to low calcium levels (Fig. 5 B) or 2) cells pretreated with TG (Fig. 6). The profound desensitization of NOA (Figs. 1 C, 2, 3 E, and 7) may help NOB refill the NOA-sensitive calcium stores, pump calcium to the external medium, or pump calcium to the nucleus. We rule out a calcium pump coupled to the electron transport chain as a potential site for NOB because the receptor can be activated under anaerobic conditions, and calcium transport coupled in this manner should be minimal. Candidate molecules for the NOB receptor include the Na+/Ca2+exchanger, the Ca2+-ATPase, and the capacitative calcium entry system (CCE). In common with many other cells, the CCE in C6 glioma cells allows refilling of ER stores from the bath following Ca2+ depletion (6, 17, 29). CCE does not appear to be a candidate for NOA, because NO and TG share common internal calcium stores (Fig. 4), and buffering external calcium levels is without effect (Fig. 3 A). However, in TG-treated cells, a block of CCE by NO may result in a decrease in cytoplasmic calcium levels as Na+/Ca2+ exchange and the Ca2+-ATPase both pump calcium out (thus causing a decrease in internal calcium levels). Therefore, CCE is a candidate for NOB. The activity of NOB may dominate when internal calcium stores are low, because such stores are depleted by TG treatment (see e.g., Ref. 6). Because we observed identical effects of NO when cells are chronically exposed to low calcium levels, we infer that such exposure reduces calcium levels within internal stores. We also suggest that internal stores may be low in the relatively rare cases of gradual increases in cytoplasmic calcium levels, which NO blocks (not shown). However, the reason for putative store depletion under control conditions is unknown, and this suggests that other effectors, yet unrecognized, regulate internal stores.
The effect of NO on cytoplasmic calcium levels appears to depend on the levels of calcium within internal stores. When cells are exposed to normal external calcium levels, NO increases cytoplasmic calcium levels. By contrast, if cells are exposed to conditions expected to diminish calcium levels within internal stores, NO evokes a decrease in cytoplasmic calcium levels. When internal calcium stores are full, the capacity of the stores to release calcium into the cytoplasm dominates the capacity of the mechanism to remove calcium from the cytoplasm.
We fitted the NO dose-response curve using the simplest ligand binding theory, presuming one receptor. However, the evidence suggests that two receptors exist and that each likely competes for cytoplasmic calcium. The fitted parameters should be viewed as qualitative descriptions of the total action of NO, rather than a quantitative description of a single receptor. Therefore, we use terms such as EC50rather than K d, and n apprather than n H (Hill coefficient). Nevertheless, the shape of the NO dose-response relationship suggests that when NOA dominates, two NO molecules bind to NOA and evoke the increase in cytosolic calcium. Ultimately, these characteristics will have to be checked when NOA is purified. Further studies are necessary to determine the dose-response curve of the second receptor.
Several NO donors failed to elicit a detectable response in C6 glioma cells otherwise responsive to the gas. The organic nitrate ISDN, the thionitrite SNAP, and the iron-containing nitrosyl SNP all failed to evoke a detectable calcium response. By contrast, the NONOate DEA/NO evoked responses identical to those of the gas alone. We did not test the ineffective donors on cells known to be responsive to these agents (i.e., a positive control), so the failure of these agents may be attributed to our handling of the compounds. However, another laboratory has reported that 30 μM SNP is ineffective in C6 glioma cells (30). It is also possible that the conditions supporting the release of NO from these agents may not exist in C6 glioma cells under our experimental conditions. Organic nitrates and thionitrites require sufficient levels of thiols for nonenzymatic NO release (9). In addition, release of NO from some thionitrites require trace amounts of Cu2+(9). SNP appears to require a membrane-bound enzyme as a cofactor for release of NO (9). In contrast to these agents, DEA/NO donor spontaneously releases NO. Interestingly, high levels of thiols inhibit release of NO from NONOates (9). The mechanism of release of NO from SIN-1 is not well known but is thought to be spontaneous (9). However, it was not effective in our hands.
Hb solutions made from lyophilized Hb evoked substantial increases in cytosolic calcium levels (not shown), which precluded its use in experiments designed to determine whether NO was required for its effect. A previous report in which lyophilized bovine Hb was used found a similar effect (see Ref. 15,methods). These authors reduced the bound iron by standard procedures before use. Because we did not do so, we suggest that the effect of Hb on cytoplasmic calcium levels is unrelated to the state of oxidation of the heme groups. After considerable work, we found that human Hb, in dH2O (stored in liquid N2 ), was without effect on cell calcium levels. Sodium dithionite (0.10–2 mM), used in the preparation of hemoglobin, was without effect on calcium levels, and 1 μM FeCl2, added in the absence of NO, failed to affect calcium levels. These results may suggest that the lyophilization of Hb causes a conformational change in the structure of Hb that C6 glioma cells and neuronal cells (15) detect. Interestingly, after the nonlyophilized Hb sample was subjected to repeated freezing and thawing, this Hb also evoked an increase in cytosolic calcium levels. Further investigation of these interesting observations is necessary.
The NO-evoked quenching of fura 2 dye in the presence of Hb (Fig.3 D) was never seen in the absence of Hb (e.g., Fig. 1). Addition of Hb alone was without effect on the fluorescence emission of fura 2 excited at 363 nm. Hb and NO are both required for the quenching, and it is possible that NO opens up a calcium-impermeable (see Fig. 4 A) pore in the plasmalemma, allowing entry of Fe2+. The affinity of fura 2 for iron is 3–10 times higher than for calcium, and iron will quench fura 2 (11). We speculate that some unbound (perhaps tens of nanomolar) iron may exist in the 10 μM Hb solution. Further investigation of this interesting finding is required. NO does open a calcium-permeable pathway in the plasmalemma of astrocytes maintained in a mixed neuronal-glial culture (38).
What are the possible physiological consequences of two NO receptors regulating cytoplasmic calcium levels? In the central nervous system, astrocytes are structurally poised to sense the gas produced by all other types of cells in the brain. Candidate sources of NO in the brain include neurons, microglia, oligodendrocytes, and endothelial cells. Because astrocytic processes surround capillaries and neuronal elements, the NO produced by endothelial cells may also be sensed by astrocytes. Increases (or decreases) in astrocyte calcium may subsequently occur, depending on the levels of calcium inside the endoplasmic reticulum. Because the extracellular space in the brain is very small (24), low extracellular calcium levels might follow massive releases of excitotoxic glutamate in pathological conditions. Low calcium levels inside the ER may also result from transient cerebral ischemia and may contribute to the pathology after the insult (24). Vertebrate retinal Muller cells (13) and primate astrocytes (37) have the endothelial form of NOS (eNOS). Therefore, NO-evoked changes in astrocyte calcium levels (up or down) might be expected to alter spatial buffering characteristics for K+ (23,28) by locally altering the activity of calcium-activated K+ channels (if present in these cells). Although NO generated in astrocytes can contribute to neuronal apoptosis in ischemia (31), we do not know whether the NO-evoked changes in astrocytic calcium levels are related. The changes in calcium levels may influence other events related to NO, such as the reduction of glutamate uptake by astrocytes (39). The details of how astrocytic calcium increases or decreases in response to NO remain to be determined. Further work is required to characterize the NO-evoked increase in calcium levels in astrocytes in primary cell culture (31) and to determine whether NO has similar effects on astrocytes in vivo.
In summary, our results suggest that C6 glioma cells have two receptors for NO and that both cause changes in internal calcium levels, but in opposite directions. Both receptors may be activated simultaneously by NO, but one NO receptor may dominate the other, depending on the levels of calcium inside internal stores. The up- or downregulation of astrocytic calcium levels (locally) by NO could be particularly suited for adjusting potassium spatial buffering properties of glial cells under normal physiological conditions or for reducing astrocyte cytoplasmic calcium concentration in pathological conditions.
We thank Dr. Robert W. Noble for his generosity and helpful discussions.
Address for reprint requests and other correspondence: C. L. Bowman, Buffalo Institute for Medical Research, Veterans Affairs Medical Center, 3495 Bailey Ave., Buffalo, NY 14215-1129 (E-mail:).
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 © 2001 the American Physiological Society