Neurovascular coupling in rat brain operates independent of hemoglobin deoxygenation

Transcription

1 & 2010 ISCBFM All rights reserved X/10 $ Neurovascular coupling in rat brain operates independent of hemoglobin deoxygenation Ute Lindauer 1,2,6, Christoph Leithner 1,3,6, Heike Kaasch 1, Benjamin Rohrer 1, Marco Foddis 1,3, Martina Füchtemeier 1, Nikolas Offenhauser 1, Jens Steinbrink 3, Georg Royl 1,3, Matthias Kohl-Bareis 5 and Ulrich Dirnagl 1,3,4 1 Departments of Neurology and Experimental Neurology, Charité Universitätsmedizin Berlin, Berlin, Germany; 2 Department of Neurosurgery, Technical University of Munich, Munich, Germany; 3 Center for Stroke Research Berlin, Berlin, Germany; 4 NeuroCure Research Center Berlin, Berlin, Germany; 5 Department of Mathematics and Technology, University of Applied Sciences Koblenz, RheinAhrCampus Remagen, Remagen, Germany Recently, a universal, simple, and fail-safe mechanism has been proposed by which cerebral blood flow (CBF) might be coupled to oxygen metabolism during neuronal activation without the need for any tissue-based mechanism. According to this concept, vasodilation occurs by local erythrocytic release of nitric oxide or ATP wherever and whenever hemoglobin is deoxygenated, directly matching oxygen demand and supply in every tissue. For neurovascular coupling in the brain, we present experimental evidence challenging this view by applying an experimental regime operating without deoxy-hemoglobin. Hyperbaric hyperoxygenation () allowed us to prevent hemoglobin deoxygenation, as the oxygen that was physically dissolved in the tissue was sufficient to support oxidative metabolism. Regional CBF and regional cerebral blood oxygenation were measured using a cranial window preparation in anesthetized rats. Hemodynamic and neuronal responses to electrical forepaw stimulation or cortical spreading depression (CSD) were analyzed under normobaric normoxia and during up to 4 ATA (standard atmospheres absolute). Inconsistent with the proposed mechanism, during, CBF responses to functional activation or CSD were unchanged. Our results show that activation-induced CBF regulation in the brain does not operate through the release of vasoactive mediators on hemoglobin deoxygenation or through a tissue-based oxygen-sensing mechanism. ; doi: /jcbfm ; published online 30 December 2009 Keywords: ATP; cerebral blood flow; cortical spreading depression; feed-forward regulation; hyperbaric hyperoxygenation; somatosensory stimulation Introduction Changes in neuronal activity are tightly coupled in both the spatial and temporal domains to changes in regional metabolism, regional cerebral blood flow (rcbf), and regional cerebral blood oxygenation (rcbo): neurovascular coupling. Modern neuroimaging methods, such as functional magnetic resonance Correspondence: Professor/Dr U Lindauer, Department of Neurosurgery, Experimental Neurosurgery, Technical University Munich, Klinikum rechts der Isar, Ismaninger Str. 22, Munich 81675, Germany. ute.lindauer@lrz.tum.de 6 These authors contributed equally to the work. This study was supported by the Deutsche Forschungsgemeinschaft, the Hermann and Lilly Schilling Foundation, and the European Union (EFRE). Received 3 September 2009; revised 13 November 2009; accepted 23 November 2009; published online 30 December 2009 imaging (fmri) and positron emission tomography, map brain activity with high spatial resolution by measuring changes in rcbf or rcbo. Although neurovascular coupling has been investigated for almost 120 years, the cellular and molecular mechanisms that couple neuronal activity to metabolism and blood supply are still incompletely understood. The metabolic hypothesis implicates vasoactive products of metabolism, which are released from activated neurons and glial cells, such as H +, or adenosine. The neuronal hypothesis, on the other hand, posits that neuronal energy demand is communicated to the vasculature (either directly or through astrocytes) in an anticipatory, feedforward manner by vasoactive neurotransmitters or products of synaptic signaling, such as K + or nitric oxide (NO) (Iadecola, 2004). The additional oxygen consumed during functional brain activation is delivered by increased

2 758 concentration of oxygenated hemoglobin (oxy-hb), which becomes deoxygenated in this process. It is tempting to ulate that hemoglobin deoxygenation might be the master regulator of organ blood flow, as such a mechanism would potentially be failsafe and independent of tissue factors. Indeed, three major hemoglobin deoxygenationbased mechanisms have recently been ulated as universal regulators of tissue perfusion (for an overview, see Figure 1): Stamler et al (1997) have shown that in the lung the potent vasodilator NO binds to the b-93 cysteine moiety (bcys93) of oxy-hb, forming S-nitroso-hemoglobin (SNO-Hb). Oxygenated blood then carries SNO-Hb to the tissue, where oxygen consumption induces allosteric transition of hemoglobin and release of NO on hemoglobin deoxygenation. Bound to thiols within the erythrocyte, NO is shuttled to the proximity of vascular smooth muscle cells, in which it induces guanylate-cyclase-dependent vasorelaxation. A second, recently described hemoglobin-deoxygenation-based mechanism to regulate organ perfusion ulates the release of NO from nitrite in an oxygen-consumption-dependent manner by the nitrite reductase activity of deoxygenated hemoglobin (deoxy-hb) (Cosby et al, 2003). Finally, organ perfusion may be regulated by ATP, a potent vasodilator, released from erythrocytes depending on the amount of conversion of oxy-hb to deoxy-hb (Dietrich et al, 2000; Gonzalez-Alonso et al, 2002). A characteristic shared by all three mechanisms is that whenever and wherever oxygen is released from hemoglobin, a vasodilator is delivered (NO or ATP). This co-release of oxygen and vasodilator(s) might adapt blood flow to local hypoxia or local changes in oxygen metabolism in any tissue of any organism with hemoglobin-like oxygen carriers, independent of local tissue factors. Whether these elegant hemoglobin deoxygenationbased mechanisms have a role in regulation of blood flow in the brain has, surprisingly, never been tested experimentally during neuronal activation. In principle, SNO-Hb measurements could provide insights into whether hemoglobin delivers NO for tissue blood flow regulation. However, even in more easily accessible organs than the brain, this has led to controversial results (Gaston and Hare, 2003; Gladwin and Schechter, 2004; Schechter and Gladwin, 2003; Singel, 2003; Stamler, 2003), and detection of local arteriovenous gradients of the proposed mediators (SNO-Hb and ATP) is not feasible in the brain. Figure 1 Schematic overview of hemoglobin deoxygenation-based mechanisms of blood flow regulation to increased oxygen demand: Recently, several elegant mechanisms of oxygen metabolism-induced delivery of the vasoactive molecules nitric oxide (NO) and ATP have been proposed for the coupling of tissue metabolism to blood flow in any tissue. Model 1: The potent vasodilator NO binds to the b-93 cysteine moiety (bcys93) of oxygenated hemoglobin (oxy-hb), forming S-nitroso-hemoglobin (SNO-Hb). Oxygenized blood carries SNO-Hb to the tissues, where increased tissue oxygen consumption induces allosteric transition of hemoglobin and release of NO on hemoglobin deoxygenation (Stamler et al, 1997). Model 2: NO is released from nitrite in an oxygen-consumptiondependent way by the nitrite reductase activity of deoxygenated hemoglobin (deoxy-hb; Cosby et al, 2003). Model 3: ATP is released by erythrocytes depending on the amount of conversion of oxy-hb to deoxy-hb (Dietrich et al, 2000; Gonzalez-Alonso et al, 2002).

3 In addition to methodological difficulties, the measurements of SNO-Hb, nitrite, or ATP would not allow researchers to establish a causal functional role for these molecules in blood flow regulation. Even recently presented data on new transgenic mouse models lacking the Hb-bCys93 residue were not able to solve the controversy (original publication Isbell et al, 2008; controversial discussion Patel and Townes, 2008; Stamler et al, 2008). We have therefore opted to use hyperbaric hyperoxygenation () to test whether neuronal activation-induced vasodilation depends on the delivery of vasodilators during hemoglobin deoxygenation. During with 100% oxygen concentration of gas at X3 standard atmospheres of absolute pressure (ATA; 1 ATA = Pa), oxygen supply to tissue is provided entirely through oxygen physically dissolved in the blood: The periarteriolar O 2 gradient is eliminated, no allosteric transition of oxy-hb to deoxy-hb occurs, and therefore SNO-Hb or nitrite remains preserved and no ATP is released by erythrocytes. When hemoglobin deoxygenation is prevented, the blood flow response should be blocked. In our experiments in anesthetized rats, we recorded CBF increases elicited by somatosensory stimulation during. As functional activation leads to only small increases of oxygen metabolism, we also used the strong stimulus of neuronal metabolism and oxygen consumption, cortical spreading depression (CSD), and measured CSD-evoked CBF increases during. We found that despite clamping brain oxy-hb at 100%, somatosensory stimulation and CSD-induced CBF increases remained unchanged, clearly ruling out any significant contribution by hemoglobin deoxygenation- or tissue-oxygen-sensor-dependent mechanisms to neurovascular coupling. Materials and methods Animals The experiments were approved by the Landesamt für Arbeitsschutz, Gesundheitsschutz und technische Sicherheit Berlin (G 0133/02) and performed according to national and institutional regulations. A total of 35 male Wistar rats (250 to 380 g) were used for the following studies: functional activation study = 20 and CSD study = 15. Animal Preparation The animals were anesthetized with isoflurane (in 70% N 2 O, 30% O 2 ) and artificially ventilated. Body temperature was continuously measured by a rectal probe and kept constant using a heating pad. The left femoral artery and vein were cannulated for arterial blood gas analysis, arterial blood pressure monitoring, and infusion of physiologic saline solution. The animals were placed in a stereotactic frame, the skull was removed, and a cranial window (dura mater intact) closed with a glass coverslip was implanted over the right somatosensory cortex. The space beneath the coverslip was filled with artificial cerebrospinal fluid (acsf). Somatosensory evoked potentials (SEPs) (functional activation study) were measured through an epidurally placed Ag/AgCl ball electrode and averaged using a commercial EEG recording unit (Nihon Kohden Neuropack Monitor, Tokyo, Japan). In the CSD study, direct current (DC) potentials were measured through an epidurally placed Ag/AgCl ball electrode connected to an amplifier (Jens Meyer, Munich, Germany). A second open cranial window (small burr hole) was rostrally implanted over the parietal cortex in the CSD study to elicit CSD by KCl application. After surgery, anesthesia was switched to the a-chloralose HBC complex (45 mg per kg body weight as intravenous (i.v.) bolus injection, followed by continuous i.v. infusion of 45 mg/kg per h as necessary). In the functional activation study, neuronal activation was induced by electrical sensory stimulation of the left forepaw (stimulation period 16 secs at 3 Hz; single stimulus duration 0.3 ms; intensity 1.2 to 1.6 ma; interstimulation interval 60 secs). In the CSD study, shortlasting trains of 3 to 5 CSDs were elicited at the frontal open window by a droplet of KCl (160 to 180 mmol/l), which was removed by repeated gentle flushing as soon as the first CSD wave was detected by DC potential shift and CBF response. was induced in a hyperbaric chamber (Decompression Chamber Systems, Dräger Safety AG & Co. KGaA, Lübeck, Germany) flushed with 100% O 2 and pressurized to 3 or 4 ATA, respectively. The hyperbaric chamber was especially constructed to afford space for the animal without additional space for the respirator or infusion pumps. Mechanical ventilation and i.v. infusion were therefore not feasible in the closed chamber during periods. The animals breathed spontaneously within seconds after disconnection from the ventilator. Before interruption of continuous i.v. infusion, a bolus of a-chloralose HBC complex (0.5 ml) was injected to maintain adequate anesthesia levels during /sham periods. Body temperature in the chamber was recorded and kept constant within the physiologic range by chamber heating. Arterial blood gases were analyzed before each stimulation block or CSD block under normobaric normoxia at baseline and after. In animals of the functional activation study, blood gases were also analyzed during by a catheter originating from the femoral artery advanced through one of the pressure-resistant ports in the chamber wall. Systemic arterial blood pressure was recorded continuously with the exception of the period of hyperbaric pressure, during which recording was technically impossible. After the experiments, the animals were killed in deep anesthesia by intravenous administration of concentrated KCl. The experimental setup is illustrated in Figures 2A 2C. While establishing the chamber, in one animal, oxygenation of arterial blood of the systemic circulation was measured in the chamber at 3 ATA in the femoral artery catheter using an oxygen electrode (solid-state small needle electrode coupled to a LICOX po 2 monitor, Integra NeuroSciences GmbH, Ratingen, Germany). In this animal, no CBF data were recorded. 759

4 760 EEG recording LDF / spectroscopy probe DC recording LDF probe window for KCI (CSD elicitation) hyperbaric chamber 100% O 2 3 or 4 ATA anesthetized rat LDF-spectroscopy / EEG-DC / KCl Arterial line / temperature probe Forepaw stimulation Experimental design (functional activation / CSD) control group (n=7 / n=4) baseline sham sham group I (n=9 / n=6) group II (n=4 / n=5) baseline 3ATA 3ATA baseline 4ATA 4ATA stimulation block I (fct activation / CSD) stimulation block II (fct activation / CSD) stimulation block III (fct activation / CSD) chloralose i.v. continuous infusion YES NO (animals anesthetized from chloralose bolus injection prior to ) YES ventilation YES NO YES Figure 2 Experimental setup. (A) Functional activation study: Localization of the cranial window over the parietal cortex within the somatosensory area of the forepaw with combined laser-doppler flowmetry (LDF)/spectroscopy probe and EEG electrode. (B) Cortical spreading depression (CSD) study: Localization of the cranial window at the parietal cortex with the LDF probe and direct current (DC) potential electrode. An additional small burr hole rostrally from the recording area served to elicit CSDs by short-lasting KCl application (160 to 180 mmol/l). (C) Hyperbaric chamber: Schematic drawing of a rat placed in the hyperbaric chamber. LDF, spectroscopy, and temperature probes, as well as the femoral artery catheter and the forepaw stimulation electrode (functional activation study) or the catheter for KCl application (CSD study), were induced through pressure-resistant ports in the chamber wall. (D) Experimental design: n = number of animals in each group (functional activation study/csd study). Measurement of Cerebral Blood Flow and Blood Oxygenation Changes in rcbf were measured through the cranial window by laser-doppler flowmetry (LDF, Perimed AB, Järfälla, Sweden) and were recorded continuously with a time resolution of 100 ms. In four animals of the functional activation study (three of the 3 ATA group and one of the 4 ATA group), hemoglobin oxygenation changes in the somatosensory cortex were recorded in addition to CBF through the cranial window with an instrument and analysis which has been described in detail elsewhere (Royl et al, 2008). The device consisted of separate channels for measurement of CBF based on LDF and of hemoglobin oxygenation based on reflection spectroscopy. Of these four animals, the LDF data of two animals with 3 ATA had to be discarded because of laser instability with significant and rapid baseline drifts. Therefore, the data on rcbf changes from seven out of nine animals in the

5 3 ATA group of the functional activation study were analyzed. Experimental Design The experimental design is illustrated in Figure 2D. Briefly, under normobaric normoxia (baseline condition, functional activation/csd block I), rcbf responses to functional activation or CSD were recorded in each group. The animals were then transferred to the chamber; the chamber was flushed with 100% oxygen for 5 to 10 mins before closing, and was pressurized to 3 or 4 ATA, within 10 to 15 mins. Regional CBF and rcbo responses to functional activation or CSD were recorded again under hyperbaric hyperoxia ( condition, functional activation/csd block II). The pressure within the chamber was then slowly brought back to baseline and rcbf responses to functional activation or CSD were recorded starting 20 mins after (- condition, functional activation/csd block III). In the control group, the animals were treated accordingly, with the exception that the chamber was flushed with room air and not pressurized (21% O 2 at 1 ATA, sham ). Data Analysis Cerebral blood oxygenation: In four animals of the functional activation study, averaged responses in oxy-hb, deoxy-hb, and total hemoglobin (total-hb; sum of oxy-hb and deoxy-hb, reflecting blood volume) at normoxia, during and after, were calculated and presented as concentration change from resting values directly before stimulation. Owing to the low number of animals, no statistical analysis of hemoglobin oxygenation changes was carried out in this data set. Table 1 Physiological parameters Experimental group MABP (mm Hg) pao 2 (mm Hg) paco 2 (mm Hg) Functional activation study Control group Baseline 113±12 126±10 38±3 7.38±0.02 Sham 111±13 92±10* 37±4 7.36±0.02 Post-sham 124±7 96±11* 37±4 7.36± ATA group Baseline 111±14 116±14 38±3 7.38±0.01 ND d 44±4* 7.34±0.02* Post- 122±9* 96±4* 39±7 7.38± ATA group Baseline 118±12 127±8 38±6 7.37±0.05 ND d 41±5 7.38±0.02 Post- 122±6 110±13 34±4 7.39±0.02 CSD study Control group Baseline 103±15 121±11 38±4 7.41±0.03 Sham 97±15 88±5* 42±1 7.35±0.01 Post-sham 110±17 94±6* 40±3 7.36± ATA group Baseline 103±10 126±10 38±2 7.40±0.03 ND ND ND ND Post- 119±2* 108±15 40±3 7.37± ATA group Baseline 105±20 133±4 38±3 7.41±0.02 ND ND ND ND Post- 123±7 103±7* 36±4 7.38±0.01*, hyperbaric hyperoxygenation; MABP, mean arterial blood pressure; ND not determined; d, exceeding upper limit of blood gas analyzer. *Significantly different to baseline within the same group (P < 0.05). ph 761 Cerebral blood flow: The rcbf responses to functional activation or CSD under hyperbaric hyperoxia () and after hyperbaric hyperoxia (-) were compared with responses elicited in each animal during baseline conditions under normobaric normoxia (baseline) in the same group. In the functional activation study, CBF data are presented as the average of 8 to 12 stimulus trains during baseline, and -, respectively. Regional CBF changes in the functional activation study are calculated as percentage increase from resting flow before stimulation onset. For quantification, functional activation-induced rcbf increases were calculated as the mean percentage increase of the period of 5 to 16 secs after stimulation onset (plateau phase of the CBF response). At the higher CBF values that occur during CSD, LDF-CBF overestimates absolute blood flow and is therefore no longer linearly correlated with CBF measured by the gold standard of autoradiography (Dirnagl et al, 1989). Therefore, rcbf changes during CSD in arbitrary units (AU) of LDF were used for further analysis. CSD-induced rcbf increases are presented as the maximum amplitude of the CBF response during CSD. Neuronal activity: Neuronal activity was assessed by SEP and DC potential recording. For comparison of functional changes of neuronal activity during baseline conditions and during as well as - periods, the absolute amplitude of the averaged SEP (P1 to N1) or the absolute value of the maximal negative shift of the DC potential during spreading depression was used for further analysis. Data normalization: For quantitative analysis, resting CBF as well as the mean responses of CBF and neuronal activity (SEP and DC potential amplitude) during and after were normalized to the values under baseline conditions in each animal, to correct for differences in resting CBF and for differences in stimulation responses in the neuronal and vascular parameters during baseline conditions between animals. Statistical Analysis For averaged time courses of LDF, oxy-hb, deoxy-hb, and total-hb, mean values were calculated from all animals tested. Quantified data are presented as vertical boxes with error bars (median, 10th, 25th, 75th, and 90th percentiles, single values included). Comparisons between baseline,, and - are conducted using Kruskal Wallis

6 762 change of resting CBF normalized to baseline Functional activation study 3 ATA group 4 ATA group * * change of resting CBF normalized to baseline CSD study 3 ATA group 4 ATA group * Figure 3 Effect of hyperbaric hyperoxygenation () on resting cerebral blood flow (CBF). In the (A) functional activation study and in the (B) cortical spreading depression (CSD) study, resting CBF increased slightly during and after at 3 or 4 ATA ( and -). Data are normalized to baseline values and are presented as box plots with median (line within the box), 25th/75th percentile (outer margin of the box) and 10th/90th percentile (lines originating from boxes), and single values are included as black dots. *P < one-way analysis of variance on ranks. When differences were detected, we used Dunn s method for multiple comparisons versus baseline as hoc test (SigmaStat 2.0, SPSS Inc, Chicago, IL, USA; statistical software). P < 0.05 was considered as statistically significant. Results Systemic Physiologic Parameters Remained Within the Physiologic Range With the exception of po 2 during, all systemic parameters were within physiologic ranges during the experiments. A slight decrease in arterial po 2 was detected in the control group during and after sham (animals transferred to hyperbaric chamber, experiment conducted at normobaric normoxia) as was a slight increase in arterial pco 2 in the groups during the period of (see Table 1). In one animal, systemic arterial po 2 measured with an oxygen electrode within the chamber at 100% oxygen and 3 ATA exceeded the upper detection limit of the electrode of 2000 mm Hg. This confirms the effectivity of flushing the chamber with 100% oxygen at a pressure of 3 ATA. The value calculated for systemic arterial po 2 under these conditions should be B2000 mm Hg, assuming an atmospheric pressure of 760 mm Hg under normal conditions at 1 ATA and considering the arterial pco 2 and the vapor pressure of water. Resting Cerebral Blood Flow Slightly Increased During and After Sham Hyperbaric Hyperoxygenation or Hyperbaric Hyperoxygenation In the functional activation study, a slight increase in resting CBF was detected in the control group (Supplementary Figure S1a, Supplementary Material) as well as in the groups (Figure 3A). In the CSD study, resting CBF showed the same tendency toward higher values during sham (Supplementary Figure S1a, Supplementary Material) or (Figure 3B). The slight increase in resting CBF was most likely attributable to a slight increase in paco 2. We could not counteract the paco 2 increase through adjustment of ventilation, because our pressure chamber does not allow for mechanical ventilation of the animals. Stimulation-Induced Neuronal and Cerebral Blood Flow Responses Remained Stable in the Control Groups Evoked neuronal activity and stimulation-induced CBF responses during functional activation and CSD remained unaltered in the control groups of both studies during sham and -sham (see Supplementary Figure S1b and c, Supplementary Material). Deoxygenated Hemoglobin Response to Functional Activation Disappeared During Hyperbaric Hyperoxia In four animals of the functional activation study, we applied optical spectroscopy in addition to LDF to measure hemoglobin oxygenation changes. In all animals tested, the deoxy-hb changes disappeared during at 3 or 4 ATA (see Supplementary Figure S2, Supplementary Material). Hyperbaric Hyperoxia Showed no Effect on Neurovascular Coupling During Functional Activation or Cortical Spreading Depression Stimulated neuronal activity (SEP: typical example in Figure 4A, quantified results in Figure 4B; DC potentials: typical example in Figure 5A, quantification in Figure 5B) did not change during at 3 or 4 ATA compared with their respective baseline. SEP or DC amplitudes and temporal dependence remained constant, showing cellular activation to functional stimulation or CSD comparable under baseline condition and. Even during 4 ATA, no

7 2.0 3 ATA group 4 ATA group 763 baseline 50 ms 4 ATA 4 ATA 10 µv SEP amplitude normalized to baseline CBF response [%] ATA group time [s] ATA group baseline time [s] mean CBF response normalized to baseline response ATA group 4 ATA group total-hb change [mm] ATA group baseline time [s] Figure 4 No effect of hyperbaric hyperoxygenation () on neuronal activity and neurovascular coupling during functional activation. (A) Typical example of the somatosensory evoked potential (SEP) changes in one animal during electrical forepaw stimulation under baseline condition (left), during at 4 ATA (middle) and after (right). Neither the amplitude nor the shape of the SEPs changed during and after. (B) Quantified data: SEP amplitudes remained unchanged during and after in the 3 ATA group and 4 ATA group. (C) Time courses of averaged CBF responses to functional activation (increase in percentage from 100% of resting flow) at baseline conditions under normobaric normoxia (blue line), during (orange line) and after (gray line) of the 3 ATA group (left) and of the 4 ATA group (right). The gray area marks the stimulation period. (D) Quantified data of stimulationinduced regional cerebral blood flow (rcbf) increases, normalized to baseline responses: CBF responses remained unchanged during and after at 3 or 4 ATA. (E) Time courses of averaged total hemoglobin (total-hb) concentration responses to functional activation (change of total-hb in mmol/l) at baseline conditions under normobaric normoxia (blue line) and during (orange line) of the 3 ATA group. The gray area marks the stimulation period. Excessive oxygen availability during (3 or 4 ATA groups) had no effect on the blood flow and corpuscular blood volume responses to electrical forepaw stimulation. (B and D) Data are normalized to baseline values and are presented as box plots with median (line within the box), 25th/75th percentile (outer margin of the box) and 10th/90th percentile (lines originating from boxes), and single values are included as black dots. *P < seizure activity was observed in the anesthetized animals. Excessive oxygen supply during up to 4 ATA had no effect on the CBF responses to functional activation or CSD. Neither the dynamic course (Figure 4C) nor the mean amplitude (Figure 4D) of the blood flow increase during functional activation showed any difference between baseline condition,, and - in any of the groups tested. In addition, corpuscular blood volume changes to functional activation remained constant during, as shown by optical spectroscopy at 3 ATA (Figure 4E). Even during the more pronounced metabolic stimulus of CSD, neither the dynamic time courses (typical example Figure 5A) nor the maximal increases of CBF (quantified results Figure 5C) changed during or after compared with their baseline conditions.

8 764 CBF [AU] control 3 ATA 3 ATA DC potential [mv] min DC potential amplitude normalized to baseline ATA group 4 ATA group * max. CBF response normalized to baseline response ATA group 4 ATA group Figure 5 No effect of hyperbaric hyperoxygenation () on neuronal activity and neurovascular coupling during cortical spreading depression. (A) Typical example of the laser-doppler flowmetry (LDF) (upper panel) and direct current (DC) potential (lower panel) changes in one animal following cortical spreading depression (CSD) elicitation by short-lasting application of KCl under baseline condition (left, blue line), during at 3 ATA (middle, orange line), and after (right, gray line). Amplitudes or shapes of DC potentials as well as amplitudes and time courses of CBF changes during and after did not differ from the respective recordings at baseline. Quantified data: (B) DC potential amplitudes as well as (C) stimulation-induced regional cerebral blood flow (rcbf) increases (peak LDF changes in absolute values in arbitrary units) remained unchanged during and after at 3 or 4 ATA. (B and C) Data are normalized to baseline values and are presented as box plots with median (line within the box), 25th/75th percentile (outer margin of the box) and 10th/90th percentile (lines originating from boxes), and single values are included as black dots. *P < Discussion In this study, we tested whether blood-borne mediators released from hemoglobin on deoxygenation might, at least partially, explain blood flow increases during increased neuronal activity and increased oxygen metabolism in the brain. To this end, we functionally activated the primary somatosensory cortex (SI) in rats by electrical sensory stimulation of the contralateral forepaw, or by elicitation of CSD, in a hyperbaric chamber filled with 100% O 2 at 3 or 4 ATA. Under these conditions, oxygen supply to brain tissue is provided entirely by physically dissolved oxygen, and no deoxygenation of hemoglobin occurs. Using LDF and optical spectroscopy, rcbf and rcbo were recorded through a closed cranial window overlying the SI cortex during normobaric normoxia, and after return to normobaric normoxia. Neuronal activity was monitored by recording of SEPs and DC potentials using epidural electrodes. Contrary to the hypothesis, did not change the rcbf or corpuscular blood volume responses to functional activation or the rcbf responses to CSD (Figures 4 and 5), and rcbf remained tightly coupled to neuronal activity during excessive oxygen availability. During Hyperbaric Hyperoxia, Physically Dissolved Oxygen Fully Supports Increases in Stimulation- Induced Oxygen Metabolism Under physiologic conditions, depending on hematocrit, per 100 ml blood around 20 ml O 2 are bound to hemoglobin and only 0.3 ml O 2 are physically dissolved. Under resting conditions, an overall amount of 5 to 6 ml O 2 /100 ml of blood is consumed in rat brain, as estimated by Vovenko (1999). These values largely depend on blood hemoglobin concentration, CBF, and CMRO 2 (cerebral metabolic rate of oxygen), and might vary considerably between different experimental setups. During hyperbaric hyperoxia at 3 ATA, systemic arterial po 2 reaches B2000 mm Hg. At partial pressures > 3 ATA, brain concentrations of oxygenated hemoglobin and thus microcirculatory hemoglobin saturation reach a maximum (Meirovithz et al, 2007), the amount of physically dissolved oxygen significantly increases up to 7 ml/100 ml blood, and tissue po 2 in the brain increases to more than 400 mm Hg (Demchenko et al, 2005; Jamieson and Van den Brenk, 1962). As shown by Demchenko et al (2005), tissue po 2 values of above 400 mm Hg were reached already at 3 ATA when hyperoxygenation-induced decrease of resting CBF was prevented by acetazolamide application,

9 which slightly elevated the arterial pco 2 up to 44 mm Hg. In this study, owing to spontaneous breathing in the hyperbaric chamber, arterial pco 2 and resting CBF were increased to values comparable with those achieved by Demchenko et al, (2005), suggesting an equally elevated tissue po 2 of > 400 mm Hg. This increase in tissue O 2 leads to a complete oxidation of cytochrome oxidase type aa3 (Hempel et al, 1977) and of mitochondrial NADH (Meirovithz et al, 2007). During, oxygen consumption during neuronal activation could therefore be matched (3 ATA) or even exceeded (4 ATA) by the amount of physically dissolved oxygen. To verify that hemoglobin deoxygenation in brain microcirculation was indeed prevented by hyperbaric hyperoxia in our setup, rcbo was recorded by optical spectroscopy in some animals. Deoxy-Hb remained unchanged in response to functional activation during in all animals tested at 3 or 4 ATA, showing that hemoglobin deoxygenation can be prevented by. When hemoglobin remains fully oxygenated, that is, clamped at 100%, no vasodilator release from hemoglobin, if existent at all, can occur. Functional Activation-Induced Neurovascular Coupling in the Brain Operates Independent of Hemoglobin Deoxygenation or Tissue-Based Oxygen Sensors Vasodilation to local tissue hypoxia may not be a direct effect of oxygen at the vascular wall. It has been shown that isolated cerebral arterioles only dilate to hypoxia when they are perfused with red blood cells (Dietrich et al, 2000). These findings suggest that erythrocytes release or bind vasodilators depending on the availability of perivascular oxygen, resulting in vasodilation or vasoconstriction. Indeed, several mechanisms of oxygen-dependent release of vasoactive substances have been proposed in the last decade, which might be responsible for the tight vascular coupling to oxygen metabolism and hence to activity and energy demand of the tissue. The common feature of these mechanisms is deoxygenation of hemoglobin during increased oxygen demand. As mentioned above, oxygen delivery to tissues occurs mainly through oxygen released from hemoglobin, as the amount of physically dissolved oxygen in blood is very low under normoxic conditions. During hemoglobin deoxygenation, either NO or ATP may be released, both of which are strong vasodilators (Cosby et al, 2003; Dietrich et al, 2000; Gonzalez-Alonso et al, 2002; Jia et al, 1996; Stamler et al, 1997). To functionally study hemoglobin deoxygenation-related mechanisms independent of specific mediators, the ideal approach is to investigate neurovascular coupling under conditions without conversion of oxy-hb to deoxy-hb. Our results showing unchanged CBF responses within the somatosensory cortex to sensory forepaw stimulation in anesthetized rats under strongly argue against a functional role of blood-borne vasodilators in neurovascular coupling. Before our findings that excessive oxygen supply has no effect on neurovascular coupling, it had been shown in human subjects that the CBF response to physiologic activation was not altered by moderate hypoxia (Mintun et al, 2001). Taken together, these results suggest that the mechanism of CBF increase during physiologic brain activation is not dependent on a shortage of oxygen. Our findings also argue against a tissue oxygen-sensing mechanism, potentially at the level of mitochondria, matching blood flow to metabolic demands of oxygen, as recently proposed (Gjedde, 2002). We have used a comparably strong forepaw stimulation that produced large increases in flow (60% to 80% increase in the 3 ATA group), because in the short period, extensive averaging of stimuli was not possible and large CBF responses can therefore reduce noise and increase the statistical power of our analysis. Therefore, one may wonder whether, during these strong hyperemic stimuli, different vasodilator mechanisms that overwhelm the purported hemoglobin-based mechanism are engaged. Owing to interindividual variations of CBF responses, a number of animals had smaller CBF responses. We have therefore reanalyzed the data of four animals ( at 3 or 4 ATA) with an average CBF response of 47% (Supplementary Figure S3, Supplementary Material), which is in good agreement with recent studies reporting blood flow responses in awake humans of 63% during visual or of 46% during motor stimulation (Donahue et al, 2009) or in anesthetized rats of 68% (fmri; Stefanovic et al, 2007) or 36% (laser speckle contrast imaging; Sakadzic et al, 2009) during electrical forepaw stimulation. In accordance with the results obtained from all animals, the CBF responses during in this subgroup of animals did not change compared with their respective baseline responses. This retrospective analysis showed no evidence that our results are confounded by the magnitude of the CBF responses. In addition, we have already shown in a previous pilot study that does not affect neurovascular coupling to forepaw stimulations with 20% to 30% increases in CBF (Leithner et al, 2005). It may be considered a limitation of our findings that we cannot completely exclude a role for Hb-S- or nitrite-derived NO or ATP under normoxic conditions that might be quantitatively compensated for by another pathway during. However, it seems unlikely that during one of the other known pathways or another still unknown pathway (Leithner et al, 2009) is more activated if hemoglobin deoxygenation-dependent vasodilators are blocked, leading to identical magnitude and time course of the CBF response. During, a slight increase of resting CBF was detected that was most likely attributable to a slight 765

10 766 increase in paco 2. In principle, the CBF response may be determined by the baseline resistance/blood flow. To rule out that the increased resting CBF altered the CBF response to functional activation, we have reanalyzed the data of five animals ( at 3 or 4 ATA) with small paco 2 changes and thus stable resting CBF during. As shown in Supplementary Figure S4, the CBF responses during in this subgroup of animals did not change compared with their respective baseline responses, which is in accordance with the results obtained from all animals. In summary, our data strongly argue against a neurovascular coupling mechanism depending on oxygen availability in the brain or blood. It is noted that this does not rule out matching oxygen delivery with tissue metabolism as an evolutionarily evolved function of neurovascular coupling, which may have developed as a feed-forward mechanism. Hemoglobin Deoxygenation Does Not Regulate the Blood Flow Response to High Metabolic Demand of Oxygen (CSD) The relatively small increase in oxygen metabolism within the somatosensory cortex of rats during functional activation (10% to 15%; Royl et al, 2008) may not be sufficient to release relevant amounts of the hemoglobin deoxygenation-dependent vasodilators. However, hemoglobin deoxygenation-based mechanisms may have a role in preventing tissue from moderate or severe hypoxia during intense metabolic stress. It was shown recently that the blood flow response in the heart and the diameter response of isolated thoracic aorta from the rabbit or mouse during severe hypoxia depend on O 2 -dependent delivery and release of NO from hemoglobin (Datta et al, 2004; Diesen et al, 2008; James et al, 2004). We therefore tested whether hemoglobin deoxygenation-mediated coupling of blood flow and oxygen metabolism might be operative during the much stronger metabolic stimulus of CSD in the cerebral cortex. The significant increase in energy need during CSD is reflected in a 200% to 300% increase in glucose metabolism (Gjedde et al, 1981; Nedergaard and Astrup, 1986) and a 50% increase in oxygen metabolism (Mayevsky and Weiss, 1991), suggesting that the high ATP demand during CSD is largely met by oxidative metabolism. Very recently, it was shown that despite a marked blood flow increase, CSD was associated with local depletion of tissue O 2 and marked neuronal swelling (Takano et al, 2007). In comparison with physiologic functional activation, CSD may therefore be even better suited to test for a hemoglobin deoxygenationbased mechanism of blood flow regulation in the brain. However, CBF responses to the intense metabolic stimulus of CSD did not change during. Evidence Against Astrocyte-Dependent Mechanisms Inducing Vasoconstriction Under Hyperoxia Within the last decade, a role for astrocytes in neurovascular coupling has been established. Astrocytic mediators released by glutamate uptake during neuronal activation may be involved in triggering the CBF response to functional activation (Filosa et al, 2004; Metea and Newman, 2006; Mulligan and MacVicar, 2004; Peng et al, 2002; Schummers et al, 2008; Zonta et al, 2003). Using a rat brain slice preparation, Gordon et al (2008) very recently showed a strong dependency of the astrocytic control of arteriolar diameters on brain tissue metabolism and oxygen supply. Their findings show that under conditions of low oxygen concentration (acsf bubbled with 20% O 2 ), astrocytes couple arteriolar tone to neuronal activation by prostaglandin E 2 -mediated and lactate- and adenosine-dependent vasodilation. In contrast, during conditions of high oxygen (acsf bubbled with 95% O 2 ), the same stimulus induced vasoconstriction. Applying these data to neurovascular coupling in vivo, a reduction in the CBF increase or even its conversion to a CBF decrease during tissue hyperoxygenation must be ulated. However, neither normobaric hyperoxia (Wolf et al, 1997) nor hyperbaric hyperoxia at 3 or 4 ATA (this study) was able to even reduce the normal vasodilation during functional activation in rats. It has to be kept in mind that studying blood flow regulation without blood flow (brain slice) excludes numerous potentially relevant effectors such as corpuscular and blood plasma flux- or flow-induced vascular tone. Conclusion In this study, we have shown that clamping brain hemoglobin oxygenation at 100% by up to 4 ATA does not change the CBF response to physiologic functional activation or transient tissue depolarization during CSD. Our results indicate that the regional coupling of blood flow in the brain to physiologic or pathophysiologic neuronal activation is independent of the level of oxygen transported to the tissue, and that shortage of oxygen is not a driving force for vasodilation during increased neuronal activity. Matching blood flow to neuronal activity in the rat brain must involve mechanisms other than hemoglobin deoxygenation-dependent delivery of vasodilators. Several mechanisms have been taken into consideration within the last decade such as neuronal or astrocytic products of synaptic signaling (K + or NO; Iadecola, 2004), activation of GABA interneurons (Kocharyan et al, 2007), or arachidonic acid metabolites from neurons (production of prostaglandin E2 by cyclooxygenase activation; Niwa et al, 2000) or astrocytes (production of epoxyeicosatrienoic acids by CYP450 epoxygenase; Peng et al, 2002). Together with previous findings on the lack of effect of hypoglycemia or hyperglycemia

11 on neurovascular coupling (Powers et al, 1996; Wolf et al, 1997), our data suggest that rcbf responses to functional activation are regulated in a feed-forward manner, and not by a shortage of energy substrates. Our results therefore strongly support the neuronal hypothesis of a neuronal or glial signal inducing vasodilation in an anticipatory manner. Conflict of interest The authors declare no conflict of interest. References Cosby K, Partovi KS, Crawford JH, Patel RP, Reiter CD, Martyr S, Yang BK, Waclawiw MA, Zalos G, Xu X, Huang KT, Shields H, Kim-Shapiro DB, Schechter AN, Cannon RO, Gladwin MT (2003) Nitrite reduction to nitric oxide by deoxyhemoglobin vasodilates the human circulation. Nat Med 9: Datta B, Tufnell-Barrett T, Bleasdale RA, Jones CJ, Beeton I, Paul V, Frenneaux M, James P (2004) Red blood cell nitric oxide as an endocrine vasoregulator: a potential role in congestive heart failure. Circulation 109: Demchenko IT, Luchakov YI, Moskvin AN, Gutsaeva DR, Allen BW, Thalmann ED, Piantadosi CA (2005) Cerebral blood flow and brain oxygenation in rats breathing oxygen under pressure. J Cereb Blood Flow Metab 25: Diesen DL, Hess DT, Stamler JS (2008) Hypoxic vasodilation by red blood cells: evidence for an S-nitrosothiolbased signal. Circ Res 103: Dietrich HH, Ellsworth ML, Sprague RS, Dacey RG, Jr (2000) Red blood cell regulation of microvascular tone through adenosine triphosphate. Am J Physiol Heart Circ Physiol 278:H Dirnagl U, Kaplan B, Jacewicz M, Pulsinelli W (1989) Continuous measurement of cerebral cortical blood flow by laser-doppler flowmetry in a rat stroke model. J Cereb Blood Flow Metab 9: Donahue MJ, Blicher JU, Ostergaard L, Feinberg DA, MacIntosh BJ, Miller KL, Gunther M, Jezzard P (2009) Cerebral blood flow, blood volume, and oxygen metabolism dynamics in human visual and motor cortex as measured by whole-brain multi-modal magnetic resonance imaging. J Cereb Blood Flow Metab 29: Filosa JA, Bonev AD, Nelson MT (2004) Calcium dynamics in cortical astrocytes and arterioles during neurovascular coupling. Circ Res 95:e73 81 Gaston BM, Hare JM (2003) Hemoglobin and nitric oxide. N Engl J Med 349:402 5 Gjedde A, Hansen AJ, Quistorff B (1981) Blood brain glucose transfer in spreading depression. J Neurochem 37: Gjedde A (2002) Cerebral blood flow change in arterial hypoxemia is consistent with negligible oxygen tension in brain mitochondria. NeuroImage 17: Gladwin MT, Schechter AN (2004) NO contest: nitrite versus S-nitroso-hemoglobin. Circ Res 94:851 5 Gonzalez-Alonso J, Olsen DB, Saltin B (2002) Erythrocyte and the regulation of human skeletal muscle blood flow and oxygen delivery: role of circulating ATP. Circ Res 91: Gordon GRJ, Choi HB, Rungta RL, Ellis-Davies GCR, MacVicar BA (2008) Brain metabolism dictates the polarity of astrocyte control over arterioles. Nature 456:745 9 Hempel FG, Jobsis FF, LaManna JL, Rosenthal MR, Saltzman HA (1977) Oxidation of cerebral cytochrome aa3 by oxygen plus carbon dioxide at hyperbaric pressures. J Appl Physiol 43:873 9 Iadecola C (2004) Neurovascular regulation in the normal brain and in Alzheimer s disease. Nat Rev Neurosci 5: Isbell TS, Sun C-W, Wu L-C, Teng X, Vitturi DA, Branch BG, Kevil CG, Peng N, Wyss JM, Ambalavanan N, Schwiebert L, Ren J, Pawlik KM, Renfrow MB, Patel RP, Townes TM (2008) SNO-hemoglobin is not essential for red blood cell-dependent hypoxic vasodilation. Nat Med 14:773 7 James PE, Lang D, Tufnell-Barret T, Milsom AB, Frenneaux MP (2004) Vasorelaxation by red blood cells and impairment in diabetes: reduced nitric oxide and oxygen delivery by glycated hemoglobin. Circ Res 94: Jamieson D, Van den Brenk A (1962) Measurement of oxygen tensions in cerebral tissues of rats exposed to high pressures of oxygen. J Appl Physiol 18: Jia L, Bonaventura C, Bonaventura J, Stamler JS (1996) S-nitrosohaemoglobin: a dynamic activity of blood involved in vascular control. Nature 380:221 6 Kocharyan A, Fernandes P, Tong X-K, Vaucher E, Hamel E (2007) Specific subtypes of cortical GABA interneurons contribute to the neurovascular coupling response to basal forebrain stimulation. J Cereb Blood Flow Metab 28: Leithner C, Sellien H, Rohrer B, Royl G, Megow D, Steinbrink J, Kohl-Bareis M, Dirnagl U, Lindauer U (2005) Allosteric release of nitric oxide from hemoglobin does not mediate neurovascular coupling. J Cereb Blood Flow Metab 25:S207 Leithner C, Royl G, Offenhauser N, Füchtemeier M, Kohl-Bareis M, Villringer A, Dirnagl U, Lindauer U (2009) Pharmacological uncoupling of activation induced increases in CBF and CMRO2. J Cereb Blood Flow Metab; e-pub ahead of print 30 September 2009 Mayevsky A, Weiss HR (1991) Cerebral blood flow and oxygen consumption in cortical spreading depression. J Cereb Blood Flow Metab 11: Meirovithz E, Sonn J, Mayevsky A (2007) Effect of hyperbaric oxygenation on brain hemodynamics, hemoglobin oxygenation and mitochondrial NADH. Brain Res Rev 54: Metea MR, Newman EA (2006) Glial cells dilate and constrict blood vessels: a mechanism of neurovascular coupling. J Neurosci 26: Mintun MA, Lundstrom BN, Snyder AZ, Vlassenko AG, Shulman GL, Raichle ME (2001) Blood flow and oxygen delivery to human brain during functional activity: theoretical modeling and experimental data. Proc Natl Acad Sci 98: Mulligan SJ, MacVicar BA (2004) Calcium transients in astrocyte endfeet cause cerebrovascular constrictions. Nature 431:195 9 Nedergaard M, Astrup J (1986) Infarct rim: effect of hyperglycemia on direct current potential and [14C]2- deoxyglucose phosphorylation. J Cereb Blood Flow Metab 6: Niwa K, Araki E, Morham SG, Ross ME, Iadecola C (2000) Cyclooxygenase-2 contributes to functional hyperemia in whisker-barrel cortex. J Neurosci 20:

12 768 Patel R, Townes T (2008) SNO-hemoglobin and hypoxic vasodilation. Nat Med 14: Peng X, Carhuapoma JR, Bhardwaj A, Alkayed NJ, Falck JR, Harder DR, Traystman RJ, Koehler RC (2002) Suppression of cortical functional hyperemia to vibrissal stimulation in the rat by epoxygenase inhibitors. Am J Physiol Heart Circ Physiol 283:H Powers WJ, Hirsch IB, Cryer PE (1996) Effect of stepped hypoglycemia on the regional cerebral blood flow response to physiologic brain activation. Am J Physiol Heart Circ Physiol 270:H554 9 Royl G, Füchtemeier M, Leithner C, Megow D, Offenhauser N, Steinbrink J, Kohl-Bareis M, Dirnagl U, Lindauer U (2008) Hypothermia effects on neurovascular coupling and cerebral metabolic rate of oxygen. NeuroImage 40: Sakadzic S, Yuan S, Dilekoz E, Ruvinskaya S, Vinogradov SA, Ayata C, Boas DA (2009) Simultaneous imaging of cerebral partial pressure of oxygen and blood flow during functional activation and cortical spreading depression. Appl Opt 48:D Schechter AN, Gladwin MT (2003) Hemoglobin and the paracrine and endocrine functions of nitric oxide. N Engl J Med 348: Schummers J, Yu H, Sur M (2008) Tuned responses of astrocytes and their influence on hemodynamic signals in the visual cortex. Science 320: Singel DJ (2003) Hemoglobin and nitric oxide. N Engl J Med 349:402 5 Stamler JS, Jia L, Eu JP, McMahon TJ, Demchenko IT, Bonaventura J, Gernert K, Piantadosi CA (1997) Blood flow regulation by S-nitrosohemoglobin in the physiological oxygen gradient. Science 276: Stamler JS (2003) Hemoglobin and nitric oxide. N Engl J Med 349:402 5 Stamler JS, Singel DJ, Piantadosi CA (2008) SNO-hemoglobin and hypoxic vasodilation. Nat Med 14: Stefanovic B, Schwindt W, Hoehn M, Silva AC (2007) Functional uncoupling of hemodynamic from neuronal response by inhibition of neuronal nitric oxide synthase. J Cereb Blood Flow Metab 27: Takano T, Tian G-F, Peng W, Lou N, Lovatt D, Hansen AJ, Kasischke KA, Nedergaard M (2007) Cortical spreading depression causes and coincides with tissue hypoxia. Nat Neurosci 10: Vovenko E (1999) Distribution of oxygen tension on the surface of arterioles, capillaries and venules of brain cortex and in tissue in normoxia: an experimental study on rats. Pflugers Arch 437: Wolf T, Lindauer U, Villringer A, Dirnagl U (1997) Excessive oxygen or glucose supply does not alter the blood flow response to somatosensory stimulation or spreading depression in rats. Brain Res 761:290 9 Zonta M, Angulo MC, Gobbo S, Rosengarten B, Hossmann KA, Pozzan T, Carmignoto G (2003) Neuron-to-astrocyte signaling is central to the dynamic control of brain microcirculation. Nat Neurosci 6:43 50 Supplementary Information accompanies the paper on the Journal of Cerebral Blood Flow & Metabolism website (