Control of analgesic and anti-inflammatory pathways by fatty acid amide hydrolase Long, James Harry

The copyright of this thesis rests with the author and no quotation from it or information derived from it may be published without the prior written consent of the author

For additional information about this publication click this link. http://qmro.qmul.ac.uk/jspui/handle/123456789/3124

Information about this research object was correct at the time of download; we occasionally make corrections to records, please therefore check the published record when citing. For more information contact [email protected]

Control of analgesic and anti-inflammatory

pathways by fatty acid amide hydrolase

James Harry Long

Thesis submitted for the degree of Doctor of Philosophy to the University of London

Translational Medicine and Therapeutics

William Harvey Research Institute

Charterhouse Square, London, EC1M 6BQ

Table of contents

Table of Contents

Declaration VIII

Acknowledgements IX

Abstract X

Abbreviations XI

Chapter 1 – Introduction 1

1.1. Pain and analgesia 2

1.1.1. Nociception 2

1.1.2. Inflammatory pain 5

1.1.3. Neuropathic pain 10

1.1.4. Analgesia 10

1.1.5. COX inhibitors 11

1.1.6. Opioid receptor agonists 12

1.1.7. Glucocorticoids 13

1.1.8. Anaesthetics 13

1.1.9. Antidepressants 14

1.1.10. Anticonvulsants 14

1.1.11. Muscle relaxants 15

1.1.12. An alternative analgesic pathway 15

1.2. Endocannabinoid system 16

1.2.1. Cannabinoid receptors 16

1.2.2. Endocannabinoids 18

1.2.3. Endocannabinoid biosynthesis 21

1.2.4. Endocannabinoid metabolism 21

1.2.5. Endocannabinoid signalling 26

1.3. Cannabinoid receptor-mediated analgesia 28

I Table of contents

1.3.1. Cannabinoid receptor knock-out and activation 28

1.3.2. Endogenous cannabinoid receptor ligands 29

1.3.3. Other cannabinoid receptor ligands 29

1.4. FAAH-mediated analgesia 30

1.4.1. The preference of FAAH over MAGL 30

1.4.2. FAAH knock-out and inhibition 31

1.4.3. The multiple mechanisms of FAAH-mediated analgesia 32

1.4.4. The action of OEA via the PPAR-α and TRPV1 receptors 32

1.4.5. The action of PEA via the PPAR-α receptor and other pathways 33

1.4.6. The action of the FAAH substrates via GPR55 33

1.4.7. FAAH as a potential source of arachidonic acid for prostaglandin 34

synthesis

1.4.8. Further analgesic effects of FAAH inhibition 35

1.5. Development of FAAH inhibitors 38

1.5.1. FAAH inhibitors 38

1.5.2. The FAAH substrates as a pharmacodynamic parameter 40

1.5.3. FAAH substrate assays 41

1.6. Paracetamol 42

1.6.1. Paracetamol inhibits COX activity 43

1.6.2. Paracetamol interacts with the endocannabinoid system 44

1.6.3. Paracetamol interacts with serotonin receptors 46

1.6.4. AM404 interacts with COX, iNOS, TRPV1 and sodium channels 47

1.6.5. Pharmacokinetics of intravenous paracetamol in the cerebrospinal 47

fluid

1.6.6. Paracetamol toxicity 49

1.7. Aims 50

II Table of contents

Chapter 2 – Inhibition of FAAH and MAGL suppresses prostaglandin E2 51 synthesis by LPS-stimulated macrophages

2.1. Introduction 52

2.2. Materials and methods 53

2.2.1. Materials 53

2.2.2. Macrophages 53

2.2.3. Prostaglandin E2 assay 54

2.2.4. Nitric oxide assay 54

2.2.5. Cell viability assay 55

2.2.6. Western blotting COX-2 55

2.2.7. Statistical analysis 55

2.3. Results 56

2.3.1. Effect of FAAH and MAGL inhibition on the cell viability of LPS- 56

stimulated RAW264.7 macrophages

2.3.2. Effect of FAAH and MAGL inhibition on LPS-induced PGE2 synthesis in 58

RAW264.7 macrophages

2.3.3. Effect of FAAH and MAGL inhibition on LPS-induced nitric oxide 59

synthesis in RAW264.7 macrophages

2.3.4. Effect of MAGL inhibition on LPS-induced COX-2 expression in 60

RAW264.7 macrophages

2.4. Discussion 61

2.4.1. Limitations of the present study 65

2.4.2. Summary 66

Chapter 3 – FAAH substrates as a pharmacodynamic response in an in 67 vivo model of inflammatory pain

3.1. Introduction 68

III Table of contents

3.2. Materials and methods 70

3.2.1. Materials 70

3.2.2. FCA-induced hypersensitivity 70

3.2.3. Drug administration 71

3.2.4. Sample preparation 71

3.2.5. FAAH substrate assay by LC-MS/MS 72

3.2.6. FAAH activity assay 73

3.2.7. Intrinsic clearance assay 73

3.2.8. Protein binding assay 74

3.2.9. Statistical analysis 76

3.3. Results 77

3.3.1. FAAH substrate assay – optimisation 77

3.3.2. FAAH substrate assay – validation 81

3.3.3. Effect of URB597 on brain and blood FAAH substrate levels in an in 85

vivo inflammatory pain model

3.3.4. Effect of the PF-04457845 and two non-covalent inhibitors, NC1 and 90

NC2, on brain and blood FAAH substrate levels in an in vivo

inflammatory pain model

3.3.5. Dose response of the non-covalent FAAH inhibitor, NC3, on brain and 92

blood FAAH substrate levels in an in vivo inflammatory pain model

3.3.6. Effect of the COX-2 inhibitor, celecoxib, on brain and blood FAAH 93

substrate levels in an in vivo inflammatory pain model

3.3.7. Oxidative metabolism of the FAAH substrates in liver microsomes 94

3.3.8. Protein binding of the FAAH substrates in brain and liver microsomes 97

3.4. Discussion 98

3.4.1. FAAH substrate assay 98

3.4.2. Effect of FAAH inhibitors on brain and blood FAAH substrate levels in 99

an in vivo inflammatory pain model

3.4.3. Oxidative metabolism and protein binding of AEA, OEA and PEA 102

3.4.4. Summary 104

IV Table of contents

Chapter 4 – Pharmacokinetics and metabolism of intravenous 105 paracetamol in the human cerebrospinal fluid and its conjugation to fatty acids via FAAH

4.1. Introduction 106

4.2. Materials and methods 108

4.2.1. Materials 108

4.2.2. Formation and detection of FAAH formed paracetamol metabolites in 108

rat brain homogenate

4.2.3. Patients 109

4.2.4. Sample collection 110

4.2.5. Protein assay 110

4.2.6. Paracetamol assay 110

4.2.7. AM404 and FAAH substrate assay 111

4.2.8. Prostaglandin E2 assay 112

4.2.9. Leukotriene B4 assay 113

4.2.10. 5HT assay 113

4.2.11. Pharmacokinetic calculations 113

4.3. Results 116

4.3.1. Detection of fatty acid paracetamol conjugates in brain homogenate 116

4.3.2. Formation of fatty acid paracetamol conjugates by FAAH in rat brain 119

homogenate

4.3.3. Collection of human CSF and plasma samples 123

4.3.4. Measurement of total protein in CSF 123

4.3.5. Preliminary investigation of a single method for quantitation of 123

paracetamol and its major and minor metabolites

4.3.6. Paracetamol assay by LC-UV – optimisation 124

4.3.7. Paracetamol assay – validation 125

4.3.8. AM404 and FAAH substrate assay – validation 127

4.3.9. 5HT assay – validation 130

4.3.10. Pharmacokinetics of paracetamol in the cerebrospinal fluid and plasma 132

V Table of contents

following intravenous administration

4.3.11. Metabolism of paracetamol to paracetamol glucuronide, paracetamol 134

sulphate and AM404 in the cerebrospinal fluid and plasma following

intravenous administration

4.3.12. Formation of oleic and palmitic acid derived metabolites of 140

paracetamol in CSF following intravenous administration

4.3.13. FAAH substrate levels in the CSF and plasma following intravenous 142

administration

4.3.14. Prostaglandin E2 levels in the CSF and plasma following intravenous 143

administration

4.3.15. Leukotriene B4 levels in the CSF and plasma following intravenous 144

administration

4.3.16. 5HT levels in the CSF and plasma following intravenous administration 145

4.4. Discussion 146

4.4.1. Conjugation of fatty acids to paracetamol via FAAH in brain 146

homogenates

4.4.2. Pharmacokinetics of paracetamol in human cerebrospinal fluid and 148

plasma

4.4.3. Metabolism of paracetamol in human cerebrospinal fluid and plasma 150

4.4.4. Metabolism of paracetamol to AM404 and other fatty acid conjugates 152

4.4.5. The transfer of paracetamol and its metabolites into the CSF 154

4.4.6. Paracetamol’s effect on endogenous pathways 155

4.4.7. Summary 158

Chapter 5 – General discussion 160

5.1. Overview 161

5.2. The expanding endocannabinoid system 162

5.2.1. New receptors in the endocannabinoid system 163

5.2.2. New ligands in the endocannabinoid system 163

5.2.3. New proteins in the endocannabinoid system 165

VI Table of contents

5.2.4. New actions of the endocannabinoid system 166

5.2.5. New interactions of the endocannabinoid system 167

5.2.6. Summary 168

5.3. The role of the endocannabinoid system in paracetamol’s 168

action

5.4. The therapeutic potential of a FAAH inhibitor 169

5.4.1. FAAH inhibitors as analgesics 170

5.4.2. Alternative uses of FAAH inhibitors 171

5.4.3. Potential adverse effects of FAAH inhibitors 172

5.4.4. Conclusion 172

References 173

Publications 195

Appendix 196

VII Declaration

Declaration

I declare that the work presented in this thesis, unless otherwise stated, is my own.

James Long

VIII Acknowledgements

Acknowledgements

Thank you to Professor David Perrett for giving me the opportunity to do a PhD. I am grateful for your supervision during my time with you. Especially I thank you for teaching me that most things do not work first time and that what you want to happen is not always what will happen.

Thank you to Dr. Nanda Nayuni for your supervision, friendship, company in the lab and support during my PhD.

Thank you to Dr. Chhaya Sharma for your help in the clinical trial and for being so efficient at providing me with the cerebrospinal fluid and blood samples. Thank you to

Dr. Essam Ghazaly for the use of your LC-MS/MS system. Thank you to Professor

Atholl Johnston for your help with the pharmacokinetic calculations and advice in interpreting the results. Thank you to Dr. Samir Ayoub for your help with the prostaglandin extraction.

Thank you to Dr. Izza Nordin for providing me with the macrophages that I used in my experiments. Thank you to Dr. Ananda Chapagain for your assistance in the

Western blot analysis.

Thank you to Dr. Andy Harris, Dr. Kim Matthews, Parmjit Sadra and Clare Bartle for your help during my placement at GSK. Also I would like to thank the various in vivo technicians at GSK that provided me with the tissue samples analysed for the various studies.

Thank you to Dr. Ajit Shah for helping in the setting up of my PhD.

I am thankful to all that I worked with at the William Harvey Research Institute .

I am grateful to the BBSRC and GSK for funding this PhD.

IX Abstract

Abstract

The endogenous cannabinoids (endocannabinoids) arachidonoyl ethanolamide (AEA) and 2-arachidonoyl glycerol (2-AG) regulate neurotransmission and inflammation by activating the CB1 cannabinoid receptors in the central nervous system and the CB2 cannabinoid receptors on immune cells. Endocannabinoids are inactivated via a 2-step process: firstly undergoing reuptake into the cell and then hydrolysis by fatty acid amide hydrolase (FAAH) for AEA or by monoacylglycerol lipase (MAGL) for 2-AG.

Inhibition of FAAH is an effective treatment for inflammatory pain and is a target for the development of novel analgesics.

FAAH and MAGL produce arachidonic acid by hydrolysis of endocannabinoids. In inflammation arachidonic acid is oxidised by cyclooxygenase 2 (COX-2) to synthesise the prostaglandins. Both FAAH and MAGL inhibition were showed to indirectly suppress prostaglandin E2 synthesis in activated macrophages. This effect was found to be dependent on the amount of lipopolysaccharide used to activate the macrophages.

The FAAH substrates, AEA, oleoyl ethanolamide (OEA) and palmitoyl ethanolamide

(PEA), were measured in samples from an in vivo model of inflammatory pain to screen various FAAH inhibitors to prove their mechanism of action. Liquid chromatography-tandem mass spectrometry methods were developed and validated for the measurement of these three substrates in brain and whole blood. Blood OEA and PEA concentrations were found to be indicative of brain FAAH activity and brain

AEA concentrations.

FAAH is capable of conjugating arachidonic acid to paracetamol to form the active metabolite AM404, an inhibitor of AEA reuptake. It was shown that FAAH can also conjugate oleic acid and palmitic acid to paracetamol, producing novel metabolites. A clinical trial was conducted to comprehensively investigate the pharmacokinetics, metabolism and mechanism of action of intravenous paracetamol in cerebrospinal fluid and plasma. Paracetamol, its major metabolites, AM404, prostaglandin E2, leukotriene

B4, serotonin and the FAAH substrates were measured. For the first time AM404 was demonstrated to be formed by humans and be present in the cerebrospinal fluid.

X Abbreviations

Abbreviations

2-AG 2-arachidonyl glycerol 2-AGE 2-archidonylglyceryl ether 2-OG 2-oleoyl glycerol 2-PG 2-palmitoyl glycerol 5HIAA 5-hydroxyindoleacetic acid 5HT 5-hydroxytryptamine AEA Arachidonoyl ethanolamide Abh4 α/β-hydrolase 4 AUC Area under the curve AM404 Arachidonoyl phenolamine BCA Bicinchoninic acid assay C Covalent FAAH inhibitor CB Cannabinoid receptor CL Clearance CLi Intrinsic clearance Cmax Maximum concentration CNS Central nervous system C0 Concentration at time zero COX Cyclooxygenase cPLA2 Cytosolic phospholipase A2 CSF Cerebrospinal fluid CV Coefficient of variation DAG Diacylglycerol DAGL Diacylglycerol lipase DMA Dimethylacetamide DMEM Dulbecco's Modified Eagle's Medium DMSO Dimethyl sulfoxide DRG Dorsal root ganglion EC Enzyme Commission number EC50 Effective concentration (half-maximal) ECL Enhanced chemiluminescence EDTA Ethylenediaminetetraacetic acid EET-EA Epoxyeicosatrienoic acid ethanolamide EP Prostanoid FAAH Fatty acid amide hydrolase FABP Fatty acid binding protein FBS Fetal bovine serum FCA Freund’s complete adjuvant FLAT FAAH-like anandamide transporter FRET Forster’s resonance energy transfer GABA γ-aminobutyric acid GSH Glutathione HETE-EA Hydroxyeicosatetraenoic acid ethanolamide HETE-G Hydroxyeicosatetraenoic acid glycerol HPETE-EA Hydroperoxyeicosatetraenoic acid ethanolamide HPETE-G Hydroperoxyeicosatetraenoic acid glycerol IC50 Inhibitory concentration (half-maximal) IL Interleukin iPLA2 Inducible phospholipase A2 i.p. Intraperitoneal in vitro Latin: Within glass in vivo Latin: Within the living iNOS Inducible nitric oxide synthase i.v. Intravenous k or λ Rate constant LC-MS/MS Liquid chromatography coupled with tandem mass spectrometry LC-UV Liquid chromatography coupled with ultra violet detection LDS Lithium dodecyl sulfate LEA Linoleoyl ethanolamide LHP Lipid hydroperoxide LOD Limit of detection

XI Abbreviations

LOQ Limit of quantification LOX Lipoxygenase LPS Lipopolysaccharide LTB4 Leukotriene B4 m/z Mass to charge ratio MAGL Monoacylglycerol lipase MAPK Mitogen activated protein kinase MC Methyl cellulose MRM Multiple reaction monitoring MD-2 Myeloid differentiation-2 mPGES Microsomal prostaglandin E synthase MTT Dimethyl thiazolyl diphenyl tetrazolium NAAA N-acylethanolamine acid amidase NADA N-arachidonyl dopamine NAPE N-arachidonoyl-phosphatidyl-ethanolamine NAPE-PLD N-arachidonoyl-phosphatidyl-ethanolamine phospholipase D NAPQI N-acetyl-p-benzoquinone imine Na v Voltage-gated sodium channel NC Non-covalent FAAH inhibitor NF-κB Nuclear factor kappa B NMDA N-methyl-D-aspartic acid NO Nitric oxide NSAIDs Non-steroidal anti-inflammatory drugs O-AEA O-arachidonoyl ethanolamide OEA Oleoyl ethanolamide PBS Phosphate buffered saline PEA Palmitoyl ethanolamide PEG 400 Polyethylene glycol 400 PGD2 Prostaglandin D2 PGE2 Prostaglandin E2 PGG2 Prostaglandin G2 PGH2 Prostaglandin H2 PG-EA Prostaglandin ethanolamide PG-G Prostaglandin glycerol PLA2 Phospholipase A2 PLD Phospholipase D PMSF Phenylmethylsulfonyl fluoride p.o. Per os Latin: By mouth PPAR-α Peroxisome proliferator-activated receptor alpha R2 R-squared s Second SD Standard deviation SEM Standard error of the mean SNRI Serotonin-norepinephrine reuptake inhibitors sPLA2 Secretory phospholipase A2 SSRI Selective serotonin reuptake inhibitors t0 Time zero t1/2 Half-life tmax Time at which concentration is half C max TBS TRIS-buffered saline TNF-α Tumour necrosis factor - alpha TLR4 Toll-like receptor 4 TRIS Tris(hydroxylmethyl)aminomethane TRP Tryptophan or Transient receptor potential TRPV1 Transient receptor potential vanilloid 1 TXA2 Thromboxane A2 ∆9-THC ∆9-tetrahydrocannabinol V Voltage VD Volume of distribution

XII Chapter 1

CHAPTER 1

Introduction

1 Chapter 1

1.1. Pain and analgesia

For the purpose of survival, all life is capable of sensing and responding to the hostilities of its environment.1 Pain, the unpleasant sensory and emotional experience caused by the presence of noxious stimuli, is one aspect of this response mechanism.2

The classification of pain is based on its cause, type, region and duration. The most common pain-causing conditions in descending order are arthritis (rheumatoid and osteoarthritis), spinal disc herniation, injury/trauma, headache/migraine, nerve injury and surgery. 3 The most common locations of pain in descending order are back, knee, head, leg, joints, shoulder, neck, hip and hands. 3

While many conditions can produce pain, its cause can be simplified as being due to noxious stimuli, inflammation or nerve injury.4 Each of these causes will produce their own types of pain, either nociceptive, inflammatory or neuropathic respectively.5

Nociceptive pain is the stimulation of peripheral nerve fibres by noxious stimuli, such as mechanical pain from pressure or movement, thermal pain from heat or cold, or chemical pain from acid. If noxious stimuli were to subsequently cause tissue injury, or if infection was to occur, then this would result in a release of inflammatory mediators that leads to inflammatory pain. Direct damage to the nervous system caused by injury to the nerves or spinal cord will result in neuropathic pain. 5

The duration of pain is classified as either acute or chronic, with acute pain being of a rapid on-set and typically of a short duration, while chronic pain potentially lasts many months and years and hence beyond the expected time of healing. Nociceptive pain is often short lasting and ends once the noxious stimulus has been removed.

Inflammatory pain lasts as long as it takes for the inflammation to resolve. The failure of inflammatory resolution can lead to chronic inflammation, and consequently chronic pain. Neuropathic pain is a longer lasting pain and may never completely cease. 6

1.1.1. Nociception

Pain begins with the detection of a stimulus by nociceptors (Figure 1.1.). These are pain specific receptors expressed on sensory neurons of the peripheral nervous system

2 Chapter 1 that express various ion channels and receptors, such as the transient receptor potential (TRP) channels and the voltage-gated sodium channels. Nociceptors are present in the skin, muscles and joints with various levels of sensitivity. 7

Figure 1.1. – Pain signalling pathways. Nociceptors detect and relay noxious stimuli via the dorsal root ganglion and spinal cord to the brain. Inflammatory mediators released by inflammatory leukocytes amplify this response. Image redrawn from Scholz et al. 5

Upon sensation of pain, a signal is sent from the periphery to the central nervous system (CNS). Connections between neurons are known synapses. It is at the synapse that signals are passed between neurons. While a signal is not being sent a resting cell membrane potential is maintained, which is characterised by high intracellular concentrations of K + (139 mM) and high extracellular concentrations of Na + (145 mM).8 During the relay of a signal an action potential, travelling along the neuron, will reaches the synapse. This causes an influx of cations, such as Na + and Ca 2+ , thereby depolarising the neuron. At resting potential the voltage-gated calcium channels are closed. 8 Depolarisation increases the Ca 2+ permeability of the presynapse, triggering vesicles filled with neurotransmitters to fuse with the membrane and release their contents into the synapse. 8 Neurotransmitters, then bind to their receptors on the

3 Chapter 1 opposing postsynapse, in a process known as neurotransmission. On completion of their action these neurotransmitters undergo reuptake back into the presynapse.

Neurotransmission results in either the continuation of the action potential along adjacent neuron, or in a longer term change such as the sensitisation of the neuron. 8

Eventually the signal is relayed to the dorsal horn, up the spinal cord and brain stem into the brain. The consequent processing of pain is complex, involving the signal transduction of the stimuli and additional cognitive and emotional processing. 9

The sensory fibres responsible for relaying pain signals from the periphery extend from the cell bodies of trigeminal ganglia and dorsal root ganglia (DRG). 6 There are a number of types of sensory fibres located in the periphery, differentiated by their size and degree of myelination. The larger Aα and Aβ-fibres do not convey pain signals, this is the role of the smaller Aδ and the smaller still C-fibres. 8 Myelin, composed of predominately lipids, water and protein, serves to insulate the nerves to prevent ion leakages and weak signalling. The conduction velocity of an unmyelinated axon is between 0.5 to 10 m/s, while myelinated axons are faster, conducting at velocities up to 150 m/s. 10 Therefore the thinly myelinated Aδ fibres relay pain the fastest, whilst the unmyelinated C-fibres are slower. 8 While some nociceptors are thinly myelinated, the majority are unmyelinated.11

Sensitivity of these fibres is not fixed and can be increased (hypersensitivity) by inflammatory mediators and decreased (diminished sensitivity), both affecting the perception of pain. Normally this pain threshold is high enough to allow for normal activity, but low enough so that noxious stimuli can be detected.5 Diminished sensitivity and hypersensitivity mean that noxious stimuli will not always cause pain and that pain will not always have been caused by noxious stimuli. Such sensitisation of neurons is due to a repeated stimulus that will cause the signal that it sends to be progressively amplified. Sensitisation is further differentiated as central and peripheral. 6 Central sensitisation is the heightened response of sensory neurons to pain. Peripheral sensitisation is a reduced threshold for the activation of sensory neurons. The sensitisation of neurons, both central and peripheral, results in allodynia and hyperalgesia. Allodynia is defined as the pain caused by a normally innocuous stimuli, while hyperalgesia is defined as an exaggerated pain in response to noxious

4 Chapter 1 stimuli due to a reduction in the pain threshold. Hyperalgesia is further differentiated into primary and secondary, with primary being a reduced pain threshold at the site of injury and secondary being a reduced pain threshold in surrounding uninjured tissue.

1.1.2. Inflammatory pain

Inflammation is triggered by infection or injury to the microvasculature, the vessels responsible for blood distribution within tissue. Inflammation was first characterised by

Celsus (25 BC – 50 AD) as the four cardinal signs: redness, swelling, heat and pain.12

Inflammation is a beneficial process, with the ultimate aims of the inflammatory response being to remove the causing factor, to prevent tissue damage, to restore normal physiological function and to develop immunity, so that future encounters are met with a faster and more specific response.13 As a response to infection/injury the tissue-resident cells, such as macrophages, lymphocytes and endothelial cells, release pro-inflammatory mediators that include peptides (e.g. cytokines, substance P and tumour necrosis factor alpha), neurotransmitters (e.g. histamine and serotonin) and lipids (e.g. leukotriene B4 and prostaglandin E2) (Figures 1.2. & 1.3.).

Figure 1.2. – Structure of Leukotriene B4. LTB4 is a chemoattractant responsible for neutrophil infiltration to the site of inflammation.

Figure 1.3. – Structure of Prostaglandin E2. PGE2 is pro-inflammatory by increasing vascular permeability, causes vasodilation and hyperalgesia, and also anti-inflammatory by inhibiting leukocyte phagocytosis and triggering production of anti-inflammatory/pro-resolution mediators.

5 Chapter 1

Arachidonic acid is oxidised in inflammation by the inducible cycloxygenase-2 (COX-2) and microsomal prostaglandin E synthase 1 (mPGES-1) enzymes to synthesise prostaglandin E2 (PGE2) (Figure 1.4.). 14, 15 COX has two active sites. The oxidising site converts arachidonic acid to prostaglandin G2 (PGG2). The peroxidase site then converts PGG2 to prostaglandin H2 (PGH2). mPGES-1 isomerises PGH2 into PGE2. mPGES-1 is closely associated with COX-2 and is also induced under inflammatory

15 conditions. Phospholipase A 2 is responsible for producing the majority of arachidonic acid used in prostaglandin synthesis. Arachidonic acid exists esterified in the membrane phospholipids of the cell membrane. Phospholipase A 2 hydrolyses the sn-2 acyl ester bond of membrane phospholipids to release arachidonic acid. 16 Disposition and calcium dependence divide the phospholipase A 2 enzymes into three classes, which include the inducible cytosolic phospholipase A 2 (cPLA 2), the secretory

2+ 17 phospholipase A 2 (sPLA 2) and the Ca -independent (iPLA 2). Most of the arachidonic acid used in PGE2 synthesis by activated macrophages is released via a Ca 2+ -

18 dependent mechanism by cPLA 2. Upon activation, cPLA 2 in macrophages translocates to the membranes of the nucleus and endoplasmic reticulum. 19 Through its activation of its four prostaglandin E (EP) receptors, PGE2 has the pro-inflammatory effect of increasing vascular permeability, 20 facilitating oedema formation and causing vasodilation (widening of the blood vessels). 21

Figure 1.4. – Prostaglandin biosynthesis. Phospholipase A 2 hydrolyses the sn-2 acyl ester bond of membrane phospholipids to release arachidonic acid. COX and prostaglandin E synthase oxidise arachidonic acid to form prostaglandin E2, which through its four EP receptors exerts its pro-inflammatory effects.

6 Chapter 1

Inflammatory cytokines such as tumour necrosis factor alpha (TNF-α), interleukin-1β

(IL-1β) and interleukin-6 (IL-6) are produced to activate the endothelium and induce the acute phase inflammatory response. 22 Additionally IL-6 can cross the blood-brain barrier to induce production of prostaglandins in the CNS, causing fever. 23 The neuropeptide substance P causes vasodilation (widening of the blood vessels) and triggers nociception. 24 The neurotransmitters/vasoactive amines histamine and serotonin causes vasodilation and increase vascular permeability, allowing for the flow of cells out of blood vessels.22 Leukotriene B4 (LTB4) is a potent chemoattractant responsible for the migration of neutrophils to the site of inflammation.25 The lipid mediator platelet-activating factor is a mediator of platelet aggregation, and also functions in leukocyte recruitment, vasodilation and increased vascular permeability. 22

These mediators cause vasodilation and increase vascular permeability by triggering a release of the free radical nitric oxide (NO) from the endothelium.26 The result of its action is increases blood flow to the inflamed tissue, producing the clinical signs of redness and heat, and facilitating exudation and the extravasation of leukocytes.

Exudation is the accumulation of fluid, mostly plasma containing leukocytes, platelets and plasma proteins. This fluid is responsible for the clinical sign of tissue swelling, forming what is known as oedema.

Through the increased permeability of the vasculature and release of chemoattractants by resident cells, leukocyte extravasation, which is the trafficking of leukocytes from the circulation towards the site of inflammation, is made possible. 27 Neutrophils are the first leukocytes to arrive at the site of inflammation,28 having migrated through blood vessels by chemotaxis. The neutrophils follow mediators released from the site of inflammation, such as the chemoattractant leukotriene B4.29 Neutrophil infiltration is followed by the arrival of monocytes, which later differentiate into macrophages. 28

Once at the site of inflammation, the neutrophils and macrophages begin to isolate and phagocytose foreign microorganisms, clear debris and release further pro- inflammatory mediators.30

Macrophages are the major source of PGE2, a key mediator in inflammatory pain. The molecular mechanisms of inflammation have been well studied in lipopolysaccharide

(LPS)-stimulated macrophages. 31 Macrophages are produced by monocyte

7 Chapter 1 differentiation. 32 In inflammation they enter an activated state, becoming more readily able to engulf and digest pathogenic cells, and also becoming more sensitive to their environment by changing the expression of their surface receptors. Macrophages are one of the first cells to respond to infection and inflammation. They migrate into tissues where they secrete pro-inflammatory mediators. There is additionally an inflammatory resolution phase macrophage that functions in the restoration of tissue homeostasis. 33 One of the most potent initiators of inflammation and activators of macrophages is LPS. It is embedded on the outer membrane of Gram-negative bacterial pathogens. Chemically it consists of a hydrophobic group (lipid A) and a hydrophilic polysaccharide chain, with the lipid A attached to the cell wall and the polysaccharide exposed to the extracellular environment. The recognition of LPS and consequent macrophage activation is mediated by a complex of three proteins: CD14, toll-like receptor 4 (TLR4) and myeloid differentiation-2 (MD-2). 34

Macrophages produce both PGE2 and TXA2, with resting macrophages producing TXA2 in favour of PGE2, as COX-2 is not expressed in unactivated macrophages. 35 Initiators of inflammation, like LPS, induce COX-2 expression through activation of nuclear factor-kB (NF-kB). 35 Expression of COX-2 is also regulated by multiple cytokines and growth factors via activation of transcriptional regulatory proteins that act on its gene’s promoter sites. PGE2 produced by activated macrophages is integral to inflammatory pain. 36 Activated macrophages additionally release numerous other mediators including cytokines and nitric oxide.

Inflammatory pain is the result of both peripheral and central processes that are initiated by the pro-inflammatory mediators PGE2 and IL-1β. Peripherally PGE2 does not directly cause pain, instead it lowers the pain threshold through their sensitisation of neurons, leading to hyperalgesia and allodynia. 37 Specifically PGE2 produces hyperalgesia by reducing the threshold for stimulation of the pain-sensing receptors transient receptor potential vanilloid 1 (TRPV1) and the voltage-gated sodium channel

38, 39 Na V1.8. Furthermore it is estimated that 15% of all C-fibres are ‘sleeping’, in that they are only activated upon exposure to mediators, such as PGE2. 7

8 Chapter 1

Pro-inflammatory mediators modulate nociceptor function by causing reversible changes in sensory neurons. Prolonged exposure of sensory neurons to pro- inflammatory mediators is thought to be the mechanism by which acute pain becomes chronic pain. 7 This results in the nociceptors being modified, as opposed to just modulated. Modification of sensory fibres triggers long lasting changes in the expression of receptors and ion channels and in interneuron connectivity. 7

Inflammatory pain also involves a central process, mediated by IL-1ß. This cytokine is capable of crossing the blood-brain barrier, allowing for it to travel into the CNS from the inflamed tissue where it was initially produced. 40 Upon entry into the CNS, IL-1ß induces production of PGE2, further increasing pain sensitivity and ultimately leading to central sensitisation. 41 Drugs that either reduce the synthesis of PGE2, such as inhibitors of COX-2 and mPGES-1, or block the action of PGE2, such as antagonists of

TRPV1, NaV1.8 and the EP receptors, have therapeutic potential in analgesia and inflammation.42

The failure of inflammation to resolve can lead to chronic inflammation, 43 which can ultimately result in auto-immunity and loss of function. 44 Continuing on from Celsus,

Galen (129 – 200 AD) made loss of function the fifth cardinal sign of inflammation. 12

Once the cause of inflammation has been neutralised the highly co-ordinated process of inflammatory resolution can begin. 45 Inflammatory resolution serves to prevent tissue damage, loss of function and auto-immunity. 27 Having completed their task the extravasted inflammatory leukocytes are eliminated. Neutrophils undergo apoptosis/necrosis and then are either phagocytosed by macrophages 46 or recirculated. 47 Removal of macrophages is instead via lymphatic drainage. 48 This elimination of inflammatory leukocytes allows for the once inflamed tissue to be repopulated with resident lymphocytes, ultimately returning to their pre-inflammation numbers and cell state. 49 At this stage there also appears to be a macrophage type unique to inflammatory resolution, which aids in both the elimination of neutrophils and the repopulation of tissue lymphocytes. 33 Resolution continues with the reversal of vasodilation and the permeability of the vasculature returning to normal.

9 Chapter 1

Resolution is controlled by specific pro-resolution lipid mediators, such as the resolvins and lipoxins. The resolvins inhibit further recruitment of leukocytes and the release of pro-inflammatory mediators.50 Lipoxins block entry of neutrophils to the site of inflammation, 51 reduce vascular permeability,52 and stimulate macrophages to ingest and clear apoptotic neutrophils. 53 PGE2 also exerts an anti-inflammatory effect by inhibiting leukocyte phagocytosis, 54 inhibiting production of pro-inflammatory mediators, 55 and triggering production of anti-inflammatory cytokines and pro- resolution mediators. 56 In addition to the on-set of inflammation, COX-2 is also induced during the resolution phase, where it synthesises the anti-inflammatory prostaglandins PGD2 and 5deoxy∆ 12–14 PGJ2.57 Induction of sPLA2 in inflammatory resolution provides the arachidonic acid required for the syntheses of these pro- resolution prostaglandins. 58

1.1.3. Neuropathic pain

Neuropathic pain results from damage to the peripheral or central nervous system.

Nerve injury leads to sensitisation of neurons and the local release of inflammatory mediators. The continuous sensitisation of the neurons increases sensitivity of the dorsal horn resulting in central sensitisation. 59 Changes in the dorsal root ganglion and dorsal horn are also seen with neuropathic pain. Following nerve injury microglia, the macrophages of the brain and spinal cord, are recruited and activated to help initiate and maintain a chronic pain state. 60 While nociceptive pain and inflammatory pain have essential protective roles, neuropathic pain is normally becomes chronic and serves no useful purpose as often persist long after the initial injury has healed. 61

Common causes of neuropathic pain include stroke, trigeminal neuralgia, multiple sclerosis and injury to the spinal cord.3

1.1.4. Analgesia

Despite the importance of pain, relief from it is often required to make injury bearable, to allow physiological function to return to normal and to improve quality of life. Not all pain is protective, such as in chronic cases where pain continues after healing is complete. 62 Analgesia, the relief of pain, is induced either by treating the cause of the

10 Chapter 1 pain, such as inflammation, or by reducing the sensory response to the pain. The majority of analgesic drugs come under two classes: the cyclooxygenase (COX) inhibitors and the opioid receptor agonists. Other classes of drugs, less specific to pain such as steroids, anaesthetics, antidepressants, anticonvulsants and muscle relaxants are also used in the treatment of pain.

1.1.5. COX inhibitors

Non-steroidal anti-inflammatory drugs (NSAID) are a class of analgesics that inhibit

COX (EC 1.14.99.1), either non-selectively inhibiting both the COX-1 and COX-2 isoforms or selectively inhibiting just COX-2. COX-1 expression is constitutive and functions under normal non-pathological conditions, while COX-2 expression is inducible by pro-inflammatory stimuli and functions in inflammation. 63 However there are examples of COX-1 being induced under inflammatory conditions in the kidney, 64 and COX-2 being constitutively expressed in parts of the CNS, 65 specifically the spinal cord. 66

COX synthesises PGH2, via the intermediate PGG2. PGH2 is a substrate for a number of synthases, including prostaglandin E2 (PGE2) synthase (EC 5.3.99.3),67, 68 prostaglandin D2 (PGD2) synthase (EC 5.3.99.2), thromboxane A2 (TXA2) synthase

(EC 5.3.99.5) and prostacyclin synthase (EC 5.3.99.4). The therapeutic effect of COX inhibition is due to the action of PGE2 being blocked, reducing the inflammatory response. By treating the cause of the pain, i.e. the inflammation, the pain is relieved.

Inhibitors of COX-1 & 2 include aspirin, ibuprofen, paracetamol, diclofenac, naproxen and indomethacin (Table 1.1.). Although mostly used for the treatment of inflammatory pain, they are also effective in treating neuropathic pain.69

Inhibitors of COX represent 68% of drugs used for treatment of chronic pain. 3 A significant disadvantage of the NSAIDs is that they can cause gastrointestinal damage, including the formation of ulcers and bleeding. 70 The reason for this is thought to be because COX-1 is expressed in the gut. Thus inhibition of COX-1 by NSAIDs blocks the formation of the protective eicosanoids, PGE2 and prostacyclin, by the gastric mucosal tissue leading to these adverse effects. 71 Prostacyclin inhibits platelet activation and

11 Chapter 1 also acts as a vasodilator. Nausea, vomiting, renal related hypertension and congestive heart failure are also associated with their use.72 By selectively inhibiting COX-2, it was thought that only production of prostaglandins at the site of inflammation would be blocked, thus avoiding the adverse effects associated with the NSAIDs. Unfortunately the COX-2 inhibitors were found to cause hypertension.73 The reason for this is thought to be because COX-2 inhibition favours production of thromboxane over prostacyclin, increasing the risk of cardiovascular thrombotic events. 74, 75

Thromboxane facilitates platelet aggregation and also acts as a vasodilator.

Table 1.1. – Selectivity of COX inhibitors for either isoform expressed as IC 50

(nmol/mL).

COX-1 COX-2 Ratio

Aspirin 1.7 ± 1.1 278 ± 56 166

Ibuprofen 0.05 ± 0.005 2.9 ± 0.4 60

Paracetamol 18 ± 13 132 ± 80 7.4

Diclofenac 1.7 + 0.7 1.2 ± 0.5 0.7

Naproxen 10 ± 4 6 ± 4 0.6

Indomethacin 0.03 ± 0.003 2.0 ± 0.2 60

Data derived from Mitchell et al. 76 Values are the concentrations required to inhibit 50% of COX activity in J774.2 murine macrophages activated with 1 µg/mL LPS (for COX-2) or without (for

COX-1).

1.1.6. Opioid receptor agonists

The opioids, often described as the most effective analgesics available, 77 are agonists of the opioid receptors and represent 28% of drugs used for treatment of chronic pain. 3 Of the three opioid receptors: µ, k and δ, it is the µ-opioid receptor that is the target of many of the opioid analgesics. The endogenous opioids for these receptors are the neuropeptides β-endorphin, leu-enkephalin, met-enkephalin and dynorphin.

The endorphins act as endogenous analgesics, being released in response to pain to help suppress its signal transduction. 5

The main opioid analgesics are morphine, codeine and methadone. Their activation of the opioid receptors initiates analgesia by reducing neuronal excitability, inhibiting neurotransmitter release and suppressing the firing of neurons. Whilst the opioids are

12 Chapter 1 powerful analgesics, they frequently cause nausea, vomiting, constipation (hence the low dose use of morphine as a treatment for diarrhea) and can alter cognitive function. 78 The target of the opioids, the µ-receptor, mediates both the analgesia and its adverse effects, including dependence. 79 A further serious risk is tolerance, the progressive decline in a drug’s ability to induce its desired effect with prolonged use.

1.1.7. Glucocorticoids

Glucocorticoids, such as cortisol and dexamethasone, are steroids that suppress the expression of pro-inflammatory cytokines and induce the expression of anti- inflammatory cytokines via the glucocorticoid receptors. 80 Additionally through the inhibition of phospholipase A2 (EC 3.1.1.4) and suppression of COX-2 expression, they reduce both prostaglandin and leukotriene biosynthesis. 81 Adverse effects of the glucocorticoids include ulcers, gastrointestinal bleeding and insulin resistance.82

Steroids represent only 2% of the drugs used for treatment of chronic pain. 3

1.1.8. Anaesthetics

Anaesthetics are used to cause a reversible loss of sensory perception, including pain sensation. This action is in contrast to that of analgesics, which are used to relieve pain with no other loss of sensation. Typically anaesthetics are intracellular Na + channel blockers, which work by decreasing the rate of depolarisation and repolarisation of sensory neurons. This is achieved through inhibition of Na + influx through voltage-gated Na + channels, which consequently prevents action potentials and the relay of signals. 83 By preventing signal transduction anaesthetics are capable of blocking the action of nociceptors.

Local anaesthetics are effective in treating the site of pain, with lidocaine being an example of such an anaesthetic that is routinely used topically to relieve pain. 84

Another routinely used class of anaesthetics are the NMDA receptor antagonists. N- methyl-D-aspartic acid (NMDA) is an amino acid derivative that acts as an agonist for the ionotropic NMDA receptor. When the receptor is open signal transfer can occur

13 Chapter 1 from neurons through to the brain. NMDA receptor antagonists are used to stop the relay of this signal through the blocking of the channel.85

1.1.9. Antidepressants

Antidepressants, whilst normally used in the treatment of depression and related disorders, are also used in the long-term treatment of chronic neuropathic pain.

Antidepressants, such as the tricyclic antidepressants, work as serotonin- norepinephrine reuptake inhibitors (SNRI). 86 By preventing reuptake of serotonin and norepinephrine, antidepressants increase the synaptic availability of these two neurotransmitters, enhancing neurotransmission and consequently increasing the pain threshold. 87 Selective serotonin reuptake inhibitors (SSRI) are often preferred as they typically have fewer adverse effects than the SNRIs. Amitriptyline is the most frequently used of the antidepressants as an analgesic. Antidepressants represent 3% of drugs used for treatment of chronic pain. 3

1.1.10. Anticonvulsants

Anticonvulsants are normally used in the treatment of seizures, but are also effective for pain. Their mechanism of action is to block voltage-gated Ca 2+ channels to reduce

Ca 2+ influx and in doing so reduce synaptic release of neurotransmitters such as glutamate, norepinephrine and substance P. 88, 89 Anticonvulsants reduce pain through this effect on signal transduction. 90 Widely used anticonvulsants for treatment of neuropathic pain include GABApentin, its more potent successor pregabalin and carbamazepine. Anticonvulsants only represent 2% of drugs used for treatment of chronic pain. 3

1.1.11. Muscle relaxants

Muscle relaxants are normally used for the treatment of spasm. The reason for their use as analgesics is due to pain often causing muscle spasm, which can result in a cycle of spasm causing more pain and pain causing more spasm. Muscle relaxants can be used to break this cycle. Routinely used muscle relaxants for pain include

14 Chapter 1 carisoprodol, cyclobenzaprine and metaxalone. 91 Muscle relaxants only represent 1% of drugs used for treatment of chronic pain. 3

1.1.12. An alternative analgesic pathway

To avoid the adverse effects associated with the various analgesics an approach known as multimodal analgesia is often used. 92 This involves a balanced combination of analgesia via multiple mechanisms, rather than through a single pathway. The concept being that instead of administering a large dose of an opioid, a lower dose could be used in combination with a second analgesic such as a COX inhibitor or paracetamol.

This should minimise the likelihood of toxicity and associated adverse effects. However this approach does not always translate into improved function and a return to normality.92

The World Health Organisation suggests a three-step analgesic "ladder" in treating pain. 93 The first step starts with paracetamol and the NSAIDs for the treatment of mild pain. This is followed by weak opioids, such as codeine, for the treatment of moderate pain. Finally strong opioids, such as morphine, are used for the treatment of severe pain. Each step suggests the use of an adjuvant to complement the analgesic, such as antidepressants, anticonvulsants, steroids or muscle relaxants. The principle of the analgesic ladder is to progress through each step if pain is still present. This approach was found to be effective in 76% of patients undergoing pain treatment, while 12% only found the approach to be satisfactory and a further 12% found it inadequate. 94

Improvement in the survival rate of major diseases, such as cancer, 95 has caused a rise in the number of patients requiring disease-related and post-surgery treatment for pain. Additionally the result of longer life spans has increased the number of age- related conditions that require pain management. The population of people requiring pain treatment is therefore growing. 5 Furthermore a combination of partial efficacy, adverse effects and potential for abuse mean that the currently used analgesics are far from ideal.

15 Chapter 1

In summary, the treatment of pain therefore often relies on balancing the effectiveness of the analgesia and the adverse effects of the therapy. There is still a need for research into new analgesics that work with less adverse effects and by mechanisms more specific to pain. A prominent alternative target for the development of novel analgesic and anti-inflammatory drug is the endocannabinoid system. Current analgesics and anti-inflammatories mostly inhibit the nociceptive and pro-inflammatory pathways that are behind the cause of the pain and inflammation. In contrast drugs that target the endocannabinoid system would allow for endogenous analgesic and anti-inflammatory pathways to instead be enhanced.

1.2. Endocannabinoid system

For millennia Cannabis has been used in the treatment of pain.96 66 bioactive components were identified by Mechoulam, with the majority of the effects of Cannabis being found to be produced by ∆9-tetrahydrocannabinol (∆9-THC).97 The binding sites of the bioactive components of Cannabis were discovered to be two G-protein coupled receptors , which were named the cannabinoid receptors.98, 99 Later the endogenous ligands of these receptors were identified,100, 101 followed by the enzymes involved in their synthesis and metabolism, 102, 103 and the carrier proteins involved in their transport. 104 Together the cannabinoid receptors, their endogenous ligands and the enzymes involved in their biosynthesis and degradation comprise what is now known as the endocannabinoid system.

1.2.1. Cannabinoid receptors

The cannabinoid receptors (Figure 1.5.) are G-protein coupled receptors, of which there are two subtypes: CB1 105 and CB2.99 CB1 is coded for by the CNR1 gene, which is 26,496 bases in length and located on human chromosome 6 at position 6q14- q15. 106 It is highly expressed in the CNS, particularly in the forebrain and to a lesser extent in the midbrain, hindbrain and spinal cord.107 CB1 is the most highly expressed of the G-protein coupled receptors in the brain,108 functioning in neurotransmission, with expression being on pre-synaptic neurons and activation inhibiting neurotransmitter release of GABA and glutamate.109

16 Chapter 1

Figure 1.5. – Structure of the cannabinoid receptor 1 (CB1). CB1 is a G-protein coupled receptor consisting of 7 transmembrane helices. It is predominantly expressed in the CNS where it is activated extracellularly by the endocannabinoids. Image reproduced from Shim et al. 110

CB2 is coded for by the CNR2 gene, which is 88,534 bases in length and located on human chromosome 1 at position 1p36.11.111 It is an important regulator of inflammation, with expression mainly being peripheral on immune cells, specifically macrophages, neutrophils, B-cells and T-cells. 112 Within the CNS CB2 is found primarily on microglia, the macrophages of the brain, 113 and within the brainstem. 114 CB2 controls immune cell mobility and production of immunomodulators,115 with its activation inhibiting the release of pro-inflammatory mediators, such as PGE2, nitric oxide and cytokines, and also inhibiting the expression of COX-2. 116

CB1 is 472 amino acid residues in length 117 and has a molecular weight of 57 kDa. 118

CB2 is smaller with 360 amino acid residues 119 and a molecular weight of 41 kDa. 120

Both receptors have an extracellular glycosylated N-terminus and an intracellular C- terminus. 119, 121 The two receptors share a sequence homology of 44%. 122 After

17 Chapter 1 activation, multiple intracellular signal transduction pathways are activated via G- protein coupling, with activation shown to attenuate pain.

The dissimilarity between CB1 and CB2 receptor activity is not just due to the different physiological locations of their expression, but also due to their differing mechanisms of signal transduction. 123 Upon activation, CB1 mediates the closing of Ca 2+ channels

+ and the opening of K channels via the G i/o protein. By a separate pathway mediated

124 by the G s protein, CB1 modulates adenylate cyclase (EC 4.6.1.1) activity. In contrast, CB2 receptor signalling involves the activation of mitogen-activated protein kinase (MAPK) (EC 2.7.11.24) via the binding of the Gi/o protein's Gα subunit. CB2 activation also stimulates β-endorphins release, which in turn activates µ-opioid receptors for inhibition of nociception.125 Despite these separate pathways, a similarity shared by both receptors is the ability to inhibit adenylate cyclase via the G i/o protein, 126 a coupling used by many G-protein coupled receptors that mediate pain transmission.127

1.2.2. Endocannabinoids

Ligands for the cannabinoid receptors are known as cannabinoids, of which five endogenous cannabinoids, the endocannabinoids, have been identified (Figure 1.6.).

These five lipids are all derivatives of arachidonic acid. Typical to lipid signalling molecules, and unlike neurotransmitters and hormones, the endocannabinoids are synthesised on demand, where they act near the site of their synthesis and are eliminated on the completion of their action.103 Unlike the endorphins that are produced in response to pain, in vivo studies have shown that the endocannabinoids can be increased, decreased and remain unaltered in various pain states. 128

Through activation of the CB1 receptor on neurons and the CB2 receptor on immune cells, the endocannabinoids are important regulators of neurotransmission and inflammation. The principal endocannabinoids are arachidonoyl ethanolamide (AEA, anandamide) and 2-arachidonoyl glycerol (2-AG). AEA, the amide of arachidonic acid and ethanolamine, was the first endocannabinoid to be identified, where it was discovered in brain tissue. 100 As well as the cannabinoid receptors, AEA is also an

18 Chapter 1 agonist for TRPV1,129 with the descending order of AEA’s affinity for its receptors being

CB1, TRPV1 and CB2. 130

Figure 1.6. – Structures of the endocannabinoids alongside arachidonic acid.

2-AG, the ester of arachidonic acid and glycerol, was later also identified in brain tissue,101, 131 and has been found to be neuroprotective, being released following brain injury. 132 Compared to AEA, 2-AG has less affinity for both of the cannabinoid receptors. 133 The most abundant of the endocannabinoids is 2-AG, which exists at a concentration approximately 1000-fold higher than AEA in brain and between 4- and

40-fold higher in plasma (Table 1.2.).

Noladin ether (2-arachidonylglyceryl ether, 2-AGE), the third endocannabinoid to be identified, is a stable analogue of 2-AG and thus has a longer biological half-life. 134 It was first identified in brain tissue, 135 although its presence in this tissue has been disputed with conflicting studies reportedly detecting 136 and not detecting the lipid.137

Noladin ether is only a weak CB2 agonist, binding preferentially to CB1. 135

NADA (N-arachidonoyl dopamine), the fourth endocannabinoid to be discovered, is the conjugate of arachidonic acid and the neurotransmitter dopamine. It was first

19 Chapter 1 synthesised as a dopamine analogue, but was found to have no affinity for the dopamine receptors. 138 Later it was identified as an endogenous component of brain tissue, 139 having affinity for the cannabinoid receptors and a 40-fold selectivity for CB1 over CB2. 138

Virodhamine (O-arachidonoyl ethanolamide, O-AEA), the latest endocannabinoid to be identified, is an analogue of AEA, where the arachidonic acid is joined to the ethanolamine by an ester linkage as opposed to the amide linkage of AEA. It was first discovered in brain tissue and functions as a CB2 agonist and a CB1 antagonist.140

Table 1.2. – The endogenous concentration ranges of the endocannabinoids.

Rat brain Most abundant brain region Ref. Human plasma Ref.

AEA 13 – 32 pmol/g Cerebellum 141 0.3 – 1.5 pmol/mL 142

2-AG 10 – 30 nmol/g Hypothalamus 141 5 – 10 pmol/mL 142

Noladin ether 4 – 65 pmol/g Thalamus 136 -

Virodhamine 4 – 20 pmol/g Cerebral cortex 140 -

NADA 0.2 – 7 pmol/g Striatum 139 -

Data combined from various referenced sources. Noladin ether, virodhamine and NADA are yet to be identified in human plasma.

AEA and 2-AG are the most studied of the five endocannabinoids. The pathways of their biosynthesis and degradation by fatty acid amide hydrolase (FAAH) and monoacylglycerol lipase (MAGL) have been fully elucidated. They are readily detectable in numerous tissues and fluids including the brain, plasma, cerebrospinal fluid, amniotic fluid and breast milk. 143

The other endocannabinoids noladin ether, virodhamine and NADA, remain undetected in plasma (Table 1.2.),142 and conflicting reports exist as to whether they are actually present in the brain.141 Additionally, unlike AEA and 2-AG, the pathways of their biosynthesis and metabolism are currently not elucidated.

20 Chapter 1

1.2.3. Endocannabinoid biosynthesis

Endocannabinoid biosynthesis is triggered by either cellular depolarisation or receptor activation. Receptors found to trigger AEA synthesis include the TRPV1, 144 dopamine

145 and muscarinic acetylcholine receptors. 146 For 2-AG, synthesis can be triggered by activation of the NMDA receptor. 147

The biosynthesis of AEA begins with the formation of the precursor N-arachidonoyl- phosphatidyl-ethanolamine (NAPE) from arachidonic acid and a glycerophospholipid by acyltransferase. This is followed by the subsequent hydrolysis of NAPE to AEA and phosphatidic acid by either NAPE-phospholipase D (NAPE-PLD) (EC 3.1.4.4), phospholipase A2 (PLA2), α/β-hydrolase 4 (Abh4) (EC 3.1.1.-) or tyrosine phosphatase

2 (EC 3.1.3.16).148, 149 There is still uncertainty as to which enzyme is responsible for the majority of AEA biosynthesis. This same mechanism has been described as the route of synthesis for the structural analogues of AEA, oleoyl ethanolamide (OEA) and palmitoyl ethanolamide (PEA), but with precursor formation being instead from oleic acid and palmitic acid respectively. 150 The generation of AEA from arachidonic acid and ethanolamine by reverse FAAH activity has also been proposed.151 Additionally mechanisms for AEA recycling exist, whereby the arachidonic acid and ethanolamine molecules produced from AEA deactivation are used for AEA re-synthesis and release. 152

While interconversion between AEA and 2-AG can occur, 153 2-AG is primarily formed as an intermediate in lipid metabolism. Phospholipase C (EC 3.1.4.3) cleaves membrane phospholipids forming diacylglycerol (DAG), which is then Ca 2+ -dependently hydrolysed by diacylglycerol lipase (DAGL) (EC 3.1.1.34) into 2-AG. 154

1.2.4. Endocannabinoid metabolism

The main degradative enzymes of the endocannabinoids are the AEA metabolising fatty acid amide hydrolase (FAAH) (EC 3.5.1.99) and the 2-AG metabolising monoacylglycerol lipase (MAGL) (EC 3.1.1.23), both enzymes of which belong to the serine hydrolase enzyme family.

21 Chapter 1

FAAH (Figure 1.7.) is an intracellular membrane-bound enzyme that is expressed abundantly in the brain and liver, and at lower levels in the kidney, spleen and leukocytes.102 It is associated with the microsomal, mitochondrial and plasma membranes. 155 FAAH is coded for by the FAAH gene, which is 19,584 bases in length and located on human chromosome 1 at position 1p35-p34. 156 The enzyme’s structure is homodimeric, composed of 579 amino acid residues and a molecular mass of approximately 63 kDa. 157 The function of FAAH is to hydrolyse AEA and its two structural analogues: OEA 158 and PEA,159 into their component fatty acid and ethanolamine. FAAH degrades AEA with a higher selectivity than for OEA and PEA. 160 A specific binding pocket exists within the active site to accommodate the long fatty acid chains present on its substrates. 161 A relatively unselective binding domain is capable of accepting various fatty acid chains, with various head groups. 162 While the main attributed function of FAAH is to hydrolyse AEA, OEA and PEA into arachidonic acid, oleic acid and palmitic acid respectively. FAAH has also been demonstrated to conjugate arachidonic acid to various small molecules, including ethanolamine, dopamine and 4-aminophenol to synthesise AEA, 151 NADA 163 and AM404. 164

Figure 1.7. – Structure of fatty acid amide hydrolase (FAAH). FAAH is a dimeric and intracellularly membrane bound enzyme responsible for AEA hydrolysis. Image reproduced from

Mileni et al. 165

22 Chapter 1

MAGL (Figure 1.8.) is monomeric and whilst mostly membrane bound, it is also present in the cytosol. 166 Most tissues express MAGL, such as the kidneys and adipose tissue where it is abundant and in the brain and adrenal glands where it is moderately expressed. 167 MAGL is coded for by the MGLL gene, which is 150,111 bases in length and located on human chromosome 3 at position 3q21.3. 168 With a length of 313 amino acid residues and a molecular mass of approximately 33 kDa, MAGL is significantly smaller than FAAH.167 Despite FAAH being capable of degrading 2-AG in vitro ,169 its in vivo degradation is principally catalysed by MAGL, whereby it is cleaved into arachidonic acid and glycerol. MAGL catalyses 85% of 2-AG hydrolysis and also functions in the hydrolysis of other monoglycerides. 166

Figure 1.8. – Structure of monoacylglycerol lipase (MAGL). MAGL is a monomeric and cytosolic enzyme responsible for 2-AG hydrolysis. Image reproduced from Bertrand et al. 170

In addition to hydrolytic metabolism of AEA and 2-AG, multiple routes for oxidative metabolism exist via cyclooxygenase (COX), lipoxygenase (LOX) (EC 1.13.11.33) and cytochrome P450 (Figure 1.9.).171 COX metabolises AEA and 2-AG to prostaglandin E2 ethanolamide (PG-EA) 172 and to prostaglandin E2 glycerol (PG-G).173 LOX metabolises

AEA and 2-AG to hydroperoxyeicosatetraenoic acid ethanolamide (HPETE-EA) 174 and

23 Chapter 1 to HPETE-glyceroyl (HPETE-G). 175 Cytochrome P450, specifically the CYP2D6 isoform, 176 metabolises AEA to hydroxyeicosatetraenoic acid ethanolamide (HETE-EA) and to epoxyeicosatrienoic acid ethanolamide (EET-EA), 177 and metabolises 2-AG to

HETE-G.178 The studies that identified the oxidative metabolites of AEA used high AEA concentrations incubated on either pure enzymes or liver microsomes. Physiological production of these metabolites in vivo under normal conditions is yet to be shown.

Figure 1.9. – The multiple routes of endocannabinoid metabolism. In addition to hydrolysis by FAAH and MAGL, the endocannabinoids can also be oxidised by COX, LOX and cytochrome P450.

Arachidonic acid is also oxidised by COX, LOX and cytochrome P450 to form the prostaglandins, thromboxane, leukotrienes, lipoxins and epoxyeicosatrienoic acids.56

These oxidative metabolites of arachidonic acid are all well known endogenous lipid mediators. 179 In tissue AEA is a lot less concentrated than arachidonic acid, for example the brain concentration of arachidonic acid is 8 µmol/g compared to 13 – 32 pmol/g for AEA. 141, 180 Therefore if such metabolites are produced physiologically it is likely that they are present only in low amounts, and would require very sensitive assays for their detection.

24 Chapter 1

A recent study failed to identify oxidative metabolites of AEA in the tissue samples of normal untreated rats. 179 The only in vivo study to have detected these metabolites used FAAH knock-out mice treated with intravenous AEA, which resulted in production of the COX derived metabolites PGE2-ethanolamide and PGF2α-ethanolamide in the liver, kidney, lung and small intestine. 181 However these metabolites were both absent in untreated control mice and only PGE2-ethanolamide was present in AEA-treated normal mice. Unfortunately this study did not investigate the brain concentrations of these metabolites. As AEA concentrations are high in the brain there should also be significant amounts of the oxidative metabolites formed here.

There is also indirect evidence for oxidative metabolism of AEA. For example COX-2 inhibition increases AEA levels 182 and the analgesic effect of COX-2 inhibitors is reversed by cannabinoid antagonists. 183 Furthermore the clinical trial of the FAAH inhibitor PF-04457845 showed that following an 8 mg dose the plasma levels of AEA,

OEA and PEA returned to normal after 10 days, while FAAH activity only returned to normal after 15 days. 184 This indicated that there were 5 days when substrate levels were being metabolised despite no FAAH activity. Together these studies suggest that under specific conditions the metabolism of AEA by COX-2 may be a physiologically occurrence. Therefore production of oxidative metabolites of AEA may occur in tissues when there are simultaneously high AEA concentrations and high expression of oxidising enzymes, although the concentration of such metabolites would be low and their physiological relevance is unlikely to be significant. Such conditions are a possibility following FAAH inhibition in inflamed tissue and the liver, where high concentrations of oxidising enzymes are present alongside significant FAAH expression.

Other endocannabinoid metabolising enzymes include N-acylethanolamine acid amidase (NAAA) (EC 3.5.1.-) 185 and the subtype fatty acid amide hydrolase-2 (FAAH-

2). 160 Their mechanisms are similar to FAAH in that they both catalyse AEA hydrolysis.

They are expressed at low levels and are not responsible for the majority of endocannabinoid metabolism, therefore they are not seen as integral enzymes in the endocannabinoid system. 128

25 Chapter 1

1.2.5. Endocannabinoid signalling

Neurotransmitters are synthesised and stored intracellularly in vesicles until release.

Upon release from presynaptic neurons they activate receptors on postsynaptic neurons, therefore the presynapse sends signals and postsynapse receives.

Endocannabinoids work differently. They are classed as retrograde transmitters as they travel against the flow of other transmitters. They are released from the postsynapse to act on the presynapse. The 6 steps of endocannabinoid signalling are

(1) trigger of biosynthesis, (2) biosynthesis, (3) postsynaptic endocannabinoid release,

(4) presynaptic receptor activation, (5) reuptake and finally (6) endocannabinoid inactivation (Figure 1.10.).

Figure 1.10. – Endocannabinoid signalling. AEA is synthesised and released on demand from the postsynapse. Once in the synaptic cleft it is free to activate the adjacent presynaptic CB1 receptor. On completion of its action it returns to the postsynapse, firstly undergoing reuptake and finally metabolism viva FAAH.

The depolarisation of postsynaptic neurons results in a Ca 2+ influx that triggers the release of a retrograde messenger. The function of this messenger is to suppress further neurotransmitter release from the presynaptic cleft.186-188 This is known as depolarisation-induced suppression of inhibition and the retrograde messenger has now been identified as the endocannabinoids. 189 As well as depolarisation, activation of

26 Chapter 1 specific receptors also triggers endocannabinoid biosynthesis and release. While the endorphins are clearly produced in response to pain, in in vivo models the endocannabinoids increase, decrease and remain unaltered under different pain states.128 The endocannabinoids are synthesised in postsynaptic neurons. Due to their lipid nature the endocannabinoids are not water soluble and therefore cannot be stored in vesicles. The endocannabinoid precursors are present in the plasma membrane, allowing for rapid diffusion into the synaptic cleft on completion of their biosynthesis.190 It is thought that they leave the cell as soon as they are synthesised. 190

The endocannabinoids have only a short range of action. Once released from their postsynaptic neuron they are free to activate cannabinoid receptors on their adjacent presynaptic neuron. The endocannabinoids then cross the cell membrane again to be inactivated by their intracellular metabolising enzymes. This process of reuptake is by high-affinity membrane transport. Extracellular AEA is transported by a selective carrier protein across the membrane in a FAAH independent process 191 before being metabolised.192 This creates a concentration gradient that drives AEA uptake. 193

AEA transport has not been fully elucidated. Transport for other lipids and transmitters, such as oleic acid,194 the prostaglandins 195 and serotonin 196 is carrier- mediated, making it likely that a dedicated transporter exists also for AEA. Additionally

AEA analogues have been synthesised that increase AEA levels without inhibiting

FAAH, hence blocking a transport mechanism. 96 A high-affinity AEA binding site,197 brain expressed fatty acid binding proteins (FABP), 198 and a FAAH variant that lacks

FAAH activity,104 have all been identified as selective AEA carrier proteins. 2-AG uptake is also by carrier mediated transport,199, 200 although a protein for this action is yet to be identified.

Finally, on completion of their action, endocannabinoids are enzymatically inactivated.

FAAH is present on postsynaptic neurons, while MAGL is on presynaptic neurons,201 thus 2-AG is inactivated presynaptically and AEA is inactivated postsynaptically.

27 Chapter 1

1.3. Cannabinoid receptor-mediated analgesia

1.3.1. Cannabinoid receptor knock-out and activation

The connection between activation of the cannabinoid receptors and pain relief is well established. CB1 knock-out mice have a reduced pain sensation to noxious stimuli alongside the adverse effects of hypomotility, catalepsy and an increased mortality. 202

Selective agonists for CB1 are effective in the treatment of nociceptive, 122 inflammatory 122 and neuropathic pain. 203 CB2 knock-out mice also exhibit a lower than normal pain threshold to noxious stimuli. 204 Due to the receptor’s important role in immunomodulation, CB2 knock-out mice lose their normal immune response. 205

Interestingly CB2 expression in the spinal cord is increased with nerve injury. 206, 207

Selective agonists for CB2 are effective in the treatment of nociceptive, 204 inflammatory 208, 209 and neuropathic pain. 210

While both cannabinoid receptors are clearly integral to the mechanism of pain, the

CB1 subtype is a major obstacle in the development of cannabinoid analgesics due to the receptor’s high expression within the CNS. Activation of this receptor is psychotropic, with the adverse effects of dysphoria, amnesia and dizziness.

Additionally, CB1 agonists induce the “tetrad of behaviours”, which have become indicative of CB1’s activation and are defined as hypothermia, catalepsy, hypomotility and the desired effect of analgesia.211 Activation of CB1 receptors in the CNS is therefore to be avoided. As CB2 is not as highly expressed in the CNS, its activation is devoid of the CB1-associated adverse effects. 212

The routinely used analgesics, the COX inhibitors, already indirectly target the cannabinoid receptors. As COX metabolises the endocannabinoids, the inhibition of

COX-2 increases AEA 182 and 2-AG 183 levels. The role of the cannabinoid receptors in the action of COX inhibitors is supported by in vivo evidence showing that CB1 antagonism reverses their analgesic effect. 183, 213 This suggests that the inhibition of

COX-mediated endocannabinoid metabolism and the consequent cannabinoid receptor activation is just as integral to the action of COX inhibitors as their inhibition of prostaglandin biosynthesis.

28 Chapter 1

1.3.2. Endogenous cannabinoid receptor ligands

The use of the endocannabinoids in pain therapeutics has been well investigated. In vivo administration of AEA (20 mg/kg i.p. ) induces analgesia with hypothermia.214 The effectiveness of this pain relief is dependent on the location of AEA injection, with more potency seen when administered at the site of pain as opposed to systemic administration. 215 2-AG (30 mg/kg i.v.) also induces analgesia in vivo , but along with the other “tetrad of behaviours”.101 There is therefore an apparent difference between cannabinoid activation by AEA and by 2-AG, with AEA being devoid of the adverse effects seen with 2-AG. The endocannabinoid, virodhamine (2.5 mg/kg i.v.) , is a CB1 antagonist, dampening/blocking receptor activation.50 In vivo virodhamine is a less potent analgesic than AEA and 2-AG.140 Noladin ether (20 mg/kg i.p. ) causes only mild anti-nociception in vivo .135 The in vivo administration of NADA has the opposite effect to the other endocannabinoids in that it induces hyperalgesia, an increased sensitivity to pain, via TRPV1. 139

1.3.3. Other cannabinoid receptor ligands

The non-endogenous cannabinoids are split into the phytocannabinoids and the synthetic cannabinoids. The phytocannabinoids are the active components of the

Cannabis plant, which includes the agonists ∆9-THC and cannabinol (Figure 1.11.). Like the endocannabinoids, the phytocannabinoids are lipophilic although they are not derivatives of arachidonic acid. These compounds activate CB1 in the CNS, explaining their psychotropic effects. 216

Figure 1.11. – Structure of cannabinol.

Many synthetic cannabinoids have been synthesised. To avoid the adverse effects associated with central CB1 activation, agonists selective for CB2 have been developed, 212 as have peripherally restricted CB1 agonists that lack the ability to cross

29 Chapter 1 the blood-brain barrier. 217 Both are effective mechanisms for inducing analgesics without the associated adverse effect. In summary, activation of the cannabinoid receptors induces an analgesic and anti-inflammatory response, but adverse effects are seen with CB1 activation in the CNS. Alternative routes have been sought to avoid this, including selective CB2 agonists and peripherally restricted CB1 agonists.

However more potent analgesia can be achieved through the inhibition of the endocannabinoid metabolising enzymes.

1.4. FAAH-mediated analgesia

Inhibition of the enzymes involved in the degradation of the endocannabinoids allows for increased activation of the cannabinoid receptors without the use of direct cannabinoid agonists. FAAH and MAGL have become potential drug targets as their inhibition increases AEA and 2-AG levels respectively, and consequently increases cannabinoid receptor activation. Inhibition should also preserve the endocannabinoid levels where cannabinoid signalling is already occurring, enhancing their action.

1.4.1. The preference of FAAH over MAGL for the development of analgesics

For the purpose of analgesia the inhibition of FAAH is preferred to MAGL for several reasons. The selective MAGL inhibitor, JZL184 (Figure 1.12), has been shown in vivo to increase the brain concentration of 2-AG 8-fold without altering levels of AEA. 218

Therefore the action of MAGL is exerted via 2-AG, not AEA.

Figure 1.12. – Structure of JZL184.

30 Chapter 1

The accumulation of 2-AG as a result of MAGL inhibition increases cannabinoid receptor activation resulting in analgesia, but also hypothermia and hypomotility.

Unlike 2-AG, activation of the cannabinoid receptors by AEA does not cause these adverse effects, highlighting a difference in action between these two endocannabinoids. This is due to a difference in receptor activation, which is demonstrated by the finding that chronic inhibition of MAGL causes desensitisation of the CB1 receptors in the brain and consequently, over time, the analgesic effect is diminished.219 Unlike MAGL, chronic inhibition of FAAH does not cause desensitisation.

MAGL is an integral enzyme for the degradation of monoglycerides, its inhibition results in an unwanted accumulation of other monoglycerides unrelated to pain.220

Again unlike MAGL, all the substrates that accumulate following FAAH inhibition are analgesic and anti-inflammatory.

1.4.2. FAAH knock-out and inhibition

FAAH knock-out mice have a 15-fold higher brain concentration of AEA than normal mice,221 without 2-AG levels being disrupted.222 This consequently causes increased activation of the cannabinoid receptors and a higher pain threshold. As 2-AG levels are not affected by FAAH inhibition, AEA is therefore the sole endocannabinoid responsible for this effect. Separately both CB1 and CB2 antagonists reduce the analgesic effect of

FAAH inhibition in vivo , and together they completely abolish it. 223 FAAH inhibitors designed to be unable to cross the blood-brain barrier are still effective analgesics, 224 demonstrating the importance of FAAH localised in the periphery not just the CNS.

These studies verify that it is AEA’s activation of both cannabinoid receptor subtypes that causes FAAH-mediated analgesia.

In vivo studies on FAAH inhibition have shown effective treatment of pain and inflammation without any disruption of motility, cognition and body temperature. 225 In vivo FAAH inhibition is an effective treatment of both inflammatory 223, 226 and neuropathic pain. 227-229 Therefore FAAH has become an attractive target for the development of novel analgesics.

31 Chapter 1

1.4.3. The multiple mechanisms of FAAH-mediated analgesia

FAAH inhibitors are devoid of the adverse effects seen with cannabinoid agonist. This is because analgesia from FAAH inhibition is not just from the cannabinoid receptors. The substrates of FAAH (Figure 1.13.), AEA, OEA and PEA, all increase in concentration with the enzyme’s inhibition.230 Despite not sharing AEA’s affinity for the cannabinoid receptors, 231 OEA and PEA also have analgesic and anti-inflammatory action, which they exert in the CNS and periphery via other non-cannabinoid pathways.

Figure 1.13. – Structures of the FAAH substrates AEA, OEA and PEA.

1.4.4. The action of OEA via the PPAR-α and TRPV1 receptors

OEA is synthesised on demand in the brain and degraded on termination of its action in a manner comparable to AEA. 232 OEA has been shown to have an analgesic effect in both visceral (internal organ) and inflammatory pain,232 through its activation of the peroxisome proliferator-activated receptor alpha (PPAR-α). This receptor belongs to the nuclear receptor family and functions in the response to extracellular lipid stimuli such as fatty acids, leukotrienes and cholesterol. 233 The main physiological action of

PPAR-α is as a transcription factor, modulating expression of specific genes that control the inflammatory response, such as those for the cytokines. 234 Other ligands of the receptor include the endocannabinoids AEA, noladin ether and virodhamine. 235 The antagonism of PPAR-α reverses the analgesic effect of FAAH inhibition, demonstrating

32 Chapter 1 an integral role for PPAR-α in the therapeutic effect of FAAH inhibition. 236 Furthermore

PPAR-α knock-out mice have an overactive inflammatory response, supporting the important role of this receptor in inflammation.237

OEA is also an agonist for another pain-associated receptor, TRPV1. 238 TRPV1 is a cation channel that is exclusively expressed on the nociceptive Aδ and C fibres of the peripheral nervous system. 239 The receptor is also activated by noxious heat (greater than 43 oC), acidity and the irritant capsaicin. The receptor can be sensitised, in that its response to noxious stimuli is exaggerated in the presence of inflammatory mediators, such as the prostaglandins. 38, 240 The significance of TRPV1 as a pain receptor has been demonstrated with TRPV1 knock-out mice, showing reduced nociception and pain sensation to inflammation and noxious stimuli. 241, 242 Interestingly both TRPV1 agonists and antagonists are effective analgesics.164, 243 In summary, OEA is analgesic and anti- inflammatory through its binding to PPAR-α and TRPV1.

1.4.5. The action of PEA via the PPAR-α receptor and other pathways

The activity of PEA is much more established than it is for OEA. PEA acts as an anti- inflammatory 244, 245 and analgesic. 246, 247 The anti-inflammatory effect is via both control of mast cell activation 244 and activation of the PPAR-α receptor. 248, 249 In addition to PPAR-α, part of the anti-inflammatory effect of PEA is by PPAR-γ. 250

Interestingly, despite lacking affinity for CB2, the action of PEA is inhibited by CB2 antagonists. 235 This suggests that PEA potentiates the endocannabinoid system by an indirect mechanism. Furthermore PEA has a role in neuroinflammation, regulating microglial migration at the site of damaged tissue, via a non-cannabinoid G i/o -protein mechanism. 251

1.4.6. The action of the FAAH substrates via GPR55

The G-protein coupled receptor GPR55 has been identified as a receptor for OEA and

PEA. The endocannabinoids and ∆9-THC are also GPR55 ligands, with AEA activating

GPR55 with the same potency as it does for CB1 and CB2. Interestingly the other

33 Chapter 1 endocannabinoids 2-AG, noladin ether and virodhamine all activate GPR55 with higher potencies than they do for the cannabinoid receptors (Table 1.3.).252

Table 1.3. – Potencies of the endocannabinoids and FAAH substrates at each receptor as EC 50 (pmol/mL).

CB1 CB2 GPR55 PPAR- α

Endocannabinoids

AEA 31 ± 6 27 ± 6 18 ± 3 28,000

2-AG 519 ± 48 618 ± 45 3 ± 1 -

Noladin ether 37 ± 5 > 30,000 10 ± 1 10,000

Virodhamine 2,920 ± 325 381 ± 34 12 ± 3 30,000

NADA 250 12,000 - -

FAAH substrates

OEA > 30,000 > 30,000 440 ± 145 120 ± 1

PEA > 30,000 19,800 ± 2,821 4 ± 1 3,100 ± 400

The potencies of the endocannabinoids and the FAAH substrates for their receptors are expressed as EC 50 values, referring to the concentration of each compound that induces a response halfway between its baseline and maximum. Data derived from Ryberg et al., 252 Lambert et al. 133 and

Sun et al. 235

GPR55 is highly expressed in the adrenal glands, and moderately expressed in parts of the CNS and gastrointestinal tract. While it is activated by the endocannabinoids and functions in the mechanism of inflammatory and neuropathic pain,253 significant differences between GPR55 and the cannabinoid receptors mean that the receptor is yet to be classified as either the CB3 receptor or as a member of the endocannabinoid system. This is because it shares only a 13.5% sequence homology with CB1 and

254 14.4% with CB2, its G-protein coupling is via G α13 instead of G i/o and considerably less expression than the other cannabinoid receptors, with expression being between three- and ten-times lower than that of CB1. 252

1.4.7. FAAH as a potential source of arachidonic acid for prostaglandin synthesis

Phospholipase A 2 produces the majority of arachidonic acid used in prostaglandin synthesis. Further potential sources of arachidonic acid for prostaglandin synthesis are the endocannabinoid metabolising enzymes FAAH and MAGL. FAAH inhibition has been

34 Chapter 1 demonstrated to reduce prostaglandin levels in microglia. 255 MAGL catalyses 85% of 2-

AG degradation.166 2-AG is the most concentrated of the endocannabinoids, being

1000-times higher than AEA in brain and 40-times higher in plasma. 141, 142

Consequently an in vivo study demonstrated MAGL inhibition to reduce unesterified arachidonic acid levels in brain by 50%. 218 Therefore inhibition of prostaglandin synthesis by FAAH and MAGL may explain part of their analgesic and anti- inflammatory action.

1.4.8. Further analgesic effects of FAAH inhibition

In addition to those described already there are multiple other analgesic and anti- inflammatory mechanisms attributed to FAAH substrates. Part of the activity of the

FAAH substrates has been attributed to an ‘entourage’ effect. As the endogenous levels of OEA and PEA are significantly higher than AEA, these two substrates compete for FAAH hydrolysis. Their higher levels mean they undergo more degradation, potentiating the action of AEA. 256 This action can also enhance the activity of AEA at the TRPV1 receptor. 257

FAAH’s synthesis of the endocannabinoid, NADA, is potentially a further way by which it may be analgesic. 163 NADA induces hyperalgesia via TRPV1, 139 thus inhibition of

FAAH should prevent NADA synthesis, blocking this action.

Pro-inflammatory stimuli, such as lipopolysaccharide, suppress expression of the AEA,

OEA and PEA synthesising enzyme NAPE-PLD on immune cells, consequently reducing their synthesis. 258 Therefore in inflamed tissue the FAAH substrates may not exert their analgesic and anti-inflammatory effect due to reduced synthesis, thus FAAH inhibition should help preserve substrate levels where synthesis is already low.

Additionally, FAAH inhibition in microglia reduces expression of inducible nitric oxide synthase (iNOS) (EC 1.14.13.39), the enzyme responsible for the synthesis of the pro- inflammatory free radical nitric oxide from L-arginine.255 Furthermore FAAH inhibition, through its indirect activation of the cannabinoid receptors, attenuates the release of the pro-inflammatory cytokines interleukin-1β and TNF-α. 259

35 Chapter 1

In conclusion the endocannabinoid system FAAH is the most attractive target for the development of novel analgesics and anti-inflammatories. In contrast to the therapeutics that inhibit nociceptive and pro-inflammatory pathways, FAAH can be utilised to bring about resolution from inflammation and pain through its control over multiple endogenous analgesic and anti-inflammatory pathways (Figure 1.14.). Instead of targeting just the cannabinoid receptors, FAAH inhibition indirectly targets many other analgesic and anti-inflammatory pathways outside of the endocannabinoid system, including the TRPV1 and GPR55 receptors via AEA, and the PPAR-α and GPR55 receptors via OEA and PEA.

36

Figure 1.14. – The multiple analgesic and anti-inflammatory pathways controlled by FAAH. AEA, OEA and PEA are agonists of the CB1, CB2, PPAR-α, TRPV1 and

GPR55 receptors. On completion of their action they are hydrolysed intracellularly by FAAH into arachidonic acid, oleic acid and palmitic acid.

Chapter 1

1.5. Development of FAAH inhibitors

1.5.1. FAAH inhibitors

Indirectly FAAH is the target of the routinely used analgesics, paracetamol and ibuprofen. Paracetamol is conjugated with arachidonic acid by FAAH to form the active metabolite, AM404, which functions as an analgesic through its action both as an AEA reuptake inhibitor 260 and as a COX-1 & 2 inhibitor. 164 As for ibuprofen, routinely used oral doses are capable of inhibiting FAAH in certain types of inflamed tissue, possibly contributing to its anti-inflammatory action. 261, 262

The active site of FAAH is comprised of an amino acid triad: serine-241/serine-

217/lysine-142. FAAH substrates and the majority of FAAH inhibitors forming an attachment to the active site’s nucleophilic serine-241 residue. 263 The nature of this attachment determines whether a FAAH inhibitor is covalent or non-covalent and reversible or irreversible. Non-covalent inhibitors that form a reversible attachment have been synthesised, 264 but due to their inhibition being reversible they have to compete with the FAAH substrates for the enzyme’s active site, 265 therefore their elevation of the FAAH substrate levels is transient. 266 As near complete inhibition of

FAAH is required for significant substrate accumulation covalent inhibitors are preferred. 267 The therapeutic effect of irreversible covalent inhibitors is long lasting due to FAAH activity only returning to normal once a sufficient number of new enzymes has been synthesised. 268

FAAH is a serine hydrolase, of which there are over 200 other members of this enzyme family with varying functions. Novel inhibitors therefore require selectivity for FAAH over the other serine hydrolases.266 Currently it is carbamate, urea and ketone based inhibitors that have been shown to be the most potent and selective. 269 To form the covalent attachments necessary for irreversible inhibition the inhibitor requires an electrophilic moiety so that it can react with the active site’s nucleophilic amino acid residue. The presence of an electrophilic group means a greater likelihood of adverse effects if the covalent attachment is not selective. To avoid the risks associated with

38 Chapter 1 covalent inhibitors, compounds with high potency are preferred so that low doses can be used. 270

Figure 1.15. – Structure of URB597.

Figure 1.16. – Structure of PF-3845.

Figure 1.17. – Structure of PF-04457845.

Most of the current FAAH inhibitors, such as URB597 (Figure 1.15.), form irreversible covalent attachments to the serine-241 residue of the active site.271 URB597 is the

222 most studied FAAH inhibitor (IC 50 = 4.6 pmol/mL). It is a cannabinoid receptor- dependent analgesic devoid of the adverse effects associated with CB1 agonists. 223

This carbamate based molecule however has selectivity for other serine hydrolases,

272 specifically peripherally expressed carboxylesterases. A more potent inhibitor (IC 50

= 0.23 pmol/mL), the urea based PF-3845 (Figure 1.16.), was later developed with

10-fold higher affinity for FAAH’s binding site than URB597, a longer duration of action and with more selectivity for FAAH over other serine hydrolases, including FAAH-2. 225

Using PF-3845 as a prototype, PF-04457845 (IC 50 = 7.2 pmol/mL) (Figure 1.17.), was developed and became the first FAAH inhibitor to enter clinical trials and be dosed to

39 Chapter 1 humans. 268, 273 The first clinical trial of a FAAH inhibitor identified this showing that 24 hours following a single 0.3 mg dose of PF-04457845, despite the drug no longer being detectable in the plasma, FAAH activity was still inhibited by ≥97%. 184 Therefore the pharmacodynamic effects of FAAH inhibitors outlast their apparent pharmacokinetics.

In summary, while FAAH can be inhibited by both covalent and non-covalent inhibitors, it is the irreversible attachment formed via covalent inhibitors that is preferred due to longevity of its action.

1.5.2. The FAAH substrates as a pharmacodynamic parameter

COX inhibitors, opioid agonists and anaesthetics all relieve pain, but by different mechanisms. The physiological response to an analgesic will only demonstrate pain relief not the drug’s mechanism of action. In the development of novel FAAH inhibitors it is important to not just demonstrate that a potential drug is analgesic but to also show that it inhibits the target enzyme in vivo . This is of importance as FAAH is a member of the serine hydrolase enzyme family, which number over 200. Hence some

FAAH inhibitors exhibit non-specific selectivity for other serine hydrolases. 266 The potency of a FAAH inhibitor is dependent on its ability to increase the central and peripheral tissue concentrations of AEA, OEA and PEA and consequently elicit its cannabinoid 223 and PPAR-α 236 dependent analgesic and anti-inflammatory effects.

Pharmacokinetics and pharmacodynamics are generally defined as what the body does to the drug and what the drug does to the body respectively. 274 Thus the changes in the tissue levels of the FAAH substrates is a physiological response to drug administration, and therefore a pharmacodynamic parameter. The measurement of the

FAAH substrates is a useful pharmacodynamic measurement for demonstrating FAAH inhibition and establishing mechanism of action.

Previous in vivo studies have measured brain and plasma AEA, OEA and PEA as a pharmacodynamic parameter, to demonstrate selectivity for FAAH and to determine the degree to which the substrates were increased. 225, 227, 273, 275 These studies investigated various classes of FAAH inhibitors, and by measuring the FAAH substrates they were able to prove that any observed in vivo efficacy seen with each inhibitor was in fact due to inhibition of FAAH. Importantly these studies all demonstrated that while

40 Chapter 1

FAAH is mostly brain expressed, FAAH inhibition will not only increase AEA, OEA and

PEA levels in the brain, but also in plasma. This was an important discovery as it allows for this approach to be translated from in vivo studies into clinical studies, where obviously measuring brain FAAH substrate levels would be impossible. Indeed the first clinical trial of the FAAH inhibitor PF-04457845 measured the FAAH substrates in plasma as a pharmacodynamic parameter to prove that the drug was in fact inhibiting FAAH in humans. 184 In this trial blood was collected at various time points following oral dosage of the inhibitor. This showed that the inhibitor elevated the plasma levels of AEA, OEA and PEA and they remained elevated for the following 7 days, proving the drug to be a potent FAAH inhibitor.

1.5.3. FAAH substrate assays

For the FAAH substrates to be used as a pharmacodynamic parameter, sensitive and selective assays for the quantitation of AEA, OEA and PEA are required. The FAAH substrates have weak UV absorption and do not fluoresce.276 Their measurement is therefore best performed by liquid chromatography-tandem mass spectrometry (LC-

MS/MS), of which a number of methods have been reported. 141, 142, 276-281 The majority of these methods have been in human plasma or serum, with a few exceptions looking at tissue samples from in vivo studies. 141 The FAAH substrates are often measured alongside the endocannabinoids, 141 and sometimes other related lipid signalling molecules. 278 The extraction of the FAAH substrates from samples is often by liquid- liquid extraction, using ethyl acetate/hexane (9:1 v/v) 141 or chloroform/methanol (2:1 v/v). 280 This is often followed by solid phase extraction, evaporation and reconstitution to further concentrate the substrates.

Separation of the substrates has mostly been on short C18 columns (50 x 2.1 mm, 1.7

µm), 142 although there are examples of longer C8 columns (100 x 2.1 mm, 3 µm) 141 and phenyl columns (50 x 2.1 mm, 5 µm) being used.279 Chromatographic resolution of the analytes is not always achieved and does not compromise the analysis. 278, 279 As positive electrospray is required for the FAAH substrates, the previously reported mobile phases have almost all been acetonitrile gradients with 0.1% formic acid to aid ionisation. Occasionally ammonium acetate is also added to the mobile phase to act as

41 Chapter 1 a buffer. 277 Run times for these assays range from 3.5 minutes 277 up to over 20 minutes. 278 In terms of MS/MS, all previously reported methods have ionised the FAAH substrates under positive electrospray and monitored the formation of the 62 m/z ethanolamine fragment ion.

1.6. Paracetamol

Paracetamol (para-acetylaminophenol, acetaminophen) was first synthesised in 1878 by Morse and first used in patients in 1887 by von Mering. 282 It has since become one of the most widely used analgesics.3 A discovery by Högestätt demonstrated that it is in part metabolised by FAAH to form the active metabolite AM404.164

Figure 1.18. – The multiple pathways of paracetamol metabolism. Paracetamol can be conjugated to glucuronide, sulphate, glutathione or arachidonic acid groups.

The majority of paracetamol metabolism occurs in the liver. Between 40–67% of the absorbed drug undergoes glucuronidation, 20–46% is sulphated and 5–15% is converted to the toxic metabolite N-acetyl-p-benzoquinone imine (NAPQI) before being rearranged and conjugated with glutathione (GSH). 283, 284 Less than 0.002% of

42 Chapter 1 paracetamol is deacetylated in the liver to 4-aminophenol, and then conjugated to arachidonic acid via brain expressed FAAH to form AM404 (arachidonoyl phenolamine)

(Figure 1.18). 164, 285

1.6.1. Paracetamol inhibits COX activity

Paracetamol is a weak COX-1 (IC 50 = 105 nmol/mL) and COX-2 inhibitor (IC 50 = 26 nmol/mL) 76, 286 that blocks prostaglandin synthesis in the brain 287 but not in peripheral tissue. 288 It does however inhibit PGE2 production in activated macrophages. 289

Despite inhibiting COX-1 & 2, paracetamol is ineffective as an anti-inflammatory. This is thought to be due to two reasons. Firstly COX-2 is induced in inflammation and as paracetamol is only a weak COX-2 inhibitor, prostaglandin synthesis is not significantly reduced in inflamed tissue. Secondly for COX to synthesise prostaglandins its active site’s tyrosine-385 residue must be oxidised. Paracetamol inhibits COX-1 & COX-2 by acting as a reducing agent, preventing the oxidation of this residue and consequently blocking prostaglandin synthesis. In inflamed tissue there are high levels of lipid hydroperoxides (LHP) that maintain oxidation of the tyrosine residue and keep COX in its active state.290 Consequently paracetamol is unable to act as a COX inhibitor in inflamed tissue and is therefore rendered ineffective as an anti-inflammatory.

However this explanation of why paracetamol lacks an anti-inflammatory effect has been disputed, as in vitro studies with macrophages have shown that paracetamol can inhibit COX-2 in both the presence and absence of hydroperoxide. 289 This is supported by the clinical finding that 1 g of paracetamol administered orally suppresses PGE2 production in the acute inflammation of peripheral tissue. 291 Therefore while evidence demonstrates that paracetamol inhibits prostaglandin production, even in inflamed peripheral tissue, the reason for its reported lack of an anti-inflammatory effect is still not understood.

43 Chapter 1

1.6.2. Paracetamol interacts with the endocannabinoid system

Paracetamol lacks affinity for the cannabinoid receptors, 292 but its analgesic effect is blocked in vivo by selective CB1 antagonists 292, 293 and is abolished in CB1 knock-out mice. 292 Paracetamol (200 mg/kg p.o. ), while analgesic in control mice, fails to exert its analgesic effect in FAAH knock-out mice. 294 Likewise the FAAH inhibitor URB597

(0.15 mg/kg i.p. ), at a low non-analgesic dose, reverses the analgesic effect of paracetamol (300 mg/kg p.o. ).292 This dependence of paracetamol on the endocannabinoid system is due to its active metabolite AM404.

Figure 1.19. – Paracetamol’s effect on endocannabinoid signalling. Paracetamol is deacylated to 4-aminophenol and then conjugated to arachidonic acid by FAAH. This forms

AM404, an AEA reuptake inhibitor that causes a synaptic accumulation of AEA and consequently increased activation of the CB1 receptor.

Like paracetamol, AM404 is neither a cannabinoid receptor agonist 295 nor a FAAH inhibitor. 260 Instead it has an indirect action on the endocannabinoid AEA. The inactivation of AEA is a 2-step process, beginning with its reuptake from the synaptic cleft into the presynapse, followed by its hydrolysis by intracellular FAAH. Paracetamol inhibits both steps of AEA inactivation. Firstly paracetamol reveres FAAH’s normal enzyme activity, so that instead of arachidonic acid being hydrolysed from ethanolamine, arachidonic acid is rather conjugated to 4-aminophenol, forming

44 Chapter 1

164, 292 AM404. Then AM404 interferes with second step by inhibiting AEA reuptake (IC 50

= 5.3 nmol/mL), thus protecting it from deactivation.104 Therefore paracetamol causes an increase in extracellular AEA in the synaptic cleft that consequently increases cannabinoid receptor activation to produce analgesia (Figure 1.19.).260

An in vivo study investigated the effect of a 300 mg/kg i.p. dose of paracetamol on the concentrations of paracetamol, 4-aminophenol and AM404 in the brain, spinal cord, blood and liver of rats. 164 Paracetamol was present in the liver at a concentration comparable to blood (both 3 µmol/g). Paracetamol was deacylated in the liver, therefore it was here where the highest concentration of 4-aminophenol occured (30 nmol/g). 4-aminophenol then transfers to the brain. About a third of the 4- aminophenol found in the liver was found in the brain (8 nmol/g). FAAH is mostly expressed in the brain 102 where there are also high concentrations of arachidonic acid

(8 µmol/g).180 It is therefore here that the majority of AM404 was synthesised (15 pmol/g), with low amounts consequently entering the spinal cord (3 pmol/g). Very little AM404 was found in blood and liver (< 0.5 pmol/g).

The concentration of AM404 required to induce analgesia is unknown. As already stated, while paracetamol (300 mg/kg i.p.) produces cannabinoid receptor-dependent analgesia in rats, only low levels of AM404 are produced in the brain (15 pmol/g). 164

Interestingly it has been separately reported that 300 mg/kg paracetamol fails to increase brain AEA levels. 294 This suggests that AM404 is active at low concentrations.

In rat models of inflammatory and neuropathic pain AM404 is still analgesic at doses as low as 1 mg/kg i.p .296 Therefore paracetamol produces cannabinoid receptor- dependent analgesia despite only low amounts of AM404 being produced in the CNS and no change in AEA levels being detected.

Analgesia from AEA reuptake inhibition is reversed in vivo by CB1 antagonists, but not by CB2 antagonists or TRPV1 antagonists. 104 AEA reuptake is therefore specific for

AEA-activation of CB1 receptors. CB1 is expressed centrally in the CNS, CB2 is expressed on immune cells and TRPV1 is expressed peripherally on nociceptors.

Therefore the analgesia produced by AM404 is a central synaptic effect, not a peripheral synaptic effect or an immune effect.

45 Chapter 1

1.6.3. Paracetamol interacts with serotonin receptors

Serotonin (5-hydroxytryptamine, 5HT) (Figure 1.20.) is a neurotransmitter synthesised from L-tryptophan by tryptophan-5-hydroxylase (EC 1.14.16.4). It acts on its seven receptors, 5HT 1-7. Upon completion of its action 5HT is oxidised by monoamine oxidase (EC 1.4.3.4) followed by aldehyde dehydrogenase (EC 1.2.1.3) to form its metabolite 5-hydroxyindoleacetic acid (5HIAA). 297 Activation of the 5HT receptors can be analgesic, for example direct activation of the 5HT 1A receptor subtype is effective in the treatment of nociceptive pain. 298

Figure 1.20. – Structure of 5HT. 5HT is a neurotransmitter integral to the function of paracetamol. Production of 5HT is increased by paracetamol and antagonism of 5HT receptors reverses paracetamol’s action.

In vivo paracetamol has been shown to increase the levels of 5HT throughout the CNS with the exception of the spinal cord. 299 The analgesic effect of paracetamol is 5HT receptor-dependent, with antagonism of the 5HT 1A , 5HT 3 and 5HT 4 receptors all blocking its action. 300-302 Interestingly paracetamol lacks affinity for all of the 5HT

303 receptors, including 5HT 1A , 5HT 3 and 5HT 4. Therefore paracetamol indirectly activates 5HT receptors by increasing 5HT levels in the CNS by an unknown mechanism.

This is clinically significant as paracetamol’s action is also reversed in humans with

5HT 3 antagonism, demonstrating that the serotonin system is integral to paracetamol’s action. 304 Interestingly the analgesic effect of CB1 agonists is also reversed by

292 antagonism of the 5HT 1A , 5HT 3 and 5HT 4 receptors. Given that both paracetamol and cannabinoid induced analgesia is 5HT receptor-dependent, AM404 could be a possible mediator between paracetamol and the 5HT receptor.

46 Chapter 1

1.6.4. AM404 interacts with COX, iNOS, TRPV1 and sodium channels

Like paracetamol, AM404 can inhibit prostaglandin synthesis by inhibiting COX-1 (IC 50

164 = 6.2 nmol/mL) and COX-2 (IC 50 = 5.7 nmol/mL). As AM404 synthesis requires arachidonic acid, this possibly reduces the amount of arachidonic acid available for prostaglandin synthesis. 164 Furthermore AM404 can decrease the expression of COX-2 and iNOS, the synthesising enzyme of nitric oxide. 296 AM404 has also been identified as a TRPV1 agonist, 305 which is interesting as activation of TRPV1 triggers AEA synthesis. 144 Evidence supports this showing that the analgesic effect of both paracetamol and AM404 is abolished in TRPV1 knock-out mice, concluding that AM404 produces TRPV1 mediated analgesia in the brain. 294 Additionally AM404 inhibits pain-

+ 306 associated voltage-gated Na channels (Na V).

1.6.5. Pharmacokinetics of intravenous paracetamol in the cerebrospinal fluid

Paracetamol is known to readily cross the blood-brain barrier and consequently penetrates rapidly into the cerebrospinal fluid (CSF). CSF is a clear and colourless fluid that fills the space around the brain and inside the spinal cord, with its function being to keep the brain buoyant and to provide mechanical protection. It flows through the brain’s ventricular system regulating its homeostasis and providing chemical stability. 307 CSF is rich in ions, with millimolar concentrations of Na +, K +, Ca 2+ , Mg 2+ and Cl -, and a total volume inside the body of approximately 140 mL. 308 It contains low levels of protein, low numbers of leukocytes and no erythrocytes. 308 Approximately

60% of CSF is produced and secreted by the choroid plexus, a ventricular structure in the brain. The remaining 40% is formed from capillary filtration. 307

To enter the CSF a drug must first cross the blood-brain barrier. This barrier consists of a highly restrictive endothelial cell network that seals the brain’s capillaries. 308

Molecules that can cross the blood-brain barrier use either transmembrane diffusion, specialised transporters, adsorptive endocytosis or extracellular pathways. 309

Transporter proteins are highly expressed on the blood-brain barrier, moving foreign molecules away from the brain, while allowing molecules such as O 2 and glucose to cross. 309 Chemically it is large, hydrophilic molecules that are restricted in crossing the

47 Chapter 1 blood-brain barrier, explaining why small and hydrophobic molecules of paracetamol molecules are capable of penetrating it. The choroid plexus, another component of blood-brain barrier, 309 consists of permeable endothelial cells. Therefore drugs that have managed to cross the blood-brain barrier can readily enter into the CSF. 308

Intravenous paracetamol is increasingly being used in post-surgery pain management as an alternative to opioids, such as morphine, to avoid the associated adverse effects of addiction, tolerance and withdrawal. Recent clinical trials in the treatment of post- operative pain have demonstrated 1 g intravenous paracetamol to provide comparable levels of analgesia to 10 mg intravenous morphine. 310-312

The intravenous route of administration for paracetamol is faster acting and more efficacious than the oral route, with the peak plasma concentration being reached immediately after dosage and with the entire administered dose reaching the systemic circulation, hence 100% bioavailability. 313 Following a 1 g intravenous dose of

314 paracetamol the plasma C max is 165 nmol/mL, whereas the same dose orally has a

315 lower C max of 90 nmol/mL. This difference between the two routes is further reflected in the fact that following an intravenous dose the onset of analgesia is reported to take approximately 8 minutes, compared to 40 minutes for an oral preparation, 316 with the maximum analgesic effect not being seen for a further 2 - 3 hours. 317

While the plasma pharmacokinetics of oral paracetamol has been well studied, the CSF pharmacokinetics of intravenous paracetamol remains to be determined in adult humans. Given that paracetamol is increasingly being dosed intravenously and that paracetamol is known to be active in the CNS, this is worth investigating. Previous studies in the CSF following intravenous paracetamol have been either in children, 318-

320 the elderly 321 or using the pro-drug of paracetamol, propacetamol. 322 Of these previous clinical trials, only the trial in children was a pharmacokinetics study.320 They showed that within 5 minutes of an intravenous injection that paracetamol is detectable in the CSF, peaking after 60 minutes and still being present after 5 hours.320 This rapid penetration into the CSF, fast accumulation and slow clearance is possibly responsible for paracetamol’s fast onset and long duration of analgesia.

48 Chapter 1

The plasma pharmacokinetics of intravenous paracetamol has only recently been

314 comprehensively studied. As mentioned earlier following a 1 g dose the C max was

165 nmol/mL, which then cleared from the plasma with a t 1/2 of 156 minutes.

Clearance was determined to be 4.2 mL/min/kg, area under the curve was 5,298 nmol.min/mL and volume of distribution was 0.9 L/kg.

1.6.6. Paracetamol toxicity

Paracetamol has a good safety profile to the extent that it is even suitable for use in neonates 323 and pregnant women. 324 The minimum fatal oral dose of paracetamol in adult humans is approximately 11 g. 325 After an overdose, post-mortem paracetamol levels in the blood are reported to be 1 – 2.5 mmol/mL, and in the liver to be 2

µmol/g. 326 Following an overdose the liver's glutathione levels are depleted allowing

NAPQI to cause damage to hepatocytes that leads to liver damage and potentially liver failure. 327 N-acetylcysteine is used as the antidote for paracetamol overdose as together with glutathione binds to NAPQI and consequently aids in its detoxification. 327

As a precaution paracetamol is often formulated with the amino acid methionine, which via cysteine is reduced to glutathione in the liver. 328 This addition of methionine does not lessen paracetamol’s analgesic effect or alter its pharmacokinetics. 328

49 Chapter 1

1.7. Aims

The aim of this research is to investigate the analgesic and anti-inflammatory pathways that FAAH is involved in. The following areas will be studied:

• The endocannabinoid inactivating enzymes FAAH and MAGL both produce

arachidonic acid, the precursor to the prostaglandins. The interaction between the

endocannabinoid and prostaglandin systems will be investigated, focusing on FAAH

and MAGL as potential sources of arachidonic acid for PGE2 synthesis in activated

macrophages.

• FAAH inhibitors are currently in development as analgesics. Their analgesic and

anti-inflammatory effect is dependent on AEA, OEA and PEA. The measurement of

these lipids as a pharmacodynamic parameter would provide a useful means for

screening novel FAAH inhibitors to establish their effectiveness at increasing the

concentration of the FAAH substrates. An assay for the FAAH substrates in brain

and blood by liquid chromatography-tandem mass spectrometry (LC-MS/MS) will

be developed and validated. Novel FAAH inhibitors will be investigated in vivo for

their ability to increase brain and blood AEA, OEA and PEA levels. Furthermore the

rates of oxidative metabolism and the degree to which each substrate undergoes

protein binding will be compared.

• The inactivation of AEA is a 2-step process beginning with its reuptake and

finishing with its hydrolysis via FAAH. Paracetamol interacts with both steps by

using FAAH to synthesis an AEA reuptake inhibitor, AM404. This active metabolite

is a conjugate of arachidonic acid and paracetamol. The conjugation of other fatty

acids to paracetamol via FAAH will be studied here. Furthermore the synthesis of

AM404 remains to be demonstrated in humans. The metabolism of paracetamol

into AM404 via FAAH will be investigated in a clinical trial looking at the

cerebrospinal fluid (CSF) and plasma of patients receiving intravenous paracetamol

for treatment of post-operative pain. Furthermore the pharmacokinetics of

paracetamol in the CSF, alongside the effect of paracetamol on the FAAH

substrates, prostaglandins, leukotrienes and serotonin will also be investigated.

50 Chapter 2

CHAPTER 2

Inhibition of FAAH and MAGL suppresses prostaglandin E2

synthesis by LPS-stimulated macrophages

51 Chapter 2

2.1. Introduction

FAAH and MAGL both produce arachidonic acid, the precursor of the prostaglandins, through their hydrolysis of the endocannabinoids AEA and 2-AG. In inflammation activated macrophages are a major source of PGE2, which is responsible for multiple pro-inflammatory and hyperalgesic effects. It is hypothesised that inhibition of FAAH and MAGL in activated macrophages should reduce the levels of arachidonic acid and consequently reduce production of PGE2.

RAW264.7, a mouse leukaemic macrophage cell line, is a frequently used macrophage cell line, 329 that will be used here to study LPS-stimulated PGE2 production. Previously

LPS-stimulated (10 ng/mL) RAW264.7 macrophages have been shown to have 10-fold higher AEA levels and increased expression of FAAH. 330 This suggests that with LPS- stimulation there is an increased metabolism of AEA and possibly an increase in FAAH formed arachidonic acid, a potential therapeutic target for inflammatory pain. For

MAGL, LPS-stimulation of macrophages has been shown to increase 2-AG levels by

2.4-fold and to down-regulate MAGL activity. 331

As both FAAH and MAGL form arachidonic acid in endocannabinoid hydrolysis, it is hypothesised that inhibition should indirectly inhibit PGE2 synthesis by LPS-stimulated

RAW264.7 macrophages. The effect of inhibiting endocannabinoid hydrolysis on LPS- stimulated PGE2 production by macrophages will be investigated using the FAAH inhibitor PF-3845 (IC 50 = 0.23 pmol/mL) and the MAGL inhibitor JZL184 (IC 50 = 8 pmol/mL). The effect of FAAH and MAGL inhibition of activated macrophages will be determined through the measurement of PGE2, nitric oxide, cell viability and COX-2 expression.

52 Chapter 2

2.2. Materials and methods

2.2.1. Materials

Dulbecco's Modified Eagle's Medium (DMEM), HEPES, phosphate buffered saline (PBS),

Lipopolysaccharide (from Escherichia coli serotype 0111:B4), JZL184, Sulfanilamide,

N-1-naphthylenediamine dihydrochloride, 3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyl- tetrazolium bromide (MTT) and protease inhibitor cocktail (containing aminoethyl benzenesulfonyl fluoride, pepstatin A, E-64, bestatin, leupeptin, and aprotinin) were from Sigma Aldrich (Poole, UK). Heat inactivated fetal bovine serum (FBS), penicillin and streptomycin mixture, lithium dodecyl sulfate (LDS) loading buffer and 4-12% bis-

TRIS gel (1.0 mm x 12 well) were from Invitrogen (Paisley, UK). Orthophosphoric acid was from BDH (Poole, UK). PF-3845 was from Cayman (Ann Arbor, Michigan, USA).

The PGE2 assay was from Cisbio (Bagnols-sur-Ceze, France). The bicinchoninic acid

(BCA) protein assay and TRIS-buffered saline (TBS) blocking buffer were from Thermo

Scientific (Runcorn, UK). Mammalian protein extraction buffer, enhanced chemiluminescence (ECL) detection reagent and Amersham Hyperfilm ECL were from

GE Healthcare (Little Chalfont, UK). Immobilon-P transfer membrane was from

Millipore (Watford, UK). COX-2 polyclonal antibody was from Cambridge Bioscience

(Cambridge, UK).

2.2.2. Macrophages

RAW264.7 macrophages (American Type Culture Collection, Manassas, Virginia, USA) were cultured in DMEM containing 100 U/mL penicillin, 100 µg/mL streptomycin, 24

o mM HEPES (pH 7) and 10% FBS, under standard culture conditions (37 C, 8% CO 2 and

92% air). Cells were seeded into 24-well plates at a density of 2 x 10 6 cells/well.

Experiments were performed on confluent cells. Before treatment, non-adherent cells were washed away with PBS. Macrophages were activated with LPS (either 10 ng/mL or 1 µg/mL). Treatments did not contain FBS. Compounds were dissolved in DMSO, with the final DMSO concentration in cell medium not exceeding 0.9%, which was shown not to affect cell viability. At the end of experiments medium was collected and frozen at -80 oC until analysis.

53 Chapter 2

2.2.3. Prostaglandin E2 assay

Prostaglandin E2 was quantified using a kit from Cisbio. The PGE2 present in the samples competes with d2-labelled PGE2 for binding to an anti-PGE2 cryptate-labelled antibody. The assay makes use of Forster’s resonance energy transfer (FRET), a mechanism by which energy can be transferred between two fluorophores: a donor and an acceptor. Excitation of the donor, which in this assay is cryptate, transfers energy to the acceptor, the d2-labelled PGE2, causing an emission of light. When PGE2 from the sample displaces the d2-labelled PGE2, the proximity between the two is increased, which in turn decreases the amount of light emitted. Therefore a high concentration of PGE2 will causes a low amount of fluorescence. On completion of each experiment media was collected and frozen at -80 oC. The assay was used in accordance with the kit’s instructions. 10 µL medium was added to 5 µL d2-labelled

PGE2 and 5 µL anti-PGE2 cryptate-labelled antibody into a 384 well plate, which was then left at room temperature over night. An 8-point PGE2 calibration line between 0–

5,000 pg/mL was prepared in the assay diluent and run in triplicate. The plate was read using a Berthold Technologies (Harpenden, UK) Mithras LB 940 FRET plate reader with FRET being measured and expressed as a ratio of 665 nm and 620 nm.

2.2.4. Nitric oxide assay

Nitric oxide levels were assayed indirectly by measuring its stable and non-volatile

- - breakdown product nitrite (NO 2 ). NO 2 reacts with sulfanilamide to form a diazonium salt, which then couples with N-1-naphthylenediamine dihydrochloride to form a chromophore that absorbs at 550 nm. 100 µL Griess reagent (1% w/v sulfanilamide,

0.1% w/v N-1-naphthylenediamine dihydrochloride and 5% v/v phosphoric acid in double distilled H 2O) was added to 100 µL of either sample or standard in a 96-well plate. A ten-point sodium nitrite calibration line between 0 – 100 nmol/mL was used.

This was prepared in DMEM to correct for any interference from the colour of DMEM.

The plate was then read at an absorbance of 550 nm using an Anthos Labtec

Instruments UV / Vis absorbance plate reader.

54 Chapter 2

2.2.5. Cell viability assay

The mitochondrial hydrogenases of living cells cleave dimethyl thiazolyl diphenyl tetrazolium (MTT) to form an insoluble purple formazan, an indicator of cell viability. 332 . Cells were aspirated and 0.4 mg/mL MTT dissolved in DMEM was added to each well (250 µL/well in a 24 well plate). Plates were incubated for 1 hour at 37 oC.

Wells were aspirated and DMSO was added (250 µL/well). 100 µL from each well was transferred into a 96-well plate with UV absorbance being measured at 550 nm.

2.2.6. Western blotting COX-2

Protein was extracted from cells using mammalian protein extraction buffer with 10

µL/mL protease inhibitor cocktail at 75 µL/well. Protein concentrations were determined using the BCA protein assay and then samples were equalised to 2 µg/µL protein. 10 µL LDS containing loading buffer was added to 30 µL of sample. Protein was denaturised at 100 oC for 5 minutes and then cooled to 4 oC. 20 µL sample was separated by gel electrophoresis (XCell Surelock, Invitrogen) on a 4-12% bis-TRIS gel

(1.0 mm x 12 well, 10 cm in length), at 200 V for 1 hour. Protein was transferred electrophoretically at 30 V for 1.5 hours onto an immobilon-P transfer membrane.

Membranes were blocked for 1 hour at 4 oC with TBS blocking buffer before being probed by COX-2 polyclonal antibody (1 µg/mL). Membranes were washed with TTBS for 15 minutes before being incubated with anti-rabbit horseradish peroxidise conjugated secondary antibody for 1 hour. Blots were visualised using an ECL detection reagent, exposed with film to light and developed.

2.2.7. Statistical analysis

Data was analysed by either ANOVA with Tukey’s post-hoc analysis or t-test. Tukey’s post-hoc analysis was used as it compares all possible pairs of groups, in a multiple group analysis. p < 0.05 was deemed statistically significant. Data are presented as mean ± SEM. Statistical analysis was by Prism 3.0 (GraphPad).

55 Chapter 2

2.3. Results

2.3.1. Effect of FAAH and MAGL inhibition on the cell viability of LPS-stimulated

RAW264.7 macrophages

At 10 ng/mL LPS had no effect on cell viability (Figure 2.1. A). At 1 µg/mL LPS decreased cell viability by 30% (Figures 2.1. B). JZL184 at concentrations between

0.01 – 100 nmol/mL did not decrease cell viability, with 100 nmol/mL actually increasing viability.

*** 125 A

100

75 Cell viability 50 (% of unstimulated control) 25

0 Control LPS DMSO 0.01 0.1 1 10 100 JZL184 (nmol/mL)

125 B 100

75 *** *** *** *** *** ***

Cell viability 50

(% of unstimulated control) 25

0 Control LPS DMSO 0.01 0.1 1 10 100

JZL184 (nmol/mL)

Figure 2.1. – Effect of the MAGL inhibitor JZL184 (0.01 – 100 nmol/mL) on the cell viability of RAW264.7 macrophages stimulated with either (A) 10 ng/mL or (B) 1

µg/mL of LPS. Macrophages were cultured as described in Section 2.2.2. Macrophages were incubated for 24 hours after treatment. DMSO treatment was at 0.9% and did not contain LPS.

Cell viability determined using the MTT assay. Data expressed as mean ± SEM (n = 6 per group).

Statistical analysis by ANOVA with Tukey’s post-hoc analysis. *** = p ≤ 0.001 in comparison with unstimulated controls.

56 Chapter 2

PF-3845 at concentrations above 10 nmol/mL was cytotoxic (Figure 2.2.), hence this inhibitor was only studied at concentrations between 0.01 – 1 nmol/mL. In dimethyl sulfoxide (DMSO) PF-3845 is soluble at concentrations up to 44 µmol/mL, while

JZL184 is soluble at concentrations up to 48 µmol/mL. The highest DMSO concentration used was 0.9% for 100 nmol/mL JZL184, which did not affect cell viability.

125 A

100

75 Cell viability 50 (% of unstimulated control) 25

*** *** 0 Control LPS DMSO 0.01 0.1 1 10 100

PF-3845 (nmol/mL)

125 B 100

75 *** *** *** ***

Cell viability 50 ***

(% of unstimulated control) 25 *** 0 Control LPS DMSO 0.01 0.1 1 10 100

PF-3845 (nmol/mL)

Figure 2.2. – Effect of the FAAH inhibitor PF-3845 (0.01 – 100 nmol/mL) on the cell viability of RAW264.7 macrophages stimulated with either (A) 10 ng/mL or (B) 1

µg/mL of LPS. Macrophages were cultured as described in Section 2.2.2. Macrophages were incubated for 24 hours after treatment. DMSO treatment was at 0.9% and did not contain LPS.

Cell viability determined using the MTT assay. Data expressed as mean ± SEM (n = 6 per group).

Statistical analysis by ANOVA with Tukey’s post-hoc analysis. *** = p ≤ 0.001 in comparison with unstimulated controls.

57 Chapter 2

2.3.2. Effect of FAAH and MAGL inhibition on LPS-induced PGE2 synthesis in RAW264.7 macrophages

Macrophages were stimulated with 10 ng/mL LPS and treated with increasing concentrations of either the FAAH inhibitor PF-3845 or the MAGL inhibitor JZL184

(Figure 2.3. A).

JZL184 PF-3845 1500 A

*** 1000 ***

*** *** 500 ***

*** LPS-stimulated PGE2 (fmol/mL cell medium) 0 0 0.01 0.1 1 10 100 Concentration (nmol/mL)

JZL184 B PF-3845

6000

4000 ***

2000 LPS-stimulated PGE2 (fmol/mL cell medium) 0 0 0.01 0.1 1 10 100 Concentration (nmol/mL)

Figure 2.3. – Effect of the FAAH inhibitor PF-3845 (0.01 – 1 nmol/mL) and the MAGL inhibitor JZL184 (0.01 – 100 nmol/mL) on the PGE2 production of RAW264.7 macrophages stimulated with either (A) 10 ng/mL or (B) 1 µg/mL of LPS. Macrophages were cultured as described in Section 2.2.2. Macrophages were incubated for 24 hours after treatment. PGE2 was quantified in cell media by ELISA. Data expressed as mean ± SEM (n = 6 per group). Statistical analysis by ANOVA with Tukey’s post-hoc analysis. *** = p ≤ 0.001 in comparison with LPS-stimulated non-treated controls.

Stimulating macrophages with 10 ng/mL LPS increased PGE2 levels by approximately

1500 fmol/mL. Both FAAH and MAGL inhibition concentration dependently reduced

58 Chapter 2

LPS-stimulated PGE2 synthesis. The EC 50 , the response halfway between the maximum and the baseline, for both inhibitors was approximately 1 nmol/mL. The effect of a higher concentration of LPS was investigated with 1 µg/mL being used (Figure 2.3. B).

At 1 µg/mL, LPS increased PGE2 levels by 6000 fmol/mL. With the exception of 100 nmol/mL JZL184, no inhibition of PGE2 production was seen. (Figure 2.3.).

2.3.3. Effect of FAAH and MAGL inhibition on LPS-induced nitric oxide synthesis in

RAW264.7 macrophages

JZL184 40 PF-3845 A

30

20 ***

10 LPS-stimulated (nmol/mLNitrite cell medium) 0 0 0.01 0.1 1 10 100 Concentration (nmol/mL)

JZL184 60 PF-3845 B 50

40 ***

30

20

10 LPS-stimulated Nitrite (nmol/mL cell medium) 0 0 0.01 0.1 1 10 100 Concentration (nmol/mL)

Figure 2.4. – Effect of the FAAH inhibitor PF-3845 (0.01 – 1 nmol/mL) and the MAGL

- inhibitor JZL184 (0.01 – 100 nmol/mL) on the NO 2 production of RAW264.7 macrophages stimulated with either (A) 10 ng/mL or (B) 1 µg/mL of LPS. Macrophages were cultured as described in Section 2.2.2. Macrophages were incubated for 24 hours after

- treatment. NO 2 was quantified in cell media using the Griess reagent. Data expressed as mean ±

SEM (n = 6 per group). Statistical analysis by ANOVA with Tukey’s post-hoc analysis. *** = p ≤

0.001 in comparison with LPS-stimulated non-treated controls.

59 Chapter 2

- Stimulating macrophages with 10 ng/mL LPS increased NO 2 concentrations to 30 nmol/mL (Figure 2.4. A). Only at the highest concentration, 100 nmol/mL, did the

- MAGL inhibitor JZL184 significantly decrease NO 2 concentrations. No effect was seen

- with PF-3845. Stimulating macrophages with 1 µg/mL LPS increased NO 2 concentrations to 50 nmol/mL (Figure 2.4. B). Again only at the highest concentration,

- 100 nmol/mL, did JZL184 significantly decrease NO 2 concentrations. No effect was seen with PF-3845. It was determined whether either PF-3845 or JZL184 would interfere with Griess assay. No change in absorbance with the Griess reagent was seen with either compound, concluding that neither PF-3845 nor JZL184 interfered with the assay.

2.3.4. Effect of MAGL inhibition on LPS-induced COX-2 expression in RAW264.7 macrophages

It was investigated as to whether JZL184 could decrease COX-2 expression. 10 nmol/mL JZL184 was not found to alter COX-2 expression in macrophages stimulated with 1 µg/mL (Figure 2.5.). As expected COX-2 was not present in unstimulated macrophages. 35 LPS-stimulation produces a 70 kDa band, corresponding to the enzyme. 333

Figure 2.5. – Effect of JZL184 on LPS-induced COX-2 protein expression in RAW264.7 macrophages. Macrophages were cultured as described in Section 2.2.2. Macrophages were incubated for 24 hours after treatment. The figure shows the results from duplicate experiments.

60 Chapter 2

2.4. Discussion

The endocannabinoid and prostaglandin systems are known to interact. Inhibition of

COX-1 and COX-2 not only reduces PGE2 levels, but also increases levels of the endocannabinoids AEA and 2-AG. 182, 183 The original aim of this study was to further investigate the interaction between endocannabinoid and prostaglandin systems, focusing on FAAH and MAGL as sources of arachidonic acid for PGE2 synthesis. A previous study demonstrated FAAH inhibition suppressed PGE2 in microglia by using the inhibitor URB597. 255 Here the more potent and selective inhibitor PF-3845 was used for FAAH inhibition and JZL184 was used for MAGL inhibition.

It was investigated as to whether inhibition of FAAH and MAGL could sufficiently lower arachidonic acid levels to the extent that prostaglandin production would also be reduced. PGE2 production was investigated using RAW264.7 macrophages stimulated with LPS. Typically the screening of COX inhibitors for their ability to inhibit PGE2 production in LPS-stimulated macrophages uses 1 µg/mL LPS, with both NSAIDs and cannabinoids inhibiting PGE2 production in this model.116, 286 Here two LPS concentrations were used: 10 ng/mL and 1 µg/mL. There is a significant difference in expression of COX-2 and the amount of PGE2 produced between these two LPS concentrations. The effect of FAAH and MAGL inhibitors were investigated under both conditions to determine whether a diminished ability to reduce PGE2 would be seen in the presence of higher COX-2 expression.

It was found that the FAAH and MAGL inhibitors PF-3845 and JZL184 concentration- dependently reduced PGE2 production at 10 ng/mL LPS (Figure 2.3). At this concentration of LPS only 1,500 fmol/mL PGE2 was produced despite significant expression of COX-2. This suggests that not much arachidonic acid was being liberated by phospholipases from the cell membrane. Therefore FAAH and MAGL were able to influence arachidonic acid levels and consequently PGE2 production. However this effect was dependent on COX-2 expression and PGE2 production.

At 1 µg/mL LPS there was higher expression of COX-2 and a 6,000 fmol/mL release of

PGE2. This higher concentration of LPS meant that the FAAH and MAGL inhibitors had

61 Chapter 2 no effect on PGE2 production. Therefore the phospholipases became a more significant source of arachidonic acid. Hence the influence of FAAH and MAGL over arachidonic acid levels and PGE2 production was diminished. These findings suggest that under certain conditions FAAH and MAGL can control arachidonic acid levels to the extent that PGE2 production is influenced. The amount of PGE2 being produced appears to be integral to the ability of FAAH and MAGL inhibitors to suppressed PGE2 production.

In addition to PGE2, nitric oxide was also measured indirectly by assaying its stable breakdown product nitrite. Nitric oxide was investigated as both LPS and PGE2 stimulate iNOS expression and NO production.255 Furthermore the FAAH inhibitor

URB597 has been shown to reduce NO production in activated macrophages. 255 It was therefore hypothesised that FAAH and MAGL inhibition might reduce LPS-stimulated

NO production in macrophages, therefore the effect of PF-3845 and JZL184 was investigated. JZL184 was the only inhibitor to reduce NO production, an effect that was only seen at the highest concentration of 100 nmol/mL (Figure 2.4.). Therefore no correlation between FAAH and MAGL inhibition on NO levels were found.

Previously cannabinoids have been demonstrated to inhibit the production of PGE2, 116 interleukin-2,334 interleukin-6, 116 TNF-α, 335 and nitric oxide 336 in macrophages.

Activation of the cannabinoid receptors therefore clearly influences macrophage activity. Therefore the observed decreases in PGE2 presented here may possibly be mediated by the cannabinoid receptors. However a recent study that showed the FAAH inhibitor URB597 to inhibit PGE2 production in microglia,255 demonstrated that CB1 and CB2 antagonists did not block this effect. Therefore it was concluded that this activity was independent of the cannabinoid receptors.

While FAAH and MAGL inhibition increases AEA and 2-AG levels respectively, it is unlikely that the observed effect on macrophage PGE2 production presented in this chapter is mediated by the cannabinoid receptors. This is supported by a study that demonstrated that an anti-oedema effect mediated by FAAH was not reversible by cannabinoid antagonism. 337 This is the first in vivo evidence of an anti-inflammatory effect that despite being mediated by FAAH was independent of the cannabinoid receptors, suggesting that this was due to FAAH’s suppression of PGE2 production.

62 Chapter 2

Following from this it would be interesting to see whether in vivo FAAH and MAGL inhibition can reduce prostaglandin production in tissue, and whether cannabinoid receptor antagonists can block prostaglandin production.

The direct effect of AEA and 2-AG on PGE2 production by LPS-stimulated macrophages has not been previously investigated due to AEA and 2-AG being analogues of arachidonic acid, and hence susceptible to metabolism by COX. 116 As this present study investigated induced COX-2 expression alongside FAAH inhibition, it is possible that COX derived metabolites of AEA were formed. Unlike the cannabinoids, the COX derived AEA metabolite PGE2-ethanolamide does not have an anti-inflammatory effect on activated macrophages. 116 The evidence from this study therefore does not support the involvement of PGE2-ethanolamide in the action of FAAH inhibitors.

While PGE2 was investigated here, FAAH and MAGL may have control of a wider range of lipids involved in inflammation, such as leukotriene and thromboxane. Recently it was demonstrated that MAGL regulates a network of lipids in tumour cells including fatty acids, fatty acid glycerols and PGE2. 338 Having demonstrated here that FAAH and

MAGL can inhibit PGE2 production, it was investigated as to the same effect could be seen with LTB4. A previous study demonstrated that LTB4 could be produced by

RAW264.7 macrophages when stimulated with the calcium ionophore (10 µM

A23187). 339 An unsuccessful attempt was made to reproduce this observation as a model for investigating the effect of FAAH and MAGL inhibitors of LTB4. This was supported by a report demonstrating only a transient production of leukotrienes following stimulation. 340 JZL184 was recently shown to inhibit 2-AG induced biosynthesis of LTB4 in human neutrophils. 341 Therefore human neutrophils would be an appropriate technique for determining whether FAAH inhibition can also suppress

LTB4 biosynthesis.

Further work is required to elucidate the relationship between arachidonic acid release and LPS-stimulation. Ideally this should look at increasing amounts of LPS and different periods of stimulation. To accomplish this an arachidonic acid assay would be required, of which LC-MS/MS methods have been reported. 342 The majority of the arachidonic acid used in PGE2 synthesis is released by cPLA2. 18 Even still MAGL

63 Chapter 2 appears to have significant control over arachidonic acid levels, as MAGL inhibition reduces unesterified arachidonic acid in the brain by 50%. 218 It is important that the relative contributions of FAAH, MAGL and phospholipase A 2 to arachidonic acid levels are determined under both physiological and pathological conditions.

The finding that the ability of FAAH and MAGL to inhibit PGE2 production is LPS dependent suggests that FAAH and MAGL are more likely to influence physiological

PGE2 production, when COX-2 is not induced, than in inflammatory production, when

COX-2 is induced. Further investigation would be required in vivo to determine whether there is any physiological relevance to this finding. Inhibition of FAAH and

MAGL is known to be anti-inflammatory and effective in the treatment of inflammatory pain in vivo . For example FAAH inhibition has previously been shown to reduce peripheral inflammation in mice. 226 Whether this effect is in part due to a suppression of PGE2 needs to be investigated. An ideal future experiment would be to repeat such a study, investigating the effect of a dose-response of a FAAH inhibitor on PGE2 concentrations.

The substrates of FAAH and MAGL have multiple analgesic and anti-inflammatory properties. It is therefore possible that the reduction in PGE2 seen here with FAAH and

MAGL inhibition is in part due to the increased concentrations of their substrates.

Therefore it would be worth determining the effects of OEA, PEA, 2-oleoyl glycerol (2-

OG) and 2-palmitoyl glycerol (2-PG) on LPS-stimulated macrophages to further investigate this.

LPS stimulation of macrophages has been shown to increase AEA levels, induce FAAH expression, 330 increase 2-AG levels and down regulate MAGL expression. 331 The development of FAAH and MAGL assays in macrophages was attempted here using fluorogenic substrates, but with little success. Future studies could achieve this by using a recently reported assay that measures FAAH leukocyte activity by monitoring the hydrolysis of a deuterated AEA analogue with LC-MS/MS. 343 Based on this method a similar MAGAL assay using deuterated 2-AG could be easily adapted for this purpose.

64 Chapter 2

2.4.1. Limitations of the present study

This study has several limitations. Firstly similar to previous studies, 35 PGE2 was measured directly in the cell medium without extraction. The PGE2 assay was carried out in accordance with the manufactures (Cisbio) instructions, which states that the assay is suitable for use with cell culture medium. However there are examples of other studies using C18 cartridges to extract prostaglandins from cell medium. 344 The benefit of such cleaner extracts is less interaction between the other components of the cell medium with the assay. PGE2 was measured using a kit, which although validated by the manufacturer, was not independently validated. In this present study cells were cultured with cell medium containing FBS, with treatments not containing

FBS. While LPS still managed to stimulate production of PGE2 and NO, in hindsight FBS should have been included in treatments as previous studies have shown serum proteins to be potentiate LPS activation of macrophages. 345, 346

The western blotting of COX-2 was part of a pilot study to investigate the effect of the

FAAH and MAGL inhibitors on COX-2 expression. While protein in samples were equalised to 2 µg/µL no loading controls, such as β-actin or tubulin, were used. JZL184 was investigated at 10 nmol/mL on macrophages stimulated with 1 µg/mL LPS. Under these same conditions no effect is seen on PGE2 levels. Therefore in the future the experiment should be reproduced instead using 10 ng/mL LPS. Furthermore like

JZL184, the effect of PF-3845 on COX-2 expression should be investigated.255

While PF-3845 and JZL184 have both been developed for the treatment of inflammatory pain, there are no reports as to whether they have any off-target selectivity for the COX enzymes. To conclusively rule out that the effects of PF-3845 and JZL184 on PGE2 are not due to off-target inhibition of COX, these inhibitors should be assayed for their binding affinity to pure COX-1 and COX-2 proteins.

65 Chapter 2

2.4.2. Summary

In summary, previously it has been shown that URB597 reduces PGE2 in microglia. 255

Here for the first time it has been shown that the more selective FAAH inhibitor PF-

3845 and the MAGL inhibitor JZL184 inhibits 10 ng/mL LPS-stimulated PGE2 production in macrophages, but not at the 100-fold higher concentration of 1 mg/mL.

To allow for a fuller understanding of the interaction between the endocannabinoid system and the prostaglandin system in macrophages, LPS stimulation now needs to be correlated with FAAH and MAGL activity, FAAH and MAGL expression, arachidonic acid levels and PGE2 levels. Furthermore whether there is an in vivo effect on PGE2 levels by FAAH and MAGL inhibitors needs to be investigated.

66 Chapter 3

CHAPTER 3

FAAH substrates as a pharmacodynamic response in an

in vivo model of inflammatory pain

67 Chapter 3

3.1. Introduction

The measurement of the FAAH substrates AEA, OEA and PEA is a useful pharmacodynamic parameter for demonstrating FAAH inhibition and thus proving the mechanism of action for potential inhibitors. The studies presented here will investigate the pharmacodynamic effects of the irreversible covalent inhibitors,

URB597 and PF-04457845, alongside three other reversible non-covalent inhibitors, referred to here as NC1, NC2 and NC3 and the COX inhibitor celecoxib. NC1, NC2 and

NC3 are novel and uncharacterised non-covalent inhibitors of FAAH. Additionally the

COX-2 inhibitor, celecoxib, will be used to investigate whether FAAH substrates would be affected by other analgesics.

To accomplish this, firstly an LC-MS/MS assay for the measurement of AEA, OEA and

PEA in both rat brain and whole blood will be developed and validated. Freund’s complete adjuvant (FCA)-induced hypersensitivity will be used as a model of inflammation in rats, with weight bearing being used to determine analgesia. FCA is an oil emulsion of dead Mycobacterium tuberculosis that is injected into the knee to induce a stable and treatable thermal hyperalgesic pain. 347, 348 The brain and blood

FAAH substrate levels and FAAH activity will be investigated before and after FCA to determine whether they are changed with inflammation. Furthermore measuring reversal of hypersensitivity will determine whether FAAH substrate levels and FAAH activity are changed with analgesia. It is assumed that a change in the enzyme’s activity should be reflected by a change in its substrates. The benefit of measuring both the enzyme’s activity and substrates is that it will provide a comprehensive picture of whether this is always the case.

The accumulation of AEA, OEA and PEA as a result of FAAH inhibition should subject the substrates to increased amounts of non-FAAH-mediated metabolism. As already discussed in Section 1.2.4. the oxidative metabolites of AEA have not been identified in vivo under normal physiological conditions. The oxidative metabolism of the other

FAAH substrates, OEA and PEA, is yet to be investigated and is possibly more likely to be a physiological occurrence due to their higher endogenous concentrations. Using liver microsomes it will be determined as to whether OEA and PEA can undergo

68 Chapter 3 oxidative metabolism by liver microsomes. The rate of oxidation for AEA, OEA and PEA by liver microsomes will be compared.

FAAH is highly expressed in brain and liver, therefore FAAH inhibition will increase the levels of protein bound AEA, OEA and PEA in these tissues. To further understand

FAAH inhibition it is therefore important that the protein binding of these lipids is investigated. Lipids strongly and reversibly bind to non-specific proteins, 349 and therefore exist as bound and unbound fractions, with only unbound lipid signalling molecules being available to activate their receptors and be oxidised. Therefore it is important that intrinsic clearance values be corrected for binding to microsomal protein. 350 Protein binding of AEA, OEA and PEA to microsomal protein will be determined, with also binding to brain protein being investigated.

Together these studies will investigate the utility of the AEA, OEA and PEA as indicators of FAAH inhibition. They will be evaluated on how readily detectable they are analytically, how abundant they are in tissue samples, the extent to which FAAH inhibition increases their concentration, how closely brain levels correlate with blood levels, the extent at which they undergo oxidative metabolism and the degree to which they bind to protein.

69 Chapter 3

3.2. Materials and methods

3.2.1. Materials

FAAH inhibitors were synthesised by chemists at GSK. AEA, OEA, PEA and PEA-d4 were from Cayman Chemicals (Ann Arbor, Michigan, USA). PEA-d4 contains four deuterium atoms, two at the 7-position and two at the 8-position. All analytes were stored in ethanol at -80 oC. URB597 was from Cayman Chemicals and stored as a solid at -20 oC. Far UV/gradient quality acetonitrile was from Romil Chemicals (Cambridge,

UK). Analytical reagent grade formic acid was from BDH (Poole, UK). Analytical reagent grade absolute ethanol was from Fisher Scientific (Loughborough, UK). TRIS-

HCl, TRIS-base, HEPES, EDTA, Glycerol and Triton X-100 were from Sigma Aldrich

(Poole, UK). Sprague Dawley rat liver microsomes were from Xenotech LLC (Kansas,

USA). Microsomes were mixed gender pools prepared at 20 mg/mL in 50 mM sodium phosphate buffer (pH 7.4). The 96-well equilibrium dialysis block and dialysis membranes were from HTDialysis LLC (Connecticut, USA).

3.2.2. FCA-induced hypersensitivity

Male Random hooded rats (Charles River Laboratories, Kent, UK) were kept on a 12 hour light/dark cycle with ad libitum access to food and water. They were fed a 5LF2 diet (LabDiet, UK). Rats used in the studies were typically between 9 - 10 weeks old and 200 – 220 g in weight.

Inflammation was induced by injection of 100 µL 1 mg/mL FCA into the left hind paw.

Weight bearing on the left hind paw was measured before and after FCA to determine

FCA-induced hypersensitivity, and then measured after the drug to determine % reversal of FCA-induced hypersensitivity. Rats were culled under CO 2 followed by cervical dislocation 1 hour after drug administration. Rat brains were frozen on collection in liquid nitrogen. Whole rat blood was collected into EDTA tubes from the rat tail vein for serial studies and using cardiac puncture once rats had been culled.

Samples were stored at -80 oC until analysis. All animal studies were ethically reviewed and carried out in accordance with Animals (Scientific Procedures) Act 1986 and the

70 Chapter 3

GSK Policy on the Care, Welfare and Treatment of Laboratory Animals. The above work was performed by technicians at GSK.

3.2.3. Drug administration

URB597 was formulated as 2 mg/mL in 2% DMSO, 65% PEG-400 and 33% H 2O. NC1,

NC2, NC3 and celecoxib were prepared in methyl cellulose. PF-04457845 was prepared in 5% dimethylacetamide (DMA) with 95% cyclodextrin, URB597 was administered at

10 mg/kg as an intraperitoneal ( i.p. ) injection. NC1 and NC2 were both administered at 10 mg/kg i.p. PF-04457845 was administered at 3 mg/kg i.p. NC3 was administered at 3, 10 and 30 mg/kg i.p. Celecoxib was administered at 10 mg/kg i.p. The above work was performed by technicians at GSK.

3.2.4. Sample preparation

Left brain hemisect was collected for FAAH substrate measurements, while the right brain hemisect was collected for a separate FAAH activity assay. An aqueous buffer

(pH 7.4, 125 mM TRIS, 0.4 mM HEPES, 1 mM EDTA, 0.2% v/v Glycerol, 0.02% v/v

Triton X-100 and 10 µM URB597) was added at 2 mL per gram of brain tissue and homogenised using an Ultra-turrax handheld high speed disperser. 100 µL homogenate was extracted and protein precipitated with 300 µL acetonitrile containing

100 ng/mL internal standard, PEA-d4. Samples were vortex mixed and centrifuged for

15 minutes at 1,500 g. 50 µL supernatant was transferred into a 96-well plate and diluted with 50 µL H2O using a Tecan liquid handling robot (Reading, UK).

Upon collection 50 µL blood was immediately deproteinated by addition of 150 µL acetonitrile containing 100 ng/mL PEA-d4 as an internal standard. Samples were frozen on dry ice and stored at -80 oC until analysis. The precipitated blood was allowed to thaw, then vortex mixed and centrifuged for 15 minutes at 1,500 g. 50 µL supernatant was transferred into a 96-well plate and diluted with 50 µL H2O using a

Tecan liquid handling robot.

71 Chapter 3

3.2.5. FAAH substrate assay by LC-MS/MS

LC-MS/MS analysis was performed on a Waters Acquity system with a Quattro Premier tandem mass spectrometer. A 5 µL injection was made onto a Waters Acquity BEH phenyl column (2.1 x 50 mm, 1.7 µm) maintained at 50 oC. Gradient elution was between mobile phase A (95% H2O, 5% acetonitrile, 0.1% formic acid) and B (95% acetonitrile, 5% H2O, 0.1% formic acid). At 0.1% formic acid the mobile phase was pH

2.6. The gradient ran from 20% B, increasing linearly to 90% over 2 minutes at a flow rate of 0.5 mL/min. This was held until 2.2 minutes, then returned to 20% and allowed to re-equilibrate. The total runtime was 2.5 minutes. Analytes were quantified by tandem mass spectrometry with positive electrospray ionisation. Multiple reaction monitoring (MRM) was performed with the [M + H]+ selected for each analyte, monitoring the transitions 348
62 m/z for AEA, 326
62 m/z for OEA, 300
62 m/z for PEA and 304
62 m/z for PEA-d4. The optimised parameters for monitoring these transitions are shown in Table 3.1.

Table 3.1. – The selected precursor and product m/z values for each analyte with their respective dwell times, cone voltages and collision energy voltages used for their identification and quantification by LC-MS/MS.

Precursor Product Dwell Cone voltage Collision energy

(m/z ) (m/z ) (s) (V) (V)

AEA 348.4 62.0 0.12 22 16

OEA 326.3 62.0 0.12 28 22

PEA 300.1 62.0 0.12 22 16

PEA-d4 304.5 62.0 0.12 22 16

Where m/z is the mass-to-charge ratio, s is seconds and V is voltage.

For the calibration line, standards of AEA, OEA and PEA were prepared at 12 concentrations from 4 pmol/g – 20,000 pmol/g in control brain tissue homogenate and

1 pmol/mL – 6,500 pmol/mL in control whole blood (Figure 3.10.). Standards were prepared using a Tecan liquid handling robot. 10 µL each standard was spiked into 100

µL brain homogenate, 5 µL was spiked into 50 µL blood. Blanks were made using H2O instead of analyte standard and total blanks were precipitated using acetonitrile without the internal standard. Due to high endogenous levels of OEA and PEA in the brain the calibration lines do not intercept the y-axis at zero.

72 Chapter 3

3.2.6. FAAH activity assay

FAAH activity was assayed using a previously reported method. 351 Brain homogenates and blood were incubated for 10 minutes at 37 oC in the presence of 2 µM [3H]-AEA.

FAAH hydrolyses this substrate into radiolabeled ethanolamine and unlabeled arachidonic acid. The unhydrolysed [3H]-AEA and arachidonic acid were removed by differential absorption to activated charcoal. The remaining radiolabeled ethanolamine was measured by liquid scintillation counting. Brain and blood from untreated rats was assumed to have 100% FAAH activity. FAAH activity was expressed as a percentage of this control group. The above work was performed by technicians at GSK.

3.2.7. Intrinsic clearance assay

Oxidation of the FAAH substrates will be investigated using the standard approach, which is the use of liver microsomes to assay intrinsic clearance. 350 Drug metabolising enzymes are abundant in the liver. Microsomes are isolated from the liver by centrifuging liver homogenate centrifuged at 9,000 g for 20 minutes. 352 At 9,000 g cells and large organelles sediment out of solution leaving cytochrome P450, other soluble enzymes and fragments of endoplasmic reticulum. The microsomes are then isolated by ultracentrifugation at 100,000 g for 1 hour. 352 This process selectively concentrates intracellular enzymes, such as cytochrome P450. 352

An automated Packard MultiPROBE II HT EX system (West Drayton, UK) was used for the determining intrinsic clearance. A 0.5 mg/mL suspension of microsomes was prepared in PBS and incubated at 37 OC for 10 minutes. The compound of interest was added to 300 µL of the suspension to give a final concentration of 0.5 µM. A co-factor solution was used to initiate the cytochrome P450 enzymes. 80 µL of the co-factor solution was added, which contained 7.8 mg/mL glucose-6-phosphate, 1.7 mg/mL

NADP and 6 U/mL glucose-6-phosphate dehydrogenase (EC 1.1.1.49) in 2% sodium bicarbonate buffer (pH 7.4). Samples of 25 µL were removed at 0, 3, 6, 9, 12, 15, 18,

24 and 30 minutes, and the proteins precipitated with 200 µL acetonitrile containing the 100 ng/mL of the internal standard PEA-d4. The sample extracts were then

73 Chapter 3 centrifuged for 15 minutes at 1,500 g. 150 µL supernatant was extracted and then diluted with 150 µL H 2O. The clearance (CLi) was calculated as:

CLi= k x 1/I x SF

where k is the rate constant, I is the compound concentration and SF is the scaling factor. CLi is expressed as mg/min/g of wet liver tissue. The scaling factor was used to predict the clearance in vivo from the clearance observed in the microsomes. For rat to human this is accepted as being 52.5. 353-355

Cytochrome P450 activity requires cofactors, which were added to the liver microsome suspension. Microsomes without co-factors were also prepared to investigate non- cytochrome P450 activity. The percentage of the compound cleared without co-factors compared to that with co-factors was calculated and is referred to here as no-cofactor loss. A value near 0% suggests the compound was mostly metabolised by cytochrome

P450, while a value near 100% means that the compound was highly metabolised via non-cytochrome P450 pathways. The experimental validity was determined by two compounds, namely 7-ethoxycoumarin and dextromethorphan, which were used as positive controls. Both are metabolised by a range of cytochrome P450 isoforms and therefore are useful probes to determine whether the enzymes are active or not. 356, 357

FAAH substrates were measured by LC-MS/MS according to Section 3.2.5. Clearance was calculated using the peak area ratio of analyte to internal standard, therefore calibration was not required.

3.2.8. Protein binding assay

The standard assay for protein binding is equilibrium dialysis. 358 A dialysis membrane is used to separate two compartments. The compound and the matrix to be investigated are placed on one side, with an appropriate buffer placed on the other.

The unbound fraction of the compound is therefore free to diffuse across the membrane. At equilibrium the unbound compound will be at the same concentration in either compartment. But as protein bound compounds are unable to permeate the membrane they will not enter the buffer compartment. Hence the total concentration

74 Chapter 3 will be higher in the matrix compartment. The concentration of unbound compound in the buffer compartment divided by the concentration of the bound and unbound compound in the matrix compartment gives the percentage protein binding.

While Random hooded rats were used in the in vivo studies, liver microsomes from this strain were not commercially available. Sprague Dawley rat brain and liver microsomes were used instead for the investigation of protein binding.

Dialysis membrane strips, obtained from HTDialysis LLC, were soaked for 1 hour in

H2O and then left in 20% ethanol / 80% H 2O until required for use. The 96-well equilibrium dialysis block was assembled with each well in the 96-well plate being separated into two compartments by the dialysis membrane. Brains were homogenised with PBS at 2mL per gram of tissue. Microsome suspensions were prepared at 0.5 mg/mL in 50 mM PBS (pH 7.4). AEA, OEA and PEA were added separately to the brain homogenate at a final concentration of 10 µM. AEA, OEA and PEA were added separately to the microsomes at a final concentration of 0.3 µM. 100 µL PBS was added to the buffer compartment of the membrane and to the other side, the matrix compartment, 100 µL of the incubation mixture was added, either the microsomes or brain homogenate. The plate was sealed with an adhesive strip to prevent evaporation.

The block was placed on a shaker inside an incubator cabinet set at 37 oC. Brain homogenates were incubated for 5 hours and microsomes for 20 hours.

As the analysis of the samples was by LC-MS/MS it was essential to ensure that all samples had a similar matrix composition and final volume to prevent matrix effects.

For the microsomes, after incubation 50 µL from the matrix compartment was precipitated with 200 µL acetonitrile and then 50 µL PBS was added. From the buffer compartment, 50 µL was precipitated with 200 µL acetonitrile and then 50 µL control microsome suspension was added. For brain, 50 µL of the matrix compartment was precipitated with 200 µL acetonitrile and then 100 µL PBS was added. From the PBS compartment 50 µL was precipitated with 200 µL acetonitrile and then 100 µL control brain homogenate was added. PEA-d4 was used as the internal standard and was present in the acetonitrile at 100 ng/mL. Samples were centrifuged for 15 minutes at

1,500 g. The supernatant was removed and diluted 1:1 with H 2O before analysis.

75 Chapter 3

FAAH substrates were measured by LC-MS/MS according to Section 3.2.5. percentage protein binding was calculated using the peak area ratio of analyte to internal standard, therefore calibration was not required. The percentage of the unbound compound was taken as the amount of unbound compound in the buffer compartment divided by the amount of bound and unbound compound in the matrix compartment.

Subtracting this percentage from 100 then gave the percentage of the protein bound compound.

3.2.9. Statistical analysis

Data was analysed by either ANOVA with Tukey’s post-hoc analysis or t-test. Tukey’s post-hoc analysis was used as it compares all possible pairs of groups, in a multiple group analysis. p < 0.05 was deemed statistically significant. Data are presented as mean ± SEM. Statistical analysis was by Prism 3.0 (GraphPad).

76 Chapter 3

3.3. Results

3.3.1. FAAH substrate assay - optimisation

The MS/MS conditions were optimised for maximum formation and detection of product ions by direct infusion of 100 ng/mL standard solutions. The FAAH substrates positively ionised under electrospray ionisation and formed a prominent 62 m/z ion when fragmented, corresponding to ethanolamine (Figures 3.1. – 3.4.).

Figure 3.1. – Product ion mass spectrum and corresponding structure for AEA (348.4

62.0 m/z ) following collision induced dissociation. Site of fragmentation is identified on the figure. MS/MS conditions are per Table 3.1.

Figure 3.2. – Product ion mass spectrum and corresponding structure for OEA (326.3

62.0 m/z ) following collision induced dissociation. Site of fragmentation is identified on the figure. MS/MS conditions are per Table 3.1.

77 Chapter 3

Figure 3.3. – Product ion mass spectrum and corresponding structure for PEA (300.1

62.0 m/z ) following collision induced dissociation. Site of fragmentation is identified on the figure. MS/MS conditions are per Table 3.1.

Figure 3.4. – Product ion mass spectrum and corresponding structure for PEA-d4 (304.5

62.0 m/z ) following collision induced dissociation. Site of fragmentation is identified on the figure. MS/MS conditions are per Table 3.1.

The analytes each lost a hydroxyl group, hence the presence of a fragment 17 m/z less than the precursor ion in each spectra. The cone voltage, collision energy and dwell times obtained from the optimisation process are shown in Table 3.1.

78 Chapter 3

A phenyl column was used as the substrates were highly retained on the C18 columns, which was not suited for the desired fast run times. Formic acid was used in the mobile phase to promote ionisation. A fast gradient was required so that the method could be high throughput. The three analytes and internal standard were not chromatographically resolved (Figure 3.5. and 3.6.), but as ion suppression was within the acceptable ±20% limit this did not need to be dealt with (Table 3.2.).

A

B

C

D

Figure 3.5. – Representative chromatograms obtained by LC-MS/MS showing endogenous AEA (A), OEA (B), PEA (C) and the internal standard PEA-d4 (D) in rat brain. Brain was extracted and analysed as described in Section 3.2.4. and Section 3.2.5.

79 Chapter 3

A

B

C

D

Figure 3.6. – Representative chromatograms obtained by LC-MS/MS showing endogenous AEA (A), OEA (B), PEA (C) and the internal standard PEA-d4 (D) in rat blood. Blood was extracted and analysed as described in Section 3.2.4. and Section 3.2.5.

As AEA, OEA and PEA are enzyme substrates it was important that FAAH activity was stopped as early in the sample preparation process as possible, to minimise analyte loss. On collection brains were frozen in liquid nitrogen and stored at -80 oC. While still frozen, brains were homogenised, then immediately protein precipitated with acetonitrile and centrifuged before extraction of the supernatant. Blood on collection was immediately protein precipitated with acetonitrile and stored at -80 oC until analysis without any prior centrifugation.

80 Chapter 3

FAAH is predominantly expressed in the brain. For the preparation of brain tissue the use of the FAAH inhibitor URB597 was investigated. The principle being that inhibiting the enzyme responsible for the degradation of the analytes would minimise analyte loss. A comparison was made between calibration lines prepared from brains homogenised with and without 10 µM URB597. The homogenates were then extracted in the same way as previously described. This was to observe the effect of URB597 on

FAAH activity in the presence of its substrates at a range of concentrations. The presence of URB597 further reduced enzyme activity and thus improved recovery of each analyte (Figure 3.7.).

1.2 R2 = 0.9995

0.8

+ URB597 - URB597

0.4

R2 = 0.9901 Peak area Peak ratio area of analyte to internal standard 0 0 4000 8000 12000 16000 AEA (pmol/g wet weight brain)

Figure 3.7. – Effect of URB597 on AEA stability in brain homogenate. Calibrations lines were prepared in brain homogenates either in the absence (open circles) or presence of 10 µM

URB597 (filled circles). Samples were extracted and analysed as described in Section 3.2.4. and

Section 3.2.5.

3.3.2. FAAH substrate assay - validation

The method was validated following the FDA approved guidance and principles of validation (Table 3.2.).359 The linearity was good, with all calibration lines having R 2 values above 0.99 (Figure 3.8). The limit of quantification (LOQ) was taken as the concentration at which the signal:noise ratio = 10. All analytes demonstrated a LOD of

2 pmol/g wet weight brain and 0.3 pmol/g blood, and a LOQ of 7 pmol/g wet weight brain and 0.9 pmol/mL blood. All samples contained levels of AEA, OEA and PEA above the LOQ.

81 Chapter 3

8 y = 0.0003x + 0.8406 2 A R = 0.9986

y = 0.0004x + 0.4983 6 R2 = 0.9989

AEA 4 OEA y = 0.0001x + 0.0169 PEA R2 = 0.9993 internal standard 2 Peak area Peak ratio area of analyte to

0 0 4000 8000 12000 16000 20000 Conc. (pmol/g wet weight brain)

8 B y = 0.0012x + 0.0318 R2 = 0.9981 y = 0.0009x + 0.0085 2 6 R = 0.9979 AEA OEA 4 PEA internal standard y = 0.0004x - 0.0061 2 R2 = 0.9954 Peak area Peak ratio area of analyte to

0 0 1000 2000 3000 4000 5000 6000 7000 Conc. (pmol/mL blood)

Figure 3.8. – AEA, OEA and PEA calibration lines prepared in (A) rat brain and (B) whole blood. Samples were extracted and analysed as described in Section 3.2.4. and Section 3.2.5.

For brain and blood the extraction repeatability was calculated for 10 different extracts

(Table 3.2.). The intra-injection repeatability was calculated for the same extract injected onto the column 10 times. The high variation seen with AEA in blood is most likely due to its lower endogenous levels and shorter biological half life compared to

OEA and PEA. 221, 360 The higher variation in blood than brain for the 3 substrates is most likely due to the endogenous levels being lower in this tissue.

82 Chapter 3

Table 3.2. – Summary of validation parameters for AEA, OEA and PEA in brain and blood by LC-MS/MS

AEA OEA PEA Brain LOD (pmol/g) 2 2 2 LOQ (pmol/g) 7 7 7 Extraction repeatability 8% 6% 8% Injection repeatability 6% 2% 3% Recovery 94 ± 7% 93 ± 8% 92 ± 4% Ion suppression -10 ± 1% -13 ± 3% -15 ± 2%

Blood LOD (pmol/mL) 0.3 0.3 0.3 LOQ (pmol/mL) 0.9 0.9 0.9 Extraction repeatability 32% 12% 4% Injection repeatability 28% 10% 6% Recovery 103 ± 6% 101 ± 7% 102 ± 11% Ion suppression 9 ± 1% 7 ± 1% 7 ± 2% LOD and LOQ defined as signal:ratio = 3 and 10 respectively. For extraction repeatability, samples were repeatedly extracted, assayed and expressed as the coefficient of variation (n=10).

For intra-injection repeatability, a single extracted sample was repeatedly assayed and expressed as the coefficient of variation (n=10). Concentrations of 100 ng/mL were used for each analyte.

Values were calculated using peak areas and expressed as the mean percentage ± SD (n = 6).

Samples were extracted and analysed as described in Section 3.2.4. and Section 3.2.5. Recovery and ion suppression were determined using peak areas. Concentrations of 50 ng/mL were used for each analyte. Values expressed as the mean percentage ± SD (n = 6). Samples were extracted and analysed as described in Section 3.2.4. and Section 3.2.5.

Recovery and ion suppression were determined for AEA, OEA and PEA in both brain and blood (Table 3.2.). For recovery, brain and blood was spiked with 50 ng/mL of the analytes and then extracted. The peak area was compared to that of brain and blood extracted first and then spiked with the analyte. Both groups were corrected for endogenous background levels and expressed as a percentage. As long as the extraction method is reproducible, recovery need not be near 100%, ideally being within a ± 20% limit. For ion suppression, a comparison was made between brain and blood that was extracted and then spiked with 50 ng/mL of the analytes, to that of a

50 ng/mL solution of the analytes. Again endogenous levels were corrected for and the result was expressed as a percentage. Ion suppression values should also be within a

± 20% limit, with the values being well within this limit.

83 Chapter 3

Brain homogenate was spiked at 100 ng/mL with either AEA, OEA or PEA. Samples were then stored on ice for different periods of time before being extracted and analysed. No significant degradation was observed over the first five hours (Figure

3.9.). In reality brain samples were extracted immediately after homogenisation. The long term stability of samples was not tested as samples were not stored long term.

Samples were analysed within a couple of days of collection.

0.6

0.4 AEA OEA PEA

0.2 Peak area Peak ratio area of to analyte internal standard 0.0 0 1 2 3 4 5 Time (hours)

Figure 3.9. – Stability of analytes in brain homogenate. 100µL brain homogenates were spiked with either 100 ng/mL AEA, OEA or PEA and stored on ice over the time course. Samples were extracted and analysed as described in Section 3.2.4. and Section 3.2.5. Data expressed as mean ± SD (n = 3).

Brains were homogenised in a buffer containing the non-ionic surfactant Triton X-100, which at a low concentration will solubilise membrane proteins to help maintain their native state structure. 361 This treatment was required for the FAAH activity assay. To confirm the utility of this solution in the sample preparation stage, it was investigated as to whether the buffer would cause any matrix effects. Calibration lines were prepared from brain homogenates treated with either H 2O or in this buffer. No difference was seen in the calibration line’s gradient and linearity for either treatment, proving that this was a suitable buffer.

84 Chapter 3

3.3.3. Effect of URB597 on brain and blood FAAH substrate levels in an in vivo inflammatory pain model

Rats injected with FCA showed a reduced weight distribution on the inflamed paw by

90 ± 10% (Figure 3.10.).

100 ***

80

60

40 hypersensitivity

% reversal ofFCA-induced 20

0 Naive Vehicle URB597

Figure 3.10. – Reversal of FCA-induced hypersensitivity by URB597 in rats. Rats (n = 18) were injected with FCA and left for 24 hours. Within each group naïve (n = 6), vehicle (n = 6) or

URB597 (10 mg/kg i.p.) (n = 6) treatments were given. Weight bearing on the left hind paw was determined and expressed as % reversal of FCA-induced hypersensitivity. Data expressed as mean ± SEM. Statistical analysis was by ANOVA with Tukey’s post-hoc analysis. *** = p ≤ 0.001 in comparison with naïve and vehicle treatments.

FCA did not change FAAH activity (Figure 3.11. A) or FAAH substrate levels in the brain

(Figure 3.12.). However it did reduce FAAH activity in the blood by 40% (Figure 3.11

B), despite blood FAAH substrates being unchanged (Figure 3.13). At 10 mg/kg,

URB597 reversed the FCA-induced hypersensitivity by 80 ± 15% (Figure 3.10.), indicating that URB597 was analgesic. FAAH activity was significantly inhibited in both brain and blood by URB597 (Figure 3.11.). URB597 significantly increased the brain and blood FAAH substrate levels in comparison to naïve and vehicle treatments

(Figures 3.12 & 3.13). This effect was seen in both the control and FCA groups.

In comparison between the FCA vehicle and FCA URB597 treatments the concentration of brain PEA increased the most (2700 ± 380 pmol/g), followed by OEA (1830 ± 220 pmol/g) then AEA (70 ± 16 pmol/g). In the blood, OEA increased the most (105 ± 20 pmol/mL), followed by PEA (50 ± 6.6 pmol/mL), then AEA (4 ± 0.6 pmol/mL).

85 Chapter 3

120 A Naive Vehicle 100 URB597 80

60

% FAAH activity FAAH % 40

20 *** 0 ***

Control FCA

120 B *** 100

80

60

% FAAH activity FAAH % 40 ** 20 ***

0 Control FCA

Figure 3.11. – Percentage FAAH activity in (A) brain and (B) blood of rats with or without FCA, either untreated (naïve), treated with vehicle or treated with the FAAH inhibitor URB597. Rats were either untreated (n = 18) or injected with FCA (n = 18), then left for 24 hours. Within each group naïve (n = 6), vehicle (n = 6) or URB597 (10 mg/kg i.p.) (n = 6) treatments were given as per Section 3.2.6. Brain and blood was collected 1 hour after treatment. % FAAH activity was determined. Data expressed as mean ± SEM. Statistical analysis was by ANOVA with Tukey’s post-hoc analysis. ** = p ≤ 0.01 and *** = p ≤ 0.001. Comparisons made between naïve and URB597 treatments. Other comparisons depicted on figures.

86 Chapter 3

200 A Naive Vehicle 150 *** *** URB597

100

50 AEA (pmol/g wet weightof brain) 0 Control FCA

3000 B *** ***

2000

1000 OEA (pmol/gwet weight of brain) 0 Control FCA

*** 4000 C ***

3000

2000

1000 PEA (pmol/g wetweight of brain) 0 Control FCA

Figure 3.12. – Brain levels of AEA (A), OEA (B) and PEA (C) in rats without (control) or with FCA that were either untreated (naïve), treated with vehicle or treated with the

FAAH inhibitor URB597. Rats were either untreated (n = 18) or injected with FCA (n = 18), then left for 24 hours. Within each group naïve (n = 6), vehicle (n = 6) or URB597 (10 mg/kg i.p.) (n = 6) treatments were given. Brain was collected 1 hour after treatment. Samples were extracted and analysed as described in Section 3.2.4. and Section 3.2.5. Data expressed as mean

± SEM. Statistical analysis was by ANOVA with Tukey’s post-hoc analysis. *** = p ≤ 0.001 in comparison with the URB597 treatment group to the naïve and vehicle treatments.

87 Chapter 3

The vehicle had no effect on hypersensitivity or FAAH activity. While in brain the vehicle had no effect on FAAH substrate levels, there was interestingly a significant increase in blood PEA levels in both the control and FCA groups (Figure 3.13. C).

8 A Naive *** *** Vehicle 6 URB597

4

AEA (pmol/mL AEA blood) 2

0 Control FCA

250 *** B *** 200

150

100 OEA(pmol/mL blood) 50

0 Control FCA

150 C *** ***

100 *** ** 50 PEA (pmol/mLPEA blood)

0 Control FCA

Figure 3.13. – Blood levels of AEA (A), OEA (B) and PEA (C) in rats untreated (naïve), treated with vehicle or treated with the FAAH inhibitor URB597. Rats were either untreated (n = 18) or injected with FCA (n = 18), then left for 24 hours. Within each group naïve

(n = 6), vehicle (n = 6) or URB597 (10 mg/kg i.p.) (n = 6) treatments were given. Blood was collected 1 hour after treatment. Samples were extracted and analysed as described in Section

3.2.4. and Section 3.2.5. Data expressed as mean ± SEM. Statistical analysis by ANOVA with

Tukey’s post-hoc analysis. ** = p ≤ 0.01 and *** = p ≤ 0.001 in comparison with the URB597 treatment group to the naïve and vehicle groups.

88 Chapter 3

For the FCA group treated with URB597 blood was collected before and after FCA, and after dosage of URB597. No difference was seen before and after FCA. AEA, OEA and

PEA were significantly increased after URB597 (Figure 3.14.).

8 A *** 6

4

AEA (pmol/mLAEA blood) 2

0 Before FCA After FCA After URB597

250 B *** 200

150

100 OEA(pmol/mL blood) 50

0 Before FCA After FCA After URB597

150 C ***

100

50 PEA (pmol/mLPEA blood)

0 Before FCA After FCA After URB597

Figure 3.14. – Blood levels of AEA (A), OEA (B) and PEA (C) in rats before and after FCA injection, and then after treatment with the FAAH inhibitor URB597. Blood was collected from rats (n = 6) before injection of FCA, 24 hours after the injection, URB597 (10 mg/kg i.p.) was dosed and blood was collected again 1 hour later. Samples were extracted and analysed as described in Section 3.2.4. and Section 3.2.5. Data expressed as mean ± SEM. Statistical analysis by paired two-tailed t-test. *** = p ≤ 0.001 in comparison with both the before FCA group.

89 Chapter 3

3.3.4. Effect of the PF-04457845 and two non-covalent inhibitors, NC1 and NC2, on brain and blood FAAH substrate levels in an in vivo inflammatory pain model

The two non-covalent FAAH inhibitors, NC1 (10 mg/kg i.p. ) and NC2 (10 mg/kg i.p. ) did not increase FAAH substrate levels in either brain or blood, with the exception of a small increase in blood OEA levels by NC2 (Figure 3.15. & 3.16.).

80 A

60

40

20 AEA (pmol/g wet weight of brain) 0 MC NC1 NC2 DMA PF-04457845

1200 *** B

800

400 OEA (pmol/g wet weight of brain) 0 MC NC1 NC2 DMA PF-04457845

3000 C 2500 ***

2000

1500

1000

500 PEA (pmol/gPEA wetweight of brain) 0 MC NC1 NC2 DMA PF-04457845

Figure 3.15. – Brain levels of AEA (A), OEA (B) and PEA (C) in rats treated with either the non-covalent FAAH inhibitors, NC1 or NC2, or the covalent FAAH inhibitor

PF-04457845. Rats were injected with FCA and 24 hours later treated with either an inhibitor or its vehicle. NC1 (10 mg/kg i.p.) and NC2 (10 mg/kg i.p.) were dosed with the vehicle methyl cellulose (MC). PF-04457845 (3 mg/kg i.p.) was dosed with the vehicle 5% DMA and 95% cyclodextrin. Brain was collected 1 hour after dosage. Samples were extracted and analysed as described in Section 3.2.4. and Section 3.2.5. Data expressed as mean ± SEM (n = 5 per group).

Statistical analysis by ANOVA with Tukey’s post-hoc analysis. *** = p ≤ 0.001 in comparison to their corresponding vehicle groups.

90 Chapter 3

The covalent FAAH inhibitor PF-04457845 (3 mg/kg i.p. ) significantly increased brain

OEA (310 ± 110 pmol/g) and PEA (710 ± 220 pmol/g) levels, and blood AEA (3.3 ±

0.7 pmol/mL), OEA (14 ± 3 pmol/mL) and PEA (18 ± 4 pmol/mL) levels. Despite these increases brain AEA levels were unchanged by inhibition.

6 A ***

4

AEA (pmol/mL AEA blood) 2

0 MC NC1 NC2 DMA PF-04457845

40 B *** 30

20 *

OEA (pmol/mL blood) 10

0 MC NC1 NC2 DMA PF-04457845

50 C *** 40

30

20 PEA (pmol/mLPEA blood) 10

0 MC NC1 NC2 DMA PF-04457845

Figure 3.16. – Blood levels of AEA (A), OEA (B) and PEA (C) in rats treated with either the non-covalent FAAH inhibitors, NC1 or NC2, or the covalent FAAH inhibitor PF-

04457845. Rats were injected with FCA and 24 hours later treated with either an inhibitor or its vehicle. NC1 (10 mg/kg i.p.) and NC2 (10 mg/kg i.p.) were dosed with the vehicle methyl cellulose (MC). PF-04457845 (3 mg/kg i.p.) was dosed with the vehicle 5% DMA and 95% cyclodextrin. Blood was collected 1 hour after dosage. Samples were extracted and analysed as described in Section 3.2.4. and Section 3.2.5. Data expressed as mean ± SEM (n = 5 per group).

Statistical analysis by ANOVA with Tukey’s post-hoc analysis. * = p≤ 0.05 and *** = p ≤ 0.001 in comparison to their corresponding vehicle groups.

91 Chapter 3

3.3.5. Dose response of the non-covalent FAAH inhibitor, NC3, on brain and blood

FAAH substrate levels in an in vivo inflammatory pain model

NC3, a non-covalent FAAH inhibitor that lacks the ability to cross the blood-brain barrier, failed to increase any of the brain FAAH substrate levels at doses 3 – 30 mg/kg (Figure 3.17. & 3.18.). A minor exception was that in blood low increases of

OEA (16 ± 7 pmol/mL) and PEA (13 ± 8 pmol/mL) were seen at the highest dose, 30 mg/kg.

80 A

60

40

20 AEA (pmol/g wet weightof brain) 0 MC 3 mg/kg 10 mg/kg 30 mg/kg

1000 B 800

600

400

200 OEA (pmol/g wet weight of brain) 0 MC 3 mg/kg 10 mg/kg 30 mg/kg

2000 C 1600

1200

800

400 PEA PEA (pmol/g wet weight of brain) 0 MC 3 mg/kg 10 mg/kg 30 mg/kg

Figure 3.17. – Dose response of a non-covalent FAAH inhibitor, NC3, on rat brain levels of AEA (A), OEA (B) and PEA (C). Rats were injected with FCA and 24 hours later treated with either the inhibitor NC3 (3, 10 or 30 mg/kg) or its vehicle methyl cellulose (MC). Brain was collected 1 hour after dosage. Samples were extracted and analysed as described in Section

3.2.4. and Section 3.2.5. Data expressed as mean ± SEM (n = 6 per group). Statistical analysis by ANOVA with Tukey’s post-hoc analysis.

92 Chapter 3

25 A 20

15

10 AEA (pmol/mL AEA blood) 5

0 MC 3 mg/kg 10 mg/kg 30 mg/kg

*** 40 B 30

20 OEA (pmol/mL blood) 10

0 MC 3 mg/kg 10 mg/kg 30 mg/kg

* 40 C

30

20 PEA (pmol/mLPEA blood) 10

0 MC 3 mg/kg 10 mg/kg 30 mg/kg

Figure 3.18. – Dose response of a non-covalent FAAH inhibitor, NC3, on rat blood levels of AEA (A), OEA (B) and PEA (C). Rats were injected with FCA and 24 hours later treated with either the inhibitor NC3 (3, 10 or 30 mg/kg) or its vehicle, methyl cellulose (MC). Brain was collected 1 hour after dosage. Samples were extracted and analysed as described in Section

3.2.4. and Section 3.2.5. Data expressed as mean ± SEM (n = 6 per group). Statistical analysis by ANOVA with Tukey’s post-hoc analysis. * = p≤ 0.05 and *** = p ≤ 0.001 in comparison with the vehicle group.

3.3.6. Effect of the COX-2 inhibitor, celecoxib, on brain and blood FAAH substrate levels in an in vivo inflammatory pain model

Celecoxib is a selective COX-2 inhibitor (IC 50 = 40 pmol/mL) that functions as an analgesic and anti-inflammatory through its inhibition of prostaglandin synthesis. 362

The selective COX-2 inhibitor, celecoxib, was used to investigate whether the FAAH substrates would be affected by an analgesic that worked independent of FAAH.

93 Chapter 3

Celecoxib (10 mg/kg i.p.), failed to increase brain or blood AEA, OEA or PEA levels

(Figure 3.19.).

Methyl cellulose 1600 A 10 mg/kg Celecoxib

1200

800

400 Conc. (pmol/g wet weight brain)

0 AEA OEA PEA

40 B 30

20

Conc. (pmol/mLblood) 10

0 AEA OEA PEA

Figure 3.19. – Brain (A) and blood (B) levels of AEA, OEA and PEA in rats treated with and without celecoxib. Rats (n=12) were injected with FCA and 24 hours later treated with either celecoxib (10 mg/kg i.p.) (n=6) or its vehicle methyl cellulose (n=6). Brain and blood was collected 1 hour after treatment. Samples were extracted and analysed as described in Section

3.2.4. and Section 3.2.5. Data expressed as mean ± SEM (n = 6 per group). Statistical analysis by ANOVA with Tukey’s post-hoc analysis.

3.3.7. Oxidative metabolism of the FAAH substrates in liver microsomes

The oxidative metabolism of the FAAH substrates was investigated using liver microsomes. The preparation of liver microsomes selectively concentrates cytosolic intracellular enzymes, therefore the membrane bound FAAH protein should be absent.

However as FAAH is highly expressed in liver it was necessary to initially investigate

FAAH activity in liver microsomes. Microsomes were incubated with 0.5 µM URB597 for

15 minutes before either AEA, OEA or PEA was added. The rate of metabolism of the

FAAH substrates by the liver microsomes was identical to those without URB597

(Figure 3.20.), demonstrating that the liver microsomes lacked FAAH activity.

94 Chapter 3

0.004 A 0.003

+ URB597 0.002 - URB597

0.001 AEA:Peak area ratio of analyte to analyte internal standard 0.000 0 5 10 15 20 25 30 Time (min)

B 0.12

0.08

0.04 OEA:Peak area ratioof analyteto internal standard 0.00 0 5 10 15 20 25 30 Time (min)

0.06 C

0.04

0.02 PEA: area Peak ratio of analyte analyte tointernal standard 0.00 0 5 10 15 20 25 30 Time (min)

Figure 3.20. – Oxidative metabolism of AEA (A), OEA (B) and PEA (C) by liver microsomes with and without URB597. AEA, OEA and PEA (0.5 µM) were added separately to

0.5 mg/mL liver microsome suspensions with (squares) and without URB597 (0.5 µM) (circles) and incubated at 37 OC over a 30 minute time course. Reactions were stopped by protein precipitation. FAAH substrates were not quantitated, with the peak area ratio to the internal standard instead being used. Intrinsic clearance was assayed as per Section 3.2.7.

Having demonstrated that FAAH does not contribute to the substrate metabolism in liver microsomes, the intrinsic clearance of AEA, OEA and PEA was investigated. No co- factor loss, t 1/2 (half life), intrinsic clearance (CLi) and the rate constant ( k) were all determined. AEA was more readily metabolised by non-FAAH enzymes than OEA and

PEA, as demonstrated by its short t 1/2 and fast clearance (Figure 3.21.). High no co- factor losses were seen for each substrate, indicating that non-cytochrome P450 metabolism was also occurring.

95 Chapter 3

0.004 A

0.003

0.002

0.001 AEA:Peak arearatio of analyteto internal standard 0.000 0 5 10 15 20 25 30 Time (min)

No co-factor loss : 95% t1/2 : 2:13 CLi : 57.2 mL/min/g liver k: 0.54

B 0.12

0.08

0.04 OEA:Peak arearatio of analyteto internal standard 0.00 0 5 10 15 20 25 30 Time (min) No co-factor loss : 95% t1/2 : 2:28 CLi : 27.4 mL/min/g liver k: 0.26

C 0.06

0.04

0.02 PEA:Peak area ratio of analyteto internal standard 0.00 0 5 10 15 20 25 30 Time (min)

No co-factor loss : 83% t1/2 : 5:16 CLi : 10.0 mL/min/g liver k: 0.09

Figure 3.21. – Kinetics of the oxidative metabolism of AEA (A), OEA (B) and PEA (C) by rat liver microsomes. AEA, OEA and PEA (0.5 µM) were added separately to 0.5 mg/mL liver microsome suspensions and incubated at 37 OC over a 30 minute time course. Reactions were stopped by protein precipitation. FAAH substrates were not quantitated, with the peak area ratio to the internal standard instead being used. Data expressed as mean ± SD (n = 3). Where no co- factor loss is the amount oxidised by non-cytochrome P450 pathways, t 1/2 is half-life, CLi is intrinsic clearance and k is the rate constant. Intrinsic clearance was assayed as per Section

3.2.7.

96 Chapter 3

3.3.8. Protein binding of the FAAH substrates in brain and liver microsomes

The protein binding of the FAAH substrates to brain homogenate and liver microsomes was investigated using equilibrium dialysis. OEA and PEA underwent significantly more protein binding than AEA in brain (Figure 3.22.). In liver microsomes PEA underwent significantly more protein binding than AEA. In both brain and liver microsomes there was no significant difference between the protein binding of OEA and PEA. Intrinsic clearance values were corrected for protein binding by taking the percentage unbound fraction and dividing the intrinsic clearance by this value (Table 3.3.). In brain, AEA is considerably less bound to protein than the other two FAAH substrates and is thus more available for intrinsic clearance. In microsomes, all three substrates have low binding.

*** AEA 100 *** OEA PEA * 90

Protein binding 80 (% bound fraction)

70 Brain Liver microsomes

Figure 3.22. – Protein binding of AEA, OEA and PEA in rat brain and liver microsomes.

AEA, OEA and PEA (0.3 µM) were added separately to suspensions of either brain or liver microsomes on one side of a microdialysis membrane, and left to equilibrate and incubate at 37 oC

(5 hours for brain, 20 hours for microsomes). Samples were collected from both sides of the membrane and analysed. Protein binding was expressed as the percentage that is protein bound.

Data expressed as mean ± SD (n = 6). Statistical analysis by ANOVA with Tukey’s post-hoc analysis. * = p≤ 0.05 and *** = p ≤ 0.001 in comparison with AEA.

Table 3.3. – Rat liver microsomal intrinsic clearance (mL/min/g liver) of FAAH substrates corrected for protein binding.

AEA OEA PEA Intrinsic clearance 57.2 27.4 10.0

Corrected intrinsic clearance 336.5 171.2 83.3 Intrinsic clearance values were corrected for protein binding by dividing the intrinsic clearance by the percentage unbound fraction.

97 Chapter 3

3.4. Discussion

3.4.1. FAAH substrate assay

An analytical method for the measurement of AEA, OEA and PEA in brain and whole blood was established. Like previous methods, 141 the FAAH substrates ionised well under positive electrospray forming a prominent 62 m/z ethanolamine ion (Figures

3.1. – 3.4.). Despite the majority of the previously reported methods using C18 columns, here a phenyl column was employed. The substrates were less retained, allowing for fast run times (Figures 3.5. & 3.6.). The fastest AEA assay currently reported method has a 3.5 minute run time, with no resolution of the 3 substrates.277

The faster gradient used here had a run time of just 2.5 minutes, with likewise all 3 substrates co-eluting. Previous methods have also lacked resolution of the analytes, with analysis not being compromised. 278, 279 Likewise the lack of resolution in this assay did not compromise analysis, with ion suppression being within ± 15% (Table

3.2.) and low limits of quantitation (7 pmol/g in brain and 0.9 pmol/mL in blood for each analyte) demonstrating good sensitivity. This is small improvement of an earlier method that had a LOQ of 10 pmol/g. 141 The fast run time made the assay high throughput, with 24 samples being run an hour.

The previously reported methods extracted the FAAH substrates using a combination of liquid-liquid and solid phase extraction. Here deproteination by acetonitrile was investigated. Recoveries were high, within ± 10% (Table 3.2.). It was found that concentrating the samples was not a requirement as the concentrations of AEA, OEA and PEA were well above the LOQ. Therefore protein precipitation by acetonitrile was an appropriate technique that contributed to the assay being high throughput.

To ensure analyte stability a novel approach was employed. The FAAH inhibitor,

URB597, was used in the sample preparation. The addition of URB597 ensured that there FAAH activity was stopped in sample preparation. This was demonstrated to significantly increase the stability of the analytes (Figure 3.7.). As non-covalent FAAH inhibitors will readily detach from the enzyme, the addition of URB597 in the sample preparation was an important step.

98 Chapter 3

Validation of the method showed good sensitivity, good linearity, analyte stability, no matrix effects, no ion suppression, good recoveries and repeatability, all demonstrating the method to be robust. This assay was used to measure the FAAH substrates in brain and blood from in vivo studies of inflammatory pain, and also in in vitro studies of intrinsic clearance and protein binding.

3.4.2. Effect of FAAH inhibitors on brain and blood FAAH substrate levels in an in vivo inflammatory pain model

The measurement of AEA, OEA and PEA in tissue samples is increasingly being used as a pharmacodynamic parameter to assess the ability of FAAH inhibitors to increase substrate levels. 225, 227, 273, 275 The approach was even used to investigate the first

FAAH inhibitor to enter clinical trials, PF-04457845.184 Typically these studies have measured AEA, OEA and PEA. Although one study additionally measured 2-AG, to demonstrate that this endocannabinoid was not affected by FAAH inhibition. 227

Furthermore another study measured linoleoyl ethanolamide (LEA), a more recently identified substrate of FAAH. 184

Here FAAH inhibitors were screened for their ability to increase AEA, OEA and PEA levels in the brain and blood from a FCA-induced hypersensitivity rat model of inflammatory pain. In addition to causing hypersensitivity, FCA decreased FAAH activity in the blood (Figure 3.11. B), while no change was seen in the FAAH substrate levels (Figure 3.13.). Decreased FAAH activity would normally equate to higher substrate levels and analgesia, which is obviously not the effect of pro-inflammatory

FCA. It therefore appears that FAAH activity, FAAH substrate levels and cannabinoid receptor-mediated analgesia cannot always be correlated. The fact that FAAH substrate levels are not changed with inflammatory pain adds to their utility, as it shows that this will not be an interfering factor.

The first inhibitor to be investigated in this study was the well characterised FAAH inhibitor URB597, which in accordance with reported results, 267 returned weight distribution to normal (Figure 3.10.), inhibited FAAH activity (Figure 3.11.) and increased AEA, OEA and PEA levels (Figures 3.12. – 3.14.). Blood substrate levels

99 Chapter 3 correlated well with brain substrate levels, FAAH activity and analgesia. The implication of this is useful clinically as it demonstrates that blood FAAH substrates levels are an easily accessible indicator of brain FAAH activity. This in vivo observation was translated to humans in the following chapter, where FAAH activity following paracetamol was investigated by measuring plasma FAAH substrate levels.

The drug vehicle used for URB597 (2% DMSO, 65% PEG-400 and 33% H2O) significantly increased blood PEA levels (Figure 3.13. C) indicating that this was a peripheral not central response. However FAAH activity and weight distribution were not changed. Therefore FAAH substrate levels again did not correlate with FAAH activity and analgesia. Similarly the substrate levels seen with the vehicle DMA were significantly higher than those seen with the other vehicle methyl cellulose (Figure

3.15.) Consequent studies accounted for this difference by comparing substrate increases between drug vehicle and drug dose groups, an approach also used in previously reported studies. 225

The initial investigation of the substrates using the inhibitor URB597 provided important information for future experimental design. Blood was no longer collected before and after FCA as no change in substrate levels were seen. Blood was no longer taken before and after drug dose, as comparisons were instead made between groups dosed with the vehicle. As there was little difference between the FCA and control groups, only FCA pain models were subsequently used. Finally as a difference between naïve and vehicle levels was seen, future studies accounted for this by not using naïve groups and instead evaluated FAAH inhibitors on the concentration difference between the drug vehicle and the dose groups.

Various FAAH inhibitors were screened for their ability to increase substrate levels, with dose responses and CNS penetration being investigated (Table 3.4.). Both covalent inhibitors, URB597 and PF-04457845, significantly increased FAAH substrate levels. In comparison the non-covalent inhibitors and the COX-2 inhibitor celecoxib failed to increase AEA, OEA and PEA levels. Like the FAAH inhibitors, celecoxib is also beneficial in the treatment of inflammatory pain. It was used here to demonstrate the specificity of measuring AEA, OEA and PEA as indicator of analgesia via FAAH than via

100 Chapter 3

COX. None of the investigated inhibitors were capable of elevating substrate levels beyond that of URB597.

Table 3.4. – The difference in concentration between vehicle and dose groups for various FAAH inhibitors in brain and blood

AEA OEA PEA Brain Blood Brain Blood Brain Blood (pmol/g) (pmol/mL) (pmol/g) (pmol/mL) (pmol/g) (pmol/mL) URB597 70 4 1831 106 2698 52 (10 mg/kg) PF-04457845 0 3.3 310 14 710 18 (3 mg/kg) NC1 2 0.7 -73 3 -334 1 (10 mg/kg) NC2 13 0.2 136 5 63 5 (10 mg/kg) NC3 -5 2 -56 7.3 -90 10 (10 mg/kg) Celecoxib -1 0 -46 3 -99 3 (10 mg/kg) Red values indicate non-significant concentration differences as determined by ANOVA with

Tukey’s post-hoc analysis.

Significant differences in the FAAH substrate levels were seen in the control groups between experiments. For example the variations in the control groups for brain OEA levels ranged from 500 – 1300 pmol/g (Figure 3.15. B & 3.12. B),. The baseline levels of brain OEA have previously been reported as approximately 200 pmol/mL respectively.225 Therefore higher concentrations than previously reported have been shown here. Quality controls would have ruled out whether that these variations were due to analytical variation, although unfortunately they were not used. The FAAH substrates in rats have been shown to be influenced by diurnal variation, with substrate concentrations increasing up to 6-fold in the presence of darkness. 363 This is a possible explanation for the variation in control levels seen between the experiments. To overcome the observed variation inhibitors were compared on the concentration difference between vehicle and dose groups, thus cancelling out the variation in baseline levels.

In summary, both covalent and non-covalent FAAH inhibitors were screened for their ability to increase brain and blood levels of AEA, OEA and PEA. Significant FAAH

101 Chapter 3 substrate increases were seen with covalent inhibitors, whereas no change in substrate levels was observed for non-covalent inhibitors.

3.4.3. Oxidative metabolism and protein binding of AEA, OEA and PEA

Oxidative metabolites of AEA have been identified in vitro using high AEA concentrations incubated on either pure enzymes or liver microsomes. 177 These metabolites have not yet been shown to be produced physiologically under normal conditions,179 but have been identified in FAAH knock outs injected with AEA. 181 The synthesis of such metabolites is only likely to be in low concentrations in tissue where there are high AEA levels and high expression of its oxidising enzymes. Such conditions are a possibility following FAAH inhibition in inflamed tissue, and are therefore worth investigating.

Here it was shown that AEA undergoes a faster rate of oxidative metabolism by liver microsomes and significantly less binding to brain protein than OEA and PEA (Figure

3.21. & 3.22.). This is the first demonstration of OEA and PEA undergoing oxidative non-FAAH metabolism. The no co-factor losses indicate that some of this metabolism is via cytochrome P450, while the majority is via other oxidising enzymes present in the liver microsomes.

Cytochrome P450 metabolises AEA to 5,6-, 8,9-, 11,12-, and 14,15- epoxyeicosatrienoic acid ethanolamide (EET-EA), of which 5,6-EET-EA is the most potent CB2 agonist. 364 The oxidative metabolites of OEA and PEA need to be determined so that it can be investigated as to whether they too have any pharmacological action. This data provides the first evidence that that OEA and PEA can be metabolised by enzymes other than FAAH, possibly cytochrome P450. To extend the understanding of their clearance the isoforms responsible would need to be determined. As oleic acid is hydroxylated by the cytochrome P450 isoform 2E1, 365 this isoform could possibly also oxidise OEA. A useful addition to this study would have been the measurement of the oxidative metabolites alongside the substrates over the time course. This would have shown the simultaneous metabolism of the substrates

102 Chapter 3 and formation of metabolites, allowing for the most abundant metabolites to be identified.

Table 3.5. – Chemical properties of the FAAH substrates alongside their rates of intrinsic clearance in liver microsomes.

Molecular Number of carbons in Number of double bonds Intrinsic clearance weight fatty acid chain in fatty acid chain (mL/min/g liver) AEA 347.5 20 4 336.5

OEA 325.5 18 1 171.2

PEA 299.5 16 0 83.3

The properties that give AEA its low protein binding probably are probably also making it susceptible to oxidative metabolism. Table 3.5. compares the molecular weights, carbon length and saturation of the FAAH substrates to the intrinsic clearance, suggesting that fatty acid length and saturation influence the metabolic stability of the

FAAH substrates.

The physiological relevance of these findings is unknown. As discussed earlier the oxidative metabolites of arachidonic acids are all well characterised, while the oxidative metabolites of AEA are yet to be detected in vivo under normal physiological conditions. The concentration of arachidonic acid in the brain is 8 µmol/g compared to

AEA being 13 – 32 pmol/g. 141, 180 Therefore any production of such AEA metabolites will only be in low amounts, and may not be of any pharmacological relevance. Given their low concentrations very sensitive assays would be required for their detection.

Interestingly the clinical trial of the FAAH inhibitor PF-04457845 showed that following dosage the plasma levels of AEA, OEA and PEA levels returned to normal 5 days before

FAAH activity returned to normal, 184 demonstrating their metabolism despite no FAAH activity Therefore when FAAH is inhibited in humans there are possibly alternative enzymatic pathways available for the metabolism of these accumulated substrates.

This clinical trial did not show any difference in the rates that these substrates were being cleared, suggesting that the finding here that OEA and PEA undergo more protein binding and less oxidative metabolism than AEA is not physiologically relevant.

In conclusion, non-FAAH metabolism of FAAH substrates probably occurs

103 Chapter 3 physiologically when FAAH is inhibited, specifically in tissues where there is high expression of oxidising enzymes, such as in the liver and inflamed tissue. However the differences in rates of oxidative metabolism and protein binding are not likely to be physiologically relevant.

3.4.4. Summary

The FAAH substrates AEA, OEA and PEA were measured as a pharmacodynamic parameter to evaluate various FAAH inhibitors in an in vivo model of inflammatory pain. These studies showed that the blood substrate levels correlate well with brain

FAAH activity and brain AEA levels. Therefore measuring these substrates in blood is a useful indicator of what is happening in the brain. Covalent FAAH inhibitors performed much better than non-covalent FAAH inhibitors, due to their irreversible inhibition of the enzyme and therefore longer pharmacodynamics. Following this the oxidative metabolism and protein binding of these substrates was investigated. OEA was the most abundant substrate in brain and blood, however PEA can be measured with the most sensitivity by LC-MS/MS, it increases the most with FAAH inhibition, it undergoes the least oxidative metabolism and it binds the most to protein. Therefore PEA is probably the optimum substrate to measure peripherally in plasma as an indicator of

FAAH activity centrally. In the following chapter the plasma FAAH substrates levels will be measured to investigate paracetamol’s affect on FAAH.

104 Chapter 4

CHAPTER 4

Pharmacokinetics and metabolism of intravenous paracetamol in the human cerebrospinal fluid and its

conjugation to fatty acids via FAAH

105 Chapter 4

4.1. Introduction

The endocannabinoids AEA is inactivated in a 2-step process, beginning with reuptake into the cell then followed by hydrolysis by FAAH. Paracetamol interferes with both steps, 164, 292 firstly using FAAH to conjugate paracetamol with arachidonic acid, forming

AM404. AM404 then inhibits AEA reuptake, blocking its entry into the cell and protecting it from inactivation. 104 As a result there is an extracellular accumulation of

AEA in the synaptic cleft, increasing activation of the cannabinoid receptors and producing analgesia. 260

Figure 4.1. – Structures of AM404, the oleic acid and palmitic acid derivatives of paracetamol. AM404 is a conjugate of arachidonic acid and 4-aminophenol that is produced in vivo in paracetamol metabolism. The oleic acid and palmitic acid conjugates of paracetamol are yet to be identified as paracetamol metabolites.

Given that an arachidonic acid conjugate of paracetamol can be formed by FAAH, it is feasible that paracetamol metabolism also forms oleic acid and palmitic acid derivatives (Figure 4.1.). The first study to show AM404 synthesis by FAAH did so by incubating 4-aminophenol with rat brain homogenate. It was demonstrated that rat brain homogenates (in 10 mM TRIS buffer containing 1 mM EDTA at 5-10 mL/g of wet tissue), incubated at 37 oC for 60 minutes with paracetamol (1 µmol/mL) synthesised

AM404 (14 pmol/g wet tissue). 164 A higher yield of AM404 (250 pmol/g) was obtained when paracetamol was replaced with 10 nmol/mL of the intermediate 4-aminophenol.

Amounts were further increased 6-fold with 100 nmol/mL arachidonic acid. This synthesis of AM404 was demonstrated to be FAAH-dependent, with the FAAH inhibitor

106 Chapter 4 phenylmethylsulfonyl fluoride (PMSF) preventing its formation. 164 Using this previously reported technique it will be investigated as to whether other fatty acids can be used by FAAH to form novel paracetamol metabolites. Based on their hypothesised structures the fragmentation and optimal mass spectrometry conditions for these potential metabolites will be extrapolated from those obtained for AM404. Finally then the synthesis of potential metabolites will be investigated in brain homogenate incubated with 4-aminophenol and either oleic or palmitic acid to determine whether these theorised metabolites are actually formed. The inhibitor URB597 will be used to show if formation of metabolites is via FAAH.

Following this initial study a clinical trial will be undertaken to comprehensively investigate paracetamol’s pharmacokinetics, metabolism and mechanism of action in both CSF and plasma following intravenous administration. The pharmacokinetics of paracetamol has yet to be studied in the CSF of adult humans. Paracetamol will be measured in both the CSF and plasma, allowing for its pharmacokinetic parameters to be derived and for the CSF:plasma ratio to be investigated. Alongside paracetamol, its metabolites will also be measured, namely its two major metabolites, paracetamol glucuronide and paracetamol sulphate, and its minor active metabolite, AM404.

Currently AM404 has only been investigated in vivo and is yet to be identified as an actual metabolite in humans. As FAAH is predominantly expressed in the brain,102

AM404 should mostly be synthesised there,164, 292 making the CSF the ideal body fluid to investigate AM404 formation in humans.

Finally the various endogenous mediators thought to be integral to the mechanism of action of paracetamol will also be measured. For COX, both PGE2 and LTB4 will be measured, the hypothesis being that COX inhibition should decrease PGE2 levels, consequently increasing the amount of arachidonic acid available for LTB4 production.

Paracetamol’s active metabolite, AM404, is an AEA reuptake inhibitor formed via FAAH.

To study the interaction with the endocannabinoid system AEA will be measured, alongside the other FAAH substrates OEA and PEA. Furthermore paracetamol is known to increase 5HT levels in the CNS and for its analgesia to be blocked by 5HT receptor antagonists. Therefore to investigate this interaction 5HT will also be measured.

107 Chapter 4

4.2. Materials and methods

4.2.1. Materials

AM404, AEA, OEA, PEA, PEA-d4, arachidonic acid and URB597 were from Cayman

Chemicals (Ann Arbor, Michigan, USA). Paracetamol, paracetamol glucuronide, paracetamol sulphate, 3-aminophenol, 4-aminophenol, indomethacin, phosphate buffered saline, oleic acid, palmitic acid, serotonin, 5-hydroxyindoleacetic acid and tryptophan were from Sigma Aldrich (Poole, UK). Ammonium acetate, potassium phosphate and EDTA were from VWR (Lutterworth, UK). Analytical grade methanol, acetonitrile, H 2O and formic acid were from Fisher Scientific (Loughborough, UK).

4.2.2. Formation and detection of FAAH formed paracetamol metabolites in rat brain homogenate

Brain homogenate was prepared from both male and female Random hooded rat brains. Brains were homogenised in a buffer containing 125 mM TRIS, 0.4 mM HEPES,

1 mM EDTA, 0.2% v/v Glycerol and 0.02% v/v Triton X-100 (pH 7.4), which was added to brain at 2 mL per gram of wet tissue using an Ultra-turrax handheld high speed disperser.

Experiments were carried out in 100 µL aliquots of homogenate incubated at 37 oC.

Reactions were started by adding 1 mM 4-aminophenol. Brain homogenates were supplemented with 1 mM of either arachidonic acid, oleic acid or palmitic acid. To stop the reaction 100 µL brain homogenate was extracted with 300 µL acetonitrile containing 100 ng/mL internal standard, PEA-d4. Samples were vortex mixed and centrifuged for 15 minutes at 1,500 g. 50 µL supernatant was transferred into a 96- well plate and diluted with 50 µL H2O.

LC-MS/MS analysis was performed on a Waters Acquity system attached to a Quattro

Premier tandem mass spectrometer. The separation was on a Waters Acquity BEH

Phenyl column (2.1 x 50 mm, 1.7 µm) maintained at 50oC. Gradient elution was between mobile phase A (95% H2O, 5% acetonitrile, 0.1% formic acid) and B (95%

108 Chapter 4 acetonitrile, 5% H2O, 0.1% formic acid). The gradient ran from 20% B, increasing linearly to 90% over 4 minutes at a flow rate of 0.5 mL/min. This was held until 4.5 minutes, then returned to 20% and allowed to re-equilibrate. The total runtime was 5 minutes. A 5 µL injection was used. Analytes were quantified by tandem mass spectrometry with positive electrospray ionisation. The analytes were detected by monitoring the transitions 396
110 m/z for AM404, 374
110 m/z for the oleic acid conjugate, 348
110 m/z for the palmitic acid conjugate and 304
62 m/z for the internal standard PEA-d4. For all analytes a dwell time of 0.15 seconds, a cone voltage of 22 V and a collision energy of 22 V were used.

4.2.3. Patients

The clinical trial was approved by the National Research Ethics Service (East London

REC 1) and the REC reference number for this study is 10/H0703/80. The ethics protocol, ethic approval and MHRA (Medicines and Healthcare products Regulatory

Agency) approval are included in the Appendix. The study was designed and carried out in conjunction with the Pain and Anaesthesia Research Centre at St Bartholomew's

Hospital, London.

The criteria for recruitment were that patients were between 18 – 80 years old and had no known blood clotting abnormalities. 30 patients were recruited from whom full written informed consent was obtained. Each patient underwent elective urological surgery under spinal anaesthesia, receiving intravenous paracetamol as a routine part of their treatment for post-operative pain. In their participation in this trial no additional risks to the patient were foreseen. The intravenous paracetamol preparation was obtained from Fresenius Kabi (Cheshire, UK). Paracetamol was formulated with

H2O, cysteine and mannitol. 30 patients received 1 g paracetamol (6.6 mmoles) intravenously as a single bolus dose. The clinical aspect to this study was carried out by an anaesthetist.

109 Chapter 4

4.2.4. Sample collection

The 30 patients were randomised into 3 groups to determine at what time their CSF and blood was to be collected. 5 mL of CSF and 10 mL of blood was then collected from each patient post dose at one of the three following time points: 15, 30 and 120 minutes. Blood was collected between 2 - 8 minutes after the CSF. CSF was collected prior to the administration of the anaesthetic into EDTA blood vacutainers containing

17.9 mg of the non-selective COX inhibitor indomethacin, so that after the addition of

5 mL CSF the final concentration of indomethacin was 10 µM. Blood was collected into

EDTA blood vacutainers containing 35.8 mg indomethacin, so that after the addition of

10 mL blood the final concentration was 10 µM. The purpose of indomethacin was to prevent further formation of prostaglandins after collection, thus ruling out any interference with the PGE2 assay. CSF and blood was centrifuged at 1,400 g at 4 oC for

15 minutes using a MSE Mistral 3000i. CSF and plasma samples were frozen and stored at -20oC until analysis.

4.2.5. Protein assay

CSF protein concentrations were determined by a bicinchoninic acid (BCA) assay kit

(Thermo Scientific, Runcorn, UK) that was used according to the manufacturer’s instructions.

4.2.6. Paracetamol assay

For CSF 100µL was diluted with 100 µL mobile phase containing the internal standard.

For plasma 100 µL was precipitated with 300 µL methanol containing the internal standard. 20 µg/mL 3-aminophenol was used as an internal standard as suggested by a previously reported assay. 366 Samples were centrifuged at 13,000 g for 10 minutes using a Heraeus Biofuge 15. 100 µL supernatant was diluted with 100 µL mobile phase.

Sample extracts were injected (20 µL for CSF and 10 µL for plasma) and separated on a Thermo Scientific C18 column (100 x 4.6mm, 5µm). For CSF, the column was eluted

110 Chapter 4 with 50 mM ammonium acetate (pH 5.5) with 1 mM EDTA and 5% methanol using a

Jasco PU-980 pumping at 1 mL/min. For plasma, 50 mM potassium dihydrogen phosphate (pH 6.5) with 1 mM EDTA and 5% methanol was used as the mobile phase.

Analytes were detected by a Jasco UV-975 UV detector set at a wavelength of 242 nm, the same wavelength used in a previously reported assay.366 Jasco Chrompass was used for control of system operation and data collection. A 6-point calibration line was run between 0 µg/mL – 250 µg/mL for paracetamol and between 0 µg/mL – 100

µg/mL for paracetamol glucuronide and paracetamol sulphate. The standards used in the calibration line were prepared as µg/mL solutions. After quantitation the determined sample concentrations were converted to nmol/mL. For CSF the calibration line was prepared in PBS, for plasma it was prepared in blank plasma. All CSF samples were run on the same day, as were the plasma samples.

4.2.7. AM404 and FAAH substrate assay

For CSF 100 µL of sample was diluted with 100 µL H2O. For plasma 100 µL of sample was precipitated with 300 µL acetonitrile containing 100 ng/mL internal standard PEA- d4. Samples were centrifuged at 13,000 g for 10 minutes using a Heraeus Biofuge 15, followed by dilution of 100 µL supernatant with 100 µL H 2O.

LC-MS/MS analysis was performed on a Thermo Scientific Accela system attached to a

Thermo Scientific TSQ Vantage mass spectrometer. The separation was on a Waters

Acquity BEH phenyl column (2.1 x 50mm, 1.7µm) maintained at 50 oC. 10 µL of sample was injected onto the column. Gradient elution was between mobile phase A (95%

H2O, 5% methanol, 0.1% formic acid) and B (95% methanol, 5% H 2O, 0.1% formic acid). The gradient ran from 5% B, increasing linearly to 95% over 3 minutes at a flow rate of 0.5 mL/min. This was held until 5 minutes, then returned to 5% and allowed to re-equilibrate. AM404, AEA, OEA and PEA were detected under positive electrospray ionisation followed by the monitoring MRM transitions in Table 4.1.

The collision gas was argon, spray voltage was 3500 V, vaporiser temperature was

300 oC, sheath gas pressure was 60 psi, capillary temperature was 270 oC, collision cell pressure was 1.5 mTorr and no declustering voltage was applied.

111 Chapter 4

Table 4.1. – The selected precursor and product m/z values for AM404 and the FAAH substrates with their dwell and collision energy used for their identification and quantification by LC-MS/MS.

Precursor ( m/z ) Product ( m/z ) Dwell (s) Collision energy (V)

AM404 396.3 110.2 0.10 19

AEA 348.3 62.2 0.10 16

OEA 326.4 62.2 0.10 16

PEA 300.3 62.2 0.10 15

PEA-d4 303.4 62.2 0.10 15

Where m/z is the mass-to-charge ratio, s is seconds and V is voltage.

A 6-point calibration line 0 ng/mL – 300 ng/mL was prepared, in either blank CSF or blank plasma. Calibration lines were prepared for both CSF and plasma. For CSF, concentrations were determined by external standard quantitation. An internal standard was not used due to the insolubility of lipids in aqueous solutions. For plasma, quantitation was by peak area ratio of analyte to the internal standard PEA- d4.

4.2.8. Prostaglandin E2 assay

PGE2 was extracted from CSF and plasma in accordance to a previously reported method 367 that has since been further optimised.288, 368 C18 cartridges (Waters,

Elstree, UK) were conditioned with 4 mL ethanol then 4 mL H 2O. 1 mL of either CSF or plasma was applied to the cartridge, which was then washed with 4 mL H 2O followed by 4 mL 15% methanol in H 2O (adjusted to pH 3 with HCl). PGE2 was eluted with 2 mL ethyl acetate, which was evaporated using a speed vac and then reconstituted in 200

µL sodium phosphate buffer (pH 7) with 1% casein for analysis. PGE2 was measured using a Cisbio kit (Bagnols-sur-Ceze, France) in accordance with the manufacturer’s instructions (see Section 2.2.3.). An 8-point calibration line was run in triplicate from 0 pg/mL – 5000 pg/mL PGE2. The plate was read using a Berthold Technologies

(Harpenden, UK) Mithras LB 940 FRET plate reader with FRET being measured and expressed as a ratio of 665 nm and 620 nm. The determined final concentrations were corrected to account for the sample preparation concentrating the sample 5-fold.

112 Chapter 4

4.2.9. Leukotriene B4 assay

CSF and plasma was extracted in the same way as for PGE2. LTB4 was quantified using a kit obtained from Cisbio (Bagnols-sur-Ceze, France), which uses the same principle as the PGE2 assay. The LTB4 present in the samples competes with XL665- labelled LTB4 for binding to an anti-LTB4 cryptate-labelled antibody. The fluorescent signal is inversely proportional to analyte concentration with the more LTB4 in the sample the lower the fluorescent signal. An 11-point calibration line was run in triplicate from 0 – 5 µg/mL LTB4. All sample concentrations fell within the range of the calibration line. The plate was read using a Berthold Technologies (Harpenden, UK)

Mithras LB 940 FRET plate reader with FRET being measured and expressed as a ratio of 665 nm and 620 nm. The determined final concentrations were corrected to account for the sample preparation concentrating the sample 5-fold.

4.2.10. 5HT assay

CSF was diluted 1:1 with 100 mM ammonium acetate (pH 5.5) containing 1 mM EDTA.

50 µL was injected and separated on a Thermo Scientific C18 column (100 x 4.6mm,

5µm) eluted with 100 mM ammonium acetate (pH 5.5) with 1 mM EDTA and 6% methanol, by a Jasco PU-980 pumping at 1 mL/min. 5HT, tryptophan and 5HIAA were monitored using a Jasco FP-920 fluorescent detector at excitation: 290nm and emission: 335nm. Jasco Chrompass was used for control of system operation and data collection. An 8-point calibration line was run between 0 µmol/mL – 100 µmol/mL.

4.2.11. Pharmacokinetic calculations

Pharmacokinetic parameters were determined using standard equations. 369 Transfer in and out of the CSF was assumed to be by passive diffusion and that the absorption into the CSF was thus by an exponential process. CSF concentration-time curves of paracetamol were fitted to a simple pharmacokinetic model to describe its absorption into the CSF. Cmax , t max and t 1/2 were derived directly from the graph, where c max is the highest observed drug concentration, tmax is the time at which the c max is reached and t1/2 is the time at which the drug concentration is half that of the c max . As a delay was

113 Chapter 4 expected in paracetamol’s penetration into the CSF following administration due to its crossing of the blood-brain barrier, an initial time lag parameter (t lag ) was incorporated into the model to account for this. The following pharmacokinetic model was used:

− − = − ka(t tlag ) Ct C0 1( e )

where C t is the paracetamol concentration in the CSF at time t, C0 is the extrapolated initial paracetamol concentration at time t lag , ka is the rate constant of paracetamol absorption into the CSF and t lag is the time lapse between administration and first appearance of paracetamol in the CSF.

Plasma concentration-time curves of paracetamol were fitted to an exponential model using the following equation:

= −λ.t Ct C0.e

where Ct is the paracetamol concentration in the plasma at time t, C0 is the extrapolated initial paracetamol concentration at time t 0 and λ is the rate constant.

Cmax , tmax and t1/2 were derived directly from the graph.

Area under the curve (AUC), the measure of the drug’s plasma exposure over a time period, was calculated using the following equation:

C AUC = 0 0−∞ λ

where AUC 0-∞ is the AUC extrapolated to infinity, C0 is the extrapolated initial paracetamol concentration at t 0 in nmol/mL and λ is the rate constant, which in this case is the exponential function derived from exponential model fitted to the graph to describe the pharmacokinetics.

Plasma clearance (CL), the measure of drug removal from plasma, was calculated using the following equation:

Dose CL = AUC0−∞

where dose is the amount of drug administered in moles and AUC 0-∞ is the AUC extrapolated to infinity.

114 Chapter 4

Volume of distribution (VD), the apparent volume of body fluid at which the drug distributes itself in at equilibrium, was calculated using the following equation:

CL V = D λ where CL is clearance and λ is the rate constant.

Pharmacokinetic parameters were calculated in Microsoft Excel 2003. Concentration- time curves were fitted using the Excel Solver tool, with R 2 being used to express correlation.

115 Chapter 4

4.3. Results

4.3.1. Detection of fatty acid paracetamol conjugates in brain homogenate

As oleic acid and palmitic acid conjugates of paracetamol were not available, their structures, their precursor and product ions and their optimal MS/MS conditions and liquid chromatography were all deduced using knowledge of the structure and chemistry of AM404 as explained below.

AM404 has a molecular weight of 395.5 and is comprised of arachidonic acid, which has a molecular weight of 304.5 and 4-aminophenol, which has a molecular weight of

109.1. To get to the final weight of 395.5, arachidonic acid loses an –OH - ion (17 m/z )

+ and 4-aminophenol loses a –H ion (1 m/z ), in a condensation reaction forming H 2O.

Arachidonic acid + 4-aminophenol  AM404

304.5 -17 + 109 -1 = 395.5

Following this it can be assumed that as oleic acid and palmitic acid have molecular weights of 282.5 and 256.6, that the molecular weights of their conjugates are 373.5 and 347.5 respectively.

Oleic acid + 4-aminophenol  Oleic acid conjugate

282.5 -17 + 109 -1 = 373.5

Palmitic acid + 4-aminophenol  Palmitic acid conjugate

256.6 -17 + 109 -1 = 347.6

From these calculations the molecular weights of oleic acid and palmitic acid conjugates of paracetamol were determined to be 373.5 and 347.6 respectively. The optimal MS/MS conditions for these compounds were extrapolated from those of

AM404 (Table 4.2.). Under direct-infusion positive electrospray ionisation, AM404’s precursor ion was 396.262 m/z , which formed a 109.931 m/z fragment corresponding to 4-aminophenol (Figure 4.2.). It was assumed that the other conjugates would also

116 Chapter 4 form this same fragment ion, therefore the product ion, dwell time, cone voltage and collision energy were kept the same.

Figure 4.2. - Product ion mass spectrum and corresponding structure for AM404 (396.3

109.9 m/z ) following collision induced dissociation by positive electrospray ionisation and fragmentation by MS/MS. Site of fragmentation is identified on the figure.

MS/MS conditions described in Table 4.2.

Table 4.2. – The precursor and product m/z value for AM404 and extrapolated values for the oleic acid conjugate and palmitic acid conjugate of 4-aminophenol with the dwell and collision energy used for their identification and quantification by LC-MS/MS.

Precursor Product Dwell Cone voltage Collision energy (m/z ) (m/z ) (s) (V) (V) AM404 396.262 109.931 0.15 22 22

Oleic acid conjugate 374.262 109.931 0.15 22 22

Palmitic acid conjugate 348.262 109.931 0.15 22 22 Where m/z is the mass-to-charge ratio, s is seconds and V is voltage.

Given that AM404 is essentially a structural analogue of AEA the same LC conditions detailed in Section 3.2.5. were used. The gradient time was extended from 2.5 minutes to 5 minutes, which is more suitable for metabolite profiling. To determine if there were endogenous compounds that shared the same MRM transitions as AM404 and the hypothesised fatty acid conjugates of paracetamol control rat brain homogenate was analysed. Rat brains were homogenised as described in Section

117 Chapter 4

3.2.4. With the exception of the oleic acid conjugate where a few minor peaks were detected, no other peaks were found at the other MRM transitions (Figure 4.3.). The

LC conditions were therefore suitable for the detection of these compounds.

A B

C

Figure 4.3. – LC-MS/MS chromatograms of control rat brain demonstrating that (A)

AM404, (B) oleic acid conjugated paracetamol and (C) palmitic acid conjugated paracetamol are not endogenous. Samples were analysed using the methodology described in

Section 4.2.2.

Brain homogenates were prepared for analysis by deproteination with acetonitrile. 100

µL rat brain homogenate was extracted with 300 µL acetonitrile containing 100 ng/mL internal standard, PEA-d4. To ensure that this was a suitable approach, recovery and ion suppression of AM404 was investigated. Recovery was 109 ± 4% and ion suppression was 18 ± 6%. Recovery and ion suppression were well within a ± 20% limit, concluding that this approach was well suited for these analytes.

118 Chapter 4

4.3.2. Formation of fatty acid paracetamol conjugates by FAAH in rat brain homogenate

It was initially verified as to whether AM404 could be synthesized ex vivo in rat brain homogenates as previously reported.164 Homogenates were incubated at 37 oC with 1

µmol/mL 4-aminophenol and 1 µmol/mL arachidonic acid, with formation of AM404 being investigated after 60 minutes (Figure 4.4.). A peak at 3.13 minutes at the 396.2

-> 109.9 m/z MRM transition confirmed the formation of AM404. URB597 prevented

AM404 formation, proving FAAH-dependent synthesis.

A

B

C

Figure 4.4. – LC-MS/MS chromatograms demonstrating arachidonic acid conjugation to

4-aminophenol in rat brain homogenates at (A) 0 minutes, (B) 60 minutes and at (C)

60 minutes in the presence of URB597. Reactions were started with the addition of 4- aminophenol (1 µmol/mL). Brain homogenates were supplemented with arachidonic acid (1

µmol/mL). For the URB597 (100 nmol/mL) treated homogenates the inhibitor was added to the buffer before control brain was homogenised, so that FAAH was already inactive when 4- aminophenol was added. Reactions were stopped by addition of acetonitrile. Samples were analysed using the methodology described in Section 4.2.2., with the 396.3
109.9 m/z transition being monitored for AM404.

119 Chapter 4

Having demonstrated the synthesis of AM404 in brain tissue homogenate, the conjugation of other fatty acids to paracetamol was investigated (Figures 4.5. and

4.6.). Brain homogenates were incubated at 37 oC with 1 µmol/mL 4-aminophenol and either 1 µmol/mL oleic acid or palmitic acid. Peaks appeared at 3.11 minutes for the oleic acid conjugate and at 3.00 minutes for the palmitic acid conjugate. URB597 prevented formation, proving a FAAH-dependent route of synthesis.

A

B

C

Figure 4.5. – LC-MS/MS chromatograms demonstrating oleic acid conjugation to 4- aminophenol in rat brain homogenates at (A) 0 minutes, (B) 60 minutes and at 60 minutes in the presence of (C) URB597. Reactions was started with the addition of 4- aminophenol (1 µmol/mL). Brain homogenates were supplemented with oleic acid (1 µmol/mL).

For the URB597 (100 nmol/mL) treated homogenates the inhibitor was added to the buffer before control brain was homogenised, so that FAAH was already inactive when 4-aminophenol was added. Reactions were stopped by addition of acetonitrile. Samples were analysed using the methodology described in Section 4.2.2., with the 374.3
109.9 m/z transition being monitored for the oleic acid paracetamol conjugate.

120 Chapter 4

A

B

C

Figure 4.6. – LC-MS/MS chromatograms demonstrating palmitic acid conjugation to 4- aminophenol in rat brain homogenates at (A) 0 minutes, (B) 60 minutes and at 60 minutes in the presence of (C) URB597. Reactions were started with the addition of 4- aminophenol (1 µmol/mL). Brain homogenates were supplemented with palmitic acid (1

µmol/mL). For the URB597 (100 nmol/mL) treated homogenates the inhibitor was added to the buffer before control brain was homogenised, so that FAAH was inactive when 4-aminophenol was added. Reactions were stopped by addition of acetonitrile. Samples were analysed using the methodology described in Section 4.2.2., with the 348.3
109.9 m/z transition being monitored for the palmitic acid paracetamol conjugate.

The formation of these conjugates, as well as AM404, was monitored over the course of an hour (Figure 4.7.). URB597 was used to demonstrate whether the formation of these conjugates was FAAH-dependent. The syntheses of the conjugates were all time- dependent. AM404 synthesis appeared to peak after 30 minutes. The synthesis of the oleic acid conjugate peaked at 40 minutes and then for an unknown reason depleted.

No conjugates could be detected in homogenate incubated with URB597, concluding that their formation was FAAH-dependent. This study showed for the first time that as well as arachidonic acid, FAAH can also conjugate oleic acid and palmitic acid to 4- aminophenol.

121 Chapter 4

1.6 A

1.2

- URB597 0.8 + URB597 AM404 AM404

0.4

(Peak area ratio area to (Peak of internalanalyte standard) 0.0 0 10 20 30 40 50 60 Time (min)

0.007 B 0.006

0.005

0.004

0.003

Oleic acid conjugate 0.002

0.001

(Peak area ratio(Peak area of to analyte internal standard) 0.000 0 10 20 30 40 50 60 Time (min)

0.005 C

0.004

0.003

0.002 Palmitic acid conjugate 0.001

(Peak area ratio of to (Peak area analyte internal standard) 0.000 0 10 20 30 40 50 60 Time (min)

Figure 4.7. – Time course of the conjugation of (A) arachidonic acid, (B) oleic acid and

(C) palmitic acid to paracetamol in rat brain homogenates. The presence of 4-aminophenol

(10 µM) caused time-dependent production of fatty acid conjugates of paracetamol in brain homogenates (filled circles) (n = 3). No conjugates were detected in brain homogenate treated with the FAAH inhibitor URB597 (100 µM) (open circles) (n = 3). Samples were analysed using the methodology described in Section 4.2.2. Data are presented as mean ± SD.

122 Chapter 4

4.3.3. Collection of human CSF and plasma samples

The clinical trial lasted 71 days. 30 patients were recruited, 2 of which self- administered paracetamol within 24 hours of participating in the study and were consequently excluded. Of the remaining 28 patients only 2 were female, and were also excluded for uniformity. The 26 remaining patients were all adult males between

34 - 78 years of age, with an average age of 65 ± 12. Their weights ranged from 52 –

125 kg, with the average of 83 ± 18 kg. Patients were split into 3 groups, with the collection of CSF and blood being either 15, 30 or 120 minutes post paracetamol dose.

Given the inevitable unpredictability of operating theatre schedules precise timing was not possible. Instead of fitting in three discrete groups, the data was continuous with the time of CSF collection following paracetamol ranging from 10 – 206 minutes and for blood ranging from 15 – 211 minutes.

4.3.4. Measurement of total protein in CSF

The average protein concentration in CSF was determined by the bicinchoninic acid assay to be 701 ± 103 µg/mL (n = 26). The low amount of protein present in CSF meant that it could simply be diluted 1:1 with an appropriate buffer to be ready for analysis.

4.3.5. Preliminary investigation of a single method for quantitation of paracetamol and its major and minor metabolites

It was initially investigated as to whether paracetamol, paracetamol glucuronide, paracetamol sulphate and AM404 could all be assayed by a single method. MS/MS readily ionised and fragmented AM404 under positive electrospray. However paracetamol and paracetamol sulphate were only weakly ionisable under positive electrospray and the glucuronide could only be ionised under negative electrospray.

Following this LC-UV was investigated as an alternative analytical approach, making use of the UV absorbance of paracetamol’s aminophenol group. Given its hydrophobicity, AM404 eluted after paracetamol. However there was not adequate sensitivity, with the LOD for AM404 being 43 nmol/mL. Separate methodologies were

123 Chapter 4 therefore required, with paracetamol and its major metabolites being measured by LC-

UV (Section 4.2.6.) and its minor metabolite AM404 being measured by LC-MS/MS

(Section 4.2.7.).

4.3.6. Paracetamol assay by LC-UV - optimisation

Different mobile phases were used for the analysis of CSF and plasma. Endogenous components in CSF eluted near paracetamol glucuronide (Figure 4.8.). These were resolved with 50 mM ammonium acetate (pH 5.5) as opposed to the 50 mM potassium phosphate buffer (pH 6.5) used for plasma. The paracetamol glucuronide peak in the plasma appears to be two unresolved peaks (Figure 4.9.). Comparison of spiked plasma with control plasma showed that this was in fact paracetamol glucuronide.

Figure 4.8. – Overlayed LC-UV chromatograms of an extracted CSF sample (red) and a standard (black) showing paracetamol and its metabolites. Peaks were identified by comparison of an extracted CSF sample post 1 g i.v. paracetamol with a standard of 83 µg/mL paracetamol, 33 µg/mL paracetamol glucuronide and 33 µg/mL paracetamol sulphate prepared in

PBS run using the methodology as described in Section 4.2.6. The internal standard, 3- aminophenol, is present in both at 20 µg/mL.

124 Chapter 4

Figure 4.9. – Overlayed LC-UV chromatograms of an extracted plasma sample (red) and a standard (black) showing paracetamol and its metabolites. Peaks were identified by comparing an extracted CSF sample post 1 g i.v. paracetamol with a standard of 100 µg/mL paracetamol, 83 µg/mL paracetamol glucuronide and 83 µg/mL paracetamol sulphate prepared in control run using the methodology as described in Section 4.2.6. The internal standard, 3- aminophenol, is present in both at 20 µg/mL.

4.3.7. Paracetamol assay - validation

The measurement of paracetamol and its major metabolites was by LC-UV with separation being on a C18 column. Before analysing clinical samples the assay for paracetamol and its major metabolites was validated (Table 4.3.) following the FDA approved guidance and principles of validation. 359

The linearity was very good, with all calibration lines having R 2 values above 0.99

(Figure 4.10.). The limit of detection (LOD) and limit of quantification (LOQ) were taken as the concentration at which the signal:noise ratio was 3 for LOD and 10 for

LOQ. With the exception of paracetamol sulphate in some CSF samples, all concentrations were above the LOQ.

125 Chapter 4

6 A 5 y = 0.0238x y = 0.0443x R2 = 1 R2 = 1 4 y = 0.0369x R2 = 0.9999

3 Paracetemol Glucuronide 2 Sulphate

1 Peak Peak ratioarea of to analyte internal standard 0 0 50 100 150 200 250 Concentration (µg/mL PBS)

1.2 B y = 0.0042x R2 = 0.9983 1.0

0.8 y = 0.0075x 2 R = 0.9990 y = 0.006x R2 = 0.9988 Paracetamol 0.6 Glucuronide Sulphate

0.4

0.2 Peak area Peak ratio area of to analyte internal standard 0.0 0 50 100 150 200 250 Concentration (µg/mL plasma)

Figure 4.10. – Calibration lines for paracetamol, paracetamol glucuronide and paracetamol sulphate prepared in (A) PBS and in (B) plasma measured by LC-UV.

Standards were analysed using the methodology as described in Section 4.2.6.

The intra-injection repeatability of the described method was determined. This was expressed as the coefficient of variation and was below 2% for all analytes. Inter- assay reproducibility was not investigated as all samples were analysed on the same day. Recovery was calculated by comparing samples spiked with standards so that the final concentration was 100 µg/mL for paracetamol and 83 µg/mL for it metabolites.

Recovery was expressed as a percentage. Recoveries fell within a ± 10% range (n =

4). Validation of the method proved that it was robust enough for the analysis of the clinical samples.

126 Chapter 4

Table 4.3. – Summary of validation parameters for paracetamol, paracetamol glucuronide and paracetamol sulphate assay by LC-UV.

Paracetamol Paracetamol Paracetamol 3-aminophenol Glucuronide Sulphate CSF LOD 54 pmol/mL 2 pmol/mL 9 pmol/mL LOQ 180 pmol/mL 6 pmol/mL 30 pmol/mL Intra-injection 0.5% 0.6% 0.6% 0.6% repeatability Recovery 99 ± 0.1% 93 ± 1% 103 ± 0.5%

Plasma LOD 0.2 nmol/mL 0.9 nmol/mL 1.2 nmol/mL LOQ 0.6 nmol/mL 2.4 nmol/mL 3.9 nmol/mL Intra-injection 0.9% 1.1% 1.2% 1.9% repeatability Recovery 92 ± 1% 87 ± 1% 93 ± 2% LOD and LOQ defined as signal:ratio = 3 and 10 respectively. For intra-injection repeatability, a single extracted sample was repeatedly assayed and expressed as the coefficient of variation

(n=10). Recovery expressed as the mean percentage ± SD (n = 6). Values were calculated using peak areas. Concentrations were 100 µg/mL for paracetamol, and 83 µg/mL for paracetamol glucuronide and paracetamol sulphate. Samples were extracted and analysed as described in

Section 4.2.6.

4.3.8. AM404 and FAAH substrate assay - validation

The measurement of AM404 and the FAAH substrates was by LC-MS/MS with separation on a phenyl column (Figure 4.11.). The assay for AM404, AEA, OEA and

PEA was based on the FAAH substrate assay described in Section 3.2.5. and the

AM404 assay described in Section 4.2.2.

127 Chapter 4

A

B

C

Figure 4.11. – Typical LC-MS/MS chromatograms of AM404 in CSF post 1 g intravenous

(A) paracetamol, (B) 75 pmol/mL AM404 standard prepared in blank CSF and (C) blank

CSF. Blank CSF was collected from a patient not dosed with paracetamol. Samples were analysed using the methodology described in Section 4.2.7. with the 396.3
110.2 m/z transition being monitored for AM404.

128 Chapter 4

The assay was validated before use (Table 4.4.) following the FDA approved guidance and principles of validation. 359 Linearity was good, with all calibration lines having R 2 values above 0.97 (Figure 4.12.). The limit of detection (LOD) and limit of quantification (LOQ) were taken as the concentration at which the signal:noise ratio was 3 for LOD and 10 for LOQ.

14000000 A 12000000

10000000

8000000 y = 39810x - 374891 2 6000000 R = 0.9779 Peak Peak area

4000000

2000000

0 0 50 100 150 200 250 300 AM404 (ng/mL CSF)

0.4 B

0.3

y = 0.0012x R2 = 0.9906 0.2

0.1 Peak area Peak ratio area of to analyte internal standard 0.0 0 50 100 150 200 250 300 AM404 (ng/mL plasma)

Figure 4.12. – Calibration lines of AM404 in (A) CSF and (B) plasma. Standards were prepared in either control CSF or plasma, and were analysed using the methodology described in

Section 4.2.7. with the 396.3
110.2 m/z transition being monitored for AM404.

The intra-injection repeatability of the described method was determined. This was expressed as the coefficient of variation and was below 10% for all analytes. Inter- assay reproducibility was not investigated as all samples were analysed on the same day. Recovery was calculated by comparing a CSF sample spiked with standards so that the final concentration was 100 ng/mL. Recovery was expressed as a percentage.

129 Chapter 4

In the case of the FAAH substrates endogenous amounts were corrected for.

Recoveries fell within a ± 10% range (n = 4). Validation of the method proved that it was robust enough for the analysis of the clinical samples.

Table 4.4. – Summary of validation parameters for the AM404 and FAAH substrate assay by LC-MS/MS

AM404 AEA OEA PEA PEA-d4 (IS) CSF LOD 0.1 pmol/mL 0.2 pmol/mL 0.7 pmol/mL 0.4 pmol/mL LOQ 0.5 pmol/mL 0.5 pmol/mL 2.3 pmol/mL 1.3 pmol/mL Intra-injection 4% 4% 8% 4% 8% repeatability Recovery 95 ± 8% 93 ± 9% 93 ± 7% 91± 9%

Plasma LOD 1.6 pmol/mL 2.0 pmol/mL 2.7 pmol/mL 2.0 pmol/mL LOQ 5.0 pmol/mL 7.0 pmol/mL 9.0 pmol/mL 7.0 pmol/mL Intra-injection 9% 9% 2% 4% 7% repeatability Recovery 101 ± 5% 100± 6% 101± 5% 91± 5% LOD and LOQ defined as signal:ratio = 3 and 10 respectively. For intra-injection repeatability, a single extracted sample was repeatedly assayed and expressed as the coefficient of variation

(n=10). Recovery expressed as the mean percentage ± SD (n = 4). Values were calculated using peak areas. Concentrations of 100 ng/mL were used for each analyte. Samples were extracted and analysed as described in Section 4.2.7.

4.3.9. 5HT assay - validation

The measurement of 5HT, its precursor and its metabolite, was by LC-fluorescence with separation on a C18 column. The assay for 5HT, 5HIAA and tryptophan was validated before use (Table 4.5.) following the FDA approved guidance and principles of validation. 359 The 3 analytes were all chromatographically resolved (Figure 4.13.).

Linearity was very good, with all calibration lines having R 2 values above 0.99 (Figure

4.14.). The limit of detection (LOD) and limit of quantification (LOQ) were taken as the concentration at which the signal:noise ratio was 3 for LOD and 10 for LOQ. All determined concentrations were above the LOQ. The intra-injection repeatability of the described method was determined. This was expressed as the coefficient of variation and was below 5% for all analytes. Inter-assay reproducibility was not investigated as all samples were analysed on the same day. Recovery was calculated by comparing a spiked sample with a standard, correcting for endogenous amounts and then

130 Chapter 4 expressing the final value as a percentage. Recoveries fell within a ± 10% range (n =

6). Validation of the method proved that it was robust enough for the analysis of the clinical samples.

Figure 4.13. – A typical LC-fluorescence chromatogram of a CSF sample post 1 g intravenous paracetamol showing 5HIAA, 5HT and tryptophan. Samples were analysed using the methodology as described in Section 4.2.10.

6000 y = 58.377x R2 = 0.9999

5000 y = 41.008x R2 = 0.9999

4000 5HIAA 5HT 3000 TRP Peak area Peak y = 13.98x 2000 R2 = 0.9999

1000

0 0 20 40 60 80 100 Concentration (nmol/mL)

Figure 4.14. – Calibration line of 5HIAA, 5HT and TRP prepared in H 2O. Standards were analysed using the methodology as described in Section 4.2.10.

131 Chapter 4

Table 4.5. – Summary of validation parameters for the 5HT assay by LC-fluorescence

5HT 5HIAA Tryptophan LOD 0.01 nmol/mL 0.03 nmol/mL 0.15 nmol/mL LOQ 0.04 nmol/mL 0.1 nmol/mL 0.5 nmol/mL Intra-injection 2% 3% 1% repeatability Recovery 96 ± 8% 103 ± 7% 106 ± 5% LOD and LOQ defined as signal:ratio = 3 and 10 respectively. For intra-injection repeatability, a single extracted sample was repeatedly assayed and expressed as the coefficient of variation

(n=10). Recovery expressed as the mean percentage ± SD (n = 4). Values were calculated using peak areas. Concentrations of 5 nmol/mL were used for each analyte. Samples were extracted and analysed as described in Section 4.2.10.

4.3.10. Pharmacokinetics of paracetamol in the cerebrospinal fluid and plasma following intravenous administration

Following intravenous administration paracetamol was detectable in the CSF as soon as 10 minutes at 5 nmol/mL and increased over time (Figure 4.15.). No clearance of paracetamol from the CSF was observed over the time course. Therefore the pharmacokinetic model fitted to the CSF data could only be used to describe absorption. The model showed a significant trend in the accumulation of paracetamol

2 in the CSF (R = 0.73). The tlag was 5 minutes. After 120 minutes the CSF paracetamol concentration appeared to plateau with a Cmax of 76 nmol/mL. The time taken to accumulate to half this concentration, 38 nmol/mL, was approximately 43 minutes.

The time course was not long enough to observe clearance of paracetamol from the

CSF therefore the t1/2 after the Cmax could not be calculated. For this same reason CL and AUC could also not be calculated. VD was determined to be 87 L, this was then corrected for the average patient body weight (83 kg), to give a volume of distribution of 1.05 L/kg (Table 4.6.).

The plasma pharmacokinetics of paracetamol was described using an exponential fit with a trend seen in its plasma clearance over time (R 2 = 0.61). The peak plasma concentration of paracetamol was immediately following intravenous administration.

Extrapolating back to t 0 determined the Cmax to be 377 nmol/mL. This then decreased over the time course with a t1/2 of 100 minutes. AUC 0-∞ was calculated to be 53,818 nmol.min/mL, CL was 122 mL/min and V D was 17.5 L. VD was then corrected for the average patient body weight (83 kg), to give a volume of distribution of 0.2 L/kg. At

132 Chapter 4 the end of the time course the concentration of paracetamol was approximately 100 nmol/mL. The CSF:plasma ratio of paracetamol was shown to increase linearly over the time course peaking at 0.6 (R 2 = 0.85).

100 A

80

60 R2 = 0.7338

40

Paracetamol (nmol/mL CSF) 20

0 0 60 120 180 Time (min) 500 B

400

300

R2 = 0.6066 200

100 Paracetamol (nmol/mL plasma)

0 0 60 120 180 Time (min) 0.7 C 0.6

0.5 y = 0.0028x + 0.0173 R2 = 0.8529 0.4

0.3

0.2

Ratio Ratio of CSF:Plasma Paracetamol 0.1

0.0 0 60 120 180 Time (min) Figure 4.15. – Time courses of paracetamol in (A) CSF, (B) plasma and (C) the

CSF:plasma ratio from 26 patients receiving a 1 g intravenous dose of paracetamol.

Samples were analysed using the methodology described in Section 4.2.6. Data shown are combined from 26 patients. The pharmacokinetic models described in Section 4.2.11. were fitted to the CSF and plasma data. For the CSF:plasma ratio a linear regression was used.

133 Chapter 4

Table 4.6. – Summary of pharmacokinetic parameters for paracetamol in CSF and plasma from patients receiving a 1 g intravenous dose of paracetamol.

CSF Plasma

C0 (nmol/mL) 5 377

Cmax (nmol/mL) 76 377 tmax (min) 200 0 t1/2 (min) before Cmax 43 - t1/2 (min) after Cmax - 100 AUC (nmol.min/mL) - 53,818 CL (mL/min) - 122

VD (L) 87 17.5

Corrected V D (L/kg) 1.05 0.2

C0 = concentration at time zero, C max = maximum concentration, t max = time of C max , t 1/2 = time at which concentration is half C max , AUC = area under plasma concentration-time curve, CL = plasma clearance, V D = volume of distribution, V D was then corrected for average patient bodyweight. All parameters were calculated using the formula described in Section 4.2.11.

4.3.11. Metabolism of paracetamol to paracetamol glucuronide, paracetamol sulphate and AM404 in the cerebrospinal fluid and plasma following intravenous administration

Paracetamol glucuronide was detectable in the CSF after 10 minutes at a concentration of 1.3 nmol/mL, while the sulphate was first detectable after 15 minutes at 0.1 nmol/mL (Figure 4.16 and 4.17.). Paracetamol glucuronide was identified in all CSF and plasma samples. Paracetamol sulphate was identified in all plasma samples, but was absent from 9 of the 26 CSF samples. Logarithmic models were fitted to the data sets but no trend was seen with concentrations over time for either the glucuronide (R2

= 0.08) or the sulphate (R 2 = 0.11). No pharmacokinetic parameters could be derived from these data sets.

The glucuronide and sulphate metabolites were at a much higher concentration in the plasma and appeared to accumulate over time. Again logarithmic models were fitted to the data sets but no trend was seen with their concentrations over time for either the glucuronide (R2 = 0.13) or the sulphate (R 2 = 0.09). No pharmacokinetic parameters could be derived from this data set. The CSF:plasma ratio for the glucuronide metabolite appears to decrease logarithmically with time (R2 = 0.48), whereas no trend was seen with the ratio for the sulphate metabolite (R2 = 0.23).

134 Chapter 4

16 A

12

8

4 y = -1.3239Ln(x) + 7.9559 R2 = 0.0792 Paracetamol glucuronide (nmol/mL CSF)

0 0 60 120 180 Time (min) 200 B

160

y = 20.556Ln(x) - 18.018 R2 = 0.1299 120

80

40 Paracetamol glucuronide (nmol/mL plasma) 0 0 60 120 180 Time (min) 0.12 C

0.10

0.08

0.06

0.04 y = -0.0263Ln(x) + 0.1429 R2 = 0.4803

0.02

Ratio Ratio of CSF:Plasma Paracetamol Glucuronide 0.00 0 60 120 180 Time (min)

Figure 4.16. – Time courses of paracetamol glucuronide in (A) CSF, (B) plasma and (C) the CSF:plasma ratio from 26 patients receiving a 1 g intravenous dose of paracetamol.

Samples were analysed using the methodology described in Section 4.2.6. Data shown are combined from 26 patients. Logarithmic curves were fitted to the three data sets.

135 Chapter 4

2.0 A

1.5

1.0

y = -0.2304Ln(x) + 1.4097 2 0.5 R = 0.1087 Paracetamol sulphate (nmol/mL CSF)

0.0 0 60 120 180 Time (min) 100 B

80

y = 8.0787Ln(x) + 8.6853 60 R2 = 0.0907

40

20 Paracetamol sulphate (nmol/mL plasma)

0 0 60 120 180 Time (min) 0.04 C

0.03

0.02

y = -0.0056Ln(x) + 0.0329 R2 = 0.2329 0.01 Ratio of CSF:Plasma Paracetamol Sulphate 0.00 0 60 120 180 Time (min) Figure 4.17. – Time courses of paracetamol sulphate in (A) CSF, (B) plasma and (C) the

CSF:plasma ratio from 26 patients receiving a 1 g intravenous dose of paracetamol.

Samples were analysed using the methodology described in Section 4.2.6. Data shown are combined from 26 patients. Logarithmic curves were fitted to the three data sets. Red points represent values at or below the limit of quantification.

136 Chapter 4

Each compound was expressed as a percentage of the total combined molar concentrations of paracetamol, glucuronide and sulphate (Figure 4.18. and 4.19.).

With the exception of the first few minutes, paracetamol was the predominant of the 3 compounds in the CSF over the time course.

100% A

y = 0.065Ln(x) + 0.6716 80% R2 = 0.2754

60%

40%

paracetamol paracetamol and metabolites 20% Percentage of Percentage CSF paracetamol of total

0% 0 60 120 180 Time (min) B 50%

40%

30%

20%

10%

total paracetamoltotal and metabolites y = -0.0631Ln(x) + 0.3147 R2 = 0.2709

Percentage of Percentage CSF paracetamol glucuronide of 0% 0 60 120 180 Time (min) 4% C

3%

2%

y = -0.0019Ln(x) + 0.0137 1% R2 = 0.0352 total paracetamoltotal and metabolites Percentage of Percentage CSF paracetamol sulphate of 0% 0 60 120 180 Time (min)

Figure 4.18. – Time course of (A) paracetamol, (B) paracetamol glucuronide and (C) paracetamol sulphate concentrations expressed as a percentage of their combined molar concentrations in CSF. Data shown are the combined values from 26 patients. Red points represent values at or below the limit of quantification. Logarithmic curves were fitted to the three data sets.

137 Chapter 4

Like CSF, paracetamol is again the dominant of the 3 compounds in plasma (Figure

4.19.). However the percentage paracetamol does not stay near 100% like in the CSF, but instead decreases to 60% as the other two metabolites are formed.

100% A

80% y = -0.1476Ln(x) + 1.2893 R2 = 0.6057

60%

40%

paracetamol paracetamol and metabolites 20% Percentage of Percentage plasma paracetamol of total 0% 0 60 120 180 Time (min)

B y = 0.0996Ln(x) - 0.2142 40% R2 = 0.4718

30%

20%

10% of of total paracetamol and metabolites

Percentage of Percentage plasma paracetamol glucuronide 0% 0 60 120 180 Time (min) 30% C

y = 0.0479Ln(x) - 0.0751 20% R2 = 0.4695

10% total total paracetamol and metabolites

Percentage of Percentage plasma paracetamol sulphate of 0% 0 60 120 180 Time (min)

Figure 4.19. – Time course of (A) paracetamol, (B) paracetamol glucuronide and (C) paracetamol sulphate concentrations expressed as a total of their combined molar concentrations in plasma. Data shown are the combined values from 26 patients. Logarithmic curves were fitted to the three data sets.

138 Chapter 4

The CSF:plasma ratio was calculated for the combined molar concentrations of the paracetamol, glucuronide and sulphate (Figure 4.20.). This value increased logarithmically over the time course (R 2 = 0.78).

0.3

y = 0.093Ln(x) - 0.2248 R2 = 0.7782 0.2 metabolites 0.1

CSF:plasma ratio of the sum of paracetamol and 0.0 0 60 120 180 Time (min) Figure 4.20. – Time course of the CSF:plasma ratio for the sum of the paracetamol, paracetamol glucuronide and paracetamol sulphate concentrations. Data shown are the combined values from 26 patients. The sum of the paracetamol, paracetamol glucuronide and paracetamol sulphate concentrations were determined for both CSF and plasma and then expressed as the CSF:plasma ratio. A logarithmic curve was fitted to the data set.

Following intravenous administration of paracetamol, AM404 was detectable in the CSF after just 10 minutes at 5 nmol/mL. AM404 was detected in 17 of the 26 CSF samples.

With the exception of one sample, AM404 ranged between 5 - 40 pmol/mL and was not changed with time (Figure 4.21.).

160

120

80 AM404 (pmol/mLAM404 CSF) 40

0 0 60 120 180 Time (min)

Figure 4.21. – Time course of the AM404 concentration in CSF from patients receiving a

1 g intravenous dose of paracetamol. CSF was analysed using the methodology described in

Section 4.2.7. Data shown are the combined concentrations from 26 patients. Red points represent values at or below the limit of quantification. No regression model could be fitted to this data set.

139 Chapter 4

AM404 was detected in 7 of the 26 plasma samples, with only 3 being above the limit of quantification ( >5 pmol/mL).

4.3.12. Formation of oleic and palmitic acid derived metabolites of paracetamol in CSF following intravenous administration

Having demonstrated that FAAH is capable of conjugating oleic acid and palmitic acid to 4-aminophenol, it was investigated as to whether, like AM404, these potentially novel metabolites of paracetamol could be formed in humans. As oleic acid and palmitic acid conjugates of paracetamol are not commercially available a compromise was required.

A precursor ion scan was used to determine the precursor ions that form the 110.2 m/z product ion corresponding to 4-aminophenol. For the oleic acid conjugate, product ions between 373.741 – 374.941 m/z were scanned, while for the palmitic acid conjugate, product ions between 347.741 – 348.941 m/z was scanned.

Chromatograms from CSF where AM404 was high and low were compared (Figure

4.22. and 4.23.). The assumption being that if significant amounts of AM404 could be formed then so could these other metabolites. Due to the number of endogenous peaks present in each chromatogram it was not possible to identify which peaks, if any, corresponded to the predicted metabolites.

140 Chapter 4

A B

Figure 4.22. – LC-MS/MS chromatograms from the precursor ion scan (373.741 –

374.941 m/z


110.2) for the oleic acid conjugate in CSF when AM404 was (A) high and (B) low.

A A

Figure 4.23. – LC-MS/MS chromatograms from the precursor ion scan (347.741 –

348.941 m/z


110.2) for the palmitic acid conjugate in CSF when AM404 was (A) high and (B) low.

141 Chapter 4

4.3.13. FAAH substrate levels in the CSF and plasma following intravenous administration

The effect of paracetamol on FAAH was investigated by measuring the FAAH substrates. AEA could not be detected in either the CSF or plasma. OEA and PEA were only detected in plasma. No correlation over the time course was found with either plasma OEA or PEA levels (Figure 4.24).

250 A

200

150

100 OEA (nmol/mL plasma)

50

0 0 60 120 180 Time (min) 800 B

600

400

PEA (nmol/mL plasma) 200

0 0 60 120 180 Time (min)

Figure 4.24. – Time courses of the OEA (A) and PEA (B) concentration in plasma from

26 patients receiving a 1 g intravenous dose of paracetamol. Samples were analysed using the methodology described in Section 4.2.7. Data shown are the combined concentrations from

26 patients. No regression model could be fitted to this data set.

142 Chapter 4

4.3.14. Prostaglandin E2 levels in the CSF and plasma following intravenous administration

The effect of paracetamol on COX was investigated by measuring PGE2. The results from the analysis showed no change in PGE2 with time in both CSF and plasma (Figure

4.25.).

20 A

15

10 PGE2 (fmol/mL CSF) 5

0 0 60 120 180 Time (min)

1000 B

800

600

400 PGE2 (fmol/mL plasma)

200

0 0 60 120 180 Time (min)

Figure 4.25. – Time courses of PGE2 levels in the (A) CSF and (B) plasma from 26 patients receiving a 1 g intravenous dose of paracetamol. Samples were analysed using the methodology described in Section 4.2.8. Data shown are the combined concentrations from 26 patients. No regression models could be fitted to these data sets.

143 Chapter 4

4.3.15. Leukotriene B4 levels in the CSF and plasma following intravenous administration

The effect of paracetamol on LTB4 was investigated. The results from the analysis showed no change in LTB4 with time in both CSF and plasma (Figure 4.26.).

200

A 150

100 LTB4 (fmol/mLLTB4 CSF) 50

0 0 60 120 180 Time (min)

3000 B

2000

1000 LTB4 LTB4 (fmol/mL plasma)

0 0 60 120 180 Time (min)

Figure 4.26. – Time courses of LTB4 levels in the (A) CSF and (B) plasma from 26 patients receiving a 1 g intravenous dose of paracetamol. Samples were analysed using the methodology described in Section 4.2.9. Data shown are the combined concentrations from 26 patients. No regression models could be fitted to these data sets.

144 Chapter 4

4.3.16. 5HT levels in the CSF and plasma following intravenous administration

The effect of paracetamol on 5HT, its precursor and metabolite, was investigated. No changes in 5HT or tryptophan levels were found in the CSF (Figure 4.27.). Only some samples contained detectable concentrations of 5HIAA.

0.8

A 0.6

0.4 5HT 5HT (nmol/mL CSF) 0.2

0.0 0 60 120 180 Time (min) 5

4 B

3

2 Tryptophan (nmol/mL CSF) 1

0 0 60 120 180 Time (min)

Figure 4.27. – Time courses of (A) 5HT and (B) tryptophan levels in the CSF from 26 patients receiving a 1 g intravenous dose of paracetamol. Samples were analysed using the methodology described in Section 4.2.10. Data shown are the combined concentrations from 26 patients. No regression models could be fitted to these data sets.

145 Chapter 4

4.4. Discussion

4.4.1. Conjugation of fatty acids to paracetamol via FAAH in brain homogenates

Two novel metabolites of paracetamol were hypothesised and shown to form through conjugation of oleic acid and palmitic acid to 4-aminophenol via FAAH in rat brain homogenate (Figure 4.1.). These two molecules are absent in control brain tissue, but are present in brain homogenates incubated with 4-aminophenol and either oleic acid or palmitic acid (Figures 4.5. & 4.6.). They are formed time-dependently and their formation is prevented by the FAAH inhibitor URB597. The metabolites have different retention times and molecular weights, but each form the same 110 m/z fragment, corresponding to 4-aminophenol. Therefore these compounds have different molecular weights and hydrophobicities, but share the same 4-aminophenol group and same

FAAH-dependent route of synthesis as AM404.

Whether formation of these two metabolites actually occurs under physiological conditions in humans is yet to be determined. Of the three fatty acids arachidonic acid, oleic acid and palmitic acid, the most abundant in the human brain is palmitic acid (20

µmol/g), followed by oleic acid (17 µmol/g) and then arachidonic acid (8 µmol/g). 180 It seems likely then that following deacylation of paracetamol in the liver, that more 4- aminophenol would be converted to the oleic and palmitic metabolites by FAAH than to the arachidonic metabolite, AM404.

Further work requires the chemical synthesis and purification of these two metabolites so that quantification can be performed using optimal MS/MS conditions. It cannot be assumed that the lower response values for the oleic acid and palmitic acid is due to less formation of these metabolites, as the method of their detection was not optimised for their structures but instead derived from those of AM404.

Inhibiting FAAH reverses the effect of paracetamol suggesting an important role for

FAAH formed metabolites. It has already been established that AM404 is integral to paracetamol’s analgesic action. Now it is important that the activity of the two novel metabolites described here also be characterised. AM404 has been shown to relieve

146 Chapter 4 pain in in vivo models of nociception, inflammatory pain and neuropathic pain. 296 The effects of these novel paracetamol metabolites should also be investigated in such in vivo models to determine whether they too are analgesic and anti-inflammatory.

Possible routes of biological action for these two metabolites can be hypothesised through comparison with the structurally similar lipids OEA and PEA and also to

AM404. The main action of AM404 is the inhibition of AEA uptake. This ability is due to

AM404 being a structural analogue of AEA, sharing an arachidonic acid side chain. As

OEA and PEA are not transported in this same fashion, it is unlikely that oleic acid and palmitic acid conjugates of paracetamol will inhibit this same transporter. It also seems unlikely that the metabolites would have affinity for the cannabinoid receptors as AM404 lacks affinity and so do OEA and PEA. Therefore any effect is likely to be independent of the endocannabinoid system.

Two possible biological targets for these metabolites are GPR55 and PPAR-α. GPR55, the putative third cannabinoid receptor, is activated by AEA, OEA and PEA. It is possible then that the two paracetamol metabolites may also activate it. Likewise,

AM404 may also have affinity for this receptor. Furthermore OEA and PEA are PPAR-α agonists. Although a connection between paracetamol and PPAR is yet to be demonstrated, the novel metabolites may be possible activators of this receptor. Of course all this depends on whether they reach pharmacologically active concentrations.

Interestingly it is paracetamol’s hydroxyl group that is sulphated and glucuronidated and its acetyl group that is conjugated with arachidonic acid. It is therefore possible that AM404 sulphate and AM404 glucuronide are also formed during paracetamol metabolism. It is also possible that other fatty acid derivatives could be formed from fatty acids such as stearic acid and linoleic acid. Furthermore the arachidonic acid side chain of AM404 may also be oxidised, like AEA, by COX, LOX and cytochrome P450.

These theoretical metabolites would all need to be investigated in future work.

However given the low concentrations of AM404 found here in the CSF, it is unlikely that AM404 glucuronide and AM404 sulphate would be detected.

147 Chapter 4

In summary, the present findings provide the first evidence that FAAH can conjugate oleic acid and palmitic acid to small molecules, and that by this mechanism two potential metabolites of paracetamol can be formed. Typically the aim of drug metabolism is to decrease lipophilicity to allow for clearance, with water soluble molecules, such as glutathione and glucuronic acid, being attached to increase solubility. Interestingly in this example FAAH does the opposite and increases the lipophilicity of paracetamol by attaching fatty acids to it.

In vivo AM404 has been shown to be formed mostly in the brain and spinal cord and to a lesser extent in liver and blood. 164 It is feasible that if the oleic acid and palmitic acid metabolites are formed in humans that they would be detectable in CSF and blood post intravenous paracetamol dose. One aspect of the subsequent clinical trial was to investigate whether these two novel metabolites are formed.

4.4.2. Pharmacokinetics of paracetamol in human cerebrospinal fluid and plasma

The three aims of the clinical trial were to study paracetamol’s pharmacokinetics, metabolism and mechanism of action. Patients were intravenously administered with 1 g paracetamol, with single samples of CSF and blood being collected afterwards. Only a single CSF sample per patient could be collected due to the clinical limitation of collecting multiple CSF samples from each patient. The average volume of CSF in humans is 140 mL, 308 thus the 5 mL collected per patient here is approximately 3.5% of the total volume. Data obtained from each patient was combined into single concentration-time plots, where CSF pharmacokinetic parameters were then derived

(Table 4.7.).

Paracetamol was shown to rapidly enter the CSF and was detectable within the first 10 minutes (Figure 4.15.). Over the time course it accumulated, peaking at 76 nmol/mL towards the end of the time course. The time taken to reach half Cmax was 43 minutes, with the C max being reach at approximately 200 minutes. The t 1/2 after the C max could not be determined as no clearance was observed over the 200 minutes (Table 4.7.).

There paracetamol reached and maintained a high concentration, that lasted for several hours following its administration.

148 Chapter 4

Table 4.7. – Comparison of the CSF paracetamol pharmacokinetics from this study with data from a previous study in children.

This study Kumpulainen et al. 314 Patient age 35 – 79 years 3 months – 12 years Dose 1000 mg i.v. 15 mg/kg mg i.v.

C0 (nmol/mL) 5 10 cmax (nmol/mL) 76 120 tmax (min) 200 60 t1/2 (min) before Cmax 43 30 t1/2 (min) after Cmax - 110

C0 = concentration at time zero, C max = maximum concentration, t max = time of C max , t1/2 = time at which concentration is half C max .

Currently only a single study on the pharmacokinetics of intravenous paracetamol in the CSF has been reported, with a 15 mg/kg dose being investigated in children.320 15 mg/kg is an approximately equivalent dose to 1 g in an adult human. Due to the significant difference in patient age little comparison can be drawn between the studies

(Table 4.7.). The study in children showed that within 5 minutes paracetamol was detectable at 10 nmol/mL and accumulated over time to 120 nmol/mL. Interestingly the clinical trial in children and the clinical trial reported here both showed paracetamol to be present in the CSF within minutes of administration in the low nmol/mL range.

However a significant difference between the two studies is that here no clearance of paracetamol from the CSF was observed over the 200 minute time course, while in the other study paracetamol clearance from CSF began as soon as 60 minutes.

In the plasma paracetamol reached a Cmax of 377 nmol/mL immediately following its administration (Figure 4.15.). The t 1/2 was reached after 100 minutes. After 200 minutes paracetamol was still presented in the plasma at a concentration of approximately 100 nmol/mL. The plasma pharmacokinetics of paracetamol has been well studied following oral administration.315 The pharmacokinetics of intravenous administered paracetamol is less well investigated, although it has recently been

314 comprehensively studied (Table 4.8.). This previous study showed that the C max following 1 g intravenous paracetamol was 165 nmol/mL, which is a lot lower than the

377 nmol/mL presented here. This was reportedly cleared from the plasma with a t 1/2 of 156 minutes, which is a lot slower than the 100 minutes presented here.

Furthermore the AUC, CL and V D values differed significantly from those reported here.

149 Chapter 4

Therefore the results from the clinical trial presented here are not consistent with previously reported intravenous pharmacokinetics. The reason for this difference is unknown.

Table 4.8. – Comparison of the plasma paracetamol pharmacokinetics from this study with data from previous intravenous and oral studies.

This study Liukas et al. 314 Hinz et al. 315 Dose 1000 mg i.v. 1000 mg i.v. 1000 mg p.o. cmax (nmol/mL) 377 165 105 t1/2 (min) 100 156 268 tmax (min) 0 0 75 AUC (mg.h/mL) 138 48 63 * CL (mL/min) 122 349 -

VD (L/kg) 0.2 0.9 -

Cmax = maximum concentration, t 1/2 = time at which concentration is half C max , t max = time of C max ,

AUC = area under plasma concentration-time curve, CL = plasma clearance, V D = volume of distribution, V D was then corrected for bodyweight. * = recalculated values from original units.

As expected the C max of 1000 mg oral paracetamol (105 nmol/mL) is a lot lower than the same dose given intravenously (377 nmol/mL). Interestingly in the clinical trial presented here paracetamol was still above 100 nmol/mL in the plasma at the end of the time course. Therefore it takes approximately 200 minutes for the plasma paracetamol level following an intravenous dose to fall to that of the Cmax of an oral dose.

In summary, the results presented here demonstrate that paracetamol rapidly enters the CSF where it accumulates over several hours. However significant differences between this study and previous studies in terms of plasma pharmacokinetics were seen. A potential reason for this is that concentration-time plots were produced from combined data from 26 patients.

4.4.3. Metabolism of paracetamol in human cerebrospinal fluid and plasma

The second aim of the clinical trial was to study the metabolism of paracetamol. The majority of intravenous paracetamol is known to be cleared as the unmetabolised drug and its major metabolites in the urine.370 Previous studies of intravenous paracetamol

150 Chapter 4 have not measured the paracetamol glucuronide and paracetamol sulphate, either in

CSF or plasma. Here the metabolism of paracetamol into its two major metabolites, paracetamol glucuronide and paracetamol sulphate, as well as its minor metabolite

AM404, was investigated in the CSF and plasma.

The major metabolites were detectable in the CSF as soon as 10 minutes post intravenous administration (Figures 4.16. & 4.17.). Their concentrations ranged from 1

- 15 nmol/mL for paracetamol glucuronide and from 0.1 – 1.7 nmol/mL for paracetamol sulphate, with the higher concentrations being seen at the beginning of the time course. No trend in their concentration over time was found, and no pharmacokinetic model could be applied. This is the first time that paracetamol glucuronide and paracetamol sulphate have been found in human CSF. The blood-brain barrier should be impermeable to large and hydrophilic compounds. 309 By an unknown mechanism low concentrations of these large and charged molecules either managed to cross the blood-brain barrier and consequently enter the CSF, or were actually synthesised somewhere within the CNS. The latter seems unlikely given the low amount of protein in the CSF and that the majority of drug metabolism occurs in the liver. A previously reported study looking at the pharmacokinetics of intravenous morphine in the CSF found morphine-6-glucuronide. 371 Therefore the presence of glucuronidated drug metabolites in the CSF is known, but the reason behind their presence here is poorly understood and requires further investigation.

Similar to the CSF, no trend between concentration and time could be seen in the plasma for either metabolite and thus standard pharmacokinetic models could not be applied. Their concentrations ranged from 10 - 160 nmol/mL for paracetamol glucuronide and 10 – 80 nmol/mL for paracetamol sulphate. These concentrations are comparable to previously reported concentrations following 1 g intravenous paracetamol, where paracetamol glucuronide peaked at 60 nmol/mL after 240 minutes, with paracetamol sulphate peaking much later after 120 minutes at 30 nmol/mL. 314 Their study demonstrated a clear trend in the accumulation and clearance of the paracetamol metabolites, and also included pharmacokinetic parameters of the metabolites. In contrast this present study showed no trend in the concentration-time plots of the metabolites and therefore no pharmacokinetic parameters could be

151 Chapter 4 derived. This is most likely because data was used from multiple patients to construct single concentration-time plots. Therefore the expected variation between patients, such as weight (which ranged from 52 – 125 kg), made it difficult for any trend to be observed.

Of the paracetamol metabolites it was paracetamol glucuronide that is the most concentrated. In the CSF 5 - 20% of paracetamol was glucuronidated compared to only 1 - 4% being sulphated (Figure 4.18.). This was also seen the in plasma, with between 20 - 40% of paracetamol being glucuronidated and 5 - 20% being sulphated

(Figure 4.19.). In summary, the major metabolites of paracetamol, its glucuronide and sulphate, are mostly formed within the plasma and as shown here for the first time are also present in the CSF, arriving rapidly.

4.4.4. Metabolism of paracetamol to AM404 and other fatty acid conjugates

The results from this study showed for the first time that AM404, the FAAH-formed active metabolite of paracetamol and AEA reuptake inhibitor, is formed in humans and is present in both the CSF (17 of 26 patients) and plasma (3 of 26 patients). AM404 concentrations ranged from 5 – 40 pmol/mL in the CSF (Figure 4.21.).

FAAH is mostly expressed in the brain. 102 AM404 should therefore mostly be formed in the brain, with a small amount crossing into the CSF. This is supported by an in vivo study that demonstrates that the majority of AM404 produced following paracetamol administration is found in the brain and to a lesser extent in the spinal cord, with negligible concentrations being found in the plasma. 164 This explains why in the clinical trial described here AM404 was found in the CSF, with only a couple of samples containing detectable amounts of AM404. The concentration of AM404 is reported to be 15-fold higher in the brain than in the spinal cord levels. 164 Therefore while low levels of AM404 were found here in the CSF, significantly higher levels of this metabolite should be found in the brain, but such a study is not possible in humans.

The dose of paracetamol used in this clinical trial was 1 g, which is equivalent to 6.6 mmoles. The total volume of CSF in the adult human body is on average 140 mL. 308

152 Chapter 4

Assuming all patients had a CSF volume of 140 mL, then the peak paracetamol CSF concentration of 60 nmol/mL seen in this present study is equivalent to a peak mass of approximately 8.4 µmoles. Therefore around 0.13% of the dosed paracetamol molecules accumulate in the CSF, whereas only 0.004% accumulates as the glucuronide and 0.0002% as the sulphate. Even less, approximately 0.0001%, was found in the CSF as AM404.

It is difficult to conclude whether the low amounts of AM404 formed in the human CNS are of any pharmacological relevance. The carrier protein responsible for AEA reuptake that AM404 inhibits, the FAAH-like anandamide transporter (FLAT), has only recently

104 been identified. AM404 inhibits this protein’s reuptake of AEA with an IC 50 of 5.3 nmol/mL. This is obviously 1000-fold higher than the concentrations found in the CSF, but AEA reuptake occurs in neurons of the brain where AM404 levels should be significantly higher.

AM404 was found to be present in the CSF at a similar concentration (5 – 40 pmol/mL) to the endogenous opioid β-endorphin (12 pmol/mL). 372 The other major endogenous opioid, met-enkephalin, is present at an even lower concentration of 79 fmol/mL. 373

This demonstrates that the endorphins, enkephalins and potentially AM404, can still be potent mediators of pain perception even at low concentrations. Therefore while

AM404 may be present in the CSF at the pmol/mL range, it does not mean it is pharmacologically inactive.

The definitive study to determine whether AM404 reaches a pharmacological relevant concentration and consequently produces analgesia via the cannabinoid receptors in humans would be by using antagonists. The 5HT receptors were demonstrated to be integral to paracetamol’s action in humans in a simple study that investigated the pain threshold of subjects receiving paracetamol compared to subjects receiving

301 paracetamol and the 5HT 3 receptor antagonist tropisetron. Repeating the same study, but replacing tropisetron with a cannabinoid receptor antagonist such as rimonabant, would demonstrate whether paracetamol’s analgesia is cannabinoid receptor-dependent. Unfortunately cannabinoid receptor antagonists, including rimonabant, reportedly cause severe depression and have now been withdrawn.374

153 Chapter 4

Although a few of the CSF samples and most of the plasma samples did not have detectable amounts of AM404, it does not mean they were not present. Due to their low concentration the assay used may have lacked the required sensitivity. A future refinement of the analytical method may be required to concentrate the samples via

SPE, for added sensitivity. However this would need to be balanced with the fact that only small volumes of CSF can be collected.

It was demonstrated in Section 4.3.2. that FAAH can conjugate other fatty acids to 4- aminophenol. While the arachidonic acid derived metabolite of paracetamol, AM404, was shown here to be formed in humans, it cannot be concluded from this study whether the oleic acid or palmitic acid derived metabolites are also formed. Given the fact that human brain contains 20 µmol/g palmitic acid, 17 µmol/g oleic acid and 8

µmol/g arachidonic acid, 180 it is possible that more of these novel metabolites should be formed than AM404. The analytical approach used here was not successful at finding these metabolites (Figure 4.22. & 4.23.). To determine whether these metabolites are actually formed an optimised analytical method to be developed, which would require the chemical synthesis of the two hypothesised metabolites.

In summary, the formation of AM404 by humans was demonstrated and detected in

CSF following intravenous paracetamol. As the mechanism of paracetamol is known to be dependent on the cannabinoid receptors in vivo , the finding that humans produce

AM404 supports this although further investigations would be required to definitively prove this. The other oleic acid and palmitic acid conjugates of paracetamol could not be detected using the methodology used here.

4.4.5. The transfer of paracetamol and its metabolites into the CSF

The CSF:plasma ratio of paracetamol increased linearly over the time course (Figure

4.15. C). After 200 minutes the ratio was 0.6, indicating that at this point there was more than half the amount of paracetamol in the CSF than in the plasma. This is in contrast to a previous study looking at intravenous paracetamol in the CSF of children, which showed twice as much paracetamol in the CSF than in the plasma after 300

154 Chapter 4 minutes. 320 The most likely explanation for this observation is the anatomical differences between children and adults.

The CSF:plasma ratio is influenced by the simultaneous accumulation of paracetamol in the CSF and its metabolism and clearance from the plasma. Due to these dynamics, equilibrium between the two compartments should not occur. However a 1:1 concentration ratio between the two compartments may occur, but would only be transient. The CSF:plasma ratio was also calculated for the combined molar concentrations of paracetamol and the glucuronide and sulphate metabolites (Figure

4.20.). This was to rule out the role of metabolism in increasing the CSF:plasma ratio, as the paracetamol moiety was investigated as opposed to just paracetamol. This demonstrated that the paracetamol moiety, whether unconjugated or conjugated, transfers into the CSF by a logarithmic process (R 2 = 0.78). The process by which this occurs begins to plateau towards the end of the 200 minute time course. Therefore while the CSF:plasma ratio of paracetamol alone appears to show that transfer into the

CSF is a linear process, combining the molar concentrations of paracetamol and its metabolites reveals that it follows a more logarithmic transfer mechanism.

In summary, paracetamol is known to be active in the CNS. This is supported by its rapid transfer into the CSF. There is still much to be investigated concerning the mechanism by which paracetamol enters the CSF, specifically related to how its metabolites manage to be transferred there.

4.4.6. Paracetamol’s effect on endogenous pathways

The third aim of the clinical trial was to comprehensively study the mechanism of action of paracetamol following intravenous administration. Paracetamol is known to inhibit COX and to interact with both the endocannabinoid and serotonin systems. To study these interactions various endogenous mediators associated with paracetamol were measured in both the CSF and plasma.

Paracetamol is a weak inhibitor of COX-1 and COX-2. It has been reported that the

IC 50 of paracetamol in human whole blood for inhibiting COX-1 is 105 nmol/mL and at

155 Chapter 4

COX-2 is 26 nmol/mL.315 This same study showed that 1 g oral paracetamol results in a C max of 105 nmol/mL and therefore is high enough to significantly inhibit COX-2, but

315 not COX-1. In this present study the C max of 1 g intravenous paracetamol was 377 nmol/mL in plasma. Therefore the drug was present at a pharmacologically relevant concentration for the inhibition of both COX-1 and COX-2 activity over the entire time course. In the CSF paracetamol reached a concentration of 76 nmol/mL, which may be pharmacologically relevant due to COX-2 being expressed in the spinal cord. 66 PGE2 levels in the brain 287, 288 and spinal cord 375 have previously been shown in in vivo studies to be decreased by paracetamol.

Given the pharmacologically relevant concentrations of paracetamol, it was hypothesised that inhibition of PGE2 synthesis should be observed in plasma. If such inhibition of PGE2 synthesis were to occur then there may also be increased arachidonic acid present for the synthesis of LTB4 via LOX. However no change in

PGE2 or LTB4 was observed in both CSF and plasma over the time course (Figures

4.25. & 4.26.). As no change was seen in PGE2, it is understandable as to why no change was also seen with LTB4. These results were supported by a similar previous study that showed orally administered paracetamol did not effect PGE2 levels in human plasma. 376 Interestingly an in vivo study showed that paracetamol (200 mg/kg) was analgesic in rats without reducing brain PGE2 levels. 294 Therefore there is evidence to support a lack of PGE2 inhibition by paracetamol it certain tissues. It should be noted that an in vivo study showed spinal cord injury triggered PGE2 production in the CSF of rats.377 Whether lumbar puncture also causes PGE2 production is not known, but it may be an issue here.

The determined concentrations of PGE2 and LTB4 were mostly similar to the known endogenous levels of these mediators. In the CSF, PGE2 levels were predominantly between 4 – 9 fmol/mL and LTB4 levels were mostly between 30 – 150 fmol/mL.

These concentrations were similar to previously reported human CSF control levels of

PGE2 (17 – 43 fmol/mL) 378 and LTB4 (60 – 180 fmol/mL). 379 In the plasma, PGE2 levels were mostly between 140 – 430 fmol/mL, and LTB4 levels were mostly between

1200 – 2400 fmol/mL. While the PGE2 concentration was similar to previously reported control levels in human plasma (100 - 130 fmol/mL), 380 the LTB4 concentrations were

156 Chapter 4 higher than previously reported levels (60 - 80 fmol/mL). 381 The reason for this discrepancy is unclear.

The presence of AM404 in human CSF following intravenous paracetamol supports the role of the endocannabinoid system in paracetamol’s action. As paracetamol’s active metabolite, AM404, is an AEA reuptake inhibitor, it was hypothesised that paracetamol may have an affect on AEA levels, particularly in the CSF. Previously reported studies assaying AEA by LC-MS/MS showed that the normal concentration of AEA in human

CSF is between 5 - 9 fmol/ml 382 and between 0.3 – 1.5 pmol/mL in plasma. 142 In this present study the LC-MS/MS assay lacked adequate sensitivity for the detection of

AEA. As previous studies concentrated the samples by SPE before analysis, it is likely that the samples required pre-concentrating to be detectable. However an in vivo study previously showed that paracetamol at a dose of 300 mg/kg was analgesic without increasing brain AEA levels. 294 Therefore even with adequate sensitivity a change in AEA levels may not have been seen.

Furthermore it was hypothesised that as FAAH is required for AM404 synthesis that paracetamol may change OEA and PEA levels. In Chapter 3 it was demonstrated that in vivo blood OEA and PEA correlated well with brain AEA and brain FAAH activity.

Previous studies have shown that FAAH substrates are present in the CSF at the fmol/mL level.382-384 The LC-MS/MS method used here had inadequate sensitivity for their quantitation. There was however adequate sensitivity to quantitate OEA and PEA in plasma, however no change in their concentrations over the time course was observed (Figure 4.24.).

Paracetamol’s analgesic action is partly dependent on its indirect activation of 5HT receptors, 300 which it achieves by increasing 5HT concentrations in certain regions of the CNS. 299 The analgesic effect of paracetamol has been shown to be blocked in humans by 5HT antagonists. 304 5HT, its precursor tryptophan and its metabolite 5HIAA are all naturally fluorescent and were measured in the CSF using a previously reported method. 385 Plasma was not assayed as 5HT has a different role in blood, being stored in platelets and released in clots to act as a vasoconstrictor (narrowing of the blood vessels).386 Despite the apparent dependence of paracetamol on the 5HT receptors, no

157 Chapter 4 change was seen with either 5HT or its precursor tryptophan (Figure 4.27.). This is not surprising as paracetamol does not change in vivo spinal cord 5HT levels. 299 The endogenous concentration of 5HT in normal human CSF has been reported to be between 0.3 – 1.3 nmol/mL. 387 Similar concentrations were therefore seen here, with

5HT levels being shown to range between 0.3 – 0.5 nmol/mL.

As already discussed, the results from the paracetamol pharmacokinetics study showed that over time the concentration of paracetamol increased in the CSF and decreased in the plasma. Therefore it appears that in the presence of increasing/decreasing concentrations of paracetamol that PGE2, LTB4, OEA, PEA and

5HT were not affected in CSF or plasma. These findings are supported by previously discussed in vivo studies that also showed no change in brain PGE2, brain AEA or spinal cord 5HT levels following paracetamol administration.

Despite no changes in PGE2, LTB4, OEA, PEA and 5HT being reported here, it cannot be concluded that these endogenous mediators are not integral to the action of paracetamol. The data from this study is unpaired in that samples were only taken from one time point per patient. No comparison before and after paracetamol dosage was made. An ideal trial would take CSF before paracetamol administration, so that baseline readings of each endogenous mediator could be made, and then at multiple time points afterwards. This would demonstrate whether paracetamol altered the concentrations of these mediators in each patient. However sampling at multiple time points may potentially cause the patient harm and additional stress and discomfort, and therefore would not be seen as ethical.

4.4.7. Summary

Paracetamol’s pharmacokinetics, metabolism and mechanism of action were studied in human CSF and plasma following intravenous administration. The data from this study was from 26 male patients of different ages, weights, heights and with various disorders. Despite this a clear trend was seen in paracetamol’s accumulation in CSF and clearance from plasma over time. Its major metabolites, paracetamol glucuronide and paracetamol sulphate, were all found to be present in the CSF, and for the first

158 Chapter 4 time the FAAH formed metabolite AM404 was shown to be produced by humans following paracetamol administration. Oleic acid and palmitic acid derivatives of paracetamol, while being shown here to be produced by FAAH in rat brain homogenates, were not detected in CSF. Furthermore the levels of the endogenous mediators PGE2, LTB4, AEA, OEA, PEA and 5HT were not shown to be changed by paracetamol over the time course.

The limitations of this study are that it was only a small study (n = 26). Drug dose was not weight adjusted. As the patient’s weights ranged from 52 – 125 kg the 1 g dose effectively ranged from 8 - 20 mg/kg. No clearance of paracetamol was observed over the time course and therefore any future studies should ideally be performed over a longer duration to investigate clearance from the CSF. AM404 was only detected in 17 of the 26 patients. This is probably not because it was absent, but because the assay was not sensitive enough to quantitate samples with AM404 concentrations lower than

0.5 pmol/mL. This is also the case for AEA, which was measured to see whether its reuptake inhibitor AM404 would change its levels. AEA is known to be present in the

CSF at the fmol/mL level. Therefore a more sensitive assay would be required for quantitation. PGE2 and 5HT are both known to be important endogenous mediators of paracetamol’s action. A possible reason as to why no change was seen in their concentration is due to the experimental design. Taking a CSF sample before and after paracetamol would therefore allow for comparisons to be made to a baseline level. A change in their concentration is more likely to be detected using this approach, but would potentially cause more discomfort to the patient. Additionally trauma to the spinal cord would cause a release of PGE2 and impair its measurement.

159 Chapter 5

CHAPTER 5

Discussion

160 Chapter 5

5.1. Overview

The research presented here has explored several areas of FAAH and its role in analgesia and inflammation. FAAH was studied in cultured activated macrophages

(Chapter 2), in an in vivo model of inflammatory pain (Chapter 3) and in human cerebrospinal fluid during a clinical trial (Chapter 4). These investigations concluded in several novel findings as detailed below.

1. FAAH and MAGL are known to produce arachidonic acid in the hydrolysis of the

endocannabinoids. It was investigated here as to whether these enzymes could

influence production of prostaglandins, which are oxidative metabolites of

arachidonic formed via COX. It was shown that inhibition of either FAAH or MAGL

reduced PGE2 production in activated macrophages (Chapter 2).

2. Measurement of the FAAH substrates are increasingly being used as a

pharmacodynamic parameter in the development of novel FAAH inhibitors.

Covalent FAAH inhibitors were shown to produce significant increases in AEA, OEA

and PEA, while such an effect was not seen with non-covalent inhibitors. Blood

OEA and PEA levels were found to be indicative of brain FAAH activity and brain

AEA concentrations (Chapter 3).

3. FAAH is already known to be capable of conjugating arachidonic acid to small

molecules, such as 4-aminophenol, ethanolamine and dopamine. Here FAAH was

demonstrated for the first time to also be capable of conjugating oleic acid and

palmitic acid to small molecules (Chapter 4).

4. The active metabolite of paracetamol, AM404, is a conjugate of arachidonic and 4-

aminophenol that is synthesised by FAAH. FAAH was shown for the first time to

also form novel oleic acid and palmitic acid conjugates of paracetamol (Chapter 4).

5. The action of paracetamol has been demonstrated to be dependent on the

endocannabinoid system. AM404, an inhibitor of AEA reuptake, is thought to

mediate this action. This FAAH formed active metabolite has only previously been

161 Chapter 5

investigated in vivo and has not been identified in humans. For the first time

AM404 was shown here to be synthesised by humans and be detectable in the

cerebrospinal fluid of patients receiving intravenous paracetamol (Chapter 4). This

offers the first evidence of paracetamol interacting with the endocannabinoid

system in humans.

6. The pharmacokinetics of paracetamol in the plasma following oral administration

has been well studied. However paracetamol’s pharmacokinetics in the CSF

following intravenous administration is not known. As paracetamol is increasingly

been administered intravenously and as the analgesic is known to be active in the

CNS, this was worth investigating. For the first time the pharmacokinetics of

intravenous paracetamol in the CSF of adult humans was determined, with

paracetamol being detectable in the CSF within 10 minutes and still being present

after 3 hours (Chapter 4).

7. Paracetamol has two major metabolites, paracetamol glucuronide and paracetamol

sulphate. These metabolites are well studied in plasma and urine, but have not

previously been measured in CSF. For the time first time the glucuronide and

sulphate metabolites of paracetamol were found to be present in the cerebrospinal

fluid of patients receiving intravenous paracetamol (Chapter 4).

The impact of these findings on research into the endocannabinoid system, FAAH inhibition and paracetamol will be discussed here alongside several areas still needing to be addressed and other recent discoveries.

5.2. The expanding endocannabinoid system

The endocannabinoid system is often simply defined as consisting of the cannabinoid receptors CB1 and CB2, the receptor’s endogenous ligands AEA and 2AG, and the ligand’s degradative enzymes FAAH and MAGL. It is now becoming clear that the players functioning in endocannabinoid signalling go beyond two receptors, two ligands and two enzymes. New receptors, ligands and enzymes connected with endocannabinoid signalling have all been identified, as has the carrier protein involved

162 Chapter 5 in AEA reuptake. Furthermore new actions for the currently known ligands and proteins have been demonstrated, as has interaction between the endocannabinoid system and other lipid signalling networks.

5.2.1. New receptors in the endocannabinoid system

The cannabinoid receptors were originally identified as the protein in the brain that the bioactive components of Cannabis highly bind to, such as ∆9-THC and cannabinol.98, 99

Endogenous lipids with the same action were later identified and named the endocannabinoids. 100, 101 Since these discoveries, endocannabinoids have also been shown to activate receptors outside of the endocannabinoid system. Beyond the cannabinoid receptors CB1 and CB2, the endocannabinoids AEA and 2AG also activate the PPAR-α and TRPV1 receptors. PPAR-α appears to be integral to the endocannabinoid system as the analgesic effect of FAAH inhibition is completely reversed by antagonism of PPAR-α.223, 236 TRPV1 is also activated by AEA, although at higher concentrations than those required for CB1. 388 However in inflammation TRPV1 expressing sensory neurons are sensitised to AEA by PGE2, thus lower concentrations are required for activation.240 It remains to be elucidated as to whether the beneficial increases of AEA by a FAAH inhibitor in inflammatory is offset by activation of sensitised TRPV1 by AEA. Additionally the uncharacterised G-protein coupled receptors

GPR18, 389 GPR55 253 and GPR199 390 have all been demonstrated to be activated by endocannabinoids. Whether any of these receptors will become the CB3 receptor remains to be seen, however it is now clear that the receptors activated by endocannabinoids outnumber just two cannabinoid receptors.

5.2.2. New ligands in the endocannabinoid system

Like the receptors, the number of lipids thought to be involved in endocannabinoid signalling is also increasing. The endocannabinoids are hydrolysed by FAAH and MAGL.

However in addition to AEA and 2AG, FAAH and MAGL have various other non- endocannabinoid substrates. FAAH metabolise a class of lipids called the fatty acid ethanolamides, while MAGL metabolises a class of lipids called the fatty acid glycerols.

Little is known about these lipids, which include myristoyl ethanolamide (14:0),

163 Chapter 5 stearoyl ethanolamide (18:0), linoleoyl ethanolamide (18:2), arachidoyl ethanolamide

(20:0), eicosapentaenoyl ethanolamide (20:5) and docosanoyl ethanolamide (22:6), and their glycerol counterparts. 278, 391 While these lipids have been reported to be endogenous, 179 their receptors need to be identified and whether like AEA, OEA and

PEA, they have analgesic and anti-inflammatory properties requires further investigation. As OEA and PEA lack affinity for the cannabinoid receptors they are not defined as endocannabinoids. Although they can potentiate the action of AEA, 257 possibly due to the higher concentrations of OEA and PEA resulting in competition for

FAAH hydrolysis.

Oxidative metabolites of the endocannabinoids have been shown to possess affinity for the cannabinoid receptor. For example the AEA metabolite 5,6-EET-ethanolamide is a potent CB2 agonist. 364 As already discussed, whether these AEA metabolites are produced physiologically is not known. Oxidative metabolism of OEA and PEA was demonstrated in Chapter 3. As OEA and PEA are significantly more concentrated than

AEA, it is more likely their oxidative metabolites are formed physiologically. While this area requires significantly more research, oxidised fatty acid ethanolamides represent a novel group of lipid signalling molecules that possibly may be formed under specific physiological conditions.

Derivatives of the haemoglobin fragment, hemopressin, have been recently identified as an endogenous peptide component of the brain capable of activating the cannabinoid receptors. 392 RVD-HPα and VD-HPα are both agonists of the CB1 receptor, while VD-HPα also has affinity for CB2. These findings open up the possibility that the cannabinoid receptors are activated by both lipid and peptide signalling molecules. In summary, the fatty acid ethanolamides, fatty acid glycerols, oxidised endocannabinoid metabolites and hemopressin derivatives all represent novel lipid and peptide classes of molecules that are in part connected to the endocannabinoid system. Further research into these lipids may reveal new therapeutic approaches based on their manipulation.

164 Chapter 5

5.2.3. New proteins in the endocannabinoid system

In addition to the identification of new receptors and new ligands, new proteins have also been identified. Specifically two potentially key players in the endocannabinoid system, namely the AEA transporter FLAT and a second FAAH subtype FAAH-2, were recently discovered by Fu et al. and Wei et al. respectively.104, 160 The reuptake of AEA into neurons has long been known, 260 with only until recently a high affinity binding site being identified.197 The reuptake of AEA into neurons has now been demonstrated to be by a FAAH variant named the FAAH-like anandamide transporter (FLAT).104 FLAT is a membrane bound carrier protein that lacks FAAH activity and is not inhibited by

FAAH inhibitors. While FAAH inhibition increases intracellular AEA, the inhibition of AEA reuptake increases the amount of AEA in the synapse. As only a fraction of the increased intracellular AEA following FAAH inhibition would make it into the synapse, inhibition of AEA reuptake is likely to become a more attractive therapeutic target than

FAAH. The best characterised inhibitor of AEA reuptake is AM404, which was shown in

Chapter 4 to be formed in the human CSF as a metabolite of paracetamol. Novel potent inhibitors of AEA reuptake based on AM404’s action now need to be developed as therapeutics. CB2 is expressed mostly on immune cells. Whether FLAT, or another

AEA carrier protein, is expressed on immune cells and therefore involved in the regulation of endocannabinoid signalling needs to be determined. Furthermore whether

FLAT is also responsible for 2-AG reuptake, or whether its own carrier protein exists needs to be determined.

The second potential new protein in the endocannabinoid system is, like FLAT, also a

FAAH variant named FAAH-2. FAAH-2 is expressed in humans but not in rodents. 160

Despite FAAH-2 sharing only 20% sequence homology with FAAH, there is significant overlap in tissue distribution and substrate selectivity. This subtype could present complications in the development of FAAH inhibitors, as in humans FAAH-2 is expressed in the heart while FAAH-1 is not.160 The recently developed FAAH inhibitors investigated in Chapters 2 & 3, PF-3845 and PF-04457845, were specifically designed not to inhibit FAAH-2, thus avoiding this issue. 225

165 Chapter 5

5.2.4. New actions of the endocannabinoid system

Alongside new receptors, ligands, enzymes and transport proteins, new actions of the currently accepted players of the endocannabinoid system have been recently demonstrated. While hydrolysis of AEA, OEA and PEA is the main attributed action of

FAAH, little is known about its reverse action of conjugating fatty acids to small molecules. This mechanism is a route of synthesis for AEA, 151 NADA 163 and AM404,164 from ethanolamine, dopamine and 4-aminophenol respectively. While previously there was only evidence to suggest that FAAH could conjugate arachidonic acid to small molecules, here in Chapter 4 it was demonstrated that the enzyme can also conjugate oleic acid and palmitic acid.

Several questions remain unanswered concerning this newly established action of

FAAH. Firstly the cause of reversed FAAH activity needs to be elucidated. The activity of some enzymes is reversed when the concentration of the enzyme’s product is higher than its substrate. For FAAH a shift in the equilibrium between the enzymes substrates

(AEA, OEA and PEA) and products (arachidonic acid, oleic acid and palmitic acid) is an unlikely explanation. This is because the brain concentrations of its products are already 1000-fold higher than the FAAH substrates. 141, 180 As well as fatty acids the other product of FAAH is ethanolamine. It is possible that specific molecules act as analogues of ethanolamine, so that once such molecules reach a critical concentration they are capable of reversing FAAH activity. Either way the reverse activity of FAAH has been now proved to be a physiological occurrence in humans, as demonstrated by the presence of AM404 in the CSF in Chapter 4.

The type of small molecule that FAAH can conjugate fatty acids to needs to be investigated. So far ethanolamine, 4-aminophenol and dopamine have all been identified. It therefore appears that the presence of an amino group is integral. FAAH then catalyses the dehydration / condensation reaction between this amino group and the carboxyl group present on fatty acids to produce a conjugate and a molecule of

H2O. Based on the understanding of this mechanism small molecules should now be screened for their ability to reverse FAAH activity. Such molecules may form the basis of novel therapeutics.

166 Chapter 5

Currently there are a number of fatty acid derivatives of neurotransmitters that have been identified, such as such as N-arachidonoyl dopamine (NADA), oleoyl dopamine, palmitoyl dopamine, arachidonoyl serotonin, oleoyl serotonin and palmitoyl serotonin. 393, 394 Additionally fatty acid derivatives of amino acids have also been discovered, such as N-arachidonoyl glycine. 395 Again these neurotransmitters and amino acids share an amino group. It is possible that some of these lipids are synthesised by reverse FAAH activity and it is therefore also possible that other FAAH formed derivatives of neurotransmitters and amino acids are in existence. Other novel fatty acid conjugates should now be identified. Furthermore whether MAGL, like FAAH, can also have its activity reversed needs to be investigated.

5.2.5. New interactions of the endocannabinoid system

It is important that the endocannabinoid system is not seen as a closed system. There are in fact a growing number of examples of overlap between the endocannabinoid system and other lipid signalling networks. Hydrolysis of the endocannabinoids produces arachidonic acid, a fatty acid present in numerous other lipids involved in pain and inflammation, such as the prostaglandins, leukotrienes and the lysophosphatidic acids. The latter is class of lipids released at the site of injury, 396-398 where they initiate pain. 399-402 It is possible that by inhibiting endocannabinoid hydrolysis that less arachidonic acid would be available for the synthesis of lipids like the prostaglandins and lysophosphatidic acids. In Chapter 2 it was shown that inhibition of FAAH and MAGL decreases PGE2 production in activated macrophages.

Whether this same effect is seen in vivo should be the subject of future research.

Similarly a recent study showed that 2-AG is capable of activating neutrophils, in that it stimulates myeloperoxidase release, kinase activation and Ca 2+ mobilisation. 341 This effect is not reversed by cannabinoid antagonists, but is instead due to 2-AG being hydrolysed by MAGL into arachidonic acid and consequently being quickly converted into the chemoattractant LTB4. The implication of this being that a part of the anti- inflammatory effect of MAGL inhibition is due to its influence over LTB4 levels and consequently neutrophil activation. Therefore due to its control of arachidonic acid, the

167 Chapter 5 hydrolysis of the endocannabinoids results in several lipid signalling networks overlapping.

5.2.6. Summary

The endocannabinoid system is expanding in the number of receptors, ligands and enzymes involved in its signalling and also in terms of previously unknown actions being revealed and interactions with other lipid signalling networks being demonstrated. The work presented here has contributed to this, highlighting interaction between endocannabinoid hydrolysis with prostaglandin synthesis and revealing a new action for FAAH in its conjugation of oleic acid and palmitic acid to small molecules. Further unravelling of the endocannabinoid system may reveal potential new mechanisms for therapeutic intervention.

5.3. The role of the endocannabinoid system in paracetamol’s action

The mechanism by which paracetamol produces analgesia is often quoted as being largely unknown. 403 Due to the adverse effects of paracetamol, such as liver toxicity, an improved version of the drug would be desirable. The screening of potential paracetamol replacements would require assays that demonstrate selectivity for its molecular targets. However until paracetamol’s mechanism is fully understood this is not possible. The reason for the confusion concerning paracetamol’s action is that it appears to involve multiple pathways including the serotonin receptors, and the prostaglandin and endocannabinoid systems.

It is well accepted that paracetamol is active in the CNS. The findings from Chapter 4 supports paracetamol being a mostly CNS active drug, demonstrating that it penetrates rapidly into the CSF where its presence significantly outlasts its presence in plasma. However despite the high concentrations reached by paracetamol in the CSF, no change in the CSF levels of PGE2 or 5HT were seen. Even so there is clinical evidence that demonstrates the interaction between paracetamol, COX and the 5HT

168 Chapter 5 receptors in humans. For example PGE2 production in the acute inflammation of peripheral tissue has been shown to be suppressed following oral administration of 1 g of paracetamol. 291 Furthermore the urinary excretion of PGE2 is reduced by 43% following a daily 4 g oral intake of paracetamol. 404 In terms of the 5HT receptors a clinical study showed that the antagonism of the 5HT 3 receptor reverses the effect of paracetamol. 304

Similar studies that conclusively support the integrity of the endocannabinoid system to paracetamol’s action are required. The finding in Chapter 4 that humans produce

AM404 clearly proves an interaction between FAAH and paracetamol. As discussed in

Section 4.4.4. the concentration required for AM404 to be analgesic is not known.

Therefore the presence of AM404 still does not answer the question as to whether there are sufficient amounts formed to inhibit AEA reuptake, activate the cannabinoid receptors and produce analgesia. The definitive experiment to prove that the cannabinoid receptors are integral to paracetamol’s action would be to use cannabinoid antagonists. Unfortunately cannabinoid antagonists are no longer in use due to the adverse effect of depression.

There is some evidence that possibly demonstrates similarity between paracetamol and cannabinoids. In humans evidence suggest that paracetamol can affect emotions, 405 while in vivo studies have shown paracetamol to affect memory. 406 These are affects not seen with other COX inhibitors, and serve as possible examples of cannabinoid receptor-mediated effects of paracetamol. In conclusion, while there is now clinical evidence for the interaction between paracetamol and FAAH, it remains to be determined as to whether the cannabinoid receptors are integral to its action in humans.

5.4. The therapeutic potential of a FAAH inhibitor

In Chapter 3 FAAH inhibitors that form a covalent attachment to the enzyme’s active site were shown to increase the levels of AEA, OEA and PEA significantly higher than inhibitors that form non-covalent attachments. Covalent attachments are now clearly the preferred mechanism for FAAH inhibitors, with PF-04457845, the first FAAH

169 Chapter 5 inhibitor to enter clinical trials, forming a covalent bond. The FAAH substrates AEA,

OEA and PEA, were investigated in Chapter 3 as a pharmacodynamic measurement to prove the action of potential inhibitors. The first clinical trial of PF-04457845 measured the FAAH substrates in plasma as a pharmacodynamic parameter.184 The study showed that the inhibitor, even at low doses, was capable of maintaining increased substrate levels and inhibition of FAAH activity (>97%) for several days after dosage.

Even after the inhibitor was no longer detectable in the body, FAAH activity was still close to 0%. Therefore the inhibitor’s pharmacodynamics outlasted its pharmacokinetics.

5.4.1. FAAH inhibitors as analgesics

Now that potent, selective and well-tolerated FAAH inhibitors have been developed, their therapeutic potential remains to be determined. While FAAH inhibitors were originally designed for analgesia, the market for novel analgesics is already a crowded one. COX inhibitors count for 68% of the analgesics used for chronic pain, with the opioids counting for 28%. 3 In Chapter 2 it was shown that FAAH inhibitors can inhibit

PGE2 synthesis in activated macrophages, but not with the same potency as other

COX inhibitors. It remains to be seen as to whether a FAAH inhibitor would be capable of competing with the already available analgesics, such as the opioid agonists, COX inhibitors and paracetamol.

An issue only recently highlighted is the AEA occupancy of cannabinoid receptors following FAAH inhibition. Despite >97% inhibition of FAAH activity by PF-04457845, the occupancy of the CB1 and CB2 receptors by AEA is increased by only 0.1% of that achieved by therapeutic doses of exogenous cannabinoids.407 FAAH inhibitors were originally seen as an alternative to synthetic cannabinoids as a means of activating the cannabinoid receptors. However it now appears that the cannabinoid receptors cannot be significantly activated by FAAH inhibitors in humans for analgesia. This does not mean that the therapeutic potential of a FAAH inhibitor lies outside of analgesia.

Currently FAAH inhibition has only been investigated clinically in relation to osteoarthritis pain.407 FAAH inhibitors may be more suited to other pain types. In vivo studies have shown that FAAH inhibition produces analgesia not just via the

170 Chapter 5 cannabinoid receptors, but also via PPAR-α.236 Therefore OEA and PEA are just as integral to analgesia as AEA, meaning that occupancy of the cannabinoid receptors may not be essential.

It may be possible for the therapeutic potential of a FAAH inhibitor to be enhanced by simultaneously targeting other pathways. FAAH could be inhibited alongside either AEA reuptake inhibition or CB2 activation. This would allow for the PPAR-α dependent anti- inflammatory effect of increased OEA and PEA to be complimented by increased activation of the cannabinoid receptors, either by AEA or a synthetic agonist. It is not unlike analgesics to target multiple pathways, such as paracetamol, which interacts with the serotonin, prostaglandin and endocannabinoid systems. Furthermore analgesics with different mechanisms are often dosed together, such as ibuprofen and codeine. The targeting of FAAH alongside another pathway could be utilised by either a single drug with dual selectivity for two targets or by simultaneously dosing a FAAH inhibitor with another drug. It has already been suggested that FAAH inhibitors be developed that with dual selectivity for either COX or TRPV1. 408 Therefore by targeting multiple pathways there still may be potential for FAAH in analgesia.

The finding that paracetamol uses FAAH to produce an active metabolite potentially opens up a new approach to targeting the enzyme. Perhaps instead of drugs being developed that inhibit FAAH, there is now a need for a drug that FAAH can conjugate fatty acids to, for the formation of analgesic active metabolites. In doing so such a drug might inadvertently block AEA, OEA and PEA hydrolysis, and thus have a similar action to a FAAH inhibitor.

5.4.2. Alternative uses of FAAH inhibitors

Outside of analgesia FAAH inhibitors are potentially well suited to numerous disorders of the CNS. FAAH inhibitors have been shown/suggested in in vivo studies to be an effective therapeutic for Alzheimer’s disease, 409 Parkinson’s disease, 410 multiple sclerosis, 411 anxiety 412 and depression. 413 To further investigate the potential of FAAH as a target for Alzheimer’s disease, inhibitors with dual selectivity for cholinesterase and FAAH have been recently developed as a means to increasing neuronal

171 Chapter 5 transmission while reducing neuroinflammation.414 Outside of disorders of the CNS,

FAAH inhibitors possibly also have potential in the cardiovascular system. AEA has been shown to cause vasodilation, modulate blood flow and pressure, and to reduce heart rate. 415 FAAH inhibitors have been demonstrated to normalise cardiovascular function in an in vivo model of hypertension. 416

5.4.3. Potential adverse effects of FAAH inhibitors

While not sharing the adverse effects seen with the cannabinoids, FAAH may possibly come with its own associated risks. Evidence suggests that FAAH inhibition induces sleep in vivo by both cannabinoid receptor and non-cannabinoid receptor mechanisms.417 This suggests that FAAH inhibitors could possibly be used in the treatment of sleep disorders. Whether this effect is seen in humans needs to be investigated as if FAAH inhibitors do cause drowsiness, then this may lessen their appeal. Importantly embryo implantation and development has been shown in vivo to partly involve a FAAH-mediated process involving regulation of AEA.418 Furthermore a clinical trial of pregnant women showed that women with low lymphocyte FAAH activity had miscarriages. 419 It remains to be determined as to whether FAAH inhibitors would have an adverse effect in pregnancy, which again would lessen its appeal as a novel analgesic.

5.4.4. Conclusion

Potent, well tolerated FAAH inhibitors are now available. Whether FAAH inhibitors will become a routinely used medicine, and if so for what conditions, is still uncertain.

FAAH inhibitors have the already available and routinely used analgesics to compete with. Furthermore it is unable to inhibit prostaglandin E2 synthesis as potently as COX inhibitors and clinically meaningful analgesia may not be obtained due to low cannabinoid receptor occupancies. There are possible routes for addressing these issues, such as inhibitors with dual selectivity for a second target. Even if FAAH inhibitors are not used as analgesics, they may have potential in disorders of the CNS and cardiovascular system. Before they can be used clinically, outstanding issues of whether FAAH inhibitors cause drowsiness and effect pregnancy need to be addressed.

172 References

References

173 References

[1] Koshland,D.E. The seven pillars of life. Science 295 , 2215-2216 (2002).

[2] Mannes,A. & Iadarola,M. Potential downsides of perfect pain relief. Nature 446 , 24 (2007).

[3] Breivik,H., Collett,B., Ventafridda,V., Cohen,R., & Gallacher,D. Survey of chronic pain in Europe: prevalence, impact on daily life, and treatment. Eur. J. Pain 10 , 287-333 (2006).

[4] Dworkin,R.H. et al. Advances in neuropathic pain: diagnosis, mechanisms, and treatment recommendations. Arch. Neurol. 60 , 1524-1534 (2003).

[5] Scholz,J. & Woolf,C.J. Can we conquer pain? Nat. Neurosci. 5 Suppl , 1062-1067 (2002).

[6] Michael-Titus,A., Revest,P., & Shortland,P. Pain and Analgesia in The Nervous System (Elsevier, 2010).

[7] Voscopoulos,C. & Lema,M. When does acute pain become chronic? Br J Anaesth. 105 Suppl 1 , i69-i85 (2010).

[8] Michael-Titus,A., Revest,P., & Shortland,P. Elements of Cellular and Molecular Neuroscience in The Nervous System (Elsevier, 2010).

[9] Julius,D. & Basbaum,A.I. Molecular mechanisms of nociception. Nature 413 , 203-210 (2001).

[10] Purves,D., Augustine,G.J., & Fitzpatrick,D. Neuroscience (Sunderland (MA),2001).

[11] Woolf,C.J. & Ma,Q. Nociceptors--noxious stimulus detectors. Neuron 55 , 353-364 (2007).

[12] Rather,L.J. Disturbance of function (functio laesa): the legendary fifth cardinal sign of inflammation, added by Galen to the four cardinal signs of Celsus. Bull. N. Y. Acad. Med. 47 , 303-322 (1971).

[13] Gilroy,D.W., Lawrence,T., Perretti,M., & Rossi,A.G. Inflammatory resolution: new opportunities for drug discovery. Nat. Rev. Drug Discov. 3, 401-416 (2004).

[14] Matsumoto,H. et al. Concordant induction of prostaglandin E2 synthase with cyclooxygenase-2 leads to preferred production of prostaglandin E2 over thromboxane and prostaglandin D2 in lipopolysaccharide-stimulated rat peritoneal macrophages. Biochem. Biophys. Res. Commun. 230 , 110-114 (1997).

[15] Murakami,M. et al. Regulation of prostaglandin E2 biosynthesis by inducible membrane- associated prostaglandin E2 synthase that acts in concert with cyclooxygenase-2. J. Biol. Chem. 275 , 32783-32792 (2000).

[16] Bingham,C.O., III & Austen,K.F. Phospholipase A2 enzymes in eicosanoid generation. Proc. Assoc. Am. Physicians 111 , 516-524 (1999).

[17] Bonventre,J.V. Phospholipase A2 and signal transduction. J Am. Soc. Nephrol. 3, 128- 150 (1992).

[18] Naraba,H. et al. Segregated coupling of phospholipases A2, cyclooxygenases, and terminal prostanoid synthases in different phases of prostanoid biosynthesis in rat peritoneal macrophages. J. Immunol. 160 , 2974-2982 (1998).

[19] Dennis,E.A. Diversity of group types, regulation, and function of phospholipase A2. J. Biol. Chem. 269 , 13057-13060 (1994).

[20] Yao,C. et al. Prostaglandin E2-EP4 signaling promotes immune inflammation through Th1 cell differentiation and Th17 cell expansion. Nat. Med. 15 , 633-640 (2009).

[21] Williams,T.J. Prostaglandin E2, prostaglandin I2 and the vascular changes of inflammation. Br J Pharmacol. 65 , 517-524 (1979).

[22] Medzhitov,R. Origin and physiological roles of inflammation. Nature 454 , 428-435 (2008).

174 References

[23] Banks,W.A., Kastin,A.J., & Gutierrez,E.G. Penetration of interleukin-6 across the murine blood-brain barrier. Neurosci. Lett. 179 , 53-56 (1994).

[24] Bossaller,C. et al. In vivo measurement of endothelium-dependent vasodilation with substance P in man. Herz 17 , 284-290 (1992).

[25] Funk,C.D. Prostaglandins and leukotrienes: advances in eicosanoid biology. Science 294 , 1871-1875 (2001).

[26] Fujii,E., Irie,K., Uchida,Y., Tsukahara,F., & Muraki,T. Possible role of nitric oxide in 5- hydroxytryptamine-induced increase in vascular permeability in mouse skin. Naunyn Schmiedebergs Arch. Pharmacol. 350 , 361-364 (1994).

[27] Rajakariar,R., Yaqoob,M.M., & Gilroy,D.W. COX-2 in inflammation and resolution. Mol. Interv. 6, 199-207 (2006).

[28] Majno,G. & Joris,I. Cells, Tissues, and Disease: Principles of General Pathology (Oxford University Press, New York, 2004).

[29] Palmblad,J. et al. Leukotriene B4 is a potent and stereospecific stimulator of neutrophil chemotaxis and adherence. Blood 58 , 658-661 (1981).

[30] Hickey,M.J. & Kubes,P. Intravascular immunity: the host-pathogen encounter in blood vessels. Nat. Rev. Immunol. 9, 364-375 (2009).

[31] Mosser,D.M. & Edwards,J.P. Exploring the full spectrum of macrophage activation. Nat. Rev. Immunol. 8, 958-969 (2008).

[32] Gordon,S. & Taylor,P.R. Monocyte and macrophage heterogeneity. Nat. Rev. Immunol. 5, 953-964 (2005).

[33] Bystrom,J. et al. Resolution-phase macrophages possess a unique inflammatory phenotype that is controlled by cAMP. Blood 112 , 4117-4127 (2008).

[34] Fujihara,M. et al. Molecular mechanisms of macrophage activation and deactivation by lipopolysaccharide: roles of the receptor complex. Pharmacol. Ther. 100 , 171-194 (2003).

[35] Ayoub,S.S., Botting,R.M., Joshi,A.N., Seed,M.P., & Colville-Nash,P.R. Activation of macrophage peroxisome proliferator-activated receptor-gamma by diclofenac results in the induction of cyclooxygenase-2 protein and the synthesis of anti-inflammatory cytokines. Mol. Cell Biochem. 327 , 101-110 (2009).

[36] Samuelsson,B., Morgenstern,R., & Jakobsson,P.J. Membrane prostaglandin E synthase- 1: a novel therapeutic target. Pharmacol. Rev. 59 , 207-224 (2007).

[37] England,S., Bevan,S., & Docherty,R.J. PGE2 modulates the tetrodotoxin-resistant sodium current in neonatal rat dorsal root ganglion neurones via the cyclic AMP-protein kinase A cascade. J. Physiol 495 ( Pt 2) , 429-440 (1996).

[38] Moriyama,T. et al. Sensitization of TRPV1 by EP1 and IP reveals peripheral nociceptive mechanism of prostaglandins. Mol. Pain 1, 3 (2005).

[39] Gold,M.S., Reichling,D.B., Shuster,M.J., & Levine,J.D. Hyperalgesic agents increase a tetrodotoxin-resistant Na+ current in nociceptors. Proc. Natl. Acad. Sci. U. S. A 93 , 1108-1112 (1996).

[40] Ek,M. et al. Inflammatory response: pathway across the blood-brain barrier. Nature 410 , 430-431 (2001).

[41] Samad,T.A. et al. Interleukin-1beta-mediated induction of Cox-2 in the CNS contributes to inflammatory pain hypersensitivity. Nature 410 , 471-475 (2001).

[42] Zeilhofer,H.U. & Brune,K. Analgesic strategies beyond the inhibition of cyclooxygenases. Trends Pharmacol. Sci. 27 , 467-474 (2006).

[43] Lawrence,T. & Gilroy,D.W. Chronic inflammation: a failure of resolution? Int. J Exp. Pathol. 88 , 85-94 (2007).

175 References

[44] Manderson,A.P., Botto,M., & Walport,M.J. The role of complement in the development of systemic lupus erythematosus. Annu. Rev. Immunol. 22 , 431-456 (2004).

[45] Serhan,C.N. et al. Resolution of inflammation: state of the art, definitions and terms. FASEB J 21 , 325-332 (2007).

[46] Savill,J. & Fadok,V. Corpse clearance defines the meaning of cell death. Nature 407 , 784-788 (2000).

[47] Hughes,J. et al. Neutrophil fate in experimental glomerular capillary injury in the rat. Emigration exceeds in situ clearance by apoptosis. Am. J Pathol. 150 , 223-234 (1997).

[48] Bellingan,G.J., Caldwell,H., Howie,S.E., Dransfield,I., & Haslett,C. In vivo fate of the inflammatory macrophage during the resolution of inflammation: inflammatory macrophages do not die locally, but emigrate to the draining lymph nodes. J Immunol. 157 , 2577-2585 (1996).

[49] Rajakariar,R. et al. Novel biphasic role for lymphocytes revealed during resolving inflammation. Blood 111 , 4184-4192 (2008).

[50] Gilroy,D.W. Eicosanoids and the endogenous control of acute inflammatory resolution. Int. J. Biochem. Cell Biol. 42 , 524-528 (2010).

[51] Serhan,C.N. et al. Design of lipoxin A4 stable analogs that block transmigration and adhesion of human neutrophils. Biochemistry 34 , 14609-14615 (1995).

[52] Takano,T., Clish,C.B., Gronert,K., Petasis,N., & Serhan,C.N. Neutrophil-mediated changes in vascular permeability are inhibited by topical application of aspirin-triggered 15-epi-lipoxin A4 and novel lipoxin B4 stable analogues. J Clin Invest 101 , 819-826 (1998).

[53] Godson,C. et al. Cutting edge: lipoxins rapidly stimulate nonphlogistic phagocytosis of apoptotic neutrophils by monocyte-derived macrophages. J Immunol. 164 , 1663-1667 (2000).

[54] Aronoff,D.M., Canetti,C., & Peters-Golden,M. Prostaglandin E2 inhibits alveolar macrophage phagocytosis through an E-prostanoid 2 receptor-mediated increase in intracellular cyclic AMP. J Immunol. 173 , 559-565 (2004).

[55] Takayama,K. et al. Prostaglandin E2 suppresses chemokine production in human macrophages through the EP4 receptor. J Biol. Chem. 277 , 44147-44154 (2002).

[56] Stables,M.J. & Gilroy,D.W. Old and new generation lipid mediators in acute inflammation and resolution. Prog. Lipid Res. 50 , 35-51 (2011).

[57] Gilroy,D.W. et al. Inducible cyclooxygenase may have anti-inflammatory properties. Nat. Med. 5, 698-701 (1999).

[58] Gilroy,D.W., Newson,J., Sawmynaden,P., Willoughby,D.A., & Croxtall,J.D. A novel role for phospholipase A2 isoforms in the checkpoint control of acute inflammation. FASEB J. 18 , 489-498 (2004).

[59] Woolf,C.J. & Mannion,R.J. Neuropathic pain: aetiology, symptoms, mechanisms, and management. Lancet 353 , 1959-1964 (1999).

[60] Wieseler-Frank,J., Maier,S.F., & Watkins,L.R. Glial activation and pathological pain. Neurochem. Int. 45 , 389-395 (2004).

[61] Dworkin,R.H. An overview of neuropathic pain: syndromes, symptoms, signs, and several mechanisms. Clin. J. Pain 18 , 343-349 (2002).

[62] Turk,D.C. & Okifuji,A. Pain terms and taxonomies in Bonica's management of pain (ed. Loeser,J.D.) 18-25 (Lippincott Williams & Wilkins, 2001).

[63] Vane,J.R., Bakhle,Y.S., & Botting,R.M. Cyclooxygenases 1 and 2. Annu. Rev. Pharmacol. Toxicol. 38 , 97-120 (1998).

[64] Hartner,A., Pahl,A., Brune,K., & Goppelt-Struebe,M. Upregulation of cyclooxygenase-1 and the PGE2 receptor EP2 in rat and human mesangioproliferative glomerulonephritis. Inflamm. Res. 49 , 345-354 (2000).

176 References

[65] Svensson,C.I. & Yaksh,T.L. The spinal phospholipase-cyclooxygenase-prostanoid cascade in nociceptive processing. Annu. Rev. Pharmacol. Toxicol. 42 , 553-583 (2002).

[66] Ghilardi,J.R., Svensson,C.I., Rogers,S.D., Yaksh,T.L., & Mantyh,P.W. Constitutive spinal cyclooxygenase-2 participates in the initiation of tissue injury-induced hyperalgesia. J. Neurosci. 24 , 2727-2732 (2004).

[67] Murakami,M., Nakatani,Y., Tanioka,T., & Kudo,I. Prostaglandin E synthase. Prostaglandins Other Lipid Mediat. 68-69 , 383-399 (2002).

[68] Scholich,K. & Geisslinger,G. Is mPGES-1 a promising target for pain therapy? Trends Pharmacol. Sci. 27 , 399-401 (2006).

[69] Ma,W. & Quirion,R. Does COX2-dependent PGE2 play a role in neuropathic pain? Neurosci. Lett. 437 , 165-169 (2008).

[70] Green,G.A. Understanding NSAIDs: from aspirin to COX-2. Clin. Cornerstone. 3, 50-60 (2001).

[71] Whittle,B.J. COX-1 and COX-2 products in the gut: therapeutic impact of COX-2 inhibitors. Gut 47 , 320-325 (2000).

[72] Page,J. & Henry,D. Consumption of NSAIDs and the development of congestive heart failure in elderly patients: an underrecognized public health problem. Arch. Intern. Med. 160 , 777-784 (2000).

[73] Tegeder,I. & Geisslinger,G. Cardiovascular risk with cyclooxygenase inhibitors: general problem with substance specific differences? Naunyn Schmiedebergs Arch. Pharmacol. 373 , 1-17 (2006).

[74] Mukherjee,D., Nissen,S.E., & Topol,E.J. Risk of cardiovascular events associated with selective COX-2 inhibitors. JAMA 286 , 954-959 (2001).

[75] Bing,R.J. & Lomnicka,M. Why do cyclo-oxygenase-2 inhibitors cause cardiovascular events? J Am. Coll. Cardiol. 39 , 521-522 (2002).

[76] Mitchell,J.A., Akarasereenont,P., Thiemermann,C., Flower,R.J., & Vane,J.R. Selectivity of nonsteroidal antiinflammatory drugs as inhibitors of constitutive and inducible cyclooxygenase. Proc. Natl. Acad. Sci. U. S. A 90 , 11693-11697 (1993).

[77] Bie,B. & Pan,Z.Z. Trafficking of central opioid receptors and descending pain inhibition. Mol. Pain 3, 37 (2007).

[78] Ersek,M., Cherrier,M.M., Overman,S.S., & Irving,G.A. The cognitive effects of opioids. Pain Manag. Nurs. 5, 75-93 (2004).

[79] Matthes,H.W. et al. Loss of morphine-induced analgesia, reward effect and withdrawal symptoms in mice lacking the mu-opioid-receptor gene. Nature 383 , 819-823 (1996).

[80] Rhen,T. & Cidlowski,J.A. Antiinflammatory action of glucocorticoids--new mechanisms for old drugs. N. Engl. J. Med. 353 , 1711-1723 (2005).

[81] Romundstad,L. & Stubhaug,A. Glucocorticoids for acute and persistent postoperative neuropathic pain: what is the evidence? Anesthesiology 107 , 371-373 (2007).

[82] Schacke,H., Docke,W.D., & Asadullah,K. Mechanisms involved in the side effects of glucocorticoids. Pharmacol. Ther. 96 , 23-43 (2002).

[83] Rang,H.P., Dale,M.M., Ritter,J.M., & Moore,P.K. General Anaesthetic Agents in Pharmacology (Churchill Livingstone, 2003).

[84] Tremont-Lukats,I.W., Challapalli,V., McNicol,E.D., Lau,J., & Carr,D.B. Systemic administration of local anesthetics to relieve neuropathic pain: a systematic review and meta-analysis. Anesth. Analg. 101 , 1738-1749 (2005).

[85] Petrenko,A.B., Yamakura,T., Baba,H., & Shimoji,K. The role of N-methyl-D-aspartate (NMDA) receptors in pain: a review. Anesth. Analg. 97 , 1108-1116 (2003).

177 References

[86] Tatsumi,M., Groshan,K., Blakely,R.D., & Richelson,E. Pharmacological profile of antidepressants and related compounds at human monoamine transporters. Eur. J. Pharmacol. 340 , 249-258 (1997).

[87] Mico,J.A., Ardid,D., Berrocoso,E., & Eschalier,A. Antidepressants and pain. Trends Pharmacol. Sci. 27 , 348-354 (2006).

[88] Gajraj,N.M. Pregabalin: its pharmacology and use in pain management. Anesth. Analg. 105 , 1805-1815 (2007).

[89] Gee,N.S. et al. The novel anticonvulsant drug, gabapentin (Neurontin), binds to the alpha2delta subunit of a calcium channel. J. Biol. Chem. 271 , 5768-5776 (1996).

[90] Dickenson,A.H., Matthews,E.A., & Suzuki,R. Neurobiology of neuropathic pain: mode of action of anticonvulsants. Eur. J. Pain 6 Suppl A , 51-60 (2002).

[91] Toth,P.P. & Urtis,J. Commonly used muscle relaxant therapies for acute low back pain: a review of carisoprodol, cyclobenzaprine hydrochloride, and metaxalone. Clin. Ther. 26 , 1355-1367 (2004).

[92] Joshi,G.P. Multimodal analgesia techniques and postoperative rehabilitation. Anesthesiol. Clin North America. 23 , 185-202 (2005).

[93] World Health Organization Cancer pain relief (World Health Organization, Geneva, 1996).

[94] Zech,D.F., Grond,S., Lynch,J., Hertel,D., & Lehmann,K.A. Validation of World Health Organization Guidelines for cancer pain relief: a 10-year prospective study. Pain 63 , 65- 76 (1995).

[95] Beral,V. & Peto,R. UK cancer survival statistics. BMJ 341 , c4112 (2010).

[96] Zias,J. et al. Early medical use of cannabis. Nature 363 , 215 (1993).

[97] Gaoni,Y. & Mechoulam,R. The isolation and structure of delta-1-tetrahydrocannabinol and other neutral cannabinoids from hashish. J Am. Chem. Soc. 93 , 217-224 (1971).

[98] Matsuda,L.A., Lolait,S.J., Brownstein,M.J., Young,A.C., & Bonner,T.I. Structure of a cannabinoid receptor and functional expression of the cloned cDNA. Nature 346 , 561- 564 (1990).

[99] Munro,S., Thomas,K.L., & bu-Shaar,M. Molecular characterization of a peripheral receptor for cannabinoids. Nature 365 , 61-65 (1993).

[100] Devane,W.A. et al. Isolation and structure of a brain constituent that binds to the cannabinoid receptor. Science 258 , 1946-1949 (1992).

[101] Mechoulam,R. et al. Identification of an endogenous 2-monoglyceride, present in canine gut, that binds to cannabinoid receptors. Biochem. Pharmacol. 50 , 83-90 (1995).

[102] Cravatt,B.F. et al. Molecular characterization of an enzyme that degrades neuromodulatory fatty-acid amides. Nature 384 , 83-87 (1996).

[103] Dinh,T.P., Freund,T.F., & Piomelli,D. A role for monoglyceride lipase in 2- arachidonoylglycerol inactivation. Chem. Phys. Lipids 121 , 149-158 (2002).

[104] Fu,J. et al. A catalytically silent FAAH-1 variant drives anandamide transport in neurons. Nat. Neurosci. (2011).

[105] Devane,W.A., Dysarz,F.A., III, Johnson,M.R., Melvin,L.S., & Howlett,A.C. Determination and characterization of a cannabinoid receptor in rat brain. Mol. Pharmacol. 34 , 605-613 (1988).

[106] Hoehe,M.R. et al. Genetic and physical mapping of the human cannabinoid receptor gene to chromosome 6q14-q15. New Biol. 3, 880-885 (1991).

[107] Tsou,K., Brown,S., Sanudo-Pena,M.C., Mackie,K., & Walker,J.M. Immunohistochemical distribution of cannabinoid CB1 receptors in the rat central nervous system. Neuroscience 83 , 393-411 (1998).

178 References

[108] Howlett,A.C. et al. International Union of Pharmacology. XXVII. Classification of cannabinoid receptors. Pharmacol. Rev. 54 , 161-202 (2002).

[109] Egertova,M. & Elphick,M.R. Localisation of cannabinoid receptors in the rat brain using antibodies to the intracellular C-terminal tail of CB. J. Comp Neurol. 422 , 159-171 (2000).

[110] Shim,J.Y., Welsh,W.J., & Howlett,A.C. Homology model of the CB1 cannabinoid receptor: sites critical for nonclassical cannabinoid agonist interaction. Biopolymers 71 , 169-189 (2003).

[111] Karsak,M. et al. Cannabinoid receptor type 2 gene is associated with human osteoporosis. Hum. Mol. Genet. 14 , 3389-3396 (2005).

[112] Galiegue,S. et al. Expression of central and peripheral cannabinoid receptors in human immune tissues and leukocyte subpopulations. Eur. J. Biochem. 232 , 54-61 (1995).

[113] Cabral,G.A., Raborn,E.S., Griffin,L., Dennis,J., & Marciano-Cabral,F. CB2 receptors in the brain: role in central immune function. Br. J. Pharmacol. 153 , 240-251 (2008).

[114] Van,S. et al. Identification and functional characterization of brainstem cannabinoid CB2 receptors. Science 310 , 329-332 (2005).

[115] Klegeris,A., Bissonnette,C.J., & McGeer,P.L. Reduction of human monocytic cell neurotoxicity and cytokine secretion by ligands of the cannabinoid-type CB2 receptor. Br. J. Pharmacol. 139 , 775-786 (2003).

[116] Chang,Y.H., Lee,S.T., & Lin,W.W. Effects of cannabinoids on LPS-stimulated inflammatory mediator release from macrophages: involvement of eicosanoids. J. Cell Biochem. 81 , 715-723 (2001).

[117] Gerard,C.M., Mollereau,C., Vassart,G., & Parmentier,M. Molecular cloning of a human cannabinoid receptor which is also expressed in testis. Biochem. J. 279 ( Pt 1) , 129-134 (1991).

[118] Xu,W. et al. Purification and mass spectroscopic analysis of human CB1 cannabinoid receptor functionally expressed using the baculovirus system. J. Pept. Res. 66 , 138-150 (2005).

[119] Cabral,G.A. & Griffin-Thomas,L. Emerging role of the cannabinoid receptor CB2 in immune regulation: therapeutic prospects for neuroinflammation. Expert. Rev. Mol. Med. 11 , e3 (2009).

[120] Filppula,S. et al. Purification and mass spectroscopic analysis of human CB2 cannabinoid receptor expressed in the baculovirus system. J. Pept. Res. 64 , 225-236 (2004).

[121] Song,C. & Howlett,A.C. Rat brain cannabinoid receptors are N-linked glycosylated proteins. Life Sci. 56 , 1983-1989 (1995).

[122] Clayton,N., Marshall,F.H., Bountra,C., & O'Shaughnessy,C.T. CB1 and CB2 cannabinoid receptors are implicated in inflammatory pain. Pain 96 , 253-260 (2002).

[123] Guindon,J. & Hohmann,A.G. Cannabinoid CB2 receptors: a therapeutic target for the treatment of inflammatory and neuropathic pain. Br. J. Pharmacol. 153 , 319-334 (2008).

[124] Hunyady,L. & Turu,G. Signal transduction of the CB1 cannabinoid receptor. J. Mol. Endocrinol. (2009).

[125] Ibrahim,M.M. et al. CB2 cannabinoid receptor activation produces antinociception by stimulating peripheral release of endogenous opioids. Proc. Natl. Acad. Sci. U. S. A 102 , 3093-3098 (2005).

[126] McAllister,S.D. & Glass,M. CB(1) and CB(2) receptor-mediated signalling: a focus on endocannabinoids. Prostaglandins Leukot. Essent. Fatty Acids 66 , 161-171 (2002).

[127] Pan,H.L. et al. Modulation of pain transmission by G-protein-coupled receptors. Pharmacol. Ther. 117 , 141-161 (2008).

179 References

[128] Sagar,D.R. et al. Dynamic regulation of the endocannabinoid system: implications for analgesia. Mol. Pain 5, 59 (2009).

[129] Smart,D. et al. The endogenous lipid anandamide is a full agonist at the human vanilloid receptor (hVR1). Br. J. Pharmacol. 129 , 227-230 (2000).

[130] Ross,R.A. Anandamide and vanilloid TRPV1 receptors. Br. J. Pharmacol. 140 , 790-801 (2003).

[131] Sugiura,T. et al. 2-Arachidonoylglycerol: a possible endogenous cannabinoid receptor ligand in brain. Biochem. Biophys. Res. Commun. 215 , 89-97 (1995).

[132] Panikashvili,D. et al. An endogenous cannabinoid (2-AG) is neuroprotective after brain injury. Nature 413 , 527-531 (2001).

[133] Lambert,D.M. & Muccioli,G.G. Endocannabinoids and related N-acylethanolamines in the control of appetite and energy metabolism: emergence of new molecular players. Curr. Opin. Clin. Nutr. Metab Care 10 , 735-744 (2007).

[134] Laine,K., Jarvinen,K., Mechoulam,R., Breuer,A., & Jarvinen,T. Comparison of the enzymatic stability and intraocular pressure effects of 2-arachidonylglycerol and noladin ether, a novel putative endocannabinoid. Invest Ophthalmol. Vis. Sci. 43 , 3216-3222 (2002).

[135] Hanus,L. et al. 2-arachidonyl glyceryl ether, an endogenous agonist of the cannabinoid CB1 receptor. Proc. Natl. Acad. Sci. U. S. A 98 , 3662-3665 (2001).

[136] Fezza,F. et al. Noladin ether, a putative novel endocannabinoid: inactivation mechanisms and a sensitive method for its quantification in rat tissues. FEBS Lett. 513 , 294-298 (2002).

[137] Oka,S. et al. Ether-linked analogue of 2-arachidonoylglycerol (noladin ether) was not detected in the brains of various mammalian species. J. Neurochem. 85 , 1374-1381 (2003).

[138] Bisogno,T. et al. N-acyl-dopamines: novel synthetic CB(1) cannabinoid-receptor ligands and inhibitors of anandamide inactivation with cannabimimetic activity in vitro and in vivo. Biochem. J. 351 Pt 3 , 817-824 (2000).

[139] Huang,S.M. et al. An endogenous capsaicin-like substance with high potency at recombinant and native vanilloid VR1 receptors. Proc. Natl. Acad. Sci. U. S. A 99 , 8400- 8405 (2002).

[140] Porter,A.C. et al. Characterization of a novel endocannabinoid, virodhamine, with antagonist activity at the CB1 receptor. J. Pharmacol. Exp. Ther. 301 , 1020-1024 (2002).

[141] Richardson,D., Ortori,C.A., Chapman,V., Kendall,D.A., & Barrett,D.A. Quantitative profiling of endocannabinoids and related compounds in rat brain using liquid chromatography-tandem electrospray ionization mass spectrometry. Anal. Biochem. 360 , 216-226 (2007).

[142] Thomas,A., Hopfgartner,G., Giroud,C., & Staub,C. Quantitative and qualitative profiling of endocannabinoids in human plasma using a triple quadrupole linear ion trap mass spectrometer with liquid chromatography. Rapid Commun. Mass Spectrom. 23 , 629-638 (2009).

[143] Marczylo,T.H., Lam,P.M., Nallendran,V., Taylor,A.H., & Konje,J.C. A solid-phase method for the extraction and measurement of anandamide from multiple human biomatrices. Anal. Biochem. 384 , 106-113 (2009).

[144] Di,M., V et al. Hypolocomotor effects in rats of capsaicin and two long chain capsaicin homologues. Eur. J. Pharmacol. 420 , 123-131 (2001).

[145] Giuffrida,A. et al. Dopamine activation of endogenous cannabinoid signaling in dorsal striatum. Nat. Neurosci. 2, 358-363 (1999).

[146] Kim,J., Isokawa,M., Ledent,C., & Alger,B.E. Activation of muscarinic acetylcholine receptors enhances the release of endogenous cannabinoids in the hippocampus. J. Neurosci. 22 , 10182-10191 (2002).

180 References

[147] Stella,N. & Piomelli,D. Receptor-dependent formation of endogenous cannabinoids in cortical neurons. Eur. J. Pharmacol. 425 , 189-196 (2001).

[148] Sugiura,T., Kobayashi,Y., Oka,S., & Waku,K. Biosynthesis and degradation of anandamide and 2-arachidonoylglycerol and their possible physiological significance. Prostaglandins Leukot. Essent. Fatty Acids 66 , 173-192 (2002).

[149] Placzek,E.A., Okamoto,Y., Ueda,N., & Barker,E.L. Membrane microdomains and metabolic pathways that define anandamide and 2-arachidonyl glycerol biosynthesis and breakdown. Neuropharmacology 55 , 1095-1104 (2008).

[150] Okamoto,Y., Morishita,J., Tsuboi,K., Tonai,T., & Ueda,N. Molecular characterization of a phospholipase D generating anandamide and its congeners. J. Biol. Chem. 279 , 5298- 5305 (2004).

[151] Ueda,N., Kurahashi,Y., Yamamoto,S., & Tokunaga,T. Partial purification and characterization of the porcine brain enzyme hydrolyzing and synthesizing anandamide. J. Biol. Chem. 270 , 23823-23827 (1995).

[152] Placzek,E.A., Okamoto,Y., Ueda,N., & Barker,E.L. Mechanisms for recycling and biosynthesis of endogenous cannabinoids anandamide and 2-arachidonylglycerol. J. Neurochem. 107 , 987-1000 (2008).

[153] McFarland,M.J., Terebova,E.A., & Barker,E.L. Detergent-resistant membrane microdomains in the disposition of the lipid signaling molecule anandamide. AAPS. J. 8, E95-100 (2006).

[154] Stella,N., Schweitzer,P., & Piomelli,D. A second endogenous cannabinoid that modulates long-term potentiation. Nature 388 , 773-778 (1997).

[155] Tiger,G., Stenstrom,A., & Fowler,C.J. Pharmacological properties of rat brain fatty acid amidohydrolase in different subcellular fractions using palmitoylethanolamide as substrate. Biochem. Pharmacol. 59 , 647-653 (2000).

[156] Lopez-Moreno,J.A., Echeverry-Alzate,V., & Buhler,K.M. The genetic basis of the endocannabinoid system and drug addiction in humans. J Psychopharmacol. 26 , 133- 143 (2012).

[157] Giang,D.K. & Cravatt,B.F. Molecular characterization of human and mouse fatty acid amide hydrolases. Proc. Natl. Acad. Sci. U. S. A 94 , 2238-2242 (1997).

[158] Rodriguez de,F.F. et al. An anorexic lipid mediator regulated by feeding. Nature 414 , 209-212 (2001).

[159] Lambert,D.M., Vandevoorde,S., Jonsson,K.O., & Fowler,C.J. The palmitoylethanolamide family: a new class of anti-inflammatory agents? Curr. Med. Chem. 9, 663-674 (2002).

[160] Wei,B.Q., Mikkelsen,T.S., McKinney,M.K., Lander,E.S., & Cravatt,B.F. A second fatty acid amide hydrolase with variable distribution among placental mammals. J. Biol. Chem. 281 , 36569-36578 (2006).

[161] Bracey,M.H., Hanson,M.A., Masuda,K.R., Stevens,R.C., & Cravatt,B.F. Structural adaptations in a membrane enzyme that terminates endocannabinoid signaling. Science 298 , 1793-1796 (2002).

[162] Patricelli,M.P. & Cravatt,B.F. Characterization and manipulation of the acyl chain selectivity of fatty acid amide hydrolase. Biochemistry 40, 6107-6115 (2001).

[163] Hu,S.S. et al. The biosynthesis of N-arachidonoyl dopamine (NADA), a putative endocannabinoid and endovanilloid, via conjugation of arachidonic acid with dopamine. Prostaglandins Leukot. Essent. Fatty Acids 81 , 291-301 (2009).

[164] Hogestatt,E.D. et al. Conversion of acetaminophen to the bioactive N-acylphenolamine AM404 via fatty acid amide hydrolase-dependent arachidonic acid conjugation in the nervous system. J. Biol. Chem. 280 , 31405-31412 (2005).

[165] Mileni,M. et al. Structure-guided inhibitor design for human FAAH by interspecies active site conversion. Proc. Natl. Acad. Sci. U. S. A 105 , 12820-12824 (2008).

181 References

[166] Blankman,J.L., Simon,G.M., & Cravatt,B.F. A comprehensive profile of brain enzymes that hydrolyze the endocannabinoid 2-arachidonoylglycerol. Chem. Biol. 14 , 1347-1356 (2007).

[167] Karlsson,M., Contreras,J.A., Hellman,U., Tornqvist,H., & Holm,C. cDNA cloning, tissue distribution, and identification of the catalytic triad of monoglyceride lipase. J. Biol. Chem. 272 , 27218-27223 (1997).

[168] Bhatia,G. et al. A covering method for detecting genetic associations between rare variants and common phenotypes. PLoS. Comput. Biol. 6, e1000954 (2010).

[169] Goparaju,S.K., Ueda,N., Yamaguchi,H., & Yamamoto,S. Anandamide amidohydrolase reacting with 2-arachidonoylglycerol, another cannabinoid receptor ligand. FEBS Lett. 422 , 69-73 (1998).

[170] Bertrand,T. et al. Structural basis for human monoglyceride lipase inhibition. J. Mol. Biol. 396 , 663-673 (2010).

[171] Burstein,S.H., Rossetti,R.G., Yagen,B., & Zurier,R.B. Oxidative metabolism of anandamide. Prostaglandins Other Lipid Mediat. 61 , 29-41 (2000).

[172] Yu,M., Ives,D., & Ramesha,C.S. Synthesis of prostaglandin E2 ethanolamide from anandamide by cyclooxygenase-2. J. Biol. Chem. 272 , 21181-21186 (1997).

[173] Kozak,K.R., Rowlinson,S.W., & Marnett,L.J. Oxygenation of the endocannabinoid, 2- arachidonylglycerol, to glyceryl prostaglandins by cyclooxygenase-2. J. Biol. Chem. 275 , 33744-33749 (2000).

[174] Hampson,A.J. et al. Anandamide hydroxylation by brain lipoxygenase:metabolite structures and potencies at the cannabinoid receptor. Biochim. Biophys. Acta 1259 , 173-179 (1995).

[175] Moody,J.S., Kozak,K.R., Ji,C., & Marnett,L.J. Selective oxygenation of the endocannabinoid 2-arachidonylglycerol by leukocyte-type 12-lipoxygenase. Biochemistry 40 , 861-866 (2001).

[176] Snider,N.T. et al. The endocannabinoid anandamide is a substrate for the human polymorphic cytochrome P450 2D6. J. Pharmacol. Exp. Ther. 327 , 538-545 (2008).

[177] Snider,N.T., Kornilov,A.M., Kent,U.M., & Hollenberg,P.F. Anandamide metabolism by human liver and kidney microsomal cytochrome p450 enzymes to form hydroxyeicosatetraenoic and epoxyeicosatrienoic acid ethanolamides. J. Pharmacol. Exp. Ther. 321 , 590-597 (2007).

[178] Chen,J.K. et al. Identification of novel endogenous cytochrome p450 arachidonate metabolites with high affinity for cannabinoid receptors. J. Biol. Chem. 283 , 24514- 24524 (2008).

[179] Buczynski,M.W. et al. Inflammatory hyperalgesia induces essential bioactive lipid production in the spinal cord. J Neurochem. 114 , 981-993 (2010).

[180] Fraser,T., Tayler,H., & Love,S. Fatty acid composition of frontal, temporal and parietal neocortex in the normal human brain and in Alzheimer's disease. Neurochem. Res. 35 , 503-513 (2010).

[181] Weber,A. et al. Formation of prostamides from anandamide in FAAH knockout mice analyzed by HPLC with tandem mass spectrometry. J Lipid Res. 45 , 757-763 (2004).

[182] Jhaveri,M.D. et al. Inhibition of fatty acid amide hydrolase and cyclooxygenase-2 increases levels of endocannabinoid related molecules and produces analgesia via peroxisome proliferator-activated receptor-alpha in a model of inflammatory pain. Neuropharmacology 55 , 85-93 (2008).

[183] Telleria-Diaz,A. et al. Spinal antinociceptive effects of cyclooxygenase inhibition during inflammation: Involvement of prostaglandins and endocannabinoids. Pain 148 , 26-35 (2010).

[184] Li,G.L. et al. Assessment of the pharmacology and tolerability of PF-04457845, an irreversible inhibitor of fatty acid amide hydrolase-1, in healthy subjects. Br J Clin Pharmacol. (2011).

182 References

[185] Tsuboi,K. et al. Molecular characterization of N-acylethanolamine-hydrolyzing acid amidase, a novel member of the choloylglycine hydrolase family with structural and functional similarity to acid ceramidase. J. Biol. Chem. 280 , 11082-11092 (2005).

[186] Llano,I., Leresche,N., & Marty,A. Calcium entry increases the sensitivity of cerebellar Purkinje cells to applied GABA and decreases inhibitory synaptic currents. Neuron 6, 565-574 (1991).

[187] Vincent,P., Armstrong,C.M., & Marty,A. Inhibitory synaptic currents in rat cerebellar Purkinje cells: modulation by postsynaptic depolarization. J. Physiol 456 , 453-471 (1992).

[188] Pitler,T.A. & Alger,B.E. Postsynaptic spike firing reduces synaptic GABAA responses in hippocampal pyramidal cells. J. Neurosci. 12 , 4122-4132 (1992).

[189] Kreitzer,A.C. & Regehr,W.G. Retrograde inhibition of presynaptic calcium influx by endogenous cannabinoids at excitatory synapses onto Purkinje cells. Neuron 29 , 717- 727 (2001).

[190] Piomelli,D. The molecular logic of endocannabinoid signalling. Nat. Rev. Neurosci. 4, 873-884 (2003).

[191] Fegley,D. et al. Anandamide transport is independent of fatty-acid amide hydrolase activity and is blocked by the hydrolysis-resistant inhibitor AM1172. Proc. Natl. Acad. Sci. U. S. A 101 , 8756-8761 (2004).

[192] Glaser,S.T., Kaczocha,M., & Deutsch,D.G. Anandamide transport: a critical review. Life Sci. 77 , 1584-1604 (2005).

[193] Glaser,S.T. et al. Evidence against the presence of an anandamide transporter. Proc. Natl. Acad. Sci. U. S. A 100 , 4269-4274 (2003).

[194] Bojesen,I.N. & Bojesen,E. Oleic acid binding and transport capacity of human red cell membrane. Acta Physiol Scand. 156 , 501-516 (1996).

[195] Kanai,N. et al. Identification and characterization of a prostaglandin transporter. Science 268 , 866-869 (1995).

[196] Wolf,W.A. & Kuhn,D.M. Modulation of serotonin release. Interactions between the serotonin transporter and autoreceptors. Ann. N. Y. Acad. Sci. 604 , 505-513 (1990).

[197] Moore,S.A. et al. Identification of a high-affinity binding site involved in the transport of endocannabinoids. Proc. Natl. Acad. Sci. U. S. A 102 , 17852-17857 (2005).

[198] Kaczocha,M., Glaser,S.T., & Deutsch,D.G. Identification of intracellular carriers for the endocannabinoid anandamide. Proc. Natl. Acad. Sci. U. S. A 106 , 6375-6380 (2009).

[199] Beltramo,M. & Piomelli,D. Carrier-mediated transport and enzymatic hydrolysis of the endogenous cannabinoid 2-arachidonylglycerol. Neuroreport 11 , 1231-1235 (2000).

[200] Bisogno,T. et al. The uptake by cells of 2-arachidonoylglycerol, an endogenous agonist of cannabinoid receptors. Eur. J. Biochem. 268 , 1982-1989 (2001).

[201] Gulyas,A.I. et al. Segregation of two endocannabinoid-hydrolyzing enzymes into pre- and postsynaptic compartments in the rat hippocampus, cerebellum and amygdala. Eur. J. Neurosci. 20 , 441-458 (2004).

[202] Zimmer,A., Zimmer,A.M., Hohmann,A.G., Herkenham,M., & Bonner,T.I. Increased mortality, hypoactivity, and hypoalgesia in cannabinoid CB1 receptor knockout mice. Proc. Natl. Acad. Sci. U. S. A 96 , 5780-5785 (1999).

[203] Scott,D.A., Wright,C.E., & Angus,J.A. Evidence that CB-1 and CB-2 cannabinoid receptors mediate antinociception in neuropathic pain in the rat. Pain 109 , 124-131 (2004).

[204] Ibrahim,M.M. et al. CB2 cannabinoid receptor mediation of antinociception. Pain 122 , 36-42 (2006).

[205] Buckley,N.E. et al. Immunomodulation by cannabinoids is absent in mice deficient for the cannabinoid CB(2) receptor. Eur. J. Pharmacol. 396 , 141-149 (2000).

183 References

[206] Wotherspoon,G. et al. Peripheral nerve injury induces cannabinoid receptor 2 protein expression in rat sensory neurons. Neuroscience 135 , 235-245 (2005).

[207] Zhang,J. et al. Induction of CB2 receptor expression in the rat spinal cord of neuropathic but not inflammatory chronic pain models. Eur. J. Neurosci. 17 , 2750-2754 (2003).

[208] Quartilho,A. et al. Inhibition of inflammatory hyperalgesia by activation of peripheral CB2 cannabinoid receptors. Anesthesiology 99 , 955-960 (2003).

[209] Nackley,A.G., Makriyannis,A., & Hohmann,A.G. Selective activation of cannabinoid CB(2) receptors suppresses spinal fos protein expression and pain behavior in a rat model of inflammation. Neuroscience 119 , 747-757 (2003).

[210] Ibrahim,M.M. et al. Activation of CB2 cannabinoid receptors by AM1241 inhibits experimental neuropathic pain: pain inhibition by receptors not present in the CNS. Proc. Natl. Acad. Sci. U. S. A 100 , 10529-10533 (2003).

[211] Cheng,Y. & Hitchcock,S.A. Targeting cannabinoid agonists for inflammatory and neuropathic pain. Expert. Opin. Investig. Drugs 16 , 951-965 (2007).

[212] Malan,T.P., Jr. et al. CB2 cannabinoid receptor agonists: pain relief without psychoactive effects? Curr. Opin. Pharmacol. 3, 62-67 (2003).

[213] Staniaszek,L.E., Norris,L.M., Kendall,D.A., Barrett,D.A., & Chapman,V. Effects of COX-2 inhibition on spinal nociception: the role of endocannabinoids. Br. J. Pharmacol. 160 , 669-676 (2010).

[214] Fride,E. & Mechoulam,R. Pharmacological activity of the cannabinoid receptor agonist, anandamide, a brain constituent. Eur. J. Pharmacol. 231 , 313-314 (1993).

[215] Calignano,A., La,R.G., Giuffrida,A., & Piomelli,D. Control of pain initiation by endogenous cannabinoids. Nature 394 , 277-281 (1998).

[216] Ameri,A. The effects of cannabinoids on the brain. Prog. Neurobiol. 58 , 315-348 (1999).

[217] Fride,E. et al. (+)-Cannabidiol analogues which bind cannabinoid receptors but exert peripheral activity only. Eur. J. Pharmacol. 506 , 179-188 (2004).

[218] Long,J.Z. et al. Selective blockade of 2-arachidonoylglycerol hydrolysis produces cannabinoid behavioral effects. Nat. Chem. Biol. 5, 37-44 (2009).

[219] Schlosburg,J.E. et al. Chronic monoacylglycerol lipase blockade causes functional antagonism of the endocannabinoid system. Nat. Neurosci. 13 , 1113-1119 (2010).

[220] Viso,A., Cisneros,J.A., & Ortega-Gutierrez,S. The medicinal chemistry of agents targeting monoacylglycerol lipase. Curr. Top. Med. Chem. 8, 231-246 (2008).

[221] Cravatt,B.F. et al. Supersensitivity to anandamide and enhanced endogenous cannabinoid signaling in mice lacking fatty acid amide hydrolase. Proc. Natl. Acad. Sci. U. S. A 98 , 9371-9376 (2001).

[222] Kathuria,S. et al. Modulation of anxiety through blockade of anandamide hydrolysis. Nat. Med. 9, 76-81 (2003).

[223] Jayamanne,A. et al. Actions of the FAAH inhibitor URB597 in neuropathic and inflammatory chronic pain models. Br. J. Pharmacol. 147 , 281-288 (2006).

[224] Clapper,J.R. et al. Anandamide suppresses pain initiation through a peripheral endocannabinoid mechanism. Nat. Neurosci. (2010).

[225] Ahn,K. et al. Discovery and characterization of a highly selective FAAH inhibitor that reduces inflammatory pain. Chem. Biol. 16, 411-420 (2009).

[226] Holt,S., Comelli,F., Costa,B., & Fowler,C.J. Inhibitors of fatty acid amide hydrolase reduce carrageenan-induced hind paw inflammation in pentobarbital-treated mice: comparison with indomethacin and possible involvement of cannabinoid receptors. Br. J. Pharmacol. 146 , 467-476 (2005).

184 References

[227] Jhaveri,M.D., Richardson,D., Kendall,D.A., Barrett,D.A., & Chapman,V. Analgesic effects of fatty acid amide hydrolase inhibition in a rat model of neuropathic pain. J. Neurosci. 26 , 13318-13327 (2006).

[228] Russo,R. et al. The fatty acid amide hydrolase inhibitor URB597 (cyclohexylcarbamic acid 3'-carbamoylbiphenyl-3-yl ester) reduces neuropathic pain after oral administration in mice. J. Pharmacol. Exp. Ther. 322 , 236-242 (2007).

[229] Chang,L. et al. Inhibition of fatty acid amide hydrolase produces analgesia by multiple mechanisms. Br. J. Pharmacol. 148 , 102-113 (2006).

[230] Clement,A.B., Hawkins,E.G., Lichtman,A.H., & Cravatt,B.F. Increased seizure susceptibility and proconvulsant activity of anandamide in mice lacking fatty acid amide hydrolase. J. Neurosci. 23 , 3916-3923 (2003).

[231] Lambert,D.M. et al. Analogues and homologues of N-palmitoylethanolamide, a putative endogenous CB(2) cannabinoid, as potential ligands for the cannabinoid receptors. Biochim. Biophys. Acta 1440 , 266-274 (1999).

[232] Suardiaz,M., Estivill-Torrus,G., Goicoechea,C., Bilbao,A., & Rodriguez de,F.F. Analgesic properties of oleoylethanolamide (OEA) in visceral and inflammatory pain. Pain 133 , 99- 110 (2007).

[233] Chawla,A., Repa,J.J., Evans,R.M., & Mangelsdorf,D.J. Nuclear receptors and lipid physiology: opening the X-files. Science 294 , 1866-1870 (2001).

[234] Desvergne,B. & Wahli,W. Peroxisome proliferator-activated receptors: nuclear control of metabolism. Endocr. Rev. 20 , 649-688 (1999).

[235] Sun,Y. et al. Cannabinoid activation of PPAR alpha; a novel neuroprotective mechanism. Br. J. Pharmacol. 152 , 734-743 (2007).

[236] Sagar,D.R., Kendall,D.A., & Chapman,V. Inhibition of fatty acid amide hydrolase produces PPAR-alpha-mediated analgesia in a rat model of inflammatory pain. Br. J. Pharmacol. 155 , 1297-1306 (2008).

[237] Cuzzocrea,S. et al. The role of the peroxisome proliferator-activated receptor-alpha (PPAR-alpha) in the regulation of acute inflammation. J. Leukoc. Biol. 79 , 999-1010 (2006).

[238] Ahern,G.P. Activation of TRPV1 by the satiety factor oleoylethanolamide. J. Biol. Chem. 278 , 30429-30434 (2003).

[239] Immke,D.C. & Gavva,N.R. The TRPV1 receptor and nociception. Semin. Cell Dev. Biol. 17 , 582-591 (2006).

[240] Singh,T.A., Santha,P., & Nagy,I. Inflammatory mediators convert anandamide into a potent activator of the vanilloid type 1 transient receptor potential receptor in nociceptive primary sensory neurons. Neuroscience 136 , 539-548 (2005).

[241] Caterina,M.J. et al. Impaired nociception and pain sensation in mice lacking the capsaicin receptor. Science 288 , 306-313 (2000).

[242] Davis,J.B. et al. Vanilloid receptor-1 is essential for inflammatory thermal hyperalgesia. Nature 405 , 183-187 (2000).

[243] Jhaveri,M.D., Elmes,S.J., Kendall,D.A., & Chapman,V. Inhibition of peripheral vanilloid TRPV1 receptors reduces noxious heat-evoked responses of dorsal horn neurons in naive, carrageenan-inflamed and neuropathic rats. Eur. J. Neurosci. 22 , 361-370 (2005).

[244] Mazzari,S., Canella,R., Petrelli,L., Marcolongo,G., & Leon,A. N-(2- hydroxyethyl)hexadecanamide is orally active in reducing edema formation and inflammatory hyperalgesia by down-modulating mast cell activation. Eur. J. Pharmacol. 300 , 227-236 (1996).

[245] Berdyshev,E., Boichot,E., Corbel,M., Germain,N., & Lagente,V. Effects of cannabinoid receptor ligands on LPS-induced pulmonary inflammation in mice. Life Sci. 63 , L125- L129 (1998).

185 References

[246] Calignano,A., La,R.G., & Piomelli,D. Antinociceptive activity of the endogenous fatty acid amide, palmitylethanolamide. Eur. J. Pharmacol. 419 , 191-198 (2001).

[247] Jaggar,S.I., Hasnie,F.S., Sellaturay,S., & Rice,A.S. The anti-hyperalgesic actions of the cannabinoid anandamide and the putative CB2 receptor agonist palmitoylethanolamide in visceral and somatic inflammatory pain. Pain 76 , 189-199 (1998).

[248] Lo,V.J. et al. The nuclear receptor peroxisome proliferator-activated receptor-alpha mediates the anti-inflammatory actions of palmitoylethanolamide. Mol. Pharmacol. 67 , 15-19 (2005).

[249] LoVerme,J., La,R.G., Russo,R., Calignano,A., & Piomelli,D. The search for the palmitoylethanolamide receptor. Life Sci. 77 , 1685-1698 (2005).

[250] Costa,B., Comelli,F., Bettoni,I., Colleoni,M., & Giagnoni,G. The endogenous fatty acid amide, palmitoylethanolamide, has anti-allodynic and anti-hyperalgesic effects in a murine model of neuropathic pain: involvement of CB(1), TRPV1 and PPARgamma receptors and neurotrophic factors. Pain 139 , 541-550 (2008).

[251] Muccioli,G.G. & Stella,N. Microglia produce and hydrolyze palmitoylethanolamide. Neuropharmacology 54 , 16-22 (2008).

[252] Ryberg,E. et al. The orphan receptor GPR55 is a novel cannabinoid receptor. Br. J. Pharmacol. 152 , 1092-1101 (2007).

[253] Staton,P.C. et al. The putative cannabinoid receptor GPR55 plays a role in mechanical hyperalgesia associated with inflammatory and neuropathic pain. Pain 139 , 225-236 (2008).

[254] Ross,R.A. The enigmatic pharmacology of GPR55. Trends Pharmacol. Sci. 30 , 156-163 (2009).

[255] Tham,C.S., Whitaker,J., Luo,L., & Webb,M. Inhibition of microglial fatty acid amide hydrolase modulates LPS stimulated release of inflammatory mediators. FEBS Lett. 581 , 2899-2904 (2007).

[256] Lambert,D.M. & Di,M., V The palmitoylethanolamide and oleamide enigmas : are these two fatty acid amides cannabimimetic? Curr. Med. Chem. 6, 757-773 (1999).

[257] Ho,W.S., Barrett,D.A., & Randall,M.D. 'Entourage' effects of N-palmitoylethanolamide and N-oleoylethanolamide on vasorelaxation to anandamide occur through TRPV1 receptors. Br. J. Pharmacol. 155 , 837-846 (2008).

[258] Zhu,C. et al. Proinflammatory stimuli control N-acylphosphatidylethanolamine-specific phospholipase d expression in macrophages. Mol. Pharmacol. 79 , 786-792 (2011).

[259] Naidu,P.S., Kinsey,S.G., Guo,T.L., Cravatt,B.F., & Lichtman,A.H. Regulation of inflammatory pain by inhibition of fatty acid amide hydrolase. J. Pharmacol. Exp. Ther. 334 , 182-190 (2010).

[260] Beltramo,M. et al. Functional role of high-affinity anandamide transport, as revealed by selective inhibition. Science 277 , 1094-1097 (1997).

[261] Fowler,C.J., Tiger,G., & Stenstrom,A. Ibuprofen inhibits rat brain deamidation of anandamide at pharmacologically relevant concentrations. Mode of inhibition and structure-activity relationship. J. Pharmacol. Exp. Ther. 283 , 729-734 (1997).

[262] Holt,S., Nilsson,J., Omeir,R., Tiger,G., & Fowler,C.J. Effects of pH on the inhibition of fatty acid amidohydrolase by ibuprofen. Br. J. Pharmacol. 133 , 513-520 (2001).

[263] Deutsch,D.G. Design of on-target FAAH inhibitors. Chem. Biol. 12 , 1157-1158 (2005).

[264] Min,X. et al. Discovery and molecular basis of potent noncovalent inhibitors of fatty acid amide hydrolase (FAAH). Proc. Natl. Acad. Sci. U. S. A 108 , 7379-7384 (2011).

[265] Swinney,D.C. Biochemical mechanisms of drug action: what does it take for success? Nat. Rev. Drug Discov. 3, 801-808 (2004).

186 References

[266] Lichtman,A.H. et al. Reversible inhibitors of fatty acid amide hydrolase that promote analgesia: evidence for an unprecedented combination of potency and selectivity. J. Pharmacol. Exp. Ther. 311 , 441-448 (2004).

[267] Fegley,D. et al. Characterization of the fatty acid amide hydrolase inhibitor cyclohexyl carbamic acid 3'-carbamoyl-biphenyl-3-yl ester (URB597): effects on anandamide and oleoylethanolamide deactivation. J. Pharmacol. Exp. Ther. 313 , 352-358 (2005).

[268] Johnson,D.S. et al. Discovery of PF-04457845: A Highly Potent, Orally Bioavailable, and Selective Urea FAAH Inhibitor. ACS Med. Chem. Lett. 2, 91-96 (2011).

[269] Vincent,F., Nguyen,M.T., Emerling,D.E., Kelly,M.G., & Duncton,M.A. Mining biologically- active molecules for inhibitors of fatty acid amide hydrolase (FAAH): identification of phenmedipham and amperozide as FAAH inhibitors. Bioorg. Med. Chem. Lett. 19 , 6793- 6796 (2009).

[270] Johnson,D.S., Weerapana,E., & Cravatt,B.F. Strategies for discovering and derisking covalent, irreversible enzyme inhibitors. Future. Med. Chem. 2, 949-964 (2010).

[271] Mor,M. et al. Cyclohexylcarbamic acid 3'- or 4'-substituted biphenyl-3-yl esters as fatty acid amide hydrolase inhibitors: synthesis, quantitative structure-activity relationships, and molecular modeling studies. J. Med. Chem. 47 , 4998-5008 (2004).

[272] Ahn,K. et al. Novel mechanistic class of fatty acid amide hydrolase inhibitors with remarkable selectivity. Biochemistry 46 , 13019-13030 (2007).

[273] Ahn,K. et al. Mechanistic and Pharmacological Characterization of PF-04457845: A Highly Potent and Selective Fatty Acid Amide Hydrolase Inhibitor That Reduces Inflammatory and Noninflammatory Pain. J. Pharmacol. Exp. Ther. 338 , 114-124 (2011).

[274] Colburn,W.A. Biomarkers in drug discovery and development: from target identification through drug marketing. J. Clin. Pharmacol. 43 , 329-341 (2003).

[275] Ezzili,C. et al. Reversible competitive alpha-ketoheterocycle inhibitors of fatty acid amide hydrolase containing additional conformational constraints in the acyl side chain: orally active, long-acting analgesics. J. Med. Chem. 54 , 2805-2822 (2011).

[276] Schreiber,D. et al. Determination of anandamide and other fatty acyl ethanolamides in human serum by electrospray tandem mass spectrometry. Anal. Biochem. 361 , 162-168 (2007).

[277] Lam,P.M. et al. Ultra performance liquid chromatography tandem mass spectrometry method for the measurement of anandamide in human plasma. Anal. Biochem. 380 , 195-201 (2008).

[278] Balvers,M.G., Verhoeckx,K.C., & Witkamp,R.F. Development and validation of a quantitative method for the determination of 12 endocannabinoids and related compounds in human plasma using liquid chromatography-tandem mass spectrometry. J. Chromatogr. B Analyt. Technol. Biomed. Life Sci. 877 , 1583-1590 (2009).

[279] Palandra,J., Prusakiewicz,J., Ozer,J.S., Zhang,Y., & Heath,T.G. Endogenous ethanolamide analysis in human plasma using HPLC tandem MS with electrospray ionization. J. Chromatogr. B Analyt. Technol. Biomed. Life Sci. 877 , 2052-2060 (2009).

[280] Ozalp,A. & Barroso,B. Simultaneous quantitative analysis of N-acylethanolamides in clinical samples. Anal. Biochem. 395 , 68-76 (2009).

[281] Jian,W. et al. Validation and application of an LC-MS/MS method for quantitation of three fatty acid ethanolamides as biomarkers for fatty acid hydrolase inhibition in human plasma. J. Chromatogr. B Analyt. Technol. Biomed. Life Sci. 878 , 1687-1699 (2010).

[282] Bertolini,A. et al. Paracetamol: new vistas of an old drug. CNS. Drug Rev. 12 , 250-275 (2006).

[283] Hendrickson,R.G. & Bizovi,K.E. Acetaminophen in Goldfrank's toxicologic emergencies (eds. Wonsiewicz,M.J., Edmonson,K.G. & Boyle,P.J.) 523 (McGraw-Hill, New York, 2006).

[284] Dong,H., Haining,R.L., Thummel,K.E., Rettie,A.E., & Nelson,S.D. Involvement of human cytochrome P450 2D6 in the bioactivation of acetaminophen. Drug Metab Dispos. 28 , 1397-1400 (2000).

187 References

[285] Ayoub,S.S., Hunter,J.C., & Simmons,D.L. Answering the burning question of how transient receptor potential vanilloid-1 channel antagonists cause unwanted hyperthermia. Pharmacol. Rev. 61 , 225-227 (2009).

[286] Greco,A., Ajmone-Cat,M.A., Nicolini,A., Sciulli,M.G., & Minghetti,L. Paracetamol effectively reduces prostaglandin E2 synthesis in brain macrophages by inhibiting enzymatic activity of cyclooxygenase but not phospholipase and prostaglandin E synthase. J. Neurosci. Res. 71 , 844-852 (2003).

[287] Flower,R.J. & Vane,J.R. Inhibition of prostaglandin synthetase in brain explains the anti- pyretic activity of paracetamol (4-acetamidophenol). Nature 240 , 410-411 (1972).

[288] Ayoub,S.S., Colville-Nash,P.R., Willoughby,D.A., & Botting,R.M. The involvement of a cyclooxygenase 1 gene-derived protein in the antinociceptive action of paracetamol in mice. Eur. J. Pharmacol. 538 , 57-65 (2006).

[289] Ayoub,S.S., Joshi,A., Chol,M., Gilroy,D.W., & Seed,M.P. Inhibition of the diclofenac- induced cyclooxygenase-2 activity by paracetamol in cultured macrophages is not related to the intracellular lipid hydroperoxide tone. Fundam. Clin. Pharmacol. 25 , 186- 190 (2011).

[290] Hanel,A.M. & Lands,W.E. Modification of anti-inflammatory drug effectiveness by ambient lipid peroxides. Biochem. Pharmacol. 31 , 3307-3311 (1982).

[291] Lee,Y.S. et al. Acetaminophen selectively suppresses peripheral prostaglandin E2 release and increases COX-2 gene expression in a clinical model of acute inflammation. Pain 129 , 279-286 (2007).

[292] Mallet,C. et al. Endocannabinoid and serotonergic systems are needed for acetaminophen-induced analgesia. Pain 139 , 190-200 (2008).

[293] Ottani,A., Leone,S., Sandrini,M., Ferrari,A., & Bertolini,A. The analgesic activity of paracetamol is prevented by the blockade of cannabinoid CB1 receptors. Eur. J. Pharmacol. 531 , 280-281 (2006).

[294] Mallet,C. et al. TRPV1 in brain is involved in acetaminophen-induced antinociception. PLoS. One. 5, (2010).

[295] Khanolkar,A.D. et al. Head group analogs of arachidonylethanolamide, the endogenous cannabinoid ligand. J. Med. Chem. 39 , 4515-4519 (1996).

[296] La,R.G. et al. Modulation of neuropathic and inflammatory pain by the endocannabinoid transport inhibitor AM404 [N-(4-hydroxyphenyl)-eicosa-5,8,11,14-tetraenamide]. J. Pharmacol. Exp. Ther. 317 , 1365-1371 (2006).

[297] Squires,L.N. et al. Serotonin catabolism and the formation and fate of 5-hydroxyindole thiazolidine carboxylic acid. J. Biol. Chem. 281 , 13463-13470 (2006).

[298] Bardin,L., Tarayre,J.P., Malfetes,N., Koek,W., & Colpaert,F.C. Profound, non-opioid analgesia produced by the high-efficacy 5-HT(1A) agonist F 13640 in the formalin model of tonic nociceptive pain. Pharmacology 67 , 182-194 (2003).

[299] Courade,J.P. et al. Effects of acetaminophen on monoaminergic systems in the rat central nervous system. Naunyn Schmiedebergs Arch. Pharmacol. 364 , 534-537 (2001).

[300] Pelissier,T., Alloui,A., Paeile,C., & Eschalier,A. Evidence of a central antinociceptive effect of paracetamol involving spinal 5HT3 receptors. Neuroreport 6, 1546-1548 (1995).

[301] Alloui,A. et al. Paracetamol exerts a spinal, tropisetron-reversible, antinociceptive effect in an inflammatory pain model in rats. Eur. J. Pharmacol. 443 , 71-77 (2002).

[302] Bonnefont,J., Alloui,A., Chapuy,E., Clottes,E., & Eschalier,A. Orally administered paracetamol does not act locally in the rat formalin test: evidence for a supraspinal, serotonin-dependent antinociceptive mechanism. Anesthesiology 99 , 976-981 (2003).

[303] Raffa,R.B. & Codd,E.E. Lack of binding of acetaminophen to 5-HT receptor or uptake sites (or eleven other binding/uptake assays). Life Sci. 59 , L37-L40 (1996).

[304] Pickering,G. et al. Analgesic effect of acetaminophen in humans: first evidence of a central serotonergic mechanism. Clin. Pharmacol. Ther. 79 , 371-378 (2006).

188 References

[305] Zygmunt,P.M., Chuang,H., Movahed,P., Julius,D., & Hogestatt,E.D. The anandamide transport inhibitor AM404 activates vanilloid receptors. Eur. J. Pharmacol. 396 , 39-42 (2000).

[306] Nicholson,R.A. et al. Sodium channel inhibition by anandamide and synthetic cannabimimetics in brain. Brain Res. 978 , 194-204 (2003).

[307] Nau,R., Sorgel,F., & Eiffert,H. Penetration of drugs through the blood-cerebrospinal fluid/blood-brain barrier for treatment of central nervous system infections. Clin. Microbiol. Rev. 23 , 858-883 (2010).

[308] Michael-Titus,A., Revest,P., & Shortland,P. Infection In The Central Nervous System in The Nervous System (Elsevier, 2010).

[309] Banks,W.A. Characteristics of compounds that cross the blood-brain barrier. BMC. Neurol. 9 Suppl 1 , S3 (2009).

[310] Bektas,F. et al. Intravenous paracetamol or morphine for the treatment of renal colic: a randomized, placebo-controlled trial. Ann. Emerg. Med. 54 , 568-574 (2009).

[311] Serinken,M. et al. Intravenous paracetamol versus morphine for renal colic in the emergency department: a randomised double-blind controlled trial. Emerg. Med. J(2011).

[312] Craig,M., Jeavons,R., Probert,J., & Benger,J. Randomised comparison of intravenous paracetamol and intravenous morphine for acute traumatic limb pain in the emergency department. Emerg. Med. J 29 , 37-39 (2012).

[313] van der Westhuizen,J., Kuo,P.Y., Reed,P.W., & Holder,K. Randomised controlled trial comparing oral and intravenous paracetamol (acetaminophen) plasma levels when given as preoperative analgesia. Anaesth. Intensive Care 39 , 242-246 (2011).

[314] Liukas,A. et al. Pharmacokinetics of intravenous paracetamol in elderly patients. Clin Pharmacokinet. 50 , 121-129 (2011).

[315] Hinz,B., Cheremina,O., & Brune,K. Acetaminophen (paracetamol) is a selective cyclooxygenase-2 inhibitor in man. FASEB J 22 , 383-390 (2008).

[316] Jack,E.S. & Baggott,M. Control of acute pain in postoperative and post-traumatic situations and the role of the acute pain service. Anaesthesia & Intensive Care Medicine 12 , 1-4 (2011).

[317] Piguet,V., Desmeules,J., & Dayer,P. Lack of acetaminophen ceiling effect on R-III nociceptive flexion reflex. Eur. J Clin Pharmacol. 53 , 321-324 (1998).

[318] Anderson,B.J., Holford,N.H., Woollard,G.A., & Chan,P.L. Paracetamol plasma and cerebrospinal fluid pharmacokinetics in children. Br J Clin Pharmacol. 46 , 237-243 (1998).

[319] van der Marel,C.D. et al. Acetaminophen in cerebrospinal fluid in children. Eur. J Clin Pharmacol. 59 , 297-302 (2003).

[320] Kumpulainen,E. et al. Paracetamol (acetaminophen) penetrates readily into the cerebrospinal fluid of children after intravenous administration. Pediatrics 119 , 766-771 (2007).

[321] Moreau,X., Le,Q.L., Granry,J.C., Boishardy,N., & Delhumeau,A. Pharmacokinetics of paracetamol in the cerebrospinal fluid in the elderly. Therapie 48 , 393-396 (1993).

[322] Bannwarth,B. et al. Plasma and cerebrospinal fluid concentrations of paracetamol after a single intravenous dose of propacetamol. Br. J. Clin. Pharmacol. 34 , 79-81 (1992).

[323] van den Anker,J.N. & Tibboel,D. Pain relief in neonates: when to use intravenous paracetamol. Arch. Dis. Child 96 , 573-574 (2011).

[324] Rudolph,A.M. Effects of aspirin and acetaminophen in pregnancy and in the newborn. Arch. Intern. Med. 141 , 358-363 (1981).

[325] Ryder,S.D. & Beckingham,I.J. ABC of diseases of liver, pancreas, and biliary system. Other causes of parenchymal liver disease. BMJ 322 , 290-292 (2001).

189 References

[326] Robinson,A.E., Sattar,H., McDowall,R.D., Holder,A.T., & Powell,R. Forensic toxicology of some deaths associated with the combined use of propoxyphene and acetaminophen (paracetamol). J Forensic Sci. 22 , 708-717 (1977).

[327] James,L.P., Mayeux,P.R., & Hinson,J.A. Acetaminophen-induced hepatotoxicity. Drug Metab Dispos. 31 , 1499-1506 (2003).

[328] Paracetamol + methionine: pameton. Drug and therapeutics bulletin 25 , 99-100 (1987).

[329] Hartley,J.W. et al. Expression of infectious murine leukemia viruses by RAW264.7 cells, a potential complication for studies with a widely used mouse macrophage cell line. Retrovirology. 5, 1 (2008).

[330] Liu,J. et al. Lipopolysaccharide induces anandamide synthesis in macrophages via CD14/MAPK/phosphoinositide 3-kinase/NF-kappaB independently of platelet-activating factor. J. Biol. Chem. 278 , 45034-45039 (2003).

[331] Di,M., V et al. Biosynthesis and inactivation of the endocannabinoid 2- arachidonoylglycerol in circulating and tumoral macrophages. Eur. J. Biochem. 264 , 258-267 (1999).

[332] Mosmann,T. Rapid colorimetric assay for cellular growth and survival: application to proliferation and cytotoxicity assays. J. Immunol. Methods 65 , 55-63 (1983).

[333] Yiangou,Y. et al. COX-2, CB2 and P2X7-immunoreactivities are increased in activated microglial cells/macrophages of multiple sclerosis and amyotrophic lateral sclerosis spinal cord. BMC. Neurol. 6, 12 (2006).

[334] Nakano,Y., Pross,S.H., & Friedman,H. Modulation of interleukin 2 activity by delta 9- tetrahydrocannabinol after stimulation with concanavalin A, phytohemagglutinin, or anti- CD3 antibody. Proc. Soc. Exp. Biol. Med. 201 , 165-168 (1992).

[335] Zheng,Z.M. & Specter,S.C. Delta-9-tetrahydrocannabinol suppresses tumor necrosis factor alpha maturation and secretion but not its transcription in mouse macrophages. Int. J Immunopharmacol. 18 , 53-68 (1996).

[336] Coffey,R.G., Yamamoto,Y., Snella,E., & Pross,S. Tetrahydrocannabinol inhibition of macrophage nitric oxide production. Biochem. Pharmacol. 52 , 743-751 (1996).

[337] Lichtman,A.H., Shelton,C.C., Advani,T., & Cravatt,B.F. Mice lacking fatty acid amide hydrolase exhibit a cannabinoid receptor-mediated phenotypic hypoalgesia. Pain 109 , 319-327 (2004).

[338] Nomura,D.K. et al. Monoacylglycerol lipase regulates a fatty acid network that promotes cancer pathogenesis. Cell 140 , 49-61 (2010).

[339] Garrido,G. et al. In vivo and in vitro anti-inflammatory activity of Mangifera indica L. extract (VIMANG). Pharmacol. Res. 50 , 143-149 (2004).

[340] Hulkower,K.I. et al. Leukotrienes do not regulate nitric oxide production in RAW 264.7 macrophages. Prostaglandins Leukot. Essent. Fatty Acids 55 , 145-149 (1996).

[341] Chouinard,F. et al. The endocannabinoid 2-arachidonoyl-glycerol activates human neutrophils: critical role of its hydrolysis and de novo leukotriene B4 biosynthesis. J Immunol. 186 , 3188-3196 (2011).

[342] Salm,P., Taylor,P.J., & Kostner,K. Simultaneous quantification of total eicosapentaenoic acid, docosahexaenoic acid and arachidonic acid in plasma by high-performance liquid chromatography-tandem mass spectrometry. Biomed. Chromatogr. 25 , 652-659 (2011).

[343] Yapa,U. et al. High-performance liquid chromatography-tandem mass spectrometry assay of fatty acid amide hydrolase (FAAH) in blood: FAAH inhibition as clinical biomarker. Anal. Biochem. 421 , 556-565 (2012).

[344] Hughes,D. et al. NAD+-dependent 15-hydroxyprostaglandin dehydrogenase regulates levels of bioactive lipids in non-small cell lung cancer. Cancer Prev. Res. (Phila) 1, 241- 249 (2008).

190 References

[345] Chen,T., Scott,E., & Morrison,D.C. Differential effects of serum on lipopolysaccharide receptor-directed macrophage activation for nitric oxide production. Immunol. Lett. 40 , 179-187 (1994).

[346] Kirikae,T., Schade,F.U., Kirikae,F., Rietschel,E.T., & Morrison,D.C. Isolation of a macrophage-like cell line defective in binding of lipopolysaccharide. Influence of serum and lipopolysaccharide chain length on macrophage activation. J Immunol. 151 , 2742- 2752 (1993).

[347] Projean,D., Lessard,E., Ducharme,M.P., & Ducharme,J. Use of Freund's Complete Adjuvant (FCA) in inflammatory pain models: consequences on the metabolism and pharmacokinetics of the non-peptidic delta receptor agonist SNC80 in the rat. Xenobiotica 37 , 870-883 (2007).

[348] Whiteside,G.T., Adedoyin,A., & Leventhal,L. Predictive validity of animal pain models? A comparison of the pharmacokinetic-pharmacodynamic relationship for pain drugs in rats and humans. Neuropharmacology 54 , 767-775 (2008).

[349] Phillips,R., Ursell,T., Wiggins,P., & Sens,P. Emerging roles for lipids in shaping membrane-protein function. Nature 459 , 379-385 (2009).

[350] Hallifax,D. & Houston,J.B. Binding of drugs to hepatic microsomes: comment and assessment of current prediction methodology with recommendation for improvement. Drug Metab Dispos. 34 , 724-726 (2006).

[351] Boldrup,L., Wilson,S.J., Barbier,A.J., & Fowler,C.J. A simple stopped assay for fatty acid amide hydrolase avoiding the use of a chloroform extraction phase. J Biochem. Biophys. Methods 60 , 171-177 (2004).

[352] Duffus,J.H., Nordberg,M., & Templeton,D.M. Glossary of terms used in toxicology, 2nd edition (IUPAC Recommendations 2007). Pure and Applied Chemistry 79 , 1153-1344 (2007).

[353] Iwatsubo,T. et al. Prediction of in vivo drug metabolism in the human liver from in vitro metabolism data. Pharmacol. Ther. 73 , 147-171 (1997).

[354] Ito,K., Iwatsubo,T., Kanamitsu,S., Nakajima,Y., & Sugiyama,Y. Quantitative prediction of in vivo drug clearance and drug interactions from in vitro data on metabolism, together with binding and transport. Annu. Rev. Pharmacol. Toxicol. 38 , 461-499 (1998).

[355] Kalvass,J.C., Tess,D.A., Giragossian,C., Linhares,M.C., & Maurer,T.S. Influence of microsomal concentration on apparent intrinsic clearance: implications for scaling in vitro data. Drug Metab Dispos. 29 , 1332-1336 (2001).

[356] Waxman,D.J. & Chang,T.K. Use of 7-ethoxycoumarin to monitor multiple enzymes in the human CYP1, CYP2, and CYP3 families. Methods Mol. Biol. 320 , 153-156 (2006).

[357] Yu,A. & Haining,R.L. Comparative contribution to dextromethorphan metabolism by cytochrome P450 isoforms in vitro: can dextromethorphan be used as a dual probe for both CTP2D6 and CYP3A activities? Drug Metab Dispos. 29 , 1514-1520 (2001).

[358] Banker,M.J. & Clark,T.H. Plasma/serum protein binding determinations. Curr. Drug Metab 9, 854-859 (2008).

[359] FDA Guidance for Industry: Bioanalytical Method Validation. Department of Health and Human Services Center for Drug Evaluation and Research (CDER) (2001).

[360] Willoughby,K.A., Moore,S.F., Martin,B.R., & Ellis,E.F. The biodisposition and metabolism of anandamide in mice. J. Pharmacol. Exp. Ther. 282 , 243-247 (1997).

[361] Makino,S., Reynolds,J.A., & Tanford,C. The binding of deoxycholate and Triton X-100 to proteins. J. Biol. Chem. 248 , 4926-4932 (1973).

[362] Penning,T.D. et al. Synthesis and biological evaluation of the 1,5-diarylpyrazole class of cyclooxygenase-2 inhibitors: identification of 4-[5-(4-methylphenyl)-3-(trifluoromethyl)- 1H-pyrazol-1-yl]benze nesulfonamide (SC-58635, celecoxib). J Med. Chem. 40 , 1347- 1365 (1997).

191 References

[363] Murillo-Rodriguez,E., Desarnaud,F., & Prospero-Garcia,O. Diurnal variation of arachidonoylethanolamine, palmitoylethanolamide and oleoylethanolamide in the brain of the rat. Life Sci. 79 , 30-37 (2006).

[364] Snider,N.T., Nast,J.A., Tesmer,L.A., & Hollenberg,P.F. A cytochrome P450-derived epoxygenated metabolite of anandamide is a potent cannabinoid receptor 2-selective agonist. Mol. Pharmacol. 75 , 965-972 (2009).

[365] Adas,F. et al. Involvement of cytochrome P450 2E1 in the (omega-1)-hydroxylation of oleic acid in human and rat liver microsomes. J. Lipid Res. 39 , 1210-1219 (1998).

[366] Vertzoni,M.V., Archontaki,H.A., & Galanopoulou,P. Development and optimization of a reversed-phase high-performance liquid chromatographic method for the determination of acetaminophen and its major metabolites in rabbit plasma and urine after a toxic dose. J Pharm. Biomed. Anal. 32 , 487-493 (2003).

[367] Powell,W.S. Rapid extraction of oxygenated metabolites of arachidonic acid from biological samples using octadecylsilyl silica. Prostaglandins 20 , 947-957 (1980).

[368] Ayoub,S.S. et al. Acetaminophen-induced hypothermia in mice is mediated by a prostaglandin endoperoxide synthase 1 gene-derived protein. Proc. Natl. Acad. Sci. U. S. A 101 , 11165-11169 (2004).

[369] Ritschel,W.A. & Kearns,G.L. Handbook of Basic Pharmacokinetics (APhA Publications,2004).

[370] Lau,G.S. & Critchley,J.A. The estimation of paracetamol and its major metabolites in both plasma and urine by a single high-performance liquid chromatography assay. J Pharm. Biomed. Anal. 12 , 1563-1572 (1994).

[371] Hain,R.D., Hardcastle,A., Pinkerton,C.R., & Aherne,G.W. Morphine and morphine-6- glucuronide in the plasma and cerebrospinal fluid of children. Br J Clin Pharmacol. 48 , 37-42 (1999).

[372] Baker,D.G. et al. Cerebrospinal fluid and plasma beta-endorphin in combat veterans with post-traumatic stress disorder. Psychoneuroendocrinology 22 , 517-529 (1997).

[373] Langemark,M., Bach,F.W., Ekman,R., & Olesen,J. Increased cerebrospinal fluid Met- enkephalin immunoreactivity in patients with chronic tension-type headache. Pain 63 , 103-107 (1995).

[374] Mitchell,P.B. & Morris,M.J. Depression and anxiety with rimonabant. Lancet 370 , 1671- 1672 (2007).

[375] Muth-Selbach,U.S., Tegeder,I., Brune,K., & Geisslinger,G. Acetaminophen inhibits spinal prostaglandin E2 release after peripheral noxious stimulation. Anesthesiology 91 , 231- 239 (1999).

[376] Seppala,E., Laitinen,O., & Vapaatalo,H. Comparative study on the effects of acetylsalicylic acid, indomethacin and paracetamol on metabolites of arachidonic acid in plasma, serum and urine in man. Int. J Clin Pharmacol. Res. 3, 265-269 (1983).

[377] Resnick,D.K., Nguyen,P., & Cechvala,C.F. Selective cyclooxygenase 2 inhibition lowers spinal cord prostaglandin concentrations after injury. Spine J 1, 437-441 (2001).

[378] Combrinck,M. et al. Levels of CSF prostaglandin E2, cognitive decline, and survival in Alzheimer's disease. J Neurol. Neurosurg. Psychiatry 77 , 85-88 (2006).

[379] Mayatepek,E., Flock,B., Zelezny,R., Kreutzer,K., & von Giesen,H.J. LTB4 and LTC4 are absent in the cerebrospinal fluid of human immunodeficiency virus type 1-seropositive persons with toxoplasmic encephalitis: evidence for inhibition of 5-lipoxygenase by Toxoplasma gondii. J. Infect. Dis. 179 , 714-716 (1999).

[380] Bassit,R.A., Curi,R., & Costa Rosa,L.F. Creatine supplementation reduces plasma levels of pro-inflammatory cytokines and PGE2 after a half-ironman competition. Amino. Acids 35 , 425-431 (2008).

[381] Shindo,K., Miyakawa,K., & Fukumura,M. Plasma levels of leukotriene B4 in asthmatic patients. Int. J Tissue React. 15 , 181-184 (1993).

192 References

[382] Giuffrida,A. et al. Cerebrospinal anandamide levels are elevated in acute schizophrenia and are inversely correlated with psychotic symptoms. Neuropsychopharmacology 29 , 2108-2114 (2004).

[383] Centonze,D. et al. The endocannabinoid system is dysregulated in multiple sclerosis and in experimental autoimmune encephalomyelitis. Brain 130 , 2543-2553 (2007).

[384] Di,F.M., Pini,L.A., Pelliccioli,G.P., Calabresi,P., & Sarchielli,P. Abnormalities in the cerebrospinal fluid levels of endocannabinoids in multiple sclerosis. J. Neurol. Neurosurg. Psychiatry 79 , 1224-1229 (2008).

[385] Bearcroft,C.P., Farthing,M.J., & Perrett,D. Determination of 5-hydroxytryptamine, 5- hydroxyindoleacetic acid and tryptophan in plasma and urine by HPLC with fluorimetric detection. Biomed. Chromatogr. 9, 23-27 (1995).

[386] Vanhoutte,P.M. Platelet-derived serotonin, the endothelium, and cardiovascular disease. J Cardiovasc. Pharmacol. 17 Suppl 5 , S6-12 (1991).

[387] Gao,H.Q., Zhu,H.Y., Zhang,Y.Q., & Wang,L.X. Reduction of cerebrospinal fluid and plasma serotonin in patients with post-stroke depression: A preliminary report. Clin Invest Med. 31 , E351-E356 (2008).

[388] Zygmunt,P.M. et al. Vanilloid receptors on sensory nerves mediate the vasodilator action of anandamide. Nature 400 , 452-457 (1999).

[389] Alexander,S.P. So what do we call GPR18 now? Br J Pharmacol. 165 , 2411-2413 (2012).

[390] Overton,H.A., Fyfe,M.C., & Reynet,C. GPR119, a novel G protein-coupled receptor target for the treatment of type 2 diabetes and obesity. Br J Pharmacol. 153 Suppl 1 , S76-S81 (2008).

[391] Di,M., V, De,P.L., Sepe,N., & Buono,A. Biosynthesis of anandamide and related acylethanolamides in mouse J774 macrophages and N18 neuroblastoma cells. Biochem. J 316 ( Pt 3) , 977-984 (1996).

[392] Gomes,I. et al. Novel endogenous peptide agonists of cannabinoid receptors. FASEB J 23 , 3020-3029 (2009).

[393] De,P.L. et al. Actions of two naturally occurring saturated N-acyldopamines on transient receptor potential vanilloid 1 (TRPV1) channels. Br J Pharmacol. 143 , 251-256 (2004).

[394] Ortar,G. et al. New N-arachidonoylserotonin analogues with potential "dual" mechanism of action against pain. J Med. Chem. 50 , 6554-6569 (2007).

[395] Parmar,N. & Ho,W.S. N-arachidonoyl glycine, an endogenous lipid that acts as a vasorelaxant via nitric oxide and large conductance calcium-activated potassium channels. Br J Pharmacol. 160 , 594-603 (2010).

[396] Eichholtz,T., Jalink,K., Fahrenfort,I., & Moolenaar,W.H. The bioactive phospholipid lysophosphatidic acid is released from activated platelets. Biochem. J. 291 ( Pt 3) , 677- 680 (1993).

[397] Watanabe,N. et al. Plasma lysophosphatidic acid level and serum autotaxin activity are increased in liver injury in rats in relation to its severity. Life Sci. 81 , 1009-1015 (2007).

[398] Tager,A.M. et al. The lysophosphatidic acid receptor LPA1 links pulmonary fibrosis to lung injury by mediating fibroblast recruitment and vascular leak. Nat. Med. 14 , 45-54 (2008).

[399] Renback,K., Inoue,M., & Ueda,H. Lysophosphatidic acid-induced, pertussis toxin- sensitive nociception through a substance P release from peripheral nerve endings in mice. Neurosci. Lett. 270 , 59-61 (1999).

[400] Renback,K., Inoue,M., Yoshida,A., Nyberg,F., & Ueda,H. Vzg-1/lysophosphatidic acid- receptor involved in peripheral pain transmission. Brain Res. Mol. Brain Res. 75 , 350- 354 (2000).

[401] Inoue,M. et al. Initiation of neuropathic pain requires lysophosphatidic acid receptor signaling. Nat. Med. 10 , 712-718 (2004).

193 References

[402] Ahn,D.K. et al. Intratrigeminal ganglionic injection of LPA causes neuropathic pain-like behavior and demyelination in rats. Pain (2009).

[403] Toussaint,K. et al. What do we (not) know about how paracetamol (acetaminophen) works? J Clin Pharm. Ther. 35 , 617-638 (2010).

[404] Prescott,L.F., Mattison,P., Menzies,D.G., & Manson,L.M. The comparative effects of paracetamol and indomethacin on renal function in healthy female volunteers. Br J Clin Pharmacol. 29 , 403-412 (1990).

[405] Dewall,C.N. et al. Acetaminophen reduces social pain: behavioral and neural evidence. Psychol. Sci. 21 , 931-937 (2010).

[406] Ishida,T. et al. Effect of acetaminophen, a cyclooxygenase inhibitor, on Morris water maze task performance in mice. J Psychopharmacol. 21 , 757-767 (2007).

[407] Li,G.L. & Nichols,D. Why analgesia for osteoarthritis pain was not achieved through modulation of endogenous cannabinoids by inhibition of fatty acid amide hydrolase? British Pharmacological Society Winter Meeting (2011).

[408] Fowler,C.J., Naidu,P.S., Lichtman,A., & Onnis,V. The case for the development of novel analgesic agents targeting both fatty acid amide hydrolase and either cyclooxygenase or TRPV1. Br J Pharmacol. 156 , 412-419 (2009).

[409] Benito,C., Nunez,E., Pazos,M.R., Tolon,R.M., & Romero,J. The endocannabinoid system and Alzheimer's disease. Mol. Neurobiol. 36 , 75-81 (2007).

[410] Johnston,T.H. et al. Fatty acid amide hydrolase (FAAH) inhibition reduces L-3,4- dihydroxyphenylalanine-induced hyperactivity in the 1-methyl-4-phenyl-1,2,3,6- tetrahydropyridine-lesioned non-human primate model of Parkinson's disease. J Pharmacol. Exp. Ther. 336 , 423-430 (2011).

[411] Pryce,G. & Baker,D. Control of spasticity in a multiple sclerosis model is mediated by CB1, not CB2, cannabinoid receptors. Br J Pharmacol. 150 , 519-525 (2007).

[412] Moreira,F.A., Kaiser,N., Monory,K., & Lutz,B. Reduced anxiety-like behaviour induced by genetic and pharmacological inhibition of the endocannabinoid-degrading enzyme fatty acid amide hydrolase (FAAH) is mediated by CB1 receptors. Neuropharmacology 54 , 141-150 (2008).

[413] Vinod,K.Y. et al. Dysfunction in fatty acid amide hydrolase is associated with depressive- like behavior in wistar kyoto rats. PLoS. One. 7, e36743 (2012).

[414] Rampa,A. et al. The First Dual ChE/FAAH Inhibitors: New Perspectives for Alzheimer's Disease? ACS Med. Chem. Lett. 3, 182-186 (2012).

[415] Garcia,N., Jr., Jarai,Z., Mirshahi,F., Kunos,G., & Sanyal,A.J. Systemic and portal hemodynamic effects of anandamide. Am. J Physiol Gastrointest. Liver Physiol 280 , G14-G20 (2001).

[416] Godlewski,G. et al. Inhibitor of fatty acid amide hydrolase normalizes cardiovascular function in hypertension without adverse metabolic effects. Chem. Biol. 17 , 1256-1266 (2010).

[417] Murillo-Rodriguez,E., Vazquez,E., Millan-Aldaco,D., Palomero-Rivero,M., & Drucker- Colin,R. Effects of the fatty acid amide hydrolase inhibitor URB597 on the sleep-wake cycle, c-Fos expression and dopamine levels of the rat. Eur. J Pharmacol. 562 , 82-91 (2007).

[418] Wang,H. et al. Fatty acid amide hydrolase deficiency limits early pregnancy events. J Clin Invest 116 , 2122-2131 (2006).

[419] MacCarrone,M. et al. Relation between decreased anandamide hydrolase concentrations in human lymphocytes and miscarriage. Lancet 355 , 1326-1329 (2000).

194 Publications

Publications

Papers in preparation

SHARMA C, LONG JH, PERRETT D, LANGFORD RM, MEHTA V: A Single Centre

Prospective Randomised Study to Investigate the Cerebrospinal Fluid (CSF)

Pharmacokinetics of Intravenous Paracetamol in Humans.

LONG JH, SHARMA C, GHAZALY, E, MEHTA V, PERRETT D: Measurement of paracetamol, its major metabolites and its active metabolite AM404 in human cerebrospinal fluid.

Posters

SHARMA C, LONG JH, PERRETT D, LANGFORD R, MEHTA V: Understanding

Acetaminophen: A Step Closer? Presented at Barts and the London School of

Anaesthesia Trainee Prize Evening, Basildon. Aug 2012.

SHARMA C, LONG JH, PERRETT D, LANGFORD R, MEHTA V: Understanding

Acetaminophen: A Step Closer? Presented at American Pain Society 31st Annual

Scientific Meeting, Hawaii. May 2012.

LONG JH, SHARMA C, GHAZALY, E, MEHTA V, PERRETT D: Measurement of paracetamol, its major metabolites and its active metabolite AM404 in human cerebrospinal fluid. Presented at the 11 th HTC - Symposium in Bruges, Belgium.

January 2012.

195 Appendix

Appendix

196

A Single Centre Prospective Randomised Study to Investigate the Cerebrospinal Fluid (CSF) Pharmacokinetics of Intravenous Paracetamol in Humans

Short Title: ‘ Paracetamol CSF Study’

Sponsor: Barts and the London NHS Trust Representative of the sponsor: Gerry Leonard Head of Resources, Joint R&D Office, 5 Walden Street London, E1 2EF Phone: 020 7882 7260 Email: [email protected]

Funder(s): Pain and Anaesthesia Research Centre

All Trust(s) and organisations: Barts and the London NHS Trust

List all the Pharmacies: Barts and The London NHS Trust Pharmacy Dept

Labs: Professor David Perrett, William Harvey Research Institute, Barts and The London, Queen Mary's School of Medicine and Dentistry, John Vane Building, Charterhouse Square, London EC1M 6BQ

Chief Investigator: Dr Vivek Mehta, Pain & Anaesthesia Research Centre Anaesthetic Laboratory, St Bartholomew’s Hospital, London, EC1A 7BE, Tel: 020 7601 7524, Fax: 020 7600 7748, Email: [email protected]

CHIEF INVESTIGATOR:

SIGNATURE ______DATE: ______

Clinical Co-Investigators: Professor Richard M Langford, Pain & Anaesthesia Research Centre, St Bartholomew’s Hospital, London, EC1A 7BE, Tel: 020 7601 7524, Fax: 020 7600 7748, Email: [email protected]

Dr Seema Shah, Pain & Anaesthesia Research Centre, St Bartholomew’s Hospital, London, EC1A 7BE, Tel: 020 7601 7524, Fax: 020 7600 7748, Email: [email protected]

Dr Samir S Ayoub, Centre for Biochemical Pharmacology, William Harvey Research Institute, Queen Mary, University of London, Charterhouse Square EC1M 6BQ, Tel: 0207 882 5817, Fax: 0207 882 6076, Email: [email protected]

Prof David Perrett, Centre for Biochemical Pharmacology, William Harvey Research Institute, Queen Mary, University of London, Charterhouse Square EC1M 6BQ

Dr Laura Fulton, Pain & Anaesthesia Research Centre, St Bartholomew’s Hospital, London, EC1A 7BE, Tel: 020 7601 7524, Fax: 020 7600 7748, Email: [email protected]

Trial Co-ordinator(s): Dr Seema Shah, Pain & Anaesthesia Research Centre, St Bartholomew’s Hospital, London, EC1A 7BE, Tel: 020 7601 7524, Fax: 020 7600 7748, Email: [email protected]

Monitors: Theresa Wodehouse, Research Scientist, Pain & Anaesthesia Research Centre, St Bartholomew’s Hospital, London, EC1A 7BE, Tel: 020 7601 7524, Fax: 020 7600 7748, Email: [email protected]

Study Statistician: Statistics review done by Xiang He, Statistician, Research and Development Office, Barts and The London NHS Trust.

Table of Contents GLOSSARY OF TERMS AND ABBREVIATIONS...... 7 1 INTRODUCTION ...... 9 1.1 BACKGROUND ...... 9 1.2 INVESTIGATIONAL MEDICINAL PRODUCT (IMP)...... 9 1.3 PRECLINICAL DATA ...... 9 1.4 CLINICAL DATA TO DATE ...... 9 1.5 RATIONALE AND RISK /B ENEFITS ...... 10 2 STUDY AIMS AND OBJECTIVES ...... 10 3 INVESTIGATIONAL PLAN ...... 10 3.1 OVERALL DESIGN ...... 10 3.2 DESIGN FORMAT ...... 10 3.3 OVERVIEW OF STUDY POPULATION ...... 10 3.4 TARGET ACCRUAL ...... 11 4 SUBJECT SELECTION ...... 11 4.1 INCLUSION CRITERIA ...... 11 4.2 EXCLUSION CRITERIA ...... 11 4.2.1 Females of childbearing potential must have a negative pregnancy test within 7 days prior to being registered for protocol therapy...... 12 4.2.2 NOTE: Subjects are considered not of child bearing potential if they are surgically sterile (they have undergone a hysterectomy, bilateral tubal ligation, or bilateral oophrectomy) or they are postmenopausal...... 12 5 STUDY PROCEDURES AND SCHEDULE OF ASSESSMENTS...... 12 5.1 INFORMED CONSENT PROCEDURES ...... 12 5.1.1 It is the responsibility of the Investigator, or a person delegated by the Investigator [the delegation log will detail who is authorised to take consent, only GCP trained individuals can take consent, and in line with sponsor SOPs) to obtain written informed consent from each subject prior to participation in this study, following adequate explanation of the aims, methods, anticipated benefits and potential hazards of the study...... 12 5.1.2 Ample time must be given for consideration by the subject before taking part. It is to be documented in the medical records when the subject information sheet (PIS) was given to the subject. The Investigator or designee must explain to the subjects that they are completely free to refuse to enter the study or to withdraw at any time during the study, for any reason. Consent will follow the process detailed in the application (IRAS) to the MainRec and MHRA, section A29. Subjects may have less than 24 hours to consider participation, however ample time will be given so that the investigator (or designated individual) is confident that the subject has had time to review, ask questions, and consent in writing...... 12 5.1.3 If new safety information results in significant changes in the risk/benefit assessment, the consent form should be reviewed and updated if necessary. All subjects, including those already being treated, should be informed of the new information, giving a copy of the revised form and give their consent to continue in the study...... 12 5.2 SCREENING AND REGISTRATION PROCEDURES ...... 12 5.3 RANDOMISATION PROCEDURES ...... 12 5.4 TREATMENT PROCEDURES ...... 12 5.5 METHOD FOR ASSIGNING SUBJECTS TO TREATMENT GROUP ...... 13 5.6 FLOW OF ASSESSMENTS ...... 13 5.7 FOLLOW UP PROCEDURES ...... 13 5.8 LABORATORY ASSESSMENTS ...... 13 5.9 RADIOLOGY OR OTHER PROCEDURE ...... 15 5.10 UNBLINDING PROCEDURES ...... 15

5.11 WITHDRAWAL ...... 15 5.11.1 Withdrawal of subjects ...... 15 5.11.2 When and How to withdraw subjects ...... 15 5.11.3 Data collection and follow-up for withdrawn subjects ...... 15 6 INVESTIGATIONAL MEDICINAL PRODUCT(S) ...... 16 DEFINITION OF EACH IMP( S) (P ARACETAMOL )...... 16 6.1 PRODUCT SOURCING MANUFACTURE AND SUPPLY ...... 16 6.2 PRE -MEDS ...... 16 6.3 PRESCRIPTION OF IMP...... 16 6.4 PREPARATION AND ADMINISTRATION OF IMP...... 16 6.5 PRIOR AND CONCOMITANT THERAPIES ...... 16 6.6 DOSE MODIFICATION /REDUCTION / DELAY ...... 16 6.7 TOXICITY PROFILES ...... 16 6.8 LABELLING AND PACKAGING ...... 16 6.9 BLINDING OF IMP...... 16 6.10 RECEIPT OF IMP SUPPLIES /S TORAGE ...... 17 6.11 DISPENSING OF IMP...... 17 6.12 RETURN OR DESTRUCTION OF IMP...... 17 7 PHARMACOVIGILANCE...... 17 7.1 GENERAL DEFINITIONS ...... 17 7.1.1 Adverse Events (AE)...... 17 7.1.2 Adverse Reaction (AR) ...... 17 7.1.3 Serious Adverse event (SAE) or Serious Adverse Reaction (SAR) ...... 17 7.1.4 Suspected unexpected Serious Adverse Reaction (SUSAR) ...... 18 7.2 INVESTIGATORS ASSESSMENT ...... 18 7.3 NOTIFICATION AND REPORTING ADVERSE EVENTS OR REACTIONS ...... 18 7.4 NOTIFICATION AND REPORTING OF SERIOUS ADVERSE EVENTS /SUSAR ...... 18 7.5 PROCEDURES FOR REPORTING BLINDED SUSAR...... 19 7.6 EXPECTED SAE S/ SAR S AND NON REPORTABLE EVENTS ...... 19 7.7 PREGNANCY AND URGENT SAFETY MEASURES ...... 19 7.8 PHARMACOVIGILANCE RESPONSIBILITIES ...... 20 8 DATA HANDLING AND RECORD KEEPING ...... 20 8.1 CONFIDENTIALITY ...... 20 8.2 STUDY DOCUMENTS ...... 20 8.3 CASE REPORT FORMS ...... 20 8.4 RECORD RETENTION AND ARCHIVING ...... 20 8.5 COMPLIANCE ...... 20 8.6 END OF STUDY DEFINITIONS ...... 21 9 CLINICAL GOVERNANCE ISSUES ...... 21 9.1 ETHICAL CONSIDERATIONS ...... 21 9.2 SUMMARY MONITORING PLAN ...... 21 9.3 AUDIT AND INSPECTION ...... 21 9.4 REPORTING OF SERIOUS BREACHES IN GCP OR TRIAL PROTOCOL ...... 21 9.5 TRIAL COMMITTEE (OVERSIGHT )...... 21 10 STATISTICS...... 21 10.1 ENDPOINTS ...... 21 10.1.1 Primary endpoints...... 22 10.1.2 Secondary Endpoints...... 22 10.1.3 Study definitions ...... 22 10.2 STATISTICAL CONSIDERATIONS ...... 22

10.2.1 Sample Size...... 22 10.2.2 Planned recruitment rate...... 22 10.3 STATISTICAL ANALYSIS ...... 22 10.4 GENERAL CONSIDERATIONS ...... 22 10.5 FREQUENCY OF ANALYSIS ...... 22 10.6 PRIMARY END POINT ANALYSIS ...... 22 10.7 SECONDARY ENDPOINT ANALYSIS...... 22 10.8 INTERIM ANALYSIS ...... 22 10.9 RANDOMISATION AND STRATIFICATION ...... 22 11 STUDY FINANCES ...... 23 11.1 FUNDING SOURCES ...... 23 11.2 SUBJECT EXPENSES +/- PAYMENTS ...... 23 12 SPONSORSHIP AND INDEMNITY...... 23 13 PUBLICATION POLICY...... 23 14 REFERENCES...... 23

Study Summary/Synopsis

Title A Single Centre Prospective Randomised Study to Investigate the CSF Pharmacokinetics of Intravenous Paracetamol in Humans

Short Title Paracetamol CSF Study

Protocol Version Protocol Version Number 1, 19/07/2010 Number and Date

Methodology Prospective randomised study

Study Duration Six months

Study Centres Single Centre

Objectives To measure the cerebrospinal levels of paracetamol and its metabolites after it has been given intravenously in humans.

Phase of the Trial IV

Number of 30 Subjects/Subjects

Main Inclusion Adult subjects aged 18 and 80 years from whom full written informed Criteria consent will be sought Subjects undergoing surgery who require a spinal anaesthetic and paracetamol as part of their routine anaesthetic Subjects who are ASA I−III (American Society of Anesthesiologists). Subjects with no known clotting abnormalities

Statistical This is a pilot study. The analysis of data will be predominantly descriptive. Methodology and The concentration time curves will be constructed for paracetamol in Analysis plasma and cerebrospinal fluid. This will also be done for the metabolites of paracetamol. The CSF:Plasma ration for paracetamol will be calculated for the different time points.

Glossary of Terms and Abbreviations

AE Adverse Event AR Adverse Reaction ASR Annual Safety Report CA Competent Authority CI Chief Investigator CRF Case Report Form CRO Contract Research Organisation CTA Clinical Trial Authorisation CTIMP Clinical Trial of Investigational Medicinal Product DMC Data Monitoring Committee EC European Commission EMEA European Medicines Agency EU European Union EUCTD European Clinical Trials Directive EudraCT European Clinical Trials Database EudraVIGILANCE European database for Pharmacovigilance GAfREC Governance Arrangements for NHS Research Ethics GCP Good Clinical Practice GMP Good Manufacturing Practice IB Investigator Brochure ICF Informed Consent Form IMP Investigational Medicinal Product IMPD Investigational Medicinal Product Dossier ISRCTN International Standard Randomised MA Marketing Authorisation MHRA Medicines and Healthcare products Regulatory Agency MRC Medical Research Council MS Member State Main REC Main Research Ethics Committee NHS R&D National Health Service Research & Development PI Principle Investigator QA Quality Assurance QC Quality Control QP Qualified Person for release of trial drug RCT Randomised Control Trial

REC Research Ethics Committee SAR Serious Adverse Reaction SAE Serious Adverse Event SDV Source Document Verification SOP Standard Operating Procedures SmPC Summary of Product Characteristics SSA Site Specific Assessment SSAR Suspected Serious Adverse Reaction Subject An individual who takes part in a clinical trial SUSAR Suspected Unexpected Serious Adverse Reaction TMG Trial Management Group TSC Trial Steering Committee

1 Introduction

1.1 Background Paracetamol is a commonly used analgesic agent for peri-operative pain management. It forms the step 1 of the WHO Pain Ladder and constitutes an essential component of balanced analgesia to produce synergistic effects while reducing the need for opioid analgesics. Although it has been used orally since early 19 th century, it has been recently formulated as a ready to use intravenous preparation in 2004. This new development is potentially beneficial to subjects with an added advantage of 100% bioavailability and a relative ease of administration in intra- and post-operative period.

There is evidence that, when administered intravenously, paracetamol has a fast onset of action within 5 −10 minutes with a peak analgesic action within 15 minutes.

Although paracetamol has been used for decades there is relative little evidence about its CSF pharmacokinetics particularly after intravenous administration in humans. Paracetamol readily penetrates into the cerebrospinal fluid in animal studies and in children. This could account for the rapid central analgesic and antipyretic action of intravenous paracetamol however this remains to be studied in adults yet.

Following deacetylation to p-aminophenol, paracetamol is conjugated with arachidonic acid to form N-arachidonoylphenolamine (AM404). AM404 activates vanilloid subtype 1 receptors and inhibits cellular anandamide uptake, allowing cannabinoids like anandamide to remain in the synaptic clefts. AM404 is also a Transient Receptor Potential Vanilloid Type 1 (TPRV1) agonist, which is activated by capsaicin.

Cannabinoids and paracetamol both probably produce descending spinal inhibition. They both lower body temperature clinically and have subjective effects of euphoria and relaxation shared by aniline analgesics. CB1 antagonists completely prevent the analgesic action of paracetamol, lending support to this proposed mechanism of action.

1.2 Investigational Medicinal Product (IMP) Intravenous paracetamol is used for the short term treatment of moderate pain, especially following surgery. The precise mechanism of the analgesic properties of paracetamol has yet to be established, it may involve peripheral and central actions.

1.3 Preclinical Data Intravenous paracetamol is already in use in humans in clinical practice. Paracetamol pharmacokinetics is linear up to 2g after single administration. The maximal plasma concentration (Cmax) of paracetamol observed at the end of 15-minutes intravenous infusion of 1g is 30 µg. The volume of distribution is 1l/kg. Paracetamol is not extensively bound to plasma proteins. Paracetamol is metabolised in the liver by glucuronic acid conjugation and sulphuric acid conjugation. The metabolites of paracetamol are mainly excreted in the urine.

1.4 Clinical Data to Date In children it has been shown that paracetamol enters the CSF readily. In healthy children, paracetamol is detected 5 minutes after administration of an intravenous dose of 15mg/kg, but the concentration is below 1mg/l with peak levels appearing to occur 60-100 minutes post administration, ranging from 11-18mg/l. In children with head injuries and intracranial pathology, peak CSF concentrations of paracetamol were found at 210 minutes post paracetamol, at levels of 20mg/ml, but nasogastric paracetamol at a dose of 40mg/kg was given.

1.5 Rationale and Risk/Benefits Paracetamol readily penetrates into the cerebrospinal fluid in animal studies and in children. This could account for the rapid central analgesic and antipyretic action of intravenous paracetamol however this remains to be studied in adults yet.

Further the exact mechanism of action of paracetamol is not fully understood and essentially involves a combination of hypotheses. It is known to have a central effect perhaps by selective inhibiton of COX. However it is hypothesised that it could also perhaps modulate the serotoninergic system and may have an indirect activation of CB 1 (cannabinoid) receptors, through its novel metabolite AM404.

This study aims to elucidate cerebrospinal levels of paracetamol and to measure the concentration of its metabolites, including AM404 in CSF. This may, in turn, give us useful information on the mechanism of action of paracetamol.

There are no foreseen extra risks to the subject as they will be undergoing elective surgery under spinal anaesthesia and receiving intravenous paracetamol as a routine part of their anaesthetic plan.

2 Study Aims and Objectives

Primary : To measure the cerebrospinal levels of paracetamol and its metabolites after it has been given intravenously in humans. Secondary :

1. To measure the effect of paracetamol on prostaglandin synthesis in the CSF 2. To measure the concentration of AM404 in the human CSF following intravenous administration of paracetamol.

3 Investigational plan

3.1 Overall Design This is a single centre prospective randomised study. Subjects will be randomised to receive intravenous paracetamol at three different time points prior to having a spinal block as a routine part of their anaesthesia.

3.2 Design Format After obtaining informed consent, 30 subjects undergoing elective surgery under spinal anaesthesia as a routine part of their anaesthetic plan will be randomised to receive intravenous paracetamol at 15 minutes, 30 minutes or 120 minutes prior to their spinal block. After the spinal tap and prior to administering any anaesthetic drugs, 5 ml of the cerebrospinal fluid will be sampled for analysis. A 10 ml of blood sample will be also obtained simultaneously to measure the plasma paracetamol and AM404 levels.

The cerebrospinal fluid and the plasma would be sent to the laboratory to be analysed. Paracetamol, AM404 levels and prostaglandin E 2 (PGE 2), used as a measure of cyclooxygenase (COX) activity would be measured.

3.3 Overview of Study population Male and female subjects, between 18 and 80 years of age, for elective surgery under spinal anaesthesia who would receive intravenous paracetamol.

3.4 Target Accrual Subjects who are undergoing surgery who meet the criteria will be identified from the elective operating theatre lists. The surgeon and anaesthetist responsible for their care will be consulted. The subjects will be provided a subject information sheet and invited to participate into the study. After obtaining informed consent they would be randomised to receive paracetamol 15 minutes, 30 minutes or 120 minutes prior to the proposed spinal anaesthetic block.

4 Subject Selection

4.1 Inclusion Criteria

• Adult subjects aged 18 and 80 years from whom full written informed consent will be sought • Subjects undergoing surgery who require a spinal anaesthetic and paracetamol as part of their routine anaesthetic • Subjects who fall into the American Society of Anesthesiologists scoring system of ASA I−III. • Subjects with no known clotting abnormalities.

4.2 Exclusion Criteria

• Subjects with any contraindication to spinal anaesthesia • Subjects with known clotting abnormalities • Pregnant or lactating women • Subjects who have known hypersensitivity to paracetamol • Patients with severe hepatocellular insufficiency • Patients who are taking a regular dose of paracetamol already • Any subjects who are deemed unsuitable by the investigators due to medical reasons

4.2.1 Females of childbearing potential must have a negative pregnancy test within 7 days prior to being registered for protocol therapy.

4.2.2 NOTE: Subjects are considered not of child bearing potential if they are surgically sterile (they have undergone a hysterectomy, bilateral tubal ligation, or bilateral oophrectomy) or they are postmenopausal.

5 Study Procedures and Schedule of assessments

5.1 Informed consent procedures

5.1.1 It is the responsibility of the Investigator, or a person delegated by the Investigator [the delegation log will detail who is authorised to take consent, only GCP trained individuals can take consent, and in line with sponsor SOPs) to obtain written informed consent from each subject prior to participation in this study, following adequate explanation of the aims, methods, anticipated benefits and potential hazards of the study.

5.1.2 Ample time must be given for consideration by the subject before taking part. It is to be documented in the medical records when the subject information sheet (PIS) was given to the subject. The Investigator or designee must explain to the subjects that they are completely free to refuse to enter the study or to withdraw at any time during the study, for any reason. Consent will follow the process detailed in the application (IRAS) to the MainRec and MHRA, section A29. Subjects may have less than 24 hours to consider participation, however ample time will be given so that the investigator (or designated

individual) is confident that the subject has had time to review, ask questions, and consent in writing.

5.1.3 If new safety information results in significant changes in the risk/benefit assessment, the consent form should be reviewed and updated if necessary. All subjects, including those already being treated, should be informed of the new information, giving a copy of the revised form and give their consent to continue in the study.

5.2 Screening and registration Procedures Subjects who are undergoing surgery who meet the criteria will be identified from the elective operating theatre lists. The surgeon and anaesthetist responsible for their care will be consulted. The subjects will be provided with a subject information sheet and invited to participate into the study. After obtaining informed consent they would be randomised to receive paracetamol 15 minutes, 30 minutes or 120 minutes prior to the proposed spinal anaesthetic block.

5.3 Randomisation procedures Randomisation will be done by sealed envelopes with the study group allocation (sampling at 15, 30 or 120 minutes) in the envelope. A sealed envelope will be selected at random by the investigator (s) or a delegated study team member, on the day of the procedure.

5.4 Treatment procedures An intravenous cannula will be inserted either on the ward or on the holding bay in theatres. Intravenous paracetamol will be given. The subject will be taken to the anaesthetic room and routine monitoring will be instituted. The spinal block will be performed. At this time, prior to injecting the local anaesthetic, 5ml of CSF will be withdrawn. 10ml of blood will also be taken. The subject will then go into theatre to have their surgical procedure.

5.5 Method for assigning subjects to treatment group Subjects will be randomised to one of the treatment groups. The randomisation process will be verified by a study statistician.

5.6 Flow of assessments a) Pre-operative visit and informed consent obtained from subject. b) Subject randomised to one of three groups, to receive IV paracetamol 15, 30 or 120 minutes prior to their planned spinal anaesthetic block. c) At the time of the spinal anaesthetic block, 5ml of CSFand 10ml blood obtained d) Subject to go into theatre for their surgery e) Follow up visit by study investigator

5.7 Follow up procedures The subject will be seen postoperatively by a study investigator to monitor if any adverse effects have occurred.

5.8 Laboratory Assessments

A copy of the trial protocol will be provided to the laboratory (laboratory manager) named in this trial. The lab manager will also be provided with approvals from the sponsor (JRO), MHRA and MainRec approvals and subsequent amendments. The lab manager (or designated individual) will maintain a laboratory trial file to show compliance with Good Clinical Practice.

The CSF and blood will be divided into 1ml aliquots for the different analyses, hence determination of the paracetamol and AM404 concentrations and PGE2 levels.

Paracetamol assay The concentration of paracetamol in plasma and cerebrospinal samples will be measured using a commercial enzyme assay. This assay will be performed according to the manufacturer’s instructions (Cambridge Life Sciences, Cambridge, UK). Five microlitres of CSF or plasma from subjects treated with intravenous paracetamol will be added to a 96 well plate. Also, 5µl of distilled water will be added to one of the wells to act as blank and 5µl of standard (containing 2mmol/L paracetamol) will be added to one of the wells. To each of the well 50µl of enzyme reagent, constituted in 10ml of assay diluent, will be added. One hundred microlitres of colour reagent A and 100µl of colour reagent B will be added to all the wells. The plate will be left at room temperature for 5min and then read at 630nm with a plate reader (GENios Tecan, Jencons).

The concentration of paracetamol in the samples will be calculated using the equation:

Paracetamol (mmol/L) = Sample Abs – blank x 2 Standard Abs - blank

Where: Sample Abs = absorbance in sample wells Blank = absorbance in wells containing water Standard Abs = absorbance in standard wells 2 = concentration of paracetamol in the standard

Prostaglandin extraction

Blood or CSF will be collected in disodium-EDTA containing 10µM indomethacin. The plasma will be separated from blood by centrifugation at 1,600g for 20 minutes at +4oC. Prostaglandins will extracted from plasma and CSF using the protocol below. and the samples will be mixed with 2N HCl at a 1:20 dilution (for example for 1ml of sample 50µl of 2N HCL will be added). The samples will be left on ice for 20 minutes. C-18 Sep-Pak cartridges will be conditioned with 4ml ethanol followed by 4ml distilled water at a flow rate of 5-10ml/min. The samples will then be applied to the columns at a flow rate of 5ml/min. The columns will then be washed in 4ml distilled water followed by 4ml of 15% (v/v) ethanol in distilled water. The samples will be eluted with 2ml of ethyl acetate at a flow rate of 5ml/min. The samples will be dried under vacuum and then stored at – 80oC ready for PGE2 measurement.

PGE 2 enzyme immunoassay This commercial enzyme immunoassay (EIA) kit will be carried out according to the manufacturer’s instructions. Briefly, the assay buffer will be diluted in 500ml distilled water to give 0.1M phosphate buffer pH 7.5 containing 0.9% (w/v) bovine serum albumin and 0.5% (w/v) kathon. Standard PGE2 will be serially diluted to give concentrations between 2.5–320 pg/well. Both the anti-PGE2 antibody and PGE2 conjugated to horseradish peroxidase will be dissolved in 6ml assay buffer and then diluted 1:3 in assay buffer. The microtitre plate coated with goat anti-mouse IgG will be set with sufficient wells to run all the standards and samples. The following will be added to each well in the volumes indicated in microlitres:

Substrate blank Non-specific Zero Standards Samples binding Assay buffer ------100 µl 50 µl ------Standard ------50 µl ------Sample ------50 µl Anti-body ------50 µl 50 µl 50 µl

Conjugate ------50 µl 50 µl 50 µl 50 µl Substrate 150 µl 150 µl 150 µl 150 µl 150 µl

The plate will be incubated at ambient temperature for 1hr on a microplate shaker. At the end of this incubation period, the plate will be washed four times with wash buffer and immediately the substrate (3,3’,5,5’tetramethylbenzidine (TMB)/hydrogen peroxide in 20% (v/v) dimethylformamide) will be added and incubated for 30 minutes. Finally, the plate will be read at 630nm and the results calculated using the following equation: %B/B 0 = (standard/sample OD - NSB OD) x 100 (B 0 OD - NSB OD) where B = Bound (peroxidase-conjugated PGE2 in samples bound to antibody) B 0 = Amount of peroxidase-conjugated PGE2 in zero standard wells OD = Optical density (absorbance values obtained from plate reader) NSB = Non-specific binding wells The OD values from all the duplicate wells will be averaged and the percentage binding (%B/B 0) for the standards and samples will be calculated using the above equation. The %B/B 0 for the standards was then plotted against the concentrations using GraphPad Prism 3.0 and the amount of PGE2 in the samples will then be read of the standard curve.

5.9 Radiology or other procedure Not applicable

5.10 Unblinding procedures Not applicable

5.11 Withdrawal

5.11.1 Withdrawal of subjects Subjects may be withdrawn from the study if any of the following occur:

1. Subject withdraws consent 2. Subjects deemed unsuitable to continue with the study protocol due to safe issues, as judged by the investigator(s) 3. Subjects unable to comply with the study protocol 4. Subjects have intolerable toxicity to the study drug 5. Subjects experience any intercurrent illness

5.11.2 When and How to withdraw subjects Subjects are at any time free to withdraw from the study without prejudice to further treatment. Such subjects will always be asked about the reason(s) and the presence of adverse events.

Adverse events will be followed up. As collection of blood and CSF samples is an integral part of this study, the subject will be withdrawn from further study participation.

5.11.3 Data collection and follow-up for withdrawn subjects Data collection will be in the form of a case report form (CRF). The withdrawn subjects and all identifiable data collected will be withdrawn from the study file and excluded in the analysis. Data that is not identifiable by the research team will be retained. Withdrawn subjects will be followed up in the same way as include subjects are at a post-operative visit. The subjects who withdraw consent or drop out will be replaced by another subject. The withdrawal will be documented in the CRF and medical notes.

The study conduct will remain unchanged. The participants are not obliged to give the reason for their withdrawal. We do not anticipate that the study procedures will cause subjects to withdraw their consent as most of the study protocol is part of a routine anaesthetic plan, except the blood and CSF sampling. Therefore, we do not anticipate that subjects withdrawing their consent will result in a protocol amendment.

6 Investigational medicinal product(s)

Definition of each IMP(s) (Paracetamol)

6.1 Product sourcing manufacture and supply http://www.mhra.gov.uk/Howweregulate/Medicines/Licensingofmedicines/Manufacturersan dWholesaleDealerslicences/index.htm ] Bristol Meyer Squibb Pharmaceuticals PL 11184/0094

6.2 Pre-meds Not applicable

6.3 Prescription of IMP The IMP will be prescribed by an anaesthetist. A prescription template will be agreed with the clinical trials pharmacist and version controlled.

6.4 Preparation and Administration of IMP The IMP is presented in a glass vial. It will be administered intravenously via a giving set over 15 minutes.

6.5 Prior and Concomitant Therapies The IMP will not be given if it has been given in the preceding 6 hours. It will not be given in conjunction with probenicid, enzyme inducing substances and oral anticoagulants.

6.6 Dose modification/reduction/ delay The dose is 1g per administration for adults weighing more than 50 kg. These are the only subjects which will be recruited in the study.

6.7 Toxicity profiles Toxicity occurs at doses of 7.5g or more in a single administration. Such doses will not be used in this study.

6.8 Labelling and packaging The standard packaging will be used. A clinical trial label will be used – this has been agreed with the clinical trial pharmacist. A copy of the trial label will be sent to the MainRec and MHRA for submission and approval.

6.9 Blinding of IMP Not applicable.

6.10 Receipt of IMP Supplies/Storage IV paracetamol is routinely stocked in theatres and given to virtually all subjects undergoing surgery. IMP management processes will be overseen by the trial team, and processes agreed with the clinical trial pharmacist.

6.11 Dispensing of IMP The IMP will be dispensed from the theatre stock.

6.12 Return or Destruction of IMP The empty vial will be disposed off in the sharps bin. The trial team is to maintain a log to detail IMP accountability and compliance. This process will be agreed with the clinical trials pharmacist.

7 Pharmacovigilance

7.1 General definitions

7.1.1 Adverse Events (AE)

An AE is any untoward medical occurrence in a subject to whom a medicinal product has been administered, including occurrences which are not necessarily caused by or related to that product. An AE can therefore be any unfavourable and unintended sign (including an abnormal laboratory finding), symptom or disease temporarily associated with the use of an Investigational Medicinal Product (IMP), whether or not considered related to the IMP.

7.1.2 Adverse Reaction (AR)

An AR is any untoward and unintended response in a subject to an Investigational Medicinal Product (IMP), which is related to any dose administered to that subject. All adverse events judged by either the reporting investigator or the Sponsor as having a reasonable causal relationship to a medicinal product qualify as adverse reactions. The expression reasonable causal relationship means to convey in general that there is evidence or argument to suggest a causal relationship.

7.1.3 Serious Adverse event (SAE) or Serious Adverse Reaction (SAR)

Serious Adverse Event (SAE)

An SAE fulfils at least one of the following criteria:

• Is fatal – results in death (NOTE: death is an outcome, not an event) • Is life-threatening • Requires insubject hospitalisation or prolongation of existing hospitalization • Results in persistent or significant disability/incapacity • Is a congenital anomaly/birth defect

Suspected Serious Adverse Reaction (SSAR)

An SSAR is an adverse reaction that is classed as serious and which is consistent with the information about the medicinal product as set out in the Summary of Product Characteristics (SmPC) or Investigator’s Brochure (IB) for that product.

7.1.4 Suspected unexpected Serious Adverse Reaction (SUSAR) The definition of a SUSAR is any suspected unexpected adverse reaction related to an IMP that is both unexpected and serious. In this case the event is not outlined in the Summary of Product Characteristics (SmPC) or Investigator’s Brochure (IB) for that product.

7.2 Investigators Assessment

Seriousness The Chief/Principal Investigator responsible for the care of the patient, or in his absence an authorised medic within the research team, is responsible for assessing whether the event is serious according to the definitions given in section 7.1.

7.2.2 Causality The Investigator must assess the causality of all serious adverse events/reactions in relation to the trial treatment according to the definition given. If the SAE is assessed as having a reasonable causal relationship, then it is defined as a SAR.

7.2.3 Expectedness The investigator must assess the expectedness of all SARs according to the definition given. If the SAR is unexpected, then it is a SUSAR.

7.2.4 Severity The Investigator must assess the severity of the event according to the following terms and assessments. The intensity of an event should not be confused with the term “serious” which is a regulatory definition based on patient/event outcome criteria.

Mild : Some discomfort noted but without disruption of daily life Moderate : Discomfort enough to affect/reduce normal activity Severe : Complete inability to perform daily activities and lead a normal life

7.3 Notification and reporting Adverse Events or Reactions If the AE is not defined as serious, the AE is recorded in the study file and the subject is followed up by the research team. The AE is documented in the subject’s notes (where appropriate and the CRF).

7.4 Notification and Reporting of Serious Adverse Events/SUSAR

7.4.1 All Serious Adverse Event (SAEs) that require reporting (see section 7.6 for cases where reporting is not required) will be recorded in the subjects’ notes, the CRF, the sponsor SAE form and reported to the Joint Research and Development Office (JRO)/ IMP provider (if applicable) within 24 hours of the CI or PI or co-investigators becoming aware of the event. Nominated co- investigators will be authorised to sign the SAE forms in the absence of the CI. Please ensure that the sponsor has been informed of these nominated co-investigators.

7.4.2 Suspected Unexpected Serious Adverse Reactions (SUSARs) that occur during the trial will be reported to the JRO within 24 hours of the CI or co-investigator becoming aware of the event. SUSARs should be reported to the sponsor (JRO Office) within 24 hours as the sponsor has a legal obligation to report this to the MHRA within 7 days (for fatal or life-threatening SUSARs) or 15 days for all other SUSARs. In the case of multicentre studies, the PI or the co-investigators at the participating site must inform the CI within 24 hours of the event. The CI or co-investigators at the co-ordinating site must inform the sponsor (JRO) immediately to allow reporting to the MHRA within the allocated timelines. The CI will need to complete the CIOMS form in conjunction with the sponsor SAE form to be sent to the MHRA by the sponsor. If warranted, an investigator alert may be issued, to inform all investigators involved in any study with the same drug (or therapy) that this serious adverse event has been reported.

The original and any subsequent follow up of Serious Adverse Event Forms and CIOMS forms (where applicable), together with the fax confirmation sheet must be kept with the TMF at the study site.

7.4.3 The Annual Safety reports (ASR) will be sent by the CI to the sponsor, the MREC and MHRA (the date of the anniversary is the date on the “notice of acceptance letter” from the MHRA) using the sponsor ASR form. The CI will carry out a risk benefit analysis of the IMPs encompassing all events having arisen on the trial.

7.4.4 The CI will send the annual progress report to the main REC (the anniversary date is the date on the MREC “favourable opinion” letter from the MREC) and to the sponsor.

7.5 Procedures for reporting Blinded SUSAR Not applicable

7.6 Expected SAEs/ SARs and non reportable events Doses higher than those recommended entail the risk of liver damage. Such high doses will not be used in the study. Paracetamol is used with caution in cases of hepatocellular insufficiency, severe renal insufficiency (creatinine clearance of less than 30 ml/min), chronic alcoholism, chronic malnutrition and dehydration. Caution is taken with the concomitant use of enzyme inducing substances and anticoagulants. Adverse drug reactions are rare (<1/1000 , >1/10000) such as malaise, hypotension, and increased levels of hepatic transaminases. Very rare reactions (1<1/10000) are hypersensitivity, thrombocytopenia, leucopenia and neutropenia. Very rare cases of hypersensitivity reactions range from simple skin rash or urticaria to anaphylactic shock have been reported and require discontinuation of treatment. Cases of erythema, flushing, pruritus and tachycardia have been reported.

7.7 Pregnancy and Urgent Safety Measures

Pregnancy Pregnant females will not be included in the study. Clinical experience of the intravenous administration of paracetamol is limited. However, epidemiological data from the use of oral therapeutic doses of paracetamol indicate no undesirable effects in pregnancy or on the health of the foetus/newborn infant.

Urgent Safety Measures The CI may take urgent safety measures to ensure the safety and protection of the clinical trial subjects from any immediate hazard to their health and safety, in accordance with Regulation 30. The measures should be taken immediately. In this instance, the approval of the Licensing Authority Approval prior to implementing these safety measures is not required. However, it is the responsibility of the CI to inform the sponsor, Main Research Ethics Committee (via telephone) and the MHRA (via telephone for discussion with the medical assessor at the clinical trials unit) of this event immediately .

The CI has an obligation to inform both the MHRA and Main Ethics Committee in writing within 3 days , in the form of a substantial amendment. The sponsor (JRO) must be sent a copy of the correspondence with regards to this matter.

7.8 Pharmacovigilance responsibilities Subjects with known hypersensitivity reactions to intravenous paracetamol will not be given the drug. In cases where paracetamol should be used with caution, such as hepatocellular insufficiency and renal failure, the subjects will be not be recruited into the study. For those subjects who are recruited in the study, there will be anaesthetist present throughout the administration of paracetamol, the spinal anaesthetic, the duration of surgery and in the recovery area. The subject will be closely monitored during all of these procedures.

8 Data Handling and Record Keeping

8.1 Confidentiality This study will be registered with the Trust Information Governance Manager at Barts and The London NHS Trust, ensuring the confidentiality of personal data. A signed declaration form (data protection form) of the group to comply with the Data Protection Act and the Department of Health Code of Confidentiality will be supplied.

8.2 Study Documents Study files will be stored in a locked Pain and Anaesthetic Research Centre (PARC) office. All electronic data will be stored in a password-protected computer in a locked office. All study data will be stored in a custom-made password protected database, which identifies the subjects only by initials and a study number. Only members from the research team will have access to the raw data. Only pooled anonymised data will be published.

The principle investigator and co-investigators will have control and act as the custodian for the data generated by the study.

8.3 Case Report Forms These will contain the following: eligibility/exclusion criteria checklist, visit details-date, drug and dose, adverse events page for toxicities, withdrawal from study, follow up of outcomes, death, prior/current medication, medical history, SAE/SUSAR form.

8.4 Record Retention and Archiving Essential study documents will be retained by the investigator and other study participants in the PARC for a minimum of two years, and thereafter archived for a longer period (at least 20 years) at Barts and the London Trust after study completion.

8.5 Compliance

The CI&PI will ensure that this study is conducted in accordance with the latest version of the “Declaration of Helsinki” ( http://www.wma.net/e/policy/b3.htm ) and the UK Legal Framework for Clinical Trials of IMPs, SI 2004/1031 and subsequent relevant amendments. The study must adhere to the principles outlined in the Guidelines for Good Clinical Practice”.

8.6 End of study definitions The end of study is defined as “the last visit of the last subject undergoing the study”.

9 Clinical governance issues

9.1 Ethical considerations The study will be performed in accordance with ethical principles that have their origin in the Declaration of Helsinki and are consistent with the ICH/Good Clinical Practice and applicable regulatory requirements.

The Informed Consent Form will incorporate wording that complies with relevant data protection and privacy legislation.

An Ethics Committee will review the final study protocol, including the final version of the Informed Consent Form and other written information to be provided to subjects. Before enrolment of any subject into the study, the national regulatory authority according to local regulations will approve the final study protocol and Informed Consent Form.

All investigators have completed Good Clinical Practice and Research Governance courses.

No conflict of interest has been declared within the study.

9.2 Summary monitoring plan The study will be monitored in accordance with the local Research and Development protocols. The Barts and The London monitoring template will be completed in a continuous fashion throughout the study and kept up to date by the co-investigators and monitors.

9.3 Audit and inspection The study will undergo audit and inspection in line with Good Clinical Practice guidelines. Audit may be carried out by the Sponsor to ensure that the trial procedures and data collected comply with the protocol and standard operating procedures, and that the data collected is correct and complete. Study monitors will have full and direct access to all source documents and other trial-related data.

9.4 Reporting of Serious breaches in GCP or trial protocol Serious breaches in GCP or trial protocol will be reported by the chief investigator to the Research and Development office (sponsor).

9.5 Trial committee (oversight) This is a single centre study. The study will be monitored in accordance with the sponsor Standard Operating Procedures (SOPs) and the trial monitor will adhere to any sponsor monitoring tools and SOPs. The CI will review the trial progress and review data on a continual basis.

10 Statistics

10.1 Endpoints

10.1.1 Primary endpoints To measure the cerebrospinal levels of paracetamol after it has been given intravenously.

10.1.2 Secondary Endpoints To measure the effect of paracetamol on prostaglandin synthesis in the CSF and to measure the concentration of AM404 in the human CSF following intravenous administration of paracetamol. The plasma levels of paracetamol post administration of intravenous paracetamol will be measured too. The plasma to CSF ratio of paracetamol concentrations can then be calculated.

10.1.3 Study definitions

10.2 Statistical considerations 10.2.1 Sample Size 30 subjects will be recruited. This is a pilot study so the sample size has not been calculated.

10.2.2 Planned recruitment rate Subjects will be recruited on the elective operative lists. It is estimated that 3-5 subjects will be recruited per week.

10.3 Statistical Analysis The concentrations of paracetamol in the cerebrospinal fluid will be plotted against time. The CSF concentrations and plasma concentrations of paracetamol will be plotted against time on a graph and the plasma to CSF ratio will be calculated. The transfer in and out of the CSF was assumed to be by passive diffusion, and therefore the rates of transfer in and out of the CSF would be dependent on concentration. The influx and efflux of drug would therefore be exponential processes with rate constants ka and l, respectively. The drug CSFconcentrations (Ct) can be fitted to a simple pharmacokinetic model.

10.4 General considerations

10.5 Frequency of Analysis The statistical analysis will take place at the end of the study by a qualified statistician.

10.6 Primary end point analysis As above

10.7 Secondary endpoint analysis As above

10.8 Interim Analysis The analysis will be done at the end of the study. However the CI will maintain an oversight of data throughout the trial.

10.9 Randomisation and Stratification Non-participating members from the Pain and Anaesthesia Research Group will prepare a randomisation table. Subjects enrolling into the study will be assigned a study number starting from 01. Once the subject has been allocated a study number, the third party (non-blinded) researcher will open the appropriate numbered envelope to reveal the pre assigned randomised study group allocation.

11 Study Finances

11.1 Funding Sources This is an internally funded study by the Pain and Anaesthesia Research Centre. 11.2 Subject expenses +/- payments Subjects will not receive further expenses or payments.

12 Sponsorship and Indemnity

Sponsorship and indemnity will be provided by Barts and The London NHS Trust

13 Publication policy

The results of this study will be published in a peer-reviewed journal. The study results may be published and/or presented at scientific meetings. All manuscripts and abstracts, which refer to data originating from the trial, must be submitted to the Sponsor before publication. The Sponsor has the right to refuse the results for registration purposes, internal presentation and promotion.

14 References

1. Kumpulainen E et al. Paracetamol (acetoaminophen) penetrates readily into the cerebrospinal fluid of children after intravenous administration. Pediatrics 2007;119(4)(4):766-771 2. Aronoff DM.Clin Pharmacol ther 2006;79: 9-19 3. Boutaud Proc Natl Acad Sci USA 2002; 99:7130-5 4. Bonnefont J. et al. Drugs 2003; 63(2), 1-4 5. Allouia A et al. Paracetamol exerts a spinal tropisetron-reversible, antinociceptive effect in an inflammatory pain model in rats. Eur J Pharmacology 2002;443 (1-3):71-7 6. Tjolsen A. Eur J Pharmacol 1991;193:193 7. Pickering G. Clin Pharmacol Ther 2006;79:371 8. Ottani A et al. Eur J Pharmacol 2006;531 (1-3):280-281 9. Hogestatt EA et al. Conversion of acetoaminophen to the bioactive N-acylphenolamine AM404 via fatty acid amide hydrolase-dependent arachidonic acid conjugation in the nervous system. J Biol Chem 2005; 280(36):31405-31412 10. Anderson BJ at al. Paracetamol plasma and cerebrospinal fluid pharmacokinetics in children. Br J Clin Pharmacol 2003; 58: 297-302 11. Mehta et al. Intravenous parecoxib rapidly leads to COX-2 Inhibitory Concentration of Valdecoxib in the Central Nervous System. Clinical Pharmacology & Therapeutics 2007;00 (0)