| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
1Department of Neuroscience, University of Minnesota, Minneapolis, Minnesota; and 2Division of Neurology, Department of Neurology and Neuroscience, Weill Medical College of Cornell University, New York City, New York
Submitted 2 March 2005; accepted in final form 17 April 2005
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
, 2001; Ebner and Chen 2003). Surface stimulation typically evokes a confined beam of increased fluorescence due to activation of parallel fibers and their postsynaptic targets (Chen et al. 1998; Dunbar et al. 2004). Sufficiently intense stimulation can trigger a propagating event in which the beam of increased fluorescence spreads at high speeds across the folium. This propagating activity is characterized by a large increase in fluorescence and an extracellular acidification that outlasts the duration of stimulation. The spread is accompanied by a powerful but transient depression of the evoked field potentials in the cerebellar cortex and is referred to as spreading acidification and depression (SAD).
The exact mechanism(s) underlying SAD is unclear. Nor is the exact relationship of SAD to classical spreading depression (SD) or its variants understood (Ebner and Chen 2003; Somjen 2001). Previous studies in rat cerebellum implicated several factors in SAD, including both the presynaptic (i.e., parallel fibers) and postsynaptic (i.e., Purkinje cells) circuitry. The evidence suggests that increased cerebellar excitability plays a critical role (Chen et al. 1999, 2001). First, greater stimulation amplitude or frequency increases the probability of evoking SAD. Generally a train of stimuli is needed to evoke SAD, indicating that a threshold level of activation needs to be reached. Second, blocking AMPA receptors and/or metabotropic glutamate receptors (mGluR) lowers the probability of generating SAD. Blocking synaptic transmission by bathing the cerebellum in Ca2+-free Ringer abolishes SAD. Conversely, blocking GABA receptors increases the probability of evoking SAD. Therefore it is likely that evoking SAD requires elevated cerebellar excitability.
Voltage-gated potassium channels control neuronal excitability by multiple mechanisms (Hille 1992). Members of the Kv1 family of delayed-rectifier K+ channels are highly expressed in the cerebellar cortex and Purkinje cells, including channels composed of Kv1.1, Kv1.2, Kv1.3, and Kv1.6 subunits (Koch et al. 1997; Veh et al. 1995). Kv1 K+ channels on Purkinje cells regulate their output, including both Na+ and Ca2+ spikes (McKay et al. 2005). Therefore it is reasonable to hypothesize that the Kv1 family of delayed-rectifier K+ channels are involved in cerebellar SAD. Furthermore, the Kv1.1-containing channels are of specific interest because of their role in episodic ataxia type 1 (EA1), which involves mutations in the gene coding for the Kv1.1 
-subunit (Browne et al. 1994, 1995; Scheffer et al. 1998). A major symptom in EA1 is transient attacks of cerebellar dysfunction (Brandt and Strupp 1997; Eunson et al. 2000). This hallmark ataxia is hypothesized to be due to changes in the biophysical properties of K+ channels containing the Kv1.1 -subunit in the cerebellar cortex. Yet it remains unclear how the changes in the Kv1.1 -subunit translate into transient cerebellar dysfunction. Emphasis has been placed on the high density of K+ channels containing Kv1.1 -subunits at basket cell axon terminals and resultant increases in spontaneous GABA release from these cells (Southan and Robertson 1998; Zhang et al. 1999). However, this mechanism does not readily explain the episodic and transient nature of the ataxia in EA1 (Kullmann et al. 2001) and disregards changes in other cerebellar neurons with heterotetrameric channels containing Kv1.1 -subunits, including granule cells and their axons, as well as Purkinje cell somas and dendrites (Chung et al. 2001; Koch et al. 1997; Veh et al. 1995; Wang et al. 1994).
Recently we proposed that the transient cerebellar symptoms of EA1 are due to evoking SAD in the cerebellar cortex (Ebner and Chen 2003). In this hypothesis, changes in K+ channels containing the altered Kv1.1 -subunit lower the threshold for SAD. The resulting transient depression in the cerebellar cortex would produce an episodic ataxia. The aims of this study were to test the role of Kv1 channels in SAD and explore the possible connection between SAD and EA1.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
All animal experimentation was approved by the Institutional Animal Care and Use Committee of the University of Minnesota and conducted in conformity with the National Institutes of Health Guide for the Care and Use of Laboratory Animals. Experimental details on the animal preparation and optical imaging techniques in the mouse have been provided in previous publications (Dunbar et al. 2004; Gao et al. 2003) and are only briefly described here. Adult FVB mice (Charles River Laboratories, Wilmington, MA), 3–10 mo of age and of either sex, were anesthetized by intramuscular injection of a cocktail of ketamine (60 mg/kg), xylazine (3 mg/kg), and acepromazine (1.2 mg/kg). Animals were mechanically ventilated using 95% O2-5% CO2. The animal was placed in a stereotaxic frame, and body temperature was feedback-regulated. The electrocardiogram was monitored to assess the depth of anesthesia, allowing anesthetics to be supplemented as needed. After the craniotomy and creation of a watertight chamber around the exposed cerebellar cortex that included Crus I and II, the chamber was filled with a gassed (95% O2-5% CO2) Ringer solution (pH = 7.4). The chamber was periodically rinsed with fresh normal Ringer. For imaging requiring dyes, the brain was stained with intraperitoneal injections of a 0.2 ml solution (35 mM) of neutral red (3-amino-7-dimethylamino-2-methylphenazine hydrochloride), which stains well into the cerebellar cortex (Chen et al. 1998). Neutral red is known to intracellularly stain most neurons (LaManna and McCracken 1984), including cerebellar Purkinje cells (Kado 1993). Also we bath applied solutions of the voltage-sensitive dye, RH-1691 (0.25 mM, 2 h). Similar to most voltage-sensitive dyes, RH–1691 is likely to stain neuronal and glial membranes (Cohen et al. 1974; Sharon and Grinvald 2002).
Two strains of transgenic mice were also used: the Kv1.1 heterozygous knockout mouse (Smart et al. 1998) and cyclin D2 null mouse (C57bl/6J) (Huard et al. 1999). The Kv1.1 mutant mice were offspring of breeding pairs (heterozygous intercrosses) purchased from JAX Mice [C3HeB.129S7(B6)-Kcna1tm1Tem/J]. Kv1.1 and cyclin D2 null mice were genotyped with PCR techniques performed on DNA isolated from tail clips (Huard et al. 1999; Zhou et al. 1998). Results from heterozygous (Kv1.1) and null (cyclin D2) mice were compared with that from their age-matched wild-type littermates.
Electrical stimulation and electrophysiological monitoring techniques
Parallel fiber stimulation (100 µs pulses, 10 Hz for 10 s) was delivered by a tungsten microelectrode (1–3 M
) placed just below the cerebellar surface. As discussed in RESULTS, the probability of evoking SAD in the mouse was quite low. Therefore each animal was initially tested with a series of stimulation amplitudes, 100, 200, and 300 µA, to establish the baseline probability prior to testing the effects of various Kvl blockers or other drugs. Extracellular recordings of the evoked field potentials were obtained from the molecular layer with glass microelectrodes (2 M NaCl, 2–5 M), using conventional electrophysiological techniques (Chen et al. 1998). The evoked field potentials were used as a measure of cerebellar cortical excitability. When needed, field potentials were recorded simultaneously with the acquisition of the images (Chen et al. 1999, 2001).
Optical imaging
The optical imaging techniques using neutral red have been described previously (Chen et al. 2001; Dunbar et al. 2004). Images of the cerebellar surface were acquired by fixing the stereotaxic frame to an X-Y stage mounted on a modified Nikon epifluorescence microscope with a Quantix 57 cooled charge coupled device camera with 12-bit digitization (Roper Scientific, Tucson, AZ). After binning the final pixel resolution was 
14 x 14 µm. Light from a 150 W mercury-xenon lamp (Hamamatsu Photonics) was passed through an excitation filter (546 ± 10 nm), and the emitted light from the preparation was filtered through a 
590 nm long-pass filter with a 575-nm dichroic mirror. A typical acquisition protocol included a series of 20 control frames followed by a series of 600–1,200 experimental frames. Frame exposure time was 100–200 ms. In some experiments, flavoprotein autofluorescence or voltage-sensitive dye (VSD) imaging were used, which employed the same procedures except for the deletion of neutral red. Autofluorescence imaging used a 455 ± 35-nm excitation filter, a 500-nm extended reflectance dichroic mirror, and a >515-nm emission filter (Reinert et al. 2004). For VSD imaging with RH-1691, the exposed cerebellar cortex was imaged with a 630 ± 10-nm excitation filter, a 650-nm dichroic mirror, and a >665-nm emission filter (Sharon and Grinvald 2002).
The first step in image analysis was to subtract a control frame from the experimental frames. The fluorescence intensity of each pixel in an image was expressed as a function of the background fluorescence, FB, that is 
F/F = (FE –FB)/FB. The average of 19 control frames was used as the background fluorescence frame (FB) and the frame of interest as the experimental frames (FE). The average F/F was calculated for various regions of interest (i.e., along a beam, entire folium, site of injection). Plots of the average F/F over time were used to show the temporal profile of the fluorescence changes.
Two types of visualization were used. The first was simply based on the F/F for a select frame, in which the gray scale depicts the change in fluorescence relative to background. The second involved a thresholded activation map in which the pixels greater than or equal to a threshold were pseudo-colored and superimposed on an image of the background fluorescence (for details, see Gao et al. 2003).
Drug administration
Most drugs were purchased from Sigma (St. Louis, MO), including: glutamate (L-glutamic acid), dendrotoxin-K (DTX-K), neutral red, 1(s),9(R)-(–)- bicuculline methochloride (bicuculline), 3-amino-2-(4-chlorophenyl)-2-hydroxypropylsulfonic acid (saclofen), 5H-dibenz[b,f]azepine-5-carboxamide (carbamazepine), and N-(5-sulfamoyl-1,3,4-thiadiazol-2-yl)acetamide (acetazolamide). Tityustoxin-K (TsTX) was obtained from Alomone Labs (Jerusalem, Israel). RH-1691 was purchased from Optical Imaging Inc. (Mountainside, NJ). For most compounds, stock solutions (1–10 mM) were made in saline, stored at –20°C and diluted to the required concentration in normal Ringer solution before each experiment. If required, the solution was prepared fresh on the day of use. Drugs were applied to the exposed cerebellar surface by superfusion of the drug solution in the chamber or intraperitoneal injection. In general, the effects of DTX-K and other Kv1 K+ channel blockers were evident within 2–5 min of being applied to the cerebellar cortex. In assessing the dose dependence on the threshold for evoking SAD, the experimental protocol was initiated 15 min after drug application to ensure uniform drug penetration. Microinjection of glutamate (10–42 mM, 0.2 µl over 0.2 s) was performed with a glass microelectrode using a pico-injection system (PLI-100, Medical System, Greenville, NY).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
The present experiments were performed in the mouse, in contrast to our previous studies on SAD that primarily used the rat (Chen et al. 1999, 2001). This allowed us to take advantage of several transgenic mouse models. In the course of these studies, it became clear that the probability of evoking SAD in the mouse by parallel fiber stimulation was markedly lower than in the rat. Based on our previous studies (Chen et al. 1999, 2001), SAD was evoked in 79% (167/212) of rats tested using a range of surface stimulation amplitudes (100–300 µA). In the present study, SAD was evoked in normal Ringer in only 12% (6/51) of the FVB mice by surface stimulation using a similar range of amplitudes. Therefore the mouse cerebellar cortex is less susceptible to SAD than the rat cerebellar cortex. Nevertheless, as detailed in the following text, the susceptibility to SAD in the mouse was greatly increased by blocking Kv1 potassium channels.
Role of Kv1 channels in SAD
To determine the role of Kv1 channels in SAD, and in particular the role of channels containing Kv1.1 -subunits, the effects of Kv1 blockers were studied. Figure 1 illustrates the effect of DTX-K on the threshold for evoking SAD. Stimulation of the cerebellar surface at 100 µA produced an optical beam due to activation of parallel fibers and their postsynaptic targets (Chen et al. 1998; Dunbar et al. 2004). Increasing the stimulation strength to 200 and 300 µA evoked wider and stronger responses but failed to evoke SAD. After superfusion of the cerebellar cortex with 50 nM of DTX-K, 100 µA stimulation readily evoked SAD with a speed of 920 µm/s (Fig. 1, A and B). Identical results were obtained in 10 animals. The significance of evoking SAD using 100-µA stimulation needs to be stressed. Surface stimulation at this amplitude did not evoke SAD in normal Ringer in 117 mice that were examined in this and other studies. Furthermore, SAD was evoked in all FVB mice when DTX-K (60 nM) was used. Therefore interfering with channels containing Kv1.1 -subunits greatly lowers the threshold for evoking SAD.
|
300 µm caudal to the electrode evoking the SAD (Chen et al. 1999, 2001). As the increased fluorescence spread over the folium, both the presynaptic field potentials (P1/N1) representing the parallel fibers and the postsynaptic response (N2) were abolished (Fig. 1, C and D). The onset of the suppression of the field potentials and the fluorescence increase were very tightly linked (Fig. 1D). The decrease in the presynaptic component persisted for 70 s and the postsynaptic depression 120 s. A similar depression was observed in eight animals in which the field potentials were monitored during 23 SAD events. The average duration of the presynaptic depression was 34 ± 25 s, and the average duration of the postsynaptic depression was 47 ± 23 s. SAD monitored by different imaging techniques
Previous studies of SAD have relied on neutral-red-based optical imaging (Chen et al. 1996, 2001). This raises the question of whether neutral red is altering the physiology of the cerebellar cortex and contributing to SAD. To answer this question, VSD and autofluorescence imaging were employed, using the same procedures including application of DTX-K but without neutral red staining (Fig. 2). After staining the cerebellar cortex with RH-1691 (Sharon and Grinvald 2002), surface stimulation evoked the expected beam of increased fluorescence (peak of 0.38% F/F). If the stimulation amplitude was sufficiently strong, this initial beam of activity propagated throughout the folium (Fig. 2A). For the example shown, the fluorescence rapidly rose to an initial peak and then with a second increase reached a peak of 2.1% F/F in 13 s and returned to baseline 100 s after the onset of the spread. The propagation speed was 1,250 µm/s. Spread monitored with VSD was also accompanied by a depression of the excitability in both the pre- and postsynaptic cerebellar cortical circuitry. In the example shown, the presynaptic component (P1/N1) was depressed for 50 s and the postsynaptic component (N2) for 70 s. Similar results were obtained in three animals. When flavoprotein autofluorescence imaging was used without application of external dyes, surface stimulation evoked a beam of increased fluorescence that spread across the entire folium (500 µm/s). Depression of the pre- and postsynaptic field potentials accompanied the spread (Fig. 2B). Similar findings were obtained in seven animals. The propagation pattern, speed, and time course of the spreading phenomenon monitored by VSD and autofluorescence imaging and the associated depression in excitability were essentially identical to those of SAD monitored with neutral red imaging. These results demonstrate that the propagating phenomenon underlying SAD is not dependent on the presence of neutral red. Furthermore, the large increases in fluorescence with flavoprotein and VSD imaging show that this phenomenon involves large changes in both oxidative metabolism and the state of neuronal/glia depolarization, respectively.
|
Blocking Kv1.1 containing channels with DTX-K not only increased the probability of evoking SAD but also resulted in spontaneous SAD. Surface stimulation in the presence of DTX-K initiated SAD in Crus II along the evoked optical beam as shown in Fig. 3 (A and B, top). However, spontaneous SAD was also observed in the same folium at a propagation speed of 760 µm/s without surface stimulation, which originated from a spontaneous beam of parallel fiber-like activity (Fig. 3, A and B, bottom). This beam occurred in a more posterior location than the stimulation evoked beam. Spontaneously occurring SAD was similar to SAD initiated by surface stimulation in its spatial pattern of propagation, speed, and time course. For both evoked and spontaneous SAD, the spread always originated from the beam and propagated perpendicular to the beam. The average propagation speed and time course were 565 ± 223 (SD) µm/s and 154 ± 71 (SD) s (same convention for all results), respectively, for surface stimulation evoked SAD and 656 ± 90 µm/s and 143 ± 64 s for spontaneous SAD. The average speed and duration were not significantly different for evoked and spontaneous SAD (P > 0.05, Student's t-test). For the example shown, the peak F/F was lower for the spontaneous SAD (15%) versus the evoked (45%). This difference in peak amplitude was within the normal variability observed for SAD evoked by surface stimulation. Spontaneous SAD was documented in 8 of 71 animals in which DTX-K was used. The percentage of spontaneous SAD is likely higher than observed here, because the experimental procedures were not designed to detect spontaneous events. Spontaneous SAD was never recorded in normal Ringer.
|
-subunits not only lowered the threshold but also created a condition in which SAD has multiple cycles. The effect of DTX-K in lowering the threshold enough to evoke SAD with multiple cycles suggests that Kv1.1 containing potassium channels play a major role in controlling excitability in the cerebellar cortex, and blocking these channels leaves the cerebellar cortex highly susceptible to SAD.
To determine the briefest interval (refractory period) at which another SAD can be evoked, the surface stimulus was repeated at varying intervals after the onset of SAD. As shown in Fig. 3D, additional surface stimulations at 100 and 150 s were ineffective but stimulation at 200 s did initiate another SAD. Similar results were obtained in five animals tested, with the average refractory period of 208 ± 30 s. This refractory period is longer than the 90 s found in rats (Chen et al. 1999), also consistent with the higher threshold for evoking SAD in mice.
Therefore blocking Kv1.1 containing potassium channels created the conditions in which SAD occurred spontaneously and displayed multiple cycles, consistent with an increase in neuronal excitability specifically involving the parallel fibers. To further test for effects on cortical excitability, we first examined the effects of DTX-K on the field potentials evoked by surface stimulation. In nine animals, 50 nM DTX-K did not produce statistically significant changes in the field potentials, similar to other findings that DTX-K did not effect extracellular field potentials (Southan and Owen 1997). A second approach to test whether DTX-K altered the excitability of the parallel fibers used microinjections of glutamate into the upper molecular layer (within 100 µ of the surface). These experiments utilized autofluorescence imaging of the unstained cerebellar cortex to take the advantage of the greater temporal resolution of this technique (Reinert et al. 2004). In normal Ringer, the injection of glutamate evoked a circular region of fluorescence increase around the tip of injection electrode (Fig. 3, E, top, and F). The signal was detected within 200 ms of injection onset, and reached a peak of 0.8% (F/F). There was a gradual increase in the intensity and diameter of the response over 1 s, consistent with the diffusion of glutamate from the injection site. In the presence of DTX-K, glutamate injection also evoked a beam-like optical response (Fig. 3E, bottom), similar to that evoked by electrical stimulation. Identical results were obtained in the five animals tested. The average signal intensity at the injection site was similar in normal Ringer and in DTX-K (P > 0.5, Student's t-test, Fig. 3F). In contrast, there was essentially no optical signal outside of the injection site in normal Ringer but a robust response with DTX-K (P < 0.01, Student's t-test). Therefore in the presence of DTX-K, parallel fiber excitability was increased.
The DTX-K used in this study also blocks Kv1.2 subunits to some degree (Sigma). To further assess the role of Kv1.2 potassium channels on SAD, a primary Kv1.2 channel blocker, TsTX (Dodson et al. 2002; Ellis et al. 2001; Hopkins 1998) was used and the effects compared with those of DTX-K (Fig. 4). Shown in Fig. 4A are the probability curves for evoking SAD at the lowest and highest stimulation amplitudes as a function of drug concentration (0–200 nM). Surface stimulation at 100 µA, which did not evoke SAD in mice in normal Ringer, evoked SAD in 50% of animals with 60 nM DTX-K. With 300 µA surface stimulation, the probability of evoking SAD was increased to 30, 90, and 100% with DTX-K concentrations of 40, 50, and 60 nM, respectively. TsTX also increased the probability of evoking SAD; however, a higher concentration was required. When stimulating at 100 µA, SAD was first evoked at 100 nM TsTX. At the highest concentration tested (200 nM), SAD occurred in only 10% of animals at low stimulation amplitude (100 µA), and in only 80% at the strongest stimulation (300 µA). 
2 analysis of the SAD frequency and drug concentrations showed a significant effect for each drug, but DTX-K has a lower effective concentration. At 100 µA stimulation, a significant increase in SAD occurred at 50 nM for DTX-K (2 = 36, df = 7 and P < 0.01) but at 100 nM for TsTX (2 = 90, df = 8 and P < 0.01). The SAD50, the concentration of a blocker at which SAD was evoked in 50% of animals, was 63 and 200 nM at 100 µA stimulation and 45 and 95 nM at 300 µA stimulation for DTX-K and TsTX, respectively. Furthermore, the SAD50 differed significantly at both 100 and 300 µA stimulation amplitudes: DTX-K < TsTX (logistic regression, 2 = 45, df = 1 and P < 0.01).
|
SAD in Kv1.1 heterozygous knockout mice
Although the Kv1.1 knockout mouse (homo- or heterozygous) does not have an obvious cerebellar phenotype consistent with EA1, the homozygous animal does have a phenotype consistent with neuronal hyperexcitability (Smart et al. 1998; Zhou et al. 1998). However, EA1 is a dominant disorder, and the patients are heterozygous. This led us to examine whether the threshold for evoking SAD was lowered in the heterozygous animal, consistent with a subclinical change in cortical excitability that does not result in a behavioral phenotype. Surface stimulation did not evoke SAD in either the wild-type littermate controls or the +/– animals in normal Ringer (Fig. 5). DTX-K increased the probability of evoking SAD for both groups, but the dose-dependent curve from the heterozygous animals was shifted to the left of the curve from the wild-type littermates (Fig. 5, A and B). At 300 µA stimulation, the lowest concentration of DTX-K required in the heterozygous animals was 25 nM compared with 50 nM in the wild-type controls. The SAD50 for the heterozygous animals was 35 nM but 65 nM for the littermate controls. The probability of SAD versus the drug concentration showed a significant difference between the heterozygous animals and their wild-type controls in the threshold concentration for evoking SAD (2 = 7, df = 1 and P < 0.05). Therefore the probability of initiating SAD was increased in the Kv1.1 heterozygous mouse compared with their wild-type littermates.
|
Both carbamazepine, a tricylic antidepressant, and acetazolamide, a carbonic anhydrase inhibitor, are used to reduce the frequency and severity of ataxia in EA1 patients (Eunson et al. 2000; Lubbers et al. 1995). If SAD is the underlying mechanism for the episodic ataxia in EA1, carbamazepine and acetazolamide would be predicted to reduce the probability and/or severity of SAD. Therefore in a series of experiments in FVB mice, the effects of these two therapeutic agents on the probability of evoking SAD were examined.
In these experiments, DTX-K (60 nM) was added to the Ringer to ensure that SAD could be evoked for both the control and drug treatments. As in previous experiments, surface stimulation amplitude was varied (100, 200, and 300 µA) to establish the threshold, and SAD was initiated in all animals in the presence of DTX-K prior to treatment. The addition of carbamazepine markedly increased the threshold for evoking SAD (Fig. 6). For the example shown in A and B, stimulation at 200 µA evoked SAD in the presence of DTX-K. However, SAD was not evoked when carbamazepine (0.5 mM) was also added to the Ringer. Stronger stimulation (300 µA) failed to elicit SAD and generated only a beam (Fig. 6, A, middle, and B). After washout of the carbamazepine, SAD was again evoked by 200 µA surface stimulation (Fig. 6, A, right, and B). The effective concentration of carbamazepine to block SAD was 0.35–0.5 mM. The generation of an optical beam in the presence of carbamazepine showed that the parallel fibers and Purkinje cells were functioning normally. There were also no significant changes in the field potentials after the application of carbamazepine at these concentrations (Fig. 6C). However, prolonged superfusion of carbamazepine at a higher concentration (>1 mM) did produce a reversible reduction in the field potentials (data not shown). The lack of an effect on the field potential and the preservation of the optical beam during the application of carbamazepine indicate that the blockade of SAD was not due to a general reduction or loss of cerebellar excitability. Carbamazepine was effective in preventing SAD in five of seven animals (Fig. 6D). Acetazolamide was also effective in reducing the probability of evoking SAD. When acetazolamide (2.5 mg/ml, 50 mg/kg) was given by intraperitoneal injection 75 min before imaging acquisition, SAD was evoked in only one of six animals, even in the presence of 60 nM DTX-K (Fig. 6D). As for carbamazepine, acetazolamide did not alter cortical excitability, since surface stimulation evoked normal field potentials and a normal optical beam (data not shown). Interestingly, topical application of acetazolamide (4 mM) was not as effective in the prevention of SAD (evoked in 4 of 5 animals). In summary, carbamazepine and acetazolamide significantly reduced the probability of evoking SAD with DTX-K to 29 and 18%, respectively, compared with 100% in the absence of the drugs (2 = 54, df = 1 and P < 0.01).
|
Both Kv1.1 and Kv1.2 -subunits are located on basket axon terminals (Wang et al. 1994), raising the question of the role of GABAergic transmission in SAD. Previous studies in the rat found that bicuculline and saclofen increased the likelihood of SAD (Chen et al. 2001). To extend these observations to the mouse, GABA receptor antagonists were used to assess the contribution of the cerebellar inhibitory circuitry to SAD in the presence of DTX-K (Fig. 7, A and B). Initially, surface stimulation produced an optical beam but not SAD. The optical beam was enhanced in width and intensity when the GABAA antagonist, bicuculline (100 µM), and the GABAB antagonist, saclofen (250 µM), were added to the Ringer, demonstrating the effectiveness of the GABAergic blockade consistent with previous studies (Chen et al. 2001; Dunbar et al. 2004). The presence of GABA receptor antagonists did not block spread because SAD was evoked after the addition of DTX-K (50 nM). In five animals tested, SAD was evoked with the addition of DTX-K, confirming that blocking GABAergic neurotransmission does not reduce the likelihood of evoking SAD in the cerebellar cortex.
|
). Surface stimulation in the presence of DTX-K (50 nM) evoked SAD in the three null and two wild-type littermates (Fig. 7C). There were no obvious differences in the onset, amplitude, and time course of SAD between the two groups. The results using GABA receptor antagonists and the cyclin D2 null mice demonstrate that intact GABAergic neurotransmission is not required for the generation of SAD. |
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
; Leao 1944; Momose-Sato et al. 2001; Newman and Zahs 1997) (for review, see Somjen 2001). Previously, SAD had only been observed using neutral red optical imaging, a technique based on pH changes (Chen et al. 1999, 2001; Ebner and Chen 2003). This raises several questions, including whether neutral red was altering cerebellar cortical physiology to generate SAD and whether non-pH-related changes were occurring. The VSD and autofluorescence imaging results conclusively demonstrate that evoking SAD is not dependent on neutral red staining. Furthermore, the increased fluorescence observed with RH-1691, which reflects depolarization of neurons and/or glia (Sharon and Grinvald 2002; Shoham et al. 1999), is consistent with the decrease in cerebellar cortical excitability. This further links the pH changes with changes in neuronal excitability. The shifts in pH are known to modulate neuronal excitability. For example, acidification depresses voltage-gated sodium channels (Tombaugh and Somjen 1996) and glutamate receptor conductance (Traynelis and Cull-Candy 1991). Conversely, a wave of depolarization would be expected to result in inactivation of voltage-gated Na+ channels, contributing to the depression in parallel fibers-Purkinje circuitry (Hille 1992). The degree of depolarization and associated ion fluxes do not appear sufficient to result in an extracellular DC shift in SAD (Chen et al. 1999). The flavoprotein autofluorescence results demonstrate that oxidative metabolism increases during SAD, likely due to the depolarization that accompanies the depression of excitability (Duchen 1992; Mayevsky and Chance 1974). Therefore changes in pH, membrane potential, and metabolism occur during this propagated wave of activity. Kv1 potassium channels in SAD
Several factors play a role in the generation and propagation of SAD, including glutamatergic synaptic transmission, parallel fibers and Purkinje cells, and cerebellar cortical excitability (Chen et al. 2001). Blocking Kv1 channels, which are essential in regulating neuronal excitability, dramatically lowered the threshold and increased the probability for evoking SAD. Both Kv1.1 and Kv1.2 blockers had similar effects, and both types of subunits are found in high densities in the cerebellar cortex (Chung et al. 2001; Koch et al. 1997).
We postulate that blocking Kv1 channels increases the probability and lowers the threshold for evoking SAD by increasing the excitability of the cerebellar cortex, consistent with the role of these channels in controlling neuronal excitability and preventing hyperexcitability (Hille 1992; Smart et al. 1998). Two findings support the hypothesis that Kv1 blockers increase parallel fiber excitability. First, injecting glutamate into the upper molecular layer results in parallel fiber beam activation in the presence of DTX-K but does not in normal Ringer. The beam is due to direct activation of the parallel fibers by the injected glutamate. Recent studies show that parallel fibers possess extrasynaptic ionotropic glutamate receptors, including AMPA and kainate subtypes, and that activation of these receptors can activate parallel fibers, in vitro (Levenes et al. 2001; Zhao et al. 1997). It is important to stress that the beam is generated within 200–600 ms of glutamate injection; this rules out the possibility that the beam is generated by granule cell activation due to glutamate diffusion. With a 0.1-µl injection, the highest glutamate concentration used, it takes glutamate 4 s to diffuse to the granular layer and reach a concentration of 100 nM. The time can be derived from the diffusion equation for glutamate in the cerebellum, using a distance of 300 µm from the injection site to the granular layer (Barbour and Hausser 1997; Popa et al. 2004). It has been estimated that glutamate at a concentration of 500 nM is not capable of binding sufficient receptors to activate granule cells (Barbour and Hausser 1997).
Second, the occurrence of spontaneous beams that develop into SAD in the presence of DTX-K also supports the concept of increased excitability of granule cells and/or parallel fibers (Fig. 3). Although spontaneous SAD was observed in only a small number of experiments, this likely reflects that the data collection protocol was not designed to detect these events. Because peripheral inputs can evoke SAD in the rat cerebellar cortex (Ebner et al. 2004), blocking Kv1.1-containing potassium channels is likely to produce a state in the cerebellar cortex in which afferent activity initiates SAD. These spontaneous beams may also be relevant to the debate on the role of the parallel fibers and whether these axons are actually activated along a folium (Bower 1997; Eccles et al. 1967). The spontaneous beams observed with DTX-K are the first demonstration that parallel fibers can be activated without direct electrical stimulation. Therefore the circuitry is capable of generating beam-like activity under certain conditions.
The preceding discussion does not imply that parallel fibers are the only elements of the circuitry involved in SAD or that Kv1 K+ channel blockers only alter parallel fiber excitability. The voltage-sensitive dye and flavoprotein autofluorescence imaging demonstrate that SAD is a complex phenomenon altering many physiological processes. As noted in the preceding text, Kv1 channels are widely expressed in the cerebellar cortex. For example, Kv1 K+ channels in Purkinje cells are critical for regulating their excitability including Na+ simple spike firing, Ca2+-Na+ bursts, and Ca2+ spikes (McKay et al. 2005). The threshold for evoking the latter two is lowered by Kv1 blockers. Also, Kv1 blockers alter the excitability of climbing fiber afferents and result in spontaneous complex spikes (McKay et al. 2005). Conceivably, climbing fiber activation with its large effects on Purkinje cell excitability could contribute to SAD (Llinas and Nicholson 1976; Llinas and Sugimori 1980; Schmolesky et al. 2002; for review, see Loewenstein et al. 2005). In the presence of DTX-K, peripheral input can evoke SAD, suggesting a role for mossy fiber and/or climbing fiber input (Ebner et al. 2004). Therefore components of the cerebellar circuit other than parallel fibers are likely also involved in SAD.
The lack of an effect on the presynaptic volley of the field potential recordings suggests that DTX-K does not change the time course of the parallel fiber action potentials. Another Kv1.1/Kv1.2 blocker, -dendrotoxin, alters basket cell transmitter release and increases the excitability of visceral neurons but does not alter the shape and amplitude of the action potential (Dodson et al. 2002; Glazebrook et al. 2002; Southan and Robertson 1998; Tan and Llano 1999). Nor does DTX-K simply increase parallel fiber-Purkinje cell synaptic transmission because the postsynaptic component of the field potential was unchanged, which is consistent with a recent report that parallel fiber excitatory synaptic input to Purkinje cells is not shaped by Kv1 K+ channels (McKay et al. 2005). Other studies have observed little or no effect on extracellularly recorded field potentials at the concentrations of DTX-K used in this study (
100 nM). For example, in the rat hippocampus, concentrations of DTX-K of 1 µM or greater are required to alter CA1 or dentate gyrus field potentials (Southan and Owen 1997).
Determining whether DTX-K or TsTX is the more effective blocker, and therefore establishing whether Kv1.1 or Kv1.2 is more critical in the generation of SAD, is challenging in these in vivo studies. Published data indicate an IC50 for TsTX (0.15–0.55 nM for Kv1.2 -subunits) that is 10–30 times smaller than for DTX-K (5.3 nM for Kv1.1 -subunits) (Hopkins 1998; Wang et al. 1999), suggesting that TsTX is the more potent antagonist. However, these IC50 values are based on testing homotetrameric channels in various expression systems. Both Kv1.1 and Kv1.2 homotetrameric channels are likely nonexistent in the cerebellar cortex, where Kv1.1 and Kv1.2 subunits are almost always found in the same channels (Akhtar et al. 2002; Koch et al. 1997). Also, the effectiveness of Kv1 blockers has been shown to be highly dependent on the CNS structure under investigation (Hopkins 1998; Robertson et al. 1996). Furthermore, Kv1.1- and Kv1.2-containing potassium channels are found on several subtypes of neurons, including Purkinje cell, granular cells, and parallel fibers (Chung et al. 2001; Koch et al. 1997; Veh et al. 1995; Wang et al. 1994). At high concentrations, DTX-K, MgTX, and other Kv1 channel toxins have indistinguishable effects on Purkinje cells (McKay et al. 2005). Therefore caution should be taken in interpreting the relative effectiveness of Kv1 blockers and the roles in SAD of the specific Kv1 channels targeted. Nevertheless, several observations suggest that DTX-K was the more effective agent in facilitating the generation of SAD. First, the dose-dependent curves and stimulation thresholds were consistent with DTX-K being highly effective. Second, DTX-K resulted in spontaneously occurring "parallel fiber" beams that progressed to SAD. Third, DTX-K also resulted in SAD with multiple cycles. Therefore blocking channels containing Kv1.1 subunits is particularly effective in creating the conditions needed for SAD.
The relationship between SAD and SD and its variants has not been fully explored. We have emphasized some of the differences in the properties and pharmacology of SAD and SD (Chen et al. 1999; Ebner and Chen 2003). The present findings implicating the malfunction of Kv1 K+ channels in SAD is of interest because K+ channels and K+ ion shifts play a major role in SD (Nicholson and Kraig 1975; Shimizu et al. 2000; Vyskocil et al. 1972). We are unaware of studies specifically examining the effects of Kv1 K+ channel blockers in SD. Such studies would add to our understanding of the similarities and differences among the various forms of propagating depression in the CNS.
Possible link between SAD and EA1
The involvement of Kv1.1 -subunits in controlling SAD suggests a connection to EA1. This connection is strengthened by the finding that the Kv1.1 heterozygous knockout mouse also had a lowered threshold for evoking SAD. Although these mice as well as the null have no obvious cerebellar phenotype, the observation showed that manipulation of Kv1.1 containing channels altered the susceptibility to SAD. Furthermore, the relative difficulty in evoking SAD in the mouse may explain the lack of a cerebellar phenotype in the knockout animal. In a recent EA1 model based on knocking in the 408A EA1 mutation (V408A/+ mice), evoking symptoms required the injection of isoproterenol to mimic severe stress (Herson et al. 2003).
Both of the commonly used therapies for EA1, carbamazepine and acetazolamide, reduced the probability of evoking SAD. Presumably carbamazepine reduces neuronal excitability, consistent with its known inhibitory action on voltage-gated Na+ channels (Kuo et al. 1997; Soderpalm 2002). The action of acetazolamide is less clear; cortical superfusion was substantially less effective than pretreatment with a systemic dose. Although not directly interacting with the Kv1.1 channel protein (Bretschneider et al. 1999), acetazolamide has a host of actions on pH homeostasis in neurons and in the extracellular space (see Chesler 2003 for review). Although speculation, it is intriguing that blocking carbonic anhydrase augments pH changes associated with neuronal activity (Chen and Chesler 1992; Chesler 2003; Chesler and Kaila 1992; Kraig et al. 1983). Potentially, the increased pH could reduce the likelihood of generating SAD by altering pH buffering in the extracellular space.
The cerebellar dysfunction in EA1 has been hypothesized to be due to an abnormal increase in GABA release from basket cells (Herson et al. 2003; Southan and Robertson 2000; Zhang et al. 1999). While an increase in the frequency of spontaneous inhibitory postsynaptic currents (IPSCs) have been described with Kv1.1 blockers, in the Kv1.1 knockout mouse and the V408A/+ mice, there was no reported effect on evoked inhibitory currents or spontaneous basket cell firing rates (Herson et al. 2003; Southan and Robertson 2000; Zhang et al. 1999). An increase in the amplitude of spontaneous inhibitory postsynaptic potentials (IPSCs) occurs in the V408A/+ mice. Reconciling the present hypothesis that SAD underlies that the episodic cerebellar symptoms of EA1 with the GABA release hypothesis is difficult. Blocking GABA receptors did not reverse the effect of DTX-K on SAD and cyclin D2 null animals lacking stellate cells have normal SAD (Fig. 7), consistent with earlier findings that GABA receptor antagonists lower the threshold for evoking SAD (Chen et al. 2001). The effectiveness of bath application of GABA antagonists in this preparation has been demonstrated in our previous studies (Chen et al. 1998, 2001; Dunbar et al. 2004). Although Kv1.1-containing channels are enriched on basket cell axon terminals at the pinceau (Wang et al. 1994), the lack of effect by GABA receptor blockers suggests that functioning GABAergic transmission is not required for SAD. In the cerebellar slice, K+ channels containing Kv1.1 -subunits are found on Purkinje cell dendrites and somas as well as granule cells and parallel fibers (Chung et al. 2001; Koch et al. 1997; Veh et al. 1995), and both parallel fibers and Purkinje cells are involved in SAD (Chen et al. 2001). The present results are consistent with this concept based on increase in parallel fiber excitability in response to glutamate injection and the spontaneous parallel fiber-like beams observed with DTX-K. The finding that the presynaptic field potential, and not just the postsynaptic component, was shut down during SAD also implies that SAD involves more than altered GABAergic transmission. Therefore the lowered threshold for evoking SAD may result from a more general increase in the excitability of cerebellar neurons. Distinguishing between or reconciling these two theories will require additional studies.
Finally, these results support the hypothesis that the transient cerebellar symptoms of EA1 are due to evoking SAD in the cerebellar cortex (Ebner and Chen 2003). The brief but profound depression in cerebellar cortical excitability in SAD disrupts cerebellar cortical function and would likely produce short paroxysms of ataxia. The attacks in EA1 generally last seconds to minutes (Brandt and Strupp 1997; Brunt and van Weerden 1990; Kullmann et al. 2001), consistent with the time course of SAD (Chen et al. 2001). Many of the alterations in Kv1.1 potassium channels in EA1 increase cerebellar cortical excitability (Adelman et al. 1995; Boland et al. 1999) and would presumably increase the likelihood of evoking SAD. Some EA1 families have an epilepsy phenotype consistent with this hyperexcitability concept (Eunson et al. 2000). With this increased neuronal excitability, stress, startle or exertion can evoke SAD that results in ataxia. The observation of spontaneous SAD in the presence of DTX-K and that peripheral inputs can evoke SAD are consistent with the clinical findings that episodic cerebellar symptoms can be similarly triggered. In summary, the hypothesis that SAD underlies the cerebellar dysfunction in EA1 is based on the finding that they share a common molecular mechanism: alteration of Kv1.1-containing potassium channels. Furthermore, this hypothesis explains the episodic nature, triggering, and time course of the attacks. If proven true, this would provide not only insights into the mechanisms of EA1 and similar episodic disorders, but also a model for testing pharmacological agents that may prevent or reduce the cerebellar dysfunction in these conditions.
|
GRANTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
|
FOOTNOTES |
|---|
Address for reprint requests and other correspondence: T. J. Ebner, Dept. of Neuroscience, University of Minnesota, Lions Research Bldg., Rm. 421, 2001 Sixth St. S.E., Minneapolis, MN 55455 (E-mail: ebner001{at}umn.edu)
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
Akhtar S, Shamotienko O, Papakosta M, Ali F, and Dolly JO. Characteristics of brain Kv1 channels tailored to mimic native counterparts by tandem linkage of alpha subunits: implications for K+ channelopathies. J Biol Chem 277: 16376–16382, 2002.
Anderson TR and Andrew RD. Spreading depression: imaging and blockade in the rat neocortical brain slice. J Neurophysiol 88: 2713–2725, 2002.
Barbour B and Hausser M. Intersynaptic diffusion of neurotransmitter. Trends Neurosci 20: 377–384, 1997.[CrossRef][ISI][Medline]
Boland LM, Price DL, and Jackson KA. Episodic ataxia/myokymia mutations functionally expressed in the Shaker potassium channel. Neuroscience 91: 1557–1564, 1999.[CrossRef][ISI][Medline]
Bower JM. Is the cerebellum sensory for motor's sake, or motor for sensory's sake: the view from the whiskers of a rat? Prog Brain Res 114: 483–516, 1997.
Brandt T and Strupp M. Episodic ataxia type 1 and 2 (familial periodic ataxia/vertigo). Audiol Neurootol 2: 373–383, 1997.[Medline]
Bretschneider F, Wrisch A, Lehmann-Horn F, and Grissmer S. Expression in mammalian cells and electrophysiological characterization of two mutant Kv1.1 channels causing episodic ataxia type 1 (EA-1). Eur J Neurosci 11: 2403–2412, 1999.[CrossRef][ISI][Medline]
Browne DL, Brunt ER, Griggs RC, Nutt JG, Gancher ST, Smith EA, and Litt M. Identification of two new KCNA1 mutations in episodic ataxia/myokymia families. Hum Mol Genet 4: 1671–1672, 1995.
Browne DL, Gancher ST, Nutt JG, Brunt ER, Smith EA, Kramer P, and Litt M. Episodic ataxia/myokymia syndrome is associated with point mutations in the human potassium channel gene, KCNA1. Nat Genet 8: 136–140, 1994.[CrossRef][ISI][Medline]
Brunt ER and van Weerden TW. Familial paroxysmal kinesigenic ataxia and continuous myokymia. Brain 113:1361–1382, 1990.
Chen G, Dunbar RL, Gao W, and Ebner TJ. Role of calcium, glutamate neurotransmission, and nitric oxide in spreading acidification and depression in the cerebellar cortex. J Neurosci 21: 9877–9887, 2001.
Chen G, Hanson CL, Dunbar RL, and Ebner TJ. Novel form of spreading acidification and depression in the cerebellar cortex demonstrated by neutral red optical imaging. J Neurophysiol 81: 1992–1998, 1999.
