Neuroscienec Letters, 162 (1993) 133-136 133 © 1993 Elsevier Scientific Publishers Ireland Ltd. All rights reserved 0304-3940/93/$ 06.00

NSL 09946

Effects of anesthetic agents on somatosensory responses of raphespinal neurons in the cat*

R o b e r t W. Blair**, Angela R. E v a n s

Department of Physiology, University of Oklahoma Health Sciences Center, PO Box 26901, Oklahoma CiO', OK 73190, USA

(Received 23 April 1993; Revised version received 12 July 1993: Accepted 3 August 1993)

K~ 3, words': a-Chloralose; Pentobarbital: Spontaneous activity; Nociceptive response

The goal of this study was to examine the effects of ct-chloralose and pentobarbital on the somatosensory responses of medullary raphespinal neurons. With ~t-chloralose, 98% of neurons responded to innocuous stimuli, particularly tapping, but only 1/64 responded selectively to nociceptive stimuli. In contrast, 73% of neurons responded selectively to noxious stimuli with pentobarbital, and none responded selectively to innocuous stimuli. In addition, raphespinal neurons studied with pentobarbital had higher spontaneous discharge rates, and a higher incidence of spontaneity, than neurons studied u.ith c~-chloralose. Thus, different anesthetics produce raphespinal neurons with different somatosensory response characteristics.

Anesthesia alters the response characteristics of neu- rons. Most studies that have examined the effects of an- esthesia on neuronal responses have compared responses in awake or conscious animals with responses in anesthe- tized animals. However, few studies have tested neurons

for differences in responses produced by different anes- thetics. Considering neurons in the medullary raphe nu- clei, a comparison of the reported somatosensory re- sponse properties suggests that different anesthetics pro- duce different proport ions of neurons responsive to in- nocuous vs. noxious stimuli; however, there also may be

differences due to species and/or to the methods used to determine receptive fields [1, 3, 6, 9, 11-13]. Light [9] suggested that lumbosacral raphespinal axons studied with pentobarbital anesthesia were more responsive to noxious stimuli than axons studied with ~-chloralose an- esthesia, but only seven axons were studied with 0t- chloralose. The purpose of the present study was to com- pare the somatosensory response properties of medul- lary raphespinal neurons in cats anesthetized with a- chloralose or pentobarbital.

Experiments were performed on 59 adult cats of either sex weighing 2.5 5.0 kg. 40 cats were tranquilized with ketamine (10-20 mg/kg i.m.) and subsequently anesthe- tized with ~-chloralose (40 mg/kg i.v.). 19 cats were anes- thetized with sodium pentobarbital (35 mg/kg i.p.). Sup- plemental doses of anesthetic were administered as re- quired to maintain anesthesia. Paralysis was initiated

*With the technical assistance of Jerry Thompson. **Corresponding author.

with a bolus dose of pancuronium bromide (Pavulon: 0.15 mg i.v.). Adequacy of anesthesia (no withdrawal re-

flex in response to noxious pinching) was tested once paralysis wore off and before the next dose of pancu- ronium was given. The cats were mechanically ventilated with room air and expiratory PCO, was maintained be- tween 3.5 and 5%. Body temperature was kept at 38 + I°C with a heating pad and heat lamp. The cat was placed in a stereotaxic device. The head was flexed for- ward approximately at a 45 ° angle, and the skull overly- ing the caudal cerebellum and brain stem was removed. The cerebellum was carefully retracted to expose the floor of the fourth ventricle. Clamps were placed on the C7 or T~ and the T 3 or T 4 spinous processes to support the spinal column, and a laminectomy was performed to expose the cord between the clamps. Concentric bipolar stainless steel electrodes (SNE 100, Rhodes) were in- serted bilaterally near the dorsolateral sulci to a depth of 1.5 2.0 m m in the spinal cord. Extracellular potentials of midline medullary neurons were recorded with tungsten or stainless steel microelectrodes with an impedance of 3-5 M£2. The recording electrode was visually placed on the midline of the floor of the fourth ventricle from 1 4

mm rostral to the obex, and potentials were recorded from 1 7 mm ventral to the brainstem surface. Neuronal activity was amplified and displayed conventionally. During the search procedure, the spinal cord was con- stantly stimulated at 1 Hz with pulses of 0.1-ms duration and 3M mA intensity from one of the spinal electrodes to stimulate axons of descending raphe neurons coursing to

134

or through the thoracic spinal cord. The midline of the

medulla was searched for antidromically activated neu-

rons. Criteria for antidromic activation of a neuron in-

cluded: (1) constant latency (range 1.0--8.4 ms) of neu-

ronal activation from the cord; (2) ability of the neuron

to follow high-frequency (-> 500 Hz) stimuli applied to

the cord; and (3) collision of orthodromic spikes with

antidromic spikes. All neurons in this report fulfilled

these criteria.

Neuronal somatosensory responses were determined

by applying various stimuli to all regions of the body.

Stimuli included noxious pinching of the skin, noxious

squeezing of the skin and underlying muscle, light tap-

ping of the skin, and movement of hairs or whiskers. The

somatic response properties of the neurons were classi-

fied as follows. Neurons that responded only to tapping

or movement of hairs were classified as low threshold

(LT). Neurons that responded only to noxious stimuli

were classified as high threshold (HT). Neurons that re-

sponded to tapping or movement of hairs, but responded with a greater intensity to noxious stimuli, were classified

as wide dynamic range (WDR). Neurons that were ex-

cited by noxious pinch but were inhibited by movement

of the overlying hairs were classified as high threshold,

inhibitory (HTi). An electrolytic lesion (50 etA for 20 s)

was made to mark the location of each neuron, and the

lesion sites were identified from the fixed (10% buffered

formaldehyde) tissue. Neuronal spikes were converted to

standardized pulses with a discriminator. Pulses were led

to a Cambridge Electronic Design (CED) 1401 A/D con- verter connected to a computer. Parametric data were

analysed with the unpaired t test. Categorical data were analysed with the X 2 or the Fisher's exact test. Statistical

calculations were performed with the Statistica (StatSoft) statistical package. Differences were considered signifi-

cant if P < 0.05. Numerical data are presented as

mean + S.E.

TABLE I

CONDUCTION VELOCITIES AND SPONTANEOUS ACTIVI- TIES OF RAPHESPINAL NEURONS

In each cell, upper numbers present mean + S.E.; lower numbers show ranges of values. These data were not obtained from one neuron stud- ied under ct-chloralose

~t-Chloralose Pentobarbital P (n = 63) (n = 22)

Conduction velocity (m/s) 65.4 + 3.0 66.8 + 4.5 N.S. 12.4- 121.0 22.2- 109.2 1.9 + 0.6 6.3 + 1.6 < 0.002

0 26 0- 30 8.6 + 1.5 27.0 + 4.9 < 0.001

0-50 1--111

Lowest spontaneous activity (spikes/s) Highest spontaneous activity (spikes/s)

The somatosensory responses of the neurons studied

under ~z-chloralose anesthesia were briefly alluded to in a

previous report [3]. The somatosensory responses of the neurons studied under pentobarbital anesthesia have nol

been previously described. A total of 64 raphespinal neu-

rons were studied under ~-chloralose anesthesia and 22 were examined under pentobarbital anesthesia. All neu-

rons were located in nucleus raphe magnus or nucleus

raphe obscurus. The antidromic conduction velocities and range of spontaneous activities for the neurons are

presented in Table I. There was no difference in the mean conduction velocity between neurons studied under ct-

chloralose (65.4 + 3.0 m/s) and pentobarbital (66.8 + 4.5

m/s) anesthesia. The spontaneous activity typically var-

ied during the time most raphespinal neurons were stud- ied. Therefore, the lowest spontaneous activity as well as

the highest activity for each neuron was recorded, and

these values were then averaged across neurons. With

c~-chloralose anesthesia, neuronal spontaneous activity

varied between a low of 1.9 + 0.6 spikes/s and a high of

8.6 + 1.5 spikes/s. Both of these mean values were signifi- cantly lower (P < 0.002 and P < 0.001, respectively) than

the mean lowest and mean highest spontaneous activities

recorded under pentobarbital anesthesia (6.3 _+ 1.6 and 27.0 + 4.9 spikes/s). In addition, the incidence of neurons

with spontaneous activity was greater (P < 0.01) with pentobarbital than with 0t-chloralose anesthesia. For this

analysis, neurons were considered not to have spontane-

ous activity if their background discharge rates were <- 1 spike/s. With ct-chloralose, 66% of the neurons had spon-

taneous activity, but with pentobarbital all but one neu-

rons (95%) exhibited spontaneous activity. The relationship of the class of response properties to

the anesthetic used is shown in Table II. The distribution of values in Table II is statistically significant (P < 0.001). No LT neurons were found in experiments

in which pentobarbital was the anesthetic agent. Fur-

TABLE II

CLASS OF RESPONSES ACCORDING TO ANESTHETIC AGENT

Value in each cell shows number of neurons with that class of somatic field, followed by the percentage of neurons with that classification under each anesthetic. LT, low threshold; WDR, wide dynamic range; HT, high threshold; Hti, high-threshold inhibitory. The distribution of values in this table is statistically significant (Z ~ test, P < 0.001)

Class of somatic field

Anesthetic LT WDR HT Hti a-Chloralose 45 (70%) 17 (27%) 1 (2%) 1 (2%) Pentobarbital 0 6 (27%) 16 (73%) 0

thermore, neurons were significantly (P < 0.001) more

likely to respond to innocuous stimuli ifcz-chloralose was

the anesthetic agent. Thus, 63/64 (98%) neurons studied with a-chloralose anesthesia responded to tapping or movement of hairs, but only 6/22 (27%) neurons studied with pentobarbital anesthesia responded to innocuous stimuli. This difference is largely accounted for by en- hanced responses to the tapping stimulus when :z- chloralose was used. Thus, all but one (98%} neurons studied with ~-chloralose anesthesia responded to tap- ping, but only 1/11 (9%) neurons responded to tapping when pentobarbital was the anesthetic (P < 0.001). A smaller percentage of neurons was responsive to move-

ment of hairs when pentobarbital was used compared with ¢z-chloralose (43 and 73%, respectively), but this dif- ference was not statistically significant.

Regarding responses to noxious stimuli, all neurons studied with pentobarbital anesthesia responded to nox- ious pinching of the skin or the skin and underlying mus- cle, but only 30% of neurons studied with cz-chloralose anesthesia responded to noxious pinching (P < 0.001). Furthermore, 16/17 (94%) of the HT neurons were found

when pentobarbital anesthesia was used. The sizes of the receptive fields were comparable when

~-chloralose and pentobarbital were used. All neurons had somatic fields encompassing most of the body.

Previous reports examining somatosensory responses of different populations of neurons also have found that

most neurons studied under cz-chloralose anesthesia re- spond to innocuous stimuli, particularly tapping, but neurons studied with pentobarbital anesthesia are more likely to respond to noxious stimuli. When studied with ~-chloralose anesthesia, from 67 98% of raphe (includ- ing raphespinal) or reticular neurons respond to tapping

[1, 6, 9, 13]. Furthermore, the majority of these neurons are classified as LT neurons, and ~25% are W D R neu- rtms, but few are HT neurons [1, 6, 9: but see Ref. 13 where 33% were HT]. Similarly, thalamic neurons stud- ied with volatile anesthetics (nitrous oxide and ha- lothane) exhibited responses to noxious stimuli, but lost their responsiveness to noxious stimuli and gained re- sponsiveness to innocuous stimuli and sometimes had larger receptive fields when chloralose was administered [7]. With pentobarbital anesthesia, all reports agree that the proport ion of raphespinal neurons, raphe neurons with unidentified projections, and/or reticular neurons

that respond to noxious stimuli is enhanced, and the pro- portion of neurons responsive to innocuous stimuli is reduced, compared with neurons studied in awake ani- mals or ~-chloralose anesthetized animals [6, 9, 11, 12]. Similar observations were made with spinal dorsal horn neurons [5]. In addition, four spinal neurons that were classified as LT when the animals were awake became

135

responsive to noxious stimuli, and were thus classified as

WDR, after subsequent administration of pentobarbital

[5]. Clearly, pentobarbital enhances neuronal respon- siveness to noxious stimuli. Perhaps this accounts for the hyperalgesic responses of humans when they receive pen- tobarbital [4, 8].

In the present study, the incidence of spontaneously active neurons was greater with pentobarbital anesthesia than with ~-chloralose anesthesia, a result opposite to a previous report [9]. The difference in results may be ac- counted for by the fact that only neurons responsive to somatosensory stimuli were included in the present re- port. In the study by Light [9], the percentage of sponta- neously active neurons was based o11 both responsive

and unresponsive neurons. In addition to a greater inci- dence of spontaneously active neurons with pentobarbi- tal anesthesia, the mean spontaneous discharge rate was higher with pentobarbital anesthesia compared with ~- chloralose anesthesia. A similar observation has not

been previously reported. Pentobarbital facilitates the opening of chloride channels by activation of GABA re- ceptors, thereby enhancing the inhibitory action of

GABA [10, 14]. Perhaps this effect reduces the activity of neurons that produce inhibitory influences on raphespi- nal neurons, resulting in increased spontaneity.

cz-Chloralose generally enhances reflexes and neuronal responsiveness, effects interpreted to indicate a removal of inhibitory influences that normally would function to

prevent neuronal responses to trivial stimuli [2]. In con- trast, pentobarbital was interpreted as reducing neuronal responsiveness because in awake rats neurons were largely classified as WDR, but with pentobarbital larger proportions of neurons were both LT and HT: that is, LT neurons lost their responsiveness to noxious stimuli, but HT neurons lost their responsiveness to innocuous stimuli [11, 12]. One might expect, then, that ~-chloralose would cause an increased neuronal spontaneity com- pared with pentobarbital, when in fact the opposite oc- curs. Thus, these two anesthetics produce two different states in raphespinal neurons. :z-Chloralose produces a state characterized by low neuronal spontaneity but en- hanced responsiveness to innocuous stimuli. Pentobarbi- tal causes enhanced neuronal spontaneity and respon- siveness to noxious stimuli, but reduced responsiveness to innocuous stimuli.

This work was supported by Grant HL-29618 from NIH. We thank S.L. Jones for helpful editorial sugges- tions.

1 Anderson. S.D., Basbaum, A.I. and Fields, It.L., Response of me- dullary raphe neurons to peripheral stimulation and to systemic opiates, Brain Res.. 123 (1977) 363 36~.

136

2 Balis, G.U. and Monroe, R.R., The pharmacology of chloralose. A review, Psychopharmacologia, 6 (1964) 1 30.

3 Blair, R.W. and Evans, A.R., Responses of medullary raphespmal neurons to electrical stimulation of thoracic sympathetic afferents, vagal afferents, and to other sensory inputs in cats, J. Neurophys- iol., 66 (1991) 2084--2094.

4 Clutton-Brock, J., Pain and the barbiturates, Anaesthesia, 16 ( 1961) 80-88.

5 Collins, J.G. and Ren, K., WDR response profiles of spinal dorsal horn neurons may be unmasked by barbiturate anesthesia, Pain, 28 (1987) 369-378.

6 Eisenhart, S.F., Jr., Morrow, T.J. and Casey, K.L. Sensory and motor properties of bulboreticular and raphe neurons in awake and anesthetized cats. In J.J. Bonica, U. Lindblom and A. lggo (Eds.), Advances in Pain Research and Therapy, Raven, New York, 1983, pp. 161-168.

7 Guilbaud, G., Peschanski, M. and Gautron, M., Functional changes in ventrobasal thalamic neurones responsive to noxious and non-noxious cutaneous stimuli after chloralose treatment: new evidence for the presence of pre-existing 'silent connections' in the adult nervous system? Pain, 11 (1981) 9- 19.

8 Harvey, S.C., Hypnotics and sedatives. The barbiturates. In L.S. Goodman and A. Gilman (Eds.), The Pharmacological Basis ~f

Therapeuticw, Fifth Edition, Macmillan, Ne~ York. 1975, pp ~it2 123.

9 Light, A.R., The spinal terminations of single, physiologically chab acterized axons originating in the pontomedullary raphe of the cat~ J. Comp. Neurol., 234 (1985) 536 548.

10 Macdonald, R.L., Rogers, C.J. and Twyman, R.E~, Barbiturate reg- ulation of kinetic properties of the GABAx receptor channel of mouse spinal neurones in culture, J. Physiol. London. 417 (1989~ 483-600.

I1 Morrow, T.J. and Casey, K.L., Suppression of bulboreticular unit responses to noxious stimuli by analgesic mesencephalic stimula- tion, Somatosens. Res., 1 (1983) 151 168.

12 Oliv~ras, J.-L., Montagne-Clavel, J. and Martin, G., Drastic changes of ventromedial medulla neuronal properties induced by barbiturate anesthesia. I. Comparison of the single-unit types in the same awake and pentobarbital-treated rats, Brain Res., 563 (1991) 241-250.

13 Tsubokawa, T., Yamamoto, T., Katayama, Y~ and Moriyasu, N., Diencephalic modulation of activities of raphe-spinal neurons in the cat, Exp. Neurol., 74 (1981) 561 572.

14 Twyman, R.E., Rogers, C.J. and Macdonald, R.L., DiffErential regulation of 7-aminobutyric acid receptor channels by diazepam and phenobarbital, Ann. Neurol., 25 (1989) 213 220.