Spinal NMDA receptor — nitric oxide mediation of the expression of morphine withdrawal symptoms in the rat

BRAIN RESEARCH ELSEVIER

Brain Research 679 (1995) 189-199

Research report

Spinal NMDA receptor nitric oxide mediation of the expression of morphine withdrawal symptoms in the rat Jerry J. Buccafusco *, Alvin V. Terry, Jr., Laura Shuster Department of VeteransAffairs Medical Center, Augusta, GA 30910, USA; Department of Pharmacology and Toxicology and Department Of" Psychiatry and Health Behavior, Medical College of Georgia, Augusta, GA 30912-2300, USA Accepted 31 January 1995

Abstract

Previous studies in this laboratory have demonstrated that cholinergic receptors within the spinal cord play an important role in the expression of naloxone-precipitated withdrawal symptoms in the morphine-dependent rat. Related cardiovascular studies in non-dependent animals have demonstrated that this spinal cholinergic system is linked to a glutamatergic, NMDA pressor pathway which also involves the participation of a nitric oxide (NO) generating system. The purpose of this study was to determine whether spinal NMDA receptors and/or NO are involved in the expression of morphine withdrawal symptoms. Rats bearing previously implanted intrathecal (IT) catheters were dependent on morphine following chronic i.a. infusion of increasing doses over 5 days. Naloxone (0.5 mg/kg) was administered via the i.a. line to precipitate withdrawal; and both cardiovascular and behavioral symptoms were recorded over 60 rain. Pretreatment 20 min before naloxone with IT injection of either of the NMDA receptor antagonists, MK-801 or AP-7 (100-200 nmol), produced a significant reduction in the expression of both the cardiovascular and behavioral symptoms of up to about 60%. IT pretreatment with the NO synthase inhibitor L-NAME - - a methyl ester derivative of L-arginine, also produced a dose-dependent, L-arginine reversible inhibition of the cardiovascular (mainly the pressor) component of withdrawal, but had no significant effect on the expression of behavioral signs. In contrast, IT pretreatment with L-NOARG and L-NMMA, non-ester analogs of L-arginine, significantly inhibited the expression of thc behavioral signs of withdrawal but did not alter the pressor component. A combined pretreatment with L-NAME and L-NOARG resulted in suppression of both pressor and behavioral components of withdrawal. The anti-withdrawal actions of either class of NO synthase inhibitor could not be attributed to blockade of local muscarinic receptors. These findings are consistent with a role for both spinal NMDA receptors and a NO generating system in the expression of both the behavioral and autonomic components of naloxone-precipitated withdrawal. They also suggest that different structural analogs of L-arginine have different profiles of activity in this regard - - opening the possibility that different isozymes of NO synthase located within the same spinal region mediate different physiological or behavioral functions.

Keywords: Morphine withdrawal; Naloxone; NMDA receptor; Nitric oxide; Spinal cord; Blood pressure; Heart rate; Behavior

I. Introduction

The naloxone-precipitated increase in cardiovascular and behavioral responses in morphine-dependent rats has been employed as a model for investigating the withdrawal phenomenon and development of physical dependence [8,14,18,19,33]. Withdrawal-associated behaviors and sympathetic activity are enhanced following administration of naloxone to various levels

* Corresponding author. Department of Pharmacology and Toxicology. Fax: (1) (706) 721-2347. 0006-8993/95/$09.50 © 1995 Elsevier Science B.V. All rights reserved SSDI 0 0 0 6 - 8 9 9 3 ( 9 5 ) 0 0 2 0 3 - 0

of the supraspinal and spinal neuraxis [3,31,34,35]. The locus coeruleus has been implicated in mediating the symptoms of abstinence, since neuronal cell firing is enhanced during withdrawal [41] and, since the antiwithdrawal agent, clonidine, suppresses withdrawal-induced neuronal activity in this brain site [53]. Also, glutamatergic receptor antagonists suppress withdrawal-induced firing of locus coeruleus neurons [2,43]. More recent studies have demonstrated, however, that direct microinjection of naloxone into the locus coeruleus in morphine-dependent rats does not provoke enhanced neuronal firing patterns in a manner similar to those observed after systemic injection of

190

J.J. Buccafusco et al. / Brain Research 679 (1995) 189-199

naloxone [2]. Therefore, the enhanced firing of locus coeruleus neurons after systemically administered naloxone appears to be mediated both by enhanced glutamatergic input to the nucleus as well as neurochemical changes intrinsic to the region [2,40,55]. Other sites which respond to local injection of opiate antagonists by initiating withdrawal symptoms include the nucleus accumbens, the amygdala and the spinal cord [34,50]. In fact, spinal pathways are important for the expression withdrawal signs following both local and systemic administration of naloxone in morphine dependent rats [10,22,34]. In non-dependent rats, intrathecal (IT) administration of cholinergic, muscarinic receptor agonists produce sympathoexcitatory responses which are similar to the responses observed during precipitated morphine withdrawal [10-12]. Interestingly, central cholinergic neurons have long been suggested to mediate many of the signs and symptoms of opiate withdrawal (for a review [10]). Also, centrally acting cholinergic agonists produce a quasi-morphine abstinence syndrome in non-dependent animals [23,56]. Despite this body of evidence, it is not clear how central cholinergic neurons mediate the constellation of signs and symptoms of withdrawal. For example, although spinal cholinergic neurons participate in mediating the increase in blood pressure associated with precipitated withdrawal, these neurons do not directly innervate preganglion cells in the intermediolateral nucleus [52]. Preliminary studies in this laboratory have demonstrated that the pressor response produced by IT injection of muscarinic receptor agonists is blocked in animals pretreated with IT injection of NMDA receptor antagonists, but was unaffected by pretreatment with non-NMDA glutamic acid receptor antagonists [20]. This observation is consistent with the recent finding that cerebroventricular injection of NMDA receptor antagonists block the expression of precipitated morphine withdrawal symptoms [43]. In fact, we have suggested that the pressor response to spinal muscarinic receptor stimulation is mediated through interaction with a descending sympathoexcitatory glutamatergic pathway [10]. The potential role of NMDA receptors in the expression of morphine withdrawal symptoms has provided the impetus for examining the possibility that this receptor pathway is linked to a nitric oxide (NO) generating system. Influx of Ca 2÷ during depolarization of the postsynaptic cell activates a calmodulin-dependent enzyme, nitric acid synthase (NOS). NO is formed from L-arginine in an NADPH-dependent reaction catalyzed by NOS which also generates citrulline. NO acts as a retrograde messenger by diffusing back to the presynaptic terminal where it activates the heme-bearing enzyme guanylate cyclase. The resultant increased level of cGMP then causes enhancement of the release of the transmitter, strengthening the synap-

tic signal [7]. Inhibitors of NO actions include arginine analogs (competitive inhibitors of NOS), hemoglobin and myoglobin (bind and inactivate NO), methylene blue (radical generation) and NADPH derivatives [30]. Interestingly, in the CNS, NO was first linked to glutamatergic synapses rather than cholinergic. In fact, NO may underlie the phenomenon of long-term potentiation (LTP) exhibited by certain hippocampal cells possibly involved in learning and memory [46]. More recently, it has been recognized that at least two isozymes of the constitutive form of the enzyme exist, each the product of different genes [48]. While both the endothelial form and brain form exist in the CNS, the brain form (bNO) is more prevalent, bNO has been linked to the actions of several neurotransmitters, including acetylcholine [6,17,39,45] and is found in high concentrations in both the hypothalamus [45] and the thoracic spinal cord, particularly in the dorsal horn and intermediolateral column [28,29,54]. Gebhart and his colleagues [36] have reported that agonists of the NMDA subtype of the glutamate receptor facilitated a thermal nociceptive spinal reflex. This facilitation was mediated by the NOS system within the spinal cord. As with the hippocampal system, both a NO generating system and a soluble guanylate cyclase system were shown to be important for nociceptive processing in the spinal cord. In their more recent study [59] they demonstrated that spinal NO release was under tonic control by a spinal muscarinic system, thus paralleling our preliminary results regarding blood pressure regulation [20]. Systemic administration of NOS inhibitors recently has been demonstrated to interfere with the expression of morphine withdrawal symptoms [1,16]. Our working hypothesis is that an intrinsic cholinergic muscarinic system within the spinal cord initiates the expression of many of the autonomic and behavioral symptoms of precipitated morphine withdrawal. Release of acetylcholine in turn directly or indirectly activates an NMDA receptor-linked NOS system in which the symptoms of withdrawal are mediated by local NO production. As a first step in testing this hypothesis, the present study was designed to determine whether IT administration of inhibitors of NMDA receptors a n d / o r inhibitors of NOS are capable of inhibiting the expression of morphine withdrawal symptoms.

2. Materials and methods

Male Wistar rats weighing 350-450g at the time of the experiment were obtained from Harlan, SpragueDawley, Indianapolis, IN. The animals were housed in our animal care facilities for at least 5 days prior to experimentation. They had access to standard laboratory rodent chow and tap water on an unlimited basis.

J.J. Buccafusco et al. / Brain Research 679 (1995) 189-199

A 12-h light/dark cycle was maintained. Procedures involving animals were reviewed by the institutional Committee for Animal Use in Research and Education before work was undertaken.

2.1. Surgical procedures Surgical procedures have been described in more detail in earlier studies [13]. Briefly, rats were anesthetized with methohexital (Brevital; 65 mg/kg) and fixed in a stereotaxic instrument in a flat skull orientation [38]. Catheterization of the spinal subarachnoid space was performed as described previously [11,12] by inserting a length of saline-filled polyethylene (PE) 10 tubing caudal to a midline nick in the atlanto-occipital window, terminating in the T l l - T 1 2 segments of the spinal cord. The length of the tubing was varied according to a predetermined scale based empirically on body weight. After recovery (at least 5 days) from the surgical procedure, the rats were again anesthetized with methohexital and a midline abdominal incision was made to expose the left iliac artery. A catheter (PE 50) filled with heparinized (30 units/ml) saline was inserted so that the tip of the catheter terminated in the base of the abdominal aorta below the origin of the renal arteries. The opposite end of the catheter was plugged with a stylet and was directed subcutaneously to emerge at the back of the neck where it was stabilized to an anchoring button. Following surgery, the animals were returned to plastic cages (45 x 25 x 20 cm) and the catheter was passed through a spring support and was connected to a water-tight swivel cannula mounted 30 cm above the cage floor. This method allowed the chronically catheterized rat unrestricted movement to all areas of the cage while a constant infusion (8 ml/day) of heparinized saline maintained the patency of the catheter.

2.2. Production of physical dependence on morphine and measurement of abstinence signs Two days after the surgery to implant the arterial line, rats were made physically dependent upon morphine according to the following schedule: morphine was added to the infusion solution to deliver a total dose of 35 m g / k g / d a y . The concentration of morphine was adjusted each morning of the next 4 days to deliver 50, 75, 100 and 100 m g / k g / d a y , respectively. This schedule was previously demonstrated to elicit a maximal degree of physical dependence on morphine [33]. On the 5th day, the catheter was connected via the swivel to a pressure transducer coupled to a hard copy device (Western Graftec 8-channel thermal array recorder) and the analog signals were amplified and digitized on a Buxco Electronics LS-14 Logging Ana-

191

lyzer. The analyzer provided 1 min averages of mean arterial pressure (MAP) and heart rate (HR) to a computer. Stable MAP and HR values were measured for at least 10 rain prior to treatments in order to obtain the baseline response. Morphine withdrawal was precipitated by an intraarterial injection of naloxone (0.5 mg/kg). Cardiovascular data were analyzed as 1 rain averages for the first 5 min, with subsequent 5 min averages for the remaining 55 min. During the 60 min observation period, the following abstinence signs were recorded: maximal increase in MAP above pre-withdrawal levels, maximal increase in HR above pre-withdrawal levels, withdrawal body shakes ('wet dog shakes'), escape attempts (vigorous attempts to leap to the top of the cage), defecation (normally formed stool), diarrhea (loosely formed stool), lacrimation (secretion of clear tears), chromodacryorrhea (secretion of reddish tears) and teeth chattering. The intensity of abstinence was quantified according to the following point system: one point for each 2 mmHg increase in MAP, maximum of 20; one point for each 20 beats/rain increase in HR, maximum of 5; one point for each withdrawal body shake, maximum of 10; one point for each escape attempt, maximum of 10; one point for each defecation, diarrhea, lacrimation, chromodacryorrhea and teeth chattering, for a potential maximum of 50 points. At the completion of the 60 min observation period, rats were replaced on chronic morphine infusion (100 m g/ kg day) for an additional 2 days. At that time a second naloxone withdrawal was initiated. Our previous studies have indicated that all components of withdrawal could be repeated using this procedure [33]. Thus, for the most part, each animal was employed for two experiments. At the completion of the second experiment, animals were euthanized. Saline controls were randomized throughout the study, being employed equally in first or second withdrawals. For each of the three main experimental series (see below) three saline control animals were run to insure reproducibility of the withdrawal response. First and second runs in a particular animal consisted of either two different doses or different drug regimens or a drug and a saline control. The same drug dose, regimen or saline administration was never repeated in the same animal.

2.3. Central injection of drugs Pretreatment drugs were dissolved in sterile saline and administered to freely moving rats. For IT injections, a 30-gauge stainless steel connector was attached to a 50 ~1 Hamilton syringe via PE 10 tubing. Injections were made into the spinal subarachnoid space by attaching the other end of the connector to the indwelling catheter. Drug solutions (5/.d) or vehicle were infused over 30 s using a constant speed syringe pump

J.J. Buccafusco et al. / Brain Research 679 (1995) 189-199

192

(Harvard Apparatus). An additional 5 ~1 of saline was infused after drug to clear the contents of the IT catheter. In each case naloxone was administered 20 rain after the pretreatment. At the end of experiments, catheter placement was confirmed by dye injection. 2.4. Drugs

Naloxone hydrochloride and atropine sulfate were purchased from Sigma Chemical Co., St. Louis, MO. Morphine sulfate was purchased from Mallinkrodt, Paris, KY, N6-Nitro-L-arginine methyl ester HCL (LNAME), N°-Nitro-L-arginine (L-NOARG), N6-Mono methyl-L-arginine acetate L-NMMA); (+)-2-amino-7phosphonoheptanoic acid (AP-7) and" (+)-MK-801 hydrogen maleate (MK-801) were purchased from Research Biologicals, Inc, Natick, MA. [3H]methylscopolamine chloride (79.5 Ci/mmol) was purchased from DuPont New England Nuclear, Boston, MA. 2.5. Muscarinic receptor binding assay

Rats were sacrificed by decapitation and the spinal cords were removed by high pressure injection of isotonic saline applied to the caudal end of the sacral vertebral column. This tissue was washed and immediately placed in 4 ml of ice cold incubation buffer (50 mM Tris-HCL, pH 7.4, in 2mM MgC1z) and homogenized (teflon/glass). The homogenate was centrifuged at 37,000 x g for 20 min. The pellet was then resuspended in 4 ml of incubation buffer and the protein content was assayed (Bio-Rad Protein Assay System). A standard displacement assay was employed to estimate the inhibition binding constants for two of the

NOS inhibitors employed in the study. Binding reactions were carried out in a reaction volume of 1 ml which contained 100 /zg membrane protein, 1 nM [3H]methylscopolamine and one of a series of increasing concentrations of NOS inhibitor or atropine (control curve) in incubation buffer. Non-specific binding was determined by performing each assay in the presence of 10/xM atropine. Each assay was performed in duplicate. After incubation at room temperature for 90 min, the reaction mixture was filtered through glass fiber filters (Schleicher and Schuell #32) and washed 3 times with 3 ml of ice cold buffer using a Brandel Cell Harvester (Gaithersburg, MD). The filters were dried and then were placed in scintillation fluid for at least 4 h prior to scintillation counting (model LS 9000, Beckman Instruments, Inc., Fullerton, CA). Binding of the drugs to spinal cord muscarinic receptors was inferred from their ability to displace the specific binding of [3H]methylscopolamine. The data points derived from the fractional specific binding (B) of [3H]methylscopolamine to spinal cord membranes were fit by nonlinear regression analysis (Sigma Plot, Jandel Scientific, Corte Madera, CA) to the mass action expression for ligand binding to a single population of non-interacting sites: B = Bm~x " C / ( C + ICso )

where C is the concentration of ligand and IC5o is the concentration of unlabeled drug which produced 50% inhibition of specific [3H]methylscopolamine binding. 2.6. Statistics

Statistical analyses of several populations were performed using ANOVA for repeated measures (in which

Table 1 Control, resting values for mean arterial pressure and heart rate for each experimental group before intrathecal naloxone treatment The plus sign indicates that L-NAME and L-ARG (L-arginine) were administered simultaneously; the arrow indicates that L-NAME was administered 20 min prior to L-NOARG. Each value represents the mean ± S.E.M., * significantly different from respective pre-drug mean, P < 0.05

Drug regimen Saline 100 nmol MK-801 200 nmol MK-801 100 nmol AP-7 100 nmol L-NAME 200 nmol L-NAME 100 nmol L-ARG + 200 nmol L-NAME 50 nmol L-NOARG 1130 nmol L-NOARG 200 nmol L-NOARG 500 nmol L-NOARG 500 mmol L-NAME ~ 500 nmol L-NOARG 200 nmol L-NAME ~ 500 nmol L-NOARG 200 nmol L-NMMA

Mean arterial pressure

Heart rate

pre-drug 105 _+ 5 117 _+ 5 126 _+ 5 122 _+ 3 111 _+ 4 104 ± 7 116 +_ 6

after drug 105 + 5 115 _+ 5 125 + 6 111 _+ 5 * 107 _+ 4 110 +_ 3 115 _+ 4

pre-drug 394 _+ 24 438 +_ 20 385 ± 24 417 ± 15 344 + 22 400 + 27 372 ± 26

pre-naloxone 386 +_ 20 450 ± 24 444 _ 16 * 427 + 23 337 + 24 397 ± 39 384 ± 22

N 9 5 7 6 6 5 6

107 108 106 117 122

109 111 108 121 128

360 356 334 364 418

356 357 343 346 390

_+ 15 _+ 31 +_ 11 ± 16 + 14

6 5 5 7 7

± 2 _+ 2 ± 2 _+ 6 ± 6

_+ 5 ± 2 ± 8 +_ 6 ± 3

± ± ± ± ±

12 28 20 17 18

123 ± 4

128 ± 5

398 ± 24

408 ± 20

7

112 ± 6

112 ± 5

361 ± 11

354 ± 14

4

].].

Buccafusco et aL / Brain Research 679 (1995) 189-199

the repeated measure was time). Although rats were generally employed for two experiments, because of the random assignments of treatments, this did not constitue a repeated measure. When comparing two groups of data, Student's t-test was performed. Differences were considered significant at P < 0.05. All data are reported as the mean _+ standard error of the mean. All of the treatment groups depicted in Figs. 1, 3 and 5 exhibited significant overal time effects by ANOVA with repeated measures.

193

35 - T

Saline 200 nmol MK-801 - - o I 0 0 nmol AP-7

-z..

t

{3

,

2° i 15

I 10

20

30

40

50

60

Time after naloxone injection (min)

3. Results

Baseline resting MAP and HR values remained within 20% of the saline control values for all of the subgroups (Table 1). For the most part, there were no significant differences between baseline (pre-drug) and pre-naloxone means for each subgroup, with pre-naloxone values remaining within 10% of respective baseline values. Care was taken to limit the dose range of all IT pretreatment regimens so as to preclude their potential for evoking hemodynamic or behavioral actions prior to naloxone withdrawal. The only exceptions were the 200 nmol dose of MK-801 which resuited in a significant increase in HR and 100 nmol of AP-7 which produced a significant decrease in MAP. However, these changes still allowed pre-naloxone values which were not out of the range for the other experimental groups.

Fig. 1. The increase in mean arterial pressure (MAP) to naloxone (0.5 m g / k g , i.a.) in morphine dependent rats. Saline pretreatment is represented by the open circles. Intrathecal (IT) pretreatment with MK-801 (filled circles) or AP-7 (squares) significantly inhibited the pressor response to naloxone. The main treatment component according to A N O V A for repeated measures was F i j 6 = 27.1, P < 0.01 and El, 9 = 9.8, P < 0.05, respectively, for the two drugs. Each value represents the mean _+S.E.M. (note: N values for each experiment are presented in Table 1).

3.2. Anti-withdrawal actions of NMDA receptor ahtagonists The non-competitive NMDA receptor antagonist MK-801 was administered by IT injection 20 min prior to i.a. injection of naloxone (0.5 mg/kg). Pretreatment with 100 or 200 nmol resulted in a dose-dependent reduction in the total withdrawal score which was significantly reduced compared to controls at the 200

3.1. The naloxone-precipitated withdrawal syndrome in control dependent rats lntraarterial injection of 0.5 mg/kg of naloxone in saline-pretreated morphine dependent animals produced an immediate increase in MAP which peaked between 5-15 min after injection and gradually declined over the remaining 45-55 min (Fig. 1). Although controls were interspaced among each of the subgroups in the study, there were no significant differences observed between control naloxone responses or between naloxone responses in which prior saline was administered either prior to the first or second withdrawal. Many of the expected withdrawal signs including body shakes, escape attempts, teeth chattering, defecation, diarrhea and chromodacryorrhea were observed in association with the cardiovascular changes following naloxone treatment. These signs were variably expressed among the animals, with the prominent behaviors being body shakes, escape attempts and teeth chattering. Withdrawal signs were most evident within the first 5-15 rain following naloxone treatment at the time of the peak cardiovascular changes.

Sal

MK-801

AP-7

L-NAME

40

30 '

~"

"

20

~-

t

*

I

10

5 0

100

21111

100

11111 20O

Dose (nmol) F i g . 2. T h e

influence

of IT

pretreatment ( 2 0

min before

naloxone)

with saline (Sal), MK-801, AP-7 or L-NAME on withdrawal associated cardiovascular and behavioral signs of naloxone-precipitated withdrawal in morphine-dependent rats. Data are presented as scored abstinence points with the combined open and solid bars representing the total combined (cardiovascular and behavioral) score and the solid bars representing only the combined behavioral score. Individual doses of each drug are indicated on the abscissa. Each value indicates the mean_+ S.E.M. * indicates a significant difference ( P < 0.05) from saline control means for respective cardiovascular or behavioral abstinence scores.

194

J.J. Buccafusco et al. / Brain Research 679 (1995) 189-199

nmol dose (Fig. 2). As indicated in Fig. 2, MK-801 also significantly inhibited the expression of behavioral signs of withdrawal. For the most part, the inhibition of the cardiovascular component of withdrawal was due to the drug's significant inhibitory effect on the postwithdrawal M A P increase (Fig. 1). For comparison with the previous literature, the maximal change in M A P after naloxone was employed in our estimation of the total abstinence score. A better estimate of the ability of the drug to inhibit the expression of the withdrawal-associated pressor response over the entire time course of the experiment is the average M A P increase (an estimate of the area under the M A P curve) [33]. In this respect, the average post-naloxone increase in M A P for control animals was 20.1 + 2.0 m m H g . MK-801 p r e t r e a t m e n t reduced the average post-naloxone M A P to 10.3 + 2.1 (49% inhibition, P < 0.01) and 5.7 + 0.87 (71% inhibition, P < 0.01) m m H g respectively, for the two doses. I T injection of 100 nmol of the competitive N M D A receptor antagonist AP-7 produced a very similar profile of inhibition as described above for MK-801 (Figs. 1 and 2). AP-7 pretreatment reduced the average post-naloxone increase in M A P to 9.3 + 1.4 m m H g (54% inhibition, P < 0.01). Higher doses of MK-801 or AP-7 were not employed in an attempt to further reduce withdrawal signs because of the potential for higher doses to evoke cardiovascular or motor disturbances (as observed in preliminary studies). 3.3. Anti-withdrawal actions o f N O S inhibitors

I T administration of the methyl ester derivative of arginine, L-NAME, to morphine dependent rats produced no lasting changes in M A P or H R and no gross alterations in ongoing behavior. The NOS inhibitor produced a dose-dependent inhibition of the total abstinence score in naloxone treated animals which was significantly different from controls at the 200 nmol dose (Fig. 2). In contrast to the effects produced by the N M D A receptor antagonists, the reduction in total abstinence score produced by L-NAME was due primarily to a reduction in the cardiovascular component of withdrawal, with little effect on the withdrawal associated behaviors. As indicated in Fig. 3, L-NAME was very effective in reducing the withdrawal associated pressor response. In fact, the average post-naloxone increase in M A P attained in the presence of L-NAME was only 12.0 + 2.1 m m H g (40% inhibition, P < 0.05). To determine whether the L-NAME-induced inhibition of the withdrawal associated pressor response was due to the specific inhibition of NOS, a parallel experiment was performed in which 100 nmol of L-arginine was administered together in the same solution with 200 nmol of L-NAME as a p r e t r e a t m e n t regimen. In this series, L-arginine almost completely reversed the ability

35

30

Saline T

200 nmol L-NAME -

o.

25

.E 2 0 -

i'

15-

o_

I

I

10

20

I 30

I 40

k

L

50

60

Time after naloxone injection (min)

Fig. 3. The increase in mean arterial pressure (MAP) to naloxone (0.5 mg/kg, i.a.) in morphine dependent rats. Saline pretreatment is represented by the open circles. IT pretreatment with L-NAME (filled circles) significantly inhibited the pressor response to naloxone. The main treatment component according to ANOVA for repeated measures was FI,14= 7.1, P < 0.05. In contrast, pretreatment with a mixture of L-arginine (L-Arg)and L-NAME (squares) did not significantly effect the naloxone-induced pressor response; F1,15 = 1.75, P > 0.05. Each value represents the mean + S.E.M.

of L-NAME to inhibit the withdrawal associated pressor response (Fig. 3). The average post-naloxone pressor response after the combination regimen was 17.1 + 1.6 m m H g , which was not significantly different from the control mean (20.1 _+ 2.0 mmHg). In an attempt to confirm these actions of L-NAME, an additional NOS inhibitor was employed. L-NOARG, like L-NAME is an arginine analog, but unlike L-NAME is not an ester derivative. During our initial experiments with L-NOARG, we were surprised to find that I T p r e t r e a t m e n t with the inhibitor was not effective in inhibiting the withdrawal associated pressor response. We therefore employed a broader range of doses, from 50-500 nmol. Pretreatment with L - N O A R G produced a dose-dependent decrease in the expression of total abstinence points (Fig. 4). Quite unexpectedly and in direct contrast to L-NAME, the reduction in total abstinence points reflected primarily a marked inhibition of the withdrawal associated behavior with virtually no effect on the cardiovascular component of withdrawal (Fig. 5). We next employed another NOS inhibitor, L-NMMA. This compound is a monomethyl derivative of arginine, but is not a methyl ester like L-NAME. L-NMMA pretreatment (200 nmol) also produced a significant reduction in total abstinence points, but like L-NOARG, the reduction was a reflection of almost a complete suppression of withdrawal associated behavior (Fig. 4). L-NMMA was without effect on the cardiovascular component of withdrawal (Fig. 5). In the next series of experiments, dependent animals were pretreated (IT) with either 200 or 500 nmol of L-NAME, followed 20 min later by 500 nmol of L - N O A R G and 20 min later by naloxone. With this regimen, withdrawal associated behavior was again sig-

ZJ. Buccafusco et al. / Brain Research 679 (1995) 189-199 Sal

L-NOARG

L-NMMA

Sal

195 L-NAME L-NOARG

40 40 i

L-NAME+ L-NOARG . . . .

MK-801

35

~

~

8

~

20

-

~

5 5O

100

2OO 5 0 0

.

5

dm 2O0

0

Dose (nmol)

5O Saline l

200 nmol L-NOARG -4D= 200 nrnol L-NMMA

c_ 30 m

~ 2O m

10

~ o --~ - - - - J 10

20

5oo

~

~

Dose (nmol)

nificantly suppressed (Fig. 6). In addition, there was a significant reduction in the expression of the withdrawal associated pressor response m comparing the data for the 500 nmol L-NOARG pretreatment against the two regimens which included L-NAME by ANOVA, F28,308 = 3.35, P < 0.01), data not shown. In fact, the combination of L-NAME and L-NOARG resulted in a degree of suppression of the expression of total abstinence points which was similar to that produced by 200 nmol of MK-801 (Fig. 6).

t

- 2OO

Fig. 4. The influence of IT pretreatment (20 min before naloxone) with saline (Sal), t.-NOARG or L-NMMA on withdrawal associated cardiovascular and behavioral signs of naloxone-precipitated withdrawal in m o r p h i n e - d e p e n d e n t rats. Data are presented as scored abstinence points with the combined open and solid bars representing the total combined (cardiovascular and behavioral) score and the solid bars representing only the combined behavioral score. Individual doses of each drug are indicated on the abscissa. Each value indicates the mean ± S.E.M. * indicates a significant difference ( P < 0.05) from saline control m e a n s for respective cardiovascular or behavioral abstinence scores.

40

*

°

0

T E

~

= o

El

~

30

I

I

I

I

30

40

50

60

Time after naloxone injection (rain)

Fig. 5. The increase in m e a n arterial pressure (MAP) to naloxone (0.5 m g / k g , i.a.) in morphine d e p e n d e n t rats. Saline pretreatment is represented by the open circles. Intrathecal (IT) pretreatment with L - N O A R G (filled circles) or L-NMMA (squares) did not inhibit the pressor response to naloxone. The main treatment c o m p o n e n t according to A N O V A for repeated m e a s u r e s was FzA3 = 3.9, P > 0.05 and Ft,t4 = 0.01, P > 0.1, respectively, for the two drugs. Each value represents the m e a n ± S.E.M.

Fig. 6. The influence of IT pretreatment (20 min before naloxone) with saline (Sal), L-NAME, L-NOARG, L-NAME followed 20 min later by L - N O A R G ( L - N A M E + L - N O A R G ) , or MK-801 on withdrawal associated cardiovascular and behavioral signs of naloxoneprecipitated withdrawal in m o r p h i n e - d e p e n d e n t rats. Data are presented as scored abstinence points with the combined open and solid bars representing the total combined (cardiovascular and behavioral) score and the solid bars representing only the combined behavioral score. Individual doses of each drug are indicated on the abscissa. For the combination regimens, the upper n u m b e r represents the L-NAME dose and the lower n u m b e r represents the L - N O A R G dose. Each value indicates the mean_+ S.E.M. * indicates a significant difference ( P < 0.05) from saline control m e a n s for respective cardiovascular or behavioral abstinence scores.

3.4. Muscarinic receptor binding affinity Certain alkyl ester derivatives of L-arginine, such as L-NAME, have been demonstrated to bind to brain muscarinic receptors at very high concentrations [15]. And, since we have demonstrated that blockade of spinal muscarinic receptors can inhibit the pressor response associated with precipitated withdrawal [10,22], we considered the possibility that L-NAME-induced inhibition of the naloxone-induced pressor response may be related to the latter mechanism. Atropine displaced [3H]methylscopolamine binding to spinal cord membranes with an IC50 of 1.5 nM. LNAME displaced [3H]methylscopolamine binding only at concentrations above 1 mM. L-NMMA was even less effective in this regard than L-NAME (Fig. 7). In view of the virtual lack of affinity of the NOS inhibitors for spinal muscarinic receptors, it is highly unlikely that spinal muscarinic receptor blockade contributed to their pharmacologic actions following IT administration.

4. D i s c u s s i o n

The nature of the neurochemical substrates underlying the expression of the morphine withdrawal syn-

J.J. Buccafusco et al. / Brain Research 679 (1995) 189-199

196 [3H] Methylscopolamlne

Displacement Curve

1.0-

N=

0.8

13 =

0.6

o

0.4

ATROPINE

•

L-NAME

o

L-NMMA

•

0.2

0.0

•

•

~

m e

r

F

r

i

i

F

-11

-10

-9

-8

-7

-6

i

i

i

-5

-4

-3

-2

log[DRUG] M

Fig. 7. Displacement of [3H]methylscopolamine binding to rat spinal membranes by atropine (solid circles), L-NAME (open circles) or L-NMMA (squares). Each point represent the mean of data derived from three separate assays (using tissue from three rats) in which each assay was performed in duplicate.

drome has been the subject of intense investigation for many years. It is not possible to review here the vast literature in this area. However, it can be stated with some confidence that several classical neurotransmitters play a role in the expression of morphine withdrawal symptoms. It also seems clear that in dependent animals, several brain and spinal regions respond with one or a variety of withdrawal symptoms when exposed to a non-selective opiate antagonist. Morphine and related drugs have a profound inhibitory action on peripheral and central cholinergic neurons and during withdrawal, the enhanced function of brain and spinal cholinergic neurons appears to contribute significantly to the expression of behavioral and autonomic withdrawal symptoms [10,22]. These withdrawal associated symptoms in large part are likely to be mediated via neuronal pathways which normally function to control the status of autonomic and sensory reflexes. In the spinal cord this might include local and descending pathways which regulate blood pressure and sensory processing. Thus, knowledge of the neurochemical substrates which mediate these regulatory pathways may provide clues to the neurochemical nature of the withdrawal phenomenon. Spinal cholinergic (muscarinic) neurons enhance descending sympathetic tone to spinal preganglionic neurons [10-12,52]. A tonic spinal cholinergic antinociceptive system also exists [59]. In both cases, muscarinic receptor-mediated pressor responses and antinociceptive responses appear to be influenced by a local NO generating system [20,59]. Moreover, another common feature appears to be the role of spinal NMDA receptors. NMDA receptor antagonists inhibit the pressor

response to spinal muscarinic receptor stimulation [20]. Also, NOS antagonists inhibit the NMDA-evoked facilitation of the thermal tail-flick reflex [36]. More direct evidence for the role of NMDA receptors and NO in mediating the symptoms of morphine withdrawal has come from studies in which antagonists of NMDA receptors or NOS have been employed in animal models of precipitated withdrawal. For example, systemic administration of NOS inhibitors in morphine dependent mice and rats has been reported to inhibit the expression of naloxone-induced withdrawal behavior [1,16,24]. Also, systemic or lateral cerebroventricular administration of non-selective glutamate receptor antagonists have been reported to inhibit the expression of morphine withdrawal symptoms [43]. Parenthetically, although blockade of brain NMDA receptors suppressed withdrawal associated behavior in these studies, there was no effect on the withdrawal associated enhanced firing of locus coeruleus noradrenergic neurons, nor was there an effect on the withdrawal associated increase in norepinephrine turnover in several brain regions [42]. Thus, in addition to the possibility that locus coeruleus noradrenergic neurons mediate the gross behavioral signs and symptoms of opiate withdrawal (e.g., [40,41]), additional new evidence suggests that other brain or spinal sites may be equally or even more important in this regard. In the present study, the IT route of administration was employed to direct drug solutions directly to spinal sites of action. Our previous experience with this specific methodology using several classes of pharmacologic agents has indicated that the spinal cord is the primary if not the exclusive site of action [9,21,22,51]. Also, it is not likely that there was significant drug redistribution to the systemic circulation after IT administration, as we observed no effect on systemic blood pressure during the period prior to naloxoneprecipitated withdrawal (Table 1). IT administration of NMDA receptor blocking drugs produced a significant and dramatic reduction in the entire spectrum morphine withdrawal symptoms. The degree of inhibition produced by either MK-801 or AP-7 was equivalent to that produced by IT injection of the muscarinic receptor (M1) antagonist pirenzepine [22]. It is possible, therefore, that some relationship exists between spinal cholinergic and glutamatergic systems in the expression of withdrawal symptoms. Since these same NMDA receptor antagonists block the pressor response to spinal muscarinic receptor stimulation [20] and, since spinal cholinergic neurons do not directly innervate preganglionic pressor neurons [52], the most likely scenario is that during withdrawal there is an enhanced release of acetylcholine which acts upon M1 receptors which may be located on glutamatergic neurons. The release of glutamate, in turn, activates NMDA receptors on spinal preganglion

J.J. Buccafusco et al. / Brain Research 679 (1995) 189-199

neurons to elevate blood pressure. The neurochemical relationship for the expression of withdrawal behaviors may or may not, be similar to that just described for the cardiovascular symptoms. Certain symptoms, such as withdrawal body ('wet dog') shakes, diarrhea and chromodacryorrhea are most likely expressed in response to withdrawal-evoked enhanced autonomic nerve discharge, whereas other behaviors may be due to activation of ascending spinal pathways, including perhaps, to the locus coeruleus. NOS containing neurons have been found throughout the central nervous system. Moreover, NOS neurons in rat brain appear to express a greater abundance of NMDA receptor mRNA than do non-NOS neurons [39]. However, in certain well defined central pathways, afferent cholinergic neurons provide the exclusive source of NOS to target neuronal populations [6]. Much less is known about the relationship of NOS to other neurotransmitter systems in the spinal cord. Each of the two known constitutive forms of NOS, bNOS and endothelial NOS (eNOS) exist in the spinal cord [48]. bNOS has been found in several spinal regions, including the dorsal horn (intrinsic and primary sensory fibers and non-sensory neurons), the central canal region (laminae X), intermediolateral cell column (thoracic level) and ventral horn (associated with motor neurons) [27,54,57]. Since spinal cholinergic pressor neurons have been localized to the central canal area (intermediomedial nucleus) [52] and NMDA receptors have been localized to preganglionic autonomic cells [5], it is possible that NOS containing cells might interact with either or both cholinergic or glutamatergic neurons in the spinal cord. In the hippocampus, NO generation appears to strengthen NMDAmediated synaptic transmission [46]. However, since IT administration of NOS inhibitors were as effective as muscarinic [22] and NMDA receptor antagonists in inhibiting the expression of withdrawal symptoms, it is likely that NO generation plays a more than modulatory role. That is, while it is clear that both acetylcholine and glutamate can stimulate the formation of NO, NO has been demonstrated to induce the release of glutamate in a cerebral cortical preparation [32] and within the spinal cord [49]. NO also appears to induce the release of glutamate and GABA in the hippocampus and striatum [47]. Further studies will be required to elucidate the precise neurochemical relationships between NO and cholinergic (muscarinic) and glutamatergic (NMDA) pathways. One additional interesting outcome of this study was the differential action of the three NOS inhibitors in blocking certain symptoms of withdrawal. In general, these compounds have exhibited only small differences (mainly in NOS inhibitory potency) in in vitro studies. However, these compounds do have certain structural and chemical differences and the results of several in

197

vivo studies have indicated that these differences may underlie their much different pharmacologic actions [26,44]. For example, L-NAME is much more lipid soluble than L-NOARG and as such may diffuse further when injected into tissue or may have access to certain cellular compartments not available to the more hydrophilic derivatives [25]. The compounds may exhibit different affinities for the different NOS isoforms [4]. In fact, the eNOS isoform, may actually be a misnomer, since eNOS may play an important role in brain function [37]. Since both isoforms exist in the spinal cord [48], it is possible that the varied profile of anti-withdrawal activity we observed for L-NAME vs. L-NOARG and L-NMMA may be due to differences in affinity for bNOS and eNOS. It is not likely, however, that these differences in anti-withdrawal profile are related to direct inhibition of spinal muscarinic receptors, since neither type of inhibitor (methyl ester analog L-NAME, non-ester methyl analog L-NMMA) exhibited any affinity for spinal muscarinic receptors. Our findings are in apparent opposition to those from an earlier study which purported that L-NAME exhibited/xM affinity for guinea pig brain muscarinic receptors [15]. Besides the obvious species difference, another potential explanation for the differing results includes our use of spinal cord rather than whole brain. The spinal cord is deficient in ml and m4 subtype specific mRNA compared with the cerebral cortex and other higher centers. On the other hand, m2 and m3 mRNAs are enriched in the spinal cord [58]. Indeed, L-NAME was much less effective in binding assays when the tissue employed was enriched in M3 muscarinic receptors [15]. Another difference between the two studies is the labeled ligand employed. The earlier study employed [3H]quinuclidinyl benzilate whereas this study employed [3H]methylscopolamine. The former ligand is essentially an irreversible muscarinic receptor antagonist and is very lipid soluble, whereas [3H]methylscopolamine is a reversible antagonist and is hydrophilic. As such, [3H]quinuclidinyl benzilate has access to internalized binding sites which may not be active functional receptors. In conclusion, the results of this study are consistent with a role for spinal NMDA receptors and a NO generating system in the expression of cardiovascular and behavioral components of the precipitated morphine withdrawal syndrome in rats. In conjunction with a spinal cholinergic muscarinic pathway, these may form the basis of a neurochemical substrate which mediates the expression of a variety of withdrawal symptoms. These results are also consistent with an increasing literature regarding the potential selectivity of several NOS antagonists for the various isoforms of NOS. As such, these compounds may be employed in future studies designed to elucidate the role of NOS isoforms in brain and spinal cord function.

198

J.J. Buccafusco et al. / Brain Research 679 (1995) 189-199

Acknowledgements T h i s s t u d y w a s s u p p o r t e d by t h e O f f i c e o f R e s e a r c h and Development, Medical Research Service of the Department of Veterans Affairs.

References [1] Adams, M.L., Kalicki, J.M., Meyer, E.R. and Cicero, T.J., Inhibition of the morphine withdrawal syndrome by a nitric oxide inhibitor, NG-nitro-L-arginine methyl ester, Life Sci., 52 (1993) PL245-PL249. [2] Akaoka, A. and Aston-Jones, G., Opiate withdrawal-induced hyperactivity of locus coeruleus neurons is substantially mediated by augmented excitatory amino acid input, Jr. Neurosci., 11 (1991) 3830-3839. [3] Baumeister A.A., Anticich, T.G., Hebert, G., Hawkins, M.F. and Nagy, M., Evidence that physical dependence on morphine is mediated by the ventral midbrain, Neuropharmacology, 28 (1989) 1151-1157. [4] Babbedge, R.C., Wallace, P., Gaffen, Z.A., Hart, S.L. and Moore, P.K., L-N6-nitro arginine p-nitroanilide (L-NAPNA) is anti-nociceptive in the mouse, Neuro Report, 4 (1993) 307-310. [5] Bazil, M.K. and Gordon, F.J., Spinal NMDA receptors mediate pressor responses evoked from the rostral ventraolateral medulla, Am. J. Physiol., 260 (1991) H267-H275. [6] Bickford, M.E., Gunluck, A.E., Guido, W. and Sherman, S.M., Evidence that cholinergic axons from the parabrachial region of the brainstem are the exclusive source of nitric oxide in the lateral geniculate nucleus of the cat, J. Comp. Neurol., 334 (1993) 410-430. [7] Bredt, D.S. and Snyder, S.H., Nitric oxide, a novel neuronal messenger, Neuron, 8 (1992) 3-11. [8] Buccafusco J.J., Cardiovascular changes during morphine withdrawal in the rat: effects of clonidine, Pharmacol. Biochem. Behav., 18 (1983) 209-215. [9] Buccafusco J.J., Participation of different brain regions in the anti-narcotic withdrawal action of clonidine in the dependent rat, Brain. Res., 513 (1990) 8-14. [10] Buccafusco J.J., Neuropharmacologic and behavioral actions of clonidine: interactions with central neurotransmitters, Intern. Rev. Neurobiol., 33 (1992) 55-107. [11] Buccafusco J.J. and Magri', V., Modification of spino-bulbar autonomic cholinergic systems by activation of alpha-adrenergic receptors, J. Auton. Nerv. Syst., 28 (1989) 133-140. [12] Buccafusco J.J. and Magri', V., The pressor response to spinal cholinergic stimulation in spontaneously hypertensive rats, Brain Res. Bull., 25 (1990) 69-74. [13] Buccafusco J.J., Makari, N.F. and Hays, A.C., Pharmacological analysis of the enhanced pressor response to central cholinergic stimulation in spontaneously hypertensive rats, Jap. J. Pharmacol., 54 (1990) 105-112. [14] Buccafusco, J.J., Marshall, D.C. and Turner, R.M., A comparison of the inhibitory effects of clonidine and guanfacine on the behavioral and autonomic components of morphine withdrawal in rats, Life Sci., 35 (1984) 1401-1408. [15] Buxton, I.L., Cheek, D.J., Eckman, D., Westfall, D.P., Sanders, D.M. and Keel, K.D., NG-nitro L-arginine methyl ester and other alkyl esters of arginine are muscarinic receptor antagonists, Circ. Res., 72 (1993) 387-395. [16] Cappendijk, S.L.T., de Vries, R. and Dzoljic, M.R., Inhibitory effect of nitric oxide (NO) synthase inhibitors on naloxone-precipitated withdrawal syndrome in morphine-dependent mice, Neurosci. Lett., 162 (1993) 97-100.

[17] Castoldi, A.F., Manzo, L. and Costa, L.G., Cyclic GMP formation induced by muscarinic receptors is mediated by nitric oxide synthesis in rat cortical primary cultures, Brain Res., 610 (1993) 57-61. [18] Delle M., Ricksten, S.-E., H~iggendal J., OIsson K., Skarphendinsson J.O. and Thorfin P., Regional changes in sympathetic nerve activity and baroreceptor reflex function and arterial plasma levels of catecholamines, renin and vasopression during naloxone-precipitated morphine withdrawal in rats, J. Pharmacol. Exp. Ther., 253 (1990) 646-654. [19] Delle M. and Thori~n P., Changes in sympathetic nerve activity during morphine abstinence in the rat, Acta Physiol. Scand., 130 (1987) 47-54. [20] Feldman, D.S. and Buccafusco, J.J., Role of glutamate and nitric oxide on the pressor response to spinal cholinergic stimulation, Neurosci. Abs., 19 (1993) 313. [21] Feldman, D.S. and Buccafusco, J.J., Localization of cholinergic neurons involved in the cardiovascular response to intrathecal injection of carbachol, Eur. J. Pharmacol., 250 (1993) 483-487. [22] Holland, L.N., Shuster, L.C. and Buccafusco, J.J., Role of spinal and supraspinal muscarinic receptors in the expression of morphine withdrawal symptoms in the rat, Neuropharmacology, 32 (1993) 1387-1395. [23] Katz, J.L. and Valentino, R.J., The opiate quasi withdrawal syndrome in rhesus monkeys: comparison of naloxone-precipitated withdrawal to effects of cholinergic agents, Psychopharmacology, 84 (1984) 12-15. [24] Kimes, A.S., Vaupel, D.B. and London, E.D., Attenuation of some signs of opioid withdrawal by inhibitors of nitric oxide synthase, Psychopharmacology, 112 (1993) 521-524. [25] Krejcy, K., Schwarzacher, S. and Raberger, G , Distribution and metabolism of NG-nitro-L-arginine and NG-nitro-L-arginine methylester in canine blood in vitro, Nauyn-Schmiedeberg's Arch. Pharmacol., 347 (1993) 342-345. [26] Lambert, L.E., Whitten, J.P., Baron, B.M., Cheng, H.C., Doherty, N.S. and McDonald, I.A., Nitric oxide synthesis in the CNS, endothelium and macrophages differs in its sensitivity to inhibition by arginine analogues, Life Sci., 45 (1991) 69-75. [27] Lee, J-H., Price, R.H., Williams, F.G., Mayer, B. and Beitz, A.J., Nitric oxide synthase is found in some spinothalamic neurons and in neuronal processes that appose spinal neurons that express Fos induced by noxious stimulation, Brain Res., 608 (1993) 324-333. [28] Lin, R.F., Lin, T-S, Tilton, R.G. and Cross, A.H., Nitric oxide localized to spinal cords of mice with experimental allergic encephalomyelitis: an electron paramagnetic resonance study, J. Exp. Med., 178 (1993) 643-648. [29] Liuzzi, F.J., Wu, W., Scoville, S.A. and Schinco, F.P., Development of nitric oxide synthase expression in the superficial dorsal horn of the rat spinal cord, Experiment. Neurol., 121 (1993) 275 -278. [30] Luscher, T., Physiology of vascular endothelium; endotheliumderived vasoactive substances. In Endothelial Vasoactive Substances and Cardiovascular Disease, Karger, Basel, 1988, pp. 17-19. [31] Maldonado, R., Stinus, L., Gold, L.H. and Koob, G.F., Role of different brain structures in the expression of the physical morphine withdrawal syndrome, J. Pharmacol. Exp. Ther., 261 (1992) 669-677. [32] Montague, P.R., Gancayco, C,D., Winn, M.J., Marchase, R.B. and Friedlander, M.J., Role of NO production in NMDA receptor-mediated neurotransmitter release in cerebral cortex, Science, 263 (1994) 973-976. [33] Marshall D.C. and Buccafusco J.J., A comparison of the cardiovascular and the behavioral changes following naloxone as measures of the degree of physical dependence on morphine in rats, Drug Devel. Res., 5 (1985) 271-280.

J.J. Buccafusco et al. / Brain Research 679 (1995) 189-199 [34] Marshall D.C. and Buccafusco J.J., Supraspinal and spinal mediation of naloxone induced morphine withdrawal in rats, Brain Res., 329 (1985) 131-142. [35] Marshall D.C. and Buccafusco J.J., Spontaneous morphine withdrawal from the rat spinal cord, Experientia, 41(1985) 1151-1152. [36] Meller, S.T., Dykstra, C. and Gebhart, G.R., Production of endogenous nitric oxide and activation of soluble guanylate cyclase are required for N-methyI-D-aspartate-produced facilitation of the nonciceptive tail-flick reflex, Eur. J. PharmacoL, 214 (1992) 93-96. [37] O'Dell, T.J., Huang, P.L., Dawson, T.M., Dinerman, J.L., Snyder, S.H., Kandel, E.R. and Fishman, M.C., Endothelian NOS and the blockade of LTP by NOS inhibitors in mice lacking neuronal NOS, Science, 265 (1994) 542-546. [38] Paxinos G. and Watson C., The Rat Brain in Steriotaxic Coordinates, Academic Press, New York, 1982. [39] Price, R.H. Jr., Mayer, B. and Beitz, A.J., Nitric oxide synthase neurons in rat brain express more NMDA receptor mRNA than non-NOS neurons, NeuroReport, 4 (1993) 807-810. [40] Rasmussen, K. and Aghajanian, G.K., Withdrawal-induced activation of locus coeruleus neurons in opiate-dependent rats: attentuation by lesions of the nucleus paragigantocellularis, Brain Res., 505 (1989) 346-350. [41] Rasmussen, K., Beitner-Johnson, D.B., Krystal, J.H., Aghajanian, G.K. and Nestler, E.J., Opiate withdrawal and the rat locus coeruleus: behavioral, electrophysiological and biochemical correlates, J. Neurosci., 10 (1990) 2308-2317. [42] Rasmussen, K., Fuller, R.W., Stockton, M.E., Perry, K.W., Swinford, R.M. and Ornstein, P.L., NMDA receptor antagonists suppress behaviors but not norepinephrine turnover or locus coeruleus unit activity induced by opiate withdrawal, Eur. J. Pharmacol., 197 (1991) 9-16. [43] Rasmussen, K., Krystal, J.H. and Aghajanian, G.K., Excitatory amino acids and morphine withdrawal: differential effects of central and peripheral kynurenic acid administation, Psychopharmacology, 105 (1991) 508-512. [44] Rees, D.D., Palmer, R.M.J., Schulz, R., Hodson, H.F. and Moncada, S., Characterization of three inhibitors of endothelial nitric oxide synthase in vitro and in vivo, Br. J. Pharmacol., 101(1990) 746-752. [45] Rettori, V., Gimeno, M., Lyson, K. and McCann, S.M., Nitric oxide mediates norepinephrine-induced prostaglandin E 2 release from the hypothalamus, Proc. Natl. Acad. Sci. USA, 891(1992) 1543-11546. [46] Schuman, E.M. and Madison, D.V., The intracellular messenger nitric oxide is required in long-term potentiation, Science, 254 (1991) 1503-1506.

199

[47] Segovia, G., Porras, A. and Mora, F., Effects of a nitric-oxide donor on glutamate and GABA release in striatum and hippocampus of the conscious rat, Neuroreport, 5 (1994) 1937-1940. [48] Sessa, W.C., Harrison, J.K., Luthin, D.R., Pollock, J.S. and Lynch, K.R., Genomic analysis and expression patterns reveal distinct genes for endothelial and brain nitric oxide synthase, Hypertension, 21 (1993) 934-938. [49] Sorkin, L.S., NMDA evokes an L-NAME sensitive spinal release of glutamate and citrulline, Neuroreport, 4 (1993) 479-482. [50] Stinus, L., Le Moal, M. and Koob, G.F., Nucleus accumbens and amygdala are possible substrates for the aversive stimulus effects of opiate withdrawal, Neuroscience, 37 (1990) 767-773. [51] Takahashi, H. and Buccafusco, J.J., The sympathoexcitatory response following selective activation of a spinal cholinergic system in anesthetized rats. Z Autonom. Nerr. Svs., 34 (1991) 59-68. [52] Takahashi, H. and Buccafusco, J.J., Excitatory modulation by a spinal cholinergic system of a descending sympathoexcitatory pathway in rats, Neuropharmacology, 31 (1992) 259-269. [53] Taylor, J.R., Elsworth, J.D., Garcia, E.J., Grant. S.J., Roth, R.H. and Redmond, D.E., Jr., Clonidine infusion into the locus coeruleus attenuates behavioral and neurochemical changes associated with naloxone-precipitated withdrawal, Psvehopharmacology, 96 (1988) 121-134. [54] Terenghi, G., Riveros-Moreno, V., Hudson, L.D., Ibrahim, N.B.N. and Polak, J.M., Immunohistochemistry of nitric oxide synthase demonstrates immunoreactive neurons in spinal cord and dorsal root ganglia of man and rat, J. Neurolog. Sci., 118 (1993) 34-37. [55] Tung, C.-S., Grenhoff, J. and Svensson, T.H., Morphine withdrawal responses of rat locus coeruleus neurons are blocked by an excitatory amino-acid antagonist, Acta Physiol. Scand., 138 (1990) 581-582. [56] Turski, W.A., Czuczwar, S.J., Turski, L., Sieklucka-Dziuba, M. and Kleinrok, Z., Studies on the mechanism of wet dog shakes produced by carbachol in rats, Pharmacology, 28 (1984) 112-120. [57] Valtschanoff, J.G., Weinberg, R.J. and Rustioni, A., NADP}t diaphorase in the spinal cord of rats. Z Cornp. Neurol., 321 (1992) 209-222. [58] Wei, J., Walton, E.A., Milici, A. and Buccafusco J.J., m l-m5 muscarinic receptor distribution in rat CNS by RT-PCR and HPLC, J. Neurochem., 63 (1994) 815-821. [59] Zhuo, M., Meller, S.T. and Gebhart, G.F., Endogenous nitric oxide is required for tonic cholinergic inhibition of spinal mechanical transmission, Pain, 54 (1993) 71-78.