311

Pain, 40 (1990) 311-322 Elsevier

PAIN 01558

Basic Section Differential responses of nociceptive vs. non-nociceptive spinal dorsal horn neurones to cutaneously applied vibration in the cat M.W. Salter *J and J.L. Henry Departments

of * Physiology,

* Research in Annesthesia (Received

and * * Psychiatty,

25 July 1989, accepted

*,** McGill University, Montreal,

25 October

Que. (Canada)

1989)

Extracellular single-unit recordings were made from dorsal horn neurones in the lumbar spinal cord of cats which were anaesthetized or were anaemically decerebrated. Each neurone was classified functionally as wide dynamic range (WDR), non-nociceptive, nociceptive specific or proprioceptive. Vibration was then applied to the hind limb using a feedback-controlled mechanical stimulator. WDR neurones had 3 distinct types of response to vibration (80 Hz; 0.3-1.0 mm): excitation, depression and a biphasic response consisting of excitation followed by depression [16]. The type of response depended upon the location of the stimulator probe. With the stimulator probe placed inside that part of the receptive field from which low intensity, non-vibrational cutaneous stimuli elicited excitation, 35 neurones were excited by the vibratory stimulation, none was depressed and 4 showed the biphasic response. On the other hand, when the probe was positioned outside the receptive field for low intensity stimuli, 7 WDR neurones were excited, 164 showed depression or the biphasic response and 7 were unaffected. On-going activity and activity evoked by iontophoretic application of glutamate were decreased during the depressant response and during the depressant phase of the biphasic response. In terms of non-nociceptive neurones, all (n = 30) were excited by vibration; depressant or biphasic responses were not observed. Excitation was elicited by placing the probe either inside or outside the receptive field for non-vibrational stimuli. All nociceptive specific neurones (n = 3) were depressed by vibration regardless of the position of the stimulus. All proprioceptive neurones (n = 12) were excited by vibration. The predominantly depressant effect of vibration on nociceptive neurones vs. the predominantly excitatory effect on non-nociceptive neurones prompts us to suggest that the increase in pain threshold and the clinical analgesia elicited by vibration may be mediated at the spinal level by a decrease in the rate of firing of nociceptive neurones and/or by excitation of non-nociceptive neurones. Summary

Key words:

Spinal dorsal

horn;

Afferent

inhibition;

Vibration;

Introduction Vibration applied to the skin has been found to relieve pain in both experimental [1,4,9,17,18] and

’ Present address: Playfair Neuroscience Unit, The Toronto Hospital, 399 Bathurst St., Toronto, Ont. M5T 2S8, Canada. Correspondence to: Dr. J.L. Henry, Department of Physiology, McIntyre Medical Sciences Building, 3655 Drummond Street, Montreal, Que. H3G lY6, Canada. 0304-3959/90/$03.50

0 1990 Elsevier Science Publishers

Somatosensory

system;

Antinociception

clinical [5-81 investigations. The use of vibration for pain relief stemmed from the pioneering electrophysiological study of Wall and Cronly-Dillon [18] who examined the vibration-induced responses of dorsal horn neurones which the authors noted were excited by all types of cutaneous mechanical stimulus; these neurones would thus be classified as wide dynamic range (WDR) neurones [lo]. The principal observation of Wall and Cronly-Dillon was that with each vibratory cycle these neurones responded with a brief period of

B.V. (Biomedical

Division)

312

excitation followed by depression of firing; in only a few cases was depression observed without excitation. Recently, we have extended these initial findings by examining the physiological [16] and pharmacological [14,15] characteristics of the vibration-induced responses of WDR neurones. We have found that WDR neurones show 3 distinct types of response to vibration: excitation, depression, and a biphasic response consisting of excitation followed by depression [16]. For the biphasic response, the excitation occurred only at the beginning of the period of vibration and the depression of firing persisted throughout the remainder of stimulation. Hence, this response is distinct from the phasic excitation/depression seen by Wall and Cronly-Dillon. We have also found by varying the site at which the vibration is applied that a given WDR neurone usually exhibits more than one type of response: the excitatory response is evoked when the vibration is applied inside the cutaneous receptive fields for other types of innocuous mechanical stimuli, and the depressant or biphasic response is elicited when the site of vibration is outside these receptive fields. The purpose of the present investigation is to extend earlier studies to include non-nociceptive neurones and to characterize and compare the vibration-induced responses of each of the different functionally identified types of neurone in the dorsal horn. Preliminary results of the present study have been reported previously [13].

Methods Animal preparation The experimental methodology has been described in detail previously [14,16]. In brief, experiments were done on adult cats (n = 57): 55 were anaesthetized with sodium pentobarbital (40 mg/kg, i.p.; supplemental doses 5 mg/kg i.v. every 3 h) and 2 were anaemically decerebrated under anaesthesia induced by halothane (Somnothane, Hoechst) which was subsequently discontinued. Spinal segments L5-L7 were exposed for recording and in all experiments the spinal

cord was transected at the first lumbar level to remove descending influences and to eliminate the possibility that the vibration-induced responses might be mediated via supraspinal structures. The animals were paralyzed with pancuronium bromide (Pavulon, Organon; 1 mg/kg i.v., repeated when required) and ventilated artificially. End-tidal CC& concentration and arterial blood pressure were monitored continuously during the experiments and were maintained within normal limits.

Recording and data acquisition Single-unit spikes were recorded extracellularly with multi-barrelled glass micropipettes (overall tip diameter 4-8 pm); the central recording barrel contained 2.7 M NaCl (impedance 4- 10 Ma). The raw data were amplified, displayed on oscilloscopes and recorded on magnetic cassette tape. Intervals between spikes and between stimuli were computed to +0.5 msec and stored on hard disk using an IBM personal computer and hardware and software developed in our laboratory. One of the outer barrels of the micropipette contained a solution of sodium L-glutamate (1 M. pH 7.4, Sigma). Only units excited by iontophoretic application of glutamate were studied in order to eliminate recordings from fibres. The excitability of selected neurones was increased by application of glutamate in order to investigate the effect of the level of excitability on the response to vibration.

Functional classification of neurones Each neurone was classified on a functional basis according to its responses to natural peripheral stimulation as described in detail elsewhere [12]. The following stimuli were used: movement of single hairs, light touch of the skin with a tissue paper, firm manual pressure (judged to be nonnoxious when applied to the experimenter), noxious pinch of the skin with a serrated forceps, noxious heating of the skin (temperature > 47”C), and passive movement of the limb. Then the neurone was classified as non-nociceptive, wide dynamic range, nociceptive specific or proprioceptive. Importantly, the response to vibration was not a criterion used in the classification.

313

Vibrational stimulation Vibration was generated by a feedback-controlled mechanical stimulator [3] to which was attached the plexiglass stimulator probe which was 1 cm in diameter. The probe was placed in contact with the skin and a constant force of up to 500 mN was applied during the periods between vibration. Care was taken to avoid positioning the probe directly over major peripheral nerves. The mechanical stimulator was driven using trains of rectangular voltage pulses (6 msec) produced by a Grass S88 electrical stimulator. The trains of pulses about 3 set in duration were given periodically at regular intervals every 20-25 sec. The frequency of the pulses within the trains was 80 Hz; this was the maximum frequency at which the probe returned completely to the rest position after each pulse and was also the frequency used in our previous neuropharmacological investigations [14,15]. The peak-to-peak displacement of the probe was 0.3-1.0 mm and was monitored during the testing to ensure that the amplitude of the vibration was consistent. The initial motion of the probe was always towards the limb. It is important to note that the term ‘receptive field’ will be used only with respect to non-vibrational stimuli. This specific use is made because receptive fields for non-vibrational stimuli are readily determined and are the convention for defining the properties of dorsal horn neurones in electrophysiological studies. These receptive fields were used as references when examining the effects of varying the position of the stimulator probe. Histological verification of sites of recording Recording sites were marked by passing inward current (5-10 PA) for 5-10 min through another barrel which contained the Pontamine Sky Blue 6BX (0.5% in 0.5 M sodium acetate, Gurr). At the end of the experiment the section of spinal cord containing the deposits was removed, fixed and embedded. The location of the deposits was determined by microscopic examination of serial sections (15-30 pm) which had been counterstained with thionin. Analysis Data

of data were analyzed

off-line.

For each testing

site, the precise classification of the response to vibration was done on the basis of a peristimulus time histogram (PSTH) which showed the average response to at least 10 consecutive stimulation trials.

Results Recordings from 238 neurones are included in these results. No difference was apparent in the effects of vibration on neurones recorded in anaesthetized (n = 220 neurones) compared with unanaesthetized (n = 18 neurones) animals, and hence data from these 2 preparations are considered together. Thirty neurones were classified as non-nociceptive, 193 as WDR, 3 as nociceptive specific and 12 as proprioceptive. The response of a given neurone to vibration at any one site was stable, varying in average magnitude less than lo%,, over periods longer than 4 h. Responses of WDR neurones to vibration As mentioned in the Introduction we have previously characterized 3 distinct types of response of WDR neurones. The results illustrated in the present report are taken from the same population as reported in our other study [16], and the data are included for purposes of comparison with the vibration-induced responses of the other functionally-identified types of neurones. The type of response of a WDR neurone and the response magnitude depended upon the location of the stimulation probe. For example the neurone shown in Fig. 1 had a weak depressant response when the probe was in position ‘a’ but stimulation with the same amplitude evoked profound depression in the rate of firing when the probe was moved to position ‘ b’; vibration caused excitation when the probe was placed inside the receptive field for low threshold stimulation (not shown). Due to the variability in the baseline rate of discharge which is apparent in Fig. 1 the classification of the type of response was done by examining PSTHs such as shown in Fig. 2. In this case vibration caused a brief excitation ( < 50 msec) followed by a prolonged depression ( > 2.5 set) of the discharge which was typical of the biphasic response (Fig. 2A).

314

a b

b

Fig. 1. Effects of vibration on a WDR neurone. Each film record shows a single sweep of the oscilloscope; vibration was applied during the period indicated by the bar above each record. The receptive fields for mechanical stimuli are illustrated in the diagram to the right. The neurone was excited by low intensity mechanical stimuli (hair movement or light touch) applied to the blackened area and by noxious pinching in both the blackened and hatched areas. The probe of the stimulator was applied in 2 different positions; the arrow in each of the diagrams and in all subsequent figures indicates the position of the centre of the probe and, in addition, the orientation of its long axis. The upper record was taken with the probe in position ‘a’ and the lower record with the probe in position ‘b.’ The stimulation amplitude was 0.3 mm. Calibration bars: vertical 250 uV, horizontal 3 sec.

Control experiments. To differentiate between effects of vibration and effects of displacement of the skin per se, a single long mechanical pulse was applied as illustrated in Fig. 2B. This long pulse was of the same duration and the same amplitude as the trains of pulses used to produce vibration. For the long pulse, the excitation at the onset was followed by depression which lasted less than 200 msec at the onset of the pulse and there was a shorter period of depression following the end of the pulse. On the other hand, for the depressant component of the biphasic response the decrease in firing rate persisted throughout the period of stimulation. Fig. 2 also provides evidence against the possibility that the responses were artefacts due to vibration: the biphasic response caused by vibration when the probe was positioned on the hind

limb ipsilateral to the recording site (Fig. 2A) was not mimicked when the probe was placed on a homologous region of the contralateral hind limb (Fig. 2C). Vibration causes depression of activity evoked bjl glutamate or by cutaneous stimuli. Iontophoretic application was used to evoke activity in silent neurones (Fig. 3) or to increase the activity of spontaneously firing neurones. In both instances this activity could be blocked during the vibration-induced depression (Fig. 3B and 3C) or during the depressant component of the biphasic response. These results suggest that the depression is produced centrally, possibly by a postsynaptic mechanism. The effects of vibration on responses to other types of cutaneous stimulus were generally not examined to avoid the possibility that vibration was causing alterations in the periphery. However, in each of the several cases tested it was clear that vibration caused depression of the responses to noxious stimuli, such as noxious radiant heat or noxious pinching, and to non-noxious stimuli such as light touch. Occurrence of the 3 types of response. Vibration applied inside the receptive fields for low threshold stimuli was tested on 39 neurones: 35 were excited and 4 showed the biphasic response. Vibration applied outside the low threshold receptive fields for low threshold stimuli was tested on 178 neurones. When the probe was placed near these receptive fields, vibration generally caused excitation and for 7 neurones vibration caused excitation at all sites of stimulation examined. However, for 164 neurones one or more sites of stimulation were found from which either the depressant or the biphasic response was evoked. These sites were typically far from the low threshold receptive fields but the distance required to elicit these responses varied greatly from neurone to neurone (for examples see Fig. 2B in Salter and Henry [16]). In the remaining 7 cases only a few sites far from the low threshold receptive fields were examined and vibration had no effect. Responses of non-nociceptive neurones to vibration Vibration-induced excitation. Non-nociceptive neurones were tested using vibration in a similar

315

Fig. 2. Effects of vibration on a WDR neurone in lamina IV are not mimicked by single long pulses or by vibration applied to the contralateral hind limb. Three PSTHs are shown (bin width: 50 msec). Each histogram was constructed using 15 trials. The bar below the histogram in A indicates the period of vibration with the stimulator probe in the position shown by the arrow in the diagram to the right. In B, the line below the histogram indicates the duration of a single long mechanical pulse applied at the same position as in A. For C, the probe was moved to a similar position on the contralateral hind limb and vibration was applied for the period indicated by the bar below the record. In all cases the stimulation amplitude was 0.5 mm. The blackened and hatched areas in the diagram on the upper right have the same significance as in Fig. 1. The location of the dye deposit at the site of recording can be seen in lamina IV in the photomicrograph.

way as were WDR neurones. Vibration only excited the non-nociceptive neurones (Fig. 4). Excitation was evoked when the stimulator probe was placed inside (Fig. 4C) or outside (Fig. 4D) the excitatory receptive fields defined for low threshold non-vibrational stimuli. The magnitude of excitation was typically greater when the probe was placed inside the receptive field. By placing the probe at increasing distances from the low threshold receptive fields it was found that there was a graded decrease in the magnitude of excitation (not shown). Stimulation of the contralateral hind limb had no effect. In no case could depressant or biphasic responses be evoked from any site either inside (12 neurones tested) or outside (21 neurones tested) the receptive fields. Amplitude-response relationships. For WDR neurones higher amplitudes of stimulation were more effective in causing depressant and biphasic responses as reported earlier [16]. Therefore, amplitude-response relationships were determined

for 5 non-nociceptive neurones (Fig. 5). At low amplitudes vibration caused only a burst (Fig. 5A) at the onset of stimulation (latency 8-12 msec). With higher amplitudes the firing rate was increased throughout the remainder of the vibration. In all cases, the magnitude of the excitation increased monotonically with the stimulation amplitude (Fig. 5B); neither depressant nor biphasic responses were evoked even at the highest amplitudes.

Effect oj increasing level of excitability by application of glutamate. In cases where vibration evoked only a burst at the onset of stimulation (for example the upper trace in Fig. 6A), it was considered possible that vibration may have been causing a biphasic response similar to that observed with WDR neurones except that this possible biphasic response was unseen because the ongoing rate of discharge was zero. In such cases it was found that application of glutamate increased the total number of spikes during the vibration-in-

316

*Or

C

Fig. 3. Activity evoked by glutamate is blocked by vibration. A: continuous time histogram showing the rate of discharge of thts neurone. The receptive field was stimulated by the movement of single hairs (H), by touching the skin (T) and by firm manual pressure (Pr). At the times indicated by the arrows, noxious pinch (Pi) was applied briefly to the receptive field; in each case the duration of the pinch was less than 1 sec. Note the prolonged afterdischarge following the pinch stimulus; this afterdischarge characterizes the response of a nociceptive neurone to noxious stimulation. Glutamate (Glu) was applied by iontophoresis during the period indicated by the long line below the record. The ejection current indicated is in nA. B: film record showing a single response to vibration during application of glutamate. The stimulator probe was applied at the position shown in D. The amplitude of stimulation was 0.5 mm. Calibration bars: vertical 150 /.LV; horizontal 3 sec. C: the continuous time histogram begins 8 min after the one shown in A; the vertical and time scales in C are the same as in A. The period of glutamate application is shown by the line below the record. At the time indicated by the arrow the ejection current was decreased from 35 to 30 nA. Vibration applications (V) are indicated by the short bars below the histogram. D: Diagram illustrating the receptive fields. The blackened and hatched areas have the same significance as in Fig. 1. The ‘x’ shows the position of the stimulator probe and indicates that it was oriented perpendicular to the foot.

Ii

C

A

T

Pr

‘Pi T

D

ILi

Heat

b

E

ba

Fig. 4. Excitation of a non-nociceptive neurone by vibration. The rate of discharge is shown in the continuous time histogram in A. The receptive field was stimulated by hair movement (H), repeated touching with a tissue paper (T) and firm pressure (Pr). At the time indicated by the arrow, the receptive field was pinched (Pi) briefly; note that while there was a brief excitation an afterdischarge was not observed and therefore the neurone was classified as non-nociceptive. The open squares below the record illustrate the period of application of noxious radiant heat to the receptive field. The receptive field is shown by the blackened area on the second toe from the left in the diagram in B. The arrows indicate the 2 positions where the stimulator probe was located. The film records in C and D show single vibration-induced responses with the probe in positions ‘a’ and ‘b,’ respectively. The average number of spikes during the period of vibration was 12.4 + 3 (10 trials) when the probe was applied outside the receptive field and was 118 + 7 (10 trials) when the probe was inside the receptive field (I = lo- ‘, P = IO-‘). With the probe in position ‘b’ a single long pulse was applied as indicated by the line above the trace in E. In C. D and E the stimulation amplitude was 1 .O mm: the calibration bars are 100 PV and 3 sec.

317 50.

A

40.

30. spikes during vibration 20.

lo-

0’

10.4

0:s Amplitude

016

0

1.0

Amplitude (mm)

10

(mm)

Fig. 5. Amplitude-response relationships for non-nociceptive neurones. The amplitude-response curve for a single non-nociceptive neurone is shown in the graph in A: the dots indicate the average number of spikes for 10 trials using stimulation amplitudes of 0.4-1.0 mm. The line is the least-squares regression line for the data points and the equation is y = 1.96x - 26.5. The error bars indicate + 1 S.E.M. The film records show examples of responses for stimulation amplitudes of 0.4,0.6 and 1.0 mm. Calibration bars: vertical 200 uV; horizontal 3 sec. The receptive field is illustrated by the blackened area in the diagram and the position of the stimulator probe is shown by the arrow. Amplitude-response curves for 5 non-nociceptive neurones are shown in the graph in B. In contrast to A, the data points are joined by line segments. The points on the y-axis indicate the average number of spikes during the control period.

duced excitation even though glutamate itself failed to cause the neurone to discharge (Fig. 6). As illustrated in Fig. 6A, when glutamate was applied vibration caused firing in the period fol-

lowing the initial burst. In no case did increasing the level of excitability reveal a biphasic response to vibration of non-nociceptive neurones.

6

ill 0

0

C

Fig. 6. Effects of increasing the level of excitability on vibration-induced excitation of a non-nociceptive neurone. Each of the 3 PSTHs in A was constructed using a series of 10 consecutive applications of vibration (amplitude of stimulation: 1.0 mm). The period of vibration is indicated by the bar below each record and the bin width for the PSTHs is 50 msec. The ejection current, in nA, for glutamate (Glu) is indicated on the right in each PSTH. To allow for stabilization, in each case the ejection began at least 20 set before the first application of vibration in the series. Using ejection currents greater than 75 nA an on-going rate discharge was elicited (not shown). The receptive field and location of the stimulator probe are illustrated in B. The film record in C shows an example of a vibration-induced response during ejection of glutamate using a current of 75 nA. Calibration bars: vertical 150 PV; horizontal 3 sec.

31X

TABLE

I

RESPONSES TO VIBRATION APPLIED OUTSIDE THE LOW THRESHOLD RECEPTIVE FIELD OF WDR VS. NON-NOCICEPTIVE (N-N) NEURONES

WDR N-N

Excitation

Depression

7 21

164 0

x’ = 143. with 2 degrees

of freedom.

or biphasic

No effect I 0

P i 0.005.

Comparison of responses of WDR vs. non-nociceptive neurones All of the non-nociceptive neurones and 90% of the WDR neurones were excited by vibration ap-

plied to the low threshold excitatory receptive fields. On the other hand, when the probe was located outside the receptive field vibration had a differential effect on neurones in the two functional groups: non-nociceptive neurones were excited whereas most WDR neurones were depressed or showed the biphasic response (Table I). Although care was taken in attempting to use similar stimulation when testing the 2 groups ot neurones, it is possible that there was a bias in the responses due to subtle differences in the location of the stimulator probe or in the parameters of stimulation. If such a bias had occurred. the conclusion may have been erroneous that vibration

Fig. 7. Comparison of responses to vibration of a WDR and of a non-nociceptive neurone. The WDR neurone A was recorded first. and when recording from this neurone had been completed the electrode was advanced to search for another single unit. The non-nociceptive neurone B was then isolated. 400 pm deeper than the WDR neurone. The stimulator probe was left undisturbed at the position indicated by the arrows in the insets. The records in a, b and c were constructed in the same way for each neurone. In each case, the continuous time histogram in ‘a’ shows the rate of discharge and illustrates the effects of hair movement (H), light 1 .O mm); each touch (T), firm pressure (Pr) and noxious pinch (Pi). The PSTHs in ‘b’ and ‘c’ show responses to vibration (amplitude histogram was made using 10 trials and the bin width is 10 msec. For b. the stimulator probe was located in the position indicated by the ’ x ’ in the diagram illustrating the receptive field. With the probe in this position the excitatory responses of both neurones were similar: there was a short latency response (8 k 0.01 msec for the non-nociceptive neurone and 9 f 0.1 msec for the WDR neurone. I = 6.7, P = lo-‘; 10 trials were used in each case to determine the latency), in both cases the burst at the onset of vibration was the discharge occurred in phase with the followed by a period of decreased excitation and during the remainder of vibration stimulation. The line below the histogram in b is a representation of the displacement output of the mechanical stimulator. For c the probe was in the position shown by the arrow. The receptive fields are illustrated in the diagrams between A and B. For the WDR neurone the receptive field is shown in the diagram to the left; this neurone was excited by mechanical stimuli applied throughout the hatched area. The receptive field for the non-nociceptive nettrone is illustrated by the blackened area in the diagram to the right. Below the diagrams is a photomicrograph showing a histological section which contains a dye deposit at the site of recording of the WDR neurone (A). The recording site for the non-nociceptive neurone (B) was determined in another section.

319

applied outside the low threshold receptive fields had a differential effect. Fig. 7 illustrates an attempt to control for possible variability in location of the probe or in stimulation parameters. In this case recording was first made from a WDR neurone (Fig. 7A,). This neurone was excited by vibration when the probe was placed inside the receptive field (Fig. 7A,) and showed the biphasic response when the probe was located outside the receptive field (Fig. 7AJ. The subsequent neurone studied was non-nociceptive (Fig. 7B,). Stimulation identical to that for the WDR neurone excited the non-nociceptive neurone with both locations of the probe (Fig. 7B, and 7B,). The receptive field of the non-nociceptive neurone was contained completely within the receptive field for innocuous mechanical stimuli of the WDR neurone. Therefore, it is unlikely that spread of vibration from the site of application into the receptive field can account for the differential excitation of the non-nociceptive neurone when the probe was applied outside the receptive field. In one other case, when it was possible to record in succession from 2 neurones of the 2 different functional types, a differential effect of vibration was again found when the stimulator probe was located outside the low threshold receptive field. In this case the WDR neurone was depressed by the vibration and the non-nociceptive neurone was excited. Therefore, the differential effect of vibration applied outside the receptive field appears to represent a real difference in the responses of WDR and non-nociceptive neurones.

Entrainment of discharge of non-nociceptive and WDR neurones The rates of firing of the WDR and of the non-nociceptive neurone illustrated in Fig. 7 could be entrained to the mechanical stimulation (Fig. 7A, and 7B,). Discharge entrained to vibrational stimulation was observed only when the stimulator probe was positioned inside the low threshold receptive field. In total, 6 of the 12 non-nociceptive neurones and 2 of the 35 WDR neurones excited by vibration exhibited a phasic discharge entrained to the stimulation; in the remaining

rt --r

Ii

TPr

11

SSS

Pi

D L!IC

6 10r bin-’ n. ” Fig. 8. Vibration-induced depression of a nociceptive specific neurone. A: the rate of discharge is shown in the continuous time histogram. The following stimuli were applied to the receptive field: movement of single hairs (H); light touch (T); firm, innocuous pressure (Pr); noxious manual squeeze (S) and noxious pinch (Pi). The receptive field is represented as the blackened area in the diagram to the right. The arrow indicates the location of the stimulator probe. B: the PSTH was constructed using 15 consecutive stimulations. The period of vibration is shown by the bar below the record. When vibration was being tested, glutamate (30 nA) was applied to induce firing of the neurone. The bin width for the PSTH is 50 msec.

cases the discharge lation.

was not entrained

to the stimu-

Responses of nociceptive specific neurones to vihration Vibration was tested on 3 nociceptive specific neurones; all were depressed by vibration regardless of the location of the stimulator probe relative to the receptive field (Fig. 8B). The time course of this depression was similar to the time course of the vibration-induced depression of WDR neurones. Neither excitation nor the biphasic response was ever seen. Responses of proprioceptive new-ones to vibration All proprioceptive neurones (n = 12) were excited by vibration. The magnitude of the excitation was directly related to stimulation amplitude. The latency and time course of this excitation were similar to those of the vibration-induced excitation of neurones with cutaneous receptive fields. As this paper focuses on neurones with cutaneous receptive fields, the responses of pro-

320

prioceptive neurones more detail.

will not

be considered

in

Correlation of site of recording with response to vibration Forty-five sites of recording (for examples see Figs. 2 and 7) were marked by deposits of Pontamine Sky Blue in a total of 30 experiments. For WDR neurones, recording sites were found in laminae I (n = 4) II (n = 11) IV (n = 10). V (n = 10) and VI (n = 5). At 39 of these 40 sites the neurones showed depressant or biphasic responses to vibration applied outside the low threshold receptive fields. The remaining site was one of those in lamina VI and this neurone was excited by vibration at all locations of the stimulator probe. Recording sites for non-nociceptive neurones were found in laminae III (n = l), IV (n = 2) and V (n = 1). The differences in the responses of WDR and non-nociceptive neurones do not appear to be due to preferential sampling of neurones in different laminae. One additional marked site of recording was that of a nociceptive specific neurone in lamina I.

Discussion The vibrational stimuli used in the present study evoked excitatory, depressant or biphasic responses in the dorsal horn neurones examined. The type of response of a given neurone was found to be critically dependent upon the location of the applied stimulation relative to the nonvibrational low threshold excitatory receptive fields and also upon the functional type of that neurone. The distinction between the functional groups was evident when the vibration was applied outside the receptive fields for low threshold stimulation: all non-nociceptive neurones were excited whereas depressant and biphasic responses were shown only by nociceptive neurones. The differential responsiveness of nociceptive and non-nociceptive neurones might arise either peripherally or centrally. The vibrational stimuli used in the present investigation likely activate a number of types of primary afferent [2] and, therefore, it is possible that the type of primary afferent mediating the responses of non-nociceptive neu-

rones is different from that mediating the responses of the nociceptive neurones. On the other hand, the excitatory responses of non-nociceptive neurones and the depressant and biphasic responses of WDR neurones appear to require mechanical transients and have similar latencies. Thus, it is possible that the different types of response result from activating the same type of primary afferent. If vibrational responses of nociceptive and non-nociceptive neurones are indeed mediated by the same type of primary afferent, then the differences in the responses may arise centrally. Two possible explanations which account for the different responses are: (i) that the network of neuronal connections between the involved primary afferents and the two types of dorsal horn neurone might be different, and/or (ii) that the chemical mediators of the responses of the dorsal horn neurones might be different. With regard to the latter, the finding that the vibration-induced depression and the depressant component of the biphasic response appear to be mediated by adenosine [14] leads to the possibility that nonnociceptive neurones show only excitatory responses because they cannot be inhibited by adenosine, due to lack of adenosine receptors or of appropriate cellular transduction mechanisms. This possibility must be rejected because exogenously administered adenosine depresses the firing rate of non-nociceptive as well as of nociceptive neurones [12]. In these previous studies it was noted, however, that exogenously applied ATP excites non-nociceptive neurones and has predominantly depressant or biphasic (excitatory/ depressant) effects on nociceptive neurones [ 121. Therefore, if ATP were released by vibration this may be sufficient to account for the differences in the responses of non-nociceptive and nociceptive neurones. For the present this explanation remains speculative and the precise basis for the different vibration-induced responses of nociceptive and non-nociceptive neurones therefore requires further investigation. Relevance to vibration-induced analgesia As mentioned in the Introduction, experimental [1,4,9,17,18] and clinical [5-81 studies indicate that

321

cutaneously applied vibration decreases pain intensity in humans. The present results prompt us to suggest that these analgesic effects of vibration may be mediated at the spinal level by depressant and biphasic responses of nociceptive neurones and/or by excitatory responses of non-nociceptive neurones. While it is tempting to focus on the depression of nociceptive neurones, the stimulation amplitudes and frequency used in the human studies are in the range expected to cause all of the neuronal responses reported here. Therefore, it is possible that excitation of non-nociceptive neurones, occurring together with depression of nociceptive neurones, may also play a role in vibration-induced analgesia. The vibration-induced depression and the depressant component of the biphasic response appear to be mediated within the dorsal horn by adenosine: these depressions are blocked by methylxanthines such as caffeine and are potentiated by the adenosine uptake inhibitor dipyridamole [14]. Thus, within the spinal cord adenosine may be an endogenous chemical mediator of vibrationinduced analgesia. Should this suggestion be substantiated by clinical investigation two main clinically relevant consequences might arise. First, it may be useful for patients receiving vibration for pain relief to limit or avoid the intake of caffeine and other methylxanthines. This is of particular concern because methylxanthines are the most extensively used pharmacological agents [ll]. Second, drugs which block adenosine uptake or degradation or which otherwise potentiate its action may be useful in increasing the efficacy of vibration. For reasons fully described elsewhere [12], similar therapeutic considerations may also apply to the analgesia caused by electrical stimulation of peripheral nerves or of dorsal columns.

Acknowledgements Funds for this project were provided by the Medical Research Council of Canada (Grant No. MA-5891) to J.L.H. M.W. Salter is a Fellow of the MRC. Thanks to Mr. R. Lamarche for expert photography. The authors also gratefully acknowledge the following generous donations: pan-

curonium bromide (Pavulon) from Organon, Hill, Ont. and halothane (Somnothane) Hoechst Pharmaceuticals, Montreal.

West from

References K.-E., Schady, W. and 1 Bini, G., Cruccu, G., Hagbarth, Torebjork, E., Analgesic effect of vibration and cooling on pain induced by intraneural electrical stimulation, Pain, 18 (1984) 239-248. 2 Burgess, P.R. and Perl, E.R., Cutaneous mechanoreceptors and nociceptors. In: A. Iggo (Ed.), Handbook of Sensory Physiology. Vol. II. Somatosensory System, Springer, New York, 1973, pp. 29-78. J.G., Small motion biological stimulator, APL 3 Chubbuck, Tech. Digest, 6 (1966) 18-23. 4 Ekblom, A. and Hansson, P., Effects of conditioning vibratory stimulation on pain threshold of the human tooth, Acta Physiol. Stand., 114 (1982) 601-604. T., Long-term results of vibratory stimulation 5 Lundeberg, as a pain relieving measure for chronic pain, Pain, 20 (1984) 13-23. R. and Ottoson, D.. Pain allevia6 Lundeberg, T., Nordemar, tion by vibratory stimulation, Pain, 20 (1984) 25-44. T.C.M., Vibratory stimulation for the allevia7 Lundeberg, tion of chronic pain, Acta Physiol. Stand., S523 (1983) 1-51. 8 Ottoson, D., Ekblom, A. and Hansson, P., Vibratory stimulation for the relief of pain of dental origin, Pain, 10 (1981) 37-45. A., Modification of human pain threshold by 9 Pertovaara, specific tactile receptors, Acta Physiol. Stand.. 107 (1979) 339-341. 10 Price, D.D. and Dubner, R., Neurons that subserve the sensory-discriminative aspects of pain, Pain 3 (1977) 307-338. 11 Rall, T.W., The xanthines. In: A.G. Gilman, L.S. Goodman and A. Gilman (Eds.), The Pharmacological Basis of Therapeutics, MacMillan, New York, 1980. pp. 592-607. 12 Salter, M.W. and Henry, J.L., Effects of adenosine 5’monophosphate and adenosine 5’-triphosphate on functionally identified units in the cat spinal dorsal horn. Evidence for a differential effect of adenosine 5’-triphospbate on nociceptive vs. non-nociceptive units, Neuroscience, 15 (1985) 815-825. 13 Salter, M.W. and Henry, J.L., Differential effects of peripheral vibration on nociceptive versus non-nociceptive neurones in the lumbar dorsal horn of the cat, Sot. Neurosci. Abst., 12 (1986) 736. 14 Salter, M.W. and Henry, J.L., Evidence that adenosine mediates the depression of spinal dorsal horn neurones induced by peripheral vibration in the cat, Neuroscience. 22 (1987) 631-650.

322 15 Salter, M.W. and Henry, J.L., Tachykinins enhance the depression of spinal nociceptive neurons caused by cutaneously applied vibration in the cat, Neuroscience. 27 (1988) 243-249. 16 Salter, M.W. and Henry, J.L.. Physiological characteristics of responses of wide dynamic range spinal neurones to

cutaneously applied vibration in the cat, Brain Res., (1990) in press. 17 Sullivan, R.. Effect of different frequencies of vibration on pain-threshold detection, Exp. Neurol., 20 (1968) 135-142. 18 Wall, P.D. and Cronly-Dillon, J.R., Pain, itch and vibration, Arch. Neurol.. 2 (1960) 365-375.