UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Intravenous self-administration of midazolam in the rat : behavioral and neurochemical characterization Finlay, Janet Mary 1989

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-UBC_1989_A1 F56.pdf [ 5.09MB ]
Metadata
JSON: 831-1.0098208.json
JSON-LD: 831-1.0098208-ld.json
RDF/XML (Pretty): 831-1.0098208-rdf.xml
RDF/JSON: 831-1.0098208-rdf.json
Turtle: 831-1.0098208-turtle.txt
N-Triples: 831-1.0098208-rdf-ntriples.txt
Original Record: 831-1.0098208-source.json
Full Text
831-1.0098208-fulltext.txt
Citation
831-1.0098208.ris

Full Text

1 I N T R A V E N O U S S E L F - A D M I N I S T R A T I O N O F M I D A Z O L A M IN T H E RAT: B E H A V I O R A L AND N E U R O C H E M I C A L C H A R A C T E R I Z A T I O N By J A N E T M A R Y F I N L A Y B.Sc. (Honours), University of Victoria, 1982 A THESIS S U B M I T T E D IN P A R T I A L F U L F I L L M E N T O F T H E R E Q U I R E M E N T S F O R T H E D E G R E E O F D O C T O R OF PHILOSOPHY in T H E F A C U L T Y O F G R A D U A T E STUDIES (Graduate Program in Neuroscience) We accept this thesis as conforming to the required standard T H E UNIVERSITY OF BRITISH C O L U M B I A October 1989 (c); Janet Mary Finlay, 1989 6 ' £ In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. 1 further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. (Department oi)/j^^U^jti, P^yj^ ,„j U The University of British Columbia Vancouver, Canada Date Oct o/r /7<r7 DE-6 (2/88) ii Abstract Benzodiazepines are one of the most commonly prescribed classes of psychoactive compounds. Given the widespread use of these drugs in clinical populations, concern has developed over the extent of their abuse potential. Experimental analysis of the abuse potential of benzodiazepines, using animal models of drug self-administration, has produced inconsistent results. The lack of robust evidence for self-administration of benzodiazepines has been interpreted as reflecting a limited reinforcement efficacy of benzodiazepines in the rat. However, the failure to observe reliable self-administration of benzodiazepines in rats may be a function of the experimental protocols that have been employed to date. Previous experiments have generally used long-acting benzodiazepines while concurrently permitting the subjects limited access to the drug. Accordingly, the present experiments assessed the extent to which naive rats would self-administer midazolam, a short-acting benzodiazepine, when given unlimited access to the drug. In addition, a series of experiments were designed to assess the role of specific neural substrates in mediating the reinforcing effects of midazolam. Naive rats given continuous drug access exhibit reliable and stable rates of responding for intravenous infusions of midazolam. Response contingent infusions of midazolam resulted in greater rates of responding than did infusions of saline (Experiment I). Midazolam also supported stable rates of responding in a two lever discriminated responding paradigm. The rats responded at greater rates on the reinforced lever than the nonreinforced lever and were able to track the changing reinforcement contingencies (Experiment II). Midazolam self-administration was characterized by an inverse relationship between dose and rate of responding during the first dose transfer session after which no consistent relationship was observed (Experiment III). Initially, rats exhibited a temporal pattern of responding for midazolam with peak responding occurring during the dark phase of the light:dark iii cycle (Experiments I, II, and III). However, rats given prolonged access to midazolam exhibited a stable pattern of responding across the light:dark cycle (Experiment II). Administration of Ro 15-1788 elicited an increase in self-administration of midazolam suggesting that blockade of central benzodiazepine receptors attenuates the reinforcing effects of the drug (Experiment IV). Previously, it has been suggested that all drugs of abuse elicit an increase in extracellular D A concentration in the nucleus accumbens and that the drug-induced increase in D A concentrations may mediate the reinforcing effects of these drugs. In view of this evidence, Experiments V and VI assessed the effects of midazolam on extracellular D A concentration in the nucleus accumbens of freely moving rats. Subcutaneous injections of midazolam elicited a decrease in extracellular D A , D O P A C and H V A concentrations (Experiments Va and VI). Repeated intravenous infusions of midazolam, delivered in a manner that mimicked the pattern of drug self-administration exhibited during peak responding for midazolam, also elicited a decrease in extracellular D A concentrations in the nucleus accumbens (Experiment Vb). In conclusion, the present experiments indicate that midazolam is reinforcing in naive rats given unlimited access to the drug and that the reinforcing effects may be mediated by central benzodiazepine receptors. In contrast to other drugs of abuse, the reinforcing effects of midazolam are not mediated by an increase in the activity of mesoaccumbens D A neurons. iv Table of Contents Abstract ii Table of Contents iv List of Tables vii List of Figures viii Acknowledgements x P A R T I. Characterization of Intravenous Self-Administration of Midazolam in the Rat Introduction 1 A. Self-Administration of Benzodiazepines in Humans i. Normal Human Subjects 1 ii. Psychiatric Patients 3 iii. Insomniacs 5 iv. Multiple Drug Abusers 5 v. Therapeutic Dose Benzodiazepine Users 7 B. Self-Administration of Benzodiazepines in Nonhuman Primates i. Intragastric Self-Administration 8 ii. Intravenous Self-Administration 9 C. Self-Administration of Benzodiazepines in Rats i. Oral and Intragastric Self-Administration 14 ii. Intravenous Self-Administration 19 E X P E R I M E N T I. Self-Administration of Midazolam versus Saline 21 Methods 22 Results 24 Discussion 32 E X P E R I M E N T II. Midazolam as a Reinforcer in a Discriminated Responding Paradigm 34 Methods 34 Results 35 Discussion 42 V E X P E R I M E N T III. Effects of Reinforcer Magnitude on Responding for Midazolam 43 Methods 44 Results 44 Discussion 49 P A R T II. Neurochemical Substrates of Midazolam Self-Administration Introduction 54 A . The GABA/Benzodiazepine-Chloride Ionophore i. G A B A 55 ii. The G A B A A Receptor 55 iii. The Benzodiazepine Binding Site 56 B. The Dopamine Hypothesis of Reward i. Anatomical Evidence 57 ii. Electrophysiological Evidence 57 iii. Biochemical Evidence 59 iv. Behavioral Evidence 60 E X P E R I M E N T IV. Effects of the Benzodiazepine Antagonist Ro 15-1788 on Intravenous Self-Administration of Midazolam 61 Methods 61 Results 63 Discussion 63 E X P E R I M E N T V . Effects of Midazolam on Extracellular Dopamine Concentrations as assessed by In Vivo Electrochemistry 66 Experiment Va Subcutaneous Midazolam Injections Methods 67 Results 71 Experiment Vb. Intravenous Infusions of Midazolam Methods 76 Results 77 E X P E R I M E N T VI. Effects of Midazolam on Extracellular Dopamine and Metabolites as Assessed by In Vivo Microdialysis 77 Methods 77 Results 84 Discussion of Experiments V & VI 92 vi G E N E R A L DISCUSSION 94 Intravenous Self-Administration of Benzodiazepines: Evidence of Reinforcing Effects 95 Intravenous Drug Self-Administration Paradigms as Experimental Models of Human Benzodiazepine Abuse 96 Is Midazolam Self-Administration a Result of Positive Reinforcement or Negative Reinforcement? 97 Neurochemical Substrates Mediating the Reinforcing Effects of Midazolam 98 The Role of the Benzodiazepine Receptor in Mediating the Reinforcing Effects of Midazolam 98 The Role of Dopamine in Mediating the Reinforcing Effects of Midazolam 99 Alternatives to the Dopamine Hypothesis of Reward 101 Summary and Future Perspectives 102 Bibliography '. 104 V l l List of Tables Table 1. A summary of the preclinical reports of Yanagita and colleagues assessing oral self-administration of benzodiazepines in monkeys (summarized from, Woods et al., 1987) Table 2. The total number of reinforced responses emitted in 24 h sessions during baseline responding for 0.05 mg midazolam per infusion and during the first session following transfer to either a) 0.0125 mg/infusion or b) 0.20 mg/infusion viii List of Figures Figure 1. The within-session pattern of midazolam self-administration as a function of the lightdark phase 27 Figure 2. The total number of reinforced responses emitted in 24 h sessions for midazolam (baseline) and saline (extinction) 29 Figure 3. The total number of reinforced responses emitted within a session during reacquisition of midazolam self-administration 31 Figure 4. The total number of responses emitted per session on a drug lever and a control lever during acquisition of discriminated responding (baseline) and subsequent reversal learning (reversal) 37 Figure 5. The within-session pattern of responding on a drug lever and a control lever during 24 h baseline and reversal sessions 41 Figure 6. The total number of responses emitted per session during acquisition (0.05 mg midazolam per infusion), dose transfer (0.0125 or 0.2 mg midazolam per infusion) and extinction (saline) 48 Figure 7. The within-session pattern of midazolam self-administration during baseline and dose transfer sessions as a function of the light:dark phase 51 Figure 8. The effect of Ro 15-1788 on the rate of midazolam self-administration within a 24 h session 65 Figure 9. The effect of midazolam (s.c.) on D A oxidation currents in the nucleus accumbens as measured by in vivo electrochemistry 73 Figure 10. The effect of midazolam (s.c), administered during the dark phase, on D A oxidation currents in the nucleus accumbens as measured by in vivo electrochemistry 75 ix Figure 11. The effect of repeated intravenous infusions of midazolam on D A oxidation currents in the nucleus accumbens as measured by in vivo electrochemistry 79 Figure 12. The effect of repeated intravenous infusions of midazolam on D A oxidation currents in the nucleus accumbens as measured by in vivo electrochemistry 81 Figure 13. The effect of midazolam (s.c.) on extracellular D A concentrations in the nucleus accumbens as measured by in vivo microdialysis 86 Figure 14. The effect of midazolam (s.c.) on extracellular D O P A C concentrations in the nucleus accumbens as measured by in vivo microdialysis 89 Figure 15. The effect of midazolam (s.c.) on extracellular H V A concentrations in the nucleus accumbens as measured by in vivo microdialysis 91 X Acknowledgments I would like to thank my supervisor, Dr. Chris Fibiger, for the encouragement, support and insights that he has provided throughout my graduate work. I also appreciate the help and support provided by Maureen Cromie, who efficiently dealt with all the administrative details associated with this dissertation. A special thanks to Carolyn Szostak for being both an excellent colleague and friend and to Danielle Wenkstern for her own special "innerface". In addition, I would like to thank Chuck Blaha and Geert Damsma for introducing me to the "in vivo techniques" and Erin Brown for his interest in and insights into the self-administration technique. A special thankyou goes to my family for their support and encouragement and in particular to my husband, Michael, who has lovingly been both my best friend and my worst critic. I would also like to thank the members of my Supervisory Committee, Drs. A . G . Phillips, E . G . McGeer and J. Sinclair for their comments and suggestions on earlier versions of this dissertation. 1 I. Characterization of Intravenous Self-Administration of Midazolam in the Rat. Benzodiazepines were introduced into clinical practice in 1960 when chlordiazepoxide (Librium) was approved for use by humans. In the past 30 years, approximately 35 benzodiazepines have been introduced and are now used extensively for their anxiolytic, muscle relaxant, sedative and anticonvulsant properties. As such, the benzodiazepines have become one of the most commonly prescribed classes of psychoactive compounds (Greenblatt and Shader, 1981). Given the widespread use of benzodiazepines in clinical populations, concern has developed over the extent to which they may have abuse potential (for review, see Woods et al., 1987). A. Self-Administration of Benzodiazepines in Humans Self-administration of benzodiazepines in humans has been assessed in normal human subjects, psychiatric patients, insomniacs, multiple drug abusers and therapeutic dose benzodiazepine users. To follow is a review of the research assessing the benzodiazepine self-administration in each of these human populations. /'. Normal Human Subjects Johanson and colleagues have assessed self-administration of benzodiazepines in normal humans (primarily students) using a double-blind choice procedure. During the first 4 days of each experiment, subjects received color-coded tablets of diazepam or placebo on alternate days. On the following five days, individuals were given the opportunity to select and ingest the color of tablet they preferred. Under these conditions, normal humans selected placebo and 2 mg of diazepam equally and exhibited a preference for placebo over 5 and 10 mg of diazepam (Johanson and Uhlenhuth, 1980). In addition this group has reported that normal humans fail to exhibit a preference for lorazepam over placebo. Specifically, given a choice between lorazepam and placebo, subjects chose placebo more often than either 1 or 2 mg of 2 lorazepam and equally as often as 0.5 mg of lorazepam (de Wit et al., 1983). Similarly, individuals given a choice between flurazepam and placebo, selected 15 mg of flurazepam as often as placebo and 30 mg of flurazepam less often than would be expected by chance (de Wit et al., 1984). Together, these experiments demonstrate that normal human subjects do not exhibit a preference for diazepam, lorazepam or flurazepam over placebo, suggesting that under those experimental conditions the drugs were not reinforcing. These experiments suggest that normal humans fail to exhibit a preference for benzodiazepines. However, there were individuals within these groups that exhibited a preference for benzodiazepines. As a result, de Wit et al. (1986a) addressed the possibility that preference for diazepam might vary with individual differences in subjective response to the drug. The subjective effects of diazepam (10 mg) in consistent choosers of diazepam were compared to the subjective effects in consistent non-choosers of the drug. The results indicate that individuals not exhibiting a preference for diazepam experienced appreciable sedative effects whereas the individuals that preferred diazepam exhibited negligible subjective drug effects. The authors also reported that the minority of subjects that chose diazepam (10 out of 37) over placebo did not report subjective effects such as euphoria that are predictive of abuse potential. The use of normal humans in assessing the abuse potential of benzodiazepines may have biased the results of the above studies. For example, in all of these experiments, individuals (primarily students) received the drug in the morning and were then required to carry on with their usual daily activities. Under these conditions, the sedative effects of benzodiazepines may have been aversive. In an attempt to examine this question, the influence of the time of drug administration on drug preference was formally examined. It was found that subjects receiving drug during either the morning or the late afternoon chose 5 mg of diazepam as often as 3 placebo and preferred placebo to 10 mg of diazepam (de Wit et al., 1985). It was also found that age did not influence preference for diazepam in normal humans. Specifically, individuals aged 21-35 yrs and 40-55 yrs both failed to exhibit a preference for diazepam (5 or 10 mg)(de Wit et al., 1985). Together, these experiments again suggest normal human subjects do not prefer benzodiazepines over placebo. de Wit et al. (1986b) subsequently assessed whether diazepam is more likely to be preferred by humans for whom the drug may have therapeutic value. They examined whether anxious individuals (based on the DSM III criteria for generalized anxiety disorder), who would likely be prescribed benzodiazepines, exhibited a greater preference for diazepam (5 and 10 mg) than normal control subjects. Anxious humans did not choose diazepam any more often than did the normal control subjects and both groups chose diazepam less often than placebo, particularly at the high drug dose. Thus, neither of the groups exhibited a preference for diazepam although the drug produced a significant decrease in anxiety scores in both groups. This research fails to support the hypothesis that excessive benzodiazepine use may reflect an attempt to self-medicate. Iii summary, these data indicate that normal human subjects do not prefer benzodiazepines over placebo and suggest that at high doses benzodiazepines may actually be aversive. In addition, factors such as time of drug administration, subject age and level of anxiety do not influence the apparent lack of preference for benzodiazepines. //. Psychiatric Patients The results of the experiment by de Wit et al. (1986) suggest that there is no relationship between the level of anxiety and preference for diazepam. However, there is evidence of a positive correlation between the level of anxiety and preference for benzodiazepines in populations of psychiatric patients. Winstead et al. (1974) allowed psychiatric patients free access to diazepam for 6 months. The patients had been 4 diagnosed as suffering from psychosis, neurosis or character disorder. Under these conditions, drug self-administration in all groups of patients was correlated with the level of anxiety. The requests for diazepam by individual subjects was highest during the first quarter of hospitalization (when anxiety would be highest) and declined over the subsequent 6 months. Similarly, Balmer et al. (1981) used a double-blind procedure to assess self-administration of diazepam in a group of psychiatric outpatients over a period of 6 months. The majority of the patients were diagnosed as suffering from neurotic disorders. They were divided into two groups, given tablets of diazepam (5 mg) or placebo, and instructed to take up to 6 tablets per day when they felt anxious or nervous. Although both placebo and diazepam intake declined over the course of the experiment, the decline was more marked in the placebo group than the diazepam group. It was also apparent that the patient population could be divided into groups exhibiting high and low tablet intake and the diazepam group had a significantly greater proportion of patients exhibiting high tablet consumption. In the diazepam group there was a positive correlation between the intensity of anxiety and drug self-administration; in contrast, there was no correlation between the severity of anxiety and tablet intake in the placebo group. In contrast to the experiment of de Wit et al. (1986), these results suggest that psychiatric patients can exhibit a preference for benzodiazepines and that this preference is correlated with the anxiety level of the patient. Hubbard and Kripke (1976) surveyed the pharmacy records of former psychiatric patients to determine whether patients who used hypnotics (flurazepam and/or chloral hydrate) as inpatients were more likely to use these drugs as outpatients than were subjects who had not previously received hypnotics. Patients were diagnosed as suffering from major psychoses, bipolar and unipolar affective disorders, chronic alcoholism or character disorders. Those who used hypnotics as inpatients were no more likely to use these drugs as outpatients than were subjects who had previously not 5 received hypnotics. However, survey reports such as this are difficult to control in that it is uncertain whether the pharmacy records accurately reflect drug self-administration. In the above experiment, pharmacy records were confirmed by contacting 38% of the patients and for all of these subjects the records were found to be indicative of drug use. The issue as to whether benzodiazepines are preferred by subjects for whom the drugs have a therapeutic value remains controversial and represents an important area for future research. iii. Insomniacs Two experiments have examined preference for benzodiazepines in patients suffering from insomnia. In both experiments individuals were given either placebo or a benzodiazepine as a sleeping aid and asked to report which treatment they preferred. Jick et al. (1966) found that subjects preferred flurazepam (15 mg) over placebo. Similarly, Fabre et al. (1976) reported that 0.5 mg triazolam was preferred over both placebo and 30 mg flurazepam whereas 0.25 mg triazolam was not preferred over 15 mg flurazepam. The significance of these experiments to the present analysis of drug self-administration behavior is limited because: 1) the information about drug preference was the result of a verbal report of preference, rather than a direct assessment of drug-taking behavior and 2) the question of preference was asked with regard to the usefulness of the treatment as a sleeping aid. iv. Multiple Drug Abusers In a series of double-blind experiments, Griffiths and collaborators have examined benzodiazepine self-administration in individuals with documented histories of sedative abuse. In their first experiment (Griffiths et al., 1979), individuals were required to cycle on a stationary bicycle in order to attain tokens that could be exchanged for tablets of diazepam (10 or 20 mg) or placebo; a maximum of 10 tablets 6 could be attained in the 7.5 h daily sessions. Diazepam was found to support higher rates of self-administration than placebo with the highest dose of diazepam supporting the highest rates of responding. However, it should be noted that diazepam self-administration declined over the 10 days of the experiment. Using a similar experimental procedure, Griffiths et al. (1976) assessed the effects of interingestion interval and dose on diazepam self-administration in sedative abusers. Increases in the interingestion interval (0, 15 or 30 min) resulted in a decrease in the number of ingestions of diazepam (10 mg), whereas increases in the dose per ingestion (2, 5 or 10 mg) produced increases in the number of drug ingestions. The results of these experiments suggest that, in sedative abusers, diazepam (10 and 20 mg) supports higher rates of self-administration than placebo and that self-administration of diazepam (2 to 10 mg) is sensitive to both interingestion interval and dose. Griffiths and collaborators have also used a double-blind choice procedure to examine drug preference in sedative abusers. In these experiments, sedative abusers were first exposed to two different tablets and subsequently given the opportunity to choose between the two alternatives. Using this procedure, the subjects receiving diazepam (200 mg) or placebo subsequently exhibited a preference for diazepam over placebo (Griffiths et al., 1980). In the same experiment, those individuals exposed to two of several different doses of diazepam (50, 100, 200, 300 or 400 mg) failed to exhibit a consistent preference for any dose of diazepam (Griffiths et al., 1980). Healey and Pickens (1983) have also examined dose preference for diazepam in a group of sedative abusers. In this experiment, the individuals were given ad libitum access to diazepam (5 or 20 mg; depending on the subject) for at least 4 days prior to the start of preference testing. To determine dose preference, the sedative abusers were given free access to the standard dose of diazepam (5 or 20 mg) and one of several doses (2, 5, 10, 20 or 40 mg) of diazepam; the only restriction on drug self-administration 7 was a 30 min time out interval imposed following a selection. Sedative abusers failed to exhibit a preference for any one dose of diazepam. Griffiths et al. (1984) have examined the relative preference for placebo, diazepam and oxazepam in sedative abusers, using the double-blind choice procedure. Individuals self-administered 80 and 160 mg diazepam more frequently than 480 mg oxazepam (62 and 92% of choices, respectively); 40 mg diazepam was chosen less often than 480 mg oxazepam. Both diazepam (40, 80, 160 mg) and oxazepam (480 mg) were preferred to placebo. These results indicate that sedative abusers prefer diazepam and oxazepam to placebo and suggest that diazepam may have a higher abuse potential than oxazepam, the latter observation perhaps being related to the more rapid onset of action of diazepam. Together, the results of these experiments indicate that sedative abusers consistently prefer benzodiazepines over placebo and yet fail to exhibit a preference for specific drug doses. In addition, there are preliminary data to suggest that, within the drug class, some benzodiazepines may be preferred over others. v. Therapeutic Dose Benzodiazepine Users It has been suggested that prolonged exposure to therapeutic doses of benzodiazepines can result in dependence (Owen and Tyrer, 1983; Petursson and Lader, 1984). Recently, interesting experimental evidence has been provided in support of this proposal. Busto et al. (1986) have reported that patients referred for assessment and treatment of benzodiazepine abuse and/or dependence appeared to form two subgroups: patients who used only benzodiazepines for long periods of time (56%) and patients who used benzodiazepines in the context of multiple drug abuse (44%). The patients using only benzodiazepines reported drug use that was within the therapeutic dose range (15 mg of diazepam or its equivalent daily) whereas multiple drug abusers used higher doses of benzodiazepines (40 mg of diazepam or its equivalent daily). Subsequently, the patients exhibiting long term use of therapeutic doses of 8 benzodiazepines participated in a double-blind experiment examining the reinforcing effects of benzodiazepines after withdrawal from therapeutic doses of benzodiazepines (Cappell et al., 1987). During withdrawal, the patients received either placebo or diazepam and the frequency of supplemental benzodiazepine use was assessed. Determination of plasma benzodiazepine concentrations indicated a greater frequency of supplementation in patients given placebo relative to individuals given diazepam (84% versus 33% of the subjects, respectively), supporting the proposal that long-term use of low doses of benzodiazepines results in drug dependence. Thus, it appears that benzodiazepines may be reinforcing both in high dose sedative abusers (previous section) and therapeutic dose benzodiazepine users. B. Self-Administration of Benzodiazepines in Nonhuman Primates Studies assessing benzodiazepine self-administration in nonhuman primates have employed either intragastric or intravenous routes of administration. Although oral intake of benzodiazepines has been extensively examined in rats (see below), this has not been a useful approach with primates. Primates appear to be suspicious of adultered food or water and will not consume water or food that tastes or smells unusual (Altshuler and Phillips, 1978). Therefore, to assess the abuse potential of benzodiazepines in primates a chronic cannula is implanted either intragastrically or intravenously and drug injections are contingent upon the primate performing a lever press response. The rate of lever press responding for the benzodiazepine is compared to that attained when responding results in a vehicle infusion. i. Intragastric Self-Administration Yanagita and collaborators have published a series of experiments evaluating the abuse potential of benzodiazepines in nonhuman primates, using an intragastric self-administration paradigm. Yanagita and Takahashi (1973) first examined the ability of 9 chlordiazepoxide and oxazolam (both 10 mg/kg/infusion*) to support lever pressing in monkeys. They found that chlordiazepoxide supported responding in monkeys (n=2) at rates greater than those attained with saline. In contrast, oxazolam supported responding in 2 of 4 monkeys but the drug intake was not high. Yanagita (1981) also found that intragastric nitrazepam (1.0) did not support responding at rates greater than those attained with saline infusions. Intragastric diazepam (0.25, 1.0 or 2.0) maintained responding, at rates that were only marginally greater than those maintained by vehicle, in 3 of 4 monkeys (Yanagita, 1983). In addition to this work, Yanagita and colleagues have examined the intragastric intake of a number of other benzodiazepines in monkeys (for review see, Woods et al., 1987). The results of these experiments are summarized in Table 1 in which the benzodiazepines are listed on the basis of their ability to support responding at rates that are above those attained for saline. Unfortunately, procedural details as well as details about drug doses are missing from a report by Weaver et al. (1975) in which it was suggested that diazepam does not support intragastric self-administration. Similarly, few details are provided in a report indicating that intragastric chlordiazepoxide supports responding in monkeys while diazepam is less likely to do so (Altshuler and Phillips., 1978). In summary, because benzodiazepines do not consistently support lever pressing in all subjects, the results of these experiments suggest that intragastrically available benzodiazepines are at best moderately reinforcing in primates (Yanagita, 1981; Yanagita and Takahashi, 1973). However, it appears that these drugs may form a continuum with some being more likely to support responding than others (i.e., see Table 1). ii. Intravenous Self-Administration In the first published report on intravenous self-administration of benzodiazepines in primates Findley et al. (1972) used a complex procedure to train *unless otherwise noted, all drug concentrations are presented as mg/kg/infusion 10 Table 1. A summary of the preclinical reports of Yanagita and colleagues assessing oral self-administration of benzodiazepines in monkeys (summarized from, Woods et al., 1987) Drug Dose (mg/kg/injection) Number of Monkeys^ H i g h 3 Chlordiazepoxide 10.0 2/2 Triazolam 0.015 to 0.25 3/4 Clobazam 0.25 to 2.0 3/4 Flutoprazepam 0.015 to 0.06 3/4 Moderate Lorazepam 1.0 to.2.5 2/4 & 1/3 2 Alprazolam 0.0075 to 0.03 2/4 Flunitrazepam 0.25 or 0.06 2/4 Lormetazepam 4.0 to 16.0 2/5 Nitrazepam 1.0 1/4 & 2/4 2 Oxazolam 10.0 2/4 Fludiazepam 1.0 1/4 Halazepam 5.0 3/4 & 1/4 2 Quazepam^ 1/4 Diazepam'* 2/4 L o w 3 Nimetazepam 1.0 or 5.0 0/4 Haloxazolazepam 0.5 or 2.0 0/4 Cloxazolazepam 1.0 to 10.0 0 Estazolam 1.0 to 5.0 0 Prazepam 0.25 and 1.0 0 Ethyl Loflazepate 0.015 to 0.25 0 Mexazolam 0.2 to 5.0 0 Clonazepam 0.125 and 0.25 0 *the number of monkeys in which the benzodiazepine supported responding at rates greater than saline over the total number of monkeys tested the number of monkeys in the original investigation and a replication 3ability to support self-administration behavior ^information on drug doses not available 11 primates to self-administer chlordiazepoxide (1.0). In order to avoid a shock, the animals were required to complete either of two fixed ratio (FR) schedules of responding on a lever. Initially, completion of both of the F R requirements resulted in an infusion of saline. When chlordiazepoxide was then substituted for saline under one of the F R conditions, the monkeys exhibited a preference for self-infusion of chlordiazepoxide over saline. Since this report, the abuse potential of intravenous benzodiazepines in primates has typically been assessed under one of two conditions: 1) in drug experienced monkeys using a drug substitution procedure and 2) in naive monkeys given unlimited drug access. In most cases, intravenous self-administration of benzodiazepines in nonhuman primates has been evaluated within the context of a drug substitution procedure. In this procedure, monkeys are trained to self-administer either a psychomotor stimulant or a barbiturate and a benzodiazepine is then substituted for the reference drug. Using a substitution procedure, Hackett and Hall (1976) trained primates to respond for codeine. When diazepam (0.05 to 0.40) was substituted for codeine, the benzodiazepine did not support responding at rates greater than those obtained with saline. Similarly, in monkeys trained to respond for cocaine, chlordiazepoxide (0.001 to 0.1) and clorazepate (0.003 to 0.1) failed to support self-administration at rates greater than those obtained with saline (Balster and Woolverton, 1982). In contrast, Bergman and Johanson (1985) reported that diazepam (0.012 to 0.40) maintained responding at rates that were greater than those obtained with saline in 3 of 11 monkeys that had previously self-administered cocaine. Responding for diazepam was maintained in a dose-response fashion with the lowest dose supporting the highest rate of responding. Similarly, in a paper that is lacking in procedural details concerning the schedule of reinforcement and the duration of drug access, Hoffmeister (1977) reported that diazepam (0.005, 0.010 and 0.05) supported responding at rates greater than those obtained with saline in monkeys initially responding for codeine. 12 However, neither 0.5 mg/kg diazepam per infusion nor chlordiazepoxide (0.05, 0.10 and 0.5) maintained responding. Griffiths and colleagues have also used a substitution procedure to examine self-administration of benzodiazepines in primates. Under the conditions of their experiments, primates were trained to respond for intravenous infusions of cocaine. Subsequently, clonazepam (0.01 to 10.0), clorazepate ( 0.01 to 5.6), diazepam (0.032 to 17.8), flurazepam (0.1 to 30.0), medazepam (0.01 to 10.0), midazolam (0.032 to 10.0) or triazolam (0.0001 to 0.32) was substituted for cocaine (Griffiths et al., 1981; Lukas and Griffiths, 1982). Clonazepam, clorazepate, diazepam, flurazepam and medazepam were found to support modest rates of self-administration that were not consistently above vehicle rates in all animals. However, both midazolam and triazolam supported consistently higher rates of self-administration than saline. On the basis of these experiments, it appears that the short-acting benzodiazepines such as midazolam and triazolam support higher rates of responding than do the longer acting compounds. The animal's previous drug history may influence drug self-administration in the substitution procedure. For example, Bergman and Johanson (1985) found that when monkeys had previously been responding for pentobarbital, diazepam (0.0025 to 0.4) supported responding above vehicle rates in each of the 5 animals tested. In contrast, only 3 of 11 monkeys that had previously been self-administering cocaine exhibited rates of responding for diazepam that were greater than those obtained with saline infusions. Differences in previous drug history may account for some of the discrepancies that have been reported using the substitution procedure. Pentobarbital has been used as a training drug prior to evaluating benzodiazepine self-administration in two other studies. Johanson (1987) examined the ability of flurazepam (0.003 to 1.0), lorazepam (0.003 to 1.0) and estazolam (0.003 to 0.3) to support responding in monkeys trained to self-administer pentobarbital. Each drug maintained responding above vehicle and saline in the majority of monkeys tested. In 13 addition, Kubota et al. (1986) have assessed the ability of triazolam, midazolam, chlordiazepoxide and flurazepam to support responding using a substitution procedure. Monkeys were trained to respond for pentobarbital and were subsequently transferred to triazolam (0.001 to 0.03), midazolam (0.01 to 0.03), chlordiazepoxide (0.1 to 1.0) or flurazepam (1.0 and 3.0). Both triazolam and midazolam maintained responding at rates that were greater than those obtained with saline. Chlordiazepoxide did not consistently maintain responding at rates that were above those obtained with saline and flurazepam did not maintain responding at all. Given that previous drug history may influence self-administration of benzodiazepines, it is important to evaluate drug self-administration in naive animals. There are two published reports in which the ability of benzodiazepines to support responding in naive primates given unlimited access to the drug has been examined. Yanagita and Takahashi (1973) found that 3 of 4 naive monkeys initiated and maintained diazepam (0.4) self-administration at rates that were greater than those obtained with saline. In addition, 3 of 4 naive monkeys initiated and maintained self-administration of chlordiazepoxide (1.0) for 4 weeks after which responding gradually decreased to rates previously attained with saline. Kubota et al. (1986) assessed the ability of midazolam and triazolam to support responding in naive monkeys given unlimited access to the drug for 2 weeks. Midazolam (0.03 or 0.1) supported responding in 3 of 4 monkeys and triazolam (0.001 to 0.01) supported responding in 2 of 4 monkeys at rates that were significantly greater than those obtained with saline. In summary, the results of experiments assessing intravenous self-administration of benzodiazepines in primates illustrate two factors that are relevant to intravenous benzodiazepine self-administration: 1) previous drug history can significantly influence subsequent drug intake; and 2) as with intragastric benzodiazepine self-administration, certain benzodiazepines may be more likely than others to support operant responding in monkeys. 14 C. Self-Administration of Benzodiazepines in Rats Self-administration of benzodiazepines in rats has been evaluated most extensively using oral self-administration paradigms in which naive rats are simply given access to water containing the drug. There are few experiments in which the intragastric route of drug administration has been used and only recently have researchers begun to assess intravenous self-administration of benzodiazepines in rats. /. Oral and Intragastric Self-Administration In general, two approaches have been used to assess oral self-administration of benzodiazepines in rats: 1) the intake of benzodiazepines has been measured in naive rats given a choice between a drug solution and water; and 2) paradigms have been developed that force or induce consumption of large volumes of a benzodiazepine solution and subsequently a preference for benzodiazepine over water is examined in a choice procedure. In a choice procedure, naive rats given unlimited access to chlordiazepoxide (0.5 or 0.2 mg/ml*) failed to exhibit a preference for chlordiazepoxide over water (Harris et al., 1968; Kamano and Arp, 1965). Amit and Cohen (1974) have reported that rats given 24 h access to diazepam exhibited a preference for a low dose of diazepam (0.015 mg/ml) over water but did not prefer a wide range of higher drug doses (0.03 to 0.18). In contrast, Fuchs et al. (1984) found that rats given 22 h access to diazepam (0.1) exhibited a strong preference for diazepam over water to the extent that 9 of 20 rats consumed 82% of their fluid intake as diazepam whereas the remaining 11 rats consumed between 20-50% of their fluid intake as diazepam. Thus, attempts to establish oral self-administration of benzodiazepines in a choice procedure have met with limited success. Since benzodiazepines are taken orally by humans, oral self-administration of benzodiazepines in rats has face validity as a technique for determining their abuse potential. However, low levels of oral *unless otherwise noted, all drug concentrations are presented as mg/ml 15 benzodiazepine self-administration may be unrelated to the abuse potential of the drug. For example, both chlordiazepoxide and diazepam solutions have a bitter taste (at least to humans) and therefore the drug solutions may have been rejected on the basis of taste. Attempts have been made to assess the gustatory and postingestional effects of chlordiazepoxide and diazepam in rats. Wolf et al. (1978) trained rats to consume all of their daily water intake in discrete 1 h sessions. They were then given chlordiazepoxide (0.2, 0.4 or 0.8) on three consecutive sessions. On the first drug session rats given a chlordiazepoxide solution of 0.2 mg/ml drank 100% of their normal daily fluid intake whereas those rats given 0.4 and 0.8 mg/ml chlordiazepoxide consumed roughly 80 and 60%, respectively, of their normal daily fluid intake. One explanation for the dose dependent decrease in fluid intake is that chlordiazepoxide may have a dose dependent gustatory aversive effect. In addition, intake of all drug doses was further decreased on the subsequent 2 sessions suggesting that chlordiazepoxide had aversive postingestional effects as well as aversive gustatory effects. Walton and Deutsch (1978) gave 2 groups of water deprived rats the opportunity to self-administer diazepam either orally (water was counterinjected intragastrically) or via intragastric counterinjection (water was consumed orally). These rats were given access to water followed by increasing concentrations of diazepam (0.25, 0.50 or 1.0). Fluid intake was stable across all conditions suggesting that diazepam has neither aversive gustatory nor postingestional effects. This limited body of evidence suggests that while chlordiazepoxide may have both aversive gustatory and postingestional effects (Wolf et al., 1978), diazepam does not (Walton and Deutsch, 1978). In this regard, it is interesting that, while rats do not exhibit a preference for chlordiazepoxide over water (Harris et al., 1968; Kamano and Arp, 1965), there is some suggestion of a preference for diazepam over water (Amit and Cohen, 1974; Fuchs et al., 1984). 16 As a result of the apparent reluctance of rats to consume significant quantities of benzodiazepines orally, a variety of techniques have been developed to induce or force drug consumption in this species. In this instance, a preference for drug or water is examined in a choice procedure after a period of forced drug consumption. The five techniques that have been used to induce consumption of benzodiazepines in rats include: restricting fluid intake to the drug solution (water deprivation), foot shock stress, stimulation of the lateral hypothalamus, making food reinforcement contingent upon drug consumption, and a scheduled feeding paradigm. Harris et al. (1968) observed that rats forced to consume chlordiazepoxide (0.5) by making it the only fluid available, failed to exhibit a preference for the drug when given a choice between chlordiazepoxide and water. Stolerman and collaborators (1971) subjected rats to an alternating schedule of 2 days forced access to chlordiazepoxide (0.5) followed by a day of choice between drug or water. This pattern continued for 46 days. Over the course of repeated choice trials rats exhibited a slight increase in drug consumption although they failed to exhibit a preference for chlordiazepoxide over water. Thus, restricted fluid intake does not appear to result in a preference for chlordiazepoxide over water. Kamano and Arp (1965) examined the effects of both shock avoidance and inescapable foot shock on free and forced intake of chlordiazepoxide (0.2). Rats given a choice of water and chlordiazepoxide and subjected to shock avoidance, failed to exhibit a preference for chlordiazepoxide. When these rats were given access to only water or chlordiazepoxide and subjected to shock avoidance, they consumed equal amounts of drug and water. However, when the rats were again given a choice between drug and water and subjected to inescapable shock, they failed to exhibit a preference for drug even though they had presumably experienced the anxiolytic effects of the drug under the previous conditions of the experiment. 17 Amit and Cohen (1974) found that stimulation of the lateral hypothalamus failed to result in a preference for a "mildly aversive" concentration of diazepam (0.06). Similarly, under a schedule of ad libitum access and periodic withdrawal of the drug (i.e., one day choice between water and drug and alternate day only water) there was no sign of an increase in diazepam intake across sessions. In another experiment it was found that if rats were trained to drink a solution of chlordiazepoxide in order to attain food reinforcement, they continued to exhibit high levels of chlordiazepoxide consumption (40% of fluid intake) when presentation of food was no longer contingent upon drinking (Harris et al., 1968). It is important to note, however, that the rats did not actually exhibit a preference for chlordiazepoxide over water. Oral self-administration of benzodiazepines has also been examined under conditions of schedule-induced polydipsia (SIP). The procedure in these experiments has been to food deprive rats and deliver food pellets at some fixed interval, usually one pellet per minute (FI-1 min), during daily sessions (1-3 h). Under these conditions, rats that are not water deprived consume large quantities of fluid during the period of food delivery. Sanger (1977) reported that rats given access to chlordiazepoxide (0.4 and 0.5) during daily SIP sessions consumed large quantities of the drug. In fact, those rats given access to the low dose of chlordiazepoxide consumed more fluid than rats given access to water during scheduled feeding. Falk and Tang (1985; 1987) have used a similar paradigm to assess oral intake of midazolam in rats. In this case, 2 groups of rats were subjected to a schedule of food delivery and given access to either water or increasingly concentrated solutions of midazolam. It is noteworthy that the rats given access to midazolam exhibited a preference for low concentrations of midazolam relative to those animals given access to water. However, the concentration of midazolam was increased until the rats were consuming approximately equal volumes of midazolam (0.05 mg/ml) or water. During 18 a subsequent test for preference, rats that had been consuming either water or midazolam both exhibited a significant preference for midazolam (0.05 mg/ml) when food was delivered as a single ration at the beginning of the session. In contrast, under a FI-1 or 3 min schedule of food delivery no preference for midazolam over water was evident. This experiment suggests that there are complex interactions between the schedule of food delivery and self-administration of midazolam because the schedule can influence the apparent preference for the drug. Together, these experiments suggest that, under the conditions of SIP, rats exhibit a preference for benzodiazepine solutions over water. However, the interpretation of these experiments is complicated by at least two factors. First, the SIP paradigm is potentially stressful and the excessive fluid intake may be a behavioral manifestation of the anxiety produced by scheduled feeding. As such, in this experimental paradigm, the preference for benzodiazepine solutions over water may be related to the anxiolytic effects of the drug solution. To determine the role of anxiety in SIP, several groups have examined the effects of benzodiazepines on the excessive water consumption induced by scheduled-feeding. Benzodiazepines have been reported to increase (Sanger and Blackman, 1976), decrease (McKearney, 1973) or have no effect (Smith and Clark, 1975) on SIP. Clearly, the role of anxiety in SIP remains controversial. Second, interpretation of these data is further complicated by the fact that benzodiazepines themselves have direct effects on food and water consumption (Cooper and Estall, 1985). Gotestam (1973) has reported that intragastric delivery of medazepam (2.5, 5.0, 10.0 mg/kg/injection) results in higher rates of responding than for saline in naive rats. In contrast, Walton and Deutsch (1978) found that water deprived rats given access to intragastric water fail to increase their fluid consumption when transferred from water to increasing concentrations of diazepam (0.25, 0.50 and 1.0), suggesting that diazepam is no more reinforcing than water under these conditions. A n important difference 19 exists between these two experiments that may explain this discrepancy; the first experiment used non-deprived rats as subjects whereas in the second experiment the rats were water-deprived. As a result, the lack of a preference for diazepam in the latter experiment may be related to a ceiling effect on fluid intake in rats highly motivated to consume fluid. The results of the previous experiments may be summarized as follows. First, rats fail to exhibit a consistent preference for benzodiazepines over water in a free choice procedure. Specifically, rats fail to exhibit a preference for chlordiazepoxide over water (Harris et al., 1968; Kamano and Arp, 1965) and yet there is some evidence that rats prefer orally available diazepam over water (Amit and Cohen, 1974; Fuchs et al., 1984). Second, while oral intake of chlordiazepoxide can be induced by water deprivation, foot shock or by making food delivery contingent upon drug consumption, rats fail to exhibit a preference for chlordiazepoxide in a subsequent test of preference (Harris et al., 1968; Kamano and Arp, 1965; Stolerman et al., 1971). Third, during periods of scheduled feeding, rats given access to solutions of chlordiazepoxide or midazolam may consume more fluid than control animals given access to water (Falk and Tang, 1985; 1987; Sanger, 1987). Fourth, naive rats given unlimited access to intragastric medazepam exhibit a preference for the benzodiazepine over water (Gotestam, 1973) whereas water deprived rats fail to exhibit a preference for diazepam over water (Walton and Deutsch, 1978). ii. Intravenous Self-Administration There have been few attempts to assess intravenous self-administration of benzodiazepines in rats. The published experiments consist of one report of intravenous self-administration of benzodiazepines under conditions of induced consumption, using a scheduled feeding paradigm, and two reports of attempts to examine self-administration under conditions of free access in naive rats. 20 As previously noted (see previous section), schedule-induced polydipsia has been used to induce oral benzodiazepine self-administration in rats. Similarly, schedule-induced self-administration of benzodiazepines has been observed when the drug is available via intravenous infusions. Pilotto et al. (1984) reported that food deprived rats exposed to a fixed interval (FI-1 min) schedule of food delivery responded at greater rates for intravenous infusions of diazepam (0.0625 to 0.5 mg/kg/infusion) than saline. Under these conditions a dose-response curve for diazepam was evident, with those rats maintained on the lowest dose of diazepam exhibiting the highest rates of responding and rats maintained on the highest dose exhibiting the lowest rates of responding. There are two published reports in which intravenous self-administration of benzodiazepines has been assessed in naive rats. Pilotto et al. (1984) reported that rats failed to respond for diazepam (0.0625 to 0.5 mg/kg/infusion) at rates that were greater than those obtained with saline infusions. However, it is important to note that Pilotto and colleagues gave these rats limited access (1 h) to diazepam. In a study by Collins et al. (1984), naive rats were given unlimited access to either diazepam or flurazepam for 5 days. Under these conditions diazepam (0.10 mg/kg/infusion) maintained responding in 1 of the 7 rats tested. Flurazepam (0.10 to 3.2 mg/kg/infusion), however, showed positive effects in up to 50% of the rats tested depending on the drug dose. Summary Although the literature pertaining to self-administration of benzodiazepines remains controversial, some general statements can be made. First, it appears that certain populations of human subjects (i.e., mulitple drug abusers, therapeutic dose benzodiazepine users and psychiatric patients) exhibit a preference for benzodiazepines over placebo suggesting that the drugs are reinforcing. Second, nonhuman primates sometimes exhibit a preference for intragastrically and intravenously available 21 benzodiazepines. In addition, some benzodiazepines appear to be more likely than others to support self-administration behavior in primates. Third, rats do not typically exhibit a preference for orally available benzodiazepines over water, an observation that may be related to aversive gustatory or postingestional effects of the drug. Experiment I. Self-administration of Midazolam versus Saline The lack of robust evidence for benzodiazepine self-administration may reflect a limited reinforcement efficacy of benzodiazepines in the rat. The failure to observe reliable benzodiazepine self-administration in the rat may, however, also be a function of the experimental protocols that have been employed to date. Previous experiments have generally assessed oral drug intake in rats and, as suggested in the Introduction, low levels of drug intake may be a function of the aversive gustatory or postingestional effects of the drug. In addition, previous experiments have generally used long-acting benzodiazepines while concurrently permitting the subjects limited access to the drug. Accordingly, the present experiment was designed to determine the extent to which naive rats would respond for intravenous infusions of midazolam, a short acting benzodiazepine (Dundee et al., 1984; Falk and Tang, 1987; Pieri et al., 1981), when given unlimited access to the drug. Midazolam was selected for use in the present experiments because it offers advantages, relative to other benzodiazepines, in terms of its solubility, stability and pharmacokinetics. First, midazolam maleate is soluble in water and as a result the drug solutions are easy to prepare and well tolerated when administered intravenously (Dundee et al., 1984; Greenblatt and Shader, 1984; Pieri et al., 1981). In the present experiments, it was critical that the intravenous drug infusions produced minimal venous irritation since aversive effects associated with drug administration may inhibit self-administration behavior. Second, because rats were given unlimited drug access in the present experiments, it was preferable to select a drug that is stable in solution for 22 at least 24 h; midazolam is stable in solution for over 1 month (Pieri et al., 1981). A third consideration in selecting midazolam over other benzodiazepines relates to the pharmacokinetic profile of the drug. When midazolam is buffered to physiological pH in vivo, the drug becomes highly lipid soluble and as a result crosses the blood-brain barrier rapidly leading to a fast onset of action. In the present experiments, the rapid onset of midazolam's psychoactive effects minimizes the delay between the operant response (lever pressing) and reinforcement (the drug effect) and may facilitate acquisition of self-administration behavior (for a discussion of delay of reinforcement, see Mackintosh, 1974). In addition, both pharmacokinetic and pharmacodynamic evidence suggests that the duration of action of midazolam is short; the elimination half-life of midazolam and its metabolites is estimated to be 1-3 h (Dundee et al., 1984; Falk and Tang, 1987; Greenblatt and Shader, 1984; Pieri et al., 1981). Rats may exhibit higher rates of responding for midazolam than for longer-acting benzodiazepines in attempting to maintain brain concentrations of the drug that are reinforcing. On the basis of these characteristics, midazolam was selected for use in the following experiments. Methods Naive, male Long-Evans rats (Charles River), weighing 300-400 g at the beginning of the experiment, served as subjects. The rats (n=6) were initially trained to bar press for food reinforcement. Following acquisition of the bar press response, each rat was anesthetized with Nembutal (50 mg/kg, intraperitoneally) and a chronic silastic cannula was surgically implanted into the right jugular vein. The cannula passed subcutaneously over the rat's shoulder and exited through a Marlex mesh assembly implanted subcutaneously at the mid-scapular region. The cannula was attached by polyethylene tubing and a protective spring to a swivel assembly (Brown et al., 1976) held by a counterbalance, and the swivel was, in turn, attached to an infusion pump (Razel). After surgery, each rat received an intramuscular injection 23 of 0.1 ml Combiotic (penicillin and dihydrostreptomycin in aqueous suspension, Pfizer Co.). For the remainder of the experiment, the rats were housed individually in standard operant boxes equipped with a house light, two removable response levers and a cue light located directly above each lever (BRS-LVE) . Each operant box was located within a sound-attenuating chamber equipped with a continuously operating fan. A l l experimental contingencies were programmed and data were recorded by a N O V A I V / X minicomputer (Data General) equipped with M A N X interface and software (GC Controls). The day after surgery, each rat was given access to one lever. Each response on the lever produced a 4.5 s infusion of midazolam (0.05 mg/0.18 ml distilled water). The cue light located directly above the lever was illuminated for the duration of the infusion. Responses on the lever during the infusion were recorded but had no programmed consequences. Each self-administration session lasted 23.75 h. During the 15 min period between the daily sessions, fresh drug was provided and the patency of the cannula was assessed by infusing 0.2 ml of the drug solution. In cases where patency was questionable, methohexital sodium (Brietal, Lilly Inc.), a short-acting anaesthetic, was infused. The cannula was accepted as patent if the infusion resulted in onset of anaesthesia within a few seconds. Methohexital sodium was administered infrequently (on average every 2-3 wks) and the anaesthetic did not affect the subsequent self-administration of midazolam. Food and water, which were available ad libitum for the duration of the experiment, were replenished daily and the litter (Sanicel) changed weekly during this time. A 12 h light:dark cycle was maintained by use of the houselight. Onset of the houselight occurred at the start of each session (10:00). 24 Following attainment of stable rates of responding for midazolam (4-5 consecutive sessions with no upward or downward trends in responding), physiological saline was substituted for the drug solution (i.e., extinction). Response contingencies during extinction were identical to those in effect during the acquisition phase. Testing under extinction conditions continued until rats displayed stable rates of responding for saline. Reacquisition of midazolam self-administration (0.05 mg/infusion) was assessed in four rats whose cannulae remained patent following extinction testing. Testing was as described for the initial acquisition of midazolam self-administration and continued until response rates were similar to asymptotic rates observed during the acquisition phase. Results Acquisition Stable rates of responding for intravenous midazolam were obtained within 16-27 days. As a result of variation between rats in the rate of acquisition, baseline was defined as the last 4 days of acquisition training. During the baseline sessions, rats administered approximately 10.5 ± 2.4 mg/kg midazolam per session (based on the mean number of drug infusions in a 24 h session and an average weight of 400 g). Inspection of the baseline data from individual rats suggested that rates of responding varied within a 24 h session. In particular, responding appeared to increase during the dark phase. To determine the temporal pattern of responding within a 24 h period, the number of reinforced responses per 2 h blocks was calculated, and expressed as a percent of the total number of reinforced responses emitted within a session. For purposes of statistical analyses, the mean of the 4 baseline sessions was calculated for each subject. A n analysis of variance ( A N O V A ) with 2 repeated measures (Lights On/Off and Blocks of 2 Hours) was conducted. Degrees of freedom appropriate for the Geisser-Greenhouse conservative F test were used for this and all subsequent 25 analyses involving repeated measures (Kirk, 1968). A significant main effect of Lights On/Off was obtained (F(l,5) = 41.104, p < .005), reflecting the higher rate of responding observed during the dark cycle (Figure 1). A l l rats exhibited increased responding during the dark phase, regardless of absolute rate of responding. Extinction Each rat exhibited a progressive decline in responding across extinction sessions (Figure 2). For purposes of statistical analysis, the total number of responses emitted per extinction session was expressed as a percent of each rat's mean baseline rate of responding. These data were then analyzed using an A N O V A with one repeated measure (Sessions). Since S 13 only received 6 extinction sessions, the mean of S 13's last 4 extinction sessions was calculated and used as an estimate of his performance for Sessions 7 and 8. A significant effect of sessions was obtained (F(l,5) = 24.47, p < .001). Correlated t-tests indicated that the number of responses emitted on the first day of extinction did not differ from baseline rates of responding (t(5) = 2.027, p > .05). Rates of responding on the last day of extinction were significantly lower than baseline rates (t(5) = 3.864, p < .025). Reacquisition As illustrated in Figure 3, a progressive increase in the total number of responses emitted across reacquisition sessions was observed (n=4). Rates of responding similar to baseline rates were reached by each subject (t(3) = 1.502, p > .10). During reacquisition, baseline rates of responding were attained in fewer sessions than required for the initial acquisition of midazolam self-administration (6-13 sessions versus 16-27; t(3) = 4.045, p < .05). 26 Figure 1. The within-session pattern of midazolam self-administrat ion obtained dur ing baseline sessions (mean o f the last 4 acquisit ion sessions). The mean number o f reinforced responses emitted in 2 h blocks was calculated as a percent of the total number o f reinforced responses emitted. Each data point represents the mean + S . E . M . (n=6). 25n (/) 9 n O Q_ (/) Ld Od I El Ld O 15 H 10 H 5H LIGHTS ON LIGHTS OFF T " 2 T " 3 4 5 6 7 8 9 BLOCKS OF TWO HOURS 10 28 Figure 2. The total number of reinforced responses emitted per session under baseline and extinction conditions are presented for 6 individual subjects. Reinforced responding during baseline (the last 4 days of acquisition of midazolam self-administration) is illustrated in the left panel of each graph. The number of responses per session obtained during extinction testing are presented in the right panel of each graph. N U M B E R O F R E I N F O R C E D R E S P O N S E S 63 30 Figure 3. The mean reinforced responses emitted during the 4 baseline sessions (B) and the total number of reinforced responses emitted during the last extinction session (E) are illustrated in the left panel of each graph for 4 individual subjects. The total number of reinforced responses per session obtained during reacquisition of midazolam self-administration are plotted in the right panel of each graph. N U M B E R O N S E S a Kj S a o o ' ' L_ o K o o V \ 32 Discussion The results of this experiment demonstrate that naive rats will lever press for intravenous midazolam, a short-acting benzodiazepine. Specifically, response-contingent infusions of midazolam yielded higher rates of responding than did saline infusions. In addition, rats tested for subsequent reacquisition of midazolam self-administration approached asymptotic levels of responding at a faster rate than during the initial acquisition phase. These results extend the observation that midazolam self-administration occurs when rats are tested for oral self-administration under conditions of schedule-induced polydypsia (Falk and Tang, 1985). In a recent paper, Kubota et al. (1986) assessed the ability of midazolam and triazolam to support responding in naive monkeys given unlimited access to the drug. It was found that both midazolam and triazolam supported responding at rates that were significantly greater than those obtained with saline. Together with the present experiments, these results indicate that the short-acting benzodiazepines midazolam and triazolam are reinforcing in naive animals given unlimited access to the drug. At the completion of the present experiment, it was not apparent if rats provided with unlimited drug access would respond for long-acting benzodiazepines as readily as for midazolam or if the present findings might only apply to short-acting benzodiazepines such as midazolam and triazolam. A recent experiment by Naruse and Asami (1987) suggests that long-acting benzodiazepines may also support responding under conditions of unlimited drug access. Using an experimental protocol that was virtually identical to the present study, Naruse and Asami (1987) gave rats unlimited access to diazepam for 30 days. Under these conditions, rats exhibited higher rates of responding for diazepam (0.5, 1.0 and 2.0 mg/kg/infusion) than saline. A dose response curve was evident such that the highest rates of responding were attained with 2.0 mg/kg/infusion. Previous experiments had suggested that intravenous diazepam does not support significant rates of self-administration in naive rats (Collins et al., 33 1984; Pilotto et al., 1984). In these cases, the failure to observe intravenous self-administration of diazepam may have been related to the use of highly limited (1 h) access to the drug (Pilotto et al., 1984) and/or the use of doses below those that support significant responding in rats. Naruse and Asami (1987) observed that 0.5 mg/kg of diazepam per infusion supported rates of responding that were only marginally greater than those obtained with saline. Neither Pilotto et al. (1984) nor Collins et al. (1984) used doses greater than 0.5 mg/kg/infusion in their attempts to assess intravenous self-administration of diazepam. In the present experiment, the rate of midazolam self-administration varied as a function of the light:dark cycle. Specifically, low but stable rates of responding were emitted during the light phase of the cycle, with an average of 2.15 responses per hour. During the dark phase, however, responding increased to an average of 6.15 responses per hour. While it is possible that the temporal pattern of responding reflects a general increase in activity normally observed in rats during the dark phase, the fluctuation in response rates may also be attributable to factors specific to midazolam. For example, cyclic variations within the neural systems mediating the pharmacological effects of midazolam may underly the pattern of responding observed in the present experiment. It is interesting that the number of benzodiazepine receptors within the frontal cortex has been reported to be reduced during the rats' dark phase (Brennan et al., 1985). As such, variations in response rate may be related to cyclic changes in these receptors or in the functional status of some other transmitter system within the central nervous system involved in the pharmacological effects of midazolam. The detection of a temporal pattern of responding for midazolam was serendipitous and made possible because rats had continuous access to the drug. The previously reported low rates of responding emitted for benzodiazepines by rats may be a function of the time of day that the rats were tested. That is, higher rates of responding may have been obtained had rats been tested during the dark (active) phase. 34 In this regard, the apparent ability of the benzodiazepines to maintain responding more readily in baboons and monkeys in comparison to rodents (for review, see Griffiths and Ator, 1981) may be an artifact of the time of day that the animals are tested. In conclusion, the present experiment provides preliminary evidence that the short-acting benzodiazepine midazolam is reliably self-administered by rats. To gain a more complete understanding of this phenomenon the profile of responding in a discriminated responding paradigm and the effect of dose titration was assessed in the following experiments. Experiment II. Midazolam as a Reinforcer in a Discriminated Responding Paradigm The observation that naive rats self-administer the short-acting benzodiazepine midazolam suggests that this drug can act as a primary reinforcer. If this is the case, then responding for midazolam may be influenced by factors known to influence responding for other primary reinforcers. In the present experiment the contribution of one such factor was examined, namely, the ability of rats to track changing reinforcement contingencies in a discriminated responding paradigm. Rats can learn a reinforcement contingency in which responses emitted to one stimulus (S+) result in reinforcement while responses emitted to another stimulus (S-) fail to result in reinforcement (cf., Mackintosh, 1974). Once the initial discrimination is established, the animals are then able to transfer responding to the appropriate stimulus if the reinforcement contingencies are altered. To date, the ability of rats to learn a discrimination and subsequently track the reinforcement contingency when lever pressing is reinforced with a short-acting benzodiazepine has not been examined. Methods Naive, male Long-Evans rats (Charles River), weighing 300-400 g at the beginning of the experiment, served as subjects (n=5). The experimental protocol was identical to that described in Experiment 1, with the following exceptions. The day 35 after surgery, each rat was given continuous access to two levers. The levers were randomly designated as either the drug (S+) or the control (S-) lever for each rat. Each response on the drug lever produced a 4.5 sec infusion of midazolam (0.05 mg in 0.18 ml distilled water) coincident with the illumination of the cue light located directly above the drug lever. Responses on the drug lever during an infusion were recorded but had no programmed consequences. Responses on the control lever did not produce an infusion but did result in a 4.5 sec illumination of the cue light located directly above the control lever, thereby controlling for the possibility that rats were responding for illumination of the cue light rather than a drug infusion. When stable levels of responding on both the drug and the control lever were attained (4-5 consecutive sessions with no upward or downward trends in responding), the reinforcement contingencies were reversed for each rat. During the reversal sessions, the lever formerly paired with drug administration became the control lever and the former control lever became the drug lever. Al l other response contingencies and infusion parameters were identical to those present during the acquisition of midazolam self-administration. Results Discrimination Learning During the acquisition of midazolam self-administration, stable rates of responding on the drug lever and control lever were attained within 16-30 sessions (Figure 4; left panel of each graph). Because of the variation between rats in the rate of acquisition of midazolam self-administration, the baseline phase was defined as the last 4 days of acquisition training. A n analysis of variance ( A N O V A ) with 2 repeated measures (Control or Drug Lever and Sessions) was conducted on the baseline data. Degrees of freedom appropriate for the Geisser-Greenhouse conservative F test were used for this and all subsequent analyses involving repeated measures (Kirk, 1968). 36 Figure 4. The total number of responses emitted on the drug and control lever within 24 h sessions are illustrated for individual rats (n=5). Responding during baseline sessions (the last 4 days of acquisition of midazolam self-administration) are illustrated in the left panel of each graph and responding during the first and last 4 reversal sessions are illustrated in the right panel. 37 I8O-1 160-uo-120-100-80-60-40-20-M UI tn WO-i z 90-o 80-a. tn 70-Hi 60-a 50-XL. 40-O 30-DC 20-U J 10-ffl S => z BASELINE S u b l e t 1 i V\ 2 7 2 8 2 9 3 0 31 32 3 3 3 4 49 5 0 51 52 sA S u b l e t 2 14 15 16 17 7 0 - | 6 0 -5 0 4 0 -3 0 2 0 -10 S u b | e r f 6 -i—X-13 U 15 16 17 18 19 2 0 28 29 3 0 31 S E S S I O N S too 9 0 8 0 7 0 -6 0 5 0 -4 0 3 0 -2 0 10 120 100 8 0 6 0 4 0 2 0 -Sub|ec< 10 2 7 2 8 29 3 0 31 32 3 3 3 4 3 7 3 B 39 4 0 S u b j e c t 9 18 19 2 0 21 2 2 2 3 24 2 5 4 6 4 7 4 8 S E S S I O N S BASELINE R E V E R S A L O 0 D R U G L E V E R 6 A • - • C O N T R O L L E V E R A A 38 During the baseline phase, rats exhibited consistently higher levels of responding on the drug lever than the control lever as supported by a significant main effect of Lever [F(l,4) = 10.93, p < 0.05]. Thus, the rats successfully discriminated between the reinforced lever and the nonreinforced lever. During this baseline phase, rats administered approximately 8.5 + 3.2 mg/kg midazolam per session (based on the mean number of infusions of midazolam per session and an average weight of 400 g). Reversal Learning During the reversal phase, each rat successfully transferred responding to the newly active drug lever (Figure 4; right panel of each graph). Due to the variability in the number of sessions required for responding to stabilize following the change in reinforcement contingency, data from the first and last 4 days of the reversal phase are presented for each rat. These data were analyzed using A N O V A with 2 repeated measures (Control/Drug Lever and Session). A significant Lever x Sessions interaction was obtained [F(l,4) = 21.62, p < 0.01), reflecting the fact that across reversal sessions responding on the control lever decreased while responding on the drug lever increased. The contribution of individual session means to the significant interaction was determined using Tukey's HSD post hoc comparisons (a = 0.05). Pairwise comparisons indicated that during the first reversal session, all rats responded at higher levels on the control lever than the drug lever. During the subsequent 4 sessions, responding on the control lever did not differ from responding on the drug lever. However, by the last 3 reversal sessions all rats had successfully tracked the change in reinforcement contingency as reflected by significantly higher number of responses emitted on the drug lever, relative to the control lever. 39 Temporal Profile of Responding To determine whether the temporal pattern of responding was consistent across experimental paradigms, the number of responses emitted in 2 h blocks during each of the 4 baseline sessions of the present experiment was calculated. The mean of the 4 baseline sessions was then calculated for each rat and these data were subjected to A N O V A with 3 repeated measures (Drug/Control Lever, Lights On/Off and Blocks of 2 h). A significant main effect of Lever [F(l,4) = 10.96, p < 0.04], confirmed that the rats exhibited higher levels of responding on the drug lever than the control lever (Figure 5a). In addition a significant main effect of Lights was obtained [F(l,4) = 18.26, p < 0.02], reflecting the higher rates of responding during the dark phase. Responding on both the control and the drug lever increased during the dark phase as indicated by the nonsignificant interaction of Lever x Lights x Blocks and Lever x Lights [F(l,4) = 4.01 and 4.14, respectively, p > 0.10]. It should be noted that all rats exhibited increased responding during the dark phase, regardless of absolute rate of responding. In order to determine the temporal pattern of responding emitted during the reversal phase, the data from the last 4 reversal sessions were subjected to the same statistical analysis described for the baseline sessions (Figure 5b). During the last 4 reversal sessions, rats responded at higher rates on the drug lever than the control lever [F(l,4) = 26.56, p < 0.01). The temporal pattern of responding across the lightdark phase was similar for both the drug and control lever, as indicated by the nonsignificant interaction of Lever x Lights x Blocks and Lever x Lights [F(l,4) = 2.43 and 0.012, respectively, p > 0.01]. However, as a group these rats did not exhibit a significant increase in responding during the dark phase, as indicated by the nonsignificant main effect of Lights [F(l,4) = 1.34, p > 0.30]. Inspection of data from individual subjects revealed that 2 rats responded at stable rates across the light:dark phase while 3 rats continued to exhibit higher rates of responding during the dark 40 Figure 5. The temporal pattern of responding on the drug and control lever during a) baseline and b) reversal (mean of the last 4 sessions of the baseline or reversal phase, respectively) is illustrated as a function of the light:dark cycle. Each data point represents the mean (+ S.E.M.) number of reinforced responses emitted in 2 h blocks (n=5). 41 A) BASELINE 12-11-10-9 -8 -7-6 -5 -4 -3 2 -1-UGHTS ON LIGHTS OFF • — i 1 r B) REVERSAL 8-i 7 -6 -5 - . 4 -3 -2 -1-~ i ~ i r i i r REINFORCED RESPONSES NONREINFORCED RESPONSES —i 1 1 1 1 r—X—| 1 1 1——•, r-1 2 3 4 5 6 7 8 9 10 11 12 TWO HOUR BLOCKS 42 phase. It is of interest that the 2 rats responding at stable levels across the light:dark phase (Subjects 01 and 09) had been exposed to midazolam for a greater number of sessions (49 and 52 sessions, respectively) than the latter group (31-40 sessions). Furthermore, during the acquisition of midazolam self-administration, these 2 rats exhibited a temporal pattern of responding with increased responding during the dark phase. Discussion The results of this experiment provide further evidence for the presence of reinforcing effects of midazolam in naive rats given unlimited access to the drug. Specifically, it was observed that rats preferred to respond on a lever that produced intravenous infusions of midazolam than on a lever that did not. Previously, it has been demonstrated that under conditions of 24 h access naive animals will self-administer midazolam, diazepam and chlordiazepoxide (Kubota et al., 1986; Narusi and Asami, 1987; Szostak et al., 1987; Yanagita and Takahashi, 1973). Together, these data confirm that benzodiazepines can act as reinforcers capable of supporting operant responding in naive animals. In addition, the reinforcing effects of benzodiazepines in drug naive rats have recently been demonstrated in the place preference conditioning paradigm (File, 1986; Spyraki et al., 1985). The results of the present experiment confirm the observation in Experiment I that rats given continuous drug access exhibit a temporal pattern of responding for midazolam. Two explanations were offered earlier for the observation that peak responding for midazolam occurs during the dark phase of the 12 h light:dark cycle. First, the temporal pattern of responding may reflect a general increase in activity normally observed in rats during the dark phase. Second, the fluctuation in the rate of responding may be attributable to cyclic variations within the neural systems that mediate the pharmacological effects of midazolam. While the present experiment does not provide unequivocal support for either of these hypotheses, the results suggest that 43 a general increase in activity may contribute to the increase in responding for midazolam during the dark phase. This suggestion is based on the observation that rats responded at higher rates on both the drug and the control lever during the dark phase. Although the present results indicate that rats given continuous access to midazolam initially exhibit peak responding for the drug during the dark phase, they also provide preliminary evidence that the temporal pattern of responding may change in rats given prolonged access to midazolam. Specifically, rats self-administering midazolam for 49 sessions or more exhibited a stable pattern of responding over the entire light:dark phase. One explanation for the appearance of stable responding in these rats may be that the chronic administration of midazolam was sufficient to produce physical dependence and withdrawal symptoms; in order to avoid withdrawal rats may have responded in a stable pattern over the 24 h period. Support for this proposal comes from an analysis of the time course of development of physical dependence on midazolam in rats given 3 h access to the drug by oral self-administration (Falk and Tang, 1987). After 42 days, only 20 percent of the subjects exhibited withdrawal signs whereas, after 84 days access, 80 percent showed signs of withdrawal. The present data raise the possibility that a majority of subjects exhibit significant withdrawal after continuous access to midazolam for 49 days or more. Experiment III. Effects of Reinforcer Magnitude on Responding for Midazolam The effects of reinforcer magnitude (drug dose) on self-administration of psychomotor stimulants and opiates have been well characterized (Baxter et al., 1974; Pickens and Harris, 1968; Pickens and Thompson, 1968; Weeks and Collins, 1979). In general, there is an inverse relationship between the dose of the drug and the rate of operant responding for that drug. Although there is preliminary evidence that responding for the benzodiazepine flurazepam may be sensitive to dose titration, this relationship has not been fully examined (Collins et al., 1984). 44 Methods Naive, male Long-Evans rats (Charles River), weighing 300-400 g at the beginning of the experiment, served as subjects (n=12). The experimental protocol was identical to that described in Experiment 1. The day after surgery, each rat was given continuous access to one lever. During the acquisition of midazolam self-administration, each response on the lever produced a 4.5 sec infusion of midazolam (0.05 mg in 0.18 ml distilled water) coincident with the illumination of a cue light located directly above the lever. Responses on the lever during an infusion were recorded but had no programmed consequences. Once stable levels of responding for 0.05 mg midazolam per infusion were obtained (4-5 consecutive sessions with no upward or downward trends in responding), the drug concentration was either increased to 0.20 mg per infusion or decreased to 0.0125 mg per infusion for one test session. The change in drug concentration was randomly determined for each rat. Al l other infusion parameters were identical to those during the initial acquisition of midazolam self-administration. In cases where the cannula remained patent following the test session (n=6), rats were given unlimited access to the new drug concentration until stable levels of responding were obtained. Following attainment of stable responding for the new drug concentration, rats were tested under extinction conditions, during which 0.9% saline was substituted for the drug solution. Results Dose Titration During the acquisition of midazolam self-administration, stable responding for 0.05 mg of midazolam per infusion was attained within 4-25 sessions. As a result of the variation between rats in the rate of acquisition of midazolam self-administration, the baseline phase was again defined as the last 4 days of acquisition training. 45 During the baseline phase, rats administered approximately 10.5 + 3.5 mg/kg midazolam per session (based on the mean number of infusions of midazolam per session and an average weight of 400 g). For purposes of statistical analyses, the mean number of reinforced responses emitted during the 4 baseline sessions was calculated for each rat. The mean number of reinforced responses emitted during the baseline phase was then compared to that attained during the first session of the transfer phase (Table 2). Following acquisition of stable responding for 0.05 mg midazolam per infusion, each rat that was transferred to a lower dose of the drug (0.0125 mg/infusion; n=5) exhibited an increase in responding whereas each rat that was transferred to a higher dose (0.20 mg/infusion; n=6) exhibited decreased responding. The inverse relationship between dose and rate of responding was consistent across all rats, regardless of the absolute rate of responding. Correlated t-tests (one-tailed) supported the observation that responding significantly increased following the reduction from 0.05 to 0.0125 mg midazolam per infusion [t(4) = 2.30, p < 0.05] and responding significantly decreased following the increase from 0.05 to 0.20 mg midazolam per infusion (t(5) = 3.46, p < 0.01). In cases where the cannula remained patent, the rats (n=6) were maintained on the transfer dose of midazolam until stable responding was attained, at which time responding under extinction conditions was examined. The number of responses emitted by individual subjects during baseline, transfer and extinction sessions are illustrated in Figure 6. As previously noted, in the first session following the change in dose, responding was consistently increased in those rats transferred to a lower dose of midazolam, whereas responding was consistently decreased in rats transferred to the higher dose. However, this pattern deteriorated when rats were maintained on the new doses of midazolam. Although responding stabilized with repeated sessions (center panel of each graph), the level at which responding stabilized was independent of the change in drug concentration and baseline rates of responding. For example, the 46 Table 2. The total number of reinforced responses emitted in 24 h sessions during baseline responding for 0.05 mg midazolam per infusion and during the first session following transfer to either a) 0.0125 mg/infusion or b) 0.20 mg/infusion. a) Baseline' Transfer Transfer as 0.05 mg 0.0125 mg % Baseline Subject 17 24 54 . 225 Subject 24 29 45 155 Subject 26 100 170 170 Subject 28 51 57 112 Subject 29 42 53 126 b) Baseline Transfer Transfer as 0.05 mg 0.20 mg % Baseline Subject 02 109 58 53 Subject 03 61 23 38 Subject 07 181 105 58 Subject 09 50 37 74 Subject 21 15 07 47 Subject 25 260 173 66 Baseline responding is represented by the mean number of responses emitted during each of the last 4 days of acquisition of midazolam self-administration for each rat. 47 Figure 6. The total number of reinforced responses emitted per session are illustrated for individual rats (n=6). Responding during baseline sessions (the last 4 days of acquisition of midazolam self-administration) is illustrated in the left panel of each graph. Responding during the subsequent dose transfer and extinction sessions are illustrated in the center and right panel, repectively. * Missing cell due to experimenter error. S E S S I O N S 49 three rats that were transferred to 0.20 mg per infusion stabilized at rates that were higher (Subject 03), lower (Subject 02) or the same (Subject 09) as those previously emitted for 0.05 mg per infusion. During the extinction phase (right panel of each graph), all rats exhibited a progressive decrease in responding across sessions. Temporal Profile of Responding The temporal pattern of responding emitted by rats during the baseline phase and the dose transfer phase was also assessed (n=6). The data were analyzed in a manner similar to that described for determining the temporal pattern of responding in Experiment II, with the exception that the data were subjected to an A N O V A with 2 repeated measures (Lights On/Off and Blocks of 2 h). Analysis of the baseline sessions, revealed a significant main effect of Lights [F(l,5) = 24.97, p < 0.01] reflecting the higher levels of responding emitted during the dark phase (Figure 7a). The dose transfer data from rats given 0.20 and 0.0125 mg midazolam per injection were combined since the concentration of midazolam per injection did not have a significant effect on the temporal pattern of responding as indicated by a 3 factor A N O V A [Dose x Lights on/off x Blocks of 2 h; all interactions and the main effect of dose were nonsignificant, p > 0.10]. During the last 4 transfer sessions peak responding occurred during the dark phase (Figure 7b), as supported by a significant main effect of Lights [F(l,5) = 7.54, p < 0.045]. Taken together, these data indicate that rats given up to 39 days access to midazolam continue to exhibit the characteristic increase in responding during the dark phase. Discussion This experiment examined the effects of dose titration on the rate of responding for intravenous midazolam. During the first dose transfer session rats trained to self-administer midazolam (0.05 mg/infusion) exhibited increased responding when transferred to a lower dose (0.0125 mg/infusion) and decreased responding when 50 Figure 7. The temporal pattern of responding during a) baseline sessions (the last 4 days of acquisition of midazolam self-administration; 0.05 mg of midazolam per infusion) and b) dose transfer sessions (the last 4 days following transfer to 0.0125 or 0.20 mg of midazolam per infusion) is illustrated as a function of the lightrdark cycle (left and right panel of each graph, respectively). Each data point represents the mean (+ S.E.M.) number of reinforced responses in 2 h blocks (n=6). 51 A) BASELINE 13 12-11-10-9 8 -7 -6 5 4H 3 2 -1-UGHTS ON LIGHTS OFT B) DOSE TRANSFER 20 n 18 16 14 12-10-8 -6 -.4-2 -1 2 3 4 5 6 7 8 9 10 TWO HOUR BLOCKS 52 transferred to a higher dose (0.20 mg/infusion) of the drug. These results are consistent with the observation of an inverse relationship between the dose of pentobarbital and psychostimulants and the rate of operant responding for the drug, as determined by a single dose transfer session (Goldberg et al., 1971; Pickens and Harris, 1968; Pickens and Thompson, 1968). While the inverse relationship between drug concentration and rate of responding in the present experiment may reflect an attempt to maintain brain concentrations of midazolam that are reinforcing, changes in responding may also be related to muscle relaxant or sedative effects of the drug. A n inverse relationship between responding and drug dose was apparent only during the first transfer session after which no consistent relationship was observed. The effects of drug dose on the within-subject rate of operant responding have most often been assessed in a single dose transfer session. There are, however, at least two previous reports of dose titration in rats maintained on the transfer drug dose for longer than a single session. Baxter et al. (1974) reported that 2 rats self-administering apomorphine exhibited a sustained decrease in responding when transferred to a higher dose of the drug for 7 sessions. The discrepancy between the present results and those obtained with apomorphine may reflect a difference in the mechanisms supporting benzodiazepine versus psychostimulant self-administration. This explanation does not account for the discrepancy between the present results and the observation that rats self-administering the benzodiazepine flurazepam exhibited increased responding when transferred to a reduced dose of the drug for 4-5 days (Collins et al., 1984). Unfortunately, the latter authors presented what appears to be the mean number of responses emitted for flurazepam during the last 3 transfer sessions and as a result it is not clear whether all rats exhibited a sustained increase in responding when transferred to a reduced dose of the drug. In the present experiment, dose titration did not result in a sustained increase or decrease in responding for midazolam, indicating that the relationship between the drug dose and the rate of responding in the long term cannot 53 be predicted on the basis of the first dose transfer session. Further research is required to determine the relationship between dose and rate of responding for other benzodiazepines. 54 II. Neurochemical Substrates of Midazolam Self-Administration The observation that naive animals self-administer benzodiazepines intravenously (Finlay et al., 1989; Kubota et al., 1986; Naruse and Asami, 1987; Szostak et al., 1987) and that benzodiazepines support place preference conditioning in naive rats (File, 1986; Spyraki et al., 1985) suggests that benzodiazepines are reinforcing. Given this evidence, it is important to determine the neural substrate(s) mediating the reinforcing effects of these drugs. To date, research has focused on the role of the central neurotransmitters 7-aminobutyric acid (GABA) and dopamine (DA) in mediating the reinforcing effects of benzodiazepines. A. The GABA/Benzodiazepine-Chloride Ionophore Benzodiazepines produce many of their behavioral effects by potentiating the effects of the inhibitory amino acid neurotransmitter G A B A (cf., Bruun-Meyer, 1987; Martin, 1987; however, see Phillis and O'Regan, 1988). Specifically, the benzodiazepines interact with a recognition site on the G A B A ^ receptor and thereby modulate the gating of the chloride channel associated with the G A B A ^ receptor. Modulation of the GABA-chloride ionophore complex by the benzodiazepines appears to be related to the drug's ability to enhance binding of G A B A to its receptor and increase the frequency of channel openings induced by G A B A , thereby increasing the current flowing through the chloride channels. Although controversial, there may be benzodiazepine sites in the brain that are not coupled to G A B A receptors (cf., Pole, 1989). The action of benzodiazepines at the proposed GABA-independent benzodiazepine sites is not known. Given the complex nature of the interaction between the benzodiazepines and G A B A , several approaches have been taken in an attempt to determine the neural substrate(s) mediating the reinforcing effects of benzodiazepines. The most general approach has been to evaluate the influence of a G A B A antagonist or agonist on the 55 reinforcing effects of the benzodiazepines. Recently researchers have also begun to assess the effects of drugs that interact specifically with either the G A B A ^ receptor or the benzodiazepine site. i. Gamma-Aminobutyric Acid Several groups have assessed the role of G A B A in mediating the reinforcing effects of benzodiazepines. In a forced drug consumption procedure, daily administration of aminooxyacetic acid, a GABA-transaminase inhibitor, reduced oral diazepam intake (Fuchs et al., 1984). The authors propose that aminooxyacetic acid may have increased central G A B A concentrations and this effect acted synergistically with the diazepam resulting in decreased drug intake. The indirect G A B A agonist sodium valproate, which increases central G A B A concentrations and facilitates coupling of the G A B A site with the chloride channel, did not significantly influence diazepam-induced place preference and failed to produce an effect when given alone (Spyraki et al., 1985). These latter findings do not conclusively support a general role for G A B A in mediating benzodiazepine-induced place preference. It is important to note that changes in extracellular G A B A concentrations will influence the activity of the neurotransmitter at G A B A g receptors as well as the GABA A /benzodiazepine receptors. Therefore, the behavioral pharmacology of GABA-specific drugs should not be expected to be identical to that of the benzodiazepines. ii. The GABA^ Receptor The indirect G A B A A receptor antagonist picrotoxin, which binds to a barbiturate site on the receptor complex, has been reported to attenuate diazepam-induced place preference (Spyraki et al., 1985). However, in this case picrotoxin itself elicited a significant place aversion. Responding for intravenous infusions of diazepam is also attenuated by the G A B A A receptor antagonist bicuculline (Pilotto et al., 1984). These results are consistent in that they suggest that antagonism of the G A B A ^ receptor 56 blocks the reinforcing effects of benzodiazepines. Again, however, these data should be interpreted with caution since there is evidence that G A B A ^ receptors are not always coupled to benzodiazepine receptors (Schock et al., 1985). iii. The Benzodiazepine Binding Site Several groups have assessed the role of the benzodiazepine receptor in mediating the reinforcing effects of benzodiazepines. Intravenous self-administration of diazepam by rats is attenuated by the benzodiazepine receptor antagonist Ro 15-1788 (Pilotto et al., 1984). Similarly, Johanson and Schuster (1986) have examined the effects of Ro 15-1788 on responding for intravenous infusions of flurazepam and lorazepam in monkeys. Ro 15-1788 produced a dose dependent effect on benzodiazepine self-administration such that low doses of the antagonist increased responding whereas high doses decreased responding. In addition, the benzodiazepine receptor antagonist CGS 8216 attenuated diazepam-induced place preference in rats without producing a significant place aversion when given alone (Spyraki et al., 1985). These results suggest that the benzodiazepine receptor antagonists attenuate the reinforcing effects of benzodiazepines. However, the results of another experiment failed to support this hypothesis. Falk and Tang (1985) food-deprived naive rats to 80% of their body weight and then subjected them to a fixed interval (FI-1 min) schedule of food delivery. During the period of food delivery, rats given access to midazolam in their drinking water consumed as much fluid as rats given water. Subcutaneous injections of Ro 15-1788 or CGS 8216 had no effect on midazolam or water intake. However, it should be noted that the concentration of midazolam selected in this experiment was such that fluid intake of midazolam and water was equated and therefore animals were not expressing a preference for the drug, making the interpretation of the effects of antagonists difficult. 57 B. The Dopamine Hypothesis of Reward Several lines of evidence have led to the proposal that central DA-containing neurons may play a role in mediating the reinforcing effects of benzodiazepines. First, the D A theory of reward proposes that the reinforcing effects of psychoactive compounds such as psychomotor stimulants and opiates may be mediated by an increase in the activity of mesolimbic dopaminergic neurons (for reviews, see Fibiger and Phillips, 1986; 1988; Wise and Bozarth, 1987). If D A represents a common neurochemical substrate mediating the reinforcing effects of psychoactive compounds, then the reinforcing effects of benzodiazepines may also be mediated by D A neurons (Wise and Bozarth, 1987). In addition, there is anatomical, electrophysiological, biochemical and behavioral evidence that benzodiazepines may influence the activity of mesolimbic D A neurons. /. Anatomical Evidence Anatomical evidence indicates that benzodiazepine receptors are distributed in regions of cell bodies and terminals of mesolimbic D A neurons. Specifically, benzodiazepine receptors are found both in the ventral tegmental area (VTA) as well as the nucleus accumbens (NAS), olfactory tubercle (OT) and medial prefrontal cortex (mPFC) of both human and rat brain (Mohler et al., 1978; Young and Kuhar, 1980). Anatomically, therefore, benzodiazepine receptors are in a position to influence the activity of the mesolimbic D A projections both at the cell body and terminal regions of these neurons. //'. Electrophysiological Evidence O'Brien and White (1987) have recently assessed the effects of benzodiazepines on the electrophysiological activity of V T A D A neurons in the rat. Intravenous diazepam had variable effects on the activity of identified D A neurons in the V T A . A n increase in firing rate was observed in 76% of the D A cells while the remaining 58 cells were either inhibited or unaffected (18% and 6%, respectively). However, diazepam produced a consistent decrease in the firing rate of non-DA cells in the V T A and this effect was reversed by Ro 15-1788. In addition, microiontophoretic administration of flurazepam and chlordiazepoxide consistently inhibited non-DA cells in the V T A but had no significant effect on D A neurons. Dopaminergic cells in the substantia nigra pars compacta (SNC) and non-DA cells in the substantia nigra pars reticulata (SNR) exhibit the same response to benzodiazepines as D A and non-DA cells in the V T A , respectively. Specifically, diazepam did not consistently alter the firing of identified D A neurons in the SNC; 50% of the cells were excited by diazepam while 30% and 20% were either unaffected or inhibited, respectively (Ross et al., 1982). However, diazepam and flurazepam had significant and consistent inhibitory effects on firing of non-DA cells in the SNR and the diazepam-induced inhibition was reversed by Ro 15-1788 (Mereu et al., 1983; Ross et al., 1982). On the basis of this body of electrophysiological evidence O'Brien and White (1987) proposed that, through a process of disinhibition, benzodiazepines indirectly increase the activity of D A neurons in both the V T A and SNC. Specifically, they suggest that D A neurons in the V T A and SNC are normally inhibited by the activity of non-DA neurons in the V T A and SNR, respectively. Benzodiazepines consistently inhibit the activity of these non-DA neurons and in turn this results in a disinhibition of D A neurons. The benzodiazepines likely produce their effects on D A and non-DA neurons by potentiating the effects of afferent GABAergic neurons that project to the V T A and SN. This proposal is supported by evidence suggesting that descending GABAergic neurons influence the activity of both D A and non-DA cells in the SN and V T A (Grace and Bunney, 1979; 1985; White and Wang, 1984; Yim and Mogenson, 1980). 59 It is noteworthy that non-DA cells in the SNR are more sensitive to the inhibitory GABAergic input and these cells may represent interneurons that normally inhibit the SNC D A neurons (Grace and Bunney 1979; 1985). In summary, the electrophysiological data suggest that, through a process of disinhibition, benzodiazepines increase the activity of central dopaminergic neurons. This observation is consistent with the proposal that the reinforcing effects of benzodiazepines may be mediated by an increase in the activity of dopaminergic neurons. iii. Biochemical Evidence Given that the benzodiazepines increase the electrophysiological activity of mesolimbic D A neurons, it might reasonably be predicted that the drug-induced increase in firing is associated with an increase in transmitter release from D A terminals. However, the results of several experiments suggest that benzodiazepines may decrease D A turnover in terminal regions of mesolimbic neurons. Diazepam, nitrazepam and brotizolam attenuate a-methyl-para-tyrosine induced depletion of D A in the O T and NAS (Fuxe et al., 1975; Ishiko et al., 1983). In addition, using ex vivo D A metabolite concentrations as an index of D A turnover, Fadda et al. (1978) found that diazepam alone had no effect on the concentration of 3,4-dihydroxyphenylacetic acid (DOPAC) in the frontal cortex or NAS and yet diazepam did attenuate the stress-induced increase in D O P A C concentrations in both brain regions. Similarly, in vivo electrochemistry has been used to assess the effects of benzodiazepines on D O P A C concentrations in the brains of freely moving rats. D'Angio et al. (1987) have reported that diazepam does not influeunce extracellular D O P A C concentrations in either the NAS or mPFC. However, tail-pinch stress induced a significant increase in D O P A C concentration in the NAS and this effect was attenuated by diazepam. 60 In summary, and in contrast to the electrophysiological results, ex vivo and in vivo biochemical evidence suggests that benzodiazepines decrease D A turnover. While there is controversy as to whether the benzodiazepines consistently decrease D A turnover in the basal condition, it is evident that under no condition did the drugs induce the increase in turnover predicted on the basis of the electrophysiological data and the D A hypothesis of reward. iv. Behavioral Evidence There is behavioral evidence to suggest that D A may mediate the reinforcing effects of benzodiazepines. Fuchs et al. (1984) found that rats exhibited a preference for solutions of diazepam over water and this preference was reduced by addition of the D A receptor antagonist haloperidol to the drug solution. In addition, intravenous self-administration of diazepam is attenuated by haloperidol (Pilotto et al., 1984). Similarly, diazepam-induced place preference is attenuated by haloperidol and by 6-hydroxydopamine lesions of dopaminergic terminals within the the nucleus accumbens (Spyraki and Fibiger, 1988). In summary, Wise and Bozarth (1987) have proposed that all drugs of abuse increase the activity of mesolimbic D A neurons and that this drug effect mediates the reinforcing properties of psychoactive compounds. The results of electrophysiological and behavioral experiments suggest that benzodiazepines increase the activity of mesolimbic D A neurons and support the proposal that the reinforcing effects of benzodiazepines may be mediated by central D A neurons. In contrast, biochemical evidence is not consistent with this proposal. 61 Experiment IV. Effects of the Benzodiazepine Antagonist Ro 15-1788 on Intravenous Self-Administration of Midazolam The results of several experiments suggest that benzodiazepine antagonists attenuate the reinforcing effects of benzodiazepines. Specifically, the rate of intravenous self-administration of diazepam, flurazepam and lorazepam is altered by Ro 15-1788 (Johanson and Schuster, 1986; Pilotto et al., 1984) and diazepam-induced place preference is attenuated by CGS 8216 (Spyraki et al., 1985). The following experiment was designed to assess the role of the benzodiazepine receptors in mediating the reinforcing effects of midazolam. To this end, the effect of the benzodiazepine antagonist Ro 15-1788 on self-administration of midazolam was examined. Ro 15-1788 is a selective benzodiazepine antagonist that displaces benzodiazepines from their central binding sites and antagonizes the effects of benzodiazepines in a variety of behavioral, electrophysiological and biochemical preparations (Hunkeler et al., 1981). Pharmacokinetic and pharmacodynamic evidence suggests that Ro 15-1788 has a rapid onset and short duration of action (d'Argy et al., 1987; Hunkeler et al., 1981; Klotz et al., 1984). For example, following an intravenous infusion of J H - R o 15-1788, high concentrations of radioactivity were observed in the mouse brain within 30-60 s of the injection whereas no radioactivity was apparent 4 h later (d'Argy et al., 1987). Methods Naive, male Long-Evans rats (Charles River), weighing 300-400 g at the beginning of the experiment, served as subjects (n=4). The experimental protocol was identical to that described in Experiment 1, with the following exceptions. In addition to an intravenous cannula, a chronic silastic cannula was implanted subcutaneously in each rat. The tip of the subcutaneous cannula was located on the abdomen. From the abdomen, the cannula passed subcutaneously to the back of the rat where it exited through a Marlex mesh assembly along with the intravenous cannula. The day after 62 surgery, each rat was given continuous access to two levers. The levers were randomly designated as either the drug or the control lever for individual rats. Each response on the drug lever produced a 4.5 sec infusion of midazolam (0.05 mg in 0.18 ml distilled water) coincident with the illumination of the cue light located directly above the drug lever. Responses on the drug lever during an infusion were recorded but had no programmed consequences. Responses on the control lever did not produce an infusion but did result in a 4.5 sec illumination of the cue light located directly above the control lever. A 12 h reverse light:dark cycle was maintained by use of a houselight within each chamber. The 12 h dark phase began at the start of each 24 h session (10:00 A M ) . When stable rates of responding on both the drug and the control lever were attained (4-5 consecutive sessions with no upward or downward trends in responding), the effects of Ro 15-1788 or vehicle injections on responding were evaluated in subsequent test sessions. During each test session, 6 injections of Ro 15-1788 (5 mg/kg/injection) or vehicle were delivered at 2 h intervals via a chronic subcutaneous cannula. The first injection was administered at the start of the session coincident with the onset of the dark phase. In this manner, repeated Ro 15-1788 or vehicle injections were delivered during the dark phase; the period during which rats typically exhibit peak responding for midazolam (Experiments I-III). Each rat received both Ro 15-1788 and vehicle treatments on separate occassions. The order of drug and vehicle injections was counterbalanced across subjects and there were at least 4 baseline sessions between each test session. During the test phase, all response contingencies and infusion parameters associated with intravenous midazolam self-administration were identical to those present during the initial acquisition of responding for midazolam. 63 Results To determine the effects of Ro 15-1788 and vehicle injections on midazolam self-administration, the number of responses emitted for midazolam in 1 h blocks of each 24 h test session was calculated. These data are illustrated in Figure 8; the time of the Ro 15-1788 and vehicle infusions are indicated by the arrows. The data acquired during the interval of Ro 15-1788 and vehicle infusions (the dark phase) were subjected to A N O V A with 2 repeated measures (Treatment with Drug/Vehicle and Blocks of 1 h). Self-administration of midazolam was signficantly influenced by repeated subcutaneous injections of the benzodiazepine antagonist as indicated by the significant 2-way interaction [F(l 1,33) = 3.72, p < 0.002]. To determine the contribution of individual means to the significant interaction, correlated t-tests (p < 0.05) were performed on the data acquired in 1 h blocks immediately following drug or vehicle injections. In the 1 h blocks following the first and second injection, Ro 15-1788 elicited a significant increase in self-administration of midazolam over control values [t(3) = 5.37 and 4.39, respectively]. The following 4 Ro 15-1788 infusions did not significantly affect responding for intravenous midazolam relative to control values [t(3) = 1.8, -0.01, 0.0 and -1.3, respectively]. Discussion In the present experiment, blockade of central benzodiazepine receptors altered the rate of midazolam self-administration. Specifically, rats given 24 h access to intravenous infusions of midazolam exhibited increased responding for midazolam following administration of the benzodiazepine antagonist Ro 15-1788. It may be that, by occupying the benzodiazepine binding site, Ro 15-1788 reduces the reinforcing effects of midazolam. Consequently, to overcome the reduction in reward magnitude rats' exhibit increased responding for the drug. 64 Figure 8. The total number of reinforced responses emitted for midazolam in 1 h blocks following Ro 15-1788 or vehicle injections. The arrows indicate the time of Ro 15-1788 or vehicle injections. Each point represent the mean ± S.E.M. (n=4). * Significantly different from vehicle control values, p < 0.05. S9 66 While benzodiazepine antagonists consistently alter the rate of benzodiazepine self-administration, there is controversy as to whether the antagonists elicit an increase (Johanson and Schuster, 1986; present results) or decrease (Johanson and Schuster, 1986; Pilotto et al., 1984) in the rate of drug self-administration. Johanson and Schuster (1986) have suggested that this discrepancy may reflect a dose-response function of the antagonist. The present results are consistent with previous experiments demonstrating that benzodiazepine antagonists alter responding for intravenous benzodiazepines (Johanson and Schuster, 1986; Pilotto et al., 1984). It has also been reported that the benzodiazepine receptor antagonist CGS 8216 attenuates the reinforcing effects of diazepam in the place preference conditioning paradigm (Spyraki et al., 1985). Together these data suggest that the reinforcing effects of benzodiazepines may be mediated by central benzodiazepine receptors. Experiment Va & b. Effects of Midazolam on Extracellular Dopamine and Metabolites as Assessed by In Vivo Electrochemistry There is substantial evidence that numerous drugs of abuse increase extracellular D A concentrations in the terminal regions of mesolimbic D A neurons (cf., Wise and Bozarth, 1987). However, until recently the evidence has been based on ex vivo analyses of drug-induced changes in D A or metabolite concentrations. The development of in vivo electrochemistry and microdialysis provide valuable tools for directly assessing the effects of drugs on extracellular D A concentrations in freely moving animals (for a review of in vivo electrochemistry or dialysis, see Justice, 1987 or Ungerstedt, 1984, respectively). In fact, recent results using in vivo microdialysis indicate that psychomotor stimulants, opiates, ethanol and nicotine preferentially increase extracellular D A concentrations in the mesolimbic regions of freely moving rats (Di Chiara and Imperato, 1988; Carboni et al., 1989), supporting the hypothesis 67 that the reinforcing effects of these drugs may be related to their ability to increase the activity of mesolimbic D A neurons. In the following experiments, in vivo electrochemistry and microdialysis (Experiments V and VI, respectively) were used to assess the effects of midazolam on extracellular D A concentrations in the NAS of freely moving rats. Experiment Va. Subcutaneous Midazolam Injections Methods In vivo electrochemical detection of D A was performed with stearate-modified graphite paste electrodes previously developed and characterized by Blaha and Lane (1983). This technique does not differentiate between D A and noradrenaline (NE) in the extracellular space, due to the similarity of their peak oxidation potentials. However, the endogenous concentration of N E in the NAS is approximately 8% of the total catecholamine content; therefore N E contributes only slightly to the oxidation current (Versteeg et al., 1976). Electrodes Stearate-modified graphite paste working electrodes were constructed from 4-5 cm lengths of Teflon-insulated stainless-steel wire (OD=l75-200 u ; Medwire Corp.). On one end of the wire, the teflon insulation was cut perpendicular to the wire axis with a fine-edged razor blade and the insulation was then forced over the end of the stainless-steel wire to form a well (0.5-1.0 mm in length). To prevent sleeving of the Teflon, the insulated wire was bent to form a single loop 8-9 mm from the tip of the well and the loop was cemented in place with dental acrylic. Immediately prior to implantation, stearate-modified graphite paste was packed into the teflon well and the electrode tip was polished and excess paste was removed by rubbing the electrode tip on Teflon tape. A smooth graphite surface was then produced by tapping the electrode 68 tip on glass. Prior to use in vivo, each electrode was examined under a microscope to ensure that there were no striations on the Teflon and that the paste surface was smooth and homogeneous in appearance. At the opposite end of the stainless-steel wire, the teflon insulation was removed from 2 mm of the wire and this end was then soldered into a gold amphenol pin. The stearate-modified graphite paste was prepared by dissolving 100 mg of stearic acid in 1 ml of paraffin oil warmed to approximately 40° C in a glass mortar. The mortar was then removed from the hot plate and 1.5 g of graphite powder (UCP-1-M, Ultra Carbon) was mixed with the paraffin oil, using a glass pestle, to form a homogeneous paste. A A g / A g C l reference electrode was prepared by removing 1.5 mm of the Teflon insulation from both ends of a length of Teflon-insulated silver wire (30 gauge; 1.5 cm in length; Medwire Corp.). One end of the silver wire was soldered into a gold amphenol pin and the other end was anodized, using direct current from a 9 V dry cell battery, in 0.9% NaCl until a homogeneous AgCl layer was apparent on the exposed tip. The chloriding process was performed immediately prior to implantation of the electrode. The auxiliary electrode was formed by removing the teflon-insulation from a 3 cm length of stainless steel wire and wrapping the stainless steel wire around a stainless steel skull screw. The free end of the wire was then soldered into a gold amphenol pin. The amphenol pins of the three electrode leads (working, reference and auxiliary electrodes) were then inserted in a miniature series-strip connector. Surgery Naive male Long-Evans rats (Charles River) weighing 430-460 g (n=5) served as subjects. Each rat was anesthetized with Nembutal (50 mg/kg, ip) and placed in a stereotaxic frame (Kopf Instruments). During anaesthesia, body temperature was monitored with a rectal probe (Yellow Springs Instruments) and maintained at 37° C 69 using a heating pad and temperature control unit (American Hospital Supply). The skull was exposed and a thin coating of bone wax was applied to reduce the bleeding. Holes were drilled in the skull for implantation of the electrodes and anchoring skull screws. After careful excision of the dura, the working electrode was slowly lowered (at a rate of approximately 1 mm/5-10 min) into the caudomedial NAS and secured in place with dental cement. The stereotaxic coordinates for implantation of the working electrode were 3.4 mm anterior to bregma, 1.5 mm lateral to the midline and 6.5 mm from dura with the incisor bar set at +5.0 mm (Pellegrino et al., 1979). The reference electrode was lowered 1 mm into the cortex and the auxiliary electrode was screwed into the skull such that it was in contact with the dura. The entire electrode assembly was then anchored to the skull with skull screws and dental acrylic. Electrochemistry Three to four days after surgery, each rat was placed in an electrically-shielded chamber and connected to a multipurpose electrochemical circuit ( G M A Instruments) by a flexible coaxial-shielded cable and a low-noise commutator-liquid swivel. Electrochemical measurements were performed using semiderivative linear sweep voltammetry (Lane et al., 1979). Oxidation currents were determined by applying a potential ramp from -100 mv to +250 mv at a rate of 10 mv per second (versus a A g / A g C l reference electrode) to the working electrode. It has previously been demonstrated that the peak oxidation current at the stearate-modified graphite paste electrode is proportional to the concentration of D A at the electrode surface and thus a change in current reflects a change in the concentration of the neurotransmitter (Blaha and Lane, 1983; Howard-Butcher et al., 1983; Lane et al., 1987). 70 Drug Administration Extracellular D A concentrations were determined at 15 min intervals until stable baseline measurements were attained (at least 4 consecutive samples in which no upward or downward trends in peak current were apparent). Rats were then given a subcutaneous injection of saline (0.5 ml/kg; n=3) or midazolam (1 or 5 mg/kg in distilled water; n=3 at each dose). In cases where a rat was given more than one injection, at least 48 h intervened between injections and the order of injections was randomized. Postinjection currents were measured at 15 min intervals until the currents returned to the preinjection baseline values. Al l injections were given during the rat's light phase, with one exception. In the previous experiments, it was observed that rats exhibit peak responding for intravenous infusions of midazolam during the dark phase of the light:dark cycle. For this reason, one rat was given an injection of midazolam (5 mg/kg; sc) 1.5 h after the onset of the dark phase. At the completion of each experiment the brain was removed, sectioned (30 micron sections) and stained (Nissl stain) for microscopic analysis of the electrode location. The data presented in the subsequent experiments were acquired from electrodes located within the caudomedial nucleus accumbens. Data Analysis Due to the variability between subjects in the peak D A oxidation current, all data are presented as a percent of baseline oxidation current (defined as those measurements performed 1 h prior to drug treatment for each rat). Pairwise comparisons were performed using two-tailed, correlated t-tests (p < 0.05). 71 Results Subcutaneous injections of midazolam (1 and 5 mg/kg) elicited a time-dependent decrease in extracellular D A concentrations in the caudomedial NAS (Figure 9). Following 1 mg/kg, D A oxidation current was maximally decreased to 64% of baseline values at 2 h postinjection. Three hours after an injection of 5 mg/kg, D A oxidation current was maximally decreased to 54% of baseline values. The maximal decrease in D A oxidation current was significantly different from the last baseline sample for both 1 and 5 mg/kg (102 ± 4.1% versus 64 ± 11.1% and 96 ± 4.9% versus 54 ± 8.2%, respectively) [t(2) = 5.58 and 8.08, respectively]. D A oxidation currents returned to baseline values by approximately 7 h postinjection. It should be noted that immediately after an injection of midazolam, extracellular D A concentrations were transiently increased above baseline values. However, subcutaneous injections of saline also increased the D A oxidation current (see inset, Figure 9) suggesting that the increase may be a result of stress associated with the injection. It has previously been observed that vehicle injections increase extracellular concentrations of the D A metabolite homovanillic acid (HVA) in the NAS (Brose et al., 1987). It was apparent (Experiments I-III) that rats exhibit the highest rates of responding for midazolam during the dark phase of the lightdark cycle. Therefore, in the present experiment, one rat was given an injection of midazolam (5 mg/kg; sc) 1.5 h after the onset of the dark phase. During the dark phase, midazolam elicited a time-dependent decrease in D A oxidation current that was not qualitatively different from that observed when the drug was injected during the light phase (Figure 10). 72 Figure 9. Dopamine oxidation currents in the nucleus accumbens following a subcutaneous injection of 1 or 5 mg/kg midazolam or saline (see inset) are presented as a percent of preinjection oxidation currents. Each data point represents the mean ± S.E.M. (n=3 per dose). The mean basal D A oxidation current (± S.E.M.) was 2.6 ± 0.4 nA; based on a mean of the 4 preinjection baseline samples for each rat. 1 0 0 — 130 120 110 elln 100 elln D m 90 L d Z 80 < Q LO 70 Q 60 50 I NUCLEUS ACCUMBENS eo O 1 m g / k g • 5 m g / k g MIDAZOLAM (s.c.) i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i 0 1 2 3 4 5 6 7 TIME AFTER INJECTION (hours) 74 Figure 10. Dopamine oxidation currents in the nucleus accumbens following a subcutaneous injection of 5 mg/kg midazolam, administered 1.5 h after onset of the dark phase, are presented as a percent of preinjection oxidation currents. The mean basal D A oxidation current was 2.5 nA; based on a mean of the 4 preinjection baseline samples (n=l). DOPAMINE (% BASELINE) 76 Experiment Vb. Intravenous Infusions of Midazolam The results of the self-administration experiments (Experiments I-III) demonstrated that rats respond for intravenous infusions of 0.05 mg midazolam per infusion. Therefore, the present experiment examined the effects of repeated intravenous infusions of midazolam on D A oxidation current in the NAS of freely moving rats. Methods Male Long-Evans rats (Charles River) weighing 380-460 g served as subjects (n=2). The methods were identical to those described for Experiment Va, with the following exception. After implantation of the electrodes, both rats were removed from the stereotaxic frame and while still anaesthetized a chronic silastic cannula was implanted into the right jugular vein. The cannula passed subcutaneously over the rat's shoulder and exited on top of the skull where it was attached to a stainless steel cannula (26 gauge). The stainless steel cannula was secured to the skull, caudal to the electrode assembly, with dental cement. Three to four days after surgery, D A oxidation currents in the NAS were assessed at 10 or 15 min intervals (depending on the subject), as previously described (Experiment Va). Following attainment of stable baseline currents, repeated intravenous infusions of midazolam (0.05 mg/infusion per 0.18 ml) or saline were delivered by the experimenter. The schedule of intravenous drug infusions was designed to mimic the pattern of responding emitted by rats self-administering midazolam, during peak responding for the drug. In this manner, a series of 4, 6, 6 and 5 infusions were delivered at 1 h intervals (each infusion was separated by 1 min). 77 Results The effects of repeated intravenous infusions of midazolam on D A oxidation currents, for individual rats, are illustrated in Figures 11 and 12. As with subcutaneous injections, repeated intravenous infusions of midazolam elicited a decrease in extracellular D A concentrations and this effect was not observed following t. repeated saline infusions (see inset, Figure 11). However, in contrast to subcutaneous drug injections, intravenous infusions of midazolam did not elicit an immediate increase in D A oxidation current. This observation provides further support for the proposal that the increase in extracellular D A concentration elicited by subcutaneous injections was an effect of the injection rather than a drug effect. Experiment VI. Effects of Midazolam on Extracellular Dopamine and Metabolites as Assessed by In Vivo Microdialysis The present experiment was designed to validate and extend the results of the previous experiment. To this end, in vivo microdialysis was used to examine the effects of midazolam on extracellular D A and D A metabolite concentrations in the NAS of freely moving rats. Methods Dialysis Probe The dialysis probe consisted of a semipermeable hollow fiber (saponified cellulose ester, OD=0.27 mm, molecular weight cut off= 10,000 Daltons; Cordis Dow Medical), a stainless steel cannula glued to the inlet end of the fiber and a tungsten wire (OD=0.15 mm) inserted through the hollow fiber for support during implantation. With the exception of those areas within the NAS (2 areas; each 1.75 mm in length), the surface of the semipermeable fiber was covered with silastic glue to prevent diffusion. 78 Figure 11. Dopamine oxidation currents in the nucleus accumbens following repeated intravenous infusions of midazolam (0.05 mg/0.18 ml/infusion) or saline (0.18 ml/infusion; see inset) are presented as a percent of preinjection oxidation currents. The number and time of infusions are indicated by the arrows. The mean basal D A oxidation current was 3.3 nA; based on a mean of the 4 preinjection baseline samples (n=l). 120 NUCLEUS A C C U M B E N S c D m L d o 110 A 100 90 A 80H 70 A 60 A 50 A I A \ \ MIDAZOLAM (0.05 mg/infusion, I.v.) 100 6 0 H I I I I l I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I 0 1 2 3 4 5 6 TIME AFTER INJECTION (hours) 80 Figure 12. Dopamine oxidation currents in the nucleus accumbens following repeated intravenous infusions of midazolam (0.05 mg/0.18 ml/infusion) are presented as a percent of preinjection oxidation currents. The number and time of infusions are indicated by the arrows. The mean basal D A oxidation current was 4.0 nA; based on a mean of the 4 preinjection baseline samples (n=l). 120 n 110 H c "65 o m L d o a 100 90 H 80 H 70 H 60 H 50 NUCLEUS ACCUMBENS 7 \ / \ / 1 v i MIDAZOLAM (0.05 mg/infusion. i.v.) I I I I I I I I I I M I I I I I I I J I I I I N M l I 0 1 2 3 4 5 6 TIME AFTER INJECTION (hours) 82 Surgery Male Wistar rats (Charles River; 250-280 g) were anaesthetized with Nembutal (50 mg/kg, ip) and placed in a stereotaxic frame (Kopf Instruments). During anaesthesia, body temperature was monitored and maintained at 37° C as previously described (Experiment V). After exposing the skull, holes were drilled on the lateral surfaces of the skull to allow for the introduction of a horizontal dialysis probe through the NAS (Imperato and Di Chiara, 1984). The coordinates for probe implantation were 2.0 mm anterior to bregma and 7.35 mm below the surface of the skull, with the incisor bar set at -3.3 mm (Paxinos and Watson, 1986). The dialysis fibre was positioned in the brain so that only the non-glued area of the dialysis membrane was exposed to the nucleus accumbens bilaterally. Subsequently, the tungsten wire was removed and a stainless steel cannula was glued to the outlet end of the dialysis fiber. Both stainless steel cannulae were then secured to the skull using dental cement and stainless steel skull screws. Microdialysis Al l experiments were carried out 18-48 hours after surgery in conscious animals during the light phase of the light:dark cycle. Microdialysis and subsequent chemical analysis of the perfusate were performed using a fully automated on-line sample injection system described previously (Westerink and Tuinte, 1986). In this system, the steel inlet cannula of the dialysis fiber was connected by polyethylene tubing (ID=0.28 mm, length=80 cm) to a perfusion pump (Carnegie Medicine). The dialysis probe in each rat was perfused with a Ringer solution (147 m M NaCl, 4 m M KC1 and 2.1 m M C a C y at a constant rate of 5 microlitres per minute. The steel outlet cannula of the dialysis fiber was connected by polyethylene tubing (as above) to the sample loop (100 microlitres) of an electrically activated injector (Valco). The injector was held in the load position for 20 minutes, during sample collection, and was then switched to the inject position for 10 seconds after which the cycle was repeated. The alternating 83 modes of the sample valve were controlled by an adjustable timer (Valco). The connection of the dialysis fiber directly to the sample loop of the high pressure liquid chromatography (HPLC) equipment resulted in a time lag of 10 min between events occurring at the rat and the time at which the dialysate associated with that event was introduced onto the H P L C system; the data presented are corrected for this time lag. Biochemical Assay The concentration of D A , D O P A C and H V A in the dialysate was quantified by H P L C in conjunction with electrochemical detection ( H P L C - E C ) (Westerink and Mulder, 1981). The mobile phase, consisting of 0.1 M acetic acid adjusted to pH 4.1 with solid sodium acetate, 0.5-0.9 m M octanesulfonic acid (Kodak), 0.01 m M N a 2 E D T A and 100-150 ml methanol per litre, was delivered by a pump ( L K B 2150) at a flow rate of 1.5 ml per minute. A pre-column (50x3 mm, Nucleosil 5 C 18) was placed between the pump and the injector (Valco). Dopamine and metabolites in the perfusate were separated on a second column by reverse phase liquid chromatography (250x4.8 mm, Nucleosil 5 C 18). Electrochemical detection was performed at a glassy carbon working electrode. The potential of the working electrode was set at +650 mV relative to a A g / A g C l reference electrode (LC4B, BAS). The chromatograms were registered on a chart recorder (Kipp). The detection limit of the assay was approximately 5 fmol/injection for D A and D O P A C and approximately 20 fmol/injection for H V A . Drug Injections Neurochemical concentrations were sampled at 20 min intervals until a stable baseline concentration of D A , D O P A C and H V A was detected in the dialysate (at least 3 consecutive samples in which no upward or downward trends in the data were apparent). Each rat was then give a subcutaneous injection of saline (1 ml/kg; n=4) or midazolam (1 or 5 mg/kg in distilled water; n=3 & 6, respectively). Postinjection 84 neurochemical concentrations were assessed at 20 min intervals until D A concentrations had returned to preinjection baseline values. After each experiment the brain was removed, sectioned (30 micron sections) and stained (Nissl stain) for microscopic analysis of the probe location. Only data acquired from rats in which the dialysis probe was located within the nucleus accumbens were included in the data analysis. Data Analysis Due to the variability between subjects in the concentration of D A , D O P A C and H V A , all data are presented as a percent of baseline values (defined as those measurements performed 1 h prior to drug treatment for each rat). Pairwise comparisons were performed using two-tailed, correlated t-tests (p < 0.05). Results Midazolam elicited a time-dependent decrease in extracellular D A concentrations in the caudomedial NAS (Figure 13). Following 1 mg/kg, D A concentrations were maximally decreased to 76% of baseline values at 60 min postinjection. At 80 min after an injection of 5 mg/kg, D A concentrations were maximally decreased to 75% of baseline values. The decrease in D A concentrations was significantly different from baseline values following an injection of 5 mg/kg [102 ± 1.7% versus 75 ± 5.3%; t(5) = 4.47] but not 1 mg/kg [97 ± 4.0% versus 76 ± 4.5%; t(2) = 2.88]. It is likely, that the effect of 1 mg/kg midazolam failed to attain statistical significance due to the small sample size. Extracellular D A concentrations had returned to baseline values by 3-4 h postinjection. Following saline injections, the maximum change in extracellular D A concentrations was an increase to 106% of preinjection baseline values at 20 min after the injection (see inset, Figure 13). 85 Figure 13. Extracellular dopamine concentration in the nucleus accumbens following subcutaneous injections of 1 or 5 mg/kg midazolam (n=3 and 6, respectively) or saline (n=4; see inset) are illustrated as a percent of preinjection baseline concentrations. Each data point represents the mean ± S.E.M. The mean basal D A concentration (± S.E.M.) in the dialysate was 9.1 ± 1.3 fmol/min; based on a mean of the 3 preinjection baseline samples for each rat (n=13). 120-110 H (D £ 100 <D (/) o m L d O Q 90 H 80 H 70H 1 0 0 -8 0 -NUCLEUS ACCUMBENS o o < / > o o o Q o o o o a 0 ^ Saline (1 m\/kgt s.c.) o 1 mgAg • 5 mgAg MIDAZOLAM (s.c.) T 1 r - — r 1 1 1 1 1 1 J ~ 0 1 2 3 TIME AFTER INJECTION (hours) i 1 r 4 87 Midazolam also elicited a decrease in extracellular D O P A C concentrations in the NAS (Figure 14). The decrease in extracellular D O P A C reached a nadir at 120 min postinjection following both 1 and 5 mg/kg (83% and 80% of baseline values, respectively). The maximum drug-induced decrease in D O P A C was significantly different from the last baseline sample following both 1 and 5 mg/kg (100 ± 1.8% versus 83 ± 3.8% and 100 ± 0.6% versus 80 ± 2.6%, respectively) [t(2) = 6.4 and t(5) = 8.5]. Extracellular D O P A C concentrations did not recover to baseline values within the 4 h of the experiment, remaining at 86% of baseline values at the time of the last sample. A progressive decrease in extracellular D O P A C concentration was observed in control rats such that by 4 h postinjection D O P A C concentrations were reduced to 92 ± 4.5% of baseline values. Subcutaneous injections of midazolam also elicited a decrease in extracellular H V A concentrations in the caudomedial NAS (Figure 15). At 100 min after an injection of 1 or 5 mg/kg, H V A concentrations were maximally decreased to 86 and 84% of baseline values, respectively. The decrease in H V A concentrations was significantly different from the last baseline value following an injection of 5 mg/kg [100 ± 2.3% versus 84 ± 4.5%; t(5) = 3.5] but not 1 mg/kg [103 ± 1.7% versus 86 ± 9.2%; t(2) = 1.94]. It is likely that the effect of 1 mg/kg failed to attain statistical significance due to the small sample size. Extracellular H V A concentrations approached baseline concentrations within the 4 h of the experiment, remaining at 90-95% of baseline concentrations during the last sample. Following saline injections, the maximum change in extracellular H V A concentrations was a decrease to 94% of preinjection baseline values at 4 h postinjection (see inset, Figure 15). In summary, the present experiment demonstrates that midazolam at both 1 and 5 mg/kg elicits a decrease in extracellular D A , D O P A C and H V A concentrations that is maximal at 1-2 h postinjection. Extracellular D A concentrations recovered to baseline values by 3-4 h postinjection. In contrast extracellular DOPA and H V A concentrations 88 Figure 14. Extracellular concentrations of 3,4-dihydroxyphenylacetic acid (DOPAC) following subcutaneous injections of 1 or 5 mg/kg midazolam (n=3 and 6, respectively) or saline (n=4; see inset) are illustrated as a percent of preinjection baseline concentrations. Each data point represents the mean ± S.E.M. The mean basal D O P A C concentration (± S.E.M.) in the dialysate was 1721.8 ± 182.3 fmol/min; based on a mean of the 3 preinjection baseline samples for each rat (n=13). 120-1 110 H NUCLEUS ACCUMBENS 100—1 80- Saline (1 ml/kg, s.c.) i i i i i i i i i' i i i i i i i 0 1 2 3 4 100 15 CO o m ( J o a 90 H 80 H 70 H MIDAZOLAM (s.c.) o 1 m g / k g • 5mgZkg 0 0 i 1 r ~~i 1 1 1 1 r— i 1 1 1 r 1 2 3 4 TIME AFTER INJECTION (hours) 90 Figure 15. Extracellular concentrations of homovanillic acid (HVA) following subcutaneous injections of 1 or 5 mg/kg midazolam (n=3 and 6, respectively) or saline (n=4; see inset) are illustrated as a percent of preinjection baseline concentrations. Each data point represents the mean ± S.E.M. The mean basal H V A concentration (± S.E.M.) in the dialysate was 1045.4 ± 144.0 fmol/min; based on a mean of the 3 preinjection baseline samples for each rat (n=13). NUCLEUS ACCUMBENS 100-8 0 - Saline (1 ml/kg, s.c.) i i i i i i i i i i i i i i i i 0 1 2 3 4 TTrrrH' r T MIDAZOLAM (s.c.) o 1 mgAg • 5 mgAg i 1 1 1 1 1 1 1 1 1 1 1 1 1 1 r 1 2 3 TIME AFTER INJECTION (hours) 92 did not recover to control values within the 4 h of the experiment. It is unclear to what extent this prolonged decrease in metabolite concentrations reflects a sustained drug effect since there was a trend for decreasing metabolite concentrations in control rats as well. Discussion Together, the results of Experiments Va & VI indicate that subcutaneous injections of midazolam (1 and 5 mg/kg) decrease extracellular D A concentrations in the NAS of freely moving rats. In addition, repeated intravenous infusions of midazolam (0.05 mg/infusion), delivered in a pattern that mimics the pattern of self-administration exhibited by rats, also decreased extracellular D A in the NAS (Experiment Vb). The results obtained using both in vivo electrochemistry and microdialysis demonstrate that subcutaneous injections of midazolam decrease extracellular D A in the NAS. However, the magnitude and the time course of the drug-induced effect obtained with the two techniques were different. In this regard, it is important to note that these techniques sample extracellular D A concentrations within different regions of the NAS. The transverse microdialysis probe samples extracellular D A concentrations throughout a horizontal region of the NAS bilaterally whereas in vivo electrochemical detection of D A is performed in a relatively discrete region of the caudomedial NAS unilaterally. The NAS may be a heterogeneous structure and this heterogeneity may be reflected in the response of subpopulations of mesoaccumbens D A neurons to pharmacological manipulations. As previously discussed, benzodiazepines increase the firing rate of some mesolimbic D A neurons (O'Brien and White, 1987). It might be predicted that an increase in cell firing should be associated with a corresponding increase in neurotransmitter release at the level of the terminals. However, the present results, combined with previous biochemical evidence (Fuxe et al., 1975; Ishiko et al., 1983), 93 suggest that benzodiazepines decrease basal extracellular D A concentrations in the terminal regions of mesolimbic D A neurons. In addition, Zetterstrom et al. (1989) have recently used in vivo microdialysis to assess the effects of the benzodiazepine flunitrazepam on extracellular D A concentrations in the NAS. Local application of flunitrazepam (the drug was added to the perfusate) elicited a decrease in extracellular D A within the NAS of freely moving rats. There are at least two explanations for the discrepancy between the electrophysiological and biochemical results. First, benzodiazepines may have different effects in anaesthetized (electrophysiological experiments) versus freely moving rats (present experiments). Support for this proposal comes from the observation that in the unanaesthetized "encephale isole" rat, intravenous midazolam elicited a decrease in multiunit activity in the SNC to approximately 42% of control values and this effect was reversed by Ro 15-1788 (Pole et al., 1981). Second, the discrepancy between the electrophysiological and biochemical data may indicate that activity at the level of the cell body of D A neurons does not consistently reflect activity at the terminal. In this regard, Trulson (1985) has performed an interesting experiment in which he simultaneously recorded electrophysiological activity in the SNC and extracellular transmitter concentrations in the striatum. This experiment revealed several dissociations between the electrophysiological activity of SNC neurons and D A concentrations in the striatum. Specifically, unit activity in the SNC was stable across the sleep-wake cycle whereas extracellular D A in the striatum decreased by 35% as the animal passed from quiet wakefulness into sleep. In addition, unit activity was observed to increase by 15-20% when the cat was engaged in episodes of tonic active movement whereas a much greater increase in striatal extracellular D A (50%) was observed during periods of movement. These data indicate that there may be qualitative and quantitative dissociations between the electrophysiological activity of D A neurons and transmitter release. Other explanations for the apparent dissociation are also possible. For 94 example, if nigrostriatal D A neurons do not respond as a homogeneous unit, electrophysiological activity of specific neurons may not necessarily reflect neurotransmitter release and vice versa. As suggested by Trulson (1985), the simultaneous determination of unit activity and transmitter release of neurons provides a more accurate assessment of the functional state of a given neurotransmitter system and as such this represents an important area for future research. The D A hypothesis of reward proposes that drugs of abuse elicit an increase in mesolimbic D A release which mediates the reinforcing effects of the drug (Wise and Bozarth, 1987). However, the present results indicate that midazolam elicits a decrease in extracellular D A in the NAS even though the drug is reinforcing in rats. In support of these results, other data suggest that intravenous diazepam is reinforcing in rats (Naruse and Asami, 1987) and yet this drug has also been reported to decrease turnover of D A in the NAS (Fuxe et al., 1975; Ishiko et al., 1983). Therefore, it appears that the reinforcing effects of benzodiazepines are not due to enhanced functional output by dopaminergic neurons. While the present data indicate that drug-induced increases in extracellular D A do not mediate the reinforcing effects of benzodiazepines, behavioral experiments indicate that D A antagonists and lesions of mesoaccumbens D A neurons can attenuate the reinforcing effects of diazepam (Fuchs et al., 1984; Pilotto et al., 1984; Spyraki and Fibiger, 1988). The reason for this discrepancy is not clear and illustrates the necessity of further work in the area of benzodiazepines and reinforcement. G E N E R A L DISCUSSION The present series of experiments was designed to assess two related questions concerning the short-acting benzodiazepine midazolam. First, is midazolam reinforcing in naive rats given unlimited access to intravenous drug infusions? Second, given that 95 midazolam is reinforcing, do dopaminergic substrates mediate these effects of the drug? Intravenous Self-Administration of Benzodiazepines: Evidence of Reinforcing Effects With regard to the first question, it is clear that naive rats self-administer midazolam under conditions of unlimited drug access and this, therefore, supports the proposal that midazolam is reinforcing. Specifically, these experiments demonstrated that: 1) response contingent infusions of midazolam result in higher rates of responding than response contingent infusions of saline (Experiment I); and 2) midazolam supports responding in a 2 lever discriminated responding paradigm in which the rats were able to track the changing reinforcment contingencies (Experiment II). In addition, responding for intravenous infusions of midazolam was characterized to illustrate that: 1) an inverse relationship between dose and rate of responding for midazolam is apparent only during the first dose transfer session after which no consistent relationship is observed (Experiment III); 2) initially, rats exhibit a temporal pattern of responding for midazolam with peak responding occurring during the dark phase of the light:dark cycle (Experiments I, II, and III); and 3) rats given prolonged access to midazolam exhibit a stable pattern of responding across the lightdark cycle (Experiment II). Together with previous research (Kubota et al., 1986; Naruse and Asami, 1987; Yanagita and Takahashi, 1973), these experiments indicate that drug-naive animals will self-administer benzodiazepines under conditions of unlimited drug access. This work provides support for the proposal that benzodiazepines are reinforcing and illustrates the utility of the self-administration technique for examining the abuse potential of benzodiazepines in naive animals. The reinforcing effects of benzodiazepines in naive rats have also been demonstrated using a place preference conditioning paradigm. In these experiments rats are given benzodiazepine injections paired with one set of environmental cues and 96 saline injections paired with a different set of environmental cues. On the test day the animal is given the opportunity to select between the drug-paired and saline-paired environment. Under these conditions, rats prefer an environment that has previously been paired with the effects of a benzodiazepine (File, 1986; Nomikos and Spyraki 1988; Spyraki and Fibiger, 1988; Spyraki et al., 1985). These results provide further support for the observation that benzodiazepines are reinforcing in naive rats. Intravenous Drug Self-Administration Paradigms as Experimental Models of Human Benzodiazepine Abuse To date, two experimental protocols have been used to demonstrate successfully intravenous self-administration of benzodiazepines in animals. Benzodiazepine self-administration has been observed in drug-naive animals given unlimited drug access (Finlay et al., 1989; Kubota et al., 1986; Naruse and Asami, 1987; Szostak et al., 1987 Yanagita and Takahashi, 1973). In addition, a drug substitution procedure has been used to demonstrate benzodiazepine self-administration in drug-experienced animals  given limited drug access (Bergman and Johanson, 1985; Griffiths et al., 1981; Johanson, 1987; Kubota et al., 1986; Lukas and Griffiths, 1982). Both experimental protocols provide evidence of the reinforcing effects of benzodiazepines and suggest that benzodiazepines have significant abuse potential. However, as experimental models of human drug-taking there may be important differences between the two techniques. In this regard, it is interesting that clinical research has provided evidence of abuse of therapeutic doses of benzodiazepines in subjects that are not multiple drug abusers (Busto et al., 1986; Cappell et al., 1987; Owen and Tyrer, 1983; Petursson and Lader, 1984), as well as benzodiazepine abuse as a component of multiple drug abuse (Griffiths et al., 1976; 1979; 1980; 1984). It may be that different mechanisms underly benzodiazepine abuse and multiple drug abuse in these clinical populations. As such, benzodiazepine self-administration in drug naive rats may provide an experimental model for assessing the neural substrate(s) mediating the reinforcing effects of 97 benzodiazepines in a population of patients that exhibits long term use of therapeutic doses of benzodiazepines. In contrast, the drug substitution procedure may provide an experimental model for assessing the substrate(s) mediating the reinforcing effects of benzodiazepines in multiple drug abusers. Is Midazolam Self-Administration a Result of Positive Reinforcement or Negative Reinforcement? As recently discussed (Wise and Bozarth, 1987), drug self-administration may occur as a result of positive reinforcement or negative reinforcement. On the one hand, drug self-administration occurs because an effect(s) of the drug is pleasurable and for that reason the subject engages in drug seeking and drug taking behavior. On the other hand, drug self-administration may occur because the drug alleviates an undesirable central state, such as anxiety. At present, it is not clear whether benzodiazepines are self-administered for their positive reinforcing effects or for the central effects that they alleviate. Attempts have been made to assess the role of the negative reinforcement in human benzodiazepine self-administration. Several groups have assessed whether benzodiazepines are more likely to be preferred by subjects for whom the drug may have therapeutic value. The results of these experiments are inconsistent. Anxious subjects, who would likely be prescribed benzodiazepines, failed to exhibit a preference for diazepam over placebo, suggesting that there is no relationship between the level of anxiety and preference for diazepam (de Wit et al., 1986). However, psychiatric patients exhibited a preference for diazepam over placebo and a positive correlation existed between drug intake and the level of anxiety (Balmer et al., 1981; Winstead et al., 1974). Similarly, one experiment has assessed the role of the negative reinforcement in benzodiazepine self-administration in rats. In this experiment, rats stressed by the presentation of either escapable or inescapable foot shock, failed to exhibit a preference for oral chlordiazepoxide over water (Kamano and Arp, 1965). Given that 98 oral chlordiazepoxide solutions may have aversive gustatory and postingestional effects (Wolf et al., 1978), these results are difficult to interpret. In the present experiments, it is not clear whether midazolam self-administration occurred as a result of positive or negative reinforcement. Given that the rats were required to live in an isolated and potentially stressful environment, midazolam self-administration may have occurred because the drug alleviated the anxiety induced by the experimental paradigm. Unfortunately, few experiments have been designed specifically to evaluate the role of negative reinforcement in benzodiazepine self-administration. This remains an important area for future research. Neurochemical Substrates Mediating the Reinforcing Effects of Midazolam Given that midazolam was found to be reinforcing in rats, other experiments were designed to determine the neurochemical substrates mediating the reinforcing effects of the drug. Specifically, the role of the benzodiazepine binding sites and the neurotransmitter D A were examined. The results of these experiments indicate that: 1) central benzodiazepine binding sites are critical in mediating the reinforcing effects of midazolam as demonstrated by the observation that the benzodiazepine antagonist Ro 15-1788 increased intravenous self-administration of midazolam (Experiment IV); and 2) the reinforcing effects of midazolam are not mediated by an increase in extracellular D A concentrations since subcutaneous and intravenous injections of midazolam elicit a decrease in extracellular D A concentrations (Experiments V and VI). The Role of the Benzodiazepine Receptor in Mediating the Reinforcing Effects of Midazolam At present, few experiments have attempted to determine the central mechanisms mediating the reinforcing effects of benzodiazepines. The present results are consistent with previous research demonstrating that benzodiazepine antagonists attenuate the reinforcing effects of benzodiazepines (Johanson and Schuster, 1986; Pilotto et al., 99 1984; Spyraki et al., 1985). Together these data suggest that the reinforcing effects of benzodiazepines are mediated by central benzodiazepine binding sites. However, benzodiazepine sites are widely distributed in the brain and as a result these drugs may potentially influence any number of central systems (Young and Kuhar, 1980). It is interesting to speculate however, that the dense distribution of benzodiazepine receptors in the limbic system may be related to their reinforcing effects. The Role of Dopamine in Mediating the Reinforcing Effects of Midazolam Wise and Bozarth (1987) have recently proposed a psychomotor stimulant theory of drug addiction. This theory rests on the assumption that all drugs of abuse elicit both an increase in locomotor activity and an increase in extracellular D A in the terminal regions of mesolimbic D A neurons. The results of the present experiments do not support the psychomotor stimulant theory of drug addiction in that midazolam is reinforcing and yet the drug elicits a decrease in extracellular D A in the NAS of freely moving rats. Several issues will be discussed that may account for the discrepancy between the present results and the psychomotor stimulant theory of addiction. These issues include the possibility that: 1) the reinforcing effects of midazolam may be mediated by an increase in extracellular D A in rats self-administering the drug chronically (the present experiments examined only the acute effects of midazolam); 2) the reinforcing effects of midazolam may be mediated by the same neural system as other drugs of abuse (in which D A is a major component) but midazolam acts "downstream" from the D A neuron; and 3) the psychomotor stimulant theory of addiction may not represent a general theory of drug addiction. In the present experiments, the effects of midazolam on extracellular D A were assessed following an acute drug injection. It remains possible that chronic midazolam self-administration may elicit an increase in extracellular D A concentrations and this increase mediates the reinforcing effects of the drug. However, this explanation is 100 unlikely since it fails to account for the fact that rats initiate benzodiazepine self-administration under conditions when the drug would decrease extracellular D A . According to the psychomotor stimulant theory of addiction the drug should not be rewarding in this circumstance. The present results may not be inconsistent with the concept of a common reward system, if midazolam acts on the same neural circuitry as stimulants and opiates but "downstream" from the D A neuron. This hypothesis should be subjected to further experimental analysis. For example, it would be interesting to examine the effects of excitotoxic lesions of the NAS on midazolam self-administration. Regardless of the results of such experiments, however, the present data indicate that all drugs of abuse  do not elicit an increase in mesolimbic D A concentrations, as proposed by the psychomotor stimulant theory of addiction. Wise and Bozarth (1987) propose that all positive reinforcers, including drugs and natural reinforcers such as food, are similar in their ability to elicit forward locomotion and increased activity of mesolimbic D A neurons. If this is taken as the principle upon which researchers are to identify positive reinforcers then: 1) all positive reinforcers should elicit these events; and 2) other stimuli (for example negative reinforcers) should not elicit these same behavioral and neurochemical consequences. With regard to the first prediction, the present results demonstrate that not all positive reinforcers elicit an increase in the activity of mesolimbic D A neurons. With regard to the second prediction, there is also independent evidence to suggest that negative reinforcers elicit the same behavioral and neurochemical consequences as positive reinforcers. While it may be true that positive reinforcers elicit approach behavior, the proposed theory does not account for the observation that negative reinforcers also elicit approach behavior in animals (Blanchard and Blanchard, 1988; Coss and Owings, 1985; Pinel et al., 1989; van der Poel, 1979). Second, Wise and Bozarth (1987) suggest that activation of mesolimbic D A neurons forms a critical component of the biological 101 mechanism responsible for approach behavior elicited by a positive reinforcer. Again, this theory does not account for the observation that aversive stimuli also increase the activity of mesolimbic D A neurons. Specifically, stress has been reported to increase extracellular D A concentrations and D A metabolite concentrations in both the NAS and frontal cortex (Abercrombie et al., 1989; D'Angio et al., 1987; Fadda et al., 1978). It appears, therefore, that approach behavior and increased activity of mesolimbic D A  neurons can be elicited bv both positive reinforcers and aversive stimuli. Therefore, approach behavior and increased activity of mesolimbic D A neurons may be associated with all behaviorally relevant stimuli in the animal's environment rather than providing information about the valence of those stimuli. In summary, a unified theory of positive reinforcement based on the proposal that all positive reinforcers produce forward locomotion and stimulation of mesolimbic D A neurons may be of limited utility because 1) not all drugs of abuse elicit an  increase in mesolimbic D A concentrations: and 2) approach behavior and increased  activity of mesolimbic D A neurons can be elicited bv both positive reinforcers and  aversive stimuli. Alternatives to the Dopamine Hypothesis of Reward Previously, it has been suggested that multiple neuronal systems may be involved in the processes underlying positive reinforcement (cf. Fibiger and Phillips, 1989). In support of this proposal, the present data suggest that the neural system mediating the reinforcing effects of midazolam may be independent of D A neurons. As a result, it is of interest to assess the role of other neurotransmitter systems and brain regions that may mediate the reinforcing effects of benzodiazepines. Spyraki and collaborators have examined the role of opiate receptors, noradrenaline and serotonin in mediating the reinforcing effects of diazepam. While naloxone attenuated diazepam-induced place preference in rats, the drug also produced place aversion. As a result it is not clear whether the attenuation of diazepam-induced 102 place preference was simply related to a confounding aversive effect of the opiate antagonist (Spyraki et al., 1985). Depletion of central noradrenergic neurons did not affect diazepam-induced place preference (Spyraki and Fibiger, 1988), suggesting that noradrenergic neurons are not involved in mediating the reinforcing effects of diazepam. Nomikos and Spyraki (1988) found that the 5-HT2 antagonist ritanserin completely blocked the reinforcing effects of diazepam and failed to have negative or positive reinforcing effects when administered alone. The results of this experiment provide preliminiary evidence that central 5 -HT neurons may form a component of a system involved in mediating the reinforcing effects of benzodiazepines. Summary and Future Perspectives The present experiments demonstrate that naive rats will self-administer the short-acting benzodiazepine midazolam, under conditions of unlimited access to drug. These results indicate that midazolam is reinforcing and suggest that the drug may have significant abuse potential. Further research is necessary to determine the extent to which other benzodiazepines will support self-administration behavior using the present experimental paradigm. The benzodiazepine anatagonist Ro 15-1788 increased responding for intravenous infusions of midazolam, suggesting that central benzodiazepine binding sites mediate the reinforcing effects of midazolam. Although controversial, there is evidence to suggest that two types of benzodiazepine sites (type I and II) are present in the brain (Dennis et al., 1988; Giorgi et al., 1989; Sadzot et al., 1989; however, see Martin, 1987). The classical benzodiazepines and Ro 15-1788 bind with equal affinity to both of these sites (cf. Martin, 1987). However, the preferential localization of type II sites in the limbic system has led to the suggestion that these sites may mediate the anxiolytic and anticonvulsant effects of benzodiazepines, whereas localization of type I sites in sensorimotor cortical areas and cerebellum implicates these sites in the sedative-hypnotic effects of benzodiazepines (cf., Dennis et al., 1988). It is possible, that the 103 dense distribution of type II sites in the limbic system may be related to the reinforcing effects of midazolam. Examining the role of type I and II benzodiazepine binding sites in the reinforcing effects of midazolam is an important area for future research. In the present experiments, midazolam was observed to elicit a decrease in extracellular D A concentrations in the NAS of freely moving rats. These data are in direct contrast to the proposal that the reinforcing effects of all drugs of abuse are mediated by an increase in the functional output of mesolimbic D A neurons. It remains to be determined whether there are multiple reward systems, one of which may be a DA-independent system responsible for mediating the reinforcing effects of midazolam and perhaps other benzodiazepines. 104 BIBLIOGRAPHY Abercrombie, E .D. , Keefe, K . A . , DiFrischia, D.S. and Zigmond, M . E . (1989) Differential effect of stress on in vivo dopamine release in striatum, nucleus accumbens, and medial frontal cortex. J. Neurochem. 52: 1655-1657. Altshuler, H . L . and Phillips, P.E. (1978) Intragastric self-administration of drugs by the primate. In B.T. Ho, D.Wi Richards and D.L . Chute (Eds.) Drug  Discrimination and State Dependent Learning. Academic Press, New York, pp. 263-280. Amit, Z . and Cohen, J. (1974) The effect of hypothalamic stimulation on oral ingestion of diazepam in rats. Behav. Biology 10: 223-229. Balmer, R., Battegay, R. and von Marschall, R. (1981) Long-term treatment with diazepam. Int. Pharmacopsvchiat. 16: 221-234. Balster, R . L . and Woolverton, W.L. (1982) Intravenous buspirone self-administration in rhesus monkeys. J. Clin. Psychiatry. 43: 34-37. Baxter, B .L. , Gluckman, M.I., Stein, L . and Scerni, R .A . (1974) Self-injection of apomorphine in the rat: positive reinforcement by a dopamine receptor stimulant. Pharmacol. Biochem. Behav. 2: 387-391. Bergman, J. and Johanson, C E . (1985) The reinforcing properties of diazepam under several conditions in the rhesus monkey. Psvchopharmacology 86: 108-113. Blaha, C D . and Lane, R.F. (1983) Chemically modified electrode for in vivo monitoring of brain catecholamines. Brain Res. Bull. 10: 861-864. Blanchard, D.C. and Blanchard, R.J. (1988) Ethoexperimental approaches to the biology of emotion. Ann. Rev. Psvchol. 39:43-68. Brennan, M.J.W., Volicer, L . , Moore-Ede M . C and Borsook, D. (1985) Daily rhythms of benzodiazepine receptor numbers in frontal lobe and cerebellum of the rat. Life Sci. 36: 2333-2337. Brown, Z.W., Amit, Z . and Weeks, J.R. (1976) Simple flow-thru swivel for infusions into unrestrained animals. Pharmacol. Biochem. Behav. 5: 363-365. Bruun-Meyer, S.E. (1987) The GABA/benzodiazepine receptor-chloride ionophore complex: nature and modulation. Prog. Neuro-Psvchopharmacol. & Biol.  Psvchiat. 11: 365-387. 105 Busto, U . , Sellers. E . M . , Naranjo, C . A . , Cappell, H.D. , Sanchez-Craig, M . and Simpkins, J. (1986) Patterns of benzodiazepine abuse and dependence. British J. of  Addiction 81: 87-94. Cappell, H . , Busto, U . , Kay, G . , Naranjo, C .A. , Sellers, E . M . and Sanchez-Craig, M . (1987) Drug deprivation and reinforcement by diazepam in a dependent population. Psvchopharmacology 91: 154-160. Carboni, E . , Imperato, A . , Perezzani, L . and Di Chiara, G . (1989) Amphetamine, cocaine, phencyclidine and nomifensine increase extracellular dopamine concentrations preferentially in the nucleus accumbens of freely moving rats. Neuroscience 28: 653-661. Collins, R.J . , Weeks, J.R., Cooper, M . M . , Good, P.I. and Russell, R.R. (1984) Prediction of abuse liability of drugs using IV self-administration by rats. Psvchopharmacology 82: 6-13. Cooper, S.J. and Estall, L .B . (1985) Behavioral pharmacology of food, water and salt intake in relation to drug actions at benzodiazepine receptors. Neurosci.  Biobehav. Rev. 9: 5-19. Coss, R . G . and Owings, D .H. (1985) Restraints on ground squirrel antipredator behavior: adjustments over multiple time scales. In Johnston and Pietrewicz (Eds.) Issues in the Ecological Study of Learning. L E A Publishers, pp. 167-195. Dennis, T. , Dubois, A . , Benavides, J. and Scatton, B. (1988) Distribution of central W j (benzodiazepine j) and w 2 (benzodiazepine2) receptor subtypes in the monkey and human brain. A n autoradiographic study with [3H]flunitrazepam and the W j selective hgand [ H]zolpidem. J. Pharmacol. Exp. Ther. 247: 309-321. D'Angio, M . , Serrano, A . , Rivy, J.P. and Scatton, B. (1987) Tail-pinch stress increases extracellular D O P A C levels (as measured by in vivo voltammetry) in the rat nucleus accumbens but not frontal cortex: antagonism by diazepam and Zolpidem. Brain Res. 409: 169-174. d'Argy, R., Persson, A . and Sedvall, G . (1987) A quantitative cerebral and whole body autoradiographic study of a intravenously administered benzodiazepine antagonist 3 H - R o 15-1788 in mice. Psvchopharmacology 92: 8-13. de Wit, H . , Johanson, C . E . and Uhlenhuth, E . H . (1983) Drug preference in humans: lorazepam. In Problems of Drug Dependence. NIDA Research Monograph Series 49: 227-232. de Wit, H . , Uhlenhuth, E . H . , Hedeker, D., McCracken, S.G. and Johanson, C . E . (1986b) Lack of preference for diazepam in anxious volunteers. Arch. Gen. Psychiatry 43: 533-541. 106 de Wit, H . , Uhlenhuth, E . H . and Johanson, C . E . (1984) Lack of preference for flurazepam in normal volunteers. Pharmacol. Biochem. Behav. 21: 865-869. de Wit, H . , Uhlenhuth, E . H . and Johanson, C . E . (1985) Drug preference in normal volunteers: effects of age and time of day. Psvchopharmacology 87: 186-193. de Wit, H . , Uhlenhuth, E . H . and Johanson, C . E . (1986a) Individual differences in the reinforcing and subjective effects of amphetamine and diazepam. Drug and  Alcohol Dep. 16: 341-360. Di Chiara, G . and Imperato, A . (1988) Drugs abused by humans preferentialy increase synaptic dopamine concentrations in the mesolimbic system of freely moving rats. Proc. Natl. Acad. Sci. 85: 5274-5278. Dundee, J.W., Halliday, N.J., Harper, K.W. and Brogden, R .N . (1984) Midazolam: a review of its pharmacology properties and therapeutic use. Drugs 28: 519-543. Fabre, L . F . , McLendon, D . M . and Harris, R.T. (1976) Preference studies of triazolam with standard hypnotics in out-patients with insomnia. J. Int. Med. Res. 4: 247-254. Fadda, F. , Argiolas, A . , Melis, M.R. , Tissari, A . H . , Onali, P.L. and Gessa, G . L . (1978) Stress-induced increase in 3,4-dihydroxyphenylacetic acid (DOPAC) levels in the cerebral cortex and in n. accumbens: reversal by diazepam. Life Sci. 23: 2219-2224. Falk, J .L. and Tang, M . (1985) Midazolam oral self-administration. Drug Alcohol.  Depend. 15: 151-163. Falk, J .L. and Tang, M . (1987) Development of physical dependence on midazolam by oral self-administration. Pharmacol. Biochem. Behav. 26: 797-800. Fibiger, H.C. and Phillips, A . G . (1986) Reward, motivation, cognition: psychobiology of mesotelencephalic dopamine systems. In F . E . Bloom and S.R. Geiger (Eds.) Handbook of Physiology - The Nervous System IV. American Physiological Society, Bethesda, pp. 647-668. Fibiger, H .C . and Phillips, A . G . (1988) Mesocorticolimbic dopamine systems and reward. In P.W. Kalivas and C.B. Nemeroff (Eds.) Annals of the New York Academy  of Sciences. Vol 537, New York, pp 206-213. File, S.E. (1986) Aversive and appetitive properties of anxiogenic and anxiolytic agents. Behav. Brain Res. 21: 189-194. Findley, J.D., Robinson, W.W. and Peregrino, L . (1972) Addiction to secobarbital and chlordiazepoxide in rhesus monkey by means of a self-infusion preference procedure. Psvchopharmacologia 26: 93-114. 107 Finlay, J . M . , Szostak, C. and Fibiger, H.C. (1989) Further characterization of midazolam self-administration in the rat. Behav. Pharmacol., in press. Fuchs, V . , Burbes, E . and Coper, H . (1984) The influence of haloperidol and aminooxyacetic acid on etonitaxene, alcohol, diazepam and barbital consumption. Drug and Alcohol Dep. 14: 179-186. Fuxe, K . , Agnati, L . F . , Bolme, P., H6kfelt, T. , Lidbrink, P., Ljungdahl, A . , Perez de la Mora, M . and Ogren, S-O. (1975) The possible involvement of G A B A mechanisms in the action of benzodiazepines on central catecholamine neurons. In E . Costa and P. Greengard (Eds.) Mechanism of Action of  Benzodiazepines. Raven Press, New York, pp. 45-59. Giorgi, O., Corda, M . G . , Gritti, I., Mariotti, M . , Ongini, E . and Biggio, G . (1989). Binding sites of [^H]2-oxo-quazepam in the brain of the cat: evidence for heterogeneity of benzodiazepine recognition sites. Neuropharmacology 28: 715-718. Goldberg, S.R., Hoffmeister, F . , Schlichting, U . U . and Wuttke, W. (1971) A comparison of pentobarbital and cocaine self-administration in rhesus monkeys: effects of dose and fixed-ratio paramenter. J. Pharmacol. Exp. Ther. 179: 277-282. Gotestam, K . G . (1973) Intragastric self-administration of medazepam in rats. Psvchopharmacologia 28: 87-94. Greenblatt, D.J. and Shader, R.I. (1981) Clinical use of the benzodiazepines. Rational  Drug Ther. 15: 1-6. Greenblatt, D.J. and Shader, R.I. (1984) Short half-life benzodiazepines. Rational Drug  Ther. 18: 1-4. Grace, A . A . and Bunney, B.S. (1979) Paradoxical G A B A excitation of nigral dopaminergic cells: indirect mediation through reticulata inhibitory neurons. Eur. J. Pharmacol. 59: 211-218. Grace, A . A . and Bunney, B.S. (1985) Opposing effects of striatonigral feedback pathways on midbrain dopamine cell activity. Brain Res 333: 271-284. Griffiths, R.R. and Ator, N.A. (1981) Benzodiazepine self-administrationin animals and humans: A comprehensive literature review. In S.I. Szara and J.P. Ludford (Eds.) Benzodiazepines: A Review of Research Results. NIDA Research Monograph 33, pp. 22-36. Griffiths, R.R., Bigelow, G . E . and Liebson, I. (1976) Human sedative self-administration: effects of interingestion interval and dose. J. Pharmacol. Exp.  Ther. 197: 488-493. 108 Griffiths, R.R. , Bigelow, G . E . and Liebson, I. (1979) Human drug self-administration: double-blind comparison of pentobarbital, diazepam, chorpromazine and placebo. J. Pharmacol. Exp. Ther. 210: 301-309. Griffiths, R.R. , Bigelow, G . E . , Liebson, I. and Kaliszak, J.E. (1980) Drug preference in humans: double-blind choice comparison of pentobarbital, diazepam and placebo. J. Pharmacol. Exp. Ther. 215: 649-660. Griffiths, R.R. , Lukas. S.E., Bradford, L . D . , Brady, J.V. and Snell, J.D. (1981) Self-injection of barbiturates and benzodiazepines in baboons. Psvchopharmacology 75: 101-109. Griffiths, R.R. , McLeod, D.R., Bigelow, G . E . , Liebson, L A . , Roache, J.D. and Nowowieski, P. (1984) Comparison of diazepam and oxazepam: preference, liking and extent of abuse. J. Pharmacol. Exp. Ther. 229: 501-507. Hackett, D. and Hall, J . M . (1976) Reinforcing properties of intravenous diazepam in rhesus monkeys (macaca mulatta) with a history of codeine self-administration. In: W.A.M. Duncan (Ed.) Clinical Toxicology - Proceedings  of the 18th Meeting of the European Society of Toxicology. Excerpta Medica, Amsterdam, pp. 308-310. Harris, R .T . , Claghorn, J .L. and Schoolar, J. (1968) Self administration of minor tranquilizers as a function of conditioning. Psvchopharmacologia 13: 81-88. Healey, M . L . and Pickens, R.W. (1983) Diazepam dose preference in humans. Pharmacol.  Biochem. Behav. 18: 449-456. Hubbard, B. and Kripke, D.F. (1976) Hypnotic and minor tranquilizer use among inpatients and after discharge. Int. J. Addict. 11: 403-408. Hoffmeister, F. (1977) Assessment of reinforcing properties of stimulant and depressant drugs in the rhesus monkey as a tool for the prediction of psychic dependence-producing capability in man. In T. Thompson (Ed.) Predicting  Dependence Liability of Stimulant and Depressant Drugs. University Park Press, Baltimore, pp. 185-201. Howard-Butcher, S., Blaha, C D . and Lane, R.F. (1983) Differential effects of xylamine on extracellular concentrations of norepinephrine and dopamine in rat central nervous system: an in vivo electrochemical study. J. Pharmacol. Exp. Ther. 233: 58-62. Hunkeler, W., Mohler, H . , Pieri, L . , Pole, P., Bonetti, E.P., Cumin R., Schaffner, R. and Haefely, W. (1981) Selective antagonists of benzodiazpines. Nature 290: 514-516. Imperato, A . , Di Chiara, G . (1984) Trans-striatal dialysis coupled to reverse phase high performance liquid chromatography with electrochemical detection: a new 109 method for the study of the in vivo release of endogenous dopamine and metabolites. J. Neurosci. 4: 966-977. Ishiko, J . , Inagaki, C. and Takaori, S. (1983) Effects of diazepam, nitrazepam and brotizolam on dopamine turnover in the olfactory tubercle, nucleus accumbens and caudate nucleus of rats. Japan J. Pharmacol. 33: 706-708. Jick, H . , Slone, D. , Dinan, B. and Muench, H . (1966) Evaluation of drug efficacy by a preference technic. New England J. Med. 275: 1399-1403. Johanson, C . E . and Uhlenhuth, E . H . (1980) Drug preference and mood in humans: diazepam. Psvchopharmacology 71: 269-273. Johanson, C . E . (1987) Benzodiazepine self-administration in rhesus monkeys: estazolam, flurazepam and lorazepam. Pharmacol. Biochem. Behav. 26: 521-526. Johanson, C . E . and Schuster, C. (1986) The effects of Rol5-1788 on anxiolytic self-administration in the rhesus monkey. Pharmacol. Biochem. Behav. 24: 855-859. Justice, J.B. (1987) Introduction to in vivo voltammetry. In J.B. Justice (Ed.) Voltammetrv in the Neurosciences. Humana Press, New Jersey, pp. 3-88. Kamano, D . K . and Arp, D.J. (1965) Chlordiazepoxide (Librium) consumption under stress conditions in rats. Int. J. Neuropsvch. 1: 189-192. Kirk , R.F. (1968) Experimental Design: Procedures for the Behavioral Sciences. Brooks/Cole, Belmont, California. Klotz, U . , Ziegler, G . and Reimann, I.W. (1984) Pharmacokinetics of the selective benzodiazepine antagonist Ro 15-1788 in man. Eur. J. Clin. Pharmacol. 27: 115-117. Kubota, A . , Kuwahara, A . , Hakkei, A . and Nakamura, K . (1986) Drug dependence tests on a new anesthesia inducer, midazolam. Folia Pharmacol. Jpn. 88: 125-158. Lane, R .F . , Blaha, C D . and Hari, S.P. (1987) Electrochemistry in vivo: monitoring dopamine release in the brain of the conscious, freely moving rat. Brain Res.  Bull. 19: 19-27. Lane, R .F . , Hubbard, A . T . and Blaha, C D . (1979) Application of semidifferential electroanalysis to studies of neurotransmitters in the central nervous system. L Electroanal. Chem. 95: 117-122. Lukas, S.E. and Griffiths, R.R. (1982) Comparison of triazolam and diazepam self-administration by the baboon. Pharmacologist 24: 133. 110 Mackintosh, N.J. (1974) The Psychology of Animal Learning. Academic Press, San Francisco. Martin, I.L. (1987) The benzodiazepines and their receptors: 25 years of progress. Neuropharmacology 26: 957-970. Mereu, G . , Fanni, B., Serra. M . , Concas, A . and Biggio, G . (1983) Beta-carbolines activate neurons in the substantia nigra pars reticulata: an effect reversed by diazepam and Rol5-1788. Eur. J. Pharmacol. 96: 129-132 McKearney, J.W. (1973) Effects of methamphetamine and chlordiazepoxide on schedule controlled and adjunctive licking in the rat. Psvchopharmacologia 30: 375-384. M6hler, H . , Okada, T. , Heitz, P.H. and Ulrich, J. (1978) Biochemical identification of the site of action of benzodiazepines in human brain by H-diazepam binding. Life Sci. 22: 985-996. Naruse, T. and Asami, T. (1987) Intravenous self-administration of diazepam in rats. Eur. J. Pharmacol. 135: 365-373. Nomikos, G . G . and Spyraki, C. (1988) Effects of ritanserin on the rewarding properties of d-amphetamine, morphine and diazepam revealed by conditioned place preference in rats. Pharmacol. Biochem. Behav. 30: 853-858. O'Brien, D.P. and White, F.J. (1987) Inhibition of non-dopamine cells in the ventral tegmental area by benzodiazepines: relationship to A10 dopamine cell activity. Eur. J. Pharmacol. 142: 343-354. Owen, R.T. and Tyrer, P. (1983) Benzodiazepine dependence. Drugs 25: 385-398. Paxinos, G . and Watson, C. (1986) The Rat Brain in Stereotaxic Coordinates. 2nd Ed. , Academic Press, London. Pellegrino, L . J . , Pellegrino A.S. and Cushman, A .J . (1967) A Stereotaxic Atlas of the Rat  Brain. 2nd Ed. , Plenum Press, New York. Petursson, H . and Lader, M . (1984) Dependence on tranquilizers. In G . Russell and E . Marley (Eds.) Maudslev Monographs. Number 28. Oxford University Press, New York, pp. 1-130. Phillis, J.W. and O'Regan, M . H . (1988) The role of adenosine in the central actions of the benzodiazepines. Prog. Neuro-Psvchopharmacol Biol. Psvchiat. 12: 389-404. Pickens, R. and Harris, W.C. (1968) Self-administration of d-amphetamine by rats. Psvchopharmacolgia 12: 158-163. I l l Pickens, R. and Thompson, T. (1968) Cocaine-reinforced behavior in rats: effects of reinforcement magnitude and fixed-ratio size. J. Pharmacol. Exp. Ther. 161: 122-128. Pieri, L . , Schaffner, R., Scherschlicht, R., Pole, P., Sepinwall, J . , Davidson, A . , M6hler, H . , Cumin, R., Da Prada, M . , Burkard, W.P., Keller, H . H . , Muller, R . K . M . , Gerold, M . , Pieri, M . , Cook, .L and Haefely, W. (1981) Pharmacology of midazolam. Arzneim-Forsch/Drug Res 31: 2180-2200. Pilotto, R., Singer, G . and Overstreet, D. (1984) Self-injection of diazepam in naive rats: effects of dose, schedule and blockade ofdifferent receptors. Psvchopharmacology 84: 174-177. Pole, P. (1989) Electrophysiology of benzodiazepine receptor ligands: multiple mechanisms and sites of action Prog.Neurobiol. 31: 349-423. Pole, P., Laurent, J.P., Scherschlicht, R. and Haefely, W. (1981) Electrophysiological studies on the specific benzodiazepine antagonist Ro 15-1788. Naunvn- Schmiedeberg's Arch. Pharmacol. 316: 317-325. Pinel, J.P.J., Mana, M.J . and Ward, J.A. (1989) Stretched-approach sequences directed at a localized shock source by rattus norvegicus. J. Comp. Psychol., in press. Ross, R.J. , Waszczak, B.L. , Lee, E . K . and Walters, J.R. (1982) Effects of benzodiazepines on single unity activity in the substantia nigral pars reticulata. Life Sci. 31: 1025-1035. Sadzot, B., Frost, J.J. and Wagner, H.N. (1989) In vivo labeling of central benzodiazepine receptors with the partial inverse agonist [ H]Ro 15-4513. Brain Res. 491: 128-135. Sanger, D.J. (1977) Schedule-induced drinking of chlordiazepoxide solutions by rats. Pharmacol. Biochem. Behav. 7: 1-6. Sanger, D.J. and Blackman, D .E . (1975) The effects of chlordiazepoxide on the development of adjunctive drinking in rats. Quart. J. Exp. Psychol. 27: 499-505. Schoch, P., Richards, J .G. , Haring, P., Takacs, B., Stahli, C , Staehelin, T. , Haefely, W. and Mdhler, H . (1985) Co-localization of G A B A A receptors and benzodiazepine receptors in brain shown by monoclonal antibodies. Nature 314: 168-170. Smith, J.B. and Clark, F .C . (1975) Effects of d-amphetamine, chlorpromazine and chlordiazepoxide on intercurrent behavior during spaced responding schedules. J. Exp. Anal. Behav. 24: 241-248. 112 Spyraki, C. and Fibiger, H.C. (1988) A role for the mesolimbic dopamine system in the reinforcing properties of diazepam. Psvchopharmacology 94: 133-137. Spyraki, C , Kanandjian, A . and Varonos, D. (1985) Diazepam-induced place preference conditioning: appetitive and antiaversive properties. Psvchopharmacology 87: 225-232. Stolerman, LP. , Kumar, R. and Steinberg, H . (1971) Development of morphine . dependence in rats: lack of effect of previous ingestion of other drugs. Psvchopharmacologia 20: 321-336. Szostak, C , Finlay, J . M . and Fibiger, H.C. (1987) Intravenous self-administration of the short-acting benzodiazepine midazolam in the rat. Neuropharmacology 26: 1673-1676. Trulson, M . E . (1985) Simultaneous recording of substantia nigra neurons and voltammetric release of dopamine in the caudate of behaving cats. Brain Res.  Bull. 15: 221-223. Ungerstedt, U . (1984) Measurement of neurotransmitter release by intracranial dialysis. In C A . Marsden (Ed.) Measurement of Neurotransmitter Release in Vivo. Wiley, New York, pp. 81-105. Van der Poel, A . M . (1979) A note on 'stretched attention', a behavioural element indicative of an approach-avoidance conflict in rats. Anim. Behav. 27: 446-450. Versteeg, D . H . G . , Van der Gugten, J . , De Jong, W. and Palkovits, M . (1976) Regional concentrations of noradrenaline and dopamine in rat brain. Brain Res. 113: 563-574. Walton, N.Y. and Deutsch, J.A. (1978) Self-administration of diazepam by the rat. Behav. Biol. 24: 533-538. Weeks, J.R. and Collins, R.J. (1979) Dose and physical dependence as factors in the self-administration of morphine by rats. Psvchopharmacology 65: 171-177. Weaver, D.D. , Phillips, P.E. and Altshler, H . L . (1975) Intragastric self-administration of sedative-hypnotic drugs by the rhesus monkey. Pharmacologist 17: 211. Westerink, B.H.C. and Mulder, T.B.A. (1981) Determination of picomole amounts of dopamine, noradrenaline, 3,4-digydroxyphenylalanine, 3,4 r dihydroxyphenylacetic acid, homovanillic acid, and 5-hydroxyindoleacetic acid in nervous tissue after one-step purification onf Sephadex G 10, using high-performance liquid chromatography with a novel type of electrochemical detection. J. Neurochem. 36: 1449-1462. 113 Westerink, B.H.C. and Tuinte, M.H.J . (1986) Chronic use of intracerebral dialysis for the in vivo measurement of 3,4-digydroxyphenylethylamine and its metabolite 3,4-dihydroxyphenylacetic acid. J. Neurochem. 46: 181-184. White, F.J. and Wang, R.Y. (1984) A10 dopamine neurons: role of autoreceptors in determining firing rate and sensitivity to dopamine agonists. Life Sci. 34: 1161-1170. Winstead, D . K . , Anderson, A . , Eilers, M . K . , Blackwell, B. and Zaremba, A . L . (1974) Diazepam on demand. Arch. Gen. Psychiatry 30: 349-351. Wise, R . A . and Bozarth, M . A . (1987) A psychomotor stimulant theory of addiction. Psychological Review 94: 469-492. Wolf, G . , Jacquet, Y. and Carol, M . (1978) Test for oral and postingestional factors mediating differential acceptability of morphine, methamphetamine, and chlordiazepoxide drinking solutions. Psvchopharmacology 60: 101-102. Woods, J .H. , Katz, J .L. and Winger, G . (1987) Abuse liability of benzodiazepines. Pharmacol. Reviews 39: 251-393. Yanagita, T. (1981) Dependence-producing effects of anxiolytics. In F. Hoffmeister (Ed.) Handbook of Experimental Pharmacology. Vol.55, Part 2, Springer-Verlag, New York, pp. 395-406. Yanagita, T. (1983) Dependence potential of zopiclone studied in monkeys. Pharmacology 27 (suppl 2): 216-227. Yanagita, T. and Takahashi, S. (1973) Dependence liability of several sedative-hypnotic agents evaluated in monkeys. J. Pharmacol. Exp. Ther. 185: 307-315. Yim, C Y . and Mogenson, G.J. (1980) Electrophysiological studies of neurons in the ventral tegmental area of tsai. Brain Res. 181: 301-313. Young, W.S. and Kuhar, M.J . (1980) Radiohistochemical localization of benzodiazepine receptors in rat brain. J. Pharmacol. Exp. Ther. 212: 337-345. Zetterstrom, T . , Pei, Q. and Fillenz, M . (1989) Measurements of dopamine and 5-hydroxtryptamine release after anxiolytic drugs using microdialysis. Current  Separations. International Symposium on Microdialysis 9: 91. Janet M . Finlay PUBLICATIONS Corcoran, M . E . , Cain, D.P., Finlay, J .M. and Gillis, B.J. (1984) Vasopressin and the kindling of seizures. Life Sciences 35: 947-952. Finlay, J . M . , Jakubovic, A . , Fu, D.S. and Fibiger, H.C. (1987) Tolerance to haloperidol-induced increases in dopamine metabolites: Fact or artifact? European Journal of Pharmacology 137: 117-121. Szostak, C , Finlay, J . M . and Fibiger, H.C. (1987) Intravenous self-administration of the short-acting benzodiazepine midazolam in the rat. Neuropharmacology 26: 1673-1687. Finlay, J . M . , Jakubovic, A . , Phillips, A . G . and Fibiger, H.C. (1988) Fentanyl-induced conditional place preference: Lack of associated conditional neurochemical events. Psychopharmacology 96: 534-540. Finlay, J .M. , Szostak, C. and Fibiger, H.C. (1989) Further characterization of intravenous self-administration of midazolam in the rat. Behavioural Pharmacology, in press. PUBLISHED A B S T R A C T S Finlay, J .M. , Jakubovic, A . , Fu, D.S., Phillips, A . G . and Fibiger H.C. (1986) Effects of haloperidol and conditional stimuli paired with haloperidol on central dopamine turnover. 9 Annual Meeting of the Canadian College of Neuropsychopharmacology. Finlay, J .M. , Szostak, C. and Fibiger, H.C. (1986) Intravenous self-administration of midazolam in rats. 16 Annual Meeting of the Society for Neuroscience. Finlay, J .M. , Jakubovic, A . , Phillips, A . G . and Fibiger, H.C. (1987) Fentanyl-induced conditional place preference: Lack of associated conditional neurochemical events. 48 Annual Meeting of the Canadian Psychological Association. Finlay, J .M. , Szostak, C , Blaha, C D . , Lane, R.F. and Fibiger H.C . (1987) Self-administration of midazolam may be associated with a decrease in dopamine release in the nucleus accumbens. 17 Annual Meeting of the Society for Neuroscience. Finlay, J . M . , Fibiger, H . C , Blaha, C D . and Phillips, A . G . (1988) In vivo electrochemical detection of dopamine within the nucleus accumbens of rats self-administering cocaine. 18 Annual Meeting of the Society for Neuroscience. Finlay, J . M . , Damsma, G . , Wenkstern, D. and Fibiger, H.C. (1989) Effects of the benzodiazepine midazolam on extracellular dopamine concentrations in the nucleus accumbens and striatum. 191*1 Annual Meeting of the Society for Neuroscience. 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0098208/manifest

Comment

Related Items