JOURNALOF

NEIJROPHYSIOLOGY 1992. Printrd

Vol. 67. No. 4, April

in U.X.4.

Modulation of Cortical Evoked Potentials by Stimulation of Nucleus Raphe Magnus in Rats KENNETH A. FOLLETT AND G. F. GEBHART Division of Nezuos~gery, University of Iowa Hospitals, and Department College ojXfedicine, University of Iowa, Iowa City, Iowa 52242 SUMMARY

AND

CONCLUSIONS

I. In pentobarbital sodium-anesthetized rats, we evaluated changes in cortical evoked potentials ( EPs) associated with electrical and chemical stimulation of nucleus raphe magnus (NRM). A condition-test (C-T) paradigm was used. Cortical EPs were produced by test stimuli delivered to a hindpaw or the thalamic ventral posterior lateral nucleus (VPL; electrical stimulation), or by photic stimulation of the eyes or electrical stimulation of contralatera1 homotopical cortex (transcallosal EPs). These test stimuli were then preceded by electrical or chemical conditioning stimulation (CS) delivered to NRM through a stereotaxically implanted electrode or injection cannula, respectively. Effects of CS on EPs produced by the test stimuli were characterized. 2. Electrical CS preceding a test stimulus delivered to the foot reduced the amplitude of EPs at thresholds as low as lo-25 PA. The magnitude of EP reduction was dependent on CS intensity, frequency, and the C-T interval. Optimal parameters were trains of 10 pulses (400 Hz) delivered at a C-T interval of 5- 10 ms. Injection of glutamate and lidocaine into NRM demonstrated that these effects were due to activation of NRM neurons and not to current spread to medial lemniscus (ML). NRM CS also reduced cortical EPs produced by test stimulation in VPL but did not alter EPs from visual stimulation or from electrical stimulation of contralateral homotopical cortex. 3. These findings suggest that NRM CS attenuates EPs by inhibiting thalamic or thalamocortical afferent activity. Because NRM CS affected all components of the cortical EPs, the effect appears to involve alteration of general sensory activity and is not nociception specific. Even in the absence of nociceptive selectivity, the effect represents modulation of supraspinal afferent activity, in addition to well-documented descending inhibitory influences, by stimulation-produced antinociception.

INTRODUCTION

Stimulation in certain brain stem sites produces potent antinociception (stimulation-produced antinociception; SPA) mediated, at least in part, by activation of descending systems that modulate afferent activity at spinal segmental levels (see Gebhart 1986 for review). An ascending system of antinociception may also exist (Andersen and Dafny 1983; Condes-Lara and Omana Zapata 1988; Hagbarth and Kerr 1954; Handwerker and Zimmermann 1972; Hernandez et al. 1977; Morgan et al. 1989; Qiao and Dafny 1988; Schieppati and Gritti 1983). Stimulation in the midbrain periaqueductal gray (PAG) elicits potent SPA and, in rats, reduces the amplitudes of cortical evoked potentials (EPs) produced by peripheral stimuli (Garkavenko and Gura 1988; Hernandez et al. 1977; Samanin et al. 1972). Several other isolated reports indicate that stimulation in 820

of Pharmacology,

the brain stem reticular formation alters cortical activity evoked by peripheral stimuli (Bremer and Stoupel 1959; Gauthier et al. 1956; Hagbarth and Kerr 1954). These effects have not been linked to antinociceptive systems, nor have they been studied systematically. Many of the antinociceptive effects of brain stem stimulation, including those produced by stimulation in PAG, are mediated in part via structures in the ventromedial medulla, especially nucleus raphe magnus (NRM) (see Gebhart and Randich 1990 for review). To study the hypothesis that SPA is associated with changes in supraspinal afferent activity, we systematically evaluated the effects of stimulation of NRM on cortical evoked activity. A preliminary report of these data has been given (Follett and Gebhart 1988). METHODS

Adult male Sprague-Dawley rats ( Biolab, St. Paul, MN), 300450 g, were initially anesthetized deeply with pentobarbital sodium (45-50 mg/kg ip) for surgery. Femoral venous and arterial and tracheal cannulas were inserted. Blood pressure and heart rate were continuously monitored and body temperature maintained with a heating pad. Rats were paralyzed with pancuronium bromide (0.2 mg iv as needed), artificially ventilated with N,O-O2 ( l-2: 1)) and given a constant intravenous infusion of pentobarbital (3-6 mg kg-’ h-l). This anesthetic regimen is sufficient to prevent cornea1 and withdrawal reflexes in nonparalyzed animals. l

l

Stimulation A condition-test (C-T) paradigm was used, with the conditioning stimulus (CS) being delivered to NRM. Test stimuli were delivered as electrical stimulation of a hindpaw, the thalamic ventral posterior lateral sensory nucleus ( VPL), contralateral homotopical cortex, or by photic stimulation of the eyes. CS of NRM was performed with the use of a concentric bipolar electrode (#17-20-2, Frederick Haer, Brunswick, ME) in either monopolar or bipolar fashion. In some experiments a 26-gauge cannula was placed stereotaxically into NRM, and a 34-gauge copper wire, insulated except at the tip, was passed through the cannula, extending 2 mm beyond the cannula. This permitted stimulation in a monopolar fashion, after which the copper wire electrode could be replaced by an injection cannula (33 gauge) for drug injection at the stimulation site. NRM electrodes were placed in the midline, lo- 11.5 mm posterior to bregma, depending on the posterior location of the transverse sinus. Depth was adjusted to the site of maximum effect of stimulation, typically 7-9 mm ventral to bregma. After initial studies to determine the optimal stimulation parameters, CS was a train of 10 pulses, 0.1 ms each, 400-Hz pulse

0022-3077/92 $2.00 Copyright 0 1992 The American Physiological Society

Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 12, 2019.

MODULATION

OF

CORTICAL

frequency (train duration, 25 ms), repeated once each 5- 15 s. The C-T interval was maintained at 10 ms to minimize stimulus artifact in the recordings except when evaluating the influence of varying C-T intervals. Test stimuli were delivered to the hindpaw contralateral to the cortical recording site through bipolar needle electrodes inserted into the plantar surface of the foot. Stimuli were 10 mA, 0.5-ms pulses. Test stimulation in VPL, ipsilateral to the cortical recording site, was performed through a concentric bipolar electrode (# 17-20-2, Frederick Haer, Brunswick, ME). Stereotaxic coordinates (referenced to bregma) for electrode placement in VPL were posterior (P) 3.0 mm and lateral (L) 3.0 mm. Electrode depth was adjusted to the site of maximum cortical evoked response, usually 6.5 mm ventral. Test stimuli in VPL were 0. 1-ms pulses at intensities that were slightly submaximal for evoking the maximum amplitude cortical potential (300-800 PA). Test stimulation of homotopical contralateral cortex was performed with the use of sideby-side bipolar silver ball electrodes; tip separation was ~2 mm. Stimulus intensities (single pulses, 0.2-0.5 ms) were slightly submaximal for evoking the maximum cortical response contralaterally (400- 1,000 PA). All electrical stimuli were constant-current cathodal squarewave pulses (Grass S88 stimulus generator with Grass PSIUG stimulus isolation units). Visual-evoked responses were produced with the use of a Grass PS2 photostimulator. The photoflash intensity (setting, 4-8) was slightly submaximal for evoking the maximum amplitude cortical response. Atropine sulfate (0.04 mg/ ml) drops were instilled in each eye to dilate the pupils.

Recording Cortical evoked responses to the test stimuli were recorded at the site of maximum amplitude in each animal with the use of a silver ball electrode. For test stimulation of the foot, VPL, or contralateral cortex, recording sites were generally located at P 2.03.0 and L 2.0-4.0 mm. Visual evoked responses were recorded in the area of P 7.0, L 4.0 mm. The signal was led into a Grass P5 1lJ preamplifier (half-amplitude settings 10 Hz and 3 kHz, 60-Hz notch filter, gain 2,000). The signal was then led simultaneously into an oscilloscope for display and into a computerized data acquisition system (Modular Instruments, Southeastern, PA) with a IO-kHz sampling frequency. Ten evoked responses were recorded and averaged for each trial ( lOO-ms duration recording). The resulting average evoked responses were stored in digital format for later reproduction and analysis of latency, amplitude, and configuration. Experimental data were considered invalid if averaged amplitudes of control (unconditioned) EPs checked immediately before and after each set of experimental manipulations varied > 10%. Amplitudes were measured as peak-to-peak values between the major negative and positive peaks.

Drug inject ions In some animals, after determining the presence of effects of electrical CS on evoked responses to foot shocks, monosodium glutamate (0.5 ~1, 100 mM, pH 6.98; Sigma Chemical, St. Louis, MO) was injected as a chemical stimulant, or lidocaine (0.5 ~1, 4%; Astra Pharmaceuticals, Westborough, MA) was injected to reversibly block the effects of subsequent electrical stimulation in NRM. The injections were made through a 33-gauge cannula inserted through the 26-gauge guide cannula that was also used to hold the monopolar stimulation electrode. Both the injection cannula and the monopolar electrode extended 2 mm beyond the tip of the guide cannula so that injections were made at the site of the electrode tip. All injections were conducted over 30 s with the use of a hand-driven microinjector. Injection was verified by watching the movement of an air bubble placed in the iniection catheter.

EVOKED

POTENTIALS

821

When the injection was followed by electrical stimulation, the injection cannula was replaced by the electrode after one additional minute.

Histology At the conclusion of each experiment, electrolytic lesions were made at electrode or injection sites. The animals were killed by intravenous overdose of pentobarbital. Brains were removed and fixed in Formalin. Coronal frozen sections (40 pm) were cut, stained with cresyl violet, and examined to verify electrode position. RESULTS

Cortical evoked responses to foot shocks CS of NRM reduced the amplitude of the cortical EPs produced by a subsequent test shock delivered to the contralateral hindfoot. There were no associated changes in blood pressure or heart rate. The effect was dependent on CS intensity, number of pulses within the CS train, pulse frequency, and the C-T interval. Reduction of EP amplitudes occurred in a graded fashion at thresholds as low as lo-25 PA with near complete attenuation at intensities as low as 50- 100 PA (Fig. 1). Variations in the slopes of the stimulus-response curves (Fig. 1 B) may be related to minor differences in electrode placement, although examination of electrode sites in histological sections failed to demonstrate any consistent differences. For 10 animals (data shown in Fig. 1) peak-to-peak amplitudes of EPs were reduced by 50 and 90% compared with control at mean CS intensities of 32 t 2 and 85 t 3 (SD) PA, respectively. The major positive and negative peaks of the EPs were both attenuated by CS. No consistent changes in latencies of the evoked responses could be measured, although latencies to the onset of the EPs and latencies of the major negative and positive peaks were often prolonged after NRM CS (e.g., Figs. 1A and 2A). For a given CS intensity, the attenuation of cortical EPs was greater with increasing pulse frequency within the lopulse stimulus train. A pulse frequency of 400 Hz caused an average reduction of EP amplitude by 83 t 9 versus 26 t 14% at 50 Hz ( 10 animals; Fig. 2 ) . EPs could be attenuated by a single pulse delivered to NRM, but typically three to five pulses were necessary. The effect was most pronounced after delivery of at least five pulses in the stimulus train (Fig. 3). The effect of NRM CS on EPs persisted unchanged for 15-20 ms after termination of the NRM stimulus and was gradually lost as the interval between the CS and the test stimulus increased (Fig. 4). Lesser effects of NRM CS persisted at least 250-500 ms after termination of stimulation. In a few animals (e.g., 3 of 10 in Fig. 4) potentiation of the evoked responses to the foot shocks was seen at C-T intervals of 50- 100 ms, resulting in increased variability and lesser effects seen in the mean values at the 50- and IOO-ms intervals in Fig. 4. Localization

of site of action

That the reduction of EPs by CS was due to activation of cells and not to current spread to adiacent medial lemniscus

Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 12, 2019.

K. A. FOLLETT

822

A

1.25-

2 +0

-.

1.00.

-.---

AND G. F. GEBHART

control

25 50 100

/LA NRM CS /LA NRM CS /LA NRM CS

-------s-s-

control 100 Hz CS 200 Hz CS

P

1

0.00

, 0

1

1

I

1

20

40

60

80

Time

1 100

-0.6J

(msec)

I 0

I 20

I 40

Time

B

B

% control

-E 3 +

.-ii E u CL w

1 80

1 100

(msec)

% control

u 3 +.n E 0

60-

I 60

6040-

k 20 1 0’

r

1

I

,

1

0

25

50

75

100

NRM

stimulus

intensity

01

( 0

I 100

1 200

Pulse

(PA)

1 300

frequency

c 400

(Hz)

FIG. 2. Effect of varying the pulse frequency of NRM CS on cortical EPs. Reduction of cortical EPs to foot shocks is greater with increasing pulse frequency within the CS train. A: example from 1 animal. B: summary data from 10 animals. CS electrode sites are comparable with those illustrated in Fig. 1C.

C

lmm FIG. 1. Effect of NRM CS on cortical EPs. Amplitudes (peak-to-peak) of cortical EPs to foot shocks are progressively reduced by CS of increasing intensity. A : example from 1 animal. B: intensity-dependent inhibition for 10 animals. C: CS electrode sites corresponding to data presented in B (adapted from Paxinos and Watson 1982) (histology not available for one animal). LC, locus coeruleus; Sp5, trigeminal nucleus; n7, facial nerve; Pyr, pyramids.

differences in injection sites, time required for diffusion to sites of action, or may reflect the presence of a transient depolarization block. To confirm that the reduction of EPs by NRM CS was not caused by current spread to ML, lidocaine was microinjetted into NRM to reversibly block the effects of electrical stimulation. In four animals NRM CS reduced foot-shock EPs to lo-40% of their unconditioned peak-to-peak ampli% control 100,

80-

(ML) was tested by injection of glutamate and lidocaine into NRM. In 11 animals, after verifying that NRM CS attenuated EPs to foot shocks, glutamate was microinjected at the site of the electrode tip. In five animals glutamate caused >50% attenuation of subsequent EPs to foot shocks ( Fig. 5 ). In five others EPs were reduced 25-50% from control amplitudes. In eight animals the onset of action of glutamate was within 2 min. The duration of effect was 15-30 min. As the effects of glutamate wore off, the EPs often reappeared in a graded fashion until they reached their preinjection amplitudes. Variations in the effects of glutamate could not be related to obvious differences between injection sites (Fig. 5C) and may have resulted from minor

s 2 -E

E

u 0,

60-

40-

W

20 I 01

I

I

1

I

I

1

0

2

4

6

8

10

Number

of

pulses

3. Effect of the number of pulses within the CS train on cortical EPs to foot shocks. Data from 6 animals. FIG.

Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 12, 2019.

MODULATION --

-0.91

1 0

1 20

control 5 msec

---

10

----

25

1 40

msec msec

I 60

Time

OF CORTICAL NRM NRM NRM

I 00

CS CS CS

1 100

(msec)

EVOKED

823

were attenuated by CS in NRM (Fig. 7). In these animals NRM CS reduced the peak-to-peak amplitudes of footshock EPs by an average maximum of 80 t 22% at stimulus intensities ranging from 30 to 600 PA (mean, 196 t 235 PA; median, 75 PA). Amplitudes of thalamocortical potentials in these same animals were reduced by 63 t 20% at NRM CS intensities of 50-600 PA (mean, 290 t 2 13 ,uA; median, 200 PA). Stimulation of VPL at intensities that were supramaximal for evoking cortical EPs lessened the magnitude of the NRM CS effect. The time course of the effect, as demonstrated by varying C-T intervals, was similar to that seen for the foot-shock EPs. No changes in latenties were apparent.

A

B %control

% control

1 1,

125

120loo-

100“a F

POTENTIALS

~~~ :

3 .-+ n E 0 0W

GO-

75.

,----t \Y( /.

50-

4

/ L

11'

,I

t

I

10

I

1

.

1

20

30

40

50

Condition-Test

interval

I

100

1

250

I

500

1

1000

B

tudes. Within 30-60 s after microinjection of lidocaine at the electrode site, NRM stimulation no longer had any effect on the EPs (Fig. 6). The unconditioned evoked response to the foot shocks (i.e., when not paired with the NRM CS) was slightly reduced for several minutes in one animal and transiently reduced (40%, < 15-20 min) in two animals. In the fourth animal, the unconditioned EP was transiently reduced by 90% (Fig. 6C). In this latter animal a previous electrode track into the medullary pyramids probably formed a channel that allowed rapid spread of lidocaine into ML. In every case, however, loss of effect of NRM CS clearly preceded any change in unconditioned EPs, indicating that the effect was originating in the immediate vicinity of the injection/stimulation site in NRM and not by alteration of activity in ML. Furthermore, the unconditioned EPs began to return to baseline within lo- 15 min, whereas the loss of effect of NRM CS persisted. Over the course of 60- 120 min, NRM stimulation became gradually effective once again in attenuating EPs, but the preinjection level of effectiveness was not regained during the observation period. To determine whether the attenuation of EPs by NRM CS involved spinal mechanisms, the test stimuli were delivered to the thalamic sensory nucleus VPL in five animals, bypassing the spinal pathways associated with the footshock EPs. Cortical EPs produced by stimulation in VPL

’\

\

.-.\

) \

,' \

‘J ;

/

\ \

*/.-.-

'

‘\

/'

,/' '\

'/'

.\./

."

'\ \

,HHYH \'

r

1

I

1

0

5

10

15

Time

% control

after

injection

(min)

120-

(msec)

FIG. 4. Effect of varying the C-T interval on cortical EPs. The magnitude of reduction of cortical EPs to foot shocks becomes less as the interval between NRM CS and the foot shock increases. A : example from 1 animal. B: summary data from 10 animals. CS electrode sites in NRM are comparable with those illustrated in Fig. 1C.

/’ / /

.'\ ;

0’

:,sq’

/

.

0

. \

i.’

'x/

25-

o-1

,Q....

. I\

100G 3 .C

80

a E u

60

20 + El

40 :L. 10

123456789

Maximum

effect

11

of glutamate

C

lmm FIG. 5. Effect of glutamate microinjection into NRM on cortical EPs to foot shocks. A : each line represents data from 1 animal. Note the general return of EP amplitude toward preinjection baseline ( 100%) over the observation period. Line numbers correspond to injection sites shown in C (adapted from Paxinos and Watson 1982). B: maximum reduction of cortical EP amplitude for each animal, from data plotted in A (x-axis numbers correspond to line numbers in A ). C: injection sites corresponding to line numbers in A. ML, medial lemniscus.

Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 12, 2019.

A

B 1.2

Pro-Inloctlon ---1-

e

0* v) 0.8 ,>, -g 0.4 +3 L=n E 0.0 Q:

control NRM CS

- - control . . . . . . . NRM CS

-0.4 0

C

20

I 60

40

Time

% control

I 80

0

1 100

,0 4 /

.-OC .gz

Time

after

injection

60

(min)

D

120b\A-*&;>::NAy

8(-J-

9

* t

Ei

/? t:-

&L 3W

40

20

preinjection

(msec)

160-

DO) iti;

% control

:’

3:

40-

,A’\ ‘\

/” \. \/

o-

1 PI-OIn/ectlon

‘\

/’

A-.-.

‘A-.-.-‘-

~ ---A

2

-/

ML

PYr

A

I 20

Time

after

1 40

injection

1 60

lmm

(min)

FIG. 6. Effect of lidocaine microinjection into NRM on cortical EPs to foot shocks by NRM CS. A: in 1 animal, before lidocaine injection, NRM CS nearly abolishes the foot shock EPs. After lidocaine injection, NRM stimulation no longer has an effect on the foot shock EPs, whereas the unconditioned EPs are unchanged from preinjection amplitude. B: data from 4 animals. Before lidocaine injection, NRM stimulation reduces EP amplitudes to lo-40% of control. Immediately after lidocaine injection, NRM stimulation fails to attenuate EPs. C: cortical EPs to foot shocks (unconditioned EPs) are transiently reduced after lidocaine microinjection into NRM, returning to baseline amplitude in 20-30 min. Data from same animals as in B. Line numbers correspond to injection sites shown in D. D: injection sites corresponding to data presented in B and C (adapted from Paxinos and Watson 1982).

C

0.0-I

B

1

I

0

20

+ 0 L -iii 3 -.a, E