All animal experiments were performed in accordance with relevant guidelines and regulations, and were approved by Hacettepe University Animal Experiments Ethics Committee (2012/53-01, 2017/05-2). A total of 140 male and female Swiss albino mice (25 to 35 g) were used. All mice were housed with ad libitum access to food and water under a fixed 12-hour light/12-hour dark cycle.
Mice were anesthetized with intraperitoneal injection of xylazine (10 mg/kg) and urethane (1.25 g/kg), which suppresses cortical excitability relatively less compared to commonly used anesthetics like isoflurane and ketamine [16, 17]. After maintaining an adequate depth of anesthesia, mice were positioned prone and fixed in stereotaxic frame (Digital Lab Standard Stereotaxic Frame, Stoelting, USA). In photic stimulation experiments, eyes were covered by lubricant gel and eyelids were closed with surgical clips to avoid corneal drying and enhance cortical sensitivity to light . The heart rate and tissue oxygen saturation were continuously monitored by a pulse oximeter (V3304 Digital Table-Top Pulse Oxymeter, Nonin, USA) throughout all experiments. 100% oxygen mixed with room air (2 l/min) was delivered to the spontaneously breathing mice to avoid hypoxia. Rectal temperature was maintained at 37.0±0.2°C using a homeothermic blanket (Harvard Apparatus, UK) for the duration of the experiment. The depth of anesthesia was checked with toe/tail pinch and/or eye blink reflex at 10-15 minute intervals and additional doses of anesthetics were injected when needed. Animals with persistent hemodynamic instability or hypoxia were excluded from the study.
A midline incision was made over the scalp to expose the skull. One or two small circular areas (diameter<1 mm) were thinned with a high-speed drill (WPI, USA) to house the electrode tips over the cranium covering the right hemisphere; one at the anterior parietal and the other one at the frontal region. During drilling process, the skull was continuously irrigated with cold saline to prevent thermal injury to the underlying cortex.
A cranial window (2 mm in diameter) was opened over the cranium covering either the visual cortex (for photic stimulation) or somatosensory barrel cortex (for whisker stimulation) without damaging the underlying dura mater. In a subset of experiments, a plastic cylindrical tube (inner diameter and height, 2 mm) was placed over the exposed intact dura and fixed to the skull by dental acrylic to encircle the cranial window. This chamber allowed topical administration of ouabain or the fluorescent potassium probe (Asante Potassium Green-4) as well as fluorescent imaging.
Photic and Whisker Stimulation
A custom-made photic stimulator was used to illuminate both eyes. The frequency, intensity, duration and pattern of the stimulation were adjustable using a stimulus isolator (World Precision Instruments A385, USA) and a computer interface (LabChart, AD Instruments, Australia). 5-mm, 2V white mini-LED bulbs were used for stimulation. The LEDs were mounted on flexible arms that allowed optimal positioning of the LEDs for each individual mouse in stereotaxic frame.
Whiskers were cut to 1 cm and automatically stimulated by using a custom-made apparatus. This device allowed stimulation of the whiskers unilaterally with a tip blunted watercolor brush (no:8) in vertical plane at an adjustable frequency (4 to 30 Hz) without touching common fur or other parts of the face of the mouse placed in the stereotaxic frame.
Laser Speckle Contrast Imaging
Cerebral blood flow (CBF) changes were detected with laser speckle contrast imaging as described before . Briefly, the region of interest was illuminated with a 785-nm laser diode (Thorlabs DL7140-201, Thorlabs, USA) and imaged under 4X magnification by using a stereomicroscope (Nikon SMZ 1000, Nikon, Japan) and a charge-coupled device camera (Basler 602F, Basler Vision Technologies, Germany). Consecutive raw speckle images were acquired at 100 Hz (an image set) at 1-second intervals, processed by computing the speckle contrast using a sliding grid of 7x7 pixels, and averaged to improve the signal-to-noise ratio. Speckle contrast images were converted to images of correlation time values, which are inversely and linearly proportional to the mean blood velocity. Image J 1.42 software (NIH, USA) was used to compute and pseudo-color the relative blood flow changes in the cortex after sensory stimulation compared to baseline values.
In a subset of experiments, fluorescent-tagged ouabain (BODIPLY® FL Ouabain, Molecular Probes, USA) was topically applied over the dura by a cotton ball soaked with 5 μl of 0.1 mM ouabain in saline. Animals were sacrificed and the brains were removed after 60 minutes. 20 μm-thick coronal sections were obtained from the frozen brains. Sections were cover-slipped with Hoechst-33258 to delineate the tissue architecture by staining cell nuclei and, imaged under a fluorescence microscope using appropriate filter sets to assess the penetration and diffusion of ouabain into the brain tissue. In preliminary experiments (n=3 mice), we also evaluated the distribution of fluorescent-tagged ouabain after intracerebroventricular (icv) administration (5 μL, 100 μM) to see whether this route could provide sufficient concentrations in the cortex. However, 45 minutes after intracerebroventricular administration ouabain, only accumulated in the vicinity (around 100 μm) of the ventricle and did not diffuse to cortex so, this route of administration was not preferred.
To record the DC potentials, one or two Ag/AgCl pellet electrodes (E205, 1 mm diameter, Warner Instruments, USA) were placed and fixed over the thinned skull as described above. The electrode tips were covered with EEG electrode gel to enhance the electrical contact with the cranial bone. A disc-shaped Ag-AgCl ground electrode (E242, 4 mm diameter, Warner Instruments, USA) was placed between the neck muscles. The signals were digitized and acquired at 1 Hz sampling rate, displayed and analyzed by PowerLab 16/35 data acquisition system (AD Instruments).
CSD induction with sensory stimulation
For investigating the temporal relationship between the photic stimulation and CSD induction, an intermittent stimulation protocol was preferred to avoid desensitization to light. Furthermore, the room was kept dark and the eyelids were closed starting with induction of anesthesia until the stimulation in order to increase light sensitivity of the visual cortex . Around-the-threshold concentration (0.1 mM) of ouabain was topically applied over the visual cortex to prime it for CSD generation. Following a 30-minute CSD-free silent period in these mice, the eyes were opened and 1-minute on-off cycles of photic stimulation (8-12 Hz) were started until a CSD was evoked or for a maximum duration of 10 minutes (5 cycles). If no CSDs were evoked, then the same stimulation pattern was repeated 10 min after the last stimulus. In case of CSD generation, the next stimulation was initiated after a 30-min silent period has passed. Eyes were closed again right after a CSD was evoked until the next stimulation or during the 10-min pauses between the photic stimulation epochs. The experiment was terminated in 5 mice (out of 22) if the experiment duration exceeded 4 hours (n=2) or repetitive CSDs were observed during 2-hour recording without a 30-min silent period between them (n=2) or the animal was hemodynamically unstable (n=1). The effect of photic stimulation on a group of α2-Na+/K+-ATPase knockdown mice was also evaluated with the same protocol.
Whisker stimulation experiments were carried out with the same protocol to the one applied for photic stimulation. For these experiments, ouabain soaked cotton ball was placed over the somatosensory barrel cortex and only female mice were used to increase the CSD susceptibility because their CSD threshold is lower . Contralateral whiskers were stimulated as detailed above. The effect of whisker stimulation on a group of α2-Na+/K+-ATPase knockdown mice was also evaluated with the same protocol.
pLL 3.7 plasmid expressing α2-Na+/K+-ATPase-shRNA was kindly provided by Dr. Gilbert Gallardo (Washington University, School of Medicine). α2-Na+/K+-ATPase-shRNA consists of 21-nucleotide inverse repeats (sense sequence: 5'-GTG GCA AGA AGA AAC AGA AAC-3') separated by a 9-nucleotide loop sequence (CAAGTTAAC). The shRNA construct in the vector was verified before use by sequencing. We used empty pLL 3.7 plasmid as a control. A total of 24 mice injected with shRNA and 9 mice with control blank plasmid.
α2-Na+/K+-ATPase-shRNA expressing plasmid or control plasmid (1 μg) was mixed with 0.12 μl in vivo-JetPEI-TM (Polyplus transfection, France) transfection reagent. Then 1 μl of this mixture was injected at two different points intracortically 1 mm deep at a rate of 0.1 μl/min under isoflurane anesthesia by a 26 gauge Hamilton syringe. The needle was kept at the injection site for 1-minute and removed slowly over 5 minutes to allow diffusion of the plasmids into cortical layers. The sites of injection in the CSD threshold (n=6/group) and photic stimulation experiments were in the right visual cortex; (-3, 1.5) mm and (-3, 3.5) mm relative to bregma (anteroposterior and lateral coordinates, respectively) (n=6). In whisker stimulation and K+ fluoroprobe imaging experiments, the sites of injection were in the right barrel cortex; (-0.5, 3.5) mm and (-1.8, 3) mm relative to bregma (n=6).
To verify knockdown with qRT-PCR, mice were sacrificed at 6, 24 and 48 hours after shRNA delivery. The injection areas and the contralateral homologous areas were removed and RNAs were extracted with RNeasy Mini Kit (QIAGEN, GERMANY, cat no: 74104) according to instructions. Eluted RNAs were stored at -80°C. cDNA synthesis was performed with random hexamer primers with RevertAid First Strand cDNA Synthesis Kit (Thermo Fisher Scientific, USA cat no: K1621) according to instructions. cDNAs were stored at -20°C. qRT-PCR was performed with Taqman probe based technology. Taqman gene expression master mix (ABI, USA cat no: 4369016), FAM-MGB labeled Taqman probes for mouse ATP1A2 gene (Assay ID: Mm00617899_m1) and mouse GAPDH gene (Assay ID: Mm9999991) were used. cDNAs were diluted to 1:32 ratio in nuclease free water. The PCR was carried out in triplicates in ABI OneStep Q RT PCR machine (ABI, USA). The relative expression values were calculated with ΔΔCt method.
Detection of CSD threshold
To investigate the CSD threshold after knockdown, plasmid injected animals were tested either at 6 h, 24 h, or 48 h after injection. Increasing concentrations of KCl-adsorbed cotton balls were consecutively applied epidurally through the burr hole over the parietal bone. Each cotton ball was allowed to stay for 5 min. The DC potential was continuously recorded over the right visual cortex and the cotton ball was replaced with the next higher concentration if no CSD wave was observed during 5 min. Concentrations of KCl starting from 0.05 M and increasing by 0.025 M at each step were used to detect the threshold. Once CSD was induced, the experiment was ended. The minimum concentration that yielded the first CSD was accepted as the threshold.
Monitoring extracellular K+ in vivo with a fluoroprobe
In order to investigate the extracellular potassium changes, a fluorescent potassium probe, Asante Potassium Green-4 (IPG-4, formerly APG-4, TefLABs, USA) was used. IPG-4 has a higher selectivity for K+ over Na+ than previous Asante potassium dyes used in vivo. IPG-4 was prepared in 2% DMSO and diluted in artificial cerebrospinal fluid (aCSF) to yield a final concentration of 250 μM. Since it was poorly soluble in aqueous media, ultrasonic bath and gentle heating was applied to increase its solubility on the day of each experiment.
Before using for stimulation experiments, the potassium selectivity of IPG-4 was characterized in order to assess its compatibility with the large ionic shifts during CSD. For this, we applied three different concentrations of K+ (0, 10, 50 mM) in the presence of high Na+ concentrations in distilled water. Fluorescence intensity showed gradual increase with rising concentrations of K+ and was similar in the presence of 77 mM or 154 mM Na+ in the medium. Next, we investigated the distribution of IPG-4 in the mouse brain following icv (n=2) injection or topical epidural application (n=3) under an upright fluorescent microscope. Following icv injection, dye was largely accumulated in the periventricular area and did not reach the cortex. However, after topical application through a chamber over the exposed dura, IPG-4 diffused into superficial cortical layers down to 200 μm, which was satisfactory because fluorescence microscopy is limited to image only the superficial 50-70 μm of the cortex in vivo. Therefore, the epidural application was used for intravital imaging of extracellular potassium dynamics in the cortex.
To detect the extracellular K+ change in vivo, a baseline image was captured for obtaining autofluorescence intensity of the tissue and the closed cranial chamber was filled with 8-10 μl of 250 μM IPG-4. The room was darkened during 30-min dye incubation. Then the dye was removed and the chamber was gently rinsed with aCSF three times at 37°C. Chamber was filled completely with aCSF to avoid any air bubble formation and closed by a clean cover glass for optimal fluorescence microscopy. The images were acquired under a Nikon SMZ 1000 stereomicroscope with a fluorescent attachment by using a CCD camera (Nikon DS-Qi1Mc, Japan) and NIS Elements Advanced Research Software v3.2. We used a fluorescence filter (HQ FITC LP) with excitation: 480/40 and emission: >510 nm, an exposure time of 250 ms and a signal averaging of (4x) for acquisition. Each imaging session lasted 7 minutes. Time-lapse recording for every 5 seconds started 1 min before whisker stimulation, continued through the 5 min of stimulation and ended 1 min after.
For the analysis of IPG-4 fluorescence, an ROI was placed in a predefined area that showed the maximum hyperemic response with whisker stimulation in preliminary experiments. The mean fluorescent intensity of the baseline image was subtracted from the time series images. A rolling average by 10 was applied to the image sequence. By using the time measurement feature of Nikon NIS-AR software, we acquired a set of signal intensity values for each session of whisker stimulation. This data set was exported for further processing and calculations to MATLAB (Mathworks, USA) software. In MATLAB (Mathworks, USA), the baseline drift that frequently accompanied the IPG-4 fluorescence was corrected by detrending and then the percent change from baseline (dF/F0 %) was calculated. To incorporate both the duration and the amplitude of stimulation-induced signal change into analysis, we calculated area-under-curve (AUC) and compared AUC between groups.
Ouabain octahidrate (Sigma-Aldrich, USA), a Na+/K+-ATPase inhibitor, was topically applied with a micropipette at a total volume of 5 μL in saline either onto the cotton ball placed over the intact dura or, for IPG-4 experiments, into the cranial window. Electrophysiological recordings were started before ouabain applications and continued 150 minutes to monitor CSD occurrence. To determine the threshold dose for CSD induction, ouabain solutions at different concentrations (0.05, 0.1, 0.15, 0.3, 0.5 mM) were tested. The time to CSD induction and CSD frequency were measured for each tested concentration in a fixed volume (5 μl).
8-Cyclopentyl-1,3-dipropylxanthine (DPCPX, Tocris, UK), a high affinity adenosine A1 receptor selective antagonist, was dissolved in DMSO as a stock solution (30 mM). On the day of each experiment, the drug was suspended in saline and gently heated at a concentration of 30 μM for intracortical injection or 0.7-1 mM for epidural application with a cotton ball.
(+)Bicuculline (Tocris, UK), a competitive antagonist of GABAA receptors, was dissolved in 1% DMSO at a concentration of 5-10 mM and applied epidurally with a cotton ball.
Data were presented as mean ± SE or median (range) or percentage (%) of the total and were analyzed using chi-square test or Mann-Whitney U test, where appropriate. To differentiate sensory stimulation induced CSDs from spontaneously occurring CSDs, a modified version of the method previously used by von Bornstadt et al. was adopted . The statistical method used to analyze each dataset is indicated in respective figure legends.