responses of c nociceptors in human skin to remote intradermal ...

18 downloads 4454 Views 249KB Size Report
Feb 4, 2004 - Digitized signals were stored on the hard drive .... and 1B units was the time to 50% recovery of latency after the end of the 2 Hz train. .... enhanced release of such substances from damaged skin, rather than to acquired.
Articles in PresS. J Neurophysiol (February 4, 2004). 10.1152/jn.00565.2003

1

TWO TYPES OF C NOCICEPTOR IN HUMAN SKIN AND THEIR BEHAVIOR IN AREAS OF CAPSAICIN-INDUCED SECONDARY HYPERALGESIA

Jordi Serra1, Mario Campero2, Hugh Bostock3 and José Ochoa. Oregon Nerve Center, Good Samaritan Hospital and Medical Center, Oregon Health Sciences University, Portland, Oregon, USA; 1Neuropathic Pain Unit, Hospital General de Catalunya, Barcelona, Spain; 2 Departamento de Ciencias Neurológicas, Universidad de Chile, Santiago, Chile; 3Sobell Department, Institute of Neurology, University College London, Queen Square, London, UK.

Running title: Silent C nociceptors in secondary hyperalgesia

Correspondence: Dr. Jordi Serra Unitat Dolor Neuropàtic Hospital General de Catalunya c. Josep Trueta s/n 08190 Sant Cugat del Vallès (Barcelona) Spain Email: [email protected]

Copyright (c) 2004 by the American Physiological Society.

2

Abstract Peripheral nociceptor sensitization is accepted as an important mechanism of cutaneous primary hyperalgesia, but secondary hyperalgesia has been attributed to central mechanisms, since evidence for sensitization of primary afferents has been lacking. In this study, microneurography was used to test for changes in sensitivity of C nociceptors in the area of secondary hyperalgesia caused by intradermal injection of capsaicin in humans. Multiple C units were recruited by electrical stimulation of the skin at 0.25 Hz, and were identified as discrete series of dots in raster plots of spike latencies. Nociceptors slowed progressively during repetitive stimulation at 2 Hz for 3 minutes. According to their response to mechanical stimulation, nociceptors could be classified as either mechano-sensitive (CM) or mechano-insensitive (CMi). These two nociceptor subtypes had different axonal properties: CMi units slowed by 2% or more when stimulated at 0.25 Hz after a 3-minute pause, whereas CM units slowed by less than 1%. This stimulation protocol was used before capsaicin injection to identify nociceptor subtype without repeated probing, thus avoiding possible mechanical sensitization. Capsaicin, injected 10 to 50 mm away from the site of electrical stimulation, had no effect on any of 29 CM units, but induced bursts of activity in 11 out of 15 CMi units, after delays ranging from 0.5 to 18 minutes. The capsaicin injections also sensitized a majority of the CMi units, so that 11 out of 17 developed immediate or delayed responsiveness to mechanical stimuli. This sensitization may contribute a peripheral C fiber component to secondary hyperalgesia.

Introduction Capsaicin application to human skin has been extensively used as a method to induce experimental pain and the development of a characteristic flare and mechanical and heat

3 hyperalgesia (Culp et al., 1989; Simone et al., 1989; LaMotte et al., 1991; Kilo et al., 1994). However, the spatial extent and underlying mechanisms of the flare, the psychophysical profile of the hyperalgesias and the neural mechanisms responsible for them remain disputed (Ali et al., 1996; Serra et al., 1998).

The primary hyperalgesia at the site of capsaicin application features increased sensory responses to both mechanical and heat stimuli, attributed to sensitization of cutaneous nociceptor terminals, both of the common polymodal type (LaMotte et al., 1982; Campbell and Meyer, 1983; LaMotte et al., 1983; Torebjörk et al., 1984; LaMotte et al., 1992), and of the more recently described mechano-insensitive type of C nociceptor (Meyer et al., 1991; Davis et al., 1993; Schmidt et al., 1995; Schmelz et al., 2000b). The neural mechanisms underlying the secondary hyperalgesia in surrounding skin remain unclear. Common polymodal C nociceptors have been reported not to become sensitized in the area of experimental secondary hyperalgesia, in both animals and humans (Campbell and Meyer, 1983; Baumann et al., 1991; LaMotte et al., 1992; Torebjörk et al., 1992). Schmelz et al. (2000b) found that application of capsaicin to one branch of a mechano-insensitive C nociceptor did not sensitize other branches not exposed to capsaicin. These findings were consistent with previous reports on the absence of spread of sensitization to neighboring branches of a stimulated C-nociceptor (Thalhammer et al., 1982; Baumann et al., 1991; LaMotte et al., 1992; Schmelz et al., 1996). However, possible sensitization of remote mechano-insensitive C nociceptors by capsaicin was not specifically tested in those studies. Partly because of the lack of evidence of receptor sensitization, it has been proposed that capsaicin-induced secondary hyperalgesia is mediated by dorsal horn neurons, sensitized by a primary C-nociceptor barrage (LaMotte et al., 1991; Simone et al., 1991; Torebjörk et al., 1992, Koltzenburg et al., 1994, Sang et al., 1996; Woolf and Mannion, 1999). More specifically,

4 Magerl et al. (2001), using combinations of cutaneous desensitization and pressure nerve blocks, found evidence that secondary hyperalgesia to pinprick is mediated by capsaicin-insensitive Aδ afferents, centrally facilitated by capsaicin-sensitive C fiber input. However, evidence has also been reported of a likely contribution of mechano-insensitive or ‘silent’ C nociceptors to the pathophysiology of secondary hyperalgesia (Serra et al., 1993; Serra et al., 1994; Serra et al., 1995; Serra et al., 1998). With the aid of improved methods to track the behavior of multiple identified C fibers, we have now confirmed that these units are activated and can become sensitized in the area of secondary hyperalgesia.

Methods

Subjects Eleven healthy adult volunteers participated in a total of 25 microneurographic recordings. There were 7 males and 4 females, with ages ranging from 18 to 40 years (mean 35). The study had the approval of the local ethics committee and conformed to the Declaration of Helsinki. All subjects gave their informed, written consent.

Microneurographic recordings Microneurography was used to record action potentials of human C fibers from cutaneous nerve fascicles of the superficial peroneal nerve at the ankle (10 subjects), or the superficial radial nerve at the wrist (1 subject). The subjects sat relaxed with the leg or arm firmly supported in a padded platform. The general technique of microneurography has been described in detail by Vallbo and Hagbarth (1968). Intraneural recordings were obtained using a 0.2 mm diameter lacquerinsulated tungsten microelectrode (MNG active/1MΩ FHC Inc., Bowdoinham, ME, USA), which

5 was inserted percutaneously into a sensory nerve. A subcutaneous reference electrode was inserted 1-2 cm outside the nerve trunk. The neural signals were amplified by a commercial differential amplifier (FHC Inc., 3+ Cell Isolated Microamplifier) and filtered with an adjustable analogue filter (band-pass 100-2000 Hz). Line interference was removed with an on-line noise eliminator (Hum Bug, Quest Scientific, North Vancouver, Canada). Signals were displayed on a Tektronix 5113a oscilloscope and digitized by a personal computer with a Data Translation DT2812 A/D board at a sampling rate of 10 kHz. Digitized signals were stored on the hard drive of the personal computer as raw data for off-line analysis. Digital filtering (band pass 300-2000 Hz) and clamping of the baseline were performed both on-line and during off-line analysis for a better visualization of the action potentials. Temperature of the skin was measured with a thermocouple placed on the skin adjacent to the receptive fields of the units under study. Skin temperature was maintained above 30°C with an infrared lamp.

Protocol of electrical stimulation of the cutaneous receptive fields

Search for the electrical receptive fields of C fibers was conducted by stimulating electrically with a pair of needle electrodes in areas of skin to which painful sensations were referred during intraneural electrical microstimulation at near threshold levels (Torebjörk and Ochoa, 1990). Stimulation was performed using rectangular pulses of 0.25-0.3 ms duration (Grass S48, stimulus isolation unit SIU 5) at a rate of 0.25 Hz. Only fibers with latencies compatible with conduction velocities in the C fiber range (< 2 ms-1) were studied. When time locked responses with such latencies were recorded at 0.25 Hz baseline stimulation, a sequence of 3-minute pause followed by 6-minute baseline and 3-minute 2 Hz train was given (see Figure 1). Stimulation at low rates

6 following a pause has been reported to differentiate mechano-sensitive from mechano-insensitive C nociceptors (i.e. CM from CMi units) in the human skin (Weidner et al., 1999), while stimulation at 2 Hz for 3 minutes differentiates patterns of slowing among functional types of C fibers (Serra et al., 1999). For convenience, and because no axonal property has yet been shown to correlate with heat sensitivity, we will here use the abbreviations CM and CMi for mechanosensitive and mechano-insensitive nociceptors respectively, without regard for heat sensitivity. CM therefore includes common polymodal CMH units, and CMi embraces both CH and CMiHi units.

Analysis of responses Responses were recorded and analyzed with QTRAC software ( Institute of Neurology, London), specially modified to track peak latencies and display them as a latency “profile” or as a raster plot. In the latency raster plots each peak in the filtered voltage signal that exceeded a specified level is represented by a dot on a plot with latency as the ordinate and elapsed time as the abscissa (see Serra et al., 1999). Depending on the level chosen the dots could represent action potentials or noise (Figure 1 A). For the raster figures in this paper, latencies of selected units with adequate signal-to-noise were remeasured from the raw data, so that each dot represents an identified single unit. These remeasured figures are referred to as modified raster plots (Figure 1 B). A full description of the method is provided in Serra et al., 1999.

FIGURE 1 NEAR HERE

7

Measurements of conduction velocity changes and assignment of axonal type Conduction velocity was estimated by dividing the conduction distance by the baseline latency at the stimulation rate of 0.25 Hz. Slowing after the 3-minute pause was measured as the percentage increase in latency from the first spike after the pause to the baseline latency. Percentages of conduction slowing relative to the baseline latency were also determined at 1 and 3 minutes after the onset of the 2 Hz stimulus train. As shown in a previous report (Serra et al., 1999), when stimulation frequency is increased from 0.25 Hz to 2 Hz, different patterns of slowing can be observed among human C fibers: “Type 1” fibers slowed progressively (average latency increase at 3 minutes = 28.3%), whereas “Type 2” fibers slowed to reach a plateau (average latency increase 5.2%) within a minute. The former were identified as nociceptors, whereas the latter have been recently identified as specific cold receptors (Campero et al., 2001). We quickly found that the 3-minute pause enabled the Type 1 fibers to be divided into 2 groups, those which were essentially unaffected by the pause, and those which showed an appreciable reduction in latency at the end of the pause. We designated the former Type 1A, and the latter Type 1B. According to this classification, the 5 units in Figure 1B can be assigned (in order of increasing latencies) to Type 2, 1A, 1A, 2, and 1B. Only Type 1A and 1B units, putative nociceptors, are considered further in this study.

Measurement of recovery The time course of recovery of conduction velocity after cessation of stimulation at 2 Hz was also measured. Following Thalhammer et al. (1994), we used two indices of the recovery rate: (1) the time necessary to reverse 50% of the activity-induced latency change, and (2) the percentage of recovery at 30 seconds after the end of the stimulus train.

8

Functional characterization of the units Identification of the receptor properties of the units under study was determined from raster plots while the unit was being electrically stimulated regularly at 0.25 Hz, using the “marking technique” as described previously (Hallin and Torebjörk, 1974; Schmelz et al., 1995). In short, natural stimuli that activate a unit induce an abrupt increase in conduction latency, due to activity-dependent slowing (Figure 2). This has been shown to be a reliable method to follow units, both at rest and during natural activation, for long periods of time. This method is able to discriminate, among several units, which ones have been activated with natural stimulation.

FIGURE 2 NEAR HERE

To check for mechanical sensitivity of the units, von Frey hairs (Semmes-Weinstein monofilaments, Stoelting Co., Chicago) exerting different bending forces ranging from 52 mN (diameter 0.36 mm, pressure 5.1 bars) to 722 mN (diameter 0.70 mm, pressure 18.6 bars) were applied to the previously identified electrical receptive field and surrounding areas. Mechanical stimulation was kept to a minimum to avoid sensitization of the unit with repeated probing. Where possible, heat responsiveness was tested by delivering a heat pulse from a resting temperature of 35ºC to 45ºC (10ºC/s), using a Peltier thermode with a contact area of 1 cm2. On most occasions, however, the location of the receptive field of the unit made it difficult to place the Peltier on top of it. The receptive field was then probed with a short contact with a heated metal rod (~45oC). To exclude that the units under study might have been sympathetic efferents, maneuvers known to activate sympathetic fibers, i.e. the Valsalva maneuver, startle by an unexpected shout, and stress caused by mental arithmetic (Delius et al., 1972, Hallin &

9 Torebjörk, 1974) were performed and sudden latency shifts, indicative of activation of that unit, were tracked. After some experiments had been performed, it became evident that the protocol combining the pause and 2 Hz stimulation allowed unambiguous identification of the unit under study as either a CM or CMi unit (see Figure 3, later). Subsequently, some of the units were never tested for mechanical or heat sensitivity before being challenged with capsaicin. Therefore, possible sensitization of the units with repeated testing before capsaicin injection was completely avoided.

Capsaicin injection Following the 2 Hz, 3-minute period, an ensuing 6-minute (or longer if required) stimulation period at 0.25 Hz was allowed until the unit had recovered to its previous stable latency. Then, flare and hyperalgesia were induced by injecting an aqueous solution containing 1% capsaicin dissolved in 7.5% Tween 80. A volume of 10 µl, containing 100 µg of capsaicin, was injected intradermally in the skin of the dorsum of the foot or the hand at a distance of at least 1 cm from the stimulating electrodes, always within the innervation territory of the implicated nerve, using a 28-gauge hypodermic needle. In order to test for effects of capsaicin injection lasting longer than the duration of a microneurographic recording, on two occasions capsaicin was injected at least 30 minutes before the start of the microneurography. At the end of the experiment, short trains of electrical stimuli (0.25 ms pulses at 20 Hz) at the maximum intensity tolerated by the subject (~50 mA) were applied with the aid of a rounded tip stimulator (as described by Meyer et al. 1991) to search for possible branches of the recorded units in the area of capsaicin injection.

10 Monitoring of responses following remote capsaicin injection Stimulation at the usual baseline 0.25 Hz frequency was continued during and after the capsaicin injection. This enabled any spontaneous activity of the units to be revealed by sudden increases in their otherwise stable latencies on the raster plot. Subsequently, the units under study were again tested to check for possible development of mechanical or heat sensitivity.

Determination of the area of visual flare and mechanical hyperalgesia The extent of the early visual flare that appeared immediately after capsaicin injection was outlined on the skin with a soft pen, and a picture taken. As previously reported (Serra et al., 1998), the extent of this early visual flare is much larger and shorter lived than the flare that remains afterwards for several hours surrounding the injection site. Also, the area of mechanical hyperalgesia was determined using a von Frey hair exerting a force of 2.02 N (diameter 1.02 mm, pressure 24.75 bars). In normal skin this stimulus is clearly suprathreshold for mechanical pain in all subjects. The filament was applied at right angles for 5 seconds to each of a series of points 0.5 centimeters apart along radial paths converging at the injection site. Care was taken to avoid stretching the skin laterally, which might activate CMi units (Schmelz et al., 2000b). Stimulation started at points well beyond the area where hyperalgesia was typically detected. The area of hyperalgesia was outlined with a soft pen on the skin, and a picture taken.

Results

General characteristics of the population of C units A total of 29 units classified as axonal Type 1B were recorded with adequate signal-to-noise ratio to be included in the study. An additional 29 units classified as axonal Type 1A with good signal-

11 to-noise ratio were also extracted from the same raster plots for comparison. The results, like the experiments, fall into 2 parts. In the first part of an experiment, as illustrated in Figure 2, the units were tested with the pause and 2 Hz train to define the axonal type and in most cases the units were also tested with mechanical and heat stimuli to characterize their responsiveness to natural stimuli. These results confirmed that axonal behavior could be used to determine receptor type without ambiguity, so that mechanical testing (which carries a slight risk of sensitizing CMi units) could be avoided. In the second part of an experiment (as illustrated in Figure 7, which is a continuation of Figure 2), capsaicin was injected at a distance and any spontaneous activity or changes in sensitivity noted.

Part 1: Axonal and receptor properties of Type 1 units

Latency changes induced by the pause and receptor responsiveness Percentages of slowing of conduction velocity following the 3minute pause are shown in Figure 3, together with the classification of the units into Type 1A and 1B. Seventeen of the 29 Type 1B units were tested for mechanical responsiveness using a von Frey hair exerting a bending force of 722 mN. None of the units tested in such a way responded to mechanical stimulation and they are indicated by I (insensitive) in Figure 3. All 29 Type 1A units were tested mechanically and all of them responded vigorously to von Frey hairs ranging from 52 mN to 722 mN (mean ± S.D. 129.3 ± 150.8). Therefore, Type 1A units were identified as CM, while Type 1B were identified as CMi, in confirmation of the report by Weidner et al. (1999). Heat responses were not tested systematically because of the difficulty of applying the thermode without disturbing the stimulating electrodes FIGURE 3 NEAR HERE

12

Further differences in conduction properties between Type 1A and 1B units Weidner et al. (1999) found that all CMi units had lower velocities than CM units. We also found that the Type 1B (CMi) units conducted significantly more slowly than the Type 1A (CM) units, but there was considerable overlap (see Figure 4A). Although both velocity and slowing after the pause were related to axonal type, there was no correlation between these variables within each type of fiber.

FIGURE 4 NEAR HERE

We also found that the extent of conduction velocity slowing during the 2 Hz train was greater on average for Type 1B than for Type 1A fibers (Figure 4B), and that the recovery of latency after the end of the train was slower for Type 1B fibers, whether measured as time to 50% recovery to baseline latency or as percentage recovery after 30 seconds (Figure 4C).

Apart from the effect of the pause, the parameter that most clearly differentiated the Type 1A and 1B units was the time to 50% recovery of latency after the end of the 2 Hz train. These two parameters are compared in Figure 4D, which shows that although these parameters were correlated amongst the Type 1 population as a whole, there was no correlation within either the Type 1A or Type 1B sub-group. Figures 4A and 4D indicate that the Type 1A and Type 1B units form two distinct classes, which differ in multiple axonal properties as well as in their receptor properties.

13 Part 2: Effects of capsaicin injection

Spontaneous activity induced by capsaicin injection Capsaicin was injected in 10 experiments while recording from 29 Type 1A and 22 Type 1B units. Distances from injection site to electrical receptive fields of the units ranged from 10 to 50 mm (median 30 mm). In agreement with previous reports (Baumann et al., 1991; LaMotte et al., 1992), none of the 29 Type 1A (CM) fibers were affected by the remote injection of capsaicin. Examples are shown in Figures 5 (unit 1) and 7 (units 1-3). In contrast, capsaicin induced spontaneous bursts of activity in 11 out of the 15 Type 1B (CMi) units for which responses were recorded, after a delay that ranged from 0.5 to 18 minutes (median 2.7 minutes). There was no correlation between time to onset of the burst and distance from injection site. Some units bursted for long periods of time, up to almost 30 minutes in some cases (Figure 6, units 1 and 2). These bursts of activity could be readily identified in the raster plots as sudden shifts in the stable latencies to the continuous stimulation at 0.25 Hz. Examples of this are shown in Figures 5 (unit 2 only), Figure 6 (units 1-3) and Figure 7 (unit 4). On some occasions a single burst of activity appeared (Figure 7), while on other occasions there was a succession of bursts, sometimes one on top of the other (Figures 5 and 6).

TABLE 1 NEAR HERE FIGURES 5, 6, 7 NEAR HERE

Altered responses following capsaicin injection After any spontaneous activity had subsided, the areas within and surrounding the site of electrical stimulation were tested again for possible development of responsiveness of Type 1B

14 units to mechanical stimuli. Mechanical sensitization occurred in 11 out of the 17 Type 1B units tested (Table 1). The 11 units that acquired mechanical sensitivity were not the same 11 units as were activated: 2 units were activated but not sensitized and 2 units were sensitized but not activated. The newly acquired responses to mechanical stimulation usually appeared several seconds after the initial application of the stimulus, and they built up with a temporal profile different from the responses of Type 1A units (e.g. unit 4 in Figures 7 and unit 3 in Figure 8). The responses could also persist for several minutes after the stimulation had ceased. However, immediate responses were also observed (e.g. both units in Figure 9). Like secondary hyperalgesia, the altered responsiveness following remote capsaicin injection was long lasting. The responses in Figure 9 occurred more than an hour after capsaicin injection. In one of the two subjects who received the capsaicin before the start of microneurography, responses to a 68 mN von Frey hair were obtained from a Type 1B unit 4 hours after the injection 50 mm away. Sensitization of Type 1A units to remote capsaicin injection in the form of lowering of thresholds or enhanced suprathreshold responses was not observed (e.g. Figure 7, units 1-3), but was not assessed systematically. Application of short trains of electrical stimuli in the area of capsaicin injection to search for branches (as described in Methods) did not activate any of the recorded Type 1B units.

FIGURES 8, 9 NEAR HERE

Areas of flare and mechanical hyperalgesia The maximal extent of the area of flare following capsaicin injection could be clearly traced in 5 of the cases. This early flush of the skin, which lasted only a few minutes, is distinct from the

15 long lasting area of redness that remains surrounding the injection site for several hours, as previously described (Serra et al., 1998). The maximal area of this initial flush was similar to the area that subsequently became hyperalgesic (e.g. Figure 10).

FIGURE 10 NEAR HERE

Discussion This study has confirmed that human C nociceptors can be separated into two functional groups by measuring activity-dependent slowing. The mechanically insensitive CMi subgroup were mostly activated by injection of capsaicin, injected 10-50 mm from where they were being stimulated electrically, and most CMi units developed a mechanical sensitivity in the area of secondary hyperalgesia. In contrast, the mechanically sensitive CM units were unaffected by injection of capsaicin at these distances. After relating these results to previous studies, we discuss the likely roles of CMi units in secondary hyperalgesia.

Classification of C units by electrical and natural stimulation Weidner et al. (1999) reported that CMi units in human skin could be recognized by the electrical properties of their axons: compared with CM units they conducted more slowly and showed more prominent activity-dependent slowing at very slow stimulation rates, so that after at least two minutes without stimulation they slowed appreciably on stimulation at only 0.125 or 0.25 Hz. Our results with a 3-minute pause in baseline stimulation at 0.25 Hz (Figure 3) are in excellent agreement, and leave no doubt that our Type 1A fibers correspond to their mechano-sensitive units and our Type 1B fibers to their mechano-insensitive units. A slight difference in the degree

16 of slowing at 0.25 Hz after a pause for the CMi units, given as averaging 6.75% in Weidner et al. (2000) and 4.5% in the present study, may be related to their different recording site at the fibular head. The combination of pause and 2 Hz train provides further differentiation of the nociceptor fibers, which all slow progressively for 3 minutes, from non-nociceptive cold fibers, which reach a latency plateau after slowing by about 5% (Type 2, see Figure 1 and Campero et al., 2001), and a rarer, unidentified third type of afferent which hardly slows at all (Type 3, Serra et al., 1999). The great advantage of this protocol in the present study was that it enabled mechano-sensitive and mechano-insensitive nociceptors to be identified unambiguously without extensive mechanical probing, which could have carried the risk of sensitizing units before the capsaicin was applied.

Delayed activation of CMi units by capsaicin in the area of secondary hyperalgesia Previous studies have shown that both CM and CMi units are activated within a few seconds by local application of capsaicin, but that CMi units give more prolonged discharges, corresponding better with the duration of burning pain (Schmeltz et al., 2000b). We have now found that when capsaicin is injected at a distance of 10-50 mm, so that the recording site becomes engulfed in the area of secondary hyperalgesia, CM units are not activated, whereas CMi units often are activated, but with appreciable delays. The mechanism of this delayed activation is uncertain. The possibility that it is caused by diffusion of capsaicin itself can be discounted: in other studies, with the same volume and concentration of capsaicin, it has been concluded that the drug does not spread more than a few millimeters from the site of injection (LaMotte et al., 1991; Baumann et al., 1991 in the monkey). Also, in this study the capsaicin-sensitive CM units were unaffected, and amongst the CMi units there was no correlation between activation delay and distance. Another possible explanation of the delayed activation could be that the activation is

17 indirect, via a cascade system (cf. Lembeck & Gamse, 1982; Lynn & Cotsell, 1991). The distances over which the CMi units were activated in this study are actually modest, in comparison with their extensive arborisations in the skin described by Schmidt et al. (2002): CMi innervation territories are made up of multiple sub-fields, extending up to 72 mm (mean 45 mm). Although we were not able to excite branches of the CMi units recorded extending into the area of capsaicin injection, we cannot exclude the possibility that there were deep branches, which the capsaicin may have taken several minutes to reach. However, the time courses of the delayed responses were hardly consistent with a slow penetration of the algogen to the receptor membrane: the most delayed responses (e.g. unit 3 in Fig. 9) had a conspicuously abrupt onset, suggesting that they may have been mediated by another cell type, which could have integrated a weak capsaicin signal over several minutes, before releasing a packet of algogen close to the CMi unit. A further possibility is that the delayed activity recorded in CMi units was caused by dorsal root reflexes set up by nociceptive inputs into the dorsal horn and recorded antidromically. Actually, Lin et al., 1999, have recorded such activity in rats, and shown that it is an important contributor to the development of flare, 15-20 mm from the site of injection of capsaicin. However, the observation that flare can still be evoked in acutely denervated human skin until axonal degeneration has completely occurred (Lewis, 1935-6) speaks against the possibility that dorsal root reflexes might have been the cause of the spontaneous activity recorded in our experiments.

Altered responsiveness of CMi units in the area of secondary hyperalgesia Whereas all the Type 1B units tested were insensitive to mechanical stimulation prior to remote capsaicin injection, a high proportion (11/17) became responsive afterwards. These mechanically

18 evoked discharges were usually very different in time course from those of the CM units. They were characteristically delayed and prolonged, and in some cases resembled a resumption of spontaneous activity. Similar delayed responses were recorded by Schmelz et al. (2000b) from CMi units sensitized by local application of capsaicin to their electrical receptive field. The fact that the responses to mechanical stimuli are delayed and build up for a few seconds suggests an indirect mechanism in the mechanical response, most probably mediation by an algogenic chemical. The newly acquired mechanical responsiveness may be due to a heightened sensitivity to algogenic chemicals released from the skin by the noxious mechanical stimuli, and/or to an enhanced release of such substances from damaged skin, rather than to acquired mechanosensitivity of the receptor membranes. However, the nature and origin of the hypothetical chemical mediator(s) are unknown. We reported previously that some units fulfilling the definition of CMi units were activated spontaneously and became sensitized in areas of secondary hyperalgesia (Serra et al., 1995), but encountered the criticism that the activated or sensitized units might have been different from the ones recorded before the capsaicin injection. The use in this study of raster plots to follow the behavior of multiple single units for long periods avoids this potential criticism. On the other hand, the continuous stimulation of the CMi units incurs a potential risk that they may become sensitized by the electrical stimulation. Our recordings provided no evidence that this occurred: neither spontaneous discharges nor mechanically evoked responses were seen in any of the CMi units prior to capsaicin injection, even though we had in some cases already been stimulating electrically for over one hour. The behavior of CMi units to distant capsaicin injection was previously studied by Schmelz et al. (2000b), who also monitored the units by continuous stimulation at 0.25 Hz. They injected capsaicin into one part of the electrical receptive field and explored the behavior of the

19 other parts of the receptive field not challenged directly by capsaicin. They found evidence of sensitization of the part of the unit where capsaicin had been injected, but no sensitization of adjacent parts of the electrical receptive field to which capsaicin had not spread. Our results appear to be in conflict, since Schmelz et al. (2000b) were testing CMi branches at distances from the capsaicin injection site well within the range where we have found frequent sensitization. The discrepancy could be due to their use of a lower dose of capsaicin (8-20µg, as against our 100 µg). The lower dose was appropriate for limiting the number of units activated directly, whereas our higher dose is consistent with the amounts used in most psychophysical experiments on secondary hyperalgesia. Both studies are in agreement in showing that sensitization of CMi units in the area of secondary hyperalgesia is not due to impulses carried by axon reflexes in branches of the same units: Schmelz et al. (2000b) specifically activated distant branches of the same units and saw no sensitization, whereas we saw a discrepancy between activation and sensitization in 4 units. We infer that the sensitization is due to changes in the skin caused mainly by activity in CMi axons other than the one being stimulated electrically.

What is the role of CMi units in secondary hyperalgesia?

CMi units and the spread of flare and hyperalgesia Several studies (Lewis, 1935-6; LaMotte et al., 1991; Serra et al., 1998) have demonstrated that anaesthetizing a narrow strip of skin can block the spread of hyperalgesia from the site of injury or capsaicin injection. Lewis (1935-6) postulated that a system of widely arborising ‘nocifensor’ fibers (probably without sensory function but with cell bodies in the dorsal root ganglion) mediated the spread by releasing chemicals in the skin. After a detailed psychophysical reinvestigation of capsaicin-induced secondary hyperalgesia, LaMotte et al. (1991) concluded

20 that although the spread of hyperalgesia is mediated by chemo-sensitive C afferents with long arborising branches (or with functional coupling between pairs of fibers), the long-lasting facilitation must occur in the spinal cord. Our new data, coupled with other recent studies of CMi units, show that they are excellent candidates to fulfill the role of Lewis’s hypothetical nocifensor fibers, or LaMotte et al.’s (1991) hypothetical chemo-sensitive fibers, even though they probably have a sensory function as well. First, CMi units are chemo-sensitive, activated directly by capsaicin (Schmelz et al., 2000b), and they arborise widely (Schmidt et al., 1998), so that they are well placed to take the capsaicin signal into the region of secondary hyperalgesia. Secondly, we have recorded from CMi units in the area of secondary hyperalgesia and found that they are activated, and the delays of 0.5 - 18 minutes are commensurate with the recorded delays in the spread of secondary hyperalgesia. Thirdly, it is likely that CMi units are responsible for mediating flare (Schmelz et al., 2000a), and there is often a remarkably detailed correspondence between the area of flare (as described by Lewis (1935-6) or measured thermographically (Serra et al., 1998)) and the area of secondary hyperalgesia.

How does CMi unit activity contribute to secondary hyperalgesia? The most conspicuous change in sensation in the area of secondary hyperalgesia is an increase in pricking pain to punctate stimuli, mediated by myelinated afferents (Ziegler et al., 1999). According to the models of LaMotte et al. (1991) and Ziegler et al. (1999) the role of CMi units in secondary hyperalgesia should be restricted to triggering or maintaining a central facilitation of Aδ pathways. One reason why LaMotte et al. (1991) discounted a role of sensitized primary afferents in the secondary hyperalgesia was because no type of afferent with the required properties had been found (e.g. Baumann et al., 1991). Our demonstration that CMi units change

21 their responsiveness in the area of secondary hyperalgesia reopens the controversial question of the extent to which peripheral mechanisms contribute to this phenomenon (Coderre & Katz, 1997). Nevertheless, LaMotte et al.’s (1991) demonstration that punctate hyperalgesia is abolished if the capsaicin is applied during a proximal nerve block (even though the secondary hyperalgesia normally lasts much longer than the nerve block), and Ziegler et al.’s (1999) evidence from pressure blocks that this type of hyperalgesia depends on myelinated afferents, appear to exclude a significant role for CMi sensitization in the punctate secondary hyperalgesia. What then are the likely sensory consequences of the changes in CMi responsiveness that we recorded? Lewis (1935-6) reported that in the area of secondary hyperalgesia, needle pricks 'give unusually intense, diffuse, and long-lasting pain'. The newly acquired responses of CMi units to mechanical stimulation were mainly delayed and prolonged, and it seems likely that they contribute to the long-lasting component of the altered sensation, as previously argued for altered sensations at the site of capsaicin injection (Schmelz et al., 2000b). Also, since the CMi units arborise very widely, it is logical that they should only contribute to diffuse sensations. Sensitization of CMi units may also contribute to pathological hyperalgesia: a recent microneurographic study of patients with pain and hyperalgesia found C afferent fibers displaying activity-dependent slowing similar to our Type 1B units that were mechanically sensitive, and this responsiveness was regarded as evidence of sensitization of normally silent nociceptors (Orstavik et al., 2003). The altered responsiveness of the CMi units observed in the area of secondary hyperalgesia implies that there was either an altered resting chemical environment in the skin, or altered release of chemicals in response to noxious stimulation. The chemical(s) responsible are not known, but they appear not to affect the neighboring CM units (e.g. Fig. 10), in line with several reports that CM units respond normally in the area of secondary hyperalgesia (Baumann

22 et al., 1991; Schmelz et al., 1996). However, it is not unlikely that some Aδ nociceptors should have a similar chemical susceptibility to the CMi units and be able to contribute to the hyperalgesia. This possibility will be difficult to test, however, since Aδ fibers are considerably more difficult to record by microneurography than C fibers, and their activity cannot be monitored so conveniently by activity-dependent slowing. This raises the possibility that there may be a secondary hyperalgesia to first pain (mediated by Aδ fibers) and a secondary hyperalgesia to second pain (mediated by C fibers, in this case, mechano-insensitive C nociceptors). As far as we are aware, nobody has studied the differential behavior of first and second pain in the area of secondary hyperalgesia.

Conclusion In conclusion, we have shown that, following intradermal capsaicin injection, the spread of hyperalgesia into surrounding skin is accompanied by delayed activation of one category of C nociceptor (CMi units). These fibers probably represent the widely arborising chemo-sensitive C afferents in the model of LaMotte et al. (1991), whose activation in the area of secondary hyperalgesia was proposed to lead to facilitation of Aδ nociceptive pathways in the spinal cord. However, we also found that many of these units become responsive to noxious mechanical stimuli. Since CMi units are thought to have a nociceptive function (Schmelz et al., 2000b; Schmidt et al., 2000), this sensitization should contribute a direct, C fiber component to the secondary mechanical hyperalgesia.

Acknowledgement This work was supported by NIH grant RO1 NS-39761

23

References

Ali Z, Meyer RA, Campbell JN. Secondary hyperalgesia to mechanical but not heat stimuli following a capsaicin injection in hairy skin. Pain 1996; 68: 401-11. Baumann TK, Simone DA, Shain CN, LaMotte RH. Neurogenic hyperalgesia: the search for the primary cutaneous afferent fibers that contribute to capsaicin-induced pain and hyperalgesia. J Neurophysiol 1991; 66: 212-27. Campbell JN, Meyer RA. Sensitization of unmyelinated nociceptive afferents in monkeys varies with skin type. J Neurophysiol 1983; 49: 98-110. Campero M, Serra J, Bostock H, Ochoa J. Slowly conducting afferents activated by innocuous low temperature in human skin. J Physiol 2001; 535: 855-65. Coderre TJ, Katz J. Peripheral and central hyperexcitability: differential signs and symptoms in persistent pain. Behav Brain Sci 1997;20:404-19.Culp WJ, Ochoa J, Cline M, Dotson R. Heat and mechanical hyperalgesia induced by capsaicin. Cross modality threshold modulation in human C nociceptors. Brain 1989; 112: 1317-31. Davis KD, Meyer RA, Campbell JN. Chemosensitivity and sensitization of nociceptive afferents that innervate the hairy skin of monkey. J Neurophysiol 1993; 69: 1071-81. Delius W, Hagbarth KE, Hongell A, Wallin BG. Manoeuvres affecting sympathetic outflow in human skin nerves. Acta Physiol Scand. 1972;84:177-86. Hallin RG, Torebjörk HE. Single unit sympathetic activity in human skin nerves during rest and various manoeuvres . Acta Physiol Scand 1974; 92: 303-17. Hardy JD, Wolff HG, Goodell H. In: Pain Sensations and Reactions. Baltimore: Williams & Wilkins Co., 1952.

24 Kilo S, Schmelz M, Koltzenburg M, Handwerker HO. Different patterns of hyperalgesia induced by experimental inflammation in human skin. Brain 1994; 117: 385-96. Koltzenburg M, Torebjörk HE, Wahren LK. Nociceptor modulated central sensitization causes mechanical hyperalgesia in acute chemogenic and chronic neuropathic pain. Brain 1994; 117: 579-91. LaMotte RH, Thalhammer JG, Torebjörk H-E, Robinson CJ. Peripheral neural mechanisms of cutaneous hyperalgesia following mild injury by heat. J Neurosci 1982; 2: 765-81. LaMotte RH, Thalhammer JG, Robinson CJ. Peripheral neural correlates of magnitude of cutaneous pain and hyperalgesia: a comparison of neural events in monkey with sensory events in humans. J Neurophysiol 1983; 50: 1-26. LaMotte RH, Shain CN, Simone DA, Tsai E-FP. Neurogenic hyperalgesia: psychophysical studies of underlying mechanisms. J Neurophysiol 1991; 66:190-211. LaMotte RH, Lundberg LER, Torebjörk H-E. Pain, hyperalgesia and activity in nociceptive C units in humans after intradermal injection of capsaicin. J Physiol 1992; 448: 749-64. Lembeck F, Gamse R. Substance P in peripheral sensory processes. In 'Substance P in the nervous system' (CIBA Symposium), Pitman: London 1982; 91: 35-54. Lewis T. Experiments relating to cutaneous hyperalgesia and its spread through somatic nerves. Clin Sci 1935-6; 2: 373-421. Lewis T. In: Pain. London: The MacMillan Press Ltd, 1942. Lin Q, Wu J, Willis WD. Dorsal root reflexes and cutaneous neurogenic inflammation after intradermal injection of capsaicin in rats. J Neurophysiol 1999; 82: 2602–2611. Lynn B, Cotsell B. The delay in onset of vasodilator flare in human skin at increasing distances from a localized noxious stimulus. Neurovascular Res 1991; 41: 197-202. Magerl W, Fuchs PN, Meyer RA, Treede RD. Roles of capsaicin-insensitive nociceptors in

25 cutaneous pain and secondary hyperalgesia. Brain 2001; 124: 1754-64. Meyer RA, Davis KD, Cohen RH, Treede R-D, Campbell JN. Mechanically insensitive afferents (MIAs) in cutaneous nerves of monkey. Brain Res 1991; 561: 252-61. Orstavik K, Weidner C, Schmidt R, Schmelz M, Hilliges M, Jorum E, Handwerker H, Torebjörk E. Pathological C-fibres in patients with a chronic painful condition. Brain. 2003;126:567-78.Pedersen JL, Kehlet H. Secondary hyperalgesia to heat stimuli after burn injury in man. Pain 1998; 76: 377-84. Raja SN, Campbell JN, Meyer RA. Evidence for different mechanisms of primary and secondary hyperalgesia following heat injury to the glabrous skin. Brain 1984; 107: 1179-88. Sang CN, Gracely RH, Max MB, Bennett GJ. Capsaicin-evoked mechanical allodynia and hyperalgesia cross nerve territories. Evidence for a central mechanism. Anesthesiol 1996; 85: 491-6. Schmelz M, Forster C, Schmidt R, Ringkamp M, Handwerker HO, Torebjörk HE. Delayed responses to electrical stimuli reflect C-fiber responsiveness in human microneurography. Exp Brain Res 1995; 104: 331-6. Schmelz M, Schmidt R, Ringkamp M, Forster C, Handwerker HO, Torebjörk H-E. Limitation of sensitization to injured parts of receptive fields in human skin C-nociceptors. Exp Brain Res 1996; 109: 141-47. Schmelz M, Schmidt R, Bickel A, Handwerker HO, Torebjork HE. Specific C-receptors for itch in human skin. J Neurosci. 1997; 17: 8003-8. Schmelz M, Michael K, Weidner C, Schmidt R, Torebjörk HE, Handwerker HO. Which nerve fibers mediate the axon reflex flare in human skin? Neuroreport 2000a; 11: 645-648. Schmelz M, Schmidt R, Handwerker HO, Torebjörk H-E. Encoding of burning pain from capsaicin-treated human skin in two categories of unmyelinated nerve fibers. Brain 2000b; 123:

26 560-71. Schmidt R, Schmelz M, Forster C, Ringkamp M, Torebjörk H-E, Handwerker H. Novel classes of responsive and unresponsive C nociceptors in human skin. J Neurosci 1995; 15: 333-41. Schmidt R, Schmelz M, Bickel A, Weidner C, Messlinger K, Handwerker HO, et al. Innervation territories of mechanoinsensitive C nociceptor units in human skin. [Abstract] Soc Neurosci Abstr 1998; 24: 383. Serra J, Campero M, Ochoa J. “Secondary” hyperalgesia (capsaicin) mediated by C-nociceptors. [Abstract] Soc Neurosci Abstr 1993; 2: 965. Serra J, Campero M, Ochoa J. Common peripheral mechanism for neurogenic flare and hyperalgesia (capsaicin) in human skin. Muscle Nerve 1994; S250. Serra J, Campero M, Ochoa J. Sensitization of "silent" C-nociceptors in areas of secondary hyperalgesia (SH) in humans. [Abstract] Neurology 1995; 45: A365. Serra J, Campero M, Ochoa J. Flare and hyperalgesia following intradermal capsaicin injection in human skin. J Neurophysiol 1998; 80: 2801-10. Serra J, Campero M, Ochoa J, Bostock H. Activity-dependent slowing of conduction differentiates functional subtypes of C fibres innervating human skin. J Physiol 1999; 515.3: 799811. Simone DA, Baumann TK, LaMotte RH. Dose-dependent pain and mechanical hyperalgesia in humans after intradermal injection of capsaicin. Pain 1989; 38: 99-107. Simone DA, Sorkin LS, Oh U, Chung JM, Owens C, LaMotte RH, et al. Neurogenic hyperalgesia: central neural correlates in responses of spinothalamic tract neurons. J Neurophysiol 1991; 66: 228-46. Thalhammer JG, LaMotte RG. Spatial properties of nociceptor sensitization following heat injury to the skin. Brain Res 1982; 231: 257-65.

27 Thalhammer JG, Raymond SA, Popitz BF, Strichartz GR. Modality-dependent modulation of conduction by impulse activity in functionally characterized single cutaneous afferents in the rat. Somatosens Motor Res 1994; 11: 243-57. Torebjörk H-E, LaMotte RH, Robinson CJ. Peripheral neural correlates of magnitude of cutaneous pain and hyperalgesia: simultaneous recordings in humans of sensory judgements of pain and evoked responses in nociceptors with C-fibers. J Neurophysiol 1984; 51: 325-39. Torebjörk H-E, Ochoa JL. New method to identify nociceptor units innervating glabrous skin of the human hand. Exp Brain Res 1990; 81: 509-14. Torebjörk HE, Lundberg LER, LaMotte RH. Central changes in processing of mechanoreceptive input in capsaicin-induced secondary hyperalgesia in humans. J Physiol 1992; 448: 765-80. Vallbo Å, Hagbarth K-E. Activity from skin mechanoreceptors recorded percutaneously in awake human subjects. Exp Neurol 1968; 21: 270-89. Weidner C, Schmelz M, Schmidt R, Hansson B, Handwerker HO, Torebjörk H-E. Functional attributes discriminating mechano-insensitive and mechano-responsive C nociceptors in human skin. J Neurosci 1999; 15: 10184-90. Weidner C, Schmidt R, Schmelz M, Hilliges M, Handwerker HO, Torebjörk H-E. Time course of post-excitatory effects separates afferent human C fibre classes. J. Physiol 2000; 527: 195-191. Woolf CJ, Mannion RJ. Neuropathic pain: aetiology, symptoms, mechanisms, and management. Lancet 1999; 353:1959-64. Ziegler EA, Magerl W, Meyer RA, Treede R-D. Secondary hyperalgesia to punctate mechanical stimuli: central sensitisation to A-fibre nociceptor input. Brain 1999; 122: 2245-2257.

28 TABLES

RESPONSE TO CAPSAICIN INJECTION

UNIT

Mechanosensitive

Distance (mm)

Activation

Time to activation

Acquired mechanosensitivity

(min)

1

No

45

No

-

Yes

[Fig. 8, 3(1B)]

2

45

No

-

No

[Fig. 8, 2 (1B)]

3

No ?

30

Yes

0.5

Lost

4

?

12

Yes

2.5

Yes

5

No

12

No

-

No

6

No

15

Yes

2.7

Yes

[Fig. 7, 4 (1B)]

7

No

30

Yes

2.7

Lost

[Fig. 6, 1 (1B)]

8

No

30

Yes

2.1

Yes

[Fig. 6 2 (1B)]

9

No

30

Yes

17.8

No

[Fig. 6,3 (1B)]

10

?

50

Not recorded

-

Yes

11

?

10

Not recorded

-

No

12

?

10

Not recorded

-

No

13

No

40

No

-

Yes

14

No

40

Yes

13.0

Yes

15

No

40

Yes

14.0

Yes

16

No

10

Yes

1.3

Yes

17

No

20

Yes

1.1

Yes

[Fig. 9, 2 (1B)]

18

?

20

Recruited

-

Yes

[Fig. 9, 1 (1B)]

19

No

20

Yes

4.8

No

[Fig. 5, 2 (1B)]

Table 1 : List of all Type 1B units challenged with injection of capsaicin. (? = not tested.). N.B. For units 10-12, capsaicin injection was made prior to microneurography, so that baseline responsiveness and activation by capsaicin could not be recorded. Unit 18 was only recruited by the electrical stimuli 49 minutes after capsaicin injection. Recordings from 9 of the units are illustrated, as indicated in the right hand column.

29 FIGURE LEGENDS

Figure 1: TOP: (A) Raster plot, showing latency changes of many units when baseline stimulation at 0.25 Hz is interrupted by a 3-minute pause and a 3-minute period of stimulation at 2 Hz. Each peak in the filtered voltage signal that exceeds a specified level is represented by a dot, with latency as the ordinate and elapsed time as the abscissa. Depending on the level chosen, the dots can represent action potentials or noise. Several units are easily distinguished from noise by their fixed latency at 0.25 Hz baseline stimulation. (B) For the raster figures in this paper, latencies of selected units with adequate signal-to-noise were remeasured from the raw data so that each dot represents an identified single unit. As in the other modified raster plots which follow, the selected units are numbered in order of increasing latency (1-5), and their types, according to their profiles of activity-dependent slowing (see text), are given within brackets. BOTTOM: Example of original microneurographic response (filtered and inverted), from 90-550 ms after stimulus applied, at elapsed time 10 minutes. The selected units are numbered in order of increasing latency (1-5) as on the raster plot at the top.

Figure 2: Type 1A and 1B units distinguished by the pause protocol and by natural stimulation. Baseline stimulation at 0.25 Hz reveals four units, numbered 1 to 4 in order of increasing latency. Pronounced slowing of conduction velocity after the pause marks unit 4 as Type 1B, while the other three units with shorter latencies display minimal or no slowing, and are therefore classified as Type 1A. Mechanical stimulation of the electrical receptive fields of the units evokes responses in the first 2 units (M1= 52 mN, M2 = 68 mN) with weaker von Frey hairs, displayed

30 as sudden shifts in the stable baseline latency (“marking technique”). Stimulation with stronger von Frey hairs (M3 = 109 mN, M5 = 307 mN) is also able to excite unit 3. Application of a hot rod evokes responses in all 3 Type 1A units. The Type 1B unit (unit 4) is unresponsive to both mechanical and heat stimulation, and is therefore classified as mechano- and heat-insensitive.

Figure 3: Relationship between percentage of slowing after the 3-minute pause and receptor properties of 58 Type 1 fibres. Each square represents a single unit and its degree of slowing of conduction velocity after the pause. Squares labelled M indicate units identified as mechano-sensitive, while squares labelled I indicate units identified as mechano-insensitive. Blank squares indicate units that were not tested for mechanical sensitivity. It is clearly shown that Type 1A units were all mechano-sensitive, while Type 1B were mechano-insensitive.

Figure 4: A: Relationship between conduction velocity and percentage of slowing after the 3-minute pause for 53 Type 1 units. Filled circles: Type 1A units. Open circles: Type 1B units. Ellipses represent 95% confidence limits for a member of each group. B: Effect of duration of repetitive stimulation on slowing in 30 Type 1A and 26 Type 1B fibres. Slowing (i.e. increase in latency, expressed as percentage of control) measured at 1 and 3 min after start of 2 Hz repetitive stimulation. Symbols connected by lines denote same unit. Open circles for Type 1B indicate units that were lost during repetitive stimulation at 3 minutes. C: Relationship between rate of recovery from repetitive stimulation and unit type. Top: time for recovery to 50 % of resting latency after stimulation at 2 Hz for 3 min. Bottom: percentage

31 recovery 30 s after same train. Horizontal bars indicate mean values. Differences between Type 1A and 1B were significant in each case. D: Relationship between rate of recovery from stimulation at 2 Hz and slowing after the pause for 23 Type 1A units (filled circles) and 19 Type 1B units (open circles).

Figure 5: Modified raster plot showing responses of 3 C units to the 3-minute pause, 2 Hz train and remote injection of capsaicin. Slowing of conduction velocity after the pause of units 2 and 3 marks them as Type 1B units (CMi), while minimal slowing of the first unit marks it as a Type 1A unit (CM). The Type 1B units are also distinguished by their slower recovery after the 2 Hz train. Capsaicin was injected 12 mm away from the electrical receptive field of the units at the arrow. After 2.5 minutes, the Type 1B (CMi) unit 2 engaged in several bursts of spontaneous activity, shown as sudden shifts of the stable baseline latency. Bursts of activity appeared one on top of the other for over 20 minutes. N.B. Following the capsaicin injection, there was an increase in skin temperature that resulted in a slight reduction in the latencies of all the units.

Figure 6: Modified raster plot showing spontaneous activity induced by remote capsaicin injection (30 mm from electrical receptive field) in 3 Type 1B (CMi) units. Bursts of continuous activity appeared in first two units between 2 and 3 minutes after capsaicin injection, and lasted for about 25 minutes. Unit 3 also engaged in spontaneous bursts of activity, which did not appear until 17 minutes after the capsaicin injection and consisted of 3 intense bursts at 4-minute intervals.

Figure 7:

32 Remote activation and sensitisation of Type 1B unit by capsaicin. The modified raster plot is a continuation of the recording in Figure 2 (time scale is the same for both figures). After the testing for mechanical and heat responses (see Figure 2), capsaicin was injected 15 mm from the electrical receptive fields. Only unit 4, the Type 1B fibre that had not responded to the previous stimuli, engaged in a single burst of spontaneous activity 3 minutes later. After some minutes, the units were again tested mechanically (M1 = 52 mN, M2 = 68 mN) with the weaker von Frey filaments used before capsaicin. Units 1 and 2 responded in a similar way as before capsaicin. Unit 3 did not respond to these weaker von Frey filaments, as before capsaicin, so there was no evidence of sensitisation of the Type 1A units. However, the Type 1B unit, which was unresponsive to mechanical stimulation before capsaicin, was now clearly activated by the weaker von Frey filaments. The newly developed mechanical responses of the Type 1B unit appeared some seconds after the initial probing, and were of a different time course to the responses of the Type 1A units.

Figure 8: Remote sensitisation without activation of Type 1B unit by capsaicin. This modified raster plot shows one Type 1A and 2 Type 1B units, identifiable by the slowing of the latter after the 3minute pause at 3 minutes. As expected, only the Type 1A unit responded to mechanical stimulation, even with the strongest von Frey hair used (M3 = 109 mN, M6 = 722 mN). After capsaicin was injected 45 mm from the electrical receptive field, none of the units engaged in spontaneous activity over a period of 20 minutes. However, following probing with the weaker von Frey hair (M3), unit 3 started a prolonged discharge. After the activity had subsided, application of a heated rod (H, ~45ûC) at 81.5 minutes provoked another prolonged discharge in this Type 1B unit, as well as a brief response from the Type 1A unit. The other Type 1B unit

33 (unit 2) was unaffected by capsaicin or any of the stimuli throughout the recording period. The jumps between two stable latencies, which started at 63 minutes, showed no relationship to the stimulation.

Figure 9: Responses of 2 sensitised Type 1B units to mechanical stimulation, showing different discharge pattern to previous sensitised units (Figs. 10, 11). In this example, capsaicin was injected at a distance of 20 mm, 66 minutes before the start of the figure. Both units were excited by a pinprick (P) at 67.5 minutes, at 87 minutes, and unit 2 was activated by von Frey hairs (M4 = 178 mN, M5 = 307 mN). On this occasion, responses to mechanical stimulation took the form of single bursts, with no prolonged afterdischarges.

Figure 10: Illustration of typical experiment, showing areas of early flare and mechanical hyperalgesia in relation to site of capsaicin injection and electrical stimulation. Distance from capsaicin injection site to stimulating electrodes = 40 mm. Recording electrodes (not depicted) were inserted into the superficial peroneal nerve at ankle level.

34

B

A

5(1B)

4(2)

3(1A) 2(1A) 1(2)

1

2

3

Serra, Campero, Bostock and Ochoa: Figure 1

4

5

35

4(1B)

3(1A)

2(1A) 1(1A) M5

Serra, Campero, Bostock and Ochoa: Figure 2

36

Serra, Campero, Bostock and Ochoa: Figure 3

37

A

B

C

D

Serra, Campero, Bostock and Ochoa: Figure 4

38

3(1B)

2(1B)

Capsaicin at 12 mm

Serra, Campero, Bostock and Ochoa: Figure 5

1(1A)

39

3(1B)

2(1B)

1(1B)

Capsaicin at 30 mm

Serra, Campero, Bostock and Ochoa: Figure 6

40

4(1B)

3(1A)

2(1A) 1(1A) M5

Capsaicin at 15 mm

Serra, Campero, Bostock and Ochoa: Figure 7

41

3(1B)

2(1B)

1(1A) Capsaicin at 45 mm M3,6

Serra, Campero, Bostock and Ochoa: Figure 8

M3

M3

42

Serra, Campero, Bostock and Ochoa: Figure 9

43

Flare

Mechanical hyperalgesia

Electrical RF Capsaicin

Serra, Campero, Bostock and Ochoa: Figure 10