- Email: [email protected]
Brain electrical activity during combined hypoxemia and hypoperfusion in anesthetized rats
Respiration Physiology 129 (2002) 375– 384 www.elsevier.com/locate/resphysiol
Brain electrical activity during combined hypoxemia and hypoperfusion in anesthetized rats B. Wuyam a, V. Bourlier b, J.L. Pe´pin a, J.F. Payen b, P. Le´vy a,* a
PRETA-TIMC, UMR CNRS 5525, Laboratoire de Physiologie, 38700 La Tronche, France b U INSERM U 438, CHU de Grenoble, B.P. 217X, 38043 Grenoble Ce´dex, France Accepted 18 September 2001
Abstract In order to investigate the effects of moderate hypoxemia on brain electrical activity and the consequences of an altered cerebro-vascular response to hypoxemia, we recorded changes in electrical activity of the brain in anesthetized rats following unilateral carotid artery ligation (UCAL). In these animals, on the clamped side, cerebral blood flow, whilst normal during normoxia, shows less augmentation during hypoxemia. Six anesthetized (Halothane) Sprague– Dawley rats with UCAL were studied during 20 min periods of baseline (FIO2 = 30%), hypoxemia (FIO2 = 9.5%) and recovery (FIO2 =30%): mean arterial pressure of oxygen (PAO2) achieved was 177.0, 37.6 and 160.1 mmHg, respectively. A significant decrease in the frequencies of the ECoG was observed bilaterally during hypoxemia: centroid frequency (fc) = 3.3790.14 and 2.85 90.13 Hz on the intact and clamped hemisphere respectively during hypoxemia versus fc = 4.0990.20 Hz (mean 9S.E.M.) during baseline, which was not reversed during recovery (3.279 0.11 Hz) (ANOVA, PB0.01). The total power of the signal (Pw) was unaffected on the intact hemisphere but diminished on the clamped side during hypoxemia. Our results show that a significant slowing of ECoG is observed during hypoxemia of moderate intensity (40 mmHg) even when cerebro-vascular response to hypoxemia is preserved and that total power of the ECoG signal is severely diminished when the cerebro-vascular response to hypoxemia is impaired. © 2002 Elsevier Science B.V. All rights reserved. Keywords: Carotid artery, ligation; Electrical activity, brain; Hypoxia, brain; Mammals, rat; Perfusion, brain hypoperfusion
1. Introduction In humans, hypoxic encephalopathy is observed in pilots or climbers exposed to high altitude and * Corresponding author: Present address: Laboratoire EFCR, Hoˆpital A. Michallon, B.P. 217 X, 38043, Grenoble Ce´dex, France. Tel.: +33-4-7676-5516; fax: +33-4-7676-5617. E-mail address: [email protected] (P. Le´vy).
in patients with cardio-respiratory diseases. Reduced oxygenation (FIO2 = 12%, PAO = 50 mmHg) may lead to progressive impairment of cerebral function, such as impaired retinal function, followed by impairment of short-term memory and cognitive functions. Finally, severe obtundation, ataxia, stupor and coma may be observed at altitudes higher than 5000 m (FIO2 = 10%, PAO = 35 mmHg) (Novotny, 1996). 2
2
0034-5687/02/$ - see front matter © 2002 Elsevier Science B.V. All rights reserved. PII: S 0 0 3 4 - 5 6 8 7 ( 0 1 ) 0 0 3 1 9 - X
376
B. Wuyam et al. / Respiration Physiology 129 (2002) 375–384
The precise mechanisms leading to the loss of brain function have been investigated in animals and humans and may depend on the severity of hypoxemia. Severe hypoxemia (PAO B25 mmHg) is accompanied by the rapid reduction in high-energy compounds (Sjesjo¨ et al., 1974) that is associated with a rapid loss of function. In more moderate levels of hypoxemia, ATP levels remain fairly constant (Sjesjo¨ , 1978) and other factors may be involved. These include: changes in membrane excitability of neurons, changes in neurotransmission or synaptic failure (Donelly et al., 1992; Gibson et al., 1981; Novotny, 1996). On the other hand, several mechanisms limit the consequences of reduced blood oxygen content. One is the increase in cerebral blood flow, which accompanies hypoxemia below 55 mmHg (Kety and Schmidt, 1948; Cohen et al., 1967; Heistad and Kontos, 1983; Shockley and LaMamma, 1988; Bereckzi et al., 1993). Furthermore, metabolic adaptations which may help to preserve energetic conditions within the brain, have been observed as an early response to cerebral hypoxia (Brehar et al., 1983). Investigations by our group (Payen et al., 1996) and others (Allen et al., 1992) suggest that early lactate accumulation may occur in response to moderate hypoxemia (PAO of 50 mmHg). Lactate is now seen as an important source of energy for neurons (Schurr et al., 1988) which may thus participate in cellular adaptation to hypoxemia. The present investigation was undertaken to assess functional correlates of such metabolic changes in rats with unilateral carotid artery ligation (UCAL). We investigated whether electrical activity of the cerebral cortex was preserved with a reduced oxygen delivery on both an intact hemisphere—where cerebro-vascular adaptation is preserved—and a clamped hemisphere. In the latter, cerebral blood flow is normal at rest due to blood flow supplementation through the Willis polygon (Levine, 1960) but fails to increase in response to hypoxemia and/or hypercapnia (De Ley et al., 1985; Salford and Sjesjo¨ , 1974).
2. Methods
2.1. Electrocorticogram electrodes placement
2
2
Electrodes were placed in OFA (Iffa Credo) female rats, 240–260 g in weight, under general anesthesia (intraperitoneal chloral hydrate, 4%). After incision of the skull, the animals were placed in a stereotactic frame in ventral decubitus position and two sets of unipolar stainless steel electrodes were placed symetrically in front of the parietal cortex at the outer surface of the dura. Electrodes were 5 mm lateral and 2 mm anterior to bregma. The rats were allowed to recover from general anesthesia and the ability of the electrodes to record valid ECoG was controlled during a period of quiet wakefulness of the animal. Recovery from anaesthesia and the normal behavior of the animal was confirmed The hypoxic experiments were performed within 1 week (2–8 days) following electrode placement procedures.
2.2. Surgical procedure Animals with pre-implanted electrodes were initially anaesthetized with halothane (5–7%) delivered via a facemask in a cage. The animals were tracheostomized and mechanically ventilated using a Edco/NEMI, type 804-small animals-respirator. Anesthesia was maintained with 5% inhaled halothane. The left carotid artery was clamped and the caudal end of the vessel was canulated to allow continuous blood pressure monitoring using a pressure transducer and recorder (Gould 8188S). Additionally, arterial blood samples were collected via the arterial canulation for blood gas analysis (Radiometer, ABL 510, Copenhagen, Danemark). Rectal temperature was continuously monitored and maintained constant (T= 37 °C9 0.5). After surgical preparation, the animals were allowed to recover for one hour during which they were maintained under light anesthesia (0.5–1.5% halothane). The same level of anesthesia was pursued for the rest of the experiment.
B. Wuyam et al. / Respiration Physiology 129 (2002) 375–384
2.3. Protocol A baseline recording of brain electrical activity was obtained during 10 min with the animal ventilated at an inspired fraction of oxygen (FIO2) of 30%. This was followed by a period of hypoxia obtained by reduction of FIO2 from 30 to 9.5% over 2–3 min and maintained for a total duration of 20 min. This was followed by a period of recovery during which FIO2 was returned progressively to 30% over 2– 3 min and maintained for a total duration of 20 min. Blood gases were determined at the end of each period of baseline, hypoxia and recovery. Since hypoxemia has a pronounced peripheral vasodilatory action in rats, an hypotensive response was observed in all animals. Care was taken to minimize the degree of hypotension. Female animals were chosen because they appeared to be less sensitive to hypoxic hypotension than males. Small doses of Dopamine (0.06 mg/ml) were infused intraperitoneally at a rate of 1– 2 ml/h throughout the experiment. This dose has minimal cerebral vasodilatory action and helped to stabilize blood pressure during the hypoxic exposure. Tidal volume and frequency of the respirator were adjusted to maintain eucapnia (target PACO2 of 35 mmHg). The FIO2 was continuously monitored (MiniOx 1, Catalyst Research Corporation).
377
band of the electrocorticogram: delta (1–4 Hz), theta (4– 8 Hz), alpha (8–12 Hz), beta (12–16 Hz) and sigma (16– 20 Hz). Signal power was expressed both as a fraction of the total signal and in absolute terms. The sum of B 1 Hz plus delta waves was referred to as the ‘percentage of low frequencies’ (0.5–4 Hz).
2.5. Statistics A three-way analysis of variance (two withinfactor: (i) condition i.e. baseline, hypoxia or recovery and (ii) time: min 1–10; one between-factor: hemisphere) was performed for the parameters of the ECoG signal: centroı¨d frequency and global power spectrum. Single parameters like blood pressure, or PAO2 and PACO2 were analyzed using one-way ANOVA for repeated measures. P values B 0.05 were considered significant.
3. Results
3.1. Biological 6ariables during hypoxemia Changes in FIO2 and the resulting changes in mean arterial blood pressure (MABP) are shown
2.4. Data acquisition and analysis Electrical signals of brain activity were continuously recorded, amplified and filtered. The filter time constant was set at 0.3 sec. This is equivalent to a ‘high-pass’ filter of 0.5 Hz. Signals were simultaneously displayed on an ECEM analogic recorder and stored on magnetic tape (Hewlett– Packard, 0–1 kHz) for further analysis. Recorded signals were digitized at a sampling frequency of 128 Hz (AD converter MacAdios). Fast Fourier Transform of the recorded signal was performed. The following parameters were derived: (i) centroid frequency (fc) (or median frequency, i.e. the frequency dividing the spectrum in two parts of equal power); (ii) total power (Pw) of the signal defined as the sum of squares of the Fourier series; and (iii) the power corresponding to each
Fig. 1. Changes in mean (S.D.) arterial blood pressure (black circles) observed at the corresponding levels of inspired fraction of oxygen (FIO2, white circles).
378
B. Wuyam et al. / Respiration Physiology 129 (2002) 375–384
end of hypoxia (group mean pH 7.359 0.04, vs. 7.439 0.04 during baseline), (one-way ANOVA, P= 0.02), which was not reversed after 20 min at an FIO2 of 30% (pH 7.369 0.03).
3.2. Electical acti6ity during baseline
Fig. 2. Changes in mean centroid frequency on the intact (white circles) and clamped hemisphere (black circles) associated with changes in inspired oxygen fraction (FIO2). Bars refer to standard error of the mean (S.E.M.).
Compared with ECoG during (i.e. without anesthesia), Mean ECoG was significantly slower during baseline (under light anesthesia), as compared with wakefulness: group mean fc of 4.099 0.08 Hz during baseline versus 5.549 0.07 Hz during unanesthetized quiet wakefulness (m9 S.E.M., PB 0.05). Mean fc was identical between hemispheres during baseline (mean fc= 4.049 0.06 vs. 4.169 0.07 Hz on the intact and clamped hemisphere, respectively). The change in fc induced by anesthesia was similar on the clamped and intact hemisphere (ANOVA, F=0.48, P= 0.88).
3.3. Electrical acti6ity during hypoxemia in Fig. 1. After a gradual decrease over 7 min (min 13–20, Fig. 1), a hypoxic gas mixture of 9.59 0.3% FIO2 (m9S.D.) was steadily inhaled for at least 10 min (min 20– 30 served as the reference for ‘steady-state’ hypoxemia). MABP decreased significantly during hypoxemia (group MABP of 61 913 vs. 10297 and 117 9 19 mmHg during baseline and steady recovery (min 40–50), respectively (one-way ANOVA, P B 0.05). The delay between the changes in FIO2 and changes in blood pressure was less than one minute in all animals. After intitial 40% decrease, a stabilization of the level of blood pressure was obtained until the end of the hypoxic exposure (Fig. 1). Group mean PAO2 achieved during the hypoxic period was 37.696.7 mmHg (from 33.8 to 50.2 mmHg). Eucapnia was maintained throughout the experiment for all the animals. Group mean PACO2 obtained was 32.894.6, 33.79 5.5 and 30.69 7.9 mmHg during baseline, hypoxia and recovery respectively (m9 S.D.), which were not significantly different (ANOVA, P = 0.38). Moderate extra-cellular acidosis was observed at the
Changes in electrical frequency of the ECoG during baseline, hypoxemia and recovery were analyzed in two ways: (1) during ‘steady-state’ conditions (min 20–30) where a stable level of FIO2 was maintained; (2) as a ‘transient’ response corresponding to the initial periods of hypoxia (min 10–20) and recovery (min 30–40) which were analyzed separately. Steady-state and transient periods were defined according to the stability of the FIO2 to which the animals were exposed.
3.4. Steady-state response Group mean changes in electrical activity during hypoxemia on both hemispheres are shown in Fig. 2. A significant slowing of the electrical activity was observed bilaterally during steady-state hypoxemia. Analysis of variance of the changes in centroid frequency (fc) indicated a significant difference between the three conditions of baseline (group mean fc= 4.099 0.08 Hz, m9 S.E.M.), hypoxemia (group mean fc= 3.1190.36 Hz) and recovery (group mean fc= 3.249 0.14 Hz;
B. Wuyam et al. / Respiration Physiology 129 (2002) 375–384
ANOVA, P B0.001). Moreover, a significant interaction between the condition and time (‘repeated measures’) was observed irrespective of the hemisphere (ANOVA, F =1.904, P B0.05). This suggests that the evolution of centroid frequency (fc) with time was mainly attributable to hypoxemia. Although slowing of the centroid frequency, particuliarely at the beginning of the hypoxemic period, tended to be more pronounced (see Fig. 2) in the clamped (mean group fc=2.85 9 0.39 Hz) than in the unclamped hemisphere (mean fc= 3.3790.19 Hz), this did not reach statistical significance (ANOVA, F =0.748, P =0.68). In absolute terms (Fig. 3), on the intact hemisphere, a shift in the frequencies of the ECoG was observed with reduction in the power of the highest frequencies (essentially 4– 8 Hz band), and a parallel increase in low frequencies: delta wave
379
(1– 4 Hz). The appearance of very low frequencies (B 1 Hz) was also noted on this side. These changes remained essentially stable during recovery (Fig. 3). Conversely, an absolute reduction in all frequencies was observed during hypoxemia on the clamped (left) hemisphere, which was greater in the ‘high’ frequencies (4–8 Hz) but also affected low (1–4 Hz) frequencies. This was partially reversed during recovery (Fig. 3). Overall, changes in the total power of the ECoG observed during hypoxemia were markedly different between the two hemispheres. As shown in Fig. 3, an important reduction in total power of the signal was observed on the clamped (left) hemisphere during hypoxemia (group mean: Pw = 9219 36, 3769 38 and 5949 94 mV2 during the baseline, hypoxemia and recovery periods, respectively, m9 S.E.M.) which was not entirely
Fig. 3. Changes in mean total signal power (black circles) and power in each frequency bands (0 – 1 Hz: white circle; 1 – 4 Hz: white polygon; 4– 8 Hz: black polygon; 8 –12 Hz: white triangle; 12 – 16 Hz: black triangles; 16 – 20 Hz: white squares) on the intact (top pannel) and clamped hemisphere (bottom). Note the global stability in the total signal power on the unclamped side.
380
B. Wuyam et al. / Respiration Physiology 129 (2002) 375–384
reversed after 20 min. Conversely, total power remained stable on the unclamped right hemisphere (760929, 767937 and 828930 mV2 during baseline, hypoxia and recovery, respectively) Statistical interaction between the hemisphere and the condition did not reach significance, however (ANOVA, F = 2.549, P = 0.10 for the betweenhemisphere difference).
3.5. Transient response A significant delay between the changes in FIO2 and the associated changes in electrical activity was observed. Fig. 3 indicates that the delay, between the initial drop in FIO2 below 21% (beginning of hypoxia) and the beginning of changes in the median frequencies, was at least 2 min. At the end of the hypoxic period and during recovery, overall slowing of ECoG (as reflected by changes in the median frequency) tended to be similar on both hemispheres. The rate of reduction in fc at the onset of hypoxemia was not different between hemispheres (mean slope calculated between the initial decrease and the minimal value (nadir) of fc: −0.2390.17 Hz/min on the unclamped versus −0.2890.27 Hz/min on the clamped side, paired t-test, P=0.58).
4. Discussion The main finding of this study is that electrical activity of the cortex is significantly and steadily modified within a few min of exposure to hypoxemia of moderate depth (40 mmHg), whether or not the cerebro-vascular response to hypoxemia is normal. This response is mainly characterized by a slowing of the frequencies observed. Impeded CBF response to hypoxemia was further associated with a severe and global reduction of the electrical power.
4.1. Experimental conditions These effects of acute hypoxemia have been observed in anesthetized tracheotomized animals with unilateral (left) carotid artery clamping. In such a preparation, CBF is minimally affected
during normoxia (due to blood flow supplementation through the Willis polygon) without detectable metabolic alterations (Hoffman et al., 1983; Salford and Sjesjo¨ , 1974; Payen et al., 1996). In our experiments, electrical activity (ECoG) was found to be identical (no significant difference in mean fc, % of low frequency and power spectrum) in both hemisphere, during normoxia. This tends to rule out the possibility that the surgical procedure per se may have affected neuronal viability and/or function. This is presumably due to the particular development of the Willis polygon in rats. We maintained successfully a constant PACO2 during the three conditions of baseline, hypoxemia and recovery. This ensured similar conditions of cerebral vasoreactivity to CO2 (Heistad and Kontos, 1983) between the three conditions and separated the effects of hypoxemia per se from those of combined hypoxemia and hypocapnia as observed in humans. This is of importance since hypocapnia, which accompanies the normal ventilatory response to hypoxemia, has been shown to add detrimental effects on cerebral function (Kraaier et al., 1988) and may lead to a significantly greater increase in the low frequencies of the EEG, than a pure hypoxic challenge (Borgstro¨ m et al., 1975; Van der Worpe et al., 1991). We believe that the cerebral blood flow response to hypoxemia was asymmetrical across hemispheres in our protocol. Although CBF is near normal at rest in such animal preparations, the CBF response to vasodilatory stimuli has been proven to be asymmetrical when one of the carotid arteries is clamped. In a study by De Ley et al., (1985) using a similar animal experiment with unilateral clamping, cerebral blood flow (CBF) was normal during normoxia but failed to increase in response to hypercapnia on the clamped side. Similarly, in response to hypoxemia, a reduced increase in CBF was observed on the clamped side in two separate experiments (Hoffman et al., 1983; Salford and Sjesjo¨ , 1974). Blood flow on the clamped side represented only 50% of the controlateral hemisphere. Using Doppler laser flow measurement, in similar hypoxic conditions, we have observed that CBF
B. Wuyam et al. / Respiration Physiology 129 (2002) 375–384
velocity response to hypoxemia was markedly different in each hemisphere: CBF velocity increased to 120% of the baseline value on the unclamped hemisphere and was reduced to 55% of baseline CBF velocity on the clamped hemisphere (Wuyam et al., 1998). We thus suggest that brain hemispheres were exposed to the same partial pressure of oxygen (PAO2) but presumably different O2 supply to brain cells as a consequence of differences in O2 delivery. The mechanisms of such an asymmetrical CBF response cannot be clarified from the present experiments. This imbalance could be either due to a steel phenomenon of blood flow by the intact side or to the loss of the autoregulation in response to hypotension when the cerebral circulation on one side relies on supplementation by the Willis polygon.
4.2. Changes in ECoG acti6ity in conditions of hypoxemia It has long been recognized that oxygen deprivation produces changes in neuronal electrical activity both in animals and humans. Slowing in spontaneous electrical activity (EEG) occurs within seconds of brain ischemia with subsequent depression of cortical evoked potentials and, ultimately, loss of function (Prior, 1985). These effects, however, combine the effects of complete oxygen deprivation and the absence of blood metabolites. There is clear evidence that neuronal response may differ between hypoxemia itself and hypoxemia combined with ‘chemical’ ischemia (Haddad and Jiang, 1992). The proper effects of hypoxic hypoxia are less well established, in particular those associated with mild to moderate reduction in O2 availability in vivo. Several reports have shown a slowing of the electrical activity associated with short-term exposure to hypoxic hypoxia in normal men (Gurvitch and Ginsburg, 1977; Kraaier et al., 1988; Rossen et al., 1961). Significant hyperventilation, however, leading to a significant degree of hypocapnia may contribute to the changes observed (Van der Worpe et al., 1991), especially after recent exposure to hypoxemia. In patients with chronic obstructive pulmonary disease (COPD) and permanent hypoxemia, Brezinova et al. (1979)
381
have reported a mean EEG frequency significantly closer to normal in patients receiving long term oxygen therapy as opposed to untreated pateints. The mean frequency of the EEG was correlated with arterial oxygen saturation (SAO2) in these patients. However, acute administration of oxygen did not significantly change the EEG frequencies in either group. This study suggests that chronic oxygen deprivation in humans may be associated with a significant slowing of the EEG which is reversed when hypoxemia is chronically (but not acutely) relieved. Block et al. (1974) have found similar findings with a partial change in the EEG with long-term oxygen therapy, which may suggest reversible neuronal injury or dysfunction during oxygen deprivation. EEG abnormalities have, similarly, been observed in elite mountain climbers 2–9 months after a sojourn above 8000 m without supplemental O2 (Cavaletti and Tredici, 1993). Similar changes in neuronal excitability could contribute to neuropsychological and/or memory impairment reported in these situations (Cavaletti and Tredici, 1993). However, both hypoxemia per se and/or changes in CBF may have contributed to the changes in electrical activity. Since neocortical neurons of rats respond similarly to those of humans when submitted to hypoxia (Haddad and Jiang, 1992), we believe that our experiments provide some valuable insight on this question. Our observations indicate that a significant slowing of the ECoG occurred on the intact hemisphere for a steady-state PAO of around 40 mmHg. We believe that the changes in frequency observed in our experiments may be due, at least in part, to changes in synaptic transmission and/ or neuronal excitability. Although the possibility exists that structures of the central nervous system (e.g. corpus callosum) may indirectly influence cortical structures through anatomical projections, there is abundant and direct evidence that neocortical neurons themselves can be affected by hypoxia (for example, see Haddad and Jiang, 1992). In tissue slice preparations, in response to hypoxic exposure, an initial hyperpolarization is observed, preceding a progressive return to prehypoxic resting membrane potentials. This hyperpolarisation can be maintained for prolonged 2
B. Wuyam et al. / Respiration Physiology 129 (2002) 375–384
382
periods of time (5– 20 min). Along with early hyperpolarization, synaptic potentials are markedly suppressed, the threshold for action potential is increased, and electrical input resistance decreased. Such changes are suggestive of changes in K+ conductance. Such changes in membrane excitability could contribute to the overall slowing of electrical activity recorded at the cortical surface in our experiments (temporal summation) and increase low frequencies recorded in the ECoG. Factors other than global membrane excitability and spontaneous discharge frequency could additionally contribute to a global slowing of the electrical activity. For example, changes in the morphology of action potentials of central neurons may be involved (Haddad and Jiang, 1992). Cummins et al. (1991) showed that, in response to brief exposures to hypoxia (i.e. pure N2 in voltage-clamp preparations), central hippocampal neurons become hyperpolarized, due to fast-inactivation of Na+ current. (Cummins et al., 1991). Such a change in Na+ channels could lead to a change in the kinetics of depolarization in neurons and contribute to an overall slowing of cortical networks. The extent to which each of these mechanisms contributes to ECoG frequency slowing is unknown. It should be emphasized, however, that changes in ECoG observed in our experiments in presence of a normal CBF response (intact hemisphere) are in accordance with biological effects of hypoxia (such as change in K conductance and inactivation of Na+ channels) on neurons themselves on tissular slice preparations. Furthermore, this effect has been observed following similar hypoxia (PO as high as 50 mmHg at the neuronal surface in slice preparation, (Jiand et al., 1991). This suggests that significant changes in neuronal excitability can be observed with relatively modest reductions in PO , unlike the classical notion that O2 deficit occurs below a PAO of 25 mmHg (Hochachka et al., 1996). In the central nervous system, however, alternative mechanisms may be implicated in the down regulation of firing rates and synaptic transmission (‘spike arrest’) (Hochachka et al., 1996). Adenosine-mediated down-regulation of excitatory amino acids (especially glutamate) with a concomitant increase in inhibitory amino acids, are likely to play a role. 2
2
2
The extent to which reduced CBF response may add detrimental effects on cortical activity is clearly shown in our experiments. The total power of the ECoG was significantly reduced for all frequencies on the clamped hemisphere, as opposed to the normal (intact) hemisphere where the increase in the power of low frequencies compensated the decrease in the power of high frequencies. Total power equals the square of the amplitudes of each frequency recorded in the spectrum. Total power is thus a parameter which reflects the global amount of electrical activity produced by cortical neurons in the region of interest. We suggest that this between-hemisphere differences in Total Power may be attributed to the differences in O2 delivery, i.e. differences in CBF response to the same reduction of O2 content. Close relationships between the magnitude in oxygen delivery (i.e. the amount of oxygen actually delivered to neurons) and the reduction in the overall ‘power’ of the EEG signal, has previously been suggested by Nioka et al. (1990) in dogs. In this study, progressive cerebral hypoxia was induced. The variations of O2 availability to neurons, and the metabolic and oxydative phosphorylation state of the brain tissue were monitored. Electrical activity of the EEG was analyzed. The ‘amplitude’ of the signal was reported (unlike the changes in the frequency bands of the spectrum). In this experiment (Nioka et al., 1990), changes in the amplitude of electrical activity strictly paralleled changes in O2 availability. When the CBF response was insufficient to compensate for the decrease in O2 content (reduced O2 availability), the O2 consumption of the brain tissue decreased. High energy compounds (PCr) were modified and glycolysis was stimulated (drop in intracellular pH). The calculated PO felt within the range of Km for cytochrome oxidase, which suggests that the kinetics of the ATP production may have been affected. In this situation, EEG amplitude fell below the EEG amplitude in control normoxic condition. Conversely, at lower grades of hypoxia, when the CBF was adequate to compensate a smaller reduction in arterial O2 content, the O2 consumption of the brain tissue was unaffected, cerebral metabolism unchanged, calculated PO was well above the Km for cy2
2
B. Wuyam et al. / Respiration Physiology 129 (2002) 375–384
tochrome oxidase, and electrical activity was preserved. Thus, using a very similar analysis of the ECoG, our results largely agree with the observations of Nioka et al. (1990) suggesting that the power of the EEG signal may be more closely related to the reduction in O2 availability in neurons than frequency analysis per se. Both experiments confirm that the maintenance of O2 delivery by an adequate CBF increase is essential for the preservation of global electrical activity. Changes in electrical activity are precipitously observed within 2–3 min of exposure, even to moderate hypoxemia. Recent evidence suggests that a common oxygen-sensing mechanism exists in virtually all cell types and can act within seconds of hypoxic exposure, i.e. independently of gene transcription or translation to a second messenger (Bunn and Poyton, 1996). However, there is no consensus on the nature of the sensor [which could be either a heme protein at the outer surface of the cellular membrane or an intracellular pathway such as the NADPH oxidase system or even the mitochondria themselves (Chandel and Schumaker, 2000)]. Whatever the sensor, the production of reactive oxygen intermediates is susceptible to provide a chemical link between intracellular oxygen concentration and the changes in proteins activity. If such a mechanism is susceptible to modulate ion channels activity in neurons, it would provide a possible explanation for the kinetics of changes in electrical activity observed in our preparation particularly on the clamped side. The link between functional deficit of central neurons (e.g. synaptic transmission in hippocampal slices) and the production of free radicals within these cells has been recently shown in glutathion peroxydase transgenic mice (Furling et al., 2000). A moderate increase in glutathion peroxydase (a free radical scavenger) observed in these animals is able to preserve synaptic transmission after short-term (10 min) hypoxia. Accordingly, the absence of change in electrical activity during re-oxygenation sould be pointed out in our experiments. This could be due either to the persistance of free radical production— which have also been implicated in the pathways of hypoxia/reoxygenation injury— or to the persitance for several hours of stress-activated kinases,
383
which may mediate cellular response to hypoxemia. Our data, in anesthetized rats, thus, indicate that relative hypoperfusion associated with hypoxemia adds an adverse effect based on ECoG. Total power of the signal, which represents global electrical activity, is significantly reduced. This is not the case when CBF response to hypoxemia is preserved (in the same animals). A slowing of the frequencies of the ECoG content is present in this case. A balanced change in the frequency composition of ECoG (i.e. reduction in high frequencies and increase in low frequencies) occurs at moderate levels of hypoxemia and are in keeping with known intracellular changes in excitability in response to an O2 sensing mechanism. The effect of the vascular response to hypoxemia on the global amount of cortical electrical activity suggests important clinical consequences in patients with impaired cerebral circulation (OSA patients) or during adaptation to high altitude. Additional studies using O2 free radicals modifiers (scavengers or transgenic animals) in our paradigm would give further insight on the mechanisms and cerebral consequences of hypoxemia.
Acknowledgements This work was supported by a grant from Re´ gion Rhoˆ ne-Alpes (Hypoxie). The authors are additionnaly grateful to Hamid Benzzouz for technical support.
References Allen, K., Busza, A.L., Cockard, H.A., Gadian, D.G., 1992. Brain metabolism and blood flow in acute cerebral hypoxia studied by MR spectroscopy and hydrogen clearance. NMR Biomed. 5, 48– 52. Bereckzi, D., Wei, L., Otsuka, T., Acuff, V., Pettigrew, K., Patlak, C., Fenstermacher, J., 1993. Hypoxia increases velocity of blood flow through parenchymal microvascular systems in rat brain. J. Cereb. Blood Flow Metab. 13, 475 – 486. Block, A.J., Castle, J.R., Keitt, A.S., 1974. Chronic oxygen therapy treatment of chronic obstructive pulmonary disease at sea level. Chest 65, 279 – 285.
384
B. Wuyam et al. / Respiration Physiology 129 (2002) 375–384
Borgstro¨ m, L., Johanson, H., Sjesjo¨ , B.K., 1975. The relationship between arterial PO2 and cerebral blood flow in hypoxic hypoxia. Acta Physiol. Scand. 93, 423 –432. Brehar, K.L., den Hollander, J.A., Stromsky, M.E., Ogino, T., Schulman, R.G., Petroff, O.A.C., Prichard, L.W., 1983. High resolution 1H nuclear magnetic study of cerebral hypoxia in vivo. Proc. Natl. Acad. Sci. USA 80, 4945 – 4948. Brezinova, V., Calverley, P.M.A., Flenley, D.C., Townsend, H.R.A., 1979. The effect of long-term oxygen therapy on the EEG in patients with chronic stable respiratory failure. Bull. Eur. Physiopathol. Respir. 15, 603 –609. Bunn, H.F., Poyton, R.O., 1996. Oxygen sensing and molecular adaptation to hypoxia. Physiol. Rev. 76, 839 –885. Cavaletti, G., Tredici, G., 1993. Long-lasting neuropsychological changes after a single high altitude climb. Acta Neurol. Scand. 87, 103 – 105. Chandel, N.S., Schumaker, P.T., 2000. Cellular oxygen sensing by the mitochondria. J. Appl. Physiol. 88, 1880 –2000. Cohen, P.J., Alexander, S.C., Smith, T.C., Reivich, M., Wollman, H., 1967. Effects of normoxia and hypocapnia in cerebral blood flow and metabolism in conscious man. J. Appl. Physiol. 23, 183 –189. Cummins, T.R., Donnelly, D.F., Haddad, G.G., 1991. Effect of metabolic inhibition on the excitability of isolated hippocampal neurons: developmental aspects. J. Neurophysiol. 66, 1471 – 1482. De Ley, G., Nshimyumuremyi, J.B., Leusen, I., 1985. Hemispheric blood flow after unilateral common carotid artery occlusion. Stroke 16, 69 –73. Donelly, D.F., Jiang, C., Haddad, G.G., 1992. Comparative responses of brainstem and hippocampal neurons to O2 deprivation: in vitro intracellular studies. Am. J. Physiol. 262, L549 – L554. Gibson, G.E., Pulsinelli, W., Blass, J.P., Duffy, T.E., 1981. Brain dysfunction in mild to moderate hypoxia. Am. J. Med. 70, 1247 – 1254. Furling, D., Ghribi, O., Lhasaini, A., Miralut, M.E., Massicotte, G., 2000. Impairment of synaptic transmission by transient hypoxia in hippocampal slices: improved recovery in glutatione peroxydase transgenic mice. Proc. Natl. Acad. Sci. USA 97, 4351 –4356. Gurvitch, A.M., Ginsburg, D.A., 1977. Types of hypoxic delta activity in animals and men. Electroencephalogr. Clin. Neurophysiol. 42, 297 –408. Haddad, G.G., Jiang, C., 1992. Oxygen deprivation in the central nervous system (CNS). Physiol. Rev. 1, 278 – 318. Heistad, D.D., Kontos, H.A., 1983. Cerebral circulation. In: Sheperd, J.T., Abboud, F.M. (Eds.), Handbook of Physiology, Section 2: The Cardiovascular System, Ch. 5, vol. III. Oxford University Press, New York, pp. 137 –182. Hochachka, P.W., Buck, L.T., Doll, C.J., Land, S.C., 1996. Unifying theory of hypoxia tolerance: molecular/metabolic defense and rescue mechanisms for surviving oxygen lack. Proc. Natl. Acad. Sci. USA Biochem. 93, 9493 – 9498.
Hoffman, W.E., Miletich, D.J., Albrecht, R.F., Anderson, S., 1983. Regional cerebral blood flow measurements in rats with radioactive microspheres. Life Sci. 33, 1075 – 1080. Jiand, S., Agulian, S., Haddad, G.G., 1991. O2 tensions in adult and nonatal brain slices under several experimental conditions. Brain Res. 58, 159 – 164. Kety, S., Schmidt, C.F., 1948. The effects of altered tensions of carbon dioxide and oxygen on cerebral circulation. J. Clin. Invest. 27, 484 – 492. Kraaier, V., Van Huffelen, A.C., Wieneke, G.H., 1988. Quantitative EEG changes due to hypobaric hypoxia in normal subjects. Electroencephalogr. Clin. Neurophysiol. 69, 303 – 312. Levine, S., 1960. Anoxic ischemic encephalopathy in rats. Am. J. Pathol. 36, 1 – 17. Nioka, S., Smith, D.S., Chance, B., Subramanian, H.V., Butler, S., Katzenberg, M., 1990. Oxydative phosphorylation systems during steady-state hypoxia in the dog brain. J. Appl. Physiol. 68, 527 – 535. Novotny, E.J. Jr, 1996. Cerebral blood flow and metabolism in hypoxia. In: Haddad, G.G., Lister, G. (Eds.), Tissue Oxygen Deprivation. From Molecular to Integrated Function. Hypoxia. Marcel Dekker, New York, pp. 653 – 667. Payen, J.F., Le Bars, E., Wuyam, B., Tropini, B., Pe´ pin, J.L., Le´ vy, P., Decorps, M., 1996. Lactate accumulation during moderate hypoxia in neocortical rat brain. J. Cereb. Blood Flow Metab. 16, 1345 – 1352. Prior, P.F., 1985. EEG monitoring and evoked potentials in brain ischaemia. Br. J. Anaesth. 57, 63 – 81. Rossen, R., Simonsen, E., Baker, J., 1961. Electroencephalograms during hypoxia in healthy men. Arch. Neurol. 5, 648 – 654. Salford, L.G., Sjesjo¨ , B.K., 1974. The influence of arterial hypoxia and unilateral carotid artery occlusion upon regional blood flow and metabolism in the rat brain. Acta Physiol. Scand. 92, 130 – 141. Schurr, A., West, C.A., Rigor, B.M., 1988. Lactate supported synaptic function in the rat hippocampal slice preparation. Science 240, 1326 – 1328. Shockley, R.P., LaMamma, J.C., 1988. Determination of rat cerebral cortical blood volume changes by capillary mean transit time analysis during hypoxia, hypercapnia and hyperventilation. Brain Res. 454, 170 – 178. Sjesjo¨ , B.K., Johanson, H., Ljunggren, B., Norberg, K., 1974. In: Plum, F. (Ed.), Brain Dysfunction in Metabolic Disorders. Raven Press, New York, pp. 75 – 86. Sjesjo¨ , B.K., 1978. Brain Energy Metabolism. Wiley, New York, pp. 380 – 446. Van der Worpe, H.B., Kraaier, V., Wieneke, G.H., Van Huffelen, A.C., 1991. Quantitative EEG during progressive hypocarbia and hypoxia. Hyperventilation-induced EEG changes reconsidered. EEG Clin. Neurophysiol. 79, 335 – 341. Wuyam, B., DeMatteis, M., Dolbec, C., Argod, J., Pe´ pin, J.L., Payen, J.F., Le´ vy, P., 1998. Effects of acute hyoxemia/ischemia of moderate intensity on brain electrical activity in anesthetised rats. Am. J. Respir. Crit. Care 157, A342.
Brain electrical activity during combined hypoxemia and hypoperfusion in anesthetized rats B. Wuyam a, V. Bourlier b, J.L. Pe´pin a, J.F. Payen b, P. Le´vy a,* a
PRETA-TIMC, UMR CNRS 5525, Laboratoire de Physiologie, 38700 La Tronche, France b U INSERM U 438, CHU de Grenoble, B.P. 217X, 38043 Grenoble Ce´dex, France Accepted 18 September 2001
Abstract In order to investigate the effects of moderate hypoxemia on brain electrical activity and the consequences of an altered cerebro-vascular response to hypoxemia, we recorded changes in electrical activity of the brain in anesthetized rats following unilateral carotid artery ligation (UCAL). In these animals, on the clamped side, cerebral blood flow, whilst normal during normoxia, shows less augmentation during hypoxemia. Six anesthetized (Halothane) Sprague– Dawley rats with UCAL were studied during 20 min periods of baseline (FIO2 = 30%), hypoxemia (FIO2 = 9.5%) and recovery (FIO2 =30%): mean arterial pressure of oxygen (PAO2) achieved was 177.0, 37.6 and 160.1 mmHg, respectively. A significant decrease in the frequencies of the ECoG was observed bilaterally during hypoxemia: centroid frequency (fc) = 3.3790.14 and 2.85 90.13 Hz on the intact and clamped hemisphere respectively during hypoxemia versus fc = 4.0990.20 Hz (mean 9S.E.M.) during baseline, which was not reversed during recovery (3.279 0.11 Hz) (ANOVA, PB0.01). The total power of the signal (Pw) was unaffected on the intact hemisphere but diminished on the clamped side during hypoxemia. Our results show that a significant slowing of ECoG is observed during hypoxemia of moderate intensity (40 mmHg) even when cerebro-vascular response to hypoxemia is preserved and that total power of the ECoG signal is severely diminished when the cerebro-vascular response to hypoxemia is impaired. © 2002 Elsevier Science B.V. All rights reserved. Keywords: Carotid artery, ligation; Electrical activity, brain; Hypoxia, brain; Mammals, rat; Perfusion, brain hypoperfusion
1. Introduction In humans, hypoxic encephalopathy is observed in pilots or climbers exposed to high altitude and * Corresponding author: Present address: Laboratoire EFCR, Hoˆpital A. Michallon, B.P. 217 X, 38043, Grenoble Ce´dex, France. Tel.: +33-4-7676-5516; fax: +33-4-7676-5617. E-mail address: [email protected] (P. Le´vy).
in patients with cardio-respiratory diseases. Reduced oxygenation (FIO2 = 12%, PAO = 50 mmHg) may lead to progressive impairment of cerebral function, such as impaired retinal function, followed by impairment of short-term memory and cognitive functions. Finally, severe obtundation, ataxia, stupor and coma may be observed at altitudes higher than 5000 m (FIO2 = 10%, PAO = 35 mmHg) (Novotny, 1996). 2
2
0034-5687/02/$ - see front matter © 2002 Elsevier Science B.V. All rights reserved. PII: S 0 0 3 4 - 5 6 8 7 ( 0 1 ) 0 0 3 1 9 - X
376
B. Wuyam et al. / Respiration Physiology 129 (2002) 375–384
The precise mechanisms leading to the loss of brain function have been investigated in animals and humans and may depend on the severity of hypoxemia. Severe hypoxemia (PAO B25 mmHg) is accompanied by the rapid reduction in high-energy compounds (Sjesjo¨ et al., 1974) that is associated with a rapid loss of function. In more moderate levels of hypoxemia, ATP levels remain fairly constant (Sjesjo¨ , 1978) and other factors may be involved. These include: changes in membrane excitability of neurons, changes in neurotransmission or synaptic failure (Donelly et al., 1992; Gibson et al., 1981; Novotny, 1996). On the other hand, several mechanisms limit the consequences of reduced blood oxygen content. One is the increase in cerebral blood flow, which accompanies hypoxemia below 55 mmHg (Kety and Schmidt, 1948; Cohen et al., 1967; Heistad and Kontos, 1983; Shockley and LaMamma, 1988; Bereckzi et al., 1993). Furthermore, metabolic adaptations which may help to preserve energetic conditions within the brain, have been observed as an early response to cerebral hypoxia (Brehar et al., 1983). Investigations by our group (Payen et al., 1996) and others (Allen et al., 1992) suggest that early lactate accumulation may occur in response to moderate hypoxemia (PAO of 50 mmHg). Lactate is now seen as an important source of energy for neurons (Schurr et al., 1988) which may thus participate in cellular adaptation to hypoxemia. The present investigation was undertaken to assess functional correlates of such metabolic changes in rats with unilateral carotid artery ligation (UCAL). We investigated whether electrical activity of the cerebral cortex was preserved with a reduced oxygen delivery on both an intact hemisphere—where cerebro-vascular adaptation is preserved—and a clamped hemisphere. In the latter, cerebral blood flow is normal at rest due to blood flow supplementation through the Willis polygon (Levine, 1960) but fails to increase in response to hypoxemia and/or hypercapnia (De Ley et al., 1985; Salford and Sjesjo¨ , 1974).
2. Methods
2.1. Electrocorticogram electrodes placement
2
2
Electrodes were placed in OFA (Iffa Credo) female rats, 240–260 g in weight, under general anesthesia (intraperitoneal chloral hydrate, 4%). After incision of the skull, the animals were placed in a stereotactic frame in ventral decubitus position and two sets of unipolar stainless steel electrodes were placed symetrically in front of the parietal cortex at the outer surface of the dura. Electrodes were 5 mm lateral and 2 mm anterior to bregma. The rats were allowed to recover from general anesthesia and the ability of the electrodes to record valid ECoG was controlled during a period of quiet wakefulness of the animal. Recovery from anaesthesia and the normal behavior of the animal was confirmed The hypoxic experiments were performed within 1 week (2–8 days) following electrode placement procedures.
2.2. Surgical procedure Animals with pre-implanted electrodes were initially anaesthetized with halothane (5–7%) delivered via a facemask in a cage. The animals were tracheostomized and mechanically ventilated using a Edco/NEMI, type 804-small animals-respirator. Anesthesia was maintained with 5% inhaled halothane. The left carotid artery was clamped and the caudal end of the vessel was canulated to allow continuous blood pressure monitoring using a pressure transducer and recorder (Gould 8188S). Additionally, arterial blood samples were collected via the arterial canulation for blood gas analysis (Radiometer, ABL 510, Copenhagen, Danemark). Rectal temperature was continuously monitored and maintained constant (T= 37 °C9 0.5). After surgical preparation, the animals were allowed to recover for one hour during which they were maintained under light anesthesia (0.5–1.5% halothane). The same level of anesthesia was pursued for the rest of the experiment.
B. Wuyam et al. / Respiration Physiology 129 (2002) 375–384
2.3. Protocol A baseline recording of brain electrical activity was obtained during 10 min with the animal ventilated at an inspired fraction of oxygen (FIO2) of 30%. This was followed by a period of hypoxia obtained by reduction of FIO2 from 30 to 9.5% over 2–3 min and maintained for a total duration of 20 min. This was followed by a period of recovery during which FIO2 was returned progressively to 30% over 2– 3 min and maintained for a total duration of 20 min. Blood gases were determined at the end of each period of baseline, hypoxia and recovery. Since hypoxemia has a pronounced peripheral vasodilatory action in rats, an hypotensive response was observed in all animals. Care was taken to minimize the degree of hypotension. Female animals were chosen because they appeared to be less sensitive to hypoxic hypotension than males. Small doses of Dopamine (0.06 mg/ml) were infused intraperitoneally at a rate of 1– 2 ml/h throughout the experiment. This dose has minimal cerebral vasodilatory action and helped to stabilize blood pressure during the hypoxic exposure. Tidal volume and frequency of the respirator were adjusted to maintain eucapnia (target PACO2 of 35 mmHg). The FIO2 was continuously monitored (MiniOx 1, Catalyst Research Corporation).
377
band of the electrocorticogram: delta (1–4 Hz), theta (4– 8 Hz), alpha (8–12 Hz), beta (12–16 Hz) and sigma (16– 20 Hz). Signal power was expressed both as a fraction of the total signal and in absolute terms. The sum of B 1 Hz plus delta waves was referred to as the ‘percentage of low frequencies’ (0.5–4 Hz).
2.5. Statistics A three-way analysis of variance (two withinfactor: (i) condition i.e. baseline, hypoxia or recovery and (ii) time: min 1–10; one between-factor: hemisphere) was performed for the parameters of the ECoG signal: centroı¨d frequency and global power spectrum. Single parameters like blood pressure, or PAO2 and PACO2 were analyzed using one-way ANOVA for repeated measures. P values B 0.05 were considered significant.
3. Results
3.1. Biological 6ariables during hypoxemia Changes in FIO2 and the resulting changes in mean arterial blood pressure (MABP) are shown
2.4. Data acquisition and analysis Electrical signals of brain activity were continuously recorded, amplified and filtered. The filter time constant was set at 0.3 sec. This is equivalent to a ‘high-pass’ filter of 0.5 Hz. Signals were simultaneously displayed on an ECEM analogic recorder and stored on magnetic tape (Hewlett– Packard, 0–1 kHz) for further analysis. Recorded signals were digitized at a sampling frequency of 128 Hz (AD converter MacAdios). Fast Fourier Transform of the recorded signal was performed. The following parameters were derived: (i) centroid frequency (fc) (or median frequency, i.e. the frequency dividing the spectrum in two parts of equal power); (ii) total power (Pw) of the signal defined as the sum of squares of the Fourier series; and (iii) the power corresponding to each
Fig. 1. Changes in mean (S.D.) arterial blood pressure (black circles) observed at the corresponding levels of inspired fraction of oxygen (FIO2, white circles).
378
B. Wuyam et al. / Respiration Physiology 129 (2002) 375–384
end of hypoxia (group mean pH 7.359 0.04, vs. 7.439 0.04 during baseline), (one-way ANOVA, P= 0.02), which was not reversed after 20 min at an FIO2 of 30% (pH 7.369 0.03).
3.2. Electical acti6ity during baseline
Fig. 2. Changes in mean centroid frequency on the intact (white circles) and clamped hemisphere (black circles) associated with changes in inspired oxygen fraction (FIO2). Bars refer to standard error of the mean (S.E.M.).
Compared with ECoG during (i.e. without anesthesia), Mean ECoG was significantly slower during baseline (under light anesthesia), as compared with wakefulness: group mean fc of 4.099 0.08 Hz during baseline versus 5.549 0.07 Hz during unanesthetized quiet wakefulness (m9 S.E.M., PB 0.05). Mean fc was identical between hemispheres during baseline (mean fc= 4.049 0.06 vs. 4.169 0.07 Hz on the intact and clamped hemisphere, respectively). The change in fc induced by anesthesia was similar on the clamped and intact hemisphere (ANOVA, F=0.48, P= 0.88).
3.3. Electrical acti6ity during hypoxemia in Fig. 1. After a gradual decrease over 7 min (min 13–20, Fig. 1), a hypoxic gas mixture of 9.59 0.3% FIO2 (m9S.D.) was steadily inhaled for at least 10 min (min 20– 30 served as the reference for ‘steady-state’ hypoxemia). MABP decreased significantly during hypoxemia (group MABP of 61 913 vs. 10297 and 117 9 19 mmHg during baseline and steady recovery (min 40–50), respectively (one-way ANOVA, P B 0.05). The delay between the changes in FIO2 and changes in blood pressure was less than one minute in all animals. After intitial 40% decrease, a stabilization of the level of blood pressure was obtained until the end of the hypoxic exposure (Fig. 1). Group mean PAO2 achieved during the hypoxic period was 37.696.7 mmHg (from 33.8 to 50.2 mmHg). Eucapnia was maintained throughout the experiment for all the animals. Group mean PACO2 obtained was 32.894.6, 33.79 5.5 and 30.69 7.9 mmHg during baseline, hypoxia and recovery respectively (m9 S.D.), which were not significantly different (ANOVA, P = 0.38). Moderate extra-cellular acidosis was observed at the
Changes in electrical frequency of the ECoG during baseline, hypoxemia and recovery were analyzed in two ways: (1) during ‘steady-state’ conditions (min 20–30) where a stable level of FIO2 was maintained; (2) as a ‘transient’ response corresponding to the initial periods of hypoxia (min 10–20) and recovery (min 30–40) which were analyzed separately. Steady-state and transient periods were defined according to the stability of the FIO2 to which the animals were exposed.
3.4. Steady-state response Group mean changes in electrical activity during hypoxemia on both hemispheres are shown in Fig. 2. A significant slowing of the electrical activity was observed bilaterally during steady-state hypoxemia. Analysis of variance of the changes in centroid frequency (fc) indicated a significant difference between the three conditions of baseline (group mean fc= 4.099 0.08 Hz, m9 S.E.M.), hypoxemia (group mean fc= 3.1190.36 Hz) and recovery (group mean fc= 3.249 0.14 Hz;
B. Wuyam et al. / Respiration Physiology 129 (2002) 375–384
ANOVA, P B0.001). Moreover, a significant interaction between the condition and time (‘repeated measures’) was observed irrespective of the hemisphere (ANOVA, F =1.904, P B0.05). This suggests that the evolution of centroid frequency (fc) with time was mainly attributable to hypoxemia. Although slowing of the centroid frequency, particuliarely at the beginning of the hypoxemic period, tended to be more pronounced (see Fig. 2) in the clamped (mean group fc=2.85 9 0.39 Hz) than in the unclamped hemisphere (mean fc= 3.3790.19 Hz), this did not reach statistical significance (ANOVA, F =0.748, P =0.68). In absolute terms (Fig. 3), on the intact hemisphere, a shift in the frequencies of the ECoG was observed with reduction in the power of the highest frequencies (essentially 4– 8 Hz band), and a parallel increase in low frequencies: delta wave
379
(1– 4 Hz). The appearance of very low frequencies (B 1 Hz) was also noted on this side. These changes remained essentially stable during recovery (Fig. 3). Conversely, an absolute reduction in all frequencies was observed during hypoxemia on the clamped (left) hemisphere, which was greater in the ‘high’ frequencies (4–8 Hz) but also affected low (1–4 Hz) frequencies. This was partially reversed during recovery (Fig. 3). Overall, changes in the total power of the ECoG observed during hypoxemia were markedly different between the two hemispheres. As shown in Fig. 3, an important reduction in total power of the signal was observed on the clamped (left) hemisphere during hypoxemia (group mean: Pw = 9219 36, 3769 38 and 5949 94 mV2 during the baseline, hypoxemia and recovery periods, respectively, m9 S.E.M.) which was not entirely
Fig. 3. Changes in mean total signal power (black circles) and power in each frequency bands (0 – 1 Hz: white circle; 1 – 4 Hz: white polygon; 4– 8 Hz: black polygon; 8 –12 Hz: white triangle; 12 – 16 Hz: black triangles; 16 – 20 Hz: white squares) on the intact (top pannel) and clamped hemisphere (bottom). Note the global stability in the total signal power on the unclamped side.
380
B. Wuyam et al. / Respiration Physiology 129 (2002) 375–384
reversed after 20 min. Conversely, total power remained stable on the unclamped right hemisphere (760929, 767937 and 828930 mV2 during baseline, hypoxia and recovery, respectively) Statistical interaction between the hemisphere and the condition did not reach significance, however (ANOVA, F = 2.549, P = 0.10 for the betweenhemisphere difference).
3.5. Transient response A significant delay between the changes in FIO2 and the associated changes in electrical activity was observed. Fig. 3 indicates that the delay, between the initial drop in FIO2 below 21% (beginning of hypoxia) and the beginning of changes in the median frequencies, was at least 2 min. At the end of the hypoxic period and during recovery, overall slowing of ECoG (as reflected by changes in the median frequency) tended to be similar on both hemispheres. The rate of reduction in fc at the onset of hypoxemia was not different between hemispheres (mean slope calculated between the initial decrease and the minimal value (nadir) of fc: −0.2390.17 Hz/min on the unclamped versus −0.2890.27 Hz/min on the clamped side, paired t-test, P=0.58).
4. Discussion The main finding of this study is that electrical activity of the cortex is significantly and steadily modified within a few min of exposure to hypoxemia of moderate depth (40 mmHg), whether or not the cerebro-vascular response to hypoxemia is normal. This response is mainly characterized by a slowing of the frequencies observed. Impeded CBF response to hypoxemia was further associated with a severe and global reduction of the electrical power.
4.1. Experimental conditions These effects of acute hypoxemia have been observed in anesthetized tracheotomized animals with unilateral (left) carotid artery clamping. In such a preparation, CBF is minimally affected
during normoxia (due to blood flow supplementation through the Willis polygon) without detectable metabolic alterations (Hoffman et al., 1983; Salford and Sjesjo¨ , 1974; Payen et al., 1996). In our experiments, electrical activity (ECoG) was found to be identical (no significant difference in mean fc, % of low frequency and power spectrum) in both hemisphere, during normoxia. This tends to rule out the possibility that the surgical procedure per se may have affected neuronal viability and/or function. This is presumably due to the particular development of the Willis polygon in rats. We maintained successfully a constant PACO2 during the three conditions of baseline, hypoxemia and recovery. This ensured similar conditions of cerebral vasoreactivity to CO2 (Heistad and Kontos, 1983) between the three conditions and separated the effects of hypoxemia per se from those of combined hypoxemia and hypocapnia as observed in humans. This is of importance since hypocapnia, which accompanies the normal ventilatory response to hypoxemia, has been shown to add detrimental effects on cerebral function (Kraaier et al., 1988) and may lead to a significantly greater increase in the low frequencies of the EEG, than a pure hypoxic challenge (Borgstro¨ m et al., 1975; Van der Worpe et al., 1991). We believe that the cerebral blood flow response to hypoxemia was asymmetrical across hemispheres in our protocol. Although CBF is near normal at rest in such animal preparations, the CBF response to vasodilatory stimuli has been proven to be asymmetrical when one of the carotid arteries is clamped. In a study by De Ley et al., (1985) using a similar animal experiment with unilateral clamping, cerebral blood flow (CBF) was normal during normoxia but failed to increase in response to hypercapnia on the clamped side. Similarly, in response to hypoxemia, a reduced increase in CBF was observed on the clamped side in two separate experiments (Hoffman et al., 1983; Salford and Sjesjo¨ , 1974). Blood flow on the clamped side represented only 50% of the controlateral hemisphere. Using Doppler laser flow measurement, in similar hypoxic conditions, we have observed that CBF
B. Wuyam et al. / Respiration Physiology 129 (2002) 375–384
velocity response to hypoxemia was markedly different in each hemisphere: CBF velocity increased to 120% of the baseline value on the unclamped hemisphere and was reduced to 55% of baseline CBF velocity on the clamped hemisphere (Wuyam et al., 1998). We thus suggest that brain hemispheres were exposed to the same partial pressure of oxygen (PAO2) but presumably different O2 supply to brain cells as a consequence of differences in O2 delivery. The mechanisms of such an asymmetrical CBF response cannot be clarified from the present experiments. This imbalance could be either due to a steel phenomenon of blood flow by the intact side or to the loss of the autoregulation in response to hypotension when the cerebral circulation on one side relies on supplementation by the Willis polygon.
4.2. Changes in ECoG acti6ity in conditions of hypoxemia It has long been recognized that oxygen deprivation produces changes in neuronal electrical activity both in animals and humans. Slowing in spontaneous electrical activity (EEG) occurs within seconds of brain ischemia with subsequent depression of cortical evoked potentials and, ultimately, loss of function (Prior, 1985). These effects, however, combine the effects of complete oxygen deprivation and the absence of blood metabolites. There is clear evidence that neuronal response may differ between hypoxemia itself and hypoxemia combined with ‘chemical’ ischemia (Haddad and Jiang, 1992). The proper effects of hypoxic hypoxia are less well established, in particular those associated with mild to moderate reduction in O2 availability in vivo. Several reports have shown a slowing of the electrical activity associated with short-term exposure to hypoxic hypoxia in normal men (Gurvitch and Ginsburg, 1977; Kraaier et al., 1988; Rossen et al., 1961). Significant hyperventilation, however, leading to a significant degree of hypocapnia may contribute to the changes observed (Van der Worpe et al., 1991), especially after recent exposure to hypoxemia. In patients with chronic obstructive pulmonary disease (COPD) and permanent hypoxemia, Brezinova et al. (1979)
381
have reported a mean EEG frequency significantly closer to normal in patients receiving long term oxygen therapy as opposed to untreated pateints. The mean frequency of the EEG was correlated with arterial oxygen saturation (SAO2) in these patients. However, acute administration of oxygen did not significantly change the EEG frequencies in either group. This study suggests that chronic oxygen deprivation in humans may be associated with a significant slowing of the EEG which is reversed when hypoxemia is chronically (but not acutely) relieved. Block et al. (1974) have found similar findings with a partial change in the EEG with long-term oxygen therapy, which may suggest reversible neuronal injury or dysfunction during oxygen deprivation. EEG abnormalities have, similarly, been observed in elite mountain climbers 2–9 months after a sojourn above 8000 m without supplemental O2 (Cavaletti and Tredici, 1993). Similar changes in neuronal excitability could contribute to neuropsychological and/or memory impairment reported in these situations (Cavaletti and Tredici, 1993). However, both hypoxemia per se and/or changes in CBF may have contributed to the changes in electrical activity. Since neocortical neurons of rats respond similarly to those of humans when submitted to hypoxia (Haddad and Jiang, 1992), we believe that our experiments provide some valuable insight on this question. Our observations indicate that a significant slowing of the ECoG occurred on the intact hemisphere for a steady-state PAO of around 40 mmHg. We believe that the changes in frequency observed in our experiments may be due, at least in part, to changes in synaptic transmission and/ or neuronal excitability. Although the possibility exists that structures of the central nervous system (e.g. corpus callosum) may indirectly influence cortical structures through anatomical projections, there is abundant and direct evidence that neocortical neurons themselves can be affected by hypoxia (for example, see Haddad and Jiang, 1992). In tissue slice preparations, in response to hypoxic exposure, an initial hyperpolarization is observed, preceding a progressive return to prehypoxic resting membrane potentials. This hyperpolarisation can be maintained for prolonged 2
B. Wuyam et al. / Respiration Physiology 129 (2002) 375–384
382
periods of time (5– 20 min). Along with early hyperpolarization, synaptic potentials are markedly suppressed, the threshold for action potential is increased, and electrical input resistance decreased. Such changes are suggestive of changes in K+ conductance. Such changes in membrane excitability could contribute to the overall slowing of electrical activity recorded at the cortical surface in our experiments (temporal summation) and increase low frequencies recorded in the ECoG. Factors other than global membrane excitability and spontaneous discharge frequency could additionally contribute to a global slowing of the electrical activity. For example, changes in the morphology of action potentials of central neurons may be involved (Haddad and Jiang, 1992). Cummins et al. (1991) showed that, in response to brief exposures to hypoxia (i.e. pure N2 in voltage-clamp preparations), central hippocampal neurons become hyperpolarized, due to fast-inactivation of Na+ current. (Cummins et al., 1991). Such a change in Na+ channels could lead to a change in the kinetics of depolarization in neurons and contribute to an overall slowing of cortical networks. The extent to which each of these mechanisms contributes to ECoG frequency slowing is unknown. It should be emphasized, however, that changes in ECoG observed in our experiments in presence of a normal CBF response (intact hemisphere) are in accordance with biological effects of hypoxia (such as change in K conductance and inactivation of Na+ channels) on neurons themselves on tissular slice preparations. Furthermore, this effect has been observed following similar hypoxia (PO as high as 50 mmHg at the neuronal surface in slice preparation, (Jiand et al., 1991). This suggests that significant changes in neuronal excitability can be observed with relatively modest reductions in PO , unlike the classical notion that O2 deficit occurs below a PAO of 25 mmHg (Hochachka et al., 1996). In the central nervous system, however, alternative mechanisms may be implicated in the down regulation of firing rates and synaptic transmission (‘spike arrest’) (Hochachka et al., 1996). Adenosine-mediated down-regulation of excitatory amino acids (especially glutamate) with a concomitant increase in inhibitory amino acids, are likely to play a role. 2
2
2
The extent to which reduced CBF response may add detrimental effects on cortical activity is clearly shown in our experiments. The total power of the ECoG was significantly reduced for all frequencies on the clamped hemisphere, as opposed to the normal (intact) hemisphere where the increase in the power of low frequencies compensated the decrease in the power of high frequencies. Total power equals the square of the amplitudes of each frequency recorded in the spectrum. Total power is thus a parameter which reflects the global amount of electrical activity produced by cortical neurons in the region of interest. We suggest that this between-hemisphere differences in Total Power may be attributed to the differences in O2 delivery, i.e. differences in CBF response to the same reduction of O2 content. Close relationships between the magnitude in oxygen delivery (i.e. the amount of oxygen actually delivered to neurons) and the reduction in the overall ‘power’ of the EEG signal, has previously been suggested by Nioka et al. (1990) in dogs. In this study, progressive cerebral hypoxia was induced. The variations of O2 availability to neurons, and the metabolic and oxydative phosphorylation state of the brain tissue were monitored. Electrical activity of the EEG was analyzed. The ‘amplitude’ of the signal was reported (unlike the changes in the frequency bands of the spectrum). In this experiment (Nioka et al., 1990), changes in the amplitude of electrical activity strictly paralleled changes in O2 availability. When the CBF response was insufficient to compensate for the decrease in O2 content (reduced O2 availability), the O2 consumption of the brain tissue decreased. High energy compounds (PCr) were modified and glycolysis was stimulated (drop in intracellular pH). The calculated PO felt within the range of Km for cytochrome oxidase, which suggests that the kinetics of the ATP production may have been affected. In this situation, EEG amplitude fell below the EEG amplitude in control normoxic condition. Conversely, at lower grades of hypoxia, when the CBF was adequate to compensate a smaller reduction in arterial O2 content, the O2 consumption of the brain tissue was unaffected, cerebral metabolism unchanged, calculated PO was well above the Km for cy2
2
B. Wuyam et al. / Respiration Physiology 129 (2002) 375–384
tochrome oxidase, and electrical activity was preserved. Thus, using a very similar analysis of the ECoG, our results largely agree with the observations of Nioka et al. (1990) suggesting that the power of the EEG signal may be more closely related to the reduction in O2 availability in neurons than frequency analysis per se. Both experiments confirm that the maintenance of O2 delivery by an adequate CBF increase is essential for the preservation of global electrical activity. Changes in electrical activity are precipitously observed within 2–3 min of exposure, even to moderate hypoxemia. Recent evidence suggests that a common oxygen-sensing mechanism exists in virtually all cell types and can act within seconds of hypoxic exposure, i.e. independently of gene transcription or translation to a second messenger (Bunn and Poyton, 1996). However, there is no consensus on the nature of the sensor [which could be either a heme protein at the outer surface of the cellular membrane or an intracellular pathway such as the NADPH oxidase system or even the mitochondria themselves (Chandel and Schumaker, 2000)]. Whatever the sensor, the production of reactive oxygen intermediates is susceptible to provide a chemical link between intracellular oxygen concentration and the changes in proteins activity. If such a mechanism is susceptible to modulate ion channels activity in neurons, it would provide a possible explanation for the kinetics of changes in electrical activity observed in our preparation particularly on the clamped side. The link between functional deficit of central neurons (e.g. synaptic transmission in hippocampal slices) and the production of free radicals within these cells has been recently shown in glutathion peroxydase transgenic mice (Furling et al., 2000). A moderate increase in glutathion peroxydase (a free radical scavenger) observed in these animals is able to preserve synaptic transmission after short-term (10 min) hypoxia. Accordingly, the absence of change in electrical activity during re-oxygenation sould be pointed out in our experiments. This could be due either to the persistance of free radical production— which have also been implicated in the pathways of hypoxia/reoxygenation injury— or to the persitance for several hours of stress-activated kinases,
383
which may mediate cellular response to hypoxemia. Our data, in anesthetized rats, thus, indicate that relative hypoperfusion associated with hypoxemia adds an adverse effect based on ECoG. Total power of the signal, which represents global electrical activity, is significantly reduced. This is not the case when CBF response to hypoxemia is preserved (in the same animals). A slowing of the frequencies of the ECoG content is present in this case. A balanced change in the frequency composition of ECoG (i.e. reduction in high frequencies and increase in low frequencies) occurs at moderate levels of hypoxemia and are in keeping with known intracellular changes in excitability in response to an O2 sensing mechanism. The effect of the vascular response to hypoxemia on the global amount of cortical electrical activity suggests important clinical consequences in patients with impaired cerebral circulation (OSA patients) or during adaptation to high altitude. Additional studies using O2 free radicals modifiers (scavengers or transgenic animals) in our paradigm would give further insight on the mechanisms and cerebral consequences of hypoxemia.
Acknowledgements This work was supported by a grant from Re´ gion Rhoˆ ne-Alpes (Hypoxie). The authors are additionnaly grateful to Hamid Benzzouz for technical support.
References Allen, K., Busza, A.L., Cockard, H.A., Gadian, D.G., 1992. Brain metabolism and blood flow in acute cerebral hypoxia studied by MR spectroscopy and hydrogen clearance. NMR Biomed. 5, 48– 52. Bereckzi, D., Wei, L., Otsuka, T., Acuff, V., Pettigrew, K., Patlak, C., Fenstermacher, J., 1993. Hypoxia increases velocity of blood flow through parenchymal microvascular systems in rat brain. J. Cereb. Blood Flow Metab. 13, 475 – 486. Block, A.J., Castle, J.R., Keitt, A.S., 1974. Chronic oxygen therapy treatment of chronic obstructive pulmonary disease at sea level. Chest 65, 279 – 285.
384
B. Wuyam et al. / Respiration Physiology 129 (2002) 375–384
Borgstro¨ m, L., Johanson, H., Sjesjo¨ , B.K., 1975. The relationship between arterial PO2 and cerebral blood flow in hypoxic hypoxia. Acta Physiol. Scand. 93, 423 –432. Brehar, K.L., den Hollander, J.A., Stromsky, M.E., Ogino, T., Schulman, R.G., Petroff, O.A.C., Prichard, L.W., 1983. High resolution 1H nuclear magnetic study of cerebral hypoxia in vivo. Proc. Natl. Acad. Sci. USA 80, 4945 – 4948. Brezinova, V., Calverley, P.M.A., Flenley, D.C., Townsend, H.R.A., 1979. The effect of long-term oxygen therapy on the EEG in patients with chronic stable respiratory failure. Bull. Eur. Physiopathol. Respir. 15, 603 –609. Bunn, H.F., Poyton, R.O., 1996. Oxygen sensing and molecular adaptation to hypoxia. Physiol. Rev. 76, 839 –885. Cavaletti, G., Tredici, G., 1993. Long-lasting neuropsychological changes after a single high altitude climb. Acta Neurol. Scand. 87, 103 – 105. Chandel, N.S., Schumaker, P.T., 2000. Cellular oxygen sensing by the mitochondria. J. Appl. Physiol. 88, 1880 –2000. Cohen, P.J., Alexander, S.C., Smith, T.C., Reivich, M., Wollman, H., 1967. Effects of normoxia and hypocapnia in cerebral blood flow and metabolism in conscious man. J. Appl. Physiol. 23, 183 –189. Cummins, T.R., Donnelly, D.F., Haddad, G.G., 1991. Effect of metabolic inhibition on the excitability of isolated hippocampal neurons: developmental aspects. J. Neurophysiol. 66, 1471 – 1482. De Ley, G., Nshimyumuremyi, J.B., Leusen, I., 1985. Hemispheric blood flow after unilateral common carotid artery occlusion. Stroke 16, 69 –73. Donelly, D.F., Jiang, C., Haddad, G.G., 1992. Comparative responses of brainstem and hippocampal neurons to O2 deprivation: in vitro intracellular studies. Am. J. Physiol. 262, L549 – L554. Gibson, G.E., Pulsinelli, W., Blass, J.P., Duffy, T.E., 1981. Brain dysfunction in mild to moderate hypoxia. Am. J. Med. 70, 1247 – 1254. Furling, D., Ghribi, O., Lhasaini, A., Miralut, M.E., Massicotte, G., 2000. Impairment of synaptic transmission by transient hypoxia in hippocampal slices: improved recovery in glutatione peroxydase transgenic mice. Proc. Natl. Acad. Sci. USA 97, 4351 –4356. Gurvitch, A.M., Ginsburg, D.A., 1977. Types of hypoxic delta activity in animals and men. Electroencephalogr. Clin. Neurophysiol. 42, 297 –408. Haddad, G.G., Jiang, C., 1992. Oxygen deprivation in the central nervous system (CNS). Physiol. Rev. 1, 278 – 318. Heistad, D.D., Kontos, H.A., 1983. Cerebral circulation. In: Sheperd, J.T., Abboud, F.M. (Eds.), Handbook of Physiology, Section 2: The Cardiovascular System, Ch. 5, vol. III. Oxford University Press, New York, pp. 137 –182. Hochachka, P.W., Buck, L.T., Doll, C.J., Land, S.C., 1996. Unifying theory of hypoxia tolerance: molecular/metabolic defense and rescue mechanisms for surviving oxygen lack. Proc. Natl. Acad. Sci. USA Biochem. 93, 9493 – 9498.
Hoffman, W.E., Miletich, D.J., Albrecht, R.F., Anderson, S., 1983. Regional cerebral blood flow measurements in rats with radioactive microspheres. Life Sci. 33, 1075 – 1080. Jiand, S., Agulian, S., Haddad, G.G., 1991. O2 tensions in adult and nonatal brain slices under several experimental conditions. Brain Res. 58, 159 – 164. Kety, S., Schmidt, C.F., 1948. The effects of altered tensions of carbon dioxide and oxygen on cerebral circulation. J. Clin. Invest. 27, 484 – 492. Kraaier, V., Van Huffelen, A.C., Wieneke, G.H., 1988. Quantitative EEG changes due to hypobaric hypoxia in normal subjects. Electroencephalogr. Clin. Neurophysiol. 69, 303 – 312. Levine, S., 1960. Anoxic ischemic encephalopathy in rats. Am. J. Pathol. 36, 1 – 17. Nioka, S., Smith, D.S., Chance, B., Subramanian, H.V., Butler, S., Katzenberg, M., 1990. Oxydative phosphorylation systems during steady-state hypoxia in the dog brain. J. Appl. Physiol. 68, 527 – 535. Novotny, E.J. Jr, 1996. Cerebral blood flow and metabolism in hypoxia. In: Haddad, G.G., Lister, G. (Eds.), Tissue Oxygen Deprivation. From Molecular to Integrated Function. Hypoxia. Marcel Dekker, New York, pp. 653 – 667. Payen, J.F., Le Bars, E., Wuyam, B., Tropini, B., Pe´ pin, J.L., Le´ vy, P., Decorps, M., 1996. Lactate accumulation during moderate hypoxia in neocortical rat brain. J. Cereb. Blood Flow Metab. 16, 1345 – 1352. Prior, P.F., 1985. EEG monitoring and evoked potentials in brain ischaemia. Br. J. Anaesth. 57, 63 – 81. Rossen, R., Simonsen, E., Baker, J., 1961. Electroencephalograms during hypoxia in healthy men. Arch. Neurol. 5, 648 – 654. Salford, L.G., Sjesjo¨ , B.K., 1974. The influence of arterial hypoxia and unilateral carotid artery occlusion upon regional blood flow and metabolism in the rat brain. Acta Physiol. Scand. 92, 130 – 141. Schurr, A., West, C.A., Rigor, B.M., 1988. Lactate supported synaptic function in the rat hippocampal slice preparation. Science 240, 1326 – 1328. Shockley, R.P., LaMamma, J.C., 1988. Determination of rat cerebral cortical blood volume changes by capillary mean transit time analysis during hypoxia, hypercapnia and hyperventilation. Brain Res. 454, 170 – 178. Sjesjo¨ , B.K., Johanson, H., Ljunggren, B., Norberg, K., 1974. In: Plum, F. (Ed.), Brain Dysfunction in Metabolic Disorders. Raven Press, New York, pp. 75 – 86. Sjesjo¨ , B.K., 1978. Brain Energy Metabolism. Wiley, New York, pp. 380 – 446. Van der Worpe, H.B., Kraaier, V., Wieneke, G.H., Van Huffelen, A.C., 1991. Quantitative EEG during progressive hypocarbia and hypoxia. Hyperventilation-induced EEG changes reconsidered. EEG Clin. Neurophysiol. 79, 335 – 341. Wuyam, B., DeMatteis, M., Dolbec, C., Argod, J., Pe´ pin, J.L., Payen, J.F., Le´ vy, P., 1998. Effects of acute hyoxemia/ischemia of moderate intensity on brain electrical activity in anesthetised rats. Am. J. Respir. Crit. Care 157, A342.