 |
||||
|
||||
|
||||
|
||||
Previous Article | Next Article
The Journal of Neuroscience, January 1, 1999, 19(1):147-158
Nitric Oxide Acutely Inhibits Neuronal Energy Production
James R. Brorson1, Paul T. Schumacker2, and He Zhang3 1 Department of Neurology and the Committees on Neurobiology and Cell Physiology, 2 Department of Medicine and the Committee on Comparative Medicine and Pathology, and 3 Section of Neurosurgery, Department of Surgery, The University of Chicago, Chicago, Illinois 60637
ABSTRACT
Top
Abstract
Introduction
Disruption of mitochondrial respiration has been proposed as an action of nitric oxide (NO) responsible for its toxicity, but the effects of NO on the energetics of intact central neurons have not been reported. We examined the effects of NO on mitochondrial function and energy metabolism in cultured hippocampal neurons. The application of NO from NO donors or from dissolved gas produced a rapid, reversible depolarization of mitochondrial membrane potential, as detected by rhodamine-123 fluorescence. NO also produced a progressive concentration-dependent depletion of cellular ATP over 20 min exposures. The energy depletion produced by higher levels of NO (2 µM or more) was profound and irreversible and proceeded to subsequent neuronal death.
In contrast to the effects of NO, mitochondrial protonophores produced complete depolarizations of mitochondrial membrane potential but depleted the neuronal ATP stores only partially. Inhibitors of mitochondrial oxidative phosphorylation (rotenone or 3-nitropropionic acid) or of glycolysis (iodoacetate plus pyruvate) also produced only partial ATP depletion, suggesting that either process alone could partially maintain ATP stores. Only by combining the inhibition of glycolytic energy production with the inhibition of mitochondria could the effects of NO in depleting energy and inducing delayed toxicity be duplicated.
These results show that NO has rapid inhibitory actions on mitochondrial metabolism in living neurons. However, the severe ATP-depleting effects of high concentrations of NO are not fully explained by the direct effects on mitochondrial activity alone but must involve the inhibition of glycolysis as well. These inhibitory effects on energy production may contribute to the delayed toxicity of NO in vitro and in ischemic stroke.
Key words: nitric oxide; ischemia; peroxynitrite; poly-(ADP ribose) polymerase; mitochondria; glycolysis
INTRODUCTION
The marked toxicity of nitric oxide (NO) to neurons is not fully explained. NO is known to have a number of cellular effects of possible pathophysiological significance (Gross and Wolin, 1995
; Iadecola, 1997
) to form peroxynitrite (ONOO
NO impairs mitochondrial respiration in isolated mitochondria (Stadler et al., 1991
Neurons are particularly dependent on mitochondrial energy production. If neuronal mitochondria are inhibited by NO, energy depletion could result, possibly leading to either necrosis or apoptosis (Pang and Geddes, 1997
MATERIALS AND METHODS
Neuronal cultures. Dissociated cultures of hippocampal neurons were prepared from day 18 embryonic Sprague Dawley or Holtzman rats (with the sperm-positive day numbered as day 1), as previously described in detail (Brorson et al., 1997
Generation of solutions of NO. Most experiments used NO donors (see below) freshly dissolved from stock solutions in saline buffer immediately before application. Some experiments were performed with gaseous NO. NO gas (5% NO in N2), applied in different proportions with an excess of 20% O2/80% N2, was bubbled in stirred saline buffer at 22°C. During NO application the pH in the buffer was monitored. At higher rates of NO gas application the pH of the solution was found to decrease, presumably as a result of nitrous acid formation (Butler et al., 1995
Measurement of NO. To make approximate measurements of the NO concentrations generated by the NO donors or by gaseous NO, we used a NO-selective electrode (IsoNO, World Precision Instruments, Sarasota, FL). Calibration was performed according to instructions by using injections of defined amounts of NaNO2 in stirred deoxygenated solutions of KI and acetic acid, generating stoichiometric amounts of NO. Highly linear responses resulted and were used to convert the electrode current to [NO]. Measurements were made in 50 ml plastic tubes placed in a constant temperature water bath at 22-23°C. Immediately after the NO donor was mixed in buffer and at 5 min intervals thereafter, the tip of the NO electrode was placed in the NO solution and gently stirred for 1 min to achieve a stable reading and then was removed to control the buffer between readings. Separate [NO] measures also were made with the tip of the NO electrode in the well of the Plexiglas perfusion chamber used in physiological experiments during 5 min superfusions with NO donor-containing (or dissolved NO gas-containing) solutions.
Fluorescence imaging with rhodamine-123. Fluorimetric digital imaging of rhodamine-123 fluorescence was used to detect the effects on mitochondrial membrane potential. Hippocampal neurons were loaded by incubation for 10 min in 10 µg/ml rhodamine-123, followed by a 5 min wash, both at room temperature (22°C) in the standard saline buffer containing (in mM): NaCl 145, KCl 3, CaCl2 2, MgCl2 1, HEPES 10, and glucose 10, pH-adjusted to 7.4 with NaOH. The coverslips were mounted on a Plexiglas superfusion chamber, placed in a Nikon Diaphot fluorescence microscope, and continuously superfused with 0.5 µM tetrodotoxin-containing buffer to which various compounds were added. For digital imaging, excitation used a 495 nm light with a 505 nm dichroic mirror and a 515 nm long-pass barrier filter, plus a 40× fluorescence objective. Image acquisition and analysis were performed by commercial image-processing hardware and software (MetaFluor, Universal Imaging, West Chester, PA). A field of healthy-appearing neurons from the central region of a coverslip was selected, and the neurons were identified by typical morphology. Areas of interest were outlined over the soma of each neuron, including nuclear areas, and over a blank area of the coverslip for background. The average fluorescence was recorded digitally for each indicated area at 2 Hz for the duration of each experiment. Neurons were rejected if the baseline fluorescence was highly unstable or if significant digital saturation occurred during the experiment. Background traces were subtracted from the records for each neuron. Quantitative analysis was performed by comparing average background-corrected fluorescence at certain time points as a fraction of the difference between the baseline value and the peak carbonyl cyanide m-chlorophenylhydrazone-evoked (CCCP) fluorescence in the same cell. The relative fluorescence values were averaged over all imaged cells, usually 5-12 neurons for each experiment. Every condition was evaluated in four or more experiments.
ATP-luciferase assays. Cellular ATP and ADP assays in cultured neurons were performed with a firefly luciferase chemiluminescence assay essentially as described by Budd and Nicholls (1996)
20°C until an assay could be performed on all samples in parallel. Reagents were obtained from Analytical Luminescence Laboratory (Ann Arbor, MI). The low background reagent (Firelight LB) was used in 20 µl amounts with 100 µl of reaction buffer. Luminescence was measured by the Monolight 1500 luminometer (Analytical Luminescence Laboratory), made available via the generosity of Dr. Harinda Singh of the Howard Hughes Medical Institute of the University of Chicago (Chicago, IL). After a determination of the luciferase signal corresponding to the ATP concentration, pyruvate kinase (2 U) and phosphoenol pyruvate (0.5 mM) were added and a second measurement of luminescence was taken, the additional signal reflecting the conversion of ADP to ATP (driven to near-completion by an excess of phosphoenol pyruvate). The signal difference gave an estimate of ADP concentration for the calculation of ATP/ADP ratios.
Total protein was measured with the Bio-Rad DC Protein Assay kit (Bio-Rad Laboratories, Hercules, CA), according to the instructions supplied for the microplate assay protocol, and, through the generosity of Dr. Anthony Reder (Department of Neurology, University of Chicago), a Molecular Devices Thermomax microplate reader (Palo Alto, CA) at a wavelength of 650 nm. Because the buffer used for luciferase assays was incompatible with any of the available colorimetric protein assays, we were unable to normalize ATP values to concurrent measures of the total cellular protein. Instead, the total ATP signal per coverslip of cultured neurons exposed to various treatments was normalized to the total ATP signal from parallel control coverslips. Figures show the mean ± SD of normalized ATP or of the ATP/ADP ratio for each condition. An estimation of ATP per unit of protein also was performed by measures from coverslips treated in parallel, with quantities calculated by using linear regression of calibration data generated with known amounts of ATP and of ovalbumin. The ATP per total protein from cultured neurons was found to be 8.7 ± 1.6 pmol/mg (mean ± SD) in controls versus 0.7 ± 0.1 pmol/mg after exposure to 30 µM SNOC.
Survival assays. Cell death and survival were assayed by using the fluorescent markers fluorescein diacetate and propidium iodide as previously described (Brorson and Zhang, 1997
Data analysis. Statistical evaluations of rhodamine fluorescence changes and of survival were by ANOVA on ranks and of normalized ATP values by repeated measures ANOVA on ranks, followed by Dunn's test of pairwise comparisons to control values (SigmaStat, Jandel Scientific, San Rafael, CA).
Materials. A 100 mM stock of S-nitrosocysteine (SNOC) was produced immediately before each use from a mixture of 100 mM L-cysteine and 100 mM NaNO2 by acidification with 5% (v/v) with 10N HCl. SNOC was added to saline buffer at room temperature, and the pH was brought back to 7.4 with 1N NaOH. It was applied within minutes of its synthesis, and no additional fresh SNOC was added subsequently during the 20 min exposures. After synthesis, stocks of SNOC were left at room temperature for several days to degrade fully to "old SNOC" for use as a control. S-nitroso-N-acetylpenicillamine (SNAP), stored in 1 M aliquots in DMSO at
-cyano-4-hydroxycinnamic acid (4-CIN) at 1 M were stored at
RESULTS
NO donors
In previous work we have shown that NO donors, including SNOC, SNAP, and SIN-1, induce acute toxicity in neurons (Brorson and Zhang, 1997
The varied physiological effects of different NO donors must be interpreted in the light of the NO concentration ([NO]) that they each produced. Using an NO-sensitive electrode, we measured the NO concentrations produced by the various NO donors in unstirred buffer over 20 min periods (Table 1). In static buffer, high initial readings from SNOC dilutions decayed over the first few minutes to substantially lower values. At 5 min the accumulations of NO ranged from 0.5 µM (for 3 µM SNOC) to nearly 10 µM (for 1 mM SNOC). The relationship between [NO] and SNOC concentration was nonlinear, reflecting more complex chemical kinetics than a simple irreversible first order decay. In contrast, SNAP (1 mM) produced much lower [NO], rising slowly to only ~0.5 µM, whereas SIN-1 (1 mM) produced no detectable [NO]. The other NO donors used, spermine-NO (300 µM) and DEA-NO (100 µM), produced micromolar concentrations of NO. Neither saline alone nor 1 mM L-cysteine plus NaNO2 generated signals (data not shown). Because the actual NO concentrations applied to the cells might be dependent on the handling of the agents and on the configuration used, they also were assayed in the perfusion chamber that was used in our experiments. The values of [NO] delivered to the perfusion chamber from solutions of the various NO donors closely matched those measured in the static buffer at the 5 min time point, although the initial high peak values generated by SNOC were blunted in the perfusion chamber (data not shown). The NO concentrations generated from gaseous NO were measured also, as cited below.
View this table: [in this window]
[in a new window]
Table 1. Concentrations of NO, in micromolars NO depolarizes mitochondria in living hippocampal neurons
To study mitochondrial function in intact hippocampal neurons, we used the fluorophore rhodamine-123 as an indicator of mitochondrial depolarization (Duchen, 1992
m, which is maintained by the proton translocation of the electron transport system. If this activity is inhibited pharmacologically or if the proton gradient is eliminated by one of the protonophores CCCP or FCCP, the membrane depolarizes and rhodamine-123 is released into the cytoplasm. We observed increases in overall cytoplasmic fluorescence with mitochondrial depolarization (Figs. 1, 2). Basal rhodamine-123 fluorescence was quite stable in neurons, fading only slowly, whereas the fluorescence in intermingled glial cells faded rapidly within the first few minutes of observation. CCCP (1 µM) produced a rapid rise in average somatic fluorescence, reaching a plateau value within 90 sec. With the washout of CCCP, fluorescence recovered to the same baseline in each cell over a 1-5 min interval. Repeated exposures to CCCP (at 0.1, 1, and 10 µM) or by FCCP (1 or 10 µM) produced approximately the same maximal rhodamine-123 fluorescence increases (data not shown). We used this peak increase in fluorescence produced by 1 µM CCCP to normalize subsequent changes produced by other agents.
View larger version (93K):[in this window]
[in a new window]
Figure 1. Mitochondrial imaging with rhodamine-123. A, Photomicrographs of rhodamine-123 fluorescence in cultured hippocampal neurons (26 DIV). Successive frames show (1) a phase contrast image, (2) baseline fluorescence with a particulate cytoplasmic distribution consistent with localization of fluorophore to mitochondria, (3) increased diffuse cytoplasmic fluorescence after 1 min of stimulation with CCCP (1 µM), and (4) recovery of the particulate fluorescence pattern with the washout of CCCP. Scale bar, 30 µm. B, Digital fluorimetric imaging of rhodamine-123 fluorescence at baseline, after 1 min of exposure to 1 mM SNOC, and after 3 min of washout, demonstrating a reversible increase in cytoplasmic fluorescence induced by SNOC (indicated by warmer colors on the pseudocolor scale).
High concentrations of NO (~9 µM) generated by 1 mM SNOC consistently produced an increase in rhodamine-123 fluorescence in most neurons (Figs. 1, 2). The increase was a variable fraction of the signal change produced by CCCP. The effect generally occurred quickly with the application of SNOC, sometimes increasing further over a 5 or 10 min exposure, and often with incomplete recovery after washout. Changes in rhodamine-123 fluorescence at 1 and 5 min, quantified as a fraction of the maximal change produced by CCCP, showed that the average fluorescence in cells perfused with buffer alone slightly decreased over 5 min (Fig. 2E). In contrast, SNOC (1 mM) produced significant average increases in fluorescence, whereas inactivated SNOC ("old SNOC") or L-cysteine plus NaNO2 (from which SNOC is synthesized) had no significant effects. Neither the NMDA antagonist MK-801 (1 µM) nor the enzyme superoxide dismutase-1 (SOD) changed the significant effects of NO released from SNOC. Electron transport inhibitors, including the complex I inhibitor rotenone (5 µM) (Budd and Nicholls, 1996
View larger version (41K):[in this window]
[in a new window]
Figure 2. Rhodamine-123 detection of mitochondrial depolarization. Cytoplasmic rhodamine-123 fluorescence (arbitrary units) versus time for seven representative hippocampal neurons for each experiment. A, In control experiments 90 sec exposures to CCCP (1 µM) resulted in rapid increases in fluorescence, followed by stable or slightly decreased signals over 10 min exposures. B, SNOC (1 mM) produced rapid fluorescence increases. C, The effects of SNOC were not blocked by the coapplication of superoxide dismutase (SOD; 100 U/ml). D, The complex I inhibitor rotenone (5 µM) produced slower increases in fluorescence in some imaged neurons, with little reversibility. E, Quantitative summary of rhodamine-123 fluorescence studies, comparing changes in relative cytoplasmic fluorescence after 1 or 5 min averaged over all neurons in each experiment. These values then were averaged over all experiments for each agent. The concentrations used were 1 mM for SNOC, "old SNOC," SIN-1, SNAP, and L-cysteine (L-Cys) plus NaNO2, 100 U/ml for SOD, 1 µM for MK-801, 5 µM for rotenone, and 1 mM for 3-nitropropionic acid (3-NP) (mean ± SEM; *p < 0.05; n = 6-10 experiments for each condition).
We performed similar experiments by applying a range of NO concentrations, using several NO donors as well as gaseous NO dissolved in buffer (Fig. 3). Agents producing several micromolar NO concentrations, including 30 µM SNOC and 100 µM DEA-NO, produced large increases in rhodamine-123 fluorescence. NO concentrations in the 1-2 µM range (produced by 3 µM SNOC or 300 µM spermine-NO) appeared to produce smaller, variable qualitative fluorescence changes (Fig. 3A) that did not reach statistical significance. NO dissolved from gas at 1.55 µM also evoked significant increases in rhodamine-123 fluorescence, whereas buffer bubbled with 20% O2/80% N2 alone showed no changes from control traces. When the mean change in rhodamine-123 fluorescence was plotted as a function of the mean NO concentrations in the perfusion chamber for each treatment, a striking NO concentration-dependent relationship emerged (Fig. 3D). Concentrations of NO rising into the micromolar range produced increasing acute rhodamine-123 signals, indicating changes of
View larger version (35K):[in this window]
[in a new window]
Figure 3. Mitochondrial effects of other NO donors. A, Representative cytoplasmic rhodamine-123 fluorescence recordings from hippocampal neurons exposed to 3 µM SNOC, 30 µM SNOC, and 100 µM DEA-NO. CCCP (1 µM) was applied before and after each agent. B, The effects of dissolved NO from gas on rhodamine-123 fluorescence. Saline buffer was bubbled continuously with an NO/O2/N2 mixture and perfused over the neurons. Buffer bubbled with O2/N2 alone (top) had no effect. The buffer containing dissolved NO produced a modest rise in the rhodamine123 fluorescence (bottom). C, Summary of average changes in relative rhodamine-123 fluorescence (mean ± SEM; *p < 0.05 compared with controls) for 3 and 30 µM SNOC, 300 µM spermine-NO, 100 µM DEA-NO, and for NO from gas at two rates, which generated dissolved [NO] values at 5 min of 0.44 ± 0.07 and 1.55 ± 0.44 µM (mean ± SEM). D, The average changes in rhodamine-123 fluorescence are plotted versus the average concentrations of NO at 5 min produced by each agent, measured in the perfused chamber (mean ± SEM). The mitochondrial depolarization, indicated by fluorescence changes, correlated with the NO concentration (r = 0.86; p < 0.01).
As a final confirmation that NO actually mediated the effects of SNOC on rhodamine-123 fluorescence, the NO scavenging agents oxyhemoglobin and PTIO (Akaike et al., 1993
View larger version (21K):[in this window]
[in a new window]
Figure 4. NO scavengers block the effects of SNOC on mitochondrial potential. Shown are the effects of the NO scavenging agents oxyhemoglobin (Hgb; 30 µm) or PTIO (50 µm) on NO-induced changes in rhodamine-123 fluorescence hippocampal neurons. A, After a 90 sec exposure to CCCP (1 µM), fluorescence (arbitrary units) remained stable during 5 min of NO exposure (1 mM SNOC) in the presence of oxyhemoglobin but began to rise after the removal of oxyhemoglobin. B, SNOC (1 mM) alone for 10 min produced rapid, sustained, and reversible fluorescence increases. C, Quantitative summary of effects of scavengers on changes in rhodamine-123 fluorescence relative to a maximal fluorescence change produced by CCCP. SNOC alone produced a rapid, sustained rise in relative fluorescence. No significant increase was produced by SNOC in the presence of either oxyhemoglobin or PTIO. PTIO alone, in the pretreatment phase, appeared to produce a small decrease in rhodamine-123 fluorescence, likely because of its absorbance by the blue-colored PTIO (mean ± SEM; *p < 0.05 compared with baseline).
It is noteworthy that the decreases in
NO rapidly depletes ATP in hippocampal neurons
Mitochondrial dysfunction might result in energy depletion. Luciferase-based assays of total ATP and of the ATP/ADP ratio served to assess neuronal energy stores after 20 min exposures to various conditions (Fig. 5A). Average NO concentrations of 5-9 µM, produced by 300 µM or 1 mM SNOC, resulted in severe depletions of total ATP as compared with controls. Again, the depletion of ATP appeared to be mediated specifically by NO, because it was not reproduced by controls of old SNOC or L-cysteine plus NaNO2. Changes in the ratio of ATP/ADP generally paralleled the effects on total ATP, with substantially greater variance, likely related to uncertainties in the ADP measurements (see Materials and Methods). Subsequent analyses of neuronal energy stores primarily examined total ATP. Qualitative changes in the ATP/ADP ratio continued to parallel the effects on total ATP.
View larger version (32K):[in this window]
[in a new window]
Figure 5. Depletion of neuronal ATP by NO. Normalized total ATP values and ATP/ADP ratios (means ± SD), were determined by luciferase assays, as described, from cultured hippocampal neurons. A, After 20 min exposures, NO produced by 300 µM or 1 mM SNOC produced a near-complete depletion of cellular total ATP stores and of the ATP/ADP ratio, whereas treatments with "old SNOC" or L-Cys+NaNO2, also at 1 mM, had no substantial effect on energy stores although the slight changes in total ATP, but not in the ATP/ADP ratio, reached statistical significance (n = 5; *p < 0.05 compared with controls). B, Concentration-response data for SNOC on cellular energy indicated highly sensitive effects on ATP and ATP/ADP (n = 4). C, The time course of ATP depletion by SNOC over a 20 min exposure period and over the subsequent 24 hr after a return to the culture medium. D, Effect of receptor and enzyme antagonists on the recovery of cellular ATP stores at 6 hr after NO exposure. The NMDA antagonist MK-801 (1 µM), the PARP antagonist benzamide (500 µM), the calpain antagonists calpain inhibitor-1 (CI-1; 3 µM) or MDL-28170 (10 µM), and superoxide dismutase (SOD; 100 U/ml) each allowed partial recovery of cellular ATP levels at 6 hr after NO exposure (n = 4; *p < 0.05 compared with controls and p < 0.05 compared with SNOC alone).
ATP depletion by SNOC exhibited a concentration-response relationship with a half-maximal effect at ~3 µM, corresponding to an [NO] of ~0.5 µM (Fig. 5B). The time course was characterized by a steady decline of total ATP over the 20 min exposure, with no significant recovery of cellular energy stores in the subsequent 24 hr (Fig. 5C). Agents that partially protect neurons from delayed death after NO exposure (Brorson and Zhang, 1997
NO produced by other agents also depleted energy stores (Fig. 6). NO at concentrations ranging from ~0.5 µM (from 3 µM SNOC or 1 mM SNAP) to 1.5 µM (from 30 µM SNOC) or 1.9 µM (from 300 µM spermine-NO) each partially reduced total ATP and the ATP/ADP ratio after 20 min. SIN-1 (1 mM), which did not produce a measurable amount of NO, nevertheless also partially depleted energy stores. NO dissolved in solution from a gaseous mixture with O2/N2 also was applied, using a continuous exchange of buffer to maintain neutral pH over the 20 min exposures. At an average [NO] of 0.55 ± 0.20 µM (mean ± SD), a modest depletion of ATP to ~60% of parallel controls occurred, whereas an [NO] of 2.98 ± 1.46 µM depleted most of the neuronal energy stores (Fig. 6B). With the exception of the anomalous energy depletion by SIN-1, the ATP depletion under various conditions correlated with the associated measured [NO], with concentrations of NO >1-2 µM resulting in severe depletion of neuronal energy stores (Fig. 6C).
View larger version (21K):[in this window]
[in a new window]
Figure 6. ATP effects of lower concentrations of NO. A, Exposures to lower concentrations of SNOC were repeated in parallel with other NO donors, producing lower NO concentrations. SNAP (1 mM), spermine-NO (300 µM), SNOC (3 or 30 µM), and SIN-1 (1 mM) each produced significant decrements of ATP and ATP/ADP (mean ± SD; *p < 0.05). B, Neurons also were exposed for 20 min to an O2/N2 mixture alone and to NO from 5% NO gas dissolved in stirred buffer at two concentrations, 0.55 ± 0.09 and 2.98 ± 0.65 µM (mean ± SEM; averaged over 20 min). O2/N2 alone produced no measurable [NO] (0.00 ± 0.01 µM), and this treatment resulted in only a slight decrement in ATP as compared with untreated control coverslips. In contrast, NO from gas produced large concentration-dependent ATP depletions (mean ± SD; n = 4; p < 0.05 compared with O2/N2 treatment alone). C, The relative ATP after 20 min exposures to various NO donors or to gaseous NO was compared with the average measured [NO] produced by each agent. A strong concentration dependence was observed, with exposure to ~2 µM NO or more producing a 50% depletion of ATP. The effect of SIN-1 (asterisk) was anomalous in that it produced significant ATP depletion without producing measurable [NO].
Our expectation was that NO was causing energy depletion in the neurons by the inhibition of mitochondrial function. To test this hypothesis, we compared the effects of NO (from 1 mM SNOC) on ATP with the effects of mitochondrial inhibitors. In contrast to the near-total depletion of ATP by 20 min exposure to NO, rotenone (5 µM), 3-nitropropionic acid (1 mM), or the mitochondrial ATP synthase inhibitor oligomycin (6 µg/ml) each only produced a partial depletion of ATP (Fig. 7). We confirmed that this concentration of oligomycin was saturating by comparing 6 with 18 µg/ml, which produced no significant additional ATP depletion over 20 min exposures (data not shown). The protonophore CCCP (1 µM) or CCCP in combination with oligomycin also produced only partial depletion of ATP over 20 min, much less than that of SNOC. Unexpectedly, it thus appeared that the toxic NO concentrations, which had been found to depolarize mitochondria only partially, were having substantially greater energy-depleting effects than the complete inhibition of mitochondrial ATP production.
View larger version (31K):[in this window]
[in a new window]
Figure 7. Effects of mitochondrial inhibitors on ATP. Effects of NO from SNOC on neuronal ATP at 20 min and 6 hr were compared with those of oligomycin (6 µg/ml), CCCP (1 µM), rotenone (5 µM), 3-nitropropionic acid (3-NP; 1 mM), and the combinations of SNOC plus oligomycin (SNOC+Oligo) or CCCP plus oligomycin (CCCP+Oligo). All treatments were applied for 20 min, and the cells were harvested for ATP assay at 20 min and 6 hr (mean ± SD; n = 6; *p < 0.05 compared with controls and
ATP depletion: Increased consumption or decreased production?
One possible explanation for the disparity between the energy depletion produced by NO and that by mitochondrial inhibitors might be that the severe effects of NO are the result of a large increase in ATP consumption as well as of the inhibition of mitochondrial ATP production. Indeed, PARP activation by NO is thought to cause toxicity in part because of increased ATP consumption (Zhang et al., 1994
Contributions of glycolysis to energy stores
Glycolytic energy production alone generally has been thought to be insufficient to meet the energy demands of neurons (Siesjö, 1978
View larger version (27K):[in this window]
[in a new window]
Figure 8. Comparison of effects of the inhibition of glycolysis and of mitochondria on ATP levels. Shown are normalized ATP values after 20 min treatments with the indicated agents (mean ± SD). A, The glycolytic inhibitor iodoacetate (1 mM) caused a substantial depletion of ATP, which was prevented mainly by the inclusion of 10 mM pyruvate (pyr) in the buffer. The monocarboxylic acid uptake inhibitor 4-CIN (10 mM) reversed the effect of pyruvate. FCCP (1 µM) partially depleted ATP; this was partially prevented by oligomycin (6 µg/ml). Combinations of iodoacetate (iodo) and oligomycin (oligo) and FCCP caused a severe depletion of ATP after 20 min (n = 6; *p < 0.05 compared with controls and
Although high concentrations of NO appeared to be inhibiting both glycolytic and oxidative energy production, the concentrations required to inhibit the two arms of energy metabolism might be quite different. To assess whether a lower [NO] (~0.5 µM) might inhibit one of the processes selectively, we applied 3 µM SNOC (Fig. 8B). Adding 10 mM pyruvate to 3 µM SNOC did not reverse its partial ATP depletion substantially, showing that the inhibitory effect of NO was not acting at glycolysis alone. Adding inhibitors of glycolysis or of mitochondrial ATP synthesis appeared to deplete ATP further, suggesting that neither arm of energy production was fully inhibited by modest [NO]. With the full inhibition of glycolysis by iodoacetate plus pyruvate (Fig. 8C), increasing SNOC concentrations induced additional energy depletion, with a broad concentration dependence. Similarly, with the full inhibition of mitochondria by FCCP and oligomycin, SNOC also produced a concentration-dependent further depletion of ATP with a somewhat higher concentration threshold. These pharmacological manipulations partially isolated the effects of NO on mitochondria from those on glycolysis and suggested that in neurons NO can inhibit both mitochondrial energy production and, somewhat less sensitively, glycolytic energy production.
Delayed toxicity of NO is not mimicked by inhibitors of mitochondria
Given the evidence for a rapid energy depletion by NO in neurons, we asked whether the energy depletion itself could be sufficient to explain the neuronal death induced by NO. We applied the various NO donors that were found to produce ATP depletion in 20 min exposures, including SNOC, SIN-1, SNAP, and spermine-NO. Each agent significantly reduced the survival of hippocampal neurons 24 hr after exposure (Fig. 9A). When the survival was compared with measures of the average [NO] over 20 min for each NO donor, a clear concentration-dependent toxicity was demonstrated, with 2 µM NO reducing survival by approximately one-third (Fig. 9B). SIN-1 produced toxicity despite its undetectable elevations of [NO]. The proportional decreases in survival for each NO donor were notably less than the relative decreases in ATP after 20 min (compare Figs. 9B and 6C). Nevertheless, the cell death induced by the various donors correlated closely with the degree of energy depletion that they produced.
View larger version (27K):[in this window]
[in a new window]
Figure 9. Neuronal survival after energy-depleting treatments. A, Neuronal survival was assayed 1 d after 20 min exposures to control buffer, SIN-1 (1 mM), SNAP (1 mM), spermine-NO (300 µM), or SNOC (3 or 30 µM). Each agent produced a modest decrease in survival (n = 4; *p < 0.05). B, The survival at 24 hr after exposure to each NO donor was correlated with the measured [NO] produced by each agent. SIN-1 (asterisk) was anomalous, producing toxicity without producing a measurable [NO] (see Table 1). C, Survival assayed 1 d after 20 min exposures to high [NO], ~9 µM, from 1 mM SNOC as compared with metabolic inhibitors (mean ± SD; n = 4; *p < 0.05 compared with controls and
If the neurotoxicity of NO depends on the energy depletion it produces, then other agents that deplete energy (presumably without other actions specific to NO or peroxynitrite) also should produce similar neuronal death. Therefore, we also tested the delayed neuronal survival after 20 min treatments with saturating concentrations of inhibitors of energy production in comparison to that after treatment with a high level of NO (~9 µM, from 1 mM SNOC). This concentration substantially decreased neuronal survival after 24 hr. In contrast, iodoacetate, iodoacetate/pyruvate, and oligomycin each produced substantially less toxicity, paralleling the lesser effects of these agents on energy depletion. Of note, the combination of iodoacetate and oligomycin, which causes complete energy depletion similar to that of 1 mM SNOC, produced as much toxicity at 24 hr, suggesting that the effects of NO on energy metabolism may be sufficient to explain the neurotoxicity it produces.
DISCUSSION
Many of the varied physiological roles of NO are mediated by cyclic GMP formation. In addition to the activation of soluble guanylate cyclase, NO is known to act on many other cellular sites by binding to heme and Fe-S centers (Butler et al., 1995
In the present study we have demonstrated that the application to cultured hippocampal neurons of toxic concentrations of NO results in mitochondrial depolarization and progressive ATP depletion. Remarkably, the ATP depletion after 20 min exposures to high NO concentrations is profound and substantially greater than that produced by inhibitors of mitochondrial energy production, suggesting that the energy-depleting action of NO is not confined to the inhibition of mitochondrial metabolism but also may involve the inhibition of glycolytic energy production.
To understand the pathophysiological relevance of these findings, it is essential to compare them with the range of the [NO] found in biological tissues. The maximal NO levels used in these experiments were high, ranging up to 10 µM (see Table 1). These high concentrations produced large depolarizations of mitochondria and near-complete depletions of energy stores after 20 min. However, lower [NO] from NO donors or from dissolved gaseous NO also produced mitochondrial depolarizations, and partial depletions of ATP were found with [NO] of <0.5 µM. How do these compare with physiological or pathophysiological NO concentrations? Resting levels of [NO] in tissue are thought to be in the low nanomolar range (Gross and Wolin, 1995
The effect of temperature must be considered in the interpretation of the present experiments. Physiological experiments studying mitochondrial membrane potential were performed at room temperature, whereas the exposures for the reported ATP and survival assays generally were performed at 37°C. However, in separate experiments there were no significant differences in the ATP depletion by SNOC exposure at room temperature when compared with those at 37°C (data not shown). Furthermore, the exposures to NO dissolved from gas were done at 22°C, and they produced ATP depletions in line with those from chemical NO donors at 37°C. Thus the differences in temperature do not alter the large effect of NO on energy metabolism substantially.
In these experiments SIN-1 produced no measurable NO accumulation. Nevertheless, it produced partial energy depletion at 20 min and neuronal death at 24 hr, consistent with its well established neuronal toxicity (Lipton et al., 1993
Is ATP depletion caused by decreased production or increased consumption?
The present studies demonstrate the occurrence of rapid mitochondrial depolarization by NO in hippocampal neurons. Energy depletion soon follows, and the facile conclusion is that decreased production of ATP is entirely responsible. However, NO also may increase ATP hydrolysis by the cell, particularly by activation of PARP (Zhang et al., 1994
Potential sites at which NO may inhibit ATP production
These experiments directly demonstrate the effect of NO exposure on mitochondrial membrane potential. Although the depolarization produced by NO appears to be incomplete as compared with that produced by CCCP, the degree of impairment of oxidative phosphorylation may be underestimated by this method. Inhibition of electron transport may depolarize the mitochondrial membrane only to the point at which turnover of the ATP synthase slows to a halt, stopping the drain of the proton gradient. During the first several minutes cellular ATP stores may remain high enough that they contribute to maintaining the
Previous work has demonstrated the suppression of mitochondrial respiration by NO. In hepatocytes NO was reported to inhibit complexes I and II and, most sensitively, mitochondrial aconitase, an enzyme of the tricarboxylic acid cycle (Stadler et al., 1991
How might NO affect glycolysis to produce the profound energy depletion reported here? The most likely target is the glyceraldehyde-3-phosphate dehydrogenase complex, which is inhibited by NO (Molina y Vedia et al., 1992
The traditional emphasis of the dependence of neurons on aerobic metabolism stems from experiments clearly demonstrating the vulnerability of the brain to hypoxia (Siesjö and Wieloch, 1985
Relevance to neurological disease processes
If energy depletion by NO is relevant to its toxicity, why did lower [NO], here found sufficient to produce substantial energy depletion, produce relatively small decreases in neuronal survival at 24 hr? It may be that only the more severe insults produce rapid neuronal death via necrosis (Bonfoco et al., 1995
FOOTNOTES Received Sept. 15, 1998; accepted Oct. 12, 1998.
This work was supported by Grant NS01630 from National Institutes of Health to J.R.B. and by the support of the Brain Research Foundation to H.Z. and J.R.B. We thank Arush Angirasa for technical assistance in performing the neuronal cultures, toxicity assays, and ATP-luciferase assays and William Thistlethwaite for assistance in measuring NO concentrations.
Correspondence should be addressed to Dr. James R. Brorson, Department of Neurology, MC2030, The University of Chicago, 5841 South Maryland Avenue, Chicago, IL 60637.
Dr. Zhang's present address: Department of Neurosurgery, The University of Mississippi Medical Center, 2500 North State Street, Jackson, MS 39216.
REFERENCES
- Akaike T, Yoshida M, Miyamoto Y, Sato K, Kohno M, Sasamoto K, Miyazaki K, Ueda S, Maesa H (1993) Antagonistic action of imidazolineoxyl N-oxides against endothelium-derived relaxing factor/NO through a radical reaction. Biochemistry 32:827-832[Medline].
- Albin RL, Greenamyre JT (1992) Alternative excitotoxic hypotheses. Neurology 42:733-738
[Abstract/Free Full Text] .- Almeida A, Heales SJR, Bolaños JP, Medina JM (1998) Glutamate neurotoxicity is associated with nitric oxide-mediated mitochondrial dysfunction and glutathione depletion. Brain Res 790:209-216[ISI][Medline].
- Beal MF (1995) Aging, energy, and oxidative stress in neurodegenerative diseases. Ann Neurol 38:357-366[ISI][Medline].
- Beal MF, Ferrante RJ, Browne SE, Matthews RT, Kowall NW, Brown RH (1997) Increased 3-nitrotyrosine in both sporadic and familial amyotrophic lateral sclerosis. Ann Neurol 42:646-654.
- Bindokas VP, Jordan J, Lee CC, Miller RJ (1996) Superoxide production in rat hippocampal neurons: selective imaging with hydroethidine. J Neurosci 16:1324-1336
[Abstract/Free Full Text] .- Bolaños JP, Heales SJR, Land JM, Clark JB (1995) Effect of peroxynitrite on the mitochondrial respiratory chain: differential susceptibility of neurones and astrocytes in primary culture. J Neurochem 64:1965-1972[ISI][Medline].
- Bolaños JP, Almeida A, Medina JM (1998) Nitric oxide mediates brain mitochondrial damage during perinatal anoxia. Brain Res 787:117-122[ISI][Medline].
- Bonfoco E, Krainc D, Ankarcrona M, Nicotera P, Lipton SA (1995) Apoptosis and necrosis: two distinct events induced, respectively, by mild and intense insults with N-methyl-D-aspartate or nitric oxide/superoxide in cortical cell cultures. Proc Natl Acad Sci USA 92:7162-7166
[Abstract/Free Full Text] .- Brorson JR, Zhang H (1997) Disrupted [Ca2+]i homeostasis contributes to the toxicity of nitric oxide in cultured hippocampal neurons. J Neurochem 69:1882-1889[ISI][Medline].
- Brorson JR, Sulit RA, Zhang H (1997) Nitric oxide disrupts Ca2+ homeostasis in hippocampal neurons. J Neurochem 68:95-105[ISI][Medline].
- Budd SL, Nicholls DG (1996) A reevaluation of the role of mitochondria in neuronal Ca2+ homeostasis. J Neurochem 66:403-411[ISI][Medline].
- Butler AR, Flitney FW, Williams DLH (1995) NO, nitrosonium ions, nitroxide ions, nitrosothiols, and iron nitrosyls in biology: a chemist's perspective. Trends Pharmacol Sci 16:18-22[Medline].
- Cassina A, Radi R (1996) Differential inhibitory action of nitric oxide and peroxynitrite on mitochondrial electron transport. Arch Biochem Biophys 328:309-316[ISI][Medline].
- Duchen MR (1992) Ca2+-dependent changes in the mitochondrial energetics in single dissociated mouse sensory neurons. Biochem J 283:41-50.
- Erecinska M, Nelson D, Vanderkooi JM (1995) Effects of NO-generating compounds on synaptosomal energy metabolism. J Neurochem 65:2699-27
