Discriminative and Reinforcing Effects of Cocaine-Levamisole Combinations

Western Michigan University ScholarWorks at WMU

Dissertations Graduate College

6-2017

Zachary J. Zimmermann Western Michigan University, [email protected]

Follow this and additional works at: https://scholarworks.wmich.edu/dissertations

Part of the Substance Abuse and Addiction Commons

Recommended Citation Zimmermann, Zachary J., "Discriminative and Reinforcing Effects of Cocaine-Levamisole Combinations" (2017). Dissertations. 3120. https://scholarworks.wmich.edu/dissertations/3120

This Dissertation-Open Access is brought to you for free and open access by the Graduate College at ScholarWorks at WMU. It has been accepted for inclusion in Dissertations by an authorized administrator of ScholarWorks at WMU. For more information, please contact [email protected]. DISCRIMINATIVE AND REINFORCING EFFECTS OF COCAINE-LEVAMISOLE COMBINATIONS

Zachary J. Zimmermann

A dissertation submitted to the Graduate College in partial fulfillment of the requirements for the degree of Doctor of Philosophy Psychology Western Michigan University June 2017

Dissertation Committee:

Alan Poling, Ph.D., Chair Lisa Baker, Ph.D. David V. Gauvin, Ph.D. Cynthia Pietras, Ph.D.

DISCRIMINATIVE AND REINFORCING EFFECTS OF COCAINE-LEVAMISOLE COMBINATIONS

Zachary J. Zimmermann, Ph.D. Western Michigan University, 2017

The behavioral and neurochemical effects of cocaine are well established, and it is one of the most widely abused illicit drugs. Illicit cocaine is often adulterated with levamisole, which is an anthelmintic that was withdrawn from the U. S. market in 2000. It has been hypothesized that levamisole, unlike other common adulterants which are added as simple bulking agents, has effects of its own which may be responsible for its use as an adulterant. Although these effects are speculative, the addition of levamisole to cocaine has become an increasing public health concern, as serious adverse effects (e.g., vasculitis, neutropenia) of levamisole-adulterated cocaine have been observed in drug users.

The present experiments were intended to provide further information about effects of levamisole that may help to explain its use as an adulterant. The aim of the first experiment was to determine if adding levamisole alters cocaine self-administration in rats. To this end, the response patterns of rats trained to self-administer cocaine were examined to determine whether they were altered by pretreatment with two doses of levamisole (1 and 10 mg/kg). Although response patterns were generally unaffected, animals consistently consumed less cocaine at all doses tested under both pretreatment conditions. This decrease in responding suggests that levamisole is not added to directly increase cocaine intake.

The aim of the second experiment was to examine how adding levamisole affected the discriminative stimulus properties of cocaine. Rats were trained to discriminate between cocaine

and saline injections to earn food in a classic drug discrimination procedure. Various doses of cocaine, levamisole, and cocaine-levamisole combinations were then tested to yield cross- generalization profiles. Levamisole alone failed to produce cocaine-appropriate responding in the majority of animals, but did so in a dose-dependent manner for a small minority. Cocaine- levamisole combinations consistently produced dose-dependent cocaine-appropriate responding, and did so to a greater extent than did the constituent cocaine or levamisole dose alone (i.e., supra-additivity).

The results of these studies are consistent with the popular notion that levamisole is added to cocaine to alter its effects, rather than simply as a bulking agent. Furthermore, these results indicate that a supra-additive interaction between cocaine and levamisole exists with regard to their discriminative stimulus effects. Further research into the behavioral and neurochemical mechanisms of action of cocaine-levamisole combinations is warranted.

ACKNOWLEDGMENTS

I am forever indebted to my wife, Abby, and my parents, Ken and Sherry, for their support throughout all of this. I would not have been able to embark upon this journey without their complete trust, guidance, and unbelievable patience.

I want to thank Dr. Al Poling for helping me to develop into the skeptical thinker and behavioral scientist I am today. The freedom and opportunities afforded to me as your student have honestly been life changing, and for that I cannot be more appreciative. I also need to thank

Dr. Lisa Baker for cultivating my interest in behavioral pharmacology years ago, and Dr. Cindy

Pietras for exposing me to behavior analysis and presenting me with some of the most important opportunities in the beginning of my graduate career. Finally, I must thank Dr. David Gauvin, for taking me under his wing and providing me with invaluable teaching experiences. Without all of your unwavering support and willingness to challenge me, I could not have successfully completed this project.

Lastly, I must acknowledge all of the previous and current lab members of both Poling and Baker labs with whom I have grown close. You all have provided me with valuable feedback, frequent learning opportunities, and a collaborative research environment for which I am grateful to have been a part of.

Zachary J. Zimmermann

ii TABLE OF CONTENTS

ACKNOWLEDGMENTS ...... ii

LIST OF TABLES ...... vi

LIST OF FIGURES ...... vii

CHAPTER

I. INTRODUCTION ...... 1

General Background ...... 1

Cocaine ...... 1

Production and distribution ...... 1

Adulteration and drug manufacturing ...... 3

Pharmacology ...... 4

Levamisole ...... 4

Pharmacology ...... 6

Complications related to levamisole-adulterated cocaine use ...... 9

Potentiation of drug effects ...... 10

Analysis of drug interactions ...... 11

Behavioral Models ...... 14

Self-administration ...... 14

Interaction effects ...... 16

Drug discrimination ...... 18

Interaction effects ...... 20

Behavioral Literature ...... 22

iii Table of Contents — Continued

CHAPTER II. EXPERIMENT 1: EFFECTS OF LEVAMISOLE ON COCAINE SELF- ADMINISTRATION IN SPRAGUE-DAWLEY RATS...... 25 Overview ...... 25

Methods...... 25

Subjects ...... 25

Apparatus ...... 27

Procedure ...... 28

Catheter maintenance ...... 28

Response acquisition ...... 29

Cocaine training ...... 29

Cocaine testing ...... 30

Levamisole testing ...... 31

Drugs ...... 31

Data analysis ...... 31

Results ...... 32

Discussion ...... 53

Conclusion ...... 59

III. EXPERIMENT 2: ASSESSMENT OF THE SUBJECTIVE EFFECTS OF LEVAMISOLE ALONE AND IN COMBINATION WITH COCAINE IN SPRAGUE- DAWLEY RATS ...... 61 Overview ...... 61

Methods...... 63

iv Table of Contents — Continued

CHAPTER

Subjects ...... 63

Apparatus ...... 64

Procedure ...... 65

Response acquisition ...... 65

Drug training ...... 65

Testing...... 67

Drugs ...... 68

Data analysis ...... 69

Results ...... 69

Drug combination tests ...... 74

Discussion ...... 81

IV. GENERAL DISCUSSION ...... 89

REFERENCES ...... 91

APPENDICES ...... 114

A. IACUC Approval Letters ...... 114

B. Individual and Stimulus-element Combination Generalization Test Results ...... 118

v LIST OF TABLES

1. Descriptive statistics of daily and 3-day averages of drug intake expressed as number of infusions during unlimited access test sessions ...... 34

2. Individual estimated ED50 values for all three treatment conditions...... 70

vi LIST OF FIGURES

1. Regional changes in dopamine and serotonin concentrations following levamisole administration ...... 8

2. Graphical depiction of isobolographic analysis ...... 13

3. Theoretical drug-interaction dose-effect curves ...... 17

4. Cocaine self-administration dose-effect curves ...... 35

5. 3-day mean number of injections of the maintenance (i.e., 0.56 mg/kg/injection) cocaine dose across all three conditions ...... 36

6. 3-day grand mean comparisons of total number of injections and total cocaine intake (mg/kg) for the maintenance dose of cocaine (0.56 mg/kg/injection) for cocaine alone, 1 mg/kg and 10 mg/kg LVM pretreatment ...... 37

7A. Representative samples of cumulative records generated for unlimited-access test sessions in the cocaine alone condition (saline) ...... 38

7B. Representative samples of cumulative records generated for unlimited-access test sessions in the cocaine alone condition (cocaine 0.01 mg/kg/injection) ...... 39

7C. Representative samples of cumulative records generated for unlimited-access test sessions in the cocaine alone condition (cocaine 0.10 mg/kg/injection ...... 40

7D. Representative samples of cumulative records generated for unlimited-access test sessions in the cocaine alone condition (0.56 mg/kg/injection) ...... 41

7E. Representative samples of cumulative records generated for unlimited-access test sessions in the cocaine alone condition (cocaine 1.0 mg/kg/injection) ...... 42

vii

List of Figures – Continued

8A. Representative samples of cumulative records generated for unlimited-access test sessions in the 1 mg/kg levamisole (LVM) condition (saline) ...... 43

8B. Representative samples of cumulative records generated for unlimited-access test sessions in the 1 mg/kg levamisole (LVM) condition (cocaine 0.01 mg/kg/injection) ...... 44

8C. Representative samples of cumulative records generated for unlimited-access test sessions in the 1 mg/kg levamisole (LVM) condition (cocaine 0.10 mg/kg/injection) ...... 45

8D. Representative samples of cumulative records generated for unlimited-access test sessions in the 1 mg/kg levamisole (LVM) condition (cocaine 0.56 mg/kg/injection) ...... 46

9A. Representative samples of cumulative records generated for unlimited-access test sessions in the 10 mg/kg levamisole (LVM) condition (saline) ...... 48

9B. Representative samples of cumulative records generated for unlimited-access test sessions in the 10 mg/kg levamisole (LVM) condition (cocaine 0.01 mg/kg/injection) ...... 49

9C. Representative samples of cumulative records generated for unlimited-access test sessions in the 10 mg/kg levamisole (LVM) condition (cocaine 0.10 mg/kg/injection) ...... 50

9D. Representative samples of cumulative records generated for unlimited-access test sessions in the 10 mg/kg levamisole (LVM) condition (cocaine 0.56 mg/kg/injection) ...... 51

10. Inter-injection intervals Quarter-life analysis of responding during test sessions ...... 52

11. Quarter-life analysis of responding during test sessions ...... 53

12. Comparison of the number of training days required to meet testing criteria...... 70

13. Cocaine alone dose-effect curves for all three dependent measures for both male and female rats ...... 71

viii List of Figures – Continued

14. Representative cumulative records for within-session responding for cocaine alone...... 72

15. Levamisole alone dose-effect curves for all three dependent measures for both male and female rats ...... 73

16. Representative cumulative records for within-session responding for cocaine-levamisole combinations ...... 75

17. Dose-combination dose-effect curves for animal 103M ...... 76

18. Dose-combination dose-effect curves for animal 104M ...... 76

19. Dose-combination dose-effect curves for animal 106M ...... 77

20. Dose-combination dose-effect curves for animal 108M ...... 77

21. Cumulative record for 5.6 mg/kg cocaine alone test session for animal 108M ...... 78

22. Dose-combination dose-effect curves for animal 109F ...... 79

23. Dose-combination dose-effect curves for animal 111F ...... 79

24. Dose-combination dose-effect curves for animal 114F ...... 80

25. Dose-combination dose-effect curves for animal 115F ...... 81

ix CHAPTER I

INTRODUCTION

General Background

Although most drugs of abuse have been studied extensively within the context of the available abuse liability paradigms, the compounds employed therein are often of laboratory grade, and thus do not necessary reflect the true effects of “street drugs”, which contain a number of other constituents unbeknownst to the consumer. Furthermore, little is known regarding the combined effects of many common adulterants or bulking agents found in street drugs, many of which can have profound influence upon the effects of these drugs (Cole et al., 2010).

Levamisole is an increasingly prevalent adulterant reportedly found in illicitly manufactured cocaine is levamisole (Abdul-Karim, Ryan, Rangel, & Emmett, 2013). Levamisole has a lengthy history as an anthelmintic for treating parasitic worm infections. However, levamisole was withdrawn from the U.S. market in 2000 due to a combination of factors relating to adverse clinical reactions. It has been hypothesized that the rise in levamisole-adulterated cocaine can be attributed to possible stimulant effects of levamisole itself (Tallarida et al., 2014; Tallarida,

Tallarida, & Rawls, 2015). Though some have cited these stimulant effects as being responsible for the rise in the incidence of levamisole-adulterated cocaine, little behavioral research has been conducted directly examining the subjective effects of levamisole-cocaine mixtures and the subsequent effects upon drug consumption. The aim of the studies described herein was to assess both of these aspects relevant to human drug taking behavior using established animal models— self-administration and drug discrimination.

Cocaine

Production and distribution. Illicit drug use and abuse costs the United States more

1 than $193 billion annually (National Drug Intelligence Center, 2011). Indeed, the American war on drugs has been repeatedly declared an utter failure. Though they may disagree on the tactics, it is the combined goal of law enforcement and policy makers to enact changes that not only prevent the initiation of drug use by Americans, but also decrease the number of current users.

Drug abuse is a global problem, and dealing with the problem as if the U.S. drug market exists within a vacuum would be naïve. In light of this, the U.S. has joined numerous international organizations and alliances to help stop the global trade of illicit drugs, combat the threat of new drugs, and curtail the crime associated with illicit drug trade and use.

The World Health Organization (WHO) has been tasked with providing expert guidance to the United Nations on international issues of schedule control and drug abuse prevention

(United Nations, 1988; WHO, 2016). In recent years (2007 to 2012), cocaine availability within the U.S. has declined. It is believed that these efforts have contributed to the dramatic recent decreases seen in cocaine use and distribution data. Although it is impossible to infer causation based on the available sampling metrics alone, both American cocaine use and global coca bush cultivation have been on the steady decline (United Nations Office on Drugs and Crime

(UNODC), 2015). Cocaine is largely an export from South American countries (e.g., Argentina,

Bolivia, Columbia, Ecuador, and Peru), with the majority of the cocaine originating in Columbia and entering the U.S. through Mexico, an international border notoriously fraught with corruption, smuggling, and other illegal activity.

The coca plant (Erythroxylon coca), the source of a drug that has plagued the international community, has enjoyed a lengthy history in both medicinal and religious activities.

The infamous stimulant properties of the plant have traditionally been produced by a drinkable preparation (mate de coca, Vin Mariani) or through oral consumption of the leaves (chewing)

2 (Stolberg, 2011). Coca leaves contain the cocaine alkaloid that can be chemically extracted to form coca paste. Through many chemical additions and purifications, coca paste is refined into powder form (cocaine hydrochloride), the form popularized in today’s drug and media culture.

After being converted into powder form, it is then packaged and is ready for bulk distribution.

Once it has been distributed across the greater U.S. to mid-level drug dealers, it is then generally mixed with a variety of compounds (i.e., “cut” or “stepped on”), sold to other dealers, and eventually directly to users. Though cocaine has limited medical use (CII), the primary domestic consumption of cocaine is of the illicit variety, with 270 tons of illicit cocaine seized in 2014

(UNODC, 2015) compared to the legally-sanctioned production of 200 kg per year (DEA, 2015).

Adulteration and drug manufacturing. Though “cutting” of street drugs with toxic substances is often the basis of urban legend, the process of adulterating street drugs is a very real and serious issue that can have a variety of unintended effects on both the drug user and the drug market. Through dilution of the initial bulk product, street dealers are able to increase their profit margin while maintaining an initial set cost of investment. Additionally, the notion that other dealers conduct themselves in the same manner leads to a perpetuation of adulterating drugs, generating what has been referred to as the “self-fulfilling prophecy” of adulteration

(DeCorte, 2001). Adulterants are also often referred to as cutting/bulking agents or diluents, though their addition may serve purposes beyond simply increasing the amount of sellable product. Certain additions to bulk product can either serve to potentiate the effects of the constituent drug itself, or act to mimic the effects of the relevant drug. One recent example of street-level adulteration is 3,4-methylenedioxy-methamphetamine (MDMA or “ecstasy”). As has been extensively documented in both the research literature (Parrott, 2004; Sherlock, Wolff, Hay,

& Conner, 1999) and the popular drug culture (DanceSafe, 2016), ecstasy often contains

3 stimulant drugs other than MDMA. Though designer drugs like ecstasy have been notorious targets for media coverage on drug adulteration, all street drugs, including cocaine and heroin, are subject to the same diluent processes as their popular designer counterparts.

Pharmacology. Vast amounts of research dollars and resources have been dedicated to examining the neuropharmacological and behavioral effects of cocaine. It is well established that the reinforcing effects of cocaine are primarily mediated through the mesolimbic dopamine system (Callahan, de La Garza, & Cunningham, 1997). In summary, cocaine is a potent small molecule psychostimulant that readily crosses the blood-brain barrier. Once in the brain, cocaine binds to the dopamine transporter (DAT), which is responsible for reclaiming dopamine and decreasing concentrations within the synapse, thus blocking reuptake of dopamine into the nerve terminal (Cooper, Bloom, & Roth, 2003). Once cocaine has bound to DAT, dopamine rapidly begins accumulating in the synaptic cleft, increasing extracellular levels of dopamine (Carrera,

Meijer, & Janda, 2004). In this way, cocaine is an indirect-acting dopamine agonist, action that is distinct from other, direct-acting dopamine agonists that are also common drugs of abuse (e.g., amphetamine). Although cocaine does affect other neurotransmitter systems, its reinforcing and discriminative effects are mediated primarily through dopaminergic activity (Baker, Riddle,

Saunders, & Appel, 1993; Cunningham & Callahan, 1991). Cocaine is a quick-acting stimulant, with a half-life of approximately 40 min in humans (Fischman, 1984), 18 min in rats (Barbieri,

Ferko, DiGregorio, & Ruch, 1992) and 20-24 min in rhesus monkeys (Saady, Bowman, & Aceto,

1995).

Levamisole

Levamisole was discovered in 1966 by Belgian scientists working at Janssen

Pharmaceutica. In an attempt to develop the next revolutionary anthelmintic drug, Janssen

4 created and pursued over 2700 chemicals before finally uncovering the racemate, tetramisole.

Tetramisole consisted of two enantiomers, dexamisole and levamisole. Though tetramisole was found to be effective in eliminating gastrointenstial nematodes in chickens, further examination eventually led to separation and development of the safer, more potent isomer levamisole

(Amery & Bruynseels, 1992; Janssen, 1976).

The primary intended use of levamisole was as an anthelmintic and immunomodulator.

During the late 1970s into the 1980s, levamisole enjoyed a skyrocketing popularity in its use and prescription as a miracle drug for ailments ranging from herpes, leprosy, rheumatoid arthritis, lupus, and colorectal cancer (Amery & Bruynseels, 1992; Luyckx et al., 1982). It appeared as though levamisole was a safe and effective treatment for a number of maladies, and it became a welcome addition to clinician’s tool belts worldwide. However, as prescription use and treatment became more prevalent, a number of adverse side effects were reported, ranging in severity from skin discolorations and minor infection to agranulocytosis and leukoencephalopathy. In light of these clinical findings, levamisole was withdrawn from the U.S. market in 2000 due to a combination of factors involving: 1) an increased availability of newer, safer replacement drugs, and 2) an increased prevalence of reports indicating a risk of serious adverse side effects in humans (e.g., agranulocytosis, vasculitis, and leukoencephalopathy). Currently, levamisole is used abroad as an anthelminitic in livestock (e.g., sheep, cattle, pigs). Due to the physical and chemical properties of the drug (i.e., relatively odorless, white powder, high melting point, limited stimulant-like discriminative stimulus properties, ability to pass common street purity tests), in addition to its widespread availability in veterinary medicine, it is increasingly being employed as an adulterant in illicitly manufactured cocaine. Indeed, as recently as 2009, the

5 DEA confirmed that over 70% of seized samples tested positive for the presence of levamisole

(Abdul-Karim et al., 2013).

Pharmacology. Although levamisole has an extensive clinical history, its pharmacological profile regarding the immunomodulatory effects is still incomplete. However, it has been hypothesized that these effects stem from a combination of the following: 1) restoration of normalized function to T-lymphocytes and phagocytes, 2) action as a prodrug for 2-oxo-3-(2- mercaptoethyl)-5-phenylimidazole (OMPI), 3) blocking the action of natural immunosuppressive factors (Avery & Bruynseels, 1992; Hodinka, Géher, Merétey, Gyódi, Petrányi, & Bozsóky,

1981), and 4) agonist activity at nicotinic receptors in nematodes (DEA, 2013; Janssen, 1976).

Although the peak plasma concentrations of levamisole in humans occur approximately 1-2 hr post-dose (Luyckx et al., 1982), the effects of levamisole in other species have been demonstrated in as little as 30 min post-dose (Tallarida et al., 2014; Tallarida, Tallarida, &

Rawls, 2015).

Levamisole is a nicotinic receptor agonist that readily crosses the blood-brain barrier and affects dopaminergic activity, resulting in weak stimulant effects (Adams, 1978; Jiménez-

González, Ros-Moreno, Moreno-Guzmán, Rodríguez-Caabeiro, 1999; Hofmaier et al., 2014;

Robertson, Bjorn, & Martin, 1999). Its actions on nicotinic acetylcholine receptors are responsible for the spastic paralysis in nematodes, which inhibits necessary movement and egg laying (Martin & Robertson, 2010). Additionally, levamisole indirectly increases extracellular dopamine concentrations via inhibition of the DAT, but also inhibits the norepinephrine transporter (NET) (Hofmaier et al., 2014). The action of levamisole at these transporters (DAT and NET) is 100- and 300-fold less potent than cocaine, and also has weak affinity for the serotonin transporter (SERT). When administered orally, levamisole is readily absorbed via the

6 gastrointestinal tract, metabolized via the liver, and excreted by the kidneys via urine (DEA,

2013). Though the metabolic processes responsible for the breakdown of levamisole are not entirely clear (Engelmann & Richardson, 1986; Galtier, Coche, & Alvinerie, 1983; Hess, Ritke,

Broecker, Madea, & Musshoff, 2013; Kouassi, Caillé, Léry, Larivière, & Vézina, 1986), two of the predominant metabolites are pemoline and aminorex.

Dissection and electrochemical detection techniques (Schmidt, Roznoski, & Ebert, 1990;

Spector, Munjal, & Schmidt, 1998) have detected significant regional and temporal peaks in dopamine and serotonin concentrations following levamisole administration in male Sprague-

Dawley rats (35 mg/kg; ip). Levamisole administration resulted in a complex regional and temporal pattern of changes in dopamine (Figure 1; left panel) and serotonin (Figure 1; right panel). Levamisole did not alter norepinephrine levels in most brain regions. However, there was a slight and transient decrease in the cerebellum and midbrain. There was a large (p < 0.001) and rapid time-dependent increase in levels of normetanephrine (NMN), a minor norepinephrine metabolite, in all brain regions. The time-dependent effects of levamisole on dopamine were especially complex and were regionally dependent. Notable peaks in dopamine were reached in the striatum within 30 min and in the midbrain within 15 min following levamisole administration. In contrast, peaks in hypothalamic serotonin concentrations were documented at

30, 60, and 120 min post-dose, and in midbrain serotonin concentrations at 120 min following a single levamisole dose (35 mg/kg; ip; Spector et al., 1998).

7 Figure 1. Regional changes in dopamine and serotonin concentrations following levamisole administration (35 mg/kg; ip) in Sprague-Dawley rats (* = p < 0.05). Figure drawn from Tables

3 and 4 of Spector, Manjal, and Schmidt (1998).

Levamisole possesses limited stimulant-like properties, though the pharmacological profile surrounding these effects has been left largely uninvestigated (Tallarida et al., 2014;

Tallarida, Tallarida, & Rawls, 2015). In recent years, limited experimental work has been carried out regarding the pharmacokinetics of levamisole in nonhumans, although it has been shown to have a half-life of approximately 3 to 4 hours in humans (DEA 2013; Luyckx et al., 1982). While levamisole itself has not directly demonstrated abuse liability, its metabolite aminorex has been shown to possess psychomotor stimulant properties that have been likened to those of amphetamine (Hofmaier et al., 2014; Woolverton, Massey, Winger, Patrick, & Harris, 1994).

Indeed, the abuse liability of its synthetic counterpart (4-methylaminorex) has been repeatedly showcased (e.g., Glennon & Misenheimer, 1990). It has been hypothesized that the metabolic

8 conversion of levamisole to aminorex is partially responsible for its use as an additive to cocaine.

Levamisole also has similar effects to cocaine regarding sodium channels. For example, it has been clearly established that local anesthetics like lidocaine and cocaine are slow on- off sodium channel blockers and fast on-off potassium blockers (Bauman & DiDomenico, 2002).

Recently, Nowak, Rosay, Czégé and Fonta (2015) recorded local cortical field potentials and reported that tetramisole reduced neuronal response amplitude in a dose-dependent manner and decreased axonal conduction velocity. Levamisole had identical effects. Several control experiments demonstrated that these actions of tetramisole were independent from this compound acting on tissue non-specific alkaline phosphatase (TNAP). In particular, tetramisole effects were not stereo-specific. The decrease of axonal conduction velocity and the intracellular data suggested that these two imidazothiazoles block voltage- dependent sodium channels. Thus, levamisole also produces local anesthetic effects that may be comparable to those produced via cocaine administration, though these effects have not been directly examined.

Complications related to levamisole-adulterated cocaine use. Although the adverse side effects of levamisole administration have been known for some time (Symoens, Veys,

Mielants, & Pinals, 1978), the relatively recent trend of adding levamisole to cocaine has shed new light on these effects in drug-using populations. An increasing number of reports of drug- induced reactions related to levamisole have been documented, spanning a wide range of conditions.

One of the most common and serious effects related to levamisole-tainted cocaine use is agranulocytosis (i.e., agranulosis or granulopenia) (Buchanan, Vogel, & Eberhardt, 2010;

Caldwell, Graham, & Arnold, 2012; Czuchlewski et al., 2010; Hodinka et al., 1981; Muirhead &

9 Eide, 2011; Thompson et al., 1980). Agranulocytosis, which has been referred to as a “chemical form of AIDS”, is a condition in which the body fails to produce white blood cells, resulting in a severe depletion and subsequent susceptibility to serious infection (VICE, 2014). Though overt symptoms are often lacking, increased incidence of minor infections (e.g., urinary tract infections) and pneumonia are often related to agranulocytosis. A number of common prescription medications (e.g., NSAIDs, antiepileptics, and antibiotics) have been linked to agranulocytosis. Another related condition affecting white blood cells is neutropenia, which specifically affects neutrophils. Agranulocytosis and neutropenia have a similar symptomatology, though the latter may also result in a number of less severe, more overt side effects, such as fevers, ulcers, diarrhea and sore throat (NORD, 2015)

Vasculitis is another disturbing adverse effect of levamisole-tainted cocaine use (Abdul

Karim et al., 2013; James, Detz, Coralic, & Kanzaria, 2013; Tichauer, Fagan, & Goverman,

2014). In general, vasculitis is the rupturing and destruction of the blood vessels due to inflammation. Unlike agranulocytosis, this condition presents with overt symptoms. One particular form of vasculitis (cutaneous vasculopathy) specifically affects the small blood vessels in the skin, resulting in large red, black, or purple skin discolorations (palpable purpura). Other symptoms indicating vasculopathy are nose bleeds, hyptertension, and bloody stool.

The potential side effects of using levamsole-tainted cocaine range from mild to severe in nature, with varying prognoses. These effects, combined with the well-established effects of long-term cocaine use, make clear the case for studying the interaction between the two drugs from a physiological and behavioral perspective.

Potentiation of drug effects. As previously noted, cocaine increases extracellular levels of dopamine, norepinephrine, and serotonin in the brain by blocking plasma membrane

10 monoamine transporters. Neuronal pathways involved in the production of these three neurotransmitters are a reasonable place to look for targets underlying the potential interaction of cocaine and levamisole. The addition of levamisole, a drug that also indirectly increases extracellular dopamine concentrations via inhibition of these same transporters (i.e., DAT,

SERT, NET) creates a collectively greater inhibition of reuptake than either alone (i.e., additive interaction). Moreover, the primary metabolite of levamisole, aminorex, directly increases dopamine concentrations via reversal DAT and exerts amphetamine-like effects. This presents the metaphorical “perfect storm” for a potential additive/supra-additive interaction.

Subsequently, dopamine reuptake is inhibited by both cocaine and levamisole binding to DAT.

Once levamisole begins conversion to aminorex, aminorex would induce an efflux of dopamine.

Thus, co-administration of cocaine and levamisole may result in potentiation of the already potent psychomotor stimulant effects of cocaine and subsequently enhance the deleterious effects of cocaine abuse (Follath, Burkart, & Schweizer, 1971; Mlczoch, Probst, Szeless, & Kaindi,

1980; Karch et al., 2014).

The number of reports of levamisole-adulterated cocaine continues to rise, and the emergency reports and case studies are consistent with the early work demonstrating some of levamisole’s adverse side effects (Abdul-Karim et al., 2013; Formeister, Falcone, & Mair, 2014;

Knowles et al., 2009; Michaud et al., 2014; Woolverton et al., 1994; Hofmaier et al., 2014). Due to the physical and chemical properties of the drug (i.e., relatively odorless, white powder, high melting point), in addition to its widespread availability in veterinary medicine, the culmination of these characteristics lends itself to use as an adulterant or bulking agent in illicit drugs.

Analysis of drug interactions. The combined use of multiple drugs is an ever increasing aspect of the human condition (Kantor, Rehm, Haas, Chan, & Giovannucci, 2015). However, it

11 is rare that an emphasis is placed on this aspect of drug consumption when developing and manufacturing novel compounds. Indeed, it is often only after a drug has been approved and widely disseminated that some of the most important types of drug interactions have been discovered (e.g., Chamsi-Pasha, 2001). Although an inarguably important aspect of pharmacology, the study of drug interactions is rarely straightforward, even less so when the dependent measures of interest are behavioral in nature.

One way of examining drug interactions is through isobolographic analysis (Loewe,

1953). Examining the interaction between two drugs in this way requires one have a common effect against which to judge each individual drug’s efficacy. For example, ataxia is a popular categorical outcome used to measure a given drug effect. The first step in conducting drug interaction studies is to measure the relative potency of each drug alone. To this end, a range of doses for both drug A and drug B are measured such that respective ED50 values are obtained.

The range of doses and their given effect is plotted along the axes of a graph, with a straight line drawn between the ED50 for drug A on the abscissa, and drug B on the ordinate (see Figure 2).

This line now denotes the effect that would theoretically be achieved if both drugs enjoy a simply additive interaction at each dose along the dose-effect function (i.e., linear isobole).

Though the terminology surrounding drug interactions has yet to be fully resolved (Wessinger,

1986; Woolverton, 2014), there are three broad types of drug interactions: infra-additivity, simple additivity, and supra-additivity. In short, infra-additivity refers to the interaction effect of two combined doses of drugs that is less than that predicted by simple additivity, and supra- additivity refers to any interaction that is greater than that predicted by simple additivity. The description of these types of interactions is simplified using arithmetic terms. In a effect- additivity model (Woolverton, 2014) of simple additivity, 1 + 1 = 2, where the effects of drug A

12 combine with the effects of drug B in an equal manner. In infra-additivity, 1 + 1 = 1.5, where the effects of drug A combine with the effects of drug B in a manner that is less than the sum of the parts. Finally, in supra-additivity, 1 + 1 = 5, where the effects of drug A combine with the effects of drug B to produce a greater effect than the sum of its parts. The classic example of supra- additivity is the interaction between ethanol and chloral hydrate (Kaplan, Jain, Forney, &

Richards, 1969; Gessner & Cabana, 1970).

Figure 2. Graphical depiction of isobolographic analysis. The solid line (line of additivity) runs through the ED50 value for Drug A and Drug B at the abscissa and ordinate, respectively. Dotted lines represent 95% confidence intervals. Effect combinations that fall below the dotted range are considered supra-additive interactions, while combinations above the dotted range are infra- additive.

13 The method of conducting proper isobolographic analyses of drug interactions is more resource intensive than that of traditional behavioral pharmacology study designs. That is, the typical within-subject, small sample experimental designs are inconsistent with isobolographic methodology, which ideally requires independent observations at often several dozen combinations of doses. However, this challenge has not deterred behavioral researchers from attempting to research drug combinations, especially employing the behavioral assays relevant to assessing abuse liability (i.e., self-administration, drug discrimination, and dependence studies)

(see Harland et al., 1989; Woolverton, Wang, Vasterling, Carroll, & Tallarida, 2008).

Behavioral Models

Self-administration. Several assays are used to assess abuse liability and are described in detail and compared elsewhere (see Ator & Griffiths, 2003; Yokel, 1987). In summary, drug self-administration is considered to be the “gold standard” for assessing abuse liability.

Conditioned place preference (CPP) and drug discrimination (DD) procedures also provide information relevant to the abuse liability of a compound. CPP, which involves primarily classical conditioning processes, utilizes distinct environmental stimuli that have been repeatedly paired with a drug stimulus to measure the potential rewarding effects of a drug. By measuring the amount of time an animal spends in a drug-paired environment relative to a vehicle- or control-paired environment, researchers are able to infer, relatively speaking, the potential rewarding effects of a drug. In a similar manner, DD also relies on interoceptive drug stimuli to aid in the measurement of drug effects. By training an animal to respond on a particular operandum when one type of drug stimulus is present and on another when a different drug stimulus is present, one is able to measure the similarity between a given drug’s subjective effects (i.e., interoceptive stimulus condition) and the subjective effects of another (e.g., novel)

14 compound. If that compound shares subjective effects with a drug that has known abuse potential, then it is reasonable to assume that it does, as well. Additionally, one is able to infer pharmacological effects of a drug based upon responding allocated to either a drug-appropriate operandum or a vehicle-appropriate operandum. Although both procedures have their advantages, examination of drug effects in both procedures relies exclusively upon response- independent (i.e., experimenter-delivered) drug administration. Indeed, if one wishes to study the effects of a compound on drug intake patterns, the only alternative is self-administration.

The primary advantages of self-administration stem from its demonstrated validity, as well as its ability to provide meaningful, predictive data. Not only does the assay have strong face validity, but the predictive validity has been demonstrated repeatedly with regard to known drugs of abuse. For example, there is a high degree of concordance with regard to the results of human-generated and animal-generated data (Horton, Potter, & Mead, 2013; Panlilio et al.,

2005), with a substantial majority of known drugs of abuse capable of maintaining the self- administration behavior of animals (Yokel, 1987).

Jim Weeks, a cardiovascular pharmacologist at Upjohn Laboratories, developed a procedure allowing for prolonged intravenous catheter placement in a freely moving rat (Weeks,

1962). Weeks, who at the time was exploring the effects of morphine, developed a method in which rats were implanted with a jugular catheter attached to an infusion pump, which allowed for near instantaneous reinforcement of operant responding through intravenous infusion of morphine. This procedure revolutionized experimental design throughout the field of addiction studies, and has since been widely adopted and advocated for by some of the top governmental agencies across the globe in assessing the abuse liability of potential therapeutic compounds

15 (Controlled Substance Staff (CSS), 2017; European Medicines Agency, 2006; ICH M3(R2),

2009).

Although self-administration procedures may utilize a variety of administration routes

(e.g., oral or intramuscular) and species (e.g., nonhuman primates, canines), the prototypical animal model involves intravenous administration in rats. Cocaine is the typical training drug used for a number of different reasons: 1) it is a robust reinforcer that readily establishes and maintains behavior, 2) its effects on motor behavior allow for sessions of relatively short duration, 3) animals trained to self-administer cocaine can easily be transitioned to other drugs of abuse and continue to maintain self-administration, and 4) regulatory guidelines require the use of a scheduled substance for submission as part of a new drug application (NDA) (Gauvin,

Dalton, & Baird, 2017).

Interaction effects. Cocaine self-administration is a direct demonstration of the reinforcing effects of cocaine, as well as a product of behavioral history. As noted above, there are several types of drug interactions, each of which are expressed differently in drug-taking behavior. Due to the biphasic nature of the dose-effect curves produced by typical self- administration procedures (Wise, 1987), supra-additivity can be observed in several ways. A supra-additive interaction could take the form of an increased quantity of drug consumed and include an increase in overall and local response rates. For example, when the dose-effect curve shifts leftward (e.g., with agonist pretreatment), a dose located on the ascending limb would now produce a greater effect. Dependent upon the location of the dose within the dose-effect curve, a supra-additive interaction could also decrease drug intake and response rates. For example, a dose located on the descending limb would result in decreased intake relative to baseline conditions.

Figure 3. Theoretical drug-interaction dose-effect curves. The self-administration of a given positive control (drug of abuse) used as a maintenance drug (black filled circles, black line graph) is plotted as a function of drug dose. The theoretical leftward shifts in the dose-effect function that would be generated with a relevant agonist pretreatment (PTMT) (red solid circles, red line graph; e.g., amphetamine pretreatment prior to cocaine self-administration), and with a relevant antagonist pretreatment (solid blue circles, blue line graph; e.g. naloxone pretreatment prior to hydrocodone self-administration) are also depicted. The theoretical rightward shift in the dose-effect function induced by antagonist pretreatments may be limited by the direct effects on response rates engendered by the antagonist if administered alone (cf. Young, 1986).

Thus, it is imperative to examine a range of doses and within-session response pattern data to determine the true nature of a given drug’s effects upon the self-administration dose-effect curve.

Decreased intake following pretreatment with a compound could indicate any of the following have occurred: 1) increased potency of the maintenance drug, 2) behavioral toxicity, 3) satiation, 17 or 4) rate-limiting direct motoric effects. Through combined inspection of overall session data

(i.e., number of injections or total intake (mg/kg)) and within-session response patterns (e.g., cumulative record data), one is in particularly good position to determine the type of interaction having occurred between the drugs in question. The complicated nature of shifts in dose-effect curves produced using self-administration procedures has been documented (see Mello & Negus,

1996), but nonetheless many researchers often examine only one dose of a drug upon consumption of another.

The clinical implications of either an infra- or supra-additive interaction with regard to behavior between cocaine and another drug are vast. If a supra-additive interaction is observed, it would suggest that levamisole alters the reinforcing efficacy of cocaine. Contrariwise, if an infra- additive interaction is observed, this could indicate that levamisole has reinforcing effects of its own, which interfere with subsequent motivation for self-administration, or perhaps grossly impair behavior.

Drug discrimination. Subjective effects are a crucial aspect of the drug taking experience. The subjective effects of a drug refer to the covert, interoceptive stimuli generated when a given dose of a drug is administered. In everyday terms, such effects refer to “how a drug makes one feel”. One approach to studying the subjective effects of drugs is the drug discrimination assay. Though several variations on the drug discrimination assay have been developed for humans and nonhumans (Glennon & Young, 2011; Young, 1991), a food- reinforced, two-lever nonhuman operant procedure is the most commonly employed arrangement in studying the subjective or interoceptive stimulus effects of a drug. In this arrangement, a food- deprived organism (e.g., rat) is trained to respond on two separate operanda (e.g., right and left levers). Training then progresses such that it is injected with a specific dose of a drug (e.g.,

18 cocaine) and trained to respond on one lever to earn food. In separate training sessions, it receives a specific dose of another drug (or saline) and is trained to respond on the alternate lever to earn food. Once an animal is reliably responding on the appropriate lever after a given drug administration (e.g., left lever—cocaine, right lever—saline) at some predetermined criteria, it has learned the discrimination; the drug is a discriminative stimulus (SD) for responding on one lever and a S-delta (SΔ) for responding on the other, while saline serves the opposite function.

Once animals have learned the discrimination according to some predetermined testing criteria, a range of training drug doses are tested to yield a dose-effect curve. This dose-effect curve encompasses a range of doses that produce no substitution (i.e., ≤ 20% stimulus- appropriate responding), partial substitution (>20 but <80% stimulus-appropriate responding), and full substitution (≥80% stimulus-appropriate responding) (Glennon & Young, 2011). Once a dose-effect curve for the training drug has been established, another compound is then tested along a range of doses to generate its own dose-effect curve, termed a cross-generalization profile. Drugs that produce full substitution are then inferred to share some common stimulus properties with that of the training drug. A number of different manipulations can then be carried out (e.g., pretreatment with select antagonist/agonists) whose effects can be used to judge the likely pharmacological mechanisms of action affecting the stimulus properties of the comparator drug.

The phenomenon of partial substitution has sparked much debate in the field of psychopharmacology (Stolerman, 1991). There have been two broad interpretations of the subject, neither of which has been completely resolved. The first is that partial substitution represents a breakdown of stimulus control, amounting to chance responding (Colpaert, 1987).

Others hold that partial substitution holds value as a concept, stating that a stimulus that partially

19 substitutes for the training stimulus share some commonality, whether it be perceptual or receptor-mediated in nature (Gauvin & Baird, 1999; Woods, Bertalmio, Young, Essman, &

Winger, 1988). Regardless, it is clear that partial substitution frequently occurs in drug discrimination experiments. Several notable experiments (Gauvin et al., 1996; Gauvin, Carl,

Briscoe, Vallett, & Holloway, 1995; Gauvin, Criado, Moore, & Holloway, 1990; Gauvin,

Harland, Michaelis, & Holloway, 1989; Harland et al., 1989) have been conducted demonstrating that drug doses that produce partial substitution in isolation can be combined to produce full substitution in a way that is consistent with an effect-additive model of drug interaction (Woolverton, 2014). The latter interpretations served as the basis for drug interaction dose selections in the study that follows.

Interaction effects. While a primary use of the drug discrimination paradigm has been in characterizing and elucidating the neurochemical and behavioral mechanisms underlying the subjective effects of drugs, the interoceptive stimulus cues produced by concomitant administration of two or more compounds have also been examined. When plotted as function of dose, dose-effect curves for drug discrimination experiments are generally sigmoidal in nature, whereby the degree of drug-appropriate responding increases as a function of dose. As with self- administration dose-effect curves, shifts in drug discrimination dose-effect curves result from: 1) potency shift, 2) behavioral sensitivity (i.e., tolerance or sensitization) (Bertalmio & Woods,

1987; Young & Sannerud, 1989), 3) direct neuropharmacological interactions at receptor processes involved in the subjective effects of the training drug (i.e., agonist/antagonist pretreatments; Colpaert, 1987; Li, Gerak, & France, 2011), 4) influence of behavioral variables

(Lotfizadeh, Zimmermann, Watkins, Edwards, & Poling, 2014; Sannerud & Griffiths, 1993;

Sannerud & Young, 1987), and 5) change in maximal drug efficacy (i.e., partial agonism)

20 Examination of the cross-generalization profiles produced by a number of drugs against that produced by the training drug represents a significant portion of the drug discrimination literature. However, the interactive effects of a number of pretreatment drugs on subsequent generalization have also been assessed using the assay. A majority of these drugs have been targeted for their specific antagonist effects, with the goal of selectively altering or blocking the interoceptive cue generated by administration of the training drug. These lines of research are of particular interest for the development of pharmacotherapies for treatment of substance use disorders.

Of relevance to the present experiments are those studies demonstrating the blockade or enhancement of the cocaine cue. Many experiments have been carried out to examine the effects of pretreatment compounds upon the cocaine cue. Generally, the focus of these studies has been upon the various dopamine and serotonin receptor subtypes, in addition to the role that norepinephrine plays in the neuropharmacological effects of cocaine. For example, pretreatment with haloperidol, a nonselective D2 antagonist, dose-dependently decreases cocaine-appropriate responding (Colpaert, Niemegeers, Janssen, 1978a). Generally, nicotinic agonists (Banks, 2014;

Mello & Newman, 2011) and various dopaminergic compounds (Callahan, de La Garza, &

Cunningham, 1997) have shown partial and full substitution for cocaine, respectively. Of particular interest are those studies demonstrating nicotinic agonists do not substitute for cocaine, but potentiate the effects of cocaine when administered concomitantly (Mello & Newman, 2011).

While the effects of cocaine on reuptake of dopamine are undoubtedly integral to the complex cocaine stimulus, cocaine’s effects upon serotonin (Walsh & Cunningham, 1997) and norepinephrine (Spealman, 1995) have also been shown to contribute to the complex

21 interoceptive stimulus produced by cocaine administration, though the effects of these manipulations are usually dose-dependent and partial in nature.

Another area of interest lies with the study of drugs that may potentiate or have supra- additive interactions regarding the discriminative effects of drugs of abuse (Gauvin et al., 1990;

Signs & Schechter, 1986; Young, Gabryszuk, & Glennon, 1998). Many studies have been conducted concerning this matter, but few have employed the isobolographic analysis outlined above (but see Harland et al., 1989; Lelas, Rowlett, & Spealman, 2001; Li, Koek, Rice, &

France, 2010; Massey & Woolverton, 1994; Rowlett, Spealman, & Platt, 2001). Hence, the results of most studies are difficult to interpret.

Behavioral Literature

To date, no human studies have been conducted examining the cocaine-levamisole interaction. However, two studies germane to the cocaine-levamisole interaction have been carried out using animal models. Tallarida et al. (2014) sought to determine the effects of several doses of levamisole and cocaine (alone and in combination) on the stereotypical movements and

CPP of planaria. Both cocaine and levamisole alone produced statistically significant concentration-dependent increases in stereotyped movements. An isobolographic analysis was used to reveal a supra-additive interaction regarding stereotypical movements, as measured by the mean number of C-shaped movements. Cocaine produced CPP when administered in isolation while levamisole did not. Statistically significant differences in preference scores were observed relative to cocaine alone conditions when cocaine was administered in combination with various concentrations of levamisole (0.1-1 µM). However, these differences in preference scores were only observed for the lowest concentration of cocaine (0.001 µM).

22 Tallarida, Tallarida, and Rawls (2015) expanded upon the work with planaria to examine the effects of the cocaine-levamisole interaction on both locomotor activity and CPP in Sprague-

Dawley rats. In the locomotor activity experiment, animals were injected (ip) with isotonic saline solution, cocaine, levamisole, or a cocaine-levamisole combination, and activity counts were measured across a 60-min session. Levamisole alone produced statistically significant increases in locomotor activity (as measured by beam breaks) at 5 mg/kg, but not at 1 or 10 mg/kg when compared to saline. Additionally, low doses of levamisole (0.1 mg/kg) increased cocaine- induced locomotor activity when combined with 10 mg/kg cocaine, but decreased activity in combination with 30 mg/kg cocaine. Similar to the locomotor activity results, levamisole itself resulted in CPP at only one dose (1 mg/kg; ip), located toward the middle of the dose range tested. When administered concomitantly, previously inactive doses of levamisole and cocaine produced CPP at doses that in isolation did not.

Given its prevalence as an adulterant in street drugs, and the increasing number of reports of levamisole-induced adverse side effects, it is imperative that a full pharmacological and behavioral profile of the drug be developed in order to understand, identify, and treat cases related to the illegal drug use. To date, experimental work regarding the effects of levamisole- tainted cocaine in the literature is scant, with behavioral research representing an extremely small minority of that limited sample (Tallarida et al., 2014; Tallarida et al., 2015). Taken together, the results of these drug interaction studies suggest that cocaine-levamisole combinations produce effects greater than either drug administered in isolation. In other words, the effects of concomitantly administered levamisole and cocaine are additive or synergistic, but only at a select dose range. The potential for increased drug intake and modification of cocaine’s subjective effects warrants further investigation. The present document describes two

23 experiments aimed at characterizing the effects of levamisole on cocaine self-administration and discrimination.

24 CHAPTER II

EXPERIMENT 1: EFFECTS OF LEVAMISOLE ON COCAINE SELF- ADMINISTRATION IN SPRAGUE-DAWLEY RATS Overview

Previous research (Tallarida et al., 2014; Tallarida et al., 2015) indicated that levamisole sometimes increased the rewarding effects of cocaine as measured by CPP, but report of the effects of levamisole on cocaine self-administration has not yet appeared. The most appropriate way to assess this type of potential cocaine-levamisole interaction is to use a self-administration procedure, whereby one is able to directly measure several aspects of drug-taking behavior. One method of assessing the effects that one drug has upon the self-administration of another is by pretreating animals with the relevant compound of interest, and measuring subsequent self- administration behavior (Mello & Negus, 1996; National Institute on Drug Abuse, 2016). The present study examined the acute effects of levamisole pretreatments on cocaine self- administration by Sprague-Dawley rats.

Methods

Subjects. Thirty-two male Sprague-Dawley rats (Crl:SD) surgically implanted with chronic indwelling jugular catheters (polyurethane with a micro-renethane tip; SAI Infusion

Technologies, Lake Villa, IL) were purchased from Charles River Laboratories (Portage, MI).

Catheters were inserted into the jugular vein and advanced to the atrium, while the exteriorized portion exited a dorsal incision site at the midscapular area and was sutured and secured with a wound clip. Each rat was fitted with a spandex jacket (Lomir Biomedical Inc., Notre-Dame-de-

I’île Perrot, Quebec, Canada) that was used to secure the catheter to a “quick-disconnect” jacket adaptor (SAI Infusion Technologies) located on the dorsal surface of the rat. Rats were approximately 4 months old (200-400g) and had previous experience (i.e., 30 days) self-

25 administering intravenous cocaine and a 3-day exposure to a novel CNS active compound under identical behavioral contingencies used in this study (i.e., comparable drug histories). Animals originally began cocaine training at 10 weeks of age. Following a two-week acclimation period, the animals were initially food-deprived (see response acquisition below) but were allowed a 10-

15 gram per month weight gain to allow for normal growth throughout the duration of the study.

A 10-day washout period was imposed following transfer from the previous study and prior to initiation of the present study.

Animals were pair-housed in polycarbonate caging with nonaromatic bedding (56 cm x

33 cm x 21 cm). To independently control food consumption and minimize the risk of catheter damage during interactions, animals were separated by a perforated metal barrier (i.e., Buddy

Barrier™; Boggiano et al., 2008) that allowed for limited physical contact and interaction (i.e., visual, olfactory stimulation) between cohorts. Standard rodent bedding was changed and caging sanitized weekly and animals were provided with both durable (Nyla rod) and nondurable enrichment (Aspen wood blocks). Cages were housed within a temperature- (20-26ºC), humidity- (30-70%), and air pressure-controlled vivarium on a 12:12 light/dark cycle

(fluorescent lights on at 0600). Lab Diet® Certified Rodent Diet #5002 (PMI Nutrition

International, Inc., Shoreview, MN) was provided throughout the study and animals were allowed ad libitum access to water via bottles affixed to the outside of each cage. All experimental procedures and husbandry practices described herein received prior approval of the appropriate Animal Care and Use Committee and were conducted in full compliance with current national and international law and in accordance with those methods described in the

Guide for the Care and Use of Laboratory Animals (National Research Council, 2011).

26 Apparatus. Two-lever operant chambers (30.5 cm × 24.1 cm × 21.0 cm) (Med

Associates, St. Albans, VT), equipped with a swivel and tether system were used throughout the experiment. Chambers (Med Associates, ENV-008CT) were equipped with both a food hopper for pellet deliveries (response acquisition) and a single-speed (0.735 ml/min) syringe drive infusion pump (Med Associates, PHM-100). A white noise generator connected to a single external speaker was used to mask the sound of injection deliveries and any extraneous noise. 16 chambers were integrated and controlled through 16 interface modules (Med Associates, DIG-

716B) connected to an IBM-based computer system. Experimental events and data collection were controlled through a Med-PC IV software system (Med Associates).

The injection volumes ranged from approximately 35 to 150 μL/injection. The injection volume was determined by the duration of pump activation, which was controlled via computer software (Med-PC IV, St. Albans, VT). The duration of pump activation was based on the bodyweight of the animal, which was measured each day immediately prior to the session.

Infusion durations ranged from 6.9-10.1 sec in duration (M = 8.4; S.E.M. = 0.02). Previous reports have demonstrated that injection onset and duration may be varied from immediate to

100 sec without significant effect on the acquisition or maintenance of self-administration, progressive ratio breakpoints, reinstatement of drug-seeking behavior, or the assessment of the reinforcing effects of intravenously administered drugs in general (Balster & Schuster, 1973;

Crombag, Ferrario, & Robinson, 2008; Kato, Wakasa, & Yanagita, 1987; Liu, Roberts, &

Morgan, 2005; Panlilio et al., 1998; Pickens, Dougherty, & Thompson, 1969; Woolverton &

Wang, 2004; Wakabayashi, Weiss, Pickup, & Robinson, 2010).

27 Procedure

Catheter maintenance. Prior to each session, the animals’ catheters were flushed with

1.0 ml sterile saline (0.9% Normal Sterile Saline for Injection [USP]) to ensure smooth, unrestricted flow into the catheter. During the flush, the animal was observed for any signs of exterior or interior catheter leakage (e.g., the presence or development of a subdermal protrusion near the catheter passage from midscapular to right jugular access to the heart). Following session termination, animals were disconnected from the swivel/tether system and their catheters subsequently locked with either 0.5 ml of a heparinized saline solution (30% heparin) or heparinized dextrose solution (30% heparin), contingent upon whether the session immediately preceded a break in the training/testing sequence (i.e., a day off).

Catheters, exteriorization sites, quick disconnect systems, and jackets were inspected and animals were observed daily for signs of disease or distress. If catheters were found disconnected or otherwise suspected to have been compromised (e.g., lack of responding during cocaine maintenance sessions, observed swelling, unusual difficulty/ease during pre-session flush), patency was tested using an acute dose of methohexital (0.5 mg/kg, Brevital™) delivered via bolus injection through the catheter, and the animal was subsequently observed for prototypic signs of acute administration of a rapid-onset, short-acting barbiturate (i.e., loss of righting reflex and lack of muscle tone). If the catheter was declared non-patent, that animal was excluded from further testing and was humanely euthanized in accordance with institutional standard operating procedures approved by the IACUC. If a patency check was conducted on an animal and no issue with the catheter was discovered, the animal was given that day off, and training was restarted the following day.

28 Response acquisition. Initially, rats were reduced to 85-95% of their individual free- feeding bodyweight to facilitate shaping of the initial lever press response. Upon initiation of a session, a single retractable lever (left lever) extended into the chamber and a corresponding stimulus lamp (located directly above the lever) and a house light were illuminated. Initially, lever press responses were reinforced on a fixed-ratio 1 (FR 1) schedule of food reinforcement

(45 mg grain-based Dustless Precision Pellets®, Bio Serv, Flemington, NJ). Under this schedule, each lever press immediately produced a food pellet. Sessions terminated after 50 reinforcer deliveries or 30 min elapsed, whichever occurred first. Subsequent training sessions were comparable, but progressed with a gradually increasing FR requirement implemented across sessions (e.g., Day 1-FR 1, Day 2-FR 2, Day 3-FR 3...), which terminated with an FR 5 for food deliveries for all rats.

Cocaine training. For the first cocaine training session, completion of the FR 5 response requirement resulted in both a pellet and an intravenous infusion of cocaine (0.56 mg/kg/injection). Following access to both food and cocaine deliveries in this single session, completing the FR 5 schedule in all subsequent training sessions resulted solely in the delivery of a single bolus of cocaine (0.56 mg/kg/injection). Injection duration times and volumes were based upon each individual animal’s daily pre-session body weight, such that each animal received 0.56 mg/kg cocaine injections. Maintenance sessions terminated following delivery of

10 reinforcers (total daily dose of 5.6 mg/kg (iv)) or 60 min, whichever occurred first. Sessions were limited to 10 injections to reduce the likelihood of tolerance developing to the reinforcing effects of cocaine (Calipari, Ferris, & Jones, 2014; Emmett-Oglesby et al., 1993) over the initial course of training and to prevent the possibility of overdose or toxicity (Gauvin, Guha, & Baird,

2015; Young & Herling, 1986). The 10 drug deliveries (total dose of 5.6 mg/kg) also provided a

29 total daily dose that was within the behaviorally active range for cocaine commonly used in drug discrimination (e.g., Craft and Stratmann, 1996; Lamas, Negus, Gatch, & Mello, 1998) and locomotor activity assays (e.g., Thomsen, 2014). During injections the stimulus lamp above the lever flashed on and off for 0.5 s and the lever was retracted from the chamber. A 10-s timeout followed each injection during which all lights were extinguished and the lever remained retracted. Following the 10-s timeout, the stimulus and house lights were illuminated and the lever inserted back into the chamber. Subsequent sessions were comparable but progressed with a gradually increasing FR requirement implemented across sessions, terminating with an FR 10 for all rats. Training sessions were conducted 5 to 7 days each week, at approximately the same time each day.

Cocaine testing. Each rat was required to maintain < 20% day-to-day variability in the total number of injections earned for three consecutive days prior to each testing sequence. When this criterion was met, a sequence comprising three consecutive daily, 1-hr sessions began.

Animals were injected (ip) with saline at a volume of 10 ml/kg 30 min prior to initiating the self- administration session. Following the injection, each animal was placed into the darkened chamber until 30 min had elapsed, at which point the test session was initiated. During test sessions, rats were allowed unlimited access to self-administer a dose of cocaine (0-1.0 mg/kg/injection) on a FR 10 schedule of reinforcement for 1 hr for three consecutive days. Test sequences always alternated with three cocaine maintenance sessions (i.e., Train-Train-Train-

TEST-TEST-TEST-Train-Train-Train, etc.). Substitution tests for saline and the maintenance dose (MD; 0.56 mg/kg/injection) for the cocaine alone condition were conducted for all animals

(n = 32). Every other dose of cocaine was tested with a random subgroup of animals (n = 6). The same six animals were tested for each pretreatment condition for the maintenance dose (0.56

30 mg/kg/injection). No levamisole dose higher than 10 mg/kg was administered due to the known direct cardiotoxic effects of higher doses of levamisole and cocaine in the rat (Onuaguluchi and

Igbo, 1990). Accordingly, no cocaine dose higher than 0.56 mg/kg/injection was tested with either levamisole pretreatment (1 or 10 mg/kg). Testing occurred in the following order: 1) cocaine alone, 2) levamisole 1 mg/kg, and 3) levamisole 10mg/kg. All doses of cocaine (in all three conditions) were tested in random order.

Levamisole testing. Training criteria to initiate levamisole testing was identical to that of cocaine testing. Upon fulfillment of these criteria, test sequences were conducted consisting of three consecutive, unlimited-access 1-hr sessions. Animals were injected (ip) with levamisole (1 or 10 mg/kg) at a volume of 10 ml/kg 30 min prior to initiating the self-administration session.

Animals were then placed into the chambers until 30 minutes had elapsed, at which point the test session was initiated.

Drugs. Cocaine hydrochloride and levamisole hydrochloride were prepared weekly by dissolving the salt in isotonic saline solution, which was filtered through a 0.22 μm polyvinylidene fluoride (PVDF) filter. Doses were calculated and expressed as the base. The standard weight of levamisole adulteration of seized cocaine samples has been reported to be 6% by weight (Casale, Corbeil, Hays, 2008); that is, 1 g of cocaine contains 60 mg of levamisole.

The standard dose established by federal statute sentencing guidelines of the US Congress is 100 mg in a drug naïve user (United States Department of Justice (USDOJ), 2002) and up to 1 gram in the experienced user (Chitwood, 1996). The 1 mg/kg levamisole dose is equivalent to 60 mg in a 60 kg man following ingestion of 1 g of bulk illicit street product.

Data analysis. Differences between group mean data for drug intake of the maintenance dose of cocaine (0.56 mg/kg/injection), the percentage of responses emitted within the first 15

31 min for all conditions (i.e., “quarter-life” analysis), and inter-injection intervals for the maintenance dose (0.56 mg/kg/injection) were analyzed for statistical significance using one- way repeated-measures ANOVA (Graphpad Prism 7, La Jolla, CA). Quarter-life analysis was conducted using a modification of the measure originally described by Herrnstein and Morse

(1957), which excluded the first 3 min of each session to control for the initial “ramp up” in responding observed within typical cocaine self-administration sessions (Willuhn, Wanat, Clark,

& Phillips, 2009).

Results

Table 1 summarizes the daily changes in the group total number of injections self- administered (mean ± S.E.M.) during test sessions. Figure 4 shows for each cocaine dose and the three pretreatment conditions (cocaine alone, 1, and 10 mg/kg levamisole) the group mean total number of injections self-administered and total cocaine intake (mg/kg) across the three sessions of each experimental condition (solid lines), and the number of injections during each 1-hr session (bars). Pretreatment with 1 mg/kg levamisole decreased the mean number of injections of cocaine at all doses except for 0.18 mg/kg/injection, and pretreatment with 10 mg/kg levamisole further reduced the total number of injections at all doses of cocaine. During cocaine alone testing, all doses of cocaine engendered extremely stable day-to-day intake over the three days of access to cocaine (M = 32.19 injections; S.E.M. = 1.71), and responding was maintained across the full tested dose range; even relatively low doses of cocaine (0.018-0.032 mg/kg/injection) maintained stable day-to-day mean intakes across the 3-day tests. There were, however, substantial differences across animals in performance at these low doses. Pretreatment with 1 mg/kg levamisole increased the overall day-to-day variability and decreased the number of cocaine injections earned (M = 28.08 injections; S.E.M. = 2.12). Pretreatment with 10 mg/kg

32 levamisole was associated with variable and unstable levels of self-administration (M = 15.24 injections; S.E.M. = 1.67.

day averages of drug intake expressed as number infusions intake as of expressed drug of during day averages

. Descriptive statistics 3 and daily Descriptive . of

Table1 sessions test unlimitedaccess

Figure 4. Cocaine self-administration dose-effect curves. The group mean number of injections administered in three consecutive daily, 1-hr unlimited access test sessions are plotted as a function of the tested dose of cocaine (bars; mean ± 1 S.E.M.). The grand mean number of injections for each test sequence (three days) is also shown as a line graph (solid lines). Each bar represents the mean number of injections earned during the test sessions. The bars and means for saline (SAL) and 0.56 mg/kg/injection cocaine were comprised of data gathered from all subjects

(n = 32). All other bars and means were comprised of data gathered from groups (n = 6).

Figure 5. 3-day mean number of injections of the maintenance (MD; 0.56 mg/kg/injection) cocaine dose across all three conditions. Levamisole (LVM) significantly and dose-dependently decreased the mean number of injections when compared with the cocaine alone mean (1 mg/kg

LVM (p = .14), 10 mg/kg LVM (p = .007); F (2, 17), F = 2.18; 8.09).

36 Figure 6. 3-day grand mean comparisons (n = 6) of total number of injections and total cocaine intake (mg/kg) for the maintenance dose of cocaine (MD; 0.56 mg/kg/injection) for cocaine alone 1 mg/kg and 10 mg/kg levamisole (LVM) pretreatment.

Although the total number of cocaine injections earned in a session is a meaningful dependent variable, it provides no information about the temporal pattern of injections, which is an important dimension of self-administration behavior (Young and Herling, 1986).

Accordingly, representative cumulative records for a rat administering saline and several doses of cocaine (0.01, 0.1, and 0.56 mg/kg/injection (i.e., maintenance dose)) under conditions in which no pretreatment, 1 and 10 mg/kg levamisole were administered are shown in Figures 7-9.

At 0.56 mg/kg/injection, responding persisted across the full 60-min sessions and without substantial day-to-day variability, regardless of levamisole pretreatment condition. 37

Figure 7A. Representative sample of cumulative records generated for unlimited-access test sessions in the cocaine alone condition (saline). Session duration is shown on the abscissa and the number of responses is shown on the ordinate.

Figure 7B. Representative sample of cumulative records generated for unlimited-access test sessions in the cocaine alone condition (0.01 mg/kg/injection). Session duration is shown on the abscissa and the number of responses is shown on the ordinate.

Figure 7C. Representative sample of cumulative records generated for unlimited-access test sessions in the cocaine alone condition (0.10 mg/kg/injection). Session duration is shown on the abscissa and the number of responses is shown on the ordinate.

Figure 7D. Representative sample of cumulative records generated for unlimited-access test sessions in the cocaine alone condition (0.56 mg/kg/injection). Session duration is shown on the abscissa and the number of responses is shown on the ordinate.

Figure 7E. Representative sample of cumulative records generated for unlimited-access test sessions in the cocaine alone condition (1.0 mg/kg/injection). Session duration is shown on the abscissa and the number of responses is shown on the ordinate.

42 Pretreatment with 1 mg/kg levamisole did not alter the within-session pattern of responding, but was associated with a substantially less injections earned during each of the three sessions of exposure.

Figure 8A. Representative samples of cumulative records generated for unlimited-access test sessions in the 1 mg/kg levamisole (LVM) condition (saline). Session duration is shown on the abscissa and the number of responses is shown on the ordinate.

Figure 8B. Representative samples of cumulative records generated for unlimited-access test sessions in the 1 mg/kg levamisole (LVM) condition (cocaine 0.01 mg/kg/injection). Session duration is shown on the abscissa and the number of responses is shown on the ordinate.

Figure 8C. Representative samples of cumulative records generated for unlimited-access test sessions in the 1 mg/kg levamisole (LVM) condition (cocaine 0.10 mg/kg/injection). Session duration is shown on the abscissa and the number of responses is shown on the ordinate.

Figure 8D. Representative samples of cumulative records generated for unlimited-access test sessions in the 1 mg/kg levamisole (LVM) condition (cocaine 0.56 mg/kg/injection). Session duration is shown on the abscissa and the number of responses is shown on the ordinate.

The effects of pretreatment with 10 mg/kg levamisole appeared to be similar to those of 1 mg/kg pretreatment, but greater. Although levamisole pretreatment reduced the overall number of responses (and injections) in a dose-dependent fashion, it did not substantially alter the

46 temporal pattern of responding within or across sessions (see Figures 8 and 9). This effect is evidenced by visual inspection of the cumulative records, the quarter-session analyses (Figure

11), and the inter-injection interval analysis for the three conditions (cocaine alone, 1 mg/kg levamisole, and 10 mg/kg levamisole) (Figure 10).

Figure 9A. Representative samples of cumulative records generated for unlimited-access test sessions in the 10 mg/kg levamisole (LVM) condition (saline). Session duration is shown on the abscissa and the number of responses is shown on the ordinate.

Figure 9B. Representative samples of cumulative records generated for unlimited-access test sessions in the 10 mg/kg levamisole (LVM) condition (cocaine 0.01 mg/kg/injection). Session duration is shown on the abscissa and the number of responses is shown on the ordinate.

Figure 9C. Representative samples of cumulative records generated for unlimited-access test sessions in the 10 mg/kg levamisole (LVM) condition (cocaine 0.10 mg/kg/injection). Session duration is shown on the abscissa and the number of responses is shown on the ordinate.

Figure 9D. Representative samples of cumulative records generated for unlimited-access test sessions in the 10 mg/kg levamisole (LVM) condition (cocaine 0.56 mg/kg/injection). Session duration is shown on the abscissa and the number of responses is shown on the ordinate.

Figure 10. Inter-injection intervals. The mean amount of time elapsed (minutes) between successive injections earned throughout a maintenance dose (0.56 mg/kg/injection) test session is displayed as a function of condition: 1) cocaine alone, 2) 1 mg/kg levamisole (LVM), and 3) 10 mg/kg LVM pretreatment.

The differences in the percentage of responses emitted within the first 15 min of a session during cocaine and levamisole sessions were not statistically significant (1 mg/kg levamisole and 10 mg/kg levamisole (p > 0.10); F(2,17); F = 2.27). Additionally, differences in the mean inter- injection intervals during maintenance dose test sessions between cocaine and levamisole conditions were not statistically significant (1 mg/kg levamisole and 10 mg/kg levamisole (p >

52 0.10); F(2, 7); F = 2.48), corroborating the trends observed in the cumulative records via visual inspection.

Figure 11. Quarter-life analysis of responding during test sessions. The graph shows a within- subject comparison for the maintenance dose across all three conditions (cocaine alone, 1 mg/kg levamisole (LVM), and 10 mg/kg LVM pretreatment).

Discussion

As others have discussed (Feltenstein & See, 2008), the self-administration assay has demonstrated robust validity and has provided an excellent tool for examining the neuropharmacological and behavioral mechanisms contributing to the reinforcing effects of drugs of abuse. Many species, including rats, reliably self-administer cocaine. As shown in

Figure 4 (above), in the present study the dose-effect relationship depicting the number of cocaine injections as a function of cocaine dose was biphasic. Some authors refer to this pattern as hormesis, which is defined as a dose-effect relationship in which there is a stimulatory response at lower doses, but an inhibitory response at high doses, resulting in an inverted U-

53 shaped dose-effect curve (IUSDEC) (Baldi & Bucherelli, 2005; Calabrese and Baldwin, 2001a) when number of injections is the dependent variable.

Calabrese and Baldwin (1999, 2000, 2001a,b) note that because of this general phenomenon (i.e., hormesis), to examine the effects of only one self-administered dose would be potentially misleading. Nonetheless, this strategy is often used to determine whether pretreatment with a given compound alters subsequent self-administration of another. For example, studies have shown that administration of the μ-opioid partial agonist, buprenorphine

(Mello, Mendelson, Bree, & Lukas 1990), the dopamine agonist, bromocriptine (Hubner &

Koob, 1990), and the NMDA antagonist, dextromethorphan (Pulvirenti, Balducci, & Koob,

1997), reduce intake of the maintenance dose of cocaine. Such findings are sometimes mistakenly taken as evidence that the pretreatment drug is a candidate for consideration for use as a pharmacotherapy for cocaine abuse.

Unfortunately, when number of drug self-administrations or total drug intake are the primary dependent variables in a study, it is difficult to determine if a pretreatment drug actually decreases the reinforcing effectiveness of another substance, such as cocaine. In the present study, pretreatment with 10 mg/kg levamisole significantly reduced the mean number of cocaine self-administrations at the maintenance (0.56 mg/kg/injection) dose when data were combined across the three days of pretreatment with levamisole.

Figure 3 (above) presents hypothetical data showing the number of self-administrations of cocaine when no pretreatment is given (black circles) and when pretreatment is arranged with relevant agonist (blue circles) and antagonist (red circles) drugs. In this figure, the maintenance dose is 0.56 mg/kg, which falls on the descending limb of the self-administration dose-effect function. Administration of a relevant agonist would increase the relative potency of this

54 maintenance dose and reduce responding (and cocaine self-administrations) at the maintenance dose. However, administration of a relevant antagonist also would reduce responding at the 0.56 mg/kg maintenance dose. If these pretreatment drugs were tested at other doses of cocaine, however, different results would be obtained. If levamisole was tested with only lower doses of cocaine, one may conclude that levamisole potentiates the effects of cocaine and would not be suitable as a potential pharmacotherapy. The dose-dependent differences in the extent and type of effect that levamisole had on behavior are illustrated by the full dose-effect functions in

Figures 4 and 5.

Although pretreatment with both doses of levamisole generally reduced cocaine self- administration in the present study, there were subtle but detectable differences in the effects of levamisole pretreatment across cocaine doses and the interaction between the two drugs did not conform to simple additivity/antagonist models of drug interaction. For example, at low cocaine doses (0.01-0.032 mg/kg/injection), pretreatment with levamisole increased the number of injections on day 1 (Figure 4), a pattern of responding that could be interpreted as an extinction burst resulting from an antagonistic effect of levamisole (Morse, 1966; Amsel, 1967; Overmann

& Denny, 1974). Pretreatment with levamisole reduced cocaine self-administration on day 2 and day 3 relative to control (i.e., no levamisole) levels, which is also consistent with an antagonistic effect, because following the initial burst responding falls to a relative low level in extinction, regardless of whether or not it is drug-induced. Given this analysis, the 3-day mean and daily response measures for these doses of cocaine (0.01-0.32 mg/kg/injection) would be indicative of levamisole interfering with the reinforcing effects of cocaine, as would be predicted by antagonism and a subsequent rightward shift in dose-effect curve. But interpreting the effects of pretreatment drugs on the ascending limb of the cocaine self-administration dose-response curve

55 is fraught with difficulty, as other researchers have emphasized (Wise, 1987; Piazza, Deroche-

Gamonent, Rouge-Pont, Le Moalet, 2000; Katz and Higgins, 2004), and for that reason researchers often examine only a single cocaine dose, or only the descending limb of the dose- response function (Wise, 1987; Ahmed & Koob, 2004).

Increases in responding observed across days at the peak of the dose-effect curve (0.056-

.18) suggest that 1 mg/kg levamisole may escalate intake over time, although the relatively small range of doses at which this was observed and the short period of testing cannot confirm this definitively. Such effects would be consistent with previous reports (Hofmaier et al., 2014;

Tallarida et al., 2015) regarding the nature of levamisole addition to illicit cocaine. Differences in overall intake at these doses with 1 mg/kg levamisole pretreatment were insignificant.

However, drug intake for high doses of cocaine (0.32 and 0.56 mg/kg/injection) with 1 mg/kg levamisole pretreatment was substantially lower than intake at other dose levels. These lower intake levels may be a product of either behavioral toxicity or satiation, but summary self- administration data cannot confirm either interpretation. Wilson and Schuster (1973) summarized this phenomenon with regard to both monkeys and rats, and provided three reasons why decreases in responding for a drug would be observed as unit dose is increased:

(1) the drug in a dose-dependent fashion may disrupt any ongoing behavior (2) at

higher dosages the drug may produce aversive effects e.g. stimulation, and these

may predominate over any increased reinforcing efficacy, or (3) perhaps drug-

related changes in neurotransmitter levels are occurring which in some manner

make the self-administration of additional quantities of the psychomotor

stimulants non-reinforcing. (p. 292)

56 Overall session data (i.e., total number of infusions) are meaningful, but to fully understand the effects of a pretreatment drug one must examine momentary changes in behavior, which can be accomplished through the use of cumulative records, a once popular method for examining the effects of a wide range of independent variables on responding (Lattal, 2004;

Poling, 1979). Figures 7A allow for comparison of response patterns for saline for all three treatment conditions (cocaine alone, 1 and 10 mg/kg levamisole, respectively). Although dose- dependent reductions in responding are evident with levamisole pretreatments, the response patterns are generally similar across conditions. These dose-dependent differences are also apparent at the lowest dose of cocaine tested. Figures 7-9B show cumulative record data for 0.01 mg/kg cocaine for all three treatments (cocaine alone, 1 and 10 mg/kg levamisole, respectively).

Figures 7B displays a response pattern typical of a drug stimulus failing to maintain self- administration (0.01 mg/kg/injection cocaine). In contrast, the within- and between-session response patterns for the low dose of cocaine (0.1 mg/kg/injection) show a high, stable rate of responding across the entire session and across all three test days (Figures 7C). Figures 7-9C showcase response patterns for 0.1 mg/kg/injection, the peak of the dose-effect function, for all treatment conditions (cocaine alone, 1 and 10 mg/kg levamisole, respectively). Response patterns for cocaine alone and 1 mg/kg levamisole for 0.1 mg/kg/injection are characteristic of an effective, mid-dose reinforcer (i.e., stable, rapid responding throughout the sessions, with little- to-no post-reinforcement pause).

One limitation of this study was the use of non-naïve rats as subjects. Although a 10-day washout period was instated following termination of the original source study for these animals, this limited experimental history could have played a role in the subsequent self-administration in the range of cocaine doses tested. For example, substitution testing with the Sponsor drug on

57 the previous study failed to maintain self-administration behavior in any animal tested. That is, all animals experienced conditions during which responding resulted in intravenous delivery of a non-reinforcer. However, this additional “extinction” period (i.e., 3-day substitution test with a non-reinforcer) was the same for all animals. Moreover, all animals also initially underwent 3- day substitution tests with saline (i.e., a non-reinforcer), but cocaine self-administration was readily re-acquired and maintained.

An additional limitation to the present study is inherent to the experimental design itself.

The present study was conducted using a standard preclinical self-administration study design, whereby animals chosen for testing are selected at random from a group of animals that has collectively met testing criteria. Each animal is repeatedly tested at varying doses (see Mello &

Negus, 1996), and subsequent performance is judged against a stable baseline of performance established during intermediate 3-day training sequences. Thus, each animal could and did have a slightly different testing history that precludes a within-subject assessment of behavior at all doses tested, and which could impact the results of subsequent testing sessions. For example, it has been shown that squirrel monkeys with a history of exposure to relatively higher doses of cocaine self-administer lower doses more readily than in earlier tests when those same animals did not have a history of high dose cocaine self-administration (Goldberg, 1973). However,

Young and Herling (1986) note that once contingent drug delivery is arranged for a given behavior, the development and maintenance of behavior appears to be controlled by prevailing access conditions, rather than by the conditions important in initially establishing the drug as a reinforcer. Furthermore, all animals in this study had exposure to a range of both high and low doses of cocaine (0.01-1 mg/kg/injection) and levamisole had comparable effects across animals.

58 Thus, it seems unlikely that the abbreviated drug histories in the present study would have influenced the present results, though it cannot be definitively excluded.

It is, of course, possible that levamisole at the doses tested generally decreases operant behavior, regardless of reinforcer type. The lack of an extinction burst for the levamisole 10 mg/kg pretreatment on day 1 and the failure to observe a subsequent decrease in responding across the remaining test sessions is likely a product of behavioral history coupled with direct behavioral effects of levamisole on response rates. To properly tease apart the variables at play in this situation would require an extended period of testing and a number of different controls

(e.g., food-maintained control group), neither of which were employed within the present design.

Cumulative records for the high dose of cocaine (0.56 mg/kg/injection) show diminished, yet stable within- and between-session response patterns across the 3-day test sequence (Figures 7-

1D). The intervals between injections (downward, diagonal blips of the pen) and the length of the post-reinforcement pauses generally increase with relatively equal spacing in direct proportion to the magnitude of the available dose of cocaine. Although comparatively little responding occurred at these doses with levamisole pretreatments (Figures 8D and 9D), responding is evenly distributed throughout the entire session, indicating that the higher doses of cocaine are indeed serving to reinforce responding. Analysis of cumulative records suggests that responding was spaced throughout the session in a manner consistent with an effective reinforcer, rather than nonspecifically reduced due to behavioral toxicity.

Conclusion

Two prior studies using a CPP procedure demonstrated that under some conditions levamisole can increase the rewarding effects of cocaine (Tallarida et al., 2014; Tallarida et al.,

2015). The present study explored whether similar effects would be observed under a cocaine

59 self-administration procedure, in terms of increased self-administration of one or more doses of cocaine. If such an effect were observed, it would seem to provide at least a partial accounting for why producers often add levamisole to illicit cocaine. Pretreatment with two doses of levamisole (1 and 10 mg/kg) never significantly increased cocaine self-administration. Instead, pretreatment generally reduced cocaine self-administration at the maintenance dose (0.56 mg/kg/injection) and across a range of other doses (0.018-0.1; 0.32-0.56 mg/kg/injection). As others have discussed at length (e.g., Balster & Lukas, 1985; Goldberg, Woods, & Schuster,

1969; Tallarida, Porreca, & Cowan, 1989; Tallarida, 2000), interpreting changes in drug- maintained responding induced by pretreatment with another substance poses conceptual challenges. These discussions make it clear that proper interpretation requires researchers to consider both the total number of responses emitted and moment-to-moment changes in responding within test sessions (Mello & Negus, 1996; Roberts, Loh, & Vickers, 1989). The present study examined both of these measures, and neither provided evidence that levamisole pretreatment increased cocaine intake. Therefore, the present data do not support the hypothesis that levamisole is added to illicit cocaine to increase the amount of cocaine self-administered by street users in individual bouts of drug use.

To date, there are no published studies of the effects of levamisole on schedule- controlled (e.g., FR) responding, and it is possible that levamisole at the doses tested generally decreases operant behavior, regardless of reinforcer type. Such an action would be incompatible with levamisole increasing cocaine self-administration under the conditions of the present study. The effects of levamisole on responding controlled by reinforcers other than cocaine are certainly worth investigating. Until such effects are known, a limitation of the present study is that its findings may be confounded by nonspecific effects of levamisole.

60 Another limitation is that levamisole pretreatment was evaluated, but human users would administer levamisole simultaneously with cocaine, which might influence how the drugs interact. Future studies examining the concomitant self-administration of cocaine-levamisole combinations may yield a more complete and accurate picture of the effects of levamisole on initiating and maintaining cocaine self-administration.

Although the metabolic processes responsible for the breakdown of levamisole are not entirely clear (Hess et al., 2013), two of the predominant metabolites are pemoline and aminorex

(schedule IV and schedule I stimulants, respectively). Pemoline and aminorex are reported to have amphetamine-like properties, hence adding levamisole to cocaine may over time increase the intensity of the stimulant-like subjective effects produced by a single drug administration, as well as the duration of those effects (Hofmaier et al., 2014). Such an action is likely to decrease the amount of cocaine administered by a lab animal or a human in a single setting (Gold &

Balster, 1996), while increasing the relative reinforcing efficacy of a given dose of cocaine. But it might well increase the likelihood that bouts of self-administration would be repeated, or the perceived value (and worth) of the substance being administered. Either of these outcomes would be desirable from the perspective of a drug dealer. Further research using other procedures (e.g., drug discrimination, demand curve analysis, reinstatement) to examine them is certainly merited.

61 CHAPTER III

EXPERIMENT 2: ASSESSMENT OF THE SUBJECTIVE EFFECTS OF LEVAMISOLE ALONE AND IN COMBINATION WITH COCAINE IN SPRAGUE-DAWLEY RATS Overview

While measurement of drug-taking behavior is inarguably crucial to understanding the behavioral profile of a given drug, the subjective or interoceptive stimulus effects of drug administration comprise an important aspect of the drug-taking repertoire (Fischman & Foltin,

1991). Drug discrimination is one effective way to measure the subjective effects of drug administration. Given that levamisole decreased intake of cocaine in the self-administration experiment, it was a natural continuance of interest to determine whether the interoceptive or subjective changes in cocaine by levamisole contributed to the change in the reinforcing efficacy in the previous study.

Recently, the role of sex as a biological variable has enjoyed increasing popularity in the world of preclinical research. This interest has largely been spurred by the recent changes for

NIH grant applicants (McCullough et al., 2014). This issue has garnered sufficient attention such that recommendations have more recently affected preclinical abuse liability assessment. Indeed, the most recent FDA Guidance document (CSS, 2017) has recommended the inclusion of male and female subjects when conducting preclinical abuse liability assessments (but see Gauvin &

Zimmermann, 2017). Other investigators (e.g., Carroll & Anker, 2010; Craft, Kalivas, &

Stratmann, 1996) have demonstrated that sex can influence the effects that drugs of abuse have upon behavior. Thus, it was also of interest to determine if sex influenced the cocaine-levamisole interaction.

To this end, the cross-generalization profiles of levamisole alone and in combination with cocaine were examined in males and females to determine if levamisole in isolation produces

62 measurable discriminative stimulus effects similar to those of cocaine, or if cocaine-levamisole combinations produce a more salient “cocaine-like” stimulus. Furthermore, earlier work

(Tallarida et al. 2014; 2015), including the previously detailed self-administration study, has demonstrated interactions between cocaine and levamisole. Accordingly, the present study was designed to assess: 1) the subjective effects of a range of cocaine doses, 2) the cross generalization profiles between cocaine and levamisole, and 3) the relative changes in discriminative stimulus effects engendered by concomitant administration of cocaine-levamisole.

While a cocaine-levamisole interaction has been demonstrated, it is unclear whether levamisole itself or its active metabolites are responsible for the relative changes observed in the

CPP and self-administration studies. Hofmaier et al. (2014) noted that it is unlikely that levamisole administration produces psychomotor stimulant effects. Rather, it is more likely that either of the two centrally active metabolites of levamisole (pemoline and aminorex) is responsible for any interaction when combined with cocaine. Thus, using the drug discrimination paradigm to determine the time-course effects of levamisole alone and in combination with cocaine is a useful way to assess this aspect of potential drug interaction. Finally, it was previously noted that data are lacking on the non-specific effects of levamisole on operant responding. Given that drug discrimination employs a food-maintained operant response (i.e., lever pressing), the present procedure provided an opportunity to assess whether levamisole or its combination with cocaine have any rate-suppressing effects on food-maintained operant responding.

Methods

Subjects. Sixteen experimentally naïve Sprague-Dawley rats (Crl:SD; 8 male, 8 female) were purchased from Charles River Laboratories (Raleigh, NC). Rats were approximately 4-6

63 weeks old (130-210g) at the initiation of the experiment. Following a 1-week acclimation period, animals were food-deprived (see response acquisition below) but were allowed a 10-15 g per month weight gain to allow for normal growth throughout the duration of the study.

Animals were pair-housed in polycarbonate caging with nonaromatic bedding (56 cm x

33 cm x 21 cm). To independently control food consumption, animals were separated by a perforated metal barrier (i.e., Buddy Barrier™; Boggiano et al., 2008) that allowed for limited physical contact and interaction (i.e., visual, olfactory stimulation) between cohorts. Standard rodent bedding was changed and caging sanitized weekly and animals were provided with both durable (Nyla rod) and nondurable enrichment (Aspen wood blocks). Cages were housed within a temperature- (20-26ºC), humidity- (30-70%), and air pressure-controlled vivarium on a 12:12 light/dark cycle (fluorescent lights on at 0600). Lab Diet® Certified Rodent Diet #5002 (PMI

Nutrition International, Inc., Shoreview, MN) was provided throughout the study and animals were allowed ad libitum access to water via bottles affixed to the outside of each cage. All experimental procedures and husbandry practices described herein received prior approval by the appropriate Animal Care and Use Committee, and were conducted in full compliance with current national and international law and in accordance with those methods described in the

Guide for the Care and Use of Laboratory Animals (National Research Council, 2011).

Apparatus. Two-lever operant chambers (30.5 cm × 24.1 cm × 21.0 cm) (Med

Associates, St. Albans, VT) were used throughout the duration of the experiment. Chambers

(Med Associates, ENV-008CT) were equipped with both a food hopper for pellet deliveries. A white noise generator connected to a single external speaker was used to mask any extraneous noise. Sixteen chambers were integrated and controlled through sixteen interface modules (Med

64 Associates, DIG-716B) connected to an IBM-based computer system. Experimental events and data collection were controlled through a Med-PC IV software system (Med Associates).

Procedure

Response acquisition. During all sessions, both levers were extended into the chamber and the corresponding stimulus lights illuminated. Sessions terminated after 50 reinforcer deliveries or 30 min elapsed, whichever occurred first. Rats were initially shaped to press either lever under a FR 1 schedule of food reinforcement, where each response immediately produced a food pellet. Reinforcement was accompanied by the two stimulus lights located above each lever flashing on and off for 0.5s during each reinforcer delivery. Following completion of the initial response shaping session, responding on only one lever was reinforced. The lever on which the least amount of responding occurred during the initial shaping session was the first to be trained in isolation. Responding on the inappropriate lever reset the FR component, and required another entire component be completed on the appropriate lever to earn reinforcement. After completing a session (i.e., earning 50 reinforcers) on an FR-1, the lever that produced reinforcement was alternated (e.g., Left, Right, Left, Right, etc.) daily. If the animal completed 50 reinforcers on this lever, the FR requirement was increased by one and the lever producing reinforcement was again alternated. This training sequence continued until the rat had completed an FR 3 on each lever, at which point drug training commenced.

Drug training. Once drug training began, the reinforced lever was alternated in a pseudo-random manner (e.g., L-R-R-L-R-R-L-L-R-L), repeating once every 10 days. This random sequence was used to minimize the development of a lever preference during drug training, and to bring responding solely under the control of the interoceptive stimuli. For each animal, responding on a given lever was only reinforced after receiving an injection of cocaine,

65 while responding on the other lever was only reinforced after receiving an injection of saline.

Responding on the inappropriate lever reset the FR component, and required another entire component be completed on the appropriate lever to earn reinforcement. If an animal successfully completed a session on a given FR component, the component was increased on an individual basis to a terminal FR-10 schedule of reinforcement. The total number of drug and saline days were balanced over the course of 30 days such that there was equal representation with regard to exposure to the training stimuli. The following is a sample weekly sequence of drug and saline training days: DSSDSSD. Training occurred 5-7 days each week at approximately the same time each day.

Rats were injected with either saline or cocaine (10 mg/kg; ip) at a volume of 1 ml/kg 15 min prior to session initiation, and placed into the darkened chambers until the pre-session injection interval (PSII) had elapsed, at which point both levers were extended into the chamber and the corresponding stimulus lamps and a house light were illuminated. Drug training sessions terminated after the delivery of 50 reinforcers or 30 min had elapsed, whichever occurred first.

Upon reaching the terminal FR 10 schedule on both levers, rats were required to complete a 5- day, double-alternation sequence (e.g., DDSSD, SSDDS) to be considered for testing.

Additionally, rats were required to emit fewer than 18 total responses prior to the first reinforcer delivery (FRF; see Colpaert, Niemegeers, & Janssen, 1978b) and have greater than 80% stimulus-appropriate responding during the previous five sessions. If decrements in stimulus control were observed on any intervening training sessions (i.e., >18 FRF or < 80% stimulus- appropriate responding), an additional 2-day sequence of D-S or S-D was required before testing recommenced.

66 Testing. Testing commenced once rats had fulfilled the aforementioned criteria. Testing sessions were identical to training sessions, with the exception that responding was reinforced on either lever. That is, completion of 10 consecutive lever responses on either lever would result in delivery of a reinforcer. Test sessions were alternated with training sessions in a simple sequence

(e.g., Drug—TEST—Saline—TEST—Drug—TEST). If during any interspersed training session responding failed to meet the FRF or overall stimulus-appropriate response requirement (%), an additional 2-day training sequence (e.g., S-D or D-S) was required to be completed before testing recommenced. A total of 10 rats (5 males, 5 females) reached and maintained criteria for the duration of the cocaine alone and levamisole alone (15- and 60-min PSII) conditions.

All test sessions consisted of a two injection (ip) procedure. For the cocaine alone dose- effect curve, rats were injected first with a dose of cocaine at 1 ml/kg in the lower right quadrant of the abdomen, then with an additional injection of saline at 10 ml/kg in the lower left quadrant.

For levamisole tests, animals were first injected with saline at 1 ml/kg in the lower right quadrant, and then with levamisole at 10 ml/kg in the lower left quadrant. For levamisole 60-min PSII, levamisole was administered 60 min prior to session initiation, with an addition saline control injection administered 15 min prior to session initiation. For cocaine-levamisole interaction tests, rats were injected first with a dose of cocaine at 1 ml/kg in the lower right quadrant of the abdomen, then with an additional injection of levamisole at 10 ml/kg 15 min prior to session initiation. Saline and varying doses of cocaine (0.1-10 mg/kg) and levamisole (1-10 mg/kg) were tested alone and in combination. The order of testing was as follows: cocaine alone, levamisole

(15-min PSII), levamisole (60-min PSII), and cocaine-levamisole interactions. All dose levels were tested in random order with the exception of the interaction phase. No levamisole or cocaine dose higher than 10 mg/kg was administered due to the known direct cardiotoxic effects

67 of higher doses of levamisole and cocaine in the rat (Onuaguluchi and Igbo, 1990).

A subset of males (n = 4) and females (n = 4) were used for all cocaine-levamisole combination studies for within-subject comparisons. Similar to the methods employed by Gauvin et al. (1986) a limited range of dose combinations were tested, encompassing combinations which in summation, would theoretically produce >80% cocaine-appropriate responding under an effect-additivity model of drug interaction. Combination tests were not scheduled with constituent doses of cocaine or levamisole that alone produced full substitution. When dose combinations failed to engender cocaine-appropriate responding, dose levels were varied in an effort to produce full generalization (>80% cocaine-appropriate responding). Cocaine and levamisole doses up to 10 mg/kg were tested. The constituent levamisole dose was chosen at random from the three highest doses that did not produce full substitution for cocaine. The first cocaine dose to be tested in combination with that levamisole dose was the highest dose of cocaine that did not produce full substitution. Once responding reached >80% cocaine- appropriate responding with a given cocaine-levamisole combination, no higher cocaine dose was tested in combination with that levamisole dose. Subsequently, decreasing cocaine doses were tested until no substitution (i.e., <20% cocaine-appropriate responding) occurred in combination with levamisole. Each dose combination was selected and tested based on an individual animal’s generalization data. No dose of cocaine or levamisole that by itself produced full substitution was tested in combination.

Drugs. Cocaine hydrochloride was prepared weekly by dissolving the salt in isotonic saline solution, which was filtered through a 0.22 μm polyvinylidene fluoride (PVDF) filter. Due to the lengthy stability of prepared levamisole solutions (Chiadmi et al., 2005), levamisole was prepared in an identical manner to cocaine once prior to initiation of levamisole tests and again

68 during the interaction tests, as needed. Doses of cocaine and levamisole were calculated and expressed as the base. The standard weight of levamisole adulteration of seized cocaine samples has been reported to be 6% by weight (Casale et al., 2008); that is, 1 g of cocaine contains 60 mg of levamisole. The standard dose established by federal statute sentencing guidelines of the US

Congress is 100 mg in a drug naïve user (United States Department of Justice (USDOJ), 2002) and up to 1 gram in the experienced user (Chitwood, 1996). The 1 mg/kg levamisole dose is equivalent to 60 mg in a 60 kg man following ingestion of 1 g of bulk illicit street product.

Data analysis. The mean (± S.E.M ) percentage of drug-appropriate responding was plotted as a function of dose for all three treatment conditions. Full substitution was defined as

≥80% of responses on the cocaine-appropriate lever, partial substitution as >20 but <80% of responses on the cocaine-appropriate lever, and no substitution as <20% of responses on the cocaine-appropriate lever. Linear regression was conducted to estimate the individual ED50 values for all relevant dose-effect curves (Graphpad Prism 7, La Jolla, CA). Doses included for the linear regression were limited to the highest dose that produced no substitution and the lowest dose that produced full substitution. Differences in the numbers of sessions to reach testing criteria and effect-additivity predictions versus obtained values (see Woolverton, 2014) were analyzed via an independent and paired samples t-tests, respectively. The mean (± S.E.M ) percent cocaine-appropriate responses response rates were expressed as responses per second and plotted as a function of dose. Differences in percentage of cocaine-appropriate responding, response rates, and FRF values were analyzed via two-way, mixed-model ANOVA, using

Dunnett’s multiple comparison test.

Results

The comparative number of training sessions required to meet criteria is shown in Figure

69 12. The differences in the required number of training sessions to meet criteria for males and females was not statistically significant (t = -1.8; p = .26).

Figure 12. Comparison of the number of training days required to meet testing criteria.

Table 2 is a summary of the estimated ED50 values for cocaine, levamisole, and cocaine- levamisole dose combination tests. Figures 13 and 14 display the dose-effect curves for the range of cocaine, levamisole (15- and 60-min PSII) and cocaine-levamisole combinations for each individual animal.

Table 2. Individual estimated ED50 values for all three treatment conditions.

70 Cocaine alone substituted for itself in a dose-dependent manner in both males and females. The two-way mixed-model ANOVA showed no statistically significant effect of sex on percentage of cocaine-appropriate responding (p = .46; F(1, 8) = .60), response rate (p = .39;

F(1, 8) = .81) or FRF values (p = .38; F(1, 8) = .89). There were statistically significant effects of dose on percentage of cocaine-appropriate responding (p < .0001; F(9, 72) = 36.72) and response rate (p < .05; F(9, 72) = 3.74), but not FRF values (p = .53; F(9, 72) = 1.38). Dunnett’s multiple comparison test indicated significant at the 3.2 mg/kg dose level for the percentage of cocaine-appropriate responding (p < .0001) and the 10 mg/kg dose level for response rate (p <

.05). The lowest doses of cocaine that produced either partial- or full-substitution produced accompanying dose-dependent decreases in response rates.

Figure 13. Cocaine alone dose-effect curves for all three dependent measures for both male (n =

5; filled circles) and female (n = 5; open circles) rats. Error bars represent ± 1 S.E.M.

A representative within-subject (111F) cumulative record is presented below (Figure 14) for three doses of cocaine which produced full (10 mg/kg), partial (5.6 mg/kg), or no substitution

(saline). The three panels display response patterns for an entire test session.

Figure 14. Representative cumulative records for within-session responding for cocaine alone.

The top panels represent drug-appropriate responding while the bottom panels represent saline- appropriate responses.

Cross-generalization profiles for levamisole were highly idiosyncratic. Levamisole fully substituted for cocaine at the 15-min PSII for one animal (104M; male) and resulted in partial substitution (24% cocaine-appropriate responding) for another (115F; female), though dose- dependent increases in the percentage of drug-appropriate responding and decreases in response rate at the group level were observed beginning at the 5.6 mg/kg dose. Levamisole (60-min PSII) at 10 mg/kg produced full substitution in only one animal (104M).

Figure 15. Levamisole alone dose-effect curves for all three dependent measures for both male

(n = 5; top panels) and female (n = 5; lower panels) rats. Filled and open squares represent 15- and 60-min PSII treatment conditions, respectively. Error bars represent ± 1 S.E.M.

Levamisole administration using 15-min PSII resulted in dose-dependent decreases in response rates. For the 15-min PSII, differences between male and female cocaine-appropriate responding (p = .54; F(1, 8) = .41), response rates (p > 0.99; F(1, 8) = 0), and FRF values (p =

.48; F(1, 8) = .54) were not statistically significant at any dose tested. Levamisole administration using the 60-min PSII resulted in dose-dependent decreases in response rates, but to a lesser extent that the 15-min PSII condition. Like the 15-min PSII condition, differences between male and female cocaine-appropriate responding (p = .93; F(1, 8) = .01), response rates (p = .68; F(1,

8) = .18), and FRF values (p = .62; F(1, 8) = .26) for the 60-min PSII condition were not statistically significant at any dose tested. Differences in response rate for the 15- and 60-min

PSII conditions were statistically significant for males (p < .0001; F(5,20) = 12.08), with significant differences between saline and the 5.6 (p < .05) and 10 mg/kg (p = .001) dose levels.

73 Differences in response rate under the 15- and 60-min PSII conditions were statistically significant for females (p < .0001; F(5,20) = 14.99), with significant differences between saline and 10 mg/kg (p < .0010). Differences between 15- and 60-min PSII conditions on measures of

FRF were statistically significant for the 10 mg/kg levamisole dose in males (p = .02) but not statistically significant for any dose tested in females (p = .23).

The conclusions supported by visual inspection of the data were corroborated by a two- way mixed-model ANOVA, which showed that differences between the three conditions on cocaine appropriate responding (p < 0.001; F(5, 45) = 31.02) and response rates (p < 0. 001; F(5,

45) = 26.89) were statistically significant for 1 mg/kg cocaine and upward when compared to all doses of levamisole for both PSII conditions, but not for any dose on the FRF measure (p = .78;

F(5, 45) = 0.49).

Drug combination tests. Differences between obtained and expected values assuming a simple effect-additivity model of drug interaction were analyzed via a paired samples t-test, showing an overall significant deviation from the model (t = 6.28; p < .0001) regardless of levamisole dose. Overall, there was no difference between male and females subjects on measures of percent cocaine-appropriate responding, response rates, or FRF values. Within- subject data for drug combinations are presented below in Figures 17-25 for all three dependent measures. Cocaine-levamisole combinations produced full substitution for cocaine, an effect that was highly dependent upon the constituent levamisole dose. All cocaine-levamisole combinations that produced partial or full substitution did so at consistently lower doses than for either cocaine or levamisole constituent dose. Across all animals, levamisole was given in combination with cocaine on 68 occasions. On 47 of those occasions (69%), a higher percentage of cocaine-appropriate responses occurred than when the same dose of cocaine or levamisole

74 was administered alone. This pattern was observed in 67, 67, 75, 62, and 71% of test sessions for

1, 1.8, 3.2, 5.6, and 10 mg/kg levamisole constituent doses, respectively. Figure 16 contains representative cumulative record within-subject comparisons for three cocaine-levamisole test combinations, one of which produced full substitution (LVM 5.6 and COC 3.2 mg/kg) while the others resulted in partial substitution (LVM 10 and COC 1.8 mg/kg; LVM 10 and COC 3.2 mg/kg).

Figure 16. Representative cumulative records for within-session responding for cocaine- levamisole combinations. The top panels represent drug-appropriate responding while the bottom panels represent saline-appropriate responses.

For animal 103M (Figure 17), only 5.6 mg/kg levamisole combinations resulted in a ½ log unit leftward shift in the dose-effect curve, as well as dose-dependent decreases and increases in response rate and FRF values, respectively. Two other doses tested (3.2 and 10 mg/kg LVM) failed to substitute when combined cocaine, but 10 mg/kg levamisole combinations produced dose-dependent decreases in responding and increases in FRF values.

Figure 17. Dose-combination dose-effect curves for animal 103M.

Cocaine-levamisole combinations for animal 104M (Figure 18) dose-dependently substituted for cocaine and did so at ½ log unit doses lower than cocaine alone. All three doses of levamisole tested (1, 1.8, and 3.2 mg/kg) in combination with cocaine had nearly identical cross- generalization profiles (percent cocaine-appropriate responding, response rate, and FRF). When originally tested, levamisole alone produced dose full substitution at 5.6 mg/kg, indicating a ¼ log unit leftward shift from the levamisole alone curve.

Figure 18. Dose-combination dose-effect curves for animal 104M.

Animal 106M (Figure 19) demonstrated a ¼ log unit leftward shift in the dose-effect curve for 5.6 mg/kg levamisole in combination with cocaine, producing full substitution at 3.2 mg/kg cocaine constituent dose, relative to the full substitution produced by the original 5.6 mg/kg cocaine alone dose. 3.2 and 10 mg/kg levamisole combinations resulted in only partial substitution (24 and 50% cocaine-appropriate responding, respectively), with both doses producing dose-dependent decreases in response rate and FRF as the constituent dose of cocaine 76 increased.

Figure 19. Dose-combination dose-effect curves for animal 106M.

All three levamisole combination conditions (3.2, 5.6, and 10 mg/kg) produced leftward shifts of the dose-effect curves relative to that of cocaine alone in animal 108M (Figure 20). 10 mg/kg levamisole combinations produced the greatest leftward shift of the dose-effect curve and dose-dependent decreases in response rate. 3.2 and 5.6 mg/kg levamisole combinations also shifted the dose-effect curve leftward, in a relatively equal manner. The differentiating dose for these two levamisole combinations was the 3.2 mg/kg cocaine constituent dose, which produced partial substitution (44 % cocaine-appropriate responding) when combined with 3.2, but not 5.6 mg/kg levamisole (14 % cocaine-appropriate responding). No appreciable effects were observed upon the FRF measure for any levamisole combination tests.

Figure 20. Dose-combination dose-effect curves for animal 108M.

Interestingly, 108M demonstrated response alternation between vehicle- to drug- to vehicle-lever switches during 5.6 mg/kg cocaine alone tests, though this pattern (Figure 21) did not persist for

77 dose combination tests which resulted in partial substitution for this animal.

LVM 5.6 + COC 3.2

Figure 21. Cumulative record for 5.6 mg/kg cocaine alone test session for animal 108M. The top panel represent drug-appropriate responding while the bottom panels represent saline-appropriate responses.

The only levamisole combination condition dose that produced full substitution in animal

109F (Figure 22) was 10 mg/kg levamisole. Only one constituent cocaine dose showed increase relative to cocaine alone conditions (1.8 mg/kg), which increased from 0 to 78% cocaine- appropriate responding under the 10 mg/kg levamisole combination condition. Response rates for the 10 mg/kg levamisole combinations were substantially lower than under all other conditions, though no such effect was apparent with the other dose combinations. No appreciable effect was observed upon the FRF measure.

Figure 22. Dose-combination dose-effect curves for animal 109F.

Levamisole combinations increased cocaine-appropriate responding for all dose combinations for animal 111F (Figure 23), except for the highest dose combination condition tested (3.2 mg/kg COC with 10 mg/kg LVM). The levamisole combination dose-effect curves were shifted leftward relative to the cocaine alone curve, with 3.2 mg/kg levamisole combination condition producing the greatest effect next to 5.6 and 10 mg/kg, respectively. The 10 mg/kg levamisole combination condition resulted in increased cocaine-appropriate responding up to the

1.8 mg/kg constituent cocaine dose, and then decreased when the cocaine dose was further increased to 3.2 mg/kg. An accompanying decrease in response rate was also observed for this dose combination. No appreciable effects were observed upon the FRF measure for any levamisole combination tests.

Figure 23. Dose-combination dose-effect curves for animal 111F.

Levamisole combinations increased cocaine-appropriate responding for all dose combinations for animal 114F (Figure 24), but only one levamisole combination resulted in full

79 substitution (5.6 mg/kg LVM combination) at the highest constituent cocaine dose tested (3.2 mg/kg). The 5.6 mg/kg levamisole combination dose-effect curve was shifted ¼ log unit leftward relative to the cocaine alone condition, with a previously ineffective cocaine dose (1.8 mg/kg) producing partial substitution (63% cocaine-appropriate responding) when combined with 5.6 mg/kg levamisole. The other two levamisole combinations tested (3.2 and 10 mg/kg LVM) produced only partial substitution (24 and 43% cocaine-appropriate responding, respectively) at the highest constituent cocaine dose tested (3.2 mg/kg). A dose-dependent decrease in response rates occurred for all three levamisole combination conditions, with the largest relative decrease in response rate observed for the 10 mg/kg levamisole combination condition (relative to cocaine alone). No appreciable effects on the FRF measure were observed for any levamisole combination tests.

Figure 24. Dose-combination dose-effect curves for animal 114F.

Animal 115F (Figure 25) displayed a unique cross-generalization profile for levamisole combination tests. Levamisole combinations produced ¾ and ½ log unit leftward shifts in the dose-effect curves for 3.2 and 5.6 mg/kg levamisole combinations, respectively. 10 mg/kg levamisole combination tests consistently produced full substitution in combination with all constituent cocaine doses tested (0.1-1.8 mg/kg). Due to lack of an observed decrease in cocaine- appropriate responding as doses were decreased, 10 mg/kg levamisole was again tested with saline, identical to the levamisole alone condition (15-min PSII). This test also resulted in full 80 substitution (97% cocaine-appropriate responding). A substantial overall decrease in responding occurred for all 10 mg/kg levamisole combination tests, with no appreciable effects on response rate for the other two levamisole combination conditions. The FRF measure was unaffected except for one constituent cocaine dose (3.2 mg/kg) under the 10 mg/kg levamisole combination condition, as well as the 10 mg/kg levamisole and saline test, which showed large increases, past the level of chance responding (i.e., >20 responses).

Figure 25. Dose-combination dose-effect curves for animal 115F.

Discussion

Studies examining the time-course effects of levamisole and its metabolites in humans have been consistent in their findings. Both Hess et al. (2013) and Bertol et al. (2011) detected peak levamisole and aminorex concentrations in human urine at 3 and 6-7 hr, respectively.

Additionally, Spector et al. (1998) have shown that acute levamisole administration in rats can increase levels of dopamine, norepinephrine, and serotonin in several areas of the brain involved in the subjective and reinforcing effects of many drugs of abuse. These two aspects of levamisole administration and metabolism are thought to be responsible for its addition to illicitly manufactured cocaine. However, to date there have been no studies directly examining the subjective or reinforcing effects of concomitant cocaine-levamisole administration, nor have there been any studies on the effects of these metabolic processes in the rat.

81 Levamisole alone failed to substitute for cocaine in the majority of tests, although dose- dependent increases in cocaine-appropriate responding were observed in those animals for which substitution was observed. With the exception of one animal, the levamisole 60-min PSII condition failed to produce a greater degree of cocaine-appropriate responding than did the 15- min PSII condition, and generally failed to produce any substitution (i.e., <20% cocaine- appropriate responding). Cross-generalization profiles for both levamisole conditions (15- and

60-mine PSIIs) were highly idiosyncratic in nature. With the exception of two animals (104M and 115F), levamisole failed to even partially substitute for cocaine.

Several factors contribute to the characteristics of the cross-generalization profiles generated in drug discrimination studies. The role of the training dose in drug discrimination has been extensively documented in the literature (Negus & Banks, 2011; Stadler, Caul, & Barrett,

2011; Stolerman, Childs, Ford, & Grant, 2011), whereby animals trained to discriminate a relatively low dose of training drug generalize to a varying dose range of the training stimulus at lower doses than do their counterparts. Moreover, animals trained using a drug dose that is systematically decreased over time will also generalize to the training stimulus at lower dose than animals whose training dose is not titrated downward. Although all animals in the present study received the same training drug dose (10 mg/kg cocaine or saline), the aspect of the complex stimulus condition which was primarily responsible for its discriminative function was likely idiosyncratic. This is evidenced in the individual differences between animals (see Figure

19) for the cocaine dose-effect curves, and most notably, in the levamisole alone cross- generalization tests. This phenomenon has been demonstrated in the basic animal learning literature, whereby animals have exhibited stimulus control by only one dimension of a compound stimulus (Reynolds, 1961), or demonstrated multiply-controlled discriminative

82 responses (Fink & Patton, 1953). Particularly relevant to the present discussion is Baron’s (1965) summary of the literature:

[E]very stimulus property that happens to be correlated with contingencies of

reinforcement requiring differential responding acquires some degree of stimulus

control; when stimulus properties are presented in combination, their effects combine in

some manner to exert an even greater degree of control. (p. 75)

While the early behavioral literature was focused almost exclusively on the physical dimensions of exteroceptive stimuli, this work has been extended to the use of interoceptive discriminative stimuli. These idiosyncratic within-session response alternation patterns may suggest that the specific stimulus condition that exerts control over behavior for a given rat changes throughout a session. The phenomenon noted by Baron has also been highlighted within the behavioral pharmacology literature (Colpaert et al., 1978b; Gauvin et al., 1989; Gauvin et al., 1995; Gauvin et al., 1996). For example, Gauvin and colleagues have demonstrated that algebraic summation of partial substitution effects, via concomitant administration of drug doses that in isolation produce partial substitution, can in fact produce full substitution via the effect-additivity model of drug interaction.

Differing experimental histories are also capable of influencing generalization test results. However, the present study provided equivalent drug histories across animals, so this is not a factor in the observed idiosyncratic cross-generalization profiles. Due to the nature of the drug discrimination training and testing procedure, it is inevitable that some animals undergo a different number of training sessions than others. Of course, it is possible that a slightly different stimulus condition gained control of responding for animals that took longer to train than others,

83 though there appears to be no discernable pattern with regard to the number of training sessions required and proportion of cocaine-levamisole combinations which produced full substitution.

Behavioral toxicity and direct motoric effects of a drug combination can also alter the outcome of generalization tests. For example, although cocaine-levamisole combinations tests sometimes produced comparable levels of drug-appropriate responding for both 3.2 and 10 mg/kg levamisole combinations, response rate data assist in the interpretation of these comparable outcomes on the percent cocaine-appropriate response measure. For example, 10 mg/kg levamisole with cocaine sometimes produced no interaction, or sometimes weaker interaction than observed with 3.2 mg/kg levamisole with cocaine. This could be due to either a lack of stimulus similarity between this condition and the cocaine administration, or due to a general lack of responding. In some instances (e.g., 106M), upon examining response rates it is apparent the rate of responding under 3.2 mg/kg levamisole with cocaine is comparable to that of cocaine or levamisole alone, while rates for 10 mg/kg levamisole with cocaine display a notable decrease. This is consistent the with 10 mg/kg levamisole alone rate data, and with the notion that this dose, both in isolation and in combination with cocaine, has rate-limiting effects. This could be due to influences upon motivating operations, direct motoric effects, or behavioral toxicity.

In any case, the fact that levamisole often decreased response rates in a drug discrimination paradigm provides support for the hypothesis that the effects observed in the cocaine self-administration study, where levamisole generally reduced self-administration, were in part attributable to nonspecific effects of levamisole on schedule-controlled responding. As noted previously, such nonspecific effects make it difficult to ascertain how, or if, levamisole affected the reinforcing value of cocaine.

84 Drugs that produce effects that are sufficiently dissimilar to those of the training drug evoke responding on the vehicle-appropriate lever. That is, responding defaults to the non-drug lever even when a behaviorally-active dose of a different drug is administered. For example, it has been repeatedly demonstrated that some classes of CNS-active drugs do not cross-generalize to other classes of CNS-active drugs. Alternatively, drugs that produce generalization but do so at a limited range of doses, may reach a ceiling effect with regard to drug-appropriate responding. For example, such responding has been observed with three-choice drug discrimination procedures, whereby percent drug-appropriate responding increases on a given low-dose operanda until the stimulus effects no longer resemble those produced by that training drug dose (i.e., inverted U-shaped dose-effect curve; Jones, Bigelow, & Preston, 1999).

Additionally, sufficiently high doses of a drug that cause non-specific receptor binding, motor impairment, or behavioral toxicity may also result in stimulus effects that are dissimilar from that of the training drug, which may evoke little-to-no responding, or responding on the vehicle- appropriate lever.

The discriminative effects of CNS-active drugs are complex. As such, stimulus control is not dependent upon one single receptor mechanism. While receptor binding characteristics of drugs may show differences in relative affinities for particular receptors, all drugs also demonstrate unbound affinities in in vitro assays. Stimulus control by cocaine, which relies upon multiple neurotransmitter systems, is not unique. Administration of dopamine releasers (e.g., methamphetamine), reuptake inhibitors (e.g., cocaine), and selective dopamine D1 and D2 receptor agonists engender partial to full substitution for the discriminative stimulus effects of cocaine (Baker et al., 1993; Callahan, Appel, & Cunningham, 1991; Costanza, Barber, & Terry,

2001; Filip & Przegaliński, 1997). Alternatively, D1 and D2 receptor antagonists block

85 generalization to cocaine. Callahan et al. (1997) have concluded that dopamine D1 and D2 receptors in the mesocorticolimbic system are involved in modulating the discriminative stimulus effects of psychostimulants, and that that nigrostiriatal dopamine system is not primarily involved. It has also been demonstrated that while serotonin receptor compounds do not substitute for cocaine, several serotonin receptor agonists (e.g., the indole derivative RU

24969, and the arylpiperazines MCPP, MK 212, TFMPP, and quipazine) differentially modulate the discriminative stimulus effects of cocaine (Filip, Bubar, & Cunninham, 2006; Callahan &

Cunningham, 1995). Cocaine itself inhibits reuptake of dopamine, norepinephrine, and serotonin, and the discriminative stimulus effects of cocaine are enhanced by monoamine reuptake inhibitors (Cunningham & Callahan, 1991). Furthermore, it has even been suggested that the histaminergic system may also modify the discriminative stimulus effects of both cocaine and methamphetamine, through H1 and H3 receptors (Mori, Norita, Onodera, & Suzuki, 2002).

Clearly, the neuropharmacological profile of cocaine’s discriminative stimulus effects is complex.

Although partial-substitution was observed repeatedly over the course of this study for all three conditions, within-session response patterns were generally consistent. For example, most animals (7 of 8) consistently demonstrated patterns of drug- to vehicle-lever response shifts under cocaine alone conditions as the sessions progressed. Cocaine-levamisole dose-effect curves were consistently shifted leftward relative to that of cocaine alone curves, but not necessarily in a dose-dependent manner. This is not inconsistent with other types drug interaction studies, which have shown increased effects relative to baseline, but that do not necessarily demonstrate consistent effects (i.e., infra- or supra-additivity) across all combinations (e.g.,

Woolverton et al., 2008).

86 The results of the present study are inconsistent with the hypothesis (Hofmaier et al.,

2014) that aminorex, rather than levamisole, is responsible for the potential interaction between cocaine and use of levamisole as an adulterant. As noted by others (Hess et al., 2013; Bertol et al., 2011), aminorex is not detected until several hours after levamisole administration.

Therefore, it is unlikely that the results obtained in the present study were due to the effects of any behaviorally-active metabolites of levamisole. Furthermore, the results of the levamisole 60- min PSII tests indicated that even less cocaine-appropriate responding occurred when presumably levamisole might begin undergoing metabolic conversion to aminorex. However, characterization of the metabolic profile of levamisole in the rat is lacking, and it is unclear at what point aminorex is present after levamisole administration. The results in the present study were obtained using a 15-min PSII, suggesting that the supra-additive interaction observed is due to the effects of levamisole itself, rather than its metabolites.

No neurochemical measurement or analysis was conducted in the present study. That is, all dependent measures were purely behavioral in nature. One limitation of the present study was the exclusion of any antagonism tests, which could have potentially confirmed or denied the role of dopaminergic activity in the ability of levamisole to substitute for cocaine. For example, pretreatment with selective dopamine receptor antagonists could help elucidate which receptor subtype may be primarily responsible for the discriminative stimulus effects of levamisole.

However, as noted earlier with regard to cocaine, drugs are complex stimuli with several pharmacological mechanisms of action. Thus, several types of antagonist combination tests are warranted to provide a more detailed picture of the neuropharmacological mechanisms underlying the effects if levamisole. It is also worth noting that complete generalization alone does not definitively prove that there is a comparable mechanism of action between two drugs.

87 Rather, that the two drugs produce a similar stimulus effect (Glennon & Young, 2011). The present results demonstrate that although levamisole alone may fail to reliably produce cocaine- like effects in all animals, the interaction of these two compounds is generally additive, and more commonly supra-additive. These results, in combination with that of the aforementioned self- administration study, provide further evidence in support of the popular notion that levamisole is added to cocaine in an attempt to potentiate or alter the reinforcing or subjective effects of cocaine itself.

88 CHAPTER IV GENERAL DISCUSSION Like many drugs, levamisole has the potential to have a profound impact upon behavior.

It remains unclear as to why drug manufacturers add levamisole to illicit cocaine. The two theories as to why cocaine may be adulterated with levamisole deal with its physical similarity to cocaine and the potential interaction with the effects of cocaine. Due to the reasons detailed above (see Chapter 1), it would appear that the former is unlikely. The experiments described herein clearly demonstrate that not only does levamisole produce measurable behavioral effects in Sprague-Dawley rats, but that a supra-additive interaction with cocaine exists with regard to these effects. Whether this dose additivity holds true with regard to all of cocaine’s effects in humans needs to be investigated. It is clear that levamisole can have a negative impact upon the physical condition of the human body, but the behavioral effects in humans remain to be examined. Many experiments related to levamisole-adulterated cocaine cite the potentiation of cocaine’s effects in humans as one of the main motives for its addition, but this has never been reported in the literature and has yet to be examined experimentally.

The studies detailed above demonstrate that the behavioral effects of cocaine in rats can be amplified when administered with levamisole, but the neuropharmacological and gross physiological effects of their concomitant administration was not examined. The existing literature demonstrates that other aspects of animal behavior (i.e., locomotor activity, CPP) measured following concomitant cocaine-levamisole administration sometimes conform to a supra-additive model of drug interaction. Thus, taken together with the results of the present studies, it appears that this type of interaction extends to several aspects of behavior. Although the assays used in the present experiments have demonstrated high degrees of validity, the behavioral effects of cocaine-levamisole combinations in humans have yet to be directly

89 reported. Therefore, it is unclear as to whether this type of interaction effect extends to human behavior. Taken together, the results of these two studies enable us to provide some meaningful statements regarding the behavioral effects of cocaine-levamisole combinations. Unfortunately, the reasons for the addition of levamisole to illicit cocaine are still largely unclear, and much research remains to be conducted to detect and prevent the untoward effects of this increasingly widespread drug combination.

90 REFERENCES

Abdul-Karim, R., Ryan, C., Rangel, C., & Emmett, M. (2013). Levamisole-induced vasculitis.

Baylor University Medical Center Proceedings, 26(2), 163-165.

Adams, J. G. (1978). Pharmacokinetics of levamisole. Journal of Rheumatology, 4, 137-142.

Ahmed S. H., & Koob, G. F. (2004). Vertical shifts in dose-injection curves reflect reward

allostasis, not sensitization. Psychopharmacology, 171(3), 354-355.

Amery, W. K., & Bruynseels, J. P. (1992). Levamisole, the story and the lessons. International

Journal of Immunopharmacology, 14(3), 481-486.

Amsel, A. (1967). Partial reinforcement effects on vigor and persistence: advances in frustration

theory derived from a variety of within-subject experiments. In K. W. Spence & J. T.

Spence (Eds.), The psychology of learning and motivation (pp. 1-65). New York:

Academic Press.

Ator, N. A., & Griffiths, R. R. (2003). Principles of drug abuse liability assessment in laboratory

animals. Drug and Alcohol Dependence, 70, S55-S72.

Baker, L. E., Riddle, E. E., Saunders, R. B., & Appel, J. B. (1993). The role of monoamine

uptake in the discriminative stimulus effects of cocaine and related compounds.

Behavioural Pharmacology, 4(1), 69-79.

Baldi, E., & Bucherelli, C. (2005). The inverted “u-shaped” dose-effect relations in learning and

memory: Modulation of arousal and consolidation. Nonlinearity in Biology, Toxicology,

Medicine, 3(1), 9-21.

91 Balster, R. L., & Lukas, S. (1985). Review of self-administration. Drug and Alcohol

Dependence, 14(3-4), 249-261.

Balster, R. L., & Schuster, C. R. (1973). Fixed-interval schedule of cocaine reinforcement: Effect

of dose and infusion duration. Journal of the Experimental Analysis of Behavior, 20(1),

119-129.

Banks, M. L. (2014). Effects of the nicotinic acetylcholine receptor antagonist mecamylamine on

the discriminative stimulus effects of cocaine in male rhesus monkeys. Experimental and

Clinical Psychopharmacology, 22(3), 266-273.

Barbieri, E. J., Ferko, A. P., DiGregorio, G. J., & Ruch, E. K. (1992). The presence of cocaine

and benzoylecgonine in rat cerebrospinal fluid after the intravenous administration of

cocaine. Life Sciences, 51(22), 1739-1746.

Baron, M. R. (1965). Metric properties of multidimensional stimulus generalization. In D. I.

Mostofsky (Ed.), Stimulus generalization (pp. 72-93). Stanford, CA: Stanford University

Press.

Bauman, J. L., & DiDomenico, R. J. (2002). Cocaine-induced channelopathies: Emerging

evidence on the multiple mechanisms of sudden death. Journal of Cardiovascular

Pharmacology and Therapeutics, 7(3), 195-202.

Bertalmio, A. J., & Woods, J. H. (1987). Differentiation between mu and kappa receptor-

mediated effects in opioid drug discrimination: Apparent pA2 analysis. Journal of

Pharmacology and Experimental Therapeutics, 243(2), 591-597.

92 Bertol, E., Mari, F., Milia, M. G., Politi, L., Furlanetto, S., & Karch, S. B. (2011). Determination

of aminorex in human urine samples by GC-MS after use of levamisole. Journal of

Pharmaceutical and Biomedical Analysis, 55(5), 1186-1189.

Buchanan, J. A., Vogel, J. A., & Eberhardt, A. M. (2010). Levamisole-induced occlusive

necrotizing vasculitis of the ears after use of cocaine contaminated with levamisole.

Journal of Medical Toxicology, 7(1), 83-84.

Calabrese, E. J., & Baldwin, L. A. (1999). The marginalization of hormesis. Toxicologic

Pathology, 27(2), 187-194.

Calabrese, E. J, & Baldwin, L. A. (2000). Chemical hormesis: its historical foundations as a

biological hypothesis. Human & Experimental Toxicology, 19(1), 2-31.

Calabrese, E. J., & Baldwin LA (2001a). The frequency of U-shaped dose responses in the

toxicological literature. Toxicological Sciences, 62(2), 330-338.

Calabrese, E. J., & Baldwin, L. A. (2001b). Hormesis: U-shaped dose responses and their

centrality in toxicology. Trends in Pharmacological Sciences, 22(6), 285-291.

Caldwell, K. B., Graham, O. Z., & Arnold, J. J. (2012). Agranulocytosis from levamisole-

adulterated cocaine. Journal of the American Board of Family Medicine, 25(4), 528-530.

Calipari, E. S., Ferris, M. J., & Jones, S. J. (2014). Extended access cocaine self-administration

results in tolerance to the dopamine-elevating and locomotor-stimulating effects of

cocaine. Journal of Neurochemistry, 128(2), 224-232.

93 Callahan, P. M., Appel, J. B., Cunningham, K. A. (1991). Dopamine D1 and D2 mediation of the

discriminative stimulus properties of d-amphetamine and cocaine. Psychopharmacology,

103(1), 50-55.

Callahan, P. M., & Cunningham, K. A. (1995). Modulation of the discriminative stimulus

properties of cocaine by 5-HT1B and 5-HT2C receptors. Journal of Pharmacology and

Experimental Therapeutics, 274(3), 1414-1424.

Callahan, P. M., de La Garza, R., II, & Cunningham, K. A. (1997). Mediation of the

discriminative stimulus properties of cocaine by mesocorticolimbic dopamine systems.

Pharmacology, Biochemistry & Behavior, 57(3), 601-607.

Carrera, M. R., Meijler, M. M., & Janda, K. D. (2004). Cocaine pharmacology and current

pharmacotherapies for its abuse. Bioorganic & Medicinal Chemistry, 12(19), 5019-5030.

Carroll, M. E., & Anker, J. J. (2010). Sex differences and ovarian hormones in animal models of

drug dependence. Hormones and Behavior, 58(1), 44-56.

Casale, J. F., Corbeil, E. M., & Hays, P. A. (2008). Identification of levamisole impurities found

in illicit cocaine exhibits. Microgram Journal, 6, 82-89.

Chamsi-Pasha, H. (2001). Sildenafil (Viagra) and the heart. Journal of Family & Community

Medicine, 8(2), 63-66.

Chiadmi, F., Lyer, A., Cisternino, S., Toledano, A., Schlatter, J., Ratiney, R., & Fontan, J. E.

(2004). Stability of levamisole oral solutions prepared from tablets and powder. Journal

of Pharmacy and Pharmaceutical Sciences, 8(2), 322-325.

94 Chitwood, D. D. (1985). Patterns and consequences of cocaine use. In N. J. Kozel & E. H.

Adams (Eds.), Cocaine use in America: Epidemiologic and clinical perspectives (pp.

111-129). Rockville, MD: U.S. Department of Health and Human Services.

Cole, C., Jones, L., McVeigh, J., Kicman, A., Syed, Q., & Bellis, M. (2010). Adulterants in illicit

drugs: A review of empirical evidence. Drug Testing and Analysis, 3(2), 89-96.

Colpaert, F. C. (1982). The pharmacological specificity of opiate drug discrimination. In F. C.

Colpaert & J. L. Slangen (Eds.), Drug discrimination: Applications in CNS

pharmacology, (pp. 3-16). Amsterdam: Elsevier Biomedical Press.

Colpaert, F. C. (1987). Drug discrimination: Methods of manipulation, measurement, and

analysis. In M. A. Bozarth (Ed.), Methods of assessing the reinforcing properties of

abused drugs (pp. 341-372). New York: Springer-Verlag.

Colpaert, F. C., Niemegeers, C. J. E., & Janssen, P. A. J. (1978a). Discriminative stimulus

properties of cocaine and d-amphetamine, and antagonism by haloperidol: A comparative

study. Neuropharmacology, 17, 937-942.

Colpaert, F. C., Niemegeers, C. J. E., & Janssen, P. A. J. (1978b). Factors regulating drug cue

sensitivity. A long-term study on the cocaine cue. In F. C. Colpaert & J. A. Rosecrans

(Eds.), Stimulus properties of drugs: Ten years of progress (pp. 281-299). Amsterdam:

Elsevier/North-Holland Biomedical Press.

Controlled Substances Staff (CSS), Center for Drug Evaluation and Research (CDER). U. S.

Department of Health and Human Services, Food and Drug Administration (2017).

Assessment of abuse potential of drugs: Guidance for industry. Available at:

https://www.fda.gov/Drugs/GuidanceComplianceRegulatoryInformation/Guidances/UC

M198650.

95 Cooper, J. R., Bloom, F. E., & Roth, R. H. (2003). The biochemical basis of neuropharmacology

(8th ed.). New York: Oxford University Press.

Costanza, R. M., Barber, D. J., & Terry, P. (2001). Antagonism of the discriminative stimulus

effects of cocaine at two training doses by dopamine D2-like receptor antagonists.

Psychopharmacology, 158(2), 146-153.

Craft, R. M., Kalivas, P. W., & Stratmann, J. A. (1996). Sex differences in discriminative

stimulus effects of morphine in the rat. Behavioural Pharmacology, 7(8), 764-778.

Craft, R. M., & Stratmann, J. A. (1996). Discriminative stimulus effects of cocaine in female

versus male rats. Drug and Alcohol Dependence, 42(1), 27-37.

Crombag, H. S., Ferrario, C. R., & Robinson, T. E. (2008). The rate of intravenous cocaine or

amphetamine delivery does not influence drug-taking and drug-seeking behavior in rats.

Pharmacology, Biochemistry & Behavior, 90(4), 797-804.

Cunningham, K. A., & Callahan, P. M. (1991). Monoamine reuptake inhibitors enhance the

discriminative state induced by cocaine in the rat. Pyschopharmacology, 104(2), 177-180.

Czuchlewski, D. R., Brackney, M., Ewers, C., Manna, J., Fekrazad, M. H., Martinez, A., . . .

Foucar, K. (2010). Clinicopathologic features of agranulocytosis in the setting of

levamisole-tainted cocaine. American Journal of Clinical Pathology, 133(3), 466-472.

DanceSafe (2016). Drug checking. Retrieved from: https://dancesafe.org/drug-checking/

DeCorte, T. (2001). Quality control by cocaine users: Underdeveloped harm reduction strategies.

European Addiction Research, 7, 161-175.

96 Do, T., Sun, Q., Beuve, A., & Kuzhikandathil, E. V. (2007). Extracellular cAMP inhibits D1

dopamine receptor expression in CAD catecholaminergic cells via A2a adenosine

receptors. Journal of Neurochemistry, 101, 619-631.

Drug Enforcement Administration (DEA) (2013). Levamisole (Ergamisol®). Drug and Chemical

Evaluation Section.

Drug Enforcement Administration (DEA) (2015). Established aggregate production quotas for

schedule I and II controlled substances and assessment of annual needs for the list I

chemicals ephedrine, pseudoephedrine, and phenylpropanolamine for 2016. Federal

Emmett-Oglesby, M. W., Peltier, R. L., Depoortere, R. Y., Pickering, C. L., Hooper, M. L.,

Gong, Y. H., & Lane, J. D. (1993). Tolerance to self-administration of cocaine in rats:

Time course and dose-response determination using a multi-dose method. Drug and

Alcohol Dependence, 32(3), 247-256.

Engelmann, G. L., & Richardson, A. G. (1986). Effects of levamisole on primary cultures of

adult rat hepatocytes. Biochemical Pharmacology, 9(1), 1547-1554.

European Medicines Agency (EMeA) (2006). Guideline on the non-clinical investigation of the

dependence potential of medicinal products.

Feltenstein, M. W., & See, R. E. (2008). The neurocircuitry of addiction: An overview. British

Journal of Pharmacology, 154(2), 261-274.

Filip, M., Bubar, M. J., & Cunningham, K. A. (2006). Contribution of serotonin (5-HT) 5-HT2

receptor subtypes to the discriminative stimulus effects of cocaine in rats.

Psychopharmacology, 183(4), 482-9.

97 Filip, M., & Przegaliński, E. (1997). The role of dopamine receptor subtypes in the

discriminative stimulus effects of amphetamine and cocaine in rats. Polish Journal of

Pharmacology, 49(1), 21-30.

Fink, J. B., & Patton, R. M. (1953). Decrement of a learned drinking response accompanying

changes in several stimulus characteristics. Journal of Comparative Physiology and

Psychology, 46(1), 23-27.

Fischman, M. W., (1984). Behavioral pharmacology of cocaine in humans. NIDA Research

Monograph, 50, 72-91.

Fischman, M. W., & Foltin. R. W. (1991). Utility of subjective-effects measurements in

assessing abuse liability in humans. British Journal of Addiction, 86(12), 1563-1570.

Follath, F., Burkart, F., & Schweizer, W. (1971). Drug-induced pulmonary hypertension? British

Medical Journal, 1(5743), 265-266.

Formeister, E. J., Falcone, M. T., & Mair, E. A. (2015). Facial cutaneous necrosis associated

with suspected levamisole toxicity from tainted cocaine abuse. Annals of Otology,

Rhinology, & Laryngology, 124(1), 30-34.

Galtier, P., Coche, Y., & Alvinerie, M. (1983). Tissue distribution and elimination of

[3H]levamisole in the rat after oral and intramuscular administration. Xenobiotica, 13(7),

407-413.

Gauvin, D. V., & Baird, T. J. (1999). Discriminative effects of compounds drug stimuli: A focus

on attention. Pharmacology, Biochemistry & Behavior, 64(2), 229-235.

98 Gauvin, D. V., Briscoe, R. J., Baird, T. J., Vallett, M., Carl, K. L., & Holloway, F. A. (1996).

Selective attention to elements of a morphine stimulus: Idiosyncratic stimulus additivity

of incidental cues. Experimental and Clinical Psychopharmacology, 4(4), 379-388.

Gauvin, D. V., Carl, K. L., Briscoe, R. J., Vallett, M., & Holloway, F. A. (1995). Cross-

generalization between a cocaine cue and two over-the-counter antihistamines. European

Journal of Pharmacology, 294(1), 281-288.

Gauvin, D. V., Criado, J. R., Moore, K. R., & Holloway, F. A. (1990). Potentiation of cocaine’s

discriminative effects by caffeine: A time-effect analysis. Pharmacology, Biochemistry &

Behavior, 36(1), 195-197.

Gauvin, D. V., Dalton, J. A., & Baird, T. J. (2017). Current strategies for abuse liability

assessment of new chemical entities. In A. S. Faqi (Ed.), A comprehensive guide to

toxicology in preclinical drug development (2nd ed.), (pp. 293-321). London, UK:

Elsevier.

Gauvin, D. V., Guha, M., & Baird, T. J. (2015). Rat self-administration. In C. G., Margraf, T. J.,

Hudzik, & D. R. Compton (Eds.), Nonclinical assessment of abuse potential for new

pharmaceuticals (pp. 49-80). London, UK: Elsevier, Inc.

Gauvin, D. V., Harland, R. D., Michaelis, R. C., & Holloway, F. A. (1989). Caffeine-

phenylethylamine combinations mimic the cocaine discriminative cue. Life Sciences,

44(1), 67-73.

Gauvin, D. V., & Zimmermann, Z. J. (2017). Conducting preclinical abuse liability screening in

only one sex: Making a case for “reasonable exclusion”. Pharmaceutical Regulatory

Affairs, 6(1), 1-16. 99 Gessner, P. K., & Cabana, B. E. (1970). A study of the interaction of the hypnotic effects of the

toxic effects of chloral hydrate and ethanol. Journal of Pharmacology and Experimental

Therapeutics, 174(2), 247-259.

Glennon, R. A., Gabryszuk, M., & Glennon, R. A. (1998). (-)Ephedrine and caffeine mutually

potentiate one another’s amphetamine-like stimulus effects. Pharmacology, Biochemistry

& Behavior, 61(2), 169-173.

Glennon, R. A., & Misenheimer, B. (1990). Stimulus properties of a new designer drug: 4-

methylaminorex (“U4Euh”). Pharmacology, Biochemistry & Behavior, 35(3), 517-521.

Glennon, R. A., & Young, R. (2011). Drug discrimination: Practical considerations. In R. A.

Glennon & R. Young (Eds.), Drug discrimination: Applications to medicinal chemistry

and drug studies (pp. 41-128). Hoboken, NJ: John Wiley & Sons, Inc.

Gold, L. H., & Balster, R. L. (1996). Evaluation of the cocaine-like discriminative stimulus

effects and reinforcing effects of modafinil. Psychopharmacology, 126(4), 286-292.

Goldberg, S. R. (1973). Comparable behavior maintained under fixed-ratio and second-order

schedules of food presentation, cocaine injection, or a d-amphetamine injection in the

squirrel monkey. Journal of Pharmacology and Experimental Therapeutics, 186(1), 18-

30.

Goldberg, S. R., Woods, J. H., & Schuster, C. R. (1969). Morphine: Conditioned increases in

self-administration in rhesus monkeys. Science, 166(3910), 1306-1307.

100 Harland, R. D., Gauvin, D. V., Michaelis, R. C., Carney, J. M., Seale, T. W., & Holloway, F. A.

(1989). Behavioral interaction between cocaine and caffeine: A drug discrimination

analysis in rats. Pharmacology, Biochemistry & Behavior, 32(4), 1017-1023.

Herrnstein, R. J., & Morse, W. H. (1957). Effects of pentobarbital on intermittently reinforced

behavior. Science, 125(3254), 929-931.

Hess, C., Ritke, N., Broecker, S., Madea, B., & Musshoff, F. (2013). Metabolism of levamisole

and kinetics of levamisole and aminorex in urine by means of LC-QTOF-HRMS and LC-

QqQ-MS. Analytical and Bioanalytical Chemistry, 405(12), 4077-4088.

Hodinka, L., Géher, P., Merétey, K., Gyódi, E. K., Petrányi, G. G., & & Bozsóky, S. (1981).

Levamisole-induced neutropenia and agranulocytosis: Association with HLA B27

leukocyte agglutinating and lymphocytotoxic antibodies. International Archives of

Allergy and Applied Immunology, 65(4), 460-464.

Hofmaier, T., Luf, A., Seddik, A., Stockner, T., Holy, M., Freissmuth, M., . . . Kudlacek, O.

(2014). Aminorex, a metabolite of the cocaine adulterant levamisole, exerts amphetamine

like actions at monoamine transporters. Neurochemistry International, 73, 32-41.

Horton, D. B., Potter, D. M., & Mead, A. N. (2013). A translational pharmacology approach to

understanding the predictive value of abuse potential assessments. Behavioural

Pharmacology, 24(5-6), 410-436.

Hubner, C. B., & Koob, G. F. (1990). Bromocriptine produces decreases in cocaine self-

administration in the rat. Neuropsychopharmacology, 3(2), 101-108.

101 ICH M3(R2) (2009). Guidance on Non-clinical Safety Studies for the Conduct of Human

Clinical Trials and Marketing Authorization for Pharmaceuticals. 2009 June 11 (ICH

Implementation Date).

James, K. T., Detz, A., Coralic, Z., & Kanzaria, H. (2013). Levamsiole contaminated cocaine

induced cutaneous vasculitis syndrome. Western Journal of Emergency Medicine, 14(5),

448-449.

Janssen, P. A. (1976). The levamisole story. Progress in Drug Research, 20, 347-383.

Jiménez-González A., Ros-Moreno R. M., Moreno-Guzmán M. J., Rodríguez-Caabeiro F.

(1999). Characterization of levamisole binding sites Trichinella spiralis. Parasitology

Research, 84, 747-759.

Jones, H. E., Bigelow, G. E., & Preston, K. L. (1999). Assessment of opioid partial agonist

activity with a three-choice hydromorphone drug discrimination procedure. Journal of

Pharmacology and Experimental Therapeutics, 289(3), 1350-1361.

Kantor, E. D., Rehm, C. D., Haas, J. S., Chan, A. T., & Giovannucci, E. L. (2015). Trends in

prescription drug use among adults in the United States from 19999-2012. JAMA,

314(17), 1818-1830.

Kaplan, H. L., Jain, N. C., Forney, R. B., & Richards, A. B. (1969). Chloral hydrate-ethanol

interactions in the mouse and dog. Toxicology and Applied Pharmacology, 14(1), 127-

137.

Karch, S. B., Defraia, B., Messerini, L., Mari, F., Vaiano, F., & Berto, E. (2014). Aminorex

associated with possible idiopathic pulmonary hypertension in a cocaine user. Forensic

Science International, 240, 7-10.

102 Kato, S., Wakasa, Y., & Yanagita, T. (1987). Relationship between minimum reinforcing doses

and injection speed in cocaine and pentobarbital self-administration in crab-eating

monkeys. Pharmacology, Biochemistry & Behavior, 28(3), 407-410.

Katz J. L., & Higgins S. T. (2004). What is represented by vertical shifts in self-administration

dose-response curves? Psychopharmacology, 171, 360-361.

Knowles, L., Buxton, J. A., Skuridina, N., Achebe, I., Legatt, D., Fan, S., . . . Talbot, J. (2009).

Levamisole tainted cocaine causing severe neutropenia in Alberta and British Columbia.

Harm Reduction Journal, 6, 30.

Kouassi, E., Caillé, G., Léry, L., Larivière, L., & Vézina M. (1986). Novel assay and

pharmacokinetics of levamisole and p-hydroxylevamisole in human plasma and urine.

Biopharmaceutics and Drug Disposition, 7(1), 71-89.

Lamas, X., Negus, S. S., Gatch, M. B., & Mello, N. K. (1998). Effects of heroin/cocaine

combinations in rats trained to discriminate heroin or cocaine from saline. Pharmacology,

Biochemistry & Behavior, 60(2), 357-364.

Lattal, K. A. (2004). Steps and pips in the history of the cumulative recorder. Journal of the

Experimental Analysis of Behavior, 82(3), 329-355.

Lelas, S., Rowlett, J. K., & Spealman, R. D. (2001). Isobolographic analysis of chlordiazepoxide

and triazolam combinations in squirrel monkeys discriminating triazolam.

Psychopharmacology, 158(2), 181-189.

Li, J-X., Gerak, L. R., & France, C. P. (2011). Drug discrimination studies in rhesus monkeys:

Drug dependence and withdrawal. In R. A. Glennon & R. Young (Eds.), Drug

103 discrimination: Applications to medicinal chemistry and drug studies (pp. 417-430).

Hoboken, NJ: John Wiley & Sons, Inc.

Li, J-X., Koek, W., Rice, K. C., & France, C. P. (2010). Differential effects of serotonin 5-HT1A

receptor agonists on the discriminative stimulus effects of the 5-HT2A receptor agonist 1-

(2,5-dimethoxy-4-methylphenyl)-2-aminopropane in rats and rhesus monkeys. Journal of

Pharmacology and Experimental Therapeutics, 333(1), 244-252.

Liu, Y., Roberts, D. C., & Morgan, D. (2005). Sensitization of the reinforcing effects of self-

administered cocaine in rats: Effects of dose and intravenous injection speed. European

Journal of Neuroscience, 22(1), 195-200.

Loewe, S. (1953). The problem of synergism and antagonism of combined drugs.

Arzneimittelforschung, 3, 285-290.

Lotfizadeh, A. D., Zimmermann, Z. J., Watkins, E. E., Edwards, T. L., & Poling, A. (2014).

Increasing food deprivation relative to baseline influences d-amphetamine dose-response

gradients. Pharmacology, Biochemistry & Behavior, 125, 65-69.

Luyckx, M., Rousseau, F., Cazin, M., Brunet, C., Cazin, J. C., Haguenoer, J. M., . . . Demaille,

A. (1982). Pharmacokinetics of levamisole in healthy subjects and cancer patients.

European Journal of Drug Metabolism and Pharmacokinetics, 7(4), 247-254.

Martin, R. J., & Robertson, A. P. (2010). Control of nematode parasites with agents acting on

neuro-musculature systems: Lessons for neuropeptide ligand discovery. Advances in

Experimental Medicine and Biology, 692, 138-154.

104 Massey, B. W., & Woolverton, W. L. (1994). Discriminative stimulus effects of combinations of

pentobarbital and ethanol in rhesus monkeys. Drug and Alcohol Dependence, 35(1), 37-

43.

McCullough, L. D., de Vries, G. J., Miller, V. M., Becker, J. B., Sandberg, K., & McCarthy, M.

M. (2014). NIH initiative to balance sex of animals in preclinical studies: Generative

questions to guide policy, implementation, and metrics. Biology of Sex Differences, 5, 15.

Mello, N. K., Mendelson, J. H., Bree, M. P., & Lukas, S. E. (1990). Buprenorphine and

naltrexone effects on cocaine self-administration by rhesus monkeys. Journal of

Pharmacology and Experimental Therapeutics, 254(3), 926-939.

Mello, N. K., & Negus, S. S. (1996). Preclinical evaluation of pharmacotherapies for treatment

of cocaine and opioid abuse using drug self-administration procedures.

Neuropsychopharmacology, 14(6), 375-424.

Mello, N. K., & Newman, J. L. (2011). Discriminative and reinforcing stimulus effects of

nicotine, cocaine, and cocaine + nicotine combinations in rhesus monkeys. Experimental

and Clinical Psychopharmacology, 19(3), 203-214.

Michaud, K., Grabherr, S., Shiferaw, K., Doenz, F., Augsburger, M., & Mangin, P. (2014).

Acute coronary syndrome after levamisole-adulterated cocaine abuse. Journal of

Forensic and Legal Medicine, 21, 48-52.

Mlczoch, J., Probst, P., Szeless, S., & Kaindl, F. (1980). Primary pulmonary hypertension:

Follow-up of patients with and without anorectic drug intake. Cor et Vasa, 22(4), 251-

257.

105 Morse, W. H. (1966). Intermittent reinforcement. In W. K. Honig (Ed.) Operant behavior: Areas

of research and application (pp. 52-108). Englewood Cliffs, NJ: Prentice-Hall, Inc.

Muirhead, T. T., & Eide, M. J. (2011). Toxic effects of levamisole in a cocaine user. The New

England Journal of Medicine, 364(24), e52.

National Drug Intelligence Center (2011). Economic impact of illicit drug use on American

society. Washington D.C.: United States Department of Justice.

National Institute on Drug Abuse (NIDA) (2016). Research programs: The addiction treatment

discovery program. Retrieved from: https://www.drugabuse.gov/about-

nida/organization/divisions/division-pharmacotherapies-medical-consequences-drug-

abuse-dpmcda/research-programs

National Organization for Rare Disorders (NORD) (2015). Cyclic neutropenia. Retrieved from:

https://rarediseases.org/rare-diseases/cyclic-neutropenia/

National Research Council (2011). Guide for the care and use of laboratory animals (8th ed.),

Washington, D. C.: National Academies Press.

Negus, S. S., & Banks, M. L. (2011). Making the right choice: Lessons from drug discrimination

for research on drug reinforcement and drug self-administration. In R. A. Glennon & R.

Young (Eds.), Drug discrimination: Applications to medicinal chemistry and drug studies

(pp. 361-388). Hoboken, NJ: John Wiley & Sons, Inc.

Nowak, L. G., Rosay, B., Czégé, D., & Fonta, C. (2015). Tetramisole and levamisole suppress

neuronal activity independently from their inhibitory action on tissue non-specific

alkaline phosphatase in mouse cortex. Sub-cellular Biochemistry, 76, 239-281.

106 Onuagulchi, G., & Igbo, I. N. A. (1990). Electrocardiographic changes induced by levamisole

hydrochloride in the rat. Archives Internationales de Pharmacodynamie et de Thérapie,

305, 55-62.

Overmann, S. R., & Denny, M. R. (1974). The free-operant partial reinforcement effect: A

discrimination analysis. Learning and Motivation, 5(2), 248-257.

Panlilio, L. V., Goldberg, S. R., Gilman, J. P., Jufer, R., Cone, E. J., & Schindler, C. W. (1998).

Effects of delivery rate and non-contingent infusion of cocaine on cocaine self-

administration in rhesus monkeys. Psychopharmacology, 137(3), 253-258.

Panlilio, L. V., Yasar, S., Nemeth-Coslett, R., Katz, J. L., Henningfield, J. E., Solinas, M., . . .

Goldberg, S. R. (2005). Human cocaine-seeking behavior and its control by drug-

associated stimuli in the laboratory. Neuropsychopharmacology, 30, 433-443.

Parrott, A. C. (2004). Is ecstasy MDMA? A review of the proportion of ecstasy tablets

containing MDMA, their dosage levels, and the changing perceptions of purity.

Psychopharmacology, 173(3-4), 234-241.

Piazza, P. V., Deroche-Gamonent, V., Rouge-Pont, F., & Le Moal, M. (2000). Vertical shifts in

self-administration dose-response functions predict a drug-vulnerable phenotype

predisposed to addiction. Journal of Neuroscience, 20(11), 4226-4232.

Pickens, R., Dougherty, J. A., & Thompson, T. (1969). Effects of volume and duration of

infusion on cocaine reinforcement, with concurrent activity recording. In: Committee on

Problems of Drug Dependence (pp. 5805-11). Washington, DC: National Academy of

Science, National Research Council.

107 Poling, A. (1979). The ubiquity of the cumulative record: A quote from Skinner and a frequency

count. Journal of the Experimental Analysis of Behavior, 31(1), 126.

Pulvirenti, L., Balducci, C., & Koob, G. F. (1997). Dextromethorphan reduces intravenous

cocaine self-administration in the rat. European Journal of Pharmacology, 321(3), 279-

283.

Reynolds, G. S. (1961). Attention in the pigeon. Journal of the Experimental Analysis of

Behavior, 4(3), 203-208.

Roberts, D. C. S., Loh, E. A., & Vickers, G. (1989). Self-administration of cocaine on a

progressive ratio schedule in rats: Dose-response relationship and effect of haloperidol

pretreatment. Psychopharmacology, 97(4), 535-538.

Robertson, A. P., Bjorn, H. E., & Martin, R. J. (1999). Resistance to levamisole resolved at the

single-channel level. FASEB Journal, 13(6), 749-760.

Rowlett, J. K., Spealman, R. D., & Platt, D. M. (2001). Similar enhancement of the

discriminative stimulus effects of cocaine and GBR 12909 by heroin in squirrel monkeys.

Psychopharmacology, 157(3), 313-319.

Saady, J. J., Bowman, E. R., Aceto, M. D. (1995). Cocaine, ecgonine methyl ester, and

benzoylecgnonine plasma profiles in rhesus monkeys. Journal of Analytical Toxicology,

19(7), 571-575.

Sannerud, C. A., & Griffiths, R. R. (1993). Tolerance to the discriminative stimulus effects of

midazolam: Evidence for environmental modification and dose fading. Behavioural

Pharmacology, 4(2), 125-133.

108 Sannerud, C. A., & Young, A. M. (1987). Environmental modification of tolerance to morphine

discriminative stimulus properties in rats. Psychopharmacology, 93(1), 59-68.

Schmidt, D., Roznoski, M. L., & Ebert, M. H. (1990). Quantitative and qualitative HPLC

analysis of monoamine neurotransmitters and metabolites in CSF and brain tissues using

reductive electrochemical detection. Biomedical Chromatography, 4(5), 215-220.

Sherlock, K., Wolff, K., Hay, A. W., & Conner, M. (1999). Analysis of illicit ecstasy tablets:

Implications for clinical management in the accident and emergency department. Journal

of Accident & Emergency Medicine, 16(3), 194-197.

Signs, S. A., & Schechter, M. D. (1986). Nicotine-induced potentiation of ethanol

discrimination. Pharmacology, Biochemistry & Behavior, 24(3), 769-771.

Spealman, R. D. (1995). Noradrenergic involvement in the discriminative stimulus effects of

cocaine in squirrel monkeys. Journal of Pharmacology and Experimental Therapeutics,

275(1), 53-62

Spector, S., Munjal, I., & Schmidt, D. E. (1998). Effects of the immunostimulant, levamisole, on

opiate withdrawal and levels of endogenous opiate alkaloids and monoamine

neurotransmitters in rat brain. Neuropsychopharmacology, 19(5), 417-427.

Stadler, J. R., Caul, W. F., & Barrett, R. J. (2001). Effects of training dose on amphetamine drug

discrimination: Dose-response functions and generalization to cocaine. Pharmacology,

Biochemistry & Behavior, 70(2-3), 381-386.

Stolberg, V. B. (2011). The use of coca: Prehistory, history, and ethnography. Journal of

Ethnicity in Substance Abuse, 10(2), 126-146.

109 Stolerman, I. P. (1991). Measures of stimulus generalization in drug discrimination experiments.

Behavioural Pharmacology, 2(4-5), 265-282.

Stolerman, I. P., Childs, E., Ford, M. M., & Grant, K. A. (2011). Role of training dose in drug

discrimination: A review. Behavioural Pharmacology, 22(5-6), 15-29.

Symoens, J., Veys, E., Mielants, M., & Pinals, R. (1978). Adverse reactions to levamisole.

Cancer Treatment Reports, 62(11), 1721-1730.

Tallarida, R. J. (2000). Drug synergism and dose-effect data analysis. Boca Raton FL: Chapman

& Hall/CRC.

Tallarida, C. S., Egan, E., Alejo, G. D., Raffa, R., Tallarida, R. J., & Rawls, S. M. (2014).

Levamisole and cocaine synergism: A prevalent adulterant enhances cocaine’s action in

vivo. Neuropharmacology, 79, 590-595.

Tallarida, R. J., Porreca, F., & Cowan, A. (1989). Statistical analysis of drug-drug and site-site

interactions with isobolograms. Life Sciences, 45(11), 947-961.

Tallarida, C. S., Tallarida, R. J., & Rawls, S. M. (2015). Levamisole enhances the rewarding and

locomotor-activating effects of cocaine in rats. Drug and Alcohol Dependence, 149(1),

145-150.

Thompson, J. S., Herbick, J. M., Klassen, L. W., Severson, C. D., Overline, V. L., Blaschke, J.

W., . . . Vogel, C. L. (1980). Studies on levamisole-induced agranulocytosis. Blood,

56(3), 388-396.

Thomsen, M. (2014). Locomotor activating effect of cocaine and scopolamine combinations in

rats: Isobolographic analysis. Behavioural Pharmacology, 25(4), 259-266. 110 Tichauer, M., Fagan, S., & Goverman, J. (2014). Levamisole-induced ANCA vasculitis and

cutaneous necrosis. Eplasty, 14, ic39.

United Nations (1988). Commission on narcotic drugs. United nations convention against illicit

trade in narcotic drugs and psychotropic substances. Retrieved from:

https://www.unodc.org/pdf/convention_1988_en.pdf

United Nations Office on Drugs and Crime (UNODC) (2015). World drug report. New York:

United Nations publication, Sales No. E.15.XI.6.

United States Department of Justice (USDOJ) (2002). Federal cocaine offenses: An analysis of

crack and powder penalties. Retrieved from:

https://www.justice.gov/archive/olp/pdf/crack_powder2002.pdf

VICE (2014). The terrifying substances people put in cocaine. Retrieved from:

https://www.vice.com/en_us/article/the-cash-is-in-the-cut

Wakabayashi, K. T., Weiss, M. J., Pickup, K. N., Robinson, T. E. (2010). Rats markedly escalate

their intake and show a persistent susceptibility to reinstatement only when cocaine is

injected rapidly. Journal of Neuroscience, 30(34), 11346-11355.

Walsh, S. L., & Cunningham, K. A. (1997). Serotonergic mechanisms involved in the

discriminative stimulus, reinforcing and subjective effects of cocaine.

Psychopharmacology, 130(1), 41-58.

Willuhn, I., Wanat, M. J., Clark, J. J., & Phillips, P. E. M (2009). Dopamine signaling in the

nucleus accumbens of animals self-administering drugs of abuse. In D. W. Self & J. K. S.

111 Gottschalk (Eds.), Behavioral neuroscience of drug addiction (pp. 29-71). Berlin:

Springer-Verlag.

Wilson, M. C., & Schuster, C. R. (1973). The effects of stimulants and depressants on cocaine

self-administration behavior in the Rhesus monkey. Psychopharmcologia, 31(4), 291-

304.

Wise, R. A. (1987). Intravenous drug self-administration: A special case of positive

reinforcement. In M. A. Bozarth (Ed.), Methods of assessing the reinforcing properties of

abused drugs, (pp. 117-141). New York: Springer-Verlag.

Weeks, J. R. (1962). Experimental morphine addiction: Method for automatic intravenous

injections in unrestrained rats. Science, 138, 143-144.

Wessinger, W. D. (1986). Approaches to the study of drug interactions in behavioral

pharmacology. Neuroscience & Biobehavioral Reviews, 10, 103-113.

WHO, World Health Organization (2016). WHO Expert Committee on Drug Dependence: 37th

report (WHO technical report series; no. 998). Geneva, Switzerland: WHO Press.

Woods, J. H., Bertalmio, A. J., Young, A. M., Essman, W. D., & Winger, G. (1988). Receptor

mechanisms of opioid drug discrimination. In F. C. Colpaert & R. L Balster (Eds.),

Transduction mechanisms of drug stimuli (pp. 95-106). New York: Springer-Verlag

Woolverton, W. L. (2014). Analysis of drug interactions in behavioral pharmacology. In T.

Thompson, P. B. Dews, & J. E. Barrett (Eds.), Advances in Behavioral Pharmacology:

Volume 6: Neurobehavioral Pharmacology (pp. 275-302). New York: Psychology Press.

112 Woolverton, W. L., Massey, B. W., Winger, G., Patrick G. A., & Harris, L. S. (1994). Evaluation

of the abuse liability of aminorex. Drug and Alcohol Dependence, 36(3), 187-192.

Woolverton, W. L., & Wang, Z. (2004). Relationship between injection duration, transporter

occupancy and reinforcing strength of cocaine. European Journal of Pharmacology,

486(3), 251-257.

Woolverton, W. l., Wang, Z., Vasterling, T., Carroll, F. I., & Tallarida, R. (2008). Self-

administration of drug mixtures by monkeys: combining drugs with comparable

mechanisms of action. Psychopharmacology, 196(4), 575-582.

Yokel, R. A. (1987). Intravenous self-administration: Response rates, the effects of

pharmacological challenges, and drug preference. In M. A. Bozarth (Ed.), Methods of

assessing the reinforcing properties of abused drugs, (pp. 1-33). New York: Springer-

Verlag.

Young, A. M. (1986). Effects of acute morphine pretreatment on the rate-decreasing and

antagonist activity of naloxone. Psychopharmacology, 88(2), 201-208.

Young, A. M. (1991). Discriminative stimulus profiles of psychoactive drugs. In N. K. Mello

(Ed.), Advances in substance abuse behavioral and biological research, Volume 4 (pp.

139-203). London: Jessica Kingsley Publishers.

Young, A. M., & Herling, S. (1986). Drugs as reinforcers: Studies in laboratory animals. In S. R.

Goldberg & I. P. Stolerman (Eds.), Behavioral analysis of drug dependence (pp. 9-67).

Orlando, FL: Academic Press, Inc.

113 APPENDIX A

IACUC Approval Letters

114

115

116

117 APPENDIX B

Individual and Stimulus-element Combination Generalization Test Results

118