Medical University of South Carolina Medical University of South Carolina

MEDICA MEDICA

MUSC Theses and Dissertations

2020

Individual Differences in Cocaine-Conditioned Avoidance Behavior Individual Differences in Cocaine-Conditioned Avoidance Behavior

and Implication for Addiction Vulnerability and Implication for Addiction Vulnerability

Maya Eid Medical University of South Carolina

Follow this and additional works at: https://medica-musc.researchcommons.org/theses

Recommended Citation Recommended Citation Eid, Maya, "Individual Differences in Cocaine-Conditioned Avoidance Behavior and Implication for Addiction Vulnerability" (2020). MUSC Theses and Dissertations. 64. https://medica-musc.researchcommons.org/theses/64

This Dissertation is brought to you for free and open access by MEDICA. It has been accepted for inclusion in MUSC Theses and Dissertations by an authorized administrator of MEDICA. For more information, please contact medica@musc.edu.

DOCTORAL THESIS

Medical University of South Carolina

College of Graduate Studies

Department of Neuroscience

Individual di�erences incocaine-conditioned avoidance behavior

and implication for addiction vulnerability

Author: Maya Eid

Supervisor: Dr. Thomas Jhou

A dissertation submitted to the faculty of the Medical University of South Carolina

in partial ful�llment of the requirements for the degree of Doctor of Philosophy in

the College of Graduate Studies

Approved by :

Thomas Jhou, PhD

Chairman, Advisory Committee

Brett Froeliger, PhD

Dorothea Jenkins, MD

Peter Kalivas, PhD

Bryan Tolliver, MD/PhD

2

Acknowledgments

I would like to express the deepest appreciation to my mentor, Dr. Thomas Jhou.

You have the attitude and substance of a genius: you continually and convincingly

conveyed a spirit of adventure in regard to research, and an excitement in regard to

teaching. I have undoubtedly grown under your direction, not only as a scientist, but

also as an economist, historian, achitect etc! But more seriously, you contributed to

my growth as a person. I consider myself lucky and honored to have had the chance to

work with you for 5 years, and the skills I learned from you will serve me throughout

my whole life. I thank my committee members; Dr.'s Brett Froeliger, Dorothea

Jenkins, Peter Kalivas and Bryan Tolliver for their helpful feedback, encouragement

and insightful suggestions, and for showing me that the work we do should have a

purpose that transcends academia. I also want to take this opportunity to thank all

the faculty members, postdocs and employees in the department of Neuroscience; we

are lucky to be surrounded by encouraging and enthusiastic individuals that provide us

with the potential to develop into accomplished scientists. I would also like to thank

the MSTP, Dr. Demore, Dr. Mennick, Dr. Halushka and Amy Connolly, as well

as the American University of Beirut, for having given me this opportunity of a life

time! Who who have thought, a few years ago, that I would become an accomplished

scientist in one of the best institutions in the world! I thank my labmates in the Jhou

Lab, Dr. Peter Vento, Dr. Je�ery Parrilla-Carrero, Dr. Hao Li, Ying Chao, Nicki

1

Pullmann, Crystal Smith, Alen Thomas and Rachel Lipat for not only mentoring

me in scienti�c methods, but also for collaborating on many of my projects. Je�rey,

so much of my thesis used your �ndings as pillars to build on, so I thank you for

making this collaboration fun and easy, and for being patient with me! Special thank

you to Ying, you were my biggest source of support during these years, and the best

listener I have ever met. Thank you for always cheering me up from the little things

you do/say to the e�ort and thoughts you put to make me feel better. And mostly

thank you all for all the fun we have had in the past 4 years! To me, you are not

just labmates but dear friends. To all my friends in Lebanon and spread out all over

the world, I am eternally grateful for your support; thank you for listening to me

boring you about my science, and thank you for putting thoughts into every advice

you gave me. I could not have gone through this journey without you. Special thank

you to Lea and Carla. Another big thank you goes to my friends in Charleston who

I shared a lot of adventures with, including the Lebanese family, speci�cally Reda,

Joe, Ali and Zahraa. And of course, a very special thank you to Roy... Last but

never least, my greatest gratitude is to my family; my parents Hassan and Leila,

brother Marwen and sister Randa, grandma, cousins uncles and aunts (speci�cally

Micha), who have unquestionably made the largest sacri�ces and contributions to my

journey. I would not be here without your love, encouragement, support and prayers.

It was extremely di�cult to leave my little nest to pursue my dreams, and I am sure

it was even more di�cult for you, but nevertheless you always always always raised

me up no matter what. To my dad, you have always believed in me since I was 3

years old, and have always seen the potential in me. I remember wanting to go to

"polytechnique" without even knowing what it meant! Thank you for teaching me

perseverance, discipline and self-reliance, all while being loving and encouraging. To

my mom, you are my role model and I would be lucky if I can become a full-rounded

2

loving smart caring kind a�ectionate mother and physician like you. You have taught

me everything, including academic skills, but more importantly, you have passed on

to me all the great qualities that constitute a beautiful soul that you carry. You are

my comfort zone, my greatest source of love and my best source of advice. I am

accomplished and complete ONLY thanks to you. To my brother and sister, you are

my best friends and I am so lucky to be the baby and have you guys to lean on. I

was a little bit discouraged that you all could not be here and attend my defense in

person, but fear not, there is another milestone in 2 years, so see you then!

3

Contents

Acknowledgments 1

List of Figures 8

List of Abbreviations 10

Abstract 13

1 General Introduction 15

1.1 Operant Runway Model of Cocaine Seeking . . . . . . . . . . . . . . 15

1.1.1 Origins and Principles . . . . . . . . . . . . . . . . . . . . . . 15

1.1.2 Runway Model and Drugs of Abuse . . . . . . . . . . . . . . . 17

1.2 The RMTg: the Anti-Dopamine Nucleus . . . . . . . . . . . . . . . . 20

1.2.1 RMTg Anatomy . . . . . . . . . . . . . . . . . . . . . . . . . . 20

1.2.2 Cellular Pro�le . . . . . . . . . . . . . . . . . . . . . . . . . . 23

1.2.3 Distinction Between RMTg and GABA Neurons of the VTA . 25

1.2.4 RMTg Function . . . . . . . . . . . . . . . . . . . . . . . . . . 26

1.3 Aversive Properties of Drugs of Abuse . . . . . . . . . . . . . . . . . . 28

1.3.1 Opponent Process Theory and Cocaine . . . . . . . . . . . . . 28

1.3.2 RMTg and Aversion . . . . . . . . . . . . . . . . . . . . . . . . 29

4

1.3.3 Accumbens and Aversion . . . . . . . . . . . . . . . . . . . . . 31

1.3.4 Implications in Addiction . . . . . . . . . . . . . . . . . . . . 33

1.4 Statement of the Problem . . . . . . . . . . . . . . . . . . . . . . . . 34

2 Methods 36

2.1 Animals and Surgeries . . . . . . . . . . . . . . . . . . . . . . . . . . 36

2.1.1 Animals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36

2.1.2 Catheter Surgeries . . . . . . . . . . . . . . . . . . . . . . . . 37

2.1.3 Stereotaxic Surgeries . . . . . . . . . . . . . . . . . . . . . . . 37

2.2 Behaviors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38

2.2.1 Runway Operant Cocaine-Seeking . . . . . . . . . . . . . . . . 38

2.2.2 Progressive Ratio and Punishment Resistance . . . . . . . . . 39

2.2.3 Novelty-Induced Locomotion . . . . . . . . . . . . . . . . . . . 39

2.2.4 Conditioned Place Preference/Conditioned Place Aversion . . 39

2.2.5 Shuttlebox Shock Escape . . . . . . . . . . . . . . . . . . . . . 40

2.2.6 Self Administration . . . . . . . . . . . . . . . . . . . . . . . . 40

2.2.7 Elevated Plus Maze . . . . . . . . . . . . . . . . . . . . . . . . 42

2.2.8 Pavlovian Training . . . . . . . . . . . . . . . . . . . . . . . . 42

2.3 In-vivo Recording . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43

2.3.1 Wire Recording . . . . . . . . . . . . . . . . . . . . . . . . . . 43

2.3.2 Fiberphotometry . . . . . . . . . . . . . . . . . . . . . . . . . 44

2.4 Histology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45

2.4.1 Immunostaining . . . . . . . . . . . . . . . . . . . . . . . . . . 45

2.4.2 Western Blot . . . . . . . . . . . . . . . . . . . . . . . . . . . 46

2.4.3 Activated RhoA Assay . . . . . . . . . . . . . . . . . . . . . . 47

2.5 Statistical Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48

5

3 Behavioral Characterization of Runway Model 49

3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49

3.2 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51

3.2.1 Individual Sprague-Dawley rats show large variation in cocaine-

avoidance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51

3.2.2 Cocaine-avoidance on the runway is speci�c to the e�ects of

cocaine and is independent of control behaviors . . . . . . . . 53

3.2.3 Cocaine has dual rewarding and aversive e�ects, with the aver-

sive component driving individual variation in drug-seeking be-

havior . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54

3.3 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56

4 Adaptations in RMTg Underlie Individual Di�erences in Cocaine

Avoidance 63

4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63

4.2 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65

4.2.1 High cocaine avoiders show greater neuronal activation in the

RMTg . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65

4.2.2 Cocaine-avoiding rats exhibit enhanced excitability of VTA-

projecting RMTg neurons. . . . . . . . . . . . . . . . . . . . . 66

4.2.3 RhoA/Rock downregulates glutamatergic signaling, and bidi-

rectionally modulates cocaine-seeking via PTEN . . . . . . . . 69

4.2.4 Protein quanti�cation . . . . . . . . . . . . . . . . . . . . . . 71

4.3 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71

5 Adaptations in NAshell Underlie Individual Di�erences in Cocaine

Avoidance 80

6

5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80

5.2 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82

5.2.1 D2-like agonist/antagonist injections into the NAshell reduce/increase

motivation to seek cocaine in a runway operant task. . . . . . 82

5.2.2 Cocaine-induced avoidance behavior correlates with impaired

LTD at excitatory synapse in D2-MSNs of the NAshell. . . . . 84

5.2.3 Cocaine-induced avoidance behavior correlates with increased

adenosine A2A receptor tone in D2-MSNs of the NAshell. . . . 85

5.2.4 RMTg inhibition reduces shock cue-induced decrease in DA lev-

els in the NAshell, without a�ecting reward or neutral cues . . 86

5.2.5 High cocaine-avoiders have reduced early acquisition of cocaine

self-administration but higher rates of reinstatement . . . . . . 87

5.3 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88

6 Discussion and Future Directions 97

6.1 Implications for Addiction and Reinstatement . . . . . . . . . . . . . 97

6.2 Identi�cation of Potential Therapeutics . . . . . . . . . . . . . . . . . 99

Bibliography 103

7

List of Figures

1.1 Representative trace of approach/retreat behavior on the runway . . . 18

1.2 Operant runway model of cocaine self-administration . . . . . . . . . 19

1.3 Topographical organization of RMTg a�erents from lHb and e�erents

to VTA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

1.4 RMTg a�erents and e�erents . . . . . . . . . . . . . . . . . . . . . . . 23

1.5 FoxP1 as a marker of RMTg . . . . . . . . . . . . . . . . . . . . . . . 24

1.6 Opponent process of cocaine . . . . . . . . . . . . . . . . . . . . . . . 30

1.7 Proposed model of RMTg�>VTA�>NAshell circuit mediating di�er-

ences in cocaine aversion. . . . . . . . . . . . . . . . . . . . . . . . . . 35

3.1 Two populations of rats observed on the runway task . . . . . . . . . 58

3.2 Individual Sprague-Dawley rats show large variation in cocaine-avoidance 59

3.3 High- and low-avoiders do not di�er in control tasks . . . . . . . . . . 60

3.4 Cocaine has dual rewarding and aversive e�ects, with the aversive com-

ponent driving individual variation in drug-seeking behavior . . . . . 61

3.5 High cocaine-avoiders have enhanced anxiety during the aversive phase

of the drug . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62

4.1 High cocaine-avoiders show greater neuronal activation in the RMTg 75

8

4.2 Glutamatergic inputs drive greater RMTg activation in brain slices

from high vs low cocaine-avoider rats . . . . . . . . . . . . . . . . . . 76

4.3 Manipulating GluR1 subunit of AMPA receptor modulates avoidance

behavior on the runway . . . . . . . . . . . . . . . . . . . . . . . . . . 77

4.4 PTEN downregulates glutamatergic signaling, and bidirectionally mod-

ulates cocaine-seeking . . . . . . . . . . . . . . . . . . . . . . . . . . 78

4.5 RhoA/ROCK/PTEN pathway is upregulated in RMTg of low-avoiders 79

5.1 D2-like agonist/antagonist injections into the NAshell reduce/increase

motivation to seek cocaine in a runway operant task . . . . . . . . . . 93

5.2 Cocaine-induced avoidance behavior correlates with impaired LTD at

excitatory synapse in D2-MSNs of the NAshell and with upregulation

of A2A receptor. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94

5.3 RMTg inhibition reduces shock cue-induced decrease in DA levels in

NAshell, without a�ecting reward or neutral cue. . . . . . . . . . . . 95

5.4 High cocaine-avoiders have reduced early acquisition of cocaine self-

administration but higher rates of reinstatement . . . . . . . . . . . . 96

9

List of Abbreviations

5-HT2CR Serotonin 2C receptor

A2A Adenosine receptor type 2

AAV Adeno-associated vector

ALDH2 aldehyde dehydrogenase

AMPA alpha-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid

ANOVA Analysis of variance

CB1 Cannabinoid receptor type 1

CP-AMPAR Calcium-permeable AMPA receptor

CPA Conditioned place avoidance

CPP Conditioned place preference

CTA conditioned taste aversion

D1R Dopamine D1 receptor

D1-MSN Dopamine D1 receptor expressing medium spiny neuron

D2R Dopamine D2 receptor

D2-MSN Dopamine D2 receptor expressing medium spiny neuron

DA Dopamine

DREADD Designer receptor exclusively activated by designer drug

eCB endocannabinoids

EPN entopeduncular nucleus

10

EPSC Excitatory postsynaptic current (eEPSC: evoked, sEPSC: spontaneous)

Foxp1 Forkhead Box P1

FR Fixed ratio

GABAAR Ionotropic (type A) GABA receptor

GABABR Metabotropic (type B) GABA receptor

GFP Green �uorescent protein

GluR1 Glutamate Receptor Ionotropic, AMPA 1

GluR2 Glutamate Receptor Ionotropic, AMPA 2

GPCR G protein-coupled receptor

iv intravenous

lHb Lateral habenula

LTD Long term depression

LTP Long term potentiation

MDMA methylenedioxymethamphetamine

MSN Medium spiny neuron

NAc Nucleus accumbens

NAcore Nucleus accumbens core

NAshell Nucleus accumbens shell

NMDA N-methyl-D-aspartate

PDZ post synaptic density protein (PSD95), Drosophila disc large tumor suppressor

(Dlg1), and zonula occludens-1 protein (zo-1)

p-GluR1-S831 GluR1 subunit at serine831 site

PPTg pedunculopontine nuclei

PPR Paired-pulse ratio

PR Progressive ratio

PTEN Phosphatase and tensin homolog

11

RG Rabies glycoprotein

RhoA Ras homolog family member A

ROCK Rho-associated protein kinase

RMTg Rostromedial tegmental nucleus

SC subcutaneous

SNpc Substantia nigra pars compacta

SNpr Substantia nigra pars reticulata

STDP-LTD spike-timing dependent plasticity LTD

VP ventral pallidum

TH Tyrosine hydroxylase

VO-OHpic VO-OHpic trihydrate

VTA Ventral tegmental area

12

Abstract

Decades of studies have examined the mechanisms by which drugs of abuse produce

strong rewarding e�ects that drive addictive behavior. However, these drugs also

produce aversive e�ects that are less well understood, but that also strongly in�uence

drug-seeking, and may contribute to individual variation in addiction liability. For

example, individuals with particularly strong aversive responses to nicotine or alcohol

may be less likely to acquire smoking or drinking habits, with these protective e�ects

arising from speci�c biological di�erences in nicotinic receptor function or alcohol

metabolism, respectively. These aversive e�ects can also be exploited therapeuti-

cally, as in widely available treatments for alcoholism and smoking. We and others

found that cocaine also produces aversive e�ects that can slow or block drug-seeking,

and that the magnitudes of these protective aversive e�ects also vary widely between

individuals. However, the molecular substrates of cocaine's aversive e�ects, and of

their variation between individuals, are almost entirely unknown. To examine these

substrates, we examined functions of the rostromedial tegmental nucleus (RMTg),

a major GABAergic input to midbrain dopamine neurons that plays key roles in

aversive learning from cocaine and other stimuli. In particular, we found that RMTg

neural activity is necessary for cocaine's conditioned aversive e�ects, and is correlated

with aversion magnitude. Furthermore, in collaboration with a slice electrophysiolo-

gist in the lab (Dr. Je�rey Parrilla-Carrero), we identi�ed several molecular drivers of

13

this activity showing that cocaine drives RMTg activation via glutamatergic signal-

ing through calcium permeable AMPA receptors (CP-AMPARs) residing on RMTg

neurons. Additional data showed that this receptor undergoes di�erential plasticity

in response to cocaine, as it is downregulated in low cocaine-avoider rats having es-

pecially low aversive responses to cocaine, relative to high cocaine avoider rats with

stronger aversive responses. We also showed that this modulation may be driven by

the phosphatase enzyme, PTEN, which has been previously implicated in synaptic

depotentiation. In particular, we found that inhibition of PTEN increases glutamate

signaling in RMTg and reduces cocaine-seeking, suggesting PTEN as a novel thera-

peutic target for cocaine addiction. Because RMTg neurons send strong inhibitory

projections to DA neurons in the ventral tegmetal area (VTA), which in turn project

to the nucleus accumbens (NAc), and because of extensive prior studies implicat-

ing D2 medium spiny neurons (D2 MSNs) in avoidance behavior, we also looked at

whether high- and low-avoiders have di�erences in D2 MSNs properties of the nu-

cleus accumbens shell (NAshell) after exposure to cocaine. We found that aversion to

cocaine in high-avoiders correlated with increased neuronal excitability, enhanced glu-

tamate transmission and impaired long term depression (LTD) at excitatory synapse

in D2-MSNs of the NAshell, and demonstrated that RMTg plays a role in modulating

dopamine levels in the NAshell in response to noxious stimuli.

14

Chapter 1

General Introduction

1.1 Operant Runway Model of Cocaine Seeking

1.1.1 Origins and Principles

Animal learning researchers have been using operant runways as a tool for the study

of goal-directed motivated behavior for several decades now [Miller, 1944, Crespi,

1942, Hull, 1934]. In such studies, the time required for the subject to traverse the

alley (i.e., Run Time) has proven to be a reliable proxy of the animal's motivation

to seek the incentive that has been paired with the goal box. Put simply, changes

in the subject's motivation to seek an incentive produce predictable and reliable

shifts in the run times required to get to the goal box. Thus the runway had some

potential as a means of measuring the motivation of animals to seek drugs of abuse.

Consequently, Ettenberg and his group designed and built a set of automated operant

runways which we adopted in our lab. In this apparatus, animals implanted with

intravenous catheters and connected to an elaborate drug delivery system, would

traverse a corridor and enter a goal box where they received an IV infusion of drug.

15

Infrared photo-emitter/detector pairs planted throughout the entire runway deliver

signals to a desktop computer that controlled the hardware of the apparatus and

delivers the outputs, and recorded the positions of the animal in real time so that

Start Latency (time to leave the start box) and Run Time (time to enter goal box)

could be recorded, along with the precise trajectory that the rat took moving between

the two ends of the runway, on every trial.

Conceptually and procedurally, the approach is modeled on the classic runway

work of Crespi [Crespi, 1942] who showed that rats traversing an alley for food showed

Run Times that re�ected the magnitude of the reward (amount of food) delivered in

the goalbox. Additionally, Crespi observed that an �unexpected� sudden change in

the magnitude of the incentive delivered in the goal box resulted in a dramatic shift

in run latencies on the next trial (a phenomenon that has come to be known as

the �Crespi e�ect�). This suggested a way to experimentally dissociate the e�ect

of manipulations on motivational versus reinforcement processes. Conceptually, if an

experimental manipulation before a given trial altered the motivational capacity of the

animal, we expect to see an immediate e�ect on Run Time on that trial; conversely,

the reinforcing impact of actually earning the drug are re�ected by the changes in run

times from one trial to the next while having no impact on test day run times (since

the animal would not be aware of the change in reinforcer magnitude until after it

had entered the goal box). Therefore, the independent measure Run Times on any

given trial provides a good proxy of the animal's motivation to seek the reinforcer on

that trial, while changes in Run Times from one trial to the next re�ects the impact

of reinforcement on subsequent motivation, or the reinforcing consequences of earning

the drug.

16

1.1.2 Runway Model and Drugs of Abuse

To date, runway self-administration has been established in rats working for a several

drug reinforcers including IV cocaine [Ben-Shahar et al., 2008, Deroche et al., 1999,

Ettenberg and Bernardi, 2006, Ettenberg and Geist, 1991, Ettenberg and Geist, 1993,

Heinrichs et al., 1998, Wakonigg et al., 2003], IV opiate receptor agonists such as

heroin, morphine, remifentanil and alfentanil [Crespo et al., 2006, Ettenberg et al.,

1996, McFarland and Ettenberg, 1995, Wakonigg et al., 2003], IV nicotine [Cohen and

Ettenberg, 2007], IV methylenedioxymethamphetamine (MDMA) [Wakonigg et al.,

2003], IV �speedball� (heroin+cocaine) [Guzman and Ettenberg, 2004], as well as

subcutaneous (SC) amphetamine [Ettenberg, 1990], SC morphine [Zernig et al., 2002],

oral ethanol [Czachowski, 1999], and both intracerebroventricular and intracranial

infusions of cocaine[Guzman and Ettenberg, 2007]. An early result of using this model

with cocaine was the unique behavioral aspect that the animals were exhibiting. In

previous studies using natural and drug reinforcers, the subjects run the alley faster

as trials progressed. In contrast, while cocaine-reinforced animals had �normal� start

latencies, they progressively slowed down on the runway over trials. Data from the

infrared emitter-detector pairs lining the base of the runway, allowed to distinguish if

those increases in latencies were not due to animals running more slowly, but rather

those animals showed more and more stop-and-retreat behavior as the experiment

progressed; animals would approach the goal box, stop, and then retreat back to the

start box [Ettenberg and Geist, 1991, Ettenberg and Geist, 1993, Jhou et al., 2013]

(Fig. 1.1).

These behaviors closely mimicked the classic �approach-avoidance� con�ict de-

scribed by Neal Miller [Miller, 1944] over 60 years ago in animals approaching a goal

box that had known mixed positive+negative characteristics like food+mild footshock

17

Figure 1.1: A representative spatio-temporal record from a single trial of an experi-enced animal running a straight alley once a day for IV cocaine. The record illustratesthe path that the subject took to reach to the goal box and earn the drug reinforcer.The locations in the runway are represented by the numbers on the ordinate, with `1'being an infrared sensor in the start box and 9 being one just outside the goal boxentry. Positions 10�12 are inside the goal box. In the sample shown, the rat madenumerous approach� avoidance `retreats' (stopping forward movement and returningback toward the goal box) which are representative of mixed positive and negativeassociations with the cocaine goal box. . Adapted from Ettenberg et. al, 2004.

[Geist and Ettenberg, 1997a]. To be clear, retreat behaviors are not strictly an index

for cocaine's negative or aversive e�ect but rather re�ect the existence of both positive

and negative associations with the goal box. Thus, the animals continue to return to

the goal area and eventually do enter and obtain their cocaine. Indeed, when the goal

box contains only negative stimuli (e.g., shock, or aversive drugs), the animals remain

18

in the start box and choose not to go to the goal box at all. Consistent with this view

is the fact that the animals showing retreats in the corridor still developed positive

preference in a conditioned place preference test, suggesting that the cocaine has re-

tained its rewarding e�ects [Ettenberg and Geist, 1991]. It is also important to note

that the approach-avoidance retreat behaviors described above cannot be explained

by aversive properties of the goal box itself since a saline control did not produce

retreats [Ettenberg et al., 1996, Ettenberg and Geist, 1993]. It would thus seem then

that cocaine's biphasic actions (initial positive �euphoric� e�ect followed by a negative

�dysphoric� e�ect) together result in the animal's development of ambivalence about

entering the goal box. The runway model has thus proven to be sensitive to both

the positive (approach) and negative (avoidance) properties of the drug in the same

animal at the same time. It then represents a valuable means for studying the neu-

robiological basis of cocaine's two opposing drug actions that together undoubtedly

determine the nature of cocaine use and abuse (Fig. 1.2).

Figure 1.2: Operant runway model of cocaine self-administration.

Although the runway task is not as widely used as lever-pressing for cocaine self-

19

administration, we have chosen to study avoidance to cocaine for several reasons.

First, the single-trial nature of the runway task greatly helps to isolate the aversive

e�ects of cocaine, since the aversive responses are not masked by subsequent doses.

Moreover, unlike the lever-press procedure where animals work to maintain their

drugged state, the runway protocol assesses the motivation of undrugged animals to

return to a location (the goal box) where on previous trials cocaine had been deliv-

ered. Thus, by testing animals a single trial at a time, the procedure permits the

measurement of drug-seeking behavior independent of the potentially confounding

direct pharmacologic e�ects of the drug. So, for example, the increases in operant

running observed over trials cannot be easily accounted for by cocaine's psychomotor

stimulant properties nor by the potential development of drug-induced sensitization

since the animals' behavior was assessed prior to the delivery of the drug reinforcer.

Another advantage of the runway task is that the limited availability of cocaine more

closely resembles early drug use in humans, where users more commonly start with

individual trial doses rather than the more unbounded amounts available in con-

ventional self-administration paradigms at close to zero cost. Finally, Spealman's

[Spealman, 1979] classic �nding that monkeys lever-pressing for cocaine on one lever,

would press a second lever in order to terminate drug availability suggest that both

measure of lever-pressing in self-administration and run times on the runway are, at

least partly, probing at similar underlying processes.

20

1.2 The RMTg: the Anti-Dopamine Nucleus

1.2.1 RMTg Anatomy

The rostromedial tegmental nucleus (RMTg), is a GABAergic nucleus �rst described

in 2009 as a major inhibitory input to neurons in the ventral tegmental area (VTA)

[Jhou et al., 2009a, Kau�ing et al., 2009]. This nucleus occupies a column of tissue

extending from the caudal edge of the VTA, to the rostral edge of the cholinergic neu-

rons of the pedunculopontine nuclei (PPTg). In the RMTg, GABAergic neurons are

embedded in the adjacent VTA, where they are close to the dopaminergic cellbodies

and dendrites; for this reason, the RMTg is also called `tail of the VTA' or `retro

VTA' [Scammell et al., 2000, Perrotti et al., 2005, Olson and Nestler, 2007, Brin-

schwitz et al., 2010].

RMTg e�erents project densely to midbrain regions where dopamine (DA) neu-

rons are found [Jhou et al., 2009a], and overwhelmingly make inhibitory-type synapse

on tyrosine hydroxylase (TH)-positive neurons in the VTA and substantia nigra. The

projections from the RMTg to the DArgic nuclei have been extensively character-

ized by tracing and electron microscopy studies and they show a topographical or-

ganization: the rostro-medial RMTg sends denser projection to the VTA, while the

latero-caudal RMTg to the SNpc.

The RMTg also sends abundant local projections to the brainstem regions impor-

tant in the control of motivation, stress, attention and arousal (VTA, SNpc, RRF,

dorsal raphe, pedunculopontinetegmental nucleus, locus coeruleus and subcoeruleus,

Barringtonnucleus)[Jhou et al., 2009b, Lavezzi and Zahm, 2011, Sego et al., 2014].

The RMTg receives dense excitatory glutamatergic inputs form the LHb, a nu-

cleus involved in aversive event signaling and decision making [Hikosaka, 2010, Proulx

et al., 2014], and the density of the projection from LHb to the RMTg is greater than

21

Figure 1.3: Connections of excitatory and inhibitory neurons associated with thedopaminergic nuclei. Adapted from Morello et. al, 2015.

other more familiar projection such as the VTA, interpeduncular nucleus, median

and paramedian raphe and laterodorsal tegmental region [Jhou et al., 2009c, Kau-

�ing et al., 2009]. LHb projections to RMTg are organized in a highly topographic

manner. Tracing studies show that habenula projections are also topographically

organized, where medial LHb regions project to the median and paramedian tegmen-

tum and involve only the most medial and rostral part of the RMTg, and lateral LHb

22

projects proportionately to more lateral and caudal parts of the RMTg, [Jhou et al.,

2009b], that in turn more strongly innervate the VTA and SNpc, respectively (Fig.

1.3) [Jhou et al., 2009b, Gonçalves et al., 2012, Yetniko� et al., 2015]. Interestingly,

this lateral part of the LHb receives its strong input from the ventral pallidum (VP)

and entopeduncular nucleus (EPN) [Herkenham and Nauta, 1977], which suggests

that a trans-habenula in�uence of the ventral and dorsal parts of the basal ganglia,

respectively, are delivered to the RMTg, via the projection from the LHb. Thus,

the RMTg acts as an intermediate target between the LHb and the VTA, converting

the excitatory stimuli from the LHb to inhibition of the DArgic neurons. Numer-

ous other forebrain and brainstem structures also project varyingly strongly to the

RMTg, including structures implicated in aversive processing such as the anterior

cingulate cortex [Vogt, 2005] and subcortical regions such as the septum [Sheehan

et al., 2004], parts of the extended amygdala [Davis et al., 2010], and the periaque-

ductal gray [McNally et al., 2011]. Inputs also arise from the medullary, pontine,

and mesencephalic reticular formation, parabrachial nucleus, laterodorsal and pedun-

culopontine tegmental nuclei, dorsal raphe, intermediate layers of the contralateral

superior colliculus, substantia nigra, lateral habenula (lHb), zona incerta, lateral and

dorsal hypothalamic areas, preoptic region, parts of the extended amygdala (with the

notable exception of the central nucleus of the amygdala), ventral striatopallidum,

and frontal cortex [Kau�ing et al., 2009, Jhou et al., 2009c, Lavezzi and Zahm, 2011]

(Fig. 1.4).

1.2.2 Cellular Pro�le

Morphological analysis of the RMTg reveals that this structure is a heterogeneous pop-

ulation of cells comprised of small and medium-sized neurons [Jhou et al., 2009b]. The

23

Figure 1.4: This schematic illustrates some of the major tVTA inputs andoutputs. (A)RMTg a�erents are widely dispersed and originate from abroad range of structureswithin the forebrain, diencephalon, midbrain and brainstem. (B) RMTg e�erents pref-erentially target dopaminergic neurons of the VTA/SNc complex and neurons of theLH. Abbreviations: BNST, bed nucleus of the stria terminalis; Cg, cingulate cortex;CPu, caudate putamen; DMTg, dorsomedial tegmental area; DP, dorsal peduncularcortex; DR, dorsal raphe; IL; infralimbic cortex; IPN, interpeduncular nucleus; LC,locus coeruleus; LH, lateral hypothalamus; LHb, lateral habenula; LPOA, lateral pre-optic area; LDTg, laterodorsal tegmental nucleus; MR, median raphe; NAc, nucleusaccumbens; Pa, parafascicular thalamic nucleus; PAG, periaqueductal gray; PMnR,paramedian raphe nucleus; Pn, pontine reticular nucleus; PPTg, pedunculopontinetegmental nucleus; PrL, Prelimbic cortex; RN, red nucleus; RRF, retrorubral �eld;SC, superior colliculus; SI, substantia innominate; SNc, substantia nigra pars com-pacta; VP, ventral pallidum; VTA, ventral tegmental area; ZI, zona incerta. Adaptedfrom Fakhoury et. al, 2018.

vast majority of these neurons ( 70�92 %) are GABAergic [Jhou et al., 2009b, Kau-

�ing et al., 2009] and expressed Fos in response to several aversive stimuli [Jhou

et al., 2009a] and psychostimulants [Perrotti et al., 2005, Kau�ing et al., 2009, Kau-

�ing et al., 2010]. With some heterogeneity, RMTg neurons are also enriched in opioid

receptor, somatostatin and prepronociceptin [Jhou et al., 2009a]. More recently, a ge-

netic marker of the RMTg, the transcription factor FoxP1 , has been identi�ed both

in mice [Lahti et al., 2016] and rats [Smith et al., 2019a], and has been shown to

be selectively expressed in the RMTg relative to surrounding areas such as the VTA

(Fig. 1.5).

In vivo, the mean �ring rate of RMTg neurons is between 11 and 18 Hz, the

24

Figure 1.5: Distribution of FoxP1 neurons across the rostrocaudal axis of the ven-tral midbrain, as compared to the distribution of dopamine (TH-expressing) andPPTg acetylcholine (NOS-expressing) neurons in rats. Graph shows number of neu-rons immunoreactive for FoxP1, TH, and NOS in 40-um coronal sections at speci�canterior-posterior (AP) distances based on the rostral edge of the interpeduncularnucleus (IPN). Adapted from Smith et. al, 2019.

interspike interval is around 73 ms, and the mean duration of their action poten-

tial measured from peak to trough is roughly 1 ms [Jhou et al., 2009a, Jalabert

et al., 2011, Lecca et al., 2011]. In vitro, RMTg neurons �re slower ( 5 Hz), and

the excitatory postsynaptic currents (EPSCs) evoked in this region are a result of

the presence of amino-3-hydroxy-5-methyl-4-isoxazolepropionate (AMPA) receptors

25

on RMTg neurons [Lecca et al., 2011].

1.2.3 Distinction Between RMTg and GABA Neurons of the

VTA

Although some studies have shown similar functions between RMTg and VTA GABA

neurons, for example that they have similar inhibitory in�uences on DA neurons in

mediating reward prediction errors [Hong et al., 2011, Jhou et al., 2009a, Stopper

et al., 2014, Eshel et al., 2015], and that activation of either RMTg or VTA GABAer-

gic neurons induce behavioral avoidance [Jhou et al., 2009a, Lammel et al., 2012, Tan

et al., 2012, Stamatakis and Stuber, 2012], several lines of evidence suggest a dis-

tinct function of these two populations. For example, mu opioid agonists are more

reinforcing when infused into the RMTg compared to VTA, and agonists of these

receptors block inhibition of DA neurons induced by stimulation of the RMTg but

not VTA interneurons [Jalabert et al., 2011, Matsui and Williams, 2011, Jhou et al.,

2012, Matsui et al., 2014, Zangen et al., 2002].

Moreoever, to address the physiological di�erences between RMTg and VTA

GABAergic neurons, a series of studies from Uchida's group have recorded geneti-

cally identi�ed GABAergic neurons in the VTA and showed that they display ramping

activities in responses to reward cues and that they remain activated until animals

receive rewards [Cohen et al., 2012, Eshel et al., 2015], contrasting with RMTg neu-

rons that were inhibited by reward cues and then quickly returned to baseline �ring

levels. Moreover, optogenetic inhibition of VTA GABA interneurons disrupts phasic

DA responses to reward cues and results in a sustained DA activation throughout

the entire presence of cues [Eshel et al., 2015], whereas RMTg lesions removed most

of the aversion-induced inhibition in VTA neurons without a�ecting the proportion

26

or responses of VTA neurons to reward cues [Li et al., 2019b]. RMTg neurons also

predominantly show negative valence encoding (i.e. activation by aversive stimuli and

inhibition by reward-predictive cues) which is not described in VTA neurons. Finally,

RMTg neurons are enriched in Pnoc and FoxP1, whereas the majority of GABAergic

neurons in the VTA lack those markers [Smith et al., 2019b].

Taken together, these results suggest that RMTg does not appear to be simply

a caudal extension of VTA GABAergic neurons, and conveys di�erent information

onto DA neurons than VTA GABAergic neurons, while potentially mediating distinct

aspects of motivated behaviors.

1.2.4 RMTg Function

Several studies have suggested an important RMTg role in conveying information

about aversive stimuli onto VTA and substantia nigra DA neurons and mediating be-

havioral responses to these aversive stimuli [Hong et al., 2011, Jhou et al., 2009a, Jhou

et al., 2013, Vento et al., 2017, Brown et al., 2017]. In fact, RMTg neurons are inhib-

ited by rewards and reward-predictive stimuli, and activated by aversive stimuli and

their predictors, as well as unexpected omission of reward [Hong et al., 2011, Jhou

et al., 2009a, Li et al., 2019a], opposing VTA DA neurons but mimicking lHb glu-

tamatergic neurons [Matsumoto and Hikosaka, 2007, Cohen et al., 2012, Jhou et al.,

2009a]. These response patterns are consistent with the RMTg playing roles in loco-

motor inhibition and aversion learning, which appear to oppose DA roles in locomotor

activation and reward learning. Indeed, evidence strongly suggests that the RMTg

plays a critical role in aversive processing and mediates aversive behaviors through its

inhibitory projections to VTA DA neurons. Our lab and others have demonstrated

that RMTg neurons are activated by several aversive stimuli of widely varying sen-

27

sory modalities and timescales, including footshock, loud siren, bright light, lihium

chloride, retraint stress and cocaine. RMTg drove VTA inhibitions [Balcita-Pedicino

et al., 2011, Bourdy et al., 2014], and conditioned place aversion (CPA) to these aver-

sive stimuli, without a�ecting VTA activation or CPP to rewarding stimuli [Li et al.,

2019a]. Moreover, lesions of the RMTg were shown to decrease open arm avoidance

on an elevated plus maze, suggesting a potential role for the RMTg in anxiety-like

behavior [Jhou et al., 2009a].

In addition to motivated behavior, the RMTg regulates motor activity trough

the inhibition of the SNpc and subsequent reduction of DA release in the dorsal

striatum [Bourdy et al., 2014]. Consistent with this, lesioning of the RMTg resulted

in disinhibition of the DArgic system, increased DA in the dorsal striatum and motor

excitation.

Combining its role in motivated behavior and locomotion, the RMTg seems to

play a strong role in promoting inhibitory avoidance behaviors in response to aversive

stimuli [Jhou et al., 2009a]. Indeed, activation or inhibition of the RMTg to VTA

pathway produces conditioned place avoidance and preference, respectively [Smith

et al., 2019a]. Furthermore, lesion or inactivation of RMTg neurons are found to

reduce many aversive behaviors including conditioned place avoidance, fear-induced

freezing behavior, nociceptive response, punishment resistance, acute and chronic

withdrawal from abused drugs [Jhou et al., 2009a, Jhou et al., 2013, Jhou et al.,

2012, Smith et al., 2019a, Vento et al., 2017, Elmer et al., 2019, Greco et al., 2008,

Kau�ing and Aston-Jones, 2015], while sparing active fear responses. For instance,

excitotoxic lesions or optogenetic inactivation of the RMTg at a period when cocaine

exhibits aversive e�ects abolish cocaine induced avoidance behaviors in a runway

operant paradigm [Jhou et al., 2013], whereas optogenetic activation of LHb inputs

to the RMTg increases behavioral avoidance and decreases positive reinforcement in

28

mice [Stamatakis and Stuber, 2012]. These �ndings lead to speculation that the

RMTg is involved in conveying information about negative reward signals and as a

result leading to behavioral inhibitory responses.

Therefore, the RMTg is implicated together with the VTA in behavioral respond-

ing to reward omission, aversive stimuli, fear-eliciting stimuli, and withdrawal from

drugs of abuse [Jhou et al., 2009a, Jhou et al., 2013, Hong et al., 2011, Lecca et al.,

2011, Lecca et al., 2012, Matsui and Williams, 2011, Glover et al., 2016, Fu et al.,

2016, Wasserman et al., 2013] and with the SNc in motor skill and motor learning

[Bourdy et al., 2014, Lavezzi et al., 2015]. It thus conveys information regarding the

aversive qualities of stimuli to DA neurons [Ungless et al., 2004, Brischoux et al.,

2009]. RMTg's response to aversive stimuli or the cue associated to that stimlus, and

the resulting inhibition of DA cells may be pivotal in promoting avoidance learning.

Indeed, this circuit could provide an aversive teaching signal that punishes actions

occurring in the recent past, and its responses to cues may inhibit the ongoing seek-

ing of rewards to be obtained in the immediate future [Vento et al., 2017]. We can

thus think of the RMTg as a converging nucleus, integrating multimodal internal and

external information, and conveys the output to DA neurons of VTA and SNc to

modulate motor/motivational responses.

1.3 Aversive Properties of Drugs of Abuse

1.3.1 Opponent Process Theory and Cocaine

Human users of cocaine describe two distinct a�ective consequences of drug adminis-

tration: an initial and profound euphoria usually followed by a dysphoric state char-

acterized by anxiety, fatigue, agitation, depression, and oftentimes cravings for more

29

cocaine [Kampman et al., 1998, Anthony et al., 1989, Mulvaney et al., 1999, Resnick

et al., 1977a, Rohsenow et al., 2007, Spotts and Shontz, 1984, Williamson et al.,

1996a]. Clinical studies report that the negative properties of the drug can even

precipitate episodes of panic attack and other related anxiogenic states [Cox et al.,

1990, Geracioti and Post, 1991, Goodwin et al., 2002, Wal�sh et al., 1990]. These

aversive e�ects may strongly modulate drug-seeking, and are observed in almost all

drugs of abuse, including psychostimulants, opioids, nicotine, and alcohol [Solomon

and Corbit, 1974, Koob et al., 1989, Panlilio et al., 1996, Ettenberg et al., 1999, Kenny

and Markou, 2001, Rothwell et al., 2009, Riley, 2011]. The mechanisms of the aver-

sive e�ects have been thought to result from �opponent� processes where aversive

e�ects work against, and outlast, the initial rewarding e�ects (Fig. 1.6) [Solomon

and Corbit, 1974].

The notion that cocaine has negative or aversive side e�ects has also been sug-

gested by the results of animal research studies using diverse paradigms such as CPP

and CPA, [Ettenberg and Bernardi, 2007, Ettenberg et al., 1999], CTA [Verendeev

and Riley, 2012], and operant paradigms [Ettenberg, 2009, Rezvani et al., 2010, Bar-

rot et al., 2012]. For example, over two decades ago, [Spealman, 1979] reported that

monkeys trained to lever-press for intravenous cocaine would also choose to press

a second lever that terminated drug availability. In more recent studies, cocaine

has been observed to decrease punished responding in a rodent con�ict test[Fontana

and Commissaris, 1989], to have anxiogenic e�ects [Prather and Lal, 1992], induce

anxiogenic-patterns of exploratory behavior in the elevated plus-maze [Rogerio and

Takahashi, 1992], to increase animals' latency to enter an open �eld [Yang et al., 1992],

and to increase latencies to reach a cocaine-paired goalbox in an operant runway task

[Ettenberg and Geist, 1993].

We now know that cocaine's motivational properties are biphasic, with rewarding

30

Figure 1.6: A schematic representation of the a�ective response to an injection ofcocaine as predicted from Opponent Process Theory. Following administration thereis an initial positive reaction that rises quickly to a short-lived peak level (time point`A'), and then decreases to a steady state level that eventually gives way and dropsback toward initial baseline (neutral) levels. The delayed onset of a second opposingnegative/aversive state is in part responsible for the reduction in positive a�ect andin time exerts a net negative impact on the subject's a�ective experience (time point`B'). Adapted from Ettenberg et. al, 2004.

e�ects most prominent within the �rst 10 minutes after individual intravenous (iv)

infusions, while aversive properties are prominent slightly later, beginning roughly 15

minutes post-infusion, suggesting that drugs of abuse are complex pharmacological

compounds with multiple stimulus e�ects and both the rewarding and the aversive

e�ects of drugs should be taken into consideration in ongoing attempts to model

drug-taking behavior.

31

1.3.2 RMTg and Aversion

Recent studies have identi�ed cocaine-induced activation of a dense, previously over-

looked GABAergic input to midbrain DA neurons arising from the RMTg. A number

of studies have implicated the RMTg in processing the aversive properties of drugs of

abuse including psychostimulants [Kau�ing et al., 2009, Kau�ing et al., 2010, Jhou

et al., 2009a, Jhou et al., 2013, Zahm et al., 2010], opiates [Lecca et al., 2011, Lecca

et al., 2012, Matsui and Williams, 2011] and alcohol [Glover et al., 2016, Fu et al.,

2016]. This activation has been demonstrated both through cfos induction [Kau�ing

et al., 2010] as well as phasic activations [Lecca et al., 2011, Li et al., 2019a] to acute

exposure to a variety of drugs of abuse. Speci�cally, RMTg has been shown to have a

biphasic response after an acute dose of cocaine, with an initial inhibition followed by

delayed excitation paralleling cocaine's shift from rewarding to aversive. Moreover,

stimulation of either the LHb or RMTg inhibits 82�97 % of DA neurons [Hong et al.,

2011, Christoph et al., 1986, Ji and Shepard, 2007, Matsumoto and Hikosaka, 2007]

and contributes to aversive conditioning [Lammel et al., 2012, Stamatakis and Stu-

ber, 2012]. Indeed, direct activation of the RMTg (by AMPA) is su�cient to produce

aversive conditioning, while cocaine-conditioned avoidance responses are abolished

by lesions of the FR or RMTg, or by inhibition of RMTg activity 15�25min after

cocaine delivery [Jhou et al., 2013]. However, despite these studies, RMTg's molecu-

lar substrates of cocaine's aversive e�ects, and of their variation between individuals,

are almost entirely unknown, and a better understanding of the role of RMTg in

modulating drug aversion's role in drug-taking is crucial for �nding new avenues for

interventions.

32

1.3.3 Accumbens and Aversion

The mesolimbic DArgic system not only plays a pivotal role in a wide range of mo-

tivation and learning [Wise, 2004, Schultz, 2007, Bromberg-Martin et al., 2010], but

its dysfunction has also been implicated in severe neuropsychiatric disorders, and

has been widely implicated in modulating drug addiction. DA neurons in the VTA

react to rewarding stimuli by phasic �ring, and the main function of this �ring is

suggested to encode reward prediction error, [Schultz et al., 1997]. In contrast to the

response to rewarding stimuli, DA neurons' response to aversive stimuli are far from

homogeneous; i.e., some DA neurons are activated in response to aversive stimuli,

whereas most others react by decreasing their �rings [Cohen et al., 2012, Ungless

et al., 2004, Brischoux et al., 2009, Matsumoto and Hikosaka, 2009]. In fact, recent

studies have revealed that optogenetic activation of GABAergic neurons and resultant

inactivation of DA neurons suppresses reward consumption and induce an aversive

response [Tan et al., 2012, van Zessen et al., 2012]. However, it has largely remained

elusive as to which mechanisms in the neural circuits are essential for the acquisition

of aversive learning following the inactivation of DA neurons in the VTA and as to

how behavioral responses are controlled toward suppressing reward consumption and

inducing aversive behaviors. The general understanding is that DA signals from the

midbrain dynamically modulate the direct and indirect pathways in opposite direc-

tions via Dopamine D1 receptor (D1Rs) and Dopamine D2 receptor (D2Rs), and this

modulation is supposed to facilitate motivational learning [Bromberg-Martin et al.,

2010, Surmeier et al., 2009]. As for the rewarding stimuli, up-regulated DA levels

induced by rewarding signals activate the D1Rs and thus excite the direct pathway

in the nucleus accumbens (NAc). On the other hand, the suppression of DA neuron

�rings in response to aversive stimuli decreases DA levels in the NAc which results

33

in increase of signal transmission in the indirect pathway through activated D2Rs.

And while VTA DA neurons project to multiple brain areas, the most prominent

pathway involved in this bidirectional encoding of reward and aversive stimuli is be-

lieved to be encoded in the ventral striatum. For example, a recent elegant study

by Roitman et al. stimulated a sensory pathway that can carry both positive and

negative information to show that opposite dopamine state in the NAshell is key

for encoding opposing valences stimuli. Particularly, Roitman et al., demonstrated

that sucrose rewards increase, while the aversive quinine suppressed the frequency of

dopamine transients in the NAshell [Roitman et al., 2008]. Several studies using the

pharmacological strategies and reversible neurotransmission blocking (RNB) method

have supported this mechanism of regulation in the NAc [Hikida et al., 2010, Hikida

et al., 2013], and it has been shown that this aversive learning is mediated through

D2. One study in particular showed that selective optogenetic inactivation of DA

neurons in the VTA decreased DA levels in the NAc which induced aversive reaction

and learning through D2 activation [Danjo et al., 2014]. Interestingly the optoge-

netic up-regulation of D2R-expressing medium spiny neurons (MSNs) in the indirect

pathway in the NAshell evokes behavioral avoidance [Kravitz et al., 2012]. To ex-

tend those �ndings to cocaine, acute cessation of cocaine intake causes accumbens

DA levels to fall below baseline [Panlilio et al., 1996] and decreased 30�40 % during

cocaine withdrawal compared with pre-session levels, whereas reduced accumbens DA

signaling in turn produces CPA [Shippenberg et al., 1991]. But despite this observa-

tion, it is still not clear whether the bi-directional encoding of reward and aversive

behavioral response in the NAshell is true for cocaine, and what are the implications

for D2 MSNs, and to what extent the di�erences in this process correlate with indi-

vidual di�erences in cocaine avoidance. The fact that the NAc also plays a role in

aversion seems to still be underappreciate and requires extensive studies for a more

34

comprehensive understanding of the circuitry of aversion.

1.3.4 Implications in Addiction

Over 90 of Americans have had some exposure to drugs of abuse, but only 15-32 of

individuals exposed to the major classes of abused drugs (nicotine, alcohol, cocaine,

heroin) go on to become addicted, with the rest presumably being able to stop on

their own, before progressing to addiction [Wagner-Echeagaray et al., 1994, Wenzel

et al., 2011, Heyman, 2013], suggesting that not all individuals are equally vulnerable

to developing addiction-like behavior [Deroche-Gamonet and Piazza, 2014]. How-

ever, little is known about the neurobiological predisposing factors regulating the

propensity toward addiction. Indeed, the reasons for these individual di�erences are

complex and likely involve multiple areas of vulnerability. Some such areas of dif-

ferential vulnerability could involve di�erences in initial responses to drugs. In fact

much basic research has been directed at understanding individual animals who have

already progressed into addiction-like behaviors, with relatively less study of what

protective factors may help prevent acquisition or escalation of drug use in the �rst

place and hence render some individuals less vulnerable to becoming addicted, one

of those factors being drug-induced aversion. Even after relatively infrequent admin-

istration of cocaine, aversive responses are powerful enough to greatly attenuate or

prevent seeking of drugs or drug cues. These responses have been posited to ex-

plain why many humans and rodents do not progress from occasional to compulsive

drug use despite the drug's strong rewarding properties [Deroche-Gamonet et al.,

2004, Wagner-Echeagaray et al., 1994]. Seminal studies from Dr. Ettenberg and

colleagues have identi�ed key attributes of cocaine's aversive properties that are suf-

�cient to condition a net aversion to cocaine, that in most (but not all) animals, is

35

strong enough to overcome cocaine's rewarding e�ects [Ettenberg et al., 1999, Et-

tenberg, 2004, Jhou et al., 2013, Ettenberg and Bernardi, 2007]. These individual

di�erences in aversive responses to drugs may contribute to di�erences in individual

propensity to develop addictive behaviors. For example, high aversive responses to

drugs may be protective against early acquisition of drug-seeking as individuals seek

to avoid drug-induced aversion, and the underlying mechanism can be exploited to

better understand why some individuals are protected whereas others are not.

Although aversion confers a protective e�ects initially, it is likely to promote drug-

seeking in later stages, as individuals seek to mitigate the aversive e�ects of recent

drug use. Indeed, it has been posited that withdrawal play an important role in the

transition of drug use to the development of dependence. Indeed, some have gone so

far as to argue that the presence of a negative a�ective state is the de�ning feature

of addiction [Koob and Le Moal, 2005, Russell, 1976, Baker, 1987], where those

begative states might substantially contribute to the maintenance of dependence.

The motivation for maintenance of compulsive drug use could now be thought of

requiring both negative reinforcement processes and positive reinforcement processes

where the individual is seeking the drug not only to retrieve the rewarding e�ect, but

also to alleviate the aversive state. Therefore, understanding the mechanisms behind

drug aversion, and the role it plays a di�erent stages, could help provide di�erent

interventions depending on the stage that is being targeted.

1.4 Statement of the Problem

Cocaine addiction is a devastating brain disorder that a�ects the health of millions

of people in Western countries and is associated with high health-care costs [Uhl

and Grow, 2004]. Despite the substantial sum of money spent on treatments for

36

substance use disorder, e�ect sizes of all available therapies o�ered are typically mod-

est with a substantial proportion of patients failing to respond. Psychostimulant in

particular, including cocaine, addiction treatment is di�cult because of the lack of

e�cient medication and other therapies [Jupp and Lawrence, 2010]. Because there

are large individual di�erences in vulnerability to cocaine addiction, novel insights

into possible mechanisms and treatments may arise from studying individuals that

are naturally resilient to cocaine-taking, but the precise di�erences in these individ-

uals has not yet been uncovered. Despite the striking in�uence of drug aversion on

behavior, little is known about its neural mechanisms nor whether these mechanisms

can be targeted for therapeutic e�ect, and this topic as a whole remains relatively un-

derstudied. Elucidation of factors underlying inter-individual di�erences in substance

use behavior, speci�cally aversive e�ects of drugs, could result in the individualiza-

tion of therapies and targeting of speci�c clusters of patients using psychotherapeutic

or pharmacological approaches, which include the prediction of therapeutic response

based on pharmacogenetic pro�les. In this study, I will examine individual di�erences

in cocaine aversion and its role in cocaine addiction vulnerability at three levels: 1)

behavioral; 2) circuitry; 3) molecular.

Particularly, I hypothesize that RMTg neurons in "low- cocaine avoiders" exhibit-

ing weak aversive responses to cocaine will be less excitable during the aversive phase

of cocaine compared to "high-cocaine avoiders" who exhibit strong aversive responses

to cocaine. The resulting di�erence in RMTg-mediated inhibition of DA neurons be-

tween the two populations will in turn result in di�erential adaptations in D2-MSNs

in the NAshell, which will lead to opposite behavioral outputs towards drug-seeking.

37

Figure 1.7: Proposed model of RMTg�>VTA�>NAshell circuit mediating di�erencesin cocaine aversion.

38

Chapter 2

Methods

2.1 Animals and Surgeries

2.1.1 Animals

Adult male Sprague Dawley (Charles River Laboratories, Raleigh, NC, USA), weigh-

ing 250�300g upon delivery from vendor (Charles River Laboratories, Raleigh, NC,

USA; ), were individually housed in standard ventilated shoebox cages within a

temperature-controlled (22 ◦C), 12/12 h light/dark cycle (lights on at 6:00am) vi-

varium, with food and water provided ad libitum, unless otherwise stated. Proce-

dures conformed to the National Institutes of Health Guide for the Care and Use of

Laboratory Animals, and all protocols were approved by Medical University of South

Carolina Institutional Animal Care and Use Committee. All animals were tested

during the light phase.

39

2.1.2 Catheter Surgeries

Rats were anesthetized with iso�urane and �tted with indwelling intravenous catheters

inserted into the right jugular vein, and subcutaneously passed to a guide cannula

exiting the animal's back. After surgery, catheter patency was maintained via daily

�ushing with 0.05 ml of Taurolidine Citrate Solution. Animals recovered for 7 days

prior to behavioral testing. Catheter patency was assessed periodically through obser-

vation of the loss of the righting re�ex after i.v. injection of, methohexital (Brevital,

2.0 mg/kg/0.1 ml). Rats that were unresponsive to Brevital were implanted with a

new catheter using the left jugular vein.

2.1.3 Stereotaxic Surgeries

Dlight: Stereotaxic viral vector injections were performed in mice anesthetized with

iso�urane (1�3 % in 95 % O2/5 % CO2 provided via nose cone at 1 L/min) as previ-

ously described (Cho et al., 2017). Following anesthesia, preparation and sterilization

of the scalp, and exposure of the skull surface, a craniotomy hole was drilled over the

LNAc (antero-posterior: 1.2 mm, medio-lateral: 1.6 mm relative to Bregma). 800 nL

of the AAV9-hSyn-dLight1.2 vector (titer: 4 x 1012 viral genomes/mL, produced at

the UC Davis Vision Center Vector Design and Packaging Core facility; Addgene #

111068) was delivered into the NAshell (antero-posterior: 1.8 mm, medio-lateral: 0.8

mm, dorso-ventral: 7.2 mm relative to Bregma) using a blunt 33-gauge microinjec-

tion needle within a 10 mL microsyringe (NanoFil, World Precision Instruments), a

WPI microsyringe pump (UMP3, World Precision Instruments), and pump controller

(Micro4, World Precision Instruments) over 10 min. Following viral injection, a 5 mm

long, 400 mm outer diameter mono �ber-optic cannula (Doric Lenses Inc) was low-

ered to the same stereotaxic coordinates and a�xed to the skull surface with C and

40

B Metabond (Parkel Inc) and dental cement.

Electrode implants: For recording experiments, drivable electrode arrays were

implanted above the RMTg (AP: -7.4mm; ML: 2.1mm; DV: -7.4mm from dura, 10-

degree angle).

Cannula implants: A doubled 6mm length guide cannula (Plastic One Inc.) was

implanted in RMTg (AP: -7.4mm; ML: 2.1mm; DV: -7.4mm, from dura, 10-degree

angle) or in NAshell (AP: 1.8mm; ML: 2.3mm; DV: -8mm, from dura, 10-degree

angle).

2.2 Behaviors

2.2.1 Runway Operant Cocaine-Seeking

Methods are similar to our previous work [Jhou et al., 2013]. Brie�y, the run-

way consists of two opaque plastic compartments (�start� and �goal�, 25x10x17 cm

length/width/height) connected by a 170cm long corridor with motorized doors be-

tween the start/goal boxes and corridor. Rats are tethered to an intravenous line,

then placed into the start box for 30 seconds, after which doors are opened and allow-

ing free exploration of the apparatus. Entry into the goal box causes doors to close

and a syringe pump to deliver cocaine (0.75mg/kg iv over 10 seconds, same dose as

in CPP and electrophysiological recording). Failure to enter the goal compartment

after 15 minutes results in a timeout. 7 trials are given, at most twice daily and at

least 4 hours apart, ensuring that animals run in a cocaine-free state, during the light

cycle over a period of 4-5days. All animals are habituated to the runway apparatus

before starting cocaine trials; all animals must have at least 2 consecutive habitua-

tion sessions under with run latency<60s before progressing to cocaine 1 session. the

41

criterion for a retreat was a reversal in direction for a distance spanning a minimum

of two photobeams.

2.2.2 Progressive Ratio and Punishment Resistance

Training procedures are identical to those we previously described [Vento et al., 2017].

Brie�y, animals are food restricted to 85 of initial body weight, trained to leverpress

for food pellets (45mg) until stable response rates were reached, then tested either on

the PR task with progressively increasing e�ort requirements (leverpresses required

per pellet) [Richardson and Roberts, 1996] or progressively increasing punishment

(i.e. a footshock of increasing intensity presented 1 second after food pellet delivery,

whose magnitude increases about 25 every 3 trials). In the PR task, rats time out

if two minutes passed without a leverpress. In the punishment task, a timeout is

de�ned by 3 consecutive trials with >30 seconds passing without a leverpress. For

both tasks, the breakpoint is the last successfully completed ratio or shock intensity

endured before timeout.

2.2.3 Novelty-Induced Locomotion

Novelty-induced locomotion was recorded in operant chambers placed inside sound-

attenuating cabinets Movement within the chamber were detected via interruptions

of an array of 4 photodetector/emitter pairs. Locomotor counts were averaged over

two sessions on consecutive days, lasted for 30 minutes each.

42

2.2.4 Conditioned Place Preference/Conditioned Place Aver-

sion

CPP/CPA was assessed using an unbiased protocol. We used a 3-chambered appa-

ratus with two conditioning chambers (having dark versus light colored walls) and a

small neutral center chamber. On day 1, animals were placed in the central chamber

and a baseline was collected for 15 min. Animals received a total of four conditioning

sessions for each task. Rats received an infusion of i.v. cocaine (0.75 mg/kg) on

the paired side (0 min post-infusion for CPP and 15 min post-infusion for CPA) or

Lithium Chloride (0 min post-infusion only) and saline on the unpaired trials (left for

15 min in the chamber). Two session were conducted daily (cocaine/LiCl and saline)

counterbalanced in both order of test (CPP vs CPA) and identity of the drug-paired

chamber (dark versus light walls). One day after �nal conditioning sessions, animals

were tested by placement into the central compartment of the CPP apparatus for 15

min, and time spent in each compartment was recorded.

2.2.5 Shuttlebox Shock Escape

Rats were placed into a shuttlebox having two equal sized compartments and a mo-

torized doorway connecting the two. At two-minute intervals, footshock was delivered

via a �oor grid, which terminated after the rat crossed through the doorway. Pho-

tobeams measured the latency to reach the other side. 10 trials were given for each

session, and average latency recorded for each rat. Shock intensity was �xed during

each session, and two sessions were conducted for each rat at each shock intensity.

Shocks were tested in ascending order: 0mA, 0.15mA, 0.25mA, 0.35Ma, 0.5mA and

0.7mA.

43

2.2.6 Self Administration

After being tested on the runway task, animals were trained on a daily 2h session of

cocaine self-administration during the light cycle, in standard Med Associates operant

chambers equipped with two levers, a house light, cue light, and tone generator.

The designated active lever delivered an infusion of cocaine (0.75 mg/kg/infusion)

dissolved in sterile 0.9 saline, along with a compound cue (light + tone) on an FR1

schedule of reinforcement. Following each infusion, a 20 s time out was signaled by

the loss of illumination of the house light where active lever presses resulted in no

programmed consequence. Responses on the inactive lever were recorded but had no

programmed consequence. Two days prior to the start of training, rats were food-

restricted to 15 g of chow per day to facilitate acquisition of SA. All rats received

ad libitum access to food once SA training was done. During self-administration

training, for rats that did not spontaneously acquire lever-pressing within 3 sessions,

the active lever was baited with a chow food pellet. Once rats reached stability (10

or more infusions) foo baiting was discontinued. After a minimum of 10 days of self-

administration >15 infusions, including the last 3 days, and discrimination between

the active and inactive levers, animals moved into extinction. During extinction

training, rats were exposed to the SA environment for a 2-h session each day across

10 consecutive days. During the extinction sessions, lever presses had no scheduled

consequences; ie no longer produced drug or presentation of the drug-paired cues. and

the responses on the active and inactive levers were recorded. Rats received extinction

training on a minimum of 7-8 consecutive days until the responding on the active

lever fell to <10 of the active lever pressing during maintenance for 3 consecutive

days. The day after achieving this criterion, subjects were returned to the operant

chambers for reinstatement testing. Reinstatement testing began following extinction

44

training. For cue reinstatement, active lever presses resulted in the presentation of

the cues (both tone and light), but no cocaine infusions. For cocaine reinstatement,

rats received a systemic cocaine injection (10 mg/kg, ip) and were immediately placed

in the chambers. Cocaine-seeking behavior was operationally de�ned as responses on

the active lever in the absence of cocaine reinforcement [Beardsley and Shelton, 2012].

2.2.7 Elevated Plus Maze

The maze consisted of two open (50 x 20 cm) and two enclosed arms (50 x 20 x 30

cm) and an open roof with the entire maze elevated 100 cm from the �oor. Each rat

was initially placed in the center of EPM (facing toward open arm) to explore freely.

Rats' behavior was recorded for 3 min by a video camera mounted centrally above the

maze. The time spent in open vs closed arms of the maze was recorded and reported

by a blind experimenter.

2.2.8 Pavlovian Training

Behavior was conducted in standard Med Associates chambers (St. Albans, VT).

Food-predictive and neutral cues were a 1 kHz tone (75 dB) and white noise (75dB),

respectively. The food-predictive cue was presented for 2s, and a food pellet (45

mg, BioServ) was delivered immediately after cue o�set. The neutral cue was also

presented for 2s, but no food pellet was delivered. The two trial types were randomly

presented with a 30s interval between successive trials. A �correct� response was

scored if the animal either entered the food tray within 2 s after reward cues or

withheld a response for 2 s after neutral cues. Rats were trained with 100 trials per

session, one session per day, until they achieved 85 % accuracy in any 20-trial block,

after which animals were then trained with one extra session in which neutral cue

45

trials were replaced by shock trials consisting of a 2s 8kHz tone (75dB) followed by a

10ms 0.7mA footshock.

2.3 In-vivo Recording

2.3.1 Wire Recording

Electrodes consisted of a bundle of sixteen 18 um Formvar-insulated nichrome wires

(A-M system) attached to a customized printed circuit board (Seeed Studio, Shen-

zhen, China). Electrodes were grounded through a 37-gauge wire attached to a gold-

plated pin (Mill-max, part # 1001-0-15-15-30-27-04-0), which was implanted into the

overlying cortex. Recordings were performed during once-daily sessions, 1-5 days af-

ter animals' last runway session. Electrodes were advanced 80-160 um at the end of

each session. A 15 minutes baseline was collected, after which animals were given a

single infusion of cocaine (0.750mg/kg, i.v.). The recording apparatus consisted of

a unity gain headstage (Neurosys LLC) whose output was fed to preampli�ers with

high-pass and low-pass �lter cuto�s of 300 Hz and 6 kHz, respectively. Analog signals

were converted to 18-bit values at a frequency of 15.625 kHz using a PCI card (Na-

tional Instruments) controlled by customized acquisition software (Neurosys LLC).

Spikes were initially detected via thresholding to remove signals less than twofold

above background noise levels, and signals were further processed using principal

component analysis performed by NeuroSorter software. Spikes were accepted only

if they had a refractory period, determined by <0.2 % of spikes occurring within

1ms of a previous spike, as well as by the presence of a large central notch in the

auto-correlogram. Neurons that had signi�cant drifts in �ring rates were excluded.

Drivable electrode arrays were implanted above the RMTg (AP: -7.4mm; ML: 2.1mm;

46

DV: -7.4mm from dura, 10-degree angle). Placements at RMTg were veri�ed with

FoxP1 stain.

2.3.2 Fiberphotometry

Fiber photometry was used to monitor �uorescent dopamine signals using a custom

system as previously described (Cho et al., 2017; Patriarchi et al., 2018), which al-

lowed for dLight1.2 excitation and emission light to be delivered and collected via the

same implanted optical �ber. Our system employed a 465 nm LED (M490F1, Thor-

labs, Inc; �ltered with FF02-472/30-25, Semrock) for �uorophore excitation and a 405

nm LED for isosbestic excitation (M405F1, Thorlabs Inc; �ltered with FF01-400/40-

25, Semrock), which were modulated at 211 Hz and 531 Hz, respectively, controlled

by a real-time processor (RX8-2, Tucker David Technologies), and delivered to the

implanted optical �ber via a 0.48 NA, 400 mm diameter mono �ber optic patch cable

(Doric Lenses Inc). The emission signal from isosbestic excitation, which has previ-

ously been shown to be calcium independent for GCaMP sensors (Kim et al., 2016;

Lerner et al., 2015), was used as a reference signal to account for motion artifacts

and photo-bleaching. Emitted light was collected via the patch cable, collimated,

�ltered (MF525-39, Thorlabs), and detected by a femto-Watt photoreceiver (Model

2151, Newport Co.) after passing through a focusing lens (62� 561, Edmunds Op-

tics). Photoreceiver signals were demodulated into dLight1.2 and control (isosbestic)

signals, digitized (sampling rate: 382 Hz), and low-pass �ltered at 25 Hz using a

second-order Butterworth �lter with zero-phase distortion. A least-squares linear �t

was applied for the 405 nm signal to be aligned with the 465 nm signal. Then, the

�tted 405 nm signal was subtracted from 465 nm channel, and then divided by the

�tted 405 nm signal to calculate DF/F values. Rats received three di�erent types

47

of trials that were randomly selected: reward trial, shock trial, and neutral trial. In

reward trials, a two-second auditory tone (1kHz) predicts a sucrose pellet delivery,

and reward delivery. In shock trials, a distinct auditory tone (8kHz) predicts a mild

footshock (0.7mA, lasting for 10ms). In neutral trials, a distinct auditory tone (1kHz)

was delivered with no consequences.

2.4 Histology

2.4.1 Immunostaining

Rats were transcardially perfused with saline and 10 % formalin. Brains were removed

and post�xed overnight in 10 % formalin, equilibrated in 20 % sucrose, sliced into 40

µm sections in 4 or 6 series then incubated in primary antibody or mounted for Nissl

stain.

Cfos: Rats that were classi�ed into high- and low-avoiders from their runway la-

tencies were transcardially perfused with saline and 10 oercent formalin 1h after giving

an i.p. injection of cocaine (10mg/kg), 2-5 days after their last runway latency. and

immunostained for c-Fos induced by cocaine using rabbit anti-c-Fos (1:10K, EMD

Millipore, MA, USA)in PBS with 0.25 % Triton-X and 0.05 % sodium azide. After-

ward, tissue was washed three times in PBS and incubated in biotinylated donkey

anti-rabbit secondary (1:1000 dilution, Jackson Immunoresearch, West Grove, PA)

for 30 min, followed by three 30s rinses in PBS, followed by 1 hour in avidin-biotin

complex (Vector). Tissue was then rinsed in sodium acetate bu�er (0.1M, pH 7.4), fol-

lowed by incubation for 5 min in 1 % diaminobenzidine (DAB). Nickel and hydrogen

peroxide (Vector) were added to reveal a blue-black reaction product.

FoxP1: Free-�oating sections were immunostained for FoxP1 by overnight incu-

48

bation in rabbit anti- FoxP1 (Abcam, ab16645, 1: 50,000 dilution) primary in PBS

with 0.25 % Triton-X and 0.05 % sodium azide. Afterward, tissue was washed three

times in PBS and incubated in biotinylated donkey anti-rabbit secondary (1:1000 di-

lution, Jackson Immunoresearch, West Grove, PA) for 30 min, followed by three 30s

rinses in PBS, followed by 1 hour in avidin-biotin complex (Vector). Tissue was then

rinsed in sodium acetate bu�er (0.1M, pH 7.4), followed by incubation for 5 min in

1 % diaminobenzidine (DAB). Nickel and hydrogen peroxide (Vector) were added to

reveal a blue-black reaction product.

GFP: Free-�oating sections were immunostained for GFP by overnight incubation

in rabbit anti- GFP (1: 1000 dilution) primary in PBS with 0.25 % Triton-X and 0.05

% sodium azide. Afterward, tissue was washed three times in PBS and incubated in

biotinylated donkey anti-rabbit secondary (1:1000 dilution, Jackson Immunoresearch,

West Grove, PA) for 30 min

2.4.2 Western Blot

Western blotting was used to analyze expression levels of total GluR1 subunit (t-

GluR1) and phosphorylated GluR1 subunit at serine831 site (p-GluR1-S831), total

GluR2, and PTEN in the RMTg ,and A2A and D2 receptors in the NAshell, of

high-avoiders, low-avoiders and control. Brie�y, rats were sacri�ced 24 hours after

their last runway session and their brains were removed rapidly, and then the region

of the brain that contains the RMTg (AP, � 6.8mm to � 7.4mm from bregma) was

sectioned and collected the RMTg and NAC shell using a punch according to the

anatomic marks [Paxinos and Watson, 2006] and stored frozen at � 80C until pro-

cessed. On the day of the experiment, tissue samples were homogenized on ice in

radio-immunoprecipitation assay bu�er supplemented with protease inhibitors. The

49

homogenates were centrifuged at 15,000rpm for 15min at 4C, and supernatants were

collected for further analysis. For all samples, protein concentration was quanti�ed

using a Pierce BCA Protein Assay Kit (Thermo Scienti�c). Samples were denatured

by boiling before being loaded (80ug) on the10 Tris-Glycine gels (Invitrogen). After

electrophoresis, the proteins were transferred to polyvinylidene �uoride membranes

(Millipore, Billerica, MA, USA) and non-speci�c protein binding was blocked with 5

% non-fat milk for 1hat room temperature, followed by an overnight incubation at

4C with the following primary antibodies: anti-t-GluR1 (rabbit monoclonal, 1:1000;

Abcam, Cambridge, MA, USA), anti-p-GluR1-S831 (rabbit monoclonal, 1:2000; Ab-

cam), anti-pTEN, anti-A2A; anti-D2 and anti-glyceraldehyde-3-phosphate dehydro-

genase (GAPDH; mouse monoclonal, 1:1000; Santa Cruz Biotechnology, Santa Cruz,

CA, USA). The next day, the membranes were washed several times and then then in-

cubated with �uorescent goat anti-rabbit or anti-mouse secondary antibodies (1:5000,

IR-dye 680 or IR-dye 800), before being imaged on an Odyssey �uorescent scanner

(Li-cor Biosciences). To ensure equal loading, GAPDH expression was used as a

loading control. Separate blots were used for each protein of interest.

2.4.3 Activated RhoA Assay

RhoA activation was determined using a G-LISA (# BK124; Cytoskeleton, Denver,

CO) as previously described [Fox et al., 2018]. Brie�y, tissue was collected 24 hours

after the last runway session from high-avoiders, low-avoiders and control animals.

RMTg was extracted and homogenized with a Pellet Pestle (Thermo Fisher Scienti�c,

Waltham, MA) in lysis bu�er, homogenates were centrifuged, and supernatants were

snap frozen at -80C. Protein concentrations were equalized with lysis bu�er, and

active RhoA determined according to manufacturer's instructions.

50

2.5 Statistical Analysis

All statistical analyses were completed using GraphPad Prism 7.00 and SAS v9.4.

We indicate p values <0.0001 as p < 0.0001 and provide exact p values for those

>0.0001. The level of signi�cance was set at p<0.05. All values were expressed as

mean ± SEM.

Since the runway latency data do not meet the assumptions of normality, the

corresponding data was log-transformed prior to hypothesis testing, which was con-

ducted based on a general linear mixed modelling (GLMM) approach. GLMMs are

especially useful for modeling longitudinal data by accounting for within-subject cor-

relation, in this case using random mouse e�ects. The GLMMs contained �xed e�ects

condition, sex, trial, and a generation by trial interaction term. For analyses of la-

tency data pooled across trials 4-7, Kruskal-Wallis tests followed by a Mann-Whitney

U test for pairwise comparisons were used. analyzed by non-parametric statistical

tests. Runway data by trial was analyzed using a repeated-measures ANOVA on

rank-transformed latency data. For analyses with one variable, Kruskal-Wallis test

followed by a Mann-Whitney U test for pairwise comparisons.

Correlations were calculated according to the Spearman method. For the rest

of the behavioral data, student's t-tests and one-way ANOVA were used to assess

di�erences in group means, and two-way ANOVA was used with more than one vari-

able. Main e�ects analyses from one-way ANOVA, two-way ANOVA or Students

t-tests were followed with Bonferroni or Tukey post hoc tests to adjust for multiple

comparisons.

51

Chapter 3

Behavioral Characterization of

Runway Model

3.1 Introduction

Although many people experiment with drugs at some point in their lives, far fewer

develop subsequent substance use disorders. Indeed, although addictive disorders,

de�ned in the Diagnostic and Statistical Manual of Mental Disorders, 5th Edition

(DSM5) as use disorders, are contingent on availability and use of an addictive agent,

people are di�erentially vulnerable at multiple stages of the cycle of addiction, includ-

ing the initial propensity to use, transition from use to abuse, long-term addiction, and

relapse [Kwako et al., 2016]. Marked variability in addictive vulnerability has spurred

strong interest in identifying predisposing factors, risk factors and physiological pro-

cesses that underlie drug abuse and addiction [Kalivas and Volkow, 2005, Koob et al.,

2004, Olmstead, 2006, Bau et al., 2001, Belin et al., 2008, Crabbe, 2010, Stead et al.,

2006], with the goal of developing interventions to avert abuse later in life. Indeed,

during the last few decades, there has been a considerable increase in the number of

52

studies assessing the neurobiological basis of individual di�erences in vulnerability to

drug abuse. However, much basic research has been directed at understanding individ-

ual animals who have already progressed into addiction-like behaviors, with relatively

less study of what protective factors may help prevent acquisition or escalation of drug

use in the �rst place and hence render some individuals less vulnerable to becoming

addicted. Nonetheless, a number of vulnerability factors have been examined. For

example, prior work from Huda Akil and others has shown that locomotor response

to novelty is heritable, and predictive of the rate of acquisition of drug-seeking and

that that individual di�erences in response to novelty have been repeatedly linked to

drug abuse [Stead et al., 2006, Cain et al., 2005]. Moreover, impulsivity is a well-

established predictor of drug self-administration in laboratory animals and is related

to abuse vulnerability in humans [Perry and Carroll, 2008]. Others have shown people

who score higher on the Zuckerman sensation-seeking questionnaire or the Cloninger

novelty-seeking questionnaire also report more frequent drug use [Zuckerman et al.,

1993] and are more sensitive to the subjective and reinforcing e�ects of drugs. How-

ever, considerable unexplained variance remains, suggesting that these phenotypes

are just some of many possible protective elements against/predictors of addiction

vulnerability. We hypothesize that drug-induced aversion also plays major roles in

drug-resistance.

Since the �rst experimental demonstration that a drug of abuse supports instru-

mental behavior, drugs have been discussed in the context of their rewarding e�ects,

which are assumed to drive and maintain drug-taking behavior. Indeed, drug reward

has been fundamental in the formulation of most models of drug use, abuse, and

addiction. Over the last several decades, however, drugs of abuse have been increas-

ingly recognized as complex pharmacological compounds producing multiple stimulus

e�ects, not all of which are rewarding [Wise et al., 1976, Stolerman, 1992, Koob and

53

Le Moal, 1997, Verendeev and Riley, 2012]. The aversive e�ects of such drugs, for

example, have been described by a number of researchers working in the �eld, al-

though few attempts have been made to investigate the role of these aversive e�ects

in drug taking. Recently, the aversive e�ects of drugs of abuse, that have been well

described for all major classes of drugs of abuse [Verendeev and Riley, 2012] and have

been found to occur independent of the drugs' rewarding e�ects [Verendeev and Ri-

ley, 2012], have also been increasingly recognized [Verendeev and Riley, 2012, Goudie,

1979, Gaiardi et al., 1991a, Riley, 2011]. In particular, cocaine's aversive e�ects are

experimentally robust, having been studied for decades by several groups [Geist and

Ettenberg, 1997b, Ettenberg et al., 1999, Ettenberg et al., 2015, Jhou et al., 2013].

Particularly elegant experiments by Ettenberg's group have shown that single doses

of cocaine produce an initial rewarding phase followed by an aversive crash about 15

minutes later [Ettenberg et al., 1999] that is su�cient to condition a net aversion to

cocaine, that in most (but not all) animals, is strong enough to overcome cocaine's

rewarding e�ects [Ettenberg, 2004, Jhou et al., 2013]. In this study, we further found

that the aversive e�ects of cocaine were more individually variable, and much better

predictors of individual di�erences in cocaine-seeking, than the rewarding e�ects, and

that the resulting avoidance response was not unique to the e�ects of the drug and

not in�uenced by confounding phenotypes.

54

3.2 Results

3.2.1 Individual Sprague-Dawley rats show large variation in

cocaine-avoidance

Prior studies noted that individual cocaine doses produce immediate rewarding ef-

fects followed a few minutes later by a delayed aversive �crash�, producing avoidance

responses on a runway operant cocaine-seeking task [Ettenberg et al., 1999, Geist

and Ettenberg, 1997b, Ettenberg, 2004]. To examine more precisely the distribution

of this variation, we tested 105 rats on the runway task, and found, as expected, a

gradual increase in latencies over the 7 trials, with average latencies over the last 4 ses-

sions varying considerably between individual rats (Fig. 3.1A). Interestingly, latency

distributions could be �tted to a bimodal distribution (Fig. 3.1B) separated by a

median of 256 seconds that also corresponds to the notch between the bimodal peaks.

Hence, we subsequently grouped animals into high- or low-avoiders using this thresh-

old. Consistent with prior reports, the average run latencies were low for trials 1-2 and

signi�cantly higher for trials 4-7, re�ecting the development of conditioned avoidance

(Wilcoxon test; low avoiders: W=132, n=20, p=0.012; high-avoiders: W=496, n=31,

p<0.0001) (Fig. 3.2A). Notably, run latencies remained low in control animals run-

ning to receive saline (two-way ANOVA, Bonferroni test, n=6-26 per group, p<0.01)

(Fig. 3.2B), indicating that increased latencies were due to cocaine exposure, rather

than habituation to the apparatus or other non-speci�c reductions in exploratory ac-

tivity. Latencies to obtain cocaine did not di�er between males and females (Mann

Whitney, n=22-26 per group, U=224.0, p=0.21) (Fig. 3.2D). In rats running for

cocaine, increased latencies to obtain cocaine were also accompanied by increasing

number of run direction reversals, re�ecting alternating approach and retreat behav-

55

iors posited to be due to cocaine's dual rewarding and aversive e�ects [Ettenberg

et al., 1999], which correlated with run latencies (n=48; r2=0.2879; p=0.0001) (Fig.

3.2C).

In all the above tests, the runway apparatus was initially novel to the animal,

potentially confounding the current results because novelty-induced locomotion has

also been shown to correlate with cocaine seeking [Stead et al., 2006]. Hence, we

had attempted to minimize novelty-related e�ects by habituating all animals to the

runway apparatus prior to cocaine trials for up to 3 trials. To further control for this

possible confound, we then explicitly measured locomotor activity in a novel oper-

ant chamber distinct from the runway apparatus and found no correlation between

locomotor activity and run latencies (males: n=25, r2=0.06, p=0.25, females: n=18,

r2=0.09, p=0.23) (Fig. 3.2E).

Prior studies had shown that high doses of cocaine can be toxic, presumably

contributing to aversive e�ects, but our dose (0.75mg/kg, or about 0.25mg per rat)

is well below such levels [Smith et al., 1991], and is similar to hundreds of other self-

administration studies. Furthermore, a cohort of animals tested on a half-size dose

(0.375mg/kg, or about 0.125mg per rat) showed a similar distribution of latencies

to obtain cocaine. In particular, neither the means nor standard deviations di�ered

between the two doses (298 ± 94 seconds vs 344 ± 103 seconds avg ± SEM for full

and half size doses, n=16-25 per group, mean: Mann Whitney, U=197.0, p=0.94;

standard deviation: p=0.44) (Fig. 3.2F).

56

3.2.2 Cocaine-avoidance on the runway is speci�c to the ef-

fects of cocaine and is independent of control behaviors

To assess if the di�erences in runway latencies between high- and low-avoiders are

driven speci�cally by the properties of cocaine, we controlled for several potentially

confounding phenotypes by testing the animals on a battery of behaviors. First,to

rule out the in�uence of exploratory activity on run latencies, we tested the ani-

mals on a locomotor task in a novel chamber and found no di�erence in novelty-

induced locomotion between high- and low-avoiders (Fig. 3.3B) (n=8-9 per group,

t(15)=0.3191, p=0.750). To further explore the possibility that di�erences in runway

behavior might be due to di�erences in generalized motivation to seek rewards, we

tested these high- and low-avoider rats in a progressive ratio food-seeking task (PR).

In this task, animals leverpress for food pellets, with the required number of presses

per pellet increasing progressively throughout the task; high- and low-avoiders had

similar breakpoints (Fig. 3.3C) (n=6 per group, t(10)=0.4404, p=0.67).

We next examined whether di�erences in runway latency (which presumably re-

�ect di�erences in the relative rewarding versus aversive qualities of cocaine) gener-

alize to other tasks in which an operant action produces both rewarding and aversive

outcomes. We used a progressive punishment task (PP) in which rats pressing for

food pellets also receive concurrent footshocks whose intensity progressively increases

throughout the task. A �shock breakpoint� is reached when leverpressing is largely

inhibited, indicating the amount of shock an animal will endure for a given reward,

and this breakpoint was not di�erent between the two o�spring groups (Fig. 3.3D)

(n=6 per group, t(10)=0.4404, p=0.67). In addition to that, testing on the elevated

plus maze showed that high- and low-avoiders had no di�erences in anxiety levels, as

measured by time spent in the open arms of the maze (n=7 per group, t(12)=0.2031,

57

p=0.84) (Fig. 3.3E).

Finally, we wanted to assess if high-avoiders' aversive response to cocaine is not

speci�c to the drug and generalizes to other aversive stimuli. We therefore tested

both groups on a lithium chloride CPA test as well as on a their rate of escape from

an acute shock in a two-way shuttlebox. We found no signi�cant di�erence in the

CPA score (Fig. 3.3F) nor in escape latencies at di�erent shock intensities between

high- and low-avoiders, indicating that high-avoiders are not simply hypersensitive

to all aversive stimuli (Fig. 3.3G). Altogether, these �ndings demonstrate that the

di�erences observed in run latencies between the two groups were not likely due to

general di�erences in exploration, motivation, learning capacity or anxiety.

3.2.3 Cocaine has dual rewarding and aversive e�ects, with

the aversive component driving individual variation in

drug-seeking behavior

Although we had hypothesized that high-avoider rats exhibit stronger aversive e�ects

of cocaine, their long latencies to reach the goal compartment would also be con-

sistent with them experiencing lower rewarding e�ects of the drug. To address this

issue, we tested high- and low-avoiders in a CPP/CPA task in which rats were either

placed into conditioning chambers immediately after receiving an infusion of i.v. co-

caine (0.75 mg/kg), or 15 minutes afterwards. We found that rats with high (> 256

s) or low (<256 s) run latencies both showed robust cocaine CPP when placed into

chambers immediately after cocaine infusions. However, high-latency rats showed

dramatically higher CPA than the low-latency rats (one-way ANOVA, Bonferroni

test, n=7-8 per group, p>0.9999 and, p=0.04, respectively) (Fig. 3.4A). Correla-

tional analysis of individual animals showed similar results, in which run latencies

58

correlated with CPA but not CPP scores (n=15, r2=0.48, p=0.004, r2=0.028, p =0.5

respectively) (Fig. 3.4B). These results are consistent with prior reports [Ettenberg,

2004] and suggest that individual variation in cocaine-seeking likely depends more on

variation in cocaine's aversive e�ects, rather than its rewarding e�ects, which were

relatively constant across animals.

Because cocaine's crash has long been known to be characterized by dysphoria,

irritability, anxiety, and cravings [Gawin, 1991, Resnick et al., 1977b, Williamson

et al., 1996b], we compared high and low avoider animals on the elevated plus maze

(EPM) 15 minutes after i.v. cocaine infusions. When tested on the EPM after

saline infusions, high and low-avoider animals did not di�er signi�cantly in time spent

exploring the open arms of the EPM. However, relative to saline, cocaine infusions

induced a signi�cant decreases in open arm time in the high-avoiders but not low-

avoiders (paired t-test: low-avoiders: n=8, t(7)=0.596, p=0.57; high-avoiders n=8,

t(9)=3.736, p=0.0047) (Fig. 3.5A), suggesting enhanced anxiety during this period

in high�avoiders only. Analysis of individual animals within groups further showed

that open arm time correlated with run latencies 15 minutes after cocaine (n=18,

r2=0.32, p=0.015) but not saline injections (n=18, r2=0.028, p=0.5) (Fig. 3.5B).

The decrease in open arm time in high avoiders was not explained by a decrease in

locomotor activity, since high- and low-avoiders did not di�er in their total arm entries

on the EPM either with or without cocaine infusion (paired t-test: Low-avoiders: n=6,

t(5)=0.515, p=0.63; high-avoiders n=6, t(5)=1.346, p=0.24) (Fig. 3.5C).

3.3 Discussion

The complexity of the addictive phenotype poses several challenges in identifying

speci�c areas of vulnerabilities, as addiction likely results from a con�uence of multi-

59

ple independent vulnerabilities, that also interact with environmental factors, drug-

induced neurobiological changes, and comorbidity with other neuropsychiatric dis-

orders and personality traits [Stone et al., 2012, Wong and Schumann, 2008, Belin

et al., 2016, Gottesman and Gould, 2003]. We posit that the current results suggest

that the challenges posed by the complexity of human behavior could be addressed by

studying animal models in which intermediate endophenotypes of speci�c behavioral

elements can be more readily isolated and discerned. Hence, we sought strategies to

examine such endophenotypes in an animal model using tasks that deconstruct the

cost-bene�t decision processes into its building block and allow us to subdivide the

complexity of the addictive phenotype into smaller constituents that could then be

investigated independently. Indeed, it is worth noting that the control tasks exam-

ined could identify phenotypes that are important components of addiction in their

own right and could in�uence di�erent stages of addiction, despite our focus being on

individual di�erences in cocaine avoidance.

To study cocaine-avoidance, we require a model that is sensitive to both positive

and negative e�ects of drug-seeking, akin to the human experience. A substantial

body of literature from the lab of Aaron Ettenberg has used the runway operant

for this purpose [Ettenberg, 2004]. Although the runway task is not as widely used

as lever-pressing for cocaine self-administration, we have chosen it in this study for

several reasons. First, the single-trial nature of the runway task greatly helps to

isolate the aversive e�ects of cocaine, since the aversive responses are not masked

by subsequent doses. Another advantage of the runway task is that the limited

availability of cocaine more closely resembles early drug use in humans, where users

more commonly start with individual trial doses rather than the more unbounded

amounts available in conventional self-administration paradigms at close to zero cost.

Using this task, we found that, even though drug addiction is commonly viewed as

60

a disorder of reward learning, it was variations in avoidance learning that explained

more of the variability in drug-seeking, where avoidance responses to cocaine slowed

the initial acquisition of drug-seeking.

In summary, the current study identi�ed an important behavior, cocaine avoid-

ance, that could in�uence addiction propensity. However, it remains unknown what

anatomical substrates and molecular targets reside within these circuits, and why

they behave so di�erently in di�erent individuals. Discovering previously unknown

molecular correlates of addiction-propensity would improve our understanding, pre-

diction, and treatment, of addiction. Such a �nding, if applicable to humans, would

have complex but important implications for our theoretical understanding of how

individual di�erences contribute to drug abuse liability.

61

Figure 3.1: Two populations of rats observed on the runway task. (A) Rats showedon average a gradual increase in latencies over 7 trials, but with average latenciesvarying considerably between individuals over the last 4 sessions . (B) Latencies werebimodally distributed with two roughly equal size peaks separated by a median of256 seconds residing in the valley between peaks. We subsequently grouped animalsinto high- or low-avoiders using this threshold.

62

Figure 3.2: Individual Sprague-Dawley rats show large variation in cocaine-avoidance(A) Individual outbred Sprague-Dawley (SD) rats showed large variation in theirincrease in run latencies after 4-7 conditioning trials. (B) On average, latencies toobtain cocaine increased progressively with the number of trials in both males andfemales, compared to animals receiving saline, which showed consistently short la-tencies across all trials. (C) Average run latencies were not signi�cantly di�erentbetween males and females (D) Latency to obtain cocaine did not correlate withnovelty-induced locomotion, for either male or female rats. (E) Animals tested ona half-size dose (0.375mg/kg, or about 0.125mg per rat) showed similar latencies toobtain cocaine (298± 94 seconds vs 344± 103 seconds avg ± SEM for full and halfsize doses).

63

Figure 3.3: High- and low-avoiders do not di�er in control tasks (A) Low- and high-avoiders do not di�er in (B) novelty-induced locomotion, (C) punishment resistance,(D) food-seeking, (E) elevated plus maze, (F) lithium chloride CPA or (G) escapelatency from shock. *** p<0.001.

64

Figure 3.4: Cocaine has dual rewarding and aversive e�ects, with the aversive com-ponent driving individual variation in drug-seeking behavior (A) Cocaine producesstrong CPP (grey bars) for both high- and low-avoiders if rats are placed into con-ditioning chambers 0-15 minutes after iv cocaine infusions. However, high-avoiders(but not low-avoiders) showed a CPA (red bars) when rats were placed into chambers15-30 minutes post-infusion. (B) Runway latencies did not correlate with CPP scoreson the 0-15 minute condition, but strongly correlated with CPA scores in the 15-30minute conditions. * p<0.05, **** p<0.0001.

65

Figure 3.5: High cocaine-avoiders have enhanced anxiety during the aversive phaseof the drug (A) Relative to saline, cocaine infusions induced a signi�cant decreases inopen arm time in the high-avoiders but not low-avoiders. (B) Analysis of individualanimals within groups further showed that open arm time correlated with run latencies15 minutes after but not saline injections. (C) high- and low-avoiders did not di�erin their locomotor counts in a novel chamber either with or without cocaine. *p<0.05,** p<0.01.

66

Chapter 4

Adaptations in RMTg Underlie

Individual Di�erences in Cocaine

Avoidance

4.1 Introduction

We have already found that cocaine's aversive e�ects vary greatly between individual

animals, with high cocaine-avoider and low cocaine-avoider animals exhibiting strong

or weak aversive responses to cocaine, respectively, and that cocaine avoidance is

powerfully protective against drug-seeking in those high-avoiders. Because naturally

occurring high-avoider rats did not exhibit locomotor impairments, resistance to pun-

ishment, lack of reward motivation or heightened anxiety, we thus sought to identify

what mechanisms drive the normal variations between high- and low-avoiders in re-

sponse to cocaine, in order to determine whether these mechanisms might drive more

selective modulation of drug-seeking. The reasons for these variations are not yet

understood, but we previously reported that RMTg lesions markedly reduce run-

67

way latencies to obtain cocaine [Jhou et al., 2013], with this e�ect being replicated by

temporally restricted optogenetic inhibition of the RMTg during a time window 15-30

minutes after cocaine infusions [Jhou et al., 2013]. Hence, RMTg activity during this

time window critically drives cocaine avoidance, prompting us to examine whether

cocaine might activate the RMTg di�erently during this window in some rats but not

others.

To address these questions, and in collaboration with a slice electrophysiologist

in the lab (Dr. Je�rey Parrilla-Carrero), we generated data indicating that RMTg

activation by cocaine depends on glutamatergic changes at the RMTg involving cal-

cium permeable AMPA receptors (CP-AMPARs) residing on VTA-projecting RMTg

neurons. Speci�cally, we found increases in glutamate release in high-avoider and

a post-synaptic depontentiation characterized by a decrease in CP-AMPAR in low-

avoiders. To further characterize the sources of this di�erence at the cellular level,

we manipulated upstream cellular e�ector that may regulate glutamate transmis-

sion. We focused on Ras homolog family member A (RhoA), Rho-associated pro-

tein kinase (ROCK) and phosphatase and tensin homolog deleted on chromosome 10

(PTEN) (RhoA/ROCKs/PTEN) cellular pathway, as it has been previously shown to

be involved in glutamate transmission and plasticity [Du et al., 2013, Jurado et al.,

2010], and all the molecules involved in this pathway are highly expressed in the

RMTg. Hence, we tested the hypothesis that di�erence in RhoA/ROCKs/PTEN

activity modulate glutamate transmission to promote/prevent heterogeneity in co-

caine avoidance behavior. As expected, we found that pharmacological manipulation

of RhoA/ROCKs/PTEN pathway activity, in fact was involved on individual behav-

ioral di�erences in cocaine seeking, suggesting that manipulation of this cellular signal

may be use in the future for therapeutic bene�t.

68

4.2 Results

4.2.1 High cocaine avoiders show greater neuronal activation

in the RMTg

Having established that cocaine avoidance di�ers greatly between animals, we next

sought to identify the neural substrates of these individual di�erences. Previous

work has implicated activation of the serially connected EPN, LHb and RMTg brain

regions as key drivers of aversive reactions to cocaine [Jhou et al., 2013, Meye et al.,

2015, Meye et al., 2016]. However, it is not known whether individual di�erences in

behavior are due to variations in one or more of these regions. Hence, in a subset of

rats screened for cocaine avoidance, we examined cocaine-induced expression of the

immediate early gene cFos in the EPN, LHb, and RMTg (Fig. 4.1A). After a single

infusion of cocaine given 1 hour before sacri�ce, cFos counts in the RMTg, EPN,

and LHb, were all higher than in rats receiving saline (one-way ANOVA, Tukey's

test: n=5-7 per group, RMTg: F=86.88, p<0.0001, LHb: F=12.53, p=0.0013 (low-

avoiders vs saline) and p=0.0009 (high-avoiders vs saline), EPN:F=19.5, p<0.0001)

(Fig. 4.1B). Furthermore, cocaine-induced cFos in the RMTg was much higher in

high- than low-avoider rats (Fig. 4.1B; one-way ANOVA, Tukey's test, n=7 per group,

F=86.88, p=0.0004). However, cFos counts unexpectedly did not di�er between high-

and low-avoider rats in the EPN and LHb (one-way ANOVA, Tukey's test, n=7

per group, LHb: F=12.53, p=0.9758, EPN: F=19.5, p=0.9789). Further analysis

of individual rats revealed that the number of cFos cells in the RMTg positively

correlated with the mean latency of the last 4 sessions in high avoiders but not low

avoiders (r2=0.7455, p=0.0123; r2 = 0.0915, p = 0.51, respectively (Fig. 4.1C). These

results suggest that although the EPN and LHb are equally activated by cocaine in

69

high- and low-avoiders, the RMTg itself is activated more strongly in high-avoiders

than in low-avoiders.

Because cFos is an indirect measure of neuronal �ring with poor temporal resolu-

tion, we further evaluated RMTg responses to cocaine using in vivo electrophysiology

in the RMTg of both high- and low-avoiders (Fig. 4.1D). We recorded 71 RMTg

neurons, with 35 neurons from 11 high-avoider rats and 36 from 13 low-avoider

rats. RMTg neurons showed decreased �ring during the �rst 15 minutes after co-

caine infusion (when cocaine is rewarding), that did not di�er between high and low

avoiders (unpaired t test, n=26-33 per group, t(57)=0.09653, p=0.9234). However,

high avoiders showed signi�cantly higher �ring relative to low avoiders 15-45 minutes

after cocaine infusion, overlapping the period when cocaine is aversive (unpaired t

test, n=26-32 per group, t(56)=3.962, p=0.0034) (Fig. 4.1D). Furthermore, as with

cFos measures, correlation analysis in individual rats revealed that RMTg neuron re-

sponses 15-45 minutes but not 0-15 minutes post cocaine infusion positively correlated

with individual's runway latency (Fig. 4.1E, r2=0.5661, p=0.015 and r2=0.0236,

p=0.6927). These results further corroborate that, although the EPN, LHb, and

RMTg are serially connected structures that all mediate the aversive properties of

cocaine, only the RMTg exhibited individual di�erences in responses to cocaine that

correlated with individual di�erences in behavioral avoidance.

4.2.2 Cocaine-avoiding rats exhibit enhanced excitability of

VTA-projecting RMTg neurons.

Although a majority (55 %) of RMTg neurons project to the VTA, some do not, and

these in vivo recordings do not indicate the target of the cocaine-activated neurons.

Hence, we further examine possible cellular changes underlying these di�erences in

70

RMTg activity, Je�rey patch clamped VTA-projecting RMTg neurons labeled by in-

jections into the VTA of the retrograde tracer CTb conjugated to the �uorescent

molecule AlexaFluor555 (CTB-555) (Fig. 4.2A). We �rst characterized excitatory

synaptic transmission in high- and low- cocaine avoider rats and found that high-

avoider rats had a higher sEPSC frequency compared to both saline controls and

low-avoiders, consistent with presynaptic alterations. To further assess presynaptic

di�erences, we used the paired-pulse ratio (PPR), and found a signi�cantly decreased

PPR (Fig. 4.2C; One-way ANOVA F(2,46) = 7.631, P = 0.0014), in the high-avoiders

in compared to both low-avoiders (Tukey's test: t(46)=5.209, p= 0.00172) and saline

controls (t(46)=3.92, p= 0.021), consistent with increased glutamate release probabil-

ity. In addition to these presynaptic e�ects, analysis of sEPSC decay times revealed

faster decay times in low-avoider rats compared to saline and high-avoider rats, con-

sistent with the possibility of altered postsynaptic function.

Our observation of slower sEPSC decay times in high-avoider rats suggest the

strong possibility of there being post-synaptic di�erences between phenotypes in the

RMTg, in addition to the pre-synaptic di�erences noted. To that end, we calculated

the recti�cation index from I�V curves constructed from pharmacologically isolated

AMPAR-mediated eEPSCs in VTA-projecting RMTg neurons (Fig. 4.2A). Surpris-

ingly, we detected a strong voltage-dependent blockade of AMPAR-eEPSCs in VTA-

projecting RMTg neurons of the saline control and high-avoider rats (recti�cation

index: saline control = 0.4513 +/- 0.0468, n = 8 cells/4 rats; high-avoiders = 0.4164

+/- 0.04615, n = 12 cells/ 5 rats; Fig. 4.2A- inset), supporting the presence of synap-

tic CP-AMPARs in these groups. In contrast, low-avoiders displayed more linear

I�V relationships in the AMPAR-eEPSC (recti�cation index: 0.8094 +/- 0.1242, n

= 6 cells/3 rats; Fig. 4.2D-inset; One-way ANOVA, F(2,23) = 6.054, p= 0.0077),

low-avoider rats compared to saline (Tukey posttest: t(23)=3.906, p=0.0287), and

71

high-avoider rats (Tukey posttest: t(23)=4.766, p=0.0071), consistent with lower

proportions of CP-AMPARs (Isaac et al., 2007; Clem and Huganir, 2010) at these

synapses. We further examined AMPAR subunit composition in VTA-projecting

RMTg neurons using a selective pharmacological blocker of AMPAR-mediated EP-

SCs, NASPM 100uM (Fig. 4.2E). We found a signi�cant NASPM-induced reduc-

tion of the amplitude of AMPA-EPSC in saline control and high-avoider, but not

low avoider rats (two-way ANOVA: phenotypes, F(2, 22)= 4.054, p=0.0317; time,

F(15, 330) = 7.089, p = 0.0001.; interaction, F(30,330)= 3.054, p=0.0001, Bonferroni

posttest: min 14 t(352)= 5.278, p=0.000647, min 15 t(352)= 5.661, p=0.000225, min

16 t(352)= 5.011, p=0.0013 for saline vs low-avoiders, and min 14 t(352)= 5.291,

p=0.00062., min 15 t(352)= 5.856, p= 0.000128, min 16 t(352)= 5.904, p=0.000111,

low-avoiders vs high-avoiders respectively).

Finally, compared with the saline-control and high-avoiders the AMPA/NMDA

ratio was signi�cantly lower in VTA-projecting RMTg neurons from low-avoider rats

(Fig. 4.2B)(One-way ANOVA: F (2,45) = 5.71, p = 0.0061; Bonferroni post-test,

saline-control vs low-avoiders t = 3.029; p = 0.0121; Low-avoiders vs high-avoiders

t = 2.665; p = 0.031, saline-control vs high-avoiders t = 0.3942; p = 0.999). This

decreased CP-AMPAR mediated transmission could increase the threshold for acti-

vation of these neurons and consequently reduce conditioned avoidance to cocaine.

In support of this hypothesis, rats that received intra-RMTg infusion of either 2-

pyrrolidinone or NASPM displayed a slower or faster latency to reach the goal box

than rats that received intra-RMTg infusion of saline, respectively (Fig. 4.3B), al-

though only signi�cant for NASPM (one-way ANOVA, n=6-11 per group, F=9.012,

p=0.0021; Tukey's test; 2-pyrrolidinone vs control: p=0.1130; NASPM vs control:

n=6-11 per group, , p=0.043) (Fig. 4.3C), further suggesting that CP-AMPA recep-

tor signaling in the RMTg is key for development of this behavior. Together, these

72

data suggest that CP-AMPARs in the RMTg in�uence constitute a neural substrate

driving variability in cocaine aversion and in turn the motivation to seek cocaine.

4.2.3 RhoA/Rock downregulates glutamatergic signaling, and

bidirectionally modulates cocaine-seeking via PTEN

Taken together, the above results show that conditioned cocaine avoi- dance requires

activation of CP-AMPAR in RMTg, and that cocaine-induced reduction in the trans-

mission of this receptor on VTA-projecting RMTg neurons may be a source that

confers variability in this e�ect of cocaine. Although many mechanisms haves been

posited to down-regulate CP-AMPAR from the synapse, the RhoA/ROCKs/PTEN

pathway has been recently showed to be a key player in this regulation. For exam-

ple PTEN's lipid phosphatase activity is able to drive depression of AMPA receptor-

mediated synaptic responses [Jurado et al., 2010], and RhoA is upregulated in stressed

mice and contributes to the loss of dendritic arbor [Fox et al., 2018]. Moreover, our

data With RNA sequencing showed that these proteins are highly expressed in the

RMTg; out of several PTEN activators, ROCK was the most highly expressed (

one-way ANOVA; n= 4 per group, F=68.27; p<0.0001; Tukey's test: ROCK vs oth-

ers: p<0.0001)(Fig. 4.5B), and out of several ROCK activators, RhoA also had the

highest expression (one-way ANOVA; n= 4 per group, F=125.4; p<0.0001; Tukey's

test: RhoA vs others: p<0.0001) (Fig. 4.5A). Therefore, due to the role of these

molecules on glutamate plasticity and their expression at the RMTg, we rationalized

that the activity of any of them may in�uence glutamate transmission to alter co-

caine avoidance behavior. To investigate the role of PTEN in avoidance behavior,

we microinfused VO-OHpic trihydrate (VO-OHpic) (2.5pmol in 0.5µL), a highly se-

lective PTEN inhibitor [Mak et al., 2010, Mak and Woscholski, 2015] in the RMTg,

73

and found that run latencies were signi�cantly increased compared to saline-infused

rats (one-way ANOVA: n=5-9 per group, F=4.504, p=0.04; Tukey's test; control

vs VO-OHpic: p=0.037). As a next step, we showed that bilateral microinfusions

of a RhoA inhibitor Rhosin (30 umol/L, 0.5 uL infusion on each side) or ROCK

activator (narciclasine) (2.5pmol in 0.5µL) into RMTg produced large increases or

reductions in run way latency, respectively, although not reaching signi�cance com-

pared to the control group (Fig. 4.4A-C). Speci�cally, narciclasine was given to a

group of high-avoiders (from their average run latencies of trials 4-7) 30 minutes be-

fore their 8th trial; narciclasine dramatically reduced latencies for that trial and the

next one (9th trial, saline microinfused), but the animals recovered on the 10th trial

(saline microinfused), suggesting that we did not inadvertently mechanically damage

the RMTg (Repeated-measure one-way ANOVA, n=4-7 per group, F=5.71, p =0.006;

Tukey's test; average run latencies vs narciclasine: p=0.0043; average run latencies vs

1st saline: p=0.002; average run latencies vs 2nd saline: p=0.61) (Fig. 4.4B). To cor-

roborate those �ndings with slice electrophysiology, we found that bath application

of narciclasine, which indirectly activates PTEN via ROCK [Fürst, 2016], reduced

amplitudes of evoked EPSCs in RMTg (Fig. 4.4E) and also reduced inward recti�ca-

tion (i.e. increased ratio of currents at +40/-70mV) (Fig. 4.4D), both consistent with

reductions in glutamatergic signaling through CP-AMPARs and in agreement with

our in vivo �ndings. Finally, to prove that modulating Rho/ROCK in our paradigm

ultimately modulates avoidance response through PTEN, and not other downstream

targets, we microinfused a cocktail of narciclasine + VO-OHpic and found that run

latencies were signi�cantly increased (one-way ANOVA: n=5-9 per group, F=4.504,

p=0.04; Tukey's test; Narciclasine vs VO-OHpic + narciclasine: p=0.02) (Fig. 4.4A).

Hence, PTEN activation appears to reduce cocaine's aversive e�ects via reduction of

glutamate signaling in the RMTg.

74

4.2.4 Protein quanti�cation

Even though manipulating the pathway of interest provides insight into what might

be di�erent between high- and low-avoiders after exposure to cocaine in the runway, it

does not necessarily mean that this pathway re�ects the biological di�erences between

these two groups. We therefore quanti�ed some proteins of interest along this pathway

using western blot or ELISA. Total Glutamate Receptor Ionotropic, AMPA 1 (GluR1)

subunit (t-GluR1) and phosphorylated GluR1 subunit at serine831 site (p-GluR1-

S831), total Glutamate Receptor Ionotropic, AMPA 2 (GluR2), and PTEN receptor

subunit protein levels were examined by western blot analysis in whole homogenate

prepared from RMTg extracted from the brains of high- vs low-avoiders 24 hours after

their last runway session. High-avoiders and low-avoiders had similar expression of t-

GluR1 and GluR2 relative to control animals. However, relative to control rats, only

low-avoiders had reduced expression of p-GluR1-S831, consistent with our previous

�ndings (Fig. 4.5C). Interestingly, we found that PTEN was not di�erent between

the two groups of animals, but its upstream activator RhoA was; we quanti�ed the

amount of active RhoA using a modi�ed ELISA protocol [Fox et al., 2018] in those 3

groups of animals and found that low-avoiders had increased active RhoA after runway

compared to low-avoiders and control (n=4-5 per group, one-way ANOVA, F=6.081;

p=0.0166; Tukey's test: high-avoiders vs low-avoiders p=0.0183) (Fig. 4.5D). The

protein quanti�cation with western blot was done with only one sample per animal

group, which makes the data very preliminary and its interpretation very limited.

4.3 Discussion

In this study, we have elucidated novel molecular mechanisms driving aversive con-

ditioning to cocaine and their individual di�erences, that �ll major gaps in our un-

75

derstanding of drug-seeking, and that someday could be exploited for therapeutic

bene�t. Indeed, aversive e�ects of drugs have already been used as therapeutics, as

in the case of disul�ram (Antabuse) which inhibits ADLH2, thereby reducing alcohol

intake by greatly potentiating its aversive e�ects. Following the same principle, we

have attempted to understand the molecular basis of cocaine aversion in an attempt

to indentify tharapeutic targets. We have shown that cocaine's aversive e�ects vary

greatly between individual animals and that these variabilities were mediated by cor-

responding di�erences in RMTg cellular responses to glutamate. We found that VTA-

projecting RMTg neurons were more excitable in high-avoiders, and showed LTD in

low-avoiders, and that this GluR1 downregulation was driven by di�erential activity

of PTEN, previously implicated in synaptic plasticity [Jurado et al., 2010]. Recently,

a role for RhoA and its e�ector kinase ROCK in positive regulation of PTEN activity

has been reported [Li et al., 2005]. Indeed, ROCK can increase the phosphorylation

and activity of PTEN, through an unknown mechanism but most likely involving a

physical interaction with PTEN [Li et al., 2005, Meili et al., 2005]. Although the

upstream signals that control this pathway RhoA/ROCK/PTEN remain unknown,

we have demonstrated that this pathway is upregulated in low-avoider animals. How-

ever, egarding the use of current PTEN inhibitor in this research, some aspects should

be taken into consideration to endorse the assumption that the observed experimen-

tal e�ects of the compounds are, in fact, due to PTEN inhibition. These include,

among others, dose-response analysis using more than one inhibitory compound; the

analysis of GluR1 content as an indirect readout of PTEN inhibition; the compara-

tive analysis of both PTEN-positive and PTEN-negative cells; and the comparative

analysis of GluR1 with PTEN-knockdown, if feasible. Building on this data, we thus

seek to further characterize these cellular mechanisms of cocaine avoidance, their dif-

ferential regulation between individuals, and the possibility of modulating them for

76

therapeutic bene�t.

Even our current �ndings, albeit still at the beginning of the work, can already

be exploited for potential therapy. Indeed the PTEN inhibitor VO-OHpic had only

minimal e�ect on RMTg membrane current when delivered alone (unpublished obser-

vations), suggesting it could potentiate cocaine aversion with relatively fewer e�ects

in cocaine's absence. This type of allosteric modulator, which regulates responses

to ligand without activating the receptor on its own have garnered increasing at-

tention in drug-discovery due to their potentially greater functional selectivity than

traditional agonists and antagonists [Anastasio et al., 2013, Nussinov and Tsai, 2013].

Moreover, in addition to its roles in cocaine avoidance, the RMTg plays broad roles

in aversive learning, anxiety and punishment [Vento et al., 2017], potentially compro-

mising attempts to modulate these pathways therapeutically. Interestingly, because

high-avoider rats did not exhibit locomotor impairments or heightened anxiety, we

posited that the biological mechanisms underlying high- versus low-avoider pheno-

types involving PTEN can apparently modulate drug-seeking without a�ecting other

non-drug behaviors, thereby constituting a more ideal mechanism to target for ther-

apeutic intervention.

However, despite PTEN being an intriguing therapeutic target, more e�orts are

required to de�ne relevant indications for a reduction of PTEN's functional dose.

Indeed, PTEN's �risky� pro�le as a tumor suppressor has to be considered. While

recent studies suggest a possible therapeutic window for PTEN dose reduction without

triggering tumorigenesis, dysfunctional outcomes of PTEN deletion in the past have

raised questions over its therapeutic suitability [Gutilla et al., 2016, Liu et al., 2010,

Pun et al., 2012, Williams et al., 2015] and global manipulation of the PIP3 pathway,

particularly targeting a tumor suppressor such as PTEN, is unlikely to be considered

for therapeutic intervention in humans. To address this issue, important e�orts have

77

already been made to develop speci�c and biologically active small molecule PTEN

inhibitors, and to validate them as suitable experimental and clinical tools. This

has provided compounds targeting PTEN that are widely used in research, as well

as the proof-of-concept that PTEN pharmacological inhibition could be bene�cial

for human welfare in speci�c diseases or conditions. Moreover, chemical tools are

required that allow selective modulation of the various PTEN functions, and there

are already groups that have already developed some of those biologically active

peptides. As a treatment for Alzheimer's disease for example, one group developed

a cell-permeable peptide that competes intracellular PTEN-PDZ interactions and

targeted PDZ-dependent interactions of PTEN (preserving its catalytic activity) to

disengage the cascade of events triggered by Abeta, therefore preventing synaptic and

cognitive dysfunction. Interestingly, and relevant to this study, they found that PTEN

was engaged by Abeta in a manner that resembles synaptically induced LTD, and

that the peptide rescued synaptic function and mouse cognitive de�cits, in parallel

with protecting from Abeta-induced NMDAR-dependent long term depression and

synaptic toxicity events, a phenomenon which was dependent on PTEN recruitment

to synaptic spines through PTEN-PDZ interactions [Knafo et al., 2016]. But again,

PTEN speci�city of these molecules is intrinsically limited due to the ubiquity of PDZ

motifs and recognition elements [Liu et al., 2010].

Alternatively, antagonists derived from PTEN interaction partners have been re-

ported such as 3L4F, a truncated peptide from an intracellular loop of the 5-HT2C

receptor (5-HT2CR), which was found to prevent PTEN-5-HT2CR complex forma-

tion and to decrease PTEN-mediated dephosphorylation both in vitro and in vivo.

The ability to selectively target this protein�protein interaction for a receptor that

is found predominately in the brain [Anastasio et al., 2013] also suggests that resul-

tant therapeutic compounds will have very limited side-e�ect pro�les. Such allosteric

78

modulation of 5-HT2CR signaling presents a new therapeutic approach in psychi-

atric disorders in which disrupted 5-HT2CR signaling is implicated. Getting a better

understanding of PTEN's structure and how it interacts with GluR1 to cause its

internalization gives hope toward potential targeted drug development.

79

Figure 4.1: High cocaine avoiders show greater neuronal activation in the RMTg.(A) 24hours after being tested on the runway, animals were split into high- and low-avoiders to examine di�erences cocaine-induced expression of the immediate earlygene cFos in the EPN, LHb, and RMTg. (B) cFos counts in the RMTg, EPN, andLHb, were all higher than in rats receiving saline. cFos in the RMTg is signi�cantlyin high-avoiders compared to low-avoider rats, but not in EPN and LHb. (C) cFoscells in the RMTg positively correlate with the mean run latency of the last 4 sessionsin high avoiders but not low avoiders. (D) RMTg neurons inhibition during therewarding phase of cocaine (0-15 minutes post-infusion), that did not di�er betweenhigh and low avoiders, but high avoiders showed signi�cantly higher RMTg �ringrelative to low avoiders 15-45 minutes after cocaine infusion, overlapping the periodwhen cocaine is aversive. (E) Correlation analysis in individual rats revealed thatRMTg neuron responses 15-45 minutes but not 0-15 minutes post cocaine infusionpositively correlated with individual's runway latency. ** p<0.01; *** p<0.001.

80

Figure 4.2: Glutamatergic inputs drive greater RMTg activation in brain slicesfrom high vs low cocaine avoider rats. (A) Patch-clamped VTA-projecting RMTgneuron triple-labeled for biocytin (blue), retrograde tracer CTB-AF555 (red), and theRMTg marker FoxP1 (green). (B-E) Relative to low-avoider rats (grey bars), VTA-projecting RMTg neurons in high-avoiders (red bars) exhibited higher AMPA/NMDAratios (B) lower paired-pulse ratios (PPR) (C) greater inward recti�cation (R.I.)(D)slower EPSC decay kinetics (E), and greater sensitivity to the CP-AMPAR blockerNASPM (F), all consistent with increased pre- and post-synaptic signaling. * p<0.05;** p<0.01.

81

Figure 4.3: Manipulating GluR1 subunit of AMPA receptor modulates avoidancebehavior on the runway (A) Cannulas placements at the RMTg (B) Microinfusing(B)2-pyrrolidinone or (C) NASPM in the RMTg increases or decreases run latenciescompared to a control group respectively. * p<0.05.

82

Figure 4.4: PTEN downregulates glutamatergic signaling, and bidirectionally mod-ulates cocaine-seeking. (A) Micro-infusing PTEN inhibitor VO-OHpic and RhoAinhibitor Rhosin in the RMTg increase run latencies vs the RhoA inhibitor narcicla-sine decreases run latencies. (B) Microinfusing narciclasine in a group of high-avoidersbefore the runway dramatically reduces run latencies, and the e�ect gets carried overto the next day, after which the latency recovers. (C) Run latencies by trial for thedi�erent manipulations. (D) Bath application of narciclasine reduces amplitudes ofevoked EPSCs in RMTg and (E) reduces inward recti�cation. * p<0.05; *** p<0.001.

83

Figure 4.5: RhoA/ROCK/PTEN pathway is upregulated in RMTg of low-avoiders.(A) RhoA had the highest RMTg mRNA expression out of several ROCK activatorsexamined. (B) ROCK had the highest RMTg mRNA expression out of several RhoAactivators examined (C) Western blot of RMTg sections from high- and low-avoiderrats showed a down-regulation of p-GluR1-S831 in low-avoider rats. (D) Active RhoAwas upregulated in low-avoider rats compared to high-avoiders and control. * p<0.05.

84

Chapter 5

Adaptations in NAshell Underlie

Individual Di�erences in Cocaine

Avoidance

5.1 Introduction

Many studies have noted that VTA DA neurons increase their activity to rewarding

stimuli, and/or decrease it in response to aversive stimuli [Ungless et al., 2004, Calipari

et al., 2016, Danjo et al., 2014, Roitman et al., 2008]. In fact, recent studies have re-

vealed that optogenetic activation of GABAergic neurons and resultant inactivation of

DA neurons suppress reward consumption and induce an aversive response [Tan et al.,

2012, van Zessen et al., 2012]. In addition to noxious stimuli, midbrain DA neurons

appear to be actively inhibited by the absence of expected food rewards (or reward

omission) [Waelti et al., 2001, Tobler et al., 2003, Matsumoto and Hikosaka, 2007],

drug withdrawal [Diana et al., 1993], conditioned stimuli associated with lithium-

induced sickness [Mark et al., 1991]. These data are consistent with the idea that

85

DA neuron inhibition encodes prediction errors involving negative incentive stimuli.

While VTA DA neurons project to multiple brain areas, the most prominent pathway

involved in this bidirectional encoding of reward and aversive stimuli is believed to

reside in the medial striatum. For example, a recent elegant study by Roitman et al.

stimulated a sensory pathway that can carry both positive and negative information

to show that opposing DA state in the NAshell is key for encoding opposing valences

stimuli [Roitman et al., 2008]. Particularly, Roitman et al., demonstrated that sucrose

rewards increase, while the aversive quinine suppressed the frequency of DA transients

in the NAshell [Roitman et al., 2008]. Indeed, while DA neurons respond to reward-

related events similarly regardless of their locations in A8, A9, and A10, DA neurons

projecting from the medial A10 to the ventromedial striatum appear to be impor-

tant for the a�ective component of stimulus-outcome learning or stimulus-incentive

learning. Speci�cally, inhibition of DA neurons selectively in medial A10 neurons

(projecting to the ventromedial striatum, including medial accumbens shell) leads to

conditioned aversion, but not in lateral A10 (projecting to ventrolateral striatum) or

A9 neurons (projecting to dorsal striatum) [Liu et al., 2008].

Moreover, several studies have shown the suppression of DA neuron �rings in re-

sponse to aversive stimuli, that in turn decreases DA levels in the NAshell, speci�cally

promote the signal transmission in the indirect pathway through activated D2Rs, re-

sulting in avoidance behavior [Hikida et al., 2010, Hikida et al., 2013, Danjo et al.,

2014, Kravitz et al., 2012]. However, it is not clear whether this bi-directional en-

coding of reward and aversive behavioral response is true for cocaine, and to what

extent the di�erences in this process correlate with individual di�erences in cocaine

avoidance. Since RMTg neurons send strong inhibitory projections to DA neurons in

the VTA, which in turn project to the NAc, we wanted to investigate if this in�uence

contributes to the encoding of aversive learning by decreasing DA and thus increas-

86

ing D2-MSNs activity in the NAshell. Indeed, it has been already shown that RMTg

excitation reduces DA levels in the NAc [Taylor et al., 2019], and that RMTg hyper-

function accounts for depressive symptoms by reducing VTA�DA transmission onto

the NAshell of alcohol-withdrawn rats [Fu et al., 2019]. Therefore, we hypothesized

that individual di�erences in RMTg in�uences on VTA DA neurons, which modulate

the level of DA release in the NAshell, in turn drive di�erences in NAshell D2-MSN

properties that contribute to cocaine avoidance.

5.2 Results

5.2.1 D2-like agonist/antagonist injections into the NAshell

reduce/increase motivation to seek cocaine in a runway

operant task.

As a preliminary attempt to examine whether a dip in DA transmission may disin-

hibit D2-MSNs to drive aversion to cocaine, we mimicked reductions or increases in D2

transmission by infusing the D2/3 receptor antagonist sulpiride (0.4 ug/0.5 ul per side)

or the D2/3 agonist quinpirole (0.4 ug/0.5 ul per side) directly in the NAshell. As ex-

pected, decreases D2R signaling with sulpiride increased, while increases D2R signal-

ing with quinpirole reduced, latencies to reach the goal box when compared to saline-

infused rats (one-way ANOVA, n=6-7 per group, F=11.8, p=0.0001; Tukey's test;

vehicle vs cocaine+quinpirole: p=0.37 ; vehicle vs cocaine + sulpiride: p=0.0221; co-

caine + sulpiride vs saline + sulpiride: p=0.0001) (Fig. 5.1E). Notably, sulpiride/quinpirole

were infused 10 minutes after the rats reached the goal box, shortly before the on-

set of RMTg activation and cocaine's aversive e�ects [Jhou et al., 2013, Li et al.,

2019a]. Furthermore, because infusions are given after animals complete each trial,

87

these e�ects are independent of any possible e�ects of the drug on motor perfor-

mance or action selection, both of which occur before rats enter the goal box. These

results also suggest that D2 signaling might ultimately alter learning and the deci-

sion of whether or not to approach the reward based on the perceived value of the

two options. To further determine whether changes in D2-MSNs excitability corre-

late with cocaine aversion phenotypes, Je�rey conducted patch clamp recordings in

saline-control, low-avoider, and high-avoider rats respectively. Strikingly, we found

that lower amounts of injected current were needed in order to evoke action potentials

in D2-MSNs from high-avoider versus low-avoider rats (Figure 3B). These results sug-

gest that D2 receptor signaling plays an important role in cocaine aversion, and that

similar to the RMTg individual di�erences in neurons excitability correlates with the

cocaine avoidance phenotypes.

To further investigate the di�erences between high and low-avoiders in D2, Je�rey

examined whether individual behavioral di�erences correlated with altered glutamate

transmission in D2-MSNs, as altered AMPA transmission and glutamate release onto

MSNs is one of the hallmarks of the transition from recreational to compulsive drug

use [Lüscher and Malenka, 2011, Britt and Bonci, 2013, Gipson et al., 2014, Grueter

et al., 2012, van Huijstee and Mansvelder, 2015, Morales and Pickel, 2012]. Hence,

we measured paired pulse ratios (PPR) and AMPA/NMDA ratios to assess changes

in pre- or postsynaptic glutamate transmission on D2-MSNs in the NAshell. Indeed,

we found that increased glutamate transmission correlated with decreased motiva-

tion to seek cocaine, consistent with prior �ndings [Bock et al., 2013] (Fig. 5.1B).

Speci�cally, we found that compared with saline-control and low-avoider rats the

AMPA/NMDA ratio in D2-MSNs was signi�cantly higher than in high-avoider rats

(One-way ANOVA: F (2,45) = 5.71, p <0.006; Bonferroni post-test, High-avoider vs

saline-control t = 3.029; p <0.05; High-avoiders vs Low-avoiders t = 2.665; p <0.05)

88

(Fig5.1C). We also found a signi�cantly decreased PPR (One-way ANOVA F(2,45)

= 8.591, p = 0.0007), in high-avoiders compared to low-avoiders (t=5.609, p<0.05)

and saline control rats (t=4.192, p<0.05), suggesting increased presynaptic gluta-

mate release in the high avoiders. Hence, decreased cocaine-seeking in high-avoider

rats correlated with both presynaptic and post-synaptic potentiation of glutamate

transmission on D2-MSNs of the NAshell.

5.2.2 Cocaine-induced avoidance behavior correlates with im-

paired LTD at excitatory synapse in D2-MSNs of the

NAshell.

In addition to changes in synaptic strength, it is well accepted that saturation or

occlusion of long-term plasticity in the form of long-term depression (LTD) and

long-term potentiation (LTP) correlates with di�erent phase of cocaine addiction

[Lüscher and Malenka, 2011, Britt and Bonci, 2013, Gipson et al., 2014, Grueter

et al., 2012, van Huijstee and Mansvelder, 2015, Morales and Pickel, 2012]. Notably,

it has been suggested that alterations in LTD in the NAc by cocaine might be a

key step in the cascade of neuroadaptations that lead to addiction. Therefore, we

examined whether LTD alterations in NAshell D2-MSNs correlate with decreased

cocaine-seeking in high-avoider rats. To induce LTD at excitatory synapses, we used

a spike-timing-dependent plasticity protocol previously known to involve D2R and

cannabinoid receptors (CB1) receptor activities in D2-MSNs of the striatum [Saun-

ders et al., 2013], and that mimics naturally occurring activity patterns in vivo,

making it a valuable tool to correlate behavior with neuronal plasticity. Interestingly,

we found that compared with saline-control rats and low-avoiders, spike-timing de-

pendent plasticity (STDP-LTD) was impaired in D2-MSNs in high-avoider rats (Fig.

89

5.2B). Because previous reports [Saunders et al., 2013] demonstrated that this type

of plasticity requires CB1 receptor activation, we further characterized whether this

e�ect in high-avoider was due to altered signaling of any of these receptors. In sep-

arate experiments, we bath applied the CB1 receptor agonist win55212 (5uM) and

found that the sensitivity of the drug to induced LTD is reduced in high avoider, as

compared to low-avoider or saline rats in NAshell D2-MSNs (Fig. 5.2C). These results

further suggest that decreased LTD in NAshell D2MSNs may drive individual di�er-

ences of the negative conditioning attributes of cocaine [Ettenberg, 2009, Ettenberg

et al., 1999, Jhou et al., 2013].

Indeed, it is known that D2 receptor activation increases the production/release of

endocannabinoid (eCB) [Giu�rida et al., 2001, Patel et al., 2003, Centonze et al., 2004]

that depress excitatory synapses in the striatum via CB1 activation [Kreitzer and

Malenka, 2005, Yin and Lovinger, 2006, Lovinger, 2008]. Therefore, we reasoned that

decreases in D2R signaling in the NAshell may drive cocaine aversion by reducing eCB

dependent depression at excitatory synapses in the NAshell. To test this hypothesis,

we infused the D2 antagonist sulpiride (0.4 ug/0.5 ul per side) into the NAshell with

or without the CB1 agonist win55212 (0.5 ug/0.5 ul per side) and demonstrated that

sulpiride alone decreases increases run latencies (consistent with greater avoidance)

and that this e�ect was blocked by the CB1 agonist (one-way ANOVA, n=6-7 per

group, F=13.03, p<0.0001; Tukey's test; cocaine + sulpiride vs cocaine + sulpiride

+ win55212: p=0.0001)(Fig. 5.2D).

90

5.2.3 Cocaine-induced avoidance behavior correlates with in-

creased adenosine A2A receptor tone in D2-MSNs of the

NAshell.

We next decided to investigate the involvement of A2A receptors in mediating cocaine

avoidance behavior, as these receptors: 1) are heavily expressed in the striatum, where

they colocalize with D2 receptors on MSNs [Fink et al., 1992, Svenningsson et al.,

1998] 2) exert tonic inhibitory control over D2 receptor signaling within the striatum

[Farrar et al., 2010, Harper et al., 2006, Nagel et al., 2003, Weber et al., 2010] 3)

reduce eCB release and 4) drive LTP [Dias et al., 2012, d'Alcantara et al., 2001].

We therefore microinfused the A2A agonist CGS 21680 (0.4 ug/0.5 ul per side) into

the NAshell with or without cocaine and found that it dramatically increased run

latencies compared to vehicle microinfusion or to saline injections (one-way ANOVA,

n=6-7 per group, F=13.03, p<0.0001; Tukey's test; cocaine + vehicle vs cocaine +

CGS 21680: p=0.048 ; cocaine + CGS 21680 vs saline + CGS 21680: p=0.0016)

(Fig. 5.2D). We furthermore quanti�ed D2 receptor and A2A receptor proteins in

the NAshell of high- and low-avoiders using western blots, and found that A2A was

upregulated in high-avoiders compared to low-avoiders or saline controls, whereas

D2 levels were not changed (Fig. 5.2E). Again, the protein quanti�cation was only

acquired from one sample, which limits the interpretation of the results.

91

5.2.4 RMTg inhibition reduces shock cue-induced decrease in

DA levels in the NAshell, without a�ecting reward or

neutral cues

To test whether RMTg modulates DA levels in the NAshell in response to a re-

ward/aversion predicting cue, we recorded the �uorescence changes of an intensity-

based genetically encoded DA sensor (dLight1.2), in freely moving rats with �ber

photometry, with or without inhibition of RMTg with baclofen/muscimol (Figure

5.3A), a cocktail of GABAA/GABAB agonists [Bowery et al., 1980, Andén et al.,

1979]. dLight1.2 enables optical readout of changes in DA concentration by cou-

pling the agonist binding-induced conformational changes in human DA receptors to

changes in the �uorescence intensity of circularly permuted (cp) GFP derived from

GCaMP6 [Patriarchi et al., 2018]. We delivered an adeno associated virus encoding

dLight1.2 (AAV9-syn-dLight1.2) in the NAshell (Figure 5.3B), followed by implan-

tation of an optic �ber for recordings. We trained the animals on a Pavlovian task

in which rats trained to associate distinct auditory cues with rewarding, neutral or

aversive outcomes (sucrose pellets, nothing or mild footshocks, respectively). We then

recorded DA levels in the NAshell in response to the cues and observed a sharp peak

in response to the reward cue and an inhibition in response to the aversive cue. How-

ever, after inhibition of RMTg with microinfusion of baclofen/muscimol, DA decrease

in response to the shock predictive cue was reduced, (Fig. 5.3 C, D) (Student t test:

n=4; t(47)=2.672; p=0.0122), whereas DA increase/decrease to the reward cue (Stu-

dent t test: n=4; t(47)=1.537; p=0.1313) (Fig. 5.3C, E) and neutral cue (Student

t test: n=4; t(47)=1.357; p=0.1811) (Fig. 5.3C, F) were not a�ected. Moreover,

�uorescent readout in no virus control showed no changes in response to any of the

three cues. These results suggest that RMTg partly modulates the levels of DA in

92

the NAshell in response to shock cue but not reward cue, consistent with previous

observations that demonstrated that RMTg inhibitions abolished VTA inhibition by

aversive stimuli, without altering their responses to reward cue [Li et al., 2019a].

5.2.5 High cocaine-avoiders have reduced early acquisition of

cocaine self-administration but higher rates of reinstate-

ment

Because we characterized cocaine avoidance thus far exclusively using an operant run-

way task, we next tested whether this phenotype predicts animals' behavior in a dif-

ferent operant task, a conventional leverpress self-administration paradigm. Interest-

ingly one study showed that resilience to develop cocaine addiction is characterized by

enhancement in glutamate transmission in the D2-MSNs of the ventral striatum [Bock

et al., 2013] while another group demonstrated as well that injecting an adenosine 2A

receptor agonist in the NAc prevented acquistion of self-administration [Knapp et al.,

2001, Wydra et al., 2018]. We therefore tested a new group of SD rats on the runway

task, then trained them to leverpress for cocaine, and found that run latencies corre-

lated negatively with lever pressing during the �rst three days of self-administration

(n=22, r2=0.49, p=0.0003, bar graph: n=9-13 per group, t(20)=5.117, p<0.0001)

(Fig. 5.4A). However, all animals eventually met the acquisition criteria, and after 2

weeks of training, the mean number of cocaine infusions no longer correlated with run-

way latency, and did not di�er between groups in the last 3 days of self-administration

(n=22, r2=0.015, p=0.59, bar graph: n=10-12 per group, t(20)=0.012, p=0.13) (Fig.

5.4B). We subsequently extinguished cocaine seeking behavior, and tested animals

on both cued-induced and drug-primed reinstatement. Surprisingly, high avoiders

were more likely to reinstate the extinguished cocaine seeking behavior compared

93

to the low avoider group for both types of reinstatement (cued reinstatement: n=15,

r2=0.57, p=0.0011, bar graph: n=8 per group, t(13)=2.925, p=0.012, cocaine-primed

reinstatement: n=8, r2=0.60, p=0.02) (Fig. 5.4C and 5.4D).

5.3 Discussion

We have demonstrated that high- and low-avoiders exhibit di�erences in D2 cells

of the NAshell after exposure to cocaine, where high-avoiders had higher D2 MSNs

excitability, upregulation of A2A receptor and had impaired eCB-mediated LTD.

Consequently, high-avoiders were less likely to acquire self-administration, consistent

with prior reports [Bock et al., 2013]. Finally we found that decreases in DA in

the NAshell in response to a shock predicting cue is mediated through RMTg. The

above results not only implicate the NAshell in mediating individual di�erences in

run latencies, but also provide a likely downstream mechanism by which the RMTg

regulates behavioral output. In this view, di�erences in neuronal excitability in the

VTA projecting RMTg pathway between high- and low-avoiders drive behavior by

in�uencing downstream targets, particularly involving dopaminergic projections to

D2 MSN of the NAshell, consistent with a role of the RMTg to VTA projections

in the negative a�ect produced by noxious stimuli (including cocaine aversion)[Jhou

et al., 2013, Smith et al., 2019b, Li et al., 2019b].

Our results are in agreement with other studies that have shown that the NAshell

mediates avoidance behavior to noxious stimuli by activating D2 through decreased

DA release. Indeed manipulating D1 vs D2 results in opposite behavioral outputs.

Rewarding stimuli elicit a burst of phasic �rings in DA neurons [Mirenowicz and

Schultz, 1994], and vastly increase synaptic concentrations of DA in the NAc [Tsai

et al., 2009, Di Chiara and Imperato, 1988]. This increase e�ectively activates the low-

94

a�nity D1 receptors and results in induction of LTP in striatonigral neurons of the

direct pathway [Shen et al., 2008, Gerfen and Surmeier, 2011, Goto and Grace, 2005].

The simultaneous activation of the high-a�nity, inhibitory D2 receptors, however,

keeps the indirect pathway inactive. D1 receptor antagonist thus disrupts cocaine-

induced and reward-based behavior in a direct pathway�speci�c manner. Aversive

stimuli on the other hand have been shown to reduce tonic �rings of most DA neurons

[Brischoux et al., 2009, Ungless et al., 2004, Mirenowicz and Schultz, 1994, Coizet

et al., 2006], where blunting the tonic DA release in the ventromedial striatum leads to

conditioned place aversion. Particularly, an electric footshock been shown to inhibit

about 90 % of DA neurons in the VTA [Tan et al., 2012] and would reduce DA levels in

the NAc. This reduction in DA relieves the inhibitory actions of the high-a�nity D2

receptors but has no e�ects on the low-a�nity D1 receptors. The resultant selective

transmission enhancement in the indirect pathway would induce aversive learning.

RMTg predominantly projects to VTA DArgic neurons and inhibits these neurons

in response to shock and shock predictive cues [Li et al., 2019a]. It is thus not

surprising to �nd that RMTg modulated DA levels in the NAshell in response to

shock cues, where RMTg inhibtion blunted the decrease of DA in response to a shock

predicting cue in the NAshell. Since acute cocaine exposure increases VTA-DA neuron

�ring and DA release in limbic structure, and acute cocaine withdrawal decreases DA

neuronal activity [Kuhar et al., 1991, Ferris et al., 2014, Sulzer, 2011, Koeltzow and

White, 2003, Weiss et al., 1992] and since RMTg is a signi�cant source of inhibitory

input to the midbrain DA neurons, increased RMTg activity in high-avoiders could

account, at least in part, for the aversive learning that results from reduced activity

of DA neurons during acute withdrawal. And although we didn't test the role of

RMTg in modulating NAshell DA in response to cocaine directly, we hypothesize that

15 minutes after individual cocaine infusions (when cocaine becomes aversive) high-

95

avoiders (but not low-avoiders) who have higher RMTg activation, would exhibit a dip

in DA transmission in the NAshell that disinhibits D2-MSNs, exerting an increased

inhibitory in�uence on the DArgic cells of VTA, resulting in a hypodopaminergic state

that disinhibits D2-MSNs cells in the NAshell nucleus to drive avoidance behavior.

In contrast, low-avoiders would lack this D2-MSNs disinhibiting mechanism, and

therefore show less avoidance of cocaine. Such an interpretation is consistent with

previous studies showing that low DA levels in the ventral striatum guide behavior to

avoid aversive outcomes by disinhibits D2-MSNs [Danjo et al., 2014]. Of course, since

we used cannulas at the RMTg, a structure that does not exclusively project to the

VTA we cannot rule out the possibility that the decreased dip in DA with shock cue

can be an indirect result of RMTg inhibition. It would be crucial to follow up those

results with a pathway speci�c manipulation, using DREADDs to deactivate VTA-

projecting RMTg cells and to record from D2-MSNs in the NAshell using calcium

imaging in D2-cre rats, with and without RMTg inhibition.

Interestingly, we found that high-avoiders animals have upregulated A2A recep-

tor, and were less likely to acquire self-administration. These results are not only

consistent with the changes we observed in D2 MSNs [Danjo et al., 2014], but also

with the role of A2A in protecting against initiation of drug seeking [Knapp et al.,

2001]. These results again suggest that adenosine A2A receptor stimulation may op-

pose DA D2 receptor signaling in the NAshell that mediates cocaine seeking. Because

A2A has been shown to decrease eCB release and promote LTP, our working model

suggests that high-avoiders who have upregulated A2A receptor, on the one hand

prevent the release of eCB and thus blunt the LTD that develops in low-avoiders, and

on the other hand contribute in promoting LTP in D2 of high-avoiders. The question

of whether adenosine signaling can be used as a potential therapy in abuse disorders

remains to be answered. For example, several A2A antagonists have been developed

96

for Parkinson's disease. The neurobiology of addiction has similarities to the neuro-

science of Parkinson's disease, where striatal DA signaling is also impaired [Jenner

et al., 2009]. A2A antagonists are being tested for Parkinson's disease speci�cally to

increase DA signaling at the D2 receptor, and several compounds have been proven

to be safe for use in humans. To date, three medications (istradefylline, preladenant,

and tozadenant) have been used in human clinical trials. Although we showed that

high-avoiders have higher cocaine reinstatement compared to low-avoiders, we haven't

looked at the neurobiological di�erences between high and low avoiders after self-

administration, extinction and reinstatement. However, it is worth mentioning some

studies that implicate A2A at those later stages. Indeed, human and animal studies

have consistently shown that addiction is accompanied by blunted DA signaling in

the striatum, and that the status of DA transmission predicts relapse or recovery from

addiction. Therefore, there is a consensus that manipulating this neurotransmission

system could be an e�cient strategy to treat addiction. In particular, converging

evidence suggests that targeting D2R - i.e. enhancing its activity - would increase

the response rate to treatment by giving patients the boost they need to shift their

habits away from drug or alcohol use [Martinez et al., 2011, Tri�lie� et al., 2013].

A promising way to increase signaling at the D2 receptor is by blocking the A2A

receptor. A2A receptors are closely linked to D2R and are able to modulate their sig-

naling [Ferré et al., 2016]. Blocking the A2A receptor potentiates the e�ect of DA at

the D2 receptor by facilitating the downstream cellular e�ects. Accordingly, studies in

rodents have shown that A2A antagonists can enhance DA-mediated processes, such

as reward-driven behaviors and motivation. More precisely, stimulation of adenosine

A2A receptors antagonizes reinstatement of cocaine seeking elicited by cocaine, DA

D2-receptor stimulation and cocaine-conditioned cues [Borroto-Escuela et al., 2015].

Based on these rodent studies, there has been a lot of interest in A2A antagonists for

97

addiction research in humans, and one experiment using A2A antagonists for cocaine

dependence with functional MR scanning showed that this medication increased ac-

tivity in brain regions known to be hypoactive in cocaine addiction [Moeller et al.,

2012]. However, despite these promising data, there is no clinical trials testing A2A

antagonists for addiction. It would nonetheless be an interesting avenue to pursue,

after gaining a better understanding of the role of this receptor in the addiction phe-

notype. Moreover, the limited distribution of the receptor in the brain makes it a

valuable target in terms of targeted delivery and restriction of global side-e�ects of

medications.

98

Figure 5.1: D2-like agonist/antagonist injections into the NAshell reduce/increasemotivation to seek cocaine in a runway operant task. (A) Schematic illustration ofthe performed ex-vivo recording on D2-MSNs of D2-cre rats, (B) D2-MSNs displayedincreased excitability and (C) increased in AMPA/NMDA ratio. (D) NAshell can-nulae placements. (E) Sulpiride increased, while quinpirole reduced run latencies. *p<0.05, *** p<0.001.

99

Figure 5.2: Cocaine-induced avoidance behavior correlates with impaired LTD atexcitatory synapse in D2-MSNs of the NAshell and upregulation of A2A receptor. (A)PPR was signi�cantly decreased in high-avoiders vs low-avoiders and control. (B) D2-MSNs displayed attenuated LTD, (C) and decreased CB1-LTD in high-avoiders ratsas compared with saline-control and low-avoiders. (D) The CB1 agonist Win55212blocked sulpiride's e�ect on the runway, and A2A agonist increased run latencies.(E) Western blot of NAshell collected from brains of high- and low-avoider animalsrevealed an upregulation of A2A receptors compared to low-avoiders and control. (F)Working model of our �ndings * p<0.05, ** p<0.01.

100

Figure 5.3: RMTg inhibition reduces shock cue decrease in DA levels in NAshell,without a�ecting reward or neutral cue. (A) dLight1.2 was injected in the NAshelland optical �ber implanted at the same site. Cannulaes were implanted at the RMTgfor baclofen/muscimol injusion. (B) Virus and �ber placements. (C) Peak �uorescentat cue onset showing (D) less DA decrease in response to shock cue when RMTg isinhibited, whereas (E) reward and (F) neutral cues were not a�ected. * p<0.05.

101

Figure 5.4: High cocaine-avoiders have reduced early acquisition of cocaine self-administration but higher rates of reinstatement. (A) Latencies to obtain cocaine(average of trials 4-7) correlated negatively with acquisition of lever pressing duringthe �rst three days of cocaine self-administration, and the number of presses on theactive lever was signi�cantly di�erent between high and low avoiders. (B) Latencies toobtain cocaine (average of trials 4-7) did not correlate with active lever pressing duringthe last three days of cocaine self-administration, and the average number of cocaineinfusions during the last 3 days did not di�er between high and low avoiders. Runwaylatencies to obtain cocaine correlated positively with both cued (C) and cocaine-primed (D) reinstatement lever pressing. * p<0.05, ** p<0.001, **** p<0.0001.

102

Chapter 6

Discussion and Future Directions

6.1 Implications for Addiction and Reinstatement

Since the �rst experimental demonstration that a drug of abuse supports instrumental

behavior, drugs have been discussed in the context of their rewarding e�ects, which

are assumed to drive and maintain drug-taking behavior. Indeed, drug reward has

been crucial in the development of most models of drug use, abuse, and addiction.

Over the last few decades, however, drugs of abuse have been increasingly recog-

nized as intricate compounds producing multiple stimulus e�ects, not all of which

are rewarding. The aversive e�ects of such drugs, for example, have been recognized

by a number of scientist working in the �eld, although not many e�orts have been

made to understand the role of these aversive e�ects in drug taking. We recognize

that the aversive e�ects of drugs should be taken into consideration in ongoing at-

tempts to model drug-taking behavior. Importantly, it has been suggested by several

researchers that drug taking may be a function of a delicate balance between drug

reward and withdrawal [Stolerman and d'Mello, 1981, Gaiardi et al., 1991b, Riley and

Simpson, 2001, Riley, 2011]. If so, the discussion of drug use and abuse requires the

103

examination of both of these drug e�ects, their interaction, and the di�erent factors

a�ecting them. In the present study, we present evidence in support of the inclusion

of the aversive e�ects of drugs in discussions of drug use and abuse. Even though we

haven't emphasized long-term adaptations occurring over weeks, months, and years,

the shorter-term e�ects examined here in�uence addictive behavior, where immediate

aversive responses to acute drug withdrawal conferred a protective e�ect against ad-

diction and helped to slow or prevent the transition from casual to compulsive drug

use [Rezvani et al., 2010].

And although our study focused on characterizing individual di�erence during

that initial phase of addiction and helped us identify protective factors, it would be

extremely useful to also understand the di�erences that mark later phases of drug

addiction. Interestingly, we found that high-avoiders, that were initially protected

against addiction, had higher reinstatement compared to low-avoiders. High-avoiders

might indeed seek the drug more because in addition to seeking the rewarding e�ects

of cocaine, they might be motivated by alleviating withdrawal symtpoms/conditioned-

withdrawal to the cue. These �ndings call to use the same approach that emphasizes

individual di�erences in earlier stages and extend it to the stages of withdrawal and

reinstatement. Those stages might involve completely di�erent molecular targets and

circuitry than the ones previously identi�ed. Indeed, although RMTg has been shown

to be activated during cocaine withdrawal [Mahler and Aston-Jones, 2012], and might

contribute to the negative a�ect in withdrawal and drug-cravings, our lab has shown

that it does not promote reinstatement. On the contrary, RMTg stimulation during

cocaine-seeking produced enduring reductions in reinstatement (unpublished data).

Alternatively, we would be interested in looking at the circuitry underlying negative

reinforcement and, as a result, active avoidance. We mention "negative reinforce-

ment" here because external stimuli associated with environmental stress or drug

104

withdrawal could be indeed viewed as negative reinforcers that contribute to instru-

mental drug seeking and consumption responses by strengthening behaviors that allow

escape from and/or avoidance of the aversive states elicited by these stimuli. One

study in particular showed that the Roman-high rats selected for rapid acquisition of

two-way active avoidance in the shuttle-box had higher reinstatement compared to

their low-avoidance rat lines which have poor acquisition of that same task [Fattore

et al., 2009], although the two lines were also di�erent on many other phenotypes,

suggesting that reinstatement could be viewed as an active avoidance phenomenon

driven by negative reinforcement. However, little is known about the brain mech-

anisms involved and better understanding of the neural basis of escape/avoidance

behavior will provide important information that will allow a deeper understanding

of the role of aversive states in substance abuse. Moreover, since the little we know

about active avoidance suggests that it shares some common circuitry with reinstate-

ment [Martinez et al., 2013], [Ramirez et al., 2015], for future studies, we will shift

our focus towards investigating the circuitry involved in active avoidance and deter-

mine if the corresponding brain areas exhibit adaptive changes that correlate with

reinstatement. Speci�cally, we will examine the contribution of the amygdala and the

contribution of anatomical outputs of key amygdala nuclei (especially the projection

from the basal amygdala to the nucleus accumbens) and the role of individual di�er-

ences in fear and avoidance and their value in predicting susceptibility to addiction

and relapse.

6.2 Identi�cation of Potential Therapeutics

As we have mentioned, aversive e�ects of drugs strongly in�uence abuse potential.

For example, nicotine and alcohol both produce aversive e�ects that are particu-

105

larly strong in some individuals, conferring on them protection against acquisition

of smoking and drinking habits. These protective e�ects are due in part to func-

tional variations in speci�c molecular targets, including alpha5 nicotinic receptors for

smoking, and an ALDH2 for drinking [Morton et al., 2018, Fowler et al., 2011, Wall

et al., 2000, Bierut et al., 2008]. Cocaine also produces aversive e�ects, which vary

between individuals, but unlike with nicotine or alcohol, the molecular processes reg-

ulating were still largely unexplored, hampering the search for novel treatments for

cocaine addiction, for which no FDA-approved medications exist [Maxwell, 2020].

Nowadays,the main treatments against psychostimulant addiction are focused on be-

havioral interventions [Phillips et al., 2014], with some pharmacological treatments

available for abstinent individuals [Tomko et al., 2018], which call for a better under-

standing of individual di�erences in cocaine addiction propensity as a means to more

personalized treatment options.

Our study took advantage of the individual di�erences in animal's responses to

one such factor that contributes to the likelihood to become addicted, aversion to

cocaine, to unveil underlying biomarkers, at di�erent circuitry levels, that could be

used as potential treatments. Indeed, the concept of using drugs that potentiate

aversion to drugs has already been explored, as in the case of disul�ram (Antabuse)

which inhibits ADLH2, thereby reducing alcohol intake. For instance in the case

of cocaine, individuals who carry the "low-aversion" pro�le can be given a PTEN

inhibitor to be taken whenever drugs are consumed. Interestingly, if these �ndings

can be extended to heroin, we could be able to apply them in a clinical settings

when giving pain medicine to patients, and take into account their vulnerability to

addiction when prescribing the drug. And �nally, if we extend this approach to the

later stages of addiction, we could identify biomarkers that are involved in withdrawal

and drug cravings in addicted individuals that are motivated by these negative a�ects,

106

and target them in treatment. For instance cocaine withdrawal in humans in the

outpatient setting is often characterized by severe depressive symptoms combined

with irritability, anxiety, and anhedonia lasting several hours to several days (i.e., the

�crash�) and may be one of the major motivating factors in the maintenance of the

cocaine-dependence cycle [Gawin and Kleber, 1986]. We could again use the pro�le of

the individual to identify what underlying mechanisms are potentially driving relapse

(drug withdrawal, depression, reward-seeking, punishment resistance etc) and provide

treatment accordingly.

Findings from this project may also help interpret genomic targets identi�ed by

another ongoing collaborative study in our lab (1U01DA044468, as of April 2018)

that tackles a similar problem (individual di�erences in cocaine avoidance) from a

very di�erent angle, by examining DNA and mRNA correlates of behavioral dif-

ferences. Combining the �ndings of this study with the next-generation DNA and

mRNA sequencing approach will help extend the �ndings from this study, and po-

tentially discover previously unknown genetic correlates of addiction-propensity. As

a consequence, we can eventually, based on the genetic pro�le of the individual, tailor

treatments to individuals depending on their genetic pro�le.

Finally, we should keep in mind that addiction results from an interaction of a

panoply of underlying phenotypes, which makes the hope for �nding a single highly

e�ective drug unlikely. Indeed, although a number of medications acting as agonists,

antagonists, or antidepressants have been evaluated in both animal models and hu-

mans, only limited success has come out of them [Carrera et al., 2001, Berger et al.,

1989, Maxwell, 2020], which suggests that pharmacological agents must be part of

a comprehensive approach toward treatment, targeting each one of the contributing

factors. This is where deconstructing addiction into its subcomponents is crucial.

Indeed, even though this study is focused on the aversive components of cocaine, our

107

lab aims at identifying DNA and mRNA correlates of several other phenotypes that

have been shown to contribute to addiction propensity, such as exploratory behavior

and resistance to punishment. This comprehensive approach to understanding indi-

vidual di�erences in drug addiction will allow us to understand molecular basis of

several phenotypes that dictate the course of addiction, and the interaction between

these phenotypes, which in turn would improve our understanding, prediction, and

treatment, of addiction, and possibly lead to personalized counseling and therapeutic

substrates.

108

Bibliography

[Anastasio et al., 2013] Anastasio, N. C., Gilbertson, S. R., Bubar, M. J., Agarkov,

A., Stutz, S. J., Jeng, Y., Bremer, N. M., Smith, T. D., Fox, R. G., Swinford, S. E.,

et al. (2013). Peptide inhibitors disrupt the serotonin 5-ht2c receptor interaction

with phosphatase and tensin homolog to allosterically modulate cellular signaling

and behavior. Journal of Neuroscience, 33(4):1615�1630.

[Andén et al., 1979] Andén, N.-E., Grabowska-Andén, M., and Wachtel, H. (1979).

E�ects of the gaba receptor agonist muscimol on the turnover of brain dopamine

and on the motor activity of rats. Acta pharmacologica et toxicologica, 44(3):191�

196.

[Anthony et al., 1989] Anthony, J. C., Tien, A. Y., and Petronis, K. R. (1989). Epi-

demiologic evidence on cocaine use and panic attacks. American Journal of Epi-

demiology, 129(3):543�549.

[Baker, 1987] Baker, T. B. (1987). Drugs: A psychobiologi-cal analysis of urges.

Alcohol and addictive behavior, 34:257.

[Balcita-Pedicino et al., 2011] Balcita-Pedicino, J. J., Omelchenko, N., Bell, R., and

Sesack, S. R. (2011). The inhibitory in�uence of the lateral habenula on midbrain

dopamine cells: ultrastructural evidence for indirect mediation via the rostromedial

109

mesopontine tegmental nucleus. Journal of Comparative Neurology, 519(6):1143�

1164.

[Barrot et al., 2012] Barrot, M., Sesack, S. R., Georges, F., Pistis, M., Hong, S., and

Jhou, T. C. (2012). Braking dopamine systems: a new gaba master structure for

mesolimbic and nigrostriatal functions. Journal of Neuroscience, 32(41):14094�

14101.

[Bau et al., 2001] Bau, C. H., Spode, A., Ponso, A. C., Elias, E. P., Garcia, C. E.,

Costa, F. T., and Hutz, M. H. (2001). Heterogeneity in early onset alcoholism

suggests a third group of alcoholics. Alcohol, 23(1):9�13.

[Beardsley and Shelton, 2012] Beardsley, P. M. and Shelton, K. L. (2012). Prime-,

stress-, and cue-induced reinstatement of extinguished drug-reinforced responding

in rats: Cocaine as the prototypical drug of abuse. Current protocols in neuro-

science, 61(1):9�39.

[Belin et al., 2016] Belin, D., Belin-Rauscent, A., Everitt, B. J., and Dalley, J. W.

(2016). In search of predictive endophenotypes in addiction: insights from preclin-

ical research. Genes, Brain and Behavior, 15(1):74�88.

[Belin et al., 2008] Belin, D., Mar, A. C., Dalley, J. W., Robbins, T. W., and Everitt,

B. J. (2008). High impulsivity predicts the switch to compulsive cocaine-taking.

Science, 320(5881):1352�1355.

[Ben-Shahar et al., 2008] Ben-Shahar, O., Posthumus, E. J., Waldroup, S. A., and

Ettenberg, A. (2008). Heightened drug-seeking motivation following extended daily

access to self-administered cocaine. Progress in Neuro-Psychopharmacology and

Biological Psychiatry, 32(3):863�869.

110

[Berger et al., 1989] Berger, P., Gawin, F., and Kosten, T. (1989). Treatment of

cocaine abuse with mazindol. The Lancet, 333(8632):283.

[Bierut et al., 2008] Bierut, L. J., Stitzel, J. A., Wang, J. C., Hinrichs, A. L., Grucza,

R. A., Xuei, X., Saccone, N. L., Saccone, S. F., Bertelsen, S., Fox, L., Horton, W. J.,

Breslau, N., Budde, J., Cloninger, C. R., Dick, D. M., Foroud, T., Hatsukami, D.,

Hesselbrock, V., Johnson, E. O., Kramer, J., Kuperman, S., Madden, P. A., Mayo,

K., Nurnberger, J., Pomerleau, O., Porjesz, B., Reyes, O., Schuckit, M., Swan, G.,

Tisch�eld, J. A., Edenberg, H. J., Rice, J. P., and Goate, A. M. (2008). Variants in

nicotinic receptors and risk for nicotine dependence. Am J Psychiatry, 165(9):1163�

1171.

[Bock et al., 2013] Bock, R., Shin, J. H., Kaplan, A. R., Dobi, A., Markey, E.,

Kramer, P. F., Gremel, C. M., Christensen, C. H., Adrover, M. F., and Alvarez,

V. A. (2013). Strengthening the accumbal indirect pathway promotes resilience to

compulsive cocaine use. Nature neuroscience, 16(5):632.

[Borroto-Escuela et al., 2015] Borroto-Escuela, D. O., O�ijan, J., Pinton, L., Wydra,

K., Narváez, M., Hansson, A., Ferraro, L., Filip, M., and Fuxe, K. (2015). Adeno-

sine a2a-dopamine d2 heteroreceptor complexes in the brain: Relevance for co-

caine addiction. In ALCOHOL AND ALCOHOLISM, volume 50. OXFORD UNIV

PRESS GREAT CLARENDON ST, OXFORD OX2 6DP, ENGLAND.

[Bourdy et al., 2014] Bourdy, R., Sánchez-Catalán, M.-J., Kau�ing, J., Balcita-

Pedicino, J. J., Freund-Mercier, M.-J., Veinante, P., Sesack, S. R., Georges, F.,

and Barrot, M. (2014). Control of the nigrostriatal dopamine neuron activity and

motor function by the tail of the ventral tegmental area. Neuropsychopharmacology,

39(12):2788�2798.

111

[Bowery et al., 1980] Bowery, N., Hill, D., Hudson, A., Doble, A., Middlemiss, D.,

Shaw, J., and Turnbull, M. (1980). (�) baclofen decreases neurotransmitter release

in the mammalian cns by an action at a novel gaba receptor. Nature, 283(5742):92�

94.

[Brinschwitz et al., 2010] Brinschwitz, K., Dittgen, A., Madai, V., Lommel, R.,

Geisler, S., and Veh, R. (2010). Glutamatergic axons from the lateral habenula

mainly terminate on gabaergic neurons of the ventral midbrain. Neuroscience,

168(2):463�476.

[Brischoux et al., 2009] Brischoux, F., Chakraborty, S., Brierley, D. I., and Ungless,

M. A. (2009). Phasic excitation of dopamine neurons in ventral vta by noxious

stimuli. Proceedings of the national academy of sciences, 106(12):4894�4899.

[Britt and Bonci, 2013] Britt, J. P. and Bonci, A. (2013). Optogenetic interroga-

tions of the neural circuits underlying addiction. Current opinion in neurobiology,

23(4):539�545.

[Bromberg-Martin et al., 2010] Bromberg-Martin, E. S., Matsumoto, M., and

Hikosaka, O. (2010). Dopamine in motivational control: rewarding, aversive, and

alerting. Neuron, 68(5):815�834.

[Brown et al., 2017] Brown, P. L., Palacorolla, H., Brady, D., Riegger, K., Elmer,

G. I., and Shepard, P. D. (2017). Habenula-induced inhibition of midbrain

dopamine neurons is diminished by lesions of the rostromedial tegmental nucleus.

Journal of Neuroscience, 37(1):217�225.

[Cain et al., 2005] Cain, M. E., Saucier, D. A., and Bardo, M. T. (2005). Novelty

seeking and drug use: contribution of an animal model. Experimental and Clinical

Psychopharmacology, 13(4):367.

112

[Calipari et al., 2016] Calipari, E. S., Bagot, R. C., Purushothaman, I., Davidson,

T. J., Yorgason, J. T., Peña, C. J., Walker, D. M., Pirpinias, S. T., Guise, K. G.,

Ramakrishnan, C., et al. (2016). In vivo imaging identi�es temporal signature of

d1 and d2 medium spiny neurons in cocaine reward. Proceedings of the National

Academy of Sciences, 113(10):2726�2731.

[Carrera et al., 2001] Carrera, M. R. A., Ashley, J. A., Wirsching, P., Koob, G. F.,

and Janda, K. D. (2001). A second-generation vaccine protects against the psy-

choactive e�ects of cocaine. Proceedings of the National Academy of Sciences,

98(4):1988�1992.

[Centonze et al., 2004] Centonze, D., Battista, N., Rossi, S., Mercuri, N. B., Finazzi-

Agro, A., Bernardi, G., Calabresi, P., and Maccarrone, M. (2004). A critical

interaction between dopamine d2 receptors and endocannabinoids mediates the

e�ects of cocaine on striatal gabaergic transmission. Neuropsychopharmacology,

29(8):1488�1497.

[Christoph et al., 1986] Christoph, G. R., Leonzio, R. J., and Wilcox, K. S. (1986).

Stimulation of the lateral habenula inhibits dopamine-containing neurons in the

substantia nigra and ventral tegmental area of the rat. Journal of Neuroscience,

6(3):613�619.

[Cohen and Ettenberg, 2007] Cohen, A. and Ettenberg, A. (2007). Motivational ef-

fects of nicotine as measured in a runway model of drug self-administration. Be-

havioural pharmacology, 18(4):265�271.

[Cohen et al., 2012] Cohen, J. Y., Haesler, S., Vong, L., Lowell, B. B., and Uchida,

N. (2012). Neuron-type-speci�c signals for reward and punishment in the ventral

tegmental area. nature, 482(7383):85.

113

[Coizet et al., 2006] Coizet, V., Dommett, E., Redgrave, P., and Overton, P. (2006).

Nociceptive responses of midbrain dopaminergic neurones are modulated by the

superior colliculus in the rat. Neuroscience, 139(4):1479�1493.

[Cox et al., 1990] Cox, B. J., Norton, G. R., Swinson, R. P., and Endler, N. S. (1990).

Substance abuse and panic-related anxiety: a critical review. Behaviour research

and therapy, 28(5):385�393.

[Crabbe, 2010] Crabbe, J. C. (2010). Consilience of rodent and human phenotypes

relevant for alcohol dependence. Addiction biology, 15(2):103�108.

[Crespi, 1942] Crespi, L. P. (1942). Quantitative variation of incentive and perfor-

mance in the white rat. The American Journal of Psychology, 55(4):467�517.

[Crespo et al., 2006] Crespo, J. A., Sturm, K., Saria, A., and Zernig, G. (2006). Acti-

vation of muscarinic and nicotinic acetylcholine receptors in the nucleus accumbens

core is necessary for the acquisition of drug reinforcement. Journal of Neuroscience,

26(22):6004�6010.

[Czachowski, 1999] Czachowski, C. L. (1999). Ethanol-seeking and self-

administration in a runway paradigm.

[d'Alcantara et al., 2001] d'Alcantara, P., Ledent, C., Swillens, S., and Schi�mann,

S. N. (2001). Inactivation of adenosine a2a receptor impairs long term potentiation

in the accumbens nucleus without altering basal synaptic transmission. Neuro-

science, 107(3):455�464.

[Danjo et al., 2014] Danjo, T., Yoshimi, K., Funabiki, K., Yawata, S., and Nakanishi,

S. (2014). Aversive behavior induced by optogenetic inactivation of ventral tegmen-

114

tal area dopamine neurons is mediated by dopamine d2 receptors in the nucleus

accumbens. Proceedings of the National Academy of Sciences, 111(17):6455�6460.

[Davis et al., 2010] Davis, M., Walker, D. L., Miles, L., and Grillon, C. (2010). Phasic

vs sustained fear in rats and humans: role of the extended amygdala in fear vs

anxiety. Neuropsychopharmacology, 35(1):105�135.

[Deroche et al., 1999] Deroche, V., Le Moal, M., and Piazza, P. V. (1999). Cocaine

self-administration increases the incentive motivational properties of the drug in

rats. European Journal of Neuroscience, 11(8):2731�2736.

[Deroche-Gamonet et al., 2004] Deroche-Gamonet, V., Belin, D., and Piazza, P. V.

(2004). Evidence for addiction-like behavior in the rat. Science, 305(5686):1014�

1017.

[Deroche-Gamonet and Piazza, 2014] Deroche-Gamonet, V. and Piazza, P. V.

(2014). Psychobiology of cocaine addiction: Contribution of a multi-symptomatic

animal model of loss of control. Neuropharmacology, 76:437�449.

[Di Chiara and Imperato, 1988] Di Chiara, G. and Imperato, A. (1988). Drugs

abused by humans preferentially increase synaptic dopamine concentrations in the

mesolimbic system of freely moving rats. Proceedings of the National Academy of

Sciences, 85(14):5274�5278.

[Diana et al., 1993] Diana, M., Pistis, M., Carboni, S., Gessa, G. L., and RoSSETTI,

Z. L. (1993). Profound decrement of mesolimbic dopaminergic neuronal activity

during ethanol withdrawal syndrome in rats: electrophysiological and biochemical

evidence. Proceedings of the National Academy of Sciences, 90(17):7966�7969.

115

[Dias et al., 2012] Dias, R. B., Ribeiro, J. A., and Sebastiao, A. M. (2012). En-

hancement of ampa currents and glur1 membrane expression through pka-coupled

adenosine a2a receptors. Hippocampus, 22(2):276�291.

[Du et al., 2013] Du, H.-C., Anastasio, N. C., Briggs, J. M., Cunningham, K. A.,

and Gilbertson, S. R. (2013). Disruption of pten: 5-ht2cr coupling: A potential

treatment for restoration of 5-ht2cr function. In ABSTRACTS OF PAPERS OF

THE AMERICAN CHEMICAL SOCIETY, volume 245. AMER CHEMICAL SOC

1155 16TH ST, NW, WASHINGTON, DC 20036 USA.

[Elmer et al., 2019] Elmer, G., Palacorolla, H., Mayo, C., Brown, P., Jhou, T., Brady,

D., and Shepard, P. (2019). The rostromedial tegmental nucleus modulates the de-

velopment of stress-induced helpless behavior. Behavioural brain research, 359:950�

957.

[Eshel et al., 2015] Eshel, N., Bukwich, M., Rao, V., Hemmelder, V., Tian, J., and

Uchida, N. (2015). Arithmetic and local circuitry underlying dopamine prediction

errors. Nature, 525(7568):243�246.

[Ettenberg, 1990] Ettenberg, A. (1990). Haloperidol prevents the reinstatement of

amphetamine-rewarded runway responding in rats. Pharmacology Biochemistry

and Behavior, 36(3):635�638.

[Ettenberg, 2004] Ettenberg, A. (2004). Opponent process properties of self-

administered cocaine. Neuroscience & Biobehavioral Reviews, 27(8):721�728.

[Ettenberg, 2009] Ettenberg, A. (2009). The runway model of drug self-

administration. Pharmacology Biochemistry and Behavior, 91(3):271�277.

116

[Ettenberg and Bernardi, 2006] Ettenberg, A. and Bernardi, R. E. (2006).

Anxiolytic-like actions of buspirone in a runway model of intravenous cocaine self-

administration. Pharmacology Biochemistry and Behavior, 85(2):393�399.

[Ettenberg and Bernardi, 2007] Ettenberg, A. and Bernardi, R. E. (2007). E�ects

of buspirone on the immediate positive and delayed negative properties of intra-

venous cocaine as measured in the conditioned place preference test. Pharmacology

Biochemistry and Behavior, 87(1):171�178.

[Ettenberg et al., 2015] Ettenberg, A., Fomenko, V., Kaganovsky, K., Shelton, K.,

and Wenzel, J. M. (2015). On the positive and negative a�ective responses to

cocaine and their relation to drug self-administration in rats. Psychopharmacology,

232(13):2363�2375.

[Ettenberg and Geist, 1991] Ettenberg, A. and Geist, T. D. (1991). Animal model

for investigating the anxiogenic e�ects of self-administered cocaine. Psychophar-

macology, 103(4):455�461.

[Ettenberg and Geist, 1993] Ettenberg, A. and Geist, T. D. (1993). Qualitative and

quantitative di�erences in the operant runway behavior of rats working for cocaine

and heroin reinforcement. Pharmacology Biochemistry and Behavior, 44(1):191�

198.

[Ettenberg et al., 1996] Ettenberg, A., MacConell, L. A., and Geist, T. D. (1996).

E�ects of haloperidol in a response-reinstatement model of heroin relapse. Psy-

chopharmacology, 124(3):205�210.

[Ettenberg et al., 1999] Ettenberg, A., Raven, M. A., Danluck, D. A., and Neces-

sary, B. D. (1999). Evidence for opponent-process actions of intravenous cocaine.

Pharmacology Biochemistry and Behavior, 64(3):507�512.

117

[Farrar et al., 2010] Farrar, A. M., Segovia, K. N., Randall, P. A., Nunes, E. J.,

Collins, L. E., Stopper, C. M., Port, R. G., Hockemeyer, J., Müller, C. E., Correa,

M., et al. (2010). Nucleus accumbens and e�ort-related functions: behavioral

and neural markers of the interactions between adenosine a2a and dopamine d2

receptors. Neuroscience, 166(4):1056�1067.

[Fattore et al., 2009] Fattore, L., Piras, G., Corda, M. G., and Giorgi, O. (2009).

The roman high-and low-avoidance rat lines di�er in the acquisition, maintenance,

extinction, and reinstatement of intravenous cocaine self-administration. Neuropsy-

chopharmacology, 34(5):1091�1101.

[Ferré et al., 2016] Ferré, S., Bonaventura, J., Tomasi, D., Navarro, G., Moreno, E.,

Cortés, A., Lluís, C., Casadó, V., and Volkow, N. D. (2016). Allosteric mechanisms

within the adenosine a2a�dopamine d2 receptor heterotetramer. Neuropharmacol-

ogy, 104:154�160.

[Ferris et al., 2014] Ferris, M. J., España, R. A., Locke, J. L., Konstantopoulos, J. K.,

Rose, J. H., Chen, R., and Jones, S. R. (2014). Dopamine transporters govern diur-

nal variation in extracellular dopamine tone. Proceedings of the National Academy

of Sciences, 111(26):E2751�E2759.

[Fink et al., 1992] Fink, J. S., Weaver, D. R., Rivkees, S. A., Peterfreund, R. A.,

Pollack, A. E., Adler, E. M., and Reppert, S. M. (1992). Molecular cloning of the

rat a2 adenosine receptor: selective co-expression with d2 dopamine receptors in

rat striatum. Molecular Brain Research, 14(3):186�195.

[Fontana and Commissaris, 1989] Fontana, D. J. and Commissaris, R. L. (1989). Ef-

fects of cocaine on con�ict behavior in the rat. Life sciences, 45(9):819�827.

118

[Fowler et al., 2011] Fowler, C. D., Lu, Q., Johnson, P. M., Marks, M. J., and Kenny,

P. J. (2011). Habenular α5 nicotinic receptor subunit signalling controls nicotine

intake. Nature, 471(7340):597�601.

[Fox et al., 2018] Fox, M. E., Chandra, R., Menken, M. S., Larkin, E. J., Nam, H.,

Engeln, M., Francis, T. C., and Lobo, M. K. (2018). Dendritic remodeling of d1

neurons by rhoa/rho-kinase mediates depression-like behavior. Molecular psychia-

try, pages 1�13.

[Fu et al., 2016] Fu, R., Zuo, W., Gregor, D., Li, J., Grech, D., and Ye, J.-H. (2016).

Pharmacological manipulation of the rostromedial tegmental nucleus changes vol-

untary and operant ethanol self-administration in rats. Alcoholism: Clinical and

Experimental Research, 40(3):572�582.

[Fu et al., 2019] Fu, R., Zuo, W., Shiwalkar, N., Mei, Q., Fan, Q., Chen, X., Li, J.,

Bekker, A., and Ye, J.-H. (2019). Alcohol withdrawal drives depressive behaviors

by activating neurons in the rostromedial tegmental nucleus. Neuropsychopharma-

cology, 44(8):1464�1475.

[Fürst, 2016] Fürst, R. (2016). Narciclasine�an amaryllidaceae alkaloid with potent

antitumor and anti-in�ammatory properties. Planta medica, 82(16):1389�1394.

[Gaiardi et al., 1991a] Gaiardi, M., Bartoletti, M., Bacchi, A., Gubellini, C., Costa,

M., and Babbini, M. (1991a). Role of repeated exposure to morphine in determining

its a�ective properties: place and taste conditioning studies in rats. Psychophar-

macology, 103(2):183�186.

[Gaiardi et al., 1991b] Gaiardi, M., Bartoletti, M., Bacchi, A., Gubellini, C., Costa,

M., and Babbini, M. (1991b). Role of repeated exposure to morphine in determining

119

its a�ective properties: place and taste conditioning studies in rats. Psychophar-

macology, 103(2):183�186.

[Gawin, 1991] Gawin, F. (1991). Gawin fh. cocaine addiction: psychology and neu-

rophysiology. science 251: 1580-1586. Science (New York, N.Y.), 251:1580�6.

[Gawin and Kleber, 1986] Gawin, F. H. and Kleber, H. D. (1986). Abstinence symp-

tomatology and psychiatric diagnosis in cocaine abusers: clinical observations.

Archives of general psychiatry, 43(2):107�113.

[Geist and Ettenberg, 1997a] Geist, T. D. and Ettenberg, A. (1997a). Concurrent

positive and negative goalbox events produce runway behaviors comparable to those

of cocaine-reinforced rats. Pharmacology Biochemistry and Behavior, 57(1-2):145�

150.

[Geist and Ettenberg, 1997b] Geist, T. D. and Ettenberg, A. (1997b). Concurrent

positive and negative goalbox events produce runway behaviors comparable to those

of cocaine-reinforced rats. Pharmacology Biochemistry and Behavior, 57(1-2):145�

150.

[Geracioti and Post, 1991] Geracioti, T. D. and Post, R. M. (1991). Onset of panic

disorder associated with rare use of cocaine. Biological psychiatry, 29(4):403�406.

[Gerfen and Surmeier, 2011] Gerfen, C. R. and Surmeier, D. J. (2011). Modulation

of striatal projection systems by dopamine. Annual review of neuroscience, 34:441�

466.

[Gipson et al., 2014] Gipson, C. D., Kupchik, Y. M., and Kalivas, P. W. (2014).

Rapid, transient synaptic plasticity in addiction. Neuropharmacology, 76:276�286.

120

[Giu�rida et al., 2001] Giu�rida, A., Beltramo, M., and Piomelli, D. (2001). Mecha-

nisms of endocannabinoid inactivation: biochemistry and pharmacology. Journal

of Pharmacology and Experimental Therapeutics, 298(1):7�14.

[Glover et al., 2016] Glover, E. J., McDougle, M. J., Siegel, G. S., Jhou, T. C., and

Chandler, L. J. (2016). Role for the rostromedial tegmental nucleus in signaling

the aversive properties of alcohol. Alcoholism: Clinical and Experimental Research,

40(8):1651�1661.

[Gonçalves et al., 2012] Gonçalves, L., Sego, C., and Metzger, M. (2012). Di�erential

projections from the lateral habenula to the rostromedial tegmental nucleus and

ventral tegmental area in the rat. Journal of Comparative Neurology, 520(6):1278�

1300.

[Goodwin et al., 2002] Goodwin, R. D., Stayner, D. A., Chinman, M. J., Wu, P.,

Tebes, J. K., and Davidson, L. (2002). The relationship between anxiety and

substance use disorders among individuals with severe a�ective disorders. Compre-

hensive psychiatry, 43(4):245�252.

[Goto and Grace, 2005] Goto, Y. and Grace, A. A. (2005). Dopamine-dependent

interactions between limbic and prefrontal cortical plasticity in the nucleus accum-

bens: disruption by cocaine sensitization. Neuron, 47(2):255�266.

[Gottesman and Gould, 2003] Gottesman, I. I. and Gould, T. D. (2003). The en-

dophenotype concept in psychiatry: etymology and strategic intentions. American

journal of psychiatry, 160(4):636�645.

[Goudie, 1979] Goudie, A. (1979). Aversive stimulus properties of drugs. Neurophar-

macology, 18(12):971�979.

121

[Greco et al., 2008] Greco, M.-A., Fuller, P. M., Jhou, T. C., Martin-Schild, S., Zad-

ina, J. E., Hu, Z., Shiromani, P., and Lu, J. (2008). Opioidergic projections to

sleep-active neurons in the ventrolateral preoptic nucleus. Brain research, 1245:96�

107.

[Grueter et al., 2012] Grueter, B. A., Rothwell, P. E., and Malenka, R. C. (2012).

Integrating synaptic plasticity and striatal circuit function in addiction. Current

opinion in neurobiology, 22(3):545�551.

[Gutilla et al., 2016] Gutilla, E. A., Buyukozturk, M. M., and Steward, O. (2016).

Long-term consequences of conditional genetic deletion of pten in the sensorimotor

cortex of neonatal mice. Experimental neurology, 279:27�39.

[Guzman and Ettenberg, 2004] Guzman, D. and Ettenberg, A. (2004). Heroin atten-

uates the negative consequences of cocaine in a runway model of self-administration.

Pharmacology Biochemistry and Behavior, 79(2):317�324.

[Guzman and Ettenberg, 2007] Guzman, D. and Ettenberg, A. (2007). Runway self-

administration of intracerebroventricular cocaine: evidence of mixed positive and

negative drug actions. Behavioural pharmacology, 18(1):53�60.

[Harper et al., 2006] Harper, L., Beckett, S., Marsden, C., McCreary, A., and Alexan-

der, S. (2006). E�ects of the a2a adenosine receptor antagonist kw6002 in the nu-

cleus accumbens in vitro and in vivo. Pharmacology Biochemistry and Behavior,

83(1):114�121.

[Heinrichs et al., 1998] Heinrichs, S. C., Klaassen, A., Koob, G. F., Schulteis, G.,

Ahmed, S., and De Souza, E. B. (1998). Corticotropin-releasing factor receptor

blockade enhances conditioned aversive properties of cocaine in rats. Psychophar-

macology, 136(3):247�255.

122

[Herkenham and Nauta, 1977] Herkenham, M. and Nauta, W. J. (1977). A�erent

connections of the habenular nuclei in the rat. a horseradish peroxidase study, with

a note on the �ber-of-passage problem. In Neuroanatomy, pages 277�299. Springer.

[Heyman, 2013] Heyman, G. M. (2013). Addiction and choice: theory and new data.

Frontiers in psychiatry, 4:31.

[Hikida et al., 2010] Hikida, T., Kimura, K., Wada, N., Funabiki, K., and Nakanishi,

S. (2010). Distinct roles of synaptic transmission in direct and indirect striatal

pathways to reward and aversive behavior. Neuron, 66(6):896�907.

[Hikida et al., 2013] Hikida, T., Yawata, S., Yamaguchi, T., Danjo, T., Sasaoka, T.,

Wang, Y., and Nakanishi, S. (2013). Pathway-speci�c modulation of nucleus ac-

cumbens in reward and aversive behavior via selective transmitter receptors. Pro-

ceedings of the National Academy of Sciences, 110(1):342�347.

[Hikosaka, 2010] Hikosaka, O. (2010). The habenula: from stress evasion to value-

based decision-making. Nature reviews neuroscience, 11(7):503�513.

[Hong et al., 2011] Hong, S., Jhou, T. C., Smith, M., Saleem, K. S., and Hikosaka, O.

(2011). Negative reward signals from the lateral habenula to dopamine neurons are

mediated by rostromedial tegmental nucleus in primates. Journal of Neuroscience,

31(32):11457�11471.

[Hull, 1934] Hull, C. L. (1934). The rat's speed-of-locomotion gradient in the ap-

proach to food. Journal of comparative psychology, 17(3):393.

[Jalabert et al., 2011] Jalabert, M., Bourdy, R., Courtin, J., Veinante, P., Manzoni,

O. J., Barrot, M., and Georges, F. (2011). Neuronal circuits underlying acute

123

morphine action on dopamine neurons. Proceedings of the national academy of

sciences, 108(39):16446�16450.

[Jenner et al., 2009] Jenner, P., Mori, A., Hauser, R., Morelli, M., Fredholm, B., and

Chen, J. (2009). Adenosine, adenosine a2a antagonists, and parkinson's disease.

Parkinsonism & related disorders, 15(6):406�413.

[Jhou et al., 2009a] Jhou, T. C., Fields, H. L., Baxter, M. G., Saper, C. B., and

Holland, P. C. (2009a). The rostromedial tegmental nucleus (rmtg), a gabaergic

a�erent to midbrain dopamine neurons, encodes aversive stimuli and inhibits motor

responses. Neuron, 61(5):786�800.

[Jhou et al., 2009b] Jhou, T. C., Geisler, S., Marinelli, M., Degarmo, B. A., and

Zahm, D. S. (2009b). The mesopontine rostromedial tegmental nucleus: a structure

targeted by the lateral habenula that projects to the ventral tegmental area of tsai

and substantia nigra compacta. Journal of Comparative Neurology, 513(6):566�596.

[Jhou et al., 2009c] Jhou, T. C., Geisler, S., Marinelli, M., Degarmo, B. A., and

Zahm, D. S. (2009c). The mesopontine rostromedial tegmental nucleus: a structure

targeted by the lateral habenula that projects to the ventral tegmental area of tsai

and substantia nigra compacta. Journal of Comparative Neurology, 513(6):566�596.

[Jhou et al., 2013] Jhou, T. C., Good, C. H., Rowley, C. S., Xu, S.-p., Wang, H.,

Burnham, N. W., Ho�man, A. F., Lupica, C. R., and Ikemoto, S. (2013). Co-

caine drives aversive conditioning via delayed activation of dopamine-responsive

habenular and midbrain pathways. Journal of Neuroscience, 33(17):7501�7512.

[Jhou et al., 2012] Jhou, T. C., Xu, S.-P., Lee, M. R., Gallen, C. L., and Ikemoto,

S. (2012). Mapping of reinforcing and analgesic e�ects of the mu opioid agonist

124

endomorphin-1 in the ventral midbrain of the rat. Psychopharmacology, 224(2):303�

312.

[Ji and Shepard, 2007] Ji, H. and Shepard, P. D. (2007). Lateral habenula stimula-

tion inhibits rat midbrain dopamine neurons through a gabaa receptor-mediated

mechanism. Journal of Neuroscience, 27(26):6923�6930.

[Jupp and Lawrence, 2010] Jupp, B. and Lawrence, A. J. (2010). New horizons for

therapeutics in drug and alcohol abuse. Pharmacology & therapeutics, 125(1):138�

168.

[Jurado et al., 2010] Jurado, S., Benoist, M., Knafo, S., Petrok, C. N., and Esteban,

J. A. (2010). Pten is recruited to the postsynaptic terminal for nmda receptor-

dependent long-term depression. The EMBO journal, 29(16):2827�2840.

[Kalivas and Volkow, 2005] Kalivas, P. W. and Volkow, N. D. (2005). The neural

basis of addiction: a pathology of motivation and choice. American Journal of

Psychiatry, 162(8):1403�1413.

[Kampman et al., 1998] Kampman, K. M., Volpicelli, J. R., Mcginnis, D. E., Alter-

man, A. I., Weinrieb, R. M., D'Angelo, L., and Epperson, L. E. (1998). Reliabil-

ity and validity of the cocaine selective severity assessment. Addictive behaviors,

23(4):449�461.

[Kau�ing and Aston-Jones, 2015] Kau�ing, J. and Aston-Jones, G. (2015). Persis-

tent adaptations in a�erents to ventral tegmental dopamine neurons after opiate

withdrawal. Journal of Neuroscience, 35(28):10290�10303.

125

[Kau�ing et al., 2009] Kau�ing, J., Veinante, P., Pawlowski, S. A., Freund-Mercier,

M.-J., and Barrot, M. (2009). A�erents to the gabaergic tail of the ventral tegmen-

tal area in the rat. Journal of Comparative Neurology, 513(6):597�621.

[Kau�ing et al., 2010] Kau�ing, J., Waltisperger, E., Bourdy, R., Valera, A.,

Veinante, P., Freund-Mercier, M.-J., and Barrot, M. (2010). Pharmacological re-

cruitment of the gabaergic tail of the ventral tegmental area by acute drug exposure.

British journal of pharmacology, 161(8):1677�1691.

[Kenny and Markou, 2001] Kenny, P. J. and Markou, A. (2001). Neurobiology of

the nicotine withdrawal syndrome. Pharmacology Biochemistry and Behavior,

70(4):531�549.

[Knafo et al., 2016] Knafo, S., Sánchez-Puelles, C., Palomer, E., Delgado, I., Dra�n,

J. E., Mingo, J., Wahle, T., Kaleka, K., Mou, L., Pereda-Perez, I., et al. (2016).

Pten recruitment controls synaptic and cognitive function in alzheimer's models.

Nature neuroscience, 19(3):443.

[Knapp et al., 2001] Knapp, C. M., Foye, M. M., Cottam, N., Ciraulo, D. A., and

Kornetsky, C. (2001). Adenosine agonists cgs 21680 and neca inhibit the initi-

ation of cocaine self-administration. Pharmacology Biochemistry and Behavior,

68(4):797�803.

[Koeltzow and White, 2003] Koeltzow, T. E. and White, F. J. (2003). Behavioral

depression during cocaine withdrawal is associated with decreased spontaneous

activity of ventral tegmental area dopamine neurons. Behavioral neuroscience,

117(4):860.

126

[Koob et al., 1989] Koob, G., Wall, T., and Bloom, F. (1989). Nucleus accumbens as

a substrate for the aversive stimulus e�ects of opiate withdrawal. Psychopharma-

cology, 98(4):530�534.

[Koob et al., 2004] Koob, G. F., Ahmed, S. H., Boutrel, B., Chen, S. A., Kenny, P. J.,

Markou, A., O'Dell, L. E., Parsons, L. H., and Sanna, P. P. (2004). Neurobiological

mechanisms in the transition from drug use to drug dependence. Neuroscience &

Biobehavioral Reviews, 27(8):739�749.

[Koob and Le Moal, 1997] Koob, G. F. and Le Moal, M. (1997). Drug abuse: hedonic

homeostatic dysregulation. Science, 278(5335):52�58.

[Koob and Le Moal, 2005] Koob, G. F. and Le Moal, M. (2005). Plasticity of re-

ward neurocircuitry and the'dark side'of drug addiction. Nature neuroscience,

8(11):1442�1444.

[Kravitz et al., 2012] Kravitz, A. V., Tye, L. D., and Kreitzer, A. C. (2012). Distinct

roles for direct and indirect pathway striatal neurons in reinforcement. Nature

neuroscience, 15(6):816.

[Kreitzer and Malenka, 2005] Kreitzer, A. C. and Malenka, R. C. (2005). Dopamine

modulation of state-dependent endocannabinoid release and long-term depression

in the striatum. Journal of Neuroscience, 25(45):10537�10545.

[Kuhar et al., 1991] Kuhar, M., Ritz, M., and Boja, J. (1991). The dopamine hypoth-

esis of the reinforcing properties of cocaine. Trends in neurosciences, 14(7):299�302.

[Kwako et al., 2016] Kwako, L. E., Momenan, R., Litten, R. Z., Koob, G. F., and

Goldman, D. (2016). Addictions neuroclinical assessment: a neuroscience-based

framework for addictive disorders. Biological Psychiatry, 80(3):179�189.

127

[Lahti et al., 2016] Lahti, L., Haugas, M., Tikker, L., Airavaara, M., Voutilainen,

M. H., Anttila, J., Kumar, S., Inkinen, C., Salminen, M., and Partanen, J. (2016).

Di�erentiation and molecular heterogeneity of inhibitory and excitatory neurons

associated with midbrain dopaminergic nuclei. Development, 143(3):516�529.

[Lammel et al., 2012] Lammel, S., Lim, B. K., Ran, C., Huang, K. W., Betley, M. J.,

Tye, K. M., Deisseroth, K., and Malenka, R. C. (2012). Input-speci�c control of

reward and aversion in the ventral tegmental area. Nature, 491(7423):212�217.

[Lavezzi et al., 2015] Lavezzi, H. N., Parsley, K. P., and Zahm, D. S. (2015). Mod-

ulation of locomotor activation by the rostromedial tegmental nucleus. Neuropsy-

chopharmacology, 40(3):676�687.

[Lavezzi and Zahm, 2011] Lavezzi, H. N. and Zahm, D. S. (2011). The mesopontine

rostromedial tegmental nucleus: an integrative modulator of the reward system.

Basal ganglia, 1(4):191�200.

[Lecca et al., 2011] Lecca, S., Melis, M., Luchicchi, A., Ennas, M. G., Castelli, M. P.,

Muntoni, A. L., and Pistis, M. (2011). E�ects of drugs of abuse on putative

rostromedial tegmental neurons, inhibitory a�erents to midbrain dopamine cells.

Neuropsychopharmacology, 36(3):589�602.

[Lecca et al., 2012] Lecca, S., Melis, M., Luchicchi, A., Muntoni, A. L., and Pistis,

M. (2012). Inhibitory inputs from rostromedial tegmental neurons regulate spon-

taneous activity of midbrain dopamine cells and their responses to drugs of abuse.

Neuropsychopharmacology, 37(5):1164�1176.

[Li et al., 2019a] Li, H., Pullmann, D., Cho, J. Y., Eid, M., and Jhou, T. C. (2019a).

Generality and opponency of rostromedial tegmental (rmtg) roles in valence pro-

cessing. Elife, 8:e41542.

128

[Li et al., 2019b] Li, H., Vento, P. J., Parrilla-Carrero, J., Pullmann, D., Chao, Y. S.,

Eid, M., and Jhou, T. C. (2019b). Three rostromedial tegmental a�erents drive

triply dissociable aspects of punishment learning and aversive valence encoding.

Neuron, 104(5):987�999.

[Li et al., 2005] Li, Z., Dong, X., Wang, Z., Liu, W., Deng, N., Ding, Y., Tang, L.,

Hla, T., Zeng, R., Li, L., et al. (2005). Regulation of pten by rho small gtpases.

Nature cell biology, 7(4):399�404.

[Liu et al., 2010] Liu, K., Lu, Y., Lee, J. K., Samara, R., Willenberg, R., Sears-

Kraxberger, I., Tedeschi, A., Park, K. K., Jin, D., Cai, B., et al. (2010). Pten

deletion enhances the regenerative ability of adult corticospinal neurons. Nature

neuroscience, 13(9):1075.

[Liu et al., 2008] Liu, Z.-H., Shin, R., and Ikemoto, S. (2008). Dual role of medial a10

dopamine neurons in a�ective encoding. Neuropsychopharmacology, 33(12):3010�

3020.

[Lovinger, 2008] Lovinger, D. M. (2008). Presynaptic modulation by endocannabi-

noids. In Pharmacology of Neurotransmitter Release, pages 435�477. Springer.

[Lüscher and Malenka, 2011] Lüscher, C. and Malenka, R. C. (2011). Drug-evoked

synaptic plasticity in addiction: from molecular changes to circuit remodeling.

Neuron, 69(4):650�663.

[Mahler and Aston-Jones, 2012] Mahler, S. V. and Aston-Jones, G. S. (2012). Fos

activation of selective a�erents to ventral tegmental area during cue-induced rein-

statement of cocaine seeking in rats. Journal of Neuroscience, 32(38):13309�13325.

129

[Mak et al., 2010] Mak, L. H., Vilar, R., and Woscholski, R. (2010). Characterisation

of the pten inhibitor vo-ohpic. Journal of chemical biology, 3(4):157�163.

[Mak and Woscholski, 2015] Mak, L. H. and Woscholski, R. (2015). Targeting pten

using small molecule inhibitors. Methods, 77:63�68.

[Mark et al., 1991] Mark, G., Blander, D., and Hoebel, B. (1991). A conditioned

stimulus decreases extracellular dopamine in the nucleus accumbens after the de-

velopment of a learned taste aversion. Brain research, 551(1-2):308�310.

[Martinez et al., 2011] Martinez, D., Carpenter, K. M., Liu, F., Slifstein, M., Broft,

A., Friedman, A. C., Kumar, D., Van Heertum, R., Kleber, H. D., and Nunes, E.

(2011). Imaging dopamine transmission in cocaine dependence: link between neuro-

chemistry and response to treatment. American Journal of Psychiatry, 168(6):634�

641.

[Martinez et al., 2013] Martinez, R. C., Gupta, N., Lázaro-Muñoz, G., Sears, R. M.,

Kim, S., Moscarello, J. M., LeDoux, J. E., and Cain, C. K. (2013). Active vs.

reactive threat responding is associated with di�erential c-fos expression in speci�c

regions of amygdala and prefrontal cortex. Learning & Memory, 20(8):446�452.

[Matsui et al., 2014] Matsui, A., Jarvie, B. C., Robinson, B. G., Hentges, S. T., and

Williams, J. T. (2014). Separate gaba a�erents to dopamine neurons mediate acute

action of opioids, development of tolerance, and expression of withdrawal. Neuron,

82(6):1346�1356.

[Matsui and Williams, 2011] Matsui, A. and Williams, J. T. (2011). Opioid-sensitive

gaba inputs from rostromedial tegmental nucleus synapse onto midbrain dopamine

neurons. Journal of Neuroscience, 31(48):17729�17735.

130

[Matsumoto and Hikosaka, 2007] Matsumoto, M. and Hikosaka, O. (2007). Lateral

habenula as a source of negative reward signals in dopamine neurons. Nature,

447(7148):1111�1115.

[Matsumoto and Hikosaka, 2009] Matsumoto, M. and Hikosaka, O. (2009). Two

types of dopamine neuron distinctly convey positive and negative motivational

signals. Nature, 459(7248):837�841.

[Maxwell, 2020] Maxwell, J. C. (2020). Is cocaine coming back? a commentary.

Substance Use & Misuse, 55(2):345�348.

[McFarland and Ettenberg, 1995] McFarland, K. and Ettenberg, A. (1995). Haloperi-

dol di�erentially a�ects reinforcement and motivational processes in rats running

an alley for intravenous heroin. Psychopharmacology, 122(4):346�350.

[McNally et al., 2011] McNally, G. P., Johansen, J. P., and Blair, H. T. (2011). Plac-

ing prediction into the fear circuit. Trends in neurosciences, 34(6):283�292.

[Meili et al., 2005] Meili, R., Sasaki, A. T., and Firtel, R. A. (2005). Rho rocks pten.

Nature cell biology, 7(4):334.

[Meye et al., 2016] Meye, F. J., Soiza-Reilly, M., Smit, T., Diana, M. A., Schwarz,

M. K., and Mameli, M. (2016). Shifted pallidal co-release of gaba and glutamate in

habenula drives cocaine withdrawal and relapse. Nature neuroscience, 19(8):1019.

[Meye et al., 2015] Meye, F. J., Valentinova, K., Lecca, S., Marion-Poll, L.,

Maroteaux, M. J., Musardo, S., Moutkine, I., Gardoni, F., Huganir, R. L., Georges,

F., et al. (2015). Cocaine-evoked negative symptoms require ampa receptor traf-

�cking in the lateral habenula. Nature neuroscience, 18(3):376�378.

[Miller, 1944] Miller, N. E. (1944). Experimental studies of con�ict.

131

[Mirenowicz and Schultz, 1994] Mirenowicz, J. and Schultz, W. (1994). Importance

of unpredictability for reward responses in primate dopamine neurons. Journal of

neurophysiology, 72(2):1024�1027.

[Moeller et al., 2012] Moeller, F. G., Steinberg, J. L., Lane, S. D., Kjome, K. L.,

Ma, L., Ferre, S., Schmitz, J., Green, C. E., Bandak, S. I., Renshaw, P. F., et al.

(2012). Increased orbitofrontal brain activation after administration of a selective

adenosine a2a antagonist in cocaine dependent subjects. Frontiers in psychiatry,

3:44.

[Morales and Pickel, 2012] Morales, M. and Pickel, V. M. (2012). Insights to drug ad-

diction derived from ultrastructural views of the mesocorticolimbic system. Annals

of the New York Academy of Sciences, 1248(1):71�88.

[Morton et al., 2018] Morton, G., Nasirova, N., Sparks, D. W., Brodsky, M., Sivaku-

maran, S., Lambe, E. K., and Turner, E. E. (2018). Chrna5-expressing neurons

in the interpeduncular nucleus mediate aversion primed by prior stimulation or

nicotine exposure. Journal of Neuroscience, 38(31):6900�6920.

[Mulvaney et al., 1999] Mulvaney, F. D., Alterman, A. I., Boardman, C. R., and

Kampman, K. (1999). Cocaine abstinence symptomatology and treatment attri-

tion. Journal of substance abuse treatment, 16(2):129�135.

[Nagel et al., 2003] Nagel, J., Schladebach, H., Koch, M., Schwienbacher, I., Müller,

C. E., and Hauber, W. (2003). E�ects of an adenosine a2a receptor blockade in

the nucleus accumbens on locomotion, feeding, and prepulse inhibition in rats.

Synapse, 49(4):279�286.

[Nussinov and Tsai, 2013] Nussinov, R. and Tsai, C.-J. (2013). Allostery in disease

and in drug discovery. Cell, 153(2):293�305.

132

[Olmstead, 2006] Olmstead, M. C. (2006). Animal models of drug addiction: where

do we go from here? The Quarterly journal of experimental psychology, 59(4):625�

653.

[Olson and Nestler, 2007] Olson, V. G. and Nestler, E. J. (2007). Topographical

organization of gabaergic neurons within the ventral tegmental area of the rat.

Synapse, 61(2):87�95.

[Panlilio et al., 1996] Panlilio, L. V., Schindler, C., and Weiss, S. (1996). Cocaine

self-administration increased by compounding discriminative stimuli. Psychophar-

macology, 125(3):202�208.

[Patel et al., 2003] Patel, S., Rademacher, D. J., and Hillard, C. J. (2003). Di�er-

ential regulation of the endocannabinoids anandamide and 2-arachidonylglycerol

within the limbic forebrain by dopamine receptor activity. Journal of Pharmacol-

ogy and Experimental Therapeutics, 306(3):880�888.

[Patriarchi et al., 2018] Patriarchi, T., Cho, J. R., Merten, K., Howe, M. W., Marley,

A., Xiong, W.-H., Folk, R. W., Broussard, G. J., Liang, R., Jang, M. J., et al.

(2018). Ultrafast neuronal imaging of dopamine dynamics with designed genetically

encoded sensors. Science, 360(6396):eaat4422.

[Paxinos and Watson, 2006] Paxinos, G. and Watson, C. (2006). The rat brain in

stereotaxic coordinates: hard cover edition. Elsevier.

[Perrotti et al., 2005] Perrotti, L. I., Bolaños, C. A., Choi, K.-H., Russo, S. J., Ed-

wards, S., Ulery, P. G., Wallace, D. L., Self, D. W., Nestler, E. J., and Barrot, M.

(2005). δfosb accumulates in a gabaergic cell population in the posterior tail of

the ventral tegmental area after psychostimulant treatment. European Journal of

Neuroscience, 21(10):2817�2824.

133

[Perry and Carroll, 2008] Perry, J. L. and Carroll, M. E. (2008). The role of impulsive

behavior in drug abuse. Psychopharmacology, 200(1):1�26.

[Phillips et al., 2014] Phillips, K. A., Epstein, D. H., and Preston, K. L. (2014). Psy-

chostimulant addiction treatment. Neuropharmacology, 87:150�160.

[Prather and Lal, 1992] Prather, P. L. and Lal, H. (1992). Protracted withdrawal:

sensitization of the anxiogenic response to cocaine in rats concurrently treated with

ethanol. Neuropsychopharmacology.

[Proulx et al., 2014] Proulx, C. D., Hikosaka, O., and Malinow, R. (2014). Reward

processing by the lateral habenula in normal and depressive behaviors. Nature

neuroscience, 17(9):1146.

[Pun et al., 2012] Pun, R. Y., Rolle, I. J., LaSarge, C. L., Hosford, B. E., Rosen,

J. M., Uhl, J. D., Schmeltzer, S. N., Faulkner, C., Bronson, S. L., Murphy, B. L.,

et al. (2012). Excessive activation of mtor in postnatally generated granule cells is

su�cient to cause epilepsy. Neuron, 75(6):1022�1034.

[Ramirez et al., 2015] Ramirez, F., Moscarello, J. M., LeDoux, J. E., and Sears, R. M.

(2015). Active avoidance requires a serial basal amygdala to nucleus accumbens

shell circuit. Journal of Neuroscience, 35(8):3470�3477.

[Resnick et al., 1977a] Resnick, R. B., Kestenbaum, R. S., and Schwartz, L. K.

(1977a). Acute systemic e�ects of cocaine in man: a controlled study by intranasal

and intravenous routes. Science, 195(4279):696�698.

[Resnick et al., 1977b] Resnick, R. B., Kestenbaum, R. S., and Schwartz, L. K.

(1977b). Acute systemic e�ects of cocaine in man: a controlled study by intranasal

and intravenous routes. Science, 195(4279):696�698.

134

[Rezvani et al., 2010] Rezvani, A. H., Sexton, H., and Levin, E. D. (2010). Persistent

high alcohol consumption in alcohol-preferring (p) rats results from a lack of normal

aversion to alcohol. Alcohol & Alcoholism, 45(3):219�222.

[Richardson and Roberts, 1996] Richardson, N. R. and Roberts, D. C. (1996). Pro-

gressive ratio schedules in drug self-administration studies in rats: a method to

evaluate reinforcing e�cacy. Journal of neuroscience methods, 66(1):1�11.

[Riley and Simpson, 2001] Riley, A. and Simpson, G. (2001). Contemporary learning

theory.

[Riley, 2011] Riley, A. L. (2011). The paradox of drug taking: the role of the aversive

e�ects of drugs. Physiology & behavior, 103(1):69�78.

[Rogerio and Takahashi, 1992] Rogerio, R. and Takahashi, R. N. (1992). Anxiogenic

properties of cocaine in the rat evaluated with the elevated plus-maze. Pharmacol-

ogy Biochemistry and Behavior, 43(2):631�633.

[Rohsenow et al., 2007] Rohsenow, D. J., Howland, J., Minsky, S. J., Greece, J.,

Almeida, A., and Roehrs, T. A. (2007). The acute hangover scale: A new measure

of immediate hangover symptoms. Addictive behaviors, 32(6):1314�1320.

[Roitman et al., 2008] Roitman, M. F., Wheeler, R. A., Wightman, R. M., and

Carelli, R. M. (2008). Real-time chemical responses in the nucleus accumbens dif-

ferentiate rewarding and aversive stimuli. Nature neuroscience, 11(12):1376�1377.

[Rothwell et al., 2009] Rothwell, P. E., Thomas, M. J., and Gewirtz, J. C. (2009).

Distinct pro�les of anxiety and dysphoria during spontaneous withdrawal from

acute morphine exposure. Neuropsychopharmacology, 34(10):2285�2295.

[Russell, 1976] Russell, M. (1976). Drugs and drug dependence.

135

[Saunders et al., 2013] Saunders, B. T., Yager, L. M., and Robinson, T. E. (2013).

Preclinical studies shed light on individual variation in addiction vulnerability.

Neuropsychopharmacology, 38(1):249.

[Scammell et al., 2000] Scammell, T. E., Estabrooke, I. V., McCarthy, M. T.,

Chemelli, R. M., Yanagisawa, M., Miller, M. S., and Saper, C. B. (2000). Hypotha-

lamic arousal regions are activated during moda�nil-induced wakefulness. Journal

of Neuroscience, 20(22):8620�8628.

[Schultz, 2007] Schultz, W. (2007). Multiple dopamine functions at di�erent time

courses. Annu. Rev. Neurosci., 30:259�288.

[Schultz et al., 1997] Schultz, W., Dayan, P., and Montague, P. R. (1997). A neural

substrate of prediction and reward. Science, 275(5306):1593�1599.

[Sego et al., 2014] Sego, C., Gonçalves, L., Lima, L., Furigo, I. C., Donato Jr, J.,

and Metzger, M. (2014). Lateral habenula and the rostromedial tegmental nucleus

innervate neurochemically distinct subdivisions of the dorsal raphe nucleus in the

rat. Journal of Comparative Neurology, 522(7):1454�1484.

[Sheehan et al., 2004] Sheehan, T. P., Chambers, R. A., and Russell, D. S. (2004).

Regulation of a�ect by the lateral septum: implications for neuropsychiatry. Brain

Research Reviews, 46(1):71�117.

[Shen et al., 2008] Shen, W., Flajolet, M., Greengard, P., and Surmeier, D. J.

(2008). Dichotomous dopaminergic control of striatal synaptic plasticity. Science,

321(5890):848�851.

136

[Shippenberg et al., 1991] Shippenberg, T. S., Bals-Kubik, R., Huber, A., and Herz,

A. (1991). Neuroanatomical substrates mediating the aversive e�ects of d-1

dopamine receptor antagonists. Psychopharmacology, 103(2):209�214.

[Smith et al., 1991] Smith, M., Garner, D., and Niemann, J. T. (1991). Pharma-

cologic interventions after an ld50 cocaine insult in a chronically instrumented

rat model: are beta-blockers contraindicated? Annals of emergency medicine,

20(7):768�771.

[Smith et al., 2019a] Smith, R. J., Vento, P. J., Chao, Y. S., Good, C. H., and Jhou,

T. C. (2019a). Gene expression and neurochemical characterization of the rostro-

medial tegmental nucleus (rmtg) in rats and mice. Brain Structure and Function,

224(1):219�238.

[Smith et al., 2019b] Smith, R. J., Vento, P. J., Chao, Y. S., Good, C. H., and Jhou,

T. C. (2019b). Gene expression and neurochemical characterization of the rostro-

medial tegmental nucleus (rmtg) in rats and mice. Brain Structure and Function,

224(1):219�238.

[Solomon and Corbit, 1974] Solomon, R. L. and Corbit, J. D. (1974). An opponent-

process theory of motivation: I. temporal dynamics of a�ect. Psychological review,

81(2):119.

[Spealman, 1979] Spealman, R. D. (1979). Behavior maintained by termination of a

schedule of self-administered cocaine. Science, 204(4398):1231�1233.

[Spotts and Shontz, 1984] Spotts, J. V. and Shontz, F. C. (1984). The phe-

nomenological structure of drug-induced ego states. ii. barbiturates and sedative-

hypnotics: Phenomenology and implications. International journal of the addic-

tions, 19(3):295�326.

137

[Stamatakis and Stuber, 2012] Stamatakis, A. M. and Stuber, G. D. (2012). Ac-

tivation of lateral habenula inputs to the ventral midbrain promotes behavioral

avoidance. Nature neuroscience, 15(8):1105.

[Stead et al., 2006] Stead, J. D., Clinton, S., Neal, C., Schneider, J., Jama, A., Miller,

S., Vazquez, D. M., Watson, S. J., and Akil, H. (2006). Selective breeding for

divergence in novelty-seeking traits: heritability and enrichment in spontaneous

anxiety-related behaviors. Behavior genetics, 36(5):697.

[Stolerman, 1992] Stolerman, I. (1992). Drugs of abuse: behavioural principles, meth-

ods and terms. Trends in Pharmacological Sciences, 13:170�176.

[Stolerman and d'Mello, 1981] Stolerman, I. and d'Mello, G. (1981). Role of training

conditions in discrimination of central nervous system stimulants by rats. Psy-

chopharmacology, 73(3):295�303.

[Stone et al., 2012] Stone, A. L., Becker, L. G., Huber, A. M., and Catalano, R. F.

(2012). Review of risk and protective factors of substance use and problem use in

emerging adulthood. Addictive behaviors, 37(7):747�775.

[Stopper et al., 2014] Stopper, C. M., Maric, T., Montes, D. R., Wiedman, C. R.,

and Floresco, S. B. (2014). Overriding phasic dopamine signals redirects action

selection during risk/reward decision making. Neuron, 84(1):177�189.

[Sulzer, 2011] Sulzer, D. (2011). How addictive drugs disrupt presynaptic dopamine

neurotransmission. Neuron, 69(4):628�649.

[Surmeier et al., 2009] Surmeier, D. J., Plotkin, J., and Shen, W. (2009). Dopamine

and synaptic plasticity in dorsal striatal circuits controlling action selection. Cur-

rent opinion in neurobiology, 19(6):621�628.

138

[Svenningsson et al., 1998] Svenningsson, P., Le Moine, C., Aubert, I., Burbaud, P.,

Fredholm, B., and Bloch, B. (1998). Cellular distribution of adenosine a2a receptor

mrna in the primate striatum. Journal of Comparative Neurology, 399(2):229�240.

[Tan et al., 2012] Tan, K. R., Yvon, C., Turiault, M., Mirzabekov, J. J., Doehner, J.,

Labouèbe, G., Deisseroth, K., Tye, K. M., and Lüscher, C. (2012). Gaba neurons

of the vta drive conditioned place aversion. Neuron, 73(6):1173�1183.

[Taylor et al., 2019] Taylor, N. E., Long, H., Pei, J., Kukutla, P., Phero, A., Hadaegh,

F., Abdelnabi, A., Solt, K., and Brenner, G. J. (2019). The rostromedial tegmental

nucleus: a key modulator of pain and opioid analgesia. Pain, 160(11):2524�2534.

[Tobler et al., 2003] Tobler, P. N., Dickinson, A., and Schultz, W. (2003). Coding

of predicted reward omission by dopamine neurons in a conditioned inhibition

paradigm. Journal of Neuroscience, 23(32):10402�10410.

[Tomko et al., 2018] Tomko, R. L., Jones, J. L., Gilmore, A. K., Brady, K. T., Back,

S. E., and Gray, K. M. (2018). N-acetylcysteine: A potential treatment for sub-

stance use disorders. Current psychiatry, 17(6):30.

[Tri�lie� et al., 2013] Tri�lie�, P., Feng, B., Urizar, E., Winiger, V., Ward, R. D.,

Taylor, K. M., Martinez, D., Moore, H., Balsam, P. D., Simpson, E. H., et al.

(2013). Increasing dopamine d2 receptor expression in the adult nucleus accumbens

enhances motivation. Molecular psychiatry, 18(9):1025�1033.

[Tsai et al., 2009] Tsai, H.-C., Zhang, F., Adamantidis, A., Stuber, G. D., Bonci, A.,

De Lecea, L., and Deisseroth, K. (2009). Phasic �ring in dopaminergic neurons is

su�cient for behavioral conditioning. Science, 324(5930):1080�1084.

139

[Uhl and Grow, 2004] Uhl, G. R. and Grow, R. W. (2004). The burden of complex

genetics in brain disorders. Archives of general psychiatry, 61(3):223�229.

[Ungless et al., 2004] Ungless, M. A., Magill, P. J., and Bolam, J. P. (2004). Uniform

inhibition of dopamine neurons in the ventral tegmental area by aversive stimuli.

Science, 303(5666):2040�2042.

[van Huijstee and Mansvelder, 2015] van Huijstee, A. N. and Mansvelder, H. D.

(2015). Glutamatergic synaptic plasticity in the mesocorticolimbic system in ad-

diction. Frontiers in cellular neuroscience, 8:466.

[van Zessen et al., 2012] van Zessen, R., Phillips, J. L., Budygin, E. A., and Stuber,

G. D. (2012). Activation of vta gaba neurons disrupts reward consumption. Neuron,

73(6):1184�1194.

[Vento et al., 2017] Vento, P. J., Burnham, N. W., Rowley, C. S., and Jhou, T. C.

(2017). Learning from one's mistakes: A dual role for the rostromedial tegmental

nucleus in the encoding and expression of punished reward seeking. Biological

psychiatry, 81(12):1041�1049.

[Verendeev and Riley, 2012] Verendeev, A. and Riley, A. L. (2012). Conditioned taste

aversion and drugs of abuse: History and interpretation. Neuroscience & Biobe-

havioral Reviews, 36(10):2193�2205.

[Vogt, 2005] Vogt, B. A. (2005). Pain and emotion interactions in subregions of the

cingulate gyrus. Nature Reviews Neuroscience, 6(7):533�544.

[Waelti et al., 2001] Waelti, P., Dickinson, A., and Schultz, W. (2001). Dopamine

responses comply with basic assumptions of formal learning theory. Nature,

412(6842):43�48.

140

[Wagner-Echeagaray et al., 1994] Wagner-Echeagaray, F. A., Schütz, C. G., Chilcoat,

H. D., and Anthony, J. C. (1994). Degree of acculturation and the risk of crack

cocaine smoking among hispanic americans. American Journal of Public Health,

84(11):1825�1827.

[Wakonigg et al., 2003] Wakonigg, G., Sturm, K., Saria, A., and Zernig, G. (2003).

Opioids, cocaine, and food change runtime distribution in a rat runway procedure.

Psychopharmacology, 169(1):52�59.

[Wal�sh et al., 1990] Wal�sh, S., Massey, R., and Krone, A. (1990). Anxiety and

anger among abusers of di�erent substances. Drug and Alcohol Dependence,

25(3):253�256.

[Wall et al., 2000] Wall, T. L., Horn, S. M., Johnson, M. L., Smith, T. L., and Carr,

L. G. (2000). Hangover symptoms in Asian Americans with variations in the alde-

hyde dehydrogenase (ALDH2) gene. J. Stud. Alcohol, 61(1):13�17.

[Wasserman et al., 2013] Wasserman, D. I., Wang, H. G., Rashid, A. J., Josselyn,

S. A., and Yeomans, J. S. (2013). Cholinergic control of morphine-induced lo-

comotion in rostromedial tegmental nucleus versus ventral tegmental area sites.

European Journal of Neuroscience, 38(5):2774�2785.

[Weber et al., 2010] Weber, M., Chang, W.-L., Breier, M. R., Yang, A., Millan, M. J.,

and Swerdlow, N. R. (2010). The e�ects of the dopamine d2 agonist sumanirole on

prepulse inhibition in rats. European Neuropsychopharmacology, 20(6):421�425.

[Weiss et al., 1992] Weiss, F., Markou, A., Lorang, M. T., and Koob, G. F. (1992).

Basal extracellular dopamine levels in the nucleus accumbens are decreased dur-

ing cocaine withdrawal after unlimited-access self-administration. Brain research,

593(2):314�318.

141

[Wenzel et al., 2011] Wenzel, J. M., Waldroup, S. A., Haber, Z. M., Su, Z.-I., Ben-

Shahar, O., and Ettenberg, A. (2011). E�ects of lidocaine-induced inactivation of

the bed nucleus of the stria terminalis, the central or the basolateral nucleus of

the amygdala on the opponent-process actions of self-administered cocaine in rats.

Psychopharmacology, 217(2):221�230.

[Williams et al., 2015] Williams, M. R., DeSpenza, T., Li, M., Gulledge, A. T., and

Luikart, B. W. (2015). Hyperactivity of newborn pten knock-out neurons results

from increased excitatory synaptic drive. Journal of Neuroscience, 35(3):943�959.

[Williamson et al., 1996a] Williamson, S., Gossop, M., Powis, B., Gri�ths, P., Foun-

tain, J., and Strang, J. (1996a). Adverse e�ects of stimulant drugs in a community

sample of drug users. Drug and Alcohol Dependence-Shannon, 44(2):87�94.

[Williamson et al., 1996b] Williamson, S., Gossop, M., Powis, B., Gri�ths, P., Foun-

tain, J., and Strang, J. (1996b). Adverse e�ects of stimulant drugs in a community

sample of drug users. Drug and Alcohol Dependence-Shannon, 44(2):87�94.

[Wise, 2004] Wise, R. A. (2004). Dopamine, learning and motivation. Nature reviews

neuroscience, 5(6):483�494.

[Wise et al., 1976] Wise, R. A., Yokel, R. A., and DeWit, H. (1976). Both positive

reinforcement and conditioned aversion from amphetamine and from apomorphine

in rats. Science, 191(4233):1273�1275.

[Wong and Schumann, 2008] Wong, C. C. and Schumann, G. (2008). Genetics of ad-

dictions: strategies for addressing heterogeneity and polygenicity of substance use

disorders. Philosophical Transactions of the Royal Society B: Biological Sciences,

363(1507):3213�3222.

142

[Wydra et al., 2018] Wydra, K., Suder, A., Frankowska, M., Escuela, D. B., Fuxe,

K., and Filip, M. (2018). E�ects of intra-accumbal or intra-prefrontal cortex mi-

croinjections of adenosine 2a receptor ligands on responses to cocaine reward and

seeking in rats. Psychopharmacology, 235(12):3509�3523.

[Yang et al., 1992] Yang, X.-M., Gorman, A. L., Dunn, A. J., and Goeders, N. E.

(1992). Anxiogenic e�ects of acute and chronic cocaine administration: neurochem-

ical and behavioral studies. Pharmacology Biochemistry and Behavior, 41(3):643�

650.

[Yetniko� et al., 2015] Yetniko�, L., Cheng, A. Y., Lavezzi, H. N., Parsley, K. P.,

and Zahm, D. S. (2015). Sources of input to the rostromedial tegmental nucleus,

ventral tegmental area, and lateral habenula compared: a study in rat. Journal of

Comparative Neurology, 523(16):2426�2456.

[Yin and Lovinger, 2006] Yin, H. H. and Lovinger, D. M. (2006). Frequency-

speci�c and d2 receptor-mediated inhibition of glutamate release by retrograde

endocannabinoid signaling. Proceedings of the National Academy of Sciences,

103(21):8251�8256.

[Zahm et al., 2010] Zahm, D. S., Becker, M. L., Freiman, A. J., Strauch, S., De-

Garmo, B., Geisler, S., Meredith, G. E., and Marinelli, M. (2010). Fos after sin-

gle and repeated self-administration of cocaine and saline in the rat: emphasis

on the basal forebrain and recalibration of expression. Neuropsychopharmacology,

35(2):445�463.

[Zangen et al., 2002] Zangen, A., Ikemoto, S., Zadina, J. E., and Wise, R. A. (2002).

Rewarding and psychomotor stimulant e�ects of endomorphin-1: anteroposterior

143

di�erences within the ventral tegmental area and lack of e�ect in nucleus accum-

bens. Journal of Neuroscience, 22(16):7225�7233.

[Zernig et al., 2002] Zernig, G., Harbig, P., Weiskirchner, I., Au�nger, M., Wakonigg,

G., and Saria, A. (2002). Reinforcing e�ect of subcutaneous morphine in a modi�ed

ettenberg runway. Journal of Molecular Neuroscience, 18(1-2):135�142.

[Zuckerman et al., 1993] Zuckerman, M., Kuhlman, D. M., Joireman, J., Teta, P.,

and Kraft, M. (1993). A comparison of three structural models for personality: the

big three, the big �ve, and the alternative �ve. Journal of personality and social

psychology, 65(4):757.

144