Involvement of NMDA receptor subtypes in cortical spreading depression in rats assessed by fMRI

Neuropharmacology 93 (2015) 164e170

Contents lists available at ScienceDirect

Neuropharmacology journal homepage: www.elsevier.com/locate/neuropharm

Involvement of NMDA receptor subtypes in cortical spreading depression in rats assessed by fMRI € hn a Artem Shatillo a, *, Raimo A. Salo a, Rashid Giniatullin a, b, Olli H. Gro a b

Department of Neurobiology, A. I. Virtanen Institute for Molecular Sciences, University of Eastern Finland, Neulaniementie 2, Kuopio 70700, Finland Lab of Neurobiology, Kazan Federal University, 18 Kremlyovskaya St., Kazan 420008, Republic of Tatarstan, Russian Federation

a r t i c l e i n f o

a b s t r a c t

Article history: Received 28 August 2014 Received in revised form 23 December 2014 Accepted 26 January 2015 Available online 14 February 2015

Cortical spreading depression (CSD) is a phenomenon implicated in migraine with aura and associated with other neurological disorders (e.g. stroke, brain trauma). Current evidence points to the essential role of NMDA receptors in CSD mechanisms. However, the roles of multiple subunits of NMDA receptors expressed in neurons, glia and blood vessels in vivo, are little explored. Using BOLD fMRI of urethane anesthetized rats as an integrative CSD readout, we tested the involvement of different NMDA receptor subtypes in CSD induction and propagation. Rats were treated with a non-selective NMDA blocker (MK801), NR2B antagonist (ifenprodil) or a NR2A selective antagonist (TCN-201). CSD was induced during fMRI scanning by application of KCl onto the cerebral cortex and fMRI data were collected by 9.4 T MRI. The non-specific NMDA antagonist MK-801 completely blocked CSD, which was not observed in the NR2A group where TCN-201 did not alter the CSD features. Unexpectedly, the NR2B specific antagonist ifenprodil largely promoted the initial negative phase of the BOLD CSD response, likely due to altered neurovascular coupling. Our data suggest key roles and differential involvement of NMDA receptor subtypes in CSD generation and propagation, highlighting an important role for the NR2B subtype. © 2015 Elsevier Ltd. All rights reserved.

Chemical compounds studied in this article: Dizocilpine, MK-801 (PubChem CID: 180081) Ifenprodil (PubChem CID: 11957579) TCN-201 (PubChem CID: 4787937) Keywords: Spreading depression BOLD fMRI Glutamate NMDA receptors

1. Introduction Cortical spreading depression (CSD) is a self-propagating wave of neuronal depolarization (Leao and Morison, 1945; Leao, 1944) with glial and vascular involvement. (Charles and Baca, 2013) CSD is associated with the massive release of various neurotransmitters, dramatic shifts in ionic gradients and essential blood flow. (Smith et al., 2006) CSD takes place in a number of major neurological disorders including stroke, brain trauma and notably, in migraine with aura. (Dreier, 2011; Lauritzen et al., 2011). The primary role of CSD in migraine with aura (MWA) was hypothesized by Leao and Morrison (1945), Milner (Milner, 1958) and later confirmed, using the planar intracarotideal 133-Xenon method, SPECT and PET (Lauritzen, 1994; Olesen et al., 1981; Woods et al., 1994).

* Corresponding author. Tel.: þ358 453111767. E-mail addresses: artem.shatillo@uef.fi (A. Shatillo), raimo.salo@uef.fi (R.A. Salo), €hn). rashid.giniatullin@uef.fi (R. Giniatullin), olli.grohn@uef.fi (O.H. Gro http://dx.doi.org/10.1016/j.neuropharm.2015.01.028 0028-3908/© 2015 Elsevier Ltd. All rights reserved.

Observation of CSD in MWA patients was also done using functional MRI (Hadjikhani et al., 2001). The excitatory neurotransmitter, glutamate, acting through via ionotropic N-methyl-D-aspartate (NMDA) receptors, is currently recognized as a key contributor to CSD induction and propagation (Pietrobon and Moskowitz, 2013; Somjen, 2001). For these reasons, NMDA receptors are a promising therapeutic target for clinical use. The NMDA receptor is one of the most abundant receptors in the mammalian brain, expressed in both neuronal and glial tissues (Verkhratsky and Kirchhoff, 2007) as well as in cerebral arteriolar endothelium (LeMaistre et al., 2012; Sharp et al., 2003). The receptor is a heterotetrameric ionotropic channel that is typically comprised of different combinations of NR1 and NR2A-D or NR3AB subunits (Monyer et al., 1992; Paoletti and Neyton, 2007). In particular, receptors with NR2A subunits normally mediate synaptic transmission whereas NR2B containing receptors are mainly expressed extrasynaptically (Sanz-Clemente et al., 2013) and in astrocytes (Dzamba et al., 2013). A growing body of evidence

A. Shatillo et al. / Neuropharmacology 93 (2015) 164e170

suggests that diverse brain disorders, including migraine, are based on the abnormal function of different NMDA receptor subtypes (Cull-Candy et al., 2001). Thus, NMDA-receptor subunits are attractive targets for CSD modulation (Lauritzen and Hansen, 1992; McLachlan, 1992; Peeters et al., 2007). Classical CSD features have been traditionally analyzed using electrophysiological and optical techniques that have revealed a very slow rate of propagation (2e8 mm/min) and circular patterns of this activity (EikermannHaerter and Moskowitz, 2008). However, conventional techniques such as optical methods, local field potentials (LFP), electrocorticography (ECoG), or electroencephalography (EEG) are mostly limited to the cortical surface and collect information from a finite number of sampling points. Another shortcoming of electrophysiological approaches is the inability to address the contribution of non-neuronal tissue components such as glia and vessels (Brennan et al., 2007). A complementary approach to study properties and mechanisms of CSD is the functional magnetic resonance imaging (fMRI) technique (Autio et al., 2014; Gardner-Medwin et al., 1994; Shatillo et al., 2013) using blood-oxygen level-dependent (BOLD) contrast (Ogawa et al., 1990). Based on the nature of the BOLD signal it is possible to indirectly evaluate complex neuronal activity in the brain through a process called neurovascular coupling (Logothetis et al., 2001). A positive BOLD signal reflects an excessive supply of oxygenated blood to satisfy metabolic demand in an activated brain region. Whereas negative BOLD signal may point to reduced neuronal activity (Shmuel et al., 2006) or inadequate blood supply, i.e. due to vasoconstriction (Ayata, 2013) vascular “stealing” phenomenon (Harel 2002) and/or local excessive energy demand. As an imaging entity, BOLD fMRI provides good spatial and temporal resolution for tracking CSD wave propagation in 3D space (see Supplementary video, S1). Supplementary data related to this article can be found online at http://dx.doi.org/10.1016/j.neuropharm.2015.01.028. The aim of this study was to test the action of three different NMDA-antagonists including the NR2B-selective drug ifenprodil, NR2A-selective TCN-201 and non-selective blocker MK-801 in vivo in a KCl-induced CSD animal model. For the first time we used whole-brain BOLD fMRI to investigate the basic spatio-temporal features of CSD in rats and the contribution of different subtypes of NMDA in this phenomenon. Our data revealed a key role of NR2B containing NMDA receptors in CSD. 2. Experimental procedures 2.1. Animal distribution Adult male Wistar rats (n ¼ 24, 10e12 weeks old, weighing 322 ± 29 g) obtained from the National Laboratory Animal Center of the University of Eastern Finland) were divided into 5 groups to be treated with intraperitoneal (i.p.) injection of MK801 (n ¼ 4, #M107, Sigma Aldrich GmbH, Seelze, Germany), ifenprodil (n ¼ 5, #I2892, Sigma Aldrich GmbH), TCN-201 (n ¼ 5, #4154, Tocris Bioscience, Bristol, UK), DMSO (n ¼ 3) and saline (n ¼ 5) prior to imaging. Ifenprodil and dizocilpine (MK801) dosage of 10 mg/kg was chosen based on the literature (Dempsey et al., 2000; Rod and Auer, 1989) We found no publications of in-vivo use of TCN-201 in rats, so we estimated its dosage at 10 mg/kg based on similarity in potency to Ifenprodil

165

Table 2 Physiological parameters of the animals before and after the CSD fMRI session (mean ± SEM, no statistically significant differences between groups in any of the parameters p > 0.2, paired t-test). Parameter

Before CSD

After CSD

pO2 pCO2 Respiration rate O2 saturation Core temperature

133.9 ± 19.9 mmHg 39.2 ± 5.1 mmHg 122.2 ± 14.2 bpm 97 ± 2.54% 36.8 ± 0.1  C

146.7 ± 16.0 mmHg 41.5 ± 9 mmHg 118 ± 18 bpm 97.4 ± 2.4% 36.9 ± 0.1  C

(TCN-201 IC50 ¼ 0.16 mM, Ifenprodil IC50 ¼ 0.11 mM (Monaghan et al., 2012). All used drugs solutions were prepared fresh at the day of imaging. CSD was evoked during the fMRI data acquisition with remote application of 10 ml of 1 M potassium chloride (Sigma Aldrich GmbH). The 6th group (n ¼ 2), also pretreated with saline, served as a double control, therefore equimolar (1 M) NaCl was applied to the cortex instead of KCl (Table 1) to test for the possible osmotic and physical effect of the application. 2.2. Animal preparation Rats were initially anesthetized with isoflurane (4% for induction and 2% for maintenance) in an inhalation mixture of 70% N2O, 30% O2 for surgical preparation. The right femoral artery was cannulated (PE10, Becton, Dickinson & Co, NJ, USA) for blood sampling during the imaging. Respiratory gases (pO2, pCO2) and pH were analyzed at the onset and immediately after the end of the fMRI scan from arterial blood samples, using a portable blood analyzer (i-Stat, Abbot Laboratories, Abbott Park, IL). In all experiments, physiological parameters such as body temperature, breathing rate, and blood gases (pCO2, pO2, and blood oxygen saturation) were maintained within normal physiological range and did not change significantly during the imaging (Table 2). A 3 mm2 cranial window was drilled into the parietal bone above the visual cortex (6 mm posterior from bregma, 2.5 mm lateral from midline) and the meningeal layers were carefully removed to expose the cortical surface. Maximum care was taken to avoid bleeding and injury of the exposed cortex. The cranial window was then covered with saline moisturized filter paper. The same polyethylene tubing as used for cannulation was glued (“Permabond 2011” cyanoacrylate, Permabond, Winchester, UK) to the skull adjacent to the exposed cortex for potassium chloride delivery. The cranial window was covered with Parafilm (Bemis, WI, USA) and the surgical wound was then filled with 2% agar gel (Sigma Aldrich GmbH) to minimize susceptibility artifacts during MRI imaging. Immediately after surgery, isoflurane was discontinued and anesthesia was switched to urethane (1.25 g/kg, i.p., Sigma Aldrich GmbH) for fMRI imaging. Urethane was selected as it provides steady and long-lasting anesthesia in single injection, has minimal effects on respiration, cardiovascular system and provides muscle relaxation (Maggi and Meli 1985). Furthermore it has been shown to be the most favorable for inducing CSD in rats (Kudo et al., 2008) and it provides robust BOLD fMRI responses during CSD (Shatillo et al., 2013; Autio et al., 2014). Body temperature was maintained at 37  C using a water heating pad (Gaymar Industries Inc., Orchard Park, NY) and breath rate was monitored using a pressuretransducing probe (SA Instruments Inc., Stony Brook, NY) throughout all experiments. All animal procedures were approved by the Animal Ethics Committee of the Provincial Government of Southern Finland, and conducted in accordance with the guidelines set by the European Commission Directive 2010/63/EEC. Rats were group housed with a preserved 12 h light/12 h dark cycle and ad libitum access to food and water. At the end of our experiments, animals were sacrificed by intravenous injection of saturated Kþ solution without discontinuing the anesthesia. All possible efforts were made to minimize both the number of animals and to optimize their welfare throughout. 2.3. MRI acquisition

Table 1 Distribution of subjects into groups according to the drug and vehicle combination for pretreatment and chemical agents used for cortical application. Group Group Group Group Group Group Group

1 2 3 4 5 6

(n (n (n (n (n (n

¼ ¼ ¼ ¼ ¼ ¼

5) 4) 5) 5) 3) 2)

Pretreatment drug

Cortical application

Saline MK-801 in saline Ifenprodil in saline TCN-201 in DMSO DMSO Saline

KCl, 1 M, ~7 ml

NaCl, 1 M, ~7 ml

All MRI experiments were performed on a 9.4 T/31 cm horizontal magnet system interfaced to a DirectDrive console (Agilent inc., Stony Brook, CA, USA). An actively decoupled volume radiofrequency coil and rat brain quadrature surface receiver coil (RAPID Biomedical, Rimpar, Germany) were used for signal transmission and reception respectively. Anatomical images were acquired using a T2 weighted fast spin-echo sequence (TR ¼ 3 s, effective TE ¼ 48 ms, echo spacing ¼ 16 ms, echo train length ¼ 8, FOV ¼ 5  5 cm2, matrix 512  512, resolution 98  98 mm2, slice thickness ¼ 0.75 mm, 10 slices). Functional MR data were acquired using a single-shot spin-echo echo-planarimaging sequence (TR ¼ 2 s, slice thickness ¼ 1.5 mm, 5 coronal slices). The field of view was 2.5 cm  2.5 cm with an acquisition matrix of 64  64 (in-plane resolution of 390  390 mm2) with 2 sagittal saturation bands placed outside the brain. After a local 3D fieldmap based shimming protocol, in the shimmed volume of

166

A. Shatillo et al. / Neuropharmacology 93 (2015) 164e170

10  6  5 mm3 (proton signal FWHM ¼ 35 ± 16 Hz), two functional MRI measurement sessions were performed in a queued manner. The first measurement consisted of 150 images (5 min) of baseline followed by either drug (MK-801, ifenprodil or TCN-201) or vehicle (saline or DMSO) i.p. injection and 1200 images (40 min) of follow up. In the second scan, after 150 baseline images (5 min) either KCl or NaCl was remotely delivered to the cortical surface and followed by a scan with 900 images (30 min). A microsyringe (100 ml, Hamilton Company, Bonaduz, GR, Switzerland) was used to consistently administer 7 ml of liquid through the tubing to the cortical surface. 2.4. Analysis

3.1. Spatio-temporal properties of CSD measured with BOLD fMRI Application of KCl to the cortical surface consistently evoked CSDs detected as a transient change in BOLD signal (e.g. Fig. 3B). Fig. 2 and supplementary video S1 shows the propagation of CSD across the cortex in different time points from the induction site in the posterior cortex (lateral secondary visual area, V2L) towards the frontal cortical regions as well as in a ventral direction. As seen on the images, the CSD mainly affected neocortical gray matter regions in ipsilateral to KCl application hemisphere. Wave was maintaining the omnidirectional propagation pattern irrespective of functional and anatomical zoning of the cerebral cortex within the affected hemisphere, never crossing the medial longitudinal fissure. As it is visible on these figures, the strongest CSD response was observed in the most superficial layers with little detectable involvement of deeper brain areas. No noticeable differences in CSD propagation pattern or involved regions were detected between the test groups.

In order to quantify CSD wave features in both spatial and temporal domains, we used two types of analysis. A group level ROI-based approach was used to describe and compare CSD wave features in single 5  5 voxel (2 mm2) ROIs drawn in the S1 cortical area (Fig. 2A, B). ROI placement site was chosen based on cortical thickness, functional specialization and distance from KCl injection site, so that the direct influence of application itself would not affect the results. Worth mentioning that BOLD response shape was similar both in more proximal and distant areas of the cortex (Fig. 5A). Then, to visualize propagation of the wave in the entire cortex, we employed individual model-free independent component analysis (ICA) to reveal all areas affected by the single CSD march. For those cases where multiple CSD waves were elicited (see the results section below), only the first wave was taken into analysis. For quantification purposes, the average timecourses from 25 adjacent voxels in the somatosensory cortex (Fig. 2B) were analyzed using AEDES (http://aedes.uef. fi/) and other in-house made Matlab routines. If more than one wave was present, always only the first evoked CSD was taken into analysis of the wave's properties. Each BOLD time series taken from specified ROIs was de-trended and normalized prior to statistical analysis. Each test group was compared against its own control group using ManneWhitney two-tailed t-tests with a significance level set at p < 0.05. To evaluate the level of motion that might contribute to the BOLD signal, we performed the center-of-mass displacement analysis of our data. In all cases, only the negligible motion (0.29 ± 0.02 of a voxel or 0.11 ± 0.01 mm, mean ± SEM) was observed in frequency encoding direction, suggesting that motion does not influence the data obtained. To estimate the in-plane speed of the CSD march, two ROIs were drawn in each subject at the same distance from one to another. The olfactory bulb edge was used as a landmark for ROI placement in the same cortical regions. The speed of CSD propagation was calculated from that distance and time between the peaks on the CSD BOLD response timecourse taken from these ROIs (Fig. 5A). For dynamic statistical mapping, the data were preprocessed using SPM8 (http://www.fil.ion.ucl.ac.uk/spm) and Matlab (Version 2011a, The Mathworks Inc, Natick, MA, USA). All volumes were first corrected for slice-timing differences, followed by motion correction and spatial smoothing with an 0.8  0.8  0.1 mm FWHM Gaussian kernel. The visualization of the spreading wave was done using two different approaches: Independent component analysis (ICA; (Comon, 1994) provided by the Group ICA of fMRI Toolbox (GIFT, version 2.0a, http://mialab.mrn.org/software/gift/) and GLM provided by SPM8. ICA separates data into subcomponents by assuming that components are statistically independent from each other. This can be used to separate the single passing wave into temporally separate components (Fig. 2A). With SPM8 analysis, the spreading pattern was described by shifting a 6-volumelong positive boxcar block over a portion of the time-series spanned by a single wave (Fig. 2B). Every block was tested separately from the others. The visualization of the wave progression on the 3D rat brain was done using rignac, France) time series visualization module Amira's (ver. 5.5.0, FEI VSG3D, Me with the t-threshold set to 3.5 (Supplementary video S1).

In the control group (n ¼ 6), CSD waves caused a relatively strong transient BOLD response with an average amplitude of 6.18 ± 0.84% (mean ± SEM, Fig. 3A) followed by a long lasting negative undershoot. Repetitive CSDs following the first wave with the interval of 4e8 min were observed in all rats tested with high [Kþ]. In contrast, the isosmotic concentration of Naþ did not evoke any CSDs (Fig. 3A), ruling out the direct osmolar effect of high concentration solution in our model. In order to test the general involvement of NMDA mechanisms we first tested the nonselective NMDA antagonist, MK-801. Application of MK-801 totally blocked all CSDs (Fig. 3C), indicating the validity of our approach and a key role of NMDA receptors in the generation of CSDs in this experimental model. The most interesting results were obtained with the NR2B selective blocker, ifenprodil. Pretreatment of rats with this NMDA antagonist reduced the number of CSD waves to just one in all animals tested (n ¼ 5, Figs. 3D, 4A). Most striking, however, was the dramatic change in the shape of the BOLD response. Fig. 3D shows that in the presence of ifenprodil the BOLD response to KCl application was initiated with a deep negative peak that was followed by a positive overshoot. Thus, ifenprodil “inverted” the BOLD signal in comparison to that of controls (compare Fig. 3B and D). In contrast to ifenprodil, the selective NR2A NMDA antagonist TCN-201 did not change the shape of the first CSD wave (Fig. 3F). As this agent was diluted in DMSO we performed a respective control experiment with injection of DMSO vehicle, which also had no effect on the CSD BOLD profile (Fig. 3E).

3. Results

3.3. Effect of NMDA antagonists on CSD propagation

To evaluate the effects of the NMDA antagonists on CSD and describe the spatio-temporal propagation of the wave, we used a 2step imaging protocol, depicted in Fig. 1.

Surprisingly, in spite of the differential effects of the tested NMDA antagonists on the initiation of the local BOLD response, these drugs did not alter the CSD propagation rate (Fig. 5B). The mean speed of

3.2. Effect of NMDA antagonists on amplitude and duration of the CSD BOLD response

Fig. 1. Timeline of the experiment with sequence of preparation procedures followed by the pharmacological (phMRI) and functional imaging.

A. Shatillo et al. / Neuropharmacology 93 (2015) 164e170

167

Fig. 2. A e Stacks of coronal MRI images of a representative rat brain with BOLD fMRI activation maps (t-maps, threshold ¼ 5), showing propagation of the wave from CSD induction site to distant areas of the brain. Application of KCL is taken as a zero time point. B e Combined map of all cortical regions affected by CSD wave progression from its induction to decay. The color bar represents the time in seconds when voxels were activated relative to KCL application. The circle represents the KCL application site and the square is the ROI used for BOLD timecourse quantification.

the CSD wave in all groups (except MK-801 group in which CSDs were not detected) was close to 6 mm/min with little variation between groups (5.9 ± 0.3 mm/min, mean ± SEM, see Fig. 5B). 3.4. Effect of drug administration on the BOLD response observed with pharmacological MRI The immediate effect of vehicle injection on the BOLD signal during our pharmacological MRI scans was negligible. The absence of any response was also observed after the injection of both MK801 and TCN-201 NMDA-antagonists and DMSO. In contrast, the injection of the NR2B selective antagonist, ifenprodil, had a prominent effect on the BOLD timecourse. It is seen as an immediate and rapid negative deflection of the BOLD signal followed by slower descent of the slope and leveling at a plateau that persisted for the entire duration of the pharmacological MRI scan (Fig. 6). Recovery of the baseline was not observed during following CSD fMRI session. It is worth noting that the effect was not localized to a particular region but observed simultaneously in the entire cerebral cortex. 4. Discussion The present study employed for the first time a BOLD fMRI approach for pharmacological investigation of CSD. Drastically different effects of NR2B and NR2A selective drugs on CSD BOLD response suggest differential involvement of NMDA receptors in this phenomenon with predominant role of NR2B containing receptors. fMRI measures brain activity indirectly through complex cascade that relates electrical brain activity to hemodynamic and blood oxygenation responses (i.e. neurovascular coupling) which eventually cause detectable BOLD signal changes. Even though it has been shown that electrical activity recorded as LFP strongly correlates with BOLD (Logothetis et al., 2001) and features of neurovascular coupling during CSD are known (Chang et al., 2010; Sonn and Mayevsky, 2000), interaction of drugs with any step in cascade leading to BOLD response may contribute to fMRI results. Therefore, in following discussion we will compare our results with other methods used in CSD research and discuss about possible effects of altered hemodynamic coupling to our results. Consistent with previous findings obtained with electrophysiological and optical imaging techniques (Peeters et al., 2007), we

also show using BOLD fMRI that the non-selective channel pore blocker MK-801 completely eliminated CSDs, confirming the key role of ionotropic NMDA glutamate receptors in the CSD mechanism and the validity of our experimental approach. Similarly, injection of MK-801 per se didn't have any dramatic effect on BOLD in phMRI portion of the imaging which is in accordance with previous studies showing neither strong disruption of CBF (Nehls et al., 1990) nor prominent activation (Roberts et al., 2008) in similar trials. It is generally accepted that two major subtypes of NMDA receptors, NR2A and NR2B are expressed in synaptic and extrasynaptic regions respectively (Loftis and Janowsky, 2003; SanzClemente et al., 2013). Moreover, while NR2A receptors can support neuronal survival, the NR2B subtype is involved in brain injury (Menniti et al., 2000) and pro-apoptotic mechanisms (Brittain et al., 2012; Hardingham and Bading, 2010). Previously, in vitro studies have shown that NR2A selective antagonists were effective in blocking CSD in chick retinas (Wang et al., 2012). However, in brain slices of mouse entorhinal cortex, the NR2B specific antagonist ifenprodil prevented CSDs from occurring (Faria and Mody, 2004). In our experiments, NR2A antagonist TCN-201 was not effective, as it did not influence the CSD BOLD response shape. This discrepancy may be attributed to special features of the models used as our readout modality is largely based on brain-specific hemodynamic responses. Regrettably, there is a lack of data on TCN-201 effects in-vivo, which would allow us to compare obtained results. However, high lipophilic properties of TCN-201 (see PubChem CID 4787937) and similar potency (Monaghan et al., 2012) suggests that the inability of the latter to modify CSDs reflects the minimal role of NR2A receptors in the current phenomenon. Unlike TCN-201, the NR2B selective antagonist, ifenprodil, significantly reduced the number of CSD waves but failed to block them completely whilst inverting the BOLD response. This is in contrast with the aforementioned results of Faria and Mody (2004) that were obtained in vitro from brain slices using electrophysiological methods, which therefore reflect only electrical neuronal effects. The nature of such modulation by ifenprodil in our experiments is not fully understood. The observed baseline BOLD signal change after Ifenprodil injection during phMRI scan and “inversion” of CSD BOLD response may indicate global cerebral hypoperfusion or hypoxic condition as observed in LDF experiments (Sukhotinsky et al., 2008). This could be possible if direct vasoconstriction caused by Ifenprodil would take place as described by

168

A. Shatillo et al. / Neuropharmacology 93 (2015) 164e170

Fig. 3. Cortical BOLD response after KCl application (highlighted by arrow) to the rat brain surface (ROI depicted in Fig. 1B). Thin blue lines are the individual cases and the thick gray line is the median of the group. A e Absence of response in saline pretreated animals with 1 M NaCl application instead of 1 M KCl. B e CSD waves detected in the BOLD signal from the control group of animals pretreated with saline. C e Complete blockade of CSD by injection of the MK-801 pore channel blocker. D e Altered CSD after NR2B antagonist ifenprodil injection, note the ”inverted” response pattern and complete absence of following secondary waves in contrast to the control group above. E e Vehicle (DMSO) control group for TCN-201 where DMSO served as a solvent solution. F e Effect of TCN-201 pretreatment on CSD (no significant difference with corresponding control).

Fig. 4. A e Group average of number of CSD waves evoked by single application of KCl to the cortex. B e BOLD response amplitude of the first peak measured from an ROI in the parietal cortex. (*p < 0.05, ***p < 0.0001).

A. Shatillo et al. / Neuropharmacology 93 (2015) 164e170

169

Fig. 5. A e In-plane CSD speed calculation was done by analyzing the timing (t) between wave peaks of the BOLD timecourses taken from 2 cortical ROIs (white squares 1, 2) drawn with a fixed distance of 5.2 mm. B e Group averaged propagation speed of the CSD wave.

studies on isolated cerebral vessels (Young et al., 1981). However, other in-vitro studies suggested the anti-vasoconstrictive effect of Ifenprodil (Arai et al., 1991) (McCool and Lovinger, 1995). These discrepancies may be attributed to different experimental models and registration methods, region-specific expression of receptors in the vessels and a complex nature of regulatory mechanisms modulating the vascular tone. Another possible explanation of such BOLD action could be based on the disruption of normal neurovascular coupling. This disruption may result from disabling the nitric oxide (NO) regulation of blood vessels (Bandettini, 2012; Toda et al., 2009), resulting in vasoconstriction promotion instead of normal CSD associated vasodilation (Ayata, 2013). Consistent with this view, it has been shown that NR2B containing NMDA receptors are linked to neuronal nitric oxide synthase (nNOS) that is activated by Ca2þ influx following NMDA receptor activation (Loftis and Janowsky, 2003). Activation of nNOS causes the release of NO, a critical agent responsible for reactive hyperemia by vasodilation and increase in CBF (Busija et al., 2007). Therefore it is likely that in our experiments, the blockade of NR2B containing NMDA receptors with ifenprodil diminished the normal dilatory vascular response to neuronal activation. This hypothesis complies with the observation of reduced rCBF response in mice lacking the nNOS gene expression (Ma et al., 1996). Thus, massive neuronal depolarization by CSD is not backed up with the immediate and adequate hemodynamic response, which causes inversion of the BOLD signal in the presence of ifenprodil. Since the CSD waveform is preserved in all other test groups, we can suggest a particular importance of NR2B containing receptors in the neurovascular response to CSD. Unlike the various changes in the shape of the BOLD response with different NMDA antagonists, neither of our NR2A nor NR2B antagonists led to alterations in the rate of propagation of CSDs. The

absence of statistically significant changes in CSD speeds in all treatment groups in our data may be explained by different mechanisms behind CSD induction (which are likely glutamatedriven) whereas propagation could be mainly based on potassium mediated mechanisms (Obrenovitch and Zilkha, 1995). Thus, the massive neuronal depolarization that occurs during CSDs triggers a transient rise in glutamate concentration (Iijima et al., 1998) followed by release of intracellular potassium. In agreement with Grafstein's “potassium theory” (Grafstein, 1956), Kþ accumulates in the extracellular space and after reaching the threshold concentration, further depolarization occurs in the surrounding cells, leading to a self-sustaining wave of spreading depression. Our data are largely consistent with this hypothesis. On contrary, recent data suggests that regenerative vesicular glutamate release mediated by presynaptic NMDA receptors (mostly of NR2B type) contributes and promotes the CSD ignition and propagation (Zhou et al., 2013). That may explain the effect of Ifenprodil on occurrence of CSD waves but maintained propagation rate across the groups is puzzling. According to the same study, CSD propagation rate is linearly correlated with logarithm of glutamate concentration and therefore should be altered in the presence of Ifenprodil. That inconsistency may be also attributed to the properties of BOLD fMRI and neurovascular coupling that limits the accuracy of propagation rate assessment. All mentioned findings and contradictions rise extremely interesting research questions, which need to be further addressed, in future. For instance, in-vivo testing and further characterization of NR2A antagonist effects in different concentrations of drugs is obviously needed. The mechanism behind Ifenprodil-induced BOLD reversal also requires comprehensive investigation. In summary, our functional testing of NMDA receptor involvement in CSD induction with fMRI BOLD signals reveals a key role for NR2B receptors in this phenomenon. If proven in human studies, these data can suggest novel strategies for the development of preventive medications for severe neurological conditions such as migraine, stroke and ischemia associated with CSD. Acknowledgments This work was supported by the Academy of Finland (252511), University of Eastern Finland and UEF-Brain strategic funding and Center for International Mobility (CIMO) TM-09-6241 fund. RG was partially supported by the program of competitive growth of Kazan Federal University.

Fig. 6. Averaged pharmacological MRI BOLD time series for ifenprodil and saline (control) treatment groups. Note the signal change in the ifenprodil group (red). Injection time is indicated by the black arrow. Note immediate decrease in the BOLD signal following the ifenprodil administration compared to the saline control (***p < 0.001, false discovery rate corrected t-test). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

References Arai, Y., Nakazato, K., Kinemuchi, H., Tadano, T., Satoh, N., Oyama, K., Kisara, K., 1991. Inhibition of rat brain monoamine oxidase activity by cerebral anti-ischemic agent, ifenprodil. Neuropharmacology 30, 809e812.

170

A. Shatillo et al. / Neuropharmacology 93 (2015) 164e170

€hn, O.H., 2014. Parenchymal spin-lock fMRI Autio, J.A., Shatillo, A., Giniatullin, R., Gro signals associated with cortical spreading depression. J. Cereb. Blood Flow Metab. 34, 768e775. Ayata, C., 2013. Spreading depression and neurovascular coupling. Stroke 44, S87eS89. Bandettini, P.A., 2012. Functional MRI: a confluence of fortunate circumstances. Neuroimage 61, A3eA11. pez-Valde s, H.E., Theriot, J., Toga, A.W., Brennan, K.C., Beltr an-Parrazal, L., Lo Charles, A.C., 2007. Distinct vascular conduction with cortical spreading depression. J. Neurophysiol. 97, 4143e4151. Brittain, M.K., Brustovetsky, T., Brittain, J.M., Khanna, R., Cummins, T.R., Brustovetsky, N., 2012. Ifenprodil, a NR2B-selective antagonist of NMDA receptor, inhibits reverse Naþ/Ca2þ exchanger in neurons. Neuropharmacology 63, 974e982. Busija, D.W., Bari, F., Domoki, F., Louis, T., 2007. Mechanisms involved in the cerebrovascular dilator effects of N-methyl-d-aspartate in cerebral cortex. Brain Res. Rev. 56, 89e100. Chang, J.C., Shook, L.L., Biag, J., Nguyen, E.N., Toga, A.W., Charles, A.C., Brennan, K.C., 2010. Biphasic direct current shift, haemoglobin desaturation and neurovascular uncoupling in cortical spreading depression. Brain 133, 996e1012. Charles, A.C., Baca, S.M., 2013. Cortical spreading depression and migraine. Nat. Rev. Neurol. 9, 637e644. Comon, P., 1994. Independent component analysis, a new concept? Signal Process.. Cull-Candy, S., Brickley, S., Farrant, M., 2001. NMDA receptor subunits: diversity, development and disease. Curr. Opin. Neurobiol. 11, 327e335. an, A., 2000. Attenuation of brain edema, bloodDempsey, R.J., Bas¸kaya, M.K., Dog brain barrier breakdown, and injury volume by ifenprodil, a polyamine-site N-methyl-D-aspartate receptor antagonist, after experimental traumatic brain injury in rats. Neurosurgery 47, 399e404 discussion 404e6. Dreier, J.P., 2011. The role of spreading depression, spreading depolarization and spreading ischemia in neurological disease. Nat. Med. 17, 439e447. Dzamba, D., Honsa, P., Anderova, M., 2013. NMDA receptors in glial cells: pending questions. Curr. Neuropharmacol. 11, 250e262. Eikermann-Haerter, K., Moskowitz, M.A., 2008. Animal models of migraine headache and aura. Curr. Opin. Neurol. 21, 294e300. Faria, L.C., Mody, I., 2004. Protective effect of ifenprodil against spreading depression in the mouse entorhinal cortex. J. Neurophysiol. 92, 2610e2614. Gardner-Medwin, A.R., van Bruggen, N., Williams, S.R., Ahier, R.G., 1994. Magnetic resonance imaging of propagating waves of spreading depression in the anaesthetised rat. J. Cereb. Blood Flow Metab. 14, 7e11. Grafstein, B., 1956. Mechanism of spreading cortical depression. J. Neurophysiol. 19, 154e171. Hadjikhani, N., Sanchez Del Rio, M., Wu, O., Schwartz, D., Bakker, D., Fischl, B., Kwong, K.K., Cutrer, F.M., Rosen, B.R., Tootell, R.B., Sorensen, A.G., Moskowitz, M.A., 2001. Mechanisms of migraine aura revealed by functional MRI in human visual cortex. Proc. Natl. Acad. Sci. U. S. A. 98, 4687e4692. Hardingham, G.E., Bading, H., 2010. Synaptic versus extrasynaptic NMDA receptor signalling: implications for neurodegenerative disorders. Nat. Rev. Neurosci. 11, 682e696. Iijima, T., Shimase, C., Iwao, Y., Sankawa, H., 1998. Relationships between glutamate release, blood flow and spreading depression: real-time monitoring using an electroenzymatic dialysis electrode. Neurosci. Res. 32, 201e207. Kudo, C., Nozari, A., Moskowitz, M.A., Ayata, C., 2008. The impact of anesthetics and hyperoxia on cortical spreading depression. Exp. Neurol.. Lauritzen, M., 1994. Pathophysiology of the migraine aura. The spreading depression theory. Brain 117 (Pt 1), 199e210. Lauritzen, M., Dreier, J.P., Fabricius, M., Hartings, J.A., Graf, R., Strong, A.J., 2011. Clinical relevance of cortical spreading depression in neurological disorders: migraine, malignant stroke, subarachnoid and intracranial hemorrhage, and traumatic brain injury. J. Cereb. Blood Flow Metab. 31, 17e35. Lauritzen, M., Hansen, A.J., 1992. The effect of glutamate receptor blockade on anoxic depolarization and cortical spreading depression. J. Cereb. Blood Flow Metab. 12, 223e229. Leao, A.A., 1944. Spreading depression of activity in the cerebral cortex. J. Neurophysiol. 7 (6), 359e390. Leao, A.A., Morison, R.S., 1945. Propagation of spreading cortical depression. J. Nerv. Ment. Dis. 102, 512. LeMaistre, J.L., Sanders, S.A., Stobart, M.J., Lu, L., Knox, J.D., Anderson, H.D., Anderson, C.M., 2012. Coactivation of NMDA receptors by glutamate and Dserine induces dilation of isolated middle cerebral arteries. J. Cereb. Blood Flow Metab. 32, 537e547. Loftis, J.M., Janowsky, A., 2003. The N-methyl-D-aspartate receptor subunit NR2B: localization, functional properties, regulation, and clinical implications. Pharmacol. Ther. 97, 55e85. Logothetis, N.K., Pauls, J., Augath, M., Trinath, T., Oeltermann, A., 2001. Neurophysiological investigation of the basis of the fMRI signal. Nature 412, 150e157. Ma, J., Ayata, C., Huang, P.L., Fishman, M.C., Moskowitz, M.A., 1996. Regional cerebral blood flow response to vibrissal stimulation in mice lacking type I NOS gene expression. Am. J. Physiol. 270, H1085eH1090. McCool, B.A., Lovinger, D.M., 1995. Ifenprodil inhibition of the 5hydroxytryptamine3 receptor. Neuropharmacology 34, 621e629.

McLachlan, R.S., 1992. Suppression of spreading depression of Le~ ao in neocortex by an N-methyl-D-aspartate receptor antagonist. Can. J. Neurol. Sci. 19, 487e491. Menniti, F.S., Pagnozzi, M.J., Butler, P., Chenard, B.L., Jaw-Tsai, S.S., Frost White, W., 2000. CP-101,606, an NR2B subunit selective NMDA receptor antagonist, inhibits NMDA and injury induced c-fos expression and cortical spreading depression in rodents. Neuropharmacology 39, 1147e1155. Milner, P.M., 1958. Note on a possible correspondence between the scotomas of ~o. Electroencephalogr. Clin. Neuromigraine and spreading depression of Lea physiol. 10, 705. Monaghan, D.T., Irvine, M.W., Costa, B.M., Fang, G., Jane, D.E., 2012. Pharmacological modulation of NMDA receptor activity and the advent of negative and positive allosteric modulators. Neurochem. Int. 61, 581e592. Monyer, H., Sprengel, R., Schoepfer, R., Herb, A., Higuchi, M., Lomeli, H., Burnashev, N., Sakmann, B., Seeburg, P.H., 1992. Heteromeric NMDA receptors: molecular and functional distinction of subtypes. Science 256, 1217e1221. Nehls, D.G., Park, C.K., MacCormack, A.G., McCulloch, J., 1990. The effects of Nmethyl-D-aspartate receptor blockade with MK-801 upon the relationship between cerebral blood flow and glucose utilisation. Brain Res. 511, 271e279. Obrenovitch, T.P., Zilkha, E., 1995. High extracellular potassium, and not extracellular glutamate, is required for the propagation of spreading depression. J. Neurophysiol. 73, 2107e2114. Ogawa, S., Lee, T.M., Kay, A.R., Tank, D.W., 1990. Brain magnetic resonance imaging with contrast dependent on blood oxygenation. Proc. Natl. Acad. Sci. U. S. A. 87, 9868e9872. Olesen, J., Larsen, B., Lauritzen, M., 1981. Focal hyperemia followed by spreading oligemia and impaired activation of rCBF in classic migraine. Ann. Neurol. 9, 344e352. Paoletti, P., Neyton, J., 2007. NMDA receptor subunits: function and pharmacology. Curr. Opin. Pharmacol. 7, 39e47. Peeters, M., Gunthorpe, M.J., Strijbos, P.J., Goldsmith, P., Upton, N., James, M.F., 2007. Effects of pan- and subtype-selective N-methyl-D-aspartate receptor antagonists on cortical spreading depression in the rat: therapeutic potential for migraine. J. Pharmacol. Exp. Ther. 321, 564e572. Pietrobon, D., Moskowitz, M.A., 2013. Pathophysiology of migraine. Annu. Rev. Physiol. 75, 365e391. Roberts, T.J., Williams, S.C., Modo, M., 2008. A pharmacological MRI assessment of dizocilpine (MK-801) in the 3-nitroproprionic acid-lesioned rat. Neurosci. Lett. 444, 42e47. Rod, M.R., Auer, R.N., 1989. Pre- and post-ischemic administration of dizocilpine (MK-801) reduces cerebral necrosis in the rat. Can. J. Neurol. Sci. 16, 340e344. Sanz-Clemente, A., Nicoll, R.A., Roche, K.W., 2013. Diversity in NMDA receptor composition: many regulators, many consequences. Neuroscientist 19, 62e75. Sharp, C.D., Fowler, M., Jackson, T.H., Houghton, J., Warren, A., Nanda, A., Chandler, I., Cappell, B., Long, A., Minagar, A., Alexander, J.S., 2003. Human neuroepithelial cells express NMDA receptors. BMC Neurosci. 4, 28. Shatillo, A., Koroleva, K., Giniatullina, R., Naumenko, N., Slastnikova, A.A., Aliev, R.R., Bart, G., Atalay, M., Gu, C., Khazipov, R., Davletov, B., Grohn, O., Giniatullin, R., 2013. Cortical spreading depression induces oxidative stress in the trigeminal nociceptive system. Neuroscience 253, 341e349. Shmuel, A., Augath, M., Oeltermann, A., Logothetis, N.K., 2006. Negative functional MRI response correlates with decreases in neuronal activity in monkey visual area V1. Nat. Neurosci. 9, 569e577. Smith, J.M., Bradley, D.P., James, M.F., Huang, C.L., 2006. Physiological studies of cortical spreading depression. Biol. Rev. Camb Philos. Soc. 81, 457e481. Somjen, G.G., 2001. Mechanisms of spreading depression and hypoxic spreading depression-like depolarization. Physiol. Rev. 81, 1065e1096. Sonn, J., Mayevsky, A., 2000. Effects of brain oxygenation on metabolic, hemodynamic, ionic and electrical responses to spreading depression in the rat. Brain Res. 882, 212e216. Sukhotinsky, I., Dilekoz, E., Moskowitz, M.A., Ayata, C., 2008. Hypoxia and hypotension transform the blood flow response to cortical spreading depression from hyperemia into hypoperfusion in the rat. J. Cereb. Blood Flow Metab. 28, 1369e1376. Toda, N., Ayajiki, K., Okamura, T., 2009. Cerebral blood flow regulation by nitric oxide in neurological disorders. Can. J. Physiol. Pharmacol. 87, 581e594. Verkhratsky, A., Kirchhoff, F., 2007. NMDA receptors in glia. Neuroscientist 13, 28e37. Wang, M., Chazot, P.L., Ali, S., Duckett, S.F., Obrenovitch, T.P., 2012. Effects of NMDA receptor antagonists with different subtype selectivities on retinal spreading depression. Br. J. Pharmacol. 165, 235e244. Woods, R.P., Iacoboni, M., Mazziotta, J.C., 1994. Brief report: bilateral spreading cerebral hypoperfusion during spontaneous migraine headache. N. Engl. J. Med. 331, 1689e1692. Young, A.R., Bouloy, M., Boussard, J.F., Edvinsson, L., MacKenzie, E.T., 1981. Direct vascular effects of agents used in the pharmacotherapy of cerebrovascular disease on isolated cerebral vessels. J. Cereb. Blood Flow Metab. 1, 117e128. Zhou, N., Rungta, R.L., Malik, A., Han, H., Wu, D.C., MacVicar, B.A., 2013. Regenerative glutamate release by presynaptic NMDA receptors contributes to spreading depression. J. Cereb. Blood Flow Metab. 33, 1582e1594.