The neuropeptide galanin modulates behavioral and neurochemical signs of opiate withdrawal Venetia Zachariou*†, Darlene H. Brunzell*, Jessica Hawes*, Diann R. Stedman*, Tamas Bartfai‡, Robert A. Steiner§, David Wynick¶, U¨lo Langel储, and Marina R. Picciotto*,** *Department of Psychiatry, Yale University School of Medicine, New Haven, CT 06508; †Faculty of Medicine, Department of Pharmacology, University of Crete, 711 10 Crete, Greece; ‡Department of Neuropharmacology, The Scripps Research Institute, La Jolla, CA 92037; §Department of Physiology and Biophysics, University of Washington, Seattle, WA 98195; ¶Department of Medicine, Bristol University, Bristol BS2 8HW, United Kingdom; and 储Department of Neurochemistry and Neurotoxicology, Stockholm University, 106 91 Stockholm, Sweden Edited by Richard D. Palmiter, University of Washington School of Medicine, Seattle, WA, and approved June 2, 2003 (received for review May 23, 2003)
Much research has focused on pathways leading to opiate addiction. Pathways opposing addiction are more difficult to study but may be critical in developing interventions to combat drug dependence and withdrawal. Galanin decreases firing of locus coeruleus neurons, an effect hypothesized to decrease signs of opiate withdrawal. The current study addresses whether galanin affects morphine withdrawal signs by using a galanin agonist, galnon, that crosses the blood– brain barrier, and mice genetically engineered to under- or overexpress galanin peptide. Galnon significantly decreased morphine withdrawal signs in C57BL兾6 mice. Further, knockout mice lacking galanin showed exacerbated morphine withdrawal signs, suggesting that endogenous galanin normally counteracts opiate withdrawal. Transgenic mice overexpressing galanin in noradrenergic neurons also showed decreased morphine withdrawal signs, suggesting a possible neuroanatomical locus for these effects of galanin. Both c-fos immunoreactivity, a marker of neuronal activity, and phosphorylation of tyrosine hydroxylase at Ser-40, a marker of cAMP levels, are decreased in the locus coeruleus by galnon treatment after morphine withdrawal, suggesting a possible molecular mechanism for the behavioral effects of galanin. These studies suggest that galanin normally acts to counteract opiate withdrawal and that small molecule galanin agonists could be effective in diminishing the physical signs of withdrawal.
T
he neuropeptide galanin is widely expressed in brain (1) and decreases neurochemical and behavioral effects related to opiate reinforcement. Galanin is neither rewarding nor aversive, but the peptide attenuates morphine place preference (2). Galanin also inhibits mesolimbic dopamine neurotransmission (3, 4), an effect critical for rewarding properties of drugs of abuse (5). In addition, galanin inhibits the firing of locus coeruleus (LC) neurons (6), which may affect expression of opiate withdrawal signs (7). Further, stress increases galanin mRNA in the LC (8), and at least one galanin receptor subtype, GalR1, is up-regulated in the LC during opiate withdrawal (9). Thus, galanin may counteract opiate withdrawal through effects on the noradrenergic system. Galanin mediates its effects through activation of at least three receptor subtypes coupled to Gi, Gq, or Go (10, 11). These receptors show distinct patterns of distribution and activate different second messenger pathways, depending on cell type (11–14). The galanin receptors and peptide are expressed in many brain regions, and all are expressed in the LC (8, 15–17). Galnon, a galanin agonist, can cross the blood–brain barrier, has moderate affinity for GalRs (4.8 M), does not distinguish among GalR subtypes, and has significantly lower affinity when screened against a broad panel of nongalanin receptors (including opiate receptors; ref. 18). Thus, this compound can be considered a relatively selective galanin agonist. The current study was designed to determine whether galnon could affect morphine withdrawal signs. In addition, the role of endogenous galanin in modulating opiate withdrawal was studied in galanin knockout mice (19). Further, potential site(s) of action for 9028 –9033 兩 PNAS 兩 July 22, 2003 兩 vol. 100 兩 no. 15
galanin-mediated effects on morphine withdrawal were examined in transgenic mice that overexpress the peptide under the control of the dopamine--hydroxylase (DH) promoter, which normally drives gene expression in noradrenergic neurons (20), and in studies of c-fos and tyrosine hydroxylase (TH) phosphorylation in the LC of wild-type mice administered galnon. Methods Animals. Galanin knockout mice. Experimental animals were mice
homozygous for a targeted mutation of the galanin gene on an inbred 129OlaHsd background (F18; Gal⫺兾⫺; ref. 19), and controls were 129OlaHsd wild-type mice (Gal⫹兾⫹). Galanin-overexpressing mice. Mice overexpressing galanin under the control of DH promoter (20) were backcrossed onto the C57BL兾6 background for seven generations. Experimental animals were generated by crossing mice heterozygous for the transgene with C57BL兾6 wild-type mice to generate littermate controls of which 50% were positive for the transgene (Gal-tg) and 50% were negative for the transgene (wild type). Pharmacological studies. Galnon was synthesized as has been described (18). C57BL兾6 mice (The Jackson Laboratory) or 129OlaHsd mice were used for pharmacological studies. For all studies, mice were used at 8–12 weeks of age and were age- and sex-matched. Opiate Withdrawal. Mice were injected i.p. with escalating morphine doses every 8 h (20, 40, 60, 80, 100, 100, and 100 mg兾kg). Two hours after the last morphine injection, mice were injected with naloxone (1 mg兾kg s.c.), and withdrawal symptoms (jumping, wet-dog shakes, backward walking, paw tremor, tremor, diarrhea, ptosis, and weight loss) were monitored for 30 min after naloxone administration. In addition to measuring individual withdrawal signs, an overall opiate withdrawal score was calculated as (no. of backward walking steps ⫻ 0.1) ⫹ (diarrhea ⫻ 2) ⫹ (no. of jumps ⫻ 0.1) ⫹ (paw tremor ⫻ 0.1) ⫹ (ptosis) ⫹ (tremor) ⫹ (% weight loss ⫻ 5) ⫹ (no. of wet-dog shakes) (modified from ref. 21). Morphine Analgesia and Tolerance Tests. For morphine tolerance studies, mice were placed on a 52°C hot plate apparatus, and latency to paw lick was measured. A cutoff time of 30 s was set to avoid tissue damage. Mice were tested once daily for 3–5 days. Data were calculated as percent maximal possible effect [MPE ⫽ (test latency ⫺ baseline latency)兾(cutoff ⫺ baseline latency)]. Morphine (10 mg兾kg) was administered s.c. 30 min before
This paper was submitted directly (Track II) to the PNAS office. Abbreviations: DH, dopamine--hydroxylase; LC, locus coeruleus; P, phosphorylated; TH, tyrosine hydroxylase. **To whom correspondence should be addressed at: Department of Psychiatry, Yale University School of Medicine, 34 Park Street, New Haven, CT 06508. E-mail: marina.
[email protected].
www.pnas.org兾cgi兾doi兾10.1073兾pnas.1533224100
c-fos Immunohistochemistry and TH Western Blotting. Four groups of
mice were used for c-fos experiments (n ⫽ 6 per group) and TH Western blotting experiments (n ⫽ 5–7 per group). The Sal–Veh group received chronic saline treatment as described for morphine withdrawal studies and an acute vehicle injection, and 15 min later received a 1 mg兾kg injection of naloxone. The Sal–Gal group received chronic saline treatment and an acute galnon injection (2 mg兾kg), and 15 min later received a 1 mg兾kg injection of naloxone. The Mor–Veh group received chronic morphine treatment and a vehicle injection, and 15 min later received a 1 mg兾kg injection of naloxone. The Mor–Gal group received chronic morphine treatment and a galnon injection (2 mg兾kg), and 15 min later received a 1 mg兾kg injection of naloxone. c-fos immunohistochemistry was performed essentially as has been described (22). Briefly, 2 h after naloxone injection, mice were injected with chloral hydrate and perfused with 0.1 M PBS for 5 min, followed by a 15-min perfusion with 4% paraformaldehyde. Brains were removed, postfixed for 24 h, and cryoprotected in 20% sucrose. Forty-micrometer coronal sections were cut at the level of the LC and periacqueductal gray (PAG) by using a Leica microtome (approximately at Bregma level ⫺5.52). Based on anatomical markers, including ventricle shape and presence of the VII cranial nerve tract, two to four sections per brain were selected for c-fos immunocytochemistry. Sections were washed in 0.1 M PBS, blocked with 5% normal goat serum, and incubated overnight with anti-c-fos antiserum (1:5,000, Santa Cruz Biotechnology) at 4°C for 48 h. After three washes in 0.1 M PBS, sections were incubated with biotinylated goat anti-rabbit Ig (1:200), washed three times in 0.1 M PBS, and incubated with biotin-streptavidin-peroxidase complex (1:200), and the peroxidase reaction product was visualized by 3,3⬘diaminobenzidine ( Vector Laboratories). Sections were mounted on electrostatic Superfrost glass slides, dehydrated, and coverslipped with prolonged antifade medium (Molecular Probes). Using NIH IMAGE software (http:兾兾rsb.info.nih.gov兾 nih-image), a defined box was placed over the center of the LC on each side of the brain (see Fig. 4) for each section, and all c-fos-positive nuclei were counted within the box. The number of positive nuclei reported represents the average of the right and left side from two sections for each mouse. All slides were quantitated at the same time under the same conditions. For TH and phosphorylated (P) TH Western blotting studies, mice were killed by rapid decapitation 15 min after naloxone administration, and 1-mm punches of the LC [close to Bregma ⫺5.52 mm (23)] were taken and immediately frozen in liquid nitrogen (24). Groups were identical to those described above for c-fos studies [Sal–Veh (n ⫽ 5), Sal–Gal (n ⫽ 5), Mor–Veh (n ⫽ 5), and Mor–Gal (n ⫽ 7)]. LC punches were homogenized in lysis 1 reagent (RPA538, Amersham Biosciences), and protein concentrations were determined (Bio-Rad Lowry reagents). Five micrograms of protein for each sample were separated on SDS兾8.5% PAGE gels and transferred to nitrocellulose membranes. For quantitation, 10 g of liver extract and 0.0125–0.1 ng of Ser-40P-TH PC12 standard (generously provided by J. Haycock, Louisiana State University, Baton Rouge) were also loaded. Each blot was cut at the 77-kDa protein marker and blocked in 5% BSA. The bottom portion of each blot was incubated overnight at 4°C in 1:200 Ser-40P-TH antibody (25), and the top portion was incubated in 1:250 PYK2 antibody (Upstate Biotechnology, Lake Placid, NY). After washing with Tris-buffered saline兾0.1% Triton X-100, blots were incubated with peroxidase-labeled 1:2,000 anti-rabbit IgG (Vector Laboratories) for 1 h at room temperature. Bands were visualized by using enhanced chemiluminescence and exposure to Kodak Zachariou et al.
Biomax MR film. The bottom portion of each blot was subsequently stripped with 0.2 M NaOH and reprobed with 1:2,000 TH antibody (Chemicon), followed by secondary 1:2,000 antirabbit IgG antibody. All bands were quantified by using NIH IMAGE. Protein loading was verified by using Ponceau staining. Band intensities were corrected for background and normalized to PYK2 levels, which had previously been shown to remain constant across treatment groups. Band intensities from the liver sample were used as an internal control for comparison between individual blots. PC12 cell extracts at known concentrations were used to generate a concentration curve and calculate the total nanograms of TH, as well as nanograms of TH phosphorylated on Ser-40 (P-Ser-40), for each LC sample. Using these standard curves, the percentage of TH phosphorylated on Ser-40 was determined ([P-Ser-40]兾[total TH] ⫻ 100). Significance was determined by using repeated-measures ANOVA and the least significant difference post hoc test (P ⬍ 0.05). Results The small molecule galanin agonist galnon has been shown to act as an anticonvulsant (18), but its effects on opiate-related behaviors have not been tested. Systemically administered galnon (2 mg兾kg i.p.) significantly attenuated several opiate withdrawal signs in C57BL兾6 mice (Fig. 1A). Backward walking, chewing, diarrhea, jumping, paw tremor, and weight loss were all significantly reduced after galnon administration. Consistent with these changes, the overall withdrawal score was significantly reduced (P ⬍ 0.01) in galnon-treated vs. vehicle-treated mice (Fig. 1B). Peripherally administered galnon had no effect on pain threshold (Fig. 1C) or morphine-induced analgesia and tolerance (Fig. 1D) at the dose used. Galnon experiments suggested that pharmacological activation of galanin receptors can decrease opiate withdrawal signs, but could not test whether activation of galanin receptors by endogenous galanin normally plays a role in protection against opiate dependence and withdrawal. Knockout mice lacking the galanin gene on an inbred 129OlaHsd genetic background were tested for severity of morphine withdrawal. Gal⫺兾⫺ mice showed significantly more withdrawal signs than Gal⫹兾⫹ mice (Fig. 2A). Backward walking, diarrhea, rearing, weight loss, and wet-dog shakes were all significantly increased in Gal⫺兾⫺ mice. In addition, overall withdrawal scores were significantly different (P ⬍ 0.01) between Gal⫹兾⫹ and Gal⫺兾⫺ mice (Fig. 2B). It is important to note that the 129OlaHsd strain shows much less jumping behavior than the C57BL兾6 strain during opiate withdrawal [the 129OlaHsd mice showed ⬍3 jumps over the course of the experiment (Fig. 2 A), whereas C57BL兾6 mice showed ⬎120 jumps (Fig. 1 A)]. A higher dose of naloxone (4 mg兾kg) results in more severe withdrawal signs in 129OlaHsd mice, but does not result in increased jumping behavior (data not shown). Thus, the lack of jumping response in this experiment is likely to be a strain-related phenomenon. To determine whether the altered morphine withdrawal seen in Gal⫺兾⫺ mice is due, at least in part, to acute effects during the expression of withdrawal signs, rather than involvement of galanin during chronic morphine administration or adaptation to effects of galanin during development, Gal⫺兾⫺ mice were treated acutely with galnon (2 mg兾kg) or vehicle 15 min before naloxone injection (1 mg兾kg). Galnon administration reversed several signs of opiate withdrawal in Gal⫺兾⫺ mice to wild-type levels, including backward walking, diarrhea, rearing, tremor, weight loss, and wet-dog shakes (Fig. 2 A), and reduced the overall withdrawal score (Fig. 2B). Paw tremor was reduced below the level seen in Gal⫹兾⫹ mice, suggesting that this sign might be particularly sensitive to exogenous galnon administration. Withdrawal signs not affected by galnon treatment in galanin knockout mice were also not strongly expressed in Gal⫹兾⫹ mice. PNAS 兩 July 22, 2003 兩 vol. 100 兩 no. 15 兩 9029
NEUROSCIENCE
testing. Galnon (or vehicle) was administered i.p. (2 mg兾kg) 15 min before morphine administration.
Fig. 1. Galnon attenuates morphine withdrawal. (A) Mice were administered increasing doses of morphine every 8 h for 3 days (20, 40, 60, 80, 100, 100, and 100 mg兾kg s.c.), and withdrawal was precipitated 2 h after the last morphine dose by using naloxone (1 mg兾kg s.c.). One group received naloxone alone (n ⫽ 8), and a second group received 2 mg兾kg galnon 15 min before naloxone injection (n ⫽ 7). After naloxone administration, withdrawal signs were scored for 30 min by an observer blind to treatment. For each withdrawal sign, comparisons were made between galnon- and vehicle-treated mice by using Student’s t test. (B) The overall withdrawal score was significantly decreased by galnon administration (P ⬍ 0.01). (C) Galnon (2 mg兾kg) alone did not alter pain threshold as measured on the 52°C hotplate. (D) No differences in initial analgesic response to morphine or the development of morphine tolerance was seen between galnon-treated and vehicle-treated control mice. *, P ⬍ 0.05; **, P ⬍ 0.01; ***, P ⬍ 0.001.
Gal⫺兾⫺ mice have previously been shown to have alterations in neuropathic pain, survival of cholinergic neurons, and overall excitability (22, 26–28). Although a small, but significant, de-
crease in nociceptive threshold was seen in Gal⫺兾⫺ mice (Fig. 2C), morphine analgesia and tolerance to morphine were not significantly different between Gal⫺兾⫺ and Gal⫹兾⫹ mice (Fig.
Fig. 2. Knockout of the galanin gene exacerbates morphine withdrawal, and this is reversed by treatment with galnon. (A) Gal⫺兾⫺ mice on the 129OlaHsd background (n ⫽ 8) and Gal⫹兾⫹ mice (n ⫽ 8) were administered increasing doses of morphine every 8 h for 3 days (20, 40, 60, 80, 100, 100, and 100 mg兾kg s.c.), and withdrawal was precipitated 2 h after the last morphine dose by using naloxone (1 mg兾kg s.c.). After naloxone administration, withdrawal signs were scored for 30 min by an observer blind to treatment. An independent group of Gal⫺兾⫺ mice was treated with 2 mg兾kg galnon 15 min before naloxone injection (n ⫽ 6). One-way ANOVA was used to compare differences in withdrawal signs between Gal⫹兾⫹ mice, Gal⫺兾⫺ mice, and Gal⫺兾⫺ mice treated with galnon. (B) A significant effect of genotype on overall withdrawal scores was identified by one-way ANOVA. Using the least significant difference post hoc test, overall withdrawal scores were significantly different between Gal⫹兾⫹ and Gal⫺兾⫺ mice (P ⬍ 0.01), and between Gal⫺兾⫺ mice treated with vehicle or galnon (P ⬍ 0.01). (C) Pain thresholds as measured on the 52°C hotplate were slightly decreased in Gal⫺兾⫺ mice as compared with Gal⫹兾⫹ mice. (D) No differences in initial analgesic response to morphine or the development of morphine tolerance was seen between Gal⫹兾⫹ and Gal⫺兾⫺ mice. *, P ⬍ 0.05; *, P ⬍ 0.01; ***, P ⬍ 0.001. 9030 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.1533224100
Zachariou et al.
2D). Whereas C57BL兾6 mice showed complete tolerance to morphine analgesia by 3 days of treatment (Fig. 1D), 129OlaHsd mice required treatment for 5 days to achieve full morphine tolerance (Fig. 2D). One method of identifying the site of action for galnon or endogenous galanin in attenuating opiate withdrawal is to determine whether localized overexpression of galanin can mimic the effect of peripherally administered galnon. Transgenic mice expressing the galanin cDNA under the control of a DH promoter that targets expression to noradrenergic neurons (20) were examined for opiate withdrawal signs. Significant decreases were seen in backward walking, chewing, diarrhea, jumping, tremor, and wet-dog shakes in Gal-tg mice as compared with their wild-type siblings (Fig. 3A), as well as in overall withdrawal scores (Fig. 3B). These data suggest that galanin effects on opiate withdrawal signs may be mediated through actions on the noradrenergic (or potentially the adrenergic) system and兾or targets of noradrenergic innervation in the brain and periphery. Pain thresholds of galanin transgenic mice were similar to those of their wild-type littermates in the hotplate test (Fig. 3C), and no differences were observed for responses to morphine or morphine tolerance (Fig. 3D) between Gal-tg mice and their wild-type siblings. Based on the transgenic mouse studies and previous physiological studies (6), a possible mechanism underlying galnon antagonism of opiate withdrawal signs could be decreased firing of noradrenergic neurons. We used c-fos as a marker of neuronal activation and compared c-fos immunoreactivity in the LC of wild-type mice treated with galnon or vehicle before naloxoneprecipitated withdrawal (Fig. 4 A and B). A decrease in c-foslabeled LC neurons was observed in mice treated with galnon alone in the absence of opiate withdrawal, but this did not reach significance. There was a significant increase in c-fos immunoreactivity in the LC after morphine withdrawal (P ⬍ 0.0004), and this increase was significantly attenuated in galnon-treated mice as compared with vehicle-treated mice (P ⬍ 0.05). Zachariou et al.
A likely mechanism underlying the decreased firing rate of LC neurons reflected in c-fos measurements is a decrease in cAMP levels as a result of activation of GalRs. It has previously been shown that Ser-40 in TH is specifically phosphorylated by cAMP-dependent protein kinase (PKA) (25). Thus, TH Ser-40 phosphorylation can be used as a measure of cellular cAMP levels in noradrenergic neurons. Further, because only noradrenergic neurons in this brain region express TH, the examination of the phosphorylation state of TH is a sensitive assay for changes in cAMP in this subclass of neurons (24). Morphine withdrawal significantly increased the stoichiometry of Ser-40 phosphorylation (P ⬍ 0.001; Fig. 4C). In contrast, galnon treatment decreased the phosphorylation of Ser-40 back to baseline levels (P ⬍ 0.05, galnon vs. vehicle). Total TH levels appeared to increase as a result of morphine withdrawal, but it was not possible to determine whether this was a true increase or reflected residual IgG bound to P-Ser-40 (blots for total TH were performed after stripping after P-Ser-40 blotting to allow measurement of total TH and P-TH from the same LC samples). Because ratios of P-Ser-40 to total TH were used for all measurements, the increase in stoichiometry after morphine withdrawal may represent an underestimate. Discussion The circuitry underlying the development of addiction and tolerance has been widely explored (for review, see refs. 29 and 30). In contrast, endogenous pathways that oppose addictive processes have been more difficult to study but may be critical in developing pharmacological interventions to combat opiate dependence and withdrawal. Using pharmacological and genetic models, we have shown that the galanin system modulates morphine withdrawal. When galanin is not expressed, as in Gal⫺兾⫺ mice, opiate withdrawal signs are more severe. In contrast, decreased opiate withdrawal is seen when galanin is overexpressed under the control of the DH promoter in transgenic mice. Systemic administration of a galanin receptor agonist also alleviates opiate withdrawal symptoms, suggesting PNAS 兩 July 22, 2003 兩 vol. 100 兩 no. 15 兩 9031
NEUROSCIENCE
Fig. 3. Galanin overexpression under control of the DH promoter can attenuate some signs of morphine withdrawal. (A) Transgenic mice overexpressing galanin under control of the DH promoter (n ⫽ 10) and their wild-type siblings (n ⫽ 9) were administered increasing doses of morphine every 8 h for 3 days (20, 40, 60, 80, 100, 100, and 100 mg兾kg s.c.), and withdrawal was precipitated 2 h after the last morphine dose by using naloxone (1 mg兾kg s.c.). After naloxone administration, withdrawal signs were scored for 30 min by an observer blind to treatment. One-way ANOVA was used to compare differences in withdrawal signs between genotype. (B) The overall withdrawal score was significantly reduced in Gal-tg mice as compared with their wild-type siblings (P ⬍ 0.01). (C) No difference in pain threshold was seen between Gal-tg and their wild-type siblings as measured on the 52°C hotplate. (D) No differences in initial analgesic response to morphine or the development of morphine tolerance was seen between Gal-tg and wild-type mice. tg, Gal-tg mice; wt, wild-type siblings; *, P ⬍ 0.05; **, P ⬍ 0.01; ***, P ⬍ 0.001.
Fig. 4. c-fos immunoreactivity and TH phosphorylation in the LC after galnon administration after naloxone-stimulated withdrawal. Mice treated with morphine and naloxone showed significantly more c-fos immunoreactivity and phosphorylation of TH Ser-40 in the LC than mice treated with naloxone alone. Galnon significantly attenuated the induction of c-fos immunoreactivity, as well as TH phosphorylation, in morphine- and naloxone-treated mice. A decrease in c-fos immunoreactivity was seen in the LC of naloxone-treated mice after galnon (2 mg兾kg) administration, but this did not reach significance. (A) Representative photomicrographs of c-fos staining at the level of the LC. A section through the LC after morphine withdrawal is shown at left to identify the areas of c-fos immunoreactivity quantified for all treatment groups. The boxed region shown at higher magnification for each treatment group represents the area counted for each section. Counts were averaged from left and right LC across two sections per mouse matched for anatomical markers (23). (B) Quantitation of c-fos immunoreactivity. There was a significant effect of naloxone-precipitated withdrawal on c-fos immunoreactivity in the LC (ANOVA, P ⬍ 0.0005). *, P ⬍ 0.05 galnon vs. vehicle. (C) Stoichiometry of TH phosphorylation at Ser-40. Representative Western blots are shown for Ser-40-P, total TH, and pyk2 (as a control for protein loading). There was a significant increase in Ser-40-P levels and stoichiometry in mice treated with morphine and naloxone that was significantly attenuated by galnon administration. S-V, saline–vehicle group; S-G, saline– galnon group; M-V, morphine–vehicle group; M-G, morphine– gainon group. *, P ⬍ 0.05 galnon vs. vehicle.
that the galanin system could be a target for the development of agents that reduce opiate withdrawal. The present study provides evidence that modulation of opiate withdrawal by galanin is mediated, at least in part, through effects in noradrenergic neurons. It is also possible that galanin and galanin receptors localized beyond the noradrenergic system could be involved in modulation of morphine dependence as well. Gal-tg mice overexpress galanin in the noradrenergic nuclei, the piriform and entorhinal cortices, and the subiculum (26), and also have increased galanin expression in the pituitary and adrenal glands (J. G. Hohmann and R.S., unpublished data). Thus, actions of galnon outside the noradrenergic system may contribute to its effects on opiate withdrawal signs. Data obtained by using galnon to alleviate opiate withdrawal signs suggest that the galanin system is critical during the expression of withdrawal signs. One concern in studies of mice with constitutive knockout or overexpression of a gene of interest is that phenotypic differences reflect a role for the protein during development and not during adulthood. For example, galanin knockout mice show deficits in development of the dorsal root ganglia (27) and basal forebrain (28). The effects seen here on opiate withdrawal are less likely to be due to an irreversible developmental defect, however. First, galnon can be administered to adult mice, and its effect is similar to the decrease in opiate withdrawal signs seen in Gal-tg mice. Further, the increased withdrawal signs seen in Gal⫺兾⫺ mice can be reversed in adulthood by administration of galnon. These pharmacological data suggest that galanin plays a role acutely in the adult mouse in the expression of opiate withdrawal signs. However, the fact that galnon did not reduce withdrawal signs other than paw tremor in Gal⫺兾⫺ mice below those of untreated Gal⫹兾⫹ mice could suggest that a higher dose of galnon might be necessary to see a baseline effect on opiate withdrawal signs in mice on the 129OlaHsd background. A previous study infused a single dose of the galanin peptide into the cerebral ventricle of rats and did not observe an effect on opiate withdrawal (31). This discrepancy suggests that local administration of galanin might not be sufficient to counteract the broad spectrum of opiate withdrawal symptoms because of limited diffusion, rapid proteolysis, or the necessity of galanin action at sites outside the brain. In the current study, galnon was administered at a high concentration and was administered peripherally. Galnon also has a longer duration of action than 9032 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.1533224100
galanin, because the peptide can be rapidly proteolyzed. In addition, mice that received galnon were not cannulated, whereas cannulated mice may have experienced stress as a result of the surgical procedure that could modulate the effects of galanin. We cannot rule out that peripheral actions of galnon are responsible for its behavioral effects, although this is unlikely given that many withdrawal signs have been shown to depend on central activity (32) and that peripherally administered galanin had no effect on withdrawal signs in the current study (data not shown). Galanin receptors are modulated during opiate dependence, because both galanin receptor levels, as measured by radioligand binding, and GalR1 mRNA levels are increased in the LC after opiate withdrawal (9). In contrast, galanin peptide mRNA levels do not change during withdrawal (31). This finding suggests that levels of galanin receptors, and their regulation, are a critical factor in determining the actions of the endogenous galanin system on opiate withdrawal. Galanin (33) and GalR1 (34) expression are also regulated after activation of the cAMP pathway. Up-regulation of the cAMP pathway results in the increased neuronal activity observed in the LC during withdrawal (35–37) and in adaptive changes in expression of a variety of genes in several brain regions (38). Because GalR1 is negatively coupled to cAMP generation (12), and this effect appears to be responsible for the ability of galanin to decrease firing rate in the LC (39), up-regulation of the particular receptors might indicate a compensatory mechanism for the increased neuronal activity during opiate withdrawal. We have shown that peripheral galnon administration attenuates the increase in LC cAMP due to morphine withdrawal. Thus, it is likely that the action of GalRs on adenylyl cyclase activity contributes to the effect of galnon on behavioral signs of opiate withdrawal. Although GalR1 is a likely candidate for mediating the effects of galanin on LC firing, both GalR1 and GalR2 are found in areas involved in the development of morphine dependence, including the LC, the periaqueductal gray, the nucleus accumbens, and the spinal cord. Recent studies have also reported that the GalR3 subtype is expressed in the LC (16). GalR3 is also negatively coupled to adenylyl cyclase and therefore might play a role in attenuating behaviors related to opiate withdrawal. Galanin affects behavioral signs of morphine dependence, but tolerance to morphine analgesia does not appear to be modulated by the peptide. Pain thresholds were normal in galaninoverexpressing mice and in mice treated with galnon. Consistent Zachariou et al.
with previous observations (22), galanin knockout mice had lower pain thresholds for thermal stimuli. In addition, transgenic overexpression of galanin in the dorsal root ganglia attenuates allodynia (40). However, neither pharmacological nor genetic manipulation of the galanin system had an effect on the development of opiate tolerance in mice in the current study. Although galnon did not have an effect on the development of opiate analgesia or tolerance at the dose used here, we cannot exclude the possibility that this compound might affect nociceptive transmission when applied in higher doses or when used in different pain models. A number of studies have suggested that galanin agonists could have pharmacological value as analgesics because intrathecal galanin administration potentiates morphine analgesia (predominantly in models of chronic pain), and thus, lower morphine doses could be coadministered with galanin to produce an analgesic effect (41–45). In addition, galanin antago-
We thank Dr. John W. Haycock for antiserum against TH Ser-40 and Christina Lee for technical assistance. This work was supported by National Institutes of Health Grants DA15425 and DA00436 and the ¨ .L. is supported by the SwedChristiane Brooks Johnson Foundation. U ish Medical Research Council.
1. Merchenthaler, I., Lopez, F. J. & Negro-Vilar, A. (1993) Prog. Neurobiol. 40, 711–769. 2. Zachariou, V., Parikh, K. & Picciotto, M. R. (1999) Brain Res. 831, 33–42. 3. Tsuda, K., Tsuda, S., Nishio, I., Masuyama, Y. & Goldstein, M. (1998) Am. J. Hypertens. 11, 1475–1479. 4. Ericson, E. & Ahlenius, S. (1999) Brain Res. 822, 200–209. 5. Koob, G. F. (2000) Ann. N.Y. Acad. Sci. 909, 170–185. 6. Sevcik, J., Finta, E. P. & Illes, P. (1993) Eur. J. Pharmacol. 230, 223–230. 7. Rasmussen, K., Beitner-Johnson, D. B., Krystal, J. H., Aghajanian, G. K. & Nestler, E. J. (1990) J. Neurosci. 10, 2308–2317. 8. Holmes, P. V., Blanchard, D. C., Blanchard, R. J., Brady, L. S. & Crawley, J. N. (1995) Pharmacol. Biochem. Behav. 50, 655–660. 9. Zachariou, V., Thome, J., Parikh, K. & Picciotto, M. R. (2000) Neuropsychopharmacology 23, 127–137. ¨ ., Bartfai, T. & Wiesenfeld-Hallin, Z. (1992) Acta Physiol. 10. Bedecs, K., Langel, U Scand. 144, 213–220. 11. Branchek, T. A., Smith, K. E., Gerald, C. & Walker, M. W. (2000) Trends Pharmacol. Sci. 21, 109–117. 12. Wang, S., Hashemi, T., Fried, S., Clemmons, A. L. & Hawes, B. E. (1998) Biochemistry 37, 6711–6717. 13. Karelson, E., Laasik, J. & Sillard, R. (1995) Neuropeptides 28, 21–28. 14. Chen, Y., Couvineau, A., Laburthe, M. & Amiranoff, B. (1992) Biochemistry 31, 2415–2422. 15. Burazin, T. C. D., Larm, J. A., Ryan, M. C. & Gundlach, A. L. (2000) Eur. J. Neurosci. 12, 2901–2917. 16. Mennicken, F., Hoffert, C., Pelletier, M., Ahmad, S. & O’Donnell, D. (2002) J. Chem. Neuroanat. 24, 257–268. 17. O’Donnell, D., Ahmad, S., Wahlestedt, C. & Walker, P. (1999) J. Comp. Neurol. 409, 469–481. 18. Saar, K., Mazarati, A. M., Mahlapuu, R., Hallnemo, G., Soomets, U., Kilk, K., Hellberg, S., Pooga, M., Tolf, B.-R., Shi, T. S., et al. (2002) Proc. Natl. Acad. Sci. USA 99, 7136–7141. 19. Wynick, D., Small, C. J., Bacon, A., Holmes, F. E., Norman, M., Ormandy, C. J., Kilic, E., Kerr, N. C., Ghatei, M., Talamantes, F., et al. (1998) Proc. Natl. Acad. Sci. USA 95, 12671–12676. 20. Steiner, R. A., Hohmann, J. G., Holmes, A., Wrenn, C. C., Cadd, G., Jure´us, A., Clifton, D. K., Luo, M. L., Gutshall, M., Ma, S. Y., et al. (2001) Proc. Natl. Acad. Sci. USA 98, 4184–4189. 21. Maldonado, R., Blendy, J. A., Tzavara, E., Gass, P., Roques, B. P., Hanoune, J. & Schutz, G. (1996) Science 273, 657–659. 22. Kerr, B. J., Cafferty, W. B., Gupta, Y. K., Bacon, A., Wynick, D., McMahon, S. B. & Thompson, S. W. (2000) Eur. J. Neurosci. 12, 793–802.
23. Franklin, K. B. J. & Paxinos, G. (1997) The Mouse Brain in Stereotaxic Coordinates (Academic, San Diego). 24. Punch, L. J., Self, D. W., Nestler, E. J. & Taylor, J. R. (1997) J. Neurosci. 17, 8520–8527. 25. Salvatore, M. F., Garcia-Espana, A., Goldstein, M., Deutch, A. Y. & Haycock, J. W. (2000) J. Neurochem. 75, 225–232. 26. Mazarati, A. M., Hohmann, J. G., Bacon, A., Liu, H., Sankar, R., Steiner, R. A., Wynick, D. & Wasterlain, C. G. (2000) J. Neurosci. 20, 6276–6281. 27. Holmes, F. E., Mahoney, S., King, V. R., Bacon, A., Kerr, N. C. H., Pachnis, V., Curtis, R., Priestley, J. V. & Wynick, D. (2000) Proc. Natl. Acad. Sci. USA 97, 11563–11568. 28. O’Meara, G., Coumis, U., Ma, S. Y., Kehr, J., Mahoney, S., Bacon, A., Allen, S. J., Holmes, F., Kahl, U., Wang, F. H., et al. (2000) Proc. Natl. Acad. Sci. USA 97, 11569–11574. 29. Koob, G. F. (1992) Trends Pharmacol. Sci. 13, 177–184. 30. Koob, G. F., Maldonado, R. & Stinus, L. (1992) Trends Neurosci. 15, 186–191. 31. Holmes, P. V., de Bartolomeis, A., Koprivica, V. & Crawley, J. N. (1995) Synapse 19, 197–205. 32. Maldonado, R. & Koob, G. F. (1993) Brain Res. 605, 128–138. 33. Corness, J. D., Burbach, J. P. H. & Ho ¨kfelt, T. (1997) Mol Brain Res. 47, 11–23. 34. Zachariou, V., Georgescu, D., Kansal, L., Merriam, P. & Picciotto, M. R. (2001) J. Neurochem. 76, 191–200. 35. Aghajanian, G. K. (1978) Nature 276, 186–188. 36. Duman, R. S., Tallman, J. F. & Nestler, E. J. (1988) J. Pharmacol. Exp. Ther. 246, 1033–1039. 37. Nestler, E. J. (1992) J. Neurosci. 12, 2439–2450. 38. Nestler, E. J., Hope, B. T. & Widnell, K. L. (1993) Neuron 11, 995–1006. 39. Nishibori, M., Oishi, R., Itoh, Y. & Saeki, K. (1988) J. Neurochem. 51, 1953–1955. 40. Holmes, F. E., Bacon, A., Pope, R. J., Vanderplank, P. A., Kerr, N. C., Sukumaran, M., Pachnis, V. & Wynick, D. (2003) Proc. Natl. Acad. Sci. USA 100, 6180–6185. 41. Przewlocka, B., Machelska, H., Rekowski, P., Kupryszewski, G. & Przewlocki, R. (1995) J. Neural Transm. 102, 229–235. 42. Wiesenfeld-Hallin, Z., Xu, X. J., Villar, M. J. & Ho ¨kfelt, T. (1990) Neurosci. Lett. 109, 217–221. 43. Wiesenfeld-Hallin, Z., Xu, X. J., Hughes, J., Horwell, D. C. & Ho ¨kfelt, T. (1991) Neuropeptides 19, Suppl., 79–84. 44. Zhang, Y. P., Lundeberg, T. & Yu, L. C. (2000) Brain Res. 852, 485–487. 45. Wang, D., Lundeberg, T. & Yu, L. C. (2000) Neuroscience 96, 767–771.
Zachariou et al.
PNAS 兩 July 22, 2003 兩 vol. 100 兩 no. 15 兩 9033
NEUROSCIENCE
nizes morphine place preference (2). We show here that a systemically administered galanin agonist also attenuates opiate withdrawal. In addition, we present evidence that the underlying mechanism for these behavioral effects could be decreased cAMP signaling in the LC (although effects on other brain areas may also contribute). Together, these findings suggest that galanin agonists would be ideal analgesics because they would potentiate morphine analgesia and decrease morphine abuse potential. Further, the current results indicate that galanin agonists could also have pharmacological value for the alleviation of opiate withdrawal in dependent individuals.
