Budai, Dénes and Howard L. Fields. Endogenous opioid peptides acting at μ-opioid receptors in the dorsal horn contribute to midbrain modulation of spinal nociceptive neurons. J. Neurophysiol. 79: 677–687, 1998. Activation of neurons in the midbrain periaqueductal gray (PAG) inhibits spinal dorsal horn neurons and produces behavioral antinociception in animals and analgesia in humans. Although dorsal horn regions modulated by PAG activation contain all three opioid receptor classes (μ, δ, and κ), as well as enkephalinergic interneurons and terminal fields, descending opioid-mediated inhibition of dorsal horn neurons has not been demonstrated. We examined the contribution of dorsal hornμ-opioid receptors to the PAG-elicited descending modulation of nociceptive transmission. Single-unit extracellular recordings were made from rat sacral dorsal horn neurons activated by noxious heating of the tail. Microinjections of bicuculline (BIC) in the ventrolateral PAG led to a 60–80% decrease in the neuronal responses to heat. At the same time, the responses of the same neurons to iontophoretically applied NMDA or kainic acid were not consistently inhibited. The inhibition of heat-evoked responses by PAG BIC was reversed by iontophoretic application of the selective μ-opioid receptor antagonists, d-Phe-Cys-Tyr-d-Trp-Orn-Thr-Pen-Thr-NH2 (CTOP) and d-Phe-Cys-Tyr-d-Trp-Arg-Thr-Pen-Thr-NH2 (CTAP). A similar effect was produced by naloxone; however, naloxone had an excitatory influence on dorsal horn neurons in the absence of PAG-evoked descending inhibition. This is the first demonstration that endogenous opioids acting via spinal μ-opioid receptors contribute to brain stem control of nociceptive spinal dorsal horn neurons. The inhibition appears to result in part from presynaptic inhibition of afferents to dorsal horn neurons.
Pain transmission is controlled by CNS circuits that operate at several levels of the neuraxis. The most extensively studied of these circuits includes the midbrain periaqueductal gray (PAG) and the adjacent nucleus cuneiformis (Cnf). When electrically stimulated, this region produces behavioral antinociception in animals and analgesia in humans (see Fields and Basbaum 1978; Fields et al. 1991 for reviews). The PAG-Cnf region controls nociceptive spinal cord dorsal horn neurons via a relay in the rostral ventromedial medulla (RVM) (Aimone and Gebhart 1986; Basbaum and Fields 1984; Dostrovsky et al. 1983; Fields and Basbaum 1978; Fields et al. 1991). The axons of RVM neurons terminate in dorsal horn laminae I, II, and IV–VI (Basbaum et al. 1978; Cho and Basbaum 1989a; Fields et al. 1995). Although this PAG-RVM spinal projection is well established, the neural mechanisms by which it modulates dorsal horn nociceptive neurons remain unclear, largely because the laminae that receive direct RVM input contain several different neural elements including the terminals of primary afferent nociceptors, projection neurons, and both excitatory and inhibitory local circuit interneurons (Light 1992; Willis and Coggeshall 1991).
An important unresolved issue is the contribution of endogenous opioids to the PAG-RVM circuit control of dorsal horn nociceptive neurons. The dorsal horn regions that receive RVM input contain μ, δ, and κ opioid receptors, as well as enkephalinergic interneurons and terminal fields (Arvidsson et al. 1995a,b; Besse et al. 1991; Cheng et al. 1996; Fields et al. 1991; Mansour et al. 1995) and RVM axon terminals contact enkephalinergic dorsal horn neurons (Cho and Basbaum 1989b; Glazer and Basbaum 1984). Release of enkephalin at spinal levels has been demonstrated (Bourgoin et al. 1988; Collin et al. 1992; Dado et al. 1993; Tang et al. 1989; Yaksh and Chipkin 1989). Furthermore, opioids directly inhibit primary afferent nociceptors (Moises et al. 1994; Taddese et al. 1995; Werz et al. 1987) and nociceptive dorsal horn neurons (Glaum et al. 1994; Grudt and Williams 1994; Omote et al. 1990; Randic et al. 1995). Finally although some report that intrathecal naloxone does not antagonize PAG-elicited antinociception (Aimone et al. 1987; Fang and Proudfit 1996; Jensen and Yaksh 1984), others found that PAG- or RVM-evoked antinociception is reduced by intrathecal naloxone (Tseng and Tang 1989; Tortorici et al. 1996; Zorman et al. 1982) or naltrexone (Morgan et al. 1991).
Despite this evidence implicating spinal endogenous opioids in the brain stem control of nociceptive dorsal horn neurons this has not been demonstrated electrophysiologically. In fact, systemic naloxone failed to reverse RVM-evoked inhibition of responses of spinal dorsal horn neurons to noxious stimuli (Duggan and Griersmith 1979) or PAG (Carstens et al. 1979). Adding to the complexity of exploring opioid mechanisms, naloxone, the standard nonselective opioid receptor antagonist was reported to either excite (Henry 1979) or inhibit (Jones et al. 1990) spinal dorsal horn neurons, including spinothalamic tract neurons (Willcockson et al. 1986) in the absence of opioid administration and to have effects unrelated to antagonism of opioids (Dingledine et al. 1978). In the current electrophysiological study we have used the μ-selective opioid receptor antagonists,d-Phe-Cys-Tyr-d-Trp-Orn-Thr-Pen-Thr-NH2 (CTOP) andd-Phe-Cys-Tyr-d-Trp-Arg-Thr-Pen-Thr-NH2 (CTAP), todemonstrate that the PAG-RVM circuit inhibits responses of dorsal horn neurons to noxious heat in part by a local spinal action of endogenous opioids acting at μ opioid receptors.
Anesthesia and surgery
Male Sprague-Dawley rats (300–450 g; Bantin and Kingman, Hayward, CA) were initially anesthetized with pentobarbital sodium (50 mg/kg, ip). A catheter was inserted into an external jugular vein for supplementary anesthetic. For single-unit recordings, the sacral spinal cord was exposed by a laminectomy and the rat was placed in a stereotaxic apparatus. Holes were drilled in the skull and the dura removed to allow microinjections into the PAG. The spinal cord was covered with a pool of warmed mineral oil. Body temperature was kept at 37°C by a water-heated blanket beneath the rat and an infrared heat lamp from above. Heart rate was monitored and maintained within normal limits for lightly anesthetized rats. Recordings were commenced no less than 1 hr after surgery. During the experiments, the animals were maintained in a lightly anesthetized state with a continuous intravenous infusion of methohexital sodium (0.3–0.6 ml/h of 20 mg/ml solution). The infusion rate was adjusted so that the tail flick reflex could be evoked by application of noxious heat (43–45°C) to the tail.
Microinjections of bicuculline into the PAG were made through a 31-gauge stainless steel injection cannula that extended 5 mm below a 25-gauge guide cannula. The guide cannula was stereotaxically positioned in contact with the surface of cortex at a point 7.6 mm caudal from bregma and 2.2 mm lateral from the midline at an angle of 15°. The injection cannula, connected to a 1.0 μl Hamilton syringe, was inserted through the guide cannula before each injection. Bicuculline was freshly dissolved in physiological saline (100 ng/μl) and volumes of 0.10–0.50 μl were manually delivered from the Hamilton syringe over a period of 2 min. PAG bicuculline (BIC) was used because it does not activate fibers of passage but produces a robust, transient inhibition of the tail flick reflex via the RVM (Roychowdhury and Fields 1996) and strongly inhibits dorsal horn nociceptive neurons (Sandkuhler et al. 1989, 1991).
Extracellular, single unit recordings were made from neurons of the dorsal horn of the sacral spinal cord. Recording/iontophoresis electrodes were constructed from a seven-barreled array of thin wall borosilicate glass capillary tubings (1.5 mm OD, 1.12 mm ID; Frederick Haer, Bowdoinham, ME). The recording center barrel contained a 7-μm carbon fiber creating a low impedance (∼1 MΩ) electrode. Drugs were iontophoretically delivered from the surrounding six barrels. Action potentials were displayed on an oscilloscope and activity of single units was isolated with a window discriminator (BAK Electronics, Germantown, MD). Collection of experimental data as well as iontophoretic delivery of drugs were automated with a multifunction instrument control and data acquisition board (NB-MIO-16, National Instruments, Austin, TX) interfaced with a Power Macintosh 8100 computer programmed in LabVIEW (National Instruments, Austin, TX). Detailed description of data acquisition hardware and software is given elsewhere (Budai 1994).
Single-unit extracellular recordings were made from selected dorsal horn neurons responding to noxious heat delivered by a projector lamp focused on the blackened ventral surface of the tail. A thermistor probe placed in contact with the heated area was used to provide feedback control of the heat stimulus. Temperature ramps were generated from a holding temperature of 35°C to a peak of 50°C at a rate of 2°C/s. Neurons were characterized as low threshold (LT), nociceptive specific (NS), or wide dynamic range (WDR) by their responses to mechanical stimuli of increasing strength. Both innocuous (brush, pressure) and noxious (pinch, squeeze that was felt as painful by the experimenter) stimuli were applied to the excitatory receptive fields of the tail.
Microiontophoresis and drugs
Microiontophoresis was performed with a five-channel Neurophore BH-2 controller with automatic current balancing (Medical Systems, Greenvale, NY). Drug barrels of the combined recording/iontophoresis electrode contained one of the following freshly made solutions: 100 mM N-methyl-d-aspartate Na (NMDA) in 100 mM NaCl (pH 8.0), 20 mM kainic acid (KA) in 180 mM NaCl (pH 8.0), 4.7 mM CTOP dissolved in distilled water (pH 5.7), 4 mM CTAP dissolved in distilled water (pH 5.7), 20 mM naloxone HCl in 180 mM NaCl (pH 5.0), 5 mM [d-Ala2,methyl-Phe4,Gly-ol5]enkephalin (DAMGO) in 160 mM NaCl (pH 5.5), 100 mM clonidine HCl in distilled water (pH 5.7), and 2% Pontamine sky blue in 100 mM sodium acetate. CTOP and CTAP are highly μ-receptor selective antagonists (Gulya et al. 1988; Kim and Cox 1993; Kramer et al. 1989). NMDA, KA, and Pontamine sky blue were delivered by negative currents, whereas naloxone, DAMGO, CTOP, CTAP, and clonidine were ejected by positive currents. All drugs were obtained from RBI (Natick, MA) except Pontamine sky blue and CTAP, which were purchased from BDH Chemicals (Poole, UK) and Peninsula Laboratories (Belmont, CA), respectively.
Recording sites were marked by ejection of Pontamine sky blue by using 1 μA negative current for 20 min. Although not all marks were found, the depth of recording was noted in all cases so that cell locations could be approximated. Sites of microinjections into the PAG were labeled by microinjections of 0.25 μl Pontamine sky blue dye through the injection cannula. At the end of the experiment, the animal was euthanized with an overdose of methohexital and intracardially perfused with physiological saline followed by 10% formalin. Recording and injection cannulae locations were verified histologically in 50 μm thin sections stained with neutral red. Positions of the Pontamine sky blue marks were established according to the stereotaxic atlas of Paxinos and Watson (1986). Histologically verified brain microinjection sites and spinal cord recording sites as well as a sample recording of neuronal responses are shown in Fig. 1.
Statistical evaluations were made by using the total number of spikes evoked during each epoch of excitation by heat stimuli or iontophoretic application of an excitatory compound. The background neuronal discharge was calculated by averaging a 15 s period of ongoing activity preceding each epoch of excitation and this value was subtracted from all evoked responses. Differences in magnitude between different response epochs of a single cell were confirmed by one-factor analysis of variance (ANOVA; with Student Newman-Keuls test for posthoc analysis) by comparing the total number of spikes per excitation period. To make data from different experiments more comparable, analysis of pooled data were done after normalizing the baseline heat-evoked response to 100%. Means ± SD of a number (n) of observations are given throughout. A P value of <0.05 was considered significant in all cases. All statistical calculations were performed with GB-Stat software for the Macintosh (Dynamic Microsystems, Silver Spring, MD).
Experiments were carried out on a total of 97 noxious heat responsive spinal dorsal horn neurons in 78 lightly anesthetized rats. Neurons were located between 100 and 600 μm from the surface of the dorsal horn, as estimated by microdrive readings. The dorsal horn neurons with histologically verified locations and the locations of the BIC microinjections in the PAG are shown in Fig. 1. In preliminary experiments, 30 ng BIC microinjected into the ventral portion of the PAG proved to be sufficient to produce a significant inhibition of the heat-evoked response in the majority of spinal cord dorsal horn neurons. All cells had excitatory receptive fields located on the tail. On the basis of their responses to mechanical stimuli of increasing intensity, including innocuous brush, pressure, noxious pinch and squeeze, 73 of the 97 cells were characterized as WDR neurons, whereas the remaining 24 cells were identified as NS neurons. With a temperature ramp from 35 to 50°C at a rate of 2°C/s to stimulate their cutaneous receptive fields on the tail, these neurons produced a stable baseline response to heat stimuli repeated at 4-min intervals (e.g., Figs. 2, 3, and 5).
Effects of bicuculline microinjection into the PAG on nociceptive spinal dorsal horn neurons
Microinjection of 30 ng BIC into the PAG significantly inhibited responses to noxious heating of the tail of 49 of the total 55 dorsal horn neurons tested (Fig. 2 A). Significant inhibition was taken as a minimum 25% decrease in the total number of spikes per heat stimulation epoch after BIC microinjection compared with preinjection baseline. The magnitude and time course of PAG-BIC effects were studied in 26 dorsal horn neurons inhibited by PAG-BIC, of which 19 were WDR, and 7 were NS. After BIC (30 ng) microinjection into the PAG, the inhibition peaked at 15–20 min and lasted for 30 to 60 min (Fig. 2 A). Thus at the 4-min intertrial testing intervals used in our experiments, peak inhibition was observed for the fourth and fifth responses after BIC administration. Consequently, to evaluate drug effects we used the fourth and fifth responses for statistical comparisons. Compared with baseline, BIC administration into the PAG reduced the fourth and fifth responses to means of42.7 ± 12.7% and 43.3 ± 14.3% of baseline preBIC control (n = 26), respectively (Fig. 4).
Reversal of descending inhibition by iontophoretic application of naloxone, CTOP, and CTAP in the dorsal horn
To explore the involvement of endogenous opioids acting at spinal μ-opioid receptors in the descending inhibitory influence of the PAG-RVM system, naloxone, CTOP, or CTAP was iontophoresed near dorsal horn nociceptive neurons that were inhibited by PAG-BIC. The PAG-BIC evoked inhibition of responses to noxious heat was apparently reduced by iontophoretic application of the nonselective opioid antagonist, naloxone (Fig. 2 B), tested in six WDR and one NS neurons. Iontophoresis parameters (current and duration of ejection) for naloxone were selected so that the released naloxone was just sufficient to antagonize the inhibitory effects of the selective μ-opioid receptor agonist, DAMGO, tested before the PAG-BIC application (see Fig. 5 B for example). Under these conditions, naloxone iontophoresis led to a reduction of inhibition by PAG-BIC of heat-evoked responses (Fig. 2 B, Table 1). However we also confirmed the previously reported excitatory actions of naloxone on basal and evoked activity of dorsal horn neurons (Henry 1979). Thus in the absence of PAG-BIC induced inhibition, iontophoretically ejected naloxone at currents that adequately antagonized DAMGO inhibition (see next paragraph) significantly increased the heat-evoked responses of 7 dorsal horn neurons (5 WDR and 2 NS) to 146 ± 17% (P < 0.05) of the prenaloxone value (Fig. 3 A). This effect on “baseline” responsiveness of dorsal horn neurons prevented us from using naloxone to conclusively demonstrate a spinal opioid component in the inhibition by PAG-BIC.
The selective μ-opioid receptor antagonists, either CTOP (n = 8) or CTAP (n = 5), were iontophoresed onto 13 dorsal horn neurons (12 WDR, 1 NS) after evoking descending inhibition of responses to noxious heat by PAG-BIC application. Ejection currents for these compounds were sufficient to antagonize inhibition of heat responses by DAMGO and were established before BIC application as shown in Fig. 5, C and D. Iontophoresis of CTOP or CTAP was carried out at the peak of the PAG-BIC induced inhibition. Figure 2, C and D, respectively illustrates the transient reduction by CTOP and CTAP of the PAG-BIC evoked descending inhibition. When the data from all experiments were pooled, the selective μ-opioid receptor antagonists significantly reduced the PAG-BIC induced inhibition during the fourth and fifth epoch of tail heating (Fig. 4, Table 1). In contrast to naloxone, in the absence of PAG-BIC inhibition, CTOP or CTAP did not change the spontaneous or heat-evoked activity of dorsal horn nociceptive neurons (Fig. 3). In fact, when applied for prolonged periods at ejection currents higher than required to reverse PAG-BIC inhibition (e.g., 100 nA for 120 s), CTOP or CTAP occasionally produced a slight (nonsignificant) inhibition of heat responses.
Specificity of the effects of naloxone, CTOP, and CTAP
Effects of the potent and selective μ-opioid receptor agonist, DAMGO, and the antagonism between DAMGO and either naloxone, CTOP, or CTAP were shown by using the responses to noxious heat of 22 nociceptive dorsal horn neurons (16 WDR and 2 NS). Iontophoresis of DAMGO near recorded dorsal horn neurons reduced their heat responses to about 30% of preDAMGO levels (Table 1). The DAMGO inhibition usually lasted for 15–30 min. As Fig. 5 demonstrates, each compound antagonized the inhibitory effects of DAMGO (see Table 1). As mentioned above however, levels of naloxone iontophoresis, but not of CTOP or CTAP, that were sufficient to antagonize DAMGO caused an increase in the basal and heat-evoked activity of dorsal horn neurons in the absence of PAG-BIC inhibition (Fig. 3).
To provide further evidence for the specificity of CTOP for reversal of opioid mediated inhibition at the spinal level, we tested CTOP against the inhibition of 5 dorsal horn neurons (3 WDR and 2 NS) produced via α2-adrenergic receptors. Iontophoretic application of the selective α2-adrenoreceptor agonist, clonidine, caused a strong inhibition of the heat-evoked responses of nociceptive dorsal horn neurons, reducing their responses to 21 ± 20% of control (mean ± SD at maximum inhibition, n = 5, P < 0.01). Responses of the same cells to iontophoretically applied NMDA or kainic acid remained unchanged at this level of ejected clonidine (Fig. 6 A). CTOP, iontophoresed at currents sufficient to block DAMGO and PAG-BIC inhibition, had no effect on clonidine's inhibitory actions on noxious heat-evoked responses (Fig. 6 B).
Effects of PAG-BIC on responses to excitatory amino acids
The mechanism of the inhibition evoked by PAG-BIC was further studied by comparing the effect of PAG-BIC on the heat-evoked responses of nociceptive dorsal horn neurons to its effect on the responses of the same neurons to iontophoretically applied NMDA and kainic acid. Twelve nociceptive dorsal horn neurons (8 WDR and 4 NS cells) were successively excited by iontophoretic application of NMDA and kainic acid and then by noxious tail heat, as shown in Fig. 7. Ejection currents for NMDA and kainic acid were selected to produce a peak neuronal response that was comparable with that of the responses to noxious heat. Microinjection of BIC (30 ng) into the PAG led to a significant decrease in the responses to noxious heat in all 12 neurons and this inhibition was partially reversed by iontophoretic application of CTOP (Fig. 7 A). In contrast,PAG-BIC had no consistent action on the excitatory amino acid (EAA) elicited responses of dorsal horn nociceptive neurons. In fact, PAG-BIC differentially affected responses to these two EAAs (Table 2). In 7 of the 12 neurons, PAG-BIC caused reciprocal changes in the two EAA responses. This was characterized by a decrease in responses to one class of excitatory amino acid receptor agonist (mainly NMDA-evoked responses), which was accompanied by an increase in responses to the other type of excitatory amino acid receptor agonist (mainly kainic acid-evoked responses; Table 2). Iontophoretic application of CTOP did not affect responses to NMDA or kainic acid nor did it alter the change of EAA-evoked responses produced by PAG-BIC (Fig. 7 A).
Similar to the effect of PAG-BIC microinjection, iontophoresis of the selective μ-opioid receptor agonist, DAMGO, inhibited heat-evoked responses more effectively than responses to iontophoretically applied EAA agonists. Figure 7 B represents experiments carried out in five WDR dorsal horn neurons. DAMGO iontophoresis, at low to medium high currents (20–30 nA), led to strong inhibition of heat responses, whereas responses to kainic acid remained relatively unchanged and some inhibition (15–25%) of the NMDA-evoked cell firing was observed in three neurons.
The major finding of this study is that iontophoretic application of the μ-selective opioid receptor antagonists, CTOP and CTAP, significantly reduces the PAG elicited inhibition of the responses to noxious heat of sacral dorsal horn neurons. The opioid receptor specificity of this effect is supported by the finding that, in the same neurons and at the same ejection currents, either CTOP or CTAP reversed the inhibitory effect of both PAG BIC and iontophoretic DAMGO but not that elicited by the α2 agonist clonidine. Iontophoretically applied naloxone appeared to have similar effects. These data strongly support the hypothesis that the PAG elicited inhibition of dorsal horn nociceptive neurons is mediated in part through the local release of an endogenous opioid acting at the μ opioid receptor. These studies are consistent with behavioral studies showing partial reversal by spinal naloxone of the behavioral antinociceptive effect of stimulating the PAG (Morgan et al. 1991) or its medullary relay to the spinal cord, the RVM (Zorman et al. 1982). On the other hand, in previous reports, naloxone has not consistently reduced PAG or RVM induced inhibition of nociceptive spinal dorsal horn neurons. For example, in the cat, intravenous naloxone had no effect on the inhibition of dorsal horn neurons by electrical stimulation of the PAG or RVM (Carstens et al. 1979, 1983; Duggan and Griersmith 1979) but appeared to partially reverse the effect of such stimulation on neurons in the trigeminal nuclear complex (Sessle and Hu 1981). In the rat, only partial and inconsistent reversal by systemic naloxone of the inhibitory action of PAG (Carstens et al. 1990) or RVM (Rivot et al. 1979) was reported. Rather than directly contradicting our results, previous inconsistencies may result from the limitations of using naloxone as an opioid antagonist. Although naloxone is a μ-receptor preferring antagonist it significantly antagonizes opioid agonist action at δ and κ receptors (e.g., Goldstein and Naidu 1989). Thus its use could be difficult to interpret if there were tonic release of endogenous ligands acting at two or more opioid receptors, particularly if there were both excitatory and inhibitory opioid actions. For example, there is evidence that an endogenous κ-opioid receptor ligand facilitates dorsal horn neurons (Ren et al. 1991). This might explain the observations that under certain conditions naloxone can have either inhibitory or excitatory effects on nociceptive dorsal horn neurons, including local circuit neurons in lamina II (Henry 1979; Jones et al. 1990; Mokha 1992; Shen et al. 1996; Willcockson et al. 1986). In some studies, iontophoretic naloxone has failed to reverse inhibitory opioid actions on dorsal horn neurons (Jones et al. 1990; Willcockson et al. 1986). In trying to determine an opioid mediated action of descending modulatory systems, the confounding issue with naloxone, as observed in the current study, is that it produces significant increases in background and nociceptor evoked activity in dorsal horn neurons in the absence of evoked descending inhibition.
The use of the μ selective opioid receptor antagonists CTOP and CTAP obviated the problem of naloxone's enhancement of spontaneous and noxious heat evoked activity in the absence of activation of descending inhibition. In the absence of PAG-BIC inhibition, iontophoretic CTOP or CTAP produced no significant change in dorsal horn neuronal activity. Thus the reduction of the PAG-BIC elicited inhibition by CTOP or CTAP cannot be attributed to an excitatory effect unrelated to descending inhibition.
The enhancement by naloxone but not CTOP or CTAP of baseline and heat-evoked activity in rat dorsal horn neurons is consistent with recent studies demonstrating a background κ and δ but not μ-opioid receptor mediated inhibition of the tonic phase of the nociceptive response to formalin (Ossipov et al. 1996). Naloxone was also reported to enhance the “spontaneous” activity of lamina V neurons in arthritic rats (Lombard and Besson 1989). It is conceivable that there is a tonic nociceptive input generated by the surgical injury required for this experiment and that this tonic input produces a κ and δ opioid receptor mediated inhibition. The observed enhancement of background activity by iontophoretically applied naloxone could result from the reversal of such tonic inhibition. It is also possible that naloxone could be acting as a functional γ-aminobutyric acid (GABA) antagonist (Dingledine et al. 1978) or as an inverse agonist at one or more of the opioid receptors in the spinal cord dorsal horn (Wang and Gintzler 1994; Wang et al. 1994).
The neural target of opioid-mediated inhibition activated by PAG-BIC
The mechanisms by which the PAG-RVM circuit controls nociceptive transmission in the dorsal horn are largely unexplored (Fields et al. 1991, 1995; Jones et al. 1990). In principle, one or more of the following could contribute: 1) direct modulation of the dorsal horn terminals of primary afferent nociceptors, 2) a postsynaptic action on second-order nociceptive projection neurons, or 3) an action on dorsal horn local circuit neurons that either directly control primary afferent nociceptor terminals or excite projection neurons.
In the current experiments, although dorsal horn neuronal responses to noxious heat were consistently inhibited,PAG-BIC did not consistently inhibit the responses of the same neurons to iontophoretically applied EAA agonists. This lack of parallelism between inhibition of noxious stimulus-evoked and EAA-evoked responses suggests that the descending opioidergic control of nociceptive dorsal horn neurons evoked by PAG-BIC is exerted directly on the dorsal horn terminals of primary afferents or on the somadendritic or axon terminal regions of local circuit neurons afferent to the recorded cells. An alternative, but less likely possibility is that the opioid control is exerted postsynaptically near the synapse activated by noxious heat but at a site on the same neurons that is spatially separate from the sites of action of the iontophoretically applied EAAs.
There is anatomic and pharmacological evidence supporting some of the possible mechanisms proposed above. Evidence supporting a postsynaptic action of endogenous opioids on second-order dorsal horn neurons is as follows. First, opioid agonists directly inhibit second-order dorsal horn neurons (Zhang et al. 1996), including putative local circuit neurons in the substantia gelatinosa (Grudt and Williams 1994; Jeftinija 1988; Yoshimura and North 1983) and trigemino-thalamic projection neurons (Wang and Mokha 1996). Both radioligand and immunocytochemical studies have demonstrated that μ-opioid receptors (MOR) are present on the somadendritic region of intrinsic dorsal horn neurons (Arvidsson et al. 1995b; Besse et al. 1992a; Mansour et al. 1994). Enkephalin-immunoreactivity was demonstrated in synapses onto the somadendritic region of dorsal horn spinothalamic tract neurons (Ruda 1982; Ruda et al. 1984). Finally, a recent ultrastructural study of rat spinal cord dorsal horn lamina I and II has demonstrated that, in addition to its location on the plasmalemma of small diameter afferent terminals, MOR immunoreactivity is present on the soma and dendrites of dorsal horn neurons (Cheng et al. 1996). Importantly, that study demonstrated that about one-fifth of the MOR labeled dendrites received synaptic terminals containing immunoreactive leu-enkephalin. These anatomic studies indicate that endogenous opioids, probably enkephalins, act at the MOR on the somadendritic region of intrinsic dorsal horn neurons. These intrinsic neurons likely include local circuit and possibly nociceptive projection cells.
Our data are more consistent with a presynaptic site of action for the MOR mediated inhibition of dorsal horn nociceptive neurons activated by the PAG-RVM circuit. There is an extensive body of evidence supporting a MOR mediated inhibition of transmitter release in superficial dorsal horn. In addition to the neurochemical and anatomic evidence establishing the presence of MOR on primary afferent terminals (Arvidsson et al. 1995b; Besse et al. 1992b; Cheng et al. 1996; Fields et al. 1980; Mansour et al. 1994), other evidence supporting a presynaptic site of MOR action includes 1) μ agonists inhibit the release of neurotransmitters from afferent terminals in the dorsal horn (Glaum et al. 1994; Grudt and Williams 1994; Macdonald et al. 1978) and 2) as in the current report, others (Glaum et al. 1994; Hope et al. 1990) have shown that μ agonists block afferent evoked but not EAA-evoked activity of superficial dorsal horn neurons. Whether these μ receptor mediated effects on transmitter release represent a direct action on primary afferent or local circuit neuron terminals afferent to projection neurons is unclear, however, because μ agonists clearly have direct actions on both primary afferents (Mudge et al. 1979; Taddese et al. 1995; Werz et al. 1987) and putative local circuit neurons in the substantia gelatinosa (Glaum et al. 1994; Grudt and Williams 1994; Yoshimura and North 1983), both neural elements are potential contributors to the endogenous opioid-mediated inhibition observed in the current experiments.
We thank C. Chiu for help in preparing this manuscript and T.-L. Lee and C. Magid for technical assistance.
This work was supported by National Institutes of Health Grants NS-21445 and DA-01949.
Address for reprint requests: D. Budai, Dept. of Neurology, University of California, San Francisco, 513 Parnassus Ave., S-784, San Francisco, CA 94143–0114.
- Copyright © 1998 the American Physiological Society