| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Department of Neuroscience and 1Graduate Programs in Neuroscience and 2Pharmacology, University of Minnesota, Minneapolis, Minnesota
Submitted 18 April 2005; accepted in final form 30 November 2005
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
nociceptors innervating inflamed skin were opiate-sensitive, and their excitability was attenuated by direct application of morphine to their receptive fields. All morphine-sensitive units were nociceptors from inflamed skin with conduction velocities <1.3 m/s. Morphine effects were concentration-dependent and naloxone-sensitive, indicating that the effects were receptor-mediated. These findings provide direct evidence that morphine acts through peripheral opioid receptors to inhibit the activity of cutaneous nociceptors under conditions of inflammation. |
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
; Stein 1995; Stein and Yassouridis 1997; Stein et al. 1999). In most of these studies, peripherally administered opiates return hyperalgesia pain responses to preinflammatory baseline levels without reducing it further. These behavioral effects are mediated by mu, delta, and kappa opioid receptors, and they have been shown to be dose-dependent, stereospecific, and antagonist reversible (Barber and Gottschlich 1992; Ferreira and Nakamura 1979; Joris et al. 1990; Levine and Taiwo 1989; Stein et al. 1989).
Opioid receptors have been detected in healthy tissues (Coggeshall et al. 1997; Stander et al. 2002; Stein et al. 1990b; Wenk and Honda 1999); however, cutaneously applied opiates have little discernable effect on nociceptive thresholds in normal animals (Stein 1993). Topically applied opiates may be acting directly on the peripheral processes and terminals of sensory neurons in inflamed skin. This is suggested by decreased behavioral response to noxious stimuli (Antonijevic et al. 1995; Hargreaves and Joris 1993; Lawrence et al. 1992; Peyman et al. 1994; Stein and Yassouridis 1997; Stein et al. 1989, 1990a, b) and reductions in Substance P release (Brodin et al. 1983; Yaksh 1988; Yonehara et al. 1988). After peripheral inflammation, radioligand-binding experiments indicate that opioid receptor accumulation at the level of the sciatic nerve is increased (Brodin et al. 1983; Hassan et al. 1993). The local effect of opiates applied directly to peripheral afferent fibers is less clear.
In the present study, we provide electrophysiological evidence for functional opioid receptors on peripheral terminals of primary afferent neurons. Using an in vitro skin-nerve preparation, we compared the effects of morphine on the response properties of single afferent fibers innervating normal and inflamed skin. Previous immunohistochemical findings (Wenk and Honda 1999) demonstrate that opioid receptors are abundant in unmyelinated peptidergic fibers and free-nerve endings in glabrous rat skin. We therefore chose to concentrate on single units that conducted in the C-fiber conduction velocity range. A smaller number of units conducting in the A
and A conduction velocity ranges were also examined for purposes of comparison. In skin treated with complete Freunds' adjuvant (CFA), morphine reduced the activity of most nociceptors in response to noxious mechanical and thermal stimulation in a concentration-dependent and naloxone-preventable manner. In contrast, morphine had no significant effect on mechanical or thermal responsiveness of nociceptors in control skin. These data provide the first direct electrophysiological evidence for functional opioid receptors on nociceptive neurons innervating inflamed skin.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
All experimental procedures were approved by the Institutional Animal Care and Use Committee at the University of Minnesota and conform to established guidelines. A total of 93 adult male Sprague-Dawley rats (200–400 g) was used in this study.
Induction of inflammation
Rats were deeply anesthetized with halothane (4% in air), and 100 µl of CFA (1:1 emulsion in saline, Sigma; St. Louis, MO) was injected subcutaneously into the plantar surface of the right hind paw. Rats recovered fully from anesthesia within a few minutes and werehoused in standard plastic cages with soft bedding. Eighteen hours after injection, the glabrous skin of the right hind paw was removed for in vitro recording. In preliminary behavioral studies, rats displayed increased thermal and mechanical sensitivity at this time point relative to baseline, in agreement with findings of other laboratories (Hylden et al. 1989; Iadarola et al. 1988).
Skin-nerve preparation
An isolated skin-nerve preparation (Reeh 1986) was used for combined electrophysiological and pharmacological study of single afferent fibers innervating plantar skin of the hind paws. Rats were anesthetized with a mixture of ketamine, acepromazine, and xylazine (6.8/0.09/0.45 mg/kg), and the entire plantar surface of the paw was dissected free along with the attached tibial and plantar nerves. The preparation was placed corium side up in a tissue chamber and constantly perfused (15–20 ml/min) with warmed (30 ± 0.5°) and oxygen-saturated synthetic interstitial fluid. The cut end of the nerve was placed on a small mirrored dissection platform and de-sheathed. Filaments were divided from the main nerve trunk using fine sharpened forceps and lifted on to a fine gold wire electrode for extracellular recording. Two different search strategies were used to isolate single units.
In the first approach, the tibial nerve was stimulated electrically, and filaments were progressively divided until single-unit activity was observed. The electrical receptive field of the unit was then delineated with a second stimulating electrode held in a micromanipulator and placed at progressively distal sites along the dividing branches of the plantar nerve. Cutaneous regions surrounding the most distal effective nerve stimulation site were then systematically searched for a site having the lowest electrical threshold. This electrical receptive field was then mechanically probed with a small glass rod.
The second approach utilized a hand-held glass rod as a mechanical search stimulus. The skin was gently probed while recording from a large nerve filament until a general area of innervation was determined. The branch of the plantar nerve nearest this area was then electrically stimulated while progressively dividing the filament until single-unit activity was evident. Regardless of which search strategy was employed, single-unit activity could be activated by electrical stimulation of the nerve and cutaneous receptive field. Mechanical and thermal stimuli were then applied to the electrically identified receptive field as described in the following sections. Units unresponsive to either mechanical or heat stimuli within the tested parameters were not studied further.
Neural signals were amplified (DAM80, World Precision Instruments, Austin, TX), filtered, then broadcast over an audio speaker, and displayed on an oscilloscope. All data were collected and analyzed with data acquisition software written on the LabView platform (National Instruments, Sarasota, FL). Software controlled thermal stimulus delivery, and it recorded neural activity, stimulus temperature, and timing of mechanical stimulus delivery. Data were analyzed on- and off-line.
Preparation of solutions
Synthetic interstitial fluid solution [SIF; containing (in mM) 123 NaCl; 3.5 KCl, 0.7 MgSO4, 2.0 CaCl2, 9.5 Na gluconate, 1.7 NaH2PO4,5.5 glucose, 7.5 sucrose, and 10 HEPES; pH 7.45 ± 0.05] (Koltzenburg et al. 1997) was prepared no more than 2 days in advance and stored at 4°C. Stock solutions of morphine sulfate and naloxone (Sigma) were diluted with SIF (pH 7.4–7.6). All solutions were saturated with oxygen before use.
Characterization of single units
Units were classified according to conduction velocity then characterized in terms of responses to mechanical and thermal stimulation. A, A, or C fibers were identified based on conduction velocity ranges obtained from whole-nerve compound action potential recordings made at the beginning of most experiments. Single units with conduction velocities that fell between the C and A wave ranges were classified as C/A units.
Mechanical response threshold was determined using calibrated von Frey filaments applied in order of increasing pressure. Threshold was defined as the pressure exerted by the smallest-diameter von Frey filament capable of evoking at least two impulses. If a mechanically -sensitive unit responded to a glass rod but not to the largest von Frey filament tested (9.3 bars), a threshold value of 10 bars was assigned for purposes of statistical analysis.
Units were also tested with constant pressures of 0.5, 1.8, 3.3, and 5.5 bars delivered with weighted probes having tip diameters of 2.5 mm. All mechanically sensitive units were tested with 1.8 and 3.3 bar stimuli to estimate stimulus response relationships spanning innocuous to noxious ranges. Most units were also tested with one or both of the remaining stimulus probes.
Feedback-controlled thermal stimuli were delivered using either a Peltier device placed in direct contact with the corium or a radiant heat stimulator directed at the epithelium through the translucent bottom of the recording chamber. The Peltier device (7 x 7 mm) delivered heat stimuli in 2°C steps of 5-s duration (rise time: 30°C/s) from a baseline temperature of 30°C. Some units were also tested for cooling and cold responses.
In later experiments, morphine was applied directly to the corium through a small metal ring (6 mm ID) placed around the receptive field to create a separate drug reservoir within the tissue bath. Because this limited access to the corium receptive field, thermal stimuli were delivered using a radiant heat source directed at the epidermal surface through the bottom of the recording chamber. A circular area of illumination (6–7 mm diam) was centered under the receptive field, and feedback control was referenced to a needle thermocouple inserted into the corium in the center of the drug reservoir ring. Feedback control was most reliable at lower thermal ramp rates, so a slope of 0.5°C/s was used in all heat trials. Even at this slow rate of heating, a temperature gradient existed between the epithelial and the corium sides of the preparation, reflecting a nonlinear rate of thermal conduction through the skin. A temperature of 35°C on the corium corresponded to an epithelial temperature of 38°C, and a corium temperature of 45°C corresponded to 50°C. A maximum temperature of 48°C (corresponding to an epithelial temperature of 53.5°C) was used in heat test trials.
Most heat-responsive units began to discharge slowly and increased in firing frequency until reaching a sustained firing rate several seconds later. Units allowed to continue maximal sustained firing for more than a few seconds frequently began to burst erratically or stopped firing completely before the end of the heat ramp stimulus was reached. These units remained thermally insensitive or responded erratically to a second heat ramp for up to an hour after initial testing. For this reason, thermal ramps were stopped manually before the maximum temperature of 48°C was reached, if the unit responded clearly to the heat stimulus and had maintained a sustained firing rate for at least one second. The maximum temperature reached in the first heat trial was then used as the ramp endpoint for all subsequent heat test trials for that particular unit. Heat stimulation trials were performed 
15 min apart. Units not responding by 48°C were considered insensitive to heat. For each conduction velocity classification group, no significant difference was found between the heat thresholds for units tested with the Peltier device or the radiant heat source. Threshold data from the two groups were therefore combined for analysis.
Spontaneous activity
To determine the rate of on-going activity in the absence of intentional stimulation, the number of spikes in a 30 s time period was recorded for each unit. Single units with a mean baseline activity level >0.1 spike/s were classified as spontaneously active. A few units temporarily developed spontaneous activity or increased their baseline firing rate after mechanical stimulation of their receptive fields. This stimulus-induced afterdischarge was not included in calculations of spontaneous activity level.
Peripheral testing with morphine
All units were first characterized as described in the preceding text. After the first 3.3 bar mechanical stimulus trial, morphine was applied directly to the corium side of skin through a small metal ring that was sealed over the cutaneous receptive field with petroleum jelly to create a separate reservoir. At least 15 min after the initial heat ramp trial, the drug reservoir was emptied of SIF with a suction pipette and filled with warmed oxygenated morphine sulfate solution. After 2 min, a second heat trial was performed in the presence of morphine. The reservoir was then emptied, and response to the 3.3 bar mechanical stimulus was tested immediately after removal of the reservoir ring from the receptive field. The postdrug mechanical trial was therefore performed after morphine had been on the corium receptive field for 
5 min.
Thermal and mechanical responses were re-tested 15–20 min after removal of the reservoir ring and subsequent commencement of drug washout. Washout trials were repeated at the same interval until responses returned to baseline (50% reversal of drug-induced response change).
Baseline responses for each unit were determined from the initial mechanical and thermal stimulation trials during receptive field characterization. The average firing rate over a 5-s stimulus was determined for mechanical trials, and response threshold and the total number of spikes during the entire heat ramp were recorded for thermal trials. Data are expressed either as total number of spikes fired or as percent of baseline response for each unit.
Naloxone reversibility of morphine sulfate effect
The effect of naloxone on morphine responses was tested on five units from inflamed skin. After characterization of thermal and mechanical response properties, units were tested with 500 nM morphine. A unit was tested with naloxone only if it responded to morphine and it demonstrated a washout recovery of 50% of the drug-induced response change. An equimolar solution (500 nM) of naloxone was then applied to the receptive field for 3 min, followed by a second application of morphine. Mechanical and thermal testing was performed as described in the preceding text.
Response variability and drug effect criteria
Even with intervals between tests of 15 min, many units exhibited a large amount of variability in their responses to repeated stimulation trials. To draw valid conclusions about whether a change in response level after morphine application was due to the action of drug or to normal response variability, it was necessary to establish an objective criterion.
A total of 10 C-fiber units were tested according to the established peripheral drug testing protocol using vehicle (SIF) instead of morphine sulfate. Each identified unit was tested twice with a 15-min interval between trials. Second trial responses were expressed as percent of first trial (baseline) response for that particular unit. Average percent response for the second trial was not significantly different in C-fiber units from inflamed and normal skin, and the two groups were combined. Responses to heat and mechanical stimuli showed no significant difference between the first and second stimulation trials. The mean percent response for a second heat trial was 98.5 ± 20.7%. Mean percent response for repeated mechanical stimulation was 95.0 ± 12.0%.
Values within 2 SD of the sample mean were considered to be within the range of normal population response variability. Therefore for an individual unit to be classified as morphine-sensitive, the percent baseline response after morphine application was required to be 2 SD below the sample mean for percent baseline response under normal conditions.
Based on results from these experiments, a positive drug effect on heat response was defined as a response after morphine application equal to
57% baseline response for that particular unit. Likewise, an increase to >140% of baseline response value for that particular unit would be defined as drug-evoked excitation. A positive drug effect on mechanical response was defined as a change to <70 or >118.5% of baseline.
Data processing and statistical analyses are described as necessary for each section of RESULTS. Unless specified otherwise, values are expressed as means ± SD.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
Compound action potentials were recorded from the tibial nerve at the beginning of most experiments to assist in the subsequent classification of single units by conduction velocity. The conduction velocity ranges for each of the three major components (A
/, A and C waves) of all recorded compound action potentials are summarized in Fig. 1. At conduction distances available in this preparation, A and A waves often overlapped. Differentiation of these two waves was accomplished by adjustment of electrical stimulus intensity and duration or by extrapolation of the falling and rising phases of the two waves. The average maximum and minimum conduction velocities for the A/ wave component were 36.7 ± 8.5 m/s and 13.2 ± 2.4 m/s respectively, with a mean conduction velocity of 24.5 ± 5.1 m/s. The A waves had a mean conduction velocity of 9.5 ± 2.1 m/s and ranged from 13.8 ± 2.8 to 6.6 ± 1.8 m/s. C waves ranged from 0.76 ± 1.2 to 0.47 ± 0.1 m/s, with a mean of 0.62 ± 0.2 m/s. The A and C wave components of the compound action potentials never overlapped. The minimum conduction velocity value recorded for the A wave was 3.0 m/s, and the fastest conduction velocity recorded for the C wave was 1.0 m/s.
|
units. Those conducting between 13.5 and 3.0 m/s were classified as A units, and all units conducting below 1.0 m/s were classified as C fibers. Units conducting between 3.0 and 1.0 m/s were assigned to a fourth category, C/A. Functional classification of primary afferent units
A total of 185 single units isolated from the tibial nerves of 93 rats was examined, 107 were from normal skin, and 78 innervated inflamed skin. Receptive fields were evenly distributed across the plantar aspect of the foot (Fig. 2). No attempt was made to measure or otherwise quantify receptive field size.
|
|
|
; Hardy et al. 1952). With heat stimulation, we analyzed response rates over a range of 35–43°C (corium surface). This corresponded to temperatures of 40–48°C on the epidermal side of the preparation and encompasses the lower range of human heat pain thresholds in psychophysical studies (LaMotte and Campbell 1978; Meyer and Campbell 1981).
Nociceptors were further classified according to conduction velocity and stimulus response modality as follows: A-mechanical nociceptive units (AM), A-mechanical nociceptive units (ADM), A-mechanoheat nociceptive units (ADMH), A-mechanocold nociceptive units (ADMC), C/A-mechanical nociceptive units (C/ADM), C/A-mechanoheat nociceptive units (C/ADMH), C/A-cold nociceptive units (C/ADC), C-mechanical nociceptive units (CM), C-mechanoheat nociceptive units (CMH), C-heat nociceptive units (CH), and C-mechanoheat-cold nociceptive units (CMHC).
Units were classified as thermoreceptors (THERM) based on the following characteristics: static discharge at baseline temperature, dynamic response to temperature changes with either a positive (warming) or negative (cooling) coefficient, mechanically insensitive/very high mechanical response threshold (Hensel and Iggo 1971; Iggo 1969). The distribution of all functional classes of afferent fibers is summarized in Table 1.
A fiber units
Mean conduction velocity for A fibers was 21 ± 4 m/s for inflamed skin and 20.9 ± 5 m/s for units innervating normal skin. Of 19 normal skin A units, 18 were classified as LTM, and 1 was classified as AM based on mechanical stimulus-response function. All 6 A units from inflamed skin were LTM. The proportion of different subtypes within the A population was not significantly correlated with experimental condition (Fisher exact test, P > 0.05).
A fiber units
A total of 25 single units (20 normal skin, 5 inflamed skin) were classified as A fibers. Of the normal skin units, 2 were classified as LTM, and 18 (90%) were classified as nociceptors as follows: 16 ADM (86%), 1 ADMH, and 1 ADMC. Of 5 A units innervating inflamed skin, 3 were classified as ADM, and 2 were classified as LTM. The distribution of functional subtypes was independent of inflamed or normal skin experimental condition (Fisher exact test, P > 0.05).
C fiber units
Nociceptors were the most frequent type of C-fiber unit encountered in both inflamed skin and normal skin groups. Of 108 C-fiber units recorded in both normal and inflamed skin, 3 were LTM, 3 were warming fibers, and the remaining 102 units (94.4%) were nociceptors.
A total of 55 C-fiber units were recorded from normal skin. Of these, 3 were warming receptors, and 2 were LTM. The remaining 50 units were classified nociceptors as follows: 19 CM (38%), 24 CMH (48%), 6 CH (12%), and 1 CMHC. Of 53 inflamed skin units, 52 were nociceptors, and 1 was LTM. Of the nociceptors, 22 were CM (42%), 19 CMH (37%), and 11 CH (21%). The distribution of nociceptor subtypes was not significantly different in units from inflamed as compared with normal skin (
2 test, P > 0.05).
C/A fiber units
A total of 14 units from inflamed skin and 13 from normal skin had conduction velocities between 3 and 1 m/s and were classified as C/A units. Normal skin C/A units had an average conduction velocity of 1.67 ± 0.5 m/s (range: 1.1–2.7). Of these four were LTM (30%), and the remaining nine units were nociceptors as follows: six C/ADM (66%), two C/ADMH (22%), and one C/ADC. C/A units from inflamed skin had an average conduction velocity of 1.8 ± 0.4 m/s (range: 1.4–2.7). Of these 1 was classified as LTM. The other 13 were classified as nociceptors as follows: 8 C/ADM (62%), 5 C/ADMH (38%).
Distribution of functional types in normal and inflamed skin
The proportion of nociceptors within each conduction velocity group did not vary significantly between normal and inflamed skin units (z test, P > 0.05). This suggests that stimulus-response relationships remained constant in the presence of inflammation. The distribution of nociceptor subtypes, based on response modality, were also unaffected by inflammation (2 test, P > 0.05).
Spontaneous activity
Significantly more units were spontaneously active in CFA-treated skin than in normal skin (2 test, P < 0.05). In normal skin, 6 of 107 (6%) units were spontaneously active: 4 of 55 C units (7%) and 2 of 20 A units (10%). In inflamed skin, 22 of 78 (28%) units were spontaneously active: 21 of 53 C-fiber units (40%) and 1 of 14 C/A units. The average firing rate was also significantly higher for spontaneously active units from inflamed skin, with an average spontaneous discharge rate of 1.3 ± 0.2 spikes/s, compared with 0.6 ± 0.3 spikes/s for control skin units (t-test, P < 0.05).
Mechanical response properties
All A, A, and C/A fibers examined in this study responded to mechanical stimulation, except for 1 cold nociceptor (C/ADC). A total of 42 C-fiber units from inflamed skin (79%) and 46 C-fiber units from control skin (83%) were mechanically responsive. The median von Frey threshold of 3.4 bars for inflamed skin units was significantly lower than the median threshold of 5.6 bars for control skin units (Mann-Whitney rank sum test, P < 0.05). There was no significant difference between median mechanical threshold for control and inflamed skin in any of the other three conduction velocity categories (Fig. 4).
|
units had an average response rate of 14.1 ± 6.0 spikes/s for control skin (n = 6) and 15.2 ± 4.6 spikes/s for inflamed skin (n = 10). A units had an average response rate of 9.9 ± 4.6 in control skin (n = 4) and 6.2 ± 3.2 in inflamed skin (n = 4). For A units, average response rate was 9.4 ± 5.0 in control skin (n = 5) and 7.1 ± 3.3 in inflamed skin (n = 5).
|
Very high threshold C-mechanical nociceptors
All single units in inflamed skin that responded to mechanical stimulation of the receptive field with a blunt glass rod also responded to stimulation with one or more von Frey filaments. In control skin, 6 of 46 mechanically sensitive C-fiber units (CM or CMH) did not respond to the largest von Frey filament tested (9.3 bars). Greater pressures were not applied to the corium to avoid damaging the receptive field. These were assigned a threshold value of 10 bars. The proportion of these very high-threshold CM nociceptors was significantly higher in control compared with inflamed skin (z test, P < 0.05).
Thermal response properties
A total of 10 A units and 25 A units from control and inflamed skin was tested for responses to thermal stimulation. Of these, one A unit responded to heat with a thermal threshold of 43°C. A total of three units (1 A, 1 C/A, 1 C) responded to cooling of the skin with the Peltier device. Three C-fiber units in control skin responded only to temperatures in the innocuous range and fit the criteria for warming receptors.
A total of 33 units from control skin (31 C-fiber, 2 C/A) and 35 units (30 C-fiber, 5 C/A) from inflamed skin responded in a graded fashion to heating over an innocuous to noxious temperature range and were classified as either CH or CMH nociceptive units. There was no significant difference in heat thresholds between control skin units and inflamed skin units (t-test, P > 0.05). Average heat threshold for C-fiber units from inflamed skin was 37.9 ± 4.0 and 37.7 ± 4.5°C for control skin units. Mean threshold for C/A units was 44.3 ± 4.0°C for control skin units and 41.2 ± 4.0°C for inflamed skin units (Fig. 6).
|
|
A total of 44 units with conduction velocities between 0.2 and 25.5 m/s was tested with morphine sulfate. Of these, 25 units innervated inflamed skin, and 19 fibers innervated normal skin. For each tissue condition, mean population responses before, during and after morphine application were compared independently for heat and mechanical responses.
Units innervating normal skin were tested with 1 µM concentrations of morphine sulfate. Following morphine, no significant differences were observed in mean response to either a standardized heat ramp (Fig. 8A) or mechanical (Fig. 8B) stimulation with 3.3 bars of pressure. This was true for the entire population of units, as well as for separate analyses of nociceptors and C-fiber units. No subpopulation of units was affected by morphine application in healthy skin.
|
, 1 A).
|
Morphine concentration-response relationships
Concentration-response curves were constructed to determine whether morphine-sensitive units responded to the drug in a concentration-dependent manner. Data were used only from units demonstrating a reduction in responsiveness to either mechanical or thermal stimuli after application of morphine sulfate to their receptive fields followed by a return toward baseline of 50% after washout of drug. A total of 13 of 31 (42%) responses recorded from 25 C-fiber nociceptive units from inflamed skin met the requirements and were included in the analysis. All 13 units were classified as nociceptors as follows: 9 CMH, 3 CM, and 1 CH.
The following data points were thus obtained: 500 nM (n = 7), 1 µM (n = 6), and 10 µM (n = 3). Thermal (n = 10) and mechanical (n = 12) responses were considered separately. None of an additional five units tested with either 200 or 20 nM morphine sulfate showed a decrease in responsiveness after application of drug. Three of the 13 units were tested with two concentrations of morphine applied in increasing order of molarity. However, two of the three units tested with 200 nM morphine responded to subsequent application of a higher concentration of the drug (5 and 2 µM, data not included in concentration-response curve). Based on these results, a value of 0% response was assigned to the 200 nM concentration value.
Mechanical and thermal responses were expressed as percent of the initial baseline trial response. Percent response inhibition was calculated, and two separate concentration-response curves were constructed, one showing percent inhibition in mechanical response to a 3.3 bar stimulus (Fig. 10A) and one showing percent inhibition in heat response (Fig. 10B). For both mechanical and heat stimuli, morphine inhibited single-unit responses in a concentration-dependent manner. Regression lines and estimated EC50 values of 426 nM (mechanical) and 450 nM (thermal) were generated using Sigma Stat statistical software (SPSS; Chicago, IL).
|
Criteria for effects of morphine on single unit responses were established based on response variability of control units to repeated stimulation (see METHODS). A unit was judged to be inhibited by morphine if its response to thermal stimulation decreased to 57% of baseline value or its response to mechanical stimulation decreased to 70% of baseline, followed by 50% return after drug washout.
None of 19 units innervating healthy control skin were affected by either 500 nM or 1 µM morphine according to the response criteria. A total of 25 units from inflamed skin was tested. Morphine (500 nM) reduced the responses of 11 of these units to mechanical (Fig. 11) and thermal (Fig. 12) stimulation. All 11 morphine-sensitive units were nociceptors with conduction velocities of 1.3 m/s. A total of 8 of the 11 units were mechanoheat nociceptors (7 CMH, 1 C/AMH). A total of 2 units were CH nociceptors (18%), and 1 was a CM unit. These 11 units accounted for 58% of all C and C/A units from inflamed skin tested with 500 nM morphine sulfate.
|
|
Naloxone pretreatment
Five morphine-sensitive units (2 CMH, 2 CM, 1 CH) were tested with morphine a second time after pretreatment with 500 nM naloxone for 3 min. In four cases, morphine did not have a significant effect after pretreatment with antagonist. One of the CMH units showed a significant decrease in heat response (61% of baseline) but not in mechanical response (150% of baseline) after the second morphine treatment. Mean percent response following the initial morphine application was 28.4 ± 16.8 and 112 ± 44% after naloxone pretreatment plus morphine (Fig. 13). This indicates that the opioid receptor antagonist blocked the effect of the second morphine application.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
Distribution of single units
During the initial electrical search for single-unit activity, emphasis was placed on units conducting in the C and C/A fiber conduction velocity ranges. More systematic surveys of the response properties of single afferent axons innervating the intact rat foot have been described elsewhere (Fleischer et al. 1983; Handwerker et al. 1987; Leem et al. 1993; Lynn and Carpenter 1982). The distribution of functional categories of units classified within the C, A, and A populations reported here are similar, but not entirely consistent, with those previously reported for the rat plantar nerve in vivo preparation. A major difference is the distribution of subclasses within the C and A fiber conduction velocity groups. We encountered a lower percentage (4%) of heat-sensitive A nociceptors than reported (18%) by Leem et al. (1993). The slow rate of skin heating (0.5°C/s) used in the present study may have decreased the probability of detecting A heat-sensitive nociceptors because rates of heating of <2°C/s have been suggested to preferentially activate C-fiber nociceptors in the saphenous nerve of the rat, whereas heating at a higher rate of 6.5°C/s activated both C and A fiber units (Yeomans and Proudfit 1996).
We also encountered a higher proportion of nociceptors among C fibers innervating normal (89%) and inflamed (98%) plantar skin compared with 77% of plantar nerve units reported by Leem et al. (1993). The remaining 23% of the units reported in the Leem study were cold receptors. In the present study, non-nociceptive C-fiber units included warm receptors, low-threshold mechanoreceptors, and a single cooling receptor. Our small yield of cooling units is most likely due to the fact that we did not test for cooling in all experiments.
Sensitization of primary afferent neurons
In the present study, peripheral inflammation enhanced the responsiveness of C-fiber units to noxious mechanical and thermal stimulation. We observed decreases in mechanical thresholds and increased responsiveness to noxious mechanical and thermal stimuli in C-fiber units from CFA-inflamed skin when compared with control skin units. We also report an increased occurrence and elevated rate of discharge of spontaneously active units from inflamed skin. Sensitization of primary afferent neurons is usually characterized by decreased response threshold and increased responsiveness to suprathreshold stimuli that can be accompanied by the developmentof spontaneous activity (Raja et al. 1999; Treede et al. 1992). Sensitized cutaneous afferent neurons have been reported to demonstrate elevations in spontaneous activity level (Kocher et al. 1987; Pogatzki et al. 2002). Sensitization of nociceptors to heat stimuli after tissue injury has been well documented (Campbell and Meyer 1983; LaMotte et al. 1982; Meyer and Campbell 1981), but reports of cutaneous afferent fiber sensitization to mechanical stimuli are less consistent. Some studies report no change in mechanical thresholds (Andrew and Greenspan 1999; Campbell et al. 1979; Reeh 1986); others report a decrease in mechanical thresholds (Hamalainen et al. 2002; Pogatzki et al. 2002) and increased responsiveness to suprathreshold mechanical stimulation (Ahlgren et al. 1992; Andrew and Greenspan 1999).
Our findings differ from a recent study (Du et al. 2003) of CFA-induced sensitization using an in vitro glabrous skin-nerve preparation that reported decreases in heat threshold, no change in the total discharge to a heat stimulus, and no change in mechanical thresholds between inflamed and control fiber populations. The lack of agreement of our findings may be due to differences in sampling approaches, volumes of CFA injected into the hindpaw, and time courses utilized for the development of inflammation. Du et al. (2003) injected 25 µl of CFA and reported increased redness, temperature and thickness of the hindpaw at 48 h, all classical signs of inflammation. In preliminary studies, we observed similar physical signs of inflammation following intraplantar injection of 50 or 100 µl of CFA. However, we only observed reliable mechanical or thermal hyperalgesia with 100 µl CFA injections. Our findings are consistent with studies reporting time- and volume-dependent development of inflammation and hyperalgesia after intraplantar injection of CFA (Fraser et al. 2000; Iadarola et al. 1988).
Our use of an electrical search stimulus may have increased the representation of units with very high mechanical thresholds in the inflamed skin population. After tissue injury or inflammation, subsets of primary afferent neurons become sensitized; this increases their responsiveness to noxious stimuli and contributes to the development of behavioral hyperalgesia (Andrew and Greenspan 1999; Campbell and Meyer 1983; Handwerker et al. 1991; Kessler et al. 1992; LaMotte et al. 1982; Meyer and Campbell 1981; Pogatzki et al. 2002). Units with very high thresholds are difficult to identify and may not be included as frequently in sample populations. Enhanced responsiveness during inflammation would therefore result in greater representation in an otherwise randomly selected experimental sample. This can only be partially corrected by the use of an electrical rather than mechanical search protocol. For each isolated unit, the electrically identified receptive field was characterized using natural stimuli. Units unresponsive to either mechanical or heat stimuli within the tested parameters were not studied further. It is therefore likely that units with very high response thresholds would be under-represented in the experimental sample.
Peripheral opioid receptors
Peripherally delivered opiates attenuate hyperalgesia in experimental models of inflammation (Ferreira and Nakamura 1979; Joris et al. 1990; Stein et al. 1989). These effects are dose-dependent, stereospecific, and antagonist-reversible for ligands acting at delta, mu, and kappa opioid receptors (Stein et al. 1989). In the present study, we used morphine to determine whether functional opioid receptors are present on peripheral processes of cutaneous sensory neurons. Morphine is a nonselective opioid receptor agonist with an affinity for mu receptors in the low nanomolar range and in the high nanomolar to micromolar range for delta and kappa opioid receptors (Raynor et al. 1994; Williams et al. 2001). High concentrations (1 µM) of morphine exert inhibitory actions on the excitability of rat dorsal root ganglia cells (Abdulla and Smith 1998; Khasabova et al. 2004), and the inhibitory effects were blocked by antagonists selective for mu and delta opioid receptors (Khasabova et al. 2004). In the present study, inhibitory actions of morphine were concentration-dependent and preventable with naloxone, a nonselective opioid receptor antagonist, providing pharmacological evidence that the effects were receptor-mediated. We tested a range of concentrations of morphine (from 20 nM to 10 µM) and determined the half-maximal concentration for morphine to inhibit the mechanical and thermal responsiveness of primary afferent neurons to be 426 and 450 nM, respectively. At concentrations used in the present study, morphine could be acting at multiple opioid receptor types in the skin. In normal tissue, each opioid receptor type has been localized to cells in rat dorsal root ganglia (Arvidsson et al. 1995a,b; Ji et al. 1995; Wang and Wessendorf 2001), and delta and mu receptors have been detected on small-diameter axons in normal rat skin (Coggeshall et al. 1997; Hassan et al. 1993; Pare et al. 2001; Truong et al. 2003; Wenk and Honda 1999). Future studies with ligands selective for mu, delta, and kappa receptors will be necessary to determine the relative contribution of each receptor type to the actions of morphine described here.
Direct electrophysiological evidence for opiate modulation of afferent neuron activity in peripheral tissues has been limited. Although several lines of behavioral and anatomical evidence support a role of opioid receptors in peripheral sensory processing, the local effect of opiates on somatic nerve axons is unclear. Early in vivo experiments reported that direct application of morphine to the sural, vagus, and phrenic nerves decreased various components of the compound action potential (Jurna and Grossman 1977). Later studies failed to reproduce this finding and attributed the earlier results to a possible nonspecific anesthetic action of preservatives in the morphine solution (Senami et al. 1986; Yuge et al. 1985). The morphine sulfate solution tested in the present experiments was free of preservatives.
Two studies using in vitro preparations have suggested that opioid receptors may attenuate neuronal activity in peripheral nerves. Mu and kappa opioid receptor agonists inhibited spontaneous activity induced by irradiation of the hindpaw (Andreev et al. 1994), and a kappa receptor agonist inhibited release of prostaglandin E2 and bradykinin after electrical stimulation (Averbeck et al. 2001). In contrast to the somatic nerve in vivo experiments, which were conducted in healthy, uninjured peripheral nerves, these findings suggest a role for peripheral opioid receptors after tissue injury or inflammation.
The focus of in vivo studies has been primarily on sensory units innervating deep tissues or viscera. Locally delivered opiates have been shown to depress spontaneous activity in afferent fibers innervating the inflamed knee joint (Russell et al. 1987), and the activity of afferent fibers in response to noxious stimulation of the bladder and colon was suppressed after delivery of kappa opioid receptor agonists into the local blood supply (Burton and Gebhart 1998; Su et al. 1997a,b). These studies highlight important differences between visceral and somatic afferent neurons in their sensitivity to opiates. Whereas delta, mu, and kappa opioid receptors appear to be active under inflammatory conditions in somatic structures (Antonijevic et al. 1995; Averbeck et al. 2001; Bartho et al. 1990; Stein et al. 1989, 1990a), kappa receptor functionality seems to predominate in the viscera (Sengupta et al. 1996; Su et al. 1997a, 1998). Moreover, opiates are effective in healthy noninflamed visceral tissue, and opiate sensitivity is unchanged after the induction of inflammation (Burton and Gebhart 1998; Su et al. 1997a,b, 2000).
Increased efficacy of peripheral morphine under inflammatory conditions
Electrophysiological characterization of opioid receptors on functionally identified somatic primary afferent neurons has not been described before now. Fifty-eight percent of identified C and C/A units from inflamed skin were sensitive to morphine, and these slowly conducting units were all nociceptors. Furthermore, no morphine effect was observed on units innervating healthy skin. Several mechanisms have been proposed to explain the enhanced peripheral efficacy of morphine observed under inflammatory conditions.
Hindpaw inflammation may result in increased axonal transport of opioid receptors to the periphery by elevation of transport rates or receptor concentration in the axoplasm. Consistent with this idea are elevations of radiolabeled -endorphin binding in sciatic nerve within 24 h of CFA injection (Hassan et al. 1993) as well as immunohistochemical labeling for opioid receptors in small-diameter cutaneous nerves within 4 days (Stein et al. 1990b). In addition, hindpaw inflammation induced an upregulation of mu opioid receptors in DRG 24 h after injection of CFA (Shaqura et al. 2004). We are unaware of any evidence that peripheral inflammatory events can influence rates of peripheral axonal transport. Considering the time that would be required for changes in protein synthesis in DRG and subsequent axonal transport to be reflected as increased numbers of opioid receptors in peripheral axonal terminals, it is reasonable to assume that increased availability of functional opioid receptors at early times (i.e., 18 h) after inflammatory insult results from local peripheral mechanisms.
We have demonstrated in the present study that peripheral processes of nociceptors are sensitive to morphine 18 h after CFA injection. Similarly, analgesic effects of peripherally applied opiates are evident within hours of inflammatory onset (Antonijevic et al. 1995; Czlonkowski et al. 1993; Wenk et al. 2003). Because opioid receptors are present in peripheral tissues under normal conditions (Antonijevic et al. 1995; Coggeshall et al. 1997; Pare et al. 2001; Wenk and Honda 1999), it is possible that early inflammatory events induce axonal changes in the functional state or availability of existing opioid receptors.
Under inflammatory conditions, opiate ligands may have improved access to receptors, resulting in enhanced function of opioid receptors. Peripheral opioid analgesia has been shown to be potentiated in normal tissue when the perineurium was artificially disrupted by the injection of hyperosmotic solutions (Antonijevic et al. 1995), suggesting that exogenous opiate ligands or locally released opioid peptides may have improved access to receptors during tissue inflammation due to breakdown of a perineurial barrier.
Opioid receptors undergo complex regulatory changes during exposure to ligand, and changes in peripheral levels of endogenous opioids occur during the development of inflammation. The dynamic relationship between receptors and their membrane insertion and internalization is a highly regulated process. Trafficking of opioid receptors is sensitive to levels of opiate ligands (Cahill et al. 2001; Gaudriault et al. 1997; Patel et al. 2002), and enhanced translocation of opioid receptors to neuronal plasma membrane occurs during inflammation (Cahill et al. 2003). Endogenous opioid peptides released by immune cells infiltrating the region of peripheral inflammation (Brack et al. 2004a,b; Machelska et al. 2003, 2004) may modulate the translocation of opioid receptors from intra-axonal storage compartments (Zhang et al. 1998) to the plasma membrane within processes and terminals of peripheral axons. Consistent with this idea, opioid receptor signaling in primary afferent neurons is enhanced after peripheral inflammation, and intracellular coupling between mu opioid receptors and G proteins in
