The Involvement of Ventral Tegmental Area Dopamine and CRF Activity in Mediating the Opponent Motivational Effects of Acute and Chronic Nicotine

Transcription

1 The Involvement of Ventral Tegmental Area Dopamine and CRF Activity in Mediating the Opponent Motivational Effects of Acute and Chronic Nicotine by Taryn Elizabeth Grieder A thesis submitted in conformity with the requirements for the degree of Doctor of Philosophy Institute of Medical Science University of Toronto Copyright by Taryn Elizabeth Grieder 2012

2 The Involvement of Ventral Tegmental Area Dopamine and CRF Activity in Mediating the Opponent Motivational Effects of Acute and Chronic Nicotine Taryn Elizabeth Grieder Doctor of Philosophy Institute of Medical Science University of Toronto 2012 Abstract A fundamental question in the neurobiological study of drug addiction concerns the mechanisms mediating the motivational effects of chronic drug withdrawal. According to one theory, drugs of abuse activate opposing motivational processes after both acute and chronic drug use. The negative experience of withdrawal is the opponent process of chronic drug use that drives relapse to drug-seeking and -taking, making the identification of the neurobiological substrates mediating withdrawal an issue of central importance in addiction research. In this thesis, I identify the involvement of the neurotransmitters dopamine (DA) and corticotropinreleasing factor (CRF) in the opponent motivational a- and b-processes occurring after acute and chronic nicotine administration. I report that acute nicotine stimulates an initial aversive a-process followed by a rewarding opponent b-process, and chronic nicotine stimulates a rewarding a-process followed ii

3 by an aversive opponent b-process (withdrawal). These responses can be modeled using a place conditioning paradigm. I demonstrate that the acute nicotine a-process is mediated by phasic dopaminergic activity and the DA receptor subtype-1 (D1R) but not by tonic dopaminergic activity and the DA receptor subtype-2 (D2R) or CRF activity, and the opponent b-process is neither DA- nor CRF-mediated. I also demonstrate that the chronic nicotine a-process is DA- but not CRF-mediated, and that withdrawal from chronic nicotine (the b-process) decreases tonic but not phasic DA activity in the ventral tegmental area (VTA), an effect that is D2R- but not D1Rmediated. I show that a specific pattern of signaling at D1Rs and D2Rs mediates the motivational responses to acute nicotine and chronic nicotine withdrawal, respectively, by demonstrating that both increasing or decreasing signaling at these receptors prevents the expression of the conditioned motivational response. Furthermore, I report that the induction of nicotine dependence increases CRF mrna in VTA DA neurons, and that blocking either the upregulation of CRF mrna or the activation of VTA CRF receptors prevents the anxiogenic and aversive motivational responses to withdrawal from chronic nicotine. The results described in this thesis provide novel evidence of a VTA DA/CRF system, and demonstrate that both CRF and a specific pattern of tonic DA activity in the VTA are necessary for the aversive motivational experience of nicotine withdrawal. iii

4 Acknowledgments It s pretty shocking that a full seven years have passed since I began the amazing experience that is graduate school. There are so many people who have influenced my life throughout this time that it s hard to decide where to begin. However, as we all know, everything ultimately began with my parents, Sherri and Gord, who have been the most loving, supportive, and motivating people throughout my studies and life as a whole. I couldn t have asked for better friends and fans. A huge thanks to my supervisor and mentor, Derek van der Kooy, aka the most knowledgeable man I ve ever met. Without the excellent guidance and gentle (ok, sometimes harsh) constructive criticism I received along the way, I would not have learned and grown into the scientist I pride myself in being today. Thanks as well to my program advisory committee members Rachel Tyndale, Bernard Le Foll and Larry Grupp for all their advice and help, and especially the final push to graduate. To my love, Oleg, thanks for the support and for putting up with me (and Mini-Me)! A special thanks to my collaborator, co-author, occasional partner in crime, and always beloved friend, Olivier George. Our many hours spent dreaming up mad scientific genius ideas have actually paid off! Thanks as well to my best friends Michelle and Katie, my sisters Jenna and Holly, Kyle, Lauren, Susan and my lacrosse teammates (FTG!), and of course my unstoppable dog Bender. Although none of you really understand what I do except that it involves nicotine and killing mice for their brains, the love and support you ve all given me has been simply amazing. No acknowledgement would be complete without mention of the amazing people I ve worked with in the van der Kooy lab throughout the years. My desk buddy Jessica, fellow motis Ryan and Drew, laboids Mary Rose, Simon, Rachel and Brenda, and of course my strangest and dearest friend and roomie, Hector Vargas-Perez. I love you, d che. Without all of you to complain, sympathize, and party with, these years would not have flown by so quickly. I dedicate this thesis to my family: Luka, Oleg, and the Grieders. I love you more than words could ever say. iv

5 Contributions For Chapter 2: Dopaminergic Signaling Mediates the Motivational Response Underlying the Opponent Process to Chronic but Not Acute Nicotine. Authors: Taryn E. Grieder, Laurie H. Sellings, Hector Vargas-Perez, Ryan Ting-A-Kee, Eric C. Siu, Rachel F. Tyndale and Derek van der Kooy. Author contributions: T.E.G., O.G., B.L.F. and D.V.D.K. designed the experiments. S.G. provided D1KO mice. T.E.G. performed the minipump surgeries and the place conditioning experiments, and H.T. and S.R.L. performed the electrophysiology. T.E.G. and O.G. analyzed the data. T.E.G., O.G. and D.V.D.K. wrote the paper. All authors discussed the results and commented on the manuscript. For Chapter 3: Phasic D1 and Tonic D2 Dopamine Receptor Signaling Double Dissociate the Motivational Effects of Acute Nicotine and Chronic Nicotine Withdrawal. Authors: Taryn E. Grieder, Olivier George, Huibing Tan, Susan R. George, Bernard Le Foll, Steven R. Laviolette and Derek van der Kooy Author contributions: T.E.G., O.G., B.L.F. and D.V.D.K. designed the experiments. S.G. provided D1KO mice. T.E.G. performed the minipump surgeries and the place conditioning experiments, and H.T. and S.R.L. performed the electrophysiology. T.E.G. and O.G. analyzed the data. T.E.G., O.G. and D.V.D.K. wrote the paper. All authors discussed the results and commented on the manuscript. For Chapter 4: Recruitment of a VTA CRF system mediates the aversive effects of nicotine withdrawal. Authors: Taryn E. Grieder, Hector Vargas-Perez, Candice Contet, Laura A. Tan, John Freiling, v

6 Laura Clarke, Elena Crawford, Pascale Koebel, Brigitte L. Kieffer, Paul E. Sawchenko, George F. Koob, Derek van der Kooy and Olivier George. Author Contributions: TEG and OG designed the experiments. TEG and HVP performed minipump, cannulation and viral vector surgeries. TEG performed place conditioning and open field testing. CC, LAT and PES performed in situ hybridization. CC performed double in situ hybridization and immunohistochemistry. TEG and LC performed rtpcr. JF and EC performed immunohistochemistry. CC, PK and BLK supplied viral vectors. TEG analyzed the data. TEG, CC, GFK, DVDK and OG wrote the paper. All authors discussed the results and read the paper. vi

7 Table of Contents Abstract... ii Acknowledgments... iv List of Figures... ix List of Abbreviations... xii Chapter General Introduction What is Motivation? The Study and Measurement of Motivation... 3 Operant conditioning procedures... 4 Classical conditioning procedures Models of Drug Motivation The Mesolimbic Dopamine Reward Hypothesis The error prediction model The incentive- sensitization theory The non- deprived/deprived hypothesis The opponent process theory The Anatomy of the Ventral Tegmental Area Neurons and Projections Nicotinic Receptors The Neurobiology of Nicotine Motivation: DA and CRF The Use and Abuse of Nicotine VTA Dopamine Research Aims and Hypotheses Chapter Dopaminergic Signaling Mediates the Motivational Response Underlying the Opponent Process to Chronic but Not Acute Nicotine Abstract Introduction vii

8 Materials and Methods Results Discussion Chapter Phasic D1 and Tonic D2 Dopamine Receptor Signaling Double Dissociate the Motivational Effects of Acute Nicotine and Chronic Nicotine Withdrawal Abstract Introduction Materials and methods Results Discussion Chapter Recruitment of a VTA CRF system mediates the aversive effects of nicotine withdrawal.. 98 Abstract Introduction Materials and methods Results Discussion Chapter General Discussion Overview Overall conclusion Future directions References Copyright Acknowledgements viii

9 List of Figures Figure 1.1. The place conditioning paradigm. 9 Figure 1.2. The opponent process theory of motivation.. 24 Figure 1.3. The allostatic state of drug addiction Figure 1.4. The VTA: Neurons, receptors, inputs and projections 30 Figure 2.1. The opponent process theory of motivation and its modeling by use of the place conditioning paradigm 46 Figure 2.2. The time course of spontaneous nicotine somatic and motivational withdrawal Figure 2.3. The opponent processes of chronic and acute nicotine and the effect of DA antagonism.. 57 Figure 2.4. Dopaminergic signaling differentially mediates the opponent motivational process after acute and chronic nicotine Figure 2.5. The D2R mediates the aversive response to chronic nicotine withdrawal Figure 3.1. Phasic DA activity mediates aversions to acute nicotine while the specific pattern of tonic DA activity mediates aversions to withdrawal from chronic nicotine Figure 3.2. The DAR agonist and antagonist have no motivational effects on their own Figure 3.3. The cannabinoid-1 receptor inverse agonist rimonabant significantly decreases phasic VTA DA activity but does not affect tonic DA activity ix

10 Figure 3.4. A specific pattern of signaling at D1Rs is required for aversions to acute nicotine in nondependent mice, while a specific pattern of D2R activity is required for aversions to nicotine withdrawal in dependent mice Figure 3.5. D1R and D2R agonists and antagonists have no motivational effects on their own at the doses used in this study, however a high dose of the D1R agonist prevents learning Figure 3.6. Manipulations of the A2AR block the aversive response to withdrawal from chronic nicotine but not acute nicotine Figure 4.1. Nicotine dependence increases CRF mrna levels in the VTA. 108 Figure 4.2. Nicotine dependence and withdrawal recruits and activates the CRF system in the VTA Figure 4.3. Nicotine dependence increased the number of cells that contain CRF mrna in the pvta but not avta Figure 4.4. Double labeling of DA neurons and CRF mrna using CRF in situ hybridization and TH immunohistochemistry Figure 4.5. Withdrawal from chronic nicotine depletes CRF peptide in the pvta Figure 4.6. Nicotine dependence and withdrawal decreases CRF peptide density in the CeA but not the PVN. 117 Figure 4.7. Downregulation of CRF mrna in the VTA by a viral vector prevents the aversive motivational response to withdrawal from chronic nicotine. 119 Figure 4.8. The silencing vector must be injected in the VTA to block withdrawal aversions. 122 Figure 4.9. CRF1R antagonism prevents the aversive motivational response to withdrawal from chronic nicotine 124 x

11 Figure The opponent motivational process occurring after acute nicotine is not blocked by CRF1R antagonism Figure 5.1. Summary of the involvement of DA and CRF in the opponent motivational responses occurring after acute and chronic nicotine xi

12 List of Abbreviations α-flu ANOVA AP avta A2AR CeA CRF CRF-BP CRF1R CRF2R CMV DA DNA DV D1R D2R EGFP ELISA GABA α-flupenthixol analysis of variance anterior-posterior anterior ventral tegmental area adenosine receptor subtype-2a central nucleus of the amygdala corticotropin-releasing factor corticotropin-releasing factor-binding protein corticotropin-releasing factor receptor subtype-1 corticotropin-releasing factor receptor subtype-2 cytomegalovirus dopamine deoxyribonucleic acid dorsal-ventral dopamine receptor subtype-1 dopamine receptor subtype-2 enhanced green fluorescent protein enzyme-linked immunosorbent assay gamma-aminobutyric acid xii

13 GAD Hz i.p. KO ML MΩ mrna ms NAc nachr NMDA PBS PFC PVN pvta RNA rtpcr s.c. sirna SEM glutamic acid decarboxylase hertz intraperitoneal knockout medial-lateral milliohm messenger ribonucleic acid millisecond nucleus accumbens nicotinic acetylcholine receptor N-methyl-D-aspartic acid phosphate buffered saline prefrontal cortex paraventricular nucleus posterior ventral tegmental area ribonucleic acid real-time polymerasechain reaction subcutaneous small interfering ribonucleic acid standard error of the mean xiii

14 TH TPP VP VTA WT tyrosine hydroxylase tegmental pedunculopontine nucleus ventral pallidum ventral tegmental area wild-type xiv

15 Chapter 1 General Introduction 1

16 1 General Introduction 1.1 What is Motivation? Broadly defined, motivation is the internal driving force that initiates, guides and maintains goal-directed behaviour. Motivation is what causes an organism to act, whether it is simply getting out of bed in the morning, finding water or food to reduce thirst or hunger, reading a book to gain knowledge, or taking a drug to feel its pleasurable effects. These are examples of conditioned motivation, because the organism learns to associate the benefits obtained from the action, leading to future motivated responses. However, motivation is not simply appetitive or approach behaviour, as organisms may also be motivated to avoid experiences, stimuli, or environments that are perceived to be bad or aversive. Motivation involves the biological, emotional and social forces that activate behaviour. In everyday usage, the term motivation is frequently used to describe why a person or animal does something. Whether the organism approaches, avoids, or is neutral toward a particular stimulus, their action depends on the motivational context within which the stimulus has been perceived. This process of attributing motivational value to a particular stimulus allows organisms to make associations between the cues they encounter and positive or negative outcomes. In this sense, motivation can be quantified using behavioural paradigms that measure an organisms appetitive or aversive response to a particular stimulus. Motivation is thus defined here as the driving force responsible for approach or avoidance behaviour. Motivated behaviour is required for an organism s survival: In the absence of basic motivation, an organism would not eat or drink and would thus perish. However, not all motivated behaviour is beneficial to the organism, as there are instances when motivation drives the actions that lead to compulsive activities. The best example of this, and the focus of this thesis, is drug addiction: the habitual psychological and physical dependence on a substance that is beyond voluntary control and persists in spite of negative consequences. Drugs of abuse such as nicotine, opiates, cocaine or ethanol are capable of producing robust appetitive responding and 2

17 subjective feelings of pleasure on the first exposure. This rewarding experience due to drug exposure is attributed a positive motivational value and leads to subsequent motivated behaviour to obtain more of the drug, eventually leading to physical dependence and withdrawal when drug use is discontinued. The aversive experience of withdrawal then leads to further motivated behaviour to avoid this negative outcome. The studies described in this thesis define the neurobiological substrates mediating the rewarding and aversive motivational responses to acute and chronic nicotine, being primarily concerned with the aversive motivational response occurring after withdrawal from chronic nicotine. 1.2 The Study and Measurement of Motivation Motivation can be studied at the psychological, physiological, sociological, or philosophical levels, however, this thesis will focus on the neurophysiological study of motivation. One of the first neurobiological demonstrations of motivation occurred almost 70 years ago by Olds and Milner after they observed that rats would repeatedly press a lever that led to electrical stimulation of certain brain regions such as the septal area, mammilothalamic tract, cingulate cortex and tegmentum (Olds and Milner, 1954). It was proposed that electrical stimulation in certain areas of the brain produced acquisition and extinction curves that compared with those produced by a primary reward. Furthermore, they observed that stimulation of electrodes placed in other areas of the brain appeared to be punishing or aversive (Olds and Milner, 1954). Subsequent studies reproduced this rewarding self-stimulation phenomena and showed that the amount of reward increased with the amount of electrical stimulation, especially in the area corresponding to a group of ascending and descending axon fibers, the medial forebrain bundle, which passes through the hypothalamic system (Olds et al., 1960). Furthermore, the rewarding effects of electrical brain stimulation in the medial forebrain bundle, but not the septal area, could overshadow the appetitive properties of food even while under food deprivation, with rats preferring to self-stimulate the medial forebrain bundle rather than to eat and satisfy their hunger (Routtenberg and Lindy, 1965). These experiments and many more that 3

18 followed were the first neurobiological demonstration of motivation, showing that stimulation of the medial forebrain bundle and other brain areas was both rewarding and drive inducing, and stimulation of certain other brain areas was aversive and induced avoidance. In the study of motivation, especially the motivational responses to abused drugs, it is most often necessary as well as beneficial to utilize experimental animal subjects. The huge amount of work and progress made in drug addiction research over recent years can be attributed to animal models of drug motivation. However, animals cannot verbalize their feelings or reactions to a drug stimulus, thus their motivation must be inferred by use of an accepted experimental procedure. Motivation has been defined here as the driving force responsible for conditioned approach or avoidance behaviour. As such, there have been a variety of behavioural paradigms developed that can infer and quantify motivation by measuring an animals appetitive or aversive response to a particular stimulus. These paradigms provide a way to assess the neurobiological and behavioural processes underlying drug motivation, factors which may not be easily tested in human subjects due to ethical restrictions. Furthermore, various animal models may be used to investigate the relationship between environmental, developmental, behavioural or neurobiological influences that are hypothesized to contribute to drug addiction. Most of these procedures can be categorized as either operant or classical conditioning. Operant conditioning procedures Operant conditioning is a form of learning where an animal s voluntary (operant) behaviour is modified and maintained by its consequences. The consequences of the animal s behaviour are reinforcement or punishment, which cause subsequent behaviour to occur with more or less frequency, respectively. When a certain behaviour leads to a reward, such as food for a hungry animal, this behaviour is reinforced and the organism s motivation is increased. In other words, the driving force behind their approach behaviour increases. Conversely, if a behaviour leads to punishment, the organism s motivation to repeat that behaviour will be decreased. Additionally, if a certain behaviour is inconsequential, it will eventually extinguish. 4

19 Operant conditioning procedures are typified by self-administration paradigms, where the subject performs an action such as a lever press or nose poke that leads to the delivery of a stimulus. The stimulus can come in many forms, such as a drug or shock delivery, access to food, or electrical stimulation of the brain. The behaviour observed by Olds and Milner that rats would work for stimulation of certain brain areas is an example of self-administration. In the context of this thesis, I will discuss self-administration in terms of psychostimulant drug delivery. The method of drug self-administration can be intravenous, through a surgically implanted catheter, or directly to a specific area of the brain. Different schedules of reinforcement may also be used in this procedure, with the simplest being a fixed ratio schedule of continuous reinforcement. On this schedule, a fixed number of lever presses leads to the delivery of one unit of the drug, which eventually leads to a stable pattern of drug selfadministration (Caine et al., 1993). Other schedules of reinforcement such as a second-order schedule use a fixed number of lever presses that lead to a brief stimulus, usually visual in nature, then the next fixed number of lever presses completed after that stimulus produces the same brief stimulus accompanied by a drug injection (Katz and Goldberg, 1987). The brief stimulus in this schedule of reinforcement will acquire its own reinforcing properties, termed secondary reinforcers (Koob and Le Moal, 2006). Another schedule of reinforcement, which is hypothesized to directly evaluate a drug s reinforcing efficacy, is a progressive-ratio schedule of reinforcement. In this schedule, the number of lever presses required for each successive drug delivery is increased. Drugs that lead to higher numbers of responses are thought to be more reinforcing. Using this method, a break point can be determined where the subject will finally cease responding, and again a higher break point is thought to be produced in drugs of abuse that are more reinforcing (Koob and Le Moal, 2006). In the self-administration paradigm, the rate of responding for a drug delivery is taken as a measure of the reinforcing efficacy of the abused drug. When a stable rate of responding has developed, which often requires initial training with food prior to the introduction of an abused drug, the effects of experimental manipulations can be measured on responding. Common manipulations are pharmacological pretreatment with agonists or antagonists, lesioning of a particular brain region of interest, or more recently genetic methods such as ribonucleic acid 5

20 (RNA) interference or optogenetic approaches (which temporarily activate or silence specific neurons) (Caille et al., 2012; Cao et al., 2011). Although there is no animal model that fully mimics the human condition, selfadministration is the closest procedure to human drug-seeking and -taking, as it allows the animal subjects to regulate their own amount of drug intake. Additionally, the animal may serve as its own control, which reduces the number of animals required per study. However, there are many drawbacks of the self-administration paradigm. For example, a number of abused drugs, such as nicotine, caffeine and marijuana, are not readily self-administered by animals. Some hallucinogenic compounds abused by humans, such as ecstasy, have never been shown to be self-administered by animals (Corrigall, 1999). Nicotine and other drug self-administration is acquired by rats and less often in mice, but usually requires extensive prior training with food self-administration after deprivation. Furthermore, nicotine self-administration does not mimic the inhalation route of administration that it is trying to model in humans (Corrigall, 1999). Self-administration also relies on the animals ability to make a motor response (press a lever or nose poke), thus the use of drugs that decrease motor activity, such as high doses of abused drugs or dopamine (DA) receptor antagonists, may not provide accurate measures of drug-taking behaviour. Similarly, if a drug increases motor activity, more lever presses may occur and an inaccurate measure of drug reward may be obtained. Throughout selfadministration procedures, the animal is under the influence of the drug(s) being studied, making the determination of whether changes in response rate are indeed due to motivational changes or to drug-induced changes in motor responses or other unconditioned motivational effects difficult. Additionally, self-administration procedures require extensive and time-consuming surgeries in addition to the prior training that is often required with food self-administration and deprivation; therefore a single self-administration experiment often requires many months to complete. Importantly, self-administration procedures cannot measure the aversive motivational properties of drugs, simply because animals will not self-administer an aversive substance. This problem could be overcome by either negatively reinforcing the self-administration behaviour, where the animal presses a lever to avoid the administration of an aversive substance, or training 6

21 the animal to self-administer food, then giving the aversive drug and observing if the motivation or obtain food decreases. However these types of procedures are complicated and not widely utilized. Finally, perhaps the most concerning drawback of self-administration paradigms is the fact that both increases or decreases in the amount of responding can be interpreted as increases in the reinforcing properties and motivation to take the abused drug being studied. When an animal increases responding, they are thought to find the outcome pleasurable and therefore want more of the drug. Likewise, when the animal decreases responding, they are thought to have responded less because each drug infusion is more rewarding or pleasurable. The use of a progressive-ratio schedule of reinforcement and the determination of a break point in responding addresses this problem to some extent, but further problems may then arise involving the relation of the rate of responding (which can be influenced by motor effects, detailed above) to break point, and these two factors are not necessarily correlated (Richardson and Roberts, 1996). In summary, although self-administration more closely models human nicotine intake (Rose and Corrigall, 1997), separating drug motivation due to its rewarding effects or the alleviation of withdrawal is more easily performed using a classical conditioning procedure (Mucha et al., 1982). The self-administration paradigm possesses a number of drawbacks that classical conditioning procedures can account for, however, these procedures also posses their own set of limitations which are detailed below. Classical conditioning procedures Classical conditioning is a form of learning in which one stimulus is associated with and becomes a signal or predictor for the occurrence of another stimulus. Behaviours conditioned via a classical conditioning procedure are not maintained by consequences as in operant conditioning, but by the ability of one stimulus to successfully predict a second stimulus. 7

22 Classical conditioning procedures are typified by place conditioning, where animals experience two similar but distinctly different neutral environments that are paired spatially and temporally with distinct drug cues. The animal is passively administered a drug by the investigator and associates the motivational effect of that drug with one of the environments. During the alternate conditioning treatment, the animal is administered a vehicle control treatment and placed in the other distinct environment. These drug- and vehicle-pairings are performed for a fixed amount of time and number of cycles (the conditioning phase), after which the animals undergo preference testing (Figure 1.1). During testing, the animals are given a choice with equal opportunity to enter and explore either environment, and the amount of time spent in the previously drug-paired environment versus the vehicle-paired environment is taken as a measure of the motivational value of the drug. If the choice is made to spend more time in the drug-paired environment relative to the vehicle-paired environment, the drug is considered to be rewarding, and the motivational effect is a conditioned place preference. Conversely, if the animal chooses to spend less time in the drug-paired environment relative to the vehicle-paired environment, the drug is considered to have an aversive motivational effect and a conditioned place aversion is observed. The behaviour observed by Olds and Milner where rats would return to or avoid a particular area of a chamber where they had received electrical brain stimulation in certain areas is an example of place conditioning. The apparatus used in place conditioning experiments typically consists of two or three distinct environments that may be differentiated from each other based on color, texture, smell and/or lighting. The environments must be distinct for conditioning to develop and be observed, as the animal associates the cues experienced in the separate environments with the motivational effects of the drug in question. The distinct conditioning environments should be selected and modified in a way such that the animal can differentiate between them, but should not exhibit a preference for either of them in the absence of any treatment. A procedure with three environments to choose between on testing adds additional controls for nonspecific effects and permits easier balancing between the two environments being used for drug- and vehicle-pairings (Koob and Le Moal, 2006). The pairing of the drug under investigation with a particular environment is fully counterbalanced, with half the animals receiving drug first, the other half 8

23 Figure 1.1. The place conditioning paradigm. Place conditioning is an example of a classical conditioning procedure where animals experience two similar but distinct neutral environments, usually differing in the wall colour and floor texture. During the conditioning phase, the animal is passively administered a drug by the investigator and placed in one of the environments. During the alternate conditioning treatment, the animal is administered a vehicle (usually saline) and placed in the other distinct environment. These drug- and vehicle-pairings are performed for a fixed amount of time and number of cycles, during which the animal associates the motivational effects of the drug in question with the environments it was repeatedly paired with. During preference testing, the animals are given a choice with equal opportunity to enter and explore either environment, and the amount of time spent in the previously drug-paired environment versus the vehicle-paired environment is taken as a measure of the motivational value of the drug. If the choice is made to spend more time in the drug-paired environment relative to the vehicle-paired environment, the drug is considered to be rewarding, and the motivational effect is a conditioned place preference. Conversely, if the animal chooses to spend less time in the drug-paired environment relative to the vehicle-paired environment, the drug is considered to have an aversive motivational effect and a conditioned place aversion is observed. 9

24 10

25 receiving vehicle first, and half of each of those groups receiving one of the two different environments first. In this sense, the place conditioning procedure uses an unbiased design. A biased design is less appealing and more time-consuming, requiring a pre-conditioning phase to assess pretest preferences, after which the drug is usually paired to the least-preferred environment for each individual animal. Using this design, increases in the amount of time spent in the drug-paired environment in comparison to pretest time are taken as a measure of the drug s rewarding properties. The behavioural work presented in this thesis was completed using an unbiased, fully counterbalanced place conditioning paradigm, which will be described in further detail in chapter 2. The most important benefit of the place conditioning procedure is that it permits measurement of both rewarding and aversive motivational effects of abused drugs. Furthermore, the motivational response can often be observed after a single conditioning session with the drug for a variety of abused drugs (Bardo et al., 1986; Grieder et al., 2010; Mucha et al., 1982), allowing for the observation of drug reward or aversion without any induction of tolerance or sensitization (Bardo and Bevins, 2000). Place conditioning procedures are relatively simple to perform in comparison to self-administration procedures, requiring no pre-training and no surgical implantation of a catheter, therefore the time required to set up and perform a place conditioning experiment is considerably less, with some place conditioning experiments being performed in just one week. Also, since the drug is usually not administered during testing, place conditioning is independent of motor responses. In this sense, drugs that reduce motor activities can be readily used in the place conditioning paradigm. However, it is also possible in some studies to examine the effects of administration of the drug during testing on the expression of motivational responses, which serves as a control for the presence of state-dependent learning (learning that is exhibited only during a specific physiological and/or mental state). Place conditioning procedures also allow the investigator to maintain precise control over the amount of drug administered throughout the conditioning phase. However, this benefit also can be considered a drawback, as the drug-seeking and -taking aspect that is modeled by selfadministration is lost when experimenters passively administer a drug during place conditioning 11

26 procedures. Another benefit of the place conditioning paradigm in terms of drug administration is that testing can be and usually is performed in a drug-free state, thus there is no satiety or motor effects on the results obtained. Furthermore, most conditioning apparatuses can measure locomotion simultaneously while recording motivational effects, providing an extra experimental measure for investigation. The place conditioning paradigm is also adaptable to a wide variety of animal models, and could even be modeled in human subjects due to the ease of experimental set up (Bardo and Bevins, 2000). A common criticism of place conditioning procedures is that the animal may have an innate motivational response to a novel environment. To address this concern, control experiments can be performed at the same time with separate groups of animals that are tested for a novelty preference or aversion. Similarly, handling the animals prior to conditioning to familiarize them with injection and handling procedures that will be carried out during the conditioning phase may control for any criticisms about stress effects on conditioning. Finally, drugs often produce motivational effects with narrower dose ranges in place conditioning versus self-administration procedures, leading to dose-response curves that are considered to be less informative. The study of drug motivation by use of operant or classical conditioning procedures represents a similar phenomenon to Olds and Milner s electrical self-stimulation in that drugs of abuse represent stimuli that are not physiologically important. Although abused drugs such as nicotine, opiates, alcohol or cocaine actually act in the exact opposite way as natural rewards such as food and water, producing detrimental effects after continued use, they are nevertheless capable of producing robust appetitive responding in a comparable way to natural rewards. Research examining the neurobiological substrates important for the reinforcing effects of drugs of abuse has shown that they are the same as those that are important for brain stimulation reward (Koob and Le Moal, 2006). This research refined the theories put forth by Olds and Milner and identified specific neurochemical and neuroanatomical pathways that are important 12

27 for the occurrence of motivational responses to drugs of abuse, leading to the development of a variety of theoretical models describing drug motivation. 1.3 Models of Drug Motivation Over the course of decades of drug addiction research, a variety of neurobiological theories have been developed that attempt to identify and explain the neurocircuitry behind the motivation to consume drugs of abuse. The following are brief summaries of the major hypotheses that have been proposed since the discovery of electrical brain self-stimulation. These theories are both complementary and contradictory, without one single theory having the ability to fully explain drug motivation. The Mesolimbic Dopamine Reward Hypothesis One of the original theories of drug actions on the brain was the DA hypothesis, which suggested that all drugs of abuse acted on dopaminergic neurotransmission in the brain reward system (Wise, 1980). This hypothesis was formulated based on the considerable body of evidence suggesting that the brain reward system, consisting of the fast-conducting myelinated fibers of the medial forebrain bundle and the VTA dopaminergic neurons projecting to the nucleus accumbens (NAc) of the ventral striatum, served as a final common pathway for the transmission of rewarding motivational information (Wise, 1980). This theory originally focused on the ventral tegmental-striatal projection, but later came to focus more broadly on the mesolimbic DA system and its various inputs (Wise, 2002). Since the development of this theory, a variety of experiments using a wide range of techniques have reported that activity of the mesolimbic DA system is necessary for the motivational response to both natural rewards and drugs of abuse. Compelling evidence for this theory comes from studies showing that abused drugs act directly on DA synapses (amphetamine and cocaine) or cell bodies (opiates) in the mesolimbic DA system (Wise, 1980). Furthermore, studies that utilized pharmacologic manipulation of DA 13

28 signaling through DA receptor (DAR) antagonist drugs reported that blockade of DA signaling strongly reduces or completely attenuates the reinforcing properties of natural reinforcers, such as food reward (Wise et al., 1978), as well as the rewarding effects of various drugs of abuse, such as morphine, nicotine, cocaine, amphetamine and ethanol (Acquas et al., 1999; Corrigall and Coen, 1991; Price and Middaugh, 2004; Yokel and Wise, 1975). Additional evidence for this theory came from studies showing that destruction of the mesolimbic DA system using 6- hydroxydopamine lesions induces an anhedonic motivational state, wherein a general loss of interest in both natural and other rewarding stimuli occurs (Wise, 1982). This suggested that inactivation of the DA system rendered the subject incapable of feeling pleasure and without motivation to seek reinforcing events or stimuli. Taken together, these studies implied that DA function and the activation of the DA system is both necessary and sufficient for the mediation of reward-related motivational signals, leading to the formulation of the mesolimbic DA hypothesis. The main idea behind this DA hypothesis was that the common feature of all drugs of abuse is their ability to activate the mesolimbic DA system to produce reward. It was thus postulated that a single neural system is the final common pathway for the transmission of all rewarding motivational information. This theory is elegantly simple, but therein lays one of the criticisms of this hypothesis: Drugs of abuse can, and do, act through other non-da neurobiological substrates. For example, studies in previously drug naive animals given acute nicotine or cocaine demonstrated that the rewarding effects of these drugs are DA-independent (Lanca et al., 2000; Laviolette et al., 2002; Mackey and van der Kooy, 1985). Furthermore, the rewarding response to acute morphine is present in both wild-type (WT) and DAR subtype-2 (D2R) knockout (KO) mice (Dockstader et al., 2001). Rewarding responses after chronic drug use can also be DA-independent, as the rewarding effects of chronic ethanol are not blocked by DAR antagonism (Ting-A-Kee et al., 2009), and DA is not even necessary for the pursuit of natural reinforcers such as food reward (Bechara and van der Kooy, 1992). Finally, mice that are DA-deficient will still develop a conditioned place preference for cocaine (Hnasko et al., 2007). Furthermore, DA has been shown in some situations to not signal reward at all, but rather that changes in DA activity signal arousing or novel sensory events, or even aversive stimuli 14

29 (Schultz et al., 1992; Shultz, 2001). This predictive quality of DA was shown to be independent of hedonic processing in monkeys that were conditioned over a very large number of trials to receive sweetened juice after a neutral stimulus: Although DA neurons initially responded to the juice reward, after conditioning the DA neurons no longer responded to the reward, but would then respond to the presence (or absence) of the neutral stimulus alone (Schultz et al., 1992). Further, some neurons showed activation following aversive, non-noxious stimuli (Schultz, 2001). Other studies have also shown that aversive footshocks increase DA release (Joseph et al., 2003) and populations of DA neurons have been found in the VTA that are strongly excited by footshocks (Brischoux et al., 2009). These results stand in direct contrast to the mesolimbic DA hypothesis, as activation of the DA system did not occur during the experience of reward, but rather during aversive experiences. Perhaps the most compelling argument against the hypothesis that the DA system is the exclusive mediator of reward signals, as well as the most relevant to this thesis, is that the mesolimbic DA system can also be activated during aversive events. For example, this hypothesis would assume that nicotine, a widely abused drug, would produce reward through a DA-mediated system. However, acute nicotine administration in nondependent subjects can produce an aversive motivational response in a place conditioning paradigm that is DAmediated, being blocked by DAR antagonists (Laviolette and van der Kooy, 2004; Tan et al., 2009; also see chapters 2 and 3) or lesions of dopaminergic afferents to the NAc (Sellings et al., 2008). DA is also released in the NAc after aversive footshocks in rats (Young et al., 1993), and extracellular DA concentrations in the forebrain are increased after footshock or anxiogenic drug administration in rats (Dazzi et al., 2001). Human imaging studies have also shown increased DA signaling in the ventral tegmentum during aversive thermal stimulation (Becerra et al., 2001). Taken together, these results suggest that increased DA does not necessarily signal reward and largely disprove the hypothesis that the mesolimbic DA pathway represents a common and sufficient system to explain reward. Another problem with the DA hypothesis that drug-taking and -seeking is performed due to its DA-increasing euphoric effects is the evidence that tolerance develops after repeated drug administration. When tolerance develops, the pleasurable effects of drugs are decreased and may 15

30 be completely absent, but the subject continues to seek out and take drugs (Koob and Le Moal, 2006). Although the DA hypothesis remains as a prominent and influential theory, it is clear that the role of DA and the mesolimbic system in motivation is far more complex than a simple reward signal. Indeed, the studies reported in this thesis, as well as other data, suggest that specific patterns of DA signaling actually mediate aversive motivational responses. Consequently, other theories that have proven more relevant and encompassing in the explanation of drug motivation have persisted. The DA reward prediction error model Both complementary and in contrast to the idea that DA signals the receipt of a reward or reinforcing motivational stimuli, Schultz and colleagues have proposed the DA reward prediction error model, whereby the slower activity of DA neurons is hypothesized to code the uncertainty associated with rewards (Schultz, 2001; Schultz, 2007; Schultz et al., 2000; Schultz et al., 2002). A series of experiments performed by Schultz and colleagues suggested that mesolimbic DA projections from the midbrain to the striatum and frontal cortex show increases in activity following primary food and liquid rewards (consistent with the DA hypothesis) as well as after the presentation of conditioned, reward-predicting stimuli or novel stimuli (Schultz, 2001). They also show depression by the attention-generating omission of reward or during aversive events (Schultz et al., 2003; Schultz, 2007). This data led to the hypothesis that DA neurons are not in general activated by salient stimuli, but are reporting rewards as far as they occur differently than predicted, producing a prediction error message that serves as a powerful teaching signal for behaviour and learning (Schultz, 2001). This short-acting, subsecond DA message was hypothesized to be different from the more long-term (ie. minutes) DA function in behavioural responses and the much more long-term DA function (ie. hours to days) that is deficient in Parkinson s disease, suggesting that DA neurons serve different functions at different time scales. The prediction error hypothesis suggests that all responses to rewards, aversive events, and reward-predicting stimuli depend on the predictability of an event. After extensive training, 16

31 Schultz s monkeys were fully trained and the DA neurons no longer responded to receipt of a liquid reward because it was fully predicted. However, the DA neurons were activated or depressed if the predicted reward occurred sooner or failed to occur, respectively, at its habitual presentation time (Schultz et al., 1993). These changes in DA activity reflect an expectation process that is based on an internal clock that measures the precise timing of a predicted reward. The DA response is hypothesized to be equal to the reward occurrence minus the reward prediction (Schultz, 2001). This theory posits that the response to a reward does not occur unconditionally, but rather codes the prediction error such that an unpredicted reward elicits activation, a fully predicted reward elicits no response, and the omission of a predicted reward (an aversive event) induces a depression (Schultz, 2007). In this sense, DA neurons do not discriminate between or indicate different rewards, rather emitting an alerting message about the surprising presence or absence of rewards. The prediction error process continues until the behavioural outcome matches the prediction and the prediction error becomes nil. Since the proposal of this theory, reward-responsive signals have been demonstrated in the tegmental pedunculopontine nucleus (TPP) (Okada et al., 2009), the lateral habenula (Bromberg-Martin et al., 2010) and the frontal cortex (Roesch and Olson, 2004). DA prediction error signals have also been reported that respond to aversive stimuli in the rostromedial tegmental nucleus (Jhou et al., 2009) and the lateral habenula (Matsumoto and Hikosaka, 2009). Furthermore, DA firing has been suggested to code not only for the predicted timing of a reward, but also for the size of a prospective reward (Bromberg-Martin and Hikosaka, 2009), leading to the hypothesis that midbrain DA neurons are involved in information processing as well as reward and aversive signaling. However, in addition to having the same major criticism of the DA hypothesis, that reward has been demonstrated in the absence of DA (see for example Bechara et al., 1992; Hnasko et al., 2007; Vargas-Perez et al., 2009), the prediction error model has its own set of drawbacks. Some have suggested that DA firing is more likely to play a central role in identifying which aspects of context and behavioural output are crucial in causing unpredicted events rather than precisely signaling the error in timing of the predicted event (Redgrave and Gurney, 2006). In support of this idea, the results mentioned above showing that DA reward- and 17

32 aversion-responsive signaling occurs in a variety of areas other than the midbrain (Bromberg- Martin et al., 2010; Jhou et al., 2009; Matsumoto and Hikosaka, 2009; Okada et al., 2009; Roesch and Olsen, 2004) and demonstrate that DA neurons use a variety of activity patterns to signal different properties of rewarding and aversive stimuli, suggesting that DA activity may indeed be signaling more than a reward prediction error. Another major drawback of this theory is that it does not address drug dependence and withdrawal, or make any suggestions or hypotheses about DA neurons signaling these important areas of the addictive process. Thus, although a variety of experiments have supported the hypothesis that DA neurons signal a prediction error and research continues in this area, other groups have pursued alternative avenues of research in an attempt to explain the complexities of drug addiction and motivation. The incentive-sensitization theory This hypothesis speaks against the fundamental importance of both pleasure (the DA hypothesis) and withdrawal (the opponent process theory, described below) in the establishment of addiction to drugs of abuse, rather suggesting that an addict s neural system wrongfully attributes salience to drugs and drug cues, leading to pathological wanting of the abused drug. There is a conceptual and neurobiological distinction made between the hedonic aspects of drugs of abuse, or liking, and the motivational factors mediating their use, or wanting (Robinson and Berridge, 2003). This distinction was hypothesized based on a variety of experiments examining facial expressions and movements in a taste reactivity paradigm designed to measure hedonic impact, where the researchers assessed rhythmic mouth movements, tongue movements and protrusions, and gapes and various accompanying facial movements in a frame-by-frame camera analysis. It was observed that lesions of the mesolimbic DA system had no effect on the hedonic impact or liking of taste stimuli, leading to the idea that the DA system has no role in the liking of a drug (Robinson and Berridge, 1993). However, the DA system remains important for the incentive salience, or perceived value, which leads to wanting a drug of abuse after repeated exposure (Robinson and Berridge, 2003). The postulates of the incentive-sensitization theory 18

33 suggest that repeated drug exposure increases the incentive salience of the drug and leads to a hypersensitivity, or sensitization, of the DA system, which results in compulsive motivation, or pathological wanting, of the drug. However, the implication is not that DA neurons themselves mediate incentive salience, rather that the attribution of incentive salience coincides with the activation of DA neurons, and that incentive salience is the reward component most directly altered by manipulations of DA systems (Berridge and Robinson, 1998). DA manipulations can reveal dissociations between liking and wanting of drug rewards, but do not reveal the full nature of the psychological process or its neurobiological substrates, thus the involvement of other systems in the rewarding effects of drugs of abuse are not ruled out by this theory (Berridge and Robinson, 1998). The incentive-sensitization theory posits that reward is a multiplex process, comprising hedonic activation (liking), associative learning of the relationship between neutral events and their hedonic consequences, and subsequent attribution of incentive salience to those events (Berridge and Robinson, 1998). DA is not needed for the hedonic or the associative prediction components, but is required for the incentive salience component of reward. Dysregulation of the DA system responsible for wanting increases the motivation to seek the stimulus, such as a drug of abuse, and causes an increased pursuit of it, leading to drug-seeking behaviour and later, the negative consequences of addiction. This theory provides a more complete picture of drug addiction than its preceding neurobiological theories of motivation, however there are still many valid criticisms of its suggestions. First, in common with the DA hypothesis and prediction error hypothesis, there is no explanation of how drug motivation and wanting occur in the absence of DA (see for example Bechara et al., 2002; Hnasko et al., 2007; Laviolette and van der Kooy, 2003; Vargas- Perez et al., 2009). The evidence that motivated behaviour for abused drugs occurs in the absence of dopaminergic activity, therefore preventing the DA-mediated attribution of incentive salience, stands in contrast to this theory. Although, as mentioned above, the involvement of other systems in drug motivation is not completely ruled out by this theory, no direct evidence for any alternative systems in the attribution of incentive salience has been reported to date by this group or others. Furthermore, some DAR antagonism studies have reported an increase in 19

34 drug-seeking behaviour after DAR activity is blocked (Ettenberg et al., 1982) or even a switch in the motivational valence of a drug, whereby administration of a DAR antagonist switches a conditioned place aversion to a conditioned place preference (Laviolette and van der Kooy, 2003; Sturgess et al., 2010). Another criticism of this theory comes from the lack of evidence of a true distinction between the two components of liking and wanting. Indeed, infusions of µ-opioid-receptor, glutamate receptor, or gamma-aminobutyric acid (GABA) receptor drugs into various brain regions are all capable of modulating the hedonic reactions to taste stimuli, or the liking component, and increasing the wanting component as well (Pecina et al., 2006; Reynolds and Berridge, 2002; Reynolds and Berridge, 2003) implying that the two components are at least partially linked, being dependent on the same neurobiological substrates. This lack of a double dissociation whereby a manipulation affects only one process and not the other indicates that further investigation of this hypothesis and its postulates is required. The non-deprived/deprived hypothesis The main idea behind this model, proposed by van der Kooy and colleagues, is that the rewarding properties of both natural and drug rewards are mediated by either a dopaminergic or a non-dopaminergic motivational system depending on the deprivation state of the animal (Bechara et al., 1992). In contrast to the theories discussed above, this hypothesis proposes that the mechanisms underlying both natural and drug rewards are not rigid, but rather are transient, and are changing depending on the current motivational state of the organism (Bechara et al., 1992). This theory is not a specific alternative to the hypotheses involving DA, but rather imposes a constraint on when DA mediates reward. In a satiated or non-deprived motivational state, reward is mediated through a non-dopaminergic system, involving the VTA GABA neurons projecting to the TPP nucleus of the brainstem (Bechara et al., 1998). In the nonsatiated, drug withdrawn or deprived motivational state, the TPP no longer is thought to mediate the rewarding effects of food or drugs, but rather the mesolimbic DA system is responsible for 20

35 motivational drive (Bechara et al., 1998). Therefore, DA can and does mediate reward, but only when the subject is in a deprived motivational state. The two separate systems postulated to mediate reward in the brain are mutually exclusive: they contribute similarly to behaviour, but are operative at different times, depending on the deprivation state. This hypothesis was developed based on a series of experiments that showed that administration of the broad-spectrum DAR antagonist α-flupenthixol (α-flu), but not lesions of the TPP, diminished food and opiate reward when rats were hungry or opiatewithdrawn, respectively, but not when they were sated or opiate-naive (Bechara et al., 1992; Bechara et al., 1995; Nader et al., 1997). Conversely, excitotoxic lesions of the TPP but not α-flu administration disrupted food and morphine reward only if rats were tested while sated or drugnaive, respectively, but not if they were hungry or in withdrawal from chronic opiate exposure (Bechara and van der Kooy, 1989; Bechara et al., 1992; Olmstead et al., 1998). More recent studies support this hypothesis, finding that brain-derived neurotrophic factor in the VTA promotes a shift from a DA-independent to a DA-dependent opiate reward system, involving a switch from inhibitory to excitatory signaling in the GABA receptors on VTA neurons (Vargas- Perez et al., 2009). Unlike the incentive salience model, this work demonstrated a double dissociation between the motivational state of the animal (non-deprived or deprived) and the mechanism responsible for mediating reward (TPP- or DA-dependent). The switch from a non-deprived, TPP-dependent, to a deprived, DA-dependent motivational state is not permanent. One study showed that opiate-dependent and withdrawn rats (whose conditioned place preferences are DA-dependent) given opiates to relieve their withdrawal are in a TPP-dependent, non-deprived motivational state, having their conditioned place preferences blocked by TPP lesions but not DAR antagonism (Bechara and van der Kooy, 1992). These results suggest that the presence or absence of withdrawal, or being in a deprived or non-deprived motivational state, respectively, determined which neurobiological substrate mediated reward. Similarly, if deprived animals are given enough time to recover fully from the effects of withdrawal, the mechanisms underlying opiate reward are again TPP-dependent (Nader et al., 1994). These results are in contrast to the incentive-sensitization theory in that the 21

36 non-deprived/deprived theory does not require that a permanent switch in brain neurochemistry occur (Robinson and Berridge, 1993). Criticisms of this hypothesis come from studies showing that the DA system is activated (DA overflow is observed in the NAc) by sex, food and drug reward in non-deprived or drugnaive rats (Di Chiara and Imperato, 1998; Fiorino et al, 1997; Martel and Fantino, 1996). This hypothesis cannot account for cocaine or amphetamine reward, and cannot completely account for nicotine motivation, as nicotine reward is DA-mediated in the deprived motivational state and many studies have shown that nicotine reward in nondeprived subjects is also DA-mediated (Acquas et al., 1989; Lecca et al., 2006; Merlo Pich et al., 1999; Pak et al., 2006; Sellings et al., 2008; Spina et al., 2006; Tanabe et al., 2008). However, the van der Kooy group and others have demonstrated DA-independent (TPP-dependent) nicotine reward (Corrigall et al., 2001; Lanca et al., 2000; Laviolette et al., 2003), suggesting that nicotine and opiates act in a similar way in terms of TPP- and DA-mediation. Much debate remains about the precise neurobiological substrates mediating nicotine reward in the non-deprived motivational state, therefore this thesis will focus on another theory of motivation, the opponent process theory, which attempts to explain the dual motivational effects of acute and chronic nicotine in both the non-deprived and deprived motivational states. The opponent process theory The opponent process theory of motivation was first described by Solomon and Corbit (1974) and later expanded and refined by Koob and colleagues. It is a two-sided hedonic hypothesis that has gone by many different names, such as positive-negative reinforcement, opponent processes, hedonic dysregulation, and reward allostasis (Koob and Le Moal, 2006; Solomon and Corbit, 1973). This theory all but abandons the idea that DA and the pleasant feelings of drug administration drive compulsive drug use, rather focusing on the unpleasantness of withdrawal as the driving force behind continued drug-taking and relapse to compulsive drug seeking. It was originally hypothesized that many hedonic, affective, or emotional states, both pleasant and aversive, are automatically opposed by neurobiological mechanisms that reduce the 22

37 intensity of the state (Solomon and Corbit, 1974). Koob and colleagues elaborated on this theory, suggesting that drugs are taken at first because they are pleasant, but with repeated use certain neuroadaptations lead to tolerance and dependence, such that drug taking is no longer pleasant and in fact, unpleasant withdrawal symptoms ensue and eventually dominate upon cessation of drug use (Koob and Le Moal, 2006). Compulsive drug taking is thus maintained simply in order to escape the negative experience of withdrawal that occurs upon cessation of chronic drug use. The opponent process theory posits that initial pleasant stimuli activate a dose-dependent, relatively short-acting a-process, which in turn triggers the activation of a longer-lasting opponent b-process. The b-process is thought to restore homeostasis in the brain, bringing the activity states back to normal, being strengthened by use and weakened by disuse (Solomon and Corbit, 1974). An initially aversive stimulus similarly activates a dose-dependent, aversive a- process, that triggers and is followed by a longer lasting, slower to decay, rewarding b-process (Figure 1.2). With repeated activation, the opponent b-process is strengthened, growing in magnitude and duration, and causes a tolerance effect to the a-process. In the example of an initially rewarding drug of abuse, the a-process is not as pleasant during subsequent drug exposures because of the adaptation of the b-process. The a-process is not affected by use, being a relatively stable, unconditioned reaction, but with repeated use, the b-process shows a shorter latency in response to the a-process, a quicker rise, and a longer decay time (Solomon and Corbit, 1974). Furthermore, unpleasant effects of withdrawal are caused when the rewarding effects of a drug wear off, as the b-process is slow to decay and opposite in direction to the a- process, thus the aversive effects of the b-process remain after the pleasantness of the a-process has worn off. Similar to the DA hypothesis, the opponent process theory postulates that the a- process is caused by mesolimbic DA activity (Koob and Le Moal, 2008). The b-process also involves DA neurotransmission and a loss of function of the brain reward system (a withinsystem neuroadaptation), but it has been hypothesized that the key player in the anxiogenic and aversive effects of withdrawal experienced during the b-process to drugs of abuse are mediated by recruitment of the corticotropin-releasing factor (CRF) brain stress anti-reward system (a between-system neuroadaptation) in the amygdala and other brain areas (Koob and Le Moal, 1997; Koob and Le Moal, 2006; Koob and Le Moal, 2008). This combination of decreased 23

38 Figure 1.2. The opponent process theory of motivation. Solomon and Corbit (1974) postulated that any stimulus would trigger an initial a-process that will closely follow the stimulus and will be fast to occur and fast to end. The initial a-process can be rewarding or aversive and will be followed by a later occurring opponent b-process that is longer lasting, slower to end and is opposite in direction to the a-process. For drugs of abuse, a rewarding dose will produce an initial rewarding a-process followed by an aversive opponent b- process (top), while an aversive dose will produce an initial aversive a-process followed by a rewarding opponent b-process (bottom). Appetitive Stimulus (Reward) a process b process Stimulus Response Aversive Stimulus b process a process Grieder et al.,

39 reward function and recruitment of an anti-reward system is hypothesized to lead to an allostatic state, or chronic deviation from the normal homeostatic state, where the a-process is less rewarding and the b-process is much more intense (Koob and Le Moal, 2001). This process provides a strong source of negative reinforcement, whereby the drug is taken to relieve the negative effects of the b-process. This progression contributes to and is hypothesized to drive relapse and addiction. Koob and colleagues hypothesized that an individual who does not frequently use a drug, allowing sufficient time between re-administering the drug, will not experience allostasis and will retain the a-process, experiencing a positive hedonic motivational state after the cessation of drug use (Koob and Le Moal, 2001). In other words, an appropriate counteradaptive opponent b- process that balances the a-process does not lead to an allostatic state. However, the changes in an individual with repeated frequent drug use represent a transition to an allostatic state in the brain reward systems and therefore a transition to drug addiction and withdrawal (Figure 1.3). In an allostatic state, the b-process never returns to the original homeostatic level before drugtaking begins again, thus creating a greater and greater allostatic state in the brain reward system (Koob and Le Moal, 2001; Koob and Le Moal, 2006). In the allostatic motivational state, the counteradaptive opponent b-process no longer balances the a-process. The opponent process theory of motivation is very sound in that it incorporates both pleasure and withdrawal in its descriptions of opponent motivational a- and b-processes. It can also account for both rewarding and aversive stimuli, leading to rewarding and aversive a- processes, respectively. The opponent process theory is unique in that the original theory did not view addiction as an abnormality, but rather as an inevitable consequence of a normally functioning system that opposes affective or hedonic states (Solomon and Corbit, 1974). However, criticisms of this theory come from data showing that previously addicted subjects often experience intense cravings and subsequent relapse to drug-taking months or years after complete abstinence, and long after the negative effects of withdrawal have subsided, the b- process having decayed (Lu et al., 2004). It appears from these studies that elimination of withdrawal symptoms does not protect against relapse. Furthermore, withdrawal is not as 25

40 Figure 1.3. The allostatic state of drug addiction. The changes in an individual after repeated drug use lead to an allostatic state, where the normal homeostatic state is chronically deviated from, and a new homeostatic point is reached. At top, the initial experience of a drug with no prior drug history is depicted, modeling the original opponent process theory of motivation put forth by Solomon and Corbit (1974). The a-process is positive and the opponent b-process is negative. The motivational state is the sum of the a- process and b-process. An individual whom experiences a positive hedonic mood state from a drug of abuse with sufficient time between re-administering the drug is hypothesized to retain the a-process and does not experience an allostatic state. At bottom, the changes in the affective state in an individual with repeated frequent drug use leading to an allostatic motivational state. In an allostatic state, the b-process does not balance the a-process because the system never returns to the original homeostatic level before drug-taking begins again. Koob and Le Moal,

41 powerful a motivator for relapse to drug-taking as was originally implied in this theory, as studies have shown that activating the a-process is far more effective at reinstating drug-seeking and -taking in recovered drug addicted rats than activating the b-process (Stewart and Wise, 1992). Additional evidence against this theory, which predicts that withdrawal motivates drugseeking and -taking, comes from studies in opiate dependent animals. In opiate withdrawn rats, no predictive relationship between the demonstration of a somatic withdrawal syndrome and the aversiveness of withdrawal measured by conditioned place aversion was observed (Mucha, 1987). Furthermore, opiates will be self-administered in the absence of withdrawal symptoms (Ternes et al., 1985). These studies suggest that withdrawal is actually not as strong a predictor of drug-seeking than postulated in this theory. As with the other theories of motivation, the opponent process theory cannot fully explain the process of drug addiction and the motivation to seek drugs of abuse. However, it is the most relevant theory to this thesis, and will be explained further in relation to acute nicotine intake and nicotine withdrawal in Chapter 2 and in the discussion. 1.4 The Anatomy of the Ventral Tegmental Area The VTA is home to one of the major populations of DA cells in the brain (Kalivas, 1993), thus it is not surprising that this area has received much attention in the study of drug motivation. The VTA lies close to the midline on the floor of the midbrain, bordered laterally by the substantia nigra, rostrally by the mammillary bodies and posterior hypothalamus, and caudally by the pons and hindbrain (Oades and Halladay, 1987). 27

42 Neurons and Projections The VTA is mainly composed of DA and GABA neurons (Kalivas, 1993). Although the exact proportions of these two major cell types have not been precisely determined, it has been suggested using a variety of methods that the VTA is approximately 55-60% DA and 5-33% GABA cell types (Kalivas, 1993; Margolis et al., 2006). There are numerous methods that may be utilized to identify DA and GABA cells, including histology, pharmacology, and electrophysiology. DA cells are very large in size, stain positively for tyrosine hydroxylase (TH, the rate limiting enzyme responsible for DA synthesis), and when measured electrophysiologically will demonstrate well established features: (I) a relatively longer action potential width (>2.5 ms); (II) a triphasic (+/-/+) waveform consisting of a notch on the rising phase followed by a delayed after-potential; (III) a characteristic low tone by audio monitoring; (IV) a slow, irregular or bursting firing pattern, and (V) a spontaneous firing rate of 2-5 Hz or less (Grace and Bunney, 1983; Tan et al., 2009;). GABA cells are smaller in size, stain positively for GABA and glutamic acid decarboxylase (GAD, an enzyme responsible for GABA synthesis), and when measured electrophysiologically will demonstrate spontaneous activity, and higher firing frequencies and shorter action potentials than DA neurons (Cameron et al., 1997; Korotkova et al., 2004). There are also glutamate neurons present in the VTA, comprising approximately 1-15% of VTA cells, that are suggested to provide local excitatory modulation of the DA and GABA neurons (Dobi et al., 2010; Yamaguchi et al., 2007). These neurons can be identified through the detection of mrna encoding vesicular glutamate transporters, which transport glutamate into synaptic vesicles at presynaptic terminals, and are known to be independent from DA and GABA cells because they do not co-stain for TH or GAD, respectively (Dobi et al., 2010; Yamaguchi et al., 2007). Dopaminergic, GABAergic and glutamatergic receptors are located throughout the VTA. Axon terminals derived from both GABAergic and glutamatergic neurons establish local intrinsic synapses as well as extrinsic inputs on dendrites of both dopaminergic and nondopaminergic neurons in the VTA (Morales and Pickel, 2012). The principal excitatory extrinsic inputs to the VTA are glutamatergic projections from prefrontal cortex, bed nucleus of the stria terminalis, amygdala, and the pontomesencephalic tegmental nuclei (TPP and lateral dorsal 28

43 tegmental nucleus) (Mao and McGehee, 2010). The pontomesencephalic tegmental projections are also cholinergic and GABAergic, and have been shown to be important for the rewarding effects of nicotine (Corrigall et al., 2002; Lanca et al., 2000) probably because of their contribution to the phasic burst firing of VTA DA neurons (Floresco et al., 2003; Lodge and Grace, 2005). The principal inhibitory inputs to the VTA are GABAergic and include local interneurons (mentioned above) and projections from the NAc, ventral pallidum, and pontomesencephalic tegmental nuclei (Mao and McGehee, 2010). However, because of the presence of inhibitory GABAergic interneurons, excitatory inputs from extrinsic sites can also cause inhibition of VTA DA firing, as in the case of lateral habenular suppression of DA neuronal activity (Matsumoto and Hikosaka, 2007). The VTA receives CRF inputs from the extended amygdala and paraventricular nucleus of the hypothalamus (PVN), from which CRF is co-released with GABA and glutamate (Tagliaferro and Morales, 2008). VTA dopaminergic and nondopaminergic neurons express both CRF1Rs and CRF2Rs (George et al., 2012; Ungless et al., 2003). The dopaminergic projections from VTA to NAc are very well established, however, some studies have demonstrated that in addition to VTA dopaminergic innervations, the NAc receives inputs from both GABAergic (Van Bockstaele et al., 1995) and glutamatergic VTA neurons (Yamaguchi et al., 2011). Mesocortical DA neurons also project to the prefrontal cortex, however, retrograde tracing studies indicate that approximately half of the mesocortical projection neurons from the VTA are nondopaminergic (Morales and Pickel, 2012). The VTA also sends dopaminergic projections to the extended amygdala and PVN (Eliava et al., 2003). A summary figure of the VTA neurons, projections, and inputs is shown in Figure 1.4. Nicotinic Receptors Ultimately, nicotine influences neuronal activity in the VTA and causes its motivational effects by binding to nicotinic acetylcholine receptors (nachrs), which are pentameric, ligandgated ion channels (Koob, 2001; Mansvelder and McGehee, 2002). There are 12 neuronal 29

44 Figure 1.4. The VTA: Neurons, receptors, inputs and projections. The VTA (blue dotted box) is composed of DA, GABA, and glutamate (Glu) neurons. Both inhibitory GABAergic and excitatory glutamatergic neurons establish local intrinsic synapses (white arrows) on both dopaminergic and non-dopaminergic neurons. Extrinsic inputs to the VTA (yellow arrows) come from the prefrontal cortex (PFC), amygdala, TPP, NAc, ventral pallidum (VP), habenula, and other regions not shown in this figure. The VTA neurons send reciprocal projections (yellow arrows) back to the areas pictured here, as well as other areas not shown in this figure. VTA dopaminergic and nondopaminergic neurons express CRF receptors (both CRF1Rs and CRF2Rs) and a variety of nachrs, as well as DA and GABA receptors, which are expressed throughout and not labeled on this diagram. 30

45 nachr subunits identified to date, the α2-α10 and β2-β4, of which only α8-α10 have not been reported as being expressed in the VTA. The nachr receptor profiles that are associated with DA and GABA neurons differ considerably, with dopaminergic VTA neurons mainly expressing α7, α4β2* and α6β2* nachrs, with the asterisk denoting the possibility of other subunits, such as α5 and β3, being incorporated into these receptors (Mao and McGehee, 2010). On nondopaminergic neurons, less than 25% have been reported to express the α3, α5, α6, and β4 subunits, thus α4β2 combinations predominate, with some α7 homomeric nachrs being found as well (Laviolette and van der Kooy, 2004; Mao and McGehee, 2010). nachr activation leads to increased cation flow through the central channel, which induces depolarization and increased excitability in the neuron (Mao and McGehee, 2010). Nicotine produces its motivational effects by first acting on its receptors, and modulating the activity of VTA neurons. 1.5 The Neurobiology of Nicotine Motivation: DA and CRF The Use and Abuse of Nicotine Nicotine is a highly toxic alkaloid, one of over 4000 chemicals that may be obtained from smoking dried tobacco leaves of the cultivated plant Nicotiana tabacum (Koob and Le Moal, 2006). Nicotine is the major reinforcing component of tobacco smoke that leads to dependence in humans (Stolerman and Jarvis, 1995). Its use by indigenous peoples of the Americas for both medicinal and ceremonial purposes has been traced back 8000 years, but the first documented practice of smoking the dried leaves of the tobacco plant has been attributed to European explorers in 1492 (Koob and Le Moal, 2006). The first cigarette-making machine was produced in 1880, and since then, cigarette production boomed due to the ease of production and distribution, the ability of the mass media to market their product, and increased demand (Akehurst, 1968). Today, tobacco smoking and the addiction to nicotine that comes from it, is a worldwide health problem. Tobacco addiction is the leading avoidable cause of disease and premature death 31

46 in North America (Fellows et al., 2002). The cost to society associated with nicotine addiction leading to health care problems that often result in death, medical costs, and human suffering is significant (CDC, 2008). The most common reason for relapse reported by quitting smokers is the desire to relieve the discomforts that come with nicotine abstinence and withdrawal (Allen et al., 2008). Similar to nicotine-withdrawn humans, nicotine-dependent experimental animals that undergo nicotine withdrawal demonstrate an observable somatic nicotine abstinence syndrome (Epping-Jordan et al., 1998; Malin et al., 1992; Stoker et al., 2008). Most animal studies involve antagonist-precipitated nicotine withdrawal (George et al., 2007; Kenny and Markou, 2001; Laviolette et al., 2008; Watkins et al., 2000), although a spontaneous withdrawal procedure would more closely model the human response to withdrawal from chronic nicotine. Rodents experiencing spontaneous withdrawal from chronic nicotine will show a conditioned aversive response to a withdrawal-paired environment in place conditioning paradigms (Merritt et al., 2008; also see Chapter 2). This withdrawal response is modeled by the opponent process theory of motivation and represents the opponent b-process to the a-process of chronic nicotine reward in dependent animals. The neurobiological substrates mediating these opponent motivational processes in nicotine-dependent animals are essentially unknown. In nondependent animals, acute nicotine administered directly into the VTA will produce both rewarding and aversive effects (Laviolette et al., 2002; Sellings et al., 2008), thus two different a-processes, one rewarding and another aversive, would be stimulated after acute nicotine administration. These acute nicotine a-processes are mediated by different neural substrates, with reward being TPPmediated and aversion being DA-mediated (Laviolette et al., 2002). In many studies over many years of research on the neurobiology of nicotine motivation and withdrawal, focus has been on a variety of neurobiological substrates, including VTA DA, the TPP, CRF, epinephrine, serotonin, brain-derived neurotrophic factor, GABA, and others. However, this thesis will focus on the DA and CRF systems involvement in nicotine s acute and chronic motivational effects. 32

47 VTA Dopamine Like most drugs of abuse, nicotine acutely produces both aversive and positive motivational effects (Grunberg, 1994; Laviolette and van der Kooy, 2004; Perkins et al., 2008) by increasing the extracellular concentration of DA in the mesolimbic system (Di Chiara and Bassareo, 2007; Grace, 2000; Nestler, 2005; Picciotto and Corrigall, 2002) as well as non- DAergic neural substrates (Fowler et al., 2011; Lanca et al., 2000; Laviolette et al., 2002; Levin et al., 1996; Picciotto and Corrigall, 2002). Nicotine given peripherally has centrally mediated effects on DA release (Seppa et al., 2000) and selectively activates DA neurons in the pvta, but not avta (Zhao-Shea et al., 2011). DA signaling has been implicated in the aversive motivational response to acute nicotine (Laviolette and van der Kooy, 2003; Tan et al., 2009); however, DA-dependent acute nicotine reward has also been demonstrated (Acquas et al., 1989; Lecca et al., 2006; Merlo Pich et al., 1999; Pak et al., 2006; Sellings et al., 2008; Spina et al., 2006; Tanabe et al., 2008). DA acts at five different receptor subtypes, D1-D5, of which only D1Rs and D2Rs are found on VTA neurons (Le Foll et al., 2009). DA neurons in the VTA exhibit burst firing that produces a fast and large DA release that mainly activates postsynaptic D1Rs, as well as population firing that produces a slower tonic DA release that mainly activates higher affinity, mostly presynaptic D2Rs (Floresco et al., 2003; Goto and Grace, 2005; Grace, 2000). A single systemic injection of acute nicotine in nondependent animals increases the firing rate and phasic burst activity of VTA DA neurons, elevating DA in the projection sites of VTA DA neurons (Mameli-Engvall et al., 2006; Zhang et al., 2009). Consistent with the idea that phasic activation leads to D1R activation, D1R antagonism blocks acute nicotine motivation in nondependent mice (David et al., 2006). Conversely, chronic exposure to nicotine decreases tonic but not phasic DA activity in the VTA (Tan et al., 2009) and spontaneous and mecamylamineprecipitated nicotine withdrawal in dependent animals is associated with a decrease in extracellular DA levels in the NAc (Carboni et al., 2000; Rahman et al., 2004.) These effects reflect changes in both DA release and reuptake (Duchemin et al., 2009) that lead to a decrease in DA signaling in the brain (Kalivas and Volkow, 2005), results that are similar to those obtained in drug-dependent human subjects, who show marked decreases in D2R availability 33

48 (Fehr et al., 2008) and thus presumably in DA release (Volkow et al., 2009). These results and many others demonstrate that VTA DA signaling through D1Rs and D2Rs is involved in both the rewarding and aversive motivational effects of acute and chronic nicotine, suggesting that different DA neurons in the same brain area may signal both reward and aversion, possibly due to the different patterns of DA activity, different timing of DA release, or different activation of DA receptors in VTA projection sites. In the striatum, the main VTA projection site, adenosine receptors subtype-2a (A2ARs) and D2Rs are colocalized (Tozzi et al., 2011) and form A2AR-D2R heteromers (Fuxe et al., 2010). The A2AR and D2R interact antagonistically, such that agonism of A2ARs decreases signaling at D2Rs (Tanganelli et al., 2004) and antagonism of A2ARs increases signaling at D2Rs (Fuxe et al., 2010). In terms of nicotine motivation, very few studies have examined the role of A2ARs, however one study found that genetic deletion of the A2AR would prevent the rewarding motivational effects of acute nicotine in nondependent mice without affecting the conditioned taste aversion for acute nicotine (Castañé et al., 2006). These results suggest that the A2AR plays a role in nicotine motivation, possibly through a DAR-mediated mechanism. Although there have been a variety of studies on DA activity, DA receptors, and less so A2ARs in nicotine motivation, the role of tonic and phasic DA activity and activation of the various DARs (and possibly of A2ARs because of their antagonistic activities) in signaling the motivational effects of both acute and chronic nicotine is essentially unknown. The studies described in Chapters 2 and 3 of this thesis will address this gap in the nicotine motivation research. Corticotropin-Releasing Factor (CRF) CRF is a 41-amino acid polypeptide isolated initially in 1981 from ovine hypothalamus (Vale et al., 1981) that controls hormonal, sympathetic and behavioural responses to stressors (Koob, 2008). CRF is present in the prefrontal cortex, extended amygdala, medial septum, hypothalamus, thalamus, cerebellum, locus coeruleus, and midbrain and hindbrain nuclei 34

49 (Swanson et al., 1983). The anatomical distribution of CRF indicates a role for the neuropeptide in responses to stress, food intake and cognition (Koob, 2008). There are two known CRF receptors, CRF1R and CRF2R, both of which can be found on VTA neurons (Sauvage and Steckler, 2001; Ungless et al., 2003). CRF release occurs during ethanol (Funk et al., 2006; Merlo-Pich et al., 1995) and opiate (Weiss et al., 2001) withdrawal in dependent rats, and CRF receptor antagonists can reduce ethanol (Rassnick et al., 1993), cocaine (Basso et al., 1999), and opiate (Stinus et al., 1995) withdrawal-induced anxiety-like behaviour. Fewer studies examining CRF and nicotine motivation have been performed, but it has been shown that CRF release is increased in the amygdala during nicotine withdrawal (George et al., 2007) and that CRF1R antagonists can block the anxiogenic effects of withdrawal from chronic nicotine (George et al., 2007; Tucci et al., 2003). The very few studies on the involvement of CRF in nicotine motivation have demonstrated that CRF plays a role in the anxiogenic effects of nicotine withdrawal, suggesting a possible role for CRF in the opponent motivational effects of chronic nicotine. Dopamine and CRF in drug motivation The mesolimbic DA system and the CRF brain stress system have been extensively studied independently and are usually considered to be mutually exclusive in terms of drug motivation. However, recent results have demonstrated that these two systems do indeed interact with each other (reviewed in George et al., 2012), suggesting that dysregulation of this newly discovered DA-CRF interaction may produce motivational effects that contribute to the development of drug addiction. There are reciprocal connections between CRF-producing neurons in the extended amygdala and PVN and the DA-producing neurons of the VTA. The VTA sends heavy dopaminergic projections to the extended amygdala and PVN (Liposits and Paull, 1989), directly innervating CRF-containing neurons in the central nucleus of the amygdala (CeA) (Eliava et al., 2003). In return, CRF neurons from the extended amygdala and PVN project to the VTA 35

50 (Rodaros et al., 2007). DARs are expressed throughout CRF-producing areas of the brain, and the VTA expresses both CRF1Rs and CRF2Rs (Sauvage and Steckler, 2001; Ungless et al., 2003). Neuroanatomical studies show that CRF is colocalized in glutamatergic and GABAergic afferents to the VTA, and that these afferents synapse with dopaminergic as well as nondopaminergic VTA neurons (Tagliaferro and Morales, 2008). Furthermore, corticotropin releasing factor-binding protein (CRF-BP), a protein that participates in the regulation and potentiation of CRF signaling at the synapse (Ungless et al., 2003), has been identified in a subset of DA and GABA neurons within the VTA (Wang and Morales, 2008). This CRF/DA mechanism in the VTA has been implicated in the process of relapse to cocaine seeking (Wang et al., 2007), suggesting that a CRF/DA mechanism operating in the VTA may be involved in drug addiction, withdrawal, and relapse. Consistent with anatomical data, CRF dose-dependently increased VTA DA neuronal firing (Hahn et al., 2009; Wanat et al., 2008) and locomotor activity (Kalivas et al., 1987), which was prevented by antagonism of CRF1Rs but not CRF2Rs, and was mimicked by CRF1R agonists (Wanat et al., 2008). CRF1R activation can also facilitate slow, D2R-mediated neurotransmission (Beckstead et al., 2009) as well as DA release (Bagosi et al., 2006; Muramatsu et al., 2006), while CRF1R knock down in the VTA reduces DA release in the prefrontal cortex (Refojo et al., 2011). CRF1R antagonism was also found to significantly increase DA neuron tonic population activity without affecting phasic burst firing, average firing rate, or NAc DA levels (Lodge and Grace, 2005). These results suggest that modulating CRF activity may have effects on DA neuronal firing activity and DA release in various brain areas. Although a variety of recent studies have demonstrated a definite interaction between the mesolimbic DA system and the CRF brain stress system, there are no known studies that have directly investigated nicotine addiction and motivation in regards to this DA-CRF interaction. Chapter 4 of this thesis will address this gap in the research. 36

51 1.6 Research Aims and Hypotheses Identification of the involvement of VTA DA and CRF activity in mediating the opponent motivational responses to acute and chronic nicotine In the following dissertation, I have investigated the role of the neurotransmitters DA and CRF in the VTA in mediating the initial aversive response and opponent rewarding response to acute nicotine, as well as the initial rewarding response to chronic nicotine and opponent aversive motivational response of withdrawal from chronic nicotine. I have explored in depth the changes in VTA phasic and tonic dopaminergic neuronal activity that occur during the administration of acute aversive nicotine versus chronic nicotine and withdrawal, and investigated which type of DA receptor activation is required for these aversive motivational responses to acute nicotine and chronic nicotine withdrawal. I have also examined the activation of a VTA CRF system during the transition from a nicotine nondependent to a dependent motivational state. Furthermore, I have studied the connection between the DA and CRF systems in the VTA of nicotine dependent and withdrawn animals. Previous studies showing that CRF- BP is present in VTA neurons (Wang and Morales, 2008) and that chronic experience with cocaine, another drug of abuse, enhances CRF-dependent potentiation of VTA DA neurons (Hahn et al., 2009) suggest that such a DA/CRF link might not be unexpected. My overall hypothesis is that nicotine dependence upregulates CRF and modifies DA activity in the VTA, and that both the increase in CRF and the specific pattern of DA activity in the VTA are necessary for the experience of the aversive motivational response to nicotine withdrawal. In the first series of experiments described in chapter 2, I tested whether the mesolimbic DA system is involved in the motivational response to nicotine withdrawal. Previous work suggests that the DA system is involved in the processing of the aversive motivational effects of opiate withdrawal (Bechara and van der Kooy, 1992). I thus hypothesized that blocking DA signaling at dopaminergic receptors would prevent the aversive response to nicotine withdrawal. I began by examining whether rodents would experience withdrawal after cessation of chronic nicotine administration. I assessed the effect of spontaneous withdrawal of chronic 37

52 nicotine administration on the expression of a somatic nicotine abstinence syndrome in rats and mice. Previous work suggests that antagonist-precipitated nicotine withdrawal induces an abstinence syndrome in rodents (Isola et al., 1999; Malin et al., 1992). I thus hypothesized that the spontaneous removal of chronic nicotine in a dependent animal would produce an observable and significant nicotine abstinence syndrome in both mice and rats. I next tested whether rodents would show a motivational response to nicotine withdrawal by examining the effect of spontaneous withdrawal from chronic nicotine in a place conditioning paradigm in rats and mice. Previous work suggests that withdrawal from chronic opiate administration in rats leads to conditioned place aversions to a withdrawal-paired environment (Bechara and van der Kooy, 1992). The opponent process theory of motivation suggests that this aversive response to withdrawal is the opponent process to chronic drug reward. I hypothesized that withdrawal from chronic nicotine would lead to conditioned place aversions in both rats and mice, and tested whether this motivational response would coincide with the demonstration of a nicotine abstinence syndrome. I also investigated whether acute aversive nicotine in nondependent mice would produce opponent motivational processes that could be measured in the place conditioning paradigm. The opponent process theory of motivation suggests that an initial aversive stimulus will lead to a later occurring and longer lasting rewarding stimulus (Solomon and Corbit, 1974), which implies that acute aversive nicotine would stimulate an initial aversive motivational response and a later occurring rewarding motivational response. The most important and novel finding in the set of experiments detailed in Chapter 2 tested the effect of administration of a DA antagonist, which disrupts dopaminergic signaling by blocking DA receptors, on the opponent motivational processes occurring after both acute and chronic nicotine. We showed using the place conditioning paradigm that acute aversive nicotine in nondependent mice produces an initial aversive response followed by a rewarding opponent motivational response, chronic nicotine in dependent mice produces a rewarding motivational response, and withdrawal from chronic nicotine in dependent mice will produce an opponent aversive motivational response. Previous results have demonstrated that DA antagonism blocks 38

53 the conditioned aversive response to acute nicotine in rats (Laviolette and van der Kooy, 2004) and that DA activity is involved in nicotine motivation in dependent animals (Bruijnzeel and Markou, 2005; Kenny and Markou, 2001; Laviolette et al., 2008; Smolka et al., 2004). I hypothesized that the mesolimbic DA system mediates the motivational response underlying the opponent process to chronic but not acute nicotine. Similarly, I also examined the effect of D2R KO on the opponent motivational effects of acute and chronic nicotine by utilizing D2R KO mice in the place conditioning paradigm. I hypothesized that if both genetic deletion of a DA receptor and antagonism of DA receptors could block the aversive motivational effects of nicotine withdrawal, a process that changes the activity of DA neurons in the VTA and the release of DA in the NAc (Hildebrand et al., 1998; Liu and Jin, 2004; Rada et al., 2001), then this suggests that the modification of DA signaling prevents a specific pattern of activity that may signal nicotine withdrawal. In chapter 3, I thoroughly investigated the specific pattern of DA signaling that mediates nicotine motivation by testing the effect of withdrawal from chronic nicotine on tonic and phasic VTA DA activity and whether the specific pattern of signaling through D1Rs and D2Rs mediates the conditioned motivational responses to nicotine withdrawal and acute nicotine. I began by examining whether increasing or decreasing dopaminergic signaling at receptors could prevent nicotine withdrawal aversions. Opiate withdrawal aversions can be blocked with DA agonist or antagonist pretreatment (Laviolette et al., 2002), leading to the hypothesis that a specific pattern of signaling at DA receptors mediates the expression of opiate withdrawal aversions. I thus tested whether increasing or decreasing DA signaling at receptors by using a DA agonist or antagonist, respectively, would prevent the expression of nicotine withdrawal aversions. I next examined the pattern of VTA DA signaling electrophysiologically. Previous research has shown that DA neurons exhibit burst- and population-firing activity that leads to phasic and tonic DA release, respectively (Floresco et al., 2003; Goto and Grace, 2005; Grace, 2000), and that acute nicotine affects phasic VTA DA activity (Mameli-Engvall et al., 2006) while chronic nicotine affects tonic VTA DA activity (Tan et al., 2009). I thus investigated both phasic and tonic VTA DA activity using in vivo electrophysiology after administration of a dose of acute nicotine that produces conditioned place aversions, and after nicotine dependence and 39

54 spontaneous withdrawal from chronic nicotine. I hypothesized that acute nicotine would modify phasic DA activity, while withdrawal from chronic nicotine would modify tonic DA activity. Furthermore, I examined whether blockade of phasic DA activity using antagonist drugs that were known to selectively modify phasic but not tonic DA activity would prevent the expression of acute nicotine aversions or nicotine withdrawal aversions, hypothesizing that only the phasic DA-mediated acute nicotine aversions would be blocked by the selective antagonists. I also examined whether modification of D1Rs and D2Rs would differentially affect acute nicotine and nicotine withdrawal aversions. Since previous research has shown that phasic DA release mainly activates D1Rs and tonic DA release mainly activates D2Rs (Floresco et al., 2003; Goto and Grace, 2005), I hypothesized that modifying D1R activity by using D1R agonists or antagonists, or genetic deletion of the D1R, would prevent the expression of acute nicotine aversions in nondependent mice but not aversions to withdrawal from chronic nicotine in dependent mice. I also hypothesized that modifying D2R activity using D2R agonists and antagonists, or genetic deletion of the D2R, would prevent aversions to withdrawal from chronic nicotine in dependent mice but not to acute nicotine in nondependent mice. Similarly, the adenosine A2AR is colocalized with D2Rs on neurons in the mesolimbic system (Tozzi et al., 2011) and acts antagonistically to the D2R (Tanganelli et al., 2004). I thus hypothesized that modifying A2AR activity would prevent aversions to chronic nicotine withdrawal, but not acute nicotine, and tested this idea using A2AR agonists and antagonists as well as A2AR KO mice. In chapter 4 I examined the involvement of the CRF system in the VTA in nicotine motivation. It is well known that the mesolimbic DA system originating in the VTA is important in mediating nicotine s motivational effects, and that CRF is involved in the negative effects of withdrawal from drugs of abuse. Previous work has demonstrated that CRF1 receptors are present in the VTA (Sauvage and Steckler, 2001), and centrally administered CRF1 receptor antagonists can mediate the motivational effects of nicotine (Bruijnzeel et al., 2009; Tucci et al., 2003). Furthermore, chronic cocaine administration leads to the recruitment of CRF1 receptors in the VTA that may control VTA activity (Hahn et al., 2009; Lodge and Grace, 2005), and CRF- BP has been identified in VTA neurons (Wang and Morales, 2008), suggesting that a CRF/DA mechanism operating in the VTA may be involved in drug addiction, withdrawal, and relapse. I 40

55 thus hypothesized that chronic nicotine and withdrawal would recruit and activate the CRF- CRF1 system in the VTA, and that this newly activated CRF would mediate the aversive response to withdrawal from chronic nicotine but not to acute nicotine. To test whether nicotine dependence and withdrawal upregulates brain CRF levels, I measured CRF mrna in the PVN, CeA and VTA using quantitative real-time polymerase chain reaction (rtpcr) and in situ hybridization, as well as CRF protein levels using immunohistochemistry. I then observed the effect on the motivational responses to acute nicotine and chronic nicotine withdrawal after blocking CRF mrna and CRF1R activation using sirna knockdown and antagonist drugs, respectively. If these modifications of CRF activity prevented the expression of nicotine withdrawal aversions, the hypothesis that CRF activity mediates nicotine withdrawal motivation would be supported. 41

56 Chapter 2 Dopaminergic Signaling Mediates the Motivational Response Underlying the Opponent Process to Chronic but Not Acute Nicotine Taryn E. Grieder, Laurie H. Sellings, Hector Vargas-Perez, Ryan Ting-A-Kee, Eric C. Siu, Rachel F. Tyndale and Derek van der Kooy This chapter is adapted from the paper published in Neuropsychopharmacology, vol. 35, p , Reprinted with permission. 42

57 Abstract The mesolimbic DA system is implicated in the processing of the positive reinforcing effect of all drugs of abuse, including nicotine. It has been suggested that the dopaminergic system is also involved in the aversive motivational response to drug withdrawal, particularly for opiates, however the role for dopaminergic signaling in the processing of the negative motivational properties of nicotine withdrawal is largely unknown. We hypothesized that signaling at dopaminergic receptors mediates chronic nicotine withdrawal aversions and that dopaminergic signaling would differentially mediate acute versus dependent nicotine motivation. We report that nicotine dependent rats and mice demonstrated conditioned place aversions to an environment paired with withdrawal from chronic nicotine that were blocked by the DA receptor antagonist α-flu and in D2R KO mice. Conversely, α-flu pretreatment had no effect on preferences for an environment paired with abstinence from acute nicotine. Taken together, these results suggest that dopaminergic signaling is necessary for the opponent motivational response to nicotine in dependent, but not non-dependent, rodents. Further, signaling at the D2R is critical in mediating withdrawal aversions in nicotine dependent animals. We propose that the alleviation of nicotine withdrawal primarily may be driving nicotine motivation in dependent animals. 43

58 Introduction Nicotine is the major reinforcing constituent of tobacco smoke that is responsible for smoking dependence in humans (Stolerman and Jarvis, 1995). Nicotine causes its motivational effects by acting on nicotinic receptors localized in the mesocorticolimbic DA system (Koob, 2001; Mansvelder and McGehee, 2002). Like most drugs of abuse, nicotine increases the extracellular concentration of DA in the mesolimbic system (Di Chiara and Bassareo, 2007; Picciotto and Corrigall, 2002). Nicotine also produces motivational effects through nondopaminergic neural systems such as the cholinergic TPP (Lanca et al., 2000; Laviolette et al., 2002; Levin et al., 1996; Picciotto and Corrigall, 2002). The mesolimbic DA system has been implicated in the processing of the acute motivational properties of nicotine (Laviolette et al., 2003, 2008; Spina et al., 2006; Tanabe et al., 2008); however, the involvement of dopaminergic signaling in the aversive response to chronic nicotine withdrawal is largely unknown. The most common reason for relapse reported by quitting smokers is the desire to relieve the discomforts of withdrawal (Allen et al., 2008). The aversive abstinence syndrome experienced by quitters as well as the ability of renewed nicotine use to relieve this syndrome likely contributes to relapse. Similar to nicotine-withdrawn humans, rodents that undergo spontaneous withdrawal show a somatic nicotine abstinence syndrome (Epping-Jordan et al., 1998; Malin et al., 1992; Stoker et al., 2008). The negative affective state of withdrawal and its alleviation by nicotine is one of the primary factors driving nicotine craving in nicotine dependent subjects. Most studies on nicotine motivation in dependent animals involve antagonist-precipitated withdrawal (George et al., 2007; Kenny and Markou, 2001; Laviolette et al., 2008; Watkins et al., 2000). However, a spontaneous withdrawal procedure more closely models human withdrawal. When a psychoactive drug triggers a motivational response, animals will experience a rebound motivational state (Koob et al., 1989; Koob and Le Moal, 2001; Robinson and Berridge, 2003; Wise, 1996) that is predicted by the opponent process theory of motivation (Solomon and Corbit, 1974). This theory postulates that any motivational stimulus activates two opposing motivational processes: The a-process has a fast onset and offset and the b-process is opposite in 44

59 direction, lasts longer and is slower to start and end (Figure 2.1a). Similar to other drugs of abuse, chronic nicotine produces a negative withdrawal syndrome that can be viewed as the opponent process to the rewarding effects of nicotine in dependent subjects (Gutkin et al., 2006; Koob and Le Moal, 1997). Acute nicotine produces both rewarding and aversive stimulus properties (Laviolette and van der Kooy, 2004; Sellings et al., 2008; Wilkinson and Bevins, 2008). Dopaminergic signaling is involved in acute nicotine aversion (Laviolette and van der Kooy, 2003; Tan et al., 2009) and chronic nicotine motivation (Bruijnzeel and Markou, 2005; Kenny and Markou, 2001; Laviolette et al., 2008; Smolka et al., 2004); however, little is known about the role of dopaminergic signaling in the opponent motivational processes of acute and chronic nicotine. We thus investigated the role of dopaminergic signaling in the acute and chronic nicotine a- and b-processes. We first studied the correlation between somatic and affective nicotine withdrawal by examining the timing of the maximal somatic withdrawal syndrome and motivational withdrawal in a place conditioning paradigm. Next, we subjected previously drug naive and nicotine dependent rodents to place conditioning after acute and chronic nicotine, respectively, and examined the opponent motivational processes after nicotine exposure. The involvement of dopaminergic signaling in the aversive a-process and rewarding b-process of acute nicotine and the rewarding a-process and aversive b-process of chronic nicotine was investigated by treatment with the DA receptor antagonist α-flu prior to conditioning. We also examined D2 receptor involvement in chronic nicotine withdrawal aversions. The D2R has been implicated in nicotine dependence (Fehr et al, 2008) and withdrawal (Laviolette et al, 2008). Our results demonstrate that acute aversive and chronic rewarding nicotine lead to opponent a- and b- processes and that dopaminergic signaling, specifically at the D2R, mediates the opponent motivational process of chronic aversive but not acute rewarding nicotine. Materials and Methods Animals 45

60 Figure 2.1. The opponent process theory of motivation and its modeling by use of the place conditioning paradigm. (a) The opponent process theory of motivation. Solomon and Corbit (1974) postulated that any stimulus would trigger an initial a-process that will closely follow the stimulus and will be fast to occur and fast to end. The a-process can be rewarding or aversive and will be followed by a later occurring b-process that is longer lasting, slower to end and is opposite in direction to the a- process. At the dose used in the present experiments, acute nicotine is aversive and the a-process is therefore negative. The acute nicotine b-process will be later occurring and positive or rewarding. We postulate that chronic nicotine elicits a rewarding a-process in dependent animals and the aversion to nicotine withdrawal is the conditioned opponent b-process. (b) The B, N, and W procedures. In procedure B, each animal experienced the effects of nicotine in one environment (Nic) and the lack of nicotine (or the effects of withdrawal) in the other environment (WD). Procedure B measures both the rewarding value of the drug itself and the aversiveness associated with drug withdrawal, and models both the a- and b-process of the opponent process theory. In procedure N, each animal was conditioned only while experiencing the effects of chronic nicotine. On the alternate day, the animals experienced withdrawal in their home cage. Procedure N measures the rewarding value of the drug itself, modeling only the a- process of the opponent process theory. In procedure W, conditioning took place only while the animals experienced withdrawal from nicotine. On the alternate day, the animal was confined to its home cage during chronic nicotine exposure. Procedure W measures only the aversive motivational effects of drug withdrawal, separate from the rewarding value of the drug itself, and is used as a model to measure the b-process of the opponent process theory of motivation. 46

61 A Appetitive Stimulus (Reward) a process b process Stimulus Response Aversive Stimulus b process a process B Nic WD B procedure Nic N procedure WD W procedure 47

62 Male WT mice were C57BL/6 (n = 329; Charles River, Montreal, Canada) weighing g. Heterozygous 5 th generation D2 breeder mice were received as a gift (Kelly et al., 1997) and crosses were bred at the University of Toronto to obtain homozygous male D2R KO mice (n = 26) and their controls (n = 20). Mice were housed individually in plastic cages in a soundattenuated room at a temperature of 22 C with lights on from 7:00 AM to 7:00 PM. Male Wistar rats (Charles River) weighing g (n = 128) were individually housed in Plexiglas cages in a room kept at a temperature of 22 C with lights on from 7:00 AM to 7:00 PM. All animals had ad libitum access to food and water except during behavioral testing. All procedures were approved by the University of Toronto Animal Care Committee in accordance with the Canadian Council on Animal Care guidelines. Chronic nicotine treatment (-)-nicotine hydrogen tartrate salt (Sigma-Aldrich, Ontario) titrated to a ph of 7.0±0.3 or saline was administered to mice (n = 263) and rats (n = 128) using osmotic minipumps (models 1002 and 2001, respectively; Alzet, Cupertino, CA). Animals were anesthetized by inhalation of 5% isofluorane in oxygen (1-2% maintenance) and the minipump placed subcutaneously between the scapulae parallel to the spine. Nicotine was administered at doses of 1.4 and 7 mg/kg/day (free base) for 13 days in mice and 1 and 3.15 mg/kg/day (free base) for 7 days in rats based on previous studies showing that these doses induce nicotine dependence with the expression of spontaneous somatic withdrawal signs (Damaj et al., 2003; Malin et al., 1992, 2006; Watkins et al., 2000). After minipump implantation, the surgical wound was sutured and treated with Polysporin antibiotic cream. Due to the faster metabolism of mice in comparison to rats (Matta et al., 2007), mice were exposed to chronic nicotine for 6 additional days. Blood Analysis Blood was collected by cardiac perfusion from nicotine dependent mice (n = 3) after 12 days of exposure at the 7 mg/kg/day dose. Samples were then analyzed by high performance 48

63 liquid chromatography as described previously (Siu and Tyndale, 2007). Somatic withdrawal assessment Wild-type mice (n = 24) were observed for somatic signs of nicotine withdrawal at 30 minutes, 4, 8, 12, 24 and 48 hours after minipump removal. A group of D2R KO mice (n = 6) was also observed for somatic signs of withdrawal at 8 hours following minipump removal. Rats (n = 24) were observed at 30 minutes, 4, 8, 12, 16, 24, 36 and 48 hours after minipump removal. A group of rats pretreated with α-flu (n = 8) were observed for somatic signs 16 hours following minipump removal. Experimenters were blind to the drug treatment of each subject. Typical abstinence signs in mice included head shakes, paw tremors, writhing, scratching, backing and jumping (Isola et al., 1999; Stoker et al., 2008). Rats were observed for body and head shakes, cheek tremors, eye blinks, ptosis, foot and genital licks, scratches, writhes and gasps (Malin et al., 1992). Place Conditioning Procedure Mice and rats were conditioned in an apparatus as described previously (Dockstader et al., 2001; Vargas-Perez et al., 2009). Briefly, mice were conditioned in an apparatus consisting of two different environments measuring 15 x 15 x 15 cm. One environment was black with a smooth Plexiglas floor that was wiped with 5% acetic acid and the other environment was white with a wire mesh floor. The boxes were separated by a removable wall that was painted with the corresponding color on each side. During preference testing, the dividing wall was removed and mice were given free access to both environments. Rats were conditioned in boxes measuring 41 x 41 x 38 cm. One environment was black with a Plexiglas bottom wiped with a 3% acetic acid solution prior to conditioning. The other environment was white with a smooth aluminum bottom covered by a mesh grid. The test cage consisted of a black and a white conditioning cage separated by a middle grey area. For preference testing, the rats were placed in the neutral grey zone and given free choice between the different environments. 49

64 Each cage was cleaned between animals and each group was fully counterbalanced. A single 10 min preference testing session was performed 3-5 days after the last conditioning day, when subjects were drug- and withdrawal-free. Behavioral testing for rats consisted of three phases: pre-exposure, conditioning and testing. The pre-exposure phase comprised a single 20 minute session in separate boxes painted grey with a grey floor. The conditioning phase comprised one to two sessions of 40 min each for rats and 1 hour for mice, depending on the procedure (B, N, or W - see below for details). All place conditioning and testing was performed between 8:30 AM and 7 PM. Procedure B (both drug and withdrawal pairing) was adapted from the method described by Bechara et al. (1992). Procedure B involved two pairings, the first having one environment paired to the administration of nicotine that was continuously delivered through a minipump. For the second pairing, the minipump was removed and the animal was paired to the other environment while experiencing withdrawal. The drug-paired environment was counterbalanced within groups. Before the withdrawal-paired conditioning, each mouse (n = 100) and rat (n = 16) underwent 8 hours and 16 hours of abstinence from nicotine, respectively. Thus, in procedure B, each animal experienced the effects of chronic nicotine in one environment and the lack of nicotine (or the effects of withdrawal) in the other environment (Figure 2.1b). The difference score for each animal was calculated by subtracting the time spent in the withdrawal-paired environment from the time spent in the nicotine-paired environment. This method of place conditioning (procedure B) measures both the rewarding value of the drug itself and the aversiveness associated with drug withdrawal (Bechara et al., 1992). Procedure N (nicotine only) was a modified place conditioning procedure for which conditioning took place in only the nicotine-paired environment of the place conditioning apparatus. As in procedure B, each animal was chronically nicotine treated and confined to one of the environments. On the alternate day, the minipump was removed and the animal experienced withdrawal in its home cage. Thus, in contrast to procedure B, the mice (n = 23) and rats (n = 10) were never allowed to experience withdrawal in the other compartment of the place conditioning apparatus. The difference score for each animal was calculated by subtracting the time spent in the non-paired environment from the time spent in the nicotine-paired environment. 50

65 This place conditioning method (procedure N) measures the rewarding value of the drug itself (Bechara and van der Kooy, 1992). In procedure W (withdrawal only), conditioning took place in only the withdrawal-paired environment of the place conditioning apparatus. The key difference between this procedure and procedure N was that withdrawal only (but not the direct effects of chronic nicotine) was paired with one compartment of the place conditioning apparatus. On the first day, each mouse (n = 60) and rat (n = 14) received a sham surgery where the minipump was removed and replaced immediately, controlling for any effects of surgery on conditioning. For the remainder of the day, the animal was confined to its home cage. On the conditioning day, the minipump was removed and when the animal was experiencing withdrawal they were confined to one of the conditioning environments. Mice were conditioned at 4 hours (n = 10), 8 hours (n = 38) and 12 hours (n = 12) following pump removal. Rats were conditioned at 16 hours following pump removal. The difference score for each animal was calculated by subtracting the time spent in the withdrawalpaired environment from the time spent in the non-paired environment. This method of place conditioning (procedure W) measures only the aversive motivational effects of drug withdrawal, separate from the rewarding value of the drug itself (Bechara and van der Kooy, 1992). Effects of α-flupenthixol Mice (n = 51) and rats (n = 12) were made nicotine dependent and conditioned according to procedure B or W as described above except that subjects were pretreated (i.p.) with either saline or α-flu (Sigma-Aldrich). This DAR antagonist has no motivational effects of its own at the doses and times used in this study (Laviolette and van der Kooy, 2003) and is known to antagonize both D1Rs and D2Rs (Creese et al., 1976). Mice (n = 12) were also conditioned according to procedure N. Mice were pretreated with 0.8 mg/kg (i.p.) at 60 minutes and rats with 0.1 mg/kg (i.p.) α-flu at 120 minutes prior to conditioning. Acute nicotine conditioning 51

66 Previously drug naive WT mice (n = 62) were given a single dose of nicotine (1.75 mg/kg free base, s.c.) in one environment and saline in the other environment. This dose of nicotine was expected to produce an acute aversive motivational response (Rauhut et al, 2008). The mice were conditioned in the B procedure in the same way as dependent and withdrawn mice (described above). To examine the acute nicotine a-process, previously drug naive mice (n = 20) were conditioned immediately following nicotine administration. To examine the effect of DA system blockade on the acute nicotine a-process, an injection of α-flu or saline was administered one hour prior to nicotine (n = 18). To examine the acute nicotine b-process, mice (n = 42) were conditioned 8 hours after nicotine administration for one hour. To examine DA system involvement in the acute nicotine b-process, α-flu or saline was administered one hour prior to conditioning (n = 24; blocking the b-process) or one hour prior to nicotine administration (n = 18; blocking the a-process) in separate groups of mice. Statistical analysis Somatic withdrawal results were analyzed with SYSTAT software using a two-way repeated measures ANOVA at each somatic withdrawal point after dependence assessment. For conditioned place preference experiments, statistical analysis was performed with a one- or twoway ANOVA. Posthoc Student-Newman-Keuls tests or Student s t-tests were performed where appropriate. P values of less than 0.05 were considered to be significant. Results Both rats and mice exhibit somatic signs upon withdrawal from chronic nicotine Discontinuing the administration of chronic nicotine after 7 days in rats and 12 days in 52

67 mice produced a spontaneous somatic nicotine abstinence syndrome. The severity of this syndrome at various time points following chronic pump removal is depicted in Figure 2.2a for mice and 2.2b for rats, shown as mean abstinence scores taking saline as 100%. A two way repeated measures ANOVA comparing mouse abstinence scores revealed a significant dose x time interaction (F 10,75 = 2.761, p < 0.05). Nicotine withdrawn mice displayed significantly increased somatic withdrawal signs compared to saline-treated mice in both the 1.4 and 7 mg/kg/d group at 8 hours following pump removal (p < 0.05), but not at 4 hours (p > 0.05) or 12 hours (p > 0.05) following pump removal. The 7 mg/kg/day nicotine dose was selected for use in subsequent mouse experiments due to the largest abstinence syndrome being observed with this dose at the 8 hour time point. To test whether DA signaling and the D2R specifically is involved in the emergence of spontaneous nicotine withdrawal after chronic nicotine exposure, a group of D2R (-/-) KO mice (n = 6) was observed for abstinence signs at 8 hours after removal of pumps containing the 7 mg/kg/day nicotine dose. D2R KO mice exhibited spontaneous somatic signs of withdrawal at a similar level as WT mice given the 7 mg/kg/day dose of nicotine (t 10 = 1.041, p > 0.05; Figure 2.2a), demonstrating that the D2R is not involved in somatic withdrawal. Nicotine withdrawn rats after 7 days of exposure displayed the largest abstinence syndrome at 16 hours following pump removal. A two way repeated measures ANOVA comparing abstinence scores in rats revealed a significant dose x time interaction (F 18,189 = 5.740, p <0.05; Figure 2.2b). Somatic withdrawal signs were significantly increased compared to saline-treated rats in both the 3.15 mg/kg/d group (p < 0.05) and 1 mg/kg/d group (p < 0.05) at 16 hours following pump removal, but not at baseline prior to pump removal (F 2,21 = 0.190, p > 0.05) nor after 48 hours (F 2,21 = 2.900, p > 0.05) following pump removal. An additional group of rats (n = 8) pretreated with α-flu was observed for somatic signs of nicotine withdrawal 16 hours after pump removal. Nicotine withdrawn rats treated with α-flu showed significant somatic withdrawal signs in comparison to saline treated animals (t 14 = 6.20, p < 0.05; Figure 2.2b), therefore it appears that somatic withdrawal is not mediated by signaling at DA receptors. 53

68 Figure 2.2. The time course of spontaneous nicotine somatic and motivational withdrawal. (A) Mice were given chronic nicotine in osmotic minipumps (7 and 1.4 mg/kg/day) for 13 days. Somatic withdrawal signs were recorded at 30 minutes, 4, 8, 12, 24 and 48 hours following minipump removal. An abstinence syndrome compared to saline-treated mice was observed that peaked at 8 hours following pump removal, suggesting that chronic nicotine and spontaneous withdrawal will induce a somatic withdrawal syndrome in mice that peaks 8 hours after the removal of chronic nicotine. A group of D2R KO mice given 7 mg/kg/day nicotine for 13 days was also observed for abstinence signs 8 hours after pump removal. The D2R KO mice showed a somatic withdrawal syndrome that did not differ from WT mice, suggesting that the D2R does not mediate the expression of a somatic nicotine withdrawal syndrome. (B) Somatic withdrawal signs were recorded in rats at 30 minutes, 4, 8, 12, 24 and 48 hours following minipump removal. After 7 days of nicotine minipumps (1 and 3.16 mg/kg/day), an abstinence syndrome compared to saline-treated rats was observed that peaked at 16 hours following pump removal. A group of rats treated with α-flu were observed for abstinence signs 16 hours after pump removal. These rats showed a somatic withdrawal syndrome similar to nicotine-dependent and -withdrawn rats, suggesting that DA receptor antagonism does not affect the expression of a somatic withdrawal syndrome. (C) Mice were trained in the W procedure at 4, 8, and 12 hours following pump removal at the 7 mg/kg/day chronic nicotine dose. A significant aversive motivational response to the withdrawal-paired environment was observed in nicotine-dependent and - withdrawn mice only at the 8-hour time point. These results suggest that the motivational response to withdrawal corresponds to the time point of somatic withdrawal. Data represent means +/- SEM (*p < 0.05). 54

69 A % Abstinence score B % Abstinence score C Difference score (s) Mice * Saline 1.4 mg/kg/d 7 mg/kg/d D2 KO 30 m 4 h 8 h 12 h 24 h 48 h Time since pump removal Rats Saline 1 mg/kg/d 3.16 mg/kg/d α-flu pretreatment 30 m 4 h 8 h 12 h 16 h 24 h 36 h 48 h Time since pump removal * * 4 h 8 h 12 h Saline Time since pump removal 55

70 Chronic nicotine exposure did not noticeably affect the subjects during the exposure period (based on observations of locomotor activity, feeding patterns and general behaviour). To compare our dose of nicotine in mice to the human condition, we analyzed plasma levels of nicotine in mice treated with the 7 mg/kg/day dose and found an average of 29.9 ng/ml ± 15.5, a result that is similar to the average maximum arterial blood concentration of human chronic smokers (range ~20-40 ng/ml) (Armitage et al., 1975; Matta et al., 2007; O Dell et al., 2006). Somatic and affective nicotine withdrawal occur coincidentally To determine if the time when maximal somatic withdrawal signs were observed corresponded to the time of motivational response in a place conditioning paradigm, chronic nicotine-treated mice were conditioned in the W procedure at 4 and 12 hours following pump removal (when few somatic withdrawal signs were observed in mice) and compared to 8 hours following pump removal (when most somatic withdrawal signs were observed in mice). Salinetreated mice conditioned at 4, 8, and 12 hours following pump removal showed no significant difference for time of conditioning (F 2,25 = 0.279, p > 0.05) and were therefore analyzed as one group. A two-way ANOVA showed a significant treatment x withdrawal time interaction (F 2,44 = 3.414, p < 0.05; Figure 2.2c). Nicotine-treated mice demonstrated a significant motivational effect only at the 8 hour time point (p < 0.05), a result which validates the use of the 8 hours following pump removal time point as our maximal withdrawal conditioning time and suggests that somatic withdrawal coincides with motivational withdrawal over time. Mice and rats exhibit conditioned place aversions to an environment paired with nicotine withdrawal To dissociate the rewarding from the aversive motivational effects of nicotine and nicotine withdrawal in dependent subjects, we performed place conditioning using the B, N and W procedures. A one-way ANOVA comparing the B, N, W procedures and saline in mice showed a significant effect of chronic nicotine treatment (F 3,47 = , p < 0.05; Figure 2.3a). Saline-treated mice in each of the B, N, and W procedures were not significantly different (F 2,13 56

71 Figure 2.3. The opponent processes of chronic and acute nicotine and the effect of DA receptor antagonism. (a) Separation of the rewarding effects of chronic nicotine from the aversive effects of withdrawal in nicotine dependent mice, as revealed by a modified place conditioning paradigm for assessing the rewarding effects of nicotine only (Nicotine procedure), the aversive effects of nicotine withdrawal (Withdrawal procedure), or both the rewarding effects of nicotine and the aversive effects of withdrawal (Both procedure). Nicotine dependent mice conditioned in all three procedures showed a significant motivational response as compared to saline-treated animals, with negative and positive difference scores representing aversive and rewarding motivational responses, respectively. (b) α-flu pretreatment attenuated the motivational response in each of the B, N, and W procedures, and had no motivational effect on its own in salinetreated mice, suggesting that both the rewarding response to chronic nicotine and the aversive response to withdrawal are DA-mediated. (c) Separation of the rewarding effects of chronic nicotine from the aversive effects of withdrawal in nicotine dependent rats. Significant effects as measured by place conditioning were observed in the Both and Withdrawal procedures, but not in the Nicotine procedure as compared to saline. (d) α-flu pretreatment attenuated the B and W effects. (e) Previously drug naive mice administered a single dose of nicotine and conditioned immediately (0 hours) in the B procedure demonstrated a significant conditioned place aversion for the nicotine-paired environment. Mice conditioned 8 hours after acute nicotine administration demonstrated a significant preference for the nicotine-paired environment. Mice conditioned 4 or 12 hours after nicotine administration showed no significant motivational effect. These results suggest that after acute nicotine administration, an aversive a-process is followed by a rewarding b-process 8 hours later. Data represent means +/- SEM (*p < 0.05). 57

72 Difference score (s) Mice Both Nicotine Withdrawal Saline Difference score (s) α-flu Both Nicotine Withdrawal Saline Procedure Procedure Difference score (s) Rats Both Nicotine Withdrawal Saline Difference score (s) α-flu Both Withdrawal Saline Procedure Procedure Difference Score (s) Time since acute nicotine administration (h) 58

73 = 0.447, p > 0.05) and were therefore analyzed as one group. Mice showed a significant aversion to the withdrawal-paired side as compared to the nicotine-paired side in the B procedure (p < 0.05), demonstrating that either nicotine is rewarding to dependent animals, or withdrawal is aversive, or both. To determine which of these motivational effects was responsible for the B procedure results, we used the N and W procedures, which separate the rewarding effects of a drug from the aversiveness of withdrawal (Bechara and van der Kooy, 1992). Nicotinedependent mice demonstrated a significant preference for the nicotine-paired side versus the nonpaired side in the N procedure (p < 0.05), suggesting that the presence of nicotine in a nicotinedependent animal is rewarding. In the W procedure, mice demonstrated a significant aversion to the withdrawal-paired side compared to the non-paired side (p < 0.05), suggesting that withdrawal from chronic nicotine is indeed aversive. To control for a bias in novelty seeking in the N and W procedures, we tested for a place preference for a novel, previously unpaired environment in saline-treated animals. Saline-treated mice did not show a preference for the novel side over a previously paired side (t 5 = 0.932, p > 0.05; data not shown), demonstrating that a novelty effect cannot account for the N and W procedure results. To evaluate whether the rewarding and aversive properties of chronic nicotine would generalize to another species we conditioned rats using the same protocols as mice. A one-way ANOVA comparing the B, N, W procedures and saline in rats showed a significant effect of procedure (F 3,44 = , p < 0.05; Figure 2.3c). Saline-treated rats in each of the B, N, and W procedures were not significantly different (F 2,21 = 0.016, p > 0.05) and were therefore analyzed as one group. Rats demonstrated a significant aversion to the withdrawal-paired side as compared to the nicotine-paired side in the B procedure (p < 0.05). In the N procedure, nicotinedependent rats did not prefer the nicotine-paired or the non-paired environment (p > 0.05). Rats tested in the W procedure demonstrated an aversion to the withdrawal-paired side compared to the non-paired side (p < 0.05), suggesting that withdrawal from chronic nicotine is indeed motivationally aversive. Results from the N and W procedures in rats suggest that the motivational effect in the B procedure may be attributable primarily to an aversive response to nicotine withdrawal. 59

74 Acute nicotine stimulates opposing motivational processes We next examined whether acute nicotine would elicit opposing motivational processes in previously drug naive mice, using the B procedure and identical conditioning times to those used above in nicotine dependent and withdrawn mice. We administered a single dose of nicotine (1.75 mg/kg, s.c.) and conditioned separate groups of mice immediately (0 hours) and 4, 8 and 12 hours after nicotine administration. A one-way ANOVA showed a significant effect of conditioning time (F 3,36 = 7.827, p < 0.05; Figure 2.3e). When drug naive mice were given an acute aversive dose of nicotine and conditioned immediately following nicotine administration, they demonstrated a significant conditioned place aversion (p < 0.05). The opponent process theory of motivation suggests that an acute aversive effect of nicotine (a-process) will set up a longer lasting opponent b-process that should manifest as a rewarding response to the environment paired with 8 hours of abstinence from acute aversive nicotine. We observed a significant preference for the nicotine-paired environment at 8 hours following nicotine administration (p < 0.05; Figure 2.3e). When mice were conditioned 4 and 12 hours after acute nicotine administration, they showed no significant motivational effect (p > 0.05). These results demonstrate that an acute aversive dose of nicotine stimulates two separate opposing motivational effects: an aversive a-process response when mice are conditioned immediately after nicotine administration and a rewarding b-process response when mice are conditioned 8 hours after nicotine administration. DA receptor blockade attenuates chronic nicotine withdrawal aversions We examined if the DA system mediates the motivational responses to nicotine and withdrawal in the nicotine dependent state. Mice were given α-flu 1 hour prior to conditioning in the B, N, and W procedures. A two way ANOVA using procedure (B, N, W, or saline) and pretreatment (α-flu or saline) as independent factors showed a procedure x pretreatment interaction (F 3,106 = , p < 0.05; Figure 2.3b). Saline-treated mice in each of the B, N, and W procedures were not significantly different (F 2,19 = 0.043, p > 0.05) and were therefore analyzed as one group. DA receptor blockade prevented the conditioned motivational effect in the B, N, and W procedures (p > 0.05). Mice conditioned in the B procedure with α-flu in one 60

75 environment and saline in the other environment showed no motivational preference (t 6 = 0.861, p > 0.05; data not shown), demonstrating that the dose of α-flu used in the present experiments does not have any motivational effects on its own. Consistent results were obtained in rats when they were pretreated with α-flu 1 hour prior to conditioning in the B and W procedures. A two way ANOVA using procedure (B, W, or saline) and pretreatment (α-flu or saline) as independent factors showed a procedure x pretreatment interaction (F 2,61 = 9.174, p < 0.05; Figure 2.3d). Saline-treated rats in the B and W procedures were not significantly different (F 1,14 = 0.032, p > 0.05) and were therefore analyzed as one group. We did not include a group of rats pretreated with α-flu in the N procedure, as no motivational effect was observed in the dependent group (Figure 2.3c). DA receptor blockade prevented the conditioned motivational effect in the B and W procedures (p > 0.05). The results from both rats and mice receiving α-flu pretreatment in the W procedure demonstrate that the DA system is mediating the aversive response to withdrawal after chronic nicotine. Dopaminergic signaling specifically mediates chronic nicotine withdrawal aversions In order to further investigate dopaminergic mediation of the rewarding effects of nicotine versus the aversiveness of withdrawal in nicotine-dependent animals, we gave α-flu to mice in either the nicotine-paired or withdrawal-paired environment in the B procedure and observed the effect on the aversive motivational response after 8 hours of abstinence from chronic nicotine. A one-way ANOVA showed a significant effect of treatment (F 2,33 = , p < 0.05; Figure 2.4a). Mice given saline on both conditioning days demonstrated an aversive motivational response to chronic nicotine withdrawal, as observed previously (saline; p < 0.05). When mice were pretreated with α-flu on the nicotine-paired side and given saline on the withdrawal-paired side, a preference for the nicotine-paired environment over the withdrawalpaired environment was observed (no delay; p < 0.05). Mice that received the opposite treatment (α-flu on the withdrawal-paired side and saline on the nicotine-paired side) did not show an aversive response to withdrawal from chronic nicotine (delay; p > 0.05). These results suggest that the dopaminergic system mediates the aversive response to withdrawal from chronic nicotine. 61

76 Figure 2.4. Dopaminergic signaling differentially mediates the opponent motivational process after acute and chronic nicotine. (a) Mice chronically treated with nicotine and pretreated with saline demonstrate a conditioned place aversion to the withdrawal-paired environment in the B procedure. α-flu given 1 hour prior to conditioning during chronic nicotine administration (no delay) did not prevent the aversive response from occurring in dependent mice. However, α-flu given prior to conditioning after 8 hours of withdrawal from chronic nicotine (delay) blocked the aversive response from occurring. These results suggest that DA activity is necessary for the b-process occurring after chronic nicotine. (b) Previously drug naive mice given acute nicotine, pretreated with saline and conditioned after 8 hours of abstinence from nicotine demonstrate a preference for the environment paired with the acute nicotine b-process. α-flu administered prior to the acute nicotine a-process (no delay) prevented this rewarding response from occurring. However, α-flu given prior to b-process conditioning after 8 hours of abstinence from acute nicotine (delay) did not prevent the occurrence of the rewarding response, suggesting that DA activity is not required for the acute nicotine b-process. Data represent means +/- SEM (*p < 0.05). 62

77 A Difference score (s) Chronic nicotine * B 200 Acute nicotine Difference score (s) * Saline α-flu, no delay α-flu, delay 63

78 DA antagonism does not block the rewarding motivational response 8 hours after acute nicotine We next investigated if the DA system mediates the positive b-process response observed 8 hours after acute nicotine in the same way as it does the aversive b-process response observed 8 hours after chronic nicotine. The immediate aversive effects of acute nicotine are blocked by pretreatment with α-flu in rats (Laviolette and van der Kooy, 2003). Similarly, we observed that the immediate aversive response to acute nicotine (the a-process) in mice was blocked by α-flu pretreatment (t 6 = 1.785, p>0.05; data not shown). To investigate the effect of DAR antagonism on the rewarding motivational b-process 8 hours after acute aversive nicotine, we administered α-flu 1 hour prior to conditioning for either the immediate aversive or the delayed rewarding effect, similar to the procedure followed with nicotine dependent mice. A one-way ANOVA showed a significant effect of treatment (F 2,39 = 4.068, p < 0.05; Figure 2.4b). Mice pretreated with saline showed a preference for an environment paired with 8 hours of abstinence from acute nicotine (the b-process), as observed previously (Saline; p < 0.05). However, α-flu pretreatment prior to acute nicotine prevented the later occurring rewarding response (no delay; p > 0.05). This result suggests that the immediate aversive a-process response to acute nicotine is required for the delayed rewarding b-process response to occur. Mice treated with α-flu prior to conditioning for the delayed rewarding response exhibited a similar conditioned place preference to saline-treated mice (delay; p < 0.05). These results demonstrate that dopaminergic signaling is required for conditioning to the immediate aversive effect of acute nicotine, but not the delayed rewarding b-process effect occurring 8 hours after acute nicotine. The D2R mediates the aversive response to withdrawal from chronic nicotine The behavior of male D2 (+/+) WT and D2 (-/-) KO littermate mice was examined using the N and W place conditioning procedures to determine the role of the D2R in mediating the motivational response to chronic nicotine and withdrawal. A two-way ANOVA using genotype and procedure (N or W) showed a genotype x procedure interaction (F 2,42 = , p < 0.05; Figure 2.5). WT mice conditioned in the N procedure showed a significant preference for an environment paired with chronic nicotine (p<0.05) that was not demonstrated by D2R KO mice (p > 0.05). WT mice conditioned in the W procedure demonstrated a significant aversion to 64

79 Figure 2.5. The D2R mediates the aversive response to chronic nicotine withdrawal. WT mice conditioned in procedure N showed a preference for chronic nicotine that was not present in D2R KO mice. D2R KO mice conditioned in procedure W did not exhibit nicotine aversions while WT mice showed normal aversions to chronic nicotine withdrawal. These results suggest that the D2R is important for signaling both the rewarding response to chronic nicotine and the aversive response to nicotine withdrawal in dependent mice. Data represent means +/- SEM (*p < 0.05). Difference Score (s) * WT D2 KO Nicotine Procedure * Withdrawal 65

80 an environment paired with withdrawal from chronic nicotine (p < 0.05), similar to the WT mice in previous experiments. However, D2R KO mice did not show an aversion to withdrawal (p > 0.05), indicating that the D2R mediates the aversive response to withdrawal from chronic nicotine. Discussion Understanding the neurobiological substrates mediating the motivational response experienced by smokers during nicotine withdrawal has important implications for improving smoking cessation. We show here that dopaminergic signaling at the D2R mediates affective but not somatic nicotine withdrawal. Moreover, we have dissociated the role of dopaminergic signaling in the opponent motivational processes of acute and chronic nicotine. Indeed, dopaminergic signaling, specifically at the D2R, is required for the delayed motivational response to chronic nicotine in dependent subjects. In contrast, dopaminergic signaling is required for the immediate but not the delayed motivational response to acute nicotine in nondependent subjects. Somatic signs observed in the present experiments resemble those described previously in mice (Isola et al., 1999; Stoker et al., 2008) and rats (Epping-Jordan et al., 1998; Malin et al., 1992; Watkins et al., 2000); however the time course of peak nicotine somatic withdrawal differs. The use of a variety of time points in our measurements of somatic withdrawal symptoms allowed us to find the most appropriate time for withdrawal motivation studies in both mice and rats. Although the dose used in mice (7 mg/kg/day) and rats (3.15 mg/kg/day) exceeds the amount of nicotine smoked by the heaviest smokers of high-yield cigarettes (Armitage et al., 1975; Epping-Jordan et al., 1998), it is important to consider that rats and mice have much higher metabolic and drug clearance rates (Matta et al., 2007) than humans. Furthermore, plasma levels of nicotine measured presently were similar to those observed in humans and measured in previous rodent studies (Guillem et al., 2005; O Dell et al., 2006). We showed that nicotine somatic withdrawal coincides with affective withdrawal, such that the largest abstinence syndrome in both mice and rats occurs when the aversive response to withdrawal can be 66

81 conditioned in the place preference paradigm. Dopaminergic signal disruption by α-flu pretreatment or genetic deletion of the D2R blocked the aversive motivational response to affective nicotine withdrawal but not the somatic signs of withdrawal. These results suggest that somatic and affective motivational withdrawal occur coincidentally, but are not causally related. Similarly, motivational withdrawal from opiates can be blocked without attenuating somatic withdrawal signs (Bechara et al., 1995), lending support to the idea that somatic withdrawal signs do not necessarily reflect the motivational impact of withdrawal (Watkins et al., 2000). Although self-administration more closely models human nicotine intake (Rose and Corrigall, 1997), separating drug motivation due to the rewarding effects of nicotine or the alleviation of withdrawal is more easily performed using a place conditioning procedure (Mucha et al., 1982). Our place conditioning experiments showed that nicotine withdrawal is aversive in both dependent mice and rats. Furthermore, results from the N and W procedures in rats suggest that the motivational effect in the B procedure may be attributable primarily to an aversive response to nicotine withdrawal, and that the motivational effects observed in the N procedure in mice might reflect the ability of nicotine to overcome the aversiveness of withdrawal. Nicotine reward in dependent animals has been previously demonstrated in conditioned place preference (Acquas et al., 1989; Sellings et al., 2008; Wilkinson et al., 2008), self administration (Tammimaki et al., 2008) and ICSS (Kenny and Markou, 2001) paradigms, however the N place conditioning procedure used in the present experiments was not sensitive enough to measure the rewarding effect of nicotine in dependent rats. An alternative hypothesis would be that the dose of nicotine (compared to mice) used in the present experiments was not sufficient to produce a motivational response, and that a higher dose of nicotine in rats would elicit a preference for the nicotine-paired environment in the N procedure. We demonstrated that dopaminergic signaling through activation of D2Rs is critical for the expression of chronic nicotine withdrawal aversions. It is unlikely that the present results are due to a learning deficit in D2R KO mice as these mice can learn morphine (Dockstader et al., 2001) and ethanol place preferences (Ting-A-Kee et al., 2009). This result confirms previous work showing that D1R and D2R antagonists block conditioned aversions to nicotine withdrawal using pharmacologically precipitated withdrawal (Laviolette et al., 2008) and further extends 67

82 these findings to spontaneous withdrawal, which more closely models the human condition. The opponent process theory of motivation (Solomon and Corbit, 1974) postulates that any motivational stimulus activates two opposing motivational processes, the a-process having a fast onset and offset and the b-process being slower to start, longer lasting and occurring in an opposite direction to the a-process (Ettenberg, 2004; Koob and Le Moal, 2008; Koob et al., 1989). We have shown that an aversive dose of nicotine will act as a negative motivational stimulus in previously drug naive mice that will manifest as a negative a-process, in turn causing the activation of a delayed positive b-process. In nicotine dependent animals, chronic nicotine will act as a rewarding motivational stimulus that will manifest as a positive a-process, in turn causing the activation of a delayed aversive b-process during withdrawal. Each of these opponent motivational effects was modeled using the place conditioning paradigm. We demonstrated that signaling at dopaminergic receptors is required for the immediate aversive response (a-process) but not for the delayed rewarding effect after acute nicotine (b-process). In nicotine dependent animals, we showed that signaling at dopaminergic receptors is required for both the immediate rewarding response to chronic nicotine (a-process) and for the delayed aversive response (bprocess) to withdrawal from chronic nicotine. The immediate aversive response to nicotine in previously drug naive mice could be due to central and/or peripherally mediated effects. Indeed, conditioned place aversions to acute nicotine could be due to peripheral effects since nicotine is known to induce nausea (Perkins et al., 2008). However, conditioned place aversions to acute nicotine have been demonstrated after intra-cerebral administration (Laviolette et al., 2003). Furthermore, nicotine given peripherally has centrally mediated effects on DA release (Seppa et al., 2000), therefore it is unlikely that the acute aversive motivational response is simply due to nausea or another peripheral effect. The role of the dopaminergic and TPP systems in the opponent motivational effects produced by nicotine bears striking resemblance to those produced by opiates. The aversive a- process is DA-mediated in both acute opiate (Zito et al., 1988) and acute nicotine motivation, as shown presently and previously (Laviolette et al., 2003). The rewarding a-process for acute opiates (Bechara et al., 1992) and acute nicotine are TPP mediated (Laviolette et al., 2002). We 68

83 have shown that the rewarding b-process after acute aversive nicotine is not DA-mediated, which resembles the DA-independent acute b-process for opiates (Bechara et al., 1992). Although the TPP appears to be a good candidate to mediate the acute b-process in both nicotine and opiate motivation, it was recently suggested that the TPP does not mediate the acute opiate b-process (Vargas-Perez et al., 2009). However, TPP involvement in the acute nicotine b-process cannot be completely ruled out as the b-process after acute aversive nicotine is rewarding, and TPP involvement in acute nicotine reward has been previously demonstrated (Laviolette et al., 2002). Results from this study showed that nicotine given acutely in a previously drug naive animal elicited a DA-mediated aversion in a place conditioning paradigm, while other groups have demonstrated DA-dependent acute nicotine reward (Acquas et al., 1989; Lecca et al., 2006; Merlo-Pich et al., 1999; Pak et al., 2006; Sellings et al., 2008; Spina et al., 2006). Acute nicotine administered directly into the brain produces both rewarding and aversive motivational effects (Laviolette et al., 2002; Sellings et al., 2008) that can be segregated within the nucleus accumbens (Sellings et al., 2008). Therefore it is not surprising that different groups have reported that different paradigms (place conditioning vs. self-administration) or routes of nicotine administration (intracerebral, subcutaneous, intravenous) produce differences in the direction of the observed motivational response. It is possible that nicotine intake may occur during withdrawal to restore previous levels of dopaminergic signaling in the dependent user s brain and therefore to blunt the negative experience of withdrawal. The N procedure results in mice may be due to the ability of nicotine to alleviate withdrawal. In support of this idea are the present data showing that the motivational response observed in the N and W procedures do not add to give the B (both) motivational effect. The argument then follows that nicotine is not actually rewarding in dependent animals, and the rewarding effect observed in mice conditioned in the N procedure is simply an alleviation of the aversiveness of withdrawal. Furthermore, when dopaminergic signaling was blocked during withdrawal-paired conditioning but left intact during nicotine-paired conditioning in the B procedure, no motivational response was observed. When the opposite experiment was performed, where the DA system was blocked during nicotine-paired conditioning but left intact during withdrawal-paired conditioning, a conditioned place aversion to the withdrawal-paired 69

84 environment was observed. These results suggest that nicotine motivation in nicotine-dependent and -withdrawn animals is driven by a DA-dependent aversion to nicotine withdrawal. Dopaminergic signaling after nicotine administration is a complex phenomenon involving tonic and phasic DA activity (Rice and Cragg, 2004; Zhang and Sulzer, 2004; Zhang et al., 2009). We hypothesize that the motivational response to withdrawal from chronic nicotine is mediated by a dysregulated pattern of DA signaling at the receptor resulting from a decreased level of DA in the NAc during withdrawal (Hildebrand et al., 1998; Rada et al., 2001; Rahman et al., 2004). We pharmacologically and genetically modified the specific dopaminergic signal that occurs during withdrawal and therefore blocked the negative affective component of withdrawal. These results suggest that the specific pattern of DA signaling mediates the aversive motivational response to nicotine withdrawal. The present study suggests that dopaminergic signaling is necessary for the opponent motivational response to nicotine in dependent, but not non-dependent, animals. Further, signaling at the D2R is critical in mediating withdrawal aversions in nicotine-dependent animals. These results suggest that different neurobiological substrates mediate the opponent motivational process for nicotine in drug dependent and non-dependent animals and that the alleviation of nicotine withdrawal primarily may be driving nicotine motivation in dependent animals. These findings may have implications in understanding motivational processes in dependent smokers and may therefore inform targeted drug development in this population. 70

85 Chapter 3 Phasic D1 and Tonic D2 Dopamine Receptor Signaling Double Dissociate the Motivational Effects of Acute Nicotine and Chronic Nicotine Withdrawal Taryn E. Grieder, Olivier George, Huibing Tan, Susan R. George, Bernard Le Foll, Steven R. Laviolette, Derek van der Kooy This chapter is adapted from the paper published in Proceedings of the National Academy of Sciences, vol. 109, issue 8, p , Reprinted with permission. 71

86 Abstract Nicotine, the main psychoactive ingredient of tobacco smoke, induces negative motivational symptoms during withdrawal that contribute to relapse in dependent individuals. The neurobiological mechanisms underlying how the brain signals nicotine withdrawal remain poorly understood. Using electrophysiological, genetic, pharmacological and behavioral methods, we demonstrate that tonic but not phasic activity is reduced during chronic nicotine withdrawal in VTA dopaminergic neurons, and that this pattern of signaling acts through D2Rs and adenosine A2ARs, but not D1Rs. Selective blockade of phasic DA activity prevents the expression of conditioned place aversions to a single injection of nicotine in nondependent mice, but not to withdrawal from chronic nicotine in dependent mice, suggesting a shift from phasic to tonic dopaminergic mediation of the conditioned motivational response in nicotine dependent and withdrawn animals. Either increasing or decreasing activity at D2Rs or A2ARs prevents the aversive motivational response to chronic nicotine withdrawal, but not to acute nicotine. Modification of D1R activity prevents the aversive response to acute nicotine, but not to nicotine withdrawal. This double dissociation demonstrates that the specific pattern of tonic dopaminergic activity at D2Rs is a key mechanism in signaling the motivational effects experienced during nicotine withdrawal, and may represent a novel target for therapeutic treatments for nicotine addiction. 72

87 Introduction Tobacco addiction is the leading avoidable cause of disease and premature death in North America (Fellows et al., 2002). Of more than 3000 chemicals present in tobacco smoke, nicotine is the main psychoactive ingredient responsible for tobacco addiction (Koob and Le Moal, 2006). Withdrawal from nicotine is hypothesized to represent a powerful source of negative reinforcement that drives relapse and compulsive tobacco use (George et al., 2007), therefore understanding the neurobiological substrates mediating the motivational properties of nicotine withdrawal is an important step in the development of new treatments for nicotine addiction. Current hypotheses suggest that nicotine withdrawal leads to a decrease in DA signalling in the brain (Kalivas and Volkow, 2005) and that DA neurons exhibit two activity states, phasic and tonic, that mediate separate aspects of goal-directed behaviour (Floresco et al., 2003; Goto and Grace, 2005; Grace, 2000). However, the role of these two activity states in the motivational effects of nicotine is unknown. Nicotine acutely produces both aversive and positive motivational effects (Grunberg, 1994; Laviolette and van der Kooy, 2004; Sellings et al., 2008) by activating the mesolimbic DA system (Grace, 2000; Nestler, 2005) as well as non-dopaminergic neural substrates (Fowler et al., 2011; Laviolette et al., 2003). DA neurons exhibit burst and population firing activity that leads to phasic and tonic DA release, respectively (Floresco et al., 2003; Goto and Grace, 2005; Grace, 2000). Burst firing produces a fast and large DA release that mainly activates postsynaptic D1Rs, while population firing produces a slower tonic DA release that mainly activates the higher affinity (Hikida et al., 2010) D2Rs (Floresco et al., 2003; Goto and Grace, 2005). The phasic and tonic activities of DA neurons are thought to mediate different aspects of goal-directed behavior; phasic activity facilitates cue-reward association and acquisition of incentive salience, whereas tonic activity is involved in response inhibition and behavioral flexibility (Floresco et al., 2003; Zweifel et al., 2009; Goto and Grace, 2005). Consistent with its motivational properties, a single systemic nicotine injection increases phasic activity in the VTA (Mameli-Engvall et al., 2006) and the release of DA in the ventral striatum (Zhang et al., 2009; Tan et al., 2009), while chronic exposure to nicotine decreases tonic but not phasic DA activity in the VTA (Tan et al., 2009). However, the role of tonic activity and the D2R vs. phasic activity 73

88 and the D1R in signaling the motivational effects of both acute nicotine and chronic nicotine withdrawal is unknown. Here we tested whether withdrawal from chronic nicotine differentially affects tonic and phasic dopaminergic activity in the VTA and whether the specific pattern of signaling through D1Rs and D2Rs mediates the expression of the aversive motivational responses to acute nicotine and chronic nicotine withdrawal. Materials and methods Animals All animal use procedures were approved by the University of Toronto Animal Care Committee, in accordance with the Canadian Council on Animal Care guidelines. Adult male Wistar rats and C57BL/6 mice (Charles River, Montreal, Canada) and D1R, D2R and A2AR knockout mice were housed in a temperature-controlled room with lights on from 6 AM to 8 PM. Heterozygous twelfth-generation D1R and fifth-generation D2R breeder mice were received as a gift from DK Grandy and MJ Low and heterozygous A2AR mice from M Schwarszchild and J Chen. Crosses were bred at the University of Toronto to obtain homozygous male D1R, D2R and A2AR knockout mice and their wild-type controls. Drugs Nicotine hydrogen tartrate salt (Sigma-Aldrich, Ontario) was dissolved in saline at ph 7.0±0.3 and administered via osmotic minipumps (chronic nicotine, 7 mg/kg/day) or s.c. injection (acute nicotine, 1.75 mg/kg). The DAR agonist apomorphine (2.5 mg/kg), DAR antagonist α-flu (0.8 mg/kg), D1R antagonist SCH23390 (0.01 mg/kg), D1R agonist A (0.1 mg/kg), D2R agonist quinpirole (0.05 mg/kg), D2R antagonist eticlopride (1.0 mg/kg), A2AR antagonist SCH58261 (0.5 mg/kg) and A2AR agonist CGS21680 (0.1 mg/kg) were purchased from Sigma-Aldrich, Ontario, dissolved in PBS and administered i.p. at 0, 60, 10, 0, 15, 20, 30 and 20 minutes prior to conditioning, respectively. The NMDA receptor antagonist CGP39551 (2.5 mg/kg) was purchased from Tocris, Missouri, dissolved in PBS and administered i.p. immediately prior to conditioning. Rimonabant (3.0 mg/kg; National Institute 74

89 on Drug Abuse) was suspended in 0.3% Tween80 in saline and administered i.p. 45 minutes prior to conditioning. Additional groups of D1R KO mice and their controls received the D1R agonist A at a dose of 1.0 mg/kg. All doses of drugs are expressed as mg of free base/kg of body weight. Doses and time of injections were selected based on previous studies (Chausmer and Katz, 2002; Fontinha et al., 2009; Grieder et al., 2010; Le Foll and Goldberg, 2004; Natividad et al., 2010; Ralph and Caine, 2005; Sotak et al., 2005). Electrophysiology Rats were subcutaneously implanted with osmotic minipumps (model 2001, Alzet, Cupertino, CA) delivering either saline (non-dependent) or nicotine (nicotine dependent; 3.14 mg/kg/day) for 7-10 days. Nicotine dependent and withdrawn rats had their minipump removed hours prior to electrophysiological recordings, a time that corresponds to peak motivational withdrawal (George et al., 2010; Grieder et al., 2010). Rats were anaesthetized with urethane (1.5 mg/kg, i.p.) and placed in a stereotaxic apparatus. A scalp incision was made and a hole was drilled in the skull overlaying the VTA. Electrodes were pulled from borosilicate glass with average impedance between 6 and 8 MΩ. Microelectrodes were filled with a 2% Pontamine Sky Blue solution and lowered in to the VTA (AP: -5.3 mm caudal to bregma; ML: ± mm; DV: mm ventral to the brain surface) (Tan et al., 2009). Extracellular signals were amplified using a MultiClamp 700B amplifier (Molecular Devices) and recorded through a Digidata 1440A acquisition system (Molecular Devices) using Clamp10 software. Extracellular recordings were typically filtered at 1 khz and sampled at 5 khz. Body temperature of the rats was monitored and maintained at 37 ± 1 by a thermostatically regulated heating pad. DA neurons were identified according to well established electrophysiological features: (I) a relatively longer action potential width (>2.5 ms); (II) a triphasic (+ /-/ +) waveform consisting of a notch on the rising phase followed by a delayed after-potential; (III) a characteristic low tone by audio monitoring; (IV) a slow, irregular or bursting firing pattern, and (V) a spontaneous firing rate of 2-5 Hz or less (Grace and Bunney, 1983; Tan et al., 2009). Phasic bursting activity of DA neurons was defined as the occurrence of two or more consecutive spikes with an interspike interval lower than 80 ms and terminating with an interspike interval greater than

90 ms. Tonic activity was defined as the baseline firing rate (2-5 Hz) of the DA neuron (Tan et al., 2009) and did not include a measure of the number of neurons firing. Place Conditioning The place conditioning apparatus was obtained from Med Associates (SOF-700RA-25 Two Chamber Place Preference Apparatus). One environment was black with a metal rod floor and the other was white with a wire mesh floor. An intermediate gray area housed a removable partition. Each cage was cleaned between animals and each group was fully counterbalanced. Mice were implanted with osmotic minipumps (model 1002; Alzet) or given acute nicotine and pretreated i.p. with saline, apomorphine, α-flu, rimonabant, SCH23390, A-77636, quinpirole, eticlopride, SCH58261, or CGS21680 and conditioned according to modified place-conditioning procedures, as described previously (Grieder et al., 2010) and hereafter. Each cage was cleaned between animals and each group was fully counterbalanced. During preference testing, the dividing partition was removed and mice were given free access to both environments. A single 10-min preference testing session was performed 3 5 days after the last conditioning day, when subjects were drug- and withdrawal-free. All place conditioning and testing was performed between 10:00 AM and 6:00 PM. Nicotine-dependent and -withdrawn mice were conditioned according to modified place conditioning procedure W. Conditioning took place in only the withdrawal-paired environment of the place conditioning apparatus, so that the motivational effects of withdrawal (but not the direct effects of chronic nicotine) were paired with that compartment. Mice were implanted with osmotic minipumps (model 1002; Alzet) that were removed 13 days later. Eight hours after pump removal, when the mouse was experiencing motivational withdrawal from chronic nicotine (Grieder et al., 2010), it was pretreated i.p. with saline, apomorphine (2.5 mg/kg), α-flu (0.8 mg/kg), rimonabant (3 mg/kg), SCH23390 (0.01 mg/kg), A (0.1 or 1.0 mg/kg), quinpirole (0.05 mg/kg), eticlopride (1.0 mg/kg), SCH58261 (0.5 mg/kg), or CGS21680 (0.1 mg/kg) and confined to one of the conditioning environments for 1 h. The difference score for each animal was calculated by subtracting the time spent in the non-paired environment from the time spent in the withdrawal-paired environment. 76

91 For acute nicotine experiments, mice were pretreated i.p. with saline, rimonabant (3 mg/kg), SCH23390 (0.01 mg/kg), A (0.1 or 1.0 mg/kg), quinpirole (0.05 mg/kg), eticlopride (1.0 mg/kg), SCH58261 (0.5 mg/kg), or CGS21680 (0.1 mg/kg) and given a s.c. injection of nicotine (1.75 mg/kg) or saline, and confined immediately to one of the conditioning environments for 1 hour. The difference score for each animal was calculated by subtracting the time spent in the saline-paired environment from the time spent in the nicotine-paired environment. Statistical Analysis Results were analyzed using a one- or two-way ANOVA or Student s t-test with alpha level of 0.05 (two-tailed). In all cases a normality test and equal variance test were performed before the ANOVA to ensure its validity. Post hoc Bonferroni or Duncan s tests were used where appropriate. Data are shown as mean ± SEM. Results Activation or blockade of DA receptors prevents the expression of chronic nicotine withdrawal aversions Pharmacological blockade of DA activity at receptors attenuates the expression of food (Sotak et al., 2005) and drug motivation (Grieder et al., 2010; Laviolette et al., 2002) in place conditioning paradigms. Interestingly, pharmacological activation of DA receptors also prevents food motivation (Sotak et al., 2005) and the expression of conditioned morphine withdrawal aversions (Laviolette et al., 2002). We hypothesized that a specific pattern of signaling at DA receptors could mediate nicotine withdrawal, and thus that either pharmacologically increasing or decreasing activity at DA receptors would prevent the expression of conditioned place aversions to nicotine withdrawal. Mice were given chronic nicotine (7 mg/kg/d) via osmotic minipumps for 13 days and subjected to place conditioning after 8 hours of spontaneous withdrawal (Grieder et al., 2010). The motivational response to a withdrawal-paired environment 77

92 was assessed after pretreatment with saline vehicle, the DA receptor agonist apomorphine (2.5 mg/kg) or the DA receptor antagonist α-flu (0.8 mg/kg). A one-way ANOVA showed a significant effect of pharmacological pretreatment (F 2,42 = 17.1, P < 0.05) (Figure 3.1a), where nicotine dependent and withdrawn mice pretreated with vehicle (n=15) showed a significant aversion to a withdrawal-paired environment (P < 0.05) that was blocked with apomorphine (n=15; P < 0.05) or α-flu (n=15; P < 0.05) pretreatment. Each drug pretreatment had no motivational effects on its own (F 2.26 = 0.26, P > 0.05) (Figure 3.2). Similar to previous results in chronic opiate withdrawn rats (Laviolette et al., 2002), these results suggest that disruption of the specific pattern of dopaminergic signaling by either increasing or decreasing activity at DA receptors prevents the expression of nicotine withdrawal aversions in dependent mice. Tonic but not phasic dopaminergic activity in the VTA is altered in nicotine dependent and withdrawn rats We next directly investigated the specific patterns of DA neuron firing that mediate the motivational response to nicotine withdrawal. Using defined criteria to measure phasic bursting activity and tonic population activity of DA neurons (Grace and Bunney, 1983; Tan et al., 2009), we measured tonic and phasic VTA DA activity with in vivo extracellular single-unit recordings in saline control, previously drug-naive given acute nicotine (1.5 mg/kg), nicotine dependent (3.14 mg/kg/day), and nicotine-dependent and spontaneously withdrawn rats (Grieder et al., 2010; George et al., 2010). Analysis of tonic DA neuron activity with one-way ANOVA revealed a significant effect of drug treatment (F 3,35 = 9.7, P < 0.05) (Figure 3.1b). Saline- treated (n = 11) and acute nicotine-treated (n = 11) rats showed no difference in tonic DA neuron activity (P > 0.05). However, similar to previous studies (Rahman et al., 2004; Tan et al., 2009), nicotine-dependent rats receiving chronic nicotine (n = 11) showed a significant decrease in tonic DA activity in comparison with both saline controls and acute nicotine-treated groups (P < 0.05). Most interesting, rats experiencing withdrawal from chronic nicotine (n = 6) showed a further decrease in tonic DA activity compared with nicotine-dependent rats that were not in withdrawal (P < 0.05). This result is consistent with the hypothesis that dependent human smokers have decreased DA activity during withdrawal (Kalivas and Volkow, 2005), and suggests that the 78

93 Figure 3.1. Phasic DA activity mediates aversions to acute nicotine while the specific pattern of tonic DA activity mediates aversions to withdrawal from chronic nicotine. (A) Both increasing or decreasing DA receptor activity prevents the expression of motivational withdrawal aversions. Nicotine-dependent and -withdrawn mice subjected to place conditioning and pretreated with vehicle showed a conditioned place aversion to the withdrawal-paired environment. This aversive motivational response was blocked in separate groups of mice that were made nicotine-dependent and -withdrawn and pretreated with the DAR agonist apomorphine or DA receptor antagonist α-flu. (B) Electrophysiological recordings of tonic and phasic VTA DA activity during acute and chronic nicotine and withdrawal show that phasic activity is modified after acute nicotine administration while tonic DA activity is modified after chronic nicotine and withdrawal. Top: Representative electrophysiological recordings from VTA DA neurons in rats treated with saline vehicle, acute nicotine, chronic nicotine (nicotine dependent) and chronic nicotine with spontaneous withdrawal. Middle: Acute nicotine did not affect tonic DA firing. Nicotine dependent rats exhibited a decrease in tonic VTA DA activity compared to saline control rats that was further significantly decreased in rats undergoing withdrawal from chronic nicotine (*p < 0.05 in comparison to all other groups). Bottom: Only acute nicotine increased phasic activity in VTA DA neurons. Chronic nicotine exposure and withdrawal from chronic nicotine did not alter phasic activity. (C) Selective blockade of phasic DA activity using antagonist pretreatment prevents conditioned place aversions to acute nicotine, but not to withdrawal from chronic nicotine. Nondependent mice that were given acute nicotine and pretreated with saline showed a conditioned place aversion to the nicotine-paired environment. Selectively blocking phasic dopaminergic activity at receptors with the cannabinoid-1 inverse agonist rimonabant or the NMDA receptor antagonist CGP39551 prevented the expression of the conditioned aversive response to acute nicotine. Nicotinedependent and -withdrawn mice pretreated with saline showed conditioned place aversions to the withdrawal-paired environment that were not blocked with rimonabant or CGP39551 pretreatment. Separate groups of mice given acute or chronic saline and pretreated with saline, rimonabant or CGP39551 showed no motivational response to any pretreatment, suggesting that 79

94 at the doses used in the present study, these drugs have no motivational effect on their own. Data represent mean ± SEM (*p < 0.05). 80

95 Figure 3.2. The DA receptor agonist and antagonist have no motivational effects on their own. Separate groups of mice given chronic saline in minipumps were place conditioned after pretreatment with saline (n=8), apomorphine (n=14) or α-flu (n=7). The saline-treated group showed no motivational response to the novel environment. The apomorphine or α-flu pretreated groups showed no motivational response to the drug-paired environment. These results suggest that novelty plays no role in the motivational responses demonstrated by mice in this study, and that apomorphine and α-flu have no motivational effects on their own at the concentrations used in the present study. Data represent mean ± SEM Difference score Saline DA receptor agonist DA receptor antagonist Pretreatment 81

96 aversive motivational state of spontaneous nicotine withdrawal is signaled by a further patterned decrease in tonic DA activity than that observed during the nicotine-dependent state. Analysis of phasic VTA DA activity with one-way ANOVA revealed a significant effect of drug treatment (F 3,35 = 5.0, P < 0.05) (Figure 3.1b). Acute nicotine increased phasic DA firing rates (P < 0.05), in comparison with nondependent, dependent, and nicotine-dependent and -withdrawn groups (all P > 0.05), suggesting that the specific pattern of phasic activity may mediate the motivational response to acute nicotine. Blockade of phasic DA activity prevents aversions to acute nicotine, but not to withdrawal from chronic nicotine To test if phasic DA activity directly mediates the aversive response to acute nicotine, but not to chronic nicotine withdrawal, we examined the effect of blocking phasic DA activity on conditioned place aversions for acute nicotine and withdrawal from chronic nicotine using the cannabinoid receptor-1 inverse agonist rimonabant (3.0 mg/kg) and the NMDA receptor antagonist CGP39551 (2.5 mg/kg). Previous studies suggested that rimonabant blocks phasic DA release without affecting baseline DA transients (Cheer et al., 2000; Cheer et al., 2007; Cohen et al., 2002); however, these were performed in vitro (Cheer et al., 2000) or measured the absolute amount of DA release using voltammetry (Cheer et al., 2007) or microdialysis (Cohen et al., 2002). We thus performed in vivo electrophysiological recordings of VTA DA neurons in drugnaive rats given rimonabant (n = 10) to test the hypothesis that the drug selectively decreases phasic but not tonic baseline DA firing. Rimonabant significantly decreased phasic DA activity (t 9 = 2.715, P < 0.05) but not tonic DA activity (t 9 = , P > 0.05) (Figure 3.3). CGP39551 is another pharmacological tool that selectively disrupts phasic DA activity without affecting tonic activity and blocks nicotine-induced VTA DA bursting (Schilström et al., 2004). A two-way ANOVA showed a significant interaction of pharmacological pretreatment and nicotine history (F 6,110 = 4.291, P < 0.05) (Figure 3.1c). Nondependent mice given acute nicotine after saline pretreatment (n = 9) showed a conditioned place aversion to a nicotine-paired environment (P < 0.05) that was blocked with rimonabant (n = 9; P > 0.05) or CGP39551 (n = 13; P > 0.05) 82

97 Figure 3.3. The cannabinoid-1 receptor inverse agonist rimonabant significantly decreases phasic VTA DA activity but does not affect tonic DA activity. Rats were given rimonabant and their VTA DA neuronal activity was recorded electrophysiologically. A representative trace from a rat VTA DA neuron is depicted. Rimonabant significantly decreased the amount of bursts per minute (28.0±5.06 to 22.9±4.96) but did not decrease tonic DA activity (4.34±0.61 to 4.50±1.22). 83

98 pretreatment. In contrast, nicotine-dependent and withdrawn mice given saline (n = 11) showed a conditioned place aversion to the withdrawal-paired environment (P < 0.05) that was not blocked by rimonabant (n = 13; P < 0.05) or CGP39551 (n = 12, P < 0.05). Mice given chronic or acute saline and saline (n = 14), rimonabant (n = 7), or CGP39551 (n = 6) showed no motivational response to the drugs (P > 0.05). These results suggest that phasic DA activity is required for the motivational response to acute nicotine, but not to withdrawal from chronic nicotine. Genetic deletion of D2Rs versus D1Rs double dissociate chronic versus acute nicotine motivation Phasic and tonic DA signalling act through D1Rs and D2Rs, respectively (Floresco et al., 2003; Goto and Grace, 2005), and we demonstrated here that acute nicotine modifies phasic DA activity, but withdrawal from chronic nicotine modifies tonic DA activity. We thus hypothesized that genetic deletion of the D2R would prevent aversions to nicotine withdrawal in dependent mice, but D1R deletion would prevent acute nicotine aversions in non-dependent mice. D1R and D2R KO mice and their WT littermates were place-conditioned after receiving acute nicotine (1.75 mg/kg) or during spontaneous withdrawal from chronic nicotine (7 mg/kg/d). One-way ANOVA revealed a significant group effect (F 2,24 = 3.43, P < 0.05) (Figure 3.4a) in dependent and withdrawn mice. WT mice in withdrawal from chronic nicotine (n = 11) showed a conditioned place aversion to a withdrawal-paired environment (P < 0.05); aversions that were shown as well by D1R KO mice (n = 9) but not by D2R KO mice (n = 7; P > 0.05). For acute nicotine-treated mice, one-way ANOVA revealed a significant group effect (F 2,27 = 8.27, P < 0.05) (Figure 3.4b). D2R KO (n = 7) and previously drug-naive WT mice (n = 14) given acute nicotine showed a significant conditioned place aversion to a nicotine-paired environment (P < 0.05) that was blocked in D1R KO mice (n = 9; P > 0.05). Taken together, these results doubly dissociate the role of D1Rs and D2Rs in nicotine motivation; D2Rs (but not D1Rs) are required for the aversive motivational response to withdrawal in nicotine-dependent mice, and D1Rs (but not D2Rs) are necessary for acute nicotine aversions in non-dependent mice. 84

99 Figure 3.4. A specific pattern of signaling at D1Rs is required for aversions to acute nicotine in nondependent mice, while a specific pattern of D2R activity is required for aversions to nicotine withdrawal in dependent mice. (A) WT mice that are made nicotine-dependent and -withdrawn and place conditioned in the W procedure will show a conditioned place aversion to the withdrawal-paired environment. This aversive motivational response to chronic nicotine withdrawal is observed in a group D1R KO mice, but is not shown by D2R KO mice. These results suggest that D2Rs but not D1Rs are required for the expression of the motivational response to nicotine withdrawal. (B) WT nondependent mice given acute nicotine prior to place conditioning will demonstrate an aversive motivational response to the nicotine-paired environment. These conditioned place aversions to acute nicotine are blocked in D1R KO mice but are shown by D2R KO mice, suggesting that D1Rs but not D2Rs are required for the expression of acute nicotine aversions. (C) Nicotinedependent and -withdrawn mice pretreated with vehicle show a conditioned place aversion to the withdrawal-paired environment that is blocked with D2R antagonist eticlopride and D2R agonist quinpirole pretreatment, but not with D1R antagonist SCH23390 or D1R agonist A pretreatment. These results suggest that either increasing or decreasing D2R but not D1R activity will prevent the expression of nicotine withdrawal aversions in dependent mice. (D) Nondependent mice given an injection of nicotine and pretreated with vehicle showed an aversion to the nicotine-paired environment that was blocked with A and SCH23390 pretreatment, but not with quinpirole or eticlopride pretreatment, suggesting that either increasing or decreasing D1R but not D2R activity prevents acute nicotine aversions. Data represent mean ± SEM (*p<0.05). 85

100 86

101 Pharmacological manipulations of D2Rs but not D1Rs block withdrawal aversions in nicotinedependent mice Modification of the specific pattern of activity at DA receptors, and D2R but not D1R deletion, blocked the aversive response to withdrawal from chronic nicotine. We thus hypothesized that either increasing or decreasing activity at D2Rs but not D1Rs would block conditioned withdrawal aversions in dependent mice. We examined the effect of the D1R agonist A (1.0 mg/kg), the D1R antagonist SCH23390 (0.01 mg/kg), the D2R agonist quinpirole (0.05 mg/ kg), and the D2R antagonist eticlopride (1.0 mg/kg) on conditioned place aversions to withdrawal. A one-way ANOVA revealed a significant effect of pharmacological pretreatment (F 4,60 = 4.11, P < 0.05) (Figure 3.4c). Nicotine-dependent and -withdrawn mice that received vehicle (n = 31) before conditioning showed an aversion to a withdrawal- paired environment (P < 0.05) that was blocked in mice that received quinpirole (n = 7; P > 0.05) and eticlopride (n = 7; P > 0.05), but not in mice that received A (n = 12; P < 0.05) or SCH23390 (n = 8; P < 0.05). No motivational response to any of the drugs on their own was observed (F 4,44 = 0.08, p>0.05) (Figure 3.5a). However, groups of mice tested with a higher dose of the D1R agonist A (10.0 mg/kg) showed a nonspecific block of learning (F 1,25 = 15.08, P < 0.05) (Figure 3.5b,c). These results suggest that either increasing or decreasing D2R activity blocks the specific pattern of signaling that mediates the aversive motivational response to chronic nicotine withdrawal, and that activity at D1Rs is not required for the experience of nicotine withdrawal aversions. Pharmacological manipulations of D1Rs but not D2Rs block aversions to acute nicotine in nondependent mice We next tested the hypothesis that either increasing or decreasing activity at D1Rs but not D2Rs would prevent the aversive motivational response to acute nicotine by examining the effect of A-77636, SCH23390, quinpirole, and eticlopride on conditioned place aversions to a nicotinepaired environment in nondependent mice. There was a significant effect of pharmacological pretreatment on acute nicotine aversions (F 4,44 = 4.99 P < 0.05) (Figure 3.4d) that occurred exactly opposite to dependent and withdrawn mice. Nondependent mice given acute aversive 87

102 Figure 3.5. D1R and D2R agonists and antagonists have no motivational effects on their own at the doses used in this study, however a high dose of the D1R agonist prevents learning. (A) Separate groups of mice given chronic saline in minipumps and place conditioned after pretreatment with saline vehicle (n=14), A (n=7), SCH23390 (n=10), quinpirole (n=8) or eticlopride (n=10) showed no motivational response to any of the drugs. These results suggest that there was no motivational response to novelty (in the saline-pretreated group) and that the D1R and D2R agonists and antagonists have no motivational effects on their own at the concentrations used in the present study. However, a higher dose of the D1R agonist A prevents learning in nicotine-dependent and -withdrawn WT or D1R KO mice (B) and in nondependent mice given acute nicotine (C). Nicotine-dependent and -withdrawn mice pretreated with saline showed an aversive motivational response to the withdrawal-paired environment that was not observed in WT or D1R KO mice pretreated with A at 10.0 mg/kg. Previously drug-naive mice given acute nicotine and pretreated with saline show an aversion to an acute nicotine-paired environment that is blocked in mice pretreated with A at 10.0 mg/kg. These results suggest that A at a dose of 10.0 mg/kg blocks learning in our paradigm. Data represent mean ± SEM (*p < 0.05). 88

103 a Difference Score Vehicle D2R antagonist Eticlopride Saline minipumps D2R agonist Quinpirole D1R antagonist SCH23390 D1R agonist A b Nicotine dependent and withdrawn Difference Score (s) * WT+ Saline WT + A D1KO + Saline * D1KO + A c Difference score (s) Acute nicotine * Saline pretreated A pretreated 89

104 nicotine and pretreated with vehicle (n = 17) showed a conditioned place aversion to a nicotinepaired environment (P < 0.05) that was blocked in mice pretreated with the D1R agonist A (n = 8; P > 0.05) and the D1R antagonist SCH23390 (n = 10; P > 0.05), but not with the D2R agonist quinpirole (n = 7; P < 0.05) or the D2R antagonist eticlopride (n = 7; P < 0.05). These results demonstrate that either increasing or decreasing activity at D1Rs blocks aversions to acute nicotine, and suggest that D2R activation is not necessary for the aversive response to acute nicotine. Pharmacological and genetic modifications of A2ARs specifically block chronic nicotine withdrawal aversions In the striatum, A2ARs and D2Rs are colocalized (Tozzi et al., 2011) and form A2AR- D2R heteromers (Fuxe et al., 2010). The A2AR and D2R interact antagonistically, such that agonism of A2ARs decreases signaling at D2Rs (Tanganelli et al., 2004) and antagonism of A2ARs increases signaling at D2Rs (Fuxe et al., 2010). If the specific pattern of activity at D2Rs is a key factor in mediating aversions to nicotine withdrawal, and colocalized A2ARs and D2Rs act antagonistically in the striatum (Fuxe et al., 2010; Tozzi et al., 2011), then genetic and pharmacological manipulation of A2ARs should also affect nicotine withdrawal aversions in dependent animals. We examined the effect of A2AR manipulation on the conditioned aversive responses to acute nicotine and withdrawal from chronic nicotine in A2AR KO mice and WT mice pretreated with the A2AR agonist CGS21680 (0.1 mg/kg) or the A2AR antagonist SCH58261 (0.5 mg/kg). One-way ANOVA revealed a significant effect of A2AR manipulation in nicotine-dependent and -withdrawn mice (F 3,51 = 6.2, P < 0.05) (Figure 3.6a) but not in nondependent mice given acute nicotine (F 2,24 = 0.06, P > 0.05) (Figure 3.6b). Dependent and withdrawn WT mice that received vehicle (n = 23) showed a conditioned place aversion to a withdrawal-paired environment (P < 0.05) that was blocked in A2AR KO mice (n = 14; P > 0.05) and in WT mice that received CGS21680 (n = 7; P > 0.05) or SCH58261 (n = 11; P > 0.05). Previously drug-naive mice given acute nicotine and pretreated with vehicle (n = 13) showed a conditioned place aversion to a nicotine-paired environment that was not blocked in mice pretreated with CGS21680 (n = 7; P > 0.05) or SCH58261 (n = 7; P > 0.05). No 90

105 Figure 3.6. Manipulations of the A2AR block the aversive response to withdrawal from chronic nicotine but not acute nicotine. (A) Nicotine-dependent and -withdrawn WT mice pretreated with vehicle and place conditioned showed an aversive motivational response to the withdrawal-paired environment. These conditioned place aversions were blocked in groups of A2AR KO mice, in mice pretreated with the A2AR agonist CGS21680 and in mice pretreated with the A2AR antagonist SCH These results suggest that any modification of the A2AR will prevent the expression of the aversive motivational response to nicotine withdrawal in dependent mice. (B) Nondependent mice given an acute injection of nicotine and pretreated with vehicle prior to place conditioning showed conditioned place aversions to the nicotine-paired environment that were not blocked with CGS21680 or SCH58261 pretreatment. These results suggest that the A2AR does not play a role in mediating the aversive motivational response to acute nicotine in nondependent mice. (C) Mice given chronic or acute saline in minipumps and pretreated with vehicle (n=6), CGS21680 (n=7) or SCH58261 (n=7) showed no motivational response to the drugs, suggesting that the A2AR agonist and antagonist have no motivational effects on their own at the doses used in the present study. Data represent mean ± SEM (*p<0.05). 91

106 Difference score (s) Difference score (s) A B * Dependent and Withdrawn WT + Saline A2A KO + Saline Saline Acute nicotine A2AR agonist CGS21680 WT + A2A agonist CGS21680 A2AR antagonist SCH58261 WT + A2A antagonist SCH58261 C Saline minipumps Difference Score (s) Saline A2A agonist CGS21680 A2A antagonist SCH

107 motivational response to the drugs on their own was observed (F 2,17 = 0.15, P > 0.05) (Figure 3.6c). These results suggest that either increasing or decreasing activity at A2ARs blocks aversions to withdrawal from chronic nicotine but not the aversive response to acute nicotine, possibly via modification of D2R activity. Discussion Withdrawal from nicotine has been hypothesized to represent a powerful source of negative reinforcement (Grieder et al., 2010; George et al., 2010) that drives relapse and compulsive tobacco use (George et al., 2007; Koob and Le Moal, 2006). Therefore, understanding the neurobiological substrates mediating the motivational properties of withdrawal from chronic nicotine is an important step in the development of new treatments for nicotine addiction. Previous reports have suggested that a neurobiological switch occurs during the transition from a drug-naive to a drug-dependent motivational state (Vargas-Perez et al., 2009). The transition from acute nicotine use to nicotine dependence has been hypothesized to result from neuroadaptive changes that produce the powerful withdrawal syndrome and negative emotional state observed upon cessation of nicotine use (George et al., 2007). The present results demonstrate that a shift in VTA DA signaling from phasic to tonic, and of receptor mediation from D1 to D2, occurs upon dependence and withdrawal from nicotine, and doubly dissociates the role of D1Rs vs. D2Rs in nicotine motivation. We suggest that phasic DA activity at D1Rs mediates acute nicotine aversions, whereas tonic DA activity at D2Rs (and indirectly, A2ARs) mediates aversions to withdrawal from chronic nicotine. Rodents experiencing spontaneous withdrawal from chronic nicotine show a conditioned place aversion to a withdrawal-paired environment in place conditioning paradigms (Grieder et al., 2010; Merritt et al., 2008). Similarly, previously drug-naive mice given a single aversive dose of nicotine will show a conditioned place aversion to the nicotine-paired environment (Grieder et al., 2010). The important difference in these two effects is the exposure to nicotine: acute exposure in nondependent animals versus chronic exposure and withdrawal in dependent 93

108 animals. The present results demonstrate that modifying activity at D2Rs prevented the expression of nicotine-withdrawal aversions in dependent animals. A previous study suggested that both increasing or decreasing DA signaling at DA receptors blocked the expression of conditioned place aversions to withdrawal from chronic morphine (Laviolette et al., 2002). Our present work suggests that a similar phenomenon occurs during nicotine withdrawal in dependent animals, such that treatment with a broad-spectrum DA receptor agonist or antagonist, or a specific D2R agonist or antagonist, prevented the expression of nicotine withdrawal aversions in dependent mice. Furthermore, genetic deletion of the D2R (but not the D1R) prevented nicotine withdrawal aversions. In nondependent animals exposed to acute nicotine, exactly the opposite phenomenon occurred: Both D1R-specific agonism or antagonism, as well as D1R deletion, selectively prevented aversions to acute nicotine in nondependent mice, without affecting aversions to withdrawal in nicotine-dependent mice. These results doubly dissociate the role of the D1R versus the D2R in nicotine motivation, such that the motivational response to withdrawal in dependent mice is D2R-mediated and acute nicotine motivation is D1R-mediated. These results are in line with previous studies showing that drug-dependent human subjects have marked decreases in D2R availability (Fehr et al., 2008) and presumably in DA release (Volkow et al., 2009), which is consistent with the hypothesis that a pattern of DA activity signals nicotine motivation and the present results showing that nicotine-dependent and -withdrawn mice have a decrease in tonic activity of VTA DA neurons. Furthermore, animal studies have shown that D1R antagonism blocks nicotine motivation in nondependent mice (David et al., 2006). A recent study showed that blockade of D1R but not D2R transmission prevented acquisition of opiatereward memory in nondependent rats, and D2R but not D1R blockade prevented opiate-reward encoding in dependent and withdrawn rats (Lintas et al., 2011). However, previous studies suggest that both acute nicotine and opiate reward are mediated by the non-dopaminergic brainstem TPP nucleus (Laviolette and van der Kooy, 2004; Laviolette et al., 2002), and thus must involve separate cells in the TPP that are thought to mediate burst-firing of VTA DA neurons (Floresco et al., 2003); burst-firing that we show here is involved in the response to acute nicotine. The present data show that only the acute aversive motivational effects of nicotine are mediated by D1Rs, leading to the suggestion that the induction of nicotine dependence switches the neurobiological substrate mediating the aversive motivational effects of 94

109 nicotine from D1R to D2R-mediated. We have suggested that a specific pattern of tonic DA activity through D2Rs signals withdrawal from chronic nicotine. The D2R system is important for learning to shift behavior in response to change in motivation (Goto and Grace, 2005). It is thus possible that animals have a tonic pattern of DA activity that does not shift with nicotine dependence and withdrawal. D2R KO mice did not demonstrate a conditioned aversive response to withdrawal, possibly because these mice never experience a change in tonic DA activity that signals withdrawal. However, this block of the motivational response is not simply because of an effect on learning, as both D1R and D2R KO mice can learn a motivational response to nicotine in our paradigm. Indeed, both hyperdopaminergic and hypodopaminergic mice can learn various tasks although their motivation is altered (Zweifel et al., 2009), suggesting that DA mediates motivation rather than learning. Our block of nicotine withdrawal aversions with both D2R agonist and antagonist drugs provides further support for the hypothesis that the specific pattern of DA release at D2Rs signals withdrawal, and that any deviation from this pattern, whether an increase or a decrease of DA activity and release, at receptors will prevent the aversive motivational response to nicotine withdrawal. The present results doubly dissociate the role of phasic and tonic dopaminergic activity in the motivational response to acute nicotine in nondependent mice and to withdrawal in nicotine dependent mice. Previous studies have shown that tonic DA activity is decreased in nicotine dependent animals (Tan et al., 2009) and that precipitated withdrawal from chronic nicotine leads to decreased DA levels (Natividad et al., 2010). Using defined electrophysiological methods to measure VTA DA activity (Grace and Bunney, 1983; Tan et al., 2009), we confirmed and extended these results to dependent animals experiencing spontaneous withdrawal, showing that tonic DA activity is further decreased during withdrawal from chronic nicotine. Acute nicotine increased phasic DA activity in nondependent animals; pharmacologically blocking phasic activity via cannabinoid-1 (Cheer et al., 2000; Cheer et al., 2007; Cohen et al., 2002) or NMDA (Schilström et al., 2004) receptor modulation prevented the aversive motivational response to acute nicotine, but not to withdrawal from chronic nicotine. Taken together, these results suggest that nicotine withdrawal is signaled by a pattern of tonic but not phasic DA 95

110 activity, and that there is a decrease in tonic DA release during withdrawal in dependent animals. Although the amount of DA released via tonic neuronal activity is small in comparison with that via phasic activity, a previous study showed that tonic DA activity is independent of burst firing and provides sufficient DA to engage behavior (Zweifel et al., 2009), thus it is plausible that a tonic DA signal mediates the behavioral response to nicotine withdrawal. A single injection of nicotine leads to the large-scale phasic release of DA (Rice and Cragg, 2004; Zhang et al., 2009); therefore, it is possible that nicotine-dependent subjects who are experiencing withdrawal may take nicotine to temporarily modulate DA levels in the brain by increasing release through phasic activation of VTA DA neurons. This hypothesis is similar to Grace s tonic/phasic model of DA system regulation (Grace, 2000). Another possibility suggested by the present results is that acute nicotine floods the DA system in a similar fashion as administration of a broad-spectrum DA receptor agonist. These manipulations of DA activity would modify the specific pattern of DA firing that signals withdrawal, and would thus prevent the aversive motivational effects of withdrawal from chronic nicotine. We suggest that modulation of D2Rs could prevent the motivational effects of nicotine withdrawal; however, directly increasing or decreasing DA activity could potentially produce schizophrenic or Parkinson-like symptoms, respectively. We demonstrate here that both increasing and decreasing activity at adenosine A2ARs blocked nicotine withdrawal aversions in dependent mice but, similar to a previous study (Castañé et al., 2006), had no effect on acute nicotine aversions in nondependent mice. These results suggest that A2AR modulation can prevent the aversive motivational response to nicotine withdrawal, possibly through an indirect disruption of the specific pattern of D2R activity that mediates withdrawal. Furthermore, we hypothesize that tonic and phasic VTA DA activity leads to effects on D1Rs, D2Rs, and A2ARs in the ventral striatum. This idea is supported by a previous study showing that intrastriatal DA receptor antagonism has similar effects to systemic antagonist administration (Laviolette and van der Kooy, 2003). Activation of striatal receptors could in turn feed back to the VTA via direct and indirect pathways (Hikida et al., 2010), and this feedback may be important in the generation of the specific pattern of tonic DA activity that signals nicotine withdrawal aversions. Taken together, our results suggest that a key mechanism signaling nicotine withdrawal is 96

111 tonic activity of VTA DA neurons, which may act through D2Rs and indirectly, A2ARs, to signal an aversive motivational state during withdrawal in dependent subjects that contributes to relapse. Pharmacological manipulation of the tonic DA signal prevents the aversive motivational state that is normally experienced during nicotine withdrawal, suggesting that modifying tonic DA activity via manipulation of D2Rs or possibly A2ARs may represent a unique target for therapeutic treatments of nicotine addiction. 97

112 Chapter 4 Recruitment of a VTA CRF system mediates the aversive effects of nicotine withdrawal Taryn E. Grieder, Hector Vargas-Perez, Candice Contet, Laura A. Tan, John Freiling, Laura Clarke, Elena Crawford, Pascale Koebel, Brigitte L. Kieffer, Paul E. Sawchenko, George F. Koob, Derek van der Kooy and Olivier George This paper is adapted from a version that currently is submitted to Neuron. 98

113 Abstract The CRF system has been hypothesized to counteract the VTA DA-mediated positive rewarding effects of drugs of abuse through upregulation of CRF and activation of CRF1Rs in the extended amygdala. This phenomenon is known as a between-system neuroadaptation that contributes to the transition to drug dependence and to withdrawal. Using animal models of nicotine dependence, rtpcr, in situ hybridization, immunohistochemistry, pharmacology and viral vector approaches, here we show that in addition to between-system neuroadaptation, CRF participates in a within-system neuroadaptation: after chronic exposure to nicotine, CRF mrna is expressed in dopaminergic posterior VTA (pvta) neurons, in the core of the brain reward system. Moreover, upregulation of CRF mrna, CRF release, and activation of CRF1Rs locally in the pvta during withdrawal directly control the motivational state of nicotine withdrawal, thus linking the brain reward and stress systems in the same neurons in the pvta. 99

114 Introduction Drug addiction has been hypothesized to be driven by two mechanisms; downregulation of the brain reward system (Volkow et al., 2007) and upregulation of the anti-reward brain stress system (Koob and Le Moal, 2008), concepts known as within- and between-system neuroadaptations, respectively (Koob and Le Moal. 1997). A prominent downregulation of the mesolimbic DA reward system originating in the VTA, and upregulation of the CRF brain stress system originating in the extended amygdala has been observed in rodents, non-human primates and humans during abstinence from drugs of abuse (Koob and Volkow, 2010). Several groups have examined the mechanisms behind this between-system DA and CRF neuroadaptation (Hahn et al., 2009; Lodge and Grace, 2005; Wanat et al., 2008; Wang et al., 2007; Wang and Morales, 2008), but how these two systems interact in drug dependence and withdrawal is essentially unknown. Here we tested the hypothesis that CRF participates in a within-system neuroadaptation during nicotine dependence by examining CRF expression in the core of the brain reward system, in dopaminergic neurons of the pvta, and assessed whether this neuroadaptation could directly control the motivational state of nicotine withdrawal. Materials and methods Animals All animal use procedures were approved by the University of Toronto Animal Care Committee in accordance with the Canadian Council on Animal Care guidelines and The Scripps Research Institute Institutional Animal Care and Use Committee and were in accordance with National Institutes of Health guidelines. Adult male C57BL/6 mice (Charles River, Montreal, Canada, or Hollister, CA, USA) were housed in a temperature-controlled room with lights on from 7:00 AM to 7:00 PM. Drugs 100

115 Nicotine hydrogen tartrate salt (Sigma-Aldrich, Ontario, Canada) was dissolved in saline, ph 7.0 ± 0.4, and administered via osmotic minipumps (chronic nicotine, 7 mg/kg/day, minipump model 1002, Alzet, Cupertino, CA, USA) or subcutaneous injection (acute nicotine, 1.75 mg/kg). Nicotine-dependent and -withdrawn mice had their minipumps removed 8 h prior to experimentation at a time that corresponded to peak motivational withdrawal (Grieder et al., 2010). The CRF1R antagonist MPZP was synthesized at The Scripps Research Institute (Richardson et al., 2008), dissolved in HBC, and administered subcutaneously 20 min prior to conditioning or at a concentration of 0.14 µg/0.3 µl over 10 min for intra-vta infusions (coordinates: AP-3.3, DV-4.4, ML±0.5). Viral vector production shrna-encoding AAV2 vectors that target the CRF transcript were generated using the same procedure as described by Darcq et al. (2011). A shrna sequence (shcrf, sense strand 5 -GGATCTCACCTTCCACCTTCT-3 ) predicted to have high silencing efficiency was selected using Block-iT RNAi Designer (Life Technologies, Carlsbad, CA). A universal scramble shrna with no homology to any transcripts was used as a control (shscr, sense strand 5 -GCGCTTAGCTGTAGGATTC-3 ). An AAV2 shuttle plasmid that encodes shcrf or shscr downstream of mu6 promoter and enhanced green fluorescent protein (EGFP) under the control of the cytomegalovirus (CMV) promoter 3 -flanked by a β-globin intron was generated using Invitrogen Gateway technology (see Figure 4.4b). Helper-free AAV2 particles were produced by the triple transfection of AAV-293 cells (Agilent Technologies Inc., Santa Clara, CA) with the AAV2 shuttle plasmid described above, a plasmid that contains AAV2 rep and cap genes, and a plasmid that encodes the adenovirus helper functions. Two days later, cells were collected and lysed by three freeze-thaw cycles, treated with benzonase, and clarified by centrifugation. Viral vectors were purified by iodixanol gradient ultracentrifugation (Zolotukhin, et al., 2002), followed by dialysis and concentration against Dulbecco phosphate-buffered saline (PBS) using centrifugal filter units (Millipore, Billerica, CA). Genomic units were quantified by rtpcr. The titers were GU/ml. 101

116 Real-time PCR RNA was isolated using a Qiagen RNeasy extraction kit (Crh, Mm _s1) with DNase to remove genomic DNA contamination, and a specified amount of cdna was reversetranscribed using SuperscriptIII (Invitrogen, Foster City, CA). Quantitative PCR was performed using Taqman Gene Expression Assays for CRF in a 7900HT Fast Real-Time PCR System (both from Life Technologies, Carlsbad, CA). Quantification was performed using the delta C t method with Hprt or GAPDH as an endogenous control. Radioactive in situ hybridization The mice were anesthetized with chloral hydrate (350 mg/kg, i.p.) and perfused via the ascending aorta with 0.9% saline followed by ice-cold 4% paraformaldehyde in 0.1 M borate buffer, ph 9.5. The brains were removed, postfixed for 3 h, and cryoprotected in 20% sucrose in 0.1 M phosphate buffer overnight at 4 C. Five one-in-five series of 30 µm-thick frozen coronal sections were cut, collected, and stored in 30% ethylene glycol and 20% glycerol in 0.1 M phosphate buffer at -20 C until processing. In situ hybridization was performed using 35 S-labeled sense (control) and antisense crna probes labeled to similar specific activities using a fulllength (1.2 kb) probe for mrna that encodes CRF (1.2 kb; Dr. K. Mayo, Northwestern University, Evanston, IL). Sections were mounted on Superfrost plus slides and dried under vacuum overnight. They were postfixed with 10% paraformaldehyde for 30 min at room temperature, digested with 10 µg/ml proteinase K for 15 min at 37 C, and acetylated for 10 min. The probes were labeled to specific activities of dpm/µg and applied to the slides at concentrations of ~10 7 cpm/ml overnight at 56 C in a solution that contained 50% formamide, 0.3 M NaCl, 10 mm Tris, 1 mm EDTA, 0.05% trna, 10 mm dithiothreitol, 1x Denhardt s solution, and 10% dextran sulfate, after which they were treated with 20 µg/ml of ribonuclease A for 30 min at 37 C and washed in 15 mm NaCl/1.5 mm sodium citrate with 50% formamide at 70 C. The slides were then dehydrated and exposed to X-ray film (Kodak Biomax MR, Eastman Kodak, Rochester, NY, USA) for 18 h. They were coated with Kodak NTB-2 liquid emulsion and exposed at 4 C for 3-4 weeks as determined by the strength of the signal on film. The slides 102

117 were developed with Kodak D-19 and fixed with Kodak rapid fixer. One series of sections that adjoined those used for analysis was stained with thionin to facilitate the accurate localization of hybridization signals. Densitometry The semiquantitative densitometric analysis of hybridization signals for CRF mrna was performed on emulsion-dipped slides. Photomicrographs were captured using a Leica light microscope with a Hamamatsu Orca charge-coupled device camera through OpenLab software (version 3.1.5) and analyzed using ImageJ software. The optical densities of hybridization signals were determined under dark-field illumination at 400 magnification with a circular ROI with a 20 µm diameter that was placed over individual neurons. The size of the ROI was chosen on the basis of the average diameter of TH-immunoreactive neurons in the pvta. The sections were analyzed at regular 160 µm intervals across the pvta. Optical densities were corrected for the average background signal that was determined by sampling 20 cell-sized areas per section in non-signal areas adjacent to the pvta. Optical density values are expressed in gray scale values of 1 to 256, corresponding to a gradation from low to high absorbance, respectively. Neurons on both sides of the brain were pooled for analysis to calculate the animal mean. Animal means were then grouped according to virus treatment, averaged, and statistically analyzed. Immunohistochemistry The mice were anesthetized with 3% halothane, pre-perfused transcardially with a solution of 0.5 ml heparin/100 ml saline for 1-2 min and then perfused with a solution of 4% paraformaldehyde in 0.1M phosphate buffer (PB), ph 7. The brains were removed from the skull, post-fixed at 4 C in the perfusate solution for 6-18 h, rinsed in several changes of PBS that contained 20% sucrose, and stored at 4 C in fresh 20% sucrose/pbs that contained 0.1% sodium azide. Coronal cryostat sections (40 µm) were obtained using a Reichert Jung cryostat. The sections were collected in strict anatomical order in a one-in-four series and stored at 4 C in PBS 103

118 0.1% azide prior to processing. The sections were incubated free-floating with shaking in multiwell plates, and all incubations were performed at room temperature unless otherwise specified. The samples were incubated for 20 min in 1% hydrogen peroxide/pbs to quench endogenous peroxidases, rinsed several times in PBS, and exposed to a blocking solution that contained PBS/Triton-X100 (0.3%), 1 mg/ml BSA, and 5% normal donkey serum (Jackson Immuno Research, West Grove, PA, USA) for a minimum of 60 min. The sections were incubated overnight at 4 C in anti-crf purified goat polyclonal antibody (Santa Cruz Biotechnology, Santa Cruz, CA, USA) diluted 1:200 in PBS, 0.5% TWEEN 20, and 5% normal donkey serum. Control sections were incubated in the antibody diluent. Following three rinses of 10 min each in PBS, the sections were incubated in Vector ImmPRESS Goat (Vector Labs, Burlingame, CA, USA) for 1 h, rinsed in PBS as above, and reacted with a DAB Substrate Kit (Vector Labs). The sections were monitored under a microscope to determine the optimal reaction time. The reaction was stopped in PBS. The sections were mounted on coated slides, air dried, dehydrated through a series of ethanol and xylene, and coverslipped with Permount. All brightfield photographs for analysis were taken with a Q Imaging Retiga 2000R color digital camera mounted on a Zeiss Axiophot microscope. Combined TH immunohistochemistry and CRF in situ hybridization Brains were snap-frozen using isopentane. Twenty µm cryostat sections were mounted onto Superfrost Plus slides. A digoxigenin (DIG)-labeled CRF riboprobe was synthesized using a commercial kit (Roche, Indianapolis, IN) from a plasmid containing full length rat CRF cdna (kind donation of Dr. K. Mayo, Northwestern University, Evanston, IL). Sections were postfixed in 4% formaldehyde for 1 min. Following phosphate-buffered saline (PBS) washes, proteins were acetylated in 0.1 M triethanolamine, ph 8.0, and 0.2% acetic acid. Following washes in saline sodium citrate buffer 2x, sections were dehydrated in a graded ethanolchloroform series. Pre-hybridization and hybridization were then performed at 70 C in a buffer containing 50% formamide, SSC 2x, Denhardt s 5x, 0.5 mg/ml sheared salmon sperm DNA, and 0.25 mg/ml yeast RNA. Probe was diluted in the hybridization buffer (800 ng/ml) and incubated overnight on slides. Post-hybridization washes were performed in 50% formamide, 104

119 SSC 2x, and 0.1% Tween-20. Slides were then blocked for 1 h and incubated with anti-dig antibody conjugated to alkaline phosphatase (Roche, 1:1000) and anti-th antibody (Millipore, Billerica, MA, AB152, 1:500) overnight at 4 C in TNT buffer (0.1 M Tris, ph 7.5, 0.15 M NaCl, 0.1% Tween-20) containing 1% blocking reagent (Roche). A donkey anti-rabbit secondary antibody conjugated to Alexa Fluor 488 (Life Technologies, Carlsbad, CA, 1:200, 2 h) was used to reveal the TH signal. Following TNT washes and incubation in 0.1 M Tris-HCl, ph 8, 0.1 M NaCl, 0.01 M MgCl 2, HNPP combined with Fast Red TR (Roche) was then used to detect alkaline phosphatase. To enhance the signal, fresh substrate was applied three times for 30 min and slides were rinsed in TNT in-between. Slides were washed, air dried, and coverslipped with Vectashield HardSet-DAPI (Vector Laboratories, Burlingame, CA). Images were taken using either epifluorescence (Zeiss Axiophot) or confocal microscopy (LaserSharp 2000, version 5.2, emission wavelengths 488, 568, and 647 nm, Bio-Rad). Place conditioning The place conditioning apparatus was obtained from Med Associates (SOF-700RA-25 Two Chamber Place Preference Apparatus; St. Albans, VT, USA). One environment was black with a metal rod floor, and the other was white with a wire mesh floor. An intermediate gray area housed a removable partition. Each cage was cleaned between animals, and each group was fully counterbalanced. During preference testing, the dividing partition was removed, and the mice were given free access to both environments. A single 10 min preference test session was performed 5 days after the last conditioning day. All place conditioning and testing were performed between 10:00 AM and 6:00 PM. The nicotine-dependent and -withdrawn groups of mice were conditioned according to modified place conditioning procedures as described previously (Grieder et al., 2010; and see Chapter 2). Conditioning occurred only during withdrawal from chronic nicotine so that the motivational effects of withdrawal but not the direct effects of chronic nicotine were paired with the place conditioning environment. Eight hours after minipump removal, when the mouse was experiencing motivational withdrawal from chronic nicotine (Grieder et al., 2010; and see 105

120 Chapters 2 and 3), it was subcutaneously pretreated with vehicle (20% HBC) or MPZP (20 mg/kg) and confined to one of the conditioning environments for 1 h. The difference score for each animal was calculated by subtracting the time spent in the non-paired environment from the time spent in the withdrawal-paired environment during preference testing. For nicotine-dependent mice (not withdrawn), conditioning occurred only during exposure to chronic nicotine so that the motivational effects of chronic nicotine in a dependent animal were paired with the place conditioning environment. The mice were subcutaneously pretreated with vehicle (20% HBC) or MPZP (20 mg/kg) and confined to one of the conditioning environments for 1 h. The difference score for each animal was calculated by subtracting the time spent in the non-paired environment from the time spent in the nicotine-paired environment during preference testing. For the acute nicotine experiments, previously drug-naive mice were subcutaneously pretreated with 20% HBC or MPZP (20 mg/kg), given a subcutaneous injection of 1.75 mg/kg nicotine, and immediately confined or confined 8 h later (Grieder et al., 2010; and see Chapter 2) to one of the conditioning environments for 1 h. The next day, the mouse was pretreated again with HBC or MPZP, given acute saline instead of nicotine, and confined to the other environment. The difference score for each animal was calculated by subtracting the time spent in the saline-paired environment from the time spent in the nicotine-paired environment. Open field testing Mice that were bilaterally injected with AAV2-shSCR or AAV2-shCRF (coordinates: AP-3.3, DV-4.4, ML±0.5) were placed in the center of a grey box measuring 41 x 41 x 38 cm for 5 minutes. The room was dark with the open field testing box illuminated by a soft red light. The locomotor activity and duration of time spent in the center square area of the box was recorded by video camera and calculated by the monitoring program (Ethovision XT, Noldus; Leesburg, VA, USA). The test box was cleaned with 70% alcohol between each mouse. 106

121 Statistical analysis The data were analyzed with Statistica software using a one- or two-way ANOVA or Student s t-test, where appropriate. In all cases, a normality test and an equal variance test were performed before the ANOVA to ensure its validity. Duncan s post hoc tests were used when appropriate. The data are expressed as mean ± SEM. Results Nicotine dependence upregulates CRF mrna in the pvta To test whether nicotine dependence upregulates CRF in the brain stress and reward systems, we first measured CRF mrna in two key regions of the CRF brain stress system, the PVN and CeA, and in the VTA of the brain reward system using quantitative real-time polymerase chain reaction (rtpcr). Groups of mice were made nicotine dependent by chronic exposure to nicotine delivered by osmotic minipumps (7 mg/kg/d) (Grieder et al., 2010; Grieder et al., 2012). Brain punches of the PVN, CeA, and VTA (Figure 4.1a) were sampled in salinetreated mice, in dependent mice with intact minipumps (dependent mice), or 8 h after removal of the minipump (withdrawn mice). CRF mrna levels were 7-15 times lower in the VTA than in the CeA and PVN (Figure 4.1b), in accordance with the fact that the CeA and PVN contain large populations of cell bodies that contain CRF mrna, whereas similar neurons have never been reported in the VTA (Swanson et al., 1983). However, we detected CRF mrna in saline-treated groups that could not be attributed to experimental noise, suggesting that a very small amount of CRF mrna is present in the VTA of nondependent animals. Chronic exposure to nicotine selectively increased CRF mrna levels only in the VTA in both the dependent and withdrawn groups of mice compared with saline-treated mice (F 4,100 = 2.7, p = 0.034), without altering CRF expression in the PVN or CeA (Figure 4.1b). CRF neurons project to and synapse with both DA and γ-aminobutyric acid (GABA) neurons in the VTA (Tagliaferro and Morales, 2008), and CRF release in the VTA is potentiated after repeated but not acute cocaine exposure (Hahn et al., 2009), suggesting that chronic exposure to abused drugs may either lead to axonal transport of 107

122 Figure 4.1. Nicotine dependence increases CRF mrna levels in the VTA. (a) Location of the brain tissue samples in the CeA, VTA, and PVN. (b) rtpcr measurement of CRF mrna levels (delta C t ) in the CeA, VTA, and PVN relative to the housekeeping gene GAPDH, expressed on a logarithmic scale, in mice chronically exposed to saline or nicotine or withdrawn from chronic nicotine. In saline-treated mice, CRF mrna levels were approximately 10 times lower in the VTA than in the CeA and PVN, in accordance with the fact that the CeA and PVN contain large populations of cell bodies that contain CRF mrna, whereas similar neurons have never been reported in the VTA. However, a very small amount of CRF mrna was detected in saline-treated groups, suggesting that a very small amount of CRF mrna is present in the VTA of nondependent animals. A significant increase in CRF mrna was observed in the VTA but not the CeA or PVN in both nicotine-dependent and -withdrawn groups of mice (*p < 0.05), suggesting that the induction of nicotine dependence leads to an approximately two-fold increase in CRF mrna selectively in the VTA. 108

123 CRF mrna to the VTA, or to the recruitment of a small population of VTA neurons that synthesize CRF mrna. Nicotine dependence recruits a population of CRF neurons in the pvta To test the hypothesis that chronic nicotine recruits a small population of CRFexpressing neurons in the VTA, we next performed CRF mrna in situ hybridization in the anterior and posterior VTA (avta and pvta respectively; Figure 4.2a, b) in a separate cohort of nicotine-dependent mice. As expected by the lack of previous reports in the literature of CRF neurons in the VTA, we observed very few CRF mrna-containing cells in the VTA in salinetreated mice (Figures 4.2c and 4.3). However, after exposure to chronic nicotine, a significant population of CRF neurons with dense CRF mrna in cell bodies could be detected bilaterally in the VTA. This increase was associated with a significant two-fold increase in CRF mrna density in the pvta (t 42 = -2.43, p = ; Figure 4.2c) as well as a non-significant increase in the avta (t 32 = -1.30, p = ; Figure 4.2c). Nicotine dependence also increased the number of neurons per section that contained CRF mrna in the pvta (t 42 = -1.9, p = 0.029) but not avta (t 32 = 0.75, p = 0.77; Figure 4.3). These results confirm the findings obtained with rtpcr and further demonstrate that exposure to chronic nicotine recruits a population of CRF mrnaexpressing neurons in the VTA, specifically in the pvta. Interestingly, a recent study demonstrated that nicotine selectively activates dopaminergic neurons in the pvta, but not avta, suggesting that newly CRF-expressing neurons in the pvta may be dopaminergic (Zhao-Shea et al., 2011). CRF mrna is expressed in dopamine neurons To test whether CRF neurons in the pvta were also dopaminergic, we performed double labeling of CRF mrna and DA neurons using CRF mrna radioactive and fluorescent in situ hybridizations coupled with tyrosine hydroxylase (TH) immunohistochemistry (Figure 4.4). CRF mrna positive neurons were located bilaterally in the pvta (Figure 4.4a) in TH enriched 109

124 Figure 4.2. Nicotine dependence and withdrawal recruits and activates the CRF system in the VTA. (a) Location of the avta (left: bregma range: to -3.16) and pvta (middle: bregma range: to -3.80) and density of CRF mrna signal in the avta and pvta in saline- and chronic nicotine-treated mice (right). Nicotine-dependent mice showed an increase in CRF mrna density compared with saline control (*p < 0.05) in the pvta but not avta. (b) Photographic representation of Nissl-stained sections that validate the proper anatomical location for corresponding in situ hybridization. Scale bars = 100 µm. (c) Representative CRF mrna in situ hybridization sections of the avta and pvta (box) in mice chronically exposed to nicotine or saline. The number of CRF mrna-containing neurons (arrows) was increased in nicotinedependent mice in the pvta (inset). 110

125 111

126 Figure 4.3. Nicotine dependence increased the number of cells that contain CRF mrna in the pvta but not avta. The average number of CRF mrna-positive cells counted per section across the entire avta or pvta in saline-treated or chronic nicotine-treated groups of mice is shown. Nicotine dependence increased CRF-positive cells in the pvta but not the avta. 112

127 Figure 4.4. Double labeling of DA neurons and CRF mrna using CRF in situ hybridization and TH immunohistochemistry. (a) CRF mrna radioactive in situ hybridization (black) demonstrates CRF-positive cell bodies (arrows) in the pvta. (b) TH immunohistochemistry (red) on the same section of the pvta shows that dopaminergic cell bodies are located bilaterally in the pvta, in TH immunoreactive areas. (c) Double fluorescent labeling of CRF mrna (red) and TH protein (green) in the pvta at high magnification shows that the majority of CRF mrna-positive neurons also express TH (CRF+/TH+ neuron: arrow, CRF-/TH+ neuron: arrowhead). (d) Confocal image of a single VTA neuron co-expressing TH (green) and CRF mrna (red). 113

128 regions (Figure 4.4b), and significantly colocalized with TH-positive neurons with an estimated 87.5% (14/16) of CRF mrna-expressing neurons being TH-positive (Figure 4.4c, d). The induction of CRF mrna in dopaminergic neurons in the pvta after the development of nicotine dependence suggests that a pool of neurons in the pvta can either synthesize CRF mrna de novo, or that it reflects a prominent upregulation over baseline levels in naive animals that were too low to be easily detected using classical detection methods (Kovacs and Sawchenko, 1996). Our results showing very low levels of CRF mrna in the VTA using rtpcr and in situ hybridization in naive rats support the latter hypothesis. Nevertheless, the upregulation of CRF mrna suggests a gain of CRF function in VTA dopaminergic neurons. Withdrawal depletes CRF peptide in the pvta Increased CRF release during drug withdrawal is associated with decreased immunodensity of CRF peptide, which has been proposed to reflect CRF depletion from synaptic vesicles subsequent to a local increase in CRF release (Merlo-Pich et al., 1995). Thus, we examined if recruitment of CRF neurons in the pvta was also associated with a local decrease in CRF peptide density during withdrawal from chronic nicotine using CRF immunohistochemistry in the avta and pvta (Figure 4.5a). Densitometry analysis of the VTA, CeA, and PVN revealed that both nicotine dependence and withdrawal from chronic nicotine decreased CRF peptide density compared with saline-treated controls in the pvta (F 2,31 = 4.4, p = 0.02; Figure 4.5b) and CeA (F 2,23 = 1.1, p = 0.36; Figure 4.6), but not avta (F 2,23 = 0.03, p = 0.97; Figure 4.5b) or PVN (F 2,21 = 0.75, p = 0.48; Figure 4.6). Background immunoreactivity outside of, but surrounding, the VTA was not significantly different between groups (F 2,31 = 3.08, p > 0.05; data not shown), demonstrating the specificity of these effects. Our combined observations that chronic nicotine exposure increases CRF mrna expression and decreases CRF peptide density in the pvta suggest that newly induced CRF-synthesizing neurons are functional and locally release CRF in the pvta. The decreased immunodensity observed in the CeA replicates results obtained previously with alcohol-dependent and -withdrawn rats (Funk et al., 2006) and extends them to nicotine-dependent and -withdrawn mice. The finding that CRF 114

129 Figure 4.5. Withdrawal from chronic nicotine depletes CRF peptide in the pvta. (a) Representative CRF immunohistochemistry sections of the areas containing the avta (left) and pvta (middle; pvta enclosed in box), and close-up of the pvta (right) in mice chronically exposed to saline (upper sections) or nicotine (lower sections). Scale bars = 100 µm. (b) CRF peptide density in the avta and pvta in mice given saline (Sal), chronic nicotine (Nic), or chronic nicotine and withdrawal (WD). The density of CRF peptide was decreased was decreased after chronic nicotine and withdrawal from chronic nicotine (*p < 0.05, vs. saline), corresponding to a proposed increase in CRF release. 115

130 116

131 Figure 4.6. Nicotine dependence and withdrawal decreases CRF peptide density in the CeA but not the PVN. Densitometry analysis on CRF immunohistochemical labeling of the CeA and PVN revealed that both chronic nicotine exposure and withdrawal from chronic nicotine decreased CRF peptide density compared with saline-treated controls (*p < 0.05, vs. saline) in the CeA (F 2,23 = 1.1, p = 0.36) but not the PVN (F 2,21 = 0.75, p = 0.48). 117

132 density was decreased in the pvta, but not avta or PVN, suggests that increased CRF release not only in the CeA but also in the pvta may contribute significantly to the motivational effects of nicotine withdrawal in dependent subjects. Blocking upregulation of CRF mrna in the pvta prevents the aversive effects of withdrawal To test whether CRF mrna-expressing neurons in the VTA play a role in nicotine dependence, we tested the causal relationship between upregulation of CRF mrna in the VTA of nicotine-dependent and -withdrawn animals and the expression of conditioned place aversion to nicotine withdrawal and anxiety-like behavior. We injected an adeno-associated viral vector that encodes a short-hairpin RNA-targeting CRF mrna (AAV2-shCRF) or a non-targeting sequence (AAV2-shSCR) bilaterally in the pvta 3-4 weeks before exposure to chronic or acute nicotine to chronically decrease CRF mrna expression during exposure to nicotine and withdrawal (Figure 4.7a, b). Quantification of CRF-positive neurons in withdrawn mice injected with the AAV2-shCRF in the pvta and subjected to in situ hybridization showed a significant downregulation (-31%; Figure 4.7c) of the number of CRF mrna-containing cells in the pvta compared with withdrawn mice injected with the AAV2-shSCR vector (t 11 = 2.737, p = ; Figure 4.7d). Moreover, in the remaining CRF-expressing pvta cells, AAV2-shCRF vectorinjected mice exhibited an 11% decrease in CRF mrna content (normalized optical density: 88.6 ± 5.5 for AAV2-shCRF vs ± 3.4 for AAV2-shSCR; p = ). Notably, the mice were sacrificed approximately 4 weeks into withdrawal after the end of behavioral testing, demonstrating that the upregulation of CRF mrna in nicotine-withdrawn mice and CRF silencing by the viral vector were both long-lasting. In the place conditioning paradigm, a twoway analysis of variance (ANOVA) revealed a significant interaction between virus injection and nicotine history (F 1,23 = 4.976, p = ; Figure 4.7e). Nicotine-dependent and -withdrawn mice infused with the AAV2-shSCR control vector in the pvta showed an aversive motivational response to the withdrawal-paired environment (p < 0.05), which was not observed in mice infused with the AAV2-shCRF vector (p > 0.05). Furthermore, nondependent mice given either the control or silencing vector in the pvta and an acute aversive injection of nicotine 118

133 Figure 4.7. Preventing upregulation of CRF mrna in the VTA by a viral vector prevents the aversive motivational response to withdrawal from chronic nicotine. (a) Timeline of experiment. Groups of mice were injected with AAV2-shCRF or AAV2-shSCR in the VTA and allowed 3-4 weeks to recover. Osmotic minipumps containing 7 mg/kg/d nicotine were implanted (D1) and left in place for 12 days. On D13, pumps were removed for withdrawn groups, or left in place for dependent groups, and conditioning was performed 8 hours later. Five days later, on D18, mice were preference tested, and then tested in the open field on D38. Subsequent to testing, all mice were sacrificed and their brains subjected to in situ hybridization. (b) DNA construct used to produce vectors for CRF silencing (AAV2-shCRF) and control vectors (AAV2-shSCR). ITR, inverted terminal repeat; CMV, cytomegalovirus; EGFP, enhanced green fluorescent protein; hgh polya, human growth hormone polyadenylation signal. (c) Atlas showing location of sections and representative images of VTA CRF mrna-containing cells. Scale bars = 100 µm. (d) The absolute number of CRF mrna-positive cells after AAV2- shcrf was significantly decreased compared to mice injected with AAV2-shSCR (*p < 0.05). (e) AAV2-shCRF blocked the conditioned place aversion to nicotine withdrawal (*p < 0.05) in nicotine-dependent and withdrawn mice given AAV2-shSCR, but not the aversion to acute nicotine in nondependent mice. (f) Nicotine-dependent and -withdrawn mice injected with AAV2-shSCR spent significantly less time spent in the central area of the open field than mice injected with AAV2-shCRF (*p < 0.05). 119

134 120

135 (1.75 mg/kg) showed an aversive response to the nicotine-paired environment, demonstrating that the lack of aversion to withdrawal in dependent mice infused with the silencing vector was not due to a general impairment of conditioned place aversion but was specific to withdrawal in dependent mice. Mice that were injected with the AAV2-shCRF vector outside of the pvta showed an aversive response to nicotine withdrawal similar to control mice (t 10 = , p = ; Figure 4.8), demonstrating the anatomical specificity of this effect. Altogether, these results suggest that CRF mrna in the pvta does not mediate the aversive response to acute nicotine, but specifically mediates aversion to nicotine withdrawal in dependent subjects. These results establish a causal relationship between the recruitment of CRF mrna-expressing neurons in the VTA during withdrawal from chronic nicotine and the aversive motivational response to nicotine withdrawal. Activation of the brain CRF-CRF1R system is associated with increased anxiety-like behavior in humans and animals (Holsboer and Ising, 2008) and is hypothesized to be responsible for the negative emotional states after protracted abstinence (George et al., 2007). To test this hypothesis, we used the same nicotine-dependent and -withdrawn mice injected with the AAV2-shCRF and AAV2-shSCR vectors and measured open field activity to evaluate anxietylike behavior during protracted abstinence (3-4 weeks). A one-way ANOVA revealed a significant effect of the viral vector on the duration of time spent in the central open area (F 3,31 = 4.148, p = 0.015; Figure 4.7f). Mice injected with the AAV2-shSCR vector spent significantly less time in the central open area of the open field than mice injected with the AAV2-shCRF vector (p < 0.05). These results demonstrate that upregulation of CRF mrna in the pvta is required for the anxiogenic-like effects of protracted nicotine abstinence. CRF1R receptor blockade prevents withdrawal aversions The present results revealed a depletion of CRF peptide in the pvta during nicotine withdrawal, which is thought to reflect a local increase in CRF release (Merlo-Pich et al., 1995). We therefore hypothesized that activation of CRF1Rs in the pvta would be necessary for the 121

136 Figure 4.8. The silencing vector must be injected in the VTA to block withdrawal aversions. Nicotine-dependent and -withdrawn mice that were injected with AAV2-shCRF silencing vector outside of the VTA showed an aversive motivational response to nicotine withdrawal similar to AAV2-shSCR control vector-injected mice, demonstrating the anatomical specificity of this effect. 122