Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Spinal administration of 5-HT and 5-HT agonists produces antinociception (Alhaider et al., 1993). The use of agonist and antagonist drugs purported to be selective for the various 5-HT receptors has provided evidence for spinal 5-HT_1A, 5-HT_1B, and 5-HT₃ receptors in antinociception in models of acute nociceptive pain (Crisp and Smith, 1989; Alhaider and Wilcox, 1993; Alhaider et al., 1993; Xu et al., 1994). 5-HT_1A and 5-HT_1B receptors appear to have reciprocal roles in the mediation of spinal reflexes (el-Yassir et al., 1988; Eide et al., 1990; Murphy and Zemlan, 1990).

In the spinal cord, 5-HT₃ receptors are located primarily on the axon terminals of the sensory primary neurons that terminate in the substantia gelatinosa of the dorsal horns (Hamon et al., 1989). In contrast, 5-HT_1B receptors are autoreceptors on the neurons of the descending inhibitory serotonergic tract from the nucleus raphe magnus to the spinal dorsal horn (Glaum and Anderson, 1988).

Several groups have examined the role of 5-HT₃ receptors in tail-flick analgesia. Glaum et al. (1990) and Crisp et al. (1991) used highly selective 5-HT₃ receptor antagonists to block the analgesia produced by i.t. injection of 5-HT or 2-methyl-5-HT, an agonist with a 5- to 10-fold selectivity for 5-HT₃ receptors. Both authors concluded that this receptor subtype is of primary importance for spinal analgesia produced by 5-HT agonists. Although the agonists used have only a 4- to 5-fold selectivity for 5-HT₃ receptors over other 5-HT receptor subtypes, the antagonists used were greater than 60-fold selective. Moreover, the doses used were similar to doses used to attribute other drug effects to mediation by 5-HT₃ receptors. The localization of the 5-HT₃ receptors on axon terminals (Hamon et al., 1989) puts this receptor subtype in a prime location for the modulation of sensory neurons. Thus, the evidence for involvement of 5-HT₃ receptors is strong.

Other investigators have examined the role of 5-HT_1B receptors in spinal analgesia (Alhaider and Wilcox, 1993; Alhaider et al., 1993; Hain et al., 1999). These authors also used agonists and antagonists to characterize the role of 5-HT_1B receptors in spinal analgesia. However, highly selective 5-HT_1B agonists and antagonists are not available. Moreover, the doses of the 5-HT_1B agonists used in early studies were much greater than necessary to elicit other 5-HT_1B-mediated effects: >40 nmol (Alhaider and Wilcox, 1993) versus 6 nmol (Gradin and Persson, 1993), and the doses of 5-HT_1B antagonists were also excessive. Moreover, if 5-HT_1B receptors functioned as autoreceptors, stimulation of this subtype would tend to inhibit the descending modulatory system. Therefore, the role for 5-HT_1B receptors in spinal 5-HT analgesia is more controversial than that of 5-HT₃ receptors.

In recent years, approaches that use antisense oligodeoxynucleotides (AODNs) to modify gene expression in vivo have been described (Pasternak and Standifer, 1995). An AODN is a synthetic single-stranded DNA molecule with a sequence complementary to a specific mRNA sequence. It is designed to specifically interact and hybridize with a particular sequence of mRNA so that the mRNA is either degraded by RNase or prevented from being translated into protein, or both. DNA is composed of a sense strand and an antisense strand. Messenger RNA is copied from the antisense strand, and its sequence is identical to the sense strand, except for the thymine bases, which are replaced by a pyrimidine base, uracil. Theoretically then, an antisense strand would be expected to hybridize with the mRNA. The specificity of the interaction between the oligonucleotide and its target mRNA is based on the Watson-Crick model of base pairing, which allows precise hybridization of nucleic acids based on base stacking and hydrogen bonding. The mRNA sequence to which an AODN is complementary must also be free of secondary structure. In vivo administration of AODNs has been used to reduce the number of opioid, -aminobutyric acid_B, neuropeptide Y, N-methyl-D-aspartate, dopamine, and _2A receptors in the central nervous system (Wahlestedt et al., 1993; Zhou et al., 1994; Pasternak and Standifer, 1995; Bilsky et al., 1996). The administration of AODNs in vivo also has been used to identify the functional importance of several receptor subtypes (Wahlestedt et al., 1993; Zhou et al., 1994). This approach is of particular merit when investigating the physiological role of novel receptors or receptor subtypes for which a selective antagonist has not been developed. With regard to 5-HT receptor subtypes, 5-HT₃ antagonists are suitable for pharmacological characterization of drug effects, but 5-HT_1B antagonists are not (Alhaider et al., 1993). Accordingly, we re-examined the issue of which receptor subtypes mediate the analgesia produced by spinal 5-HT, the 5-HT_1B agonist CGS-12066A, and the 5-HT₃ agonist 2-methyl-5-HT, using an in vivo antisense approach.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Drugs and Reagents. CGS-12066A was purchased from Research Biochemicals Inc. (Natick, MA). Tris was purchased from Bio-Rad (Richmond, CA). Radioligands were purchased from NEN Life Science Products (Boston, MA). All other drugs and reagents were purchased from Sigma (St. Louis, MO). All doses of drugs were calculated as the salt form.

Subjects. All procedures were approved by the Louisiana State University Health Sciences Center Institutional Animal Care and Use Committee and adhered to the International Association for the Study of Pain Guidelines on Animal Experimentation. Male CD-1 mice (Charles River, Boston, MA) were housed 10 to a cage and maintained on a 12-h light/dark cycle with ad libitum access to food and water. Intrathecal injections were made using a 10-µl Hamilton syringe fitted to a 30-gauge needle with PE10 tubing (Hylden and Wilcox, 1980). Doses of agonist drugs were selected from preliminary dose-response curves as a dose that produced analgesia in approximately 60 to 70% of mice.

Antisense Design, Synthesis and Treatment. AODN sequences spanned the initiation codon for each receptor. All unmodified (phosphodiester) ODNs were synthesized by the LSU Health Sciences Center Core Labs using an Expedite nucleic acid synthesis system (Applied Biosystems, Foster City, CA) and purified using reverse-phase liquid chromatography. The oligos were diluted in sterile water, and concentrations were confirmed by spectrophotometry.

Tail-Flick Analgesia. Analgesia was assessed quantally using the tail-flick assay (D'Amour and Smith, 1941). The latency to withdraw the tail from a focused light stimulus was measured electronically using a photocell. Baseline latencies (3.0-4.0 sec) were determined for all animals before 5-HT agonist treatments as the mean of two trials. 5-HT agonists were injected i.t. 10 min before testing. Mice having a tail-flick latency that was at least double their predrug baseline were considered to be analgesic. A 12-s cutoff was used to minimize tissue damage.

5-HT_1B Receptor Binding. To prepare tissue for the 5-HT_1B receptor binding assay, mice used in the tail-flick experiments were killed by cervical dislocation; then, spinal cords were rapidly removed and placed in 50 mM treated Tris-HCl buffer (50 mM Tris-HCl, 1 mM EDTA, 100 mM NaCl) at a ratio of 1:50 (w/v). Five spinal cords were pooled and homogenized with a Tekmar Tissumizer (Tekmar, Cincinnati, OH) for 20 s at setting 50. The homogenate was incubated at room temperature with 0.1 mM phenylmethylsulfonyl fluoride (protease inhibitor) for 15 min and then centrifuged at 10,000g for 40 min at 4°C. The pellets were resuspended with 0.32 M sucrose (1:6, w/v) and stored at 75°C until use.

5-HT₃ Binding Assay. For the 5-HT₃ receptor binding assay, spinal cords of mice used in the behavioral experiments, and an additional 10 mice per treatment group, were rapidly removed, weighed, and placed in ice-cold 50 mM Tris/Krebs buffer (50 mM Tris, 118 mM NaCl, 4.75 mM KCl, 1.2 mM KH₂PO₄, 1.2 mM MgSO₄, 2.5 mM CaCl₂, 25 mM NaHCO₃, and 11 mM glucose, pH 7.6) containing 0.2 M EDTA and 4 M NaCl. Nine to 10 spinal cords were pooled and homogenized for each tissue preparation. Phenylmethylsulfonyl fluoride (100 mM) was added, and the homogenate was allowed to sit at room temperature for 15 min. It was then centrifuged twice at 48,000g for 20 min. The pellet was resuspended at a ratio of 1:50 in the Tris/Krebs buffer.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Pretreatment of mice with the AODN for 5-HT_1B receptors produces a 70% attenuation (p < 0.001) of [¹²⁵I]iodocyanopindolol binding for this subtype (Fig. 1). The AODN for 5-HT₃ receptors produced a slight reduction in 5-HT_1B binding (p > 0.05). This result was similar to that produced by the two mismatch treatments. Pretreatment of mice with the AODN for 5-HT₃ receptors reduced [³H]zacopride binding to undetectable levels (p < 0.0001; Fig. 1). Similar to the results with [¹²⁵I]iodocyanopindolol binding, the AODN for 5-HT_1B receptors produced a slight reduction in 5-HT₃ binding (p > 0.05). This result was similar to that produced by the two mismatch treatments. For each of the in vivo experiments, results represent pooling of data from two separate experiments that produced similar results.

View larger version (44K): [in this window] [in a new window]   Fig. 1.   Antisense knockdown of spinal 5-HT1B and 5-HT3 receptors. Groups of mice were treated on days 1, 3, and 5 with saline, AODN for 5-HT1B receptors, mismatch ODN for 5-HT1B receptors, AODN for 5-HT3 receptors, or mismatch ODN for 5-HT3 receptors. On day 6, spinal cords were rapidly removed and processed for 5-HT1B receptor binding with 100 pM 125I-cyanopindolol or 5-HT3 receptor binding with 3 nM [3H]zacopride. Specific binding in control spinal cords was 12,330 dpm/mg tissue for 5-HT1B binding and 260 dpm/mg tissue for 5-HT3 binding. Data represent mean percent control specific binding ± S.E.M. Only the AODN for 5-HT1B receptors significantly reduced 125I-cyanopindolol binding. Likewise, only the AODN for 5-HT3 receptors significantly reduced [3H]zacopride binding.

Spinal administration of 5-HT (70.5 nmol, i.t.) produced analgesia in 64% of mice pretreated with saline (Fig. 2). At higher doses, 5-HT produced spontaneous tail-flicks. Pretreatment with the AODN for 5-HT₃ receptors significantly attenuated analgesia produced by i.t. 5-HT. Pretreatment with the AODN for 5-HT_1B receptors did not affect spinal 5-HT analgesia significantly. Likewise, pretreatment with the mismatch oligos for these two AODNs had no significant effect on analgesia.

View larger version (40K): [in this window] [in a new window]   Fig. 2.   Effect of antisense treatment on intrathecal 5-HT-induced analgesia. Groups of mice (n  18) were treated on days 1, 3, and 5 with saline, AODN for 5-HT1B receptors, mismatch ODN for 5-HT1B receptors, AODN for 5-HT3 receptors, or mismatch ODN for 5-HT3 receptors. On day 6, all mice were injected with 5-HT (70.5 nmol, i.t.). Ten minutes later, all mice were tested for tail-flick analgesia. Only the AODN for 5-HT3 receptors produced a significant attenuation of 5-HT analgesia.

Spinal administration of CGS-12066A (44.4 nmol, i.t.) produced analgesia in 50% of saline-pretreated control mice (Fig. 3). At doses higher than this, spontaneous tail-flicks interfered with the testing procedure. Surprisingly, none of the oligo treatments had a significant effect on CGS-12066A-produced analgesia.

View larger version (38K): [in this window] [in a new window]   Fig. 3.   Effect of antisense treatment on intrathecal CGS-12066A-induced analgesia. Groups of mice (n  18) were treated as described in Fig. 3. On day 6, all mice were injected with CGS-12066A (44.4 nmol, i.t.). Ten minutes later, all mice were tested for tail-flick analgesia. None of the treatments significantly affected CGS-12066A-induced analgesia.

Spinal administration of 2-methyl-5-HT (49 nmol, i.t.) produced analgesia in 67% of mice pretreated with saline (Fig. 4). Doses higher than this also produced spontaneous tail-flicks. Like animals tested with 5-HT, pretreatment with the AODN for 5-HT₃ receptors significantly attenuated analgesia produced by i.t. 2-methyl-5-HT, whereas pretreatment with the AODN for 5-HT_1B receptors had no significant effect. Pretreatment with the mismatch oligos for the two AODNs had no significant effect on analgesia.

View larger version (44K): [in this window] [in a new window]   Fig. 4.   Effect of antisense treatment on intrathecal 2-methyl-5-HT-induced analgesia. Groups of mice (n  18) were treated as described in Fig. 3. On day 6, all mice were injected with 2-methyl-5-HT (49 nmol, i.t.). Ten minutes later, all mice were tested for tail-flick analgesia. Only the AODN for 5-HT3 receptors produced a significant attenuation of 2-methyl-5-HT analgesia.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

A valuable axiom in pharmacology is that confidence in your data should be only as strong as the selectivity of your drugs. In light of pharmacological characterization using putative selective agonists and antagonists, several investigators have implicated both 5-HT_1B and 5-HT₃ receptors in the mediation of spinal 5-HT analgesia (Eide et al., 1990; Glaum et al., 1990; Alhaider and Wilcox, 1993; Alhaider et al., 1993; Hain et al., 1999). Although the selectivity of 5-HT₃ receptor antagonists used in these studies appears adequate, the selectivity of the 5-HT_1B agonists and antagonists was insufficient for definitive attribution of effects to this receptor subtype (Alhaider et al., 1993). In preliminary testing, we found that although i.t. CGS-12066A was analgesic, another 5-HT_1B receptor agonist, CP93129, was not (data not shown). Likewise, Crisp et al. (1991) found no elevation of tail-flick latency by intrathecally administered trifluoromethylphenyl piperazine. Moreover, doses of 5-HT_1B receptor agonists used in both our studies (44.4 nmol) and those of Alhaider et al. (1993) (45 nmol) appeared to be much higher than those used to characterize other spinal 5-HT_1B receptor-mediated effects (6 nmol; Gradin and Persson, 1993).

The AODN approach is of particular merit when investigating the physiological role of novel receptors or receptor subtypes for which a selective antagonist has not been developed. For example, the AODN strategy has been used to block effects of 5-HT_1A and 5-HT₆ receptor stimulation (Bourson et al., 1995; Sleight et al., 1996). Because of the lack of selective compounds, functional studies of 5-HT₆ receptors have used the AODN approach almost exclusively (Bourson et al., 1995; Sleight et al., 1996). Following this strategy, we have demonstrated the usefulness of a 5-HT₃ receptor AODN to block the analgesia produced by intrathecal administration of 5-HT or the 5-HT₃ receptor agonist, 2-methyl-5-HT. In contrast, AODN treatment that reduced 5-HT_1B receptors by 70% did not block analgesia produced by spinal administration of 5-HT, 2-methyl-5-HT, or the 5-HT_1B receptor agonist, CGS-12066A. These results are evidence that analgesia produced by 5-HT and 2-methyl-5-HT is primarily attributable to an action at the 5-HT₃ receptor subtype, but the analgesia produced by CGS-12066A is not attributable to either the 5-HT_1B or the 5-HT₃ receptor subtype. This latter finding is consistent with the results of Pickard et al. (1996), who demonstrated that the effects of CGS-12066A on light-induced phase shifts of the circadian activity rhythm and induction of c-fos expression in the suprachiasmatic nucleus were not antagonized by the 5-HT₁ antagonist, methiothepin. An alternative mechanism of the analgesic effects of CGS-12066A is suggested by the results of Durham and Russo (1999). These authors found that administration of CGS-12066A inhibits the release of calcitonin gene-related peptide from sensory primary afferent neurons. Release of this neuropeptide has been correlated with the mediation of nociception (Van Rossum et al., 1997).

The findings that the AODN for 5-HT_1B receptors did not significantly attenuate 5-HT₃ receptor binding and conversely, the AODN for 5-HT₃ receptors did not significantly attenuate 5-HT_1B receptor binding, attest to the selectivity of these two binding assays. ¹²⁵I-Cyanopindolol could not be binding to a subpopulation of 5-HT₃ receptors, and [³H]zacopride could not be binding to a subpopulation of 5-HT_1B receptors.

Because 30% of 5-HT_1B receptors remained following AODN treatment, it could be argued that if there are a considerable number of spare receptors in a spinal 5-HT_1B analgesia system, a significant reduction in analgesia would not be expected. We feel that this interpretation is unlikely to be true because CGS-12066A is considered to be a low-efficacy agonist (Neale et al., 1987). Consequently, this stimulation of all or nearly all of the spinal 5-HT_1B receptors would be required for maximum analgesia. A 70% reduction in receptors would be expected to produce a significant reduction in this drug's analgesic effect.

Using the most selective agonists and antagonists available, several laboratories have concluded that both 5-HT_1B and 5-HT₃ receptors are involved in serotonin-produced analgesia (Glaum et al., 1990; Crisp et al., 1991; Alhaider and Wilcox, 1993; Alhaider et al., 1993; Hain et al., 1999). Our results using an antisense strategy are inconsistent with this interpretation. We propose that 5-HT₃ receptors are important, but 5-HT_1B receptors are of relatively little importance, at least in this model of analgesia. These results do not preclude involvement of other 5-HT receptor subtypes, although only these two subtypes have been implicated in previous studies. In contrast, Hain et al. (1999), using a pharmacological approach, have proposed that the involvement of 5-HT_1B receptors is strain-dependent. DBA/2 mice, which have relatively high levels of this receptor, were sensitive to i.p. administration of CGS12066. In contrast, C57BL/6 mice, which are deficient in functional 5-HT_1B receptors, were less sensitive. Unfortunately, these authors did not demonstrate antagonism of the CGS12066 analgesia with either an antagonist or antisense. In addition, Crisp et al. (1991) reported that the 5-HT_1B receptor agonist trifluoromethylphenyl piperazine was analgesic in a pindolol-reversible manner in the hot-plate test, but not the tail-flick test, in rats. Therefore, it would be of interest to assess the relative importance of these two receptor subtypes in other species and in other analgesic assays not involving a spinal reflex. For example, 5-HT_1B receptors may be important in the modulation of antinociception in the rat hot-plate test.

Nevertheless, a role for 5-HT_1B receptors in the modulation of nociception does not fit with neuroanatomical and physiological data regarding this subtype. If 5-HT_1B receptors are autoreceptors on the neurons of the descending inhibitory serotonergic tract from the nucleus raphe magnus to the spinal dorsal horn (Glaum and Anderson, 1988), then stimulation of these receptors would be expected to attenuate the activity of the descending inhibitory system. The expected action would be hyperalgesia or allodynia, rather than analgesia.

In summary, the analgesia produced by spinal injection of 5-HT in CD-1 mice appears to be mediated primarily by 5-HT3 receptors, as does the analgesia produced by 2-methyl-5-HT. Because neither AODN treatment significantly attenuated the analgesia produced by CGS-12066A, this effect is not mediated by 5-HT_1B receptors or 5-HT₃ receptors. However, it is not clear whether this analgesic effect is mediated by the release of calcitonin gene-related peptide, stimulation of another 5-HT receptor subtype, or some other receptor that may be stimulated by CGS-12066A.