Neuroscience Letters 336 (2003) 113–116 www.elsevier.com/locate/neulet
Trigeminal antinociception induced by bicuculline in the periaqueductal gray (PAG) is not affected by PAG P/Q-type calcium channel blockade in rat Yolande E. Knight, Thorsten Bartsch, Peter J. Goadsby* Headache Group, Institute of Neurology, Queen Square, London WC1N 3BG UK Received 6 September 2002; received in revised form 18 October 2002; accepted 21 October 2002
Abstract We have recently shown that injection of the P/Q-type (Cav2.1/a1A) calcium channel blocker, v-agatoxin IVA, into the periaqueductal gray (PAG) facilitates meningeal dural stimulation-evoked trigeminal nociceptive processing. We injected the GABAA antagonist bicuculline into the PAG in addition to the agatoxin and observed bicuculline’s effect on neurons responding to dural stimulation recorded in the trigeminal nucleus caudalis of rats in order to determine if P/Q channel-mediated changes acted through GABAergic mechanisms. The inhibition of trigeminal nociceptive neurons characteristic of bicuculline administered into the PAG was maintained in the presence of blocked PAG P/Q-type calcium channels. This suggests the PAG descending pain modulatory pathway is not affected by P/Q-type calcium channel blockade at the postsynaptic GABAergic inhibitory interneuron and the facilitation produced by agatoxin is mediated by another mechanism. These findings have implications for disorders involving the PAG or P/Q-type channels, such as migraine, in particular for the development of preventative treatments, suggesting GABAergic and voltage-gated calcium channels could be separately modulated. q 2002 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Electrophysiology; Periaqueductal gray; P/Q-type calcium channel; Bicuculline; Trigeminal; Nociception
Various animal models demonstrate modulation of nociceptive processing by the periaqueductal gray matter (PAG) [1]. This study examined the vlPAG, which is the site of reciprocal projections with the spinal trigeminal nucleus [2], is activated by trigeminovascular nociceptive afferents [8], and can exert inhibitory effects on specifically afferent trigeminovascular traffic [10]. It has recently been shown that blockade of P/Q-type calcium channels in the periaqueductal gray (PAG) facilitates trigeminal nociception in dorsal horn neurons [9]. P/Qtype calcium channels are found in the PAG [4] and are irreversibly blocked by v-agatoxin IVA (agatoxin) [12]. GABAergic mechanisms play an important role in the regulation of descending anti-nociceptive systems that arise from the PAG [16]. To investigate a possible mechanism by which PAG P/Qtype calcium channels could mediate nociceptive facilitation, we have studied the effect of bicuculline injection into the vlPAG on agatoxin-induced trigeminal facilitation. We used a model of cranial nociception based on experimental * Corresponding author. Fax: 144-207-813-0349. E-mail address:
[email protected] (P.J. Goadsby).
observations of migraine [6]. Migraine may involve a P/Qtype calcium channel dysfunction and is a disease whose pathophysiology includes a role for the PAG [7] Experiments were carried out under a project license issued by the UK Home Office under the Animals (Scientific Procedures) Act 1986. Sixteen male Sprague–Dawley rats weighing 334 ^ 11 g (mean ^ SD) were anaesthetized (Sagatal w {pentobarbitone sodium} 65 mg/kg i.p. induction, maintenance with a-chloralose 15 mg/kg i.v.) and, during electrophysiological recording, paralysed (Pavulon w {pancuronium bromide} 1 mg/kg initially, then 0.4 mg/kg maintenance). In the unparalysed state, absence of corneal blink reflex and flexor reflex to pinching of the forepaw were used to determine the depth of anaesthesia. When paralysed, depth of anaesthesia was determined by the absence of significant fluctuations of blood pressure and heart rate in response to noxious stimulation. Body temperature was maintained at 37 8C. Under artificial ventilation the end-tidal CO2 was maintained between 3.5–4.5%. The parietal dura mater adjacent to the middle meningeal artery (MMA) was exposed through a small cranial window and stimulated. A multibarrelled microinjection unit with
0304-3940/02/$ - see front matter q 2002 Elsevier Science Ireland Ltd. All rights reserved. doi:10.1016/S03 04 -3 94 0( 02 ) 01 250 - 8
114
Y.E. Knight et al. / Neuroscience Letters 336 (2003) 113–116
total tip diameter of no more than 100 mm was positioned in the ipsilateral vlPAG: 1.36 mm rostral and 4.2 mm dorsal from the interaural point, 0.5–0.7 mm left of the midline [14]. Drugs were pressure injected over a period of 30–120 s in a volume range of 50–600 nl. The brain stem at the level of the caudal medulla was exposed and tungsten microelectrodes were lowered into the trigeminal nucleus caudalis. The MMA was stimulated with a pair of bipolar electrodes using electrical square-wave stimuli (0.5–0.6 Hz) of 0.5–2 ms duration up to 10 V. After determining the threshold of electrical stimulus required to evoke action potentials, trains of 20 electrical stimuli were applied to the dura mater. Each round of stimulation was 5 min apart. Post stimulus histograms were constructed from the 20 stimulation sweeps, 100 ms duration per sweep. Stimulation latencies corresponding to Adand C-fibre responses were 0–30 ms and 30–100 ms. These latencies are calculated from Ad- and C-fibre conduction velocities and the distance from the middle meningeal artery to the trigeminal nucleus caudalis. To determine the rate of spontaneous activity a rate histogram of background firing was generated and displayed as spikes per second (Hz). Spontaneous activity was measured for a 30 s period within 1 min preceding the next round of middle meningeal artery stimulations. Histograms were displayed and analysed using Spike 2.01 (Cambridge Electronic Designs, UK). Functional projection between the region of PAG microinjection and the recorded trigeminal nucleus caudalis (TNC) neuron was established by search injection of the GABAA antagonist, bicuculline, into the vlPAG, where inhibition of evoked TNC activity after bicuculline injection was considered a functional connection. As agatoxin is an irreversible P/Qtype calcium channel blocker [12], only one TNC neuronal response in each animal was tested after the application of agatoxin into the PAG. Bicuculline methiodide 0.4 mM (Sigma) and v-agatoxin IVA 0.1 mM (Scientific Marketing Association) were prepared fresh on the day of the experiment. At the completion of the experiment TNC recording site and PAG injection site were marked by an electrical lesion or a deposit of Pontamine Sky Blue, respectively. The sequence of experimental events were as follows: (1) Three baseline collections of MMA-evoked TNC activity. (2) Bicuculline injection into PAG. (3) Collections at 1, 5, 10, 15, 20 and 25 min post bicuculline injection. (4) Three new baseline collections. (5) Agatoxin injection into the PAG. (6) Collections at 5, 10, 15, 20, 25, 30, 40, 50, 60 min post agatoxin injection. (7) Bicuculline injection into the PAG, preceded by top-up injection of agatoxin into the PAG. (8) Collections at 1, 5, 10, 15, 20 min post bicuculline injection. Effects on trigeminal activity were calculated as a percentage of the mean of the three responses immediately prior to intervention in the PAG expressed as 100%. Data for spontaneous activity were analysed raw (Hz). Two analyses of variance for repeated measures were used to independently determine the time course of significant drug interventions
for bicuculline, agatoxin and bicuculline with agatoxin, and to determine a significant difference between the effect of bicuculline alone compared with bicuculline in the presence of agatoxin. At the time of maximal effect, two-sample Student’s t-test for post-hoc analysis were used to evaluate statistical significance of bicuculline or agatoxin compared to baseline, or bicuculline in the presence of agatoxin compared to agatoxin at the last three samplings. Data are expressed as mean ^ SEM for a number (n) of observations. Statistical significance was set at P , 0:05. Recordings were made from 16 neurons (14 WDR/2 NS) responsive to dural stimulation with cutaneous receptive fields restricted to the ophthalmic division of the trigeminal nerve, including the cornea and pericranial muscles. Neurons were found in the deep layers of the dorsal horn of the C1/TNC transition zone at a mean depth of 990 mm (Fig. 1). PAG microinjection sites were localised to the ventrolateral PAG (vlPAG; Fig. 1). Saline control injections in the PAG (n ¼ 5), and test drugs injected outside the PAG in the superior colliculus did not affect responses to dural stimulation (n ¼ 16). To identify functional inhibitory projections from the PAG to the TNC, bicuculline was injected into the vlPAG. Injection of bicuculline (50–300 nl) into the vlPAG produced inhibition of the trigeminal nociceptive response to dural stimulation. In Ad-fibre responses maximum inhibition was observed at 5 min post bicuculline injection (F5;7 ¼ 9:59; P , 0:0001). Maximum inhibition
Fig. 1. Histological reconstruction of PAG microinjection sites in the midbrain (A) and ipsilateral TNC recording sites in the dorsal horn at the C1 and medulla transition (B). Ophthalmic division (V1) facial receptive field of a sample experiment (C) and the response properties to ophthalmic division brush or pinch, corneal brush or prodding of the frontalis muscle. Abbreviations: Aq, aqueduct; C1, cervical spinal cord division 1; TNC, trigeminal nucleus caudalis; PAG, periaqueductal gray; vl, ventrolateral. Illustration for PAG adapted from Paxinos and Watson [14] and TNC from Molander and Grant [13].
Y.E. Knight et al. / Neuroscience Letters 336 (2003) 113–116
Fig. 2. Effect on TNC activity of bicuculline (bic), agatoxin (aga) or bicuculline with agatoxin (bic 1 aga) injection into the PAG (black circle). Response to saline injection in the PAG after verification with bicuculline (open circle). Response over time; Ad-activity (A). Response over time; spontaneous activity (B). Data expressed as mean ^ SEM. *P , 0:05 and **P , 0:01 indicate significant differences compared with baseline, or compared with activity before injection in the case of bicuculline with agatoxin.
of mean spontaneous activity (MSA) was also at 5 min (F5;13 ¼ 22:40; P , 0:0001) (Fig. 2). Ad-fibre responses were inhibited by 39 ^ 4% of baseline, ranging from 8– 67% (P , 0:0001; Table 1). C-fibre inhibition was
115
55 ^ 8% (P , 0:0001). Baseline mean spontaneous activity was 55 ^ 4 Hz, ranging from 21–102 Hz. Bicuculline significantly inhibited MSA to 24 ^ 3 Hz, ranging from 4 to 95 Hz (P , 0:0001). Bicuculline in the vlPAG elicited a characteristic transient decrease in blood pressure lasting 60–100 s and fluctuating between 15–20 mmHg. Injection of 400–600 nl of agatoxin into the vlPAG produced a facilitation of the trigeminal nociceptive response to dural stimulation and to spontaneous activity. In Ad-fibre responses facilitation became significant at 15 min post agatoxin injection and was 120 ^ 5% (F9;11 ¼ 5:49; P , 0:0001). Maximal facilitation of Adresponses was at 60 min post injection and ranged from 116–216% (142 ^ 8%; P , 0:0001; Fig. 2). C-fibre facilitation was 188 ^ 41% (P ¼ 0:001). Baseline spontaneous activity after bicuculline was 55 ^ 3 Hz (range 20–90 Hz). Spontaneous activity facilitation became significant at 10 min post agatoxin injection and was 65 ^ 4 Hz (F9;16 ¼ 16:23; P , 0:0001). Maximal facilitation was at 60 min post injection and was 90 ^ 6 Hz, ranging from 31–209 Hz (P , 0:0001) (Fig. 2). In many cases a notable characteristic of the increased spontaneous activity after agatoxin injection into the vlPAG was that after an initial tonic increase of activity the firing changed to a burst-like pattern which lasted throughout the 60 min observation period. Agatoxin also affected blood pressure by inducing fluctuations over a 30–40 mmHg range throughout the observation period. Bicucilline injected into the PAG following agatoxin produced inhibition of trigeminal activity. In Ad-fibre responses maximum inhibition was observed at 5 min post bicuculline injection (F5;6 ¼ 7:41; P , 0:0001). Maximum inhibition of mean spontaneous activity (MSA) was also at 5 min (F5;16 ¼ 23:11; P , 0:0001; Fig. 2). Ad-fibre responses were inhibited by 42 ^ 5% of baseline, ranging from 8–77% (P , 0:0001). C-fibre inhibition was 58 ^ 13% (P ¼ 0:003). Bicuculline with agatoxin significantly inhibited mean spontaneous activity from 90 ^ 4 Hz to 35 ^ 6 Hz, ranging from 3 to 111 Hz (P , 0:0001). There was no significant difference between inhibition produced by bicuculline alone compared with bicuculline in the presence of agatoxin in Ad-fibre responses (F1;3 ¼ 0:01; P ¼ 0:917), C-fibre responses (P ¼ 0:99) or spontaneous activity (F1;14 ¼ 1:27; P ¼ 0:28). After the
Table 1 Peak effect† on TNC Ad-fibre, C-fibre or spontaneous activity responses to bicuculline, agatoxin or bicuculline with agatoxin injection in the vlPAG a Percentage change from baseline
Bicuculline
Agatoxin
Bicuculline and agatoxin
Ad-fibre C-fibre Spontaneous (firing rate Hz)
39 ^ 4*** 55 ^ 8*** 47 ^ 4*** (26 ^ 3 Hz)
142 ^ 8*** 188 ^ 41** 182 ^ 20*** (100 ^ 6 Hz)
42 ^ 5*** 58 ^ 13*** 48 ^ 7*** (48 ^ 6 Hz)
a
Group data of normalised values indicating percentage change from baseline. Mean ^ SEM. Comparison with baseline, P , 0:0001***, P , 0:001**, P , 0:01*. †Baseline firing rate 55 ^ 4 Hz.
116
Y.E. Knight et al. / Neuroscience Letters 336 (2003) 113–116
effect of bicuculline had worn off spontaneous activity began to return to a level similar to that induced by agatoxin, where spontaneous activity at 20 min post bicuculline injection with agatoxin was 67 ^ 6 Hz and Ad-responses were 90 ^ 9%. Bicuculline in the presence of agatoxin in the vlPAG elicited the same characteristic transient decrease in blood pressure of 15–20 mmHg over 60–100 s post injection as was seen with bicuculline alone. The new data demonstrate that inhibition of trigeminal nociception induced by bicuculline injection in the vlPAG is unaffected by P/Q-type calcium channel blockade. Our results further demonstrate that bicuculline injection in the PAG inhibits nociceptive transmission in the trigeminal nucleus caudalis and that agatoxin injection in the PAG facilitates trigeminal nociception. These latter findings are consistent with previous studies demonstrating that PAG neurons can modulate nociceptive traffic from the periphery [11]. PAG agatoxin-induced facilitation of trigeminal nociception indicates a P/Q-type calcium channel mechanism since v-agatoxin IVA is an irreversible blocker of these channels [12]. An important caveat is that in some cells agatoxin has a small effect on T-type channels [15] that depends on the concentration of agatoxin at the synapse. Because of the difficulty in accurately controlling toxin concentration at all synapses in complex tissue [5] we cannot exclude the possibility that non-P/Q-type channels were also blocked by agatoxin. The population of T-type channels in the PAG is only moderate [3], but we must consider that some small portion of the facilitation may be attributable to T-type channel blockade. Following injection of agatoxin into the PAG we observed a gradual increase in both spontaneous activity and evoked Ad- and C-fibre responses. Given that P/Q-type calcium channel blockade produced an increase in evoked and, especially, spontaneous activity it seems likely the channels were located on actively firing neurons involved in tonic modulation of the trigeminovascular nociceptive response. Subsequent application of bicuculline onto neurons with already blocked P/Q-type calcium channels produced inhibition of nociceptive responses. This effect was similar in magnitude and duration to the response to bicuculline without agatoxin. Maintenance of the bicuculline-induced inhibition in the presence of agatoxin indicates that P/Q-type calcium channel blockade does not affect the site of bicuculline action, probably a postsynaptic GABAA receptor mediated inhibition of inhibitory interneurons in the PAG. Furthermore, P/Q-type calcium channel blockade does not affect the pre-synaptic GABA release subsequently blocked postsynaptically by bicuculline. Based on known sites of action of bicuculline in the PAG, our results suggest that GABAergic disinhibition does not play an important role in P/Q-type calcium channel effects in the PAG. The authors wish to thank Simon Akerman, R. James Storer, Michele Lasalandra, Paul Hammond, Bridget
Lumb, Simon McMullan, David Bulmer and Alexandra V Gourine for technical advice and support. The Wellcome Trust and the Migraine Trust supported this work. The Deutsche Forschungsgemeinschaft supported Thorsten Bartsch. [1] Behbehani, M.M., Functional characteristics of the midbrain periaqueductal gray, Prog. Neurobiol., 46 (1995) 575–605. [2] Beitz, A.J., Wells, W.E. and Shepard, R.D., The location of brainstem neurons which project bilaterally to the spinal trigeminal nuclei as demonstrated by double flourescent retrograde tracer technique, Brain Res., 258 (1983) 305–312. [3] Craig, P.J., Beattie, R.E., Folly, E.A., Bannerjee, M.D., Reeves, M.B., Priestley, J.V., Carney, S.L., Sher, E., PerezReyes, E. and Volsen, S.G., Distribution of the voltagedependent calcium channel alpha1G subunit mRNA and protein throughout the mature rat brain, Eur. J. Neurosci., 11 (1999) 2949–2964. [4] Craig, P.J., McAinsh, A.D., McCormack, A.L., Smith, W., Beattie, R.E., Priestley, J.V., Yip, J.L., Averill, S., Longbottom, E.R. and Volsen, S.G., Distribution of the voltagedependent calcium channel alpha(1A) subunit throughout the mature rat brain and its relationship to neurotransmitter pathways, J. Comp. Neurol., 397 (1998) 251–267. [5] Dunlap, K., Luebke, J.I. and Turner, T.J., Exocytotic Ca 21 channels in mammalian central neurons, Trends Neurosci., 18 (1995) 89–98. [6] Feindel, W., Penfield, W. and McNaughton, F., The tentorial nerves and localization of intracranial pain in man, Neurology, 10 (1960) 555–563. [7] Goadsby, P.J., Lipton, R.B. and Ferrari, M.D., Migraine: current understanding and treatment, N. Engl. J. Med., 346 (2002) 257–270. [8] Hoskin, K.L., Bulmer, D.C.E., Lasalandra, M., Jonkman, A. and Goadsby, P.J., Fos expression in the midbrain periaqueductal grey after trigeminovascular stimulation, J. Anat., 197 (2001) 29–35. [9] Knight, Y.E., Bartsch, T., Kaube, H. and Goadsby, P.J., P/Qtype calcium channel blockade in the PAG facilitates trigeminal nociception: a functional genetic link for migraine? J. Neurosci., 22 (2002) 1–6. [10] Knight, Y.E. and Goadsby, P.J., The periaqueductal gray matter modulates trigeminovascular input: a role in migraine? Neuroscience, 106 (2001) 793–800. [11] Malmberg, A.B. and Yaksh, T.L., Voltage-sensitive calcium channels in spinal nociceptive processing- blockade of Ntype and P-type channels inhibits formalin-induced nociception, J. Neurosci., 14 (1994) 4882–4890. [12] Mintz, I.M., Venema, V.J., Swiderek, K.M., Lee, T.D., Bean, B.P. and Adams, M.E., P-type calcium channels blocked by the spider toxin omega-Aga-IVA, Nature, 355 (1992) 827– 829. [13] Molander, C. and Grant, G., Spinal cord cytoarchitecture, In G. Paxinos (Ed.), The Nervous System, Academic Press, San Diego, CA, 1995. [14] Paxinos, G. and Watson, C., The Rat Brain in Stereotaxic Coordinates, Academic Press, London, 1998. [15] Rusin, K.I. and Moises, H.C., Mu-Opioid receptor activation reduces multiple components of high-threshold calcium current in rat sensory neurons, J. Neurosci., 15 (1995) 4315–4327. [16] Sandkuhler, J., Willmann, E. and Fu, Q.G., Characteristics of midbrain control of spinal nociceptive neurons and nonsomatosensory parameters in the pentobarbital-anesthetized rat, J. Neurophysiol., 65 (1991) 33–48.