Acupuncture-evoked responses of subnucleus reticularis dorsalis neurons in the rat medulla

Neuroscience Vol. 44, No. 3, pp. 693-703, 1991 Printed in Great Britain

0306-4522/91 $3.00 + 0.00 Pergamon Press plc

© 1991 IBRO

ACUPUNCTURE-EVOKED RESPONSES OF SUBNUCLEUS RETICULARIS DORSALIS NEURONS IN

THE RAT MEDULLA

Z. BING,* L. VILLANUEVA and D. LE BARSt INSERM U. 161, 2 rue d'A16sia, 75014 Paris, France

A~traet--Recordings were made from neurons in subnucleus reticularis dorsalis of the rat. Two populations of neurons could be distinguished: those with total nociceptive convergence which were driven by activating A6- and C-fibers from any part of the body and those with partial nociceptive convergence which were driven by activating A 6-fibers from any part of the body or C-fibers from some, mainly contralateral, regions. The effects on subnucleus reticularis dorsalis neurons of manual acupuncture, performed by a traditional Chinese acupuncturist at the "Renzhong' , "Sousanli", "Changqiang", and "Zusanli" acupoints and at a non-acupoint next to "Zusanli", were studied.

Acupuncture stimulation for 30 s at the acupoints or the non-acupoint strongly excited all the total nociceptive convergence neurons tested; these neurons responded with a discharge of rapid onset which was often followed by after-discharges lasting for approximately 30~50 s. The majority but not all of the partial nociceptive convergence neurons responded to 30 s of acupuncture stimulation at the acupoints or the non-acupoint. This was especially the case when the stimulus was applied to contralateral or midline parts of the body. The potency of acupuncture as a means of activating subnucleus reticularis dorsalis neurons varied significantly with the area of the body being stimulated such that: contralateral > midline > ipsilateral areas. The levels of induced activity were of similar magnitude to those evoked by noxious mechanical stimuli applied under identical experimental conditions. No differences were found between the capacities to activate subnucleus reticularis dorsalis neurons of the ~Zusanli'" point and the adjacent non-acupoint, no matter whether these were stimulated ipsi- or contralaterally; this suggests a lack of topographical specificity in the activation of these neurons.

Since subnucleus reticularis dorsalis neurons are activated exclusively or preferentially by noxious inputs, it is concluded that the signals elicited by manual acupuncture travel through pathways responsible for the transmission of nociceptive information. Since acupuncture, a manoeuvre which is known to elicit widespread extrasegmental antinociceptive effects, activates subnucleus reticularis dorsalis neurons which, anatomically, send dense projections to the dorsal horn at all levels of the spinal cord, we would suggest that this structure may be involved not only in signalling pain but also in modulating pain by means of spino-reticulo-spinal feed-back mechanisms.

The existence of physical techniques to relieve pain has been known for centuries 22"33'46 but the under lying neurophysiological mechanisms are poorly under- stood. Amongs t these techniques are those kinds of acupunc ture which are delivered ei ther by low frequency, high intensity electrical st imulation through needles or by man ipu la t ion of the needles. These procedures elicit extrasegmental , widespread and non-selective hypoalgesia (see Refs in 20, 28, 31, 33).

In a previous study in the rat, we reported that manua l acupunc ture elicits s t rong inhibi tory effects on nocicept ive-evoked activities of t r igeminal con- vergent neurons; 7 such neurons are known to play an i m p o r t a n t role in the t ransmiss ion of nociceptive signals towards higher centers in the brain. 4"14"27'34'45"48 These acupuncture-media ted inhibi tory effects had many features in common with the spino-bulbo-spinal-

*Present address: Institute of Acupuncture and Moxibustion, Academy of Traditional Chinese Medicine, 18 Beixincang Street. Dongzhimennei, 100700, Beijing, China.

tTo whom correspondence should be addressed. Abbreviations: DNIC, diffuse noxious inhibitory controls;

PNC, partial nociceptive convergence; PSHs, post-stimulus histograms; SRD, subnucleus reticularis dorsalis; TNC, total nociceptive convergence.

media ted inhibi tory mechanisms which affect the whole popula t ion of spinal and tr igeminal covergent neurons and which have been termed diffuse noxious inhib i tory controls (DNIC) . 13'25'26 Both D N I C and the acupunc ture -evoked inhibi t ions are extrasegmental , have magni tudes and t ime-courses of the same order, exhibi t a lack of topographica l specificity and involve an opioidergic link. 7

In the present study, we investigated the effects of manua l acupunc ture on the activities of medul lary neurons located in subnucleus reticularis dorsalis (SRD) of the rat. 41 We chose S R D neurons because they may play an i m p o r t a n t role in the processing of nociceptive in format ion in view of the facts tha t (i) they are unresponsive to visual, auditory, or pro- prioceptive s t imulat ion, and are act ivated exclusively by cutaneous A6- or A r - and C-fiber per ipheral volleys f rom any par t of the body; 42 (ii) they encode the intensity of electrical, thermal and mechanical cu taneous stimuli within noxious ranges; ~3 (iii) their activities are depressed by systemic morph ine in a dose related and naloxone-reversible fashion; 5 and (iv) the signals responsible for their act ivat ion travel in the lateral par ts of the ventrola tera l quadran t . 6 Finally, using the an te rograde tracer Phaseo lus t, ul-

693

694 Z. BINGe/ al.

garis leucoagglutinin, we have been able to show that SRD neurons send ascending project ions to several structures, including the parafascicular and ventro- medial thalamic nuclei and also descending projections th rough the dorsola tera l funiculus which terminate in the dorsal ho rn at all levels of the spinal c o r d ) Since neurons with similar characteris t ics exist in the medul lary reticular fo rmat ion of the monkey, ~ it is likely tha t S R D is a link of spino-re t iculo- thalamic pa thways and conta ins links for spino-ret iculo-spinal loops. Thus, SR D neurons may par t ic ipate in bo th the t ransmiss ion and modu la t ion of pain signals.

We recorded the activities of S R D neurons and observed the effects on these of manua l acupunc ture per formed by a t radi t ional Chinese acupunctur i s t at different points on the body cor responding to the "classical" acupoints: " R e n z h o n g " , "Sousan l i " , " C h a n g q i a n g " and "Zusan l i " , and at a non-acupo in t next to Zusanli . The exper imental procedure allowed the effects of acupunc ture on the responses of two types of SRD neurons to be compared for s t imula t ion of different areas of the body.

EXPERIMENTAL PROCEDURES

The methods were essentially similar to those described previously. 42

Animal preparation Experiments were performed on 25 Sprague-Dawley

male rats weighing 220-300 g. Following an intraperitoneal injection of 100/~g atropine sulphale, the animals were deeply anesthetized with 2% balothane in a N20-O 2 mixture (2/3: 1/3). A tracheal cannula was inserted, the jugular vein cannulated and the animals were paralysed by intravenous injection of gallamine triethiodide (Flaxedil) and artificially ventilated; tidal CO 2 was monitored using a capnometer (Traverse Medical Monitors, MI) and the rate (70-80 stokes/ rain) and volume of ventilation were adjusted to maintain a normal acid-base equilibrium. Heart rate was monitored continuously and core temperature maintained at 37 + 0.5°C by means of a homeothermic blanket system.

The animals were mounted in a stereotaxic frame with the head fixed in a ventroflexed position by means of a metallic bar cemented to the skull, and the caudal medulla was then exposed by removing the overlying musculature, atlanto- occipital membrane and dura matter.

After surgery, the level of hatothane was reduced to 0.5% to achieve a level of anesthesia which was adequate for ethical purposes but did not excessively depress neuronal responses to noxious stimuli. In this respect, we have reported previously that this anesthetic regime allows a stable level of anaesthesia under which neither electroencephalogram arousal nor cardiovascular reactions are observed during the application of strong stimuli. 2'47

Recordings Unitary extracellular recordings were made with glass

micropipettes (I 0-15 MII) filled with a mixture of 5 % N aCl and Pontamine Sky Blue dye, The single unit activity was amplified and fed into a window discriminator, the output of which was connected to a tape-recorder and a multichannel analyser (Tracor TN 1710), to allow further processing of the data.

The micropipettes were inserted on the left side of the medulla, 1.0-2.0mm caudal to the obex and 0.5-1.5 mm lateral to the midline. Stability for the recordings was achieved by placing over the surface of the medulla, a glass frame which was held in position with a micromanipulator

and 2% Ringer-agar gel. Non-noxious and noxious electrical or mechanical search stimuli were used to help isolate unitary activity, and neurons were classified on the basis of their characteristic responses to different stimuli applied to their peripheral receptive fields. Once a cell had been identified, the extent of its receptive field was determined.

Experimental design As described previously, two populations of SRD

neurons were recorded: neurons with total and neurons with partial nocicpetive convergence (TNC and PNC neurons, respectively). These neurons responded to natural and electrical stimulation of widespread areas of the body. A detailed characterization of their responses was systemat- ically performed using percutaneous electrical stimulation.

Electrical stimuli were delivered through pairs of stainless steel stimulating electrodes inserted subcutaneously into the cheeks, the extremities of the limbs, and the tail. The effects of the repeated application of single, square-wave stimuli (50 trials, 0.66 Hz, 2-ms duration) were analysed using post- stimulus histograms (PSHs) built by means of the multi- channel analyser. One or two components were generally revealed by such analysis and these always had fixed latencies. As described previously, 42 these components are due to the activation of peripheral A6- or Ar- and C-fibers. All the SRD neurons responded with an early (A6) peak of activity from all the areas of the body tested using suprathreshold percutaneous electrical stimuli. The cells were classified as TNC neurons when two peaks of activity (A6, C) were elicited from all the areas of the body. When one or several areas of the body gave rise to only an early (A6) component, the neurons were classified as PNC. Two types of noxious natural stimuli were also employed viz. thermal (immersion of the limbs or tail in a 48°C waterbath) and mechanical (calibrated pinch of the paws or tail).

Acupuncture We chose to compare the effects on SRD neurons of

stimulating acupoints on different parts of the body. These were the "Renzhong", "Sousanli", "changqiang" and "Zusanli" acupoints, as well as a non-acupoint next to "Zusanli", as shown in Fig. 1. "Zusanli" is also called St 36, which belongs to the "Stomach channel of foot-Yangming'" and in man is located 6 cm below the patella and 1 cm below the anterior crest at the tibia. Taking this into account, the equivalent point in the rat is located below the capitulum fibulae and lateral to the tibia (Fig. 1). The "non-acupoint" was chosen beside "Zusanli" in front of the tibia. The "Sousanli" point (the "large intestine channel of hand- yangming", L.I. 10) is the forelimb homologue to "Zusanli" and lies between the ulna and radius. The "Renzhong" point is located in the midline at the junction of the nose and the upper lip. The "Changqiang" point is located on the base of the tail, between the last lumbar vertebra and the coccyx. The depth to which the acupuncture needle was inserted was approximately 0.5-1.0 cm below the entry to the skin. A traditional Chinese acupuncturist performed manual acupuncture by lifting, thrusting and rotating the needle in a clockwise and anti-clockwise fashion (2-4 Hzl,

In each case, quantitative analyses were made by calcu- lating the mean firing rate observed during the 30-s periods of acupuncture. Afterdischarges were expressed in terms of durations and mean firing rates.

Statistical analyses ANOVA analyses were used for statistical purposes, fol-

lowed by post hoe tests when necessary. Data are presented as means_+ S.E.M.

Histological controls At the end of the experiments, selected recorded sites were

marked by electrophoretic deposition of Pontamine Sky Blue to enable histological localization in 100-#m thick frozen

Acupuncture effects on SRD neurons 695

Renzhong ~"

Radius Humerui~a ~

Sousanli

Changqiang . ~ acupoint ~ ~ -

T/l/z ao,i poio, ,,bu,a

Tibia---~ i

Fig. 1. Schematic representation of the experimental design. TNC or PNC neurons were recorded in the left SRD. Anatomical localizations of the acupoints and non-acupoint used in the experiments are shown on the midline ("Renzhong", "Changqiang") or right limbs ("Sousanli", "Zusanli", non-acupoint).

In these latter cases, contra- and ipsilateral points were stimulated (see text).

sections of the lower medulla which were Nissl-stained with Cresyl Violet. Recording sites were then determined by histo- logical examination and camera lucida drawings were made (see Fig. 2).

RESULTS

General characteristics o f the recorded units

A total of 48 units were studied. Histological recov- ery of the dye spots made at the end of the experiments showed that these cells were located within the SRD (Fig. 2). 4' As previously reported, 42 the S R D neurons were activated by percutaneous electrical stimulation of any part of the body. They could be divided into two categories: neurons with TNC, i.e. those neurons driven by A6- and C-fiber activation anywhere on

the body (n = 28) and neurons with PNC, i.e. those neurons driven by A6-fiber activation anywhere on the body, but by C-fiber activation from only some areas (n = 20).

The majority (68%) of T N C neurons displayed spontaneous activity (mean 3.8 _+ 1.0 spikes/s), which consisted of irregular discharges. Forty-three per cent of PNC units were also spontaneously active (mean 2.9 ± 1.2 spikes/s).

All the T N C neurons responded to suprathreshold percutaneous electrical stimuli (2-ms duration) with A6- and C-fiber components no matter which part of the body was stimulated (Fig. 3). The mean threshold for obtaining the C-fiber components from the contralateral hindpaw was 2.9 ± 0.3 mA. The whole populat ion of T N C neurons responded to noxious

696 Z. BING el al.

'~ ~ ' " " i - - " l/ f ~ . o - ~ \ ~ ooi

,~O[lffl

Fig. 2. Locations of the two neuronal types recorded within SRD. Each neuronal type is presented in a single schematic representation of a coronal section of the medulla, 1.5mm caudal to the obex. (A) Location of neurons with TNC--mainly in the dorsomedial part of the SRD. (B) Location of neurons with PNC--mainly in the ventrolateral part of the SRD. (C) Key for anatomic structures: CC, central canal; Cu, nucleus cuneatus; N. caud. V, trigcminaI nucleus caudalis; SRV, subnucleus reticularis ventralis:

ST, solitary tract; Pyr, pyramidal deccusation.

N N

7 5 -

50

25

\

' ~ o ' 0 2 500 ms

Fig. 3. Individual example of responses of a TNC neuron following repetitive supramaximal pcrcutaneous electrical stimulation (0.66 Hz, 2-ms duration, 15 mA) of different areas of the body (arrows). PSHs were built from responses to 50 trials. Note that large A~ and C-fiber responses were evoked from all the areas.

Acupuncture effects o n SRD neurons 697

N N

la_

75-

50

25

• _ --~_tL.a\

0 0 2 0 500 ms

Fig. 4. Individual example of responses of a PNC neuron following repetitive supramaximal percutaneous electrical stimulation (0.66 Hz, 2-ms duration, 25 mA) of different areas of the body (presentation as in Fig. 3). Note that Af-fiber responses were evoked from all the stimulated areas while C-fiber responses

were not evoked from the ipsilateral hindpaw.

mechanical (pinch), and thermal (48 °) stimulation of any part of the body.

All the PNC neurons responded to suprathreshold percutaneous electrical stimulation with an A6-fiber component , no matter which part of the body was stimulated. In addition they all responded with a C-fiber component , at least from some parts of the body, especially those lying contralaterally (Fig. 4); their mean threshold for C-fiber activation from the contralateral hindpaw was higher than that for the T N C neurons (4.0 + 0 . 3 m A ) . The PNC neurons responded to noxious mechanical and thermal stimuli applied to some, especially contralateral, parts of the body.

Responses o[ total nociceptive convergence neurons to acupuncture at acupoints and non-acupoints

A total of 28 T N C neurons were studied during stimulation of several acupoints and the non-acupoint.

All responded to acupuncture stimulation at all these points. An individual example of these responses is shown in Fig. 5: note that no matter whether it was applied to the "Zusanl i" points or non-acupoints of the ipsi- or contralateral hindlimbs, the "Sousanl i" points of the ipsi- or contralateral forelimbs, the "Renzhong" point or the "Changq iang" point, the acupuncture gave rise to discharges with a rapid onset which lasted throughout the period of stimulation. In several cases these responses were followed by long- lasting after-discharges; this was particularly obvious when the contralateral acupoints or non-acupoints were stimulated but was also clear from other points, for example the acupoints of the midline.

At a quantitative level, the responses to the appli- cation of acupuncture were evaluated during the 30-s period of stimulation. As shown in Table I A, the mean firing was in the 10-16 spikes/s range with a tendancy for larger responses with contralateral stimuli. There

698 Z. BING et al.

25 Hz

/ . . ¸

Renzhong ~X~ 1 min

- - f - - Sousanli Sousanli

Z'--"~sanli / I~ ~ Zusanli

m n

Non-acupoint Changcliang Non-acupoint Fig. 5. Individual example of responses of a TNC neuron, elicited by 30 s o f acupuncture stimulation at acupoints and non-acupoints located in different parts of the body (arrows). Note that needle stimulation at all the points (including the non-acupoints) induced strong neuronal activity followed by long-tasting

after-discharges.

Table 1. Comparison of the responses of total nociceptive convergence (A) and partial nociceptive convergence (B) neurons to various acupuncture conditioning stimuli

(A) TNC neurons (B) PNC neurOns

Points ipsi- contra- ipsi- contra Renzhong 13.9 + 1.2 10.7 4- 1.4 Sousanli 12.1 + 1.6 16.4 _ 2.4 5.2 -t- 1.8 15.7 4- 2.6 Zusanli 11.3 + 2.7 13.7 + 1.8 3.5 _ 0.7 13.0 +__ 1.8 Non-acupoint 10.5 + 2.1 15.4 _+ 1.7 3.5 __+ 1.0 14.1 + 1.8 Changqiang 10.2 + 1.3 7.4 4- 1.6

Results are expressed in terms of the mean firing rate (spikes/s) observed during the 30-s duration needling. Needling induced more firing when stimuli were applied to the contralateral points than to their ipsilateral counterparts, especially for PNC neurons (P < 0.014).001). All the responses were higher for TNC neurons than for PNC neurons with the corresponding stimuli, especially when the ipsilateral points were stimulated (see Table 2 for global statistical analysis).

Acupuncture effects on SRD neurons 699

1 min j ~ 25 Hz f

~ Renzhong

Sousanli Sousanli

~ L zusanl i

m m m

Non-acupoint Changqiang Non-acupoint Fig. 6. Activation of a PNC neuron by acupuncture stimulation for 30 s at acupoints and non-acupoints in different areas of the body (arrows). Note that the responses were weak when the stimulus was applied to an ipsilateral point; by contrast, long-lasting evoked responses were obvious when contralateral or

midline points were stimulated.

were no significant differences between the activity evoked from the "Zusanl i" and that from the non- acupoints ipsi- or contralaterally. The after-discharges which followed the cessation of acupuncture lasted 30-60 s when the contralateral acupoints or non- acupoint were stimulated (mean discharges in the 2 5 spikes/s range) and 5-30 s for the other acupoints (mean discharges of 2-3 spikes/s).

Responses o f partial nociceptive convergence neurons to acupuncture at acupoints and non-acupoints

A total of 20 PNC neurons were studied. The majority responded to acupuncture from some acu- points, especially those located on the contralateral and midline parts of the body. Figure 6 provides an individual example in which acupuncture applied to the contralateral "Zusanl i" point or "non-acupoin t" ,

the "Sousanl i" point, the "Renzhong" or "Chang- qiang" points resulted in high levels of neuronal firing. An identical stimulus applied ipsilaterally to the "Zusanli", "Sousanli" or "non-acupoint" gave rise to only a very small number of spikes. After-discharges were observed only when the contralateral acupoints and non-acupoints were stimulated. By contrast, either rapidly adapting, small responses or no responses at all were observed when acupuncture was applied to the ipsilateral acupoints or non-acupoints.

The responses to the application of acupuncture were evaluated at a qualitative level during the 30-s periods of conditioning stimulation. As shown in Table 1B, the results were heterogeneous in relation to the part of the body being stimulated: firing was in the 13-16, 7 10 and 3-5 spikes/s ranges, respectively when contralateral, medial and ipsilateral parts of the

700 Z. BING et aL

i i

i

i

SOUlll'dl

l, loni

20,

i

Fllmlhong

Ol~mllqhnll

!

i

i

l~._-~,__-~. II

Zu~nll

Fig. 7. Summary of the global output from SRD neurons during acupuncture. Histograms represent the mean firing of the whole population of SRD neurons recorded (n = 48) during acupuncture applied to body areas ipsilateral, medial or contralateral to the recording site. Note that SRD activation presents

the following order of magnitude: contralateral > midline > ipsilateral.

body were stimulated. Again, there were no signiticant differences between the levels of activity evoked from "Zusanli" and from the non-acupoints. The after- discharges which followed the cessation of acupunc- ture lasted about 30 s for the contralateral acupoints and non-acupoint (discharges in the 2-5 spikes/s range). In most cases, the after-discharges were not obvious following stimulation of the ipsilateral acupoints or non-acupoint.

It is important to note that all these responses were weaker than the corresponding values observed for

TNC neurons (Table 1) especially for the ipsilateral points (>2-fold) (P < 0:05 or P < 0.01).

Global output of subnucleus reticularis dorsalis The global output of SRD in terms of the mean

firing rates of all the recorded neurons is summarized in Fig. 7. Post hoc tests following ANOVA analysis (Table 2) revealed that the potency of acupuncture to activate SRD neurons varies significantly with the part of the body being stimulated in the following order: contralateral > midline > ipsilateral areas. By

Acupuncture effects on SRD neurons 701

Table 2. ANOVA of the effects of acupuncture on subnucleus reticularis dorsalis neurons Sources of Sum of Degrees of Mean variation squares f r e e d o m squares F Significance Type ofcell 730016.29 1 730016.29 18.67 P < 0.001 Part ofthe body 1791639.32 2 895819.66 22.91 P < 0.001 Points 202777.79 4 50694.45 1.29 n.s. Error 7232872.61 185 39096.61 Total 9957306.01 195 Three factors of variation were considered: (i) the type of cell, namely TNC or PNC

neurons; (ii) the part of the body stimulated, namely the ipsilatera/, midline or contralateral areas and (iii) the five points stimulated (see Fig. 1).

contrast, there was no significant difference between points to activate SRD neurons (Table 2).

DISCUSSION

The present study has demonstrated that, in the rat, a large population of bulbar neurons located within SRD conveys information set up by the stimulation of acupoints and non-acupoints on different parts of the body.

We have recently suggested that SRD neurons might play an important role in the processing of cutaneous and visceral nociceptive information. 37'42 While SRD neurons do not respond to heterosensory (flashes, whistle sounds) or proprioceptive stimuli and very few respond to innocuous cutaneous stimuli, 42 they are all activated either preferentially or specifically by noxious inputs from all areas of the body; in addi- tion they encode precisely the strength of peripheral stimuli within the noxious range. 43

In the present study all the neurons fell into one of the two already defined categories: neurons with TNC, i.e. those neurons driven by A6- and C-fibers from the whole body or neurons with PNC, i.e. those neurons driven by At-fibers from the whole body, and by C-fibers from some parts of the body. Comparison of the effects of acupuncture and applications of noxious pinch on homologous parts of the body under identical experimental conditions 42 lead to the conclusion that these two types of stimulation have similar potencies in evoking SRD neuronal activities.

We are therefore dealing with neurons which re- spond both to noxious stimuli and to the manipulation of acupuncture needles, which suggests that the SRD may be involved in both pain transmission and pain modulation. In keeping with the present results, a Chinese group 24 has reported that in the cat, neurons which they termed as convergent pain-sensitive cells but which would seem to be equivalent to SRD neurons in the rat, were activated by stimulating not only the dental pulp and others parts of the body including the sciatic nerve but also some acupoints; interestingly, these neurons were driven by A6 and C-fibers from some parts of the body.

Several authors have reported that for a reliable pain-relieving effect to occur, acupuncture stimulation must be as strong as the patient will tolerate, and possibly even painful. ~'j6'3236 The aim with such stimu-

lation is to elicit the "needling sensation" or "techi"-- a combination of the ache which one might experience in muscular fatigue, with numbness, distention and a heaviness sometimes associated with soreness, tingling or warmth) '3~ In this respect, acupuncture stimulation activates both A t / C cutaneous and groups III and IV muscular afferent fibers) °'23 These clinical and exper- imental observations support the proposal that signals produced by acupuncture manipulations are trans- mitted through thin peripheral fibers identical or simi- lar to those responsible for nociceptive inputs. ~'~'2t'4~ The fact that such signals then travel through central pathways responsible for the transmission of noci- ceptive information from the spinal cord to the brain is suggested by several lines of evidence; in particular, the anterolateral quadrant has been shown to be essential for triggering analgesia by means of acu- puncture.~'39 In keeping with this notion, the ascend- ing pathways in the spinal cord involved in the activation of SRD neurons by noxious cutaneous stimuli are confined to the anterolateral quadrant. 6

There were obvious differences amongst the re- sponses of SRD neurons to acupuncture stimulation but these differences were based only on the types of neuron (i.e. TNC vs PNC neurons) and the parts of the body being stimulated (i.e. ipsi- versus contra- lateral). The greater activation of TNC than of PNC neurons was repeatedly observed especially when the ipsilateral or midline parts of the body were stimulated.

The responses of TNC neurons were slightly weaker from ipsilateral parts of the body than from their contralateral counterparts. A similar but much more marked lateralization was found for the PNC neurons: their responses to contralateral stimuli were three to five times greater than to the corresponding ipsilateral stimuli. These differences are very much in keeping with the following facts: (i) the ascending fibers which activate SRD neurons travel mainly in the contralateral part of the ventrolateral quadrant; 6 (ii) a majority of PNC neurons do not respond to the activation of C-fibers from ipsilateral parts of the body; (iii) thresholds for obtaining A6- and C-fiber responses to electrical stimulation are lower for TNC than for PNC neurons, with the difference being particularly marked for ipsilateral parts of the body; (iv) the responses of PNC neurons to electrical, thermal and mechanical stimuli are weaker than those

702 Z. BING et al.

of TNC neurons, especially when ipsilateral parts of the body are stimulated. 42

An interesting observation emerges from the com- parison of the responses of SRD neurons to stimu- lation of the "Zusanl i" point and of the surrounding non-acupoint: no matter whether stimuli were applied to the ipsi- or contralateral side of the body or whether TNC or PNC neurons were concerned, we never found any difference between the responses to the two types of stimulus. Interestingly, we have also previously found a lack of topographical specificity for the anti- nociceptive effects of acupuncture on trigeminal con- vergent neurons.7 Interestingly, a lack of topographical specificity for the analgesic effects of acupuncture has been reported in controlled clinical trials: pain was relieved to similar extents regardless of whether acu- puncture was applied to an acupoint or to a close, non-acupoin t ) 2,~5'17'1s In addition, controlled clinical studies have not found any difference between the efficacy of acupuncture applied near to the painful focus and acupuncture applied to a distant extra- segmental region of the body. 19'29'35 In this context, as was stated by MacDonald, 3~ it is difficult to find a site which is not close to an acupoint, the number of which has grown with time and is presently 747 in man. 38 Our results therefore contribute to a growing body of evidence, which suggests that acupoints are non-specific in terms of efficacy, at least as far as hypoalgesia is concerned.

At first sight, the convergence of both acupuncture- induced and nociceptive information onto single SRD neurons might appear puzzling. However, one can suggest a role for this neuronal type in both nocicep- tion and antinociception by means of spino-reticulo- spinal feedback loops and there are anatomical data

to support this hypothesis. In keeping with Lima, 3'j we recently observed the labeling of numerous dorsal horn neurons at various levels of the spinal cord following small injections of a retrograde tracer within the SRD (unpublished observations), thus providing evidence for the ascending limb of the loop. Small injections of the anterograde tracer, Phaseolus vulgaris leucoagglutinin, showed that SRD neurons send descending projections through the dorsolateral funiculus which terminate principally in laminae V, VI, VII and X at all levels of the spinal cord; 3 these data provide evidence for the descending limb of the loop. In keeping with these results, it has been shown that the inhibitory effect of acupuncture on viscero- somatic reflexes is abolished by spinal lesions only if they include the dorsolateral funiculusfl 9 Interest- ingly, some of the laminae reached by SRD projections contain large numbers of convergent neurons which convey nociceptive information and are inhibited both by heterotopic noxious s t imula t ion - -DNIC ~s'2s'26 and by acupuncture whether applied at an acupoint or a non-acupoin t )

Obviously, further studies will be necessary to estab- lish the existence and significance of these feedback loops. In particular one intriguing question remains, namely, what type(s) of interaction occur between acupuncture-induced and nociceptive signals onto individual SRD neurons?

Acknowledgements--We thank Dr S. W. Cadden for advice in the preparation of the manuscript, J. Carrou6 for the histology, E. Dehausse for drawings and photography and M. Cayla for secretarial help. This work was supported by INSERM and la Direction des Recherches et Etudes Techniques (DRET).

REFERENCES

1. Andersson S. A. (1979) Pain control by sensory stimulation. In Advances in Pain Research and Therapy (eds Bonica J. J., Liebeskind J. C. and Albe-Fessard D.), Vol. 3, pp. 569-585. Raven Press, New York.

2. Benoist J. M., Kayser V., Gautron M. and Guilbaud G. (1984) Letter to the editor. Pain 18, 410-411. 3. Bernard J. F., Villanueva L., Carroue J. and Le Bars D. (1990) Efferent projections from the subnucleus reticularis

dorsalis (SRD): a Phaseolus vulgaris leucoagglutinim study in the rat. Neurosci. Lett. 116, 257-262. 4. Besson J. M. and Chaouch A. (1987) Peripheral and spinal mechanisms of no¢iception. Physiol. Rev. 67, 67-186. 5. Bing Z., Villanueva L. and Le Bars D. (1989) Effects of morphine upon Ar- and C-fibre evoked activities of subnucleus

reticularis dorsalis (SRD) neurones in the rat medulla. Eur. J. Pharmac. 164, 85-92. 6. Bing Z., Villanueva L. and Le Bars D. (1990) Ascending pathways in the spinal cord involved in the activation of

subnucleus reticularis dorsalis neurons in the medulla of the rat. J. Neurophysiol. 63, 424-438. 7. Bimg Z., ViUaneava L. and Le Bars D. (1990) Acupuncture and diffuse noxious inhibitory controls: naloxone-reversible

depression of activities of trigeminal convergent neurons. Neuroscience, 37, 809-818. 8. Burgess P. R. and Perl E. R. (1973) Cutaneous mechanoreceptors and nociceptors. In Handbook o f Sensory Physiology

(ed. Iggo A.), pp. 29-78. Springer, Bedim. 9. Campbell J, N., Raja S. N., Cohen R. H., Manning D. C., Khan A. A. and Meyer R. A. (t989) Peripheral neural

mechanisms of nociception. In Textbook o f Pain (eds Wall P. D. and Melzack R.), pp. 24-45. Churchill Livingstone, Edinburgh.

10. Chen L., Tang J. and Yan J. (1981) Analysis of afferent fibers for the impulses of acupuncture analgesia. Kexue Tongbao 26, 564-569.

11. Chiang C. Y., Liu J. Y., Chu T. H., Pai Y. H. and Chang S. C. (1975) Studies on spinal ascending pathway for effect of acupuncture analgesia in rabbits. Sci. Sin. XVIII, 651-658.

12. Co L. L., Schmitz T. H., Havdala H., Reyes A. and Westerman M. P. (1979) Acupuncture: an evaluation in painful crises of sickle cell anaemia. Pain 7, 181-185.

13. Dickenson A. H., Le Bars D. and Besson J. M. (1980) Diffuse noxious inhibitory controls (DNIC). Effects on trigeminal nucleus caudalis neurones in the rat. Brain Res. 200, 293-305.

14. Dubner R. and Bennett G. J. (1983) Spinal and trigeminal mechanisms of nociception. A. Rev. Neurosci. 6, 381-418.

Acupuncture effects on SRD neurons 703

15. Edelist G., Gross A. E. and Langer F. (1976) Treatment of low back pain with acupuncture. Can. Anaesth. Soc. J. 23, 303-306.

16. Eriksson M. and Sjolund B. (1976) Acupuncture-like electroanalgesia in TNS-resistant chronic pain. In Sensory Functions of the Skin in Primates (ed. Zotterman Y.), pp. 575-581. Pergamon Press, Oxford.

17. Gaw A. C., Chang L. W. and Shaw L. C. (1975) Efficacy of acupuncture on osteoarthritic pain. New Eng. J. Med. 293, 375 378.

18. Ghia J. N., Mao W., Toomey T. C. and Gregg J. M. (1976) Acupuncture and chronic pain mechanisms. Pain 2, 285 299. 19. Godfrey C. M, and Morgan P. (1978) A controlled trial of the theory of acupuncture in musculoskeletal pain. J. Rheum.

5, 121-124. 20. Han J. S. and Terenius L. (1982) Neurochemical basis of acupuncture analgesia. A. Rev. Pharmac. Toxic. 22, 193 220. 21. Iggo A. (1985) Sensory receptors in the skin of mammals and their sensory functions. Rev. Neurol. 141, 599 613. 22. Kane K. and Taub A. (1975) A history of local electrical analgesia. Pain 1, 125 138. 23. Kawakita K. and Funakoshi M. (1982) Suppression of the jaw-opening reflex by conditioning A-delta fiber stimulation

and electroacupuncture in the rat. Expl Neurol. 78, 461 465. 24. Kirin Medical College, Changchun (1977) The inhibition effect and the mode of action of electroacupuncture upon

discharges from the pain-sensitive cells in spinal trigeminal nucleus. Sci. Sin. XX, 485-501. 25. Le Bars D., Dickenson A. H. and Besson J. M. (1979a) Diffuse noxious inhibitory controls (DNIC). 1--Effects on dorsal

horn convergent neurones in the rat. Pain 6, 283-304. 26. Le Bars D., Dickenson A. H. and Besson J. M. (1979b) Diffuse noxious inhibitory controls (DNIC). II--Lack of effect

on non convergent neurones, supraspinal involvement and theoretical implications, Pain 6, 305 327. 27. Le Bars D., Dickenson A. H., Besson J. M. and Villanueva L. (1986) Aspects of sensory processing through convergent

neurons. In Spinal Afferent Processing (ed. Yaksh T. L.), pp. 467 504. Plenum Press, New York. 28. Le Bars D., Willer J. C., De Broucker T. and Villanueva L. (1989) Neurophysiological mechanisms involved in the

pain-relieving effects of counter-irritation and related techniques including acupuncture. In Scientific Bases of Acupuncture (eds Pomerantz B. and Stux G.), pp. 79-112. Springer, Berlin.

29. Lee P. K., Andersen T. W., Modell J. H. and Saga S. A. (1975) Treatment of chronic pain with acupuncture. J. Am. reed. Ass. 232, 1133-1135.

30. Lima D. (1990) A spinomedullary projection terminating in the dorsal reticular nucleus of the rat. Neuroscience 34, 577- 590.

31. MacDonald A. J. R. (1989) Acupuncture analgesia and therapy. In Textbook of Pain (eds Wall P. D. and Melzack R.), pp. 906-919. Churchill Livingstone, Edinburgh.

32. Mann F. (1974) Acupuncture analgesia, report of 100 experiments. Br. J. Anaesth. 46, 361-364. 33. Melzack R. (1989) Folk medicine and the sensory modulation of pain. In Textbook of Pain (eds Wall P. D. and Melzack

R.), pp. 897-905. Churchill Livingstone, Edinburgh. 34. Melzack R. and Wall P. D. (1965) Pain mechanisms: a new theory. Science 150, 971-979. 35. Mendelson G., Selwood T. S., Kranz H., Kidson M. A. and Scott D. S. (1983) Acupuncture treatment of chronic pain.

A double blind placebo controlled trial. Am. J. Med. 74, 49-55. 36. Nathan P. W. (1978) Acupuncture analgesia. Trends" Neurosci. 1, 21 23. 37. Roy J. C., Bing Z., Villanueva L. and Le Bars D. (1990) Activation of subnucleus reticularis dorsalis neurones (SRD)

of the rat medulla by colorectal distention. Pain suppl. 5, S166. 38. Shanghai College of Traditional Medicine (1981) Acupuncture: A Comprehensive Text. Eastland Press, Chicago. 39. Shen E., Tsai T. T. and Lan C. (1975) Supraspinal participation in the inhibitory effect of acupuncture on

viscero-somatic reflex discharges. Chin. Med. J. 6, 431 440. 40. Torebjork E. (1985) Nociceptor activation and pain. Phil. Trans. R. Soc. Lond. 308, 227 234. 41. Valverde F. (1962) Reticular formation of the albino rat's brain stem. Cytoarchitecture and corticofugal connections.

J. comp. Neurol. 119, 25-49. 42. Villanueva L,, Bouhassira D., Bing Z. and Le Bars D. (1988) Convergence of heterotopic nociceptive information onto

subnucleus reticularis dorsalis neurones in the rat medulla. J. Neurophysiol. 60, 980 1009. 43. Villanueva L., Bing Z., Bouhassira D. and Le Bars D. (1989) Encoding of electrical, thermal and mechanical noxious

stimuli by subnucleus reticularis dorsalis neurons in the rat medulla. J. Neurophysiol. 61, 391 402. 44. Villanueva L., Cliffer K. D., Sorkin L. S., Le Bars D. and Willis W. D. (1990) Convergence of heterotopic nociceptive

information onto neurons of caudal medullary reticular formation in monkey (Macaca fascicularis). J. Neurophysiol. 63, 1118--1127.

45. Wall P. D. (1978) The gate control theory of pain mechanisms. Brain 101, 1 18. 46. Wand-Tetley J. I. (1956) Historical methods of counter-irritation. Ann. Phys. Med. 3, 90-98. 47. Weil-Fugazza J., Godefroy F. and Le Bars D. (1984) Increase in 5-HT synthesis in the dorsal part of the spinal cord,

induced by a nociceptive stimulus: blockade by morphine. Brain Res. 297, 247-264. 48. Willis W. D. and Coggeshall R. E. (1978) Sensory Mechanisms of the Spinal Cord. Plenum Press, New York.

(Accepted 22 April 1991)