- Email: [email protected]
Mechanisms of antinociception of spinal galanin: how does galanin inhibit spinal sensitization?
Neuropeptides Neuropeptides 39 (2005) 211–216 www.elsevier.com/locate/npep
Special Issue on Galanin
Mechanisms of antinociception of spinal galanin: how does galanin inhibit spinal sensitization? X.-Y. Hua a,*, K.F. Salgado a, G. Gu a, B. Fitzsimmons a, I. Kondo a, T. Bartfai b, T.L. Yaksh a a
b
Anesthesia Research Laboratory, Department of Anesthesiology, University of California, San Diego, 9500 Gilman Drive, La Jolla, CA 92093-0818, USA The Harold Dorris Neurological Institute, The Scripps Research Institute, La Jolla, CA, USA Received 30 November 2004; accepted 2 December 2004 Available online 19 March 2005
Abstract Galanin by a spinal action has been shown to have an antihyperalgesic action. Thus, in rats with lumbar intrathecal (IT) catheters, the thermal hyperalgesia evoked by carrageenan paw injection was blocked by IT delivery of galanin1–29 (Gal1–29) and galanin2–11 (Gal2–11) with the rank order of activity being Gal1–29 > Gal2–11. We sought to determine whether this spinal action reflects an effect upon afferent transmitter release, e.g., substance P (SP), and/or on secondary neurons, e.g., signaling postsynaptic to neurokinin 1 (NK1) receptor activation. To address the question on afferent release, we investigated the effect of IT administration of galanin on tissue injury-induced spinal NK1 internalization (an indicator of SP release). Noxious stimulation (paw compression) produced an increase in NK1 internalization in dorsal horn lamina I. IT pretreatment of rats with Gal1–29 and Gal2–11 significantly attenuated the evoked NK1 internalization, with the rank order of activity being Gal1–29 > Gal2–11 > saline. To address the question of postsynaptic action, we examined the effects of IT galanin upon IT SP-induced thermal hyperalgesia and spinal PGE2 release. Application of SP (30 nmol) directly to spinal cord led to a decrease in thermal thresholds and a profound increase in PGE2 concentration in spinal dialysates. Both phenomena were reversed by Gal1–29 and Gal2–11 (10 nmol, IT). These findings suggest that the antihyperalgesic effect of spinal galanin is due to its action on sites both presynaptic (inhibition of SP release) and postsynaptic (blockade of SP-evoked hyperalgesia and PGE2 production) to the primary afferents. 2005 Elsevier Ltd. All rights reserved. Keywords: Galanin; Antinociception; Intrathecal; Substance P; Neurokinin 1 receptor; Prostaglandins
1. Introduction Tissue injury and inflammation results in an increased sensitivity to subsequent noxious stimulation, indicating hyperalgesia. Evidence suggests that hyperalgesia arises in large part from facilitated processing of noxious input at spinal level. The injury-evoked afferent input leads to spinal release of excitatory amino *
Corresponding author. Tel.: +1 619 543 3597/2007; fax: +1 619 543 6070. E-mail address: [email protected] (X.-Y. Hua). 0143-4179/$ - see front matter 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.npep.2004.12.024
acids (e.g., glutamate), and peptides (e.g., substance P, SP), which, via activation of the respective receptors and subsequent signaling pathway, generate spinal sensitization. It has been shown that intrathecal (IT) delivery of galanin produce antihyperalgesic effects in a number of animal pain models (Hua et al., 2004; Liu and Hokfelt, 2002; Xu et al., 2000). The mechanism of these effects is not clear. Galanin immunoreactivity is expressed in dorsal root ganglia (DRG) (Hokfelt et al., 1987). In spinal dorsal horn, a significant proportion of the galanin-immunoreactivity is present in primary afferent fibers (Zhang et al., 1995a, 1993) with
212
X.-Y. Hua et al. / Neuropeptides 39 (2005) 211–216
the remainder present in a subpopulation of lamina II interneurons, which also contain GABA and enkephalins (Simmons et al., 1995; Zhang et al., 1995b). Three galanin receptor subtypes, GalR1, GalR2 and GalR3, have been cloned, and they belong to the superfamily of G-protein coupled receptors (for review see (Branchek et al., 2000)). Activation of either GalR1 or GalR3 produces hyperpolarization via Gi/o and inhibits adenylyl cyclase. GalR2 activation leads to stimulation of phospholipase C via Gq/11, producing calcium mobilization, diacylgycerol formation and subsequent activation of protein kinase C (Branchek et al., 2000). All three receptor transcripts are present in DRG and spinal cord (Waters and Krause, 2000). GalR1 mRNA is predominantly present in larger DRG cells, while GalR2 is highly expressed in small and intermediate neurons (OÕDonnell et al., 1999). The motif of Gi protein coupled receptors presynaptic on primary afferents suggest the possibility that activation of these receptors may alter spinal nociceptive processing by regulating primary afferent terminal release and reducing the post synaptic excitability of the nociceptive processing (Yaksh, 1999). GalR1 mRNA has also been found in lamina II local neurons (OÕDonnell et al., 1999; Parker et al., 1995), and galanin opens K+ channels leading to hyperpolarization of neurons (Ren et al., 2001); this suggests that a postsynaptic action of galanin may be also involved. In the present study, we investigated (i) whether IT galanin (Gal1–29 vs. Gal2–11) at antinociceptive doses would alter tissue injury-induced spinal SP release (as measured by NK1 receptor internalization), and (ii) whether IT galanin will affect direct activation of spinal NK1 receptor (by IT SP)-mediated hyperalgesia and PGE2 release.
the internalization of the neurokinin 1 (NK1) receptor, a marker of substance P release from small afferent terminals (Mantyh et al., 1995; Marvizon et al., 1997). Rats received IT injection of Gal1–29, Gal2–11 (30 nmol each, Biopeptide, San Diego, CA) or saline (10 ll). After 10 min, the rats were anesthetized with sodium pentobarbital (100 mg/kg). A single hind paw was then subjected to a brief intense compression with a pair of non-serrated forceps for 1 min. Five min later, the rats were perfused with saline followed by 4% paraformaldehyde in 0.1 M phosphate buffer (pH 7.4, PBS). The lumbar spinal cord was removed and post-fixed in the same fixative overnight followed by 30% sucrose for cryoprotection. Immunostaining was performed on 30 lm coronal sections of the lumbar spinal cord cut by a cryostat. NK1 receptors were identified using an anti-NK1 polyclonal antibody (Advanced Targeting Systems, San Diego, CA) diluted of 1:3000 in PBS with 10% normal goat serum and 0.2% Triton X-100. Sections were incubated with primary antibody for 24 h at room temperature followed by incubation with secondary antibody (Alexa-488 conjugated goat anti-Rabbit IgG) at 1:1000 in PBS, 5% normal goat serum and 0.2% Triton X-100 for 2 h, then mounted and cover-slipped with Prolong (Molecular Probes, Eugene, OR). The magnitude of NK1 receptor internalization was quantified by counting at 60X NK1 immunoreactive neurons in lamina I ipsilateral and contralateral to stimulation. Neuronal somas and contiguous proximal dendrites with ten or more endosomes were considered to have internalized receptors (Mantyh et al., 1995). Data are expressed as mean percent of NK1 positive neurons showing the receptor internalization in multiple sections from Lumbar 5–6. 2.3. Thermal paw withdrawal test
2. Methods All experiments were carried out according to protocols approved by the Institutional Animal Care Committee of University of California, San Diego. 2.1. Animals Male Holtzman Sprague–Dawley rats (300–350 g) were prepared under isoflurane anesthesia with chronic lumbar intrathecal catheters (Yaksh and Rudy, 1976) or triple lumen intrathecal dialysis loop catheters (Koetzner et al., 2004) that permit concurrent dialysis and intrathecal injection. 2.2. Paw compression injury and NK1 immunohistochemistry To determine whether spinal galanin receptors altered small afferent transmitter release, we examined
Thermally evoked paw withdrawal response was assessed by using a commercially available Hargreavestype device (Dirig et al., 1997). Briefly, rats were placed on a 30 C glass surface and a thermal nociceptive stimulus, originating from a projection bulb below the glass surface, was projected separately to each hind paw. Gal1–29 or Gal2–11 was given IT 10 min prior to carrageenan injection or IT SP (30 nmol, Sigma). Thermal paw withdrawal tests were carried out on rats that received carrageenan (2%, 100 ll, s.c., unilateral plantar) paw injection, or IT SP. 2.4. Intrathecal dialysis and PGE2 assay To define the effects of spinal galanin on downstream spinal cascade initiated by spinal NK1 activation, we examined the effects of intrathecal galanin on PGE2 release evoked by IT SP. Dialysis was undertaken in rats with triple lumen lumbar intrathecal dialysis catheters
X.-Y. Hua et al. / Neuropeptides 39 (2005) 211–216
213
at 10 ll/min (Koetzner et al., 2004). Two baseline samples were collected following a 30-min washout, and an additional four fractions were collected after IT injection of SP. Gal1–29 or Gal2–11 was given IT 10 min prior to IT SP. PGE2 in spinal dialysate was measured by an ELISA selective for PGE2 with less than 2.0% crossreactivity to PGF1a, PGF2a, 6-ketoPGF1a, PGA2 or PGB2, but cross-reacts with PGE1 and PGE3 (Assay Designs 90001, Assay Designs, Ann Arbor, MI).
3. Results 3.1. Paw carrageenan-induced thermal hyperalgesia Baseline latencies before injection of carrageenan were 12.7 ± 0.3 s for the right paw (contralateral) and 12.9 ± 0.8 s for the left paw (ipsilateral) in IT saline treated group (N = 6). After carrageenan injection into the plantar surface of the left paw, a progressive reduction in time to paw withdrawal was observed. The withdrawal latency of the inflamed paw of saline rats decreased to 4.3 ± 0.8 s at 120 min and maintained the same level for next 2 h (Fig. 1). Pretreatment with Gal1–29 (30 nmol, IT, N = 7) significantly attenuated carrageenan-induced thermal hyperalgesia (Fig. 1A and C), while Gal2–11 (N = 6) at the same dose displayed a small inhibition (Fig. 3B and C), but the difference did not reach statistical significance (Fig. 3C). The response latency of the non-inflamed paw remained unchanged during the 4 h examination period, and there was no difference between the groups of IT saline, Gal1–29 and Gal2–11. 3.2. Paw compression-induced NK1 internalization NK1 receptor bearing neurons were prominent in the superficial dorsal horn. As previously described (Kondo et al., 2004), 5 min after unilateral compression of the hind paw, approximately 50% of the ipsilateral NK1 positive neurons in lamina I of the L5–L6 segments displayed internalization as compared to 10% on the contralateral side in intrathecal saline-treated (N = 5) animals (Fig. 2). Gal1–29 (N = 5) and Gal2–11 (N = 4) both at the dose of 30 nmol (IT) significantly decreased the evoked NK1 internalization with the ordering of activity being Gal1–29 > Gal2–11 > saline = 0 (p < 0.05, Fig. 3). 3.3. Intrathecal SP evoked thermal hyperalgesia Baseline escape latency to a thermal stimulus applied to the hind paw was 10.1 ± 0.4 s in control group (SP + saline, N = 7). Gal1–29 or Gal2–11 (10 nmol, IT) had no effect alone upon the thermal escape latency as compared to control. Ten minutes after IT bolus injec-
Fig. 1. Effects of IT Gal1–29 (A) and Gal2–11 (B) at the dose of 30 nmol on carrageenan-induced thermal hyperalgesia. Paw withdrawal thresholds in seconds (s) to thermal stimulation in each paw plotted against time before and after IT injection of saline, Gal1–29 or Gal2–11. Carrageenan was injected s.c. to plantar surface of left paw. (C) Antinociceptive effects of Gal1–29 and Gal2–11 on carrageenan-induced hyperalgesia are presented as area under curve (AUC, paw thermal threshold · 4 h). The data are presented as means ± S.E.M. of 6–7 rats per groups. *p < 0.05, ANOVA.
tion of SP (30 nmol), the paw withdrawal latencies were reduced to 7.1 ± 0.3 s (Algesic index: 30 ± 3, Fig. 4A). The SP-evoked thermal hyperalgesia was reversed by IT pretreatment with Gal1–29 or Gal2–11 (10 nmol, N = 5, p < 0.05, Fig. 4B). 3.4. Intrathecal SP-induced spinal PGE2 release Baseline levels of PGE2 in spinal dialysates were 251 ± 64 fmol/100 ll (N = 12). Following IT injection of SP (30 nmol), there was a rapid and profound (5–6 fold) increase in PGE2 concentration in the dialysates. IT Gal1–29 (N = 7) and Gal2–11 (N = 9) at 10 nmol efficiently inhibited IT SP-evoked release of spinal PGE2 (p < 0.05, Fig. 5).
214
X.-Y. Hua et al. / Neuropeptides 39 (2005) 211–216
Fig. 2. Confocal images show lamina I NK1 immunoreactive neurons in ipsilateral (ipsi) and contralateral (contra) dorsal horn in the paw compression model in rats receiving IT saline (10 ll) or Gal1–29 (30 nmol). NK1 positive neurons in ipsilateral cord show the internalization in a rat with IT saline (A), but not with IT Gal1–29 (C). The NK1 internalization was minimal in contralateral sides (B and D).
(Yaksh et al., 1999). The present work demonstrates that the antinociceptive effect of spinal galanin may be mediated by actions at sites both presynaptic and postsynaptic to the primary sensory afferent. 4.1. Galanin effect upon SP release
Fig. 3. The bar graph presents the difference between the percent of neurons displaying NK1 internalization in the ipsilateral vs. contralalateral dorsal horn after ipsilateral paw compression in rats receiving IT saline, Gal1–29 (30 nmol) or Gal2–11 (30 nmol). N = 4–5 rats per group. *p < 0.05, ANOVA.
4. Discussion Following peripheral tissue injury there is a behaviorally defined hyperalgesia. Current thinking indicates that this hyperalgesia reflects repetitive activation of small afferents which releases excitatory neurotransmitters, such as glutamate and SP. This release initiates a spinal cascade that leads to a facilitated state by the activation of several receptors including the NK1 receptors
The presence of galanin receptors GalR1 and GalR2 on DRG and the spinal terminals of primary afferent fibers respectively suggest possible inhibitory and/or facilitatory effects of galanin upon primary afferent terminal excitability. Our data on the effects of IT galanin on NK1 internalization indicate that galanin presynaptically inhibits the release of SP from central terminals of small afferents. This finding is consistent with the ability of galanin to block the opening of voltage sensitive calcium channels (Parsons et al., 1998). In this case, the galanin effects are similar to that produced by other G protein coupled receptors with presynaptic binding (e.g., l/o and a2 receptors) which have been shown to modulate small afferent release and are believed to produce in part their antinociceptive effects at the spinal level (Yaksh, 1999). 4.2. Postsynaptic effect of galanin We have previously demonstrated that NK1 receptor activation will initiate PGE2 release though activation of
X.-Y. Hua et al. / Neuropeptides 39 (2005) 211–216
a COX-2 isozyme (Yaksh et al., 2001). Importantly, this PGE2 release mediates, in part, the observed NK1-mediated hyperalgesia, as these behavioral effects are also prevented by spinal COX inhibition (Malmberg and Yaksh, 1992). Given that current data do not support
215
a pre-afferent terminal effect of NK1 receptors, IT SP initiated events are believed to represent a direct activation of a post-primary afferent terminal mechanism. Accordingly, the ability of galanin to block the NK1evoked PGE2 release and affiliated hyperalgesia is consistent with the hypothesis that galanin also exerts a direct effect postsynaptic to the primary afferents. 4.3. Receptor pharmacology of spinal galanin An important question relates to the identity of the galanin receptor subtypes which mediate these several effects of spinal galanin. In the present work, Gal1–29 was more efficacious than an equimolar dose of Gal2–11 in reducing SP release. Conversely, in blocking the NK1-mediated thermal hyperalgesia and PGE2 release, these molar doses were equieffective. It has been argued that Gal1–29 is equally effective at both GalR1 and GalR2 sites (Branchek et al., 2000), while Gal2–11 is more effective at GalR2 than GalR1 (Liu and Hokfelt, 2002). As we observed in two animal pain behavioral models, i.e., formalin test (Hua et al., 2004) and carrageenan model (the present study), the antihyperalgesic effects of IT Gal1–29 are more potent than Gal2–11. Based on this relative efficacy and the reported G protein coupling, the present results would be consistent with the hypothesis that GalR1 mediates both pre and post synaptic effects. As regards the role of the GalR2, this receptor is considered to be
Fig. 4. (A) Time course of effects of IT pre-treatment with Gal1–29 and Gal2–11 (10 nmol each) on IT SP (30 nmol)-induced thermal hyperalgesia. The data are present as algesic Index = (baseline latency posttreatment latency)/baseline latency · 100%. (B) The anti-hyperalgesic effects of Gal1–29 and Gal2–11 are presented as are under curve (AUC: Algesic index · 60 min). *p < 0.05, ANOVA. N = 5–7.
Fig. 5. Bar graph shows PGE2 concentration measured in CSF collected by in vivo spinal dialysis of conscious rats before and after IT injection of saline (10 ll), SP (30 nmol), SP + Gal1–29 (30/10 nmol each) or SP + Gal2–11 (30/10 nmol each), and presented as % baseline. Gal1–29 and Gal2–11 were given 10 min prior to IT SP. *p < 0.05, SP vs. saline, SP + Gal1–29 or SP + Gal2–11 at 15, 30, 45 and 60 min, ANOVA. N = 7–12.
Fig. 6. Schematic of proposed connectivity accounting for the observed actions of intrathecal galanin acting through GalR1 and GalR2 on small afferent evoked dorsal processing. Released SP is believed to act via an NK1 receptor to initiate downstream cascades that lead to hyperalgesia, in part through the spinal release of prostaglandins. As discussed in the text, based on the present physiological and behavioral effects, we hypothesize that GalR1 serves to exert an inhibitory effect at sites pre and post synaptic to the primary afferent, while GalR2 may activate a local inhibitory circuit. The role of GalR3 cannot be asserted at the present time due to a lack of definitive pharmacology.
216
X.-Y. Hua et al. / Neuropeptides 39 (2005) 211–216
excitatory (Liu and Hokfelt, 2002), although the relative selectivity of Gal2–11 for GalR2 is controversial. GalR2 receptors are believed to be preterminal on primary afferents (OÕDonnell et al., 1999), Gal1–29 or Gal2–11 alone, however, had at most only a minimal stimulatory effect upon basal SP release. Moreover, neither peptide had any effect upon the acute thermal threshold (data not shown), suggesting that unlike SP, there was little evoked facilitation by galanin peptides in the present model. These observations are consistent with possibility that GalR2 may be excitatory on inhibitory interneurons (Simmons et al., 1995; Zhang et al., 1995b). The organization of the galanin receptor subtypes on spinal circuitry is suggested in Fig. 6. Further definition of this pharmacology and mechanism awaits selective agonists and antagonists. In conclusion, the present findings indicate that antinociception of IT galanin is due to its inhibitory action on sites both presynaptic and postsynaptic to the primary afferents, which prevent development of spinal sensitization and hyperalgesia. Inhibition of release of SP from small afferents (via GalR1), and blockade of spinal neuron activation-induced PGE2 production (via GalR1/2) are likely two potential mechanisms.
Acknowledgment This work was supportd by NIH NS 41954.
References Branchek, T.A., Smith, K.E., Gerald, C., Walker, M.W., 2000. Galanin receptor subtypes. Trends Pharmacol. Sci. 21, 109–117. Dirig, D.M., Salami, A., Rathbun, M.L., Ozaki, G.T., Yaksh, T.L., 1997. Characterization of variables defining hindpaw withdrawal latency evoked by radiant thermal stimuli. J. Neurosci. Methods 76, 183–191. Hokfelt, T., Wiesenfeld-Hallin, Z., Villar, M., Melander, T., 1987. Increase of galanin-like immunoreactivity in rat dorsal root ganglion cells after peripheral axotomy. Neurosci. Lett. 83, 217– 220. Hua, X.Y., Hayes, C.S., Hofer, A., Fitzsimmons, B., Kilk, K., Langel, U., Bartfai, T., Yaksh, T.L., 2004. Galanin acts at GalR1 receptors in spinal antinociception: synergy with morphine and AP-5. J. Pharmacol. Exp. Ther. 308, 574–582. Koetzner, L., Hua, X.Y., Lai, J., Porreca, F., Yaksh, T.L, 2004. Nonopioid actions of intrathecal dynorphin evoke spinal excitatory amino acid and prostaglandin E2 release mediated by cyclooxygenase-1 and -2. J. Neurosci. 24, 1451–1458. Kondo, I., Salgado, F., Hua, X.-Y., Yaksh, T.L., 2004. Analgesic doses of l and o opiates inhibit noxious stimulation-induced spinal NK1 receptor internalization. Society for Neuroscience, San Diego, 862.8. Liu, H.X., Hokfelt, T., 2002. The participation of galanin in pain processing at the spinal level. Trends Pharmacol. Sci. 23, 468–474.
Malmberg, A.B., Yaksh, T.L., 1992. Hyperalgesia mediated by spinal glutamate or substance P receptor blocked by spinal cyclooxygenase inhibition. Science 257, 1276–1279. Mantyh, P.W., DeMaster, E., Malhotra, A., Ghilardi, J.R., Rogers, S.D., Mantyh, C.R., Liu, H., Basbaum, A.I., Vigna, S.R., Maggio, J.E., 1995. Receptor endocytosis and dendite reshaping in spinal neurons after somatosensory stimulation. Science 268, 1629–1632. Marvizon, J.C., Martinez, V., Grady, E.F., Bunnett, N.W., Mayer, E.A., 1997. Neurokinin 1 receptor internalization in spinal cord slices induced by dorsal root stimulation is mediated by NMDA receptors. J. Neurosci. 17, 8129–8136. OÕDonnell, D., Ahmad, S., Wahlestedt, C., Walker, P., 1999. Expression of the novel galanin receptor subtype GALR2 in the adult rat CNS: distinct distribution from GALR1. J. Comp. Neurol. 409, 469–481. Parker, E.M., Izzarelli, D.G., Nowak, H.P., Mahle, C.D., Iben, L.G., Wang, J., Goldstein, M.E., 1995. Cloning and characterization of the rat GALR1 galanin receptor from Rin14B insulinoma cells. Brain Res. Mol. Brain Res 34, 179–189. Parsons, R.L., Mulvaney, J.M., Merriam, L.A., 1998. Galanin activates an inwardly rectifying potassium conductance and inhibits a voltage-dependent calcium conductance in mudpuppy parasympathetic neurons. Ann. N.Y. Acad. Sci. 863, 156–169. Ren, J., Hu, H.Z., Starodub, A.M., Wood, J.D., 2001. Galanin suppresses calcium conductance and activates inwardly rectifying potassium channels in myenteric neurones from guinea-pig small intestine. Neurogastroenterol. Motil. 13, 247–254. Simmons, D.R., Spike, R.C., Todd, A.J., 1995. Galanin is contained in GABAergic neurons in the rat spinal dorsal horn. Neurosci. Lett. 187, 119–122. Waters, S.M., Krause, J.E., 2000. Distribution of galanin-1, -2 and -3 receptor messenger RNAs in central and peripheral rat tissues. Neuroscience 95, 265–271. Xu, X.J., Hokfelt, T., Bartfai, T., Wiesenfeld-Hallin, Z., 2000. Galanin and spinal nociceptive mechanisms: recent advances and therapeutic implications. Neuropeptides 34, 137–147. Yaksh, T.L., 1999. Spinal systems and pain processing: development of novel analgesic drugs with mechanistically defined models. Trends Pharmacol. Sci. 20, 329–337. Yaksh, T.L., Rudy, T.S., 1976. Chronic catheterization of the spinal subarachnoid space. Physiol. Behav. 17, 1031–1036. Yaksh, T.L., Hua, X.Y., Kalcheva, I., Nozaki-Taguchi, N., Marsala, M., 1999. The spinal biology in humans and animals of pain states generated by persistent small afferent input. Proc. Natl. Acad. Sci. USA 96, 7680–7686. Zhang, X., Nicholas, A.P., Hokfelt, T., 1993. Ultrastructural studies on peptides in the dorsal horn of the spinal cord – I. Co-existence of galanin with other peptides in primary afferents in normal rats. Neuroscience 57, 365–384. Yaksh, T.L., Dirig, D.M., Conway, C.M., Svensson, C., Luo, Z.D., Isakson, P.C., 2001. The acute antihyperalgesic action of nonsteroidal, anti-inflammatory drugs and release of spinal prostaglandin E2 is mediated by the inhibition of constitutive spinal cyclooxygenase-2 (COX-2) but not COX-1. J. Neurosci. 21, 5847–5853. Zhang, X., Ji, R.R., Nilsson, S., Villar, M., Ubink, R., Ju, G., Wiesenfeld-Hallin, Z., Hokfelt, T., 1995b. Neuropeptide Y and galanin binding sites in rat and monkey lumbar dorsal root ganglia and spinal cord and effect of peripheral axotomy. Eur. J. Neurosci. 7, 367–380. Zhang, Q., Shi, T.J., Ji, R.R., Zhang, Y.Z., Sundler, F., Hannibal, J., Fahrenkrug, J., Hokfelt, T., Zhang, Y., 1995a. Expression of pituitary adenylate cyclase-activating polypeptide in dorsal root ganglia following axotomy: time course and coexistence. Brain Res. 705, 149–158.
Special Issue on Galanin
Mechanisms of antinociception of spinal galanin: how does galanin inhibit spinal sensitization? X.-Y. Hua a,*, K.F. Salgado a, G. Gu a, B. Fitzsimmons a, I. Kondo a, T. Bartfai b, T.L. Yaksh a a
b
Anesthesia Research Laboratory, Department of Anesthesiology, University of California, San Diego, 9500 Gilman Drive, La Jolla, CA 92093-0818, USA The Harold Dorris Neurological Institute, The Scripps Research Institute, La Jolla, CA, USA Received 30 November 2004; accepted 2 December 2004 Available online 19 March 2005
Abstract Galanin by a spinal action has been shown to have an antihyperalgesic action. Thus, in rats with lumbar intrathecal (IT) catheters, the thermal hyperalgesia evoked by carrageenan paw injection was blocked by IT delivery of galanin1–29 (Gal1–29) and galanin2–11 (Gal2–11) with the rank order of activity being Gal1–29 > Gal2–11. We sought to determine whether this spinal action reflects an effect upon afferent transmitter release, e.g., substance P (SP), and/or on secondary neurons, e.g., signaling postsynaptic to neurokinin 1 (NK1) receptor activation. To address the question on afferent release, we investigated the effect of IT administration of galanin on tissue injury-induced spinal NK1 internalization (an indicator of SP release). Noxious stimulation (paw compression) produced an increase in NK1 internalization in dorsal horn lamina I. IT pretreatment of rats with Gal1–29 and Gal2–11 significantly attenuated the evoked NK1 internalization, with the rank order of activity being Gal1–29 > Gal2–11 > saline. To address the question of postsynaptic action, we examined the effects of IT galanin upon IT SP-induced thermal hyperalgesia and spinal PGE2 release. Application of SP (30 nmol) directly to spinal cord led to a decrease in thermal thresholds and a profound increase in PGE2 concentration in spinal dialysates. Both phenomena were reversed by Gal1–29 and Gal2–11 (10 nmol, IT). These findings suggest that the antihyperalgesic effect of spinal galanin is due to its action on sites both presynaptic (inhibition of SP release) and postsynaptic (blockade of SP-evoked hyperalgesia and PGE2 production) to the primary afferents. 2005 Elsevier Ltd. All rights reserved. Keywords: Galanin; Antinociception; Intrathecal; Substance P; Neurokinin 1 receptor; Prostaglandins
1. Introduction Tissue injury and inflammation results in an increased sensitivity to subsequent noxious stimulation, indicating hyperalgesia. Evidence suggests that hyperalgesia arises in large part from facilitated processing of noxious input at spinal level. The injury-evoked afferent input leads to spinal release of excitatory amino *
Corresponding author. Tel.: +1 619 543 3597/2007; fax: +1 619 543 6070. E-mail address: [email protected] (X.-Y. Hua). 0143-4179/$ - see front matter 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.npep.2004.12.024
acids (e.g., glutamate), and peptides (e.g., substance P, SP), which, via activation of the respective receptors and subsequent signaling pathway, generate spinal sensitization. It has been shown that intrathecal (IT) delivery of galanin produce antihyperalgesic effects in a number of animal pain models (Hua et al., 2004; Liu and Hokfelt, 2002; Xu et al., 2000). The mechanism of these effects is not clear. Galanin immunoreactivity is expressed in dorsal root ganglia (DRG) (Hokfelt et al., 1987). In spinal dorsal horn, a significant proportion of the galanin-immunoreactivity is present in primary afferent fibers (Zhang et al., 1995a, 1993) with
212
X.-Y. Hua et al. / Neuropeptides 39 (2005) 211–216
the remainder present in a subpopulation of lamina II interneurons, which also contain GABA and enkephalins (Simmons et al., 1995; Zhang et al., 1995b). Three galanin receptor subtypes, GalR1, GalR2 and GalR3, have been cloned, and they belong to the superfamily of G-protein coupled receptors (for review see (Branchek et al., 2000)). Activation of either GalR1 or GalR3 produces hyperpolarization via Gi/o and inhibits adenylyl cyclase. GalR2 activation leads to stimulation of phospholipase C via Gq/11, producing calcium mobilization, diacylgycerol formation and subsequent activation of protein kinase C (Branchek et al., 2000). All three receptor transcripts are present in DRG and spinal cord (Waters and Krause, 2000). GalR1 mRNA is predominantly present in larger DRG cells, while GalR2 is highly expressed in small and intermediate neurons (OÕDonnell et al., 1999). The motif of Gi protein coupled receptors presynaptic on primary afferents suggest the possibility that activation of these receptors may alter spinal nociceptive processing by regulating primary afferent terminal release and reducing the post synaptic excitability of the nociceptive processing (Yaksh, 1999). GalR1 mRNA has also been found in lamina II local neurons (OÕDonnell et al., 1999; Parker et al., 1995), and galanin opens K+ channels leading to hyperpolarization of neurons (Ren et al., 2001); this suggests that a postsynaptic action of galanin may be also involved. In the present study, we investigated (i) whether IT galanin (Gal1–29 vs. Gal2–11) at antinociceptive doses would alter tissue injury-induced spinal SP release (as measured by NK1 receptor internalization), and (ii) whether IT galanin will affect direct activation of spinal NK1 receptor (by IT SP)-mediated hyperalgesia and PGE2 release.
the internalization of the neurokinin 1 (NK1) receptor, a marker of substance P release from small afferent terminals (Mantyh et al., 1995; Marvizon et al., 1997). Rats received IT injection of Gal1–29, Gal2–11 (30 nmol each, Biopeptide, San Diego, CA) or saline (10 ll). After 10 min, the rats were anesthetized with sodium pentobarbital (100 mg/kg). A single hind paw was then subjected to a brief intense compression with a pair of non-serrated forceps for 1 min. Five min later, the rats were perfused with saline followed by 4% paraformaldehyde in 0.1 M phosphate buffer (pH 7.4, PBS). The lumbar spinal cord was removed and post-fixed in the same fixative overnight followed by 30% sucrose for cryoprotection. Immunostaining was performed on 30 lm coronal sections of the lumbar spinal cord cut by a cryostat. NK1 receptors were identified using an anti-NK1 polyclonal antibody (Advanced Targeting Systems, San Diego, CA) diluted of 1:3000 in PBS with 10% normal goat serum and 0.2% Triton X-100. Sections were incubated with primary antibody for 24 h at room temperature followed by incubation with secondary antibody (Alexa-488 conjugated goat anti-Rabbit IgG) at 1:1000 in PBS, 5% normal goat serum and 0.2% Triton X-100 for 2 h, then mounted and cover-slipped with Prolong (Molecular Probes, Eugene, OR). The magnitude of NK1 receptor internalization was quantified by counting at 60X NK1 immunoreactive neurons in lamina I ipsilateral and contralateral to stimulation. Neuronal somas and contiguous proximal dendrites with ten or more endosomes were considered to have internalized receptors (Mantyh et al., 1995). Data are expressed as mean percent of NK1 positive neurons showing the receptor internalization in multiple sections from Lumbar 5–6. 2.3. Thermal paw withdrawal test
2. Methods All experiments were carried out according to protocols approved by the Institutional Animal Care Committee of University of California, San Diego. 2.1. Animals Male Holtzman Sprague–Dawley rats (300–350 g) were prepared under isoflurane anesthesia with chronic lumbar intrathecal catheters (Yaksh and Rudy, 1976) or triple lumen intrathecal dialysis loop catheters (Koetzner et al., 2004) that permit concurrent dialysis and intrathecal injection. 2.2. Paw compression injury and NK1 immunohistochemistry To determine whether spinal galanin receptors altered small afferent transmitter release, we examined
Thermally evoked paw withdrawal response was assessed by using a commercially available Hargreavestype device (Dirig et al., 1997). Briefly, rats were placed on a 30 C glass surface and a thermal nociceptive stimulus, originating from a projection bulb below the glass surface, was projected separately to each hind paw. Gal1–29 or Gal2–11 was given IT 10 min prior to carrageenan injection or IT SP (30 nmol, Sigma). Thermal paw withdrawal tests were carried out on rats that received carrageenan (2%, 100 ll, s.c., unilateral plantar) paw injection, or IT SP. 2.4. Intrathecal dialysis and PGE2 assay To define the effects of spinal galanin on downstream spinal cascade initiated by spinal NK1 activation, we examined the effects of intrathecal galanin on PGE2 release evoked by IT SP. Dialysis was undertaken in rats with triple lumen lumbar intrathecal dialysis catheters
X.-Y. Hua et al. / Neuropeptides 39 (2005) 211–216
213
at 10 ll/min (Koetzner et al., 2004). Two baseline samples were collected following a 30-min washout, and an additional four fractions were collected after IT injection of SP. Gal1–29 or Gal2–11 was given IT 10 min prior to IT SP. PGE2 in spinal dialysate was measured by an ELISA selective for PGE2 with less than 2.0% crossreactivity to PGF1a, PGF2a, 6-ketoPGF1a, PGA2 or PGB2, but cross-reacts with PGE1 and PGE3 (Assay Designs 90001, Assay Designs, Ann Arbor, MI).
3. Results 3.1. Paw carrageenan-induced thermal hyperalgesia Baseline latencies before injection of carrageenan were 12.7 ± 0.3 s for the right paw (contralateral) and 12.9 ± 0.8 s for the left paw (ipsilateral) in IT saline treated group (N = 6). After carrageenan injection into the plantar surface of the left paw, a progressive reduction in time to paw withdrawal was observed. The withdrawal latency of the inflamed paw of saline rats decreased to 4.3 ± 0.8 s at 120 min and maintained the same level for next 2 h (Fig. 1). Pretreatment with Gal1–29 (30 nmol, IT, N = 7) significantly attenuated carrageenan-induced thermal hyperalgesia (Fig. 1A and C), while Gal2–11 (N = 6) at the same dose displayed a small inhibition (Fig. 3B and C), but the difference did not reach statistical significance (Fig. 3C). The response latency of the non-inflamed paw remained unchanged during the 4 h examination period, and there was no difference between the groups of IT saline, Gal1–29 and Gal2–11. 3.2. Paw compression-induced NK1 internalization NK1 receptor bearing neurons were prominent in the superficial dorsal horn. As previously described (Kondo et al., 2004), 5 min after unilateral compression of the hind paw, approximately 50% of the ipsilateral NK1 positive neurons in lamina I of the L5–L6 segments displayed internalization as compared to 10% on the contralateral side in intrathecal saline-treated (N = 5) animals (Fig. 2). Gal1–29 (N = 5) and Gal2–11 (N = 4) both at the dose of 30 nmol (IT) significantly decreased the evoked NK1 internalization with the ordering of activity being Gal1–29 > Gal2–11 > saline = 0 (p < 0.05, Fig. 3). 3.3. Intrathecal SP evoked thermal hyperalgesia Baseline escape latency to a thermal stimulus applied to the hind paw was 10.1 ± 0.4 s in control group (SP + saline, N = 7). Gal1–29 or Gal2–11 (10 nmol, IT) had no effect alone upon the thermal escape latency as compared to control. Ten minutes after IT bolus injec-
Fig. 1. Effects of IT Gal1–29 (A) and Gal2–11 (B) at the dose of 30 nmol on carrageenan-induced thermal hyperalgesia. Paw withdrawal thresholds in seconds (s) to thermal stimulation in each paw plotted against time before and after IT injection of saline, Gal1–29 or Gal2–11. Carrageenan was injected s.c. to plantar surface of left paw. (C) Antinociceptive effects of Gal1–29 and Gal2–11 on carrageenan-induced hyperalgesia are presented as area under curve (AUC, paw thermal threshold · 4 h). The data are presented as means ± S.E.M. of 6–7 rats per groups. *p < 0.05, ANOVA.
tion of SP (30 nmol), the paw withdrawal latencies were reduced to 7.1 ± 0.3 s (Algesic index: 30 ± 3, Fig. 4A). The SP-evoked thermal hyperalgesia was reversed by IT pretreatment with Gal1–29 or Gal2–11 (10 nmol, N = 5, p < 0.05, Fig. 4B). 3.4. Intrathecal SP-induced spinal PGE2 release Baseline levels of PGE2 in spinal dialysates were 251 ± 64 fmol/100 ll (N = 12). Following IT injection of SP (30 nmol), there was a rapid and profound (5–6 fold) increase in PGE2 concentration in the dialysates. IT Gal1–29 (N = 7) and Gal2–11 (N = 9) at 10 nmol efficiently inhibited IT SP-evoked release of spinal PGE2 (p < 0.05, Fig. 5).
214
X.-Y. Hua et al. / Neuropeptides 39 (2005) 211–216
Fig. 2. Confocal images show lamina I NK1 immunoreactive neurons in ipsilateral (ipsi) and contralateral (contra) dorsal horn in the paw compression model in rats receiving IT saline (10 ll) or Gal1–29 (30 nmol). NK1 positive neurons in ipsilateral cord show the internalization in a rat with IT saline (A), but not with IT Gal1–29 (C). The NK1 internalization was minimal in contralateral sides (B and D).
(Yaksh et al., 1999). The present work demonstrates that the antinociceptive effect of spinal galanin may be mediated by actions at sites both presynaptic and postsynaptic to the primary sensory afferent. 4.1. Galanin effect upon SP release
Fig. 3. The bar graph presents the difference between the percent of neurons displaying NK1 internalization in the ipsilateral vs. contralalateral dorsal horn after ipsilateral paw compression in rats receiving IT saline, Gal1–29 (30 nmol) or Gal2–11 (30 nmol). N = 4–5 rats per group. *p < 0.05, ANOVA.
4. Discussion Following peripheral tissue injury there is a behaviorally defined hyperalgesia. Current thinking indicates that this hyperalgesia reflects repetitive activation of small afferents which releases excitatory neurotransmitters, such as glutamate and SP. This release initiates a spinal cascade that leads to a facilitated state by the activation of several receptors including the NK1 receptors
The presence of galanin receptors GalR1 and GalR2 on DRG and the spinal terminals of primary afferent fibers respectively suggest possible inhibitory and/or facilitatory effects of galanin upon primary afferent terminal excitability. Our data on the effects of IT galanin on NK1 internalization indicate that galanin presynaptically inhibits the release of SP from central terminals of small afferents. This finding is consistent with the ability of galanin to block the opening of voltage sensitive calcium channels (Parsons et al., 1998). In this case, the galanin effects are similar to that produced by other G protein coupled receptors with presynaptic binding (e.g., l/o and a2 receptors) which have been shown to modulate small afferent release and are believed to produce in part their antinociceptive effects at the spinal level (Yaksh, 1999). 4.2. Postsynaptic effect of galanin We have previously demonstrated that NK1 receptor activation will initiate PGE2 release though activation of
X.-Y. Hua et al. / Neuropeptides 39 (2005) 211–216
a COX-2 isozyme (Yaksh et al., 2001). Importantly, this PGE2 release mediates, in part, the observed NK1-mediated hyperalgesia, as these behavioral effects are also prevented by spinal COX inhibition (Malmberg and Yaksh, 1992). Given that current data do not support
215
a pre-afferent terminal effect of NK1 receptors, IT SP initiated events are believed to represent a direct activation of a post-primary afferent terminal mechanism. Accordingly, the ability of galanin to block the NK1evoked PGE2 release and affiliated hyperalgesia is consistent with the hypothesis that galanin also exerts a direct effect postsynaptic to the primary afferents. 4.3. Receptor pharmacology of spinal galanin An important question relates to the identity of the galanin receptor subtypes which mediate these several effects of spinal galanin. In the present work, Gal1–29 was more efficacious than an equimolar dose of Gal2–11 in reducing SP release. Conversely, in blocking the NK1-mediated thermal hyperalgesia and PGE2 release, these molar doses were equieffective. It has been argued that Gal1–29 is equally effective at both GalR1 and GalR2 sites (Branchek et al., 2000), while Gal2–11 is more effective at GalR2 than GalR1 (Liu and Hokfelt, 2002). As we observed in two animal pain behavioral models, i.e., formalin test (Hua et al., 2004) and carrageenan model (the present study), the antihyperalgesic effects of IT Gal1–29 are more potent than Gal2–11. Based on this relative efficacy and the reported G protein coupling, the present results would be consistent with the hypothesis that GalR1 mediates both pre and post synaptic effects. As regards the role of the GalR2, this receptor is considered to be
Fig. 4. (A) Time course of effects of IT pre-treatment with Gal1–29 and Gal2–11 (10 nmol each) on IT SP (30 nmol)-induced thermal hyperalgesia. The data are present as algesic Index = (baseline latency posttreatment latency)/baseline latency · 100%. (B) The anti-hyperalgesic effects of Gal1–29 and Gal2–11 are presented as are under curve (AUC: Algesic index · 60 min). *p < 0.05, ANOVA. N = 5–7.
Fig. 5. Bar graph shows PGE2 concentration measured in CSF collected by in vivo spinal dialysis of conscious rats before and after IT injection of saline (10 ll), SP (30 nmol), SP + Gal1–29 (30/10 nmol each) or SP + Gal2–11 (30/10 nmol each), and presented as % baseline. Gal1–29 and Gal2–11 were given 10 min prior to IT SP. *p < 0.05, SP vs. saline, SP + Gal1–29 or SP + Gal2–11 at 15, 30, 45 and 60 min, ANOVA. N = 7–12.
Fig. 6. Schematic of proposed connectivity accounting for the observed actions of intrathecal galanin acting through GalR1 and GalR2 on small afferent evoked dorsal processing. Released SP is believed to act via an NK1 receptor to initiate downstream cascades that lead to hyperalgesia, in part through the spinal release of prostaglandins. As discussed in the text, based on the present physiological and behavioral effects, we hypothesize that GalR1 serves to exert an inhibitory effect at sites pre and post synaptic to the primary afferent, while GalR2 may activate a local inhibitory circuit. The role of GalR3 cannot be asserted at the present time due to a lack of definitive pharmacology.
216
X.-Y. Hua et al. / Neuropeptides 39 (2005) 211–216
excitatory (Liu and Hokfelt, 2002), although the relative selectivity of Gal2–11 for GalR2 is controversial. GalR2 receptors are believed to be preterminal on primary afferents (OÕDonnell et al., 1999), Gal1–29 or Gal2–11 alone, however, had at most only a minimal stimulatory effect upon basal SP release. Moreover, neither peptide had any effect upon the acute thermal threshold (data not shown), suggesting that unlike SP, there was little evoked facilitation by galanin peptides in the present model. These observations are consistent with possibility that GalR2 may be excitatory on inhibitory interneurons (Simmons et al., 1995; Zhang et al., 1995b). The organization of the galanin receptor subtypes on spinal circuitry is suggested in Fig. 6. Further definition of this pharmacology and mechanism awaits selective agonists and antagonists. In conclusion, the present findings indicate that antinociception of IT galanin is due to its inhibitory action on sites both presynaptic and postsynaptic to the primary afferents, which prevent development of spinal sensitization and hyperalgesia. Inhibition of release of SP from small afferents (via GalR1), and blockade of spinal neuron activation-induced PGE2 production (via GalR1/2) are likely two potential mechanisms.
Acknowledgment This work was supportd by NIH NS 41954.
References Branchek, T.A., Smith, K.E., Gerald, C., Walker, M.W., 2000. Galanin receptor subtypes. Trends Pharmacol. Sci. 21, 109–117. Dirig, D.M., Salami, A., Rathbun, M.L., Ozaki, G.T., Yaksh, T.L., 1997. Characterization of variables defining hindpaw withdrawal latency evoked by radiant thermal stimuli. J. Neurosci. Methods 76, 183–191. Hokfelt, T., Wiesenfeld-Hallin, Z., Villar, M., Melander, T., 1987. Increase of galanin-like immunoreactivity in rat dorsal root ganglion cells after peripheral axotomy. Neurosci. Lett. 83, 217– 220. Hua, X.Y., Hayes, C.S., Hofer, A., Fitzsimmons, B., Kilk, K., Langel, U., Bartfai, T., Yaksh, T.L., 2004. Galanin acts at GalR1 receptors in spinal antinociception: synergy with morphine and AP-5. J. Pharmacol. Exp. Ther. 308, 574–582. Koetzner, L., Hua, X.Y., Lai, J., Porreca, F., Yaksh, T.L, 2004. Nonopioid actions of intrathecal dynorphin evoke spinal excitatory amino acid and prostaglandin E2 release mediated by cyclooxygenase-1 and -2. J. Neurosci. 24, 1451–1458. Kondo, I., Salgado, F., Hua, X.-Y., Yaksh, T.L., 2004. Analgesic doses of l and o opiates inhibit noxious stimulation-induced spinal NK1 receptor internalization. Society for Neuroscience, San Diego, 862.8. Liu, H.X., Hokfelt, T., 2002. The participation of galanin in pain processing at the spinal level. Trends Pharmacol. Sci. 23, 468–474.
Malmberg, A.B., Yaksh, T.L., 1992. Hyperalgesia mediated by spinal glutamate or substance P receptor blocked by spinal cyclooxygenase inhibition. Science 257, 1276–1279. Mantyh, P.W., DeMaster, E., Malhotra, A., Ghilardi, J.R., Rogers, S.D., Mantyh, C.R., Liu, H., Basbaum, A.I., Vigna, S.R., Maggio, J.E., 1995. Receptor endocytosis and dendite reshaping in spinal neurons after somatosensory stimulation. Science 268, 1629–1632. Marvizon, J.C., Martinez, V., Grady, E.F., Bunnett, N.W., Mayer, E.A., 1997. Neurokinin 1 receptor internalization in spinal cord slices induced by dorsal root stimulation is mediated by NMDA receptors. J. Neurosci. 17, 8129–8136. OÕDonnell, D., Ahmad, S., Wahlestedt, C., Walker, P., 1999. Expression of the novel galanin receptor subtype GALR2 in the adult rat CNS: distinct distribution from GALR1. J. Comp. Neurol. 409, 469–481. Parker, E.M., Izzarelli, D.G., Nowak, H.P., Mahle, C.D., Iben, L.G., Wang, J., Goldstein, M.E., 1995. Cloning and characterization of the rat GALR1 galanin receptor from Rin14B insulinoma cells. Brain Res. Mol. Brain Res 34, 179–189. Parsons, R.L., Mulvaney, J.M., Merriam, L.A., 1998. Galanin activates an inwardly rectifying potassium conductance and inhibits a voltage-dependent calcium conductance in mudpuppy parasympathetic neurons. Ann. N.Y. Acad. Sci. 863, 156–169. Ren, J., Hu, H.Z., Starodub, A.M., Wood, J.D., 2001. Galanin suppresses calcium conductance and activates inwardly rectifying potassium channels in myenteric neurones from guinea-pig small intestine. Neurogastroenterol. Motil. 13, 247–254. Simmons, D.R., Spike, R.C., Todd, A.J., 1995. Galanin is contained in GABAergic neurons in the rat spinal dorsal horn. Neurosci. Lett. 187, 119–122. Waters, S.M., Krause, J.E., 2000. Distribution of galanin-1, -2 and -3 receptor messenger RNAs in central and peripheral rat tissues. Neuroscience 95, 265–271. Xu, X.J., Hokfelt, T., Bartfai, T., Wiesenfeld-Hallin, Z., 2000. Galanin and spinal nociceptive mechanisms: recent advances and therapeutic implications. Neuropeptides 34, 137–147. Yaksh, T.L., 1999. Spinal systems and pain processing: development of novel analgesic drugs with mechanistically defined models. Trends Pharmacol. Sci. 20, 329–337. Yaksh, T.L., Rudy, T.S., 1976. Chronic catheterization of the spinal subarachnoid space. Physiol. Behav. 17, 1031–1036. Yaksh, T.L., Hua, X.Y., Kalcheva, I., Nozaki-Taguchi, N., Marsala, M., 1999. The spinal biology in humans and animals of pain states generated by persistent small afferent input. Proc. Natl. Acad. Sci. USA 96, 7680–7686. Zhang, X., Nicholas, A.P., Hokfelt, T., 1993. Ultrastructural studies on peptides in the dorsal horn of the spinal cord – I. Co-existence of galanin with other peptides in primary afferents in normal rats. Neuroscience 57, 365–384. Yaksh, T.L., Dirig, D.M., Conway, C.M., Svensson, C., Luo, Z.D., Isakson, P.C., 2001. The acute antihyperalgesic action of nonsteroidal, anti-inflammatory drugs and release of spinal prostaglandin E2 is mediated by the inhibition of constitutive spinal cyclooxygenase-2 (COX-2) but not COX-1. J. Neurosci. 21, 5847–5853. Zhang, X., Ji, R.R., Nilsson, S., Villar, M., Ubink, R., Ju, G., Wiesenfeld-Hallin, Z., Hokfelt, T., 1995b. Neuropeptide Y and galanin binding sites in rat and monkey lumbar dorsal root ganglia and spinal cord and effect of peripheral axotomy. Eur. J. Neurosci. 7, 367–380. Zhang, Q., Shi, T.J., Ji, R.R., Zhang, Y.Z., Sundler, F., Hannibal, J., Fahrenkrug, J., Hokfelt, T., Zhang, Y., 1995a. Expression of pituitary adenylate cyclase-activating polypeptide in dorsal root ganglia following axotomy: time course and coexistence. Brain Res. 705, 149–158.