Cerebral Cortex, April 2016;26: 1580–1589
doi: 10.1093/cercor/bhu336
Advance Access Publication Date: 16 January 2015
Original Article
ORIGINAL ARTICLE
Cytoarchitecture-Dependent Decrease in Propagation
Velocity of Cortical Spreading Depression in the Rat
Insular Cortex Revealed by Optical Imaging
Satoshi Fujita1,2, Naoko Mizoguchi1,3, Ryuhei Aoki1,4, Yilong Cui5,
Noriaki Koshikawa1,2, and Masayuki Kobayashi1,2,5
1
Department of Pharmacology, 2Division of Oral and Craniomaxillofacial Research, Dental Research Center,
Nihon University School of Dentistry, 1-8-13 Kanda-Surugadai, Chiyoda-ku, Tokyo 101-8310, Japan, 3Division of
Physiology, Department of Human Development and Fostering, 4Division of Oral and Maxillofacial Surgery,
Department of Diagnostic and Therapeutic Sciences, Meikai University School of Dentistry, 1-1 Keyakidai,
Sakado, Saitama 350-0283, Japan, and 5Molecular Dynamics Imaging Unit, RIKEN Center for Life Science
Technologies, 6-7-3 Minatojima-minamimachi, Chuo-ku, Kobe, Hyogo 650-0047, Japan
Address correspondence to Dr Satoshi Fujita, Department of Pharmacology, Nihon University School of Dentistry, 1-8-13 Kanda-Surugadai, Chiyoda-ku,
Tokyo 101-8310, Japan. Email: [email protected]
Abstract
Cortical spreading depression (SD) is a self-propagating wave of depolarization accompanied by a substantial disturbance of the
ionic distribution between the intra- and extracellular compartments. Glial cells, including astrocytes, play critical roles in
maintenance of the extracellular environment, including ionic distribution. Therefore, SD propagation in the cerebral cortex
may depend on the density of astrocytes. The present study aimed to examine the profile of SD propagation in the insular cortex
(IC), which is located between the neocortex and paleocortex and is where the density of astrocytes gradually changes. The
velocity of SD propagation in the neocortex, including the somatosensory, motor, and granular insular cortices (5.7 mm/min),
was higher than that (2.8 mm/min) in the paleocortex (agranular insular and piriform cortices). Around thick vessels, including
the middle cerebral artery, SD propagation was frequently delayed and sometimes disappeared. Immunohistological analysis of
glial fibrillary acidic protein (GFAP) demonstrated the sparse distribution of astrocytes in the somatosensory cortex and the IC
dorsal to the rhinal fissure, whereas the ventral IC showed a higher density of astrocytes. These results suggest that cortical
cytoarchitectonic features, which possibly involve the distribution of astrocytes, are crucial for regulating the velocity of SD
propagation in the cerebral cortex.
Key words: astrocytes, insula, middle cerebral artery, paleocortex, piriform cortex
Introduction
Spreading depression (SD) is an electrophysiological phenomenon characterized by a self-propagating wave of transient
neuronal depolarization spreading throughout the cerebral
hemisphere. SD is followed by a long-lasting suppression of neuronal activity (Leão 1944) and has been implicated in various
neurological disorders such as migraine, stroke, epilepsy, and
trauma (Gorji 2001; Cui et al. 2009; Zhang et al. 2010). The principal structures where SD develops are the gray matter of the
central nervous system, for example, the cerebral cortex, hippocampus, and cerebellum. SD appears first in a restricted region
and spreads out in all directions at a velocity of 3–5 mm/min
(Hadjikhani et al. 2001; Cui et al. 2003; Tfelt-Hansen 2010). SD
© The Author 2015. Published by Oxford University Press. All rights reserved. For Permissions, please e-mail: [email protected]
1580
Anisotropic Propagation of CSD in Insular Cortex
propagation in the cerebral cortex is likely induced by release and
diffusion of certain excitatory substances, most likely potassium
and glutamate, into the interstitial space (Hansen et al. 1980; Holland et al. 2010; Dreier 2011; Grafstein 2011). Glial cells, including
astrocytes, buffer extracellular K+ (Gardner-Medwin 1981) and
take up glutamate (John et al. 2012). Several neurological reports
have demonstrated that SD propagation ends at the border of the
gray and white matter, and at the edge of glial-fibrous scars left by
injury or infarction (van Harreveld et al. 1956; Hull and van Harreveld 1964; Somjen 2001). Therefore, the distribution pattern of
astrocytes could be involved in the regulation of SD propagation.
The cerebral cortex is evolutionarily divided into 2 categories:
the neocortex, which basically represents 6 layers, and the paleocortex, which shows a heterogeneous laminar structure (Sarma
et al. 2011). The insular cortex (IC) is located between the somatosensory cortex, a part of the neocortex with 6 distinct layers, and
the piriform cortex, a part of the paleocortex consisting of 3 layers
(Reep and Winans 1982; Chen et al. 2007; Kobayashi 2011; Sarma
et al. 2011). The dorsal IC, the granular part of the IC (GI), shows a
similar laminar structure to the somatosensory cortex, whereas
the ventral IC, the agranular part of the IC (AI), lacks the granule
cell layer (layer IV). The dysgranular part of the IC (DI), between
the GI and AI, shows an incomplete layer IV. The IC also has,
as another anatomical feature, graduation, the density of
astrocytes. The GI and AI exhibit low and high densities of glial
fibrillary acidic protein (GFAP)-immunoreactive astrocytes,
respectively (Zilles et al. 1991). Given these findings, the IC is considered to be a suitable cortical area to examine the structural
features that are critical for propagation of cortical SD.
The present study investigated how cortical cytoarchitecture
affects the velocity of SD propagation. To address the issue, we
developed an in vivo optical imaging system with a voltagesensitive dye to visualize cortical neural activity. This enabled
us to observe the propagation of cortical SD intuitively with
high spatio-temporal resolution. In addition, GFAP-immunohistological analysis was performed to examine the distribution
pattern of astrocytes in the IC and surrounding areas.
Materials and Methods
The experiments were approved by the Animal Experimentation
Committee of Nihon University and were performed in accordance with the institutional guidelines for the care and use of experimental animals described in the National Institutes of Health
Guide for the care and use of laboratory animals. All efforts were
made to minimize animal suffering and reduce the number of
animals used.
Fujita et al.
| 1581
The anesthetized animal was mounted on a custom-made
stereotaxic snout frame (Narishige, Tokyo, Japan), which was
then tilted 60° laterally to make the left IC accessible to the CCD
camera system and electrodes. The left temporal muscle and
zygomatic arch were carefully removed, and a craniotomy was
performed to expose the IC and surrounding cortices.
The method of optical imaging with a voltage-sensitive dye has
been described previously (Chen et al. 2010; Fujita et al. 2010, 2011,
2012; Kobayashi et al. 2010; Mizoguchi et al. 2011; Adachi et al.
2013), and only a brief account of the methods employed will be
given here. The voltage-sensitive dye, RH1691 (1 mg/mL, Optical
Imaging, NY, USA), in 0.9% saline was applied to the cortical surface for approximately 1 h, residual dye was rinsed with saline
for approximately 30 min, and fluorescent changes in RH1691
were measured by a CCD camera system (MiCAM02, Brainvision,
Tokyo, Japan) mounted on a stereomicroscope (Leica Microsystems, Wetzler, Germany). The cortical surface was illuminated
through a 632-nm excitation filter and a dichroic mirror by a tungsten halogen lamp (CLS150XD, Leica Microsystems), and the fluorescent emission was captured through an absorption filter
(λ > 650 nm long-pass, Andover, Salem, NH, USA). The IC and surrounding cortices were imaged (Fig. 1). The CCD-based camera had
a 6.4 × 4.8 mm2 imaging area consisting of 184 × 124 pixels. The
sampling interval was 200 ms. Acquisition time was set at 8 min.
Induction of Cortical Spreading Depression
To evoke cortical SD, a cotton ball (1 mm diameter) soaked with
1 M KCl (Wako Pure Chemical, Tokyo, Japan) in distillated water
was placed on the pial surface for approximately 2 s. The stimulation point was set at the caudal end of the imaged area (Fig. 1,
black arrowhead) or the dorsal part of the somatosensory area
(Fig. 1, white arrowhead). In the latter experiments, another
small craniotomy (2 mm diameter) was performed on the parietal part of skull. Some rats were used in both experiments.
The stimulation was applied every 20 min. Between acquisitions,
0.9% saline was applied to moisten the cortical surface and wash
away excess KCl.
Local Field Potentials
In some imaging experiments, local field potentials (LFPs) were
simultaneously recorded to confirm that optical signals were
Optical Imaging
Five- to six-week-old male Sprague-Dawley rats (Sankyo Labo,
Tokyo, Japan) weighing 164.3 ± 3.1 g (n = 21) received atropine methyl bromide (5 mg/kg, intraperitoneal injection, Sigma-Aldrich,
St Louis, MO, USA) and were anesthetized with urethane (1.5 g/kg,
intraperitoneal injection, Sigma-Aldrich). The adequacy of anesthesia was gauged by the absence of the toe pinch reflex and
heart rate changes to pinching the toe. Anesthesia was maintained throughout the experiments by injecting additional urethane. The heart rate was maintained at physiological levels
(350–460 beats/min). Body temperature was monitored using a
rectal probe (BWT-100, Bio Research Center, Tokyo, Japan) and
was maintained at approximately 37°C using a heating pad. Animals received a tracheotomy and intubation, with lidocaine (2%
gel) applied to the incisions to ensure complete analgesia.
Figure 1. A schematic drawing of the preparation for optical imaging of the insular
cortex (IC). To induce cortical spreading depression (SD), a cotton ball soaked with
1 M KCl was applied at the edge of the caudal side in the imaging area in the
caudal–rostral experiments (black arrowhead) or in an opening of another
craniotomy on the parietal part of skull in the dorso-ventral experiments (white
arrowhead; modified from Fujita et al. 2010).
1582
| Cerebral Cortex, 2016, Vol. 26, No. 4
correlated with electrical activity. A monopolar tungsten electrode
(impedance = 0.25 MΩ; FHC, Bowdoin, ME, USA) was inserted
0.3 mm from the cortical surface, and LFPs were amplified (band
pass: 0–100 Hz; ER-1, Cygnus Technology, Delaware Water Gap,
PA, USA), digitized, and stored on a computer hard disk (Micro
1401 MK2, Cambridge Electronic Design, Cambridge, UK).
Data Analysis for Optical Imaging
The change in the intensity of fluorescence (ΔF) in each pixel relative to the initial fluorescence intensity (F) was calculated (ΔF/F).
A digital bleaching filter (Brain Vision Analyzer, Brainvision,
Tokyo, Japan) was used to compensate for dye bleaching
(Fig. 2A,C). The methods regarding velocity calculation and activation map analysis were described previously (Efimov et al.
2004). Briefly, images were processed with a derivative (DERIV) filter (±3 frames) to detect the frontline of the SD propagation, that
is, the rising portion of the depolarization, from the slow depolarization traces of SDs. A spatial filter (9 × 9 pixels) and a low-pass
filter (0.390 Hz) were applied to obtain the apparent frontline of
SDs and reduce the noise (Fig. 2B,C; Brain Vision Analyzer). An activation map was constructed by superimposition of the frontlines of an SD (Fig. 2D). The velocity of SD propagation was
calculated by line analysis (Fig. 2E; Brain Vision Analyzer). To calculate velocities, 3 SDs from a rat were averaged. In the rostral
part of piriform cortex, the velocity of SD propagation was calculated only from the cases in which an SD was observed in that
area. Data are expressed as means ± SEM. Velocities of SD propagations were compared using one-way ANOVA followed by
Tukey’s post hoc test (SPSS version 15, Chicago, IL, USA). A level
of P < 0.05 was considered statistically significant.
Immunohistochemistry
Male Sprague-Dawley rats (6 weeks old, n = 5) were deeply anesthetized with pentobarbital (100 mg/kg) and perfused transcardially with 0.9% saline (50 mL) followed by 4% paraformaldehyde
in 0.1 M phosphate buffer (PB, pH7.4, 200 mL). The brain was
removed and postfixed in 4% paraformaldehyde at 4°C. After an
overnight fixation, tissues were transferred to 15–30% sucrose
in phosphate-buffered saline (PBS) for several days for cryoprotection. Thirty-micrometer-thick coronal (n = 3) or horizontal
(n = 2) sections were cut with a cryostat and collected in PBS.
The protocols for immunohistochemistry were described previously (Li et al. 2004). Briefly, the sections were rinsed with 0.3%
Triton X-100 in 10 mM phosphate-buffered saline (PBST) and
then treated with 0.3% H2O2 for 1.5 h at room temperature to suppress endogenous peroxidase activity. After rinsing in 10 mM
PBST, the sections were incubated with Goat anti-Glial Fibrillary
Acidic Protein serum (GFAP, 1:1000; Santa Cruz Biotechnology,
Santa Cruz, CA, USA) overnight at 4°C. After primary antibody
incubation, sections were washed in 10 mM PBST. For light microscopic visualization with diaminobenzidine (DAB) as the chromagen, sections were incubated with an anti-Goat biotinylated
secondary antibody (1:200; Vector, Burlingame, CA, USA) for 1 h
at room temperature. After further rinsing in 10 mM PBST, the
sections were then incubated in avidin-biotinylated enzyme
complex solution (ABC reagent; 1:100, Vector, Burlingame, CA,
USA) for 1 h. After rinsing in 10 mM PBST, the reaction product
was detected using 0.03% 3-3′-DAB (Sigma-Aldrich) and 0.01%
H2O2 in 50 mM Tris–HCl at room temperature. The sections
were mounted on gelatin-coated slides and coverslipped. Control
sections that were processed without primary antiserum
displayed no specific staining.
Quantitative Analysis for GFAP
A microscope (Eclipse E600, Nikon, Tokyo, Japan) equipped with a
motorized stage (Lucivid, MicroBrightField, Colchester, VT, USA)
and Stereo Investigator software (Neurolucida, ver. 10.40, MicroBrightField) was used to count immunocytochemically labeled
astrocytes (Chen et al. 2010; Kobayashi et al. 2010). In this
study, multipolar-shaped structures with a densely stained cell
body and 3 ≥processes (Fig. 8C, black arrowheads) were counted
with a ×20–40 objective.
We chose coronal sections located near the intersection of the
middle cerebral artery (MCA) and rhinal fissure (RF) to compare
the density of astrocytes between dorsal and ventral regions.
We generated contour plots (Origin 8, OriginLab, MA, USA) in
horizontal sections to visualize the distribution of astrocytes.
The contour lines were smoothed using a software function
(Smooth Matrix values = 0.01).
Results
Frontlines of Cortical Spreading Depression
Topical application of 1 M KCl (Ayata et al. 2006; Chen et al. 2006;
Zhang et al. 2010) to the pial surface of the caudal end of the imaged area increased the amplitude of the optical signal (ΔF/F),
which was followed by a long-lasting signal decrease as shown
in Figure 2. An increase or decrease in optical signals is considered to reflect neuronal depolarization or hyperpolarization,
respectively (Farkas et al. 2008).
Simultaneous recording of the field potential demonstrated
that an approximately −20-mV negative DC potential shift was
observed being synchronously accompanied by a depolarization
phase of optical signals, as is characteristic of the cortical SD
(Fig. 2C; Herreras and Somjen 1993; Cui et al. 2013). In agreement
with a previous study comparing the signals of LFP and voltagesensitive dye in SD propagation (Farkas et al. 2008), the signals
consistently showed good concordance in the depolarization
phase but not in the hyperpolarization phase following depolarization. Application of a DERIV filter allowed clear visualization
of the sharp frontlines of SD propagations (Fig. 2B). The frontlines
of SD propagated in a concentric manner (Supplementary Movie 1).
A line perpendicular to the frontline of the SD was drawn to measure the velocity of SD propagation (Arrow in Fig. 2D). By applying
the stripe map function (see Materials and Methods; Brain Vision
Analyzer), a spatio-temporal map along the perpendicular line
was generated (Fig. 2E). Movement of the frontline of the SD was represented as a color-coded signal in the map (Dotted line in Fig. 2E).
The calculated velocity of SD was 4.8 mm/min. Similar to a previous study (Farkas et al. 2008), LFP in the AI showed a similar
temporal excitation curve to optical signals obtained from the voltage-sensitive dye (Fig. 2C); this indicated that the frontlines of
optical signals represented the onset to peak of the depolarization
phase of SD.
Propagation Velocity of SD in the IC Area
We evoked cortical SD by stimulating the caudal edge of the imaging area to investigate the propagation characteristics of SD in
the IC in a caudal-to-rostral direction, which involves a representative transition of cytoarchitectonic laminar structures (stars in
Figs 2A and 3A,B). The activation map showed distinct characteristics of SD between the dorsal and ventral part of the IC (Fig. 3A,
B; also see Supplementary Movie 1). Fast SD propagation was observed in dorsal part of the imaging area, including the motor
cortex (MC), primary sensory cortex (SI) and GI (Fig. 3A,B).
Anisotropic Propagation of CSD in Insular Cortex
Fujita et al.
| 1583
Figure 2. The detection of the frontline of cortical SD. (A) Color-coded optical signals of cortical SD are shown. A template of cortical subdivisions around the IC is shown in
the left upper panel. A blue circle in the agranular part of the IC (AI) indicates the location of the region of interest (ROI). The asterisk indicates the site of the recording
electrode. The red star indicates the stimulation site. In this figure, long-term bleaching was corrected by a beaching filter. Note that in the SD, biphasic waves of potent
depolarization and hyperpolarization can be observed. (B) The frontline of the SD is clearly visible due to the use of derivative (DERIV), spatial, and low-pass filters. (C) The
frontline corresponded to the rising phase of depolarization of the SD. Note that traces of local field potential (LFP; black line) from the recording electrode and optical
signals (blue line; digital bleaching filtered) from an ROI in (A) are very similar. A DERIV filter revealed the rise phasing of the SD (red line). The vertical lines in upper part
correspond to timings of panels in (A) and (B). (D) Example of activation map analysis. The red lines indicate a frontline traveling from 10 to 180 s at 10 s intervals after 1 M
KCl stimulation. The area filled by black indicates the shadow of the LFP recording electrode. (E) A spatio-temporal map of the movement of a frontline of an SD wave along
the arrow in (D) is shown. AI, agranular IC; DI, dysgranular IC; DLO, dorsolateral orbital cortex; Fr3, frontal cortex (area 3); GI, granular IC; Pir, piriform cortex; MC, motor
cortex; SI, primary somatosensory cortex; SII, secondary somatosensory cortex.
No significant difference in propagation velocity was observed
among the MC, SI, and GI (Fig. 4). A lower velocity of SD propagation was observed in the AI and piriform cortex, which are characterized by typical paleocortical cytoarchitecture, that is, <6 cell
layers and a high density of glial cells. In DI, which is located between GI and AI, SD propagated with medium velocity (Fig. 4B;
F5,66 = 19.83, P < 0.001, one-way ANOVA).
To confirm the characteristic profile of SD propagation, that is,
that AI and piriform cortex show a lower velocity relative to the
dorsal part of IC and somatosensory cortex, the pial surface of
the parietal cortex was stimulated with a cotton ball soaked
with 1 M KCl, and SD propagation velocity was evaluated in a
dorsal-to-ventral direction (Fig. 5). Consistent with the above
hypothesis, the propagation velocity gradually decreased from
1584
| Cerebral Cortex, 2016, Vol. 26, No. 4
Figure 3. Two examples of cortical SDs in the caudal–rostral experiments. The frontline travels (thin lines) from 18 to 50 s at 2 s intervals (A) and from 26 to 74 s at 2 s
intervals (B) after 1 M KCl stimulation. A template of cortical subdivisions around the IC is shown in the panels. Thick dotted and solid lines indicate the middle
cerebral artery (MCA) and the rhinal fissure (RF). Note that the distance between thin lines is large on dorsal side, whereas it is narrow on the ventral side. The
frontlines on the rostral side of the MCA become orthogonal to the MCA at the arrowhead.
Figure 4. The velocity of cortical SD propagation varied depending on the area. (A) Velocities were measured from the MC, SI, GI, DI, AI, and Pir. Arrows indicate examples of
actual measurement sites. Black thin lines indicate the frontline traveling from 5 to 140 s at 5 s intervals after 1 M KCl stimulation. (B) The velocities were lower in the DI, AI,
and Pir compared with the SI, GI, and MC. (*P < 0.05, **P < 0.01, ***P < 0.001, Tukey’s test).
SI/SII and GI to the ventral IC. A significant decrease of propagation velocity was observed in DI and AI relative to SI (Fig. 5B; F3,48
= 22.42, P < 0.001, one-way ANOVA).
the MCA and RF and reached the ventral area. In the remaining
2 cases, SDs terminated completely at the RF.
Distribution Profile of Astrocytes
Disturbance of SD Propagation by the MCA and Rhinal
Fissure
Interestingly, SD propagation seems to be disturbed by the MCA
and RF. As shown in Figures 3 and 4, SD passed across the dorsal
MCA but not across the ventral MCA (see Supplementary Movie
1). Therefore, the direction of SD propagation changed near the
MCA in the ventral part of IC (Figs 3 and 4A). On the other
hand, regardless of the thickness of the MCA, such a strong disturbance of SD propagation was not found in the dorsal area, including DI and SI (Fig. 3A,B); SD sometimes traveled with a short
time delay onto the MCA (solid arrowhead in Figs 5B and 6, but
not outline arrowhead). In addition to the MCA, SD propagation
was sometimes disturbed by the caudal part of the RF. In the
dorso-ventral experiments, 29 of 39 SDs propagated through
the RF (Fig. 7A), whereas 10 of 39 SDs did not traverse the RF directly (Fig. 7B). Seven of the latter 10 SDs detoured caudally outside
of the imaging area and propagated recurrently in the caudal to
rostral direction. Each of these 7 SDs stopped at the ventral
MCA (Fig. 7B). Only 1 SD detoured around the intersection of
In agreement with previous studies using computer-assisted
image analysis of GFAP-immunoreactive astrocytes (Zilles et al.
1991), our immunohistological analysis showed sparse astrocytes in SI/SII and IC dorsal to the RF, whereas ventral IC and piriform cortex showed a higher density of astrocytes (Fig. 8A–E). In
SI/GI, distribution patterns were heterogeneous; superficial and
deeper layers generally showed higher densities of astrocytes
(Fig. 8A–D; Zilles et al. 1991). Density of astrocytes in the middle
layers gradually increased from SI/SII and GI to RF (Fig. 8E). Density of astrocytes in the ventral region 1–2 mm from RF was 2.3fold higher relative to the dorsal region 4–5 mm from the RF (n = 3,
P < 0.05, ANOVA with Tukey’s test, Fig. 8E). In addition, the distribution pattern of astrocytes was affected by thick vessels. A
dense distribution of astrocytes was observed around RF
(Fig. 8F). The MCA effected a hollow in the cortical surface,
which exhibited a higher density of astrocytes (Fig. 8I–K). On
the other hand, thin vessels had little effect on morphology of
the cortical surface, though the density of astrocytes was slightly
higher than in adjacent regions (Fig. 8G,H).
Anisotropic Propagation of CSD in Insular Cortex
Fujita et al.
| 1585
Figure 5. The observation of cortical SD induced by parietal cortex stimulation revealed a gradual decrement from dorsal areas to the ventral areas. (A) An example of
movement of frontlines in the dorso-ventral SD experiment is shown. Thin lines indicate the frontline traveling from 140 to 285 s at 5 s intervals after 1 M KCl
stimulation. (B) An example of a spatio-temporal map based on the travel of the frontline (thick black band) of the SD wave along the arrow in (A) is shown. The
dotted white line indicates the frontline. In this case, the vessels delayed the propagation of the SD for a moment, and this part was removed from velocity
measurements (arrowhead). (C) The velocity of the cortical SD decreased gradually from the GI to the AI (*P < 0.05, **P < 0.01, ***P < 0.001, Tukey’s test).
Figure 6. The MCA sometimes briefly delayed the propagation of the SD. The left panels show a frontline traveling from the caudal to rostral direction. A red star indicates
the stimulation site. The right panel shows a spatio-temporal map of the movement of a frontline along the arrow in the first left panel. Note that there is a slight shift
shown by the two dotted lines because the propagation was temporally delayed at the solid arrowhead, but not at the outlined arrowhead.
Discussion
The present study investigated whether SD propagation patterns
are dependent on cortical cytoarchitecture, using an in vivo voltage-sensitive dye imaging technique. We demonstrated that AI
and piriform cortex exhibited a lower SD propagation velocity
relative to SI/SII and GI. The velocity change likely relates to cytoarchitectonic differences between SI/SII/GI (neocortex) and AI/
piriform cortex ( paleocortex). We also demonstrated that AI/
piriform exhibited a higher density of astrocytes than SI/SII/GI
(Fig. 8; Zilles et al. 1991). In addition, we found that propagation
of SD was disturbed at thick vessels, such as the MCA and RF,
where the density of astrocytes was higher than in surrounding
regions. Therefore, it is likely that the buffering mechanism in
astrocytes may play a critical role in propagation of cortical SD.
which consists of cytoarchitectonically heterogeneous laminar
structures, that is, the neocortex (GI) and paleocortex (AI;
Cechetto and Saper 1987; Shi and Cassell 1998; Kobayashi 2011;
Maffei et al. 2012). The piriform cortex is the most ventral area
imaged in this study, and it consists of 3 layers (Chen et al.
2007; Sarma et al. 2011). We found a gradual decrement in SD
propagation velocity from SI and GI to AI (Fig. 5B). Several studies
have demonstrated variation in SD propagation velocity. Bowyer
et al. (1999) showed faster SD propagation in gyri than that in
sulci in the pig brain in vivo. In vitro, hippocampal, and entorhinal cortical slice preparations exhibited the highest velocity
(5.4 mm/min) in CA3 region and the slowest velocity (2.7 mm/
min) in cortical regions (Buchheim et al. 2002). These observations suggest that cytoarchitectonic features contribute to the
propagation of SD velocity.
Anisotropic Propagation Patterns of SD in IC and
Surrounding Cortices
Astrocytes Possibly Regulate SD Propagation Velocity
The main finding of the present study is that different velocities
of SD propagation in the temporal cortex depend on the different
cytoarchitectonic laminar patterns. Since the discovery of cortical SD as a self-propagating wave of depolarization, cortical
SD has been considered to spread in a concentric manner (Tomita et al. 2002). To the best of our knowledge, propagation of SD
has been studied in tissue of uniform cytoarchitecture. In this
study, we demonstrate anisotropic propagation of SD in IC,
It has been proposed that SD propagation is caused by the release
and diffusion of excitatory chemical mediators, including potassium and glutamate, into the extracellular space. Therefore, it
has been hypothesized that glia, especially astrocytes in the
brain, contribute to SD propagation by buffering extracellular potassium and excitatory amino acids (Gardner-Medwin 1981;
Somjen 2001). Astrocytes remove potassium ions from the extracellular space by several mechanisms: 1) inwardly rectifying K+
1586
| Cerebral Cortex, 2016, Vol. 26, No. 4
Figure 7. The RF sometimes prevented the propagation of the SD into the ventral area. (A) The SDs (29/39 cases of the dorso-ventral experiments) usually propagated
through the caudal part of the RF (arrowheads). Directions of arrowheads correspond to directions of SD propagation. Note that the blue-colored zone (asterisk) is due
to transition (repolarization) from depolarization phase to hyperpolarization phase (see Fig. 2). (B) The SD (7/39 cases) propagations detoured caudally outside the
imaging area and spread to the area below the RF from the caudal to rostral direction (black arrowheads). Note that the cortical SD stopped at the MCA in all cases
(white arrowheads).
channels; 2) Na+/K+-ATPase; 3) Na+-K+-Cl− co-transporters; and 4)
gap junctions (Kofuji and Newman 2004; Olsen and Sontheimer
2008; Larsen et al. 2014). Astrocyte-dependent buffering of extracellular glutamate is mediated by membrane transporters, such
as excitatory amino acid transporter 1 and excitatory amino
acid transporter 2. In agreement with the above hypothesis, ventral IC, which has a higher density of astrocytes, showed slower
SD propagation relative to dorsal IC, where astrocytes are sparse
(Fig. 8; Zilles et al. 1991).
Relative to the decrease in velocity of SD propagation, the
density of astrocytes abruptly decreased from dorsal to ventral
IC. A possible explanation is that the density map was made by
plotting cell bodies, rather than their abundant branches. Therefore, the functional region of an astrocyte is likely to be much larger than its plot, which may cause gradual alternation of velocity.
Interestingly, SD propagation from caudal IC was slowed or
stopped near the MCA at the ventral part of IC (Fig. 7). In addition,
dorso-ventral propagation of SD showed a delayed frontline near
the MCA, indicating a decrease in propagation velocity (Fig. 5A).
Astrocytes are densely localized around brain blood vessels
(Fig. 8F–H; Maddahi et al. 2009), where abundant expression of inwardly rectifying K+ channels is observed (Newman 1984; Hibino
et al. 2004; Hibino and Kurachi 2007). Taking into account the role
of inward rectifying K+ channels, that is, removing extracellular
K+, astrocytes near blood vessels in the brain may disturb SD
propagation by buffering extracellular K+ and glutamate. Previous studies using optical imaging techniques have reported
that such disturbances by blood vessels are not found in the rat
neocortex (Tomita et al. 2002; Chen et al. 2006; Brennan et al.
2007). The reason for the discrepancy between our findings and
these reports may be differences in cortical areas examined.
The present study demonstrated that SD propagation was
stopped or slowed at the MCA only in IC or piriform cortex, but
not in SI. Recently, Santos et al. (2014) demonstrated that morphology critically affects SD propagation in pig gyrencephalic brain;
the pial vessels acted as physical barriers that blocked SD propagation. Our findings in rat cerebral cortex are in line with their observation. We propose that the density of astrocytes in the region
surrounding the MCA and RF in the paleocortex is higher than in
the neocortex and, therefore, the buffering effect is larger in IC
than in SI.
The astrocyte-induced buffering of extracellular K+ is considered to suppress SD propagation, and our present results support
this hypothesis. In addition to the functional contribution of astrocytes, other components, including neurons and microglia,
may also play a role in regulating SD propagation. Indeed, Ca2+
influx causes K+ efflux to the extracellular space via Ca2+dependent K+ channels, and the increase in [K+]o is considered
to trigger cortical SD and increase velocity of SD propagation
(Torrente et al. 2014), suggesting that high neural activity facilitates SD. In addition, a recent study reported that SD requires
microglia (Pusic et al. 2014). Depletion of microglia in slice cultures inhibits initiation of SD, and microglial polarization state
modulates SD threshold both in vitro and in vivo.
Anisotropic Propagation of CSD in Insular Cortex
Fujita et al.
| 1587
Figure 8. The characteristics of astrocyte distributions. (A) An example of glial fibrillary acidic protein (GFAP)-immunoreactive astrocytes in a coronal section. An
arrowhead indicates the RF. (B) An example of the transition area where astrocytes density changes, which is boxed with red-dotted lines in (A). Note that the dorsal
part shows a lighter reaction, whereas the ventral part is darkened. (C) An example of astrocyte counting (arrowheads). (D) An example of plots of astrocytes and
compartments where density was quantified. (E) A summary of the distribution of astrocytes (n = 3). (F) A contour plot from blue-dotted area in (D). Note that
astrocytes were dense around the RF (arrowhead). (G) An example of plots of astrocytes around thin arteries (arrowheads) in a coronal section. (H) A contour plot from
(G). (I) An example of GFAP-immunoreactive astrocytes around a thick MCA in the horizontal section. “r” and “c” indicate rostral and caudal, respectively. (J) The plots from
(I). (K) A contour plot on from (J). Note that the sparse astrocyte area was narrow around the MCA, as if it was being compressed by the MCA.
SD is considered to cause migraine aura symptoms. Our
results suggest that the paleocortex, such as AI and piriform cortex, is resistive to SD propagation. Taking into account that IC
and piriform cortex process gustatory, visceral, and olfactory
information, such senses are less likely to be affected by SD.
Indeed, in contrast to visual and somatosensory aura, gustatory
and olfactory aura are minor symptoms of migraine in humans
(Ferrari 1998).
1588
| Cerebral Cortex, 2016, Vol. 26, No. 4
Technical Advantage
Cortical SD is characterized by a self-propagating wave of intense depolarization. Optical imaging techniques with a voltage-sensitive dye are potent tools for investigating the kinetics
of a direct current potential with high temporal and spatial resolution (Farkas et al. 2008; see also Fig. 2). Although the SD signal
was intense and showed a high signal/noise ratio, cortical SD
exhibited very slow and wide development, which makes for
difficulty in detecting its onset. To resolve this problem, we differentiated between signal intensity and time. As a result, onset
of SD propagation, that is, the rising component of depolarization, is clearly detected as a frontline. This derivation process
has been widely used in investigation of cardiac physiology.
For optical imaging of the heart, fluorescence derivatives are
used to detect time points of myocardial activation and repolarization (Efimov et al. 2004).
Inhomogeneous SD propagation due to the MCA results in inaccurate velocity measurements. Therefore, the short time delay
(solid arrowhead, but not the outlined arrowhead in Fig. 6) in the
dorsal part of the MCA was excluded from velocity measurements (Fig. 5B). In addition to change in SD propagation velocity,
a diffraction-like phenomenon (i.e., direction of change of SD
propagation by the MCA in the ventral part of IC) also makes
the accurate measurement of SD velocity difficult (Figs 3 and
4A; see Supplementary Movie 1). A line perpendicular to the
wave of SD was required to calculate the accurate velocity of SD
propagation. Therefore, as part of caudal-to-rostral experiments,
in the ventral part of the MCA, we divided SD propagation measurement in IC into segments caudal and rostral to the MCA.
Conclusions
In the present study, we demonstrated for the first time that the
propagation velocity of cortical SD gradually decreased from dorsal to ventral cortical areas of the IC. In this area, the laminar
cytoarchitecture and the volume distribution of glial cells are
known to change gradually from neocortex to paleocortex.
These results suggest that these cytoarchitectonic features, especially the density of glial cells, are crucial for regulating SD propagation among cortical areas.
Supplementary Material
Supplementary material can be found at http://www.cercor.
oxfordjournals.org/ online.
Funding
This work was supported by Japan Society for the Promotion of
Science KAKENHI (22791802 to S.F., 25293379 to M.K., 26463192
to N.M.); the Ministry of Education, Culture, Sports, Science and
Technology—supported Program for the Strategic Research
Foundation at Private Universities, 2013–2017, to M.K.; the Sato
Fund, Nihon University School of Dentistry to S.F.
Notes
We thank Prof. John L. Waddington for critical comments on the
manuscript. We also thank Drs Kenji Tsubokura and Kiyo Murano
for developing the software for optical imaging analysis. Conflict
of Interest: None declared.
References
Adachi K, Fujita S, Yoshida A, Sakagami H, Koshikawa N,
Kobayashi M. 2013. Anatomical and electrophysiological mechanisms for asymmetrical excitatory propagation in the rat
insular cortex: in vivo optical imaging and whole-cell patchclamp studies. J Comp Neurol. 521:1598–1613.
Ayata C, Jin H, Kudo C, Dalkara T, Moskowitz MA. 2006. Suppression of cortical spreading depression in migraine prophylaxis.
Ann Neurol. 59:652–661.
Bowyer SM, Tepley N, Papuashvili N, Kato S, Barkley GL, Welch KM,
Okada YC. 1999. Analysis of MEG signals of spreading cortical
depression with propagation constrained to a rectangular cortical strip. II. Gyrencephalic swine model. Brain Res. 843:79–86.
Brennan KC, Beltrán-Parrazal L, López-Valdés HE, Theriot J,
Toga AW, Charles AC. 2007. Distinct vascular conduction with
cortical spreading depression. J Neurophysiol. 97:4143–4151.
Buchheim K, Weissinger F, Siegmund H, Holtkamp M,
Schuchmann S, Meierkord H. 2002. Intrinsic optical imaging
reveals regionally different manifestation of spreading depression in hippocampal and entorhinal structures in vitro.
Exp Neurol. 175:76–86.
Cechetto DF, Saper CB. 1987. Evidence for a viscerotopic sensory
representation in the cortex and thalamus in the rat. J Comp
Neurol. 262:27–45.
Chen S, Fujita S, Koshikawa N, Kobayashi M. 2010. Pilocarpineinduced status epilepticus causes acute interneuron loss
and hyper-excitatory propagation in rat insular cortex.
Neuroscience. 166:341–353.
Chen S, Kobayashi M, Honda Y, Kakuta S, Sato F, Kishi K. 2007.
Preferential neuron loss in the rat piriform cortex following
pilocarpine-induced status epilepticus. Epilepsy Res. 74:1–18.
Chen S, Li P, Luo W, Gong H, Zeng S, Luo Q. 2006. Time-varying
spreading depression waves in rat cortex revealed by optical
intrinsic signal imaging. Neurosci Lett. 396:132–136.
Cui Y, Kataoka Y, Li QH, Yokoyama C, Yamagata A, MochizukiOda N, Watanabe J, Yamada H, Watanabe Y. 2003. Targeted
tissue oxidation in the cerebral cortex induces local prolonged
depolarization and cortical spreading depression in the rat
brain. Biochem Biophys Res Commun. 300:631–636.
Cui Y, Li QH, Yamada H, Watanabe Y, Kataoka Y. 2013. Chronic degeneration of dorsal raphe serotonergic neurons modulates
cortical spreading depression: a possible pathophysiology of
migraine. J Neurosci Res. 91:737–744.
Cui Y, Takashima T, Takashima-Hirano M, Wada Y, Shukuri M,
Tamura Y, Doi H, Onoe H, Kataoka Y, Watanabe Y. 2009. 11CPK11195 PET for the in vivo evaluation of neuroinflammation
in the rat brain after cortical spreading depression. J Nucl Med.
50:1904–1911.
Dreier JP. 2011. The role of spreading depression, spreading depolarization and spreading ischemia in neurological disease.
Nat Med. 17:439–447.
Efimov IR, Nikolski VP, Salama G. 2004. Optical imaging of the
heart. Circ Res. 95:21–33.
Farkas E, Pratt R, Sengpiel F, Obrenovitch TP. 2008. Direct, live imaging of cortical spreading depression and anoxic depolarisation using a fluorescent, voltage-sensitive dye. J Cereb Blood
Flow Metab. 28:251–262.
Ferrari MD. 1998. Migraine. Lancet. 351:1043–1051.
Fujita S, Adachi K, Koshikawa N, Kobayashi M. 2010. Spatiotemporal dynamics of excitation in rat insular cortex: intrinsic
corticocortical circuit regulates caudal-rostro excitatory
propagation from the insular to frontal cortex. Neuroscience.
165:278–292.
Anisotropic Propagation of CSD in Insular Cortex
Fujita S, Kitayama T, Mizoguchi N, Oi Y, Koshikawa N,
Kobayashi M. 2012. Spatiotemporal profiles of transcallosal
connections in rat insular cortex revealed by in vivo optical imaging. Neuroscience. 206:201–211.
Fujita S, Koshikawa N, Kobayashi M. 2011. GABAB receptors accentuate neural excitation contrast in rat insular cortex.
Neuroscience. 199:259–271.
Gardner-Medwin AR. 1981. Possible roles of vertebrate neuroglia
in potassium dynamics, spreading depression and migraine.
J Exp Biol. 95:111–127.
Gorji A. 2001. Spreading depression: a review of the clinical relevance. Brain Res Rev. 38:33–60.
Grafstein B. 2011. Subverting the hegemony of the synapse: complicity of neurons, astrocytes, and vasculature in spreading
depression and pathology of the cerebral cortex. Brain Res
Rev. 66:123–132.
Hadjikhani N, Sanchez del Rio M, Wu O, Schwartz D, Bakker D,
Fischl B, Kwong KK, Cutrer FM, Rosen BR, Tootell RB, et al. 2001.
Mechanisms of migraine aura revealed by functional MRI in
human visual cortex. Proc Natl Acad Sci USA. 98:4687–4692.
Hansen AJ, Quistorff B, Gjedde A. 1980. Relationship between
local changes in cortical blood flow and extracellular K+ during spreading depression. Acta Physiol Scand. 109:1–6.
Herreras O, Somjen GG. 1993. Analysis of potential shifts associated with recurrent spreading depression and prolonged
unstable spreading depression induced by microdialysis of
elevated K+ in hippocampus of anesthetized rats. Brain Res.
610:283–294.
Hibino H, Fujita A, Iwai K, Yamada M, Kurachi Y. 2004. Differential assembly of inwardly rectifying K+ channel subunits,
Kir4.1 and Kir5.1, in brain astrocytes. J Biol Chem. 279:
44065–44073.
Hibino H, Kurachi Y. 2007. Distinct detergent-resistant membrane microdomains (lipid rafts) respectively harvest K+ and
water transport systems in brain astroglia. Eur J Neurosci.
26:2539–2555.
Holland PR, Akerman S, Goadsby PJ. 2010. Cortical spreading depression-associated cerebral blood flow changes induced by
mechanical stimulation are modulated by AMPA and GABA
receptors. Cephalalgia. 30:519–527.
Hull CD, van Harreveld A. 1964. Absence of conduction of spreading depression through cortical region damaged by asphyxiation. Am J Physiol. 207:921–924.
John CS, Smith KL, Van’t Veer A, Gompf HS, Carlezon WA Jr,
Cohen BM, Öngür D, Bechtholt-Gompf AJ. 2012. Blockade of
astrocytic glutamate uptake in the prefrontal cortex induces
anhedonia. Neuropsychopharmacology. 37:2467–2475.
Kobayashi M. 2011. Macroscopic connection of rat insular cortex:
anatomical bases underlying its physiological functions. Int
Rev Neurobiol. 97:285–303.
Kobayashi M, Fujita S, Takei H, Song L, Chen S, Suzuki I,
Yoshida A, Iwata K, Koshikawa N. 2010. Functional mapping
of gustatory neurons in the insular cortex revealed by pERKimmunohistochemistry and in vivo optical imaging.
Synapse. 64:323–334.
Kofuji P, Newman EA. 2004. Potassium buffering in the central
nervous system. Neuroscience. 129:1045–1056.
Larsen BR, Assentoft M, Cotrina ML, Hua SZ, Nedergaard M,
Kaila K, Voipio J, MacAulay N. 2014. Contributions of the
Na+/K+-ATPase, NKCC1, and Kir4.1 to hippocampal K+ clearance and volume responses. Glia. 62:608–622.
Leão AAP. 1944. Spreading depression of activity in the cerebral
cortex. J Neurophysiol. 7:359–390.
Fujita et al.
| 1589
Li QH, Nakadate K, Tanaka-Nakadate S, Nakatsuka D, Cui Y,
Watanabe Y. 2004. Unique expression patterns of 5-HT2A
and 5-HT2C receptors in the rat brain during postnatal development: western blot and immunohistochemical analyses.
J Comp Neurol. 469:128–140.
Maddahi A, Chen Q, Edvinsson L. 2009. Enhanced cerebrovascular
expression of matrix metalloproteinase-9 and tissue inhibitor
of metalloproteinase-1 via the MEK/ERK pathway during cerebral ischemia in the rat. BMC Neurosci. 10:56.
Maffei A, Haley M, Fontanini A. 2012. Neural processing of gustatory information in insular circuits. Curr Opin Neurobiol.
22:709–716.
Mizoguchi N, Fujita S, Koshikawa N, Kobayashi M. 2011. Spatiotemporal dynamics of long-term potentiation in rat insular
cortex revealed by optical imaging. Neurobiol Learn Mem.
96:468–478.
Newman EA. 1984. Regional specialization of retinal glial cell
membrane. Nature. 309:155–157.
Olsen ML, Sontheimer H. 2008. Functional implications for Kir4.1
channels in glial biology: from K+ buffering to cell differentiation. J Neurochem. 107:589–601.
Pusic KM, Pusic AD, Kemme J, Kraig RP. 2014. Spreading depression requires microglia and is decreased by their M2a
polarization from environmental enrichment. Glia.
62:1176–1194.
Reep RL, Winans SS. 1982. Afferent connections of dorsal and
ventral agranular insular cortex in the hamster Mesocricetus
auratus. Neuroscience. 7:1265–1288.
Santos E, Schöll M, Sánchez-Porras R, Dahlem MA, Silos H,
Unterberg A, Dickhaus H, Sakowitz OW. 2014. Radial, spiral
and reverberating waves of spreading depolarization occur
in the gyrencephalic brain. Neuroimage. 99:244–255.
Sarma AA, Richard MB, Greer CA. 2011. Developmental dynamics
of piriform cortex. Cereb Cortex. 21:1231–1245.
Shi CJ, Cassell MD. 1998. Cortical, thalamic, and amygdaloid connections of the anterior and posterior insular cortices. J Comp
Neurol. 399:440–468.
Somjen GG. 2001. Mechanisms of spreading depression and hypoxic spreading depression-like depolarization. Physiol Rev.
81:1065–1096.
Tfelt-Hansen PC. 2010. History of migraine with aura and cortical
spreading depression from 1941 and onwards. Cephalalgia.
30:780–792.
Tomita Y, Tomita M, Schiszler I, Amano T, Tanahashi N,
Kobari M, Takeda H, Ohtomo M, Fukuuchi Y. 2002. Repetitive
concentric wave-ring spread of oligemia/hyperemia in the
sensorimotor cortex accompanying K+-induced spreading
depression in rats and cats. Neurosci Lett. 322:157–160.
Torrente D, Mendes-da-Silva RF, Lopes AA, González J,
Barreto GE, Guedes RC. 2014. Increased calcium influx triggers
and accelerates cortical spreading depression in vivo in male
adult rats. Neurosci Lett. 558:87–90.
van Harreveld A, Terres G, Dernburg EA. 1956. Cortical discontinuity and propagation of spreading depression. Am J
Physiol. 184:233–238.
Zhang X, Levy D, Noseda R, Kainz V, Jakubowski M, Burstein R.
2010. Activation of meningeal nociceptors by cortical spreading depression: implications for migraine with aura.
J Neurosci. 30:8807–8814.
Zilles K, Hajós F, Kálmán M, Schleicher A. 1991. Mapping of glial
fibrillary acidic protein-immunoreactivity in the rat forebrain
and mesencephalon by computerized image analysis. J Comp
Neurol. 308:340–355.