| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
REPORT
1Zentrum für Physiologie und Pathophysiologie, Abteilung Neuro- und Sinnesphysiologie, 2Zentrum Biochemie, and 3Deutsche Forschungsemeinschaft Research Center for Molecular Physiology of the Brain, Georg-August-Universität Göttingen, Göttingen, Germany
Submitted 28 September 2005; accepted in final form 4 April 2006
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
; Duchen 1999; Nicholls and Budd 2000). Several central neurons react rapidly to oxygen shortage, but the mechanisms triggering their early response and the exact involvement of mitochondria in these events are still unclear. Because mitochondria require a continuous oxygen supply, they might be the primary sensors of hypoxia, and their function might be impaired early. This is suggested by previous studies, which reported rapid depolarization of mitochondria in hypoxic/anoxic hippocampal slices (Bahar et al. 2000; Schuchmann et al. 2000). Therefore one might suspect that mitochondria act as metabolic sensors, preparing their host cells for metabolic compromise and adapting neuronal network activity to metabolic supply and demand. Such mitochondrial sensor functions have been discussed for arterial chemoreceptors (Lopez-Barneo et al. 2001; Waypa et al. 2001). Also, we earlier found evidence that mitochondria apparently initiate the anoxic response of dorsal vagal neurons (Müller et al. 2002).
The four complexes of the mitochondrial respiratory chain are at the core of cellular energy production. While transferring electrons from NADH (nicotinamide adenine dinucleotide) to oxygen, three of the four respiratory complexes (complexes I, III, and IV) extrude protons from the mitochondrial inner space (matrix) into the cytosol. The resulting inwardly directed proton gradient across the inner mitochondrial membrane—also referred to as proton motive force—gives rise to the mitochondrial membrane potential (
m), and it drives the mitochondrial ATP synthase (also termed F0F1 ATPase or complex V) (Mitchell 1961). Therefore short-circuiting the proton gradient by protonophores or inhibition of the respiratory chain, e.g., by lack of oxygen will inevitably affect mitochondrial ATP synthesis. During such mitochondrial failure, the only source available for ATP production remains glycolysis, which is usually not sufficient to maintain undisturbed cellular function. Yet ATP is not the only mitochondria-derived messenger. Mitochondrial metabolism also generates reactive oxygen species (Boveris and Chance 1973) and modulates the cytosolic redox couples NAD/NADH and FAD/FADH2 (flavin adenine dinucleotide) (Schuchmann et al. 2001; Shuttleworth et al. 2003), all of which may potentially act as cellular signaling molecules (Dröge 2002; Park et al. 1995).
To assess to what extent neuronal tolerance and responses to hypoxia are modulated by mitochondrial (dys-)function, we blocked mitochondrial metabolism at various sites and probed for changes in the susceptibility of hippocampal slices to hypoxic spreading depression (HSD). HSD resembles a synchronized massive depolarization of neurons and glial cells that gives rise to a negative deflection of the extracellular DC potential and results in the loss of neuronal excitability (for recent reviews, see Somjen 2001, 2004). Once ignited focally, it slowly spreads out in neural tissue such as cortex (Basarsky et al. 1998), hippocampus (Aitken et al. 1998; Basarsky et al. 1998), retina (Martins-Ferreira et al. 2000), cerebellum (Nicholson 1984), and brain stem (Richter et al. 2003) at a velocity of a few millimeters per minute. Spreading depression is generated by the concerted activation of several types of cation channels acting as parallel pathways and mediating massive neuronal Na+ and Ca2+ influx (Kager et al. 2002; Müller and Somjen 1998, 2000a,b; Somjen 2001). The contribution of mitochondria to the process is not clear.
Mitochondria immediately react to anoxia (Schuchmann et al. 2000), and they depolarize strongly during spreading depression (Bahar et al. 2000), yet details of their responses and their role in signaling are poorly understood. To analyze the impact of the severity and the site of mitochondrial inhibition, we applied various inhibitors selectively targeting respiratory complexes I, II, III, or IV: rotenone, 3-nitropropionic acid (3-NPA), antimycin A, and cyanide, respectively. Alternatively, mitochondria were depolarized by the uncoupler carbonyl cyanide 4-(trifluoromethoxy) phenylhydrazone (FCCP) or mitochondrial ATP synthesis was targeted by the inhibitor of the FoF1 ATPase, oligomycin. Under these conditions of mitochondrial impairment, we analyzed changes in the characteristic HSD parameters and correlated them to changes in the mitochondria-controlled parameters (cellular ATP, NADH, FAD, and m) induced by the very drug treatment.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
Hippocampal tissue slices (400 µm thickness) were prepared from ether anesthetized, 4- to 8-wk-old Sprague-Dawley rats (150–300 g body wt, mostly males) as described earlier (Hepp et al. 2005; Müller and Somjen 2000a). Recordings were performed in an interface recording chamber (Oslo style) at 35–36°C which was aerated with 95% O2-5% CO2 (400 ml/min) and perfused with oxygenated artificial cerebrospinal fluid (ACSF; flow rate 4 ml/min). Hypoxia was induced by switching the chamber’s gas supply to 95% N2-5% CO2; the aeration of ACSF with 95% O2-5% CO2 was continued during hypoxia. Reoxygenation was started 20 s after HSD onset.
Solutions
Chemicals—unless otherwise mentioned—were obtained from Sigma-Aldrich. The ACSF contained (in mM) 130 NaCl, 3.5 KCl, 1.25 NaH2PO4, 24 NaHCO3, 1.2 CaCl2, 1.2 MgSO4, and 10 dextrose; aerated with 95% O2-5% CO2 to adjust pH to 7.4. Rotenone, diphenyleneiodonium (DPI, Tocris), and FCCP were dissolved in dimethylsulfoxide (DMSO) as 10–12.5 mM stocks and stored at 4°C. Antimycin A (MW 
532 g/mol), 3-NPA, and oligomycin ABC (MW 791 g/mol) (see Cho et al. 1997) were dissolved in absolute ethanol as 20 mM, 500 mM, and 10 mg/ml stocks, respectively and kept at –20°C. Rhodamine 123 was dissolved in absolute ethanol (20 mg/ml). Cyanide (NaCN) was prepared as aqueous 1 M stock and kept frozen; CN– containing solutions were prepared freshly shortly before each use. Aqueous stocks of NADP-disodium salt (20 mM, Boehringer Mannheim) and glucose-6-P dehydrogenase (ammonium sulfate suspension 5 mg/ml; Roche) were kept at 4°C. Hexokinase was purchased as ammonium sulfate suspension (1,000 U/1.4 ml). Final DMSO and EtOH concentrations of all solutions used in the experiments were 
0.2%, and control experiments verified that neither DMSO (0.1%) nor EtOH (0.2%) induced any significant changes in HSD parameters (Fig. 1B).
|
Extracellular DC potentials were recorded with a locally constructed DC potential amplifier using thin-walled borosilicate glass (GC150TF-10, Harvard Apparatus) microelectrodes. Electrodes were filled with ACSF, and their resistance was 5 M
. Optical recordings of the mitochondrial membrane potential (m), NADH/FAD autofluorescence and intrinsic optical signal (IOS), were performed with a fluorescence imaging system composed of a Xenon light source (Polychrome II; Till Photonics, Gräfelfing, Germany) and a highly sensitive CCD camera (Imago QE; PCO Imaging, Kelheim, Germany) attached to an upright microscope (Axiotech Vario; Zeiss, Oberkochen, Germany).
Changes in mitochondrial membrane potential were quantified using the fluorescent dye rhodamine 123 (Rh123) (Duchen 1999; Emaus et al. 1986). Rh123 was excited at 485 nm and fluorescence was recorded using a 505-nm beam splitter and a 535/35-nm band-pass filter. NADH and FAD were excited alternately at 360 nm (NADH) and 445 nm (FAD), and their autofluorescence was recorded using a 450-nm beam splitter and a 510/40-nm band-pass filter (Hepp et al. 2005). This beam splitter/emitter constellation absorbed a large part of NADH autofluorescence, but it allowed for the simultaneous detection of NADH and FAD autofluorescence. For Rh123 and NADH/FAD imaging experiments slices were submerged (33–35°C) and a x40, 0.8NA water-immersion objective (Achroplan, Zeiss) was used.
The characteristic IOS of spreading depression is an increase in light scattering that coincides with the extracellular DC potential shift of normoxic and hypoxic spreading depression (Aitken et al. 1999; Andrew et al. 1999; Basarsky et al. 1998; Kreisman et al. 2000; Müller and Somjen 1999). We measured the IOS by quantifying changes in light reflectance in the CA1 region. Interfaced slices were illuminated by white light at an angle of 40–45o. Images were taken at 2-s intervals and 0.2-s exposure time using a x5, 0.13-NA objective (Epiplan, Zeiss). Hypoxia- and drug-induced reflectance changes were visualized by off-line image subtraction and referred to an image taken before oxygen withdrawal (Müller and Somjen 1999). They are displayed in a 256 gray-scale mode covering a range of ±20% brightness change and were quantified in a rectangular region of interest in CA1 stratum radiatum close to the recording electrode. The spreading velocity of HSD was calculated from the progression of the wave front of the reflectance increase within st. radiatum of the CA1 region (the main direction of propagation is always parallel to the pyramidal cell layer). The area of the hippocampal formation invaded by HSD was determined at the height of the reflectance increase by counting those pixels whose brightness had increased by 
5%. Image processing and analysis was performed with Tillvision 4.0 (Till Photonics) and MetaMorph Off-line 6.1 (Universal Imaging).
Biochemical determination of cellular ATP levels
Cellular ATP levels were determined spectrophotometrically by quantifying the reduction of NADP in a coupled reaction with glucose-6-phospate dehydrogenase (Lamprecht and Trautschold 1974; Wilken et al. 2000)
 |
 |
). The supernatant was neutralized (pH 7–8) with KHCO3 (pure salt), centrifuged once again (Lamprecht and Trautschold 1974; Wilken et al. 2000), and frozen overnight.
The determination of ATP levels is based on the formation of NADPH2, which was detected as increased absorption at 366 nm (1101M photometer, Eppendorf Gerätebau). The 200-µl sample volume was added to a 0.5-ml quartz cuvette and supplemented with 200 µl test solution [composition: 91.5 mM triethanolamin-hydrochlorid (AppliChem), 1 mM NADP, 100 mM D-glucose, 20 mM MgCl2]. First, 1 µl 1:20 diluted stock of glucose-6-phosphate dehydrogenase (corresponds to 0.25 µg protein) was added to metabolize traces of glucose-6-P, and the reaction was then started by addition of 1 µl hexokinase (0.7 units) while NADPH2 absorption was measured continuously (Fig. 7A). NADPH2 concentrations were determined using an NADPH2 extinction coefficient of 
366 = 3.4 cm2/µMol (Lamprecht and Trautschold 1974). Due to the 1:1 stoichiometry they report directly the ATP content of the sample which was normalized to protein content (nmol ATP/µg protein).
|
The data were obtained from 109 rats (92 males, 17 females), using up to five slices from each brain and performing each experiment on at least three different rats. Obvious differences in the characteristic HSD parameters in male and female rats were not observed. Data are given as means ± standard deviations. Statistical significance of the observed changes was tested by a two-tailed, unpaired Student’s t-test using a significance level of 5%. In the case of paired observations, a one-sample t-test was used to compare normalized drug effects against pretreatment control conditions. Significant changes are marked by asterisks (*P < 0.05; **P < 0.01).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
Under control conditions, HSD occurred within 154.1 ± 59.7 s on oxygen withdrawal. The associated extracellular DC potential shift (Vo) had an amplitude of –16.4 ± 5.4 mV, and, measured at the half-amplitude level (Fig. 1A), a duration of 47.2 ± 11.3 s (n = 158). As shown previously, HSD can be induced repeatedly in a given slice, if oxygen is readmitted in time. Such repeated hypoxia hardly affects the characteristic HSD parameters: the time to HSD onset and the DC-potential amplitude tend to decrease and HSD duration tends to increase (Müller and Somjen 1998). For the first three HSD episodes, these changes are not statistically significant (Müller and Somjen 1998). Nevertheless, to account for these tendencies, any drug-induced changes in HSD parameters were referred to a control HSD induced in each slice before drug administration and were then statistically compared with the second HSD induced in either untreated or solvent (DMSO, EtOH)-treated control slices (control series data are shown in Fig. 1B).
As working concentrations for the various drugs, we chose those used in previous studies with acute brain tissue slices: 20 µM rotenone, 20–25 µM DPI, 20 µM AMC-A, 0.1–1 mM CN, 1 mM 3-NPA, 1 µM FCCP, and 10 µg/ml oligomycin (Allen et al. 2005; Müller et al. 2002; Schuchmann et al. 2000; Wei et al. 2004). Interfaced slices are exposed to flowing bathing solution on one surface only. Therefore to ensure sufficient diffusion into the tissue, drugs were applied for 20–25 min before HSD was induced.
Uncoupling of mitochondria by FCCP (1 µM) or block of the mitochondrial F0F1 ATPase by oligomycin (10 µg/ml, 12.5 µM) failed to trigger spontaneous spreading depression (SD) episodes. Inducing HSD by oxygen withdrawal during oligomycin treatment did not affect HSD at all (n = 8). In contrast, FCCP hastened the onset of HSD by 31.0 ± 11.6%, reduced its amplitude by 28.3 ± 15.5%, and decreased its duration by 28.0 ± 25.0% (n = 7, Fig. 1). In addition, the DC potential showed a transient positive overshoot on reoxygenation (Fig. 1, 
). Because the F0F1 ATPase may reverse its direction of operation in the presence of uncouplers and hydrolyze ATP to stabilize the mitochondrial membrane potential (Duchen 1999; Nicholls and Budd 2000), we combined oligomycin and FCCP. Even on inhibition of the ATP-synthase by oligomycin, however, FCCP still hastened HSD onset by 28.0 ± 15.4% (n = 7, Fig. 1). Alternatively, to account for the quite different molecular weights of oligomycin and FCCP (791 and 254 g/mol, respectively) and the likely differences in tissue penetration, we preincubated slices first for 15 min with 10 µg/ml oligomycin, followed by the combined application of oligomycin and FCCP for 20–25 min. Under these conditions, the onset of HSD was still hastened by 20.9 ± 12.8%. HSD amplitude was not affected and HSD duration was slightly reduced, as previously seen with FCCP alone (n = 5, Fig. 1B, data set Oligom, FCCP+Oligom).
Next we tested the effects of impaired mitochondrial respiration. The complex I blocker rotenone (20 µM) hastened HSD onset by 26.4 ± 19.3% and decreased HSD duration by 21.4 ± 18.9% (n = 10; Fig. 2). DPI (25 µM), a less specific complex I blocker, also hastened HSD onset by 16.7 ± 20.0% and decreased HSD duration by 16.5 ± 18.7% (n = 8). HSD amplitude was not significantly reduced by either rotenone or DPI, but both drugs induced a transient positive overshoot of the DC potential on reoxygenation (see Fig. 2A, ). The complex II inhibitor 3-NPA (1 mM) did not significantly affect the time to onset or the amplitude of HSD, but it increased HSD duration by 72.1 ± 66.4% (n = 11, Fig. 2). Blocking complex III by antimycin A (20 µM) slightly decreased the time to HSD onset, by 17.7 ± 19.3%, and shortened HSD duration by 24.3 ± 25.8%, but did not affect HSD amplitude (n = 8). Antimycin A also caused a transient positive shift of the DC potential on reoxygenation (Fig. 2A). Complex IV block by CN– (100 µM) hastened HSD onset by 27.8 ± 10% and increased HSD duration by 26.7 ± 33.3%; HSD amplitude was not affected (n = 6, Fig. 2). Combined application of rotenone and antimycin triggered spontaneous SD episodes in three of six slices tested; in the other three slices, HSD onset was hastened by 23.3 ± 9.6%, its duration was decreased by 24.8 ± 6.8%, and the amplitude was reduced by 21.6 ± 9.6% (n = 3). An overview of the effects induced by the various mitochondrial blockers is presented in Table 1.
|
|
|
|
; Andrew et al. 1999; Kreisman et al. 2000; Müller and Somjen 1999). Having the unique advantage of being detectable by a noninvasive approach, not requiring any fluorescent markers, and yielding two-dimensional spatiotemporal information, the IOS visualizes the site of SD ignition, its spreading velocity and range. The IOS signals of K+-induced normoxic SD and HSD have been shown earlier to be similar (Aitken et al. 1999), but SD induced by CN–, azide, and FCCP in the presence of oxygen turned out to differ from drug-free HSD (Fig. 4). During drug-free HSD the reflectance increased by 11.6 ± 3.3%, reached its maximum 44 s after HSD onset, and spread at a velocity of 6.5 ± 3.2 mm/min (n = 14). In the case of CN–-, azide-, and FCCP-evoked SD, the reflectance increase also coincided with the onset of the DC potential shift, but the peak intensities of 14.5 ± 4.9% (n = 7), 12.9 ± 3.0% (n = 8), and 21.8 ± 7.4% (n = 9) were reached more slowly, 122, 70, and 70 s after SD onset, respectively (Fig. 4C). The highest IOS intensity was seen with the FCCP-induced SD (Fig. 4C). The IOS of the drug-induced SDs recovered more slowly and only incompletely, especially in the case of CN–- and azide-induced SD, as were the associated DC potential shifts (Fig. 3). In the absence of drugs, when hypoxia was continued after HSD onset for an additional 5 min, the recovery of the IOS became similarly incomplete (data set hypoxia +5 min plotted in Fig. 4C). As can be seen from the slice pictures in Fig. 4B which illustrate the maximal extension of IOS during various SD episodes, the hippocampal area invaded by the IOS was larger during CN–- and azide- than during hypoxia-induced SD, whereas it did not differ in the case of FCCP-induced SD. The IOS propagation velocity of HSD, CN–-, azide-, and FCCP-induced SD did not differ (Fig. 4A).
Changes in mitochondrial membrane potential (m) and metabolism
The effects of the mitochondrial inhibitors on the characteristic HSD parameters were unexpected because some of the drugs hastened the onset of HSD, but at the same time, shortened its duration, i.e., improved the recovery from HSD, which involves normalization of membrane potentials and ion distributions and therefore requires ATP (Rosenthal and Sick 1992). Because the preparation was consistently re-oxygenated 20 s after HSD onset, the shortening of HSD duration cannot be attributed to the earlier HSD onset. As analyzed in 130 slices, HSD duration and the time to HSD onset are not correlated (correlation coefficient = 0.049, Fig. 5). Also the molecular weight of the drugs used, and the resulting differences in tissue penetration, cannot hold as a general explanation for the different effects on the characteristic HSD parameters. Therefore to understand the drug-induced modulation of HSD, we also analyzed the effects of the various drugs on mitochondrial membrane potential (m), mitochondrial metabolism (NADH/FAD autofluorescence) and ATP levels. An overview of all findings is presented in Table 1.
|
; Emaus et al. 1986; Müller et al. 2005). The isolated hippocampal formation was bulk loaded in 4–5 µg/ml Rh123 for 25–30 min, and the changes in Rh123 fluorescence resulting from mitochondrial inhibition were quantified for a region of interest in st. radiatum of the CA1 region (see Fig. 6A for summary). These experiments were performed in submerged slices (34°C) to improve the detection efficiency of fluorescence emission. Due to the faster drug-penetration in submerged slices, drugs were applied for 15 min; CN– acted rapidly and was applied for 5 min only. A complete recovery of Rh123 levels on washout could only observed for CN– (Fig. 6A) because most drug effects were only poorly reversible or Rh123 levels even kept increasing very slowly after drug withdrawal (rotenone and AMC-A, see Fig. 6A). A clear depolarization of mitochondria was induced by rotenone (22.8 ± 8.2%, n = 5), DPI (35.7 ± 7.9%, n = 5), antimycin A (15.4 ± 5.4%, n = 5), CN– (1 mM 34.3 ± 11.3% n = 5; 100 µM 12.2 ± 9.0% n = 6), and FCCP (35.3 ± 14.7%, n = 6). In contrast, 3-NPA and oligomycin caused only moderate changes in mitochondrial membrane potential (4.8 ± 2.0%, n = 6 and –2.6 ± 1.9%, n = 7, respectively, Fig. 6A).
|
; Mills and Jöbsis 1972; Shuttleworth et al. 2003) (Fig. 6B). Experiments were performed similarly to the Rh123 recordings—except that the recording of autofluorescence does not require dye loading. The two compounds contributing to cellular autofluorescence are reduced NADH and oxidized FAD. Therefore changes in mitochondrial substrate utilization cause opposite changes in these two measures. An increase in NADH levels and the corresponding decrease in FAD levels occurred in response to rotenone and CN–, indicating inhibition of mitochondria respiration (Fig. 6B); antimycin A clearly increased NADH levels but hardly affected FAD. As expected from mitochondrial uncoupling, FCCP decreased NADH and increased FAD levels, indicating stimulation of mitochondrial respiration. Against expectation, DPI also decreased NADH and increased FAD levels. Unclear is, why oligomycin slightly decreased NADH and increased FAD levels (Fig. 6B). Changes in cellular ATP levels
Cellular ATP levels may affect the time to HSD onset (Allen et al. 2005; Roberts and Sick 1992) and depletion of ATP—as mimicked by pharmacological inhibition of Na+/K+-ATPase—triggers SD despite the presence of oxygen (Balestrino et al. 1999). We found HSD onset to be hastened by 34.0 ± 15.2% in 4 mM glucose-containing ACSF (n = 4), whereas HSD did not occur in 30 mM glucose-containing ACSF during 10 min of severe hypoxia (n = 5; data not shown). It should be kept in mind, however, that such HSD postponement by high glucose levels seems—at least in part—to arise from pronounced lactate-formation, with the resulting more pronounced acidosis (Li and Siesjö 1997) depressing voltage-activated Na+ and high-voltage-activated Ca2+ currents (Somjen 2004; Tombaugh and Somjen 1996).
To decide whether our observed changes in HSD parameters—especially the hastened onset—could be a consequence of a severe reduction in ATP levels or even ATP depletion, we compared the ATP levels in slices kept in control solution (ACSF) with those slices exposed to the various mitochondrial inhibitors (Fig. 7). The cellular ATP level of untreated slices averaged 11.4 ± 3.8 nmol ATP/mg protein (n = 22), which corresponds to the values determined by others for hippocampal slices of adult rats by using the luciferin/luciferase bioluminescence assay (Galeffi et al. 2000) (13.3 ± 1.1 nmol ATP/mg protein) or HPLC (Paschen and Djuricic 1995; Riepe et al. 1997) (13.2 ± 1.1 and 9.4 ± 0.5 nmol/mg protein, respectively). The various mitochondrial drugs had only moderate effects on cellular ATP levels, and most importantly only the high concentration of 1 mM CN–-depleted cellular ATP. Compared with control slices, 20–25 min pretreatment of slices with rotenone (20 µM), or oligomycin pretreatment (15 min, 10 µg/ml) followed by combined application of oligomycin and FCCP (1 µM) significantly reduced cellular ATP levels to 6.4 ± 1.9 nmol/mg protein (n = 8) and 8.1 ± 1.6 nmol/mg protein (n = 6), which correspond to 56.1 and 71.1% of control ATP levels, respectively (Fig. 7B). Antimycin A (20 µM), FCCP (1 µM) or oligomycin (10 µg/ml) just showed a tendency to reduce ATP levels, which did however not reach statistical significance. DPI (25 µM), 3-NPA (1 mM), and CN– (100 µM) did not noticeably affect ATP levels (Fig. 7B). Cellular ATP was completely depleted, however, when slices were exposed to 1 mM CN– for 20–25 min. This was very probably the result of SD, which is induced within 4–5 min by this concentration of CN– (Fig. 3).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
Mitochondrial impairment does not necessarily decrease cellular ATP levels
Inhibition of any one of the respiratory complexes affects electron transfer within the entire mitochondrial respiratory chain (Berg et al. 2002; Nicholls and Budd 2000) and therefore may potentially interfere with mitochondrial metabolism and ATP production. Similarly, mitochondrial uncoupling by FCCP rapidly depolarizes mitochondria and potentially interferes with mitochondrial ATP synthesis that is driven by the proton gradient (proton motive force) across the inner mitochondrial membrane (Mitchell 1961). Inhibition of the ATP synthase by oligomycin directly targets mitochondrial ATP production yet without affecting the function of the respiratory chain. Although inhibition of the proximal parts of the respiratory chain (complexes I and II) can at least partly be compensated for, inhibition of the distal parts (complexes III and IV) cannot. Inhibition of complex I by, e.g., rotenone or DPI, still does allow FADH2 to be utilized and its electrons to be shuttled into the respiratory chain via complex II (Berg et al. 2002). Inhibition of succinate dehydrogenase in complex II by 3-NPA prevents the utilization of FADH2 but does not interfere with the utilization of NADH and the resulting flow of electrons from complex I to complex III via complex II (Berg et al. 2002). This may explain, why 3-NPA failed to hasten HSD onset and to markedly affect mitochondrial membrane potential, NADH/FAD levels, or cellular ATP content. The prolonged HSD duration in the presence of 3-NPA suggests that the recovery after reoxygenation was delayed in the presence of 3-NPA, but the underlying mechanisms are not clear. A desynchronization of the anoxic depolarization in single neurons is unlikely because this should also have reduced the HSD amplitude.
HSD onset was hastened only by those blockers that directly short-circuit the proton gradient (FCCP) or target the proton-pumping complexes I, III, and IV (rotenone, antimycin A, and CN–). Therefore one might suspect that the hastened HSD onset reflects accelerated ATP depletion because cellular ATP levels, determined by glucose content and the efficiency of anaerobic glycolysis, do affect the time to HSD onset (Allen et al. 2005; Roberts and Sick 1992). Of course ATP depletion will eventually occur during hypoxia, but as reviewed by Lipton, hippocampal slices (CA1 region, 36–37°C) exposed to anoxia for 2, 3, and 5 min still maintain 50, 60, and 35% of their ATP content, respectively (see Table 1 in Lipton 1999). A marked reduction in ATP levels occurs, however, as soon as HSD is ignited (see Mitani et al. 1994 and discussion by Somjen 2004, p. 353), all of which proves that HSD is usually triggered before cellular ATP is depleted. Unfortunately, a more accurate correlation of HSD ignition and cellular ATP levels is not available, obviously, because a continuous detection of ATP levels at cellular resolution is not yet possible.
Our ATP measurements show that none of the drugs that hastened HSD onset severely reduced cellular ATP levels before hypoxia was induced (Fig. 7B). If cellular ATP levels were affected at all, they were not reduced <56% of control ATP levels (rotenone incubation). The exception was 1 mM CN–, which triggered spontaneous SD episodes and thereby intensified ATP consumption. More importantly, the ATP measurements prove that not all the drugs that hastened HSD onset also reduced cellular ATP levels (Table 1). Application of 100 µM CN– caused the most pronounced hastening of HSD onset but did not affect ATP levels at all (Fig. 7B). In contrast, oligomycin showed a clear tendency to reduce cellular ATP content (71.9% of control levels) but did not affect HSD. Our ATP measurements report the ATP level at the time point where hypoxia would have been induced; ATP levels during hypoxia before HSD onset were not measured, and due to the short time to HSD onset—on average 2.7 min of hypoxia—this had been difficult to realize. Because the entire hippocampal slice was used for ATP determination and the different hippocampal subfields may have been affected by the drugs to a different degree, the decrease in ATP levels might have been underestimated. Yet in this case, the true reduction of ATP levels by oligomycin would have been even more pronounced, but still oligomycin did not affect HSD onset. The fact is that, for the various drugs tested, a consistent correlation of hastened HSD onset and cellular ATP levels was not observed.
Further arguments against ATP depletion as a general cause for the hastening of HSD onset is the observation that FCCP, rotenone, DPI, and antimycin A also shortened the duration of HSD, i.e., they accelerated the ATP-demanding recovery of CA1 neurons and the normalization of ion levels (Rosenthal and Sick 1992). If cellular ATP had been depleted, one would have expected a slow and incomplete recovery, similar to what was observed in the case of CN–- and azide-induced SDs or the continuation of hypoxia after HSD onset (Figs. 3 and 4C). Furthermore, the hastening of HSD onset by FCCP persisted in the presence of oligomycin—even when oligomycin treatment was started 15 min ahead of FCCP application—indicating that accelerated ATP depletion due to mitochondrial ATP consumption by reversed F0F1-ATPase (Duchen 1999; Nicholls and Budd 2000) cannot be its cause. Neither did the ATP levels markedly differ when FCCP and oligomycin were applied either alone or in combination (Fig. 7B). We therefore conclude that—rather than cellular ATP levels—other factors arising from mitochondrial depolarization/dysfunction seem to be involved in the facilitated generation of HSD.
The assay of ATP levels does not differentiate between mitochondrially and glycolytically derived ATP. Therefore it is possible that the impaired mitochondrial ATP production was in part compensated by increased glycolytic ATP production (Nicholls and Budd 2000). Such a compensation as well as the additional utilization of glycogen stores and phosphocreatine have recently been demonstrated for the CA1 region of rat hippocampal slices (Allen et al. 2005). These experiments of Allen and coworkers differ from those reported here, as juvenile rats, submerged slices, and a somewhat lower temperature (33–34°C) were used. Besides, oxygen-glucose deprivation (ischemia-like condition) was necessary to induce the anoxic depolarization (i.e., HSD). Due to these differences and the obviously lower metabolic demand of their slices, mitochondrial inhibition by hypoxia or 1 mM CN– was not sufficient to induce an SD in their experiments.
Changes in cytosolic redox state
Inhibiting mitochondrial respiration modulates cytosolic redox state by releasing reactive oxygen species (ROS) (Bindokas et al. 1996; Boveris and Chance 1973; Nicholls and Budd 2000) and by increasing cellular NADH levels (Foster et al. 2005; Mills and Jöbsis 1972; Schuchmann et al. 2001). Mitochondrial uncoupling by, e.g., FCCP activates mitochondrial respiration maximally (Duchen 1999), thereby decreasing NADH levels (Rex et al. 1999) and increasing ROS production (Bindokas et al. 1996). Accordingly, FCCP causes an oxidative shift in cytosolic redox state. Mimicking such oxidizing conditions by application of H2O2 or the sulfhydryl oxidizing agent DTNB (dithionitrobenzoic acid), we recently reported HSD onset to be postponed due to activation of BK channels (Hepp et al. 2005). In contrast, the intrinsic formation of ROS early during hypoxia is apparently not a major contributor to the ignition of HSD. Combined application of the radical scavengers ascorbic acid (1 mM) and trolox (0.75 mM) did not affect HSD at all (n = 9) (unpublished data F. J. Gerich and M. Müller). In support of our data, others also reported that the time to onset of the ischemia-induced anoxic depolarization in hippocampal CA1 neurons is not affected by the radical scavenger Mn(III)tetrakis(4-benzoic acid)porphyrin (MnTAP) (Allen et al. 2005). However, we found reducing conditions as mimicked by DTT (dithiothreitol) to favor the onset of HSD (Hepp et al. 2005), yet the very molecular targets involved still need to be identified. Therefore the reducing shift resulting from the inhibition of the respiratory chain might be considered a putative intracellular signal, mediated by, e.g., increased NADH, FADH2, and reduced glutathione levels.
Both anoxia and the induction of HSD (as well as normoxic SD) modify cellular NADH levels. Imaging tissue NADH changes during normoxic SD in gerbil cortex in vivo, Hashimoto and coworkers (Hashimoto et al. 2000) observed a biphasic NADH signal, consisting of an initial increase followed by a temporary decrease below baseline levels. Although the initial NADH increase could be verified biochemically, the secondary decrease could not, and they concluded that it rather reflects changes in cerebral blood flow that interfere with the detection of NADH autofluorescence (Hashimoto et al. 2000). However, when correcting for such changes in blood flow (by monitoring changes in tissue reflectance), an undershoot of the NADH baseline—referred to as mitochondrial hyperoxidation—is still found during electrical stimulation, normoxic spreading depression as well as seizures in in vivo cat cortex (Rosenthal and Somjen 1973). It therefore seems that besides a possible interference with hemoglobin-content and -oxygenation, changes in tissue oxygenation levels as a result of increased cerebral blood flow and altered O2 consumption contribute to the hyperoxidation of NADH (Rosenthal et al. 1995). Also, intracellular derangement disturbing mitochondrial metabolism possibly with a contribution of reactive oxygen species have been discussed (Perez-Pinzon et al. 1997; Rosenthal et al. 1995). By contrast, inducing HSD in vitro (acute hippocampal slices) causes a clear NADH increase already before HSD onset, and NADH reduction then becomes maximal as soon as HSD is triggered (Foster et al. 2005). Hyperoxidation of NADH does not occur unless anoxia is continued for several minutes after HSD onset. Under these conditions, tissue pO2 levels rise as a result of irreversible neuronal injury and the associated decreased oxygen consumption (Foster et al. 2005).
A clear correlation of HSD onset and reducing conditions, i.e., NADH/FAD levels, was not found in our experiments (see Table 1 for overview). Although HSD onset was hastened by FCCP, rotenone, DPI, antimycin, and CN–, a clear reducing shift, i.e., increase in NADH levels and/or decrease in FAD levels, was only induced by rotenone, antimycin A, and CN– (Fig. 6B). In contrast, FCCP and DPI decreased NADH and increased FAD levels, which should rather shift cytosolic redox balance to oxidizing conditions. However, we evaluated only two of the major redox couples within the cytosol. Glutathione, which may also modify sulfhydryl residues of redox-sensitive proteins (Dröge 2002; Lipton et al. 2002), has not been quantified so far. Due to the interaction of all these redox couples, using a general marker of cytosolic redox state would also be advantageous. Such markers have recently developed in the form of redox-sensitive fluorescent proteins (Hanson et al. 2004; Ostergaard et al. 2001), and they even allow for the continuous and dynamic detection of cytosolic redox changes, but they are not yet commercially available.
A clear correlation was found, however, between mitochondrial membrane potential and the time to HSD onset. Each drug that hastened HSD onset (rotenone, DPI, antimycin A, CN–, and FCCP) also caused a clear depolarization of mitochondria (Fig. 6A). Could it be the depolarization of mitochondria that facilitates the triggering of HSD onset? But how can the mitochondrial depolarization be transmitted to the plasmamembrane channels, and what messengers might be involved? Among others, mitochondrial depolarization is expected to reduce the Ca2+ sequestration by mitochondria (Nicholls 1978; von Lewinski and Keller 2005), and it may modulate cytosolic pH (Kaila et al. 1989).
DPI hastened HSD onset, just as the complex I blocker rotenone, but otherwise it behaved quite differently. ATP levels were not affected by DPI and the observed decrease in NADH and the increase in FAD levels do not match the DPI-mediated mitochondrial depolarization. The NADH/FAD response rather indicates a stimulation of mitochondrial respiration, possibly in response to the DPI-mediated mitochondrial depolarization. Because DPI does not only target complex I but also blocks various oxidases such as NADH- and xanthine oxidase as well as NO synthase (Li and Trush 1998; Stuehr et al. 1991), additional, mitochondria-independent mechanisms might have contributed.
Intrinsic optical signals
The intrinsic optical signals of the hypoxia- and drug-induced SD episodes were identical in terms of the direction of the reflectance changes, yet their intensities and time course differed (Fig. 4). The very source of the scattering increase is still unclear. According to our earlier studies, it depends on the occurrence of the electrical signs of HSD (Müller and Somjen 1998), is Cl–-dependent and sensitive to anion transport blockers. Nevertheless, cell swelling had to be excluded as a major source for the reflectance increase; possibly changes in cytoarchitecture such as swelling of mitochondria and other organelles might be involved (Aitken et al. 1999; Bahar et al. 2000; Müller and Somjen 1999). Because the most pronounced scattering increase occurs in the dendritic layers, others have suggested microstructural damage ("beading") of dendrites (Andrew et al. 1999). Such changes are, however, irreversible, and their occurrence may be mostly related to ischemic conditions or maintained hypoxia. Also the curvature of the surface of interfaced slices has been suggested to contribute (Kreisman et al. 1995). Fluoroacetate poisoning of glial cells increases the intensity of the IOS during HSD, suggesting that the generation of the scattering increase does not rely on viable glial cells (Müller and Somjen 1999). In contrast, glial poisoning by fluoroacetate or mitochondrial inhibition by rotenone were found to decrease the characteristic IOS—a scattering decrease—associated with evoked neuronal activity in hippocampal slices (Buchheim et al. 2005), once more pointing out the different nature of intrinsic optical signals associated with neuronal activity and HSD.
We have now demonstrated that the intrinsic optical changes during HSD persist in the presence of drugs targeting mitochondrial metabolism and membrane potential (FCCP, CN–, azide); this indicates that severe mitochondrial depolarization and/or failure of mitochondrial metabolism can be excluded as a major source for the generation of the IOS associated with HSD. To what degree mitochondrial swelling is affected by the tested drugs cannot be decided at present.
Concluding remarks
In conclusion, a clear correlation exists for the depolarization of mitochondria and the hastening of HSD onset, proving a crucial role of mitochondria during the early phase of hypoxia and the determination of HSD onset. The cellular levels of ATP, NADH, and FAD did not show a clear correlation with the drug-induced changes in HSD. Therefore neither ATP depletion nor a reducing shift due to NADH, FAD accumulation can serve as a general explanation for the hastening of HSD onset in response to mitochondrial inhibition. Whether these changes may contribute in the one or other case has to be clarified in further experiments at the single-cell level. In view of the central role of mitochondrial membrane potential, altered intracellular Ca2+ levels as a consequence of disturbed mitochondrial Ca2+ sequestration (Nicholls 1978; von Lewinski and Keller 2005), Ca2+ release from malfunctioning mitochondria (Schuchmann et al. 2000), and subtle changes in cytosolic pH due to the depolarizing mitochondria (Kaila et al. 1989) have to be considered as putative signaling mechanisms. Unveiling the very mechanisms involved will require further detailed investigations quantifying and correlating the early changes in mitochondrial membrane potential and its effects on other cytosolic redox couples such as reduced/oxidized glutathione.
|
GRANTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
|
FOOTNOTES |
|---|
Address for reprint requests and other correspondence: M. Müller, Zentrum Physiologie und Pathophysiologie, Universität Göttingen, Humboldtallee 23, D-37073 Göttingen, Germany (E-mail: mike{at}neuro-physiol.med.uni-goettingen.de)
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
Aitken PG, Tombaugh GC, Turner DA, and Somjen GG. Similar propagation of SD and hypoxic SD-like depolarization in rat hippocampus recorded optically and electrically. J Neurophysiol 80: 1514–1521, 1998.
Allen NJ, Karadottir R, and Attwell D. A preferential role for glycolysis in preventing the anoxic depolarization of rat hippocampal area CA1 pyramidal cells. J Neurosci 25: 848–859, 2005.
Andrew RD, Jarvis CR, and Obeidat AS. Potential sources of intrinsic optical signals imaged in live brain slices. Methods 18: 185–196, 1999.[CrossRef][ISI][Medline]
Bahar S, Fayuk D, Somjen GG, Aitken PG, and Turner DA. Mitochondrial and intrinsic optical signals imaged during hypoxia and spreading depression in rat hippocampal slices. J Neurophysiol 84: 311–324, 2000.
Balestrino M, Young J, and Aitken P. Block of (Na+,K+)ATPase with ouabain induces spreading depression-like depolarization in hippocampal slices. Brain Res 838: 37–44, 1999.[CrossRef][ISI][Medline]
Basarsky TA, Duffy SN, Andrew RD, and MacVicar BA. Imaging spreading depression and associated intracellular calcium waves in brain slices. J Neurosci 18: 7189–7199, 1998.
Berg JM, Tymoczko JL, and Stryer L. Biochemistry. New York: Freeman, 2002.
Bindokas VP, Jordan J, Lee CC, and Miller RJ. Superoxide production in rat hippocampal neurons: selective imaging with hydroethidine. J Neurosci 15: 1324–1336, 1996.
Boveris A and Chance B. The mitochondrial generation of hydrogen peroxide. General properties and effect of hyperbaric oxygen. Biochem J 134: 707–716, 1973.[ISI][Medline]
Bradford MM. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal Biochem 72: 248–254, 1976.[CrossRef][ISI][Medline]
Buchheim K, Wessel O, Siegmund H, Schuchmann S, and Meierkord H. Processes and components participating in the generation of intrinsic optical signal changes in vitro. Eur J Neurosci 22: 125–132, 2005.[Medline]
Chance B and Williams GR. Respiratory enzymes in oxidative phosphorylation. II. Difference spectra. J Biol Chem 217: 395–407, 1955.
Cho JH, Balasubramanyam M, Chernaya G, Gardner JP, Aviv A, Reeves JP, Dargis PG, and Christian EP. Oligomycin inhibits store-operated channels by a mechanism independent of its effects on mitochondrial ATP. Biochem J 324: 971–980, 1997.[Medline]
Cooper JM and Schapira AH. Mitochondrial dysfunction in neurodegeneration. J Bioenerg Biomembr 29: 175–183, 1997.[CrossRef][ISI][Medline]
Dröge W. Free radicals in the physiological control of cell function. Physiol Rev 82: 47–95, 2002.
Duchen MR. Contributions of mitochondria to animal physiology: from homeostatic sensor to calcium signalling and cell death. J Physiol 516: 1–17, 1999.
Emaus RK, Grunwald R, and Lemasters JJ. Rhodamine 123 as a probe of transmembrane potential in isolated rat-liver mitochondria: spectral and metabolic properties. Biochim Biophys Acta 850: 436–448, 1986.[Medline]
Foster KA, Beaver CJ, and Turner DA. Interaction between tissue oxygen tension and NADH imaging during synaptic stimulation and hypoxia in rat hippocampal slices. Neuroscience 132: 645–657, 2005.[CrossRef][ISI][Medline]
Galeffi F, Sinnar S, and Schwartz-Bloom RD. Diazepam promotes ATP recovery and prevents cytochrome c release in hippocampal slices after in vitro ischemia. J Neurochem 75: 1242–1249, 2000.[CrossRef][ISI][Medline]
Hanson GT, Aggeler R, Oglesbee D, Cannon M, Capaldi RA, Tsien RY, and Remington SJ. Investigating mitochondrial redox potential with redox-sensitive green fluorescent protein indicators. J Biol Chem 279: 13044–13053, 2004.
Hashimoto M, Takeda Y, Sato T, Kawahara H, Nagano O, and Hirakawa M. Dynamic changes of NADH fluorescence images and NADH content during spreading depression in the cerebral cortex of gerbils. Brain Res 872: 294–300, 2000.
, duration of hypoxia. B: summary of the effects of respiratory chain inhibition on HSD. *, significant changes as compared with the 2nd HSD induced in an untreated control slice, or if applicable, in the presence of the respective solvent used (see 
