Teleantagonism: A pharmacodynamic property of the primary nociceptive neuron

Abstract

Previous work from our group showed that intrathecal (i.t.) administration of substances such as glutamate, NMDA, or PGE₂ induced sensitization of the primary nociceptive neuron (PNN hypernociception) that was inhibited by a distal intraplantar (i.pl.) injection of either morphine or dipyrone. This pharmacodynamic phenomenon is referred to in the present work as “teleantagonism”. We previously observed that the antinociceptive effect of i.t. morphine could be blocked by injecting inhibitors of the NO signaling pathway in the paw (i.pl.), and this effect was used to explain the mechanism of opioid-induced peripheral analgesia by i.t. administration. The objective of the present investigation was to determine whether this teleantagonism phenomenon was specific to this biochemical pathway (NO) or was a general property of the PNNs. Teleantagonism was investigated by administering test substances to the two ends of the PNN (i.e., to distal and proximal terminals; i.pl. plus i.t. or i.t. plus i.pl. injections). We found teleantagonism when: (i) inhibitors of the NO signaling pathway were injected distally during the antinociception induced by opioid agonists; (ii) a nonselective COX inhibitor was tested against PNN sensitization by IL-1β; (iii) selective opioid-receptor antagonists tested against antinociception induced by corresponding selective agonists. Although the dorsal root ganglion seems to be an important site for drug interactions, the teleantagonism phenomenon suggests that, in PNNs, a local sensitization spreads to the entire cell and constitutes an intriguing and not yet completely understood pharmacodynamic property of this group of neurons.

In rats, intrathecal (i.t.) administration of glutamate, NMDA, or PGE₂ induces sensitization of mechanically stimulated nociceptive C-fibers (i.e., hypernociception) in the hind paw, characterized by the expression of Na_v1.8 sodium channels (1–3). Notably, hypernociception induced by i.t. administration of PGE₂ is NMDA-dependent and is similar to that induced by intraplantar (i.pl.) PGE₂ injection into the hind paw, in both time course and magnitude (2). In addition, it has been suggested that glutamate released at the spinal level can act retrogradely on presynaptic NMDA receptors expressed on primary nociceptive neuron (PNN) terminals in the dorsal horn, leading to maintenance of the hind paw mechanical hypernociception (1). Thus, it is conceivable that PNN hypernociception is a phenomenon involving the whole neuron, an idea supported by the fact that i.pl. injection of morphine or dipyrone inhibits mechanical hypernociception of the hind paw induced by NMDA, glutamate, or PGE₂ injected i.t. (that is, in proximity to terminals at the opposite extremity of the PNNs) (1, 3).

The opioids, a class of strong centrally acting analgesics, induce several supraspinal side effects, including respiratory depression and addiction, which restrict their clinical usefulness. To avoid these problems, i.t. or epidural administration of opioids is widely used as a clinical alternative. The discovery that opioids can also exert a peripheral-local analgesic effect in inflamed tissue (4) and the development of opioids devoid of central side effects (4–6) have enabled new approaches to the clinical treatment of inflammation (7). It is essential to understand that opioids only display peripherally mediated antinociceptive activity when PNNs are already sensitized, thus explaining why opioids have an enhanced efficacy against inflammatory pain.

In fact, sensitization of the PNN (nociceptors) is a common component of inflammatory response. In 1979, we proposed that PNN sensitivity was fine-tuned by a “yin–yang” mechanism, whereby increases in neuronal cAMP and Ca²⁺ promote hypernociception, whereas stimulation of the cGMP signaling pathway blocks ongoing sensitization (8).

Nowadays, the “yin” side of our updated model assumes that PGE₂ sensitizes C-fibers via activation of a protein kinase A- (PKA) and a protein kinase C- (PKC) mediated pathway (9–11) to modulate the voltage-sensitive tetrodotoxin-resistant sodium (3, 10, 12) and/or potassium currents (13–15), thereby facilitating the induction and conduction of PNN action potentials. At the time of the initial proposal, we suggested that morphine counteracted hypernociception directly by inhibition of PGE₂-sensitive adenylyl cyclase (16), allowing us to discover the peripheral effect of opioids (4). The “yang” side of our hypothesis stated that increasing intraneuronal cGMP levels would antagonize ongoing inflammatory hypernociception (8). When the NO signaling pathway was discovered (17–20) and their pharmacological inhibitors were made accessible, it was demonstrated that the analgesic effects of i.pl. injections of acetylcholine, morphine, dipyrone, and other peripherally acting analgesics results from the stimulation of the NO signaling pathway (21–23).

In fact, the antagonism of mechanical hypernociception induced by i.t. injections of NMDA or PGE₂ by morphine or dipyrone applied peripherally, i.e., at a site anatomically and neuronally distant from the i.t space, was called “teleantagonism.” The objective of the present investigation was to determine whether this teleantagonism phenomenon is specific for inhibitors of the NO signaling pathway during peripheral analgesia or is a general property of PNNs.

Teleantagonism was investigated by comparing the effects of pairs of test substances when both were injected locally into the same site (i.t. plus i.t. or i.pl. plus i.pl.) with those observed when each substance in the pair was delivered to opposite ends (i.e., distal and proximal terminals) of the PNNs (i.pl. plus i.t. or i.t. plus i.pl.) (see Fig. 1). In the present series of experiments, the intensity of hypernociception was quantified by our modification of the Randall–Selitto behavior test (see Materials and Methods).

Fig. 1.

Diagrammatic illustration of a PNN showing the sites of i.pl. or i.t., between L4 and L5, injections. For teleantagonism, the injections of agonists or enzyme inhibitors were administered to opposite neuronal sites. Note also that drugs delivered to the subarachnoidal space by i.t. injection can diffuse into the CSF, which bathes the spinal cord, the dorsal roots, and part of the dorsal root ganglion (DRG).

We sought, initially, to assess the contribution of the NO signaling pathway to the inhibitory effect of i.t.-administered opioids on hind-paw hypernociception induced by i.pl.-injected PGE₂. Given that NO is a gas and can readily diffuse throughout the neuron, we examined the effect of L-NMMA [a nonselective inhibitor of NO synthase (NOS)] or ODQ [an inhibitor of soluble guanylyl cyclase (sGC)], administered either i.pl. or i.t., against antihypernociception induced by opioid agonists given by the i.t. or i.pl. route, respectively. Indeed, the antihypernociceptive effect of either i.t. or i.pl. opioid agonists was teleantagonized by L-NMMA or ODQ. To further explore this neuronal pharmacodynamic property, we investigated whether opioid receptor (OR) antagonists (naloxone, the μ-OR antagonist cyprodime and the κ-OR antagonist norBNI) displayed teleantagonism against their antinociceptive agonists. In addition, we investigated whether there is teleantagonism by the nonselective inhibitor of cyclooxygenase (COX) indomethacin or by antagonists of prostaglandin EP1/EP2 or dopamine D1/D5 receptors against the hypernociception induced by IL-1β, PGE₂, or dopamine, respectively. Finally we assessed whether ³H-labeled naloxone could diffuse throughout the length of the PNN nerves to reach the spinal cord or the sciatic nerve and blood, respectively, during the 3-h period of our behavioral experiments.

Results

We first investigated whether analgesia induced by morphine (a preferential μ-OR agonist) or the κ-OR agonist ICI 204,448, given i.t., also involved activation of the NO signaling pathway. Given alone by i.pl. or i.t. routes (see scheme in Fig. 1), neither the NOS inhibitor L-NMMA nor the sGC inhibitor ODQ altered the baseline threshold of the hind paw to mechanical stimulation or the intensity of PGE₂-induced hypernociception (data not shown), indicating that the NO signaling pathway is not tonically active in these conditions. In contrast, prior i.pl. or i.t. treatment with L-NMMA or ODQ significantly antagonized the antinociceptive effects of opioids injected into the same site (Fig. 2 A and B). These results not only confirm that i.pl. opioids stimulate the NO signaling pathway to induce their peripheral antinociceptive effects (21) but reveal that this intracellular signaling system also contributes to antinociception induced by i.t. opioids. More importantly, blockade of opioid-induced antinociception was also observed when the opioids and the inhibitors of the NO signaling pathway were injected, respectively, into distinct sites located at opposite ends of the PNNs (i.e., i.pl. and i.t. or vice versa; Fig. 2 C and D). Thus, teleantagonism of opioid-induced antinociception by L-NMMA and ODQ was effective regardless of whether these treatments were administered by i.pl. or i.t. routes.

Fig. 2.

Ipsilocal or teleadministration of NO signaling pathway inhibitors prevents opioid antinociception. The ordinate shows the intensity of hypernociception measured 3 h after i.pl. injection of PGE₂ into the right (A, B, and D) or both (C) hind paws (100 ng, see Materials and Methods). Open bars show the control PGE₂ hypernociception. Black bars represent the antinociceptive effect of morphine (MPH, 6 μg) or the κ-OR agonist ICI 204,448 (40 μg), injected 1 h before the behavioral endpoint by either the i.pl. (B and D) or i.t. (A and C) route. Gray bars show the effects of NO signaling pathway inhibitors L-NMMA (50 μg) or ODQ (8 μg), injected at ipsi- (A and B) or tele- (C and D) locations 30 min before the opioid injection. In C, the i.t. drug injections exerted a bilateral paw effect, so that maintenance of opioid-induced antinociception in the contralateral paw (CONTRA, hatched gray bars) served as an additional control for local blockade of the response by i.pl. administration of NO signaling pathway inhibitors in the ipsilateral paw. Each value represents the mean ± SEM. of five animals. Asterisks indicate a significant difference compared with the PGE₂ control group, and “&” indicates a significant difference compared with the opioid plus saline (SAL) group (P < 0.05). In C, “#” indicates a significant difference induced by the pretreatment with L-NMMA or ODQ in relation to the contralateral paw, whereas “ns” indicates that the value is not significantly different from that assessed in rats given saline in the ipsilateral paw and opioids intrathecally (black bars).

We next investigated whether this telephenomenon could also be extended to the pharmacological interaction between agonists and competitive antagonists of membrane-bound opioid receptors. As expected, i.pl. injection of the nonselective opioid antagonist naloxone, the μ-OR antagonist cyprodime, or the κ-OR antagonist norBNI each attenuated antinociception induced by ipsilateral i.pl. morphine, whereas only norBNI blocked the corresponding effect of the κ-OR agonist ICI 204,448 (Fig. 3A). Delivery of these opioid receptor agonists and antagonists via i.t. administration yielded similar results (Fig. 3B). When the antagonists and agonists were given at distant sites, similar patterns of inhibition were observed (Fig. 3 C and D), although a relatively less-intense teleantagonism of antinociception induced by i.pl. norBNI was observed after i.t. administration of the κ-OR agonist ICI 204,44 (Fig. 3C). Thus, the selective action of receptor antagonists on the effect of their respective agonist clearly indicates that teleantagonism can be expressed at the level of specific pharmacological neuronal receptors.

Fig. 3.

Ipsilocal or teleadministration of opioid receptor (OR) antagonists prevents opioid-induced antinociception. The ordinate shows the intensity of hypernociception measured 3 h after i.pl. injection of PGE₂ into both (A) or only the right (B, C, and D) hind paws (100 ng, see Materials and Methods). Open bars show the control PGE₂ hypernociception. Black bars represent the antinociceptive effect of morphine (MPH, 6 μg) or the κ-OR agonist ICI 204,448 (40 μg), injected 1 h before the behavioral endpoint by either the i.pl. (A and D) or i.t. (B and C) route. Gray bars show the effects of OR antagonists, cyprodime (CY, 80 μg), norBNI (NOR, 2 μg), or naloxone (NX 1 μg), injected at ipsi- (A and B) or tele- (C and D) locations 30 min before the opioid injection. In A, MPH was injected i.pl. into both hind paws, but OR antagonists were injected only into the right hind paw, so that maintenance of opioid-induced antinociception in the contralateral paw (CONTRA, hatched gray bars) served as an additional control for lack of systemic effects of i.pl. OR antagonist administration. Each value represents the mean ± SEM. of five animals. Asterisks indicate a significant difference compared with the PGE₂ control group, and “&” indicates a significant difference compared with the opioid plus saline (SAL) group (P < 0.05). In A, “#” indicates a significant difference in relation to value measured in the contralateral hind paw, whereas “ns” indicates that the value is not significantly different from that assessed in the paw treated with opioid plus saline (black bars).

A third series of experiments investigated whether teleantagonism could be extended to other enzyme inhibitors unrelated to NO signaling pathway activation. In inflammation, prostaglandins generated in response to IL-1β are considered one of the major sensitizers of PNNs. Preliminary experiments showed that hind-paw hypernociception induced by i.pl. or i.t. IL-1β administration was abolished by prior injection, into the same site, of the nonselective COX inhibitor indomethacin (data not shown). Because IL-1β injected into the rat hind paw is known to cause mechanical hypernociception via prostaglandin release (22), this result strongly suggests that its acute effect after i.t. injection also depends on eicosanoid release. We further found that although i.t. injection of IL-1β caused bilateral hypernociception, i.pl. injection of indomethacin selectively blocked the development of the hypernociception in the injected paw only, without affecting that of the contralateral hind paw (Fig. 4A), thus indicating that teleantagonism occurred in the absence of a systemic effect. Furthermore, switching the sites of injection of the COX inhibitor and the cytokine, i.t. indomethacin inhibited the unilateral hypernociception induced by i.pl. IL-1β (Fig. 4B). Moreover, because indomethacin did not affect the hypernociception induced by i.pl. or i.t. PGE₂ injection, most likely its teleantagonistic effect against IL-1β-induced hypernociception was due to COX inhibition (Fig. 4 A and B).

Fig. 4.

Pretreatment with indomethacin given i.pl. or i.t., at the opposite end of PNNs in relation to the site of IL-1β administration, prevented the development of hypernociception. (A) The ordinate indicates the intensity of hypernociception induced in both paws (open bars: right paw, RP and left paw, LP, respectively) by i.t. injection of IL-1β (0.16 pg). Indomethacin (INDO 100 μg; i.pl.) was injected into one paw 30 min before i.t. injection of IL-1β or PGE₂. Hatched bars show the contralateral paws that received vehicle. (B) Effect of i.t. indomethacin (INDO 100 μg) pretreatment 30 min before the i.pl. injection of IL-1β or PGE₂. Open bars show the hypernociception induced by intraplantar administration of IL-1β (0.5 pg) or PGE₂ (100 ng) in independent groups of rats. In both cases, the behavioral test was conducted 3 h after i.t. or i.pl. injection of IL-1β or PGE₂. Each value represents the mean ± SEM. of five animals. Asterisks indicate a statistically significant difference compared with the corresponding vehicle-treated IL-1β- or PGE₂-challenged groups, and “ns” indicates a nonsignificant difference with respect to the vehicle-treated contralateral paw (P < 0.05). The effect of indomethacin (A, i.t. route or B, i.pl. route) was tested on PGE₂-induced hypernociception (100 ng) to discard any antinociceptive effect other than that resulting from COX inhibition.

Another series of experiments assessed whether the hypernociception induced by PGE₂ and dopamine, two major hypernociceptive mediators released during carrageenan-induced hind-paw inflammation in the rat (22, 23), could be teleantagonized by antagonists of the receptors they act on to enhance mechanosensitivity. Fig. 5A shows that i.pl. or i.t. administration of the prostaglandin EP1/EP2 receptor antagonist AH6809 prevented hypernociception induced by PGE₂ into the same site. However, when PGE₂ and AH6809 were injected into distinct sites, teleantagonism was observed only when the antagonist was administered via the i.pl. route (Fig. 5A). On the other hand, although the dopamine D1/D5 receptor-antagonist SCH23390 antagonized hypernociception induced by dopamine injection into the same i.pl. or i.t. site, no evidence for teleantagonism was found when these drugs were injected to opposite sites of PNN (Fig. 5B).

Fig. 5.

Teleantagonism by selective receptor antagonists was only partial for PGE₂ and was absent for dopamine-induced hypernociception. (A) The control hypernociception induced by i.pl. or i.t. administration of PGE₂ (100 ng) is depicted by open bars. The prostaglandin EP1/EP2 receptor antagonist AH6809 (7.5 ng) was administered at the same site as PGE₂ injection (i.pl. + i.pl. or i.t. + i.t.) or at opposite ends (distant sites) of the PNN (i.pl. + i.t. or i.t. + i.pl.), 30 min before administration of PGE₂. (B) Control hypernociception induced by i.pl. or i.t. administration of dopamine (10 μg) is depicted by open bars. The dopamine D1/D5 receptor antagonist SCH23390 (SCH, 0.375 ng), was administered at the same site as dopamine injection (i.pl. + i.pl. or i.t. + i.t.) or at opposite ends (distant sites) of the PNN (i.pl. + i.t. or i.t. + i.pl.), 30 min before the administration of dopamine. All experimental groups were challenged with PGE₂ (A) or dopamine (B) 3 h before the behavioral test. Each value represents the mean ± SEM. of five animals. Asterisks indicate that the treatment group showed a significant difference (P < 0.05) with respect to the corresponding saline + PGE₂ group (A) or saline + dopamine group (B).

We next tested the possibility that teleantagonism is brought about by diffusion overlap of the interacting drugs along the PNNs. These experiments followed exactly the same protocol as those previously performed to assess teleantagonism between morphine and naloxone. The ³H-labeled naloxone was injected i.pl. or i.t., and its distribution in different tissues was assessed 90 min later. At this time point after i.pl. injection, ³H-labeled naloxone was detected in the sciatic nerve, L5 DRG, and the lumbar enlargement of the spinal cord but not in the fore paw or blood (Fig. 6A). After i.t. injection, ³H-labeled naloxone was found in both the right and left hind paws and their respective sciatic nerves but not in the blood (Fig. 6B). These findings indicate that ³H-labeled naloxone, injected either via i.pl. or i.t. routes, distributes throughout the entire length of the PNN within 90 min, the same time frame in which teleantagonism between naloxone and morphine was observed in the behavioral tests.

Fig. 6.

Detection of ³H-labeled naloxone along the PNNs after i.pl or i.t. injection. The ³H-labeled naloxone was administered into the hind paw (1 μg, i.pl.; A) or the intrathecal space (0.2 μg, i.t.; B) according to the same protocol (PGE₂ + ³H-labeled naloxone + morphine, teleadministration) and behavioral tests as illustrated in Fig. 1. The ³H-labeled naloxone was administered 1.5 h after PGE₂ injection and 30 min before morphine injection. Samples were harvested 1 h after the morphine injection, corresponding to the maximum naloxone effect in the behavioral test. The radioactivity is expressed in cpm/μg of protein. Radioactivity values in the injected paw (A) or DRG and spinal cord (B) were not plotted because they were extremely high. When ³H-labeled naloxone was injected into the right hind paw, ≈0.4% of the total dose (100%) was found in the spinal cord. On the other hand, 22% of the total dose of the i.t.-injected ³H-labeled naloxone was found in the right hind paw. Each value represents the mean ± SEM. of five samples. Asterisks indicate a statistically significant difference compared with the blank sample (P < 0.05).

Discussion

In this study, evidence was presented that the PNN has an intriguing pharmacodynamic property, here called teleantagonism. This term was coined to describe an antagonistic interaction between the effects of two substances on PNNs when they are each administered to cellular domains that are distant from one another. In other words, teleantagonism applies to contexts in which a change in PNN sensitivity to sensory stimulation, induced by injection of substance to one end of the fiber is blocked from a distance by administration of a competitive or noncompetitive antagonist to the opposite end. The occurrence of this phenomenon was clearly evidenced in the blockade of: (i) morphine-induced antinociception by distant injection of naloxone (or cyprodime and norBNI, competitive μ-OR and κ-OR antagonists, respectively) or of inhibitors (L-NMMA or ODQ) of the NO signaling pathway, which is activated by opioids in PNNs or (ii) IL-1β-induced prostaglandin-dependent hypernociception by distant administration of the COX-inhibitor indomethacin. On the other hand, teleantagonism of the PGE₂-induced hypernociception by the prostaglandin EP1/EP2 antagonist AH6909 was only partial and was entirely absent when the dopamine D1/D5 antagonist SCH23390 was tested against dopamine-induced hypernociception. Because PGE₂ and dopamine are the direct inducers of PNN hypernociception, the resistance of their effects to teleantagonism may constitute a special characteristic of this group of mediators.

It is important to note that the examples of teleantagonism presented in this study were obtained under carefully controlled conditions. Thus, we selected the respective doses of each drug in a given combination to ensure that the interaction between them, when both drugs were injected into a single i.pl. or i.t. site, would result in blockade of the antinociceptive or hypernociceptive effect accordingly. Also, the contribution of systemic drug actions, after their local i.pl. injection at the doses used, was excluded by the absence of effects in the contralateral hind paws. Finally, the fact that antinociception induced by i.pl. or i.t. ICI 204,448, a κ-OR agonist that does not cross the blood–brain barrier (24), was amenable to teleantagonism by norBNI further strengthens the argument in favor of the phenomenon.

The possible correlation between drug lipophilicity and the teleantagonism effect was also verified. There is no correlation between the teleantagonism effect and the values of logD [Advanced Chemistry Development (ACD/Labs) Software V8.14 for Solaris (1994–2008 ACD/Labs)] (data not shown).

Teleantagonism of opioid-induced antinociception by inhibitors of NOS and sGC was independent of the site of opioid injection. This pharmacological phenomenon could occur either by the inhibitors crossing the cell membrane at the site of the injection and spreading inside the PNNs themselves or by diffusing inside the nerves (through the interneuronal spaces) and gradually crossing into the PNN fibers along the way. The second possibility resembles the volume transmission described for paracrine neurotransmission in the CNS (25). Both alternatives imply that inhibition of the NO signaling pathway by L-NMMA and ODQ should require their dispersion throughout the whole PNN to block antinociception. Indeed, NOS has been detected in the DRG and the dorsal horn of the spinal cord (26, 27).

As illustrated in Fig. 1, drugs administered via the i.t. route can directly contact large extensions of the central axonic projections of the L3–L6 DRG PNNs that innervate the hind paws. The teleantagonism of i.t.-injected opioids indicates a critical involvement of the PNN in their analgesic effects. This observation challenges the generally accepted assumption that the main site of antinociceptive action of i.t.-injected opioids is the spinal cord (28). In addition, we found that hypernociception induced by IL-1β, which is well established to be mediated by prostanoids (22, 23, 29), is susceptive to teleantagonism by another enzyme inhibitor, indomethacin. The acute hypernociception induced by IL-1β probably involves posttranslational modifications, even though PGE₂ can also activate nuclear receptors coupled to nuclear calcium signals and gene transcription (30, 31). The cells that produce PNN-sensitizing eicosanoids during inflammation or in response to i.pl. injection of IL-1β are currently unknown. However, our finding that hypernociception induced by the latter is susceptible to teleantagonism by i.t. indomethacin strongly suggests that the sensitizing eicosanoids are produced by PNN itself. In fact, COX is present in the neuronal cells and is localized in the peripheral nerve fibers (32). In this context, prostaglandins synthesized in the neuronal cytosol must cross the cell membrane to exert their autocrine pharmacological effect (33). To inhibit neuronal PGE₂ synthesis, indomethacin should diffuse throughout the entire PNN, including the DRG, as already discussed above in relation to teleantagonism of opioid-induced antinociception by inhibitors of the NO signaling pathway.

Independent of the mechanism involved, teleantagonism implies an unexpected rapid diffusion of substances throughout the PNN, which is supported by the short distribution time of ³H-labeled naloxone after its i.t. or i.pl injection. It did not escape our attention that the DRG is strategically located to be a site for the action of drugs injected either into the cerebrospinal fluid (CSF) or into the hind paw. This structure may be easily accessible to i.t.-injected agents, because, in the rat, the CSF directly bathes the entire dorsal roots of the spinal nerves, as well as the proximal (central) half of the DRG (see Fig. 1). This assumption is supported by much indirect experimental evidence, such as the demonstration that an i.t. injection of a δ-OR ligand induces δ-OR internalization in the cell bodies of small and medium-sized capsaicin-sensitive PNNs in lumbar DRG (34). Indeed, these cells respond to OR agonists and antagonists in cultures of DRG neurons (35), and Ferrari et al. (36) demonstrated that PGE₂-induced hypernociception in rats is inhibited by intraganglionar injection of morphine into the L5 DRG. The idea that the DRGs are the site of interaction of both i.t. and peripherally administered opioids is supported by the fact that a direct intraganglionar injection of naloxone inhibited the antinociceptive effect of i.pl morphine injection by ≈70% (data not shown). Our results point to the DRGs as a potentially important site for teleantagonism of the effects induced by i.pl. or i.t. administration of opioids and other agents.

In summary, the current study made use of a model of mechanical hypernociception induced by inflammatory mediators (IL-1β, PGE₂, or dopamine) to examine a pharmacodynamic phenomenon referred to as teleantagonism. Partial or no teleantagonism was observed with receptor antagonists of hypernociceptive mediators, whereas robust teleantagonism of the antinociceptive effects of opioids was found with receptor antagonists or with enzyme inhibitors of the NO signaling pathway administered at either central or peripheral sites of the PNN. The teleantagonism seen with these antagonists and inhibitors provides compelling evidence for the participation of the PNN in antinociception induced by i.t. opioids during acute hypernociception associated with injury or inflammation. On the other hand, the teleantagonism of IL-1β-induced hypernociception by the COX inhibitor indomethacin provides strong evidence that this cytokine stimulates PNNs to generate prostaglandins, which then sensitize these neurons by acting in an autocrine-like fashion on specific receptors located on the cell membrane. We do not yet know whether the teleantagonism is a pharmacological property of all peripheral somatic and visceral nociceptive neurons, but this unexpected pharmacological phenomenon may stimulate further research directed to understanding its underlying mechanisms and its physiopathological relevance.

Materials and Methods

Animals.

Male Wistar rats (180–200 g) were housed in temperature-controlled rooms (22–25°C) with an alternating 12-h light/dark cycle. Water and food were available ad libitum. All experiments were conducted in accordance with National Institutes of Health Guidelines for the Welfare of Experimental Animals (37) and with the methodology approved by the Ethics Committee of the School of Medicine of Ribeirão Preto (University of São Paulo). Each animal was used only in a single experimental group.

Drugs.

The agents used in this study were obtained as follows: PGE₂, dopamine, SCH23390 (R(+)-7-chloro-8-hydroxy-3-methyl-1-phenyl-2,3,4,5-tetrahydro-1H-3-benzazepine hydrochloride), AH6809 (6-isopropoxy-9-oxoxanthene-2-carboxylic acid), naloxone hydrochloride dehydrate, norBNI (nor-binaltorphimine dihydrochloride), ICI 204,448 ((±)-1-[2,3-(dihydro-7-methyl-1H-inden-4-yl)oxy]-3-[(1-methylethyl)amino]-2-butanol hydrochloride R,S-[3-[1-[[(3,4-dichlorophenyl)acetyl]methylamino]-2-(1-pyrrolidinyl)ethyl] phenoxy]-acetic acid hydrochloride), and cyprodime were from Sigma; ³H-labeled naloxone ([N-allyl-2,3-³H]naloxone) was from Amersham; morphine sulfate was from Cristália; indomethacin was from Prodrome; ODQ was from Tocris Cookson; L-NMMA was from RBI; and IL-1β was from the National Institute for Biological Standards and Control, U.K. The specific activity of the IL-1β was 100,000 IU μg⁻¹ ampoule⁻¹. AH6809, ICI 204,448, cyprodime, and ODQ were diluted with saline and 2% dimethyl sulfoxide (Sigma). PGE₂ was dissolved in saline and 1% ethanol (Merck). SCH23390, naloxone hydrochloride dehydrate, norBNI, IL-1β, morphine sulfate, and L-NMMA were diluted in saline. Indomethacin was diluted in Tris·HCl buffer (pH 8.0).

Drug Administration.

The drugs tested were administered via either the i.t. or i.pl. route.

The i.t. injections were performed under light halothane anesthesia (1–2%). The dorsal fur of each rat was shaved, the spinal column was arched, and a 26-gauge needle was directly inserted into the subarachnoid space, between the L4 and L5 vertebrae (38). Correct i.t. positioning of the needle tip was confirmed by manifestation of a characteristic tail flick response. A 20-μl volume containing the test agent was slowly injected with a 100-μl Hamilton microsyringe. The entire injection procedure, from the induction of anesthesia until recovery of consciousness, took ≈4 min.

The i.pl. injections (100 or 50 μl) were performed in conscious animals with a 27-gauge hypodermic needle introduced in the s.c. tissue near the third digit, with the needle tip reaching the middle of the plantar area.

Mechanical Hypernociception Test.

The latency of the nociceptive response was measured by the constant-pressure rat paw test, as described (39). In this method, a constant pressure of 20 mmHg (measured by a manometer) was applied (via a syringe piston moved by compressed air) to a 15-mm² area on the dorsal surface of the hind paw and discontinued when the rat presented a typical “freezing reaction” (nociceptive endpoint). The time elapsed until onset of the freezing reaction (i.e., reaction time) was measured before and after experimental treatment. The intensity of mechanical hypernociception was quantified as the decrease in reaction time, calculated by subtracting the latency, in seconds, after treatment from that measured before treatment (Δ reaction time). Typically, the reaction time before treatment was 31.5 ± 1 s (mean ± SEM; n = 5). This method has been used extensively in our previous studies over the years, where the results have been replicated by other laboratories and by us, using the same or other nociceptive behavioral tests (11, 40, 41). To choose the single dose used for the agonists, receptor antagonists, and enzyme inhibitors, these agents were previously tested in pilot dose–response studies performed before the experiments described.

Radioactivity Assay.

To examine the possible diffusion of opioid receptor ligands throughout the PNNs, ³H-labeled naloxone was injected into the hind paw (1 μg, 155 μl, i.pl.) or into the subarachnoid space (0.2 μg, 30 μl, i.t.) by following the same protocol used in the behavioral tests. At 1.5 h after i.pl. or i.t. ³H-labeled naloxone injection (i.e., 1 h after morphine administration) rats were anesthetized with 10% ketamine (0.1 ml/100 g, i.p.) plus 2% xylazine (0.07 ml/100 g, i.p.) and samples of the sciatic nerve, fifth lumbar DRG, lumbar enlargement of the spinal cord, plantar nerves carefully dissected from both fore and hind paws, and venous blood were harvested for radioactivity assay. All tissues were conditioned in 400 μl of PBS, macerated, and centrifuged (2°C, 31,000 × g, Beckman, JA-17, 123,0) for 15 min. Each resulting supernatant (300 μl) was mixed with 3 ml of scintillation liquid (ICN Biomedicals), and radiation was measured in a previously calibrated liquid scintillation analyzer (1500 Tri-Carb; Packard). To quantify the concentration of protein in the samples, 10 μl of each supernatant were mixed with 300 μl of Coomassie blue G-250 (Coomassie Protein Assay reagent; ICI Americas), and absorbance was measured in a spectrophotometer (Spectramax 250). The amount of radioactivity was expressed in cpm/μg of protein.

Data Analysis.

Results are presented as means ± SEM of groups with five animals. The statistical analysis used was one-way ANOVA, followed by the Bonferroni test. Differences were considered statistically significant at P < 0.05.