Transient Hypoperfusion to Ischemic/Anoxic Spreading Depolarization is Related to Autoregulatory Failure in the Rat Cerebral Cortex

Old, adult, male Sprague–Dawley rats (Charles River Laboratories, 18 months old, 645 ± 100 g, n = 12) were used in this study. Standard rodent chow and tap water were supplied ad libitum. The animals were housed under constant temperature, humidity, and lighting conditions (23 °C, 12:12 h light/dark cycle, lights on at 7 a.m.). Animals were anesthetized with 1.5–2% isoflurane evaporated with N₂O:O₂ (70% and 30%, respectively) and allowed to breathe spontaneously through a head mask throughout the experiment. Body (core) temperature was kept at 37 °C with a feedback-controlled heating pad. Atropine (0.1%, 0.1 ml) was administered intramuscularly shortly before surgical procedures to avoid the production of airway mucus. Mean arterial blood pressure (MABP) was monitored continuously and arterial blood gases were checked regularly (i.e., during baseline, under ischemia, and under anoxia) via a catheter inserted into the left femoral artery. Level of anesthesia was controlled with the aid of MABP displayed live as experiments were in progress.

A midline incision was made on the neck and both common carotid arteries were gently separated from the surrounding tissue and the vagal nerves. Lidocaine (1%) was applied topically before opening the tissue layers during preparation. A silicone coated fishing line used as occluder was looped around the common carotid arteries for the later induction of cerebral ischemia.

For electrophysiological data acquisition (n = 11), two craniotomies (4–5 mm apart) were prepared in the right parietal bone by using a dental drill. The dura in each craniotomy was carefully removed, and the exposed brain surface was regularly rinsed with artificial cerebrospinal fluid (mM concentrations: 126.6 NaCl, 3 KCl, 1.5 CaCl₂, 1.2 MgCl₂, 24.5 NaHCO₃, 6.7 urea, and 3.7 glucose bubbled with 95% O₂ and 5% CO₂ to achieve a constant pH of 7.4). For imaging (n = 1), the rostral cranial window was enlarged (4.5 × 4.5 mm) and closed by a microscopic cover glass to serve in vivo optical imaging, as described earlier [23]. The caudal craniotomy was created to have the opportunity to trigger SD with the topical application of 1 M KCl, should no spontaneous depolarization occur [24]. Retrospectively, this proved to be an unnecessary precaution, as all SDs analyzed in this study evolved spontaneously.

Isoflurane anesthesia for data acquisition was adjusted to and kept at 1.0–1.2%. Incomplete global forebrain ischemia was induced by the permanent occlusion of both common carotid arteries (two-vessel occlusion [2VO]), which readily gave rise to a single spontaneous SD (ischemic SD [iSD]). Oxygen was withdrawn from the anesthetic gas mixture 20–40 min later, creating a brief episode of anoxia, which produced anoxic/ischemic depolarization (aSD). The model has been intended to recapitulate the conversion of the ischemic penumbra to the irreversibly injured core region by the passage of the second SD (i.e., the aSD). The second event (the aSD) was triggered to worsen the metabolic challenge and promote the transformation of penumbra-like tissue to a severely injured state [25]. The condition of anoxia was resolved within 5 min by reoxygenation to avoid cardiac arrest, and then data acquisition was continued for 25 min.

Local field potential (LFP), [K⁺]_e, and local changes to the CBF were acquired from the rostral craniotomy. Potassium sensitive microelectrodes were prepared, as reported earlier [20]. In brief, glass capillary microelectrode tips (outer diameter 10–12 μm) were filled with a liquid K⁺-ion exchanger membrane (Potassium ionophore I, cocktail A; Sigma), and the microelectrode shank was backfilled with 100 mM of KCl solution. Microelectrodes were calibrated in standard solutions of known K⁺ concentrations (1, 3, 5, 10, 30, 50, and 100 mM). K⁺-sensitive microelectrodes were lowered into the cortex together with microelectrodes (tip diameter = 20 μm) filled with 150 mM of NaCl and 1 mM of HEPES (4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid). The signal of these microelectrodes was filtered in direct current (DC) potential mode (< 1 Hz) and served as the reference LFP signal for the K⁺-sensitive microelectrodes. An Ag/AgCl electrode placed under the skin of the animal’s neck was used as common ground. Microelectrodes were connected to custom made dual-channel electrometers (including AD549LH; Analog Devices, Norwood, MA) via Ag/AgCl leads. The reference electrode signal (i.e., the LFP) was subtracted from that of the K⁺-sensitive microelectrode by dedicated differential amplifiers and associated filter modules (NL106 and NL125; NeuroLog System; Digitimer Ltd, United Kingdom), which yielded potential variations related to changes in [K⁺]_e. The recorded analog signals were converted and displayed live using a dedicated analog-to-digital converter (MP 150, Biopac Systems, Inc) at a sampling frequency of 1 kHz. Changes of [K⁺]_e were expressed in mV to be translated into mM concentration offline, using least squares linear regression [20].

CBF changes were monitored with a laser Doppler needle probe (Probe 403 connected to PeriFlux 5000; Perimed AB, Sweden) positioned next to the penetration site of the K⁺-sensitive microelectrode. The flow signal was digitized and displayed together with the DC potential and K⁺ signals, as described above (MP 150 and AcqKnowledge 4.2.0; Biopac Systems Inc, CA). The completed preparations were enclosed in a Faraday cage.

A voltage-sensitive dye (VS-dye, RH-1838; Optical Imaging Ltd, Rehovot, Israel) was dissolved in artificial cerebrospinal fluid and circulated in the closed cranial window to saturate the cortical tissue, as reported earlier [23]. For live imaging, the cortex was illuminated in stroboscopic mode (100 ms/s) with a high-power light emitting diode (625 nm peak wavelength; SLS-0307-A; Mightex, Pleasanton, CA, equipped with an emission filter 620–640 nm bandpass; 3RD620-640; Omega Optical Inc, Brattleboro, VT) and with a laser diode (HL6545MG; Thorlabs Inc, NJ; 120 mW; 660 nm emission wavelength) driven by a power supply (LDTC0520; Wavelength Electronics Inc, Bozeman, MT) set to deliver a 160-mA current [26]. Two identical CCD (Charged Coupled Device) cameras (resolution 1024 × 1024 pixels, Pantera 1M30; DALSA, Gröbenzell, Germany) were attached to a stereomicroscope (MZ12.5; Leica Microsystems, Wetzlar, Germany) equipped with a 1:1 binocular/video-tube beam splitter. The VS-dye fluorescence was captured with a camera equipped with a band pass filter (3RD 670–740; Omega Optical Inc). To create CBF maps by laser speckle contrast analysis (LASCA), raw speckle images were captured by the second CCD camera (1 frame/second; 2 ms for illumination and 100 ms for exposure). A dedicated program written in LabVIEW environment synchronized the illumination and camera exposures. CBF maps were generated from the corresponding raw images, as reported earlier [24]. Changes in VS-dye fluorescence intensity and CBF with time were extracted by placing regions of interest (ROIs) of 19 × 19-pixel size (~ 70 × 70 μm) at selected sites on the surface of cerebral cortex.

All variables (i.e., LFP, [K⁺]_e, laser doppler flow signal, and MABP) were simultaneously acquired, displayed live, and stored using a personal computer equipped with a dedicated software (AcqKnowledge 4.2 for MP 150; Biopac Systems Inc, CA). Data analysis was conducted offline and was assisted by the inbuilt tools of AcqKnowledge 4.2 software. Three rats were excluded from analysis because of artifacts on the [K⁺]_e or the DC potential trace.

To characterize iSD and aSD, several variables were measured in the DC potential and the synchronous [K⁺]_e traces. Note that the DC potential signature of SD can be used as a surrogate marker of [K⁺]_e changes, and the two traces were taken into consideration (as confirmation) for variables expressed in time (e.g., event latency) or to determine aSD onset (Fig. 1a). First, the latency of the SD events with respect to 2VO or anoxia initiation was determined. The onset of SD superimposed on a negative DC potential drift was taken at a sharp (> 1 mV/s) negative shift that occurred within minutes after ischemia or anoxia induction (Fig. 1a). Next, the peak amplitude of the events, the duration at half amplitude, the rates of depolarization, and the concomitant rise of [K⁺]_e were taken, as previously reported [23]. The rate of the slow DC potential (mV/s) and [K⁺]_e (mM/s) drift prior to aSD was calculated from the deflection from the level baseline to the onset of aSD. Finally, the amplitude of the hypoperfusion transients with iSD and aSD, as well as the duration of the hypoperfusion transients at half amplitude, were measured.

The status of cerebral autoregulation during the CBF responses to iSD and aSD were determined by calculating the cerebrovascular autoregulatory index (rCBFx) in MATLAB with Signal Processing and Image Processing Toolboxes (R2018b; MathWorks, Natwick, MA), as described in detail elsewhere [21]. Briefly, CBF and MABP were downsampled to 1 Hz. Then, consecutive 8-s averages were calculated with a sixth order median filter for each to eliminate high-frequency noise. Afterward, a Pearson correlation coefficient was calculated for every 40 data points yielding rCBFx in a temporal resolution of 320 s. Then, discrete data were transformed to continuous variable using linear regression. Section averages (e.g., baseline, before, and during SD) were measured on this time series of 1 Hz resolution. Intact cerebrovascular autoregulation was indicated by rCBFx < 0.3 [27]. Local CBF variations obtained by LDF or LASCA were expressed relative to baseline by using the average CBF value of the first 5 min of baseline (100%) and the recorded biological zero obtained after terminating each experiment (0%).

Quantitative data are given as mean ± standard deviation. Statistical analysis was conducted with the software SPSS (IBM SPSS Statistics for Windows, Version 22.0; IBM Corp, Armonk, NY). Data sets were evaluated by a one-way analysis of variance, followed by a Bonferroni post hoc test when appropriate. Level of significance was set at p < 0.05* or p < 0.01**. Appropriate statistical methods are provided in each Figure legend in detail.

Successful ischemia induction was confirmed by a sudden drop of CBF to 22.2 ± 11.4% of baseline shortly after 2VO. CBF then stabilized around 37.8 ± 19.8% prior to anoxia initiation and was sustained at 33.5 ± 21.3% following the insult. Partial pressure of arterial blood O₂ (pO₂) varied in the normal range during ischemia (117.5 ± 31.7 vs. 107.7 ± 23.1 mm Hg, ischemia vs. baseline), and profoundly decreased to 34.4 ± 10.5 mm Hg under the episode of anoxia. The partial pressure of arterial CO₂ (pCO₂) varied within the physiological range throughout the experimental protocol (35.4 ± 12.9 vs. 41.3 ± 4.1 vs. 37.37.8 ± 4.0 mm Hg, anoxia vs. ischemia vs. baseline; f = 1.633, p = 0.211). MABP varied around 86 ± 19 mm Hg during baseline and displayed a gradual elevation over the ischemic period to 99 ± 18 mm Hg. The subsequent anoxia was promptly followed by a considerable reduction of MABP to 64 ± 24 mm Hg, whereas reoxygenation produced a transient MABP increase more than baseline, peaking at 105 ± 24 mm Hg. MABP then settled around 96 ± 22 mm Hg.

Spreading depolarization events occurred spontaneously in response to ischemia and anoxia initiation (Fig. 1a). The latency with respect to the onset of the triggering insult proved to be longer for aSD than for iSD (277 ± 84 vs. 155 ± 33 s, respectively) (Fig. 1b). The onset of iSD was indicated by a sharp rise of [K⁺]_e and a concomitant sudden drop of the DC potential. In contrast, aSD was preceded by and started off from a slow, gradual elevation of [K⁺]_e, (relative amplitude 11.8 ± 2.09 mM, duration 166 ± 46 s), accompanied by a matching negative drift of the DC potential (relative amplitude 2.5 ± 1.3 mV) (Fig. 1a, gray shading). Because of the slow elevation of [K⁺]_e, aSD took off from a considerably higher [K⁺]_e than iSD (19.8 ± 1.7 vs. 11.7 ± 1.7 mM, respectively), whereas the absolute peak amplitude of the [K⁺]_e and DC potential shift with aSD and iSD were similar ([K⁺]_e 47.3 ± 4.1 and 44.1 ± 3.7 mM, respectively) (Fig. 1c). The rate of depolarization was significantly slower with aSD than with iSD ([K⁺]_e 0.62 ± 0.38 vs. 2.08 ± 0.53 mM/s, respectively; DC potential: − 0.55 ± 0.47 vs. − 1.17 ± 0.74 mV/s, respectively). Finally, aSD lasted characteristically longer than iSD (516 ± 311 vs. 208 ± 191 s, respectively) (Fig. 1d).

Laser Doppler flowmetry in combination with electrophysiology showed that iSD and aSD were associated with a brief hypoperfusion (Fig. 2a). The CBF minimum reached during the event-associated hypoperfusion transient was significantly lower with aSD than with iSD (7.4 ± 3.6 vs. 11.9 ± 2.9%, respectively) (Fig. 2b), although they were of similar duration (202 ± 113 and 189 ± 80 s, respectively). Detailed analysis of CBF changes in relation to MABP revealed meaningful differences between the flow variations with iSD and aSD. In particular, the reduction of CBF was independent of MABP with iSD, whereas the drop of CBF tightly followed that of MABP reduction with aSD (Fig. 2c). Cerebral autoregulation, which was functional before and under ischemia, was weakening after anoxia initiation prior to aSD. This was reflected by rCBFx (0.27 ± 0.14) approaching the threshold of autoregulatory failure (0.3) (Fig. 2c). Autoregulatory failure became obvious shortly before and under aSD, which was substantiated by rCBFx (i.e., 0.61 ± 0.24 during aSD) exceeding 0.3, taken as the upper limit of intact autoregulation [21]. With iSD, CBF dropped in the face of increasing MABP, yielding negative rCBFx values (− 0.38 ± 0.14 during iSD).

At a high resolution, the sequence of iSD and aSD and the associated flow reduction proved to be opposing. While the CBF drop occurred at a short delay after iSD (18 ± 4 s), CBF reduction preceded aSD by minutes (− 162 ± 33 s) (Fig. 3a–c). A series of CBF maps obtained with LASCA (Fig. 3d) and the spatio temporally matching VS-dye image sequence (Fig. 3e), confirmed that the hypoperfusion transient with aSD preceded the depolarization, rather than evolving as a consequence of aSD, and was simultaneous in the field of view (Fig. 3f). Any subsequent flow increase appeared to be also passive and follow the elevation of MABP minutely.