<%BANNER%>

PK/PD Studies of Fentanyl Induced Physical Dependence and Analgesic Tolerance in Rats

Permanent Link: http://ufdc.ufl.edu/UFE0024229/00001

Material Information

Title: PK/PD Studies of Fentanyl Induced Physical Dependence and Analgesic Tolerance in Rats
Physical Description: 1 online resource (157 p.)
Language: english
Creator: Liu, Jiang
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: analgesic, dependence, fentanyl, modeling, pharmacodynamics, pharmacokinetics, rats, tolerance
Pharmacy -- Dissertations, Academic -- UF
Genre: Pharmaceutical Sciences thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: PK/PD STUDIES OF FENTANYL INDUCED PHYSICAL DEPENDENCE AND ANALGESIC TOLERANCE IN RATS Fentanyl is a potent mu-opioid receptor agonist. It has been widely used to relieve chronic pain, which is the most costly health problem in the U.S. Most research on this drug has been conducted in nai umlautve patients, but for those patients who have been exposed to the drug continuously, drug dependence and drug tolerance are significant concerns that require very careful dose adjustments. The aims of these studies were to verify and characterize the correlations among fentanyl pharmacokinetics, physical dependence and analgesic tolerance. From these studies, we expected to derive a clearer explanation of drug-effect relationships and a more efficient therapeutic design based on the administration route and dose regimen to minimize dependence and tolerance. In order to study the concentration-effect relationship, animal experiments were chosen and conducted in lab. Three doses (0.02, 0.04, and 0.08 mg/kg) of fentanyl were tested through subcutaneous administration for pharmacokinetic (PK) assessment. In the dependence studies, four doses (0.6, 0.3, 0.06, and 0.006 mg/kg/day) of fentanyl or saline vehicle were chronically administered via osmotic minipumps. A discrete-trial intracranial self-stimulation procedure was used to measure the brain reward threshold. Somatic signs were recorded from an opioid abstinence sign checklist. In analgesic tolerance studies, a tail-flick assay was employed to assess the nociceptive response. The PK of fentanyl could be well described with a linear two-compartment model. For dependence studies: 1) fentanyl dose-dependent physical dependence was verified; 2) the correlation between fentanyl PK and physical dependence was characterized by PK/PD models; and 3) intermittent dosing strategies were evaluated by simulation studies and found not to be suitable for chronic drug application. For analgesic tolerance studies: 1) a fentanyl dose-dependent, rapid development of analgesic tolerance and a fentanyl-induced hyperalgesia rebound after discontinuation of fentanyl were observed; and 2) the correlation between fentanyl PK and analgesic tolerance was well characterized by a counter response PK/PD model. These studies provided new insights to explain the drug-effect relationships of fentanyl physical dependence and analgesic tolerance. We expect our studies to improve fentanyl therapeutic design with respect to selection of the appropriate patient, product, and dosing regimen.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Jiang Liu.
Thesis: Thesis (Ph.D.)--University of Florida, 2009.
Local: Adviser: Derendorf, Hartmut C.
Local: Co-adviser: Hochhaus, Guenther.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2010-05-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0024229:00001

Permanent Link: http://ufdc.ufl.edu/UFE0024229/00001

Material Information

Title: PK/PD Studies of Fentanyl Induced Physical Dependence and Analgesic Tolerance in Rats
Physical Description: 1 online resource (157 p.)
Language: english
Creator: Liu, Jiang
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: analgesic, dependence, fentanyl, modeling, pharmacodynamics, pharmacokinetics, rats, tolerance
Pharmacy -- Dissertations, Academic -- UF
Genre: Pharmaceutical Sciences thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: PK/PD STUDIES OF FENTANYL INDUCED PHYSICAL DEPENDENCE AND ANALGESIC TOLERANCE IN RATS Fentanyl is a potent mu-opioid receptor agonist. It has been widely used to relieve chronic pain, which is the most costly health problem in the U.S. Most research on this drug has been conducted in nai umlautve patients, but for those patients who have been exposed to the drug continuously, drug dependence and drug tolerance are significant concerns that require very careful dose adjustments. The aims of these studies were to verify and characterize the correlations among fentanyl pharmacokinetics, physical dependence and analgesic tolerance. From these studies, we expected to derive a clearer explanation of drug-effect relationships and a more efficient therapeutic design based on the administration route and dose regimen to minimize dependence and tolerance. In order to study the concentration-effect relationship, animal experiments were chosen and conducted in lab. Three doses (0.02, 0.04, and 0.08 mg/kg) of fentanyl were tested through subcutaneous administration for pharmacokinetic (PK) assessment. In the dependence studies, four doses (0.6, 0.3, 0.06, and 0.006 mg/kg/day) of fentanyl or saline vehicle were chronically administered via osmotic minipumps. A discrete-trial intracranial self-stimulation procedure was used to measure the brain reward threshold. Somatic signs were recorded from an opioid abstinence sign checklist. In analgesic tolerance studies, a tail-flick assay was employed to assess the nociceptive response. The PK of fentanyl could be well described with a linear two-compartment model. For dependence studies: 1) fentanyl dose-dependent physical dependence was verified; 2) the correlation between fentanyl PK and physical dependence was characterized by PK/PD models; and 3) intermittent dosing strategies were evaluated by simulation studies and found not to be suitable for chronic drug application. For analgesic tolerance studies: 1) a fentanyl dose-dependent, rapid development of analgesic tolerance and a fentanyl-induced hyperalgesia rebound after discontinuation of fentanyl were observed; and 2) the correlation between fentanyl PK and analgesic tolerance was well characterized by a counter response PK/PD model. These studies provided new insights to explain the drug-effect relationships of fentanyl physical dependence and analgesic tolerance. We expect our studies to improve fentanyl therapeutic design with respect to selection of the appropriate patient, product, and dosing regimen.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Jiang Liu.
Thesis: Thesis (Ph.D.)--University of Florida, 2009.
Local: Adviser: Derendorf, Hartmut C.
Local: Co-adviser: Hochhaus, Guenther.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2010-05-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0024229:00001


This item has the following downloads:


Full Text

PAGE 1

1 PK/PD STUDIES OF FENTANYL INDUCED PHYSICAL DEPENDENCE AND ANALGESIC TOLERANCE IN RATS By JIANG LIU A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2009

PAGE 2

2 2009 Jiang Liu

PAGE 3

3 To my parents, my wife and my daughter

PAGE 4

4 ACKNOWLEDGMENTS Many people helped in the completion of this thesis in many different ways. First and foremost, I would like to express my gratitude a nd great appreciation to my advisor, Dr. Hartmut Derendorf, for opening the door of pharmaceutics for me, supporting and guiding me through my research and study, and helping me grow up in this field. It is a great honor to be one of his graduate students. I truly believe this superb experience of working with Dr. Derendorf will benefit my lifetime career. Also, I would like to thank my dissertation co mmittee, namely Dr. Guenther Hochhaus and Dr. Veronika Butterweck in our Department of Pharmaceutics, Dr. Adrie Bruijnzeel and Dr. Mark Gold in the Department of Psychiatry, and Dr. Andre Mauderli in the Department of Prosthodontics, for their broad knowledge, intel ligent guidance and generous support in lab space and equipments necessary for my experime nts. Without their great support, my studies would not have been conducted. Special thanks go to the research groups of Dr. Adrie Bruijnzeel, Dr. Fong Wong (in the Department of Prosthodontics), Dr Edwin Meyer (in the Departme nt of Pharmacology) and Dr. Rongling Wu (in the Department of Statistics) for letting me join their research groups to conduct research on my project and providing me w ith new input and equipment. It has really been a wonderful experience to learn from them and work with them. I also would like to extend my thanks to our postdoctoral fellows, Dr. Hao Pan, Dr. Vipul Kumar, Dr. Sreedharan N. Sabari nath and Dr. Michael Bewernitz and our department GLP lab manager, Ms. Yufei Tang for their great help, tec hnical support, and valuable discussion. I also would like to take this oppor tunity to express my gratit ude to Dr. Jeff Hughes and our department administrative staffs, including Mrs. Patricia J. Khan, Mr. Marty Rhoden and Ms. Robin Keirnan-Sanchez, for their great help and program support th roughout my Ph.D. study.

PAGE 5

5 Last but not least, I w ould like to give my special thanks to my family, friends, lab mates, our department faculty, staff, graduate students and post-doc fellows for their support and friendship. Thanks from the bottom of my heart to my parents, my sister, my in-laws, my supportive wife, Ke Ren, and my lovely da ughter, Ellen Liu, for their unconditional understanding, support, patience and encouragement. I could never complete my work without their love.

PAGE 6

6 TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................4 LIST OF TABLES................................................................................................................. ..........9 LIST OF FIGURES................................................................................................................ .......10 ABSTRACT....................................................................................................................... ............12 CHAPTER 1 INTRODUCTION................................................................................................................. .14 Background..................................................................................................................... ........14 Fentanyl....................................................................................................................... ...........16 Pharmacokinetic and Pharmacodynamic Approaches............................................................18 Pharmacokinetics: Concentration vs. Time.....................................................................18 Pharmacodynamics: Effect vs. Concentration.................................................................19 PK/PD Modeling.............................................................................................................19 Overview of PK/PD of Drug Dependence and Tolerance......................................................20 Mechanisms of Tolerance and Withdrawal.....................................................................21 Overview of PK/PD Modeling of Tolerance and Withdrawal........................................22 Hypothesis and Objectives.....................................................................................................2 4 2 PHARMACOKINETICS OF FENTANYL FOLLOWING SUBCUTANEOUS ADMINISTRATION IN RATS.............................................................................................36 Introduction................................................................................................................... ..........36 Materials and Methods.......................................................................................................... .36 Subjects....................................................................................................................... .....36 Drugs.......................................................................................................................... .....37 Jugular Vein Cannulation Surgery..................................................................................37 Drug Analysis.................................................................................................................. 38 Sample preparation...................................................................................................38 LC-MS/MS method..................................................................................................39 Method validation....................................................................................................39 Experimental Design.......................................................................................................40 Data Analysis.................................................................................................................. .40 Results........................................................................................................................ .............44 Noncompartmental PK Analysis.....................................................................................44 Compartmental PK Analysis...........................................................................................45 Discussion..................................................................................................................... ..........46

PAGE 7

7 3 EFFECTS OF FENTANYL DOSE AND EXPOSURE DURATION ON THE AFFECTIVE AND SOMATIC SIGNS OF FENTANYL WITHDRAWAL IN RATS........60 Introduction................................................................................................................... ..........60 Materials and Methods.......................................................................................................... .62 Subjects....................................................................................................................... .....62 Drugs.......................................................................................................................... .....62 Cranial Electrode Implantation Surgery..........................................................................62 Minipump Implantation Surgery.....................................................................................63 Apparatus and Assessing Brain Reward Function..........................................................64 Intracranial Self-Stim ulation Procedure..........................................................................64 Somatic Withdrawal Signs..............................................................................................66 Blood Sample Collection.................................................................................................66 Plasma Fentanyl Levels...................................................................................................67 Experimental Design.......................................................................................................67 Experiment 1: Effect of fentanyl dose on precipitated and spontaneous fentanyl withdrawal............................................................................................................67 Experiment 2: Effect of fentanyl ex posure duration on precipitated fentanyl withdrawal............................................................................................................68 Statistical Analyses..........................................................................................................6 9 Results........................................................................................................................ .............70 Experiment 1: Effect of Fentanyl Dose on Precipitated and Spontaneous Fentanyl Withdrawal...................................................................................................................70 Experiment 2: Effect of Fentanyl Ex posure Duration on Precipitated Fentanyl Withdrawal...................................................................................................................72 Discussion..................................................................................................................... ..........73 4 EFFECTS OF DOSE AND EXPOSURE DURATION ON THE DEVELOPMENT OF ANALGESIC TOLERANCE OF FENTANYL IN RATS....................................................85 Introduction................................................................................................................... ..........85 Materials and Methods.......................................................................................................... .86 Subjects....................................................................................................................... .....86 Drugs.......................................................................................................................... .....86 Thermally-Induced Response Observation Procedure....................................................87 Training....................................................................................................................88 Data collection..........................................................................................................88 Tail-Flick Measurement..................................................................................................89 Experimental Design.......................................................................................................89 Statistical Analyses..........................................................................................................9 0 Results........................................................................................................................ .............90 Thermally-Induced Response Observation.....................................................................90 Tail-Flick Measurement of Fentanyl Analgesic Effect Over Time.................................91 Discussion..................................................................................................................... ..........92

PAGE 8

8 5 PHARMACOKINETIC-PHARMACODYNAMIC MODELING OF PHYSICAL DEPENDENCE AND ANALGESIC TOLERANCE OF FENTANYL IN RATS.............100 Introduction................................................................................................................... ........100 Materials and Methods.........................................................................................................1 01 PK/PD Modeling of Physical Dependence of Fentanyl................................................101 PK modeling of fentanyl administra ted subcutaneously with osmotic minipump............................................................................................................101 PK/PD modeling of fentanyl affective dependence...............................................102 PK/PD modeling of fentanyl somatic dependence.................................................106 PK/PD Modeling of Analgesic Tolerance of Fentanyl..................................................108 Evaluation of Intermittent Dosing St rategies with Simulation Studies.........................110 Results........................................................................................................................ ...........110 Characterization of the PK/PD M odels of Physical Dependence.................................110 The receptor sensitization model char acterized affective dependence best...........111 The induction of the response model is more suitable for somatic dependence....111 Characterization of the PK/PD Model of Analgesic Tolerance....................................112 Impacts of Intermittent Dosing on Physical Dependence and Analgesic Tolerance.....113 Discussion..................................................................................................................... ........114 6 CONCLUSIONS.................................................................................................................. 134 APPENDIX A PHARMACOKINETIC MODELING OF SUBCUTANEOUS FENTANYL IN NONMEM......................................................................................................................... ...139 B SEQUENTIAL RECEPTOR SENSITIZATION MODELING OF AFFECTIVE DEPENDENCE FOLLOWING SUBCUTANEOUS FENTANYL PUMP ADMINISTRATION IN SCIENTIST.................................................................................141 LIST OF REFERENCES............................................................................................................. 143 BIOGRAPHICAL SKETCH.......................................................................................................157

PAGE 9

9 LIST OF TABLES Table page 1-1 Summary of approaches in modeling tolerance and rebound............................................27 2-1 Noncompartmental analysis of the observed fentanyl plasma concentration data with WinNonlin...................................................................................................................... ....48 2-2 Comparison of objective function values a nd Akaike information criterion values of fentanyl NONMEM PK analysis between oneand twocompartment models...............48 2-3 Population parameter estimates of the 2-compartment PK model using the standard two-stage approach............................................................................................................4 9 2-4 Parameter estimates of the final pop ulation 2-compartment PK model using NONMEM and the stability of the parameters using the bootstrap resampling procedure...................................................................................................................... ......49 3-1 Effect of fentanyl dose and exposur e period on plasma fentanyl levels............................79 3-2 Effect of fentanyl dose on the res ponse latencies associ ated with naloxoneprecipitated fenta nyl withdrawal........................................................................................80 3-3 Effect of fentanyl dose on the respon se latencies associated with spontaneous fentanyl withdrawal...........................................................................................................8 0 3-4 Effect of fentanyl dose and exposur e period on plasma fentanyl levels............................81 3-5 Effect of fentanyl dose and exposure period on the respons e latencies associated with naloxone-precipitated fe ntanyl withdrawal........................................................................81 4-1 Effect of testing temperatur e on escape and return latency...............................................96 5-1 Comparison of the goodness-of-fit statis tics of the PD models for affective dependence..................................................................................................................... ..120 5-2 Parameter estimates of the receptor sens itization model for a ffective dependence.........120 5-3 Comparison of the goodness-of-fit statis tics of the PD models for somatic dependence..................................................................................................................... ..121 5-4 Parameter estimates of the induction of the response model for somatic dependence....121 5-5 Comparison of the goodness-of-f it statistics of the PD models for analgesic tolerance.122 5-6 Parameter estimates of the counter-r esponse model for analgesic tolerance...................122

PAGE 10

10 LIST OF FIGURES Figure page 1-1 Chemical structures of the opioids.....................................................................................31 1-2 The scheme of the development of tolerance and depende nce to the chronic administration of a drug.....................................................................................................32 1-3 The scheme of the drug-response pro cess in the body and interrelationship among PK, PD and PK/PD modeling............................................................................................33 1-4 The general scheme of PK/PD tolerance modeling...........................................................34 1-5 The scheme of morphine -NO tolerance PK/PD model......................................................35 2-1 Mass spectrum product ions for fentanyl...........................................................................50 2-2 Chromatograms of plasma fentanyl...................................................................................51 2-3 The calibration curve of fe ntanyl rat plasma samples.......................................................52 2-4 The observed fentanyl plasma concentra tion time courses of three dosesdministered subcutaneously................................................................................................................. ..53 2-5 The linear regression of AUCin f (from NCA analysis) vs. dose.......................................54 2-6 The schemes of one-compartment PK m odel and two-compartment PK model with first-order drug absorption.................................................................................................55 2-7 The time courses of observed fentanyl plas ma concentrations of all individuals and the population 2-compartment PK model pred ictions using the standard two-stage approach....................................................................................................................... ......56 2-8 The diagnostic plot of the correlation among ETAs of the population 2-compartment PK model....................................................................................................................... ....57 2-9 The time courses of observed fentanyl plas ma concentrations of all individuals and the population 2-compartment PK model predictions using NONMEM..........................58 2-10 The diagnostic plots of goodness of fit of the population 2-comp artment PK model.......59 3-1 Effect of fentanyl dose on the elevations in brain reward thresholds and increases in somatic withdrawal signs associated with naloxone-precipitated fentanyl withdrawal....82 3-2 Effect of fentanyl dose on the elevations in brain reward thresholds and increases in somatic withdrawal signs associated w ith spontaneous fentanyl withdrawal...................83

PAGE 11

11 3-3 Effect of fentanyl dose and duration of exposure on the elevations in brain reward thresholds and increases in somatic wit hdrawal signs associated with naloxoneprecipitated fenta nyl withdrawal........................................................................................84 4-1 The TIROS (Thermally-Induced Response Observation System).....................................97 4-2 The complete time courses of observed antinociception of saline and different fentanyl treatments............................................................................................................ .98 4-3 Effect of fentanyl dose and exposure duration on rat an algesia during drug exposure and hyperalgesia after drug withdrawal.............................................................................99 5-1 The complete time courses of affective dependence and somatic dependence together with fentanyl PK..............................................................................................................1 23 5-2 The schemes of the PD models for affective dependence...............................................124 5-3 The schemes of the PD models for somatic dependence.................................................125 5-4 The complete time courses of antinoc iception together with fentanyl PK......................126 5-5 The model schemes of analgesic tolerance......................................................................127 5-6 The PK/PD model scheme implemented in Trail Simulator for the intermittent dosing simulation studies.................................................................................................128 5-7 The fittings to the complete time-courses of affective dependence.................................129 5-8 The fittings to the complete time-courses of somatic dependence..................................130 5-9 The fittings to the complete time-courses of the antinociceptive response.....................131 5-10 Impacts of intermittent dosing of fentanyl on the development of affective dependence, somatic dependence and analgesic tolerance..............................................132 5-11 Impacts of intermittent dosing of fentanyl on the steady state of affective dependence, somatic dependence and analgesic tolerance..............................................133

PAGE 12

12 Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy PK/PD STUDIES OF FENTANYL INDUCED PHYSICAL DEPENDENCE AND ANALGESIC TOLERANCE IN RATS By Jiang Liu May 2009 Chair: Hartmut Derendorf Cochair: Guenther Hochhaus Major: Pharmaceutical Sciences Fentanyl is a potent mu-opioid receptor agonist. It has been wi dely used to relieve chronic pain, which is the most costly health problem in the U.S. Most research on this drug has been conducted in nave patients, but for those pati ents who have been exposed to the drug continuously, drug dependence and drug tolerance are significant c oncerns that require very careful dose adjustments. The aims of these studies were to verify and characterize the corre lations among fentanyl pharmacokinetics, physical dependence and analgesic tolerance. From these studies, we expected to derive a clearer explanation of drug-effect relationships and a more efficient therapeutic design based on the administration route and dose regimen to minimize dependence and tolerance. In order to study the concentration-effect re lationship, animal experiments were chosen and conducted in lab. Three doses (0.02, 0.04, and 0.08 mg/kg) of fentanyl were tested through subcutaneous administration for pharmacokinetic (PK) assessment. In the dependence studies, four doses (0.6, 0.3, 0.06, and 0.006 mg/kg/day) of fent anyl or saline vehi cle were chronically administered via osmotic minipumps. A discrete-t rial intracranial self-stimulation procedure was

PAGE 13

13 used to measure the brain reward threshold. Somatic signs were recorded from an opioid abstinence sign checklist. In analgesic tolerance st udies, a tail-flick assay was employed to assess the nociceptive response. The PK of fentanyl could be well described with a linear two-compartment model. For dependence studies: 1) fentanyl dose-depende nt physical dependence was verified; 2) the correlation between fentanyl PK and physical dependence was characterized by PK/PD models; and 3) intermittent dosing strategies were eval uated by simulation studies and found not to be suitable for chronic drug applic ation. For analgesic toleran ce studies: 1) a fentanyl dosedependent, rapid development of analgesic to lerance and a fentanyl-induced hyperalgesia rebound after discontinuation of fentanyl were obse rved; and 2) the correlation between fentanyl PK and analgesic tolerance wa s well characterized by a counter response PK/PD model. These studies provided new insi ghts to explain the drug-effect relationships of fentanyl physical dependence and analgesic tolerance. We expect our studies to improve fentanyl therapeutic design with respect to selection of the appropr iate patient, product, and dosing regimen.

PAGE 14

14 CHAPTER 1 INTRODUCTION Background Chronic pain is one of the most underestimate d health care problems in the world. It has been reported to be the most costly health probl em in America. More th an one-quarter of adult Americans have had a problem with pain of some sort that persisted for more than 24 hours. Estimated annual costs, including direct medi cal expenses, lost income, lost productivity, compensation payments, and legal charges, sum up to over $100 billion. St atistics also reveal that pain is one of the most common symptoms in ambulatory persons with cancer/AIDS (NIH, 1998; NCHS, 2006). Opium, the dried juice of the seedpod of th e opium poppy, has been used for the relief of pain and suffering for several thousand years. Opioids, chemical compounds have a morphinelike action in the body, generally include: a) natural opiates (e.g., morphi ne, codeine, and thebaine); b) semi-synthetic opioids (e.g., heroin and buprenorphine); c) synthetic opioids (e.g., fentanyl, alfentanil, remifentanil, and meth adone) (Figure 1-1); and d) endogenous opioid peptides (e.g., endorphins, enkephalins, dynprphi ns, and endomorphins). These drugs have proven to provide cost-effective pa in medication for the treatment of moderate to severe pain when the underlying cause is cancer or AIDS. Th ese agents work by binding to opioid receptors, which are found principally in th e central nervous system (CNS) and the gastrointestinal tract. Opioid receptors are a group of G-protein couple d receptors. There are four major subtypes of opioid receptors: mu (), delta ( ), kappa ( ), and opioid-receptor-lik e receptor 1 (ORL1) (Corbett et al., 2006). The effect of an opioid de pends on which receptor it binds and its affinity for that receptor. At the cellular level, after an opioid agonist binds to the receptor, adenylyl cyclase (AC) activity is decr eased, which leads to a less cyclic adenosine monophosphate

PAGE 15

15 (cAMP) production. In tur n, this causes a reduced Ca2+ influx and K+ efflux at ion channels. This is responsible for a hyperpolarization of th e cell, which stops the transfer of afferent nociceptive signals and causes an algesia (Harrison et al., 19 98a; Raith and Hochhaus, 2004). However, opioid drugs are comm only underutilized in chroni c pain patients (Nicholson, 2003; Auret and Schug, 2005), due primarily to concerns of physical dependence and drug toleran (e.g., 61% of primary care physicians concerned about depe ndence and tolerance occurring when managing chronic pain (Bhamb et al., 2006)). Physical Dependence is generally described as a state in which a con tinuous drug presence is required to maintain the nor mal physiological or behavioral pr ocesses. Withdrawal symptoms (e.g., severe dysphoria, sweating, nausea, rhinorre a, depression, severe fatigue, vomiting and pain in humans) as well as sensitization re bound are characterized in the cessation of the drug use. The potential for physical dependence can be generally thought as the opposite effects produced by drug withdrawal. Clinical and la boratory observations have converged on the hypothesis that dependence and with drawal represent the pathologi cal neuroadaptive changes of rewardand stress-related circ uitries (Bruijnzeel et al., 2004). Drug Tolerance is described by a decreased effect with continuous or repeated drug exposure such that a higher dose is required subsequently to ma intain an equivalent effect (Figure 1-2). The phenomenon of tolerance can be divided into “appa rent tolerance” and “pharmacodynamic tolerance”. Apparent tolerance may be caused by drug PK changes or disease progression. Pharmacodynamic tolerance, also called functional tole rance, directly links the reduction in effect-intensi ty at a given concentration to the drug exposure. Opioid pharmacodynamic tolerance has been related to ne uroadaptive changes (e .g., desensitization, upregulation of cAMP, activation of N-methyl -D-aspartate (NMDA) receptors via second

PAGE 16

16 messenger mechanisms, and down-regulation spinal glutamate transporters (Mitra and Sinatra, 2004)). Tolerance develops to the analgesic, e uphoric, sedative, ventil ation depression and emetic effects of opioids but not to their e ffects on miosis and bowel motility (constipation). The developments of dependence and tolerance are commonly associated, i.e., physical dependence is always accompanied by tolerance, but not vice versa. Therefore, long-term administration of opioids could be associated wi th increased risk of abuse and addiction. The past few years have seen a marked increase in abuse of prescription opio id medications in the United States. All of these problems have created a difficult situation for health ca re providers who are considering using opioids to tr eat their patients for pain. On one hand, it is their goal to minimize suffering, and opioids in most cases are the most powerful analgesics available. On the other hand, they would like to prevent or minimize the negative health consequences of treatment of pain with opioids, including opioid overdose, abuse and addiction. Research is warranted in this area to better inform clinicians about the risk s and benefits of prescribing opioid medications for treating pain, as well as on how to prevent or minimize abuse or addiction in the clinical pain populati on treated with opioids. Fentanyl Fentanyl (N-(1-phenethyl-4-pip eridyl)-N-phenyl-propanamide) is one of the most popular opioid analgesics in modern anesthesia and pain therapy. Fentanyl infu sion is also commonly used for the prolonged sedation and analgesia that is often necessa ry for treating critically ill children. Fentanyl was first introduced into medi cal practice by Dr. Paul Janssen, the founder of Janssen Pharmaceutica (currently one of the most profitable divisions of Johnson & Johnson) in the1960s. There is a marked and continuing increase in the clinical use of fentanyl over the past twenty years (Joranson et al., 2000; Gilson et al., 2004; Novak et al., 200 4). Duragesic (the

PAGE 17

17 fentanyl patch marked by Johnson & Johnson) ha s experienced remarkable growth that has resulted in sales of $1.2 billion in 2002 and $1.7 billion in 2003 (Stanl ey, 2005). It is used extensively as an anesthetic fo r out-patient operative procedures and is gaining popularity for use in pain management. Fentanyl is a potent synthetic op ioid agonist with activity on 1and -opioid receptors. It is 100 times more potent than morphine as an anal gesic. It is highly li pophilic, crosses the blood brain-barrier (BBB) easily, accumulates in fatty ti ssues, and causes less histamine release than morphine (Simons and Anand, 2006). Thus, fentanyl has minimal depressant effects on the heart and is frequently used in anesthetic practice fo r patients undergoing heart su rgery or for patients with poor heart function. Its th erapeutic range is 0.2-2.5 ng/mL fo r nave patients (Dale et al., 2002). Fentanyl has a rapid onset a nd short duration of ef fects. It is a high -extraction drug and primarily metabolized by the redox cytochrome P450 3A4 in the liver. Less than 8% of i.v. administered drug is eliminated unchanged, with approximately 6% appearing in urine and 1% excreted in the stool (Feierman and Lasker, 1996; Labroo et al., 1997; Palkama et al., 1998). Its major metabolite, norfentanyl, is inactive. The half-life of fentanyl in humans is about 4 hours (Scholz et al., 1996). Fentanyl has high prot ein binding (80-85%) and binds mainly to 1-acid glycoprotein (Lotsch et al., 2002). Fentanyl has high stability at room temperat ure. Fentanyl pharmaceuticals are currently available in the dosage forms of oral transm ucosal lozenges, commonly referred to as the fentanyl "lollipops" (Actiq), tr ansdermal patches (Duragesic) and injectable formulations. Oral transmucosal lozenges are used for the manage ment of breakthrough can cer pain in patients who are already receiving opi oid medication for their underl ying persistent pain. The transdermal patches are used in the manageme nt of chronic pain in patients who require

PAGE 18

18 continuous opioid analgesia for pa in that cannot be managed by le sser means. Fentanyl citrate injections are administered intr avenously or intramuscularly for potent narcotic analgesia. Fentanyl tolerance and physical depende nce are common consequences of chronic treatment and are closely related (Adriaensen et al., 2003). Approximately half of the cancer patients converted to transdermal fentanyl from other opioid agents required increased dosages after initial applic ation of the patch (Muijsers and Wagsta ff, 2001). Evidence indicates that there has been a strong increase in abuse of fentanyl ov er the past years (Joranson et al., 2000; Gilson et al., 2004; Novak et al., 2004). Abuse of fentanyl and its analogs has been associated with a large number of drug overdose deaths, which mi ght be contributed by their high potency and respiratory depressant effects (Kronstrand et al., 1997; Kuhlma n et al., 2003; Lilleng et al., 2004). Meanwhile, tolerance of fentanyl has furt her complicated its therapeutic application. Recently, the FDA issued several warnings regardi ng the use of fentanyl after reports of deaths and other adverse events. The d eaths reported were th e result of improper se lection of patients, dosing, or improper product substitution (FDA, 2007a; FDA, 2007b). Pharmacokinetic and Pharmacodynamic Approaches Pharmacokinetics: Concentration vs. Time The term pharmacokinetics (PK) was first intr oduced by F. H. Dost in 1953 (DOST, 1953). Literally, it refers to the application of kinetic s to pharmakon (the Greek word for drugs). Hence, it is a study of the change of drugs over time in the body. It studies what the body does to drugs and often describes the concentration-time course of drugs and metabolites in different bodily fluids (e.g., plasma). PK is commonly divi ded into several areas including the extent and rate of drug absorption, distribution, metabolism and excretion. This is ofte n referred to as the ADME scheme.

PAGE 19

19 Pharmacodynamics: Effect vs. Concentration Pharmacodynamics (PD) is the science of drug action on the body and the relationship between drug concentration and effect (Lees et al., 2004). It details wh at the drug does to the body and characterizes the intensity of desired or undesired effects resulting from certain drug concentrations at effect sites. A drug effect can be defined as any drug-induced change in a physiological parameter when compared to the resp ective pre-dose or baseline value. PD is very important because the change in drug effect is usually not prop ortional to the change in drug concentration. At the most fundamental level, the actions of most drugs are produced by mechanisms whereby drugs interact with specific receptive target molecules or ‘receptors’ (in general we can call all target molecules ‘receptors’). Following target attachment but preceding the response, there is an ensuing linkage mechan ism or transduction pathway, leading to a time lag for the response. The duration of the time lag depends on the nature of the transduction pathway varying from milliseconds to hours. The pa rameters that characterize the drug-receptorresponse relation include affinit y, efficacy, potency and sensitivity. PK/PD Modeling PK/PD modeling uses mathematical functions to describe changes in drug concentration and changes in drug-effects. PK modeling us ually describes drug pl asma concentration ( C ) as a mathematical function of time ( t ): 11(;) Cft (1-1) where 1 is the set of PK parameters (e.g., fo r absorption, distribution, metabolism and elimination). PD modeling describes drug effect ( E ) as a mathematical function of drug concentration ( C ): 22(;) EfC (1-2)

PAGE 20

20 where 2 is the set of PD parameters (e.g., efficacy, potency, sensitivity). Ideally, concentrations should be measured at the effect site (the biophase) where the interaction with the respective biological receptor system takes pl ace. However, in most cases it is impractical to measure the effect site drug concentration. Thus, plasma drug concentrations are frequently used to establish these relationships under the assumption that the unbound (free) drug con centration in plasma reflects the unbound concentration at the e ffect site at steady-state condition. PK/PD modeling links the change in concentr ation over time as assessed by PK to the intensity of the observed respons e, regarding a certain concentr ation at the effect site, as quantified by PD. Thus, it a llows descriptions of the complete time course (e.g., onset, magnitude, duration) of the effects in response to a dosage regimen: 222112(;)((;);,) (;) EfCfftft (1-3) where 12(,) f ff is the combination of PK f unctions and PD function and12,) is the set of combined PK/PD parameters. This ap proach has proven to be effective in explaining the drug-response relation, and to help dose optimi zation to improve drug efficiencies and reduce the risk of undesired effects (Figure 1-3) (Derendorf and Meibohm, 1999; Derendorf et al., 2000). For instance, Porchet et al. (Porchet et al., 1988), Heinzen et al. (Heinzen and Pollack, 2004), and Fung et al. (Bauer et al., 1997b) have successfully applied PK/PD approaches in nicotine, morphine, and nitroglyce rin therapeutic studies to help the design of dosage regimens that avoid tolerance development while elucida ting factors that influence the development or regression of tolerance, as well as mechanisms of disease and drug action. Overview of PK/PD of Drug Dependence and Tolerance As shown in Figure 1-2, drug tolerance a nd dependence are classical PK/PD topics characterized by a time-variant dose -response pattern, and th ese generally have i ndirect-link (i.e.,

PAGE 21

21 the time dissociation between the plasma concentr ation and the biological response is caused by a distributional delay of drug to th e effect site), indirect-respons e (i.e., the temporal dissociation between the plasma concentration and the biologi cal response is characterized by describing the observed effects as a secondary result from a pr evious, time-consuming synthesis or degradation of an endogenous substance), and time-invariant at tributes (i.e., the effect regarding to drug concentrations is a function of exposure time). Much evidence has shown that drug dose, exposure duration, frequency of dosing, previous/c urrent use of cross de pendent/tolerant drugs, and the physiological condition of the subject are the PK/PD factors that determine the severity of dependence and tolerance (Lau and Sun, 2002; Li u et al., 2008). Theoretically, both tolerance and physical dependence are predictable pharm acological effects in response to a drug administration. However, the predictive power of current models is limited since they are commonly derived on empirical grounds, because of a lack of knowledge about the mechanisms underlying tolerance and dependence, and because of the difficulties in ap propriately simplifying complex physiological proce sses (Grdmark et al., 1999). Mechanisms of Tolerance and Withdrawal Research of opioid tolerance and dependence has been undertaken for decades. The first popular hypothesis suggested cellula r opioid receptor adaptation in the CNS was responsible for both opioid tolerance and the drug withdrawal syndrome (Collier, 1965). In vivo studies show conflicting data of receptor regulation (Bharg ava and Gulati, 1990a; Bhargava and Gulati, 1990b; Abdelhamid and Takemori, 1991; Rothman et al., 1991). Now we see receptor regulation as an oversimplification to explain drug depende nce and tolerance, and neuroadaptive changes involving many pathways of CNS became recognized (Harrison et al., 1998b; Bruijnzeel et al., 2004; Raith and Hochhaus, 2004). The mechanisms can be classified into three major sets:

PAGE 22

22 Cellular receptor regulation: may involve receptor downor up-regulation from the synthesis of messenger RNA and protein levels (Gies et al., 1997), receptor oligomerization (He et al., 2002) and receptor trafficking (desensitization) from receptor decoupling/coupling, phosphoryla tion, and internaliz ation levels (Crain and Shen, 1996; Chakrabarti et al., 1997; Harris on et al., 1998b). This may change the number of receptors at the effect site, and affect the efficacy and potency of drugs. Feedback regulation via depletion/pr oduction of endogenous modulators: in the indirect drug-response, certain endoge nous molecules may be involved. Those endogenous compounds may include cofactors/ enzymes, second messages, or certain receptor modulators (e.g., the vasoconstrictive effect of nitroglycerin (Wang et al., 2004; Visser et al., 2006) and the clomethiazole-i nduced hypothermia (Sallstrom et al., 2005; Visser et al., 2006)). Counter-response: may involve the opposite effects of drugs (e.g., the biphasic effects of alfentanil, alprazolam, midazolam, and clozap ine (Usune et al., 1988; Mandema and Wada, 1995; Lau and Heatherington, 1997; Lau et al., 1998) or affecting/producing the endogenous inhibitor, antagonist, partial agonist, and inverse a gonist (e.g., up-regulation of the cAMP pathway; activation of NMDA r eceptors via second messenger mechanisms, and down-regulation spinal glutamate transporte rs; elevations in neur onal nitric oxide, NO) (Liu and Anand, 2001; Heinzen and Pollack, 2004; Mitra and Sinatra, 2004; Heinzen et al., 2005). Overview of PK/PD Modeling of Tolerance and Withdrawal PK/PD models of dependence or dependence-pr ogression models associated with addictive substances suffer from the lack of reliable and qu antifiable biomarkers of dependence. C. E. Lau and colleagues successfully applied PK/PD appr oaches to study the effects of cocaine on reinforcement behavior (Lau et al., 2000) a nd later elucidated that the minimum plasma concentration ( Cmin) was the PK determinant of the freque ncy and pattern of intravenous cocaine self-administration in rats by PK modeling (Lau and Sun, 2002). Moreover, cytokines such as IL-6, CRP and TNF-alpha may serve as potential biomarkers of stress, a known predictor of relapse and addiction, in precl inical and clinical studies. The detailed mechanism of to lerance involved in a specifi c drug situation is drug and response dependent. Differences in tolerance to distinct drug-effects during the same drug treatment suggest differences in mechanisms of tolerance and receptor subtypes involved for

PAGE 23

23 various drug effects (Fattinger et al., 1997). Moreover, the de velopment of tolerance for a specific effect depends on drug route, dosage, an d dose regimen. Currently, all of the models are considered empirical or semi-physiological. They have proven to be useful in some specific situations, but none is general enough to describe every situation even for the same given drug (Grdmark et al., 1999). Theref ore, any extrapolations and pr edictions from models require special caution and validation. Si nce most models are developed from animal data and the interspecies differences in tolerance-developm ent can be significant, the application of a tolerance-model to human situations also must be verified. The first generation of tolerance models simply modifies the PD parameters as a function of time (e.g., by including a time-de pendent exponential decrease in Emax (e.g., max K tEe) or increase in EC50 (e.g., 50 K tECe ) (Chow et al., 1985)). Similar approaches were also developed to overcome the disadvantage of the predicted full tolerance of the above models. They incorporated a maximum allowed change in Emax or EC50 (e.g., maxmax 50[(1)]XY K tn nnEEeC E ECC or max 5050[(1)]XYn K tnnEC E ECECeC ). However, those kinds of models cleave the association between the tolerance developmen t and the dose exposure extent. Semi-mechanistic models have been proposed in recent years. Figure 1-4 describes the general scheme of this modeling. Generally spea king, the effect site co ncentration affects the response directly or indirectly through a modulat or or tolerance compartment. The development of tolerance is dependent on both exposure time and drug concentration; it may be represented as a hypothetical or real noncompetitive antagonist, competitive antagonist, partial agonist, inverse agonist, or direct/indirect modulator. Morphine is an exampl e of a drug whose tolerance models

PAGE 24

24 have been studied in detail during the last te n years. Heinzen and Pollack established the temporal relationship between morphine-induced el evations in neuronal n itric oxide (NO) and the development of antinociceptiv e tolerance in rats (Heinzen and Pollack, 2004). As shown in Figure 1-5, the stimulating effect of morphine on NO production was treated as an indirect response; a hypothetical effect compartment was in corporated to account for an evident delay in both NO production and the NO-associated decrease in antinociceptive effect. Not only did this model not only correctly predicte d the time course of the devel opment of tolerance at various dosages, but also helped in el ucidating a strong, time-dependent relationship between morphineinduced stimulation of NO production and tolerance development. Their work suggested that NO is a key mediator of antinociceptive toleranc e development, which is consistent with their later investigation that NO alte red in the -opioid receptor f unction (Heinzen et al., 2005). “While tolerance phenomena are well establ ished, relatively few attempts to model the time course of tolerance and its subsequent decay have appeared in the literature” (Gabrielsson and Weiner, 2006). Table 1-1 summarizes some of the exciting work over the last two decades along this line. Hypothesis and Objectives Considering the importance, th e representative, and the good PK properties of fentanyl accordingly, a dose optimization study focusing on fe ntanyl dependence and tolerance is likely: to provide important information about the drug ’s clinical applicati on, perhaps leading to mechanisms underlying its physiological actions, and serve as a useful reference for other opioids. According to the FDA, the fatal accidents invo lving fentanyl were the results of improper selection of patients, dosing, or improper product substitution (FDA, 2007a; FDA, 2007b). Selection of dosing and products mainly depends on the drug’s PK/PD properties which in this

PAGE 25

25 case are highly affected by the subject’s depende nce/tolerance profile. As mentioned before, most research on this drug has been conducted in nave patients, but the drug is mainly used in chronic treatment. For this special populat ion, drug dependence and drug tolerance are significant concerns that require very careful dose adjustments. Due to the powerful abilities of PK/PD modeling to explain the drug-response rela tionships and to optimize treatment regimens to improve drug efficiencies and reduce undesire d effects, our research hypothesis is that rigorous PK/PD studies targeting physical depe ndence and analgesic tolerance will benefit selection of patients, products, a nd dosing during fentanyl treatments. The PK and PD properties of physical dependen ce and analgesic tolerance of fentanyl have never been studied together. The goals of thes e studies are to: 1) inve stigate the impacts of fentanyl dosing and exposure duration on th e developments of physical dependence and analgesic tolerance; and 2) desi gn and simulate a more efficient therapeutic strategy based on the administration route and dose regimen to minimi ze dependence and tolerance. These studies will investigate and model th e PK profile of this drug and its dose-dependent development of dependence and tolerance over time in rats. Th rough the construction of the PK/PD models of physical dependence and analgesic tolerance, it is expected that the result s will lead to a clearer explanation of drug-effect relati onships which will contribute to improve fentanyl therapeutic design with respect to selecti on of the appropriate patient, product, and dosing regimen. To achieve our study goals, the followi ng specific aims were investigated: Specific Aim 1: Investigate the PK prof ile of fentanyl following subcutaneous administration in rats. This will allow us to confirm the linear pharmacokinetics of fentanyl and develop a PK structure model in this species With this quantitative information, drug concentrations will be predictable over time for any desired dose regimen.

PAGE 26

26 Specific Aim 2: Investigate the impacts of fentanyl-dose and exposure duration on the development of physical dependence in rats. Quanti tative measures of the deficit in brain reward function and in somatic syndrome associated with precipitated and spontaneous fentanyl withdrawal in rats will be conducted. Specific Aim 3: Investigate the impacts of fentanyl-dose and exposure duration on the development of analgesic tolerance in rats. Qu antitative measures will be determined for the attenuation in analgesic eff ect over time with different doses of fentanyl exposure. Specific Aim 4: Develop PK/PD models to describe physical dependence and analgesic tolerance following fentanyl exposure in rats. Mechanism-based PK/PD structure models will link the PK and PD information and enable us to predict the dependence and tolerance over time for any designed dose regimen. Simulation studies will be conducted with the constructed PK/PD models to optimize the drug application.

PAGE 27

27Table 1-1. Summary of approaches in modeling tolerance and rebound Mechanism base PD formula Drug PD endpoints Subject Dose protocol Reference Receptor regulation Desensitization Receptor downregulation __ 50_ __ __ __(1) (1)mn inRmoutRmm Rmn inRmfrenonpoutR n onpNNren m inEmnoutEmm inEmoutEdRDR kkR dtICDR dR kRRkDRkCRkR dt dDR dDR kCRkDRkDRkDR dtdt dE kSDRk dt dE kEk dt Methylprednisolone Tyrosine aminotransferase (TAT) Induction Male rat Long i.v. Infusion (Ramakrishnan et al., 2002) Tolerance compartment Noncompetitive Antagonist 0 501/p TSC EE CTC Nicotine Caffeine Heart rate Mean arterial Pressure Healthy men Constant i.v. Infusion (multiple dose) P.O., i.v. Infusion (Porchet et al., 1988) (Shi et al., 1993) 50 50 max 5050/ / 11 1/1/p T TTCEC CEC EE CTCTCC ARA (adenosine receptor agonist) Antilipolytic Effects Healthy men I.v. Infusion (Zannikos et al., 2001) Competitive Antagonist max 50 50(1)p T pEC E C E CC TC Morphine Antinociception (Tail-flick latencies) Male rat Long i.v. Infusion (Ouellet, 1995)

PAGE 28

28Table 1-1. Continued Mechanism base PD formula Drug PD endpoints Subject Dose protocol Reference Tolerance compartment Partial agonist max50max50 50505050ETTE E TETTEpT TpECTCTCEC E E CTCECCTCC Morphine Antinociception (Tail-flick latencies) Male rat Long infusion (Ouellet, 1995) Inverse agonist maxmax 5050 0 50501 E T ETp T p TC C ET ECTC EE C C ECTC Benzodiazepine EEG Young and elderly man Computer controlled infusion (Ihmsen et al., 2004) Counterresponse Inhibitor 50()() ():() ():1/(1,()EeTT Eeeee T TTTTTEfCfC fCaCorsC C fCbCorsC TC wheresrepresentsanaturalcubicsplinefunction Nicotine Systolic blood pressure, Diastolic blood pressure, Heart rate, Plasma epinephrine, Engergy expenditure, Plasma free fatty acids Human Constant infusion (single dose) (Fattinger et al., 1997) max max 5050E T E ETTp T p TEC TC E E CCTCC Ephedrine Systolic blood pressure Overweig ht human Oral (Persky et al., 2004) Counterresponse max max 0 5050E T E ETTp T pTEC TC EE E CCTCC Alprazolam Midazolam Shorter response Male rat I.V. bolus S.C. (Lau and Heatherington, 1997; Lau et al., 1998)

PAGE 29

29Table 1-1. Continued Mechanism base PD formula Drug PD endpoints Subject Dose protocol Reference Hypothetical counteracting metabolite Hypothetical counteracting metabolite maxmax 5050 max 10 (1)EM EEMMee ee ee onoffeee meECIM E ECCICM VCdM dM kkMkMkM dtKCdt Morphine Antinociception (electrical stimulation vocalization) Male rat Infusion (Heinzen and Pollack, 2004) Regulating counter effect Counter regulation max 0 50 0E EEp T p T TTTEC EEE ECC dE kEkE dt Alfentanil EEG Male rat Infusion (Mandema and Wada, 1995) Indirect counter regulation 0 0 PT T TTTinPTEESCE dE dM kMkEkSCkM dtdt Nitroglycerin Left ventricular end diastolic pressure (LVEDP) Rat model of congestive heart failure Infusion (Bauer et al., 1997a) Negative feedback via a moderator Indirect moderator max 50(1)(1) ()inout tolIER dE kkEM dtICER dM kEM dt Furosemide Diuresis Natriuresis Healthy man Multiple short infusion (Wakelkamp et al., 1996)

PAGE 30

30Table 1-1. Continued Mechanism base PD formula Drug PD endpoints Subject Dose protocol Reference Negative feedback via a moderator Moderator depletion max•• (oneoff e inout medE kCkE dt VC dM kkM dtKC Nitrioglycerin Peak Mean arterial pressure (MAP) Rat Repeated bolus vs. continuous infusion (Wang et al., 2004) Moderator production __max_ __(1) 1prl inprloutprlprl p I inTprloutTdC dEM kkC C dtdt M K dM kCk dt Chlorprothixene Prolactin-SecretingHealthy male Short infusion (Bagli et al., 1999) Pool model Unidirectional flow () ()intol toloutdP kkPHC dt dE kPHCkE dt Histamine and forskolin Acid secretion Fr og Repeated bolus vs. continuous infusion (Ekblad and Licko, 1984) Bidirectional flow 12 12( (intol outomedP kkkPkE dt dE kPkEkkC dt Omeprazole Gastric acid secretion Dog Short infusion (Abelo et al., 2000)

PAGE 31

31 MorphineCodeine HeroinBuprenorphine FentanylMethadone A. B. C. MorphineCodeine HeroinBuprenorphine FentanylMethadone A. B. C. Figure 1-1. Chemical structures of the opioids: natu ral opiates (morphine and codeine) (A), semi-synthetic opioids (heroin and bupr enorphine) (B), and synthetic opioids (fentanyl and methadone) (C)

PAGE 32

32 Figure 1-2. The scheme of the development of tolerance and depe ndence to the chronic administration of a drug

PAGE 33

33 Figure 1-3. The scheme of the drug-response pr ocess in the body (A) and interrelationship among PK, PD and PK/PD modeling (B) B.

PAGE 34

34 Effect C EC C E E 50 max dependent time EC E50,max Pharmacokinetics PharmacodynamicsTime Variant Em a xm o d e l C k10D ka E k1eke0 Modulator Ce Effect C EC C E E 50 max dependent time EC E50,max Pharmacokinetics PharmacodynamicsTime Variant Em a xm o d e l C k10D ka E k1eke0 C k10D ka E k1eke0 Modulator Ce Figure 1-4. The general scheme of PK/PD tolerance modeling. Th e effect site concentration affects the response directly or indir ectly through a modul ator or tolerance compartment.

PAGE 35

35 Figure 1-5. The scheme of morphine-NO tole rance PK/PD model. A temporal relationship between morphine-induced increases in ne uronal NO and loss of the antinociceptive effect is indicated (H einzen and Pollack, 2004).

PAGE 36

36 CHAPTER 2 PHARMACOKINETICS OF FENTANYL FOLLOWING SUBCUTANEOUS ADMINISTRATION IN RATS Introduction Fentanyl is a high-extracti on drug and primarily metabolized by the redox cytochrome P450 3A4 in the liver. Its major metabolite, nor fentanyl, is inactive. The pharmacokinetics (PK) of fentanyl was well studied in humans (Wille ns and Myslinski, 1993; Scholz et al., 1996; Ariano et al., 2001; Foster et al ., 2008) and mainly reflected a linear two-compartment kinetics following intravenous (i.v.) admini stration. The terminal eliminati on half-life is about four hours in humans (Scholz et al., 1996). The PK of fentanyl in rats has also been investigated, either through intravenous bolus or infusion administrati on. The PK models have varied from oneto two-compartment among those studies. The volume of distribution (Vd) values varied from several hundred to three thous and mL/kg, the clearance (CL) varied from 16 to 42 mL/min/kg, and the terminal elimination half-life (T1/2) from ~ 40 to 75 minutes (Cox et al., 1998; Yassen et al., 2005; Yassen et al., 2006; Ohts uka et al., 2007; Yassen et al ., 2008). The purpose of this study was to investigate the PK profile of fenta nyl with subcutaneous administration in rats. This allowed us to confirm the linear PK propert y of fentanyl and to develop a PK structure model in this species. With this quantitative information, drug concentrations can be predicted over time for any desired dose regimen. Materials and Methods Subjects Male Wistar rats (Charles River, Raleigh, NC) weighing 300-350 g at the beginning of the experiments were used. The rats were shipped to the UF Animal Care Services (ACS) center a week before the study started in order to reduce stress and to allo w them adapt to the research environment. Animals were socially housed (2 per cage) in a temperatureand humidity-

PAGE 37

37 controlled vivarium and maintained on a 12 hour light-dark cycle (lights off at 6 p.m.). All experiments were performed during the light phase of the cycle. Food and water were available ad libitum in the home cages. All subjects were treate d in accordance with th e National Institutes of Health guidelines regarding the principles of animal care. Animal facilities and experimental protocols were in accordance with the Associ ation for the Assessment and Accreditation of Laboratory Animal Care (AAALAC ) and approved by the Universi ty of Florida Institutional Animal Care and Use Committee. Drugs Fentanyl citrate was purchased from Sigma (Sigma-Aldrich, St. Louis, MO, USA) and dissolved in sterile saline (0.9% sodium chloride). The doses and concentrations of fentanyl are expressed as the free base. Fentanyl-d5 (100 ug/mL in methanol) was from Cerilliant (Pound Rock, TX, USA). Isoflurane (USP) was distribu ted by Webster Veterinary Supply, Charlotte, NC Jugular Vein Cannulation Surgery Two days before the PK experiment, indwel ling cannulae were implanted surgically. The cannulae were made from pyrogen-free, nonsterile polyethylene tubing (inner diameter of 0.3 mm and outer diameter of 0.7 mm). One day before surgery, cannul ae were disinfected in a 1% benzalkonium chloride solution. Surgery was carried out under anesthesia. A rodent anesthesia machine (Isotec-4 isoflurane va porizer: SurgiVet/ Smiths Medi cal, Waukesha, WI), with an induction chamber using an isoflurane vaporizer at an induction rate of 5% isoflurane with 2 l/min flow of O2 was used. Anesthesia was maintained by using a nose cone at a rate between 2.5% to 3% isoflurane with 1 l/min flow of O2. Before the surgery, the hair at the surgical area was shaved to prevent contamination at the surgic al site. The surgical site was disinfected with betadine and 70% alcohol. The rat was kept nor mothermic on a “circulating water” heating pad. The cannulae inserted in the right jugular vein were tunneled subcutaneo usly and fixed at the

PAGE 38

38 back of the neck with a rubber ring. The skin in the neck and throat was stitched with normal suture. To prevent clotting and cannula obstructi on, the cannulae were filled with physiological saline containing 40 IU/mL heparin (Baxter He althcare Corporation, Deerfield, IL, USA) Drug Analysis Fentanyl plasma concentrations were an alyzed using HPLC coupled to tandem mass spectrometer (LC/MS/MS) (Quatt ro, Micromass, Manchester, UK) (Day et al., 2003; Huynh et al., 2005). Sample preparation Standards and quality control samples: Working solutions were prepared in methanol at 10, 1.0, and 0.1 ng/L of fentanyl. Calibration curves were prepared daily in drug-free plasma containing heparin from working solution. The calib ration curve was fortified with fentanyl at 0.025, 0.05, 0.1, 0.25, 0.5, 1.0, 2.5, 5.0, 10.0, and 25 ng/mL. Each calibrator was analyzed in duplicate and quality control speci mens were analyzed in triplic ate. A second set of working solutions was prepared from separate lot numbers of reference material and used to fortify quality control samples. Quality control samples were prepared at 0.1, 0.5, and 2.5 ng/mL concentrations. Drug-free plasma was extracte d in each batch in duplicate and analyzed as negative controls. Working solutions were stable for up to 2 months at –20 C. Sample preparation and extractio n: A plasma volume of 100 L was used for the assay of rat samples, standards, and quality control samples. All plasma samp les were spiked with 5 uL of internal standard (20 ng/mL fentanyl-d5). Afte r adding 25 L of concentrated ammonia, the samples were extracted for 30 min by liquid/liquid extraction with 2.5 mL of methyl tert-butyl ether. After centrifugatio n at 1,000 g for 10 minutes, the organic phase was transferred to a clean silanized vial, evaporated to dryness, and r econstituted in 100 L of 0.1% formic acid in

PAGE 39

39 water/methanol/acetonitrile (50: 25:25, v/v/v). A volume of 50 L was injected onto the HPLC column. LC-MS/MS method The mobile phase consisted of 0.1% formic aci d in water:methanol:a cetonitrile (50:25:25, v/v/v). The pump was operated isocratically at 0.25 mL/min and ambient temperature. Chromatographic separation of analytes was achieved on a Phenomenex 30-mm 2.00-mm Luna 5 C18(2) column (Phenomenex USA, Torra nce, CA). The ESI source was operated with the spray voltage at 4.5 kV with 80 psi of sheath gas (high purity N2). The source and probe temperatures were maintained at 80 and 250 C. Positive precursor ions for fentanyl (m/z 337) and fentanyl-d5 (m/z 342) were selected to pass through the first quadr upole. In the second quadrupole, collision-induced dissociation (CID) was achieved using argon as the collision gas (at 3mTorr) and collision voltages of –30 volts for fentanyl. Product ions monitored in the third quadrupole were m/z 188.4 (fentanyl and fentanyl -d5) (Figure 2-1). The scan time was 0.2 s/scan. Method validation The method was validated according to Guid ance for Industry, Bioanalytical Method Validation from FDA (FDA, 2001). The accuracy a nd the precision of the proposed method were determined by analysis of the QC samples. Th e intra-day accuracy and precision were assessed from the results of six replicate analyses of QC samples (0.1, 0.5, and 2.5 ng/mL) on a single assay day. The inter-day accuracy and precision we re determined from the same QC samples analyzed on three consecutive days. The precision is expressed as % relati ve standard deviation (%R.S.D.), while inaccuracy (%) is expressed as [(cal culated amount predicted amount)/predicted amount 100]. The intraassay precisions at concentrations of 0.1, 0.5, and 2.5 ng/mL were 6.8 %, 3.9% and 4.7%, respectively. The in terassay precisions at concentrations of

PAGE 40

40 0.1, 0.5, and 2.5 ng/mL were 9.2%, 6.8% and 7.4%, respectively. Inaccuracy was less than 15.4% measured as the percentage difference from nominal value. The limit of detection (LOD) (10 pg/mL) was defined as the lowest detectab le concentration, taking into consideration a signal-to-noise ratio of 3. Limit of qu antification (LOQ) (25 pg/mL) (Figure 2-2) was determined at the lowest concentration at which the precision was less than 20% and the inaccuracy was less than 20%. Calibration curves were linear (r2>0.990) in the measured range between 25 and 25,000 pg/mL, using weighted (1/x2) linear regression (Figure 2-3). Experimental Design The study was performed with the approval of the Institutional Animal Care and Use Committee (IACUC) of the University of Florida (Protocol # E656). To minimize the influence of circadian rhythms, all expe riments were started between 8: 30 and 10:30 AM. Animals were randomly assigned to the three fentanyl dos e groups (0.02, 0.04, and 0.08 mg/kg, n = 6 per group). Fentanyl was administered through subcutane ous (s.c.) injection. Before the start of the drug administration, a blank blood sample ( 200 L) was withdrawn. Each blood sample withdrawn was replaced by an equal volume of heparinized 0.9% saline (20 IU heparin/mL). This procedure has minimal effects on the PK (Y assen et al., 2005). Serial blood samples were collected in heparinized microtubes at 3, 8, 16, 30, 60, 120, 180, and 240 minutes after drug administration. Plasma (100 L) was separated from the blood by centrifugation at 1,000 g for 15 minutes and frozen at -80 C until analysis (less than a month). Data Analysis Noncompartmental analysis (NCA) of the da ta was conducted with WinNonlin version 5.2 (Pharsight Corporation, Mountain View, CA). To determine the basic structural PK for fe ntanyl, both oneand two-compartment models were tested (Figure 2-6). The population PK parameters of fent anyl were estimated using both

PAGE 41

41 the standard two-stage approach (Ette and Williams, 2004) with WinNonlin version 5.2 and the nonlinear mixed-effects modeling as implemen ted in the NONMEM software version VI 1.2 (Beal and Sheiner, 1999). The population PK anal ysis approach in NONMEM, which takes into consideration both intraand inte rsubject variability, was undert aken by using the first-order conditional estimation method with interaction (FOCE interaction). All fitting procedures were performed on IBM-compatible computer s running under Windows XP with the Fortran compiler Compaq Visual Fortan version 6.5. S-PL US version 6.2 (Insightful Corp., Seattle, WA) was used for data processing and graphically presenting. Wings for NONMEM version 614 (WFN for NONMEM VI) was used for NONM EM running and bootstrap (Holford, 2008). The one-compartment model was parameterized in terms of the first-order absorption rate constant (ka), clearance (CL), and the volumes of distribution (V) by use of the ADVAN2 TRANS2 subroutines in NONMEM. The two-comp artment model analysis was performed by use of the ADVAN4 TRANS4 subroutines in NONMEM That is, the PK parameters, the firstorder absorption rate constant (ka), clearance (CL), the inter-compart mental clearance (Q), and the volumes of distribution of compartments 1 and 2 (V1 and V2) were estimated. The inter-subject variability of all model parameters was assumed to follow a log normal distribution over the population. Therefore an exponential distribution model was used to account for inter-subject variability: exp()itviPP (2-1) where Pi is the individual value of model parameter P, Ptv is the typical value (population mean value) of parameter P in the population, and i is the independent and identically distributed normal (i.i.d.-normal) inter-subject rando m variable with mean zero and variance 2. For

PAGE 42

42 example, the inter-subject variability of the two-compartment model parameters was modeled using the following exponential error model: exp()iitvkakaka (2-2) exp()iitvCLCLCL (2-3) exp()iitvQQQ (2-4) exp()iitvV1V1V1 (2-5) exp()iitvV2V2V2 (2-6) Selection of an appropriate residual error model was base d on inspection of the goodnessof-fit plots. On this basis a pr oportional error model was proposed to describe residual error in the plasma drug concentration: ,,(1)obsijpredijijCC (2-7) where Cobs,ij is the observed jth concentration of individual i, Cpred,ij is the predicted jth concentration of individual i, and ij is the i.i.d. normally distributed residual random variable with mean zero and variance 2. The residual error term contains all the error terms which cannot be explained by the inter-subject PK difference and refers to, for example, measurement and experimental error and structural model missp ecification. With the Bayesian POSTHOC option, NONMEM obtained the individual parameter estimates (posterior) using the population mean estimate of parameters (prior). From the indi vidual parameter estimates, individual predicted concentrations (IPRED) were obtained.

PAGE 43

43 To refine the stochastic mode l, the correlation between PK parameter estimates was tested by conducting covariance matrix analysis (OMEGA BLOCK option in NONMEM) and explorative graphical analysis with S-PLUS. Model selection and identifica tion was based on the likelihoo d ratio test, PK parameter point estimates and their resp ective confidence intervals, parameter correlations, and goodnessof-fit plots. For the likelihood ratio te st, the significance level was set at = 0.01, which corresponds to a decrease of 6.6 or 9.2 points in th e objective function value (OFV, which in turn corresponds to the -2X log maximu m likelihood value of the model) after the inclusion of one or two parameters respectively in the model, under the assumpti on that the difference in OFV between two nested models is 2 distributed. Also the Akaike in formation criterion (AIC) value was calculated as: 2 A ICOFVp (2-8) where p = total number of parameters in the m odel (structural + error). For hierarchy model comparison, the model with lower value of AIC was considered better (Burnham and Anderson, 1998; Wagenmakers, 2003). The time courses of observed fentanyl plas ma concentrations and population predicted (PRED) concentrations of all individuals were plotted, which allowed an overall description of the model fitting to the data. The following goodnessof-fit plots were also subjected to visual inspection to detect systemic deviations from the model fits (Ette, 1998): individual observed concentration (DV) versus PRED or individual predicted (IPRED) values, residuals (RES) versus PRED, and weighted residuals (WRES) versus PRED. Ideally the observed plasma fentanyl concentrations would be more randomly distri buted across the line of PRED, the points of DV versus PRED/IPRED should uniformly distribute along the identity line, the points of RES

PAGE 44

44 versus PRED should evenly distribute along the tw o sides of unity line, and the points of WRES versus PRED should homogenously distribute along the unity line. To demonstrate the accuracy, precision and stab ility of the model, the final population PK model was subjected to an internal validati on (Ette et al., 2003). The validation procedure consisted of a bootstrap validation analysis. For the bootstrap valida tion procedure, 1000 bootstrap replicates were generated randomly by sampling from the original data set with replacement. Subsequently, the fi nal population PK models were fitt ed to the bootstrap replicates one at a time. Finally, the mean, th e standard error of the mean (S .E.M.), coefficient of variation (CV%), and 95% confidence intervals (C.I.) of all model parameters were calculated and compared with parameter values ob tained from the NONMEM estimation. Results The experimental plasma concentration tim e courses of fentanyl following 0.02, 0.04 and 0.08 mg/kg s.c. administration were plotted in Figure 2-4. Data were expressed as mean standard deviation (SD) (n = 6). The drug had a fast absorption and also disappeared from plasma quickly. Noncompartmental PK and compartm ental PK analysis results were presented in the following. Noncompartmental PK Analysis Noncompartmental analysis (NCA) was conduc ted with WinNonlin 5.2. The results were listed in Table 2-1. Data were expressed as m ean standard error of mean (S.E.M.). The time of the maximum drug concentration (tmax) after dosing was approximately 20 minutes. The maximum concentrations (Cmax) were 1.84 0.19, 3.68 0.38, and 7.07 0.44 ng/mL for 0.02, 0.04 and 0.08 mg/kg dosing respectively, which declin ed with a terminal elimination half-life (T1/2) of 65.4 6.8, 61.5 7.2 and 74.0 10.9 minutes. The estimated total body clearances (CL_F) were 88.3 2.8, 90.5 5.0 and 88.1 3.9 mL/min respectively. The respective areas

PAGE 45

45 under the curve to the last measured time point AUCall were 207 28, 401 43, and 811 56 ng/mL*min respectively. The areas under the curve to infinity AUCINF (AUC) were 233 25, 422 47, and 9241 61 ng/mL*min. The dose normalized AUCs (AUCINF_D) were 11671 1250, 11051 1175, and 11552 763 ng/mL*min/mg respectively, which suggested a linear PK property of fentanyl together with Cmax and Cl_F (Figure 2-5). From analyses of variance (ANOVA), there is no significant diffe rence among the dose normalized AUCs (F2,15=0.29, P=0.75). Compartmental PK Analysis The PK data were analyzed by both oneand two-compartment models (Figure 2-6) using both the standard two-stage approach (Ette and Williams, 2004) with WinNonlin5.2 and the nonlinear mixed-effects modeling as implemente d in the NONMEM VI 1.2 (Beal and Sheiner, 1999). The comparison of the objective function valu es (OFV) and Akaike information criterion (AIC) values of fentanyl NONMEM PK analysis between oneand twocompartment models were listed in Table 2-2. As shown in this table, the AIC value (-168.1) from the twocompartment model was lower than the AIC va lue (-135.5) from the one-compartment model, and the OFV decreased from -149.5 of one-compa rtment model to -188.1 of two-compartment model. This is significant at the = 0.01 level (~ 2(3,0.01) = 11.4) using the likelihood ratio test under the assumption that the difference in OFV between two nested models is 2 distributed. Therefore, the two-compartment model structure was considered to be a better model structure for describing the plasma fentanyl PK data in rats. Table 2-3 listed th e population estimation of the fentanyl PK parameters by using the standard two-stage method. The simultaneous predictions of fentanyl concentration time cour ses of all three doses with the population mean parameters matched the observed data reasonably (Figure 2-7). Using the estimates from the standard two-stage method as th e initial values fo r the nonlinear mixed-effects modeling, the

PAGE 46

46 fitting converged quickly. Table 2-4 listed th e parameter estimations from NONMEM. All structural parameters could be adequately estimated, and inter-subject variability was estimated for V1_F, Q_F, V2_F and CL_F. The model was then evaluated by the bootstrap analysis. 934 out of 1000 bootstrap replicates minimized successf ully, indicating our model was not sensitive to the data variation. The NONMEM estimates were in agreement with the mean values of the bootstrap, indicating the PK model was stable. The correlation am ong inter-subject variability of parameters (ETAS) was diagnosed via a visual predictive check (VPC) and model comparison. No significant correlation was detected (Figure 2-8). So a diagonal inter-subject variancecovariance matrix (OMEGA matrix) was accepted in the final population PK model. To examine the goodness of fit of the final fentanyl population PK model, the diagnostic plots were visually inspected. Figure 2-9 indicated the population PK model was able to describe the observed concentrations time courses for all three dos es. In Figure 2-10, th e points of observed concentrations vs. individual (IPRED) or popula tion (PRED) predictions uniformly distributed along the identity lines (the soli d line) (A), the points of resi duals (RES) versus PRED evenly distributed along the two sides of the unity line and the points of weighted residuals (WRES) against PRED homogeneously distributed along th e unity line. Therefore these visual plots further validated the final fe ntanyl population PK model. Discussion In this study, the PK of fentanyl in ma le Wistar rats with doses of 0.02, 0.04 and 0.08 mg/kg administrated subcutaneously were i nvestigated. A population two-compartment linear PK model was proposed. PK model validation de monstrated the accurateness and precision of the developed population PK mode l. All PK parameters, including the estimated stochastic model parameters, were estimated precisely as indicated by the obtained S.E.M.s. The terminal elimination half-life values cal culated from NCA for fentanyl were 65.4 minutes for 0.02 mg/kg

PAGE 47

47 dose, 61.5 minutes for 0.04 mg/kg dose and 74.0 minutes for 0.08 mg/kg dose, which were consistent with what had been reported 40 to 75 minutes for rats (Cox et al., 1998; Yassen et al., 2005; Yassen et al., 2006; Ohtsuka et al., 2007; Yassen et al., 2008). Fe ntanyl is a highextraction drug and primarily metabolized by the cy tochrome P450 3A4 in the liver. Its reported clearance via i.v. bolus or infu sion varied from 16 to 42 mL/min/kg, which approximated the rat liver blood flow of 13.8 mL/min/0.25kg (Davies and Morris, 1993; Bjorkman and Redke, 2000). In this study, the estimated apparent clearan ce (clearance over bioavailability, CL/F) was 90 mL/min/kg, which was approximately two times of the reported clearance values and the rat liver blood flow. Fentanyl is a highl y lipophilic compound. A high st eady state tissue/blood partition ratio was reported which indicated that fentanyl was extensively distributed to tissues (Bjorkman et al., 1990). The estimated apparent volume of distribution (Vss/F) 5.5 l/kg was approximately one fold higher than the estimates from Cox’s and Ohtsuka’s studies (C ox et al., 1998; Ohtsuka et al., 2007). These indicated the drug substance, fentanyl citrate, administrated subcutaneously had not been fully absorbed into blood circulat ion. The estimation in this study is comparable with the reported approximately 30% transdermal absorption of tritiated fe ntanyl citrate in man (Sebel et al., 1987). Therefore, the final two-co mpartment linear PK mode l constructed in this study is reasonably powerful with respect to predicting fentanyl drug concentrations over time for any desired dose regimen and further will help to explain the drug-response relationship and design a more efficient therapeutic application.

PAGE 48

48 Table 2-1. Noncompartmental analysis (NCA) of the observed fentanyl plasma concentration data with WinNonlin v5.2 Dose (mg/kg) Tmax (min) Cmax (ng/mL) T1/2_z (min) Cl_F (mL/min)AUCall (ng/mL*min) AUCINF (ng/mL*min) AUCINF_D (ng/mL*min/mg) 0.02 25.67.2 1.840.19 65.46.8 88.32.8 20728 23325 116711250 0.04 17.64.9 3.680.38 61.57.2 90.55.0 40143 42247 110511175 0.08 23.18.1 7.070.44 74.010.9 88.13.9 81156 92461 11552763 Data are expressed as means ( S.E.M.). Table 2-2. Comparison of objec tive function values (OFV) and Akaike information criterion (AIC) values of fentanyl NONMEM PK analysis between oneand twocompartment models. Model number of parameters OFV AIC 1-compartment model 7 -149.5 -135.5 2-compartment model 10 -188.1* -168.1 *Significant decreased in OFV co mpared with 1-compartment model using the likelihood ratio test at = 0.01 level (~ 2(3,0.01) = 11.4).

PAGE 49

49 Table 2-3. Population parameter estimates of th e 2-compartment PK model using the standard two-stage approach Parameter mean CV% ka (/min) 0.03 47 V1_F (mL/kg) 1096 89 Q_F (mL/min/kg) 323 117 V2_F (mL/kg) 4215 56 CL_F (mL/min/kg) 89 11 Table 2-4. Parameter estimates of the fina l population 2-compartm ent PK model using NONMEM and the stability of the parameters using the bootstrap resampling procedure Original data 1,000 bootstrap replicates Parameter Structural model estimate ( ,(CV%)) Inter-individual variability ( ,(CV%)) Structural model estimate ( ,(CV%)) Inter-individual variability ( ,(CV%)) ka (/min) 0.024 (50.4) 0 (Fixed) 0.022 (35.3) 0 (Fixed) V1_F (mL/kg) 995 (47.5) 0.159 (99.2) 947 (34.4) 0.123 (73.7) Q_F (mL/min/kg) 158 (79.7) 0.352 (76.5) 152 (57.5) 0.421 (51.9) V2_F (mL/kg) 4450 (35.5) 0.312 (70.3) 3798 (36.0) 0.389 (52.1) CL_F (mL/min/kg) 90 (2.7) 0.084 (47.6) 88 (2.4) 0.077 (29.6) Residual Proportional error ,(95% CI ) 0.132 (0.090, 0.174) 0.127 (0.104, 0.150)

PAGE 50

50 Figure 2-1. Mass spectrum product ions for fentanyl Inset in each spectrum is the structure of fentanyl and the proposed frag mentation (Day et al., 2003).

PAGE 51

51 Figure 2-2. Chromatograms of plasma spiked with 0.025 ng/mL fentanyl and 1 ng/mL I.S.

PAGE 52

52 Figure 2-3. The calibration curve of fentanyl rat plasma samples in the range of 0.025 – 25 ng/mL Compound 2 name: Fentanyl Method File: 12-05-07 Coefficient of Determination: 0.997729 Calibration curve: 0.172581 x + 0.00965366 Response type: Internal Std ( Ref 1 ), Area ( IS Conc. / IS Area ) Curve type: Linear, Origin: Exclude, Weighting: 1/x, Axis trans: None 0.05.010.015.020.025.0 ng/ml 0 4.44 Response

PAGE 53

53 S.C. Fentanyl 0 50 100 150 200 250 0 2 4 6 8 Dose (0.02 mg/kg) Dose (0.04 mg/kg) Dose (0.08 mg/kg) Time (min)Cp (ng/ml) Figure 2-4. The observed fentanyl plasma con centration time courses of three doses (0.02, 0.04, and 0.08 mg/kg) administered subcutaneous ly. Data were expressed as mean standard deviation (SD) (n = 6).

PAGE 54

54 Figure 2-5. The linear regression of AUCinf (f rom NCA analysis) vs. dose. Data were expressed as mean standard erro r of the mean (S.E.M.) (n = 6). y = 11364x R2 = 0.9948 0 200 400 600 800 1000 1200 00.020.040.060.080.1 Dose (mg/kg)AUCInf(ng/ml*min)

PAGE 55

55 Figure 2-6. The schemes of one-compartment PK model (A) and twocompartment PK model (B) with first-order drug absorption

PAGE 56

56 Time (min) 050100150200250Cp (ng/ml) 0 2 4 6 8 10 obs. (D = 0.02 mg/kg) pred. (D = 0.02 mg/kg) obs. (D = 0.04 mg/kg) pred. (D = 0.04 mg/kg) obs. (D = 0.08 mg/kg) pred. (D = 0.08 mg/kg) Figure 2-7. The time courses of obs erved fentanyl plasma concentr ations of all individuals and the population 2-compartment PK model pred ictions using the standard two-stage approach

PAGE 57

57 Figure 2-8. The diagnostic plot of the correlation among ETAs of the population 2-compartment PK model ETA.V1 -1.0 -0.5 -0.0 0.5 -0.2 -0.1 0.0 0.1 -0.2-0.1-0.00.10.2 -1.0-0.5-0.00.5 ETA.Q EAT.V2 -0.5-0.3-0.10.10.30.5 -0.2 -0.1 -0.0 0.1 0.2 -0.2-0.10.00.1 -0.5 -0.3 -0.1 0.1 0.3 0.5 EAT.CL

PAGE 58

58 Time (min) 050100150200250Cp (ng/ml) 0 2 4 6 8 10 obs. (D = 0.02 mg/kg) pred. (D = 0.02 mg/kg) obs. (D = 0.04 mg/kg) pred. (D = 0.04 mg/kg) obs. (D = 0.08 mg/kg) pred. (D = 0.08 mg/kg) Figure 2-9. The time courses of obs erved fentanyl plasma concentrations of all individuals and the population 2-compartment PK model predictions using NONMEM

PAGE 59

59 0246810 0246810 0246810 0246810 Predicted concentration, ng/mLObserved concentration, ng/mL PRED IPRED 02468 -4-2024 PRED, ng/mlRES 02468 -4-2024 PRED, ng/mlWRES Figure 2-10. The diagnostic pl ots of goodness of fit of the popul ation 2-compartment PK model: A) observed concentration vs. population/i ndividual predictio n, B) residual vs. population prediction, and C) weighted residual vs. population prediction B. C. A.

PAGE 60

60 CHAPTER 3 EFFECTS OF FENTANYL DOSE AND EXPO SURE DURATION ON THE AFFECTIVE AND SOMATIC SIGNS OF FENT ANYL WITHDRAWAL IN RATS1 Introduction Fentanyl is a synthetic narc otic analgesic that is used as preanesthetic medication, anesthetic adjunct during general anesthesia, an d as an analgesic for the treatment of severe chronic pain (Linneman et al., 2000; Muijsers an d Wagstaff, 2001; Twite et al., 2004). Fentanyl is a specific mu-opioid recepto r agonist and is part of a class of compounds known as the phenylpiperidines (Skaer, 2006). Fentanyl is approximately 200 times more potent than morphine in rodent analgesia tests and 100 times more potent in drug discrimination procedures (Meert and Vermeirsch, 2005). Evidence indicates th at there has been a strong increase in the medical use and abuse of fentanyl over the past twenty years (Joranson et al., 2000; Gilson et al., 2004; Novak et al., 2004). Abuse of fentanyl and its analogs has been as sociated with a large number of drug overdose deaths, which might be due to their high potency and respiratory depressant effects (Kronstrand et al., 1997; Kuhlman et al., 2003; Lilleng et al., 2004). Clinical and preclinical studies indicate that the discontinuation of fent anyl administration induces a withdrawal syndrome, and a delayed discharg e from the hospital compared to when other analgesic narcotics are used (Franck et al., 1998 ; Dominguez et al., 2003; Fr anck et al., 2004). A large body of clinical evidence indicates that di scontinuation of chronic fentanyl administration results in a severe somatic withdrawal syndrome (Arnold et al., 1990; Franck and Vilardi, 1995; Franck et al., 1998; Dominguez et al., 2003; Franck et al., 2004). In addition, we demonstrated that fentanyl withdrawal is associated with el evations in brain reward thresholds in a rat 1 This chapter is a rewritten of our published paper (Neu ropharmacology 55 (2008) 8 12–818) with permission from Elsevier http://www.elsevier.com/copyright

PAGE 61

61 intracranial self-stimulation (ICSS) procedure (Bruijnzeel et al., 2006), and these elevations in brain reward thresholds can be prevented by pret reatment with the par tial mu-opioid receptor agonist buprenorphine (Bruijnzeel et al., 2007a). Elevations in brain reward thresholds are interpreted as a deficit in brain reward f unction (e.g., anhedonic-state) as higher current intensities are required to maintain responding for rewardi ng electrical stimuli (Barr and Markou, 2005). Elevations in brain reward thresholds have also been reported after the discontinuation of the administration of drugs of abuse such as cocaine, amphetamine, alcohol, and nicotine (Kokkinidis and McCarter, 1990; Schu lteis et al., 1995; Epping-Jord an et al., 1998; Cryan et al., 2003). The positive reinforcing effects of drugs of a buse play an important role in the initiating of drug taking behavior, however, it has been su ggested that the negative affective state associated with drug withdrawal provides a major motivational force for the continuation of drug abuse (Koob et al., 1997; Bruijnzeel and Gold, 2005). Therefore studies that investigate the effect of fentanyl dose and e xposure duration on the negative a ffective aspects of fentanyl withdrawal are warranted. Based on previous st udies by Easterling and Holtzman (Easterling and Holtzman, 1997), it was hypothesized that the deficit in brain re ward function depends on the dose of fentanyl administered and not on the ex posure duration. The aim of the first experiment was to investigate the effect of the dose of fentanyl on the affective and somatic signs of naloxone-precipitated and spontaneo us fentanyl withdrawal. The aim of the second experiment was to investigate the effect of the fentanyl administration pe riod on the affective and somatic signs of precipitated fentanyl withdrawal. Studies that provide information about the effects of fentanyl dose and exposure durati on on affective and somatic wit hdrawal signs may aid in the development of treatment regimen that prov ide maximal pain relief and minimal opioid withdrawal symptoms after the di scontinuation of the treatment.

PAGE 62

62 Materials and Methods Subjects Male Wistar rats (Charles River, Raleigh, NC) weighing 300-350 g at the beginning of the experiments were used. Animals were socially housed (2 per cage) in a temperatureand humidity-controlled vivarium and maintained on a 12 hour light-dark cycle (lights off at 6 p.m.). All testing occurred during th e light phase of the cycle. Food and water were available ad libitum in the home cages. All subjects were treated in accordance with the National Institutes of Health guidelines regarding the principles of animal care. Animal facil ities and experimental protocols were in accordance with the Association for the Assessment and Accreditation of Laboratory Animal Care (AAALAC) and approved by the Univer sity of Florida Ins titutional Animal Care and Use Committee. Drugs Naloxone hydrochloride and fentan yl citrate were purchased from Sigma (Sigma-Aldrich, St. Louis, MO, USA) and dissolv ed in sterile saline (0.9% sodi um chloride). Fentanyl was administered subcutaneously (sc) using osmo tic minipumps (Durect Corporation, Cupertino, CA) for 14 days. Naloxone and fentanyl doses are expressed as the salt. Cranial Electrode Implantation Surgery For the electrode implantations, the rats were anesthetized with an isoflurane / oxygen vapor mixture (1-3% isoflurane), and placed in the surgery area and secured in a Kopf stereotaxic frame (David Kopf Instruments, Tu junga, CA) with the incisor bar set 5 mm above the interaural line. Core body temperature was maintained at 35C by placing an adjustable temperature controlled heating pad beneath the ra t. The incision area of the cranium was given final prepping using 70% alcohol and betadine in succession 2 more times then a small sterile drape is placed over the incision site. A 1cm midline incision was made through the skin; the

PAGE 63

63 skin was then retracted, allowing adequate exposure of the bone surface. Us ing a micro drill, 4 0.5mm holes were drilled through the cranial surf ace; sterile micro screws were then threaded into the holes, (these screws were necessary to help anchor the electrode and only penetrated the thickness of the bone, approximately 1mm). An additional hole (AP -0.5 mm; ML 1.7 mm) was drilled for the electrode. The 11 mm stainless steel bipolar elect rode (model MS303/2 Plastics One, Roanoke, VA) was implanted in the medial forebrain bundle at the level of the posterior lateral hypothalamus (anteriorposterior 0.5 mm; medial-later al 1.7 mm; dorsal-ventral 8.3 mm from dura). After the implantation of the electrode, dental cement (Pharmaceutical grade Dental Acrylic Cement, CO-Oral-Ite Dental Mfg. Co.), was carefully applied around the electrode and screws to permanently secure th e electrode to the skull. The dental cement hardened and created a mushroom like lip for the skin to be sutured with sterile non-absorbable monofilament thread. Following completion of surgery, betadine ointment (10% povidoneiodine) and local pain reliever ointment (betacain e gel, contains 5% lidoc aine) were applied to the incision areas. The rat was then placed in a recovery cage and closely monitored for at least 1 hour. Heating pads placed under the recovery ca ges were maintained at ~25C. In addition, Flunixin (non-steroidal anti-inflammatory drug (NSAID), 1.25 mg/kg) wa s administered every 12 hours for the first 48 hours after the su rgery to relief postoperative pain. Minipump Implantation Surgery In order to induce dependence, rats were pr epared with osmotic minipumps (Alzet model 2ML2, 14 day pumps) that delivered fentanyl. The control rats were prep ared with pumps that delivered vehicle (saline). A rodent anesthesia machine, with an induction chamber using an isoflurane vaporizer at an induction rate of 5% isoflurane with 2L/min flow of O2 was used. Maintaining anesthesia using a nose cone was at a rate between 2.5% to 3% isoflurane with 1L/min flow of O2. The pumps were implanted subcutaneously in an aseptic manner. One side

PAGE 64

64 of the flank of the rats was clipped and then su rgically prepped using 70 % alcohol and a betadine solution. A 1.5 cm skin incision wa s made posterior of the ribcag e. Using a sterile blunt tissue forceps, a small pocket was created by separating th e muscle and subcutaneous layer for ease of the insertion of the sterile minipump. The mi nipumps was implanted under the skin and the incision was closed with a non-absorbable, simp le interrupted suture. Following completion of surgery, betadine ointment (10% povidone-iodine) and local pain reliever ointment (betacaine gel, contains 5% lidocaine) were applied to the incision area. Apparatus and Assessing Brain Reward Function The experimental apparatus consisted of tw elve Plexiglas chambers (30.5 x 30 x 17 cm; Med Associates, Georgia, VT), each housed in a sound-attenuating melamine chamber (Med Associates, Georgia, VT). The operant chambers consisted of a metal grid floor and a metal wheel (5 cm wide) centered on a sidewall. A phot obeam detector was attached next to the response wheel and recorded every 90 degrees of rotation. Brain stimulation was delivered by constant current stimulators (Model 1200C, Stimtek, Acton, MA). Subjects were connected to the stimulation circuit through bipolar leads (Plastics One, Roanoke, VA) attached to goldcontact swivel commutators (model SL2C Plastic s One, Roanoke, VA). A computer controlled the stimulation parameters, data coll ection, and all test session functions. Intracranial Self-Stimulation Procedure Rats were trained on a modified discrete-t rial ICSS procedure (Kornetsky and Esposito, 1979), as described previously (Bruijnzeel et al ., 2007b). The subjects were trained initially to turn the wheel on a fixed ratio 1 (FR1) schedule of reinforcement. Each quarter turn resulted in the delivery of a 0.5 second train of 0.1 millis econd cathodal square-wave pulses at a frequency of 100 Hz. After the successful acquisition of re sponding, defined as 100 reinforcements within 10 minutes, the rats were gradually trained on a di screte-trial current-thr eshold procedure. Each

PAGE 65

65 trial began with the delivery of a non-contingent electrical stimulus, followed by a 7.5 second response window within which the animal could re spond to receive a second contingent stimulus identical to the initial non-c ontingent stimulus. A response du ring this 7.5 second response window was labeled a positive response, while th e lack of a response was labeled a negative response. During a 2-second period immediately af ter a positive response additional responses had no consequence. The inter-trial interval (I TI), which followed either a positive response or the end of the response window, had an averag e duration of 10 seconds (7.5 12.5 seconds). Responses that occurred during the ITI resulted in a further 12.5 second delay of the onset of the next trial. During training on th e discrete-trial procedure, th e duration of the ITI and delay periods induced by time-out responses were gradually increased until animals performed consistently at standard test parameters. The rats were subsequently tested on the currentthreshold procedure in which stimulation inte nsities varied according to the classical psychophysical method of limits. A test session consis ted of four alternating series of descending and ascending current intensities st arting with a descending series. Blocks of three trials were presented to the subject at a given stimula tion intensity, and the intensity was altered systematically between blocks by 5 A steps. The initial stimulus intensity was set 40 A above the baseline current-threshold for each animal. Each test session typically lasted 30-40 minutes and provided two variables: brai n reward thresholds and response latencies. The brain reward threshold for a descending series was defined as the midpoint between stimulation intensities that supported responding (i.e., positive responses on at least two of the three trials), and current intensities that failed to support responding (i.e ., positive responses on fewer than two of the three trials for two consecutive blocks of tria ls). The threshold for an ascending series was defined as the midpoint between stimulation in tensities that did not support responding and

PAGE 66

66 current intensities that supported responding for two consecutive bloc ks of trials. Four threshold estimates were recorded and the mean of these va lues was taken as the brain reward threshold for a specific subject. The response la tency was defined as the time interval between the beginning of the non-contingent stimulus and a positive re sponse. The response latency for each test session was defined as the mean response latenc y on all trials during which a positive response occurred. Somatic Withdrawal Signs The rats were habituated to the Plexiglas observation chambers (10" x 10" x 18"; L x W x H) for 5 minutes per day on three consecutive days prior to the test sessions. During the test sessions the individual somatic withdrawal signs were reco rded for 10 minutes. A total withdrawal score was calculated using the modified Gellert a nd Holtzman rating scale (Gellert and Holtzman, 1978; Bruijnzeel et al., 2006). This scale consists of graded signs (score dependent on frequency of app earance) and checked signs (score independent of frequency of appearance). Graded signs included escape attempts (n = 2 4, score 1; n = 5 9, score 2; n = 10 or more, score 3) and abdominal constrictions (a score of 1 per 2 constrictions). Checked signs included diarrhea (score 3), facial fasciculations or teeth chattering (score 2 if n > 5), ptosis (score 2 if n > 5), eye blinks (score 2 if n > 10), wet dog shakes (score 2 if n > 2), swallowing (score 2 if n > 2), abnormal posture (score 3 if n > 2), erection or ejaculation (score 3 if n > 2) and irritability (score 3). Blood Sample Collection The rat was completely anesth etized by putting in an isoflurance induc tion chamber for 5 minutes before the collection procedure. The to ngue was extended in front with a cotton-tipped applicator stick and was carefully pulled forward with a forceps. One of the sublingual veins was punctured with a 23 gauge needle. After successfu lly puncturing, the rat was turned back into a

PAGE 67

67 prone position and the blood was allowed to drop into a plasma collection tube. As soon as the required volume (~450 l) of blood had been co llected successfully, potential hemorrhaging from the needle puncture was stopped by applying a cotton-tipped applicator soaked with a 50% solution of iron chloride (Zeller et al., 1998; Mahl et al., 2000). Plasma Fentanyl Levels A validated liquid chromatography–tandem mass spectrometry (LC/MS/MS) method was used to determine plasma fentanyl levels. First, 6 l of internal standard (25 ng/mL, fentanyl-d5) was added to 200 l plasma aliquots. Fifty l of concentrated ammonia was added to the samples and subsequently the samples were extracted by liquid / liquid extraction using 3 mL of methyl tertiary-butyl ether. Then the samples were r econstituted in 200 l of 0.1% formic acid in a water/methanol/acetonitrile (50: 25:25, v/v/v) solution. Fifty l was injected into a highperformance liquid chromatography (HPLC) syst em that was connected to a Quattro mass spectrometer (Micromass, Manchester, UK). The pump was operated isocratically at 0.25 mL/min at ambient temperature. Chromatographic separation of the analytes was achieved on an 30 mm 2.00 mm Luna 5 C18(2) column (P henomenex, USA, Torrance, CA). The minimal limit of quantification (LOQ) was 25 pg of fentanyl per milliliter. Experimental Design The study was performed with the approval of the Institutional Animal Care and Use Committee (IACUC) of the University of Florida (Protocol # E613). Experiment 1: Effect of fentanyl dose on p recipitated and spontaneous fentanyl withdrawal After implantation of the elec trodes and a 7-day recovery pe riod the rats were trained on the ICSS procedure. After stable baseline brain rewa rd thresholds were achieved (defined as less than 10% variation within a 5 day period), the rats were prepared with 14-day osmotic minipumps that contained either saline or fentanyl (0.006, 0.06, 0.3, 0.6 mg/kg/day, n = 6 per

PAGE 68

68 group) dissolved in saline. For all experiments, brain reward thresholds, response latencies, and body weights were measured daily between 9:00 am and 12:00 noon. Naloxone (0.003–0.03 mg/kg, sc) injections were administered accordin g to a within-subjects Latin square design and started at least 6 days after the implantation of the minipumps. After the administration of naloxone, the rats were returned to their home cage for 5 minutes. Then the rats were placed in the observation cages and somatic withdrawal signs were recorded for 10 minutes by an experienced observer who was blind to the treatment conditions. Im mediately after the observation period the rats were pl aced in the ICSS test chambers and brain reward thresholds and response latencies were assessed. After 14 days, the minipumps were removed and somatic signs, reward thresholds, and response late ncies were assessed 6, 12, 24, 36, 48, 72, 96, and 120 hours after explantation of the minipumps. Blood samples were collected from the sublingual vein 7 and 14 days after the implantation of the fentanyl pumps. The plasma fentanyl levels were determined using quantitative LC/MS/MS. Experiment 2: Effect of fentanyl exposure dur ation on precipitated fentanyl withdrawal Electrodes were implanted in the medial fo rebrain bundle and the ra ts were stabilized (defined as less than 10% variat ion within a 5 day period) on th e ICSS procedure. The rats were then prepared with 14-day minipumps that contai ned saline (n = 5) or fentanyl (0.6 mg/kg/day, n=5; 0.3 mg/kg/day, n = 6 per group). In order to investigate the development of opioid dependence, naloxone (0.03 mg/kg, sc) was ad ministered 8, 24, 48, 72, 120, and 168 hours after the implantation of the minipumps. After the administration of na loxone, the rats were returned to their home cages for 5 minutes and then so matic withdrawal signs were recorded for 10 minutes in the observation cages. Immediately afte r recording the somatic signs, the rats were placed in the ICSS test chambers and brain re ward thresholds and response latencies were assessed. Blood samples were collected from the sublingual vein immediately after ICSS testing

PAGE 69

69 at the 8, 24, 48, 72, and 120 hour time-points. The plas ma fentanyl levels were determined using quantitative LC/MS/MS. Statistical Analyses For experiment 1 (precipitated withdrawal), brain reward thre sholds and response latencies were expressed as a percentage of the pre-test day values. Soma tic signs and percent changes in brain reward thresholds and response latencies were analyzed using two-way repeated-measures analyses of variance (ANOVA), with the dose of naloxone as the within-subjects factor and pump content (saline and various fentanyl doses) as the between -subjects factor. For experiment 1 (spontaneous withdrawal), ICSS parameters we re expressed as percentages of the values obtained on the day prior to minipump explanta tion. Somatic signs and pe rcent changes in ICSS parameters were analyzed using two-way re peated-measures ANOVA with pump-content as between-subjects factor and time as the within -subjects factor. A two-way repeated-measures ANOVA was used to compare plasma fentanyl levels 7 days and 14 days after the implantation of the minipumps. The between-subjects factor wa s pump-content and the w ithin-subjects factor was time. For experiment 2, somatic signs and percent changes in ICSS parameters were analyzed using a two-way repeated-measures A NOVA. The between-subjects factor was pumpcontent and the within-subject s factor was time. A two-way repeated-measures ANOVA was also used to compare plasma fentanyl levels 8, 24, 48, 72 and 120 hours after the implantation of the minipumps. The between-subjects factor was pump-content and the within-subjects factor was time. For all experiments, statistically si gnificant results in the ANOVA were followed by a Tukey multiple pairwise comparison test. The criterion for significance was set at 0.05. The statistical analyses were performed us ing SAS for Windows software, version 9.13.

PAGE 70

70 Results Experiment 1: Effect of Fentanyl Dose on Precipitated and Spontaneous Fentanyl Withdrawal Mean ( S.E.M.) absolute brain reward thre sholds before pump-implantation for salineand fentanyl-treated (0.006, 0.06, 0.3, 0.6 mg /kg/day) rats were 101.53 3.88, 101.29 3.12, 110.80 3.19, 100.88 5.31, and 100.00 4.38 A, respectively [F4,25=0.19, n.s.]. Mean ( S.E.M.) absolute response latencies for salineand fentanyl-treated (0.006, 0.06, 0.3, 0.6 mg/kg/day) rats were 3.12 0.05, 3.14 0.04, 3.23 0.04, 3.06 0.04, and 3.13 0.03 seconds, respectively [F4,25=0.39, n.s.]. Statistic al analyses indicated that th ere was an effect of the dose of fentanyl on plasma fent anyl levels (Table 3-1; Treatment: F3,20=508.78, P<0.001). The plasma fentanyl levels were the same at da y 7 and day 14 after minipump implantation (Time: F1,20=3.26, n.s.). Statistical analyses indicated th at naloxone induced an elevati on in brain reward thresholds in the rats treated with fent anyl and naloxone induced greate r elevations in brain reward thresholds in rats treated with high doses of fentanyl than in rats treated with low doses of fentanyl (Figure 3-1A; Dose x Treatment in teraction: F12,75=8.59, P<0.001). Posthoc analyses indicated that fentanyl had a dose-dependent effect on the naloxone (0.03 mg/kg) induced elevations in brain reward thresh olds. The brain reward thresholds of the rats chronically treated with 0.6 mg/kg of fentanyl and acutely treated with treated with 0.03 mg /kg of naloxone were elevated compared to those of the rats chronica lly treated with 0.3 mg/kg of fentanyl and acutely treated with the same dose of naloxone. In addi tion, the brain reward thresholds of the rats chronically treated with 0.3 mg/kg of fentanyl and acutely treate d with treated with 0.03 mg/kg of naloxone were elevated compared to t hose of all other fentanyl groups (0.0, 0.006, 0.06 mg/kg) acutely treated with the same dose of naloxone. Naloxone (0.03 mg /kg) did not elevate

PAGE 71

71 the brain reward thresholds of the rats treate d with the lowest doses of fentanyl (0, 0.06, 0.06 mg/kg/day). Statistical analyses indicated that naloxone increa sed somatic signs in the rats treated with fentanyl and naloxone induced a greater increase in soma tic signs in rats treated with high doses of fentanyl than in rats treated w ith low doses of fentanyl (Figure 3-1B; Dose x Treatment interaction: F12,75=15.24, P<0.001). Posthoc analyses indicated that fentanyl had a dose-dependent effect on the naloxone (0.01 a nd 0.03 mg/kg) induced increases in somatic withdrawal signs. The ad ministration of naloxone did not aff ect the response latencies of the fentanyl or saline-treated ra ts (Table 3-2; Dose: F4,25=1.31, n.s.; Treatment: F3,75=0.73, n.s.; Dose x Treatment interaction: F12,75=0.46, n.s.). This indicates that naloxone did not induce motor impairments in the fentanyl dependent rats. Explantation of the minipumps induced an eleva tion in the brain reward thresholds in the fentanyl-treated rats. Discon tinuation of fentanyl administ ration induced higher and more prolonged elevations in brain reward thresholds in rats treated with high doses of fentanyl than in rats treated with low doses of fentanyl (Fi gure 3-2A; Dose x Time interaction: F28,175=4.94, P<0.001). Posthoc analyses indicated that ex plantation of the minipumps induced a dosedependent elevation in brain reward threshol ds. Discontinuation of fentanyl administration induced elevations in brain reward thresholds in rats treated with high doses of fentanyl (0.3 and 0.6 mg/kg/day), but not in the rats treate d with low doses of fentanyl (0.0, 0.006, 0.06 mg/kg/day). In addition, posthoc anal yses indicated that the elevati ons in brain reward thresholds lasted 36 hours in the rats chr onically treated with 0.6 mg/kg of fentanyl per day and only 12 hours in rats chronically treated with 0.3 mg/kg of fentanyl per da y. The dose of fentanyl did not affect the response latencies asso ciated with spontaneous fentanyl withdrawal (Table 3-3; Dose: F4,25=1.83, n.s.;). There was an effect of time on the response latencies (Time: F7,175=3.79,

PAGE 72

72 P=0.0007), but this effect was not dependent on the treatment conditions (Dose x Time interaction: F28,175=0.99, n.s.). Di scontinuation of fentanyl ad ministration induced more somatic signs that lasted for a l onger duration in rats treated with high doses of fentanyl than in rats treated with low doses of fentanyl (Fi gure 3-2B; Dose x Time interaction: F28,175=13.98, P<0.001). Posthoc analyses (level of significance, P<0.05) indicated that the discontinuation of fentanyl administration increased somatic sympto ms in rats treated with the high doses of fentanyl (0.06, 0.3, 0.6 mg/kg/ day) but not in the rats treated w ith a low dose of fentanyl (0.0 and 0.006 mg/kg/day). The increases in somatic sympto ms lasted 72 hours in the rats that were chronically treated with 0.6 mg/kg of fentanyl per day, 48 hours in rats treated with 0.3 mg/kg of fentanyl per day, and only 12 hours in rats tr eated with 0.06 mg/kg of fentanyl per day. Experiment 2: Effect of Fentanyl Exposure Dura tion on Precipitated Fentanyl Withdrawal Mean ( S.E.M.) absolute brain reward th resholds before minipump-implantation for salineand fentanyl-treated (0.3 or 0.6 mg/kg/day) rats were 143.25 20.84, 111.71 14.11, and 97.47 7.21 A, respectively [F2,13=3.45, n.s.]. Mean ( S.E.M.) absolute response latencies for salineand fentanyl-tre ated (0.3 or 0.6 mg/kg/day) rats were 3.20 0.10, 3.08 0.07, and 3.27 0.13 seconds, respectively [F2,13=0.74, n.s.]. Statis tical analyses indicated that there was an effect of fentanyl dose (Table 3-4; Treatment: F1,9=35.10, P<0.001) and time after pump implantation (Time: F4,36=14.17, P<0.001) on plasma fe ntanyl levels. Plasma fentanyl levels increased up to the 24 hour time point and then stab ilized. Statistical analys es indicated that there was a significant effect of fentanyl dose and duration of exposure on naloxone-induced elevations in brain reward th resholds (Figure 3-3A; Time x Treatment interaction: F10,65=4.37, P<0.001) and increases in somatic withdrawal signs (Figure 3-3B; Time x Treatment interaction: F10,65=20.10, P<0.001). Posthoc analyses (level of significance, P<0.0 5) revealed that at the 8 hour time point naloxone elevated the brain reward th resholds of the rats ch ronically treated with

PAGE 73

73 fentanyl (0.3 and 0.6 mg/kg/day) compared to the rats chronically treated with vehicle. At this time point, naloxone elevated the brain reward th resholds of the rats treated with 0.3 and 0.6 mg/kg/day of fentanyl to a similar degree. Af ter the 8 hour time point (24 -168 hours), naloxone induced higher elevations in brain reward thres holds in rats treated w ith 0.6 mg/kg of fentanyl per day compared to rats trea ted with 0.3 mg/kg of fentanyl per day. In both fentanyl groups, naloxone induced higher elevations in brain reward thresholds at the 24 hour time point than at the 8 hour time point. The naloxone-i nduced elevations in brain rewa rd thresholds did not further increase after the 24 hour time point. Maximum pl asma fentanyl levels and maximum naloxoneinduced elevations in brain rewa rd thresholds were achieved at the same time point (24 hours after pump implantation). Thus, a more prol onged exposure period di d not potentiate the naloxone-induced elevations in br ain reward thresholds. In both fentanyl groups, the number of naloxone-induced somatic signs gradually increa sed up to the 120 hour time point. Naloxone did not increase the somatic signs in sa lineor fentanyl-treated rats at the 8 hour time point. In the rats treated with fentanyl, nal oxone induced more somatic signs at the 24 hour time point than at the 8 hour time point and more signs at the 120 hour time point than at the 24 hour time point. From the 24 hour time point to the 168 hour time point, naloxone induced more somatic signs in the rats treated with 0.6 mg/kg of fentanyl per day than in the rats treated with 0.3 mg/kg of fentanyl per day. There was no effect of fentanyl dose and exposure period on the response latencies during naloxone-precipit ated fentanyl withdrawal (Table 3-5; Dose: F2,13=1.42, n.s.; Time: F5,65=2.34, n.s.; Dose x Ti me interaction: F10,65=1.68, n.s.). Discussion The results presented in this study indicate th at the negative affective state and somatic withdrawal signs associated with naloxone-precipitated and spontan eous fentanyl withdrawal are dose dependent. Discontinuation of the administration of high doses of fentanyl leads to a more

PAGE 74

74 severe negative emotional state and increased so matic withdrawal signs compared to when low doses of fentanyl are administered. In additi on, it was shown that the duration of fentanyl administration does not have an effect on th e negative emotional state associated with precipitated fentanyl withdrawal These findings suggest that th e dose (or more precisely, the plasma concentration) of fentanyl better pred icts the negative emotional state of fentanyl withdrawal than the administration period. To our knowledge, this is the first study to inves tigate the effect of th e dose of fentanyl on the affective and somatic fentanyl withdrawal signs. In the first experiment, the minipumps delivered 0.006, 0.06, 0.3, or 0.6 mg/kg of fentanyl pe r day for 14 days which resulted in steady state plasma fentanyl levels of approxim ately 0.02, 0.3, 1.3, 3.2 ng/mL, respectively. Clinical studies indicate that the minimum effective concen tration of fentanyl that reduces acute pain in non-dependent patients ranges from 0.2 – 1.2 ng/mL (Gourlay et al., 1988). Treatment of patients with a single transdermal fentanyl patch, 100 g/ hour, leads to plasma fe ntanyl levels of ~ 1.8 ng/mL (Varvel et al., 1989). The development of to lerance to the analgesi c effects of fentanyl can lead to a rapid escalation in the dose of fentanyl administere d. Fentanyl levels ranging from 20 170 ng/mL have been reported in patients chroni cally treated with fent anyl (Leuschen et al., 1993; Bleeker et al., 2001). The plasma fentanyl levels in the presen t study are similar to those in patients who have been treated for pain with fent anyl for a short amount of time and have little to no tolerance to the analgesi c effects of fentanyl. The results of the first experi ment indicated that the defi cit in brain reward function associated with precipitated and spontaneous fe ntanyl withdrawal is dependent on the dose of fentanyl administered. The hi ghest dose of naloxone, 0.03 mg/kg, elevated the brain reward thresholds of the rats treated w ith the highest doses of fentanyl (0.3 and 0.6 mg/kg/day), but not

PAGE 75

75 of the rats treated with lower doses of fenta nyl (0.006 and 0.06 mg/kg/day) or vehicle. The same pattern of results was detect ed after the discontinuation of fentanyl administration. Discontinuation of the administration of the highe st doses of fentanyl (0.3 and 0.6 mg/kg/day) elevated the brain reward thresholds and discont inuation of the administration of the lower doses (0.0, 0.006 and 0.06 mg/kg/day) did not elevate brain reward thresholds It is interesting to note that somatic signs were detected at lower fent anyl doses than the aff ective withdrawal signs. Naloxone (0.03 mg/kg) increased somatic signs in rats treated with 0.06 mg/kg of fentanyl and the discontinuation of the administration of 0.06 mg /kg of fentanyl per day increased the somatic withdrawal signs. Affective withdrawal signs were not detected in rats treated with 0.06 mg/kg of fentanyl per day. This is in line with a prev ious study in which we reported that a very low dose of naloxone (0.003 mg/kg) increases somati c signs in rats treated with 1.2 mg/kg of fentanyl per day, but does not elev ate brain reward thresholds in rats treated with the same dose of fentanyl (Bruijnzeel et al., 2006). However, it should be note d that Schulteis and colleagues reported that naloxone disrupt s behavioral responses (e.g., operant responding and ICSS responding) in morphine dependent animals at lower doses than those at which it induces somatic withdrawal signs (Schulteis et al., 1994). In addition, Higgins and Sellers concluded that naloxone disrupts operant responding at lower doses than those at which it induces the expression of somatic withdrawal signs (Higgins and Sellers, 1994). This pattern of results suggests that there is a differen tiation in the relative sensitivity of naloxone -precipitated somatic and behavioral withdrawal signs in morphine and fentanyl dependent animals. However, comparative studies that utilize morphine and fentanyl dependent animals are needed before firm conclusions can be drawn.

PAGE 76

76 The duration of the affective and the somatic withdrawal signs associated with the discontinuation of fentanyl admi nistration were also dependent on the dose of fentanyl. The deficit in brain reward function la sted 36 hours in the rats that received 0.6 mg /kg of fentanyl and only 12 hours in the rats that received 0.3 mg/kg of fentanyl. Higher fentanyl doses induce an even more prolonged fentanyl wi thdrawal syndrome. In a previ ous study we reported that the brain reward thresholds in rats are elevat ed for 48 hours after the discontinuation of the administration of 1.2 mg/kg of fentanyl per day for 14 days (Bruijnzeel et al., 2006). Taken together, the data presented here indicate that the affective and so matic fentanyl withdrawal signs are dose dependent; high fentanyl doses indu ce a more severe and prolonged withdrawal syndrome than low doses of fentanyl. The second experiment investigated the effect of the dose of fentanyl and the exposure duration on the development of fentanyl dependen ce. The results of this study confirmed that high doses of fentanyl mediate a more severe opioid withdrawal syndrome than low doses of fentanyl. Naloxone (0.03 mg/kg) in duced higher elevations in brai n reward thresholds and more somatic signs in rats treated w ith 0.6 mg/kg of fentanyl per day than in rats treated with 0.3 mg/kg of fentanyl per day. The results of this st udy also indicated that plasma fentanyl levels increased during the first 24 hours and then remained stable for the remainder of the experiment. A similar time-course was observed for the eff ect of naloxone on brain reward thresholds. Naloxone induced maximum elevati ons in brain reward thresholds at the same time point as when maximum plasma fentanyl levels were achieved. Prolonged exposure to fentanyl did not further increase the naloxone-induced elevations of brain reward th resholds. This suggests that fentanyl induces acute dependence and that the nega tive affective state of fe ntanyl withdrawal is mainly determined by the dose of fentanyl and not the duration of fentanyl administration. The

PAGE 77

77 maximum naloxone-induced deficit in brain reward function and the maximum naloxoneinduced increases in somatic si gns were detected at different time points. Maximum naloxoneinduced elevations in brain rewa rd thresholds were detected at the 24 hour time-point and the maximum number of naloxone-induced somatic wit hdrawal signs was det ected at the 120 hour time-point. This observation is in line with the results of a study by Easterling and Holtzman that investigated the effects of acute versus chronic morphine administration on naloxoneprecipitated affective and somatic withdrawal signs (E asterling and Holtzman, 1997). They showed that naloxone induced a similar decrease in response ra tes in the ICSS auto-titration procedure in rats acutely and chronically treated with morphi ne. However, naloxone induced a more severe somatic withdrawal syndrome in rats chronically treated with morphine than in rats acutely treated with morphine. A possible explan ation for the observation that maximum somatic withdrawal signs were detected 4 days after the maximum affective withdrawal signs were detected is that different brai n areas mediate somatic and aff ective withdrawal signs. Evidence indicates that the locus coeruleus and the periaque ductal gray at least partly mediate the somatic opioid withdrawal signs (Bozarth and Wise, 1984; Taylor et al., 1988; Maldonado and Koob, 1993). The affective opioid withdrawal signs have been suggested to be mediated by the central nucleus of the amygdala, bed nucleus of the stria terminalis, and the nucleus accumbens (Delfs et al., 2000; Gracy et al., 2001). The administration of opioids has b een shown to result in rapid changes in opioid receptor levels, G-protein coupling, and intracellular changes such as the upregulation of the cAMP pathway (Nestler, 2004; Bailey and Connor, 2005). It might be possible that opioids induce brai n site specific effects on opioid signaling pathways. This is for example supported by the observatio n that morphine differentiall y affects -opioid receptor desensitization in the locus coeruleus a nd pituitary cells (Bailey and Connor, 2005).

PAGE 78

78 In conclusion, the results presented in this study indicate that the dur ation and the severity of the dysphoria associated with the discontinuation of fentanyl administration are dependent on the dose of fentanyl administ ered. In addition, it was shown that fentanyl induces acute dependence and increasing the e xposure duration does not lead to a further increase in the naloxone-induced negative emotional state. These st udies suggest that when fentanyl is used for the treatment of chronic pain the dose of fe ntanyl should be minimized and increasing the treatment period does not further increase the negative emotional st ate associated with fentanyl withdrawal. Additional studies are needed that aid in the development of pain treatment regimen that provide maximal pain relief and mini mal dysphoria after the discontinuation of the treatment.

PAGE 79

79 Table 3-1. Effect of fentanyl dose and exposure period on plasma fentanyl levels (ng/mL). Fentanyl (mg/kg/day) Time (days) 0.006 0.06 0.3 0.6 7 0.02 0.002 0.24 0.02 1.27 0.08 3.10 0.11 14 0.02 0.003 0.24 0.02 1.37 0.11 3.20 0.13 Data are expressed as means ( S.E.M.).

PAGE 80

80 Table 3-2. Effect of fentanyl dose on the respon se latencies (percentage of baseline) associated with naloxone-precipitate d fentanyl withdrawal. Fentanyl (mg/kg/day) Naloxone (mg/kg) 0 0.006 0.06 0.3 0.6 0 98.94.5 103.14.6 90.54.7 93.15.5 90.54.2 0.00 88.72.2 101.92.8 91.54.9 93.02.7 91.93.7 0.01 95.13.9 97.57.4 91.43.7 93.44.1 95.83.3 0.03 97.98.1 102.21.8 92.72.8 94.76.1 97.02.9 Data are expressed as means ( S.E.M.). Table 3-3. Effect of fentanyl dose on the respon se latencies (percentage of baseline) associated with spontaneous fentanyl withdrawal. Fentanyl (mg/kg/day) Time (h) 0 0.006 0.06 0.3 0.6 6 105.2 2.4 95.2 2.7 106.8 2.4 101.6 3.5 106.2 3.2 12 100.7 3.6 95.2 3.6 109.7 1.6 105.2 3.1 102.6 2.6 24 104.2 2.8 93.3 3.1 103.2 4 105.7 6.3 102.0 4.7 36 99.1 3.4 96.4 4.1 103.4 1.2 102.9 4.5 96.6 6.1 48 94.4 3.1 91.1 4.6 101.1 2.1 99.0 2.9 94.9 2.5 72 97.2 2.8 97.4 5.2 93.9 5.9 99.6 2.6 95.3 3.0 96 97.5 2.0 105.5 3.8 103.3 2.1 105.0 3.4 99.2 1.5 120 97.9 2.5 95.0 4.7 104.1 1.2 103.2 2.5 100.1 1.8 Data are expressed as means ( S.E.M.). The osmotic pumps we re explanted after 14 days.

PAGE 81

81 Table 3-4. Effect of fentanyl dose and exposure period on plasma fentanyl levels (ng/mL). Fentanyl (mg/kg/day) Time (h) 0.3 0.6 8 0.34 0.09 1.21 0.39 24 1.56 0.36 2.62 0.17 48 1.42 0.28 2.41 0.11 72 1.20 0.05 2.53 0.20 120 1.26 0.06 2.52 0.19 Data are expressed as means ( S.E.M.). Table 3-5. Effect of fentanyl dose and exposure peri od on the response latenc ies (percentage of baseline) associated with naloxone-p recipitated fentanyl withdrawal. Fentanyl (mg/kg/day) Time (h) 0 0.3 0.6 8 99.8 3.9 98.3 4.3 109.7 3.9 24 97.4 4 95.0 5.8 98.0 3.7 48 102.8 1.9 94.6 4.9 111.2 5.4 72 102.8 4 101.4 3.8 105.8 5.2 120 100.7 3.4 97.6 4.7 101.4 3.2 168 104.4 4.1 93.8 2.3 91.5 7.5 Data are expressed as means ( S.E.M.). Naloxone (0.03 mg/kg, sc) was administered approximately 15 minutes prior to each time point.

PAGE 82

82 0 0.003 0.01 0.03 60 80 100 120 140 160 180 200 220 Saline Fentanyl_0.006 Fentanyl_0.06 Fentanyl_0.3 1A.A A BFentanyl_0.6 Naloxone (mg/kg, sc)Reward thresholds (%) 0 0.003 0.01 0.03 0 2 4 6 8 10 12 14 16 18 20 22 241B.A B C A B BNaloxone (mg/kg, sc)Somatic withdrawal score Figure 3-1. Effect of fentanyl dose on the el evations in brain rewa rd thresholds (A) and increases in somatic withdrawal signs (B ) associated with naloxone-precipitated fentanyl withdrawal (n = 6 / group). Brain reward thre sholds are expressed as a percentage of the pretest day values. Lette r signs (A, B, C, Tukey post-hoc multiple pairwise comparison) indicate: A is si gnificantly higher than B and C; B is significantly lower than A and significantly hi gher than C; C is significantly lower than A and B; all letters indicate a significan t elevation in brain re ward thresholds or increase in somatic signs compared to the vehicle group treated with the same dose of naloxone. The criterion for significance was 0.05. Data are expressed as mean (S.E.M.).

PAGE 83

83 6 12 24 36 48 72 96 120 60 80 100 120 140 160 180 200 220 Saline Fentanyl_0.006 Fentanyl_0.06 Fentanyl_0.3 2A.Fentanyl_0.6 A A A A* B BHours post pump explantationReward thresholds (%) 6 12 24 36 48 72 96 120 0 2 4 6 8 10 12 14 16 18 20 22 242B.A A A A A A* B B B B B C CHours post pump explantationSomatic withdrawal score Figure 3-2. Effect of fentanyl dose on the el evations in brain rewa rd thresholds (A) and increases in somatic withdrawal signs (B ) associated with spontaneous fentanyl withdrawal (n = 6 / group). Brain reward th resholds are expressed as a percentage of the reward thresholds obtained on the da y prior to minipump explantation. Letter signs (A, B, C, Tukey post-hoc multiple pairwise comparison) indicate: A is significantly higher than B a nd C; B is significantly lo wer than A and significantly higher than C; C is significantly lower than A and B; all letters indicate a significant elevation in brain reward th resholds or increase in somatic signs compared to the vehicle group at the same time point post pump explantation. A* i ndicates elevations in brain reward thresholds or increases in the somatic withdrawal signs compared to rats treated with vehicle, 0.006 mg/kg of fentanyl or 0.06 mg/kg of fentanyl. The criterion for significance was 0.05. Data are expressed as mean (S.E.M.).

PAGE 84

84 0 20 40 60 80 100 120 140 160 180 60 80 100 120 140 160 180 200 220 Saline Fentanyl_0.3 Fentanyl_0.6 3A.A A,# A A,# A,# A,# A,# B,# B B B,# BHours post pump implantationReward thresholds (%) 0 20 40 60 80 100 120 140 160 180 0 2 4 6 8 10 12 14 16 18 20 22 24 3B.A,# B,# A,# A,*A,+ A,+ B,# B,# B,+ B,+Hours post pump implantationSomatic withdrawal score Figure 3-3. Effect of fentanyl dose and duration of exposure on th e elevations in brain reward thresholds (A) and increases in somatic withdrawal signs (B) associated with naloxone-precipitated fentanyl withdrawal (vehicle and 0.6 mg/kg of fentanyl per day, n = 5 per group; 0.3 mg/kg of fentanyl per day, n = 6). Brain reward thresholds are expressed as a percentage of the pretes t day values. Letter signs (A and B, Tukey posthoc multiple pairwise comparison) indicat e: A is significantly higher than B; both A and B are significantly different from the vehicle control group. Pound signs (#) indicate significant increase s in the withdrawal signs compared to the 8 hour time point; asterisks (*) indicate an increase compared to the 24 and 48 hour time points and fewer signs compared to the 120 and 168 hour time points; plus signs (+) indicate an increase compared to the 24, 48 and 72 hour time points. The criterion for significance was 0.05. Data are expr essed as mean (S.E.M.).

PAGE 85

85 CHAPTER 4 EFFECTS OF DOSE AND EXPOSURE DURATION ON THE DEVELOPMENT OF ANALGESIC TOLERANCE OF FENTANYL IN RATS Introduction In patients who are continuously exposed to fe ntanyl, some level of analgesic tolerance and hyperalgesia is frequently observed (Arnold et al., 1990; Anand and Ar nold, 1994; Cooper et al., 1997; Chia et al., 1999; Eilers et al., 2001; Cho et al., 2007; Du mont et al., 2007). Fentanyl tolerance and hyperalgesia has also been widely studied in rats (Novack et al., 1978; Paronis and Holtzman, 1992; Thornton et al., 1997; Carter et al., 2000; Choe and Smith, 2000; Laulin et al., 2002) and in mice (Nakamura et al., 2008; Sirohi et al., 2008). Following seven days of fentanyl infusion, a dose-dependent reduction in mu-opioid receptor density was observed, which was shown to affect the magnitude of tolerance (Sirohi et al., 2008 ). Furthermore, a correlation between tolerance to fentanyl in animals and humans has b een suggested (Choe and Smith, 2000). The development of tolerance to the analgesi c effects of fentanyl can lead to a rapid escalation in the dose of fentanyl required to achieve appropriat e pain management (Milligan et al., 2001). Approximately half of the cancer patien ts converted to transdermal fentanyl from other opioid agents required incr eased dosages after initial application of the patch (Muijsers and Wagstaff, 2001). Fentanyl levels ranging from 20 170 ng/mL have been reported in patients chronically treated with fenta nyl (Leuschen et al., 1993; Bleeker et al., 2001), which is much higher than the nave therapeutic range of 0.2-2.5 ng/mL (Dale et al., 2002). Fentanyl tolerance and physical dependence are closely related (Adriaensen et al., 2003). They complicate fentanyl’s therapeutic applic ation and may contribute to a large number of drug overdose accidents.

PAGE 86

86 As expected, animal studies have suggested that the dose of fenta nyl administrated and exposure duration of the drug played an importa nt role on the development of tolerance. However, a detailed time course of tolerance de velopment versus drug-ex posure is still lacking. This study investigated the impact of fenta nyl-dose and the duration of drug exposure on the development of analgesic tolerance in rats. Ba sed on the results obtained, there will be a clearer description of the analgesic response and to lerance development following fentanyl use. Furthermore, this study will pave the way for further research aime d at designing a more efficient therapeutic design to provide maximal pain relief. Materials and Methods Subjects Male Wistar rats (Charles River, Raleigh, NC) weighing 300-350 g at the beginning of the experiments were used. The rats were shipped to the Animal Care Services (ACS) center a week before the study started in order to reduce stre ss and to allow them adapt to the research environment. The animals were socially housed (2 per cage) in a temperatureand humiditycontrolled vivarium and maintained on a 12 hour light-dark cycle (lights off at 6 p.m.). All testing occurred during the light phase of the cycle. Food and water were available ad libitum in the home cages. All subjects were treated in accord ance with the National In stitutes of Health guidelines regarding the principles of animal care. Animal facil ities and experimental protocols were in accordance with the Association for the Assessment and Accreditation of Laboratory Animal Care (AAALAC) and approved by the Univer sity of Florida Ins titutional Animal Care and Use Committee. Drugs Fentanyl citrate was purchased from Sigma (Sigma-Aldrich, St. Louis, MO, USA) and dissolved in sterile saline (0.9% sodium chloride). Fentanyl was administered s.c. using osmotic

PAGE 87

87 minipumps (Durect Corporation, Cupertino, CA ) for 14 days. The doses of fentanyl are expressed as the salt weight. Thermally-Induced Response Observation Procedure This operant assay measures a cortically medi ated response which is thought to be a better predictor of cerebral manifestations of pain than reflexes or neur onal firing patterns in the spinal dorsal horn. Thermal pain sensitivity was m easured with the Thermally-Induced Response Observation System (TIROS) invented and constructed by Dr. Andre Mauderli in the Department of Prosthodontics, University of Fl orida (Figure 4-1). The system consists of hardware (test chamber with thermally cont rolled floor, signal pr ocessing module, and temperature control modules) and software. It is a shuttle-box-like test that measures escape behavior induced by thermal nociceptive stimulati on (Mauderli et al., 2000; Vierck et al., 2005). One side of the operant escape apparatus consis ts of a thermally controlled surface and the opposite side of an escape area which is at a neut ral temperature. The animals are free to move between the two areas. Five minutes after each es cape a bright light is turned on in the escape area (3000 ft-candles). The bright light is aversive to rodents and prompts them to return to the darker area with the thermal floor. The light is turned off when the animals occupy the stimulus surface. Previous work has shown that the anim als will shuttle between the floor that provides the thermal stimulus and the neutral escape area under these conditions. In this study, each of the individual escape latencies and the average across the entire 15 minutes trial period for each stimulus temperature were used as the basis for data analysis. It is important to assure that the measured responses are indeed escape and not avoidance or regular traveling behavior. Avoidance is defined as preemp tive (based upon learning from previous experience) utilization of the neutral floor, independen t of whether the thermal floor currently elicits pain or not. A voidance learning is minimized by randomly changing the stimulus

PAGE 88

88 temperature and by the frequent in clusion of non-painful stimulus temperatures. The temperature of the thermal floor is controlled by Peltier devices (solid state h eat pumps), which is altered in a rapid and unpredictable pa ttern while the animal o ccupies the escape area. The unpredictability of the stimulus condition and the fact that prol onged occupation of the escape area is made unattractive by the bright light mi nimize avoidance behavior. Up to four different temperatures can be used throughout each 15 minute trial. Therefore, each test period provides a stimulusintensity response function. Abse nce of a stimulus-intensity-response relationship can be an indicator of avoidance behavior or an animal’s regular travel ing behavior. A shift of the stimulus-intensity-response curve is interpreted as an increase or decrease in pain sensitivity. Training The animals learned to respond in the operant test over the course of approximately 10 training sessions. A large portion of training occurred with the s timulus surface at nociceptive cold temperatures (10-15 C). The use of cold ra ther than hot temperatures eliminates the risk of burn injuries to animals which ha ve not yet learned to escape. H eat stimuli were introduced once the animals had learned how to escape the stimulus. Training was complete once the learning curve had flattened, i.e., escape latencies at a gi ven temperature remained similar from session to session. Data collection Test sessions were 15 minutes each and incl uded neutral (32-36 C) and nociceptive (4446 C) temperatures. Each test included one neutral temperature and typically one nociceptive temperature. The unoccupied pl ate was assigned to the thermal stimulus after the animal had settled down (i.e., there was no trav eling for 30 seconds). Data were automatically collected into a spreadsheet file by a computer. Each animal underwent a test session of 15 minutes.

PAGE 89

89 Tail-Flick Measurement A tail-flick nociceptive assay was also employed (Tail Flic k Analgesia Meter, 0570–001L, Columbus Instruments Int Corp, Columbus, OH) to assess the nociceptive response in the control and drug-treated rats (D'Amour and Smith, 1941) Radiant heat was ap plied using a shuttercontrolled lamp as a heat source focused on a spot located 6.5 to 7.5 cm from the tip of the tail. Thirty minutes before each experiment, the rats will be moved to the experimental room to acclimatize them to the surroundings. The measurement of tail-flick latency was defi ned as the latency for removal of the tail from the onset of the radiant heat. A digital response time indicator with a resolution of 0.1 second measured the time between activation of the light beam and the tailflick. The intensity of the stimulus was adjusted so that the baseline will be approximately 2.5 5 seconds. A maximum latency (cut-off time) of 15 seconds was used to avoid tissue damage of the tail. The baseline was determined on day 1 before the administration of the drug by using the median of 3 tail-flick tests separated by 20 seconds for each rat. The antinociception data obtained by tail-flic k tests were expresse d as a percent of maximum possible effect (%MPE), which is calcul ated: [(T1-T0)/(T2-T0)] 100 where T0 is the tail-flick latency time before the treatment on day 1, T1 is the test latency time, and T2 is the cutoff time, arbitrarily set at 15 seconds. The difference between the cut-off time (T2) and the animal’s baseline reaction time (T0) was taken to be 100% effect. Experimental Design The study was performed with the approval of the IACUC of the University of Florida (Protocol # E656). To minimize the influence of circadian rhythms, all experiments began between 8:30 and 10:30 AM. The rats were prep ared with 14-day minipumps that contained saline (n = 10, but one was lost during pump implantation surgery, and another was lost during

PAGE 90

90 pump explantation surgery) or fentanyl (0.6 mg/kg/day or 0.3 mg/kg/day, n=10 per group). In order to investigate the development of analgesic tolerance, pain sensitivity of the animal was measured 8, 12, 20, 32, 48, 72, 96, 144, and 192 hours after implantation of the minipumps. After the analgesic responses stabilized (8 da ys postpump implantation), the minipumps were removed to investigate fentanyl-induced hype ralgesia 6, 12, 24, 36, 48, 72, and 96 hours after explantation of the minipumps. Statistical Analyses The nociception data were analyzed usi ng a two-way repeated-measures ANOVA. The between-subjects factor was pump -content and the within-subjects factor was time. Statistically significant results in the ANOVA were followed by a Tukey multiple pairwise comparison test. The criterion for significance was set at 0.05. The st atistical analyses were performed using SAS for Windows software, version 9.13. Results Thermally-Induced Response Observation TIROS is a novel invention constructed by Dr Andre Mauderli, which is expected to significantly advance pain-sensitiv ity measurements (Mauderli et al., 2000; Vierck et al., 2005). Therefore, a careful initial evaluation of the system was conducted with the prototype available to us. Small software flaws of TIROS were no ticed while pre-testing the machine. For the behavior study, in order to assure that the measured responses were indeed escape and not avoidance or regular traveling behavior, multiple temperature stimuli were tested on all animals before the main analgesic study started to determ ine their impact on the escape behavior (Table 4-1). In summary, no avoidance behavior was de tected. The mean return latencies from the neutral temperatures to the stimulus temperatur es varied from 32.7 to 43.4 seconds for the testing protocols. There was no significant difference in the return latenc ies among the different testing

PAGE 91

91 protocols [F5,154=2.05, n.s.]. The testing animals we re sensitive to the cold stimulus. The escape latency from 10 C stimulus was 18.3 1.6 seconds (mean S.E.M.) which was significantly [F5,154=2.89, P=0.016] shorter than the escape latencies (the means varied from 26.9 to 33.1 seconds) from other stimulus temperat ure (from 15 to 46 C). For hot temperature stimulus (44, 45, and 46 C), we did observe a trend of decreasing in the escape latency (28.3 2.9, 26.9 4.2, and 27.7 2.1 seconds respectively) co mpared to the escape latencies, 30.2 2.9 and 33.1 3.5 seconds, of the neutral (36 C) or mild cold (15 C) stimuli respectively. But unfortunately, within the thermal nociceptive te sting temperature range (from 15 to 46 C), no significant stimulation effect of the hot temperatures (44, 45, and 46 C) on the escape latency was detected [F4,131=0.63, P=0.642]. Therefore, we suspected that TIROS, used at the current setting, was not powerful enough to differentiate the escape beha vior caused by a hot painful stimulus from the regular traveling behavior. Furt her adjustment may be needed to clarify this issue. Therefore, TIROS was not used in the further drug-response tests. Tail-Flick Measurement of Fentanyl Analgesic Effect Over Time Mean ( S.E.M.) absolute tail-flick latencie s before minipump-implantation for salineand fentanyl-treated (0.3 or 0.6 mg/kg/day) rats were 4.30 0.11, 3.91 0.22, and 4.11 0.20 seconds, respectively [F2,27=1.10, n.s.]. The complete time courses of ob served antinociception of saline and different fentanyl doses (0.3 and 0.6 mg/kg/day) treatments were expressed as a percent of maximum possible effect (%MPE) (Figure 4-2). Saline and fentanyl were administered s.c. via mini-pump implantation at time 0 and drug withdrawal was conducted via pump explantation at 192 hour post-pump implanta tion. Data were expressed as mean standard deviation (S.D.). The antinociceptive effects of fentanyl reached maximum at 12 hour post-pump implantation, and then it declined rapidly a nd reached a plateau with in the next two days. Fentanyl-induced hyperalgesia afte r discontinuation of drug was also suggested in the rats treated

PAGE 92

92 with high dose (0.6 mg/kg/day) of fentanyl. In addition, other severe somatic withdrawal syndromes (e.g., irritability, diarrhea, abnormal posture) were noticed. Statistical analyses indicated that there was a signi ficant effect of fentanyl dos e and duration of exposure on rat nociceptive responses during the drug treatment (Figure 4-3A; Time x Treatment interaction: F16,208=17.38, P<0.001). Posthoc analyses (level of si gnificance, P<0.05) revealed that at the 8 hour time point antinociception of the rats chronically treate d with fentanyl (0.3 and 0.6 mg/kg/day) was significant compar ed to the rats chronically treat ed with vehicle. At this time point, there was no difference in th e degree of antinociception obser ved in the rats treated with 0.3 and 0.6 mg/kg/day of fentanyl. At the 12 a nd 24 hour time points, an tinociception in rats treated with 0.6 mg/kg of fentanyl per day was highe r compared to rats treated with 0.3 mg/kg of fentanyl per day. After the 24 hour time poi nt, there was no significant difference in antinociception between the two fentanyl doses. Furthermore, after the 48 hour time point, antinociception of the rats tr eated with fentanyl (0.3 a nd 0.6 mg/kg/day) was no longer significant compared to the rats treated with vehi cle. Statistical analyses of the nociception data after discontinuation of fentanyl did not reveal a signi ficant effect of the previous fentanyl treatment on rat’s nociceptive responses (Fig ure 4-3B; Time x Treatment interaction: F12,150=1.28, n.s.), even though the pattern of hyperalg esia for about one day was suggested in the rats treated with high dose (0.6 mg/kg/day) fentanyl. Discussion The results presented in this study indicate th at analgesic tolerance of fentanyl during continuous drug administration is dependent on both dose and duration of exposure. Discontinuation of the drug in rats receiving hi gh dose fentanyl also suggests hyperalgesia accompanied with other withdrawal syndromes. Continuous s.c. infusion of 0.3 – 0.6 mg/kg/day fentanyl only produced a short an algesic effect in rats. Maximu m antinociception appeared to

PAGE 93

93 occur at 12 hour post-pump implantation, which was in line with when fentanyl plasma levels reached steady state (Table 3-4). Continuous expos ure to fentanyl induced the rapid development of drug tolerance; no significant analgesia was pr esent after 48 hours of pump implantation. This observation is in agreement with previous st udies that continuous s. c. infusions of 1.2– 2.4 mg/kg/day fentanyl rendered infa nt rats tolerant to the analge sic effects of fentanyl and no analgesia was present from the fentanyl released from the pumps 72 hours after implantation of the pumps (Thornton et al., 1997; Thornton and Smith, 1998; Choe and Smith, 2000). The literature also suggested that fe ntanyl tolerance was not limited to a single preweanling age and that might a lower infusion dose of fentanyl was required to induce tolerance as rats aged (Choe and Smith, 2000). The steady state plasma fentanyl levels in this study were approximately 1.5 and 3 ng/mL for the 0.3 and 0.6 mg/kg/day fentanyl minipump treatment respectively. Clinical studies indicate that the minimum effective concen tration of fentanyl that reduces acute pain in non-dependent patients ranges from 0.2 – 1.2 ng/mL (Gourlay et al., 1988). Treatment of patients with a single transdermal fentanyl patch, 100 g/ hour, leads to plasma fe ntanyl levels of ~ 1.8 ng/mL (Varvel et al., 1989). Interestingly, the pl asma fentanyl levels in the present study matched those reported in nave patients who would very likely re quire increased dosages after initial application of the patch (Muijsers and Wagstaff, 2001). Similar correlations between fentanyl animal tolerance and clinical toleranc e were also suggested by Choe’s study (Choe and Smith, 2000). Fentanyl-induced hyperalgesia was s uggested, though not statis tically significant, after discontinuation of the high dose of fentanyl. This is consistent with Celerier’s observation that long-lasting hyperalgesia was induced by fent anyl in rats (Celerie r et al., 2000) and other commonly observed cases of fentanyl induced post operative pain in human patients (Chia et al., 1999).

PAGE 94

94 Sedation syndrome was noticed in rats during th e early stage of fenta nyl treatment. Severe somatic withdrawal syndromes, such as irri tability, diarrhea, abnor mal posture, abdominal constrictions, and escape attempts were also noticed in rats after discontinuation of 0.6 mg/kg/day fentanyl. The influence of thos e syndromes and stress on the behavioral measurements should be cautiously evaluated (Ki ng et al., 2003; Rodrigues et al., 2005; Vierck et al., 2005). Tail-flick is a reflex measurement and may be less sensitive than operant measurement to the impacts of these syndromes. Motor coordination tests should be conducted in the future to dissociate the sedati ve effect and sedative tolerance of fentanyl. Reflex responses to high rates of noxious cutaneous he ating are mainly mediated by A -fiber nociceptors; however, chronic pain appears to be mainly mediated by C-fiber nociceptors (Yeomans et al., 1996; Yeomans and Proudfit, 1996; Kimura and Kontani, 2008; Xiao and Bennett, 2008). A differential opioid modulation of t hose nociceptors has also been suggested. Reflex responses are considered to be insensitive to the attenuati on of nociceptive activity in laboratory animals by low doses of opioids (Vierck et al., 2002; King et al., 2007). Therefore, the operant testing as implemented in TIROS remains a future direction to evaluate analgesia of fentanyl in chronic treatment, even though further modifications, such as the abilities to differentiate sedation from analgesia and differentiate the escape behavior from the normal traveling behavior, are needed. In conclusion, the results presented in this study indicate that analgesic tolerance of fentanyl during continuous drug exposure is dependent on dose and duration of exposure. Discontinuation of the high dose of fentanyl may cause hyperalgesia. Operant escape pain measurement is still an appropriate method to preci sely evaluate the analgesic effects of chronic opioid treatments. These studies suggest that when fentanyl is used for the treatment of chronic pain, the dose of fentanyl should be carefully adjusted according to the development of drug

PAGE 95

95 tolerance. To assist the dose adjustment, PK/PD modeling focusing on physic dependence and analgesic tolerance is warranted.

PAGE 96

96 Table 4-1. Effect of testing temperatur e on escape and return latency (seconds) Testing Protocol (Stimulus Temperature vs. Neutral Temperature) Latency 10 vs. 32 C 15 vs. 32 C36 vs. 32 C44 vs. 36 C45 vs. 36 C 46 vs. 36 C Escape Latency 18.3 1.6 33.1 3.5 30.2 2.9 28.3 2.9 26.9 4.2 27.7 2.1 Return Latency 38.1 3.4 43.4 3.4 32.7 2.3 35.4 2.4 39.3 3.3 34.5 1.9 Data are expressed as means ( S.E.M.) (n=30).

PAGE 97

97 Figure 4-1. The TIROS (Thermally-Induced Re sponse Observation System) consists of hardware (test chamber with thermally c ontrolled floor, signa l processing module, and temperature control modules) and softwa re (A). It uses a shuttle box with two compartments. The floor of one compartment c onsists of a thermal plate which is held at the desired test temperat ure by internal water circulat ion. The floor of the other compartment is thermally neural. A bright light (aversive to rodents) is activated when the escape area is occupied (B). Data collection is fully automated by computer.

PAGE 98

98 1 2 2 4 36 4 8 60 7 2 8 4 96 1 0 8 120 13 2 1 4 4 156 1 6 8 180 19 2 2 0 4 216 2 2 8 240 25 2 2 6 4 276 288 -30 -20 -10 0 10 20 30 40 50 60 70 80 Saline Fentanyl (0.3 mg/kg/day) Fentanyl (0.6 mg/kg/day) Time (hr)MPE (%) Figure 4-2. The complete time courses of obser ved antinociception of saline and different fentanyl (0.3 and 0.6 mg/kg/da y) treatments. Saline and fe ntanyl were administered subcutaneously via mini-pump implantati on at time 0 and drug withdrawal was conducted via pump explantation at 192 hour post-pump implantation. Antinociception was expressed as a percen t of maximum possible effect (%MPE). Data were expressed as mean standard deviation (S.D.) (n = 10 per group). Drug withdrawal

PAGE 99

99 24 48 72 96 120 144 168 192 -10 0 10 20 30 40 50 60 70 Saline Fentanyl_0.3 Fentanyl_0.6 3A.A A A B B A A A A*Hours post pump implantation%MPE 0 12 24 36 48 60 72 84 96 -15 -10 -5 0 5 10 15 20 3B.Hours post pump explantation%MPE Figure 4-3. Effect of fentanyl dose and exposure duration on rat analgesia during drug exposure (A) and hyperalgesia after dr ug withdrawal (B). Antinocic eption was expressed as a percent of maximum possible effect (%MPE) Letter signs (A and B, Tukey Posthoc multiple pairwise comparison) indicate: A si gnificantly differs from B; both A and B are significantly different from the vehicl e control group; A* significantly differs from vehicle. The criterion for significan ce was 0.05. Data are expressed as mean (S.E.M.) (n = 10).

PAGE 100

100 CHAPTER 5 PHARMACOKINETIC-PHARMACODYNAMIC MODELING OF PHYSICAL DEPENDENCE AND ANALGESIC TOLERANCE OF FENTANYL IN RATS Introduction Fentanyl physical dependence and toleran ce are common consequences of chronic drug exposure and are closely related (Adriaensen et al., 2003). They account for the strong increase in abuse of fentanyl over the past years as fe ntanyl claims its popularity in pain management (Joranson et al., 2000; Gilson et al., 2004; Novak et al., 2004). Drug dependence, tolerance and the wide concerns of the potential drug overu se and abuse complicate the drug’s application. Improper selection of patients, products, and dosi ng has been associated with a large number of drug overdose deaths (Kuhlman et al., 2003; Li lleng et al., 2004; FDA, 2007a; FDA, 2007b). Selection of dosing and products mainly depends on the drug’s PK/PD properties, which in this case are highly affected by the patient’s dependence/tolerance pr ofile. PK/PD modeling approaches have proven to be effective in expl aining the drug-response relation, and to help dose optimization to improve drug efficiencies and re duce the risk of undesire d effects (Derendorf and Meibohm, 1999; Derendorf et al., 2000). Fenta nyl PK, its PD of physical dependence, and its PD of analgesic tolerance in ra ts have been investig ated in our previous studies (Chapters 2, 3, and 4). The purpose of this study was to develop PK/PD models to describe physical dependence and analgesic tolerance following fentanyl exposu re in rats. Mechanism-based PK/PD structure models would link the PK and PD information and enable us to pred ict the dependence and tolerance over time for any designed dose regi men. Simulation studies with the constructed PK/PD models would allow us to evaluate any therapeutic design of this drug which will help in selection of the appropriate pa tient, product, and dosing regimen.

PAGE 101

101 Materials and Methods PK/PD Modeling of Physical Dependence of Fentanyl The methods and results of physical dependen ce time courses of fentanyl were described previously in Chapter 3. To describe the time course of physical dependence for a given dose regimen, a sequential PK/PD modeling approach was applied. PK modeling of fentanyl administrated subcutaneously with osmotic minipump To describe the observed plas ma fentanyl levels during pump implantation, a lag-time parameter describing the delay in drug release fro m the implanted pump was integrated into the linear first-order absorption two-compartment PK model constructed in Chapter 2. The structural PK parameters, i.e., the fi rst-order absorption rate constant (ka), the intercompartmental clearance (Q), and the volumes of distribut ion of compartments 1 and 2 (V1 and V2), were fixed as the values estimated from the s. c. injection data in Ch apter 2 (Table 2-4). The drug clearance (CL) parameter was refitted with the infusion data. The concentration-time course of fentanyl administrated s.c. with minipump was then modeled using the following differential equations: () dAP I NFRIND dt (5-1) (0) (0) dA I NFRINDkaA dt (5-2) (1) -()(1) (0)-()(1)()(2) 112 dACLQQ A kaAAA dtVVV (5-3) (2) ()(1)()(2) 12 dAQQ A A dtVV (5-4) (1) 1p A C V (5-5)

PAGE 102

102 where A(P), A(0), A(1), and A(2) are the drug amounts in the mi nipump, absorption site, central compartment, and peripheral compartment respectively; INFR is the drug release rate from the minipump; IND is the indicator variable for the drug rel ease and defined as 1 if drug release (i.e., implantation duration is longer than the la g time) is indicated and 0 otherwise. PK/PD modeling of fentanyl affective dependence The complete time courses of affective dependen ce together with fentanyl PK were plotted in Figure 5-1A under the assumption that the affective withdrawal syndromes induced by 0.03 mg/kg naloxone precipitation or by drug disconti nuation completely reflected the degree of rat affective dependence. As shown in the plot, th e development of affective dependence was rapid and closely correlated to the drug plasma c oncentration, and the spontaneous withdrawal syndrome after discontinuation of dosing lasted much longer than the drug predicted by our PK model. To relate the changes in fentanyl concen trations to the changes in affective dependence, various structurally different PK /PD models were tested. Since we were more interested in the basic model structure and the mean data could be tter reveal the true profile pattern, the mean time courses of drug-responses were used in th is study for model constr uction. The Scientist 3.0 software (Micromath, Saint Louis, MO) was used for the model f itting. The goodness-of-fit statistics (e.g., the correla tion between observed and predic ted data, the coefficient of determination, and the model sel ection criterion) reported from the software were used to evaluate the suitability of a model. The coeffici ent of determination is defined by the formula: 22 11 2 1()() ()iii inn obsobsobspred ii n obsobs iYYYY YY (5-6)

PAGE 103

103 where n is the number of data points, iobsY and ipredY are the ith observed and predicted data point respectively, and obsY is the mean of the observed data. The coefficient of determination is a measure of the fraction of the total varian ce accounted for by the model. It is a more appropriate measure of the goodness-o f-fit than either correlation or R-squared according to the user manual of Scientist software. The model selection criterion (MSC) reported by Scientist software is a modified AIC and is defined by the following equation: 2 1 2 1() 2 ln ()i iin obsobs i n obspred iYY p MSC n YY (5-7) where n, iobsY, i p redY, and obsY are defined as above, and p is the number of parameters in the model. The MSC gives the same ranking between m odels as the AIC and has been normalized so that it is independent of the scaling of the data points. Furthermore, the most appropriate model will be that with the largest MS C when evaluating the suitability of a model. In the following, three indirect models were studie d in detail for their appropriate ness to describe the effect of fentanyl on affective dependence (Figure 5-2): The two-effect-compartment indirect-link model (biophase equilibration model with a sigmoid Emax PD model) : As shown in Figure 5-2A, this model describes the dissociation between affective dependence and plasma drug con centration with a biophase (i.e., the site of action) distribution. Drug distribut ion to the effect site was ch aracterized on the basis of a hypothetical two-effect compartment model. Sp ecifically, the rates of the changes in the biophase drug concentrations are describe d by the following differential equations: 1 11221()()e eopeeeedC kCCkCC dt (5-8)

PAGE 104

104 2 2112()e eeedC kCC dt (5-9) where keo is a first order elimination rate consta nt in the central effect compartment, ke12 and ke21 are the first order distribution rate constant s between the two effect compartments, and Cp, Ce1, and Ce2 represent the drug concentrations in plas ma, the central effect compartment, and the peripheral effect compartment respectively. The fi nal affective dependence is then characterized with the sigmoid Emax PD model, which is of the form: max1 0 501 h e hh e E C EE E CC (5-10) where E0, which represents the baseline intrinsic response without drug presence, is set as 100 theoretically, Emax is the maximum activity of the drug, EC50 is the effect site concentration yielding half-maximal affective dependence, and h is a slope parameter (Hill factor). The indirect-response model : As shown in Figure 5-2B, this model describes the dissociation between affective dependence and plasma drug concentration via a hypothetical inhibitory neurotransmitter. The drug can reduce the inhibition effect of the neurotransmitter by stimulating its elimination. Specifically, the rate of the change in the hypothetical neurotransmitter is described by the following differential equation: max 50(1)h i inouti hhdTEC kkT dtECC (5-11) where Emax is the maximum stimulation activity of drug on the elimination of the neurotransmitter, EC50 is the drug concentration yi elding half-maximum effect, h is the Hill factor, kin is the zero-order constant for production of the neurotransmitter, kout defines the firstorder rate constant for loss of the neurotransmitter, and Ti represents the inhibitory

PAGE 105

105 neurotransmitter concentration. The final affec tive dependence is then characterized with the antagonistic sigmoid Emax PD model, which is of the form: max int 50 i i I T EE I CT (5-12) where Eint is the maximum intrinsic response without the neurotransmitter presence, Imax is the maximum inhibition activity of the neurotransmitter, IC50 is the neurotransmitter concentration yielding half-maximum inhibition effect, and is the Hill factor. Furt her, the maximum intrinsic response Eint can be set according to the theoretical baseline: max int0 50(/) (/)inout inoutIkk EE ICkk (5-13) where E0, which represents the baseline intrinsic response without drug presence, is set as 100 theoretically. The receptor sensitization model : As shown in Figure 5-2C, this model describes affective dependence as an outcome of the change in the total amount of hypothetical neuroreceptors. Drug molecules bind to the free receptors to form the drug-receptor complexes which will be eliminated with the first-order rate constant kDout. Concomitantly, the presence of drug can stimulate the synthesis of the free receptors Specifically, the ra tes of the changes in the hypothetical neuroreceptors are described by the followi ng differential equations: (1)inoffonoutdR ksCkDRkCRkR dt (5-14) onoffoutdDR kCRkDRkDR dt (5-15) where kin is the zero-order constant for the synthesis of the hypothetical free receptors, kout is the first-order rate constant fo r loss of the free receptors, kon and koff are the first order drug-receptor

PAGE 106

106 association/dissociation rate constants, kDout is the first-order elimination rate constant of drugreceptor complexes, s is the scale factor of th e drug stimulation activity, and R and DR represent the amounts of the free receptors and the drug-re ceptor complexes, respectively. The final affective dependence is then ch aracterized with the linear PD model, which is of the form: tERRDR (5-16) where Rt is the total amount of receptors. Based on the theoretical baseline, the relation between kin and kout can be further set as: 0100in outk ER k (5-17) where E0, which represents the baseline intrinsic response without drug presence, is set as 100 theoretically. PK/PD modeling of fentanyl somatic dependence The complete time courses of somatic dependenc e together with fentanyl PK were plotted in Figure 5-1B. As shown in the plot, somatic dependence, which is characterized by mixed syndromes, developed slower and further dissociated from drug concentration compared to affective dependence. The development seemed to display a two-phase pattern. However, similar to affective dependence, the spontaneous withdrawal syndrome lasted much longer than the circulating drug levels after discontinuation of dosing. To relate the changes in fentanyl concentrations to the changes in somatic depe ndence, various structur ally different PK/PD models were tested. In the fo llowing, two indirect-response mode ls were studied in detail for their appropriateness to describe the effect of fentanyl on somatic dependence (Figure 5-3): The two-phase-response model: Based on the assumption that development of somatic dependence displays a two-phase pattern as suggested by our observation, a receptorsensitization model combined with a delayed induction of the synthesis of a hypothetical

PAGE 107

107 excitory neurotransmitter was investigated (Figure 5-3A). In this model, drug molecules bind to the free receptors to form the drug-receptor comp lexes which will be eliminated with the firstorder rate constant kDout. Simultaneously, the presence of drug can stimulate the synthesis of the free receptors. At the formation of the drug-receptor complexes, the complexes can induce the production of a hypothetical excito ry neurotransmitter. The dela y in the synthesis of the neurotransmitter can be characte rized by a Weibull or logistic de lay function. Specifically, the rates of the changes in the hypothetical neurorec eptors and neurotransmitters are described by the following differential equations: (1)inoffonoutdR ksCkDRkCRkR dt (5-18) onoffoutdDR kCRkDRkDR dt (5-19) (/) _max 50(1)bta e inoutedT DR kTekTT dtETDR (5-20) where kin is the zero-order constant for the synthesis of the hypothetical free receptors, kout is the first-order rate constant fo r loss of the free receptors, kon and koff are the first order drug-receptor association/dissociation rate constants, kDout is the first-order elimination rate constant of drugreceptor complexes, s is the scale factor of the drug stimulation activity, kTin_max is the zero-order constant for the production of the neurotransmitter that can be maximally induced by the drugreceptor complex, kTout is the first-order rate constant for loss of the neurotransmitter, a and b are the scale parameter and the shape parameter of the Weibull delay function respectively, ET50 is the drug-receptor concentration yielding half-maximum induction effect on the neurotransmitter, and R DR and Te represent the amounts of the free recep tors, the drug-receptor complexes, and the hypothetical neurotransmitters respectively. The final somatic dependence is then characterized with the linear PD model, which is of the form:

PAGE 108

108 0()eErRDRRT (5-21) where r is the scale factor of th e receptor contribution; and R0, which represents the baseline free-receptor amount without drug presence, can be further defined according to kin and kout as: 0 in outk R k (5-22) The induction of the response model: As shown in Figure 5-3B, this model describes the dissociation between somatic depe ndence and plasma drug concentration via an induction of a hypothetical excitory neurotransmitter. Specifically, the rate of change in the hypothetical neurotransmitter is described by the following differential equation: _max 50 e inoutedTC kkT dtECC (5-23) where kin_max is the zero-order constant for production of the neur otransmitter that can be maximally induced by the drug, EC50 is the drug concentration yielding half-maximum induction effect, kout defines the first-order rate constant for loss of the neurotransmitter, and Te represents the excitory neurotransmitter amount. The final somatic dependence is th en characterized with the linear PD model, which is of the form: eET (5-24) PK/PD Modeling of Analgesic Tolerance of Fentanyl The methods and results of the development of analgesic tolerance of fentanyl were described previously in Chapter 4. The comp lete time courses of rat nociceptive responses together with fentanyl PK were plotted in Figure 5-4. As show n in the plot, antinociception of fentanyl reached a maximum along w ith the drug concentration, and then it declined rapidly and returned almost to the baseline with constant dr ug presence. Fentanyl-i nduced hyperalgesia after discontinuation of drug was also suggested and lasted much longer than drug after

PAGE 109

109 discontinuation of fentanyl. To re late the changes in fentanyl c oncentrations to the changes in antinociception, various structur ally different PK/PD models we re investigated. Those models were based on the three major categories review ed in Chapter 1: i.e., the cellular receptor regulation model; the feedback model via depl etion or production of endogenous modulators; and the counter-response model. Among them, a delayed counter-response model was studied in detail (Figure 5-5). The delayed counter-response model: Based on the mechanism that analgesic tolerance and rebound are the consequences of a hypothetical counter response, an indirect-response model was evaluated in which the drug-induced anti nociceptive response was combined with a tolerance compartment due to a delayed inducti on of a counter respons e (Figure 5-5A). The delay in the production of the counter res ponse can be characterized by an exponential cumulative distribution function. Specifically, th e rates of the changes in the antinociceptive response and the hypothetical c ounter response are described by the following differential equations: _max 50 inoutdRC kkR dtECC (5-25) (/) _max 50(1)ta inoutdTC kTekTT dtTCC (5-26) in which kin_max is the zero-order constant for the pr oduction of the antinociception that can be maximally induced by the drug, kout is the first-order rate constant for loss of the free receptors, kTin_max is the zero-order constant fo r the production of the hypothetical counter response that can be maximally induced by the drug, kTout is the first-order rate cons tant for loss of the counter response, a is the rate parameter of the expone ntial distribution delay function, EC50 and TC50 are the drug concentrations yielding half-maximum induction of antinoci ception and the counter

PAGE 110

110 response respectively, and R and T represent the degrees of antinociception and the counter response respectively. The final nociceptive response is then characterized w ith the linear PD model, which is of the form: ERT (5-27) Evaluation of Intermittent Dosing Strategies with Simulation Studies In the pharmacological studies described above, the impacts of continuous fentanyl exposure on physical dependence an d analgesic tolerance were i nvestigated. To evaluate the impacts of intermittent dosing strategies on thes e responses, simulation studies were performed. The same daily doses (i.e., 0.3 or 0.6 mg/kg/ day) as in the s.c. infusion studies were administrated by s.c. injection. Once a day (QD, = 24 h), twice a day (BID, = 12 h), three times a day (TID, = 8 h), and four times a day (QID, = 6 h) dose regimens were compared. The same exposure period (two weeks) was implemented. The linear two-compartment PK model constructed in Chapter 2 was adopted to simulate the concentra tion-time courses. The selected PD models were adopted for the physical dependence and analgesic tolerance simulation. As shown in Figure 5-6, the simulations were co nducted by using Trail Simulator version 2.2 (Pharsight Cor poration, Mountain View, CA). Results Characterization of the PK/PD Models of Physical Dependence As shown in Figure 5-1, the observed concen tration-time courses of plasma fentanyl during pump implantation can be described by the linear first-order absorption two-compartment PK model constructed in Chapter 2 w ith a drug release delay parameter, tlag. The populationmean estimates of tlag and refitted CL are 6.9 hours (with the 2.0% of C.V.) and 102 mL/min (with the 1.8% of C.V.). Base d on this PK model, the correl ation between the drug plasma concentration and physical dependence were quantitativel y investigated.

PAGE 111

111 The receptor sensitization model charac terized affective dependence best The fittings to the complete time-courses of affective dependence with the two-effectcompartment indirect-link model, the eliminati on of an inhibitor indirect-response model, and the receptor sensitization model are displayed in Figure 5-7. The tw o-effect-compartment indirect-link model contains 6 parameters. Th e model could successfully capture the prolonged spontaneous withdrawal response, but it slightly u nderestimated the rapid development of affective dependence during the early stage of drug exposure (F igure 5-7A). The indirectresponse model in which the drug stimulates the elimination of a hypothetical inhibitory neurotransmitter contains 8 parameters. It could also successfully capture the prolonged spontaneous withdrawal response, but it suggested that different doses would cause different developing rates and the 0.6 mg/kg/day dose would reach plateau earlier, which seems inconsistent with our observed data (Figure 5-7B). For the re ceptor sensitization model, the estimates of the rate constants kon and koff which characterize receptor association and dissociation kinetics were large, suggesting that the receptor asso ciation and dissociation kinetics are fast. This is consistent with data obtained from in vitro receptor binding studies (Boas and Villiger, 1985). Therefore the binding was further simplified with an equilibrium constant K. The model contains 4 parameters. It could su ccessfully capture the prolonged spontaneous withdrawal as well as the rapi d development of dependence (F igure 5-7C). Goodness-of-fit statistics of these three models were compared in Table 5-1. Form the model comparison, the receptor sensitization model appears to be the most suitable. Its parameter estimates were listed in the Table 5-2. The model is also the most stab le one due to the fewest number of parameters. The induction of the response model is more suitable fo r somatic dependence The fittings to the complete time-courses of somatic dependence with the two-phaseresponse model and the induction of the response model are displayed in Figure 5-8. The two-

PAGE 112

112 phase-response model contains 12 parameters. The model could successfully capture the prolonged spontaneous withdrawal response as well as the twophase development of somatic dependence (Figure 5-8A). However, the statisti cal reports of the model parameters suggested the model was over-parameterized to the observe d data. The induction of the response model is much simpler and contains only 3 parameters. The model could successfully capture the prolonged spontaneous withdrawal response as well as the prol onged development of somatic dependence (Figure 5-8B). Goodne ss-of-fit statistics of these tw o models were compared in Table 5-3. From the model comparison, the induc tion of the response model seems more suitable for the description of somatic dependence. Its pa rameter estimates were listed in the Table 5-4. The model is also much more stable according to variation of the parameter estimates. Characterization of the PK/PD Model of Analgesic Tolerance The fittings to the complete time-courses of the antinociceptive responses with the delayed counter-response model is displayed in Figure 5-9A The model contains 6 parameters. It could successfully capture the acute antinociceptive effect of the drug as well as the rapid development of analgesic tolerance during co ntinuous drug exposure. It also suggested hyperalgesia after discontinuation of the drug. In this counter -response model, an exponential cumulative distribution function was adopted to characterize the delayed anta gonistic function of the drug. To explain this delay physiologi cally, a drug-induced depletion of an inhibitory modulator of the counter response was hypothesized (F igure 5-5B). Therefore, the rate of the change in the hypothetical counter response is de scribed by the following differen tial equations modified from Equation 5-26: 501dMC kMkMM dtMCC (5-28)

PAGE 113

113 _max 50(1)inoutdTC kTMkTT dtTCC (5-29) where kTin_max, kTout, TC50, and T are defined the same as in Equation 5-26, kM is the rate constant for the production or loss of th e hypothetical inhibitory modulator, and MC50 is the drug concentrations yielding half-maximum suppres sion of the production of the modulator. The model fit the data almost identically to the e xponential delayed model (Figure 5-9B). The fitting suggested a complete suppression of the produc tion of the inhibitory modulator under our treatment conditions. Therefore, the MC50 was arbitrarily set as 0.1 ng/mL, which is approximate one tenth of the fentanyl steady state concentration of the low dose (0.3 mg/kg/day) treatment. Goodness-of-fit statistics of these two models were compared in Table 5-5 and parameter estimates of the models were listed in the Tabl e 5-6. Since the drug-i nduced depletion of an inhibitory modulator model was integrated with a potential physiological mechanism, it was selected for the simulation study. Impacts of Intermittent Dosing on Phys ical Dependence and Analgesic Tolerance With the selected PK/PD models, the impacts of intermittent dosing strategies on physical dependence and analgesic tolerance were evaluate d by simulation. The direct comparison of the impacts of QID and TID dosing ar e displayed in Figure 5-10. The plasma fentanyl concentration reached steady state over couple doses (Figure 510A). However, it took much longer time for physical dependence and analgesic tolerance to re ach the plateaus. The dos ing interval did not affect the developing rates of physical dependence and analgesic tolerance, which were also comparable to continuous dosing. However, TID dosing caused more severe affective dependence and more fluctuations than QID (Figure 5-10A). Conversely, the longer dosing interval caused slightly less of somatic de pendence (Figure 5-10B). For both affective dependence and somatic dependence, the discharg e time after discontinuation of dosing was also

PAGE 114

114 not affected by the dosing strategi es. From the tolerance study, TI D dosing reserved more of the analgesic effect of fentanyl and more fluctu ations of antinocicepti on than QID. Moreover, fentanyl-induced hyperalgesia suggested by our model after discontinuation of dosing was less with the longer dosing interval treatment (F igure 5-10C). The complete comparison of the impacts of dosing regimen on the steady state responses of affective dependence, somatic dependence, and analgesic tolera nce is displayed in Figure 5-11. It suggests that longer dosing interval will render more severe affective dependence, and less somatic dependence and analgesic tolerance. Discussion Mechanism-based PK/PD modeling for physi cal dependence and drug tolerance is considered to be difficult due to the lack of re liable and quantifiable bi omarkers of dependence, the lack of knowledge about the mechanisms underlying tolerance and dependence, and the difficulties in appropriately simplifying these comp lex physiological processes. Currently, all of the models are considered empirical or semiphysiological (Grdmark et al., 1999). Our previous studies demonstrated the connections between fent anyl plasma levels pl us exposure duration and the development of physical de pendence or analgesi c tolerance. The current study presented several semi-mechanism based PK/PD approaches to characterize the relationships among the temporal changes in affective dependence, so matic dependence, analge sic tolerance and the fentanyl plasma concentrations in up to a 14-day period of continuous drug exposure. The disassociation between the biological effect-intens ity and drug plasma concentration seen in this study usually involves a combina tion of different physiological processes such as biophase distribution, receptor kinetics, receptor regulat ion and downstream signal transduction pathways. Often it is difficult to extract and discriminate among these processes from the available PK/PD data (Jusko et al., 1995; Verotta and Sheiner, 1995; Castaneda-Hernand ez and Granados-Soto,

PAGE 115

115 2000; Yassen et al., 2005). Accordingly, various PK/PD models based on different hypothetical mechanisms were investigated. The two-effect-compartment i ndirect-link PD model, the el imination of an inhibitor indirect-response PD model, and the receptor se nsitization PD model were compared for their suitability in describing affective dependence. The two-effect-compartme nt indirect-link model explains the dissociation between affective dependence and plasma drug concentration with a biophase distribution whic h is commonly true for CNS drugs. It fit the data reasonably well except the slight estimation of the rapid development of affec tive dependence during the early stage of drug exposure. However, due to high lipophilicity of fent anyl, the distribution half-life of this drug to the rat brain is only about 6 minutes (Lotsch, 2 005). This short distribution is negligible during our long-term infusion study. The drug half-life in the hypothetical effect compartments suggested in our model is much longer (approximate 20 hours) and it also governs the time to reach the steady state of affective dependence and the ti me to return to baseline. The model suggests that the time to reach steady stat e with drug presence and th e time to return back to the baseline after drug w ithdrawal are same which was not well supported by our results. Therefore, this model may oversimplify the mech anism of affective dependence. In any of the classic indirect-response models, th e change rate of the response is governed by the elimination rate constant kout (Gabrielsson and Weiner, 2006). Sin ce the developing rate of affective dependence in the presence of drug was faster than the recovering rate of this dependence after drug removal, it seemed that the presence of dr ug should enhance the elimination rate in the indirect-response model. However, after fitting the model to the data, it showed that the 0.6 mg/kg/day dose treatment would reach plateau si gnificantly earlier than the 0.3 mg/kg/day dose which again seems contrary to our observati ons. The connection between physical dependence

PAGE 116

116 and receptor sensitization has been suggested in morphine dependent animals (Abdelhamid and Takemori, 1991; Rothman et al., 1991; Crain an d Shen, 1998; Vekovischeva et al., 2001). The receptor sensitization model constr ucted in this study related a ffective dependence proportionally to the number of hypothetical receptors. This model pr ovided the best fit to the data and was also the most stable based on the variances of the parameter estimates due to the fewest number of parameters. The two-phase-response model and the induction of the response model were compared for their suitability in describing somatic depende nce. Somatic dependence was characterized by mixed withdrawal syndromes. From our observation, some withdrawal syndromes such as facial fasciculation, blinking and ptos is developed quickly and appeared one day after the infusion started, while other syndromes such as abdomin al constrictions, escape attempts, and diarrhea appeared 2-3 days later. A possible explanati on for theses observations is that multiple mechanisms and effect sites are involved in me diating somatic withdrawal signs. The two-phaseresponse model was developed based on this reas oning. It successfully captured the two-phase development of somatic dependence as suggested by our observed data and its overall fit was also slightly better than the simple induction of the response model. However, the two-phaseresponse model involved 12 parame ters; therefore the parameter es timates were not stable. The induction of the response model was in contrast very simple and stable It also captured the overall development of somatic dependence well. Therefore, we selected this model for the simulation study. The counter response has now been considered to play the most important role in opiate tolerance (Harrison et al., 1998b; Heinzen a nd Pollack, 2004; Raith and Hochhaus, 2004). The delay in the counter response can be described by an expone ntial cumulative distribution

PAGE 117

117 function. Similar empirical approaches had b een used to characterize the time-dependent attenuation of responses (Chow et al., 1985; Hammarlund et al., 1985). However, this approach oversimplified the disassociation between the dr ug and response as a simple time-dependent function which ignored the drug eff ect on the delay. As emphasized by Gabrielsson, this kind of approaches is preferable for single dose admini stration (Gabrielsson and Weiner, 2006). To give a physiologically meaningful explanation of this delay, drug-induced deple tion of an inhibitory modulator of the counter response was integrated in the counter-response model. The modified model provided a comparable fit to the original model and expanded the model’s suitability for multiple dosing simulations. With the selected PK/PD models, the impacts of intermittent dosing strategies on physical dependence and analgesic tolerance were evaluated. Since the half-l ife of fentanyl in rats is approximate one hour, there is almost no accumulati on of the drug when the dosing interval is 6 hour or longer. However, the emergence of phys ical dependence and an algesic tolerance are significant since the drug-induced adaptive changes need much longe r time than drug itself to be eliminated. The results of our si mulation studies suggest that for the same daily dose the longer dosing interval causes more severe affective dependence. This inte resting suggestion is consistent with the common belief that intermittent rather than continuous administration causes drug craving in addicts (Post, 1980; Robinson and Becker, 1986; Stewart and Badiani, 1993; King et al., 1994; Laura E. O' Dell, 2004; Tomie et al., 2006 ). Drug craving is highly correspondent to affective dependence. Our re sults agree with Post’s observations that intermittent stimulation of a given system may produce larger output measures than continuous stimulation; and the time interval between stim ulations is only one vari able among many that is important in determining the direction and ma gnitude of adaptive res ponse following repeated

PAGE 118

118 presentations (Post, 1980). From our simulation, the longer dosing interval caused less somatic dependence. This is also in ag reement with the obse rvation that the opio id dihydroetrophine produced overt somatic signs of physical dependen ce in the rat when give n continuously rather than intermittently (Aceto et al., 2000). For the analgesic tolerance study, the longer dosing interval reduced analgesic tole rance, which is consistent w ith commonly accepted theory in clinical practice (Freye and Latasch, 2003). However, the reduc tion of analgesic tolerance in our simulation study is not significant. Considering affective dependence, somatic dependence, and analgesic tolerance together, the longer dosing in terval will enhance drug carving associated with affect affective dependence, and during each long dosing interval the patient will suffer from spontaneous withdrawal. Therefore, intermittent dosing strategies, particularly with a long dosing interval, do not appear to be suitable for chronic fentanyl treatment. In order to avoid tolerance opioid rotation and multim odal analgesia may be the direction. The comparisons among the results of aff ective dependence, somatic dependence and analgesic tolerance published in the literature again suppor t our modeling hypothesis that different mechanisms may be underlying these th ree responses. Moreover, the comparisons seem to imply that the PK propertie s of the addictive substance, especially the distribution and elimination properties play an important role in affecting the time-interval effect on physical dependence and drug tolerance. Fu ture systemic comparison studies of different classes of addictive drugs are warranted to clarify this issue. In conclusion, various structurally different PK/PD models were evaluated for affective dependence, somatic dependence and analgesic tolerance. The receptor sensitization model characterized affective dependence best; the indu ction of the response model was more suitable for somatic dependence; and the counter-respons e model with drug-indu ced depletion of an

PAGE 119

119 inhibitory modulator captured the development of analgesic tolerance well. Intermittent dosing strategies have been evaluated by simulation studie s and suggested to be not suitable for chronic fentanyl application.

PAGE 120

120 Table 5-1. Comparison of the goodness-of-fit statis tics (i.e., the data vs. prediction correlation, the data vs. prediction coefficient of determ ination, and the mode l selection criterion (MSC)) of the PD models for affective dependence Model Number of parameters Correlation Coefficient of determination MSC 2-effect-compartment 6 0.93 0.86 1.61 Indirect response (elimination of the inhibitor)8 0.94 0.87 1.58 Receptor sensitization 4 0.93 0.86 1.75 Table 5-2. Parameter estimates of the recepto r sensitization model fo r affective dependence Parameter mean CV% kin (/hr) 2.19 17.3 K 5.70 44.0 s 1.56 45.3 kDout (/hr) 0.16 37.3

PAGE 121

121 Table 5-3. Comparison of the goodness-of-fit statistics of the PD models for somatic dependence Model Number of parameters Correlation Coefficient of determination MSC 2-phase response 12 0.99 0.97 2.76 Induction of the response 3 0.98 0.95 2.78 Table 5-4. Parameter estimates of the induction of the response model for somatic dependence Parameter mean CV% kin_max (/hr) 1.50 32.7 EC50 (ng/mL) 5.68 44.2 kout (/hr) 0.021 6.0

PAGE 122

122 Table 5-5. Comparison of the goodness-of-fit statis tics of the PD models for analgesic tolerance Model Number of parameters Correlation Coefficient of determination MSC Delayed counter response 7 0.97 0.94 2.42 Counter response with depl etion of an inhibitor 8 0.97 0.94 2.40 Table 5-6. Parameter estimates of the c ounter-response model for analgesic tolerance Parameter Delayed counter response modelCounter response with depletion of an inhibitor model kin_max (/hr) 48.5 (35.0) 50.1 (37.3) kout (/hr) 0.31 (25.8) 0.31 (25.8) EC50 (ng/mL) 1.70 (56.5) 1.88 (58.0) kTin_max (/hr) 30.8 (34.4) 30.9 (34.3) kTout (/hr) 0.23 (21.3) 0.23 (21.3) TC50 (ng/mL) 1.37 (69.3) 1.29 (72.1) a (hr) 15.1 (23.2) km(/hr) 0.066 (24.2) MC50 (ng/mL) 0.10 (Fixed) Data are expressed as means (CV%).

PAGE 123

123 Hours post pump implantation 0100200300400Reward thresholds (%) 80 100 120 140 160 180 200 Fentanyl Plasma Concentration (ng/ml) 0 2 4 6 8 10 threshold (.3 mg/kg/day) threshold (.6 mg/kg/day) Cp obs. (.3 mg/kg/day) Cp obs. (.6 mg/kg/day) Cp pred. (.3 mg/kg/day) Cp pred. (.6 mg/kg/day) A. Hours post pump implantation 0100200300400Somatic withdrawal score 0 2 4 6 8 10 12 14 16 18 20 22 24 Fentanyl Plasma Concentration (ng/ml) 0 2 4 6 8 10 score (.3 mg/kg/day) score (.6 mg/kg/day) Cp obs. (.3 mg/kg/day) Cp obs. (.6 mg/kg/day) Cp pred. (.3 mg/kg/day) Cp pred. (.6 mg/kg/day) B. Figure 5-1. The complete time courses of aff ective dependence (A) and somatic dependence (B) together with fentanyl PK Drug withdrawal Drug withdrawal

PAGE 124

124 Figure 5-2. The schemes of the two-effect-c ompartment indirect-li nk PD model (A), the elimination of an inhibitor indirect-r esponse PD model (B), and the receptor sensitization PD model (C ) for affective dependence

PAGE 125

125 Figure 5-3. The schemes of the two-phase-res ponse PD model (A) a nd the induction of the response PD model (B) for somatic dependence

PAGE 126

126 Hours post pump implantation 024487296120144168192216240264288%MPE -40 -30 -20 -10 0 10 20 30 40 50 60 70 Fentanyl Plasma Concentration (ng/ml) 0 2 4 6 8 10 12 %MPE (.3 mg/kg/day) %MPE (.6 mg/kg/day) Cp obs. (.3 mg/kg/day) Cp obs. (.6 mg/kg/day) Cp pred. (.3 mg/kg/day) Cp pred. (.6 mg/kg/day) Figure 5-4. The complete time courses of antinociception together with fentanyl PK Drug withdrawal

PAGE 127

127 ERT _max 50 (/) _max 50(1)inout ta inoutdRC kkR dtECC dTC kTekTT dtTCC C CL I n d i r e c t R e s p o n s e T kTin_max kTout 75 6 12 11 10 84 2 1 93 R kin_max kout I n d i r e c t R e s p o n s e II. S(C) II. S(C) ERT _ma 50 _max 50 x 501 (1)i inout noutdMC kMkMM dtMCC dTC kTMkTT d dRC kkR dtEC T C tCC R kin_max kout II. S(C) T kTin_max kTout II. S(C) M km km I. I(C) I. I(M) Figure 5-5. The model schemes of analgesic to lerance: the delayed c ounter-response PD model (A) and a hypothesis of drug-induced depleti on of an inhibitory modulator to account for the counter-response delay (B) A. B.

PAGE 128

128 PK PD Fentanyl Cp_Obs population Abs Abs. Cpt. Plasma Central Cpt. Periph Periph. Cpt. Elim Elim. Cpt. PK_Theta Model Variables PK_Eta Multivariate Distr. PK_Eps Multivariate Distr. ka_i Expression V1_i Expression CL_i Expression Q_i Expression V2_i Expression Cp_Obs_in Expression threshold Procedure threshold_Theta Model Variables s Expression kDout Expression kin Expression K Expression threshold_Eta Multivariate Distr. PD_Eps Multivariate Distr. PD_obs_in Expression threshold_ob Response kout Expression Ka Q CL Figure 5-6. The PK/PD model scheme implemen ted in Trail Simulator for the intermittent dosing simulation studies

PAGE 129

129 Figure 5-7. The fittings to the complete time-courses of affective dependence with the twoeffect-compartment indirect-link model (A), the elimination of an inhibitor indirectresponse PD model (B), and the r eceptor sensitization model (C) A. B. C.

PAGE 130

130 Figure 5-8. The fittings to the complete time-c ourses of somatic dependence with the two-phaseresponse model (A) and the inducti on of the response model (B) A. B.

PAGE 131

131 Figure 5-9. The fittings to the complete time -courses of the antinociceptive response with the delayed counter-response model (A) and the model integrated w ith a hypothesis of depletion of an inhibitory modulator to account for the counter-response delay (B) A. B.

PAGE 132

132 QIDTime (hr) 024487296312336360384408432456480Reward threshold (%) 90 110 130 150 170 190 210 Cp (ng/ml) 0 5 10 15 20 25 30 TIDTime (hr) 024487296312336360384408432456480Reward threshold (%) 90 110 130 150 170 190 210 Cp (ng/ml) 0 5 10 15 20 25 30 R (0.3 mg/kg/day) R (0.6 mg/kg/day) C (0.3 mg/kg/day) C (0.6 mg/kg/day) QIDTime (hr) 024487296120144288312336360384408432456480Somatic withdrawal score 0 2 4 6 8 10 12 14 16 18 TIDTime (hr) 024487296120144288312336360384408432456480Somatic withdrawal score 0 2 4 6 8 10 12 14 16 18 QIDTime (hr) 024487296312336360384408432%MPE -10 0 10 20 30 40 50 60 70 80 TIDTime (hr) 024487296312336360384408432%MPE -10 0 10 20 30 40 50 60 70 80 Figure 5-10. Impacts of intermittent dosing of fentanyl on the development of affective dependence (A), somatic dependence (B) and analgesic tolerance (C) A. B. C.

PAGE 133

133 A.Fentanyl daily dose (mg/kg, sc) 0.30.6Reward threshold (%) 80 100 120 140 160 180 200 220 240 260 280 300 B.Fentanyl daily dose (mg/kg, sc) 0.30.6Somatic withdraw score 0 2 4 6 8 10 12 14 16 18 20 22 24 C.Fentanyl daily dose (mg/kg, sc) 0.30.6%MPE 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 Cont. QID TID BID QD Figure 5-11. Impacts of intermittent dosing of fentanyl on the steady state of affective dependence (A), somatic dependence (B) and analgesic tolerance (C)

PAGE 134

134 CHAPTER 6 CONCLUSIONS Fentanyl is one of the most popular opioid analge sics in modern an esthesia and pain therapy. It has been widely used to relie ve chronic pain. There is no known opioid more efficacious than fentanyl in reducing cancer pain, which makes it the first choice for breakthrough cancer pain. Fenta nyl infusion is also commonly used for the prolonged sedation and analgesia that is ofte n necessary for treating cr itically ill children. Fentanyl tolerance and physic al dependence are the most ch allenging issues in fentanyl chronic therapy. Approximately half of the can cer patients converted to transdermal fentanyl from other opioid agents required increased dos ages after initial application of the patch (Muijsers and Wagstaff, 2001). This complicates th e therapeutic applicati ons of fentanyl with respect to patient selection, product selec tion and dosing selection (FDA, 2007a; FDA, 2007b). Currently, fentanyl is classified as a Schedule II drug in the Un ited States due to its high abuse rate. Abuse of fentanyl and its analogs has been associated with a large number of drug overdose deaths, which might be contribu ted by their high potency and re spiratory depressant effects (Kronstrand et al., 1997; Kuhlman et al., 2003; Lilleng et al., 2004). PK/PD approaches have proven to be effec tive in explaining drug-response relationships, and to in dose-optimization to improve drug effici encies and reduce the risk of undesired effects (Derendorf and Meibohm, 1999; Dere ndorf et al., 2000). The aims of the present studies were to verify and characterize the corr elations among fentanyl PK, phys ical dependence and analgesic tolerance, and then to derive a clearer explanation of drug-effect relationships. This is likely to lead a more efficient therapeutic design based on the administration route and dose regimen to minimize dependence and tolerance.

PAGE 135

135 Fentanyl PK following s.c. administration in rats was linear in the testing range. A population two-compartment PK model was deve loped that could describe the observed concentration-time courses for all doses. With this quantitative information, we are now better able to predict drug concentrations ov er time for any desired dose regimen. Another aim of these studies was to investig ate the impacts of fentanyl-dose and exposure duration on the development of physical depend ence in rats. ICSS brain reward threshold measurement provided us a reliable and quantif iable biomarker of affective dependence, and scoring the somatic withdrawal syndromes from a table gave us quantitative information of somatic dependence. The results in this study indicate that fent anyl-induced physical dependence is dependent on the dose administer ed. For both affective dependence and somatic dependence, the spontaneous w ithdrawal syndrome lasted much longer than the drug after discontinuation of dosing. In addition, it was shown that fentanyl induces acute affective dependence, the development of which was highl y correlated with fentanyl concentration. However, somatic dependence gradually developed over time and the development was dissociated from drug concentration. These obs ervations suggest that different mechanisms mediate affective and somatic dependence. The results of the tail-flick tests indicate that analgesic tolerance of fentanyl during continuous drug exposure also depends on the dose and exposure-duration of fentanyl. In our studies, continuous s.c. infusion of 0.3 – 0.6 mg/ kg/day fentanyl only prod uced a short analgesic effect in rats. Antinociception of fentanyl re ached maximum 12 hours after pump implantation. Continuous exposure to fentanyl rendered rapid loss of the drug effect; and analgesia dropped to plateau 72 hours after pump implantation. Fent anyl-induced hype ralgesia was suggested after discontinuation of dosing. These observations are in agreement with prev ious observations that:

PAGE 136

136 1) continuous s.c. infusions of fentanyl rendered in fant rats tolerant to the analgesic effects of fentanyl; and 2) no analgesia was present from the fentanyl released from the pumps 72 hours after implantation of the pumps (Thornton et al., 1997; Thornton and Smith, 1998; Choe and Smith, 2000). All of these observatio ns came from the tail-flick te st, which is a measurement of the reflex or neuronal firing in th e spinal dorsal horn, and the re flex responses to high rates of noxious cutaneous heating are mainly mediated by A -fiber nociceptors. However, in the clinical situation, chronic pains are cerebr al manifestations which mainly are cortical responses mediated by C-fiber nociceptors (Yeo mans et al., 1996; Yeomans and Proudfit, 1996; Kimura and Kontani, 2008; Xiao and Bennett, 2008). A differ ential opioid modulati on of those nociceptors has also been suggested and reflex responses are considered to be insensitive to the attenuation of nociceptive sensitivity of laboratory animals by low doses of opioids (Vierck et al., 2002; King et al., 2007). Keeping that in the mind, we may need to carefully evaluate the almost complete loss of fentanyl analgesic effects obser ved in the tail-flick. Therefore, TIROS remains an important future direction to evaluate pain sensitivity in chronic treatments. PK/PD modeling links PK and PD information and enable us to pr edict the dependence and tolerance over time for any designed dose regimen. Various semi-mechanism based models have been evaluated for their suitability in characterizing affective dependence, somatic dependence, and analgesic tolerance. Among th em, the receptor sensitization model, which simply relates affective dependence proportionally to the number of hypot hetical receptors, can capture the overall time courses of affective dependence. The induction of the response model captures the overall time courses of somatic dependence and is very simple. The counterresponse model with drug-induced depletion of an inhibitory modulator capt ures the overall time courses of analgesic well. With these selected PK/PD models, the impacts of intermittent dosing

PAGE 137

137 strategies on physical dependen ce and analgesic tolerance were evaluated. The results of the simulation studies suggest that for the longer dosing interval causes more severe affective dependence but slightly less somatic dependence and analgesic tolerance. This is consistent with the common observations that intermittent rath er than continuous administration causes drug craving in addicts but less drug tolerance and somatic withdrawal syndrome and the time interval plays an important role in determining the dir ection and magnitude of adaptive response (Post, 1980; Robinson and Becker, 1986; Stewart and Badi ani, 1993; King et al., 1994; Aceto et al., 2000; Freye and Latasch, 2003; Laura E. O'Dell, 2004; Tomie et al., 2006). The comparisons among the results of affective dependence, soma tic dependence and analgesic tolerance again support our modeling hypothesis th at different mechanisms underlie these three responses. Moreover, the comparisons among the different test ed addictive substances imply that the PK properties of the addictive substa nce, especially the distribution and elimination properties play an important role in affecting the time-interval effect on physical dependence and drug tolerance. Future systemic comparison studi es of different classes of a ddictive drugs are warranted to clarify this issue. More importantly, during each of the lengthy dosing intervals the patient will suffer from spontaneous withdrawal, so that inte rmittent dosing strategies, particularly with a long dosing interval, are not recommende d for chronic fentanyl treatment. In conclusion, these studies demonstrated th e connections among fentanyl plasma levels, drug-exposure duration and the development of phys ical dependence or analgesic tolerance. The PK/PD models developed in thes e studies were able to descri be the overall time courses of affective dependence, somatic dependence and an algesic tolerance respectively in the testing situations. Comparisons between the simulation results of intermittent versus continuous dosing and the literature observations support our modeling hypothesis that different mechanisms

PAGE 138

138 underlie these three responses. We expect that the modeling exerci ses presented in these studies will profitably contribute to th e formulation of hypotheses of physical dependence and analgesic tolerance mechanisms. Specific trials will be required for this purpose, particularly those evaluating potential causal relationships betw een opioid receptor modulation and antinociceptive actions of the drug. Comparison studies of di fferent addictive drugs, both with similar and different mechanisms of action, to investigate th e impacts of their PK properties on the influence of intermittent dosing on physical dependence and analgesic tolerance may be another future direction. From a clinical perspective, these studies suggest that: 1) when fe ntanyl is used for the treatment of chronic pain, the dos e of fentanyl should be minimi zed; 2) increasing the treatment period does not further increase the negative emotional state associated with fentanyl withdrawal; and 3) continuous inst ead of intermittent dosing is preferred for long term treatment.

PAGE 139

139 APPENDIX A PHARMACOKINETIC MODELING OF SUBCUTANEOUS FENTANYL IN NONMEM $PROB Fentanyl PK S.C. study in 18 rats ;DATE = 08/23/07 PROGRAMMER=JIANG LIU ;Units: Time=min, AMT=mg/kg, Con centration=ng/mL, Volume=mL, CL=mL/min, eliminationRate=/min $DATA Rat_sc_MS.CSV IGNORE # $ABB COMRES=6 $INPUT ID AMT TIME DV $SUBROUTINE ADVAN4 TRANS4 $PK KA = POPKA*EXP(ETA(1)) V2 = POPV2*EXP(ETA(2)) Q = POPQ*EXP(ETA(3)) V3 = POPV3*EXP(ETA(4)) CL = POPCL*EXP(ETA(5)) S2 = V2/1000000 ;Sn refers to scaling fact or to respective compartment volume ;(Eg.Dose is in mg, DV (Concentration) is in ng/mL, ;V2 will be in mili-litres ie., 1/1000000 scaling factor required) $INFN IF (ICALL.EQ.3) THEN OPEN(50,FILE='cwtab1.est') WRITE(50,*) 'ETAS' DO WHILE(DATA) IF (NEWIND.LE.1) WRITE (50,*) ETA ENDDO WRITE(50,*) 'THETAS' WRITE(50,*) THETA WRITE(50,*) 'OMEGAS' WRITE(50,*) OMEGA(BLOCK) WRITE(50,*) 'SIGMAS' WRITE(50,*) SIGMA(BLOCK) ENDIF ;$THETA specifies initial estimates and bounds for structural model parameters $THETA (0,0.03) ;POPKA $THETA (500,1096) ;POPV2 $THETA (10,323) ;POPQ $THETA (1000,4215) ;POPV3 $THETA (10,89) ;POPCL (30,100) ;$OMEGA specifies initial estimates of the variance of between subject variability $OMEGA 0 FIX ;ETA_KA

PAGE 140

140 $OMEGA 0.09 ;ETA_V2 $OMEGA 0.09 ;ETA_Q $OMEGA 0.09 ;ETA_V3 $OMEGA 0.09 ;ETA_CL ;$SIGMA specifies initial estimates of the vari ance of (residual) Within subject variability $SIGMA 0.09 ;ERRSD_PRO ;ERR1-Proportional error of 30% ;$SIGMA 0.0004 ;ERRSD_ADD ;ERR2-Additive error = 50 pg/mL-Normally the LOQ of the assay method $ERROR IPRED = F IRES=DV-IPRED IF(IPRED.LE.0) THEN DEL=1 ELSE DEL=0 ENDIF IWRES=IRES/(IPRED+DEL) Y=F+F*(ERR(1));+ERR(2) ;$ERROR block specifies a model fo r (residual) Within subject variability ;Combined error model is used ie., both additive and pr oportional error model ;F = Modeled value of the dependant variable or indi vidual predicted concentrations "LAST COM(1)=G(1,1) COM(2)=G(2,1) COM(3)=G(3,1) COM(4)=G(4,1) COM(5)=G(5,1) COM(6)=HH(1,1) $ESTIMATION METH=1 INTERACTI ON MAXEVAL=9999 PRINT=10 POSTHOC MSFO=Rat_sc_8cox.msf $COV $TABLE ID TIME DV IPRED IWRES NOPRINT ONEHEADER FILE=sdtab1 $TABLE ID KA V2 Q V3 CL ETA1 ETA2 ETA3 ETA4 ETA5 NOPRINT ONEHEADER FILE=patab1 $TABLE ID COM(1)=G11 COM(2)=G21 COM(3)=G31 COM(4)=G41 COM(5)=G51 COM(6)=H11 IPRED MDV NOPRIN T ONEHEADER FILE=cwtab1.deriv

PAGE 141

141 APPENDIX B SEQUENTIAL RECEPTOR SENSITIZATION MODELING OF AFFECTIVE DEPENDENCE FOLLOWING SUBCUTANEOUS FENTANYL PUMP ADMINISTRATION IN SCIENTIST // Micromath Scientist Model File IndVars: T DepVars: Ed1,Ed2 Params: kin,K,s,kDout Infstart= T-Tlag Infstop = TTstop IND = 0 Infstart = IFGEZERO(T-Tlag,1,0) Infstop = IFGEZERO(TTstop,1,0) IND = Infstart -Infstop A1d1y'=INFR1*IND-KA*A1d1y A2d1y'=-(CL/V2)*A2d1y+ KA*A1d1y-(Q/V2)*A2d1y+(Q/V3)*A3d1y A3d1y'=(Q/V2)*A2d1y-(Q/V3)*A3d1y Cpd1y = A2d1y/V2*1000000 A1d2y'=INFR2*IND-KA*A1d2y A2d2y'=-(CL/V2)*A2d2y+ KA*A1d2y-(Q/V2)*A2d2y+(Q/V3)*A3d2y A3d2y'=(Q/V2)*A2d2y-(Q/V3)*A3d2y Cpd2y = A2d2y/V2*1000000 kout = kin/100 Rad1y'=kin*(1+s*Cpd1y) +K*Rsd1y-Cpd1y*Rad1y-kout*Rad1y Rsd1y'=-K*Rsd1y+Cpd1y*Rad1y-kDout*Rsd1y Rad2y'=kin*(1+s*Cpd2y) +K*Rsd2y-Cpd2y*Rad2y-kout*Rad2y Rsd2y'=-K*Rsd2y+Cpd2y*Rad2y-kDout*Rsd2y Ed1 = Rad1y + Rsd1y Ed2 = Rad2y + Rsd2y //independent variable initial T=0 // constant values Tstop = 336 INFR1 = 0.00795 INFR2 = 0.0159 KA = 1.44

PAGE 142

142 V2 = 995 CL = 6120 Q = 9480 V3 = 4450 Tlag = 6.9 // Parameter values kin = 2.19 K = 5.7 s = 1.56 kDout = 0.16 // Initial conditions A1d1y=0 A2d1y=0 A3d1y=0 A1d2y=0 A2d2y=0 A3d2y=0 Rad1y = 100 Rsd1y = 0 Rad2y = 100 Rsd2y = 0

PAGE 143

143 LIST OF REFERENCES Abdelhamid EE and Takemori AE (1991) Characte ristics of mu-opioid a nd delta-opioid bindingsites in striatal slices of morphine-tolerant and morphine-dependent mice. European Journal of Pharmacology 198:157-163. Abelo A, Eriksson UG, Karlsson MO, Larsson H a nd Gabrielsson J (2000) A turnover model of irreversible inhibition of gastric acid secretion by omeprazole in the dog. J Pharmacol Exp Ther 295:662-669. Aceto MD, Scates SM, Harris LS and Ji Z ( 2000) Dihydroetorphine: phy sical dependence and stereotypy after continuous infusion in the rat. European Journal of Pharmacology 387:3137. Adriaensen H, Vissers K, Noordui n H and Meert T (2003) Opioid tolerance and dependence: An inevitable consequence of chronic treatment? Acta Anaesthesiologica Belgica 54:37-47. Anand KJS and Arnold JH (1994) Opioid tolera nce and dependence in infants and children. Critical Care Medicine 22:334-342. Ariano RE, Duke PC and Sitar DS (2001) Popula tion pharmacokinetics of fentanyl in healthy volunteers. Journal of Clinical Pharmacology 41:757-763. Arnold JH, Truog RD, Orav EJ, Scavone JM and Hershenson MB (1990) Tolerance and dependence in neonates sedated with fe ntanyl during extracorporeal membraneoxygenation. Anesthesiology 73:1136-1140. Auret K and Schug SA (2005) Underutilisation of opi oids in elderly patients with chronic pain approaches to correcting the problem. Drugs and Aging 22:641-654. Bagli M, Suverkrup R, Quadflieg R, Hoflich G, Kasper S, Moller HJ, Langer M, Barlage U and Rao ML (1999) Pharmacokinetic-pharmacodynamic modeling of tolerance to the prolactinsecreting effect of chlorprothixene afte r different modes of drug administration. Journal of Pharmacology and Experimental Therapeutics 291:547. Bailey CP and Connor M (2005) Opioids: Cellul ar mechanisms of tolerance and physical dependence. Current Opinion in Pharmacology 5:60-68. Barr AM and Markou A (2005) Psychostimulant withdr awal as an inducing condition in animal models of depression. Neuroscience and Biobe havioral Reviews 29:675-706. Bauer JA, Balthasar JP and Fung H-L (1997a) Application of Pharmacodynamic Modeling for Designing Time-Variant Dosing Regimens to Overcome Nitroglycerin Tolerance in Experimental Heart Failure. Pharmaceutical Research 14:1140. Bauer JA, Balthasar JP and Fung HL (1997b) Application of pharmacodynamic modeling for designing time-variant dosing regimens to overcome nitroglycerin tolerance in experimental heart failure. Pharmaceutical Research 14:1140.

PAGE 144

144 Beal SL and Sheiner LB (1999) NONMEM User's Guide. NONMEM Project Group, University of California at San Francisc o, San Francisco, CA., in. Bhamb B, Brown D, Hariharan J, Anderson J, Balousek S and Fleming MF (2006) Survey of select practice behaviors by primary care physicia ns on the use of opioids for chronic pain. Current Medical Research and Opinion 22:1859-1865. Bhargava HN and Gulati A (1990a) Down-regula tion of brain and spinal-cord mu-opiate receptors in morphine tolerant-dependent rats. European Journal of Pharmacology 190:305-311. Bhargava HN and Gulati A (1990b) Down-regulati on of central mu-opiate and delta-opiate receptors in morphine tolerant-dependent rats. European Journal of Pharmacology 183:206-207. Bjorkman S and Redke F (2000) Cl earance of fentanyl, alfentan il, methohexitone, thiopentone and ketamine in relation to estimated hepa tic blood flow in several animal species: Application to prediction of clearance in man. Journal of Pharmacy and Pharmacology 52:1065-1074. Bjorkman S, Stanski DR, Verotta D and Harash ima H (1990) Comparativ e tissue concentration profiles of fentanyl and alfe ntanil in humans predicted from tissue/blood partition data obtained in rats. Anesthesiology 72:865. Bleeker CP, Bremer RC, Dongelmans DA, van Do ngen RTM and Crul BJP (2001) Inefficacy of high-dose transdermal fentanyl in a patient with neuropathic pa in, a case report. European Journal of Pain 5:325-329. Boas RA and Villiger JW (1985) CL INICAL ACTIONS OF FENTANYL AND BUPRENORPHINE: The Signifi cance of Receptor Binding. Br. J. Anaesth. 57:192-196. Bozarth MA and Wise RA (1984) Anatomically di stinct opiate receptor fields mediate reward and physical-dependence. Science 224:516-517. Bruijnzeel AW and Gold MS (2005) The role of corticotropin-releasing f actor-like peptides in cannabis, nicotine, and alcohol dependence. Brain Research Reviews 49:505-528. Bruijnzeel AW, Lewis B, Bajpai LK, Morey TE, Dennis DM and Gold M (2006) Severe deficit in brain reward function associated with fentanyl withdrawal in rats. Biological Psychiatry 59:477-480. Bruijnzeel AW, Marcinkiewcz C, Isaac S, Boot h MM, Dennis DM and Gold MS (2007a) The effects of buprenorphine on fent anyl withdrawal in rats. Psychopharmacology 191:931941. Bruijnzeel AW, Repetto M and Gold MS (2004) Neurobiological mechanisms in addictive and psychiatric disorders. Psychiatric Clinics of North America 27:661-+.

PAGE 145

145 Bruijnzeel AW, Zislis G, Wilson C and Go ld MS (2007b) Antagonism of CRF receptors prevents the deficit in brain reward func tion associated with precipitated nicotine withdrawal in rats. Neuropsychopharmacology 32:955-963. Burnham KP and Anderson DR (1998) Model selec tion and inference: A practical informationtheoretic approach. Model selection and inference: A practical information-theoretic approach:xx+353p. Carter BL, Tiffany ST and Conklin CA (2000) Associative and non-associative fentanyl tolerance in the rat: Evaluation of cross to lerance with mu-and ka ppa-specific opioids. Psychopharmacology 148:384-392. Castaneda-Hernandez G and Granados-Soto V ( 2000) Considerations on pharmacodynamics and pharmacokinetics: Can everythi ng be explained by the extent of drug binding to its receptor? Canadian Journal of Physiology and Pharmacology 78:199. Celerier E, Rivat C, Jun Y, Laulin JP, Larche r A, Reynier P and Simonn et G (2000) Long-lasting hyperalgesia induced by fentanyl in ra ts: Preventive effect of ketamine. Anesthesiology 92:465-472. Chakrabarti S, Yang WL, Law PY and Loh HH (1997) The mu-opioid receptor down-regulates differently from the delta-opioid receptor: Requirement of a high affinity receptor G protein complex formation. Molecular Pharmacology 52:105-113. Chia YY, Liu K, Wang JJ, Kuo MC and Ho ST ( 1999) Intraoperative high dose fentanyl induces postoperative fenta nyl tolerance. Canadian Journal of Anaesthesia 46:872-877. Cho HH, O'Connell JP, Cooney MF and Inchio sa MA (2007) Minimi zing tolerance and withdrawal to prolonged pediat ric sedation: case report and review of th e literature. J Intensive Care Med 22:173-179. Choe CH and Smith FL (2000) Se dative tolerance accompan ies tolerance to the analgesic effects of fentanyl in infant rats. Pediatric Research 47:727-735. Chow MJ, Ambre JJ, Ruo TI, Atkinson Jr AJ, Bows her DJ and Fischman MW (1985) Kinetics of cocaine distribution, eliminati on, and chronotropic effects. Clinical Pharmacology and Therapeutics 38:318. Collier HOJ (1965) A general theory of genesis of drug dependence by induction of receptors. Nature 205:181-&. Cooper DW, Lindsay SL, Ryall DM, Kokri MS, El dabe SS and Lear GA (1997) Does intrathecal fentanyl produce acute crosstolerance to i.v. morphine? Br. J. Anaesth. 78:311-313. Corbett AD, Henderson G, McKnight AT and Pate rson SJ (2006) 75 years of opioid research: the exciting but vain quest for the Holy Grail. British Journal of Pharmacology 147:S153S162.

PAGE 146

146 Cox EH, Kerbusch T, Van Der Graaf PH and Danhof M (1998) Pharmacokineticpharmacodynamic modeling of the electroencephalogr am effect of synthetic opioids in the rat: Correlation with the interact ion at the mu-opioid receptor. Journal of Pharmacology and Experimental Therapeutics 284:1095-1103. Crain SM and Shen KF (1996) Modulatory effect s of Gs-coupled excitatory opioid receptor functions on opioid analgesia, tolerance, and dependence. Neurochemical Research 21:1347-1351. Crain SM and Shen KF (1998) Modulation of opi oid analgesia, tolerance and dependence by G(s)-coupled, GM1 ganglioside-regulated opioid receptor functions. Trends in Pharmacological Sciences 19:358-365. Cryan JF, Hoyer D and Markou A (2003) Withdr awal from chronic amphetamine induces depressive-like behavioral effects in rodents. Biological Psychiatry 54:49-58. D'Amour FE and Smith DL (1941) A method for determing loss of pain sensation. J Pharmacol Exp Ther 72:74-79. Dale O, Hjortkjaer R and Kh arasch ED (2002) Nasal admini stration of opioids for pain management in adults. Acta Anaesthesiologica Scandinavica 46:759-770. Davies B and Morris T (1993) Physiological pa rameters in laboratory animals and humans. Pharm Res 10:1093-1095. Day J, Slawson M, Lugo RA and Wilkins D (2003) Analysis of fentanyl and norfentanyl in human plasma by liquid chromatography-tand em mass spectrometry using electrospray ionization. Journal of Analytical Toxicology 27:513-516. Delfs JM, Zhu Y, Druhan JP and Aston-Jones G ( 2000) Noradrenaline in the ventral forebrain is critical for opiate withdrawal-induced aversion. Nature 403:430-434. Derendorf H, Lesko LJ, Chaikin P, Colburn WA, Lee P, Miller R, Powell R, Rhodes G, Stanski D and Venitz J (2000) Pharmacokinetic/pha rmacodynamic modeling in drug research and development. J Clin Pharmacol 40:1399-1418. Derendorf H and Meibohm B ( 1999) Modeling of pharmacokinetic/pharmacodynamic (PK/PD) relationships: concepts and perspectives. Pharm Res 16:176-185. Dominguez KD, Lomako DM, Katz RW and Kelly HW (2003) Opioid withdrawal in critically ill neonates. Annals of Pharmacotherapy 37:473-477. DOST FH (1953) Der Blutspiegel. Kinetik de Konzentrat ionsablufe in der Kr eislaufflssigkeit. Leipzig: Thieme. Dumont H, Guntz E, Sosnowski M, Talla G, Roman A and Segers B (2007) Opioid-induced hyperalgesia. European Journal of Anaesthesiology 24:205-207.

PAGE 147

147 Easterling KW and Holtzman SG (19 97) Intracranial self-stimulation in rats: Sensitization to an opioid antagonist following acute or chr onic treatment with mu opioid agonists. Journal of Pharmacology and Experimental Therapeutics 281:188-199. Eilers H, Philip LA, Bickler PE, McKay WR an d Schumacher MA (2001) The reversal of fentanyl-induced tolerance by admini stration of "small-dose" ketamine. Anesthesia and Analgesia 93:213-214. Ekblad EB and Licko V (1984) Invariant relati on between total acid secr etion and secretagogue exposure: secretory dynamics in bullfrog. Am J Physiol Gastrointest Liver Physiol 246:G325-330. Epping-Jordan MP, Watkins SS, Koob GF and Ma rkou A (1998) Dramatic decreases in brain reward function during nicotine withdrawal. Nature 393:76-79. Ette EI (1998) Statistical graphics in pharmacokinetics a nd pharmacodynamics: a tutorial. Ann Pharmacother 32:818-828. Ette EI and Williams PJ (2004) Populati on pharmacokinetics II: Estimation methods. Annals of Pharmacotherapy 38:1907-1915. Ette EL, Williams PJ, Kim YH, Lane JR, Liu MJ and Capparelli EV (2003) Model appropriateness and populati on pharmacokinetic modeling. Journal of Clinical Pharmacology 43:610-623. Fattinger K, Verotta D and Benowitz NL (1997) Pharmacodynamics of acute tolerance to multiple nicotinic effects in humans. J Pharmacol Exp Ther 281:1238-1246. FDA (2001) Guidance for industry, bioanaly tical method validation, in (Food and Drug Administration FDA, Center for Drug Eval uation and Research (CDER), Center for Veterinary Medicine (CVM), USA ed). FDA (2007a) FDA issues second safety warning on fentanyl skin patc h deaths and serious injuries from improper use in. FDA (2007b) FDA warns of potential serious side effects with breakthrough cancer pain drug in. Feierman DE and Lasker JM (1996) Metabolism of fentanyl, a synthetic opioid analgesic, by human liver microsomes Role of CYP3A4. Drug Metabolism and Disposition 24:932939. Foster D, Upton R, Christrup L and Popper L (2008) Pharmacokinetics and pharmacodynamics of intranasal versus intravenous fentanyl in patients with pain after oral surgery. Ann Pharmacother 42:1380-1387. Franck L and Vilardi J (1995) Asse ssment and management of opioi d withdrawal in ill neonates. Neonatal network : NN 14:39-48.

PAGE 148

148 Franck LS, Naughton I and Winter I (2004) Opioid and benzodiazepine withdrawal symptoms in paediatric intensive care patients. Intensive and Critical Care Nursing 20:344-351. Franck LS, Vilardi J, Durand D and Powers R (1998) Opioid withdrawal in neonates after continuous infusions of morphine or fe ntanyl during extracorporeal membrane oxygenation. Am J Crit Care 7:364-369. Freye E and Latasch L (2003) Development of opioid tolerance Molecular mechanisms and clinical consequences. Anasthesiologie Intensivmedizin Notfallmedizin Schmerztherapie 38:14-+. Gabrielsson J and Weiner D (2006) Pharmacokinetic and pharmac odynamic data analysis: concepts and applications. Grdmark M, Brynne L, Hammarlund-Udenaes M and Karlsson MO (1999) Interchangeability and predictive performance of empirical tolerance models. Clinical Pharmacokinetics 36:145. Gellert VF and Holtzman SG (1978) Developmen t and maintenance of morphine-tolerance and dependence in rat by scheduled acces s to morphine drinking solutions. Journal of Pharmacology and Experimental Therapeutics 205:536-546. Gies EK, Peters DM, Gelb CR, Knag KM and Peterfreund RA (1997) Re gulation of mu opioid receptor mRNA levels by activation of protein kinase C in human SH-SY5Y neuroblastoma cells. Anesthesiology 87:1127-1138. Gilson AM, Ryan KM, Joranson DE and Dahl JL ( 2004) A reassessment of trends in the medical use and abuse of opioid analgesics and im plications for diversion control: 1997-2002. Journal of Pain and Symptom Management 28:176-188. Gourlay GK, Kowalski SR, Plummer JL, Cousin s MJ and Armstrong PJ (1988) Fentanyl bloodconcentration analgesic res ponse relationship in the treat ment of postoperative pain. Anesthesia and Analgesia 67:329-337. Gracy KN, Dankiewicz LA and Koob GF (2001) Opiate withdrawal-induced Fos immunoreactivity in the rat extended amygdala parallels the development of conditioned place aversion. Neuropsychopharmacology 24:152-160. Hammarlund MM, Odlind B and Paalzow LK (1985) Ac ute tolerance to furosemide diuresis in humans. Pharmacokinetic-pharmacodynamic modeling. J Pharmacol Exp Ther 233:447453. Harrison C, Smart D and Lambert DG ( 1998a) Stimulatory effects of opioids. British Journal of Anaesthesia 81:20-28. Harrison LM, Kastin AJ and Zadina JE (1998b) Opiate tolerance and dependence: Receptors, Gproteins, and antiopiates. Peptides 19:1603-1630.

PAGE 149

149 He L, Fong J, von Zastrow M and Whistler JL ( 2002) Regulation of opioid receptor trafficking and morphine tolerance by receptor oligornerization. Cell 108:271-282. Heinzen EL, Booth RG and Pollack GM (2005) Neuronal nitric oxide modulates morphine antinociceptive tolerance by enhancing cons titutive activity of the -opioid receptor. Biochemical Pharmacology 69:679-688. Heinzen EL and Pollack GM (2004) Pharmacodynamics of morphi ne-induced neuronal nitric oxide production and antinocicep tive tolerance development. Brain Research 1023:175184. Higgins GA and Sellers EM (1994) Antagonist-precipitated opioid w ithdrawal in rats evidence for dissociations between phys ical and motivational signs. Pharmacology Biochemistry and Behavior 48:1-8. Holford N (2008) Wings for NONMEM (WFN) http://wfn.sourceforge.net/index.html in. Huynh NH, Tyrefors N, Ekman L and Johansson M (2005) Determination of fentanyl in human plasma and fentanyl and norfentanyl in human urine using LC-MS/MS. Journal of Pharmaceutical and Biomedical Analysis 37:1095-1100. Ihmsen H, Albrecht S, Hering W, Schuttler J and Schwilden H ( 2004) Modelling acu te tolerance to the EEG effect of two benzodiazepines. British Journal of Clinical Pharmacology 57:153-161. Joranson DE, Ryan KM, Gilson AM and Dahl JL (2000) Trends in medical use and abuse of opioid analgesics. Journal of the American Medical Association 283:1710-1714. Jusko WJ, Ko HC and Ebling WF (1995) Converg ence of direct and indirect pharmacodynamic response models. J Pharmacokinet Biopharm 23:5-8; discussion 9-10. Kimura S and Kontani H (2008) Separate recordin g of a-delta and C fiber-mediated nociceptive flexor reflex responses of mouse hindlimb using electromyography and the characteristics of wind-up appearing in the responses. Journal of Pharmacological Sciences 108:172-178. King CD, Devine DP, Vierck CJ, Mauderli A an d Yezierski RP (2007) Opioid modulation of reflex versus operant responses following stress in the rat. Neuroscience 147:174-182. King CD, Devine DP, Vierck CJ, Rodgers J and Yezierski RP (2003) Di fferential effects of stress on escape and reflex responses to nociceptive thermal stimuli in the rat. Brain Research 987:214. King GR, Joyner C and Ellinwood EH, Jr. ( 1994) Continuous or intermittent cocaine administration: effects of flupent hixol treatment during withdrawal. Pharmacol Biochem Behav 49:883-889.

PAGE 150

150 Kokkinidis L and McCarter BD (1990) Postcocaine depression and sensitization of brainstimulation reward analysis of reinforcement and performance effects. Pharmacology Biochemistry and Behavior 36:463-471. Koob GF, Caine SB, Parsons L, Markou A and Weiss F (1997) Opponent process model and psychostimulant addiction. Pharmacology Biochemistry and Behavior 57:513-521. Kornetsky C and Esposito RU (1979) Euphorigenic drugs: Effects on the rewa rd pathways of the brain. Federation Proceedings 38:2473-2476. Kronstrand R, Druid H, Holmgren P and Rajs J (1997) A cluste r of fentanyl-re lated deaths among drug addicts in Sweden. Forensic Science International 88:185-195. Kuhlman JJ, McCaulley R, Valouch TJ and Be honick GS (2003) Fentanyl use, misuse, and abuse: A summary of 23 postmortem cases. Journal of Analytical Toxicology 27:499-504. Labroo RB, Paine MF, Thummel KE and Kharasch ED (1997) Fentanyl metabolism by human hepatic and intestinal cytochrome P450 3A4: im plications for interindividual variability in disposition, efficacy, and drug Interactions. Drug Metab Dispos 25:1072-1080. Lau CE and Heatherington AC (1997) Phar macokinetic-pharmacodynamic modeling of stimulatory and sedative effects of al prazolam: timing performance deficits. J Pharmacol Exp Ther 283:1119-1129. Lau CE and Sun L (2002) The pharmacokinetic dete rminants of the frequency and pattern of intravenous cocaine self-administration in rats by pharmacokinetic modeling. Drug Metab Dispos 30:254-261. Lau CE, Sun L, Wang Q, Simpao A and Falk JL (2000) Oral cocaine pharmacokinetics and pharmacodynamics in a cumulative-dose regimen: pharmacokinetic-pharmacodynamic modeling of concurrent operant and spontane ous behavior within an operant context. J Pharmacol Exp Ther 295:634-643. Lau CE, Wang Y and Ma F (1998) Pharmac okinetic-pharmacodynamic modeling of the coexistence of stimulatory and se dative components for midazolam. European Journal of Pharmacology 346:131. Laulin JP, Maurette P, Corcuff JB, Rivat C, Chauvin M and Simonnet G (2002) The role of ketamine in preventing fentanyl-induced hype ralgesia and subsequent acute morphine tolerance. Anesthesia and Analgesia 94:1263-1269. Laura E. O'Dell AJRRTSGFK (200 4) Enhanced Alcohol Self-Administration after Intermittent Versus Continuous Alcohol Vapor Exposure. Alcoholism: Clinical and Experimental Research 28:1676-1682. Lees P, Cunningham FM and Elliott J (2004) Principles of pharmacodynamics and their applications in vete rinary pharmacology. Journal of Veterinary Pharmacology and Therapeutics 27:397-414.

PAGE 151

151 Leuschen MP, Willett LD, Hoie EB, Bolam DL Bussey ME, Goodrich PD, Zach TL and Nelson RM (1993) Plasma fentanyl levels in infants undergoi ng extracorporeal membraneoxygenation. Journal of Thoracic an d Cardiovascular Surgery 105:885-891. Lilleng PK, Mehlum LI, Bachs L and Morild I (2 004) Deaths after intravenous misuse of transdermal fentanyl. Journal of Forensic Sciences 49:1364-1366. Linneman PK, Terry BE and Burd RS (2000) Th e efficacy and safety of fentanyl for the management of severe procedural pain in patients with burn injuries. Journal of Burn Care & Rehabilitation 21:519-522. Liu J, Pan H, Gold MS, Derendorf H and Bruijnz eel AW (2008) Effects of fentanyl dose and exposure duration on the affective and somatic signs of fentanyl withdrawal in rats. Neuropharmacology 55:812-818. Liu JG and Anand KJS (2001) Prot ein kinases modulate the cellular adaptations associated with opioid tolerance and dependence. Brain Research Reviews 38:1. Lotsch J (2005) Pharmacokineticpharmacodynamic modeling of opioids. Journal of Pain and Symptom Management 29:S90-S103. Lotsch J, Skarke C, Tegeder I and Geisslinger G (2002) Drug interactions with patient-controlled analgesia. Clinical Pharmacokinetics 41:31-57. Mahl A, Heining P, Ulrich P, Jakubowski J, B obadilla M, Zeller W, Bergmann R, Singer T and Meister L (2000) Comparison of clinical pat hology parameters with two different blood sampling techniques in rats: retrobulba r plexus versus sublingual vein. Laboratory Animals 34:351-361. Maldonado R and Koob GF (1993) De struction of the locus-ceruleu s decreases physical signs of opiate withdrawal. Brain Research 605:128-138. Mandema JW and Wada DR (1995) Pharmacodynami c model for acute tolerance development to the electroencephalographic effect s of alfentanil in the rat. Journal of Pharmacology and Experimental Therapeutics 275:1185. Mauderli AP, Acosta-Rua A and Vierck CJ (200 0) An operant assay of thermal pain in conscious, unrestrained rats. Journal of Neuroscience Methods 97:19. Meert TF and Vermeirsch HA (2005) A preclin ical comparison between different opioids: antinociceptive versus adverse effects. Pharmacology Biochemistry and Behavior 80:309326. Milligan K, Lanteri-Minet M, Borchert K, He lmers H, Donald R, Kress HG, Adriaensen H, Moulin D, Jarvimaki V and Haazen L (2001) Ev aluation of long-term efficacy and safety of transdermal fentanyl in the tr eatment of chronic noncancer pain. Journal of Pain 2:197204.

PAGE 152

152 Mitra S and Sinatra RS (2004) Perioperative mana gement of acute pain in the opioid-dependent patient. Anesthesiology 101:212-227. Muijsers RBR and Wagstaff AJ (2001) Transderma l fentanyl An updated review of its pharmacological properties and therapeutic e fficacy in chronic cancer pain control. Drugs 61:2289-2307. Nakamura A, Narita M, Ito H, Ozeki A, Okutsu D, Takigawa Y, Suzuki M, Kihara T, Kat A and Suzuki T (2008) Studies on drug dependence (Rep t. 429): Characterizatio n of tolerance to anti-hyperalgesic effects induced by chroni c treatment with morphine, oxycodone and fentanyl under a neuropathic pain-like state in mice. Journal of Pharmacological Sciences 106:92P-92P. NCHS (2006) National Ce nter for Health Statistics. Health United States, 2006 With Chartbook on Trends in the Health of Americans. Hyattsville, MD: 68-71., in. Nestler EJ (2004) Molecular mech anisms of drug addiction. Neuropharmacology 47:24. Nicholson B (2003) Responsible pr escribing of opioids for the management of chronic pain. Drugs 63:17-32. NIH (1998) National Institutes of Health. NIH Guid e: New Directions in Pain Research I., in. Novack GD, Bullock JL and Eisele JH (1978) Fe ntanyl: cumulative effects and development of short-term tolerance. Neuropharmacology 17:77-82. Novak S, Nemeth WC and Lawson KA (2004) Tre nds in medical use and abuse of sustainedrelease opioid analgesics: A revisit. Value in Health 7:299-300. Ohtsuka H, Fujita K and Kobayashi H (2007) Pharm acokinetics of fentanyl in male and female rats after intravenous administration. Arzneimittel-Forschung-Drug Research 57:260-263. Ouellet DMC (1995) A pharmacokinetic-pharmac odynamic model of tolerance to morphine analgesia during infusion in rats. Journal of Pharmacokinetics and Biopharmaceutics 23:531. Palkama VJ, Neuvonen PJ and Olkkola KT (1998) The CYP 3A4 inhibitor itraconazole has no effect on the pharmacokinetics of i.v. fentanyl. British Journal of Anaesthesia 81:598-600. Paronis CA and Holtzman SG (1992) Development of tolerance to the anal gesic activity of mu agonists after continuous infu sion of morphine, meperidine or fentanyl in rats. Journal of Pharmacology and Experimental Therapeutics 262:1-9. Persky AM, Berry NS, Pollack GM and Brouw er KLR (2004) Modelli ng the cardiovascular effects of ephedrine. British Journal of Clinical Pharmacology 57:552-562. Porchet HC, Benowitz NL and Sheiner LB ( 1988) Pharmacodynamic model of tolerance: application to nicotine. J Pharmacol Exp Ther 244:231-236.

PAGE 153

153 Post RM (1980) Intermittent versus continuous stimulation: effect of time interval on the development of sensitization or tolerance. Life Sci 26:1275-1282. Raith K and Hochhaus G (2004) Drugs used in th e treatment of opioid tolerance and physical dependence: A review. International Journal of Clinical Pharmacology and Therapeutics 42:191-203. Ramakrishnan R, DuBois DC, Almon RR, Pyszczynski NA and Jusko WJ (2002) Fifthgeneration model for corticosteroid pharmacodyna mics: application to steady-state receptor down-regulation and enzyme i nduction patterns during seven-da y continuous infusion of methylprednisolone in rats. Journal of Pharmacokine tics and Pharmacodynamics 29:1. Robinson TE and Becker JB (1986) Enduring changes in brain a nd behavior produced by chronic amphetamine administration: a review and ev aluation of animal models of amphetamine psychosis. Brain Res 396:157-198. Rodrigues AC, Verne GN, Schmidt S and Mauderl i AP (2005) Hypersensitivity to cutaneous thermal nociceptive stimuli in irritable bowel syndrome. Pain 115:5-11. Rothman RB, Long JB, Bykov V, Xu H, Jacobs on AE, Rice KC and Holaday JW (1991) Upregulation of the opioid receptor complex by th e chronic administration of morphine a biochemical marker related to the deve lopment of tolerance and dependence. Peptides 12:151-160. Sallstrom B, Visser SAG, Forsberg T, Peleti er LA, Ericson AC and Gabrielsson J (2005) A pharmacodynamic turnover model capturing as ymmetric circadian baselines of body temperature, heart rate and blood pressure in rats: Challenges in terms of tolerance and animal-handling effects. Journal of Pharmacokinetics and Pharmacodynamics 32:835. Scholz J, Steinfath M and Schulz M (1996) Clinical pharmacokinetics of alfentanil, fentanyl and sufentanil An update. Clinical Pharmacokinetics 31:275-292. Schulteis G, Markou A, Cole M and Koob GF (1995) Decreased brai n reward produced by ethanol withdrawal. Proceedings of the National Academy of Sciences of the United States of America 92:5880-5884. Schulteis G, Markou A, Gold LH, Stinus L and Koob GF (1994) Relative sensitivity to naloxone of multiple indexes of opiate withdrawal a quantitative dose-response analysis. Journal of Pharmacology and Experimental Therapeutics 271:1391-1398. Sebel PS, Barrett CW, Kirk CJC and Heykants J ( 1987) Transdermal absorption of fentanyl and sufentanil in man. European Journal of Clinical Pharmacology 32:529-531. Shi J, Benowitz NL, Denaro CP and Sheine r LB (1993) Pharmacokinetic-pharmacodynamic modeling of caffeine: Tolerance to pressor effects. Clinical Pharmacology and Therapeutics 53:6.

PAGE 154

154 Simons SHP and Anand KJS (2006) Pain control: Opioid dosing, population kinetics and sideeffects. Seminars in Fetal & Neonatal Medicine 11:260-267. Sirohi S, Dighe SV, Walker EA and Yoburn BC (2008) The analgesic efficacy of fentanyl: Relationship to tolerance and mu -opioid receptor regulation. Pharmacology Biochemistry and Behavior 91:115-120. Skaer TL (2006) Transdermal opioids for cancer pain. Health and Quality of Life Outcomes 4:24. Stanley TH (2005) Fentanyl. Journal of Pain and Symptom Management 29:S67-S71. Stewart J and Badiani A (1993) To lerance and sensitization to the behavioral effects of drugs. Behav Pharmacol 4:289-312. Taylor JR, Elsworth JD, Garcia EJ, Grant SJ, Roth RH and Redmond DE (1988) Clonidine infusions into the locus coeruleus attenua te behavioral and ne urochemical changes associated with naloxone-p recipitated withdrawal. Psychopharmacology 96:121-134. Thornton SR and Smith FL (1998) Long-term altera tions in opiate antinoci ception resulting from infant fentanyl tolerance and dependence. European Journal of Pharmacology 363:113119. Thornton SR, Wang AF and Smith FL (1997) Ch aracterization of neonatal rat morphine tolerance and dependence. European Journal of Pharmacology 340:161-167. Tomie A, Miller WC, Dranoff E and Pohorecky LA (2006) Intermitte nt presentations of ethanol sipper tube induce ethanol drinking in rats. Alcohol Alcohol. 41:225-230. Twite MD, Rashid A, Zuk J and Friesen RH ( 2004) Sedation, analgesi a, and neuromuscular blockade in the pediatric intensive care uni t: survey of fellowshi p training programs. Pediatr Crit Care Med 5:521-532. Usune S, Katsuragi T and Furukawa T (1988) Interaction between Bay-K-8644 and various relaxant agents on K+-depolarized c ontracture of guinea pig taenia coli. Gen Pharmacol 19:113-116. Varvel JR, Shafer SL, Hwang SS, Coen PA and Stanski DR (1989) Absorption characteristics of transdermally administered fentanyl. Anesthesiology 70:928-934. Vekovischeva OY, Zamanillo D, Echenko O, Seppala T, Uusi-Ouka ri M, Honkanen A, Seeburg PH, Sprengel R and Korpi ER (2001) Morphine -induced dependence and sensitization are altered in mice deficient in AMPA-type glutamate receptor-A subunits. Journal of Neuroscience 21:4451. Verotta D and Sheiner LB (1995) A genera l conceptual model for non-steady state pharmacokinetic/pharmacodynamic data. J Pharmacokinet Biopharm 23:1-4.

PAGE 155

155 Vierck CJ, Acosta-Rua A, Nelligan R, Test er N and Mauderli A (2002) Low dose systemic morphine attenuates operant escape but facil itates innate reflex responses to thermal stimulation. Journal of Pain 3:309. Vierck CJ, Mauderli AP and Wiley RG (2005) A ssessment of pain in laboratory animals: a comment on Mogil and Crager (2004). Pain 114:520-523. Visser SAG, Sallstrom B, Forsberg T, Peletier LA and Gabrielsson J (2 006) Modeling drugand system-related changes in body temperature: Application to cl omethiazole-induced hypothermia, long-lasting tolerance developm ent, and circadian rhythm in rats. Journal of Pharmacology and Experimental Therapeutics 317:209. Wagenmakers EJ (2003) Model selection and multimodel inference: A practical informationtheoretic approach. Journal of Mathematical Psychology 47:580-586. Wakelkamp M, Alvan G, Gabrielsson J and Pa intaud G (1996) Pharmacodynamic modeling of furosemide tolerance after multiple intravenous administration. Clinical Pharmacology & Therapeutics 60:75-88. Wang E, Balthasar J and Fung H-L (2004) Pharmac odynamics of in vivo nitroglycerin tolerance in normal conscious rats: effect s of dose and dosing protocol. Pharmaceutical Research 21:114. Willens JS and Myslinski NR (1993) Pharmacodynamics, pharmacokinetics, and clinical uses of fentanyl, sufentanil and alfentanil. Heart & Lung 22:239-251. Xiao WH and Bennett GJ (2008) C-fiber spontaneous discharge evoked by chronic inflammation is suppressed by a long-term infusion of lidocaine yielding nanogram per milliliter plasma levels. Pain 137:218-228. Yassen A, Kan J, Olofsen E, Suidgeest E, Dahan A and Danhof M (2006) Mechanism-based pharmacokinetic-pharmacodynamic modeling of th e respiratory-depressant effect of buprenorphine and fentanyl in rats. J Pharmacol Exp Ther 319:682-692. Yassen A, Olofsen E, Dahan A and Danhof M (2005) Pharmacokinetic-pharmacodynamic modeling of the antinociceptive effect of buprenorphine and fe ntanyl in rats: Role of receptor equilibration kinetics. Journal of Pharmacology and Experimental Therapeutics 313:1136-1149. Yassen A, Olofsen E, Kan J, Dahan A and Danhof M (2008) Pharmacokinetic-pharmacodynamic modeling of the effectiveness and safety of buprenorphine and fentanyl in rats. Pharmaceutical Research 25:183-193. Yeomans DC, Pirec V and Proudfit HK (1996) Nocicep tive responses to hi gh and low rates of noxious cutaneous heating are mediated by diffe rent nociceptors in the rat: behavioral evidence. Pain 68:133-140.

PAGE 156

156 Yeomans DC and Proudfit HK (1996) Nociceptive responses to high and low rates of noxious cutaneous heating are mediated by different nociceptors in the ra t: electrophysiological evidence. Pain 68:141-150. Zannikos PN, Rohatagi S and Jensen BK ( 2001) Pharmacokinetic-pharmacodynamic modeling of the antilipolytic effects of an adenos ine receptor agonist in healthy volunteers. J Clin Pharmacol 41:61-69. Zeller W, Weber H, Panoussis B, Burge T a nd Bergmann R (1998) Refinement of blood sampling from the sublingual vein of rats. Laboratory Animals 32:369-376.

PAGE 157

157 BIOGRAPHICAL SKETCH Jiang Liu was born in Guiyang, P.R. China. In fluenced by his parents, he developed strong interests in both health science and mathema tics. He received his bachelor’s degree in microbiology and immunology from Wuhan Universi ty and then his master’s degree from Peking Union Medical College in 1995. To combin e his mathematical interests with his health science career, he first achieved a Master of Bioinformatics from University of Waterloo, Canada, in 2004. In 2005, Jiang entered the Ph.D. program in the Department of Pharmaceutics, College of Pharmacy, University of Florida, under the supervision of Dr. Hartmut Derendorf. The door of Pharmacometrics started to open to hi m. Jiang enjoyed his study and research very much in Dr. Derendorf’s group and received his Doctor of Philosophy degree in pharmaceutics in May 2009.