INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Numerous retrograde and anterograde anatomic data have clearly demonstrated that the parabrachial (PB) areaa group of neurons that surround the superior cerebellar peduncle at the ponto-mesencephalic junctionis one of the major projection targets for the neurons located in the most superficial laminae (I and II outer, or to simplify: lamina I) of the dorsal horn (Bernard et al. 1989, 1995; Blomqvist et al. 1989; Burton et al. 1979; Cechetto et al. 1985; Craig 1995; Feil and Herbert 1995; Hylden et al. 1989; Kitamura et al. 1993; Lima and Coimbra 1989; Menétrey and De Pommery 1991; Menétrey et al. 1992; Panneton and Burton 1985; Slugg and Light 1994; Wiberg and Blomqvist 1984; Wiberg et al. 1987; Yamada and Kitamura 1992; see also Menétrey et al. 1982; Saper 1995; Swett et al. 1985).

The superficial laminae of the spinal cord receive heavy and direct input from the thin primary afferents A and C fibers and contain primarily nociceptive neurons some of which being identified as projecting to the thalamus (Besson and Chaouch 1987; Brown and Culberson 1981; Cervero and Connell 1984; Culberson and Brown 1984; Fitzgerald 1984; Light and Perl 1979; Mizumura et al. 1993; Sugiura et al. 1986, 1988, 1989, 1993; Woolf and Fitzgerald 1983, 1986). More recently, several studies also have demonstrated that a high number of nociceptive neurons in the superficial laminae project to the PB area (Hylden et al. 1985, 1986a,b; Light et al. 1987, 1993; see also McMahon and Wall 1983).

These studies, as a whole, strongly support a substantial noxious input from the spinal cord to the PB area. The effect of this noxious input was observed directly at the PB level. Several electrophysiological studies have recorded the unitary activity of PB neurons, a large number of which respond specifically to somatosensory stimuli in the noxious range irrespective of whether these neurons were antidromically driven from the central nucleus (Ce) of the amygdala or from the ventromedial (VMH) nucleus of the hypothalamus or whether their projecting target was not determined (Bernard and Besson 1988, 1990; Bernard et al. 1994; Bester et al. 1995b; Hayashi and Tabata 1990; Huang et al. 1993; Matsumoto et al. 1996; Menendez et al. 1996; Slugg and Light 1990). Further indirect evidence for the involvement of the PB area (Bellavance and Beitz 1996; Bullitt 1990; Clement et al. 1996; Hamba et al. 1994; Hermanson and Blomqvist 1996; Lantéri-Minet et al. 1993, 1994; Pertovaara et al. 1993) and spino-PB pathway (Bester et al. 1997b; Buritova et al. 1998) in nociception was obtained from c-Fos studies. Because the nociceptive region of the PB area is involved in numerous autonomic regulatory functions and densely projects to the Ce and the hypothalamus, the spino-PB pathway is thought to be involved in autonomic and emotional/aversive reactions to painful stimuli (see refs and discussion in Bernard and Besson 1990; Bester et al. 1995b).

Paradoxically, at least in the rat, the properties of PB neurons now are known more precisely than those of their main input, the spino-PB neurons. More specifically, their precise encoding properties to mechanical and thermal stimuli and the effect of stimuli applied outside the excitatory receptive field (heterotopic stimuli) of a consistent subgroup of PB neurons have been investigated in some detail (Bernard and Besson 1990; Bernard et al. 1994; Bester et al. 1995b). To date, these properties have not been investigated for the spino-PB neurons. Although a study in the rat investigated the main characteristics of cells projecting through the dorsolateral funiculus (McMahon and Wall 1983), it did not give details concerning spino-PB neurons. This issue has been considered in the cat but neither the encoding properties nor the effect of heterotopic stimuli were determined (Hylden et al. 1985, 1986a,b; Light et al. 1987, 1993). Indeed, the encoding properties of superficial laminae neurons only have been investigated precisely for spino-thalamic neurons or for neurons without identified projection in the cat and monkey (Bushnell et al. 1984; Chung et al. 1979; Craig and Kniffki 1985; Craig and Serrano 1994; Ferrington et al. 1987; Hayes et al. 1981; Hoffman et al.,1981; Kenshalo et al. 1979; Light et al. 1993; McHaffie et al. 1994; Price et al. 1976, 1978). The aim of the present study was to provide a detailed electrophysiological characterization of the responses of the lamina I spinal neurons backfired antidromically from the contralateral PB area. Furthermore the present investigation should permit an accurate comparison between spino-PB and PB neurons because the studies were made under identical experimental conditions. Part of these results have been presented in abstract form (Bester et al. 1995a).

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Animal preparation

Experiments were performed on 53 Sprague-Dawley male rats weighing 250-275 g. In an anesthetizing box, animals were deeply anesthetized with 2% halothane (Belamont, France) in a nitrous oxide-oxygen mixture (2/3-1/3). This anesthesia regimen was maintained throughout the surgery procedures. A tracheal cannula was inserted, the jugular vein was cannulated, and then animals were paralyzed with a continuous intravenous injection of gallamine triethiodide (Flaxedil) and artificially ventilated using a Palmer pump. The expiratory CO₂ was monitored continuously using a Capnomac (Datex Instruments) throughout the experiment, and the end-tidal CO₂ was maintained at ~4%. Neither blood pressure nor electreoencephalogram (EEG) were monitored because these same conditions have been found to provide reproducibly stable and reliable anesthesia. Use of these conditions enables comparisons across all prior studies in this laboratory. The level of halothane, O₂, and N₂O also was checked systematically during the experimental procedure. The heart rate was monitored continuously, and the core temperature maintained at 37 ± 0.5°C using a homeothermic blanket system.

The animals were mounted in a stereotaxic frame. The head was fixed in a dorsiflexed position with the incisor bar elevated 10 mm above the standard position (Paxinos and Watson 1986). A craniotomy was performed at the parieto-occipital level on the right side for the passage of the stimulating electrodes targeting the PB area. A second small craniotomy was performed rostrally in the left frontal bone to implant the reference electrode for central stimulation. Vertebras L_1-3 were exposed and fixed in "spinal hooks." A laminectomy then was performed to expose the L_4-5 spinal segments. After surgery, the halothane level was reduced to 0.5%, and the gas mixture of nitrous oxide-oxygen was maintained at 2/3-1/3. This level of anesthesia did not excessively depress neuronal responses to noxious stimuli and was adequate in terms of ethical purposes (Benoist et al. 1984; Le Bars et al. 1980). These conditions were in agreement with the ethical guidelines of the International Association for the Study of Pain (IASP) for investigations of experimental animals (Zimmermann 1983) and received the approval from the local ethical committee.

Recordings

Extracellular unit recordings made with metal microelectrodes (9-12 M), tungsten or stainless steel (Frederic Haer). Single-unit activity was fed into a window discriminator, the output of which was stored on a digital audiotape (Biologic DTR 1200). A multichannel analyzer was used to perform further data processing.

The electrode was inserted vertically in the left spinal cord close to the surface. From this zero point, the search of a neuron was started by driving the electrode through the dorsal horn. Each time only one spike or any change in the background activity was observed, the response to electrical stimuli delivered in the PB area was tested. If no response was detected, the effects of high-intensity current delivered on different distal sites of the hindpaw were tried rapidly. If the presence of a unit was confirmed, antidromic identification was tested by applying electrical stimuli in the PB area at different depths from each of the three stimulating electrodes (see following text). When antidromic activation was obtained, the cell responses to different cutaneous stimuli (electrical, mechanical, and thermal) were tested. Only one neuron was studied per animal to avoid sensitization.

Response to PB area stimulation

Stimulations of the PB area were applied with a linear array of three concentric monopolar electrodes (model No. SNE-300 modified; insulated contact protrudes from shaft, 120-µm diam, 5-mm length; exposed contact, 100-µm diam, 150-µm length; Rhodes Medical Instruments) separated by 600 µm from each another. The set of three electrodes was inserted into the right PB area with the following coordinates for the middle electrode: 1.9 mm rostral to the lambda and 1.8 mm lateral to the midline. The depth was 7,000 µm from the brain surface. The three center contacts could be independently stimulated (1-mA square wave negative current, 0.2-ms duration).

When, in the spinal cord, a unit was backfired from the PB area, the stimulation thresholds were measured for each electrode. The site of minimum threshold was determined by repeating this measure at different depths (above and below). Finally, the set of electrodes was positioned at the depth where the lowest threshold had been observed, and we used the electrode from which this lowest threshold was obtained. Then the intensity of the central stimulation was set at twice the threshold to test the criteria of antidromic activation: 1) the lack of variance (<100 µs) of the spike latency (t), 2) the ability of the evoked response to follow a train of high-frequency stimulation (>300 Hz, for 5 stimuli). This test was coupled with the measure of the refractory period (R), determined as the shortest interval between two stimuli required to produce two antidromic spikes. The refractory period always corresponded to a higher frequency (1/r > 500 Hz). And 3) the observation of a systematic collision between an orthodromic spike (spontaneous or electrically evoked from the receptive field) and the antidromically evoked response in the 2t + R period.

Response to transcutaneous electrical stimulation

Electrical square-wave stimuli (2-ms duration) were delivered through pairs of stainless steel stimulating electrodes inserted subcutaneously in the toes. We usually observed that a high-intensity stimulation evoked two periods of activation. The threshold of each peak of activation was determined by a progressive increase in the intensity of stimulation from a 0 starting point. The effects of repetitive electrical stimuli (16 trials at a frequency of 0.5 Hz) at an intensity corresponding to threefold the highest threshold (i.e., that evoking the late peak of activation) were analyzed by building poststimulus time histograms (PSTHs) off-line on a MacIntosh 7100, with the Spike2 software and CED 1401 hardware (CED, Cambridge England). The length of the sciatic nerve was measured at 15 cm from the tip of the toes (III-V) to the dorsal horn entry zone of the L_4-5 lumbar segments. The conduction velocity of the peripheral fibers then was calculated as follows: 150 mm/[observed latency of the first spike in each individual peak  1 ms]. The subtracted 1 ms corresponds to an estimate of the synaptic delay, assuming that spino-PB neurons were monosynaptically activated. All in all, our results happened to be in good agreement, a posteriori, with data obtained directly from peripheral fibers in the rat foot (Leem et al. 1993a,b), validating thus the assumptions and the methods.

Response to natural cutaneous stimulation

INNOCUOUS STIMULI. Innocuous stimuli were applied to the receptive field of the recorded neuron. They consisted of mechanical light touch and brush (paintbrush Gerand Séries SD 1, No. 6) or thermal stimuli applied with a thermostatically controlled water jet (20-40°C). To apply thermal stimuli, a 50-ml syringe was filled with water just above the desired temperature. When ejected through a 12-gauge needle, the stimulating water jet maintained a temperature (measured from a thermocouple placed next to the receptive field) at the desired level over the 20 s of stimulation. From 20 to 30°C, 5°C steps were used, then 2°C steps were used from 30 up to 50°C. In the absence of an evoked response, intervals between the discrete stimuli were of 2 min. All stimuli were applied during a 20-s period.

NOXIOUS MECHANICAL STIMULI. With a calibrated forceps (strain gauge on the arm connected to an amplifier: Hottinger, Baldwin Messtechnick, Darmstadt, Germany), we manually applied serially increasing pressures of 20-s duration (5, 10, 20, 40, 60, and 80 N · cm²; 1 N · cm² = 10 kPa  100 g · cm²) on the receptive field of the spino-PB neurons. The stimulating area was 10 mm². The level (intensity) of the applied pressure was read, on-line, on the digital screen (4 digits) of the apparatus, allowing the trained experimenter to maintain, with a feedback reaction, an almost constant pressure (variation <10%) over the stimulating period. Between two stimulations, an interval of 5 min was observed, to avoid fatigue and/or sensitization of peripheral receptors.

NOXIOUS THERMAL STIMULI. Both noxious heat and noxious cold stimuli were applied using a thermostatically controlled water jet, directly on the receptive field, for 20 s periods. With the approach used for innocuous thermal stimuli, scanning the effect of heat stimuli toward the noxious range consisted of applications of discrete water jets of increasing temperatures from innocuous 30°C up to highly noxious 52°C in 2°C steps. Similarly, cold stimuli consisted of applications of discrete water jets of decreasing temperatures from innocuous 30°C down to highly noxious 15°C, with a 5°C step. In this case, the stimulating fluid was an alcohol-water mixture that was set just below the desired temperature. This procedure allowed the desired temperature to be achieved at the level of the receptive field (measured with a thermocouple) for the 20-s stimulation period. For both heat and cold stimuli, a minimum interval of 5 min (and 10 min for higher intensity stimuli) was observed, to avoid fatigue and/or sensitization of peripheral receptors.

Effects of heterotopic noxious stimulation on a noxiously evoked response

On 16 spino-PB neurons, we have tested the effects of a noxious stimulus applied far outside the receptive field (heterotopic stimuli) on the noxiously evoked responses of the spino-PB neurons. Briefly, on a standard response to noxious stimuli (50°C, or 60 N · cm²) applied on the receptive field of a spino-PB neuron (toe or pad), we have tested the effects of a noxious stimuli (50°C, or 60 N · cm²) applied far outside the receptive field, basically, a forepaw (and in some cases the tongue). The heterotopic stimulation started 10 s before the noxious stimulation of the receptive field, continued throughout this stimulation (20 s), and was prolonged 10 s after the end of the noxious stimulation of the receptive field (total duration of the heterotopic stimulus: 40 s).

Histological controls

For all the experiments, the location of the PB electrode tip that provided the lowest threshold to obtain an antidromic spike was marked by passing a 10 µA DC for 30 s (positivity to the electrode). This procedure allowed a small electrophoretic iron deposit at the tip of the stimulating electrode.

In a first series of 28 experiments, we used stainless steel recording electrodes. At the end of each experiment, the depth of the electrode was noted, and a small deposit of iron was made at the tip of the electrode by passing a 10 µA DC current (positivity to the electrode) during 30 s.

After the iron deposits, animals were killed by an overdose of intravenous sodium pentobarbitone. The brain and lumbar spinal cord were removed and fixed for 5 days in a mixture that would label the iron with a Prussian blue color (1 volume of 1% potassium ferrocyanide in 10% formalin solution added to 2 volumes of 2% acetic acid in 95% alcohol solution). The tissues then were cut into 100-µm thick sections, mounted on slides, Nissl stained with safranin, and coverslipped. Because the depth of electrode was related very closely to a Prussian blue point in the most superficial laminae of the dorsal horn (lamina I), it appeared reasonable to use only the depth measurement to locate the recordings when we used tungsten microelectrodes, given that all the experiments were performed with identical paradigms.

Both the recording and the stimulating sites were drawn individually through an optical microscope, using an attached camera lucida. The stimulation sites were plotted on a series of 12 camera lucida drawings of the PB area (150 µm apart). The spinal recording sites were plotted on camera lucida drawings of the L_4-5 levels because no recordings were made outside these lumbar segments. For illustration purposes, the stimulation sites in the PB area were collated on four representative sections (2 pontine, 2 mesencephalic), and the recording sites on a representative section passing through the L₄ segment.

Data analysis

The magnitude (or intensity) of the responses was defined as the mean spike frequency increase during the whole period of stimulation (20 s), i.e., mean spike frequency during the stimulation minus mean spike frequency during the 20 s preceding the stimulation. The responsiveness of a neuron was defined as the magnitude of the highest response observed for the considered modality.

In the case of heterotopic stimulation, the effect (inhibition) was defined as the percentage of the residual response as compared with the control test value. The results therefore were expressed as percentages of control responses.

Unless noted otherwise, the results were expressed as means ± SE. Statistical ANOVA were made using the Fisher's PLSD (protected least significant difference) tests. However, in some cases, we used the 10th, median, and 90th percentile because the distribution of the data were obviously not normal.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Fifty-three dorsal horn neurons were recorded extracellularly and unitarily at the lumbar segment level of 53 rats. These neurons were selected by the following criteria: antidromically backfired from the contralateral PB area and located in the most superficial laminae of the dorsal horn.

General findings

The spontaneous activity of the spino-PB neurons was extremely low (10th percentile < median <90th percentile): 0.001 < 0.064 < 1.88 Hz (n = 53), being often very close to silent. All of these neurons were clearly nociceptive, i.e., responses were markedly increased to stimuli in the noxious range, irrespective of whether or not they responded to innocuous stimuli. Most of them (75%, see following text) were nociceptive specific (NS), responding only to stimuli in the noxious range. Almost all of them (92%, 49/53) responded to two modalities of noxious stimuli (mechanical and thermal). Generally, they had small receptive fields including a restricted portion of the paw. All of the spino-PB neurons responded to transcutaneous electrical stimulation applied to the receptive field with, at least, two distinct early and late periods of activity.

Antidromic activation of the spino-PB neurons

All of the spino-PB neurons were driven antidromically from the contralateral PB area and satisfied the antidromic criteria (Fig. 1): the latency of the antidromic spike was stable (Fig. 1B), i.e., with a variation <100 µs, or <1% of the latency; it was always possible to obtain a collision between the antidromic spike and a spontaneous or electrically evoked orthodromic spike (Fig. 1, C and C'); the antidromic spike could follow a high-frequency stimulation (>500 Hz; Fig. 1, D and D'). The refractory period of the spino-PB neurons was short with a mean of 0.7 ± 0.2 ms (mean ± SD, n = 53).

View larger version (24K): [in this window] [in a new window]   Fig. 1. Criteria for antidromicity. A: single antidromic spike. B: superimposition of 15 sweeps showing the stability of the antidromic spike. C and C': collision test. C: antidromic spike (); C': orthodromic spikes (), one of which have collided with the antidromic spike that does not appear where it should occur (). D and D': high-frequency testing. D: antidromic spike was evoked 5 times by 5 electrical stimuli delivered with a frequency of 714 Hz; D': antidromic spike was evoked only 4 times by 5 electrical stimuli delivered with a frequency of 770 Hz. , electrical stimuli (delivered in the parabrachial area). E: drawing of a coronal section of the parabrachial (PB) area passing close to the junction between its pontine and mesencephalic divisions. This also shows an electrode tract and the different intensities required along the tract to elicit an antidromic spike in the lamina I neuron.

The mean latency of the antidromic spikes was 9 ± 6.1 ms (mean ± SD, n = 53). The mean of the calculated conduction velocities for the spino-PB neurons (~100 mm distance between the L₄ segment and the contralateral PB area divided by the spike latency) was 15.3 ± 7.3 m/s (mean ± SD, n = 53). The lowest conduction velocity was 2.8 m/s (Fig. 2), indicating that all of the spino-PB neurons recorded here had a thin myelinated axon.

View larger version (10K): [in this window] [in a new window]   Fig. 2. Histogram of conduction velocities of spino-parabrachial axons. Abcissa: conduction velocities range of each class in m/s, left and right brackets indicate inclusion and exclusion of the value, respectively. Ordinate: number of neurons in each class.

The stimulation site was always located in or in the close vicinity of the PB area (Fig. 3, A-E). The thresholds needed to evoke an antidromic spike were often low (150-300 µA) or very low (5-150 µA): 26% (14/53) and 57% (30/53), respectively (Fig. 3, A-E). A few thresholds (17%, 9/53) were higher (300-500 µA) when a systematic search for the lowest threshold was not achieved. In numerous experiments, a low threshold could be obtained from only one of the three stimulating electrodes, whereas from the other two, even a current >1 mA was not always sufficient.

View larger version (62K): [in this window] [in a new window]   Fig. 3. Parabrachial stimulation sites. A-D: caudal to rostral schematic drawing of the parabrachial are in coronal sections. Antidromic stimulation site with threshold in the range of 5-150 µA (), 150-300 µA () and >300 µA (). E: photomicrograph of a coronal section of the right pontine parabrachial area. Arrow indicates the Prussian blue point located between the pPBdl and pPBlcr subnuclei of the parabrachial area. Scale bars = 1 mm. bc, Brachium conjunctivum; Cnf, Cuneiformis nucleus; IC, inferior colliculus; KF, Kölliker-Fuse nucleus; LC, Locus cruleus; mPBcl, mPBel, mPBem, mPBil, mPBm, mPBsl, central lateral, external lateral, external medial, internal lateral, medial, and superior lateral subnuclei of the mesencephalic parabrachial area; pPBcl, pPBdl, pPBel, pPBem, pPBil, pPBm, pPBvl, central lateral, dorsal lateral, external lateral, external medial, internal lateral, medial and ventral lateral subnuclei of the pontine parabrachial area; sct, Spino-cerebellar tract.

Location of the spino-PB neurons

The depths of the recording sites of the 53 spino-PB neurons were in the 118- to 363-µm range with a mean of 240 ± 12 µm (n = 53), i.e., corresponding to the superficial dorsal horn. The relevance of this measurement was confirmed further in a subgroup of 28 spino-PB neurons for which the recording sites were labeled with a Prussian blue point (Fig. 4A) in the superficial laminae (Fig. 4B). The mean depth of the recordings corresponding to these labeled sites was 260 ± 23 µm (n = 28). The mean depth of the 25 neurons recorded with tungsten microelectrodes was similar 233 ± 14 µm (n = 25), indicating that sites without histological confirmation were similarly located close to the lamina I. 

View larger version (71K): [in this window] [in a new window]   Fig. 4. Spinal cord recording sites. A: photomicrograph of the left dorsal horn in coronal section at L4 lumbar enlargement. Arrow indicates the Prussian blue point. B: schematic representation of the L4 dorsal horn. : Prussian blue points corresponding to the location of recorded spino-PB neurons. Scale bars = 0.5 mm.

Receptive field

Overall the spino-PB neurons had small receptive fields. In most cases (72%, 38/53), the receptive field was extremely small (Fig. 5), including only one toe (n = 33) or one pad (n = 5). In some cases (28%, 15/53), the receptive field extended to two toes (n = 10), two pads (n = 2), or one toe and one pad (n = 3).

View larger version (11K): [in this window] [in a new window]   Fig. 5. A: receptive field and responses of a spino-parabrachial neuron. Toes in dark and light gray, parts of the receptive field in which noxious stimuli evoked a heavy and lower responses, respectively. Histogram in black: firing rate; the mechanical (40 N · cm2) and heat (50°C) noxious stimuli were applied, between the 2 arrows, in the corresponding toe. B and C: examples of the responsiveness reproducibility of 2 different spino-PB neurons. Two 48°C, as well as the 2 50°C stimuli were applied with a 5-min interval. Note the clear reproducibility of the responsiveness. Bin width of histograms = 2 s.

Characterization of spino-PB neurons by electrical stimuli

All the spino-PB neurons responded to suprathreshold transcutaneous electrical stimulation of the receptive field with at least two periods of activation (Fig. 6). The mean intensity threshold of these early and late periods of activation were 0.4 ± 0.1 and 3.1 ± 0.5 mA, respectively (n = 53). The mean latency (±SD) of the early and late periods were 7.7 ± 3 ms (range 3.9-20 ms, n = 53) and 162 ± 54 ms (range 51-290 ms, n = 53), respectively. These latencies correspond to calculated mean conduction velocities (±SD) of 26 ± 9 m/s (range 7.9-52 m/s, n = 53) and 1.1 ± 0.6 m/s (range 0.5-2.9 m/s, n = 53), respectively. These ranges of conduction velocities correspond with those of A and C fibers, respectively. Intense electrical stimulation (three times the C-fibers threshold) evoked 3.7 ± 0.4 and 14 ± 1.4 spikes (n = 53), in the early (A fibers) and late (C fibers) periods of activation, respectively.

View larger version (33K): [in this window] [in a new window]   Fig. 6. Response of a spino-parabrachial neuron to transcutaneous electrical stimuli. A: single sweep recording showing 3 (1 early and 2 late) peaks of firing evoked by the stimulation of 1 toe. B: poststimuli time histogram (PSTH) made from responses to repetitive electrical stimuli (16 trials, 0.5 Hz). Note 1 early and 2 late components of response well individualized. Bin size is 0.5 ms.

The distribution of the A-fiber conduction velocities, calculated from the latency of the earliest spike in the early peak of activation for each single spino-PB neuron appeared normal (Fig. 7A). The majority of the conduction velocities (58%, 31/53) were in the 8- to 26-m/s range, i.e., clearly in the range of conduction velocities of A fibers. The remaining 42% (22/53) corresponded to conduction velocities between either 26-38 or 38-60 m/s (Fig. 7A), i.e., to conduction velocities of peripheral fibers that were either at the border between A and A fibers (n = 18) or within A-fiber (n = 4) ranges, respectively. Even in neurons with a clear A-fiber activation, the duration of the first period of activation was such that most of the evoked spikes, and more specifically the relatively late ones, were likely to be triggered by A fibers. Furthermore in 12 cases, two peaks of activation, with mean conduction velocities of 28 ± 6 and 13.6 ± 2 m/s, respectively (mean ± SD, n = 12), were identified. These clearly correspond to A/A-fiber and A-fiber conduction velocity ranges, respectively.

View larger version (20K): [in this window] [in a new window]   Fig. 7. Histogram of the peripheral conduction velocities. A: conduction velocities of peripheral fibers triggering the early period of activation. B: conduction velocities of the peripheral fibers triggering the late period of activation. Abcissa: conduction velocities range of each class in m/s; left and right brackets indicate inclusion and exclusion of the value, respectively. Ordinate: number of neurons in each class.

The distribution of the C-fiber conduction velocities, calculated from the latency of the earliest spike in the late peak of activation for each single spino-PB neuron (Fig. 7B) was also normal and clearly separated from the A-fiber group (Fig. 7A). The great majority of the C fibers (87%, 47/53) activating the spino-PB neurons had very slow conduction velocities, between 0.3 and 1.4 m/s, the few remaining cases corresponded to intermediate weaker peaks (Fig. 7A). Furthermore, as clearly illustrated in the single sweep of Fig. 6, numerous spino-PB neurons responded to suprathreshold electrical stimuli with two peaks in the late period of activation. In 26 cases where the two late peaks were clearly separated, their mean conduction velocities were 1.2 ± 0.6 and 0.6 ± 1 m/s, respectively (mean ± SD, n = 26).

Figure 8 shows that in the late period corresponding to C-fiber activation, for 36 representative spino-PB neurons stimulated over 16 trials at 0.5 Hz with the same intensity (3 times the threshold), the number of spikes evoked by the first stimulus increased slightly after a few stimulations and remained stable thereafter. Thus the wind-up was relatively weak in the present population of superficial spino-PB neurons.

View larger version (14K): [in this window] [in a new window]   Fig. 8. Number of spikes in the late component (C fibers) during repetitive transcutaneous electrical stimuli. Mean C-fiber component increased moderately during the 1st 8 stimuli (0.5 Hz) before plateauing.  and error bars: mean ± SE (n = 36).

Characterization of the spino-PB neurons by natural stimuli

RESPONSIVENESS. All the spino-PB neurons responded with a strong and sustained discharge to thermal and/or mechanical noxious stimuli. From a very low or absent spontaneous activity (see preceding text), the responses of the spino-PB neurons to noxious stimuli had a rapid onset that could reach values ~100 Hz (Figs. 5, 10, and 13), giving to the initial part of the response a phasic aspect. The response then decreased but remained strong and sustained throughout the period of stimulation, giving to the second part a tonic pattern. The majority of neurons had an afterdischarge, especially for the highest noxious intensities of stimulation (Figs. 5, 10, and 13). Overall, the spino-PB neurons had a high responsiveness, the mean of the maximum response (whatever the modality) was 47 ± 5 Hz (n = 53). Furthermore the response to noxious stimuli was clearly reproducible, the variation of responsiveness to two noxious heat stimulation of the skin (5-20 min apart) being in the 0.5-25% range (n = 29). Figure 5, B and C, shows two individual examples of the reproducibility of responsiveness to heat stimuli.

2

View larger version (12K): [in this window] [in a new window]   Fig. 9. Stimulus-response curves of some spino-parabrachial neurons to cold stimuli. Decreasing temperatures (30 to 20°C in 5°C step) were applied on the receptive field, each of them during 20 s with delays without stimuli >3 min between each temperature. Abcissa: stimulus temperature. Ordinate: mean frequency of response (see METHODS).

ENCODING PROPERTIES TO MECHANICAL STIMULI. The spino-PB neurons increased their firing rate when the pressure stimulus increased in the noxious range as shown in Fig. 10, A-C, for three neurons having different thresholds (5, 10, and 20 N · cm²) close to and in the noxious range. The progressive increase in neuronal firing paralleled the increase in intensity of the noxious pressure, this was true for the whole tonic excitation throughout the stimulation period and the afterdischarge where present (Fig. 10, B and C).

View larger version (13K): [in this window] [in a new window]   Fig. 10. Mechanical encoding properties of spino-parabrachial neurons. A-C: examples of encoding properties of 3 spino-parabrachial neurons having different thresholds (5, 10, and 20 N · cm2, respectively). Graded mechanical stimuli were applied on the receptive field during the 20 s between the arrows; binwidth of histograms = 2 s. D: histogram of mechanical thresholds. Mean threshold was 11.5 ± 1.5 N · cm2 (n = 39).

2

2

2

2

2

2

2

2

2

View larger version (39K): [in this window] [in a new window]   Fig. 11. Stimulus response curves to mechanical stimuli gouped in accordance to the thresholds. A-D: stimulus response curves of spino-PB neurons with a low (5 N · cm2), medium (10 N · cm2), high (20 N · cm2), and very high (40 N · cm2) mechanical threshold, respectively. Insets: histogram of the thresholds, the class of threshold the curves presented being in black for each group. Thin lines: curves of individual neurons; thick line mean curve of the corresponding group.

2

2

2

2

2

2

2

View larger version (9K): [in this window] [in a new window]   Fig. 12. Mean stimulus-response curve to mechanical stimuli. Increasing mechanical stimuli were applied on the receptive field. : mean response, error bars: SE. Abcissa: pressure in N · cm2. Ordinate: mean frequency of response (see METHODS).

ENCODING PROPERTIES TO HEAT STIMULI. The firing rate of spino-PB neurons increased with increasing temperature of stimulation in the noxious range as shown for three individual examples (Fig. 13, A-C) of neurons with different noxious thresholds (44, 46, and 48°C, respectively). The response increase, primarily in the 46-50°C range, was maintained over the whole period of stimulation in addition to the afterdischarge in some cases (Fig. 13B).

View larger version (17K): [in this window] [in a new window]   Fig. 13. Thermal encoding properties of spino-parabrachial neurons. A-C: examples of encoding properties of 3 spino-PB neurons having different thresholds (44, 46, and 48°C, respectively). Graded temperatures were applied on the receptive field during 20 s between the arrows; binwidth of histograms = 2 s. D: histogram of heat thresholds. Mean threshold was 43.6 ± 0.5°C (n = 38).

View larger version (25K): [in this window] [in a new window]   Fig. 14. Stimulus response-curves to heat stimuli gouped in accordance to the heat thresholds. A-C: stimulus response curves of spino-PB neurons with low (38-40°C), medium (42-44°C), and high (46-48°C) thermal thresholds, respectively. Responses of individual neurons are presented with thin lines, the average stimulus-response curve is the thick line. In each case, the insert represents the histogram of the thresholds distribution, with in black the thresholds concerned by the curves.

View larger version (11K): [in this window] [in a new window]   Fig. 15. Mean stimulus-response curve to heat stimuli. Increasing temperatures were applied on the receptive field. : mean response, error bars: SE. Abcissa: temperature in °C. Ordinate: mean frequency of response (see METHODS).

RELATIONSHIPS BETWEEN ENCODING PROPERTIES OF NEURONS TO NOXIOUS HEAT AND MECHANICAL STIMULI. Neurons were divided into two groups with respect to mechanical thresholds, low (5 N · cm², n = 8) and high (20 N · cm², n = 9), thus corresponding to the commonly adopted classification of WDR and NS cells. The encoding properties of these two neuronal groups to heat stimuli were compared. Low mechanical threshold neurons had a tendency to lower heat thresholds (42 ± 1.3°C; n = 8), whereas high mechanical heat threshold neurons had higher heat thresholds (45 ± 0.8°C; n = 9). In addition, the mean stimulus-response curves of the two groups to heat stimuli were significantly different (Fig. 16; ANOVA, F_1,151 = 5, P = 0.027). Thus the low mechanical threshold neurons (WDR) had a higher responsiveness than the high mechanical threshold neurons (NS) for all the graded temperatures in the noxious range (44°C).

View larger version (13K): [in this window] [in a new window]   Fig. 16. Mean stimulus-response curve to heat stimuli of subgroups of wide dynamic range (WDR) and nociceptive specific (NS) neurons.  and  correspond to WDR (low mechanical threshold 5 N · cm2, n = 8) and to NS (mechanical threshold 20 N · cm2, n = 9) neurons, respectively. Error bars: SE. Abcissa: temperature in °C. Ordinate: mean frequency of response (see METHODS).

Inhibition of spino-PB neurons by concomitant heterotopic noxious stimuli

Noxious heat stimulation (50°C) applied to the receptive field of the recorded lumbar neuron could be considerably inhibited by a noxious mechanical stimuli (60 N · cm²) of the forepaw (Fig. 17A, 1 and 2). Heterotopic noxious stimuli (stimulation applied outside the receptive field of the recorded lumbar neuron) clearly inhibited the response to the control noxious stimulation of the great majority of the spino-PB neurons (13/16) studied. The percentage of inhibition was clearly different between neurons (Fig. 17B), but in most cases (9/13) the neuronal response was reduced to <50% of the control value. The mean intensity of the responses during heterotopic noxious stimulation was 31.7 ± 6.1% of control responses (n = 13).

View larger version (12K): [in this window] [in a new window]   Fig. 17. Effects of heterotopic noxious stimuli on nociceptive responses of spino-PB neurons. A1: control response of a spino-PB neuron to noxious heat (50°C) applied during 20 s (between arrows) on the receptive field (1 toe). A2: response of the same neuron to the same stimuli when a 2nd mechanical noxious stimuli (60 N · cm2) was applied simultaneously in another part of the body (the forepaw). Mechanical stimuli, started 10 s before, and finished 10 s after the heat stimuli. In this case (A2) we observed a marked diminishing of the response to heat stimuli; binwidth of histograms = 2 s. B: histogram of the effect of heterotopic noxious stimuli. Abcissa: remaining response in percent of the control. Ordinate: number of neurons in each class.

This phenomenon of heterotopic inhibition was maximum during concomitant stimulation, but decreased rapidly thereafter, rarely lasting more than a couple of minutes after cessation of the heterotopic stimulation. The level of inhibition did not seem to be dependent on the intensity of the control response, which indicates that whatever the amplitude of the initial responsiveness, the neurons seemed to have the same susceptibility to inhibition by heterotopic noxious stimuli.

We did not find any evidence for a difference between the different categories of spino-PB neurons (low or high mechanical threshold) when considering their susceptibility to inhibition by heterotopic stimuli.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Results were obtained from 53 neurons, all recorded in or close to the lamina I at the lumbar level of the spinal cord, and backfired antidromically from the contralateral PB area. All spino-PB neurons were excited by and encoded the intensity of noxious mechanical and/or thermal stimuli applied to their small receptive field. None of the spino-PB neurons were activated by innocuous stimuli only. However, some neurons were lightly activated by innocuous mechanical stimuli, as well as responding to noxious stimuli. A distinction between NS (75%) and WDR (25%) spino-PB neurons was made. All neurons had a very low spontaneous activity, close to silent, and exhibited two distinct periods of activation, corresponding to A- and C-fiber-evoked activity, respectively, in response to suprathreshold transcutaneous electrical stimuli.

Methodological considerations

ANESTHETIC REGIMEN AND IMMOBILIZING DRUGS. The anesthetic conditions used in the present study were the same as those used in all of the previous electrophysiological studies of the PB area in our laboratory (e.g., Bernard and Besson 1990; Bernard et al. 1994; Bester et al. 1995b). This point is of critical importance to ensure accurate comparison between responses of spino-PB and PB neurons. It must be emphasized that the precise measurement of the halothane concentration together with the endtidal PCO₂ allowed good control of the stability and reproducibility of the anesthesia depth. Furthermore in our hands, this anesthetic regimen did not seem too depressive in terms of neuronal responses. This indeed allowed us to observe clear, and often strong, neuronal responses to light and/or noxious stimuli in different spinal laminae and several supraspinal nuclei (present results; see refs in Bernard et al. 1996; Besson and Chaouch 1987). Despite numerous advantages, our anesthesia regimen, like any other, inevitably induces some changes in comparison to the awake state. This issue is probably even more critical with other types of anesthesia regimen where the physiological assessments of the anesthesia quality and levels of drug delivery cannot be constantly monitored and accurately adapted.

ANTIDROMIC ACTIVATION AND LOCATION OF SPINO-PB NEURONS. The PB area has been shown to be the densest field of projection of contralateral spinal lamina I neurons in rats, as demonstrated recently with the Phaseolus vulgaris leucoagglutinin (Bernard et al. 1995; Feil and Herbert 1995; Slugg and Light 1994). The selectivity of the projection origin from lamina I is supported further by studies using retrograde tracers (Bernard et al. 1989; Cechetto et al. 1985; Hylden et al. 1989; Kitamura et al. 1993; Lima and Coimbra 1989; Menétrey and De Pommery 1991; Menétrey et al. 1982, 1992; Swett et al. 1985; Yamada and Kitamura 1992). In the present study, most of the thresholds (83%) needed to evoke an antidromic spike were low (300 µA) or very low (<150 µA), and all the stimulating sites were clearly located in the PB area, it is therefore highly probable that spino-PB neurons terminals were primarily stimulated. This confidence in the specificity of the antidromic stimulation for spino-PB neurons is reinforced by the personal observation that in some cases a current 1,000 µA did not spread up to 600 µm (the distance between 2 monopolar electrodes). This short distance of current spread is a little lower than indicated by Ranck (1975). Furthermore identifying well-isolated spikes that lasted long enough (1 ms) and had a high enough amplitude (2 mV) supports the somatic origin of the recordings (Lipski 1981). Consequently, as validated by the location of the Prussian blue points in the lamina I and sometimes in the Lissauer's tract, we can state that our recordings of neurons driven antidromically from the contralateral PB area were those of lamina I spino-PB neurons. However, because the axonal trajectory of spinothalamic neurons could be in the vicinity of the PB area (unpublished anatomic data from PHA-L injection centered in lamina I) (see also Mehler et al. 1960), one cannot exclude the possibility that the higher currents have stimulated spinothalamic neurons.

SELECTION CRITERIA. Neurons were selected primarily on the basis of the fulfillment of antidromic activation from the PB area. Consequently we have investigated the physiological characteristics of all of the recorded spino-PB neurons, and we were able to determine, with minimal bias, that all lamina I spino-PB neurons studied were excited by noxious stimuli. With the aim of avoiding central and peripheral sensitization, we have described only one neuron per animal, thus allowing us to report a homogenous neuronal population with homogenous properties.

Physiological properties of spino-PB neurons

CENTRAL CONDUCTION VELOCITIES. A wide range of central conduction velocities was observed. Because the slowest fiber we recorded had a conduction velocity of 2.8 m/s, we can assume that all the axons of the spino-PB neuronal population were myelinated. Previous studies have also described a wide range of central conduction velocities of spino-PB axons; however, a higher representation of low to very low conduction velocities were then reported (Hayashi and Tabata 1989; Hylden et al. 1985, 1986a,b; Light et al. 1993; McMahon and Wall 1985). Furthermore one study (Light et al. 1993), suggested that a substantial proportion (~40%) of axons of NS spino-PB neurons had very low conduction velocities in the unmyelinated range. The basis for these discrepancies may be due to species differences (Hayashi and Tabata 1989; Hylden et al. 1985, 1986a,b; Light et al. 1993). Furthermore in our experimental condition the few spikes evoked with very