Transgenic mouse models in migraine
Ven, R.C.G. van de
Citation
Ven, R. C. G. van de. (2007, November 6). Transgenic mouse models in migraine. Retrieved from https://hdl.handle.net/1887/12473
Version: Corrected Publisher’s Version
License: Licence agreement concerning inclusion of doctoral thesis in the Institutional Repository of the University of Leiden
Downloaded from: https://hdl.handle.net/1887/12473
Note: To cite this publication please use the final published version (if applicable).
FHM1 S218L Knockin Mouse Model to study Hemiplegic
Migraine, Ataxia and Mild Head Trauma-triggered Sudden
Death
O F
U R
CHAPTER 4
R.C.G. van de Ven1#, D. Pietrobon2#, S. Kaja3#, T. Pizzorusso4#, F.E.
Hoebeek5, D. Engel6, B. Todorov1, L.A.M. Broos1, †M. Fodor7, C.I.
de Zeeuw5, R.R. Frants1, N. Plesnila6, J.J. Plomp8,9, M.D. Ferrari8, A.M.J.M. van den Maagdenberg1,2
#Authors contributed equally
†Dr. M. Fodor is deceased
Departments of 1Human Genetics, 7Anatomy & Embryology, 8Neurology and 9Molecu- lar Cell Biology – Neurophysiology, Leiden University Medical Centre, Leiden, The Netherlands
2Department of Biomedical Sciences, University of Padova, Padova, Italy
3Michael Smith Laboratories, The University of British Columbia, Vancouver, Canada
4Scuola Normale Superiore, Instituto Neuroscienze CNR, Pisa, Italy
5Department of Neuroscience, Erasmus Medical Center, Rotterdam, The Netherlands
6Department of Neurosurgery and Institute for Surgical Research, University of Mu- nich Medical Centre – Grosshadern, Munich, Germany
Manuscript in preparation
Summary
Familial hemiplegic migraine type 1 (FHM1) is a Mendelian subtype of migraine with aura that is caused by mutations in Cav2.1 Ca2+channel subunit gene CACNA1A. We introduced FHM1 mutation S218L - associated with hemiplegic migraine, ataxia, seizures and mild head trauma-triggered brain edema - into mice using homologous recombination.
Homozyous S218L knockin mice exhibited ataxia and spontaneous head-trauma induced mortality associated with seizures, closely resembling the human phenotype. Mutation S218L revealed multiple gain-of-function effects, including increased Cav2.1 current density in cerebellar neurons, enhanced neurotransmission at the neuromuscular junction and a reduced threshold and increased velocity of cortical spreading depression, all with a clear allele-dosage effect. Thus, we could demonstrate that dysfunctional Cav2.1 channels result in increased susceptibility for cortical spreading depression and neuronal hyperexcitability, and are the underlying cause of these neurological phenotypes. The Cacna1a S218L knockin mouse is a valuable model for the understanding and treatment of severe diseases as FHM, ataxia, epilepsy, and trauma.
Introduction
Familial hemiplegic migraine (FHM) is a rare autosomal dominant subtype of migraine with aura characterized by hemiparesis during attacks.1 Attacks of FHM can sometimes be more severe with prolonged aura, fever and loss of consciousness, often provoked by mild head trauma. In a majority of families, FHM is caused by mutations in CACNA1A (FHM1)2, encoding the pore-forming subunit of Cav2.1 calcium channels that play an important role in controlling neurotransmitter secretion3,4 and, postsynaptically, neuronal excitability.5 In at least 20% of FHM patients, there are additional paroxysmal and progressive neurological features such as cerebellar ataxia and atrophy, epilepsy, and mental retardation.6 Our recent FHM1 knockin (KI) model harbouring the R192Q mutation – resulting in pure FHM in patients2, revealed that increased susceptibility for cortical spreading depression (CSD, the underlying cause of migraine aura)7 may be due to cortical hyperexcitability.8
Heterozygous FHM1 mutation S218L produces a particularly severe phenotype of ataxia, hemiplegic migraine triggered by mild head trauma associated with cerebral edema, and sometimes fatal coma, rarely preceded by a generalized seizure.9-11 Mild head trauma can result – especially in children - in devastating neurological consequences of impaired consciousness and convulsions that are sometimes fatal.12 Consequences of mutation S218L on Cav2.1 channel function in expression systems are most severe of all FHM1 mutation studied13-15, with a large negative shift in activation of single channels, a large increase in whole cell current density, an increased rate of recovery from inactivation, and a large component of slowly inactivating current, for instance at conditions of long depolarization as they occur during CSD.15 The unique properties of S218L channels may cause a longer duration of CSD and a severe clinical phenotype.
Here we determined the in vivo consequences of mutated S218L Cav2.1 channels to understand the cause of the complex phenotype in S218L patients, by generating knockin mice with the FHM1 S218L mutation. Homozygous S218L mice have a severe phenotype that closely resembled the human phenotype, with ataxia, spontaneous and trauma-induced mortality, due to sometimes lethal seizures. Functional analysis revealed gain-of-function effects on Ca2+ channel current, neurotransmission, and CSD in an allele-dosage manner. Our results suggest that increased Cav2.1 channel activity is the underlying mechanism that causes these clinical symptoms
Methods
Generation and basic characterisation of S218L mice
The Cacna1a gene was modified such that exon 5 contained the human FHM1 S218L mutation. Transgenesis was performed as described previously.8 In short, in the targeting vector the original TCA triplet codon 218 was changed into TTA by site-directed mutagenesis, creating the S218L mutation. Upstream of exon 5, a PGK- driven neomycin cassette flanked by LoxP sites was introduced. Embryonic stem cells (E14) were electroporated and clones were selected for homologous recombination by Southern blot analysis using external probes as indicated in Figure 1A. The presence of the S218L mutation was confirmed by PCR of exon 5 using primers 271 5’- CTCCATGGGAGGCACTTG-3’and 272 5’-ACCTGTCCCCTCTTCAAAGC-3’ and subsequent digestion with restriction enzyme VspI as well as direct sequence analysis of PCR products. Correctly targeted ES cells were injected into C57Bl/6J blastocysts to create chimaeric animals. Offspring that was heterozygous for the S218L+Neo allele was bred with mice of the EIIA-Cre strain16 to remove the neomycin cassette. S218L mice in which the neomycin cassette was successfully deleted were backcrossed to C57Bl/6Jico mice. Heterozygous (S218L/+) and homozygous (S218L/+) and wild-type littermates of the third generation were used for all analyses (~87.5% C57Bl/6Jico and ~12.5%
129Ola background). All animal experiments were performed in accordance with the guidelines of the respective Universities and national legislation. For all experiments, the investigator was blinded for genotype. Genotyping of mice for all experiments was performed by PCR as described above.
RNA analysis
Total RNA was isolated from cerebellum as described previously.8 For RT-PCR, first- strand cDNA was synthesized using random primers and subsequent PCR was performed using Cacna1a and Cyclophilin specific primers. For Northern blot analysis, 10 µg of cerebellar RNA was separated on a 1% agarose gel and subsequently transferred to a Hybond-N+ membrane (Amersham Biosciences, Buckinghamshire, UK). 32P-labeled PCR products of Cacna1a or Cyclophilin cDNA were used as probes using standard hybridization and washing conditions.
Western blot analysis
Cerebellar membrane fractions were prepared and Western blot analysis was performed as described previously.8 In brief, protein fractions were separated by SDS-PAGE and transferred onto a nitrocellulose membrane. Subsequently, blots were incubated with
primary Cav2.1 (α1) antibody (Chemicon, Temecula, CA, USA) (1:200) together with primary β-actin antibody (A2066, St Louis, MO, USA) (1:1000). Secondary peroxidase- labelled goat anti-rabbit antibody (Jackson ImmunoResearch Laboratories, West Grove, PA, USA) (1:1000) incubation was performed and Western blotting was done according to enhanced chemiluminescence protocols (Amersham Biosciences, Piscataway, NJ, USA).
Histology
Brains from adult mice (~3 months-old) were fixed in 4% PFA for 2 hours. Paraffin sections (5 µm) were prepared and HE and Klüver-Barrera staining was performed using standard protocols.
Immunohistochemistry
Brains of adult mice (~3 months-old) were obtained after perfusion with PBS followed by 4% PFA fixation. Post-fixation was done for 1 to 2 hours in 4% PFA followed by o/n incubation in 10% sucrose/0.1 M phosphate buffer at 4 ºC. Subsequently, membranes were removed, tissue was embedded in 10% sucrose/11% gelatine, and gelatin was fixed with 30% sucrose/10% formaline for 2.5 hours at room temperature, followed by o/n incubation in 30% sucrose/0.1 M phosphate buffer at 4 ºC. After freezing down, coronal sections of 40µm were processed free-floating during whole procedure. For immunohistochemistry, antigen retrieval was performed for 30 min at 80 ºC in 10 mM citrate buffer (pH 5.5).
Sections were incubated in 10% heat-inactivated FCS/0.5%TX100/TBS for 2 hours and then incubated with primary Cav2.1 α1 antibody (#ACC-001, Alomone Labs, Jerusalem, Israel) (1:200 diluted in 2% heat-inactivated FCS/ 0.4%TX100/TBS at 4 ºC. Secondary biotin-labeled goat anti-rabbit antibody (Vector Laboratories, Burlingame, CA) (1:200 diluted in the same buffer) incubation was performed for 2 hours at room temperature.
Finally, for detection, sections were incubated with the avidin-biotin kit (Vector Laboratories, Burlingame, CA) for 2 hours at room temperature, washed, and developed in 0.1 mg/ml diaminobenzidine with 0.005% H2O2.
Golgi Cox staining
Golgi Cox staining was performed as described by Glaser and Van der Loos.17 In brief, Golgi Cox solution was prepared by mixing 5% K2Cr2O7, 5% HgCl2 and 5% K2CrO4 (Sigma-Aldrich, Zwijndrecht, the Netherlands), ‘ripened’ for 5 days and the precipitate that is formed was removed. Animals were deeply anesthetized with Nembutal and perfused intracardially with PBS. The brains were immediately removed from the cranium and incubated 3 weeks in Golgi Cox solution in dark. Brains were embedded
into celloidin and 200 µm sections were prepared. The sections were developed by sequential incubation in H2O, 16% NH4OH, H2O, 1% Na2S2O3, H2O, 70% EtOH, 96%
EtOH, butanol and Histo-clear® (National Diagnostics USA, Atlanta, GA). Finally, sections were mounted with HistomountTM (National Diagnostics USA, Atlanta, GA) and stored in dark.
Survival Analysis
Survival curves were constructed using the method of Kaplan and Meier.18 The survival curves were compared by log rank test. Survival times were calculated as the date of birth to date of death of unknown cause, or date of censor. Mice that died of unknown cause were considered cases. Sacrificed animals were censored. For statistical analysis SPSS 10.0 software was used.
Electroencephalography
EEG electrodes were surgically implanted under isoflurane anesthesia in 12-week-old mice. Six stainless steel crew electrodes (1 mm diameter) were positioned as follows: one in each frontal bone, one in each medial bone. The remaining two electodes (temporal bone) served as reference (medial) and ground (lateral) electrodes. Isolated cupper wire was used to attach the electrodes to a miniconnector that was cemented to the skull with dental acrylic. The first two hours after surgery no recording was performed. During recording, animals were placed in a sound-proof, lighted cage, which acted as a Faraday cage. Electric signals were amplified, filtered (CyberAmp, Cambridge Electronic Devices (CED), Cambridge, UK), digitized (1401Plus, CED), and stored for off-line analysis (Spike2, CED). EEG signals were recorded between 0.1 – 30 Hz and sampled at 500 Hz. Synchronously, behaviour of free-moving mice was continuously monitored and recorded at 25Hz by custom-made software (Labview, National Instruments, Austin, Tx, USA). Recording sessions took up to 48 hours during which food and water were available ad libitum.
Traumatic brain injury
Mice (25-28 g) were anaesthetized with 1.2% halothane, 30% O2, and 69% N2O. A large craniotomy was performed above the right parietal cortex leaving the dura mater intact.
Moderate brain trauma was performed using a controlled cortical impact (CCI) device as described previously.19 Thereafter the craniotomy was closed with the initially removed bone flap and dental acrylic and the skin was closed. Animals were transferred to an incubator heated to 35°C where they recovered within less than 5 minutes. Thereafter mice were returned to their cages and observed for 24 hours during which food and water
were available ad libitum
Inverted Screen
The inverted screen was a 43-cm square of wire mesh consisting of 12 mm squares of 1 mm diameter wire. It was surrounded by a 4-cm wooden bead to prevent the mouse from climbing on to the other side. Each mouse was placed in the centre of the wire mesh screen and the screen rotated to the inverted position over 2 s, with the head of the mouse declining first. The screen was held steadily ±30 cm above a padded surface. The time at which the mouse fell off was noted, or the mouse removed when the criterion time of 60 s was reached. Performance was scored according to the following scheme: falling off between 0 and 10 s = 1; 11 and 25 s = 2; 26 and 60 s = 3; achieving 60 s = 4.
Grip Strength test
Muscle strength was measured using a grip strength meter for mice (600 g range;
Technical and Scientific Equipment GmbH, Bad Homburg, Germany), connected to a laptop computer. The test was carried out essentially as originally described for rats.20 The peak force of each trial was considered the grip strength. Each mouse performed five trials, each about 30 s apart. The mean value of the five trials was used for statistical analysis.
Horizontal bar
The apparatus consisted of a 0.1-cm thick and 40 cm long metal wire which was horizontally held 60 cm above the padded bench surface by two supports. Each mouse was held by the tail, placed with its front paws on the bar, and rapidly released so that it grasped the horizontal bar at the central point. The criterion point was either a fall from the bar before the mouse reached one of the end supports of the bar, or the time till one forepaw touched the support. Performance was scored according to the following scheme: falling off between 0 and 5 s = 1; 6 and 10 s = 2; 11 and 20 s = 3; 21 and 30 s = 4; achieving 30 s or reaching a support = 5.
Rotarod
The accelerating rotarod test was performed essentially as described by Kistler et al.21 Briefly, an accelerating rotarod (UGO Basile S.R.L., Commerio VA, Italy) with a 4 cm diameter horizontal rotating rod was used. The test was performed in the semi-dark room with a light source placed at the bottom to prevent mice from jumping off the rotarod.
Mice were tested in groups of five, selected blindly to the genotypes. Following a training period (mice were placed on the rotarod moving with a low constant speed of 5 rpm for 5
minutes) and a 30-min resting period, mice were subjected to six trails. Each trail started with the rotarod moving at a constant speed of 5 rpm and after 10 seconds the speed was gradually increased to 45 rpm over a period of 5 minutes. The time each mouse managed to stay on the rod was recorded (e.g endurance). The endurance per trail per genotype was presented as means ± SEM and was plotted on a graph.
Cerebellar granule cell electrophysiology
Cerebellar granule cells were grown in primary culture from 6-day-old mice as described previously.22 Experiments were performed on cells grown from 6 to 7 days in vitro.
Whole-cell patch-clamp recordings were performed at room temperature as in Fletcher et al.22 External recording solution (in mM): BaCl2 5, tetraethylammonium (TEA)–Cl 148, Hepes 10 (pH 7.4 with TEA-OH) and 0.1 mg/ml cytochrome C. Internal solution (in mM): Cs-methanesulfonate 100, MgCl2, 5, Hepes 30, EGTA 10, ATP 4, GTP 0.5 and c-AMP 1 (pH 7.4 with CsOH). Currents were sampled at 5 kHz and low-pass filtered at 1 kHz. Compensation (typically 70%) for series resistance was generally used.
Current-voltage (I-V) relationships were obtained only from cells with a voltage error of <5 mV and without signs of inadequate space clamping such as notch-like current discontinuities, slow components in the decay of capacitative currents (in response to hyperpolarizing pulses) or slow tails not fully inhibited by nimodipine. The average normalized I-V curves were multiplied by the average maximal current density obtained from all cells. I-V curves were fitted with Eq. 1: I = G (V-Erev) (1+exp((V1/2 -V)/k))-1 using a non-linear regression method based on the Levenberg-Marquardt algorithm. The liquid junction potentials were such that a value of 12 mV should be subtracted from all voltages to obtain the correct values of membrane potential in whole-cell recordings.22 Averages are given as mean ± SEM.
All drugs were stored as stock solutions at –20 °C: 250 µM ω-conotoxin-GVIA (ω- CgTx-GVIA, Bachem, Budendorf, Switzerland) and 250 µM ω-conotoxin-MVIIC (ω- CTx-MVIIC, Bachem, Budendorf, Swizerland ) in distilled water, 10 mM nimodipine (gift from Dr. Hof, Sandoz, Basel, Switzerland) in 95% ethanol.
Ex vivo neuromuscular junction electrophysiology
Mice (~2 months of age) were euthanized by carbon dioxide inhalation. Phrenic nerve- hemidiaphragms were dissected and mounted in standard Ringer’s medium (in mM:
NaCl 116, KCl 4.5, CaCl2 2, MgSO4 1, NaH2PO4 1, NaHCO3 23, glucose 11, pH 7.4) at room temperature and continuously bubbled with 95% O2 / 5% CO2. The number of mice tested was expressed as n, and 7-15 NMJs were tested per muscle.
Intracellular recordings of miniature endplate potentials (MEPPs, the spontaneous
depolarizing events due to uniquantal ACh release) and endplate potentials (EPPs, the depolarization resulting from nerve action potential-evoked ACh release) were made at NMJs at 28 °C using standard micro-electrode equipment, as described previously.23 At least 30 MEPPs and EPPs were recorded at each NMJ, and typically 7-15 NMJs were sampled per experimental condition per muscle. Muscle action potentials, were prevented by 3 µM of the selective muscle Na+ channel blocker µ-conotoxin-GIIIB (Scientific Marketing Associates, Barnet, Herts, UK). In order to record EPPs, the nerve was electrically stimulated with supramaximal pulses of 100 µs duration at 0.3 Hz and 40 Hz. Amplitudes of EPPs and MEPPs were normalized to –75 mV, assuming 0 mV as the reversal potential for ACh-induced current.24 The normalized EPP amplitudes were corrected for non-linear summation according to McLachlan and Martin25 with an ƒ value of 0.8. The quantal content at each NMJ, i.e. the number of ACh quanta released per nerve impulse, was calculated by dividing the normalized and corrected mean EPP amplitude by the normalized mean MEPP amplitude. Broadness of EPPs was determined by measuring the width of the EPP at 50% of its peak amplitude (half-width).
Measurements were also made following a 15 min pre-incubation with specific Cav channel blockers: ω-Agatoxin-IVA (200 nM; Cav2.1), ω-conotoxin-GVIA (2.5 μM; Cav2.2) and SNX-482 (1 μM; Cav2.3). Toxins were from Scientific Marketing Associates. Experiments involving 3,4-diaminopyridine (DAP, Sigma, Zwijndrecht, The Netherlands) were made following a 30 min pre-incubation.
Possible statistical differences were analyzed with paired or unpaired Student’s t-tests or analysis of variance (ANOVA) with Tukey’s HSD post-hoc test, where appropriate, on grand mean values with n as the number of mice tested, and 7-15 NMJs tested per muscle. P<0.05 was considered to be statistically significant. Data is presented as mean
± SEM.
Cortical spreading depression.
CSD was recorded as described.8 Briefly, mice (20-30 g) were anaesthetised with urethane (20% in saline; 6 ml/kg i.p.). Animals, mounted on a stereotaxic apparatus were continuously monitored for adequate level of anaesthesia, temperature, heart rate and nociceptive reflexes. To record CSD three holes were drilled in the skull over the left hemisphere. The first corresponded to the occipital cortex and was used for electrical stimulation. The second hole, at the parietal cortex (1 mm M-L, 1 mm behind bregma) and the third hole, at the frontal cortex (1 mm M-L, 1 mm anterior to bregma), were used for CSD recording. The steady (DC) potential was recorded with glass micropipettes 200 μm below the dural surface. An Ag/AgCl reference electrode was placed subcutaneous above the nose. Cortical stimulation was conducted using a copper bipolar electrode
placed on cortical tissue after removing the dura. Pulses of increasing intensity (10 up to 800 mA) were applied for 100 ms at 3 min intervals by using a stimulus isolator/
constant current unit (WPI, USA) until CSD was observed. The minimal stimulus intensity at which CSD was elicited was recorded as the CSD threshold. When CSD was elicited, recordings were continued for 1 hr to assess recurrent CSDs. To estimate CSD propagation velocity, the distance between the recording electrodes divided by the time elapsed between the CSD onset at the first and second recording sites. CSD amplitude was measured subtracting form the minimal voltage value during CSD the baseline
Wild-type allele Targeting vector S218L+Neo allele
S218L allele
Neo H
E1 H X E1
E1 E1
K
K Neo
H E1
E1 K
E4 E5 E6
K K
K K
X
X
E1 K X E1
E4 E6
K
K E1
S218L S218L
S218L 5’ probe
5’ probe
5’ probe
3’ probe 3’ probe 3’ probe
H E
E E
E1E E
E E
a
7.7 kb 9.9 kb KpnI; Probe 5’
b Eam1105I; Probe 3’
8.5 kb 10.7 kb
+/+ S218L/+ S218/S218L
Cyclophilin Cacna1a
+/+ S218L/+ S218L/S218L 190 kD Actin
Cacna1a
42 kD
+/+ S218L+Neo/+ S218L+Neo/S218L+Neo +/+S218L/+S218L/S218L S218L{Neo/S218L+Neo
+/+ S218L+Neo/+ +/+
S218L/+
S218L/S218L
e f
S218L KI Wild-type
200 �m
400 �m 400 �m
d c
200 �m P
M G w
PM G w
CA
DG Ctx
CA DG
Ctx
g h
Figure 1. Generation of S218L knockin (KI) mice. (A) Genomic structure of the wild-type Cacna1a allele, targeting vector and predicted structure after homologous recombination (S218L+Neo allele), and after Cre-mediated deletion of the neomycin cassette (S218L allele). LoxP sites are indicated by triangles. Black numbered boxes indicate respective exons, with the S218L mutation in exon 5. Thick black lines indicate probes for Southern blot analysis. Primers used for genotyping and confirmation of the S218L mutation are depicted schematically by a arrows. Restriction sites: E, Eam1105I; EI, EcoRI; H, HindIII, K, KpnI; X, XbaI.
(B) Southern blot for all genotypes of S218L+Neo and S218L allele carriers showing genomic DNA digested with EcoRI- and Eam1105I and tested for the 5’- and 3’-probe. (C) Northern blot of mRNA isolated from adult brain of S218L genotypes hybridized with Cacna1a and Cyclophilin cDNA probes. (D) Semi-quantitative Western blot of cerebellar membrane protein extracts of S218L genotypes strain tested with Cav2.1 and actin antibody. Equal levels of Cav2.1-α1 RNA and protein are present in wild-type, S218L/+ and S218L/S218L mice. Cav2.1-α1 protein immunostaining in wild-type (E) and S218L/218L (G) cerebellum. Immunostaining of Cav2.1-α1 protein of coronal section showing a relatively high expression in the hippocampus and low overall staining in the cortical regions of wild-type (F) and S218L/S218L (H) brain. No obvious differences in expression level and pattern of Cav2.1-α1 protein was observed between both genotypes. Scale bars are depicted. CA, hippocampal cornu ammonis; Ctx, cortex; DG: dentate gyrus; G, granule cell layer; M, molecular layer; P, Purkinje cell layer; W, white matter.
voltage value. CSD duration was measured at halfpeak amplitude.
Statistical analysis
Unless stated otherwise, possible statistical differences were analyzed with paired or unpaired Student’s t-tests or analysis of variance (ANOVA) with Tukey’s HSD post-hoc test, where appropriate, on grand mean values. p<0.05 was considered to be statistically significant. All data is presented as mean ± SEM.
Results
Generation and basic characterization of S218L knockin mice
KI mice were generated by introducing human FHM1 mutation S218L into the orthologous Cacna1a gene. The S218L KI allele was obtained after in vivo deletion of the neomycin cassette by crossing mice carrying the S218L+Neo allele with Cre- deleter mice (Fig. 1A).16 Southern blot analyses confirmed correct recombination events before and after deletion of the neomycin cassette in all three genotypes (Fig. 1B). The presence of mutation S218L was confirmed by direct sequencing on DNA and cDNA (data not shown). Levels of total RNA and protein level for S218L Cav2.1-α1 subunit are similar to wild-type (Fig. 1C,D). Immunohistochemistry showed a normal cerebellar expression pattern for Cav2.1-α1 protein in homozygous S218L KI and wild-type mice, with a high expression in the molecular cell layer and in Purkinje cells (Fig. 1E,F).
Representative coronal sections show a relatively high expression of Cav2.1 α1 protein in the hippocampal areas and a low overall staining in, for instance, the cortical regions (Fig. 1G,H).
Genotyping by PCR analyses (data not shown) of offspring from heterozygous intercrosses (total n = 384) revealed 19.25% homozygous S218L mice which is a small, but significant, deviation from Mendelian law (p < 0.035), indicating a slight decreased survival before weaning. Long-term follow-up using Kaplan-Meier curve analysis revealed significantly decreased survival, only of S218L/S218L mice (log rank test; p <
0.001) (Fig. 2A). Daily inspections of S218L mice indicated that death occurred during the night with no indications of prior disease. Post-mortem examinations did not indicate heart, liver or kidney pathology, but pulmonary edema, suggesting cardiac failure (data not shown).
Increased mortality in homozygous S218L mice related to seizures
Seizure-activity is a known cause of sudden death in humans26 and animal models.27,28 Seizures can be provoked by severe hemiplegic migraine attacks in FHM or have been shown to co-occur in FHM1 patients, without severe clinical consequences.29 Although seizures are a predominant phenotype in several natural Cacna1a mouse mutants it never results in such severe clinical outcome.30 However, a S218L patient that experienced mild head trauma and collapsed after a lucid interval had a generalized seizure, went into a deep coma and died shortly thereafter.11 To assess whether S218L mice were more prone to develop seizures we performed long-term EEG recordings and synchronous video observation (Fig. 2B,C). For each S218L/S218L mouse we were able to make pre-
300
a
100
+/+ (n = 140) S218L/+ (n = 240)
S218L/S218L (n = 87)
10 20 40 60 80
200 100Postnatal days
Percentage Alive
0
b
1
2
3
4
5
6
7
0.4 m V
4 s M
C
T I
I
II
III
IV
c
10 s
0.2 mV
Figure 2. Increased mortality and seizure susceptibility in homozygous S218L mice. (A) Survival of S218L mice. Kaplan Mayer curve of S218L/S218L (n = 87) (solid black line), S218L/+ (n = 240) (dashed line) and wild-type (n =190) (solid gray line) mice showed significantly decreased survival of only S218L/S218L mice (p < 0.001). Censored observations are shown by vertical bars. (B) Example of an EEG recording of a ‘grand mal’ seizure in a S218L/S218L mouse. Different chronological phases (1-7) are depicted: 1) normal EEG abnormalities prior to seizure; 2) Attacks of severe generalized seizures started with myoclonic jerks (M), coinciding with EEG spike discharges and polyspikes; 3) Myoclonic jerks gradually developed into generalized clonic (C) and tonic seizures; 4-6) Behaviors observed during the tonic phase (T) included exaggerated flexion or extension of the trunk, falling (with inability to regain the upright posture), clonic movements of the limbs, or tonic extension or retraction of the limbs and were accompanied by spike wave discharges of up to 10-15 Hz; 7) Generalized seizures were followed by a period of slow wave rhythmic activity during which the mouse was lethargic, except for occasional shivering coinciding with polyspikes on the EEG that can result in the death of the mouse. (C) Example of an EEG recording of a ‘petit mal’ (absence) seizure in a S218L/S218L mouse. Traces I-IV correspond to simultaneous recording from electrodes in right medial, right frontal, left medial, and left frontal bones, respectively.
seizure recordings that did not reveal EEG abnormalities (Fig. 2B1). All three S218L/
S218L mice experienced spontaneous severe generalized seizures; 1, 2 and 3 seizures over a period of 4, 18 and 48 hours, respectively. Attacks had a similar sequence of events and all started with myoclonic jerks (several per minute), coinciding with EEG spike discharges and polyspikes (1-5 sec duration, Fig. 2B2). Myoclonic jerks gradually developed into generalized clonic and tonic seizures that prolonged, with relapses, up to 50 min. The behavior observed during the tonic phase included exaggerated flexion or extension of the trunk, falling (with inability to regain the upright posture), clonic movements of the limbs, or tonic extension or retraction of the limbs and were accompanied by spike wave discharges of up to 10-15Hz in all four electrodes (Fig.
2B3-B6). Generalized seizures were followed by a longer post-ictal period (~15 min.) of slow wave rhythmic activity during which the mouse was responsive to pain stimuli but otherwise remained inactive, except for occasional shivering registered as polyspikes on the EEG (one per 2-3 minutes, Fig. 2B7) that can result in the death of the mouse, normally associated with respiratory insufficiency. In each animal the last seizure was lethal. Neither such attacks nor lethality was observed in S218L/+ mice. In addition to these ‘grand mal’ seizures, the two longest surviving S218L/S218L mice showed a few rhythmic spike wave discharges (1-3Hz) (Fig. 2C), which coincided with immobility of the animals lasting up to 3 min, resembling absence seizures. There seems no specific sequence of ‘petit-mal’ and ‘grand-mal’ seizures as severe seizures were only in one case preceded by a small seizure. Only in 1 out of 3 heterozygous animals we observed EEG patterns of very short (15-30 s) bilateral rhythmic sharp activity (1-3Hz) that occurred in clusters (not shown), which may be related to the relatively short time allowed for recovery after the surgery.
Traumatic brain injury causes seizures and mortality in homozygous S218L mice Mild head trauma causes severe attacks in S218L patients that can include generalized seizures. To determine whether mild head injury can cause similar attacks in S218L mice we subjected these mice to moderate brain trauma using a controlled cortical impact (CCI) device.19 Wild-type mice did not show any apparent behavioral disturbances or neurological deficits during the first 24 hours following experimental traumatic brain injury, a finding corroborated by our previous findings in other mouse strains.19 Three out of ten S218L/+ mice, however, displayed 1-2 fast tonic-clonic sequences and loss of the righting reflex followed by loss of consciousness for up to one hour following non-fatal brain trauma. All animals regained consciousness and survived the observation period (24 h) without any apparent neurological deficits. Seizure activity and mortality were dramatically enhanced in S218L/S218L knockin mice. All investigated homozygous
animals (n = 5) had multiple ‘grand-mal’ seizures as described above (total n of seizures
= 19). Four out of five homozygous animals died during such an attack between one and 20 hours after trauma.
Homozygous S218L mice are ataxic
In addition to attacks, S218L patients also display permanent cerebellar ataxia. Frequent falling of S218L/S218L mice when standing on their hind paws suggested motor function defects. Whereas S218L/S218L mice performed normal in inverted screen and grip strength tests (both being tests for muscle strength), they showed specific deficits in the horizontal bar test. All wild-type animals passed the test, whereas only 37.5% of the S218L/S218L mice did (p < 0.05). As S218L/S218L mice did not fall more often than wild-type mice (p = 0.680), but were to slow to pass the test in time, the underlying
a
b c
�
��
���
���
���
���
� � � � � �
�����
Time on rotarod (s) S218L/S218L
S218L/+
+/+
Figure 3. Ataxia in homozygous S218L mice. Motor coordination and Purkinje cell abnormalities. (A) Rotarod test in wild-type (n = 11), S218L/+ (n = 11) and S218L/S218L (n = 10) mice reveal ataxia as exemplified by deficit in maintaining on the accelerating rod. Golgi impregnation of individual cerebellar Purkinje show dendritic abnormalities in adult S218L/S218L mice (B), whereas branching in wild-type Purkinje cells is very intense with increasing number of branching points near the distal tips (C). Purkinje cells in S218L/S218L mice show decreased branching and down-turned ends of their dendritic tree (arrows).
problem suggests motor coordination and not lack of muscle strength. In rotarod tests, S218L/S218L mice showed clear deficits at higher velocities indicating that these mice are ataxic (Fig. 3A). Heterozygous S218L mice do not show abnormalities in any of these tests.
Histological analysis did not reveal gross structural cerebellar abnormalities (results not shown). However, Golgi-Cox staining in 2-month old and 1 year-old homozygous S218L mice revealed significantly reduced dendritic arborization of Purkinje cells (Fig. 3B,C). Similar Purkinje cell dendritic abnormalities have been implicated as the underlying cause of ataxia in natural Cacna1a mutant rocker.31
Allele-dosage effects of the S218L mutation on neuronal Ca2+ current.
To investigate the consequences of the S218L mutation on neuronal Cav2.1 channels we measured the Cav2.1 current density as a function of voltage in cerebellar granule cells in primary cultures of S218L mice (Fig. 4). To isolate the component of whole-cell Ca2+
current that is due to Cav2.1 channels, ω-conotoxin MVIIC (3 μM, inhibits both Cav2.2 and Cav2.1, respectively) was applied after subsequent additions of saturating concentrations of nimodipine (5 μM; Cav1 blocker) and ω-CgTx-GVIA (1μM; Cav2.2 blocker) (Fig.
4A).8 This protocol enabled measurement of the L-, N- and R-type components of the Ca2+ current in addition to the P/Q-type component. The Cav2.1 current density in S218L neurons was larger than in wild-type neurons over a broad voltage range (Fig.
4B). The relative increase in S218L current density was larger at lower voltages close to the threshold of channel activation. The voltage dependence of the relative increase in current density is consistent with mutant Cav2.1 channels being activated at more negative voltages than wild-type channels. Fitting of the current-voltage relationship curves gave half-voltage of activation values (V1/2) of -25.1 ± 0.3 mV, -20.4 ± 0.4 mV and -15.6 ± 0.6 mV for the Cav2.1 current of S218L/S218L, S218L/+, and wild-type, respectively. The shift to lower voltages of Cav2.1 channel activation and the gain-of- function of the Cav2.1 current were about twice as large in homozygous as compared to heterozygous S218L mice, revealing an allele-dosage effect consistent with dominance of the mutation in FHM1 patients.
At voltages higher than 0 mV, where Cav2.1 channels have maximal open probability (po,max), the whole-cell Cav2.1 current density (I) in neurons of S218L mice was similar to that in wild-type neurons. Because mutant and wild-type have identical single channel current, i, this implies similar densities of functional Cav2.1 channels (N) (given that I
= N i po).15 The densities of the L-, N- and R-type components of the Ca2+ current were not significantly different in neurons of S218L and wild-type mice (with the exception of a small decrease in the L-type component in S218L/S218L mice) (data not shown). At
-10 mV, the current densities for the L-type component were: 10.5 ± 0.7 (n = 24), 12.7
± 0.7 (n = 27) and 13.8 ± 1.1 (n = 29) pA/pF in S218/S218L, S218L/+ and wt neurons, respectively; for the N-type: 8.0 ± 0.7 (n=27), 8.8 ± 0.8 (n = 29) and 6.5 ± 0.8 (n = 29) pA/pF; for the R-type: 23.0 ± 1.6 (n=28), 20.4 ± 1.0 (n = 43) and 22.9 ± 1.1 (n = 35) pA/
pF. The sum of the three components (total non-P/Q current density) was similar for all three genotypes (41.5, 41.9 and 43.2 pA/pF). Therefore, the gain-of-function of Cav2.1 channels in neurons of S218L mice was not compensated by alterations in other types of Ca2+ channels.
a
b
2 4 6 8 10 12 14
0 200 400 600 800 1000
c-d
d c
d cb a
Current (pA)
d c
b a
I-V I-V
Time (min)
MVIIC GVIA
Nimodipine
-30 -25 -20 -15 -10 -5
-40 -20 0 20
+/+S218L/+
S218L/S218L
V (mV)
I (pA/pF)
-30 mV -20 mV
-10 mV
-30 mV -20 mV -10 mV
Figure 4. S218L mutation increases the Cav2.1 current density in cerebellar granule cells of S218L mice. (A) Peak whole-cell Ba2+ current versus time recorded from a cerebellar granule cell in primary culture from a heterozygote S218L/+ mouse during depolarizations at -10 mV every 10 s before and after application of the indicated drugs (nimodipine: 5 uM, GVIA: 1 µM, MVIIC 3 µM). Vh= -80 mV; Ba2+ = 5mM. Inset:
representative current traces taken at times a, b, c, d and difference trace representing the P/Q-type component (c-d). On the right: representative traces at increasing voltage from -50 to -10 mV, taken during I-V measured at times c and d. Scale bars:
20 ms, 200 pA. (B) Cav2.1 current density (I) as a function of voltage in cerebellar granule cells of wild-type, S218L/+ and S218L/S218L mice. Average normalized I- V curves (n = 14 for wild-type, n = 12 for S218L/+ and n = 10 for S218L/S218L mice) were multiplied by the average maximal current density ((n = 35 for wild-type, n = 44 for S218L/+ and n = 28 for S218L/S218L mice). Solid lines are fits by equation: I = G (V-Erev) (1+ exp((V1/2-V)/k))-1 with V1/2 = -15.6 mV for wild-type, V1/2 = -20.4 mV for S218L/+ and V1/2 = -25.1 mV for S218L/
S218L mice. Inset: pooled wild-type and S218L/S218L (top) or S218L/+ (bottom) Cav2.1 current traces at -30, -20 and -10 mV. Scale bars: 20 ms, 5 pA/pF.
Increased acetylcholine release spontaneous transmitter release at S218L neuromuscular junctions
To investigate synaptic effects of the S218L mutation we studied neuromuscular junctions (NMJs) in diaphragm/phrenic nerve preparations, where ACh release is almost exclusively mediated by Cav2.1 channels.32 Spontaneous uniquantal ACh release,
measured as MEPP frequency, was increased ~12-fold at S218L/S218L NMJs (1.24 ± 0.06 s-1 at wild-type and 15.50 ± 1.18 s-1 at S218L/S218L NMJs, n = 6 muscles, p<0.001;
Fig. 5A). S218L/+ NMJs showed an intermediate MEPP frequency of 10.56 ± 1.17 s-1 (n = 6 muscles, p<0.01). The MEPP amplitude was ~1 mV under all conditions and did not differ between genotypes (n=6 muscles, p=0.17), neither did MEPP rise times, decay times and half-width values (data not shown). Specific Cav2.1 channel blocking toxin ω-agatoxin-IVA (200 nM) reduced MEPP frequency by ~60% at wild-type and by
>90% at both S218L/+ and S218L/S218L NMJs (n = 3 muscles, p<0.05; Fig. 5A). Upon phrenic nerve stimulation at 0.3 Hz, the EPP amplitude was ~28 mV for all genotypes (n = 6 muscles, p=0.28). The quantal content was ~40 in all groups and did not differ
Figure 5. Increased ACh release and broadened endplate potentials at S218L NMJs. (A) MEPP frequency was increased ~10-fold at heterozygous and ~15-fold at homozygous S218L NMJs, when measured in 2 mM Ca2+ (n = 6 muscles). The selective Cav2.1 channel blocking toxin ω-agatoxin-IVA reduced MEPP frequency by ~60% at wild-type and by >90% at both heterozygous and homozygous S218L NMJs (n = 3 muscles). MEPP amplitude, rise- and decay times did not differ, as indicated by single representative MEPP traces. Representative 1 s traces of MEPP recordings clearly show the increase in MEPP frequency at S218L NMJs. (B) Quantal contents at 0.2 mM Ca2+ were increased by ~95% in S218L/+ (although not statistically significantly, p = 0.28) and 155% in homozygous S218L mice, compared to wild-type (n = 3 muscles).
Representative EPP traces are shown. Triangles indicate moment of nerve stimulation. (C) EPP half-width was increased at homozygous S218L NMJs, not however at S218L/+ NMJs (p = 0.83), compared to wild- type (n = 6-9 muscles). (D) EPP half-width at S218L synapses was sensitive to application of 200 ng/ml of the selective K+ channel blocker 3,4-diaminopyridine (DAP) (~100% increase, n = 5 muscles), whereas at wild-type synapses it did not change significantly (n = 3 muscles, p = 0.27). Representative EPP recordings obtained before application and in the presence of DAP are shown. Triangles indicate moment of nerve stimulation. * p<0.05, † p<0.01, ‡ p<0.001.
between genotypes (38.0 ± 2.5, 40.5 ± 2.6 and 39.8 ± 3.7 at wild-type, S218L/+ and S218L/S218L NMJs, respectively; n = 6 muscles, p=0.82). Quantal content was reduced by >95% by 200 nM ω-agatoxin-IVA in all genotypes (n = 3 muscles, P < 0.05; data not shown). Quantal content remained unaffected by application of 2.5 μM ω-conotoxin- GVIA (Cav2.2 blocker) and 1 μM SNX-482 (Cav2.3 blocker), suggesting absence of compensatory contribution of non-Cav2.1 channels to ACh release at S218L/S218L NMJs.
At physiological extracellular Ca2+ concentration, Ca2+ sensors of the neurotransmitter release machinery may approach saturation during action potential-induced Ca2+ influx.33 Therefore, neurotransmitter release measured at lower extracellular Ca2+ concentration may better reflect the behavior of (mutant) Cav2.1 channels. We, therefore, measured ACh release at 0.2 mM Ca2+ and found an allele-dosage effect on 0.3 Hz evoked EPP amplitude and quantal content. EPP amplitudes were increased ~2-fold and ~3-fold at S218L/+ and S218L/S218L NMJs, respectively, compared to wild-type (Fig. 5B).
Quantal contents were 7.2 ± 1.5, 13.9 ± 0.7 and 18.3 ± 1.7 at wild-type, S218L/+ and S218L/S218L NMJs, respectively (n = 3 muscles, p<0.01; Fig. 5B).
At S218L/S218L NMJs, we occasionally encountered broadened EPPs. Half-width of the first EPP measured at 0.3 Hz stimulation was increased by ~45% (p<0.05) at S218L/S218L NMJs (3.14 ± 0.15, 3.48 ± 0.28 and 4.54 ± 0.48 ms at wild-type, S218L/+
and S218L/S218L NMJs, respectively; n = 6-11 muscles). Broadening of EPPs may be caused by prolonged Cav2.1 channel opening, resulting in longer lasting ACh release.
3,4-Diaminopyridine (DAP) blocks presynaptic K+ channels causing broadening of the presynaptic nerve impulse, resulting in prolonged opening of Cav2.1 channels.34 We hypothesized that S218L EPP half-width should be more sensitive to application of DAP than wild-type. Indeed, 200 ng/ml DAP resulted in an almost doubling of EPP half- widths (from 5.54 ± 0.78 to 10.97 ± 0.96 ms, n = 5 muscles, p<0.05; Fig. 5D), whereas wild-type NMJs were not affected (n = 3 muscles, p=0.27; Fig. 5D).
Increased susceptibility of S218L knockin mice to Cortical Spreading Depression in vivo.
We investigated the consequences of the S218L mutation on in vivo threshold for initiation, rate of propagation, and duration of CSD (Fig. 6). We found that S218L KI mice were more susceptible to CSD induction than wild-types (Fig. 6A). More specifically, significantly less charge was needed to elicit CSD in S218L/S218L and S218L/+ animals than in wild-type mice. Interestingly, we observed an allele-dosage effect of the S218L mutation on CSD threshold as the charge needed to elicit CSD in S218L/S218L mice was about half (significantly lower than) the value for S218L/+ animals (Fig. 6A). Similar allele- dosage effects were found when propagation velocity of CSD was analyzed (Fig. 6B).
CSD velocity significantly increased from 3.2 ± 0.19 mm/min in wild-type to 7.5 ± 0.96
b
29.4%
70.6%
64%
36%
65.4%
34.6%
S218L/S218L S218L/+
+/+
recurrent CSD single CSD
c
0 5 10 18 20
Propagationvelocity(mm/min)
S218L/S218L S218L/+
+/+
*
*
*
*
a
10 20 30 40 50
CSD threshold(microC)
0
S218L/S218L S218L/+
+/+
*
*
*
1 10
0 20 40 60 80 100
CSD threshold (microC)
% of animals S218L/+
+/+
S218L/S218L
Figure 6. S218L mutation facilitates the induction and the propagation of CSD in S218L mice. (A) Wild-type and S218L mutants have different thresholds for CSD induction (one-way ANOVA, p<0.001). S218L/S218L were more sensitive to CSD induction than wt and S218L/+ mice as significantly less charge was needed to elicit CSD (S218L/S218L: 6.9 ± 0.8 microC, n = 18; S218L/+ 11.9 ± 1.2 microC, n = 32; wt 18.5 ± 2.0 microC, n = 26; post-hoc Tukey test p<0.05). Heterozygous mice had also a lower threshold than wt (post-hoc Tukey test p<0.05). Asterisks indicate a significant difference. Right: cumulative distribution of CSD threshold in homozygous, heterozygous S218L mutant and wt mice. The plot displays the fraction of animals showing a CSD after electrical stimulation of intensity minor or equal to the corresponding value on the abscissa.
The distribution of mutants is shifted to the left, indicating that CSD was more easily inducible in S218L/
S218L and in S218L/+ than in wt (Kolmogorov-Smirnov test, p<0.01). The distributions of S218L/+ and S218L/S218L are also significantly different (Kolmogorov-Smirnov test, p<0.01). (B) The S218L mutation resulted in an increase of CSD velocity from 3.2 ± 0.19 mm/min in wt to 7.5 ± 0.96 mm/min in S218L/S218L animals (one way ANOVA p<0.001; post-hoc Tukey test p<0.05). The CSD velocity measured in S218L/+
was higher then in wt animals (4.7 ± 0.29 mm/min in S218L/+ vs. 3.2 ± 0.19 mm/min in wt; post-hoc Tukey test p<0.05). Furthermore the velocity of CSD propagation was significantly higher in homozygous than S218L/+ in animals (4.7 ± 0, 29 mm/min in S218L/+ vs. 7.5 ± 0.96 mm/min in S218L/S218L; post-hoc Tukey test p<0.05). Asterisks indicate a significant difference. (C) S218L/S218L showed an increased frequency of recurrent CSDs occurring after the first electrically induced CSD. In wt group recurrent CSDs (dark grey) were recorded in 34.6% of the animals and similarly, in S218L/+, the incidence was 36%. However in S218L/
S218L the incidence rose significantly up to 70.6% if compared with wt and S218L/+ values (Fisher exact test p=0.031).
mm/min in S218L/S218L mice. S218L/+ mice showed an intermediate CSD propagation velocity of 4.7 ± 0.29 mm/min. In some cases more than one CSD was elicited by the electrical stimulus. Noteworthy, S218L mice showed a significantly higher frequency of induction of recurrent CSDs than wild-type and S218L/+ mice (Fig. 6C).
Other parameters of CSD were not altered by the S218L mutation. In particular, no difference was noted in CSD amplitude (first recording site: wild-type: 20.1 ± 1.5 mV, S218L/+: 17.8 ± 1.5 mV, S218L/S218L: 16.5 ± 1.6 mV, one way ANOVA p=0.66) and duration (wild-type: 42.3 ± 3.5 s S218L/+: 51.6 ± 8.1 s, S218L/S218L 40.6 ± 6.1 s, one way ANOVA p=0.33). Similar results have been detected in CSD recorded at the second site (amplitude: wild-type: 22.5 ± 1.8 mV, S218L/+: 18.3 ± 1.6 mV: S218L/S218L: 21.6
± 2.0 mV, one way ANOVA p=0.51; duration: wild-type: 42.4 ± 5.6 sec, S218L/+: 53.7
± 6.8 sec, S218L/S218L: 42.6 ± 4.9 sec, one way ANOVA p=0.39).
Discussion
Here we describe the generation of knockin mice with the human FHM1 S218L mutation introduced into the Cacna1a gene. Patients carrying mutation S218L suffer from a combination of clinical symptoms such as permanent cerebellar ataxia, typical attacks of hemiplegic migraine triggered by mild head trauma.12 Sometimes, after a lucid interval, attacks involve generalized seizures and result in deep coma that can be fatal. Its social impact is particularly dramatic since complications of mild head trauma occur most often in children.12,35 It is unknown how mild head trauma events can result in such devastating clinical outcome.
The identification of mutation FHM1 S218L in this severe disease made the Cav2.1 channel a direct target for trauma-related mechanisms. S218L-mutated Cav2.1 channels had unique characteristics compared to other FHM1 mutant channels, namely a large negative shift in activation of single channels, a large increase in whole-cell current density, an increased rate of recovery from inactivation, and a large component of slowly inactivating current, for instance at conditions of long depolarizations as occur during CSD.15 We hypothesized that the unique properties of S218L channels may cause a longer duration of CSD and the particularly severe clinical phenotype. Therefore, we investigated the functional consequences of S218L Cav2.1 channels on neuronal Ca2+
current, neurotransmission, and CSD in S218L mice.
Seizure-related mortality in S218L FHM1 mice
Homozygous S218L mice exhibit a phenotype that closely resembles the phenotype in S218L patients. These mice show increased mortality that could not be related to a
‘standard trigger’ (e.g. handling, cleaning cage, dropping, restraining in small space).
Deaths seemingly occurred at random at all ages. Of note, a considerable number of mice became older than 1 year. EEG recordings revealed that homozygous S218L mice had at least one severe, spontaneous, ‘grand mal’ seizure. All mice died within 48 hours of a seizure. In addition to ‘grand mals’ also absence seizures could be noticed. A controlled cortical impact device for induction of moderate brain trauma caused ‘grand mals’ that resulted in the death of almost all homozygous S218L mice within 24 hours. Importantly, in both experiments, wild-type mice were without any neurological problems. It can be concluded that mild head injury causes brain trauma, seizures and ultimately the death of these animals. Likely, the many spontaneous deaths also are related to seizure events precipitated by mild head trauma or other mild triggers.
Hyperexcitability and CSD
Functional consequences of mutation S218L for neuronal Ca2+ current, neurotransmission, and CSD show clear gain-of-function effects in an allele-dosage manner. Firstly, Ca2+ influx through both single Cav2.1 channels and the whole soma is increased as a consequence of mutant channels that open more readily and at lower voltages than wild-type channels.15 Also, in granule cells from S218L mice there is an increased Ca2+
influx, especially at lower voltages close to the threshold of channel activation and close to the resting potential of many neurons. Unlike in transfected neurons, in neurons from homozygous and heterozygous S218L mice the number of functional channels is equal to that of the wild-type. Secondly, neurotransmission at the NMJ is increased in conditions where saturation of the synaptic Ca2+ sensor was not reached and the release probability was low (evoked release at low [Ca2+]and spontaneous release). A significant broadening of EPPs reflects a prolonged opening of S218L Cav2.1 channels and further increase in neurotransmitter release. Last but not least, susceptibility to CSD is increased, as indicated by the lowered threshold for induction, the increased velocity of propagation, and a significant higher frequency of recurrent CSDs in homozygous S218L mice. Our findings indicate that neuronal hyperexcitability and CSD are important in migraine and support clinical findings of altered brain excitability in migraine patients.36 In addition, neuronal hyperexcitabilty and CSD may also be important in the pathogenesis of lethal brain edema upon mild head trauma in children.12
R192Q versus S218L KI mice
Results of functional consequences of FHM1 mutation S218L mutation in our KI mice are in perfect agreement with our previous findings in R192Q KI mice that also produced clear gain-of-function effects on Ca2+ influx, neurotransmission and susceptibility to
CSD.8 An important difference is that the consequences in S218L mice are much more severe, which is in line with the more severe phenotype in S218L patients. Moreover, for most parameters functional consequences were already present in heterozygous mice.
At low voltages close to the threshold of activation of mutant channels, the gain-of- function in homozygous S218L mice tended to be larger than in homozygous R192Q mice: i.e. in S218L mice the Cav2.1 current at -40, -30 mV (corresponding to -52, -42 mV after junction potential correction) was 6-7 times larger than in wild-type whereas in R192Q mice it was 4 times larger.8 Also more drastic effects were observed at the neuromuscular junction. Spontaneous release was 12-fold increased in homozygous S218L mice, while in R192Q mice the increase was only 2-fold.8 Interestingly, evoked ACh release (0.3 Hz), measured in quantal content under conditions of low Ca2+ was not different between homozygous S218L and R192Q mice (18.3 ± 1.7 vs. 17.1 ± 4.1, respectively). A specific feature of homozygous S218L mice are broadened EPPs that were frequently observed at 4.5 months of age. EPP broadening is a feature of inhibition or absence of acetylcholinesterase at the NMJ.37,38 If this were the case, due to some indirect effect of the mutation, the halfwidth of uniquantal responses (MEPPs) should, however, also have been found increased. Since this was not the case, the most likely explanation for EPP broadening is prolonged openening of S218L-mutated Cav2.1 channels upon activation by a nerve action potential. EPP halfwidth at S218L NMJs of 2 months-old mice being more potentiated by DAP than wild-types supports this hypothesis.
Also facilitation of CSD was larger in S218L than R192Q KI mice. Particularly striking is the larger increase in the rate of propagation of CSD, that is similar between S218L/+ (4.7 mm/min) and homozygous R192Q mice (4.6 mm/min).8 The CSD threshold shows a similar trend, but without statistically significant differences between either heterozygous S218L (11.9 ± 1.2 μC) and homozygous R192Q KI ( 7.9 ± 2 μC) mice or between homozygous S218L (6.9 ± 0.8 μC) and R192Q KI mice.8
How does the S218L FHM1 mutation cause ataxia?
In addition to the attacks, homozygous mice show permanent cerebellar ataxia. Notably, cerebellar ataxia is also a prominent clinical feature in S218L patients. Although we did not see gross cytoarchitectural abnormalities in the cerebellum of homozygous S218L mice, we could demonstrate clear abnormalities in branching of the dendritic tree of Purkinje cells. The abnormalities were already present at 2 months of age and were similar in mice of 1 year, indicating that disturbed Purkinje cell development rather than degeneration of the dendritic tree with age is the cause. It has been shown that formation of ‘weeping willows’ in Purkinje cells could be introduced by destruction of parallel fibers that originate from cerebellar granule cells at the third postnatal
