- Email: [email protected]
The induction and maintenance of central sensitization is dependent on N-methyl-d-aspartic acid receptor activation; implications for the treatment of post-injury pain hypersensitivity states
293
Penn, 44 (1991) 293-299 .I) 1991 Elscvier Science Publishers ADONIS 0304395991~928
B.V. 0304.3959/91/~03.50
PAIN 01746
Basic Section The induction and mainterkmce of central sensitization is dependent on N-methyl-D-aspartic acid receptor activation; implications for the treatment of post-injury pain hypersensitivity states
Depurtmen~
Clifford J. Woolf and Stephen W.N. Thompson ofAnaromv and Deoelopmentoi Biology, Untversity College London, London (Received
2 July 1990, revision
received
28 August
1990, accepted
31 August
WCIE
6BT (U. K.)
1990)
Repetitive stimulation of small diameter primary afferent fibres produces a progressive increase in Summary action potential discharge (windup) and a prolonged increase in the excitability of neurones in the spinal cord following the stimulus. Previous studies have demonstrated that windup is the consequence of the temporal summation of slow synaptic potentials and that the slow potentials and windup are reduced by pretreatment with ~-methyl-D-aspartic acid (NMDA) antagonists. We have now examined whether primary afferent induced hypersensitivity states in flexor motoneurones are also dependent on the activation of NMDA receptors and whether windup is a possible trigger for the production of the central hypersensitivity. Both a non-competitive (MK-801) and a competitive (D-CPP) NMDA antagonist, at doses that did not modify the baseline reflex, reduced the facilitation of the flexor reflex produced by either brief electrical stimulation of the sural nerve (1 Hz for 20 set at C-fibre strength), or by the cutaneous application of the chemical irritant mustard oil. These antagonists also prevented windup from occurring in the motoneurones. When the the MK-801 and the D-CPP were administered once a state of central facilitation had been induced by prior treatment with mustard oil, they returned the facilitated reflex to its pretreatment level. These results indicate that NMDA receptors are involved in the induction and maintenance of the central sensitization produced by high threshold primary afferent inputs. Because central sensitization is likely to contribute to the post-injury pain hypersenstivity states in man, these data have a bearing both on the potential role of NMDA antagonists for pre-emptive analgesia and for treating established pain states. Key words: Post-injury
pain;
Hypersensitivity
states;
Analgesia;
Introduction The hyperalgesia that follows peripheral tissue injury results from an increase in the sensitivity of primary afferent nociceptors in the vicinity of the injury (peripheral sensitization) [1,2,23], and from an increase in the excitability of neurones in the spinal cord (central sensitization) [35,38]. Central sensitization is triggered by nociceptive afferent inputs [30,41] and manifests as a prolonged reduction in the threshold, an expansion of the extent, and an increase in the responsiveness of the
Correspondence IO: Dr. C.J. Woolf, Department of Anatomy and Developmental Biology. University College London, Gower Street, London WClE 6BT. U.K.
NMDA
antagonists;
(Rat)
cutaneous receptive fields of dorsal horn neurones [3,6,14,25,40]. Low threshold afferent inputs have no similar effects [6]. These changes parallel the post-injury hypersensitivity of the flexion withdrawal reflex in animals [35,37], and of pain sensation in man [16]. A distinctive feature of those afferents that can induce central sensitization is their capacity to produce slow synaptic potentials in spinal neurones [26,43]. Following low frequency repetitive activation of the afferents, these slow potentials summate to produce a cumulatively increasing postsynaptic depolarization [26]. This leads to a progressive increase in action potential discharge on repeated stimulation, a phenomenon known as windup [20]. Extracellular studies of dorsal horn neurones in vivo have demonstrated that windup is prevented by administration of NMDA antagonists [7,9].
We have found. using intracellular recordings from flexor motoneurones in vitro, that the prolonged postsynaptic depolarization produced by a single high intensity stimulus, the cumulative summation of these potentials and windup, are all reduced by the NMDA antagonist d-APV [26]. What is not known is whether an NMDA receptordependent windup of spike discharge is a necessary prelude to the generation of central sensitization. In order to investigate this we have examined whether a highly specific competitive NMDA receptor antagonist (D-CPP, 3(( R)-2-carboxypiperazin-4-yl)-propyl-I -phosphonic acid) [5,10] and a non-competitive anatogonist (MK-801, dizocilpine) which binds to the PCP site in the NMDA receptor-gated ion channel [34], can prevent the establishment both of windup and of central sensitization. Both of these drugs block NMDA depolarizations in the rat spinal cord [5,8]. The approach used has been to examine alterations in the excitability of the rat hind limb flexion withdrawal reflex. The flexion reflex in man is linearly correlated with the perception of pain [33] and in animals provides a convenient way of monitoring the processing in the spinal cord of sensory information from high threshold afferent fibres [35]. In the absence of any disturbing stimulus, the discharge of action potentials in hamstring flexor alpha motoneurones elicited by a standard nontissue-damaging noxious mechanical stimulus applied at regular intervals to the toes is very stable in the decerebrate-spinal rat preparation [30,41]. A transient ( < 10 min) facilitation of the reflex can be induced by a brief (20 set) low frequency (1 Hz) train of stimuli to the sural nerve, provided that C-fibre strength stimuli are used [30]. These same stimulus parameters also produce windup [20]. A much longer facilitation can be produced by stimulation of a muscle nerve or by the cutaneous application of the chemical irritant, mustard oil [30,41]. Mustard oil produces an intense burning sensation when applied to the skin in man and a mechanical and thermal hyperalgesia in rats that lasts up to 24 h [37]. The mustard oil produces a burst of activity in C-nociceptors in the rat that decays within a few min [41]. Following its application to the skin, alterations in the receptive field properties of dorsal horn neurones [39] and a facilitation of the flexion reflex [41] occur that last for over 1 h. This chemical irritant provides a useful model to study the prolonged central effects of a brief afferent barrage in chemosensitive fibres.
Methods Experiments were performed on adult (200-300 g) Sprague-Dawley rats. The rats were anaesthetized with a 0.3-0.4 ml bolus injection of Althesin (alphaxalone/ alphadalone, Glaxo) into a tail vein. The carotid artery
was cannulated and anaesthesla maintalncd with >+rn,~ll (0.1 ml) intermittent i.a. doses c>f tht anaesthetlc ~vt~r~ 5- 10 min. The trachea was cannulated and the arumal decerebrated by aspiration of all crania! g:ontenth rostra! to the mesencephalon. A spinaliLatiort ~‘a> performed via a laminectomy at T3 or T4. The anaesthetic ~3s then discontinued, the animals paralyhed with gallamlnc (10 mg) and artificially ventilated. Rectal tsmperature. expired pCOz. heart rate and ECG were monitored and maintained within normal physiological limit+ The nerve to the posterior biceps frnlorls/semrtendinosus muscles was exposed on one hide In the popliteal fossa, dissected free, split into smal! filaments and covered with mineral oil. One of the filaments was placed on a silver recording electrode and CxtracellulaI recordings made of flexor alpha motoneuronr activity as described in detail before [30.41]. In \c)mt’ animals the sural nerve was dissected free and placed on a pair of silver stimulating electrodes. Action potential drscharge was counted using pulse intergrators providing both peristimulus histograms and total counts of the action potential discharge elicited by a standard mechanical stimulus. A standard pinch was applied to the ipsi- and contralateral middle 3 toes using a pair of calibrated forceps (100 g ~~- contact area 1X mm’ ). The standard mechanical stimulus was applied to each foot every 5 min and the total number of spikes elicited counted. Because the number of spikes evoked. although stable for a given animal. varied between animals, the data were transformed as described below to enable comparisons between animals to be made. A baseline value for the reflex, prior to any experimental manipulation, was calculated by determining the mean number of spikes generated by at least 8 successive stimuli at 5 min intervals, commencing 1 h after the spinalization. All individual results for :! particular animal were then calculated as a percentage of this mean baseline value. Mustard oil (allyl-isothiocyanate) 5% in mineral 011 was applied to a 4 mm’ patch of skin on the dorsum of the foot. The MK-801 (Merck, Sharp and Dohme) and D-CPP (Tocris Neuramin) were dissolved in saline and administered by slow (3 min) infusion via the carotid artery
Results Actions of D-CPP and MK-801 on the baseline reflex In the absence of any treatment the standard mechanical stimulus produced a very stable response, measured in terms of the total number of action potentials generated, when repeated at 5 min intervals over prolonged periods as reported previously [30,41]. The MK801 failed to alter the baseline reflex when administered at a dose of 0.1 mg/kg or lower. At doses of 0.5 and 1.0
295
mg/kg a depression of the reflex, commencing within 15 min of treatment, did develop (Fig. 1). The most sensitive aspect of the reflex was the small discharge of action potentials elicited by a contralateral stimulus, but at doses of 1.0 mg/kg or higher a substantial and prolonged depression of the ipsilateral reflex also occurred. The D-CPP also produced a reduction in the baseline reflex at a dose of 5.0 mg/kg (n = 2). This developed over a slower time course than the depression produced by the MK-X01 with peak falls in the action potential discharge manifesting 15-25 min after administration of the drug. At a dose of 2.5 mg/kg, the D-CPP produced no significant modification of the reflex (n = 4).
O.lmgkf’
i 0
I
I
20
10
30
40
50
60
Time (mln) Fig. 1. Alterations in the excitability of the flexion withdrawal reflex produced by the systemic administration of the non-competitive NMDA ion channel blocker MK-801. Excitability is expressed as the percentage deviation from the baseline (predrug treatment) response to a standard (150 g) mechanical stimulus applied for 3 set to the ipsilateral middle 3 toes. For each stimulus the number of spikes elicited in biceps femoris/semitendinosus flexor alpha motoneurones were counted using a pulse integrator. The different doses were administered i.a. at the time indicated by the arrow. Each point represents mean i SE., with n = 6 for 0.1 mg/kg, n = 5 for 0.5 mg/kg and n = 3 for 1 .O mg/kg MK-801.
Windup of motoneurone action potentiul discharge Stimulation of the sural nerve at C-fibre strength (5 mA, 500 psec) at 1 Hz produced a progressive increase in the number of action potentials elicited by the motoneurones per stimulus, over a 20 set period of stimulation. This windup can best be represented as a cumulative spike count which shows the incrementing
125
100
i 600 7.5 500 % 5c
400
2 2 v
300
01 1 :: .: 0 z ;
25 2oc
0 IOC
PRE
0
5
10
CPP 15
20
Time(s) Fig. 2. A spike count illustrating the cumulative increase in the number of action potentials elicited per stimulus in hamstring flexor motoneurones during stimulation of the sural nerve at 1 Hz for 20 set at C-fibre strength (5 mA, 0.5 msec), before (solid lines) and 20 min after 0.5 mg/kg MK-801 (dotted lines). The presence of windup is indicated by the progressive upward deviation of the curve in the control situation, while its absence after MK-801 is indicated by the straight line.
POST
-
PRE
POST
MK- 80t
Fig. 3. Changes in the peak facilitation of the hamstring flexion reflex produced 1 min after a sural nerve conditioning stimulus (1 Hz. 5 mA, 0.5 msec for 20 set) before and after treatment with D-CPP (2.5 mg/kg) and MK-801 (0.5 mg/kg). Increased excitability is expressed as the percentage increase above the baseline (predrug treatment) of the number of spikes generated by a standard mechanical stimulus as described in Fig. 1. Both D-CPP and MK-801 significantly reduced the sural induced facilitation at 1 and at 5 min after the conditioning stimulus (P < 0.01. n = 5 for MK-801 and n = 4 for D-CPP), while saline (0.2 ml) was without effect (n = 2).
125
loo
-75
I 10
’ 0
I 20
1 30
1 40
Twne Fig.
I 50
I 60
I 70
1 80
J
90
( mln)
4. The effect of pretreatment
with 0.1 mg/kg MK-801 and 2.5 mg/kg D-CPP on the excitability Jncrease in flexor motoneurones produced by the cutaneous application of mustard oil to the hairy skin on dorsum of the foot. The solid line shows the excitability increase that occurs Jn the absence of MK-801 or D-CPP (n = 6). The dotted line shows the result achieved by pretreatment with MK-801 (n = 6) I5 min before mustard oil application while the dashed line shows the effect of the D-CPP treatment 30 min before the mustard oil (n = 5). Each point Js the mean + S.E.
response over the period of the stimulation (Fig. 2). Systemic injections of MK-801 (0.5 mg/kg. n = 5) (e.g., Fig. 2) and of D-CPP (2.5 mg/kg. n = 4) prevented the
windup from occurring in the flexor motoneurones. Before MK-801 treatment the action potential discharge per stimulus after 20 set of sural nerve stimulation had
200
175
0 / 0
150
125
100
0
IO
20
I 30
I
MK.801
MO
-75
40 Tune
I 50
I 60
70
I
80
1 90
(mm)
Fig. 5. The reduction of the excitability of the flexion reflex to baseline levels produced by the administration of MK-801 (0.1 mg/kg, n = 5, dotted line) and D-CPP (2.5 mg/kg, n = 5 dashed line). once a state of central sensitization had been induced by the cutaneous application of mustard oil (MO) 30 min earlier. The duration of the facilitation of the reflex produced by mustard oil in the absence of MK-801 or D-CPP is shown by the solid line (n = 6). l indicates P < 0.05 and * * P < 0.01 compared tcr the control facilitation
297
built up from 24 ISI5 spikes for the first stimulus to 39 + 12 spikes for the last (P -C0.05, n = 5). An incrementing discharge failed to develop after MK-801 treatment, after 20 set of sural nerve stimulation the number of spikes evoked by the last stimulus was 22 + 7 spikes compared with 20 f 6 for the first.
The sural nerve stimulation that produces a windup of spikes in the flexor motoneurones of control animals also produces a period of reflex facilitation after the sural nerve stimulation has ceased [30]. This facilitation is maximal 1 min after the conditioning stimulus and reduces to baseline levels within 10 min 1301. Previous studies have shown that the facilitation only occurs when C-fibre strength conditioning stimuli are used [30]. MK-801 and D-CPP pretreatment, in addition to reducing the windup, effectively abolished the sural nerve induced facilitation of the reflex with no facilitation of reflex discharge above baseline at 1 (Fig. 3) or 5 min.
A prolonged period of increased excitability of the flexion reflex is produced by the application of mustard oil to the dorsum of the foot 1411 {Fig. 4). Pretreatment with 0.1 mg/kg MK-801 or with 2.5 mg/kg D-CPP prevented the mustard oil from producing this facilitation (Fig. 4). Because the depressant effects of the D-CPP on the baseline reflex at 5.0 mg/kg developed at a slower rate than that produced by the high doses of MK-801. the D-CPP was administered 30 min before the mustard oil as opposed to 15 min for the MK-801 (Fig, 4). Saline (n = 2) had no effect on the facilitation. When the MK-801 (0.1 mg/kg) and the D-CPP (2.5 mg/kg) were given 30 min after the development of a mustard oil induced f~~cilitation of the flexion reflex, the elevated excitability was significantly reduced, within 5 min for the MK-801 and within 15 min for the D-CPP (Fig. 5). In both cases the reflex returned to or below the baseline level of the reflex. A similar effect could be demonstrated for the prolonged facilitation that follows a 20 set C-fibre strength stimulation of the nerve to the gastrocnemius muscles. Such stimulation, in contrast to sural nerve stimulation, produces a prolonged facilitation of the flexor reflex [30]. The administration of 0.1 mg/kg MK-801 20 min after the induction of such a facilitation returned the reflex to baseline levels (n = 2).
Discussion The activation of small diameter primary afferent neurones is capable of producing a prolonged alteration in the excitability of neurones in the spinal cord, modifying the way these neurones respond to subsequent
inputs and thereby generating a state of central sensitization [35,373. These changes can be produced by electrical stimulation of peripheral nerves. with muscle afferents producing longer lasting effects than cutaneous afferents [30,41], by peripheral injury or localized inflammatory lesions [14,24.35], and by activating chemosenitive afferents with chemical irritants such as mustard oil 111,411 or capsaicin 1251. What is the importance of central sensitization? We have argued previously that while the capacity to detect and react to non-tissue-damaging noxious stimuli is the consequence of the activation of functionally specified high threshold nociceptors, which generate “physiological” pain, clinical or “pathological” pain is in contrast, the consequence of the induction of peripheral and central sensitization in the somatosensory system [38]. The changes represented by central sensitization include a situation where low threshold afferents begin to produce pain. something they never normally do, and this helps to explain why low intensity or innocuous stimuli are painful, contributing to the allodynia that occurs in the vicinity of a peripheral injury. Such a change in the sensory responses to low threshold mechanoreceptors has recently been demonstrated in human subjects in whom a state of secondary mechanical hyperalgesia was induced by the administration of capsaicin to the skin [28]. The mechanism involved seems to be the recruitment of low threshold mechanoreceptor afferent input to cells which prior to the production of central sensitization had high threshold receptive fields [39,403. The ability of NMDA receptor antag(~nists to block the development and maintenance of central sensitization implies that it should be possible to differentiate pharmacologically between physiological and pathological pain states in the sense that NMDA antagonists, by returning an abnormally exctible spinal cord to its normal level of excitability, will return a pathological situation where low and high intensity stimuli produce pain to one where there is a clear distinction between noxious and innocuous stimuli in terms of the different sensations they elicit. The clinical significance of central sensitization lies in the implicit prediction that can be made that preemptive analgesia, by preventing the establishment of central sensitization. will reduce the pain experienced following a peripheral injury. In other words if the afferent input that produces central sensitization is prevented from gaining access to the CNS or if its central effects are prevented, then less pain will result. The easiest way of testing this clinically is to use local anaesthetics prior to a surgical intervention. Such an approach has been found to reduce postoperative pain compared to patients who received a general anaesthetic without local block at the site of the surgery [29]. An experimental demonstration of preemptive “analgesia” using opioids. has shown that low doses of morphine
29x
prevented the development of an afferent induced reflex hypersensitivity but were ineffective in reducing the reflex ilypersensitivity once it was established [42]. Doses of morphine an order of magnitude higher were required to reduce the reflex facilitation to baseline levels [42]. This finding may relate to the reports of the effect of opioid premeditation on postoperative pain [17]. The present data show that NMDA antagonists differ from morphine in that the same dose of the NMDA antagonist that is required to prevent the establishment of central sensitization also abolishes it once it has been induced. This may mean that both morphine and an NMDA antagonist such as ketamine could be useful as pre-emptive analgesic agents but that NMDA antagonists may be more efficacious in reducing established central sensitization. These data also draw attention to the fact that a drug may have analgesic activity in terms of reducing a pathological post-injury pain hypersensitivity state, while leaving the normal physiological response to noxious stimuli intact. Earlier suggestions that the NMDA receptor may be involved in nociception relate to the effects of a variety of NMDA antagonists on behavioural tests [4] and on the responses of dorsal horn cells to peripheral noxious stimuli [32]. It is likely though that these studies were looking more at the suppressant effects of high doses of NMDA antagonists than on a blockade of central sensitization. Indeed conventional screening for analgesic drugs is designed to examine ‘physiological pain’ and not pathological hypersensitivity states. More relevant is the report of the reduction of postoperative pain produced by ketamine in man [18], and the hyperalgesia that is generated in laboratory animals by intrathecal injections of NMDA [31]. The cellular and molecular mechanisms underlying the contribution of the NMDA receptor-ion channel complex to the induction and maintenance of central sensitization are not known. The finding that windup is also reduced by NMDA antagonists suggests that the cumulative depolarization that results from the summation of slow potentials [26] is an important first step in the induction of central sensitization. The reason for this is likely to be the voltage-dependent block of the NMDA ion channel by Mg2+ [19,22]. A depolarization generated by the summation of EPSPs, by removing or decreasing this Mg2+ block, will increase inward currents through the ion channel amplifying the depolarization and increasing synaptic efficacy. Neuropeptides, particularly the neurokinins which have the capacity to produce prolonged depolarizations [21] and appear to be able to act as volume transmitters over wide areas [12], may be important in recruiting an NMDA component to any postsynaptic response. The maintenance of the state of central sensitization by the NMDA receptor--ion channel complex could be the result of uncovering a new polysynaptic pathway
involving NMDA receptors or a tonic. rclcabc :>f endtrg~nous transmitter from terminals producing a tom& activation of NMDA receptors. It is Intriguing to consider the possible role of changes in s<:cond messengers in mediating the prolonged effects of NMDA receptor activation. The same afferents that induce central sensetization can increase protein phosphorvinttr\rt in the dorsal horn in an N‘~DA-depeildeIlI fashjon 1361 and can initiate the expression of the immediate-cariy gene product c-Fos 1131. although whether this change in gene expression can be modified by NMI)A antagonist5 is controversial [15,27]. In conclusion, the demonstration that ;t competitive and a non-competitive NMDA receptor antagonist can prevent the induction of central sensitization and abolish the hypersensitivity once it is established, suggests that this class of drugs has potential both for the prevention and treatment of acute pain. The data alao reinforce the view that the pathogenesis of pain is in part a reftection of the capacity of central neurones to undergo dynamic aiterations in their response properties ;Ind rhe fact that nociceptive afferents can trigger such LLfuncttonal plrtaticity [401.
Acknowledgements
We thank the MRC and the Wellcome Trust for financial support. Merck. Sharp and Dohme for the supply of the MK-801 and J. Middleton for technical assistance.
References Ressou. P. and Perl. E.R., with unmyeiinated (1969)
Response
of cutaneous
fihres to noxious stimuli.
wnsory
units
.I. Ncurophysiol..
??.
1025-1043.
Campbell.
J.N..
myelinated
Neurophysiol.. C‘ervero.
F..
noxtous
Meyer,
and LaMotte.
afferents
42 (1979) of
R.H..
that innervate
oi
Sensttlzation monkey
hand
J.
1669- 1679.
Handwerker.
stimulation
properties
R.A.
nociceptive
H.O. the
and
rat’s
Lalrd.
tail:
J.M..A..
responses
Prolonged
and
of dorsal horn neurons. J. Physwl. (Land.).
encoding 404 (1988)
419.-436. C‘ahusac, Srmth.
A.M..
D.A.%
rweptor
Evans.
R.H..
Hill.
The hehavioural
antagonist
following
R.G..
Rodriguez.
application
IO the lumbar
cord of conscious rats. Neurophannacology. Childs,
A.M..
Evans. R.H.
and Watkins,
cai selectwity of three NMDA (1988)
18.-86.
(‘wk.
A.J.. Woolf.
rcceptlve primary
C‘.J.. Wall.
field plasticity afferent
receptors
23 (1984)
dorsal
325 (1987)
in “wind-up”
IS1
for Involvement
J.A
J.D.. Aram.
C-
of N-meth-
of class 2 neurones
Dawes.
D.. Millar.
following
153
402-406
S.N.. Martm.
145
S.B., Dynamic
horn
dorsal horn of the rat. Brain Res.. 424 (1987) I.odpc. I>.. IXffcrcnces
spinal
719- 724
fiur. .f. Pharmacol..
P.D. and McMahon.
in rat spinal
Dawes. S.N. and Lodge, D., Evidence ylaspartate
and
J.C.. The pharmacologt-
antagonists.
Input. Nature.
R.E.
effects of an hi-methylaspartate
. C‘hurch.
tn the J. and
tin results from in viva and In v!tr
on the use dependency of N-methylaspartate antagonism by MK801 and other phencyclidine receptor ligands. Eur. J. Pharmacol., 145 (1988) 141-151. 9 Dickenson. A.H. and Sullivan, A.F., Evidence for a role of the NMDA receptor in the frequency dependent potentiation of deep dorsal horn neurones following C-fihre stimulation. Neuropharmacology, 26 (1987) 1235-1238. IO Harris, E.W., Ganong. A.M.. Monoghan. D.T., Watkins, J.C. and Cotman. C.W.. Action of 3-(( k )-2-carboxypiperazin-4-yl)-propyll-ph(~sphoric acid (CPP) a new and highly potent antagonist of N-methyl-D-aspartate receptors. Brain Res., 382 (1986) 174-182. 11 Hoheiael, U. and Mense. S.. Long-term changes in discharge behaviour of cat dorsal horn neurones following noxious stimulation of deep tissues. Pain, 36 (1989) 2399247. 12 Hope. P.J., Schaible, H.G.. Jarrott, B. and Duggan, A.W., Release and persistence of imnlun~~reactive neurokinin A in the spinal cord is associated with chemical arthritis. Pain, Suppl. 5 (1990) S230. 13 Hunt, S.P.. Pim. A. and Evan, G.. Induction of C-fos-like protein in spinal cord neurones following sensory stimulation. Nature, 328 (1987) 6322634. 14 Hylden, J.L.K., Nahin. R.L., Trauh. R.J. and Dubner. R., Expansion of receptive fields of spinal lamina I projection neurons in rats with unilateral adjuvant-Induced Inflammation: the contribunon of dorsal horn mechanisms. Pain, 37 (1989) 229-243. 15 Kehl, L.J., Basbaum, A.I., Pollock, C.H., Mayes, M. and Wilcox, G.L.. The NMDA receptor antagomst MKXOl reduces noxious stimulus-evoked FOS expression in the mammalian spinal dorsal horn. Pain. Suppl. 5 (1990) S165. 16 LaMotte, R.H., Thalhammer, J.G. and Robinson, C.J.. Peripheral neural correlates of magnitude of cutaneous pain and hyperalgesia: a comparison of neural events in monkey with sensory Judgements in human. J. Neurophysiol., 50 (1983) l-26. 17 McQuay. H.J.. Carroll, D. and Moore, R.A., Postoperative orthothe effect of opiate premeditation and local paedic paln anaesthetic blocks. Pain. 33 (19X8) 291-295. 18 Maurset. A.. Skoglund. L.A., Hustveit, 0. and Oge, I., Comparison of ketamine and pethidine in experimental and postoperative pain. Pain, 36 (1989) 37-42. I9 Mayer. M.L.. Westbrook, G. and Guthrie. P.B.. Voltage-dependent block by Mg2+ of NMDA responses in spmal cord neurones. Nature, 309 (19X4) 261-263. 20 Mendell, L.M.. Modifiability of spinal synapses. Physiol. Rev., 64 ( 1984) 260- 324. 21 Murase, K. and Rand& M., Actrons of substance P on rat spinal cord and dorsal horn neurones. J. Physiol. (Lond.), 346 (1984) 203-217. 22 Nowak, L., Bregestovjski, P., Ascher, P., Herbert. A. and Prochianty, A.. Magnesium gates glutamate activated channels in mouse central neuronea. Nature, 307 (1984) 462-465. 23 RaJa. S.. Campbell. J.N. and Meyer, R.A., Evidence for different mechanisms of primary and secondary hyperalgesia following heat injury to the glabrous skin. Brain. 107 (1984) 1179-1188. 24 Schaible. H.G.. Schmidt, R.F. and Willis. W.D.. Enhancement of the responses of ascending tract cells in the cat spinal cord by acute inflammation of the knee joint. Exp. Brain Res.. 66 (1987) 489-499. 25 Simone. D.A.. Snrkin. L.S., Chung. J.M.. Willis, W.D. and
LaMotte. R.H., Spinothalamic tract cells hyperalgesia following rntradetmal capsaicin.
signal the pain and Sot. Neurosci. Ahst.,
14 (1988) 559. 26 Thompson. S.W.N., King, A.E. and Woolf. C.J.. Activity-dependent changes in rat ventral horn neurones in vitro; summation of prolonged afferent evoked postsynaptic depolarizations produce a d-APV sensitive windup. Eur. J. Neurosci., 2 (1990) 6388649. 27 Tolle, T.R.. Castro-Zopes, J.M.. Coimbra, A. and Zieglglnsberger, W.. C-Fos induction in the spinal cord following noxious stimulation: prevention by opiates but not NMDA antagonists. Pain. Suppl. 5 (1990) 5229. 28 TorebJiirk. E.. Lundberg. L. and LaMotte. R.. Neural mechanisms for capsaicin-induced hyperalgesia. Pain, Suppl. 5 (1990) S114. 29 Tverskoy, M.. Cozacov. C., Ayache, M.. Bradley. E.L. and Kissin. 1.. Postoperative pain after inguinal herniorrhaphy with different types of anesthesta. Anesth. Analp., 70 (79%)) 29-35. 30 Wall. P.D. and Woolf. C.J., Muscle but not cutaneous C-afferent input produces prolonged increases rn excitabiiity of the flexion reflex tn the rat. J. Physiol. (Land.). 356 (1984) 44-458. 31 Wilcox, G.L. and Aansen. L.M.. Nocicepttve action of excitatory amino acids. J. Pharmacol. Exp. Ther.. 243 (1987) 9-19. 32 Wilcockson. W.S., Chung. J.M.. Hon. Y.. Lee. K.H. and Willis. W.D., Effects of i~~nt(~phoretically released amino acids and amines on primate spinothal~til~ic tract neurons. J. Neurosci.. 4 (1984) 732-740. 33 Wilier. J.C.. Comparative study of perceived pain and nociceptive flexion reflex in man. Pain. 3 (1979) 69-80. 34 Wang. E.H.F.. Kemp. J.A.. Priestly, T.. Knight. A.R., Woodruff. G.N. and Irvine, L.L., The anticonvulsant MK801 is a potent W-methyl-n-aspartate antagomst. Proc. Nat. Acad. Sci. (U.S.A.). 83 (1986) 7104-7108. 35 Woolf. C.J.. Evidence for a central component of postinjury pain hypersensitivity. Nature, 308 (1983) 686-688. 36 Woolf. C.J., Excitatory amino acids increase glycogen phosphorylase activity in the rat spinal cord. Neurosci. Lett., 73 (1987) 209-214. 37 Woolf, C.J.. Long term alterations in the excitability of the flexion reflex produced by peripheral tissue injury in the chronic decerebrate rat. Pain, 18 (1984) 3255343. 38 Woolf. C.J.. Recent advances in the pathophysiology of acute pain. Br. J. Anaesth.. 63 (1989) 139-146. 39 Woolf. C.J. and King. A.E.. Subthreshold components of the receptive fields of dorsal horn neurones in the rat. J. Neurophysiol.. 62 (2989) 9077916. 40 Woolf. C.J. and King. A.E., Dynamic alterations in the cutaneous mechanoaensitive receptive fields of dorsal horn neurons in the rat spmal cord. J. Neurosci., 10 (1990) 2717-2726. 41 Woolf. C.J. and Wall. P.D.. The relative effectiveness of C primary afferent fibres of different origins in evoking a prolonged facilitation of the flexor reflex in the rat. J. Neurosci.. 6 (1986) 143331443. 42 Woolf. C.J. and Wail. P.D.. Morphine sensitive and morphine msensitive actions of C-fibre input on the rat spinal cord. Neuroser. Lett.. 64 (1986b) 221-225. 43 Yoshimura, M. and Jessell. T.M., Primary afferent evoked synaptic responses and slow potential generation in rat subatantta gelattnosa neurons in vitro, J. Neurophysiol.. 62 (1989) 96-108.
Penn, 44 (1991) 293-299 .I) 1991 Elscvier Science Publishers ADONIS 0304395991~928
B.V. 0304.3959/91/~03.50
PAIN 01746
Basic Section The induction and mainterkmce of central sensitization is dependent on N-methyl-D-aspartic acid receptor activation; implications for the treatment of post-injury pain hypersensitivity states
Depurtmen~
Clifford J. Woolf and Stephen W.N. Thompson ofAnaromv and Deoelopmentoi Biology, Untversity College London, London (Received
2 July 1990, revision
received
28 August
1990, accepted
31 August
WCIE
6BT (U. K.)
1990)
Repetitive stimulation of small diameter primary afferent fibres produces a progressive increase in Summary action potential discharge (windup) and a prolonged increase in the excitability of neurones in the spinal cord following the stimulus. Previous studies have demonstrated that windup is the consequence of the temporal summation of slow synaptic potentials and that the slow potentials and windup are reduced by pretreatment with ~-methyl-D-aspartic acid (NMDA) antagonists. We have now examined whether primary afferent induced hypersensitivity states in flexor motoneurones are also dependent on the activation of NMDA receptors and whether windup is a possible trigger for the production of the central hypersensitivity. Both a non-competitive (MK-801) and a competitive (D-CPP) NMDA antagonist, at doses that did not modify the baseline reflex, reduced the facilitation of the flexor reflex produced by either brief electrical stimulation of the sural nerve (1 Hz for 20 set at C-fibre strength), or by the cutaneous application of the chemical irritant mustard oil. These antagonists also prevented windup from occurring in the motoneurones. When the the MK-801 and the D-CPP were administered once a state of central facilitation had been induced by prior treatment with mustard oil, they returned the facilitated reflex to its pretreatment level. These results indicate that NMDA receptors are involved in the induction and maintenance of the central sensitization produced by high threshold primary afferent inputs. Because central sensitization is likely to contribute to the post-injury pain hypersenstivity states in man, these data have a bearing both on the potential role of NMDA antagonists for pre-emptive analgesia and for treating established pain states. Key words: Post-injury
pain;
Hypersensitivity
states;
Analgesia;
Introduction The hyperalgesia that follows peripheral tissue injury results from an increase in the sensitivity of primary afferent nociceptors in the vicinity of the injury (peripheral sensitization) [1,2,23], and from an increase in the excitability of neurones in the spinal cord (central sensitization) [35,38]. Central sensitization is triggered by nociceptive afferent inputs [30,41] and manifests as a prolonged reduction in the threshold, an expansion of the extent, and an increase in the responsiveness of the
Correspondence IO: Dr. C.J. Woolf, Department of Anatomy and Developmental Biology. University College London, Gower Street, London WClE 6BT. U.K.
NMDA
antagonists;
(Rat)
cutaneous receptive fields of dorsal horn neurones [3,6,14,25,40]. Low threshold afferent inputs have no similar effects [6]. These changes parallel the post-injury hypersensitivity of the flexion withdrawal reflex in animals [35,37], and of pain sensation in man [16]. A distinctive feature of those afferents that can induce central sensitization is their capacity to produce slow synaptic potentials in spinal neurones [26,43]. Following low frequency repetitive activation of the afferents, these slow potentials summate to produce a cumulatively increasing postsynaptic depolarization [26]. This leads to a progressive increase in action potential discharge on repeated stimulation, a phenomenon known as windup [20]. Extracellular studies of dorsal horn neurones in vivo have demonstrated that windup is prevented by administration of NMDA antagonists [7,9].
We have found. using intracellular recordings from flexor motoneurones in vitro, that the prolonged postsynaptic depolarization produced by a single high intensity stimulus, the cumulative summation of these potentials and windup, are all reduced by the NMDA antagonist d-APV [26]. What is not known is whether an NMDA receptordependent windup of spike discharge is a necessary prelude to the generation of central sensitization. In order to investigate this we have examined whether a highly specific competitive NMDA receptor antagonist (D-CPP, 3(( R)-2-carboxypiperazin-4-yl)-propyl-I -phosphonic acid) [5,10] and a non-competitive anatogonist (MK-801, dizocilpine) which binds to the PCP site in the NMDA receptor-gated ion channel [34], can prevent the establishment both of windup and of central sensitization. Both of these drugs block NMDA depolarizations in the rat spinal cord [5,8]. The approach used has been to examine alterations in the excitability of the rat hind limb flexion withdrawal reflex. The flexion reflex in man is linearly correlated with the perception of pain [33] and in animals provides a convenient way of monitoring the processing in the spinal cord of sensory information from high threshold afferent fibres [35]. In the absence of any disturbing stimulus, the discharge of action potentials in hamstring flexor alpha motoneurones elicited by a standard nontissue-damaging noxious mechanical stimulus applied at regular intervals to the toes is very stable in the decerebrate-spinal rat preparation [30,41]. A transient ( < 10 min) facilitation of the reflex can be induced by a brief (20 set) low frequency (1 Hz) train of stimuli to the sural nerve, provided that C-fibre strength stimuli are used [30]. These same stimulus parameters also produce windup [20]. A much longer facilitation can be produced by stimulation of a muscle nerve or by the cutaneous application of the chemical irritant, mustard oil [30,41]. Mustard oil produces an intense burning sensation when applied to the skin in man and a mechanical and thermal hyperalgesia in rats that lasts up to 24 h [37]. The mustard oil produces a burst of activity in C-nociceptors in the rat that decays within a few min [41]. Following its application to the skin, alterations in the receptive field properties of dorsal horn neurones [39] and a facilitation of the flexion reflex [41] occur that last for over 1 h. This chemical irritant provides a useful model to study the prolonged central effects of a brief afferent barrage in chemosensitive fibres.
Methods Experiments were performed on adult (200-300 g) Sprague-Dawley rats. The rats were anaesthetized with a 0.3-0.4 ml bolus injection of Althesin (alphaxalone/ alphadalone, Glaxo) into a tail vein. The carotid artery
was cannulated and anaesthesla maintalncd with >+rn,~ll (0.1 ml) intermittent i.a. doses c>f tht anaesthetlc ~vt~r~ 5- 10 min. The trachea was cannulated and the arumal decerebrated by aspiration of all crania! g:ontenth rostra! to the mesencephalon. A spinaliLatiort ~‘a> performed via a laminectomy at T3 or T4. The anaesthetic ~3s then discontinued, the animals paralyhed with gallamlnc (10 mg) and artificially ventilated. Rectal tsmperature. expired pCOz. heart rate and ECG were monitored and maintained within normal physiological limit+ The nerve to the posterior biceps frnlorls/semrtendinosus muscles was exposed on one hide In the popliteal fossa, dissected free, split into smal! filaments and covered with mineral oil. One of the filaments was placed on a silver recording electrode and CxtracellulaI recordings made of flexor alpha motoneuronr activity as described in detail before [30.41]. In \c)mt’ animals the sural nerve was dissected free and placed on a pair of silver stimulating electrodes. Action potential drscharge was counted using pulse intergrators providing both peristimulus histograms and total counts of the action potential discharge elicited by a standard mechanical stimulus. A standard pinch was applied to the ipsi- and contralateral middle 3 toes using a pair of calibrated forceps (100 g ~~- contact area 1X mm’ ). The standard mechanical stimulus was applied to each foot every 5 min and the total number of spikes elicited counted. Because the number of spikes evoked. although stable for a given animal. varied between animals, the data were transformed as described below to enable comparisons between animals to be made. A baseline value for the reflex, prior to any experimental manipulation, was calculated by determining the mean number of spikes generated by at least 8 successive stimuli at 5 min intervals, commencing 1 h after the spinalization. All individual results for :! particular animal were then calculated as a percentage of this mean baseline value. Mustard oil (allyl-isothiocyanate) 5% in mineral 011 was applied to a 4 mm’ patch of skin on the dorsum of the foot. The MK-801 (Merck, Sharp and Dohme) and D-CPP (Tocris Neuramin) were dissolved in saline and administered by slow (3 min) infusion via the carotid artery
Results Actions of D-CPP and MK-801 on the baseline reflex In the absence of any treatment the standard mechanical stimulus produced a very stable response, measured in terms of the total number of action potentials generated, when repeated at 5 min intervals over prolonged periods as reported previously [30,41]. The MK801 failed to alter the baseline reflex when administered at a dose of 0.1 mg/kg or lower. At doses of 0.5 and 1.0
295
mg/kg a depression of the reflex, commencing within 15 min of treatment, did develop (Fig. 1). The most sensitive aspect of the reflex was the small discharge of action potentials elicited by a contralateral stimulus, but at doses of 1.0 mg/kg or higher a substantial and prolonged depression of the ipsilateral reflex also occurred. The D-CPP also produced a reduction in the baseline reflex at a dose of 5.0 mg/kg (n = 2). This developed over a slower time course than the depression produced by the MK-X01 with peak falls in the action potential discharge manifesting 15-25 min after administration of the drug. At a dose of 2.5 mg/kg, the D-CPP produced no significant modification of the reflex (n = 4).
O.lmgkf’
i 0
I
I
20
10
30
40
50
60
Time (mln) Fig. 1. Alterations in the excitability of the flexion withdrawal reflex produced by the systemic administration of the non-competitive NMDA ion channel blocker MK-801. Excitability is expressed as the percentage deviation from the baseline (predrug treatment) response to a standard (150 g) mechanical stimulus applied for 3 set to the ipsilateral middle 3 toes. For each stimulus the number of spikes elicited in biceps femoris/semitendinosus flexor alpha motoneurones were counted using a pulse integrator. The different doses were administered i.a. at the time indicated by the arrow. Each point represents mean i SE., with n = 6 for 0.1 mg/kg, n = 5 for 0.5 mg/kg and n = 3 for 1 .O mg/kg MK-801.
Windup of motoneurone action potentiul discharge Stimulation of the sural nerve at C-fibre strength (5 mA, 500 psec) at 1 Hz produced a progressive increase in the number of action potentials elicited by the motoneurones per stimulus, over a 20 set period of stimulation. This windup can best be represented as a cumulative spike count which shows the incrementing
125
100
i 600 7.5 500 % 5c
400
2 2 v
300
01 1 :: .: 0 z ;
25 2oc
0 IOC
PRE
0
5
10
CPP 15
20
Time(s) Fig. 2. A spike count illustrating the cumulative increase in the number of action potentials elicited per stimulus in hamstring flexor motoneurones during stimulation of the sural nerve at 1 Hz for 20 set at C-fibre strength (5 mA, 0.5 msec), before (solid lines) and 20 min after 0.5 mg/kg MK-801 (dotted lines). The presence of windup is indicated by the progressive upward deviation of the curve in the control situation, while its absence after MK-801 is indicated by the straight line.
POST
-
PRE
POST
MK- 80t
Fig. 3. Changes in the peak facilitation of the hamstring flexion reflex produced 1 min after a sural nerve conditioning stimulus (1 Hz. 5 mA, 0.5 msec for 20 set) before and after treatment with D-CPP (2.5 mg/kg) and MK-801 (0.5 mg/kg). Increased excitability is expressed as the percentage increase above the baseline (predrug treatment) of the number of spikes generated by a standard mechanical stimulus as described in Fig. 1. Both D-CPP and MK-801 significantly reduced the sural induced facilitation at 1 and at 5 min after the conditioning stimulus (P < 0.01. n = 5 for MK-801 and n = 4 for D-CPP), while saline (0.2 ml) was without effect (n = 2).
125
loo
-75
I 10
’ 0
I 20
1 30
1 40
Twne Fig.
I 50
I 60
I 70
1 80
J
90
( mln)
4. The effect of pretreatment
with 0.1 mg/kg MK-801 and 2.5 mg/kg D-CPP on the excitability Jncrease in flexor motoneurones produced by the cutaneous application of mustard oil to the hairy skin on dorsum of the foot. The solid line shows the excitability increase that occurs Jn the absence of MK-801 or D-CPP (n = 6). The dotted line shows the result achieved by pretreatment with MK-801 (n = 6) I5 min before mustard oil application while the dashed line shows the effect of the D-CPP treatment 30 min before the mustard oil (n = 5). Each point Js the mean + S.E.
response over the period of the stimulation (Fig. 2). Systemic injections of MK-801 (0.5 mg/kg. n = 5) (e.g., Fig. 2) and of D-CPP (2.5 mg/kg. n = 4) prevented the
windup from occurring in the flexor motoneurones. Before MK-801 treatment the action potential discharge per stimulus after 20 set of sural nerve stimulation had
200
175
0 / 0
150
125
100
0
IO
20
I 30
I
MK.801
MO
-75
40 Tune
I 50
I 60
70
I
80
1 90
(mm)
Fig. 5. The reduction of the excitability of the flexion reflex to baseline levels produced by the administration of MK-801 (0.1 mg/kg, n = 5, dotted line) and D-CPP (2.5 mg/kg, n = 5 dashed line). once a state of central sensitization had been induced by the cutaneous application of mustard oil (MO) 30 min earlier. The duration of the facilitation of the reflex produced by mustard oil in the absence of MK-801 or D-CPP is shown by the solid line (n = 6). l indicates P < 0.05 and * * P < 0.01 compared tcr the control facilitation
297
built up from 24 ISI5 spikes for the first stimulus to 39 + 12 spikes for the last (P -C0.05, n = 5). An incrementing discharge failed to develop after MK-801 treatment, after 20 set of sural nerve stimulation the number of spikes evoked by the last stimulus was 22 + 7 spikes compared with 20 f 6 for the first.
The sural nerve stimulation that produces a windup of spikes in the flexor motoneurones of control animals also produces a period of reflex facilitation after the sural nerve stimulation has ceased [30]. This facilitation is maximal 1 min after the conditioning stimulus and reduces to baseline levels within 10 min 1301. Previous studies have shown that the facilitation only occurs when C-fibre strength conditioning stimuli are used [30]. MK-801 and D-CPP pretreatment, in addition to reducing the windup, effectively abolished the sural nerve induced facilitation of the reflex with no facilitation of reflex discharge above baseline at 1 (Fig. 3) or 5 min.
A prolonged period of increased excitability of the flexion reflex is produced by the application of mustard oil to the dorsum of the foot 1411 {Fig. 4). Pretreatment with 0.1 mg/kg MK-801 or with 2.5 mg/kg D-CPP prevented the mustard oil from producing this facilitation (Fig. 4). Because the depressant effects of the D-CPP on the baseline reflex at 5.0 mg/kg developed at a slower rate than that produced by the high doses of MK-801. the D-CPP was administered 30 min before the mustard oil as opposed to 15 min for the MK-801 (Fig, 4). Saline (n = 2) had no effect on the facilitation. When the MK-801 (0.1 mg/kg) and the D-CPP (2.5 mg/kg) were given 30 min after the development of a mustard oil induced f~~cilitation of the flexion reflex, the elevated excitability was significantly reduced, within 5 min for the MK-801 and within 15 min for the D-CPP (Fig. 5). In both cases the reflex returned to or below the baseline level of the reflex. A similar effect could be demonstrated for the prolonged facilitation that follows a 20 set C-fibre strength stimulation of the nerve to the gastrocnemius muscles. Such stimulation, in contrast to sural nerve stimulation, produces a prolonged facilitation of the flexor reflex [30]. The administration of 0.1 mg/kg MK-801 20 min after the induction of such a facilitation returned the reflex to baseline levels (n = 2).
Discussion The activation of small diameter primary afferent neurones is capable of producing a prolonged alteration in the excitability of neurones in the spinal cord, modifying the way these neurones respond to subsequent
inputs and thereby generating a state of central sensitization [35,373. These changes can be produced by electrical stimulation of peripheral nerves. with muscle afferents producing longer lasting effects than cutaneous afferents [30,41], by peripheral injury or localized inflammatory lesions [14,24.35], and by activating chemosenitive afferents with chemical irritants such as mustard oil 111,411 or capsaicin 1251. What is the importance of central sensitization? We have argued previously that while the capacity to detect and react to non-tissue-damaging noxious stimuli is the consequence of the activation of functionally specified high threshold nociceptors, which generate “physiological” pain, clinical or “pathological” pain is in contrast, the consequence of the induction of peripheral and central sensitization in the somatosensory system [38]. The changes represented by central sensitization include a situation where low threshold afferents begin to produce pain. something they never normally do, and this helps to explain why low intensity or innocuous stimuli are painful, contributing to the allodynia that occurs in the vicinity of a peripheral injury. Such a change in the sensory responses to low threshold mechanoreceptors has recently been demonstrated in human subjects in whom a state of secondary mechanical hyperalgesia was induced by the administration of capsaicin to the skin [28]. The mechanism involved seems to be the recruitment of low threshold mechanoreceptor afferent input to cells which prior to the production of central sensitization had high threshold receptive fields [39,403. The ability of NMDA receptor antag(~nists to block the development and maintenance of central sensitization implies that it should be possible to differentiate pharmacologically between physiological and pathological pain states in the sense that NMDA antagonists, by returning an abnormally exctible spinal cord to its normal level of excitability, will return a pathological situation where low and high intensity stimuli produce pain to one where there is a clear distinction between noxious and innocuous stimuli in terms of the different sensations they elicit. The clinical significance of central sensitization lies in the implicit prediction that can be made that preemptive analgesia, by preventing the establishment of central sensitization. will reduce the pain experienced following a peripheral injury. In other words if the afferent input that produces central sensitization is prevented from gaining access to the CNS or if its central effects are prevented, then less pain will result. The easiest way of testing this clinically is to use local anaesthetics prior to a surgical intervention. Such an approach has been found to reduce postoperative pain compared to patients who received a general anaesthetic without local block at the site of the surgery [29]. An experimental demonstration of preemptive “analgesia” using opioids. has shown that low doses of morphine
29x
prevented the development of an afferent induced reflex hypersensitivity but were ineffective in reducing the reflex ilypersensitivity once it was established [42]. Doses of morphine an order of magnitude higher were required to reduce the reflex facilitation to baseline levels [42]. This finding may relate to the reports of the effect of opioid premeditation on postoperative pain [17]. The present data show that NMDA antagonists differ from morphine in that the same dose of the NMDA antagonist that is required to prevent the establishment of central sensitization also abolishes it once it has been induced. This may mean that both morphine and an NMDA antagonist such as ketamine could be useful as pre-emptive analgesic agents but that NMDA antagonists may be more efficacious in reducing established central sensitization. These data also draw attention to the fact that a drug may have analgesic activity in terms of reducing a pathological post-injury pain hypersensitivity state, while leaving the normal physiological response to noxious stimuli intact. Earlier suggestions that the NMDA receptor may be involved in nociception relate to the effects of a variety of NMDA antagonists on behavioural tests [4] and on the responses of dorsal horn cells to peripheral noxious stimuli [32]. It is likely though that these studies were looking more at the suppressant effects of high doses of NMDA antagonists than on a blockade of central sensitization. Indeed conventional screening for analgesic drugs is designed to examine ‘physiological pain’ and not pathological hypersensitivity states. More relevant is the report of the reduction of postoperative pain produced by ketamine in man [18], and the hyperalgesia that is generated in laboratory animals by intrathecal injections of NMDA [31]. The cellular and molecular mechanisms underlying the contribution of the NMDA receptor-ion channel complex to the induction and maintenance of central sensitization are not known. The finding that windup is also reduced by NMDA antagonists suggests that the cumulative depolarization that results from the summation of slow potentials [26] is an important first step in the induction of central sensitization. The reason for this is likely to be the voltage-dependent block of the NMDA ion channel by Mg2+ [19,22]. A depolarization generated by the summation of EPSPs, by removing or decreasing this Mg2+ block, will increase inward currents through the ion channel amplifying the depolarization and increasing synaptic efficacy. Neuropeptides, particularly the neurokinins which have the capacity to produce prolonged depolarizations [21] and appear to be able to act as volume transmitters over wide areas [12], may be important in recruiting an NMDA component to any postsynaptic response. The maintenance of the state of central sensitization by the NMDA receptor--ion channel complex could be the result of uncovering a new polysynaptic pathway
involving NMDA receptors or a tonic. rclcabc :>f endtrg~nous transmitter from terminals producing a tom& activation of NMDA receptors. It is Intriguing to consider the possible role of changes in s<:cond messengers in mediating the prolonged effects of NMDA receptor activation. The same afferents that induce central sensetization can increase protein phosphorvinttr\rt in the dorsal horn in an N‘~DA-depeildeIlI fashjon 1361 and can initiate the expression of the immediate-cariy gene product c-Fos 1131. although whether this change in gene expression can be modified by NMI)A antagonist5 is controversial [15,27]. In conclusion, the demonstration that ;t competitive and a non-competitive NMDA receptor antagonist can prevent the induction of central sensitization and abolish the hypersensitivity once it is established, suggests that this class of drugs has potential both for the prevention and treatment of acute pain. The data alao reinforce the view that the pathogenesis of pain is in part a reftection of the capacity of central neurones to undergo dynamic aiterations in their response properties ;Ind rhe fact that nociceptive afferents can trigger such LLfuncttonal plrtaticity [401.
Acknowledgements
We thank the MRC and the Wellcome Trust for financial support. Merck. Sharp and Dohme for the supply of the MK-801 and J. Middleton for technical assistance.
References Ressou. P. and Perl. E.R., with unmyeiinated (1969)
Response
of cutaneous
fihres to noxious stimuli.
wnsory
units
.I. Ncurophysiol..
??.
1025-1043.
Campbell.
J.N..
myelinated
Neurophysiol.. C‘ervero.
F..
noxtous
Meyer,
and LaMotte.
afferents
42 (1979) of
R.H..
that innervate
oi
Sensttlzation monkey
hand
J.
1669- 1679.
Handwerker.
stimulation
properties
R.A.
nociceptive
H.O. the
and
rat’s
Lalrd.
tail:
J.M..A..
responses
Prolonged
and
of dorsal horn neurons. J. Physwl. (Land.).
encoding 404 (1988)
419.-436. C‘ahusac, Srmth.
A.M..
D.A.%
rweptor
Evans.
R.H..
Hill.
The hehavioural
antagonist
following
R.G..
Rodriguez.
application
IO the lumbar
cord of conscious rats. Neurophannacology. Childs,
A.M..
Evans. R.H.
and Watkins,
cai selectwity of three NMDA (1988)
18.-86.
(‘wk.
A.J.. Woolf.
rcceptlve primary
C‘.J.. Wall.
field plasticity afferent
receptors
23 (1984)
dorsal
325 (1987)
in “wind-up”
IS1
for Involvement
J.A
J.D.. Aram.
C-
of N-meth-
of class 2 neurones
Dawes.
D.. Millar.
following
153
402-406
S.N.. Martm.
145
S.B., Dynamic
horn
dorsal horn of the rat. Brain Res.. 424 (1987) I.odpc. I>.. IXffcrcnces
spinal
719- 724
fiur. .f. Pharmacol..
P.D. and McMahon.
in rat spinal
Dawes. S.N. and Lodge, D., Evidence ylaspartate
and
J.C.. The pharmacologt-
antagonists.
Input. Nature.
R.E.
effects of an hi-methylaspartate
. C‘hurch.
tn the J. and
tin results from in viva and In v!tr
on the use dependency of N-methylaspartate antagonism by MK801 and other phencyclidine receptor ligands. Eur. J. Pharmacol., 145 (1988) 141-151. 9 Dickenson. A.H. and Sullivan, A.F., Evidence for a role of the NMDA receptor in the frequency dependent potentiation of deep dorsal horn neurones following C-fihre stimulation. Neuropharmacology, 26 (1987) 1235-1238. IO Harris, E.W., Ganong. A.M.. Monoghan. D.T., Watkins, J.C. and Cotman. C.W.. Action of 3-(( k )-2-carboxypiperazin-4-yl)-propyll-ph(~sphoric acid (CPP) a new and highly potent antagonist of N-methyl-D-aspartate receptors. Brain Res., 382 (1986) 174-182. 11 Hoheiael, U. and Mense. S.. Long-term changes in discharge behaviour of cat dorsal horn neurones following noxious stimulation of deep tissues. Pain, 36 (1989) 2399247. 12 Hope. P.J., Schaible, H.G.. Jarrott, B. and Duggan, A.W., Release and persistence of imnlun~~reactive neurokinin A in the spinal cord is associated with chemical arthritis. Pain, Suppl. 5 (1990) S230. 13 Hunt, S.P.. Pim. A. and Evan, G.. Induction of C-fos-like protein in spinal cord neurones following sensory stimulation. Nature, 328 (1987) 6322634. 14 Hylden, J.L.K., Nahin. R.L., Trauh. R.J. and Dubner. R., Expansion of receptive fields of spinal lamina I projection neurons in rats with unilateral adjuvant-Induced Inflammation: the contribunon of dorsal horn mechanisms. Pain, 37 (1989) 229-243. 15 Kehl, L.J., Basbaum, A.I., Pollock, C.H., Mayes, M. and Wilcox, G.L.. The NMDA receptor antagomst MKXOl reduces noxious stimulus-evoked FOS expression in the mammalian spinal dorsal horn. Pain. Suppl. 5 (1990) S165. 16 LaMotte, R.H., Thalhammer, J.G. and Robinson, C.J.. Peripheral neural correlates of magnitude of cutaneous pain and hyperalgesia: a comparison of neural events in monkey with sensory Judgements in human. J. Neurophysiol., 50 (1983) l-26. 17 McQuay. H.J.. Carroll, D. and Moore, R.A., Postoperative orthothe effect of opiate premeditation and local paedic paln anaesthetic blocks. Pain. 33 (19X8) 291-295. 18 Maurset. A.. Skoglund. L.A., Hustveit, 0. and Oge, I., Comparison of ketamine and pethidine in experimental and postoperative pain. Pain, 36 (1989) 37-42. I9 Mayer. M.L.. Westbrook, G. and Guthrie. P.B.. Voltage-dependent block by Mg2+ of NMDA responses in spmal cord neurones. Nature, 309 (19X4) 261-263. 20 Mendell, L.M.. Modifiability of spinal synapses. Physiol. Rev., 64 ( 1984) 260- 324. 21 Murase, K. and Rand& M., Actrons of substance P on rat spinal cord and dorsal horn neurones. J. Physiol. (Lond.), 346 (1984) 203-217. 22 Nowak, L., Bregestovjski, P., Ascher, P., Herbert. A. and Prochianty, A.. Magnesium gates glutamate activated channels in mouse central neuronea. Nature, 307 (1984) 462-465. 23 RaJa. S.. Campbell. J.N. and Meyer, R.A., Evidence for different mechanisms of primary and secondary hyperalgesia following heat injury to the glabrous skin. Brain. 107 (1984) 1179-1188. 24 Schaible. H.G.. Schmidt, R.F. and Willis. W.D.. Enhancement of the responses of ascending tract cells in the cat spinal cord by acute inflammation of the knee joint. Exp. Brain Res.. 66 (1987) 489-499. 25 Simone. D.A.. Snrkin. L.S., Chung. J.M.. Willis, W.D. and
LaMotte. R.H., Spinothalamic tract cells hyperalgesia following rntradetmal capsaicin.
signal the pain and Sot. Neurosci. Ahst.,
14 (1988) 559. 26 Thompson. S.W.N., King, A.E. and Woolf. C.J.. Activity-dependent changes in rat ventral horn neurones in vitro; summation of prolonged afferent evoked postsynaptic depolarizations produce a d-APV sensitive windup. Eur. J. Neurosci., 2 (1990) 6388649. 27 Tolle, T.R.. Castro-Zopes, J.M.. Coimbra, A. and Zieglglnsberger, W.. C-Fos induction in the spinal cord following noxious stimulation: prevention by opiates but not NMDA antagonists. Pain. Suppl. 5 (1990) 5229. 28 TorebJiirk. E.. Lundberg. L. and LaMotte. R.. Neural mechanisms for capsaicin-induced hyperalgesia. Pain, Suppl. 5 (1990) S114. 29 Tverskoy, M.. Cozacov. C., Ayache, M.. Bradley. E.L. and Kissin. 1.. Postoperative pain after inguinal herniorrhaphy with different types of anesthesta. Anesth. Analp., 70 (79%)) 29-35. 30 Wall. P.D. and Woolf. C.J., Muscle but not cutaneous C-afferent input produces prolonged increases rn excitabiiity of the flexion reflex tn the rat. J. Physiol. (Land.). 356 (1984) 44-458. 31 Wilcox, G.L. and Aansen. L.M.. Nocicepttve action of excitatory amino acids. J. Pharmacol. Exp. Ther.. 243 (1987) 9-19. 32 Wilcockson. W.S., Chung. J.M.. Hon. Y.. Lee. K.H. and Willis. W.D., Effects of i~~nt(~phoretically released amino acids and amines on primate spinothal~til~ic tract neurons. J. Neurosci.. 4 (1984) 732-740. 33 Wilier. J.C.. Comparative study of perceived pain and nociceptive flexion reflex in man. Pain. 3 (1979) 69-80. 34 Wang. E.H.F.. Kemp. J.A.. Priestly, T.. Knight. A.R., Woodruff. G.N. and Irvine, L.L., The anticonvulsant MK801 is a potent W-methyl-n-aspartate antagomst. Proc. Nat. Acad. Sci. (U.S.A.). 83 (1986) 7104-7108. 35 Woolf. C.J.. Evidence for a central component of postinjury pain hypersensitivity. Nature, 308 (1983) 686-688. 36 Woolf. C.J., Excitatory amino acids increase glycogen phosphorylase activity in the rat spinal cord. Neurosci. Lett., 73 (1987) 209-214. 37 Woolf, C.J.. Long term alterations in the excitability of the flexion reflex produced by peripheral tissue injury in the chronic decerebrate rat. Pain, 18 (1984) 3255343. 38 Woolf. C.J.. Recent advances in the pathophysiology of acute pain. Br. J. Anaesth.. 63 (1989) 139-146. 39 Woolf. C.J. and King. A.E.. Subthreshold components of the receptive fields of dorsal horn neurones in the rat. J. Neurophysiol.. 62 (2989) 9077916. 40 Woolf. C.J. and King. A.E., Dynamic alterations in the cutaneous mechanoaensitive receptive fields of dorsal horn neurons in the rat spmal cord. J. Neurosci., 10 (1990) 2717-2726. 41 Woolf. C.J. and Wall. P.D.. The relative effectiveness of C primary afferent fibres of different origins in evoking a prolonged facilitation of the flexor reflex in the rat. J. Neurosci.. 6 (1986) 143331443. 42 Woolf. C.J. and Wail. P.D.. Morphine sensitive and morphine msensitive actions of C-fibre input on the rat spinal cord. Neuroser. Lett.. 64 (1986b) 221-225. 43 Yoshimura, M. and Jessell. T.M., Primary afferent evoked synaptic responses and slow potential generation in rat subatantta gelattnosa neurons in vitro, J. Neurophysiol.. 62 (1989) 96-108.