www.elsevier.com/locate/brainres

Brain Research 1013 (2004) 188–193

Research report

GABAA receptor modulation of trigeminovascular nociceptive

neurotransmission by midazolam is antagonized by flumazenil

Robin James Storer, Simon Akerman, Kevin G. Shields, Peter J. Goadsby*

Headache Group, Institute of Neurology and The National Hospital for Neurology and Neurosurgery, Queen Square, London WC1N 3BG, UK

Accepted 9 March 2004

Abstract

Studies of the pharmacology of trigeminocervical neurons with input from intracranial pain-producing structures have enhanced the

understanding of the basic neurobiology of primary headache, such as migraine. Clinical observations of the treatment of migraine with

medicines acting at the g-aminobutyric acid (GABA) GABAA receptor have lead to studies of their effects on models of trigeminovascular

nociception. Extracellular recordings were made from neurons in the trigeminocervical complex activated by supramaximal electrical

stimulation of superior sagittal sinus (SSS) in the cat. Intravenous administration of the benzodiazepine receptor agonist midazolam, resulted

in a dose-dependent inhibition of superior sagittal sinus evoked trigeminocervical nucleus activity. The inhibition at 50 Ag/kg midazolam was

65F 11% compared to the baseline response (n = 11). Intravenous administration of the benzodiazepine receptor antagonist flumazenil,

resulted in a dose-dependent recovery of superior sagittal sinus evoked trigeminocervical nucleus activity. At a dose of 50 Ag/kg, there was a64F 5% recovery (n = 6). The data demonstrate a potent, reproducible effect of facilitation of GABA transmission at the GABAA receptor

that results in inhibition of trigeminovascular nociceptive transmission. These data are consistent with the useful clinical effects reported with

compounds that can augment GABAergic transmission in the central nervous system (CNS).

D 2004 Elsevier B.V. All rights reserved.

Theme: Sensory systems

Topic: Pain modulation: pharmacology

Keywords: Trigeminal; Dural; Migraine; Headache

1. Introduction

Migraine is an episodic brain disorder that results in

significant morbidity [34] for between 10% and 15% of the

general population [27,38]. Modulation of trigeminal trans-

mission in order to alleviate acute migraine requires an

understanding of the pharmacology of the second order

trigeminal neurons subserving nociception [14]. In recent

years, the pharmacology of this synapse has been studied

with most attention on serotonergic and glutamatergic

transmission [13].

g-Aminobutyric acid (GABA) is well known as an inhib-

itory amino acid neurotransmitter in the central nervous

system (CNS) [39] and may modulate nociceptive response

in the spinal cord [40]. Understanding the role of GABAergic

influences on trigeminovascular nociceptive transmission

0006-8993/$ - see front matter D 2004 Elsevier B.V. All rights reserved.

doi:10.1016/j.brainres.2004.03.068

* Corresponding author. Tel.: +44-207-829-8749; fax: +44-207-813-

0349.

E-mail address: peterg@ion.ucl.ac.uk (P.J. Goadsby).

within the trigeminal nucleus is important to develop a

comprehensive description of these neurons. It has been

shown that trigeminovascular nociceptive input can be mod-

ulated through GABAergic mechanisms [46]. GABA, when

microiontophoresed onto neurons in the trigeminocervical

complex responding to dural perivascular input, specifically

stimulation of the superior sagittal sinus, inhibits trigeminal

transmission. This effect can be blocked by bicuculline, a

GABAA receptor antagonist, and mimicked by the GABAA

receptor agonist muscimol. In the same series of experiments,

the GABAB receptor agonist baclofen affected only a small

number of neurons with trigeminovascular input and the

GABAB receptor antagonist, 2-hydroxysaclofen, was much

less likely to inhibit trigeminal neurons [46]. More recently it

has been shown that topiramate may inhibit transmission in

the trigeminocervical complex [44], and this compound is

known to modulate GABAergic mechanisms, among other

effects [41]. Midazolam acts at the benzodiazepine modula-

tory site on the GABAA receptor complex to facilitate the

actions of GABA by prolonging channel opening time. We

R.J. Storer et al. / Brain Research 1013 (2004) 188–193 189

sought to determine the effect of benzodiazepine site activa-

tion on the trigeminovascular system in vivo by employing

the specific benzodiazepine site antagonist flumazenil to

reverse the inhibitory effects of midazolam [48].

2. Materials and methods

All studies were conducted and terminated under general

anaesthesia in accordance with a project licence issued by the

Home Office of the United Kingdom under the Animals

(Scientific Procedures) Act, 1986. Cats (n = 11) weighing

3.50F 0.53 kg (meanF S.D.) were anaesthetised with a-

chloralose (60 mg/kg i.p., Sigma, St. Louis, MO) and

prepared for physiological monitoring. Halothane (Rhone

Merieux, Essex, UK; 0.5–3% in a 40% oxygen/air mixture)

was administered from an anaesthetic machine during surgi-

cal procedures and then discontinued during experimental

protocols. Catheters were placed in femoral arteries for

arterial blood sampling and continuous measurement of

blood pressure (DTXplus transducer, Ohmeda, Madison,

WI; PM-1000 amplifier, CWE Instruments, Ardmore, PA).

Catheters were also placed in femoral veins allowing fluid

(saline for intravenous infusion BP (0.9% w/v), Baxter

Healthcare, Norfolk, UK) and drug administration. Supple-

mentary doses of a-chloralose in 2-hydroxypropyl-h-cyclo-dextrin (RBI, Natick, MA) were given i.v. at a rate of 5–10

mg/kg per h [45]. The cats were intubated after local

anaesthesia with lidocaine hydrochloride (Intubeaze,

Arnolds, Shrewsbury, UK) and fixed in a stereotaxic frame

(Kopf Instruments, Tujunga, CA).

Jackson/Foley urethral catheters were inserted to drain the

cat bladders, providing more even temperature regulation,

more stable control of blood pressure through control of

bladder distension, and monitoring of urine output. Core

temperature was monitored and maintained between 37 and

39 jC using a rectal thermistor probe and a low-radio-noise

emitting homeothermic heater blanket system (Harvard Ap-

paratus, Holliston, MA). Cats were ventilated with a 40%

oxygen/air mixture (Harvard Apparatus), end-tidal CO2 was

maintained between 2% and 4% and expired oxygen contin-

uously monitored (Datex-Ohmeda, Helsinki, Finland). Heart

rate was monitored by electrocardiogram (CT-1000; CWE

Instruments) or derived from blood pressure changes. The

depth of anaesthesia was monitored periodically throughout

the experiment by testing for sympathetic (pupillary and

cardiovascular) responses to noxious stimulation and with-

drawal reflexes in the absence of neuromuscular blockade.

2.1. Surgery

Midline craniotomies (approximately 20 mm diameter)

and C1/C2 laminectomies were performed allowing access to

the superior sagittal sinus and the area for recording neuronal

activity in the spinal cord. The superior sagittal sinus (SSS)

was isolated by dissecting the dura and falx cerebri adjacent

to them over approximately 15 mm. Small polyethylene

sheets were inserted under the isolated sinuses and laid over

the outlying dura mater then tucked under the edges of the

craniotomy to secure them. To prevent dehydration and to

provide electrical insulation to the cortex, a cylindrical

polypropylene dam was sealed to the bone around the

craniotomy with dental acrylic (Vertex, Zeist, Netherlands)

and filled with liquid paraffin (BDH Laboratory Supplies,

Poole, UK). Possible artefacts from arterial pulsation and

respiratory movement were reduced by bilateral pneumo-

thoraces, held patent with polypropylene tubes; immobiliza-

tion of the spine by clamping a thoracic spinous process to the

stereotaxic frame; clamping the C1 transverse processes to

auxiliary ear bar holders on the frame, and clamping the

remaining caudal portion of the dorsal C2 spinous process to

the frame.

2.2. Stimulation and recording

Isolated SSSs were gently lifted onto a pair of bipolar

platinum hook electrodes connected to a stimulus isolation

unit (SIU5A; Grass Instruments, West Warwick, RI). To

activate primary trigeminal afferents, the SSSs were supra-

maximally stimulated with stimulus-isolated (Grass SIU)

square wave pulses from a Grass S88 stimulator (250 As,110–150 V, 0.3–0.5 Hz) after neuromuscular blockade with

gallamine triethiodide (Concord, Essex, UK), initially 5–10

mg/kg i.v. and maintained with 5–10 mg/kg per h. The dura

mater above the recording regions on the surface of the spinal

cord was reflected after a midline incision and held to the

edges of the laminectomy with N-butyl-cyanoacrylate, fur-

ther stabilizing movement of the cord. Extracellular record-

ings were made using tungsten electrodes. Recording

electrode impedances were typically 0.4–2.1 MV when

measured at 1 kHz in 0.9% saline. After local removal of

the pia mater, the electrodes were lowered into the cord

substance caudal to the C2 roots in the area of the dorsal root

entry zone. The electrodes were advanced or retracted in the

cord substance in 5 Am steps using a microelectrode posi-

tioner consisting of a piezoelectric motor (IW-711, Burleigh

Instruments, Harpenden, UK) and ultra-low-noise controller

(6000ULN) attached to a micromanipulator (Kopf 1760-61).

Tissue culture grade agar (3% (w/v) in saline; Sigma) was set

over the exposed cord after electrode insertion to further

reduce cardiovascularly related movements. Signal from the

recording electrode attached to a high impedance headstage

preamplifier (NL100AK; Neurolog, Digitimer, Herts, UK)

was fed via an AC preamplifier (Neurolog NL104,

gain� 1000) through filters (Neurolog NL125; bandwidth

approximately 300 Hz to 20 kHz) and a 50 Hz noise

eliminator (Hum Bug, Quest Scientific, North Vancouver,

BC) to a second stage amplifier (Neurolog NL106) providing

variable gain (� 20– � 90). This signal (total gain approx-

imately � 20,000– � 90,000) was fed to a gated amplitude

discriminator (Neurolog NL201) and analogue-to-digital

converter (Cambridge Electronic Design, Cambridge, UK)

Fig. 1. Response of a wide dynamic range (WDR) unit to brush and pinch

stimuli on its cutaneous receptive field.

R.J. Storer et al. / Brain Research 1013 (2004) 188–193190

to a microprocessor-based (Pentium III, Intel, Santa Clara,

CA) computer where the signal was processed and stored.

Filtered and amplified action potentials were fed to a loud-

speaker via a power amplifier (Neurolog NL120) for audio

monitoring and displayed on oscilloscopes to assist the

isolation of single unit activity from adjacent cell activity

and noise. Data were collected as counts/bin in the 50 ms

following stimulation of the superior sagittal sinus.

In order to record the response of single units to

stimulation, post-stimulus histograms were constructed

on-line and saved to disk. During experiments electrophys-

iological data, including blood pressure, core temperature,

ECG, heart rate and end-tidal CO2 were processed and

recorded on VHS magnetic tape (Pulse Code Modulator;

Vetter, Rebersburgh, PA) for documentation and later

review.

The position of the recording electrodes was controlled

by use of a heavy-duty stereotaxic micropositioner (Kopf

1760-61) with reference to the mid-point of the C2 dorsal

roots. Together with the depth of the recording electrode tip

with respect to the surface of the spinal cord at the dorsal

root entry zone, as determined by the distance travelled

display on the ULN6000 pizoelectric motor controller

(Burleigh Instruments), this provided the coordinates of

the recording sites. The location of selected recording sites

were either microiontophoretically marked with Pontamine

Sky Blue dye (Gurr 6BX, BDH Laboratory Supplies; 2.5%

in 100 mM sodium acetate) using a � 2.00 AA current for

5–10 min or marked by thermocoagulation via an electro-

lytic lesion (20–50 AA, 10–30 s). Animals were euthan-

ised with sodium pentobarbital (400 mg), followed by 10%

KCl (5 ml). After termination of experiments, the sections

of spinal cord containing the recording sites were removed,

fixed with neutral buffered 10% formalin, and sectioned

(40 Am). Pontamine Sky Blue marks were counterstained

with neutral red, a Nissl procedure that allowed identifica-

tion of the laminae of the grey matter. Electrolytic lesion

marks were located after cresyl violet Nissl staining. The

position of the recording sites within the cord were

determined from histologically identified lesion or dye

marks and unmarked recording sites located by reference

to dye marks and the coordinates of recording electrode

positions.

2.3. Receptive fields

Cells responding to superior sagittal sinus (SSS) stimu-

lation were characterized as receiving low threshold mech-

anoreceptor (LTM) input if they responded non-noxious

input, such as brush or stroke on cutaneous receptive fields

on the face or forepaws. They were characterized as

nociceptive specific (NS) if they responded to noxious

mechanical stimuli, such as pinch or pricking with a needle,

or wide dynamic range (WDR) if they responded to both

[18]. These cells usually increased their firing in response to

noxious stimuli (Fig. 1).

2.4. Test compounds

Midazolam (8-chloro-6-(2-fluorophenyl)-1-methyl-4H-

imidazo[1,5][1,4]benzodiazepine; HypnovelR, Roche, Hert-fordshire, UK or Phoenix Pharma, Gloucester, UK) 2 mg/ml

was diluted in saline for injection (BP) and given i.v. at 10,

50 and 100 Ag/kg as a cumulative dose. Flumazenil

(ethyl-8-fluoro-5,6-dihydro-5-methyl-6-oxo-4H-imidazo

[1,5-a][1,4]benzodiazepine-3-carboxylate; AnexateR,Roche) 0.1 mg/ml was diluted in saline for injection

(BP) and administered i.v. at 10, 20, 50 and 100 Ag/kg as

a cumulative dose after midazolam to reverse its effects.

Not every dose was given to every animal when inhibi-

tion or reversal was marked at low doses.

2.5. Analysis and statistics

Data collected as counts/bin over 50 ms epochs were

analysed off-line. Units were considered linked to stimula-

tion of the SSS if their probability of firing exceeded 0.3

based on the Poisson distribution of count data [1], and the

critical ratio test [35]. A computer program was used to

highlight the firing latency of interest post-stimulus and

determine neuronal responses to 50 or 100 consecutive

sweeps. Three such baseline determinations were conducted

and then a cumulative dose-response for midazolam was

performed. Inhibition refers to the reduction in firing after

midazolam administration when compared to the pretreat-

ment baseline and recovery after flumazenil is again com-

pared to the pre-midazolam baseline. Physiological

parameters of the cohort are summarised as meanF S.D.

and responses to treatment are summarised as mean

F S.E.M. A logistic regression after log transformation of

the drug dose was used as this has excellent properties for

modelling ligand-receptor interactions (SPSS version 11,

R.J. Storer et al. / Brain Research 1013 (2004) 188–193 191

Chicago IL). The effect of midazolam was examined using

regression with dose versus percentage reduction in counts/

bin after superior sagittal sinus stimulation. Similarly, the

effect of flumazenil was tested using a regression of dose

with percentage recovery to the baseline response. P < 0.05

was considered significant.

Fig. 3. Dose-dependent inhibition of superior sagittal sinus evoked

trigeminocervical nucleus activity by intravenous midazolam. The ordinate

represents the percentage reduction in neuronal firing to stimulation of the

superior sagittal sinus with increasing doses of midazolam when compared

to the baseline pre-midazolam firing responses. Data are plotted as mean

with standard error bars.

3. Results

Animals from which data is reported had cardiorespira-

tory parameters that were normal for the anaesthetised cat.

Blood gas levels were measured at intervals throughout the

experiment and were within normal limits: arterial blood pH

7.37F 0.03 and pCO2 3.62F 0.28 kPa.

3.1. Localization and neuronal characteristics

Extracellular recordings were made and data collected

from 11 neurons in the trigeminocervical complex of cats

[20]. Cells were located 4 mm rostral to 4 mm caudal to the

midpoint of the C2 rootlets, 0–150 Am lateral to the dorsal

root entry zone at a depth of approximately � 1400 Am to

around � 3000 Am below the (dorsal) cord surface (Fig. 2).

Cells responded to electrical sagittal sinus stimulation with

latencies consistent with A-y fibres (fibre input/afferents;

typically 8–10 ms). Cells received wide dynamic range

(WDR) or nociceptive specific (NS) mechanoreceptor input

Fig. 2. Localization of recording sites. A transverse section through the

spinal cord at the level of C2 is represented. Dye or lesion marked recording

sites (.) were identified histologically. The positions of unmarked or

unrecovered sites were identified by reference to the surface of the cord or

the position of dye marks made in electrode tracts and electrode tip

coordinates at the recording site (o). Although the positions of the recorded

units are mapped to only one side of the cord in the figure, they represent

results obtained from both the left-hand side and right-hand side of the

spinal cord. The scale bar represents a distance of 1 mm in both directions.

from cutaneous VI or VII receptive fields on the face and

forepaws.

3.2. Benzodiazepine receptor agonist

Intravenous administration of midazolam resulted in a

dose-dependent inhibition of superior sagittal sinus evoked

trigeminocervical nucleus activity (Fig. 3). The inhibition at

50 Ag/kg was 65F 11% compared to the baseline response

(n = 11). The overall dose effect was significant (F1,24 =

13.1, P < 0.002, R2 = 0.84).

3.3. Benzodiazepine receptor antagonist

Intravenous administration of flumazenil resulted in a

dose-dependent recovery of superior sagittal sinus evoked

Fig. 4. Dose-dependent recovery of superior sagittal sinus evoked

trigeminocervical nucleus activity by intravenous administration of

flumazenil. The ordinate is the response after flumazenil as a percentage

of the pre-midazolam treatment baseline data, plotted as the mean with

standard error bars.

R.J. Storer et al. / Brain Research 1013 (2004) 188–193192

trigeminocervical nucleus activity (Fig. 4). At a dose of 50

Ag/kg, there was a 64F 5% recovery (n = 6). The dose effect

of flumazenil was significant (F1,20 = 30.5, P < 0.001,

R2 = 0.94).

4. Discussion

This study demonstrates that the benzodiazepine receptor

agonist midazolam can dose-dependently inhibit nociceptive

neurons in the trigeminocervical complex with input from

the dura mater. Furthermore, to demonstrate the receptor

specificity of this effect, the inhibition by midazolam is

substantially and dose-dependently reversed by the benzo-

diazepine receptor antagonist flumazenil. Given that mid-

azolam has its effect on GABAergic transmission through

allosteric modulation of the GABAA receptor, these new

data are consistent with our previous characterization of the

importance of GABAA transmission in the trigeminocervi-

cal complex [46], and reinforce that adequate study of that

mechanism is necessary to fully understand trigeminovas-

cular nociceptive transmission.

Disordered central metabolism of the inhibitory neuro-

transmitter GABA has been implicated in migraine patho-

genesis [47]. Increases in the plasma concentrations of the

excitatory amino acids L-glutamate and L-aspartate during

migraine episodes have been reported [11]. Butalbital,

which is commonly used for migraine and tension-type

headache [10,37], may act through allosteric barbiturate

binding sites on GABAA receptors. Progesterone and related

steroids have been reported to be effective in a limited way

in migraine treatment [4,29,43]; although these approaches

are not used widely in clinical practice [26]. These com-

pounds could exert their effects through GABAA receptor

neurosteroid binding sites [25]. Moreover, clinical studies in

migraine have also indicated the efficacy of potentially

GABAergic drugs, including valproate [16,19,21,32], top-

iramate [5,33], propofol [24] and baclofen [17].

Both GABAA and GABAB receptors are located at both

peripheral and central sites. In the CNS, GABA immuno-

reactivity has been demonstrated in the spinal cord [6,31].

Both GABAA and GABAB receptors are present in the

dorsal horn of rats [3,36] and humans [3]. Immunoreactivity

to glutamic acid decarboxylase (GAD), the biosynthetic

enzyme for GABA, has been demonstrated in the trigeminal

nucleus caudalis (TNC) of the cat spinal dorsal horn [2].

Ultrastructural localisation has indicated that inhibitory

GABAergic controls in the trigeminal nucleus caudalis

involve both pre- and post-synaptic mechanisms, and are

probably mediated via direct contacts onto ascending pro-

jection neurons, as well as via synaptic contacts onto

nociceptive primary afferent fibres [2]. GABAA receptors

are hetero-oligomeric transmembrane proteins that act as

ligand-gated chloride ion channels. They contain distinct

binding sites for benzodiazepines, barbituates, and other

allosteric modulators of chloride ion flux [30,42]. The

action of valproate may be mediated through GABAA

receptors [7]. Expression of the proto-oncogene c-fos as a

marker of nociceptive neuronal activity within the trigem-

inal nucleus caudalis is reduced by valproate [8] and

allopregnanolone, a neurosteroid progesterone metabolite,

which modulates GABAA receptor activity through an

allosteric binding site [7].

It seems likely that GABAergic modulation of trigemi-

novascular nociceptive transmission forms part of the brain-

stem modulatory systems that normally gate head pain. It is

known that activation of neurons in the periaqueductal grey

(PAG) matter can inhibit trigeminovascular nociceptive

transmission [22]. At the level of the PAG, GABAA receptor

activation is important [23]. In the trigeminocervical com-

plex locally ejected muscimol, a GABAA receptor agonist,

inhibits neurons responding to trigeminovascular input and

to L-glutamate microiontophoresis [46]. These observations

are consistent with a post-synaptic GABAA receptor and

that GABAA receptors are usually not found in pre-synaptic

locations [28]. It has been shown that GABAA receptors

located in the spinal cord have an antinociceptive effect

when activated [9,12,15]. The new data seem consistent

with a general principle effect of GABAA modulation on

nociceptive afferent traffic.

In conclusion, we show that midazolam, a benzodiaze-

pine receptor agonist that facilitates GABAA receptor mod-

ulation, can dose-dependently inhibit trigeminal neurons

activated by stimulation of normally pain-producing trige-

minovascular afferents. Furthermore, the benzodiazepine

receptor antagonist flumazenil dose-dependently inhibits

the effect of midazolam. Taken together, the data provide

further support for a potent and important GABAA receptor

mediated modulation of trigeminovascular nociceptive

transmission that may explain some of the clinical effects

of preventive anti-migraine treatments.

Acknowledgements

The authors thank Paul Hammond and Michele Lasa-

landara for their excellent technical assistance. This work

has been supported by the Wellcome Trust, The Guarantors

of Brain (KGS) and the Patrick Bertould Trust (KGS). PJG

is a Wellcome Senior Research Fellow.

References

[1] P. Armitage, G. Berry, Statistical Methods in Medical Research, 3rd

edn., Blackwell, Oxford, 1994.

[2] A.I. Basbaum, E.J. Glazer, W. Oertel, Immunoreactive glutamic

acid decarboxylase in the trigeminal nucleus caudalis of the cat:

a light- and electron-microscopic analysis, Somatosens. Res. 4

(1986) 77–94.

[3] N.G. Bowery, A.L. Hudson, G.W. Price, GABAA and GABAB recep-

tor site distribution in the rat central nervous system, Neuroscience 20

(1987) 365–383.

R.J. Storer et al. / Brain Research 1013 (2004) 188–193 193

[4] W.G. Bradley, P. Hudgson, J.B. Foster, D.J. Newell, Double-blind

controlled trial of a micronized preparation of flumedroxone (Demi-

gran) in prophylaxis of migraine, BMJ 3 (1968) 531–533.

[5] J.L. Brandes, D. Jacobs, W. Neto, S. Bhattacharaya, Topiramate in the

prevention of migraine headache: a randomized, double-blind, place-

bo-controlled, parallel study (MIGR-002), Neurology 60 (Suppl. 1)

(2003) A238.

[6] S.M. Carlton, E.S. Hayes, Light microscopic and ultrastructural anal-

ysis of GABA-immunoreactive profiles in the monkey spinal cord,

J. Comp. Neurol. 300 (1990) 162–182.

[7] F.M. Cutrer, M.A. Moskowitz, The actions of valproate and neuro-

steroids in a model of trigeminal pain, Headache 36 (1996) 579–585.

[8] F.M. Cutrer, V. Limmroth, G. Ayata, M.A. Moskowitz, Attenuation

by valproate of c-fos immunoreactivity in trigeminal nucleus caudalis

induced by intracisternal capsaicin, Br. J. Pharmacol. 116 (1995)

3199–3204.

[9] E.J. Drower, D.L. Hammond, GABAergic modulation of nociceptive

threshold: effects of THIP and bicuculline microinjected in the ventral

medulla of the rat, Brain Res. 450 (1988) 316–324.

[10] A.H. Elkind, Drug abuse and headache, Med. Clin. North Am. 75

(1991) 717–732.

[11] M.D. Ferrari, J. Odink, K.D. Box, M.J.A. Malessy, G.W. Bruyn,

Neuroexcitatory plasma amino acids are elevated in migraine, Neu-

rology 40 (1990) 1582–1586.

[12] H.L. Fields, M.M. Heinricher, P. Mason, Neurotransmitters in nocicep-

tive modulatory circuits, Annu. Rev. Neurosci. 14 (1991) 219–245.

[13] P.J. Goadsby, The pharmacology of headache, Prog. Neurobiol. 62

(2000) 509–525.

[14] P.J. Goadsby, R.B. Lipton, M.D. Ferrari, Migraine—current under-

standing and treatment, N. Engl. J. Med. 346 (2002) 257–270.

[15] M.M. Heinricher, H.J. Kaplan, GABA-mediated inhibition in rostral

ventromedial medulla: role in nociceptive modulation in the lightly

anesthetized rat, Pain 47 (1991) 105–113.

[16] R. Hering, A. Kuritzky, Sodium valproate in the prophylactic treat-

ment of migraine: a double-blind study versus placebo, Cephalalgia 12

(1992) 81–84.

[17] R. Hering-Hanit, Baclofen for prevention of migraine, Cephalalgia 19

(1999) 589–591.

[18] J.W. Hu, J.O. Dostrovsky, B.J. Sessle, Functional properties of neu-

rons in cat trigeminal subnucleus caudalis (medullary dorsal horn). I.

Responses to oral-facial noxious and nonnoxious stimuli and projec-

tions to thalamus and subnucleus oralis, J. Neurophysiol. 45 (1981)

173–192.

[19] R. Jensen, T. Brinck, J. Olesen, Sodium valproate has a prophylactic

effect in migraine without aura: A triple-blind, placebo-controlled

crossover study, Neurology 44 (1994) 647–651.

[20] H. Kaube, K.A. Keay, K.L. Hoskin, R. Bandler, P.J. Goadsby, Ex-

pression of c-Fos-like immunoreactivity in the caudal medulla and

upper cervical cord following stimulation of the superior sagittal sinus

in the cat, Brain Res. 629 (1993) 95–102.

[21] J. Klapper, On behalf of the Divalproex Sodium in Migraine Prophy-

laxis Study Group, Divalproex sodium in migraine prophylaxis: a

dose-controlled study, Cephalalgia 17 (1997) 103–108.

[22] Y.E. Knight, P.J. Goadsby, The periaqueductal gray matter modulates

trigeminovascular input: a role in migraine? Neuroscience 106

(2001) 793–800.

[23] Y.E. Knight, T. Bartsch, P.J. Goadsby, Trigeminal antinociception

induced by bicuculline in the periaqueductal grey (PAG) is not af-

fected by PAG P/Q-type calcium channel blockade in rat, Neurosci.

Lett. 336 (2003) 113–116.

[24] J.C. Krusz, V. Scott, J. Belanger, Intravenous propofol: unique ef-

fectiveness in treating intractable migraine, Headache 40 (2000)

224–230.

[25] J.J. Lambert, D. Belelli, C. Hill-Venning, H. Callachan, J.A. Peters,

Neurosteroid modulation of native and recombinant GABAA recep-

tors, Cell. Mol. Neurobiol. 16 (1996) 155–174.

[26] J.W. Lance, P.J. Goadsby, Mechanism and Management of Headache,

6th edn., Butterworth-Heinemann, London, 1998.

[27] R.B. Lipton, W.F. Stewart, S. Diamond, M.L. Diamond, M. Reed,

Prevalence and burden of migraine in the United States: data from the

American Migraine Study II, Headache 41 (2001) 646–657.

[28] H. Liu, H. Wang, M. Sheng, L.Y. Jan, Y.N. Jan, A.I. Basbaum,

Evidence for presynaptic N-methyl-D-aspartate autoreceptors in the

spinal cord dorsal horn, Proc. Natl. Acad. Sci. U. S. A. 91 (1994)

8383–8387.

[29] P.O. Lundberg, Prophylactic treatment of migraine with flumedrox-

one, Acta Neurol. Scand. 45 (1969) 309–326.

[30] R.L. Macdonald, R.W. Olsen, GABAA receptor channels, Annu. Rev.

Neurosci. 17 (1994) 569–602.

[31] R. Magoul, B. Onteniente, M. Geffard, A. Calas, Anatomical distri-

bution and ultrastructural organization of the GABAergic system in

the rat spinal cord. An immunocytochemical study using anti-GABA

antibodies, Neuroscience 20 (1987) 1001–1009.

[32] N.T. Mathew, J.R. Saper, S.D. Silberstein, L. Rankin, H.G. Markley,

S. Solomon, A.M. Rapoport, C.J. Silber, R.L. Deaton, Migraine pro-

phylaxis with divalproex, Arch. Neurol. 52 (1995) 281–286.

[33] N.T. Mathew, J. Schmitt, D. Jacobs, W. Neto, Topiramate in the

migraine prevention (MIGR-001): effect on migraine frequency, Neu-

rology 60 (Suppl. 1) (2003) A336.

[34] M. Menken, T.L. Munsat, J.F. Toole, The global burden of disease

study—implications for neurology, Arch. Neurol. 57 (2000) 418–420.

[35] J. Nagler, N. Conforti, S. Feldman, Alterations produced by cortisol in

the spontaneous activity and responsiveness to sensory stimuli of

single cells in the tuberal hypothalamus of the rat, Neuroendocrino-

logy 12 (1973) 52–66.

[36] J.M. Palacios, J.K. Wamsley, M.J. Kuhar, High affinity GABA recep-

tors—autoradiographic localization, Brain Res. 222 (1981) 285–307.

[37] N.H. Raskin, Headache, Churchill-Livingstone, New York, 1988.

[38] B.K. Rasmussen, J. Olesen, Symptomatic and nonsymptomatic head-

aches in a general population, Neurology 42 (1992) 1225–1231.

[39] E. Roberts, Disinhibition as an organizing principle in the nervous

system—the role of the GABA system. Application to neurologic and

psychiatric disorders, in: E. Roberts, T.N. Chase, D.B. Tower (Eds.),

GABA in Nervous System Function, Raven Press, New York, 1976,

pp. 515–539.

[40] L.A. Roberts, C. Beyer, B.R. Komisaruk, Nociceptive responses to

altered GABAergic activity at the spinal cord, Life Sci. 39 (1986)

1667–1674.

[41] R.P. Shank, J.F. Gardocki, A.J. Streeter, B.E. Maryanoff, An over-

view of the preclinical aspects of topiramate: pharmacology, phar-

macokinetics, and mechanism of action, Epilepsia 41 (Suppl. 1)

(2000) S3–S9.

[42] W. Sieghart, GABAA receptors: ligand-gated Cl� ion channels mod-

ulated by multiple drug-binding sites, Trends Pharmacol. Sci. 13

(1992) 446–450.

[43] I. Singh, I. Singh, D. Singh, Progesterone in the treatment of mi-

graine, Lancet i (1947) 745–747.

[44] R.J. Storer, P.J. Goadsby, Topiramate inhibits trigeminovascular traf-

fic in the cat: a possible locus of action in the prevention of migraine,

Cephalalgia 23 (2003) 726.

[45] R.J. Storer, P. Butler, K.L. Hoskin, P.J. Goadsby, A simple method,

using 2-hydroxypropyl-h-cyclodextrin, of administering a-chloralose

at room temperature, J. Neurosci. Methods 77 (1997) 49–53.

[46] R.J. Storer, S. Akerman, P.J. Goadsby, GABA receptors modulate

trigeminovascular nociceptive neurotransmission in the trigeminocer-

vical complex, Br. J. Pharmacol. 134 (2001) 896–904.

[47] K.M.A. Welch, E. Chabi, K. Bartosh, V.S. Achar, J.S. Meyer, Cere-

brospinal fluid g aminobutyric acid levels in migraine, BMJ 3 (1975)

516–517.

[48] J.G. Whitwam, Flumazenil and midazolam in anaesthesia, Acta

Anaesthesiol. Scand. 39 (Suppl. 108) (1995) 15–22.