permeation across the blood- brain and blood-cerebrospinal fluid barriers

Item Type text; Dissertation-Reproduction (electronic)

Authors Abbruscato, Thomas John, 1970-

Publisher The University of Arizona.

Rights Copyright © is held by the author. Digital access to this material is made possible by the University Libraries, University of Arizona. Further transmission, reproduction or presentation (such as public display or performance) of protected items is prohibited except with permission of the author.

Download date 04/10/2021 05:24:05

Link to Item http://hdl.handle.net/10150/282429 INFORMATION TO USERS

This manuscript has been reproduced from the microfihn master. UMI fihns the text direct^ from the original or copy submitted. Thus, some thesis and dissertation copies are in typewriter &ce, while others may be from any type of computer printer.

The quality of this reproductioii is dependent upon the quality of the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleedthrough, substandard margins, and improper alignment can adversely affect reproduction.

In the unlikely event that the author did not send UMI a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion.

Oversize materials (e.g., maps, drawings, charts) are reproduced by sectioning the original, b^inning at the upper left-hand comer and continuing from left to right in equal sections with small overlaps. Each original is also photographed in one exposure and is included in reduced form at the back of the book.

Photographs included in the original manuscript have been reproduced xerographically in this copy. Higher quality 6" x 9" black and white photographic prints are available for any photographs or illustrations appearing in this copy for an additional charge. Contact UMI directly to order. UMI A BeU A Howell Infonnation Compai^ 300 North Zeeb Road, Ann Aibor MI 48106-1346 USA 313/761-4700 800/521-0600

OPIOID PEPTIDE PERMEATION ACROSS THE BLOOD-BRAIN AND BLOOD- CEREBROSPINAL FLUID BARRIERS

by

Thomas John Abbruscato

A Dissertatioa Submitted to the Faculty of the

COMMITTEE ON PHARMACOLOGY AND TOXICOLOGY (Graduate)

In Partial Fulfillment of the Requirements For the Degree of

DOCTOR OF PHILOSOPHY

In the Graduate College

THE UNIVERSITY OF ARIZONA

1997 UMX Number: 9806814

UMI Microform 9806814 Copyright 1997, by UMI Company. All rights reserved.

This microform edition is protected against unauthorized copying under Title 17, United States Code.

UMI 300 North Zeeb Road Ann Arbor, MI 48103 9

THE UNIVERSITY OF ARIZONA ® GRADUATE COLLEGE

As members of the Final Examination Committee, we certify that we have

read the dissertation prepared by Thomas John Abbruscato

entitled PERMEATION ACROSS THE BLOOD-

BRAIN AND BLOOD-CEREBROSPINAL FLUID BARRIERS

and recommend that it be accepted as fulfilling the dissertation

requirement for the Degree of Doctor of Philosophy

Final approval and acceptance of this dissertation is contingent upon the candidate's submission of the final copy of the dissertation to the Graduate College.

I hereby certify that I have read this dissertation prepared under my direction and recommend that it be accepted as fulfilling the dissertation requirement.

W Dissertation Director Date 3

STATEMENT BY AUTHOR

This dissertation has been submitted in partial flilfiUment of requirements for an advanced degree at The University of Arizona and is deposited in the University of Library to be made available to borrowers under rules of the Library.

Brief quotations from this dissertation are allowable without special permission, provided that accurate acknowledgment of source is made. Requests for permission for extended quotation from or reproduction of this manuscript in whole or in part may be granted by the head of the major department or the Dean of the Graduate College when in his or her judgement the proposed use of the material is in the interests of scholarship. In all other instances, however, permission must be obtained from the author..

SIGNED: 4

Acknowledgements

I would first like to thank my advisor Dr. Thomas P. Davis for his support and encouragement throughout my undergraduate and graduate studies. He is an excellent role model and has influenced me both scientifically and personally. I thank him for his consistent drive to develop all aspects of my scientific method. I would also like to thank the members of my committee, namely Dr. Henry Yamamura and Dr. Robert Dorr for their pharmacological expertise, and Dr. A. Jay Gandolfi and Dr. Klaus Brendel for their chemical and toxicological assistance with my project.

I have also been fortunate enough to work with many people in the Davis lab. Dr. Steve Weber played a major role in teaching me what it takes to be a research scientist. Dr. Elizabeth Brownson exposed me to many aspects of neuroscience and helped to apply them to my project. As well, she introduced me to my wife Robyn. Dr. Sarah Thomas nee' Williams trained me to proficient in the in situ brain perfusion technique and helped me to develop my scientific writing. I will always be gratefiil for her attention Dr. Richard Egleton was a key influence in helping me finished my dissertation with his expertise in the area of the blood-brain barrier. Dr. Chris Konkoy helped me immensely with his extensive pharmacology background. Terry Gillespie provided assistance with the analytical and enzymatic aspects of my project. His expertise in peptide research extended the scope of my project. Dr. Matthew Rounseville also helped me also with his experience in molecular biology. I would also like to thank a long time fiiend. Brad Merrill, who was instrumental in getting me started in biomedical research.

I would also like to thank the other members of the lab that I have worked with over the years; Dr. Steve Waters, Dr. Barbara Mania-Famell, Dr. Dave Clark, Dr. Mary Oakes, Dr. Pierre Konings, Craig Mayr, Sharon Hom and Vincent Hau. The graduate students also deserve great thanks for making the past four years enjoyable, especially C.J. Kovelowski, Art Riegal, Marc Oshiro, Mark Bowers, Sandi Wegert, Kim Mitchell and Jack Adams.

I would especially like to thank my family for their love and support: mom (Jackie), dad (Joseph), Joe and Tim.

Lastly, I would like to thank my wife, Robyn, who has supported me throughout my graduate studies and understands my love for science. She brings out the best characteristics in myself and I love her for that. She is my life long best fiiend. 5

Dedication

This work is dedicated to the woman I love, my wife Robyn, who has always encouraged and supported me. 6

TABLE OF CONTENTS

LIST OF FIGURES 9

LIST OF TABLES 11

ABSTRACT. 13

1- General Introduction 15 Anatomy and physiology of blood-brain and blood-CSF barriers IS History of blood-brain barrier. 15 Site of the blood-brain barrier. 15 Anatomy and physiology of the blood-brain barrier. 16 Types of transport across the blood-brain barrier. 21 Periendothelial structures...... 22 Cerebral spinal fluid. 24 Anatomy and physiology of the blood-CSF barrier. 26 Techniques to study blood-CNS permeability and stability of drugs 31 Brain uptake index...... 31 Intravenous bolus injection 32 Brain perfusion studies. 33 Positron emission tomography. 34 Octanol/saline partition coefficient. 35 HPLC capacity factor. 35 Isolated cultures of cerebral capillary endothelial cells. 36 Endogenous opioid and opioid pharmacology ....37 History ofopioids. 37 Endogenous opioid peptides. 38 Midtiple types. 39 Regional CNS distribution of opioid receptors. 41 Mechanism of opioid inhibitory action on neurotransmission 42 Peptide drug delivery to the brain 44 Peptides as neuropharmaceuticals. 44 Present study 47 Hypothesis. 47 Specific aims. 47

2- Functional and Enzymatic Characterization of In Vitro Bovine Brain Microvessel Endothelial Cell Model of the In Vivo Blood-Brain Barrier. 49 Introduction 49 Methods 55 In vitro BMEC method. 55 Specific enzymatic assays...... 61 In vivo brain uptake 61 HPLC analysis. 62 Data and statistics. 63 Results 64 I. Functional...... 64 In vitro BMEC studies. 64 In vivo BBB studies. 64 H. Enzymatic aspects of BMECs 68 In vitro enzymatic activity of AP, APM, NEP and ACE. 68 In vitro BBB permeability. 68 Discussion 74

3- Defining the Mechanism of DPDPE Central Nervous System Entry. 78 Introduction 78 Methods 81 In situ brain perfiision studies. 81 Calcidation of kinetic constants. 84 Capillary depletion analysis. 86 Protein binding. 87 Octanol/saline partition coefficient. 88 HPLC analysis of ^HJDPDPE brain and perfusate extraction 88 Results 91 Multiple time uptake analysis of [^HJDPDPE compared to ['*C]sucrose 91 Self inhibition studies. 91 Contribution of vasadar component to ^HJDPDPE brain uptake 92 Octanol /saline partition coefficient. 92 Discussion 101

4- Effects of Chloro-Halogenation on CNS Entry 106 Introduction 106 Methods 110 In situ brain perfiision IIO In vitro BMEC. 110 Brain extraction after in situ vascular brain perfiision 114 Results 119 Midtiple time uptake analysis of radiolabeled biphalin, chloro-biphalin and sucrose 119 In vitro permeability coefficient, octanol /saline partition coefficient, and R litsue percent for radiolabeled biphalin, chloro-biphalin and sucrose..Ill 8

Brain extraction of radiolabeled peptide...... ,...... A22 Half time disappearance in brain...... 131 Discussion...... —.—...— 140

5- CNS Entry and Bioavailability of a Mu-Opioid Receptor Antagonist, CTAF: Comparison with . 140 Introduction...... 140 Methods 145 In situ brain perfusion 145 Octanol / saline partition coefficient. 148 Protein binding. 150 Results...... 152 Multiple time uptake analysis. 152 Self-inhibition studies. 153 R tissue ^ octanol / saline partition coefficient. 154 In vitro stability and protein binding. 154 Discussion 163

6- Biphalin Potency From a Blood-CNS Biodistribution Perspective. 169 Introduction 169 Methods 173 In vivo brain and spinal cord distribution 173 In situ brain perfitsion studies. 175 Acute modulation of transport and energy mechanisms. 175 Results 180 Time course in vivo brain and spinal cord uptake 180 Antagonist studies. 181 Self inhibition experiments. 181 Acute modulation of transport. 182 Inhibition of energy deriving mechanisms. 182 Discussion 198

7- General Discussion General Conclusions 207 Conclusions of present research 213

REFERENCES 215 LIST OF FIGURES

1.1a The structure of the cerebral capillary endothelium, a cross-sectional view 20

Lib Possible transport mechanisms for drugs across the BBB 23

1.2 The location of the blood-CSF barriers 28

1.3 The structure of the choroid capillary endothelium 29

1.4 Sites of interaction and exchange between the blood, brain extracellular fluid and CSF 30

1.5 Mechanism of action for morphine 43

2.1 The amino acid structure of the conformational constrained peptide DPDPE and methionine with the enzymatic cleavage sites responsible for methionine enkephalin degradation 54

2.2 Outline of transendothelial transport study design 59

2.3 Linearity of transendothelial passage of methionine enkephalin (with (5 MM) to protect against degradation by E.G. 3.4.24.11) 71

2.4 Reversed-phase HPLC analysis of methionine enkephalin from BMEC assay chambers 73

3.1 A schematic diagram of the perfusion circuit 83

3.2 Brain and CSF uptake of [^H]DPDPE 94

3.3 The relative uptake of [^H]DPDPE into the brain (closed squares) and CSF (open squares) measured as a function on unlabelled DPDPE concentration 96

3.4 Contribution of the vascular component to total brain uptake of [^H]DPDPE 99

3.5 HPLC / Flo-One Radioactive Detector chromatograms of [^H]DPDPE from the arterial perfusion medium, venous outflow and brain extraction 100

4.1 Multiple time uptake plots of [^^^I]biphalin, [^-^I][p-Cl-Phe'^''^']biphalin, and ['"'Cjsucrose into the brain and CSF 124 10

4.2 Contribution of the vascular component (pellet) to the brain uptake of [125i]biphaUn and p-Cl-PheM'[125i]biphaIin 126

4.3 HPLC Flo-One Radioactive Detector chromatograms of [l^^Ijbiphalin and of p- Cl-Phe^''* [^^Sijbiphaiin from the arterial inflow and the venous outflow 127

4.4 HPLC Flo-One Radioactive Detector chromatograms of [125i]biphalin and of p-Cl-Phe"*'* [^25l]biphalin TFA extracts after a 20-minute vascular perfusion...128

4.5 The percent recovery of intact biphalin and p-[Cl-Phe4,4']biphalin over a 360 minute time course in brain homogenate 130

5.1 Multiple-time uptake plots of [^HJCTAP and [l^C]inulin into brain and CSF of the in situ perfused rat 155

5.2 HPLC Flo-One Radioactive Detector Chromatograms of TFA extracts of [^H]CTAP from the brain after a 20 minute vascular perfusion 157

5.3 The percent recovery of intact CTAP in rat brain and serum over a 240 minute time course 158

5.4 The uptake of [^H]CTAP in the absence and presence of 100 (iM unlabeled CTAP 159

5.5 Rfirain percent represents the ratio of homogenate, supernatant or pellet to plasma radioactivities 161

6.1 Regional brain, spinal cord and circumventricular organ (CVO) distribution of [^125i.Tyr^]biphaIin after a 20 minute i.v. injection 184

6.2 Twenty minute regional brain, spinal cord and circumventricular organ distribution of [^25j_7yrljbiphalin after a 20 minute i.v. administration with , and CTAP pretreatment 188

6.3 The relative uptake of [^^Sj.Xyi-ljbiphalin into the brain and CSF measured as a function of unlabelled biphalin concentration 191

6.4 The contribution of the saturable and non-saturable components to the brain and CSF influx plotted against unlabelled biphalin concentrations 192 11

UST OF TABLES

1.1 Concentration of various solutes (mEq/l, except glucose, mg/lOOml) in the CSF and plasma of sheep 25

2.1 Protocol for isolation of bovine brain microvesssel endothelial cells 58

2.2 Specific inhibitors of the methionine enkephalin peptidases 60

2.3 Rank order of BMEC permeability coeflBcient, % LV. dose and capacity factor..66

2.4 Analysis of variance coupled with Newman-Keuls test for significance between permeability coeflBcient (PC) values 67

2.5 Specific enzyme activity in BMEC monolayers 70

2.6 Eflfect of specific enzyme inhibitors on the permeability coefficient (PC) of methionine enkephalin across BMEC monolayers 72

3.1 The calculated unidirectional transfer constants (Ki„), initial volumes of distribution (Vi) for [^H]DPDPE and ["C]sucrose into the brain and CSF 95

3 .2 Michaelis-Menten kinetic parameter for [^H]DPDPE influx into the CNS determined from the self-inhibition experiments 97

3.3 CNS uptake of [^H]DPDPE after in situ brain perfusion in the presence of known inhibitor of transport mechanisms 98

4.1 In vitro BBB permeability coefiBcients (PC) determined for biphalin, [p-Cl- Phe'*-'*']biphalin, [p-F-Phe''''*']biphaIin and ['''C]sucrose 123

4.2 The calculated unidirectional transfer constants (kjn), initial volumes of distribution (Vi) and cerebrovascular permeability constants (P) for ['^I]biphalin, ['"l][p-Cl-Phe'*'''']biphalin and [''*C]sucrose 125

5.1 The calculated unidirectional transfer constants (Kin) initial volumes of distribution (Vj) for [^HJCTAP and [^^CJinulin 156

5.2 Percent of [3H]CTAP bound to protein in the perfiision medium or rat serum... 160

^Tissue octanol/saline partition coeflBcient for [3H]CTAP, [^HJmorphine and [^^C]inulin 162 12

6.1 Analysis of variance coupled with the Newman-Keuls test for significance between CNS uptake values (percent injected dose / gram tissue) for 20 minute time point 187

6.2 The kinetic parameters for [^^Si.jyrljbiphalin influx into the brain and CSF determined from the vascular brain perfusion data 194

6.3 Competitive inhibition of saturable uptake of [l^SiTyj-ljbiphailn at the BBB 195

6.4 Inhibition of energy mechanisms at the BBB 196

6.5 Percent of [^2^I-Tyr^]biphalin bound to protein in the perfusion medium or rat serum 197 13

ABSTRACT

The passage of peptides across the blood-brain or blood-cerebrospinal fluid barrier is extremely limited in. Peptides can be hindered from entering the central nervous system due to the hydrophilic nature of peptides and their susceptibility to enzymatic degradation by various peptidases. Tliis limitation can be overcome through chemical modifications of opioid peptides with the goal of increasing biological stability and blood-central nervous system permeation. In the present studies, an in vitro bovine brain microvessel endothelial cell model of the blood-brain barrier was characterized both functionally and enzymatically. This primary culture model was found to be reflective of the in vivo blood- brain barrier in reference to predicting a peptides relative lipophilicity. Bovine brain microvessel endothelial cells were also found to be quite active enzymaticaily as far as the peptidases known to be involved in the degradation of methionine enkephalin. The conformationally stable analog of methionine enkephalin, DPDPE, was also characterized for its ability to enter the CNS using the in situ brain perfusion technique. DPDPE was found to enter the brain by both saturable and non-saturable uptake mechanisms. Chloro- halogenation was also found to significantly improve the central nervous system entry as well as biological stability of a potent opioid agonist, biphalin. In addition, the mu-opioid receptor selective antagonist, CTAP, was also evaluated for its ability to enter the CNS.

The amount of CTAP that crossed both the blood-brain and blood-cerebrospinal fluid barrier was quantitatively comparable to the mu-selective agonist, morphine. Biphalin was 14 found to enter both spinal and supra-spinal sites that have been shown previously to express mu and delta opioid receptors. In situ brain perfusion experiments identified a saturable component that contributes to the brain entry of [^^Sj.i-yj-ljijiphalin. Further experiments revealed that [^^Si.jyrijbiphaiin was entering the CNS by the large neutral amino acid transporter and not by the leucine enkephalin uptake system or DPDPE transport system. This research has provided important preliminaiy work for the characterization of peptide transport into the brain. The importance of using neuropharmaceutical drug delivery vectors in modem medicine needs attention for the evolution of successful drug design targeted for CNS entry 15

Chapter 1. Introduction

Anatomy and Physiology of Blood-Brain and Blood-CSF Barriers

History of the blood-brain barrier.

The notion of the blood-brain barrier (BBB) arose from early studies using dyes

injected systemically into animals (Ehrlich 1885, Goldmann, 1909;1913). The first

systematic experiments were those of Goldmann in 1909. These experiments consisted of

two main parts. The first part consisted of intravenous injection of trypan blue dye into an animal resulting in a profiise blue staining of the peripheral tissues but no staining of the brain. In addition, the cerebrospinal fluid (CSF) was colorless and the choroid plexus was

heavily stained blue. The second part consisted of injecting the dye directly into the subarachnoid space of the animal. Upon examination it was observed that the brain was

heavily stained and the systemic tissues remained colorless. These observations allowed

Goldmann to deflne a barrier between the blood and brain, and a similar barrier

between the blood and the CSF, where, there was no restriction noted between the brain and CSF.

Site of the BBB.

The dyes used for the above study presented a problem because they produce a diffuse stain microscopically, which is difficult to localize to any particular tissue site. 16

Therefore, since the discovery of the BBB by Goldmann, there have been arguments over whether the astrocytic end feet or the capillary endothelium comprise the BBB. Using electron microscopy, Reese and Kamovsky (1967) used horseradish peroxidase (MW

39,800) to visualize the staining of the BBB. Intravenous injections of horseradish peroxidase failed to reach the brain extracellular fluid. Intracerebralventricular injection into the CSF stained the brain extracellular fluid. Electron microscopy revealed that the horseradish peroxidase difllised past the astrocytic end feet and was stopped by the endothelial cells the line the capillaries. This experiment confirmed that the tight junctions between the endothelial ceils comprise the BBB. Thus, the cerebral capillary endothelium is the structural site where the blood is separated from the brain.

Anatomy and physiology of the BBB

The BBB is at the level of the capillary endothelium. There are several unique attributes of cerebral capillary endothelial cells that make them distinctive from non- cerebral capillary endothelial cells. Cerebral capillary endothelial cells contain tight junctions that join them together to form a continuous blood vessel. These tight junctions connect adjacent cells and consist of intramembranous ridges. In peripheral capillaries the junctions are not circumferential and allow for free paracellular transport of even large molecular weight compounds. In contrast, cerebral endothelial cells are joined by continuous belts of tight junctions or zonuiae occludentes, which form rows of 17 extensive, overlapping occlusions (Brightman and Reese 1969) which block out the intercellular route of solute entry.

These tight junctions are responsible for the high electrical resistance measurements observed across the capillary endothelium. Electrical resistance can be used as an indicator of cellular and paracellular ionic permeability. Transendothelial resistance has been measured across frog and rat pial vessels and has been valued at around 1870 and 1462 H.cm^ respectively (Crone and Olesen 1982; Butt et al., 1990).

These values are quite high in comparison to peripheral endothelial capillaries such as mesenteric capillaries which have a resistances of 1 to 2 Q.cm^ (Crone and Christensen

1981).

Another characteristic of cerebral capillary endotheiia is the attenuation of intracellular vesicles, endothelial fenestrations and transendothelial vesicular transport of protein (Reese and Kamovsky 1967). These vesicular transport mechanism may be decreased on the cerebral endothelial surface, yet they still may play a role in the transport of needed substances across the BBB. Three main possibilities for transendothelial vesicular transport have been postulated. These include the transcytosis

(Palade 1960), the vesicular channels hypothesis (Simionescu et al., 1975) and the fiision- fission hypothesis (Clough and Michel 1981).

With advanced cytochemical methods, several enzyme systems have been identified on cerebral capillary endothelial cells that are absent in peripheral capillaries.

Monamine oxidase and dopa decarboxylase have been identified on the vascular wall of 18

mouse cerebral capillaries (Bertler et al., 1966) which provide a protective enzymatic

barrier to L-dopa and 5-hydroxytryptophan. There are also enzyme systems present at the

BBB, such as alkaline phosphatases, that do not have a clearly defined substrate and are

involved in hydrolyzing a class of phosphate esters. Brain endothelial cells have also been

shown to be rich in y-glutamyltranspeptidase and aminopeptidase.

Another feature of cerebral capillary endothelial cells is the increased energy

requirement for active transport of nutrients to the brain from the blood. This energy is

provided by large numbers of mitochondria in cerebral capillary endothelial cells. It has

been estimated that there are 5-6 times more mitochondria per capillary cross-section in

the rat brain than in rat skeletal muscle (Oldendorf and Brown 1975). Because the

mitochondrion is the source of most cellular energy, these findings support the hypothesis

that the brain capillary cell is performing substantially more metabolic work than muscle capillary cells. This suggests that this apparent extra brain capillary work capacity is, in some part, related to energy-dependent transcapillary transport.

One group of researchers performed comparative morphometric analysis of cerebral and non-cerebral capillaries (Coomber and Stewart, 1985). They revealed a 39% decrease in the wall thickness of brain capillaries compared to muscle vessels. In addition, the number of pinocytotic vesicles that are thought to be associated with vascular

permeability were seven times higher in muscle vessels compared to brain vessels. It is postulated that the 39% decrease in the wall thickness of brain capillaries could be an adaptation to the restricted permeability of the BBB, allowing nutrients delivered by 19 carrier-mediated transport processes into the endothelial cell a shorter difiiision time to cross the cytoplasm and enter the brain extracellular fluid. (Figure I. la) 20

Morphology of the Blood-Brain Barrier

basement membrane astrocytic end feet

alkaline phospliatase

P-glycoprotem

gamma-glutamyl transpeptidase transrerrm receptor

mitochondria

aminopcptidase

GIut-1

Lumen

ZO-l Protein

Brain

tight junction

Fig 1.1a The structure of the cerebral capillary endothelium, a cross-sectional view. 21

Type of transport across the BBB.

Compounds use a number of different mechanisms at the BBB for their transport into the brain (ZIokovic 1995; Banks and Kastin 1996; Pardridge 1995). These consist of both saturable and non-saturable mechanisms. Diffusion can contribute to the transport of molecules across the BBB. DifTusion can be divided into paracellular difTusion which is between endothelial cells and transcellular difTusion which is across the endothelial cell. Both of these diffusive mechanisms are non-competitive and non­ saturable. Endocytosis can be divided into both fluid phase, specific and non-specific.

Fluid phase endocytosis is non-competitive, non-saturable, temperature and energy dependent. Specific endocytosis is competitive, saturable, temperature and energy dependent. Non-specific endocytosis is saturable, energy and electrostatic charge dependent. Carrier-mediated transport also contributes to the transport of molecules such as hexoses and amino acids. Carrier-mediated transport can be both energy and non- energy dependent. Receptor-mediated transport is another mechanism by which compounds such as transferrin and insulin use to gain access to the CNS. Another type of transport mechanism at the BBB are efflux pumps. These transporters are involved in either extruding drugs from the endothelial cells (i.e., P-glycoprotein) or transporting compounds fi'om the brain back into the blood (i.e., PTSl-5). These transporters can be energy dependent and saturable. (Figure 1.1b) 22

Periendothelial structures

The periendothelial accessory structures include the astrocytic endfeet, phagocytic pericytes and the endothelial basement membrane that surround the capillary (Figure I.la) Cerebral capillaries are unique because they are almost completely surrounded by end feet of the brain glial cells known as astrocytes. The capillary endothelium is separated from the astrocytic end-feet by the periendothelial basement membrane. Since there are 20 nm gaps between adjacent astrocytes which horseradish peroxidase can readily difiuse (Brightman and Reese 1969), they may not contribute greatly to the physical barrier, yet the close contact between astrocytes and endothelial cells does suggest a functional interaction. It is proposed that astrocytes may be involved in secreting soluble factors that induce unique morphological characteristics upon cerebral capillary endothelial cells (Janzer and Raff 1987).

A discontinuous layer of intramural pericytes is present in the basement membrane that surrounds the cerebral capillary wall. The exact function of pericytes is still unclear, although the presence of lysosomal bodies in pericytes suggest that they play a phagocytic role (Coomber and Stewart 1985). Pericytes may prevent macromolecules, which have passed through the endothelial cell barrier, from entering into the brain. 23

Drug

Lumen

Cell

I T ti Brain 12 34 5 6 7

Fig 1.1b Possible transport mechanisms for daigs across the BBB. (1= paraceliular diffusion, 2 = transcellular diffusion, 3 = fluid phase endocytosis, 4 = non-specific endocytosis, 5 = specific endocytosis, 6 = carrier mediated transport and 7 = efflux pump). 24

Cerebrospinalfluid.

The cerebro-spinal fluid (CSF) is located within the ventricles, spinal canal and the subarachnoid spaces. The total volume of CSF in humans is given by Weston (1916) as 140 ml. It has been estimated that the total volume of CSF in mammals varies between

10 and 20% of brain weight (Bradbury, 1979). The main sources of CSF are the choroid plexuses of the lateral, third and fourth ventricles. The CSF moves within the ventricles and subarachnoid spaces under the influence of hydrostatic pressure generated by its production. Its drainage is mostly through the arachnoid villi (Figure 1.3). CSF also supports the brain by reducing its weight 30-fold and acting as a buoy (Spector and

Johanson, 1989).

The composition of various substances in the CSF is significantly different from those in the plasma (Table 1.1). The brain concentration of most molecules is greater than those in CSF forming a physiological gradient between the two compartments.

The CSF is therefore more likely to act as a "sink" to the brain than the brain act as a

"sink" to the CSF (Davson et al., 1961). 25

Table 1.1. Concentration of various solutes (mEq/l, except glucose, mg/lOOmI) in the CSF and plasma of sheep, adapted from PoUay et al., 1972. RCSF is the ratio of CSF to plasma concentration.

Substance Plasma CSF RCSF Na 153.6 154.4 1.01 K 4.1 2.7 0.66 Mg 1.9 2.0 1.03 Ca 4.5 2.9 0.64 CI 116.4 132.5 1.14 Glucose 99.1 59.5 0.60 Osmolarity 299.3 304.0 1.02 dH 7.4 7.3 — Total protein 5.7 0.04 O.Ol 26

Anatomy and physiology of the blood-CSF harrier.

Goldmann in 1913 first demonstrated the existence of a barrier between the blood and the CSF. He injected trypan blue dye systemically into animals and observed that the CSF remained clear like the brain tissue. It was found that by changing the size and properties of the dyes used it was found that some could enter the CSF proposing that the barrier between the blood and CSF was selective rather than absolute (Bradburj'

1979).

The choroid plexus and the arachnoid membrane act together as barriers between the blood and CSF. The CSF bathes the exterior of the central nervous system and fills the four ventricles inside the brain. The arachnoid membrane is generally impermeable to water-soluble substances, and its role in forming a blood-CSF barrier is largely passive. The choroid plexus, however, actively regulates the concentration of molecules in the CSF and makes the blood-CSF barrier a selective one. [n human beings and other mammals, the choroid plexus consists of several small, reddish tufls or patches of tissue. The majority of the choroid plexus is distributed throughout the fourth ventricle near the base of the brain and the lateral ventricles inside the right and left cerebral hemispheres. Approximately one tenth of the choroid plexus is in the centrally located third ventricle. (Figure 1.2) In most adult mammals, the choroid plexus weight is approximately 0.25 % as much as the entire brain (Spector and Johanson 1989). The capillaries of the choroid plexus are the fenestrated, non-continuous type, with gaps between the capillary endothelial cells (Machen et al., 1972). This allows free movement 27 of small molecules such as amino acids, glucose and electrolytes out of the vascular space into the extracellular fluid of the choroidal epithelium, with the basement laminal

preventing the loss of plasma protein from blood into the choroid plexus extracellular space. Most macromolecules are eSectively prevented from passing from passing into

the CSF from the blood by tight junctions between the adjacent choroidal epithelial cells near the CSF side (Brightman, 1968) (Figure 1.3). However, these epithelial like cells have a low resistance, approximately 200 n.cm^ between blood and CSF (Saito and

Wright, 1983). The main function of the choroid plexus is the secretion of CSF by which it draws the necessary nutrients, ions and other ingredients from the blood plasma.

The double layered arachnoid membrane is another site for the blood-CSF barrier.

The brain is covered by the meninges which are composed of the dura, the arachnoid and the pia. The meninges form a potential interface with the external surface of the brain, however dura contain fenestrated capillaries, so that large molecules can pass from the blood into the dura. Further passage into the brain is prevented by the layer of the arachnoid, which contain cells with tight junctions thereby forming a barrier between blood and CSF in the subarachnoid space (Nabeshima et al., 1975). (Figure 1.4) 2X

«I»AA*CMRIR>IFI - '4JP»A«3A'-*ARTAI ^ / • CUR*UTA*IC'RTA^ACIIRTOTO AKATITRAIIUUtMC*-*'*

K1 ;U 'NIIFC VR*I:»ICITOf RWTW' N R*TIS OF L*(T9AL VTFLT»:TLR rMO»otoMUAIM VFFITFM.IRrtr»tjs rr

SR»FIAI C'.?!TN

i

Fig 1.2 The location of the blood-CSF barriers. The choroid plexus continuously secretes CSF, which cushions the CNS, carries some nutrients to the brain and cleanses the brain of waste. The choriod plexus sites are within the brain ventricles. Arrows show direction of flow of CSF in the ventricles, from Spector and Johanson (1989). 29

[Viorphoiogy of the Blood -CSF Barrier

FcncslraLcd lypc capillary • c Capillary lumen • c ECF Choroidal Gpithclial cclls 'iglu junction CSF

Fig 1.3 The structure of tlie choroid capillary endothelium. Capillary consists of fenestrated, non-continuous tight junctions. The choroidal epithelial cells are actual barrier, expressing complex tight junctions that are impermeable to many molecules. "*0

3unA

I=IPIALVESSSL AnACiinoio PI A

CONTEX

istnacYic EMortET

CEnEDRAL CAPILLAniES FPF'IDYMA \ / \/

LEIKY \ EFETLOTUA ; rov

WT vEMintClE cHonoio LVilUS 'Q

CI'CNOIOAI./ VESSELS

Fig 1.4 Sites of interaction and cxchangc between the blood, brain extracelhilar fluid and CSF. (I) Capillar)' cndotliclium of the brain (TBV) sealed by tight junctions (TJ) with a low permeability. (2) Arachnoid membrane with tight junctions in the outer leaflet obslnicling free diffusion form the dutnl extracellular fluid with its leaky capillaries (5). (3) Choroid plexus wiih tigiit junctions between the epithelial cells, forming the blood-CSF barrier. (6) Tlie capillaries of Ihc choroid plexus arc of llic fenestrated t}!^: with open patliways between Uie endothelial cells (FBV). (4) Tlic circumventricular organs with lealcj' capillaries and tight junctions between the cells of the cpcndyma forming a barrier between the extracellular fluid of liiis organ and liic CSF. 31

Techniques to Study BIood-CNS Permeability and Stability of Drugs

The BBB is a complex structure that consists of both active and passive elements, thus any technique used to measure its activity has its limitations. Several models are employed to study the permeability characteristics of this structure and can be divided broadly into two main categories depending on whether the method uses the whole animal (in vivo), or where isolated cerebral capillaries are harvested from brain, and endothelial cells are grown in tissue culture to form a pure barrier (in vitro).

Brain uptake index.

A single injection technique was developed (Oldendorf 1970) that allows regional uptake measurement by an animal brain of a ^H-labeled or "C-labeled test drug. This technique involves a 0.2 ml bolus of a buffered Ringer's solution containing either a "H- labeled or "C-labeled test compound and either a ^H-labeled or "C-labeled reference compound injected into the common carotid artery of an anesthetized animal. The reference compound is a rapidly penetrating solute such as diazepam, "H water or butanol. The bolus passes through the brain within 2 s after the single injection and the animal can be decapitated 5-15 s after injection. It has been shown that there is minimal efflux of the test or reference compound after this short time period (Pardridge and Fierer

1985). The ratio of to "C in brain tissue relative to the ratio of to in the original mixture determines the amount of test substance lost to brain tissue on a single passage through brain microcirculation. 32

The advantages of the brain uptake index are that the experimental time is quite short and many different experiments can be performed in a short time period. An obvious disadvantage to this technique is that the brain extraction is measured over a short time period allowing for more error in quantification. Without proper measurement of cerebral blood flow and the extraction of the reference compound a researcher may end up measuring the uptake of experimental artifacts. Another drawback to this technique is that it can only measure uptake of drugs which have rapid transport mechanisms.

Intravenoits bolus injection.

In this technique (Ohno et al., 1978), a femoral vein is cannulated and the radiolabeled test compound is injected. At various times after injection, groups of animals are sacrificed and arterial blood is collected. Alternately, a femoral artery may be cannulated and blood samples may be obtained at various times. The plasma and brain radioactivity are determined and a brain volume of distribution for the test compound can be computed from the ratio of dpm/g brain divided by dpm/|.il plasma at the terminal time period. A BBB permeability surface area product can then be calculated.

The intravenous injection technique has two main advantages. One is that it docs not require access of a carotid artery and the other is the increased sensitivity of the technique. The main disadvantage of the single intravenous injection technique is that there may be extensive metabolism of the test compound by peripheral tissues and this 33 may confuse interpretation of the brain uptake data and prevent accurate computation of the BBB permeability surface area product. Thus, computation of the BBB permeability surface area product using the intravenous injection technique can be dangerous when there is a measurable metabolic degradation of the test compound during the experimental time period. This can oftentimes be corrected for by using HPLC extraction of test compounds, although it is not always possible to determine if the radiolabeled metabolite is produced in the brain or the periphery. In vitro metabolism studies can sometimes clarify this issue.

Brain perfusion studies.

The intravenous injection technique and the brain uptake index have been adapted to allow a researcher to measure longer experimental time periods. Carotid artery infusion techniques have been described for both the rat (Takasota et al., 1984) and guinea pig

(Zlokovic et al., 1986). The internal carotid artery infusion technique involves cannulating an anesthetized animal at both carotid arteries. The jugular veins are then sectioned and a carotid perfusion is initiated at a constant flow rate equal to the animals cerebral blood flow. The perfusate consists of a physiologically balanced mammalian Ringer solution. At the end of the experiment a sample of the perfusate is taken and the animal is decapitated.

This technique can be extended to utilize CSF sampling to assess the ability of a compound to cross the blood-CSF barrier. Following decapitation, the amount of 34 radiolabeled test drug entering the brain is determined and can be expressed as a ratio of drug entering the brain or CSF to that in the perfiisate.

Brain perfusioa studies are more sensitive than brain uptake index techniques since the experimental time can be extended out to 30 minutes. This allows researchers to assess the BBB permeability of slowly penetrating compounds that otherwise would not be able to be quantified. This technique also has the advantage that there is no systemic exposure of the test compound prior to transport through the BBB and thus, metabolism artifacts caused with this technique are restricted to metabolic events that occur within the brain microcirculation. Also, this technique can be coupled to HPLC analysis to assess metabolism.

Positron emission tomography

It is possible to measure changes in BBB permeability in humans using quantitative positron emission tomography (PET). PET allows the in vivo quantification of substances labeled with positron-emitting isotopes in a nomnvasive manner. BBB permeability has been measured successfully in vivo by combining PET technology with theoretical and physiological information obtained by animal studies. A PET scanner can accurately measure the activity in a region of interest but it cannot differentiate the compartments

(blood or brain tissue), in which the activity is distributed. This blood compartment can be subtracted out by independently measuring the cerebral blood volume to estimate the intravascular compartment. [68Ga]-EDTA and [82Rb] are the two most frequently used 35 tracers in PET evaluation of the BBB (lonnotti 1992). Using PET evaluation of BBB permeability will provide data on the pathophysiology of the barrier in vivo and may open new perspectives on disease states such as ischemia, multiple sclerosis and

Alzheimer's disease.

Octanol / saline partition coefficient.

Oftentimes, a compounds ability to difiuse across a membrane can be easily predicted by its ability to partition between two phases. One can predict lipophilicity by measuring an octanol / saline partition coefficient, which is the ratio of the amount of test compound in the octanol phase to that in the saline phase. This in vitro technique is useful to predict the ability of compounds to diffuse across a membrane, but is in no way reflective of the ability to utilize a specific carrier to cross a membrane. Oftentimes partition coefficients are used as an initial screening to see the effects of simple structural modifications on peptide lipophilicity.

HPLC capacity factor

Capacity factors calculated from HPLC retention times are often used to assess lipophilicity of various organic compounds. Capacity factors {k) can be determined on a pre-established HPLC gradient by the following equation:

Capacity factor = k = (tr-tej/to 36

where tr is the retention time of the retained peak and to is the retention time of the

unretained peak. This oftentimes is a good predictor of lipophilicity.

Isolated cultures of cerebral capillaries.

It is usually accepted that most of the in vivo methods for measuring BBB

permeability necessitate a radiolabeled test drug for analysis. The reason for this is that

the required methodologies for serum and brain extractions of the drug followed by sensitive detection methods have not been developed. In many situations it is desirable to measure the BBB permeability of a series of compounds which are not available in

radiolabeled form. In vitro test systems allow a researcher to do this. In vitro BBB model systems have been developed wherein primary cultures of brain capillary endothelial cells are grown on filters that are pre-coated with collagen and fibronectin and transport of test compounds across the brain endothelial monolayer is measured (Audus and Borchardt

1986). These in vitro systems are useful for screening large numbers of drug analogs without requiring use of radioactivity. One drawback of this system is that there is extensive de-differentiation of the brain capillary endothelial monolayer (Pardridge et al., 1990). It has been shown that several nutrient transport systems are downregulated as much as 100-fold. Thus, the in vitro BBB model systems can potentially underestimate

BBB permeability for compounds that utilized a carrier-mediated mechanism. Also, for compounds that utilize lipid mediated transport across the BBB, in vitro BBB models can 37

result in a considerable over-estimation of the in vivo BBB permeability. This results from

the increase in the diSiision due to the decrease in tight junction formation apparent with

cell culturing. The endothelial tight junctions do exist in the in vitro models, but they are

greatly reduced (Brightman et al., 1993). For example, the trans-endothelial electrical

resistance ranges from 40-100 Q.cm^ (Audus and Borchardt 1986) which is much lower

than the in vivo value of 1462 Q.cm~ (Butt et al., 1990). Oftentimes the tight junction

structure and trans-endothelial electrical resistance can be improved by co-culture with

astrocytes or adding cAMP analogues to the culture media. Even with the above

limitations, primary cultured endothelial cells are still useful for initially screening of

drug analogs so lead compounds can be identified for fijrther characterization by more

complex in vivo methods.

Endogenous Opioid Peptides and Opioid Pharmacology

History of

Opium is the Greek word for "juice" which is obtained from the poppy, . The first reference to this poppy juice is found in the writings of

Theophrastus in the third century B.C.. The drug was introduced into the Orient by

Arabian traders where it was utilized for the control of dysenteries. The use of

became popular in Europe around 1493-1541. In 1806 Sertumer reported the isolation of a pure substance in opium that he called morphine, after Morpheus, the Greek god of dreams. In the United States the use of opium was increased with the invention of the 38

hypodermic needle, which allowed for parentarel use of morphine resulting in more

addictive behavior. The societal problem of drug addiction in the United States has

resulted in the search for potent that exert less of the unwanted side effects of

addiction, tolerance, constipation and respiratory depression.

Endogenous opioid peptides

The discovery that the brain contains endogenous opioid peptides has greatly

increased our understanding of the role of the opioid systems in the modulation of pain.

There are three classes of endogenous opioid peptides. The first class are the which were discovered by Hughes et ai., (1975). They isolated LWO small

peptides from pig brain that were shown to have potent opioid agonist activity by being able to inhibit electrically induced contractions of the mouse vas deferens which could be antagonized by the nonselective opioid receptor antagonist naloxone. A is the pro-hormone that gives rise to a large number of biologically active peptides. The proenkephalin A protein sequence contains six copies of Met-enkephalin and one copy of

Leu-enkepahlin (Davis and Konings, 1993). These are the sequences that contain the activity. The second class belongs to the (POMC) family

(Watson et al., 1977). The POMC precursor is expressed in the pituitary (Drouin and

Goodman, 1980) and hypothalamus (Gee et al., 1983), and its peptide products are released into the blood stream in response to stress. POMC give rise to 3-endorphin, melanocyte-stimulating hormone, adrenocorticotropin hormone and P-Iipotropin . The 39 third class belongs to the family (Goldstein et al., 1979). gi%'es

rise to dynorphin, a tridecapeptide, which contains the Leu-enkephalin sequence.

Dynorphin is a potent K- receptor agonist. The name dynorphin comes the word dynamis which is Greek for power. Additional peptides with opioid activity have been discovered, but most contain the sequence Tyr-Gly-Gly-Phe-Met or Leu..

Multiple opioid receptor types

The complex interaction of morphine and drugs with mixed agom'st / antagonist properties led Martin to propose the existence of multiple types of opioid receptors

(Martin and Sloan, 1977). Presently, three major types of opioid receptors (|I, 5 and K

) have been characterized based on extensive in vitro and in vivo pharmacological and biochemical studies (Gilbert and Martin, 1976; Lord et al., 1977). Most of the clinically used opioids are selective for the p. receptors (morphine), but when given at sufficiently higher doses may these drugs may not be as selective thus changing the pharmacological profile of the drug. This can oftentimes be the case when tolerance develops to a specific drug. Both respiratory depression and constipation due to inhibition of gastrointestinal transit are responses thought to be mediated by |j, receptors. DAMGO is also a [i- selective agonist. Selective antagonist for the opioid receptor have been developed,

CTAP and CTOP(GuIya et al., 1986), which are analogs. The enkephalins are the endogenous ligands for 5 receptors. Selective agonists and antagonists have been developed for this receptor. DPDPE (Mosberg et al., 1983; Toth et al., 1990) is a 5- 40 selective agonist that contains a conformationally constrained enkephalin backbone yielding greater enzymatic stability (Weber et al., 1991, 1992). Naltrindole is the 5- selective antagonist (Porteghese et al., 1988). The role of 5 mediated analgesia as far as development of tolerance is still being debated (Mathes et al., 1996 and Kest et al., 1997), but it is believed that antinociception mediated through the 5-receptor may potentially produce less of the unwanted side effects as seen with |i receptor analgesia, which provides the clinical utility for the development of 5-receptor selective agonist.

Dynorphin A is the endogenous ligand that is selective for K receptors. Kappa opioid receptors are selectively labeled by the agonist U50,488H and antagonized by nor-BNI.

Kappa-selective agonists increase psychotomimesis (hallucinations and dysphoria) and diurresis (Goodman and Gillman, 1996) which are both unwanted side effects.

Recently, Zadina et al. (1997), have identified an endogenous agonist that displays preference for the |i-opioid receptor. They recently reported the discovery and isolation from the brain, endomophin-l(Tyr-Pro-Trp-Phe-NH2), which has a high affinity

(Kp360pM) and selectivity (4,000-15,000-fold preference over the ii and k receptors) for the receptors. This peptide is more effective than the |.i-selective analog, DAMGO, in vitro and it produces potent and prolonged analgesia in mice. Also, a second peptide, -2 (Tyr-Pro-Phe-Phe-NH2) was also isolated. These new peptides have the highest specificity and affinity for the |i-receptor compared to any endogenous substance.

Thus it is hypothesized that they may be natural ligands for the |a-opioid receptor. 41

With advances in molecular biology the structures of each opioid receptor type has been determined. Opioid receptors contain a 7 transmembrane domain structure and agonist binding stimulates GTPase activity (Koski and iGee, 1981), that is regulated by guanine nucleotide (Blume, 1978) and inhibits adenylyl cyclase. The three opioid receptors have all been cloned for the human: receptor (Wang et al., 1994), 5 receptor

(Knapp et al., 1994) and K receptor (Mansson et al., 1994). The homology between three opioid receptor types is approximately 65% and they have little sequence similarity to other G protein-coupled receptors, except receptors for somatostatin (Reisine and Bell, 1993). The regions of highest similarity lie in the seven transmembrane- spanning regions and the intracellular loops. The regions of amino acid sequence divergence are the amino and carboxy termini and the second and third extracellular loops. The human opioid receptor genes have multiple introns, subtypes of |.i, 5 and K may result from alternate mRNA splicing.

Regional CNS distribution of opioid receptors.

With the advent of selective receptor ligands it has become possible to study the regional distribution of opioid receptors by autoradiographic methods. This can then be compared to the regional mRNA expression for each receptor type. High density of delta and mu opioid receptor mRNA and binding sites shown to be expressed in the nucleus accumbens, frontal cortex and caudate putamen (Mansour et al., 1995).

Kappa opioid receptors have been mostly localized to spinal sites. 42

Mechanism of opioid inhibitory action on neurotransmission.

The |i, 5 and K opioid receptors in vivo are coupled to inhibition of adenylyl cyclase activity via pertussis toxin-sensitive GTP binding proteins. The mechanism of action of the classical opioid agonist, morphine involves morphine binding to the |j. opioid receptor. This results in a modulation of a G-protein (G;) and a decrease in the activity of adenylate cyclase, resulting in a decrease in the production of cyclic adenosine monophosphate (cAMP). In addition, la-opioid receptor activation results in

efllux and cellular hyperpolarization. Both result in a decrease in cAMP and the enhanced efflux leads to decreased Ca^* entry and lower free intracellular levels of

Ca^^. This results in opioid blockade of neurotransmitter release and pain transmission (Figure 1.5). The effect of morphine on pain perception is different from the effect of a local anesthetic, which interferes with pain perception by decreasing axonal conduction. Opioids are limited in their ability to decrease neuronal conduction. In addition, opioid drugs do not have antiinflammatory activity, unlike aspirin and the nonnarcotic analgesics. Actually, opioid drugs can cause a mild inflammatory reaction due to the ability of opioids to cause release of histamine. 43

morphine rcccptor

Caiciurn Calcium \ciilry blocked (Cali l-

—Decreased release of neurotransmitter

Fig 1.5 Mechanism of action of morphine. When the agonist morphine binds the G- protein is modulated causing a decrease in the activity of adenylate cyclase, resulting in decreased production of cAMP, enhanced K"^ efflux and decreased Ca^*" entry. 44

Peptide Drug Delivery to the Brain

Peptides as neuroplvxrmaceuticals

The multiple biological actions and extreme potency of peptides in the brain

suggest that these agents have the potential to be utilized as neuropharmaceuticals in the

treatment of a variety of disorders of the brain as well as drugs for the alleviation

of pain. However, as with any potential neuropharmaceutical, peptides must be able to

undergo permeation into the brain from the blood. investigations in the

mid and late 1970s underscored the importance of a thorough understanding of the

mechanisms by which peptides are transported between blood and brain. After the

discovery of peptides such as TRH and endorphin (Li and Chung 1976), these agents were

infused intravenously into humans for the treatment of psychiatric disorders such as

schizophrenia or depression (Barchas et al., 1986). No significant effects on the brain

were found with these agents in humans, although in rats P-endorphin had profound

effects on the CNS following injection of the peptide directly into the ventricular compartment (Bloom et al., 1978). No central effects were observed when P-endorphin was administered systemically to rats or humans. The failure of any central actions of P- endorphin following peripheral administration is due to the presence of a blood-brain barrier described in Chapter 1 of this dissertation. However, as shown in Chapters 2-6 of this dissertation, it is possible for certain circulating peptides to traverse the BBB if 45 these peptides are biologically stable and can utilize specific carrier-mediated transport systems localized within the brain capillary endothelial cell wall.

The immense obstacle of designing a peptide neuropharmaceutical to actually cross the BBB and enter the CNS to exert a pharmacological effect can be best explained by the historical example of dopamine drug delivery. In the 1960s, the neutral amino acid L- dihydroxyphenylalanine (L-dopa) was developed for the treatment of Parkinson's disease, a neurodegenerative condition where the caudate putamen region of the brain forms inadequate amounts of dopamine (Lloyd et al., 1975). Circulating L-dopa is transported into the brain via the large neutral amino acid carrier which is expressed at the

BBB. After this transport, L-dopa is converted to dopamine by brain aromatic amino acid decarboxylase. Unfortunately, peripheral tissues also contain aromatic amino acid decarboxylase, and this results in the peripheral conversion of L-dopa to dopamine. This peripheral metabolism can be stopped with the use of adjunct dmgs that inhibit the enzyme but do not cross the blood brain barrier. A parallel to this idea will be described in

Chapter 2 of this dissertation in reference to the ability of methionine enkephalin to traverse the BBB in the presence of specific inhibitors of peptidases involved in the degradation of methionine enkephalin.

The use of L-dopa in Parkinson's disease illustrates the importance of two principles applicable to the use of peptides as neuropharmaceuticals. The first is that

CNS delivery of requires a drug delivery system. In the case of L- dopa, this drug delivery system is the utilization of the large-neutral amino acid carrier 46 which is expressed at the BBB. This concept will be described in Chapter 6 of this dissertation in reference to the CNS delivery of a potent opioid analgesic. The second principle illustrated by the L-dopa paradigm is the need to slow the rapid enzymatic inactivation of the neuropharmaceuticals so they can reach the brain at therapeutic doses. This concept will be addressed in this dissertation mainly through the structural modification of endogenous enkephalin backbones by skilled peptide chemists of Dr.

Victor Hruby's laboratory. These structural modifications will make the peptides studied in this dissertation more biologically stable and resistant to peptidase degradation by a variety of tissue compartments. 47

Present Study

Hypothesis: Structural modifications increase biological stability and blood-CNS penetration so that opioid, peptide analogs can enter the brain to bind to opioid receptors.

Specific Aims;

1. Characterize the /w vitro bovine brain microvessel endothelial cell model of the BBB both functionally and enzymatically.

2. Determine the contribution of saturable and diflusional components to the CNS entry ofDPDPE.

3. Examine the effects of chloro-halogenation on the CNS entry of a double enkephalin analog, biphalin.

4. Explain biphalin analgesic potency from a blood-CNS biodistribution perspective.

As mentioned in this dissertation, the passage of peptides across the BBB is extremely limited in comparison to other compounds such as diazepam, butanol, or .

This is due to the hydrophilic nature of most peptides and their susceptibility to enzymatic degradation by various peptidases. To overcome this problem, structural modifications were incorporated into endogenous peptides with the overall goal to increase both biological stability and blood-CNS penetration. All peptide analogs used in this study were designed by the chemists of Dr. Victor Hruby's laboratory. The studies of Chapter 2 48 will attempt to characterize an in vitro model of the BBB both enzymatically and functionally so that it can be used to identify successful structural modifications incorporated into methionine enkephalin. The enzymes measured are ones that have been shown in the literature to be involved in the degradation of enkephalin and have not all been previously shown to be expressed in cerebral capillary endothelial cells.

In Chapter 3, the mechanism of DPDPE CNS entry wUl be investigated. DPDPE is an enzymatically stable analog of methionine enkephalin that was shown to cross the

BBB better than methionine enkephalin (Chapter 2). Diffusional as well as saturable mechanisms to cross the blood-brain and blood-CSF barriers will be investigated using the in siUi brain perfusion technique.

In Chapter 4, the effects of halogenation of the residue on BBB permeability will be investigated. Biphalin and [p-Cl-Phe''''*']biphalin will be compared as far as their individual blood-brain and blood-CSF permeability.

In Chapter 5, a mu-selective antagonist, CTAP will be investigated for its ability to cross the BBB from a systemic route of administration. CTAP will be compared to

["Clmorphine as far as its ability to enter the brain and bind to the vascular component of the CNS. This mu-selective antagonist will be used to characterize the CNS entry of biphalin in Chapter 6.

In Chapter 6, the analgesic potency of biphalin will be explained by its regional

CNS biodistribution. In addition, a mechanism of CNS entry will be explained for this potent opioid receptor agonist, so that it may be utilized as a drug delivery system. 49

Chapter 2. Functional and Enzymatic Characterization of In Vitro Bovine Brain

Microvessel Endothelial Cell Model of the In Vivo Blood-Brain Barrier.

Introduction

Various in vivo methods have been developed to help elucidate mechanisms of transport across the BBB for various compounds. The advantage of these in vivo methods is that the physiological conditions are close to the norm, but the results can be skewed due to limitations of the protocol (i.e., metabolism or changes in cerebral blood flow). Another common disadvantage of studying BBB permeability by in vivo methods is that the limits of detection in most laboratories make a radiolabeled compound necessary.

Laboratories are involved in the rapid synthesis of numerous drug analogues and radiolabeling each drug analog is not feasible. In addition, data obtained from the use of a radiolabeled compound must be carefully interpreted because of the possibility of enzymatic degradation and / or radiolabel exchange (i.e., "H / water exchange). For the above reasons, an in vitro model of the blood-brain barrier that has the potential to screen a large number of drug analogs (without requiring the use of radiolabeled compound) would provide great use for identifying good lead compounds with enhanced BBB permeability.

An in vitro model of the BBB that uses bovine brain microvessel endothelial cells

(BMEC) grown in primary culture has been used extensively to study the BBB permeability of drug analogs (Audus and Borchardt, 1986; Audus and Borchardt, 1987, 50

Bowman et al., 1983). This in vitro model has been extensively characterized morphologically, biochemically and immunohistochemically and found to have tight junctions, attenuated pinocytosis and no fenestra (Borchardt, 1990). In addition, specific

EBB enzyme markers (i.e., gamma-glutamyl transpeptidase and alkaline phosphatase), endothelial cell markers (i.e.. Factor VHI antigen), catecholamine degrading enzymes (i.e., monamine oxidase A and B, cytosolic catechol 0-methyltransferase and phenol sulfotransferase) (Audus and Borchardt 1986, Baranczyk-Kuzma et al, 1986; 1989, Scriba and Borchardt 1989). Primary cultures of BMEC are considerably "leakier" than the in vivo BBB, as measured with membrane-impermeant markers (i.e., sucrose) and electrical resistance (Pardridge et al., 1990).

Several studies have examined the ability of peptides to cross the BBB (see reviews. Banks and Kastin 1996, Zlokovic 1995, Begley 1994). Although numerous peptides have been shown to cross the BBB, the passage of peptides across the BBB is extremely limited in comparison with other compounds such as diazepam, butanol, or heroin. The inability of peptides to readily cross such membranes barriers as the small intestine or the BBB has hindered their development as clinically useful drugs. Numerous structural modifications have been incorporated into peptides to increase lipophilicity or biological stability. Banks and Kastin (1985) have previously shown a correlation between the lipophilicity of peptides from several classes and penetration across the BBB. In addition, Pardridge et al., (1990) have shown that lipid mediated, but not carrier-mediated drug passage across BMEC correlated well with in vivo transport across the BBB. This 51 work will, in part, characterize the functional utility of the BMEC model of the BBB for predicting the passage of several analogs of methionine enkephalin that are designed to increase BBB penetration as well as enzymatic stability. Investigations will also be made to understand what role lipophilicity may play in passage across the BBB.

Peptides not only have difficulty crossing the BBB because they are highly polar but they are also susceptible to enzjmiatic degradation. The BBB not only forms a physical barrier that prevents solute entry from the blood to the CNS but also can form an enzymatic barrier (Rapoport 1976). It is highly possible that degradative enzymes expressed at the BBB play a regulatory role in the passage of small opioid peptides across the BBB. Thus, development of clinically usefiil opioid peptide drugs requires a thorough understanding of the enzymatic composition of the BBB. By understanding the enzymatic environment of the BBB, it would then be possible to develop an enzymatically stable peptide with enhanced passage across the BBB (Weber et al., 1991; 1992). Another possibility would be to use specific enzymatic inhibitors of peptidases to enhance the passage of peptides across the enyzmatic BBB. This study will also investigate the in vitro BMEC model enzymatically to evaluate both the presence of enkephalin inactivating peptidases at the BBB and the modulatory role these peptidases play in the permeability of methionine enkephalin across the BBB.

Methionine enkephalin (Tyr'-GIy^-Gly^-Phe''-Met^) has two primary sites of enzymatic hydrolysis; the Tyr^-Gly^and Gly^-Phe"' bonds (Figure 2.1). The two peptidases responsible for in vivo inactivation of methionine enkephalin are aminopeptidase and neutral endopeptidase (NEP, E.C.3.4.24.11). converting enzyme (ACE;

E.C.3.4.15.1) also has been implicated in the cleavage of methionine enkephalin at the

Gly^-Phe'* bond (Malfroy et al., 1978; Erdos et al., 1978; Shibanoki et al., 1991). These

enzymes have been shown to be expressed in a variety of tissues that a peptide drug would

be exposed to after a given route of delivery. Aminopeptidase activity has been reported

in pancreas (Terashima et al., 1992), stomach wall (Bunnet et al., 1990), intestines (Kuno

and Oka, 1987), plasma (Shibanold, 1991), nasal mucosa (Hussain, 1990), cerebrospinal

fluid (Renter et al., 1990), CNS (Dyer et al., 1990; Gibson et al., 1991; Hersh, 1981),

isolated brain microvessel (Partridge and Mietus 1991) and cultured brain microvessel

endothelial cells (Baranczyk-Kuzma and Audus, 1987). Specific enzyme inhibitor studies suggest that NEP and ACE activity were not present in CSF, (Benter et al., 1990) although, ACE activity has been measured in rat and human CSF (Schweisflirth and

Schioberg-Schiegnitz, 1984; Yoshida and Nosaka, 1990). ACE activity has also been

measured in kidney (Matsas et al., 1984), intestine (Bai, 1993), plasma (Shibanoki, 1991) and brain microvessel endothelial cells (Baranczyk-Kuzma et al., 1986). NEP activity was detected in peripheral vascular endothelial cells (Llorens-Cortes et al., 1992), pancreas

(Terashima et al., 1992), kidney (Matsas et al., 1984), intestine (Bai 1993) and CNS

(Matsas et al., 1983). In order for a peptide analgesic to be clinically efficacious it must be resistant to the above tissue or enzymatic barriers. These barriers can drastically alter the bioavailability of a peptide drug thus decreasing the BBB permeability. 53

The focus of this study is three fold. The first aim is to characterize the BMEC model of the BBB functionally for its ability to predict peptide passage across the BBB.

This will be accomplished by making in vivo / in vitro correlations. The second aim is to characterize the presence of specific methionine enkephalin inactivating peptidases in bovine brain, capillary endothelial cells. To achieve this aim total aminopeptidase, aminopeptidase M (APM; E.C.3.4.11.2), NEP and ACE activity will be measured in the

BMEC monolayers. The third aim will be to develop strategies to reduce methionine enkephalin peptidase degradation and thereby increase the passage of methionine enkephalin across the BBB. DPDPE Tyr-D-Pen-Gly-PIic-D-Pcn

MET-ENK Tyr—Gly—Gly—Phc—Met A A

Aniinopcptidasc | 1 Angiotensin Converting Enz>'mc Neutral Endopcptidasc 3.4.24.11

Fig 2.1 The amino acid structure of the conformationally constrained peptide DPDPE and methionine enkephalin with the enzymatic cleavage sites responsible for methionine enkephalin degradation. 55

Methods

In Vitro BMEC studies.

Fresh bovine brains (2-3) were obtained from a local slaughterhouse. Brain microvessel endothelial cells were then isolated from the gray matter of the cerebral cortex and cryo-preserved, as previously described (Audus and Borchardt, 1986; Audus and

Borchardt, 1987; Weber et al., 1993). Table 2.1 describes a detailed outline of the procedure for isolating BMECs.

The isolated cells were seeded at a cell density of 50,000 cells/cm^ onto tissue culture dishes, which had been pre-coated with rat tail collagen and fibronectin and contained 25-mm Costar Nucleopore polycarbonate membrane filters (10 wm pore size;

Costar Corp., Cambridge, MA). The culture media contained a final concentration of 45%

MEM culture powder, 45% HAMS F-12 culture powder, 10 mM HEPES, 10% equine serum, 13mM sodium bicarbonate, 100 j/g/ml penicillin G, 100 ug/ml streptomycin, 50 z/g/ml amphotericin B, 100 wg/ml Heparin, 50 ug/ml Polymixin B and Gentamicin 50 ug/ml.

After formation of confluent monolayers (10-12 days) the BMECs were used for transendothelial transport studies. The polycarbonate membranes plus the confluent endothelial cells were placed in a glass Side-Bi-Side diffusion cell (Crown Glass Co.,

Somerville, NJ) that was maintained at 37°C. Both sides of the assay chamber contained the physiologically balanced phosphate-buffered saline (PBS) (122 mM NaCl, 3 mM KCl, 56

25 mM Na2P04, 1.3 mM K2HPO4, 1.4 mM CaCl2, 1.2 mM MgS04, 10 mM glucose

and 10 mM HEPES, pH 7.4), which was continuously stirred. The test peptide (500 z/M),

together with the membrane impermeant marker, ["C]sucrose, were added to the donor

chamber at time 0. After the set time points (0, 15, 30, 60, 90 and 120 min), 200 II\

aliquots were removed from the receiver chamber. An equal volume of 50/50 acetonitrile/water was then added to each sample, which were then stored at -37°C for later HPLC analysis and radioactive scintillation counting (Figure 2.2).

For methionine enkephalin (500 WM) permeability experiments the specific

inhibitors coincubated with methionine enkephalin were (10 KM), thiorphan

(0.075 and 5 Z^M) or captopril (4 and 40 Z/M) (see Table 2.2). A cocktail of inhibitors was tested with DPDPE. The cocktail included amastatin (10 MM), bestatin (50 MM),

puromycin (6 MM), phosphoramidon (6 MM), thiorphan (5 MM) and captopril (4 //M).

Radiolabelled sucrose, a BBB impermeant molecule, was added simultaneously with the

peptide to ensure the integrity of the BMEC monolayers. All experiments discussed in this dissertation had sucrose penetration ranging from 7 to 11% (n-15) with a mean ± standard

error of 8.72 ± 0.40. These data show that the various treatments had no effect on the

integrity of tight junction formation since sucrose penetration was not increased during the experiment. All peptide-plus-inhibitor passages were compared to a peptide-without- inhibitor passage performed on the same day and with membranes from the same confluent monolayer of BMEC's. All treatments were performed in triplicate. 57

Permeability coefficients (PC) were calculated on the basis of diffijsion of the peptides across the BMEC by means of the following equation;

PC = X/(AxtxCD)

where PC is measured in cm/min, X is the amount of substance in moles in the receptor chamber after correction for sampling error and paracellular passage based on sucrose levels at time, t, in minutes, A is the diffusion area (0.636 cm^) and CQ is the concentration of substance in the donor chamber in moLcm-^ (C^ remains > 90% of the initial value over the time course of the experiments). 5S

Table 2.1.

Protocol for Isolation of Bovine Brain Microvesssel Endothelial Cells

Clean brains with phosphate buffered saline PBS with 3x antibiotics.

Remove surface vessels and meninges from 2-3 brains (Brain material bathed in N'Cnimum Essential Medium; MEM pH 7.4).

Aspirate cerebral gray matter from cerebral cortex using a vacuum.

Dilute gray matter to 500 ml (w/MEM) containing dispase (final concentration 0.5%). Incubate 3 hr at 37°C in a shaking water bath (adjust pH after first 30 minutes).

Centrifiige 1000 x g for 10 min, discard supernatant, resuspend pellets in 5000 ml of 13% dextran (avg. MW 70,000).

Centrifiige 5,8000 x g for 10 min, discard supernatant fat, cell debris and myelin floating on dextran. Resuspend crude microvessel pellet in 20 ml of collagenase / dispase (final concentration 1 mg/ml). Incubate for 5 hr at 37°C in a shaking water bath.

Centrifiige 1000 x g for 10 min and discard supernatant. Resuspend microvessels in 8 ml MEM. Layer this suspension over a 50 % Percoll gradient.

Centrifiige 1000 x g for 10 min. Remove band 2 from the gradient and wash with culture medium. Resuspend cell suspension in freezing medium with 20 % equine serum and 10 % DMSO. Aliquot for storage overnight at -70°C. Cells are then transferred to (N2)(l) for storage.

Band I—cell debris Band 2--reddish clumps of microvessel endothelial cells

Band 3—erythrocytes (discard) 59

Isolated Oram Microvossols

Tissue Culture Dish (50,000 calls/cm^)

HPLC Analysis

BMEC Scintillation Spectromotry Monolayers Raoenerated Cellulose or Polycarbonate Membranes Mambrane 100 Sample Support [jruo IncubatioA C10 ~ M days) ixn?^ 3 ml

Water balh (37°C) Recoiver Donor

Stir bar Transendothelial Assay Buffer

122 mM NaCi 10 mM HEPES, pH 7.4 3 mM KCI 25 mM NaHCOa 1.4 mM CaClj 10 mM Glucose 1.2 mM MgS04 0.4 mM KzHPO^

Fig 2.2 Outline of transendothelial transport study design. Brain microvessels are seeded and grown to confluent monolayers on dishes precoated with collagen and fibronectin. Once confluence is achieved cell monolayer is mounted in a Side-Bi-Side diffusion cell. 60

Table 2.2 .

Specific inhibitors of the methionine enkephalin peptidases.

Percent Enzyme Inhibitor Concentration Inhibition Aminopeptidase M Amastatin 10 mM >98 % Angiotensin Captopril 4 wM; 40 wM >98 % Converting Enzyme Thiorphan 5wM 90% Thiorphan 0.075 uM 20% Neutral Thiorphan 5J/M 98% Endopeptidase Thiorphan 0.075 wM 75% Percent inhibition based on literature values for porcine kidney (amastatin; Gillespie, et al., 1992; thiorphan: Matsas et al., 1984) and rabbit lung (captopril: Dubreul, 1989). 61

Specific enzymatic assay.

BMEC's grown to a confluent monolayer were harvested by gently scraping the cells off the dishes. Cells were centrifiiged at 2000 rpm for 10 min then washed twice with culture media and resuspended in 20 mM Tris, pH 7.4 for analysis of enzyme activity and determination of protein concentration by the Lowry (1951) or Bradford (1976) method. Cells were then disrupted by sonification for 35 seconds at an output of 3 and a

20 % cycle (Branson 450, Danbury, Conn.). The remaining cell suspension was centrifuged for 45 min at 20,000g. The supernatant was poured off and used for the soluble fraction. The remaining pellet was used for the membrane fraction. The Bradford method was used because it has negligible reaction to collagen and ftbronectin which were coharvested with the endothelial cells. The specific activity of total aminopeptidase and aminopeptidase M was determined by the method of Gillespie et al (1992). NEP activity was determined as described by Bateman and Hersh (1987) with the exception that a 6

|iM phosphoramidon concentration was used and the incubation time was extended to 6 hr due to the low level of NEP detected in BMEC monolayers. ACE activity was measured by the method of Yang and NefF (1972).

In vivo brain uptake.

On the day of the studies male CDl mice (26-28g, n = 5-7) were anesthetized with sodium pentobarbital (80 mg/kg) and administered by tail-vein; radiolabelled peptides

(1.5-2.0 |iCi per animal). After 10 min the chest cavity was opened and a blood sample 62

was immediately taken from the right ventricle of the heart. The brain was then perflised

with 0.9% saline for 1 min, resulting in brain blanching. Immediately following perfiision

the brain was removed, blotted dry, immediately weighed and solubilized by addition of

tissue solubilizer. Tissue solubilizer was also added to 50 ml samples of blood. Following

solubilization, 50 fil of glacial acetic acid was added to the samples to eliminate

chemiluminescence. Ten ml of scintillation cocktail were added and the samples counted

on a Beckman LS 5000. Results are expressed as a percent of total i.v. dose administered.

Purity of all radiolabeled peptides was checked before all studies using the HPLC system

described below which was coupled to an A200 Flo-One® Radioactive Detector

(Radiomatic Instruments & Chemical Co., Inc., Meriden, CT) equipped with a 2.5 ml flow

cell.

HPLC Analysis.

Samples from all diffusion studies were analyzed on a reverse-phase HPLC system consisting of Waters Associates WISP 71OB Autoinjector, 2 Model 6000A Solvent

Delivery Pumps, Automated Gradient Controller (Waters Associates, Milford MA),

Perkin Elmer LC-65T Detector/Oven (210 nm; Perkin Elmer, Norwalk, CT), Hewlett-

Packard 3390A Integrator (Hewlett-Packard Co., Avondale, PA), and a Vydac 2I8TP54 column (4.6 x 250 mm; Vydac, Hesperia, CA) or Beckman Ultrasphere^^ ODS column

(4.6 X 250 mm; Fullerton, CA) a previously described by Davis (1990). Samples were eluted using a linear gradient of acetonitrile against O.IM NaH2P04 buffer (pH 2.4). The 63 flow rate was maintained at 1.5 ml/min and tlie column temperature at 40°C. As a measure of lipophilicity, the above system and conditions were used to determine capacity factors {k) on a gradient of 5-45% acetonitrile against O.IM NaH2P04 buffer (pH 2.4) in

40 min.

Capacity factor = ^ = (tr - to)/to

where tr = the retention time of the retained peak and to = the retention time of an unretained peak.

Data Analysis and Statistics.

Analysis of the regression line and correlation coefficient was determined using the

Pharmacologic Calculation System™ (Version 4.0; Tallarida and Murray, 1987).

Analysis of variance coupled with the Newman-Keuls test was used to determine significance (^0.01 and /7<0.05) between permeability coefficients. 64

Results

I. Functional Aspects of BMEC

In vitro BMEC studies.

BMECs grew to confluence in 10-12 days after seeding. As shown in table 2.3, permeability coefficients ranged from 14.34 to 92.C0 (x 10"*) cm/min. B-Glc-DCDCE was found to have the lowest permeability coefficient, and test of the permeability coefficients revealed that passage across the BMEC of several peptides was significantly greater than that of others (Table 2.4). In particular, passage of [p-Cl-Phe"*'"* jbiphalin, [p-Cl-

Phe'*]DPDPE and reduced DPDPE were significantly (P <0.01) greater than passage of the other peptides examined.

When comparing in vitro permeability coefficient data to the HPLC capacity factor data, analysis of the regression line yielded a correlation coefficient of 0.745 which was significant (P < 0.01).

In vivo BBB studies.

The study of BBB passage using radiolabeled peptides was performed in an attempt to determine whether the BMEC model is useful for predicting in vivo penetration. The whole brmn levels of radiolabeled DPDPE, [p-Cl- Phe'^JDPDPE, biphalin 65 and [GIu'*] were examined 10 min after i.v. administration. Table 2.3 shows that the rank order of BMEC permeability coeflScient and the percent of total injected dose of the peptides entering the brain are the same.

When comparing in vitro permeability coeflBcient data to the in vivo EBB data, analysis of the regression line yielded a correlation coeflBcient of 0.998 which was significant (P < 0.01). 66

Table 2.3.

Rank Order of BMEC Permeability Coefficient, % I.V. Dose and Capacity Factor

BMEC Permeability % of Total I.V. Capacity Factor (A-)

Coefficient (PC xlO"^) Dose r = 0.75 r = 0.998

fp-ClPhe-^-'^'lBiphalin 92.00±5.88 10.90

b-ClPhe^lDPDPE 82.76±3.46 0.178 ±0.030 9.54

acyclic reduced DPDPE 76.22±3.80 11.11

PheO-DPDPE 62.00±4.06 10.07

acetvlated Phe®-DPDPE 56.76±3.98 12.50

DPLCE 56.34±2.00 10.87

Biphaiin 55.00±4.98 0.089 ±0.008 8.03

fL-Ala^lDPDPE 54.00±2.76 8.07

DPDPE 49.24±2.78 0.064 ±0.012 8.25

descarboxy DPCE 47.84±1.24 8.28

rMet^lenkephalin 46.44±4.28 6.50

fGlu'^lDeltorphin 38.00±6.30 0.038 ± 0.009 8.25

B-Glc-DCDCE 14.34±1.94 3.82

Capacity factor = (tr - tO)/tO, where tr = the retention time of the retained peak and tO = the retention time of an unretained peak. See Methodology section for HPLC conditions which were identical for all peptides. Apparent permeability coefficients were calculated by the following equation: PC (cm/min) = X/(A x t x Cj)) where PC is the apparent permeability coefficient in cm/min, X is the amount of substance in moles in the receptor chamber at time, t in min, A is the diffijsion area (ie., 0.626 cm^), and Cq the concentration of substance in the donor chamber in mol cm"^. ( r = correlation coefficient, calculated from linear regression) 67

Table 2.4.

Analysis of Variance Coupled with Newman-Keuls test for Significance Between Permeability Coefficient (PC) Values.

1 2 3 4 5 6 7 8 9 10 11 12 1. DPDPE 2. fn-CIPhe'^lDFDPE •• 3. reduced DPDPE •• 4. ac. PheO-DPDPE •• 5. PheO-DPDPE • •« «• 6. descarboxy DPCE 7. DPLCE •• 8. DPADPE • 9. Met-Enk 10. Biphalin • • •• 11. rp-ClPheM'lBiphalin 12. Deltorphin 11

** Denotes significance at ;7<0.01 and * denotes significance at p<0.05 68 n. Enzymatic Aspects of BMECs.

In vitro enzyme activity.

Enzyme levels for AP, APM, ACE and NEP were measured in BMEC grown to confluent monolayers (Table 2.5). Both membrane-bound and soluble enzyme activities in

BMEC's were measured for total aminopeptidase, APM and ACE. The membrane-bound fractions of AP, APM and ACE showed high levels of activity. No activity was detected in the soluble fraction for APM or NEP. NEP is solely a membrane-bound, ectoenzyme, thus no cytosolic activity was measured for NEP. Low membrane-associated levels for

NEP were detected.

In vitro BMEC permeability.

To ensure the physiology of the endothelial cell monolayer was not disrupted.

Figure 2.3 shows that the passage of methionine enkephalin was linear with time. Thus, all time points were used to calculate mean PC ± SEM values shown in table 2.6. The passage of methionine enkephalin across BMEC monolayers was significantly (p <

0.0001) increased approximately four-fold in the presence of the specific inhibitors to the metabolizing enzymes APM (amastatin) and ACE (5 MM thiorphan ad 40 J/M captopril).

Methionine enkephalin permeability was significantly (p < 0.0001) enhanced approximately two-fold with the lower concentration of captopril (4 i/M). The low 69

concentration of thiorphan (0.075), which is more specific for NEP inhibition, had no

effect on methionine enkephalin permeability.

A common question concerning CNS delivery of therapeutic agents is the integrity

of the parent drug after crossing the BBB. The integrity of the peptides studied in the

BMEC system was assayed through HPLC chromatograms (Figure 2.4). The chromatogram of the 120 min receiver chamber sample clearly showed a single peak that co-migrated with the sample taken at 0 min (which was not exposed to BMEC). This ensures that intact peptide fi^om the receiver side of the diffusion chamber was quantified

by the HPLC.

Table 2.6 also shows the PC values for DPDPE with and without inhibitors and methionine enkephalin with and without inhibitors. The PC for DPDPE was significantly larger compared to methionine enkephalin (p < 0.0001). This is approximately a four-fold increase for DPDPE compared to methionine enkephalin. These results are comparable to the changes observed in methionine enkephalin permeability when co-incubated with specific inhibitors to APM and ACE. There was no enhancement of DPDPE passage across the endothelial cell monolayer in the presence of the inhibitor cocktail. Inhibitors were pre-incubated for 20 minutes prior to peptide addition. Experiments were performed in triplicate. 70

Table 2.5.

Specific enzyme activity in BMEC monolayers.

Angiotensin Neutral Aminopeptidase Aminopeptidase Converting Endopeptidase M Enzyme Membrane 4,383 ± 144 1,618 ±155 3,219 ±246 115±2 Soluble 1,414 N.D. 2,817 ±228

Data represented as mean ± SEM pmol / mg protein / min or not detected (N.D.). 71

1,800 Met-Enk + Thiorphan {5|jlVl) 1,600 (r=0.987) o §1,400

•^^1,200 cn §1,000 o CJ 000 (UCO < GOO O Q. 400 X 200

0 I f t t * « ' 20 40 60 00 100 120 Time (min)

Fig 2.3 Linearity of transendothelial passage of methionine enkephalin (with tliiorphan (5 ifM) to protect against degradation by E.G. 3.4.24.11). Data shown is mean ± SEM. (n = 3 independent measurements. 72

Table 2.6.

Effect of specific enzyme inhibitors on the permeability coefficient (PC) of methionine enkephalin across BMEC monolayers

Peptide and Inhibitor Enzyme Targeted Permeability Coefficient (x 10"^ ± SEM cm/min)

[Met^]Enkephalin none 12.0 ± 1.0

[Met^JEnkephalin aminopeptidase M 49.2 ±2 (4.1 fold T) (APM) >98% (10 MM amastatin)

[Met^JEnkephalin angiotensin converting 39.0 ± 2 (2.3 fold t) (4 //M captopril) enzyme (ACE) >98%

[Met^JEnkephalin angiotensin converting 64.0 ± 4 (4.3 fold t) (40 j/M captopril) en^one (ACE) >98%

[Met^]Enkephalin angiotensin converting 48.0 ±2 (4.0 fold t) (5 MM thiorphan) enzyme (ACE) >90%

[Met^]Enkephalin neutral endopeptidase 14.0 ±1 (1.0 fold t) (0.075 ?/M thiorphan) 24.11 (NEP) >75%

DPDPE none 45.0 ± 3

DPDPE (APM, ACE and NEP) 32.0 ±2 (0.7 fold i) *(cocktail of inhibitor) Donor Chamber

EC o & (D cO s 8

T' IRI ^ 1—rV-r

Time (min)

Receiver Chamber

e c o (D cO no

Time (min)

Fig 2.4 Reversed-phase HPLC analysis of methionine enkephalin from BMEC assay chambers. Chromatogram of methionine enkephalin sample taken from donor chamber at time 0 min and a sample taken from receiver chamber at time 120 min. Samples analyzed using a Vydac 218TP54 column (4.6 .x 250 mm), a gradient of 0-30% acetonitrile vs. 0. IM NaH2P04 in 30 min at 37°C. 74

Discussion

Several in vitro BBB models have been developed to examine the passage of drugs and related compounds across the in vivo BBB. As with all models these techniques all have limitations including sensitivity, set-up time and necessity for radiolabeled drug. The need for an adequate model to screen numerous nonradioactive drug analogs necessitates the characterization of our in vitro BMEC model both functionally and enzymatically.

Since the drugs described in this work are peptides it was necessary to assess the activity of specific peptidases in the BMEC system.

Since peptides are hydrophilic compounds, oftentimes an extremely limited concentration of peptides cross the BBB (Rapoport et al., 1980). This characterization began by investigating structural modifications of a single peptide, methionine enkephalin that were designed to enhance lipophilicity and presumably BBB passage. It is generally agreed that any increase in lipophilicity should, in theory, increase its passage across biological membranes, as long as specific, saturable transport mechanisms do not solely control the CNS entry of the drug of interest. The data presented in this study supports this theory. A significant correlation between HPLC capacity factor (a measure of lipophilicity) and the permeability coefficient of the peptides examined suggests that the two are related in passage across the BMEC monolayer (Table 2.3). When comparing in vitro permeability coefficient data to the HPLC capacity factor data, analysis of the regression line yielded a correlation coefficient of 0.745 which was significant (P < 0.01).

In addition, the rank order, from highest to lowest BBB penetration, of the four peptides 75 examined in vivo was consistent with the in vitro BMEC permeability coefficients

calculated for the same peptide analogs. Also, when comparing in vitro permeability

coefficient data to the in vivo BBB data, analysis of the regression line yielded a

correlation coefficient of 0.998 which was significant (P < 0.01). This suggest that the

BMEC confluent monolayer is reflective of the in vivo BBB, at least for these four peptide analogs. This study also reports that structural modifications designed to increase lipophilicity can be correlated with passage across the BMEC monolayer.

The BMEC model was also used in this study to investigate the presence of

peptidases which degrade methionine enkephalin. It has been reported from the literature

that systemic administration of methionine enkephalin does not produce analgesia (Klee,

1977). This is most likely due to systemic metabolism by blood-borne peptidases or actual

metabolism at the endothelial cell component that comprises the BBB. A possible explanation for the inability of methionine enkephalin to cross the BBB is that cerebral endothelial peptidases specific for methionine enkephalin degradation may prevent the

passage of enkephalin fi-om the blood to the brain. This work showed that high levels of

membrane bound enzymatic activity were measured for total aminopeptidase, APM and

ACE and low amounts for NEP in confluent monolayers of endothelial cells (Table 2.5).

No APM or NEP activity have been investigated in the BMEC model until this present study.

It was encouraging to observe that high levels of APM and ACE activity were measured in the membrane fraction of BMEC (Table 2.5),and the specific inhibitors of 76 both APM and ACE significantly enhanced passage of methionine enkephalin across confluent BMEC monolayers (Table 2.6) (p < 0.0001). Whereas, low levels of NEP were measured in the BMEC monolayers and specific inhibition of NEP had no effect on methionine enkephalin permeability. These result support the idea that the activity of a given degradative en2yme at the BBB can have a dramatic effect on enkephalin permeability. It was interesting to observe that a higher than expected concentration of captopril was required to enhance methionine enkephalin permeability four-fold. High levels of captopril may be required, because it has been shown to be chemically unstable and able to undergo rapid degradation in biological fluids (Vlasses et al., 1982). From analysis of enzyme activity at the BMEC (Table 2.5) and inhibition of these enzymes on transendothelial permeability (Table 2.6) it is evident that the two primary peptidases involved in the blood-brain enkephalin barrier are APM and ACE.

DPDPE is a conformationally constrained and enzymatically stable analogue of enkephalin with D-penicillamines incorporated in the 2 and 5 position (Hruby et al., 1991).

The BMEC permeability coeflRcient for DPDPE was observed to be four-fold greater than methionine enkephalin and was not increased by peptidase inhibitors. DPDPE permeability was equivalent to the permeability coefficient of methionine enkephalin in the presence of APM and ACE inhibitors. These data support the hypothesis that peptidases active at specific hydrolytic cleavage sites of enkephalin can be affected not only by chemical inhibitors but also by amino acid substitutions, leading to improved BBB permeability. 77

Overall, these results have characterized an in vitro BMEC model of the BBB both functionally and enzymatically. This model will be useful for screening large numbers of drugs for potential enhanced BBB penetration, so that potential lead compounds can be further characterized by usmg more precise techniques. Based on Table 2.3, it can be observed that both DPDPE, biphalin and their chlorinated analogs have good potential

BBB penetration. These analogs will be discussed further in this dissertation in Chapters

3-7. 78

Chapter 3. Deflning the Mechanism of DPDPE Central Nervous System Entry.

Introduction

[D-Penicillamine2,5]enkephalin (DPDPE; tyrosine-D-penicillamine-glycine- phenylalanine-D-penicillamine) is a 5-opioid receptor selective agonist that is enzymatically stable due to its chemical structure (Mosberg et al., 1983, Weber et al.,

1991, 1992). DPDPE is enzymatically stable due to the D-penicillamine in the 2 and the 5 positions as well as the disulfide bond between these two positions. Aminopeptidase can not recognize the active site for hydrolysis, thus, DPDPE has extended half life values of greater than 500 minutes in blood or brain homogenate. DPDPE was designed with the goal of producing a 5-receptor selective analgesic that could be used clinically to produce antinociception without eliciting the unwanted side effects such as respiratory depression, constipation, dependence and tolerance (Mosberg et al., 1983; Hruby et al., 1991). It was also hoped that DPDPE would provide an invaluable research tool for pharmacologist to uncover the physiological roles of the opioid receptor types (|.i, K and 5) (Martin et al.,

1976). Today DPDPE is sold by a wide range of suppliers of basic biomedical research compounds.

Our laboratory has previously investigated the absorption, distribution, metabolism and excretion of [^HJDPDPE after intraperitioneal, intravenous, subcutaneous, and oral administration (Weber et al., 1991, 1992). It was shown that there was significant CNS 79 uptake of [^HJDPDPE that could be displaced by naloxone. Several other studies have shown that intravenous, intracerebroventricular, and intrathecal administration of DPDPE could produce significant analgesia, suggesting that CNS uptake does occur (Galligan et al., 1984; Porreca et al., 1984; Heyman et al., 1987; Weber et al., 1991). In addition, as shown in chapter 2 of this dissertation, DPDPE does cross the in vitro BMEC model of the BBB to a significantly greater extent than ["Qsucrose and methionine enkephalin (P <

0.01). When DPDPE was administered subcutaneously, no analgesic effect was elicited, thus it was inferred that this peptide does not cross the BBB and enter the brain very effectively (Shook et al., 1987). This observation may be due to the fact that the route of administration plays an important role in the bioavailability of drugs to the brain and it is possible that the volume of distribution of the drug may be altered after a subcutaneous route. The amount of DPDPE actually entering the circulation may be quite small from this subcutaneous route of administration due to tissue sequestration.

The aim of this study was to investigate the CNS uptake of [^H]DPDPE using an extensively characterized method called the in situ brain perfijsion technique in the anesthetized rat (Zlokovic et al., 1986; Preston et al., 1995). This technique is especially useful for studying the uptake of slowly permeating molecules such as peptides across both the BBB and the blood-CSF barrier. The systemic contribution to CNS uptake is eliminated so the interaction with the test drug and the brain endothelium can be more specifically investigated. This technique can also classify the uptake of a given dnig into both saturable and difRisional components. Hopefiilly, a clearer understanding of the 80

passage of across biological membranes will clarify their roles in drug addiction and develop new therapeutic approaches. 81

Methods

In Situ Brain Perfitsion Studies.

The protocol described below was approved by the Institutional Animal Care and

Use committee (lACUC) at the University of Arizona. Adult Sprague-Dawley rats (250-

300g) were anaesthetized with sodium pentobarbital (64.8 mg.kg"^) and heparinized

(10,000 U.kg"^). The jugular veins were located and the common carotid arteries were

cannulated using fine silicone tubing connected to a perfusion system as previously

described (Takasato et al., 1985). (Figure 3.1)

Perfusion was performed with a thoroughly oxygenated (p02 = 642-727 mm Hg)

mammalian Ringer (37°C) solution. The erythrocyte-free perfusion fluid (Preston et al.,

1995) consisted of a modified Krebs-Henseleit Ringers's solution (117.0 mM NaCl,

4.7mM KCl, 0.8 mM MgS04.3H20, 24.8 mM NaHCOj, 1.2 mM KH2PO4, 2.5 mi\l

CaCl2"6H20, 10 mM D-glucose, 39 gL"^ of dextran (MW 70,000), and 1 g L-I of bovine serum albumin). Once the desired perfusion pressure and rate were achieved

(approximately 100 mmHg and 3.1 mls.min"^ respectively), the right jugular vein was cut and allowed to drain. The contralateral carotid was cannulated and perfused in a similar manner as described above. [^H]DPDPE and ["Cjsucrose in the presence or absence of varying concentration of cold DPDPE (0, 2.5, 5, 10, 50 and 100 joM), was infused using a slow-drive syringe pump into the inflowing, mammalian Ringer. Once the set perfusion time (0-30 min) was achieved, a cistema magna cerebrospinal fluid CSF sample was taken 82 with a glass cannulae. The animal was then decapitated and the brain was removed. The choroid plexuses were excised and the brain was dissected. The perfusion outflow was collected from the carotid cannulae at the end of the time point to serve as a reference.

Brain tissue samples (-50 mg wet weight) together with the CSF and 100-H1 perfusate samples were prepared for liquid scintillation counting. All samples were treated in the same manner to ensure uniformity. They were treated with 1 ml of tissue solubilizer

(TS-2; Research Products, Mount Pleasant, EL, U.S.A.). After solubilization, 100 M1 of 30

% glacial acetic acid was added to each sample to eliminate chemiluminescence.

Approximately 4 ml of Budget Solve Liquid Scintillation Cocktail (Research Products) was added, and the samples were counted for radiactivity using a beta counter (model LS

5000 TD counter; Beckman Instruments, Fullerton, CA, U.S.A.). The and activities were converted from cpm to dpm using internal stored quench curves. S3

Peristaltic Heater Pump

Bubble Ringer Trap

Pressure Transducer

Common Carotid artery cannulation

Isotope ^ Infusion iz zz iL±

Fig 3.1 A schematic diagram of the perfusion circuit. The brain is perfused via both common carotid arteries with oxygenated mammalian Ringer (Gas=95%02 and 5%C02). Radiolabelled compounds can be introduced via a slow drive syringe. Both jugular veins are sectioned or to allow outflow of the perfusate. 84

Expression of Results.

The amount of radioactivity in the brain and CSF (Cxissug; dpm.g"^ or dpm.mi~^) was expressed as a percentage of that in the artificial perfiisate (Cpj; dpm.ml"l) and termed the Rxissue

^Tissue ~ CTissiie X 100 (1) Cpl

Unidirectional transfer constant, (Kin) and the initial volume of distribution C^-'i) were determined graphically form the multiple-time uptake data (2.5-30 min) as described previously by ZIokovic et al. (1986).

(T) / Cp,3^ (T) = Kin T + V: (2a)

Where Cuaue (T) and Cpuma (T) are radioactivities per unit weight of tissue and plasma at time T (perfusion in minutes). The above equation describes a straight line, where the slope is Kin (milliliters per minute"' per gram"') and y-intercept is Vi (milliliters per gram- l). Any brain-to-blood movement of the test compound can be observed as a departure from linearity of the experimental points. To determine blood-to-CSF transfer constants, a two compartment / singe-time uptake analysis was used.

Unidirectional rate constants (Kj^ |il.min"^g"^) were also determined by single time-point analysis, as previously described (ZIokovic et al., 1986), where T is the time in minutes:

Kin ~ CTissue (T") (2b) 85

Blood-to-brain unidirectional transfer constants determined in this manner were corrected for vascular space by subtracting ['^C]sucrose (Rerain) [^H]DPDPE(RBrain)-

Unidirectional transfer constants determined from these experiments, represent cerebrovascular permeability surface area products, PA (ml.min'.g"') (Gjedde, 1988;

Zlokovic et al., 1989;1990). Thus, if [^H]DPDPE is being studied in the presence of increasing uniabelled concentrations of DPDPE, can be defined as:

PA = BC. = Vmax + (Knx + Ccap) (3) where is the maximal transport rate of the saturable component; Kj^ is the half- saturation constant; is the constant of non-saturable diSusion and is the mean capillary concentration of DPDPE. Under the experimental conditions described above, the difference between C^ap and the concentration of DPDPE in the perfusion medium

(Cpi), becomes negligible, since the flow to the brain (F) is always greater than 1 ml.min*

^.g"l, which is much greater than the highest measured and the equation can be simplified (Gjedde, 1983,; Smith et al., 1984; Zlokovic et al., 1989; 1990):

Kin ~ Vmax + K^} (4) 3 (Km + Cpi) Unidirectional [ H]DPDPE flux (Jjjj; nmol.min~^.ml'^) into the brain and CSF can then be calculated as:

Ji„ = F(l-e-Kin/F)Cp, (5) and since F »Kin (Takasato et al., 1985) this equation approximates to;

•^in ~ KinCp[ (6) 86

Unidirectional Flux of [^H]DPDPE can be related to V^j^x and K^j by the following equation:

Jin = Vmax Cpl + Kj Cpi (7) (^m ^pl)

Estimates of the best fit values for (iunol.min"'.g"'), (mM) and Kj (ml.min"^g"

^) were obtained by fitting this equation to tlie brain vascular perfusion data by the method of least squares with statistical weighting (Enzfitter program from Biosofl, Cambridge,

UK).

Capillary Depletion.

Measurement of the vascular contribution to total brain uptake was performed using a capillary depletion step as previously described (Triguero et al., 1990; Zlokovic et al., 1992). Briefly, the brain was removed and choroid plexuses excised. The brain tissue

(500 mg) was homogenized (Polytron homogenizer, Brinkmann Instruments, Westbury

NY) in 1.5 mis of physiological buffer kept on ice (10 mM HEPES, 141 mM NaCl, 4 mM

KCl, 2.8 mM CaCl2, I niM MgS04, I mM NaH2P04 and 10 mM D-glucose, pH 7.4)

Two ml of ice-cold 26% dextran (MW 60,000) were then added and homogenization was performed again. Two aliquots of homogenate were taken and centrifuged at 5,400 x g for 15 min in a microfiige (Beckman Instruments Inc.). The capillary-depleted supernatant was then separated from the vascular pellet. All the above homogenization procedures were performed within 2 min. The homogenate, supernatant, and pellet were then 87

aliquoted and prepared for radioactive counting (Beckman 5500 beta counter).

Homogenate, Pellet, and supernatant were then solubilized, as described above, before the

addition of scintillation fluid to enable radioacitve counting. Contamination of the

supernatant with blood vessels was monitored by measuring for the specific vascular

enzyme, alkaline phosphatase, using a Sigma assay kit (Procedure # 104). The alkaline

phosphates supernatant / pellet ratio was found to be 4.61 ± 0.86 %, which is in

agreement with the value published by Tiguero et al. (1990). This indicates that little

contamination of the supernatant by cerebral capillary endothelial cells exists.

Protein Binding Studies.

The amount of [^HJDPDPE binding to either bovine albumin in the perfusion

medium or proteins in rat serum was determined by ultrafiltration centrifugal dialysis

(Paulus H., 1969). Rat serum was obtained by harvesting blood from Sprague-Dawley

rats and allowing the blood to clot for 30 minutes on ice and 30 minutes at room

temperature. The whole blood was then centrifliged (Sorvall RC2-B centrifuge; Dupont

Medical Products, Wilmington, DE) at 20,000 x g for 20 minutes to produce a serum

supernatant. [^HJDPDPE was dissolved in either perfusion medium or rat serum warmed

to 37^0 and ultrafiltrated using a Centrifree™ micropartition device (Amicon, Beverly,

MA). The total concentration (T) of [^H]DPDPE introduced into the system and found in the ultrafiltrate (F) was determined by counting on a Beckman 5500 gamma counter. The

percentage of [^H]DPDPE bound to either albumin in the perfusion medium or proteins in 88 the rat serum was expressed as [(T-F) / T ] x 100. To verify that bovine albumin was not found in the ultrafiltrate, the protein concentration was determined by the method of

Lowry et al. (1951).

Octanol/salinepartition coefficients.

Octanol / saline partition coefficients give a measure of lipophilicity. Partition coefficients were determined for [^H]DPDPE and ["C]sucrose by the method of Collins et al. (1988). This involves adding approximately 1 uCi of the test compound in 0.9% sterile saline solution (pH 7.4) at room temperature. An equal volume of octanol is added to the test solution and vortexed for 5 minutes. The mixture is then centrifiiged at 1,000 g for 5 min. The upper phase is the octanol phase and lower is the aqueous phase. These are separated using a pasture pipette and each phase is analyzed by liquid scintillation counting. A partition coefficient is determined by taking the ratio of labeled substance in the octanol phase to the concentration in the aqueous phase. Triplicate determinations were made for each test compound and the date is expressed as the mean ± SEM.

HPLC analysis for ^HJDPDPE extraction experiments.

The arterial perfusate, venous outflow, and brain were analyzed using a series 410

HPLC gradient system (Perkin-Elmer, Norwalk, CT, U.S.A.) as previously described

(Davis, 1990). This established the integrity of the tritiated label to DPDPE and the stability of DPDPE. Perfusion medium samples were prepared for analysis by addition of 89 an equal volume of acetonitrile, mixing and centrifugation at 13,000 g for 5 min. The supernatant was removed and diluted to produce a sample for HPLC analysis with a final acetonitrile concentration of less than 10 %. A modified method of Erchegyi et al. (1991) was used for brain extractions. Rats were perfused with [^H]DPDPE as described above and after 30 minutes the animal was decapitated, and the brain was removed and placed in

7.5 ml of ice-cold 10 % trifluoroacetic acid. The sample was then homogenized (Polytron homogenizer) and centrifiiged at 27,820 g for 20 min (Sorvall RC2-B centrifuge; Du Pont

Medical Products, Wilmington, DE, U.S.A.). The supernatant was then saved and an equal volume of ether was added to extract out lipids. The samples were recentrifiiged at

200 g for 20 min, and the supematants discarded. The samples were then pooled, lyophilized and diluted to 500 ul with 10 % acetonitrile before being taken for HPLC analysis.

HPLC analysis was performed using a 0.46- x l5-cm Inertsil 0DS-2h column

(Metachem Technologies, Torrance, CA, U.S.A.) with a linear gradient of 10-40% 0.1% trifluoroacetic acid in acetonitrile versus 0.1% aqueous trifluoroacetic acid over a 20-min period at 1.5 ml/min and 37°C. Following separation on the column, the HPLC outflow was routed to an A200 Flo-One Radioactive Detector (Packard Radiomatic Instruments and Chemicals, Tampa Bay, FL, U.S.A) and mixed with Flo-Scint 111 (Packard). The sample then passed through a 0.5-ml capacity flow-cell for on-line analysis of radioactive samples. 90

Data.

All experiments are presented as mean ± SEM values. The correlation coefficients

(r), slopes (Kin), and intercepts (V;) of the curves determined by least squares linear regression analysis, and the slopes were compared by ANOVA. Student's t test was used for the comparison of the two means. Statistical significance was taken as p < 0.01 or

0.05. 91

Results

Multiple time uptake analysis.

Multiple perfusion times were performed for both [^H]DPDPE and ["Cjsucrose

(2.5, 5, 10, 15, 20, 30 min). Figure 3.2 illustrates multiple-time uptake plots for

[^H]DPDPE and [^•'C]sucrose into both the brain and CSF. These results show a

progressive uptake of [^H]DPDPE into the CNS, which is significantly greater than that

for ["CJsucrose into the brain (p <0.01) and the CSF (p < 0.05). Also, after subtracting

out the vascular space, [^H]DPDPE uptake into the brain is approximately 57% greater

than that into the CSF (p < 0.01).

Table 3.1 represents the unidirectional transfer constants (ki„) and the initial

volume of distribution of [^H]DPDPE and [^'^Cjsucrose, which were calculated from the

computer generated lines of regression in Figure 3.2. Table 3.1 also shows the

unidirectional transfer constants derived by a two-compartment / singe-time uptake model.

It is interesting to note that the Ki„ value for [^H]DPDPE into the CSF is only 68% of its

rate of uptake into the brain. Also, the initial volume of distribution for [^H]DPDPE is

approximately two times higher than that calculated for ["C]sucrose.

Self-inhibition studies. •)2

The uptake of [^H]DPDPE into the CNS was studied under conditions in which

the brain was perfused with various concentrations (1 to 100 i/M) of unlabelled DPDPE

(Figure 3.3). These experiments revealed a saturable component that contributed to the

brain uptake of [^H]DPDPE. At an unlabelled concentration of 10 wM DPDPE in the

perfusion medium, the uptake of [^HJDPDPE into the brain was inhibited by

approximately 23%. Brain uptake of [^H]DPDPE was not significantly inhibited further

eitheither 50 or 100 um unlabelled DPDPE. [^H]DPDPE uptake into the CNS was not

significantly affected by the presence of l-lOO uM unlabelled DPDPE. Table 3.2

represent the Michaelis-Menten type kinetic values determined for [^H]DPDPE transport

into the brain and CSF.

Capillary depletion analysis.

Figure 3.4 represents the contribution of the vascular component to the total brain uptake of [^H]DPDPE. The amount of radioactivity detected in the vascular-enriched

pellets was small and significantly lower than that detected in the supematants.

Partition coefficient.

The octanol / saline partition coefficients for [^H]DPDPE was 0.076 ± 0.002 and

['"*C]sucrose was 0.00050 ± 0.00003. The protein binding of [^H]DPDPE to bovine serum albumin was 2.95 ± 0.29 % and ['"'CJsucrose was 0.25 ± 0.05 %. 93

Brain extraction.

Figure 3.5a represents the extraction of [^DPDPE in the perfusion medium before (arterial inflow) and after (venous outflow) it had passed through the cerebral circulation. As shown in figure 3.5a, both samples eluted as a singe peak with identical retention times. The brain extraction of [^H]DPDPE was also performed to ensure that the multiple-time uptake data actually represented intact compound and not metabolites or free tritium (Figure 3.5b). After a 30 minute perfusion time approximately 90 % of the brain extraction sample eluted as intact [""HIDPDPE. 94

10 DRAIN r-0.a2 n-2T p<0.06

-r

e 10 20 30 am c C8F

4

r"0.80 3 o"lS p<0.05

2

1 r-0.47 a-l6 P-N8 0 0 10 20 flO Tims (minutos}

Fig 3.2 Brain and CSF uptake of [^tljDPDPE (open squares) and ['"'Cjsucrose (closed triangle) over a time course of 30 minutes. Uptake is expressed as a percent ratio of tissue (brain or CSF) to perfusate radioactivities (Reissue; ml.g"' or ml.ml''). Data are mean ± SEM values of 3-7 animals. 95

Table 3.1.

The calculated unidirectional transfer constants (Kin), initial volumes of distribution (V,) for f H]DPDPE and ['''CJsucrose into the brain and CSF.

Brain CSF Isotope Kin 0'I.rnin'\g"') Vi Cml.lOOg-^) Kin (?/l.min"'.g'') r^Clsucrose 0.16 ±0.08' 1.66 ±0.15 — — 0.70 ± 0.26*' r'HlDPDPE 1.46 ±0.31' 3.60 ±0.58 — 2.26 ± 0.20" 0.99 ± 0.29'

Values ± SEM were determined from ' the computer generated lines of regression in Figure 10 and'' single-time uptake analysis at a perfiision time of 30 minutes (brain value is corrected for vascular space. (n=27) •)6

1 -

II fl _ Salutalilc :3 Conipoiieiil m1 1 -L -1 M n

i Non-salurabic '-T Conipoiienl

JL Vascular Space

1 1 1 1 1 1 1 1 1 ' 0 10 2t) 30 'ID 50 W) 70 St) VO 100 IID

Concciilnitioii {uM)

Fig 3.3 The relative uptake of [""HIDPOPE into the brain (closed squares) and CSF (open squares) measured as a function on unlabelled DPDPE concentration. Uptake is divided into saturable and non-saturable components. The proportion of ["HjDPDPE brain uptake that represents vascular space is also shown. Values are mean ± SEM. (n=3-4 rats at each concentration) 97

Table 3.2.

Michaelis-Menten kinetic parameter for [^HjDPDPE influx into the CNS determined from the self-inhibition experiments.

[^H]DPDPE V^x BCa (mM) (pmol.min'^g'^) (jd.nun"'.g*') Brain 0.0455 ± 0.0276 51.13 ± 13.23 0.56 ± 0.26 CSF — — 0.89 ± 0.07

Values are ± SEM derived from weighted non-linear regression analysis of six Jb observations based on 41 individual experiments. 98

Table 3.3

CNS uptake of [^HJDPDPE after in situ brain perfusion in the presence of known inhibitor of transport mechanisms

[^HIDPDPE Rbrain RCSF %

only 3.03 ±0.37 1.79 ±0.19

+ Leucine Enkephalin 4.29 ±0.18 1.86 ±0.30 (ImM)

+ BCH (lOmM) 5.17 ±0.41 2.21 ±0.33

Values are expressed as mean ± SEM and corrected for vascular space. 9'J

Contribution of the vascular component (pellet) to the brain uptake of [3HjDPDPE.

0 -

llomogantt* Suparnittnl Pall«t

Fig 3.4 Contribution of the vascular component to total brain uptake of [''H]DPDPE. RiiMuc % represents the ratio of tissue to plasma radioactivities x 100. Supernatant is brain homogenate devoid of pellet. Data are mean ± SEM. Perfusion time is 20 min. 100

llandarcJ Ailetlal Inflov/

Ciafu Venous Ouldow

h ——1 R'' • ' • •• •-I-' 5 TO 15 S 10 15 Hmo {rninutcs) nrno (inbiulcsl

Fig 3.5 HPLC / Flo-One Radioactive Detector cliromatograms of (A) [^HjDPDPE from the arterial perfusion medium and the venous outflow. Venous outflow samples were taken at 20 minutes after the start of perfusion. (B) represents HPLC / Flo-One Radioactive Detector chromatograms of a radioactive standard and an acid / ether extraction of pooled brains (n=2 rats) that had been perfused with [^HjDPDPE for 30 minutes. lOl

Discussion

Control experiments were performed to ensure the integrity of the BBB was not disrupted and the measurement of peptide uptake into the brain and CSF was accurate.

Figure 3.2 shows that the uptake of [''*C]sucrose into the brain and CSF was quite low.

This ensures that the highly water soluble compound, ["C]sucrose mainly remains in the vascular space. These control experiments are in fiill agreement with both Zlokovic et al.

(1992) and Ohno et al. (1978) where the brain unidirectional transfer constant. Kin, and cerebral blood volume/vascular space were determined as 0.28 ± 0.08 ul.min*'.g'' and 1.14

± 0.22 ml. IOOg-1, respectively. This ensures that the BBB was not compromised during the 30 minutes of cerebral-vascular perfiision. Perfusion pressure (PO2) was also measured and yielded values in agreement with Preston et al. (1995). The perfusion pressure was measured at 640-730 mm Hg which is adequate for a erythrocyte free mammalian ringer. Figure 3.2 also shows that the uptake of [^H]DPDPE into the brain and CSF was linear with time and significantly greater than the vascular space marker,

['''CJsucrose.

The CNS uptake of [^H]DPDPE can be separated into both brain and CSF uptake.

Ailer considering the vascular space, the brain uptake of [^H]DPDPE was -57% greater than the CSF uptake. The uptake of [^H]DPDPE into the CSF was quite small in comparison the brain uptake although it was significantly greater than the CSF uptake of the vascular space marker, ["C]sucrose. From these data, it appears that the blood-CSF barrier (choroid plexus) is not providing a significant route of CNS entry for this 102 enkephalin analog. The CSF uptake values may actually represent the diffusion of

[^H]DPDPE from the brain compartment to the CSF. This conclusion can be deduced because it has been calculated that the surface area of the BBB is 5000 times greater than that of the blood-CSF barrier at the choroid plexuses, suggesting that the BBB would have a much greater influence on controlling the brain microenvironment (Bradbury, 1979;

Pardridge 1983; Keep and Jones, 1990)

Table 3.1 also shows that the initial volume of distribution (V;) is approximately two times higher for [^H]DPDPE (3.60 ml.IOOg"') compared to ["CJsucrose (1.66 ml.IOOg"'). It appears that this difference may be the result of differences in lipid solubility. The octanol saline partition coefficient for ["HjDPDPE was 0.076 ± 0.002 compared with a value of0.00050 ± 0.00003 for ['"'CJsucrose.

Self inhibition experiments revealed a concentration dependence of [^H]DPDPE transport into the brain. This brain uptake consisted of both saturable and non-saturable components, that could be described by Michaelis-Menten type kinetics with a Km of 46 ±

28 uM, Vnux of 51.13 ± 13.23 pmoI.min"'.g"' and Ka of 0.56 ± 0.26 J/I.min''.g'' (Table

3.2). These kinetic values represent the presence of a saturable uptake system at the BBB with a very low capacity and a relatively large affinity. Leucine-5-enkephalin has been previously shown to enter the brain by a saturable uptake system with a similar affinity (k^

= 34-41 ;/M) and capacity (0.14-0.16 nmol.min'.g'^) (Zlokovic et al., 1989).

Interestingly, I mM leucine-5-enkephalin failed to inhibit both the brain and CSF uptake of

[""HJDPDPE, showing that [^HJDPDPE does not share the same carrier to cross the BBB 103

(Figure 3.3). In contrast to the brain uptake of [^H]DPDPE, the CSF entry was found to be purely non-saturable with a difiusion constant of 0.89 ± 0.07 wl.min Vg"' (Table 3.2).

Structurally, DPDPE contains an N-terminal tyrosine that is necessary for peptide binding to the opioid receptor. This allows DPDPE to potentially utilize the large neutral amino acid carrier to cross either the blood-brain or blood-CSF barriers. However, BCH, a non-metabolizable analog, which is known to be specific for this transporter, failed to inhibit the uptake of [^H]DPDPE into either the brain or the CSF (Table 3.3). This indicates that [""HIDPDPE does not use the large neutral amino acid carrier to gain access to the CNS.

Protein binding can play a major role in the blood-brain pharmacokinetics and bioavailability of peptide uptake into the CNS. Ultrafiltration centrifugal dialysis on

[^H]DPDPE and ['"'Clsucrose was performed. The percentage of protein-bound

[^H]DPDPE was found to be quite small (2.95 %). This shows that protein binding does not limit the CNS entry of [^HjDPDPE. In addition, the vascular space marker,

['•*C]sucrose, was found to have negligible protein binding.

It is important to also monitor for drug binding to the endothelial component of the CNS vasculature. The capillary depletion step allows a researcher to monitor for potential anchoring of a test drug to endothelial cell components or actual intracellular trapping of the drug molecule, preventing transport from the luminal side to the abluminal compartment. The capillary depletion step was developed by Triguero et al. (1992) and used previously by Zlokovic et al. (1992), in combination with the />; si/u brain perfusion 104 technique it provides an accurate assessment of this component of tissue trapping. It was observed that after 20 minutes of vascular perfusion the amount of radioactivity in the brain homogenates was not significantly different from that obtained in the supematants.

The amount of radioactivity present in the capillary rich pellets was also negligible (Figure

3.4). These results show that the brain accumulation of [^HTjDPDPE is solely due to brain uptake and not endothelial cell trapping.

The in situ brain perflision technique provides several advantages for a researcher.

The possibility of metabolism that would influence [^H]DPDPE uptake data is reduced.

This is accomplished by using an enzyme free artificial plasma perflisate and also preventing recirculation of the perfusion medium. Control experiments were performed that monitored for potential [^HjDPDPE metabolism or tritium exchange with water.

HPLC analysis confirmed that there was little tritium exchange with water (Figure 3.5a) after passage through the cerebral circulation. Brain extraction experiments were also performed that showed that the majority of [^H]DPDPE (90 %) was intact afler a 30 minute cerebral vascular perfusion. It is interesting to observe that only a small radioactive metabolite, i.e., 3% was detected at 15.5 minutes in the 30 minute brain sample. The majority of the radioactivity not associated with intact [^HjDPDPE eluted with the solvent front, i.e., 7%, which suggest some tritium exchange with water may have occurred. However, it is not known where this displacement of tritium occurred and it is possible that the tritium could have been displaced once it had reached the brain. 105

In conclusion, this work has demonstrated that ["HjDPDPE can cross the BBB.

This observation is in contrast to that of Shook et al. (1987), where no antinociception was elicited after subcutaneous administration of DPDPE (10 mg/kg) and it was suggested that this peptide could not cross the BBB. The analgesia studies of Shook et al. (1987) was performed in mice and this presented work was conducted in rats suggest that species variation may occur. Although, Weber et al. (1991) observed that significant analgesia was elicited when high doses of DPDPE was administered intravenously (30 and 60 mg/kg), proposing that some intact DPDPE was entering the CNS. This work also supports the observation that DPDPE can cross confluent monolayers of cerebral capillary endothelial cells shown in Chapter 2 of this dissertation. Thus, it is likely that the lack of antinociception observed by Shook et al. (1987) may be due to a decreased plasma concentration of DPDPE due to differences in bioavailability firom the different routes of administration of drug. In addition, these studies have shown that [^H]DPDPE can enter the brain by both saturable and non-saturable uptake mechanisms that can be described by

Michaelis-Menten type kinetics with a of 46 ± 28 uM, Vmuc of 51.13 ± 13.23 pmol.min'

'.g"' and Kd of 0.56 ± 0.26 «l.min*'.g"'. The CSF uptake of [^H]DPDPE was also shown to not be self-inhibited (BCj 0.89 ± 0.07) thus it enters the CSF purely through diflfiision. 106

Chapter 4. Blood-to-CNS entry and stability of biphalin, a unique double- enkephalin analog, and its halogenated derivatives

Introduction

The entry of molecules into the central nervous system (CNS) is dependent on their size, charge, hydrophobicity and/or ability to utilize carriers present at the blood- brain and blood-cerebrospinal fluid (CSF) barriers. Specific peptides have been shown to enter the CNS using both saturable and non-saturable mechanisms and several reviews have been published (Audus et al., 1992; Banks et al., 1992; Brownlees and Williams,

1993; Ermisch et al., 1993; Poduslo et al., 1994). Some peptides, for example the enkephalins, are somewhat limited in there access to the CNS due to the presence of peptidases in the blood, at the blood-CNS interface or within the endothelial and/or epithelial cytosolic compartments (Pardridge and Mietus, 1981; Masuzawa and Sato,

1983; Thompson and Audus, 1994). Another important consideration affecting the CNS distribution of peptides is the presence of a brain-to-blood efflux system for N-tyrosinated peptides such as methionine- and leucine- enkephalin (Banks and Kastin, 1986). This overall lack of bioavailability, resulting from enzymatic degradation and the presence of a saturable brain-to-blood efflux system, is a limitation in the use of peptides as successful therapies in a wide variety of clinical situations. 107

With the identification of multiple opioid receptors (Gilbert and Martin, 1976;

Lord et al., 1977) and the discovery of their endogenous ligands (Hughes et al., 1975), many peptide analogs have been developed, with the potential to be used clinically to produce favorable pharmacological effects such as analgesia (Mosberg et al., 1983; Clark et al., 1986; Toth et al., 1990). Our drug development program at the University of

Arizona has concentrated on analogs that have greater blood-to-CNS distribution and biological stability, as well as potency and selectivity for the opioid receptor types {mit, delta, kappa). [D-Pen^, D-Pen^]enkephalin (DPDPE) is a delta opioid selective, conformationally constrained and enzymatically stable analog that has been fiilly characterized, in terms of blood-to-CNS pharmacokinetics, by our laboratory (Weber et al., 1991; 1992). It is assumed that DPDPE has one active pharmacophore that binds to the delta receptor (Knapp et al., 1992).

In contrast, biphalin is a unique analog that has two enkephalin sequences linked by a hydrazide bridge giving the summary formula:

TyrI-D-Ala2-Gly3-Phe4-NH I 1 1 Tyrl'-D-Ala^'-Gly^'-Phe'^'-NH

The N-terminal tyrosine of opioid peptides has been shown to be critical for successful binding to the opioid receptor and therefore essential for the compounds pharmacological activity (Goodman and Stueben, 1953). Thus, biphalin has two biologically active lOS pharmacophores, which can bind to either mu or delta opioid receptor types (Lipkowski et al., 1982; 1987; Horan et al., 1993). In addition, incorporation of the D-Ala in the two position has been shown to enhance activity and maintain selectivity of enkephalin analogs for the opioid receptors (Lipkowski et al., 1982; Dooley et al., 1994). Furthermore, the

C-terminus of biphalin is protected from enzymatic hydrolysis by the hydrazide bridge

(Lipkowski et al., 1982). However, in the event of enzymatic degradation, it is possible to have one remaining enkephalin fragment that may still bind to the opioid receptor.

Biphalin has been shown to be an extremely potent analgesic; when administered intracerebroventricularly, it was shown to be 6.7- and 257- fold more potent than and morphine, respectively, in eliciting antinociception (Horan et al., 1993). In addition, intrathecal biphalin has been shown to be more potent than morphine (Silbert et al., 1991). However, after intraperitoneal and subcutaneous administration, it would appear that only a small fraction of biphalin enters the brain, since reduced analgesia was elicited (Horan et al., 1993).

Halogenation of enkephalin analogs has been shown to increase the brain uptake after systemic administration (Weber et al., 1991) and the specificity for delta-opioid receptors (Toth et al., 1990). Thus, in order to improve CNS entry of biphalin, our research group synthesized analogs, which had chloro- or fluoro- halogens on the para position of phenylalanine-4,4' residues. Since CNS penetration and biological stability appear to be deciding factors for the potency of biphalin after systemic administration 109

(Horan et al., 1993), the aim of this present study was to characterize and compare the

CNS uptake and stability of biphalin, [p-Cl-Phe'^'^']biphaIin and [p-F-Phe^'^']biphaIin.

The initial screening examined the passage of these opioid analogs across a well characterized in vitro blood-brain barrier (BBB) model, which employs primary cultured bovine brain microvessel endothelial cells (BMEC) (Audus and Borchardt, 1986; 1987).

The CNS uptake and stability of biphalin and [p-CI-Phe^'^']biphalin was further examined using an in situ brain perfusion technique followed by HPLC analysis (Zlokovic et al.,

1986; Williams et al., 1993; Preston et al., 1995). In addition, investigations were made as to the importance of the physicochemical characteristic of lipid solubility to CNS entry of biphalin. 110

Methods

In Vitro BMEC studies.

Fresh bovine brains were obtained from a local slaughterhouse. Brain microvessel endothelial cells were then isolated from the gray matter of the cerebral cortex and cryo- preserved, as previously described (Audus and Borchardt, 1986; Audus and Borchardt,

1987). The isolated cells were seeded at a cell density of 50,000 cells/cm^ onto tissue culture dishes, which had been pre-coated with rat tail collagen and Sbronectin and contained 25-mm Costar Nucleopore polycarbonate membrane filters (10 //m pore size;

Costar Corp., Cambridge, MA).

After formation of confluent monolayers (10-12 days) the BMECs were used for transendothelial transport studies. The polycarbonate membranes plus the confluent endothelial cells were placed in a glass Side-Bi-Side diffusion cell (Crown Glass Co.,

Somerville, NJ) that was maintained at 37°C. Both sides of the assay chamber contained the physiologically balanced phosphate-buffered saline (PBS) (122 mM NaCl, 3 mM KCl,

25 mM Na2P04, 1.3 mM K2HPO4, 1.4 mM CaCl2, 1.2 mM MgS04, 10 mM glucose and 10 mM HEPES, pH 7.4), which was continuously stirred. The test peptide (500 MM), together with the membrane impermeant marker, [^^C]sucrose, were added to the donor chamber at time 0. After the set time points (0, 15, 30, 60, 90 and 120 min), 200 ii\ aliquots were removed from the receiver chamber. An equal volume of 50/50 Ill acetonitrile/water was then added to each sample, which were then stored at -37°C for later HPLC analysis and radioactive scintillation counting.

Permeability coefficients (PC) were calculated on the basis of diffiision of the peptides across the BMEC by means of the following equation:

PC = X/(A X t X CD)

where PC is measured in cm/min, X is the amount of substance in moles in the receptor chamber after correction for sampling error and paracellular passage based on sucrose levels at time, t, in minutes, A is the diffusion area (0.636 cm-) and C[) is the concentration of substance in the donor chamber in mol.cm"^ (CQ remains > 90% of the initial value over the time course of the experiments).

In Silii Brain Peijhision Studies.

The protocol described below was approved by the Institutional Animal Care and

Use committee (lACUC) at the University of Arizona. Adult Sprague-Dawley rats (250-

300g) were anaesthetized with sodium pentobarbitol (64.8 mg.kg"^) and heparinized

(10,000 U.kg"^). The jugular veins were located and the right common carotid artery was cannulated with fine silicone tubing that was connected to a perfijsion system. Perfusion was performed with a mammalian Ringer as previously described (Williams et al., 1995a).

The perfijsion Ringer was warmed to 37°C and thoroughly oxygenated with 95% O2 and 112

5% C02. In full agreement with Preston et al. (1995), the p02 values ranged from 642-

727 mm Hg. At the start of perfusion, the right jugular vein was cut to allow drainage of the perfusion Ringer Once the desired perfusion pressure and rate were achieved

(approximately 100 mmHg and 3.1 mls.min"^, respectively), the contralateral carotid was cannulated and perfused in the same manner as described above. The remaining left jugular vein was then cut. Either [l^C]sucrose, or [I25n[p-Cl-

Phe^»^ Jbiphalin were infused, using a slow-^rive syringe pump (Model 22: Harvard

Apparatus, South Natick, MA), into the inflowing mammalian Ringer. Once the set perfusion time was reached (2.5-20 min), a cistema magna CSF sample (50 i/1) was taken with a glass cannulae. The animal was then decapitated and the brain was removed. The choroid plexuses were excised and the brain was homogenized. The perfusion Ringer containing the radiolabeled test solute was collected from the carotid cannulae at the end of each time point to serve as a reference. The brain and CSF samples were then weighed and prepared for radioactive counting. The amount of radioactivity was determined by counting on a Beckman 5500 gamma counter or a Beckman LS 5000 TD beta counter

(Beckman Instruments, Inc., Fullerton, CA).

Capillary Depletion.

Measurement of the vascular component to total brain uptake was performed using a capillary depletion step as previously described (Zlokovic et al., 1992). Briefly, the brain was removed and choroid plexuses excised. The brain tissue (500 mg) was homogenized 113

(Polytron homogenizer, Brinkmann Instruments, Westbury NY) in 1.5 mis of physiological buffer kept on ice (10 mM HEPES, 141 mM NaCl, 4 mM KCl, 2.8 mM

CaCl2, I mM MgS04, 1 mM NaH2P04 and 10 mM D-glucose, pH 7.4) Two ml of ice- cold 26% dextran (MW 60,000) were then added and homogenization was performed again. Two aliquots of homogenate were taken and centrifiiged at 5,400 x g for 15 min in a microflige (Beckman Instruments Inc.). The capillary-depleted supernatant was then separated from the vascular pellet. All the above homogenization procedures were performed within 2 min. The homogenate, supernatant, and pellet were then taken for radioactive counting (Beckman 5500 gamma counter).

Expression of Results.

The amount of [^^^I]biphalin, [^25ij[-p.d_phg4 4'jijiphalin and [^^Cjsucrose in the whole brain, CSF, homogenate, supernatant and pellet was expressed as the percentage ratio of tissue (Cxissue dpm g"^ or dpm.ml"^) to plasma activities (Cpjasma

(dpm.ml"^) and expressed as RTissue o*" ml.ml"^).

^Tissue ^ ~ ^Tissue^^Plasma 100

The unidirectional transfer constant, (Kj^) and the initial volume of distribution

(V;) were graphically determined from the multiple-time uptake data (2.5-20 min) as previously described by Zlokovic et al. (1986). lU

^ Tissue C')! ^ Plasma (^) ~ (^in) (^) ^i

Where C Tissue C'^) ^ Plasma C^) radioactivities per unit weight of tissue and plasma at time T (perfiision in minutes). The above equation describes a straight line, where the slope is Kfn (ml.min'l.g"^) and y-intercept is Vj (ml.g"^). Any brain to blood movement of the test compound can be observed as a departure from linearity of the experimental points. To determine blood-to-CSF transfer constants a two compartment/single-time uptake analysis was used. This can be performed by using the following equation:

Kin = RCSF%' T

Blood-to-brain unidirectional transfer constants were also determined in this manner. The vascular space was corrected for by subtracting [^^C]sucrose (RBrain) ^^e test drug.

Extraction of Radiolabeled Peptide.

Brain extractions were performed using a modified method of Erchegyi et al.

(1991). Briefly, rats were perfused with either a syringe containing or

[125j][p_d-Phe4>4']biphaIin in perfusion medium as described previously. At the end of a

20 min perfusion period the animal was decapitated and the brain removed and 115 immediately placed in 7.5 ml ice-cold 10% trifluoroacetic acid (TFA). Each sample was then homogenized (Polytron homogenizer) and centrifiiged at 20000 x g for 20 min. The supematants were collected and an equal volume of ether was added. The ether phase was discarded and the samples were then lyophilized to dryness. The samples were then diluted to 500 |il with 10% acetonitrile and taken for HPLC analysis. Preliminary experiments ensured that both [^^^IJbiphalin and [p-CI-Phe^''^ ]biphalin were stable in 10% TFA.

In Vitro Brain Stability Studies.

Mouse brain homogenates were prepared by a modified method of Davis and

Culling-Berglund (1985). The protein concentration was determined to be 6.8 mg.ml"^ by method of Lowry et al. (1951). Aliquots (180 MI) of re-suspended, twice-washed 15% mouse brain homogenate were placed into 1.5 ml centrifiige tubes and, together with a buffer control, warmed to 37°C In a rolling water bath incubator. At time 0, biphalin and

[p-CI-Phe^'^']biphalin were added to each tube to achieve a final concentration of 100 LIM and incubated for 0, 30, 60, 120, 240 and 360 min. At the end of the set incubation period, enzyme activity was terminated by the addition of 200 z/1 of acetonitrile and the tubes were on ice. Each tube was then centrifuged at 3000 x g, and 300 wl of the supernatant was transferred to a clean 1.5 ml conical tube. An equal volume of water was added to reduce the final acetonitrile concentration to 25%, and the sample was taken for

HPLC analysis as described below. 116

lodination of Biphalin.

Biphalin and [p-Cl-Phe'^'^']biphalin were prepared by methods similar to those

previously reported (Misicka et al., 1997). Biphalin and [p-Cl-Phe^'^']biphaIin were mono-iodinated on the tyrosine^ residue using a standard chloramine-T procedure as previously described (Bolton, 1986). Purification of iodinated biphalin and [p-Cl-

Phe^'^']biphalin was performed using the HPLC system described below. The purity of iodinated biphalin and [P-Cl-Phe^'^']biphalin was > 99% pure and eluted at approximately

2.5 min after the cold peptide standards. The specific activity was 2200 Ci/mmol in both cases.

HPLC Analysis.

In vitro BMEC samples were analyzed on a reverse-phase HPLC system consisting of a Waters Associates WISP 71OB Autoinjector, 2 Model 6000A Solvent Delivery

Pumps, Automated Gradient Controller (Waters Associates, Milford, MA), Perkin Elmer

LC-65T Detector/Oven (210 nm; Perkin Elmer; Norwalk, CT), Hewlett-Packard 3390A

Integrator (Hewlett-Packard Co., Avondale, PA) and a Vydac 218TP54 column (4.6 x

250 mm; Vydac, Hesperia, CA). Samples were eluted using a linear gradient of acetonitrile against 0.1 M NaH2P04 buffer (pH 2.4). The flow rate was 1.5 ml/min and the column temperature was maintained at 37°C. 117

The stability of the peptide analogs and the integrity of the labels to biphalin and [p-CI-Phe^'^']biphaiin in the in situ brain perfusion Ringer, before and after it had passed through the cerebral circulation, were analyzed using a Series 410 HPLC gradient system (Perkin-Elmer). Samples were eluted from an Inertsil ODS-2 |.i column ( 4.6 x 150 mm; Metachem Technologies Inc., Torrance, CA) with a curvilinear gradient of 0.1%

TFA in acetonitrile 20-50% vs. 0.1% aqueous TFA in 30 min at 1.5 ml/min and the column temperature was maintained at 37°C. After separation on the HPLC column the outflow was routed to the on-line A200 Flo-One Radioactive Detector equipped with a

2.5 ml flow cell (Packard Radiomatic Instruments & Chemicals, Meriden, CT).

Brain extractions of [l-^I]biphalin and [^^-I] [p-Cl-Phe^''^']biphalin were analyzed the same as above, except separation was performed on a Beckman Ultrasphere column (4.6 x 250 mm; Beckman Instruments Inc.).

The peptide stability studies in brain homogenate were analyzed using a reversed- phase Perkin Elmer 250 HPLC gradient system, a model 71 OB WISP autoinjector (Waters

Associates), a Perkin Elmer LC-15 UV absorbance detector, a Hewlett-Packard model

3396A integrator and a Vydac 218TP54 column (4.6 x 250 mm). Samples were eluted using a curvilinear gradient of 20-50% CH3CN vs. 0.1% NaH2P04 pH 7.4 in 30 min.

The flow rate was 1.5 ml/min and the column temperature was maintained at 37°C.

Purification of the iodinated biphalin analogs was performed on a ODS Cjg

Ultrasphere column (4.6 x 250 mm) with a linear gradient of 0.1% TFA in acetonitrile

(20-50%) vs. 0.1% aqueous TFA over 30 min at 1.5 ml/min and the column temperature 118 was maintained at 21°C. After separation on the column, the outflow from the HPLC was routed to an A200 Flo-One Radioactive Detector as described above.

OctanoUSaline Partition Coefficients.

Partition coefficients for [^25l]biphalin, [l-^I][p-Cl-Phe^'^']biphaiin and

[^'^C]sucrose were expressed as the ratio of labeled substance found in the octanol phase to that found in the aqueous phase. Triplicate determinations were made.

Data Analysis.

All experiments are expressed as means ± S.E. of the mean. Analysis of variance

(ANOVA) was used to compare the slopes, determined by least squares linear regression analysis of the multiple-time uptake data. Student's t-test was used for the comparison of the two means and statistical significance was taken as P < 0.05. 119

Results

In vitro BMEC studies.

The passage of the opioid peptide analogs and the membrane impermeant molecule, [^'^CJsucrose, were examined across the in vitro BBB and found to be linear with time. Biphalin, [p-Cl-Phe^'^']biphaIin and [p-F-Phe'^''^']biphaIin, all crossed the confluent BMEC monolayers to a significantly greater extent when compared to

[^^C]sucrose (P < 0.01; Table 4.1). In addition, [p-Cl-Phe'^»'^']biphalin had a permeability coefficient (cm/min x 10"^) of 92.00 ± 5.88, wliich was found to be significantly greater than that obtained for biphalin, 55.00 ± 4.98 (P < 0.01). Interestingly, as can also be seen in table 4.1, fluoro-halogenation caused a significant decrease in the permeability coefficient for biphalin ([p-F-Phe^'^']biphaIin, 23.21 ± 3.76 cm/min x lO'^; P < 0.01).

In situ brain perfusion studies.

Multiple-time uptake plots for [^^^IJbiphalin and [^-^I][p-Cl-Phe'^''^']biphalin into both the brain and CSF are shown in Figure 4.1. The brain uptake of both [^-^I]biphalin and [^2^I][p-Cl-Phe^'^']biphalin were significantly greater than the vascular space marker,

[^'^C]sucrose (P < 0.05 in both instances). In addition, the CSF uptake of [^25i]b,-phaiin and [^25[j[p_d.phe4.4']biphalin was significantly greater than [^^Cjsucrose (P < 0.05 and P < 0.01, respectively). However, once vascular space had been considered, the 120 uptake into the brain of [^-^qbiphalin was significantly smaller than that into the CSF (P

< 0.05). In contrast, brain uptake of [12^I][p-CI-Phe^'^']biphaIin was not significantly different when compared to CSF uptake. The uptake of [^-^I]biphalin into the brain was significantly smaller than [12^I][p-CI-Phe^»^']biphalin (P < 0.05), however there was no significant difference in the uptake into the CSF of these enkephalin analogs. Table 4.2 shows the unidirectional transfer constants (Kin) initial volume of distribution (Vj) for

[125qijiphalin, [^25i][^p.d_phe4.4']biphalin, and [I'^Cjsucrose which were determined from linear regression analysis of the experimental points in Figure 4.1.

In Table 4.2 it can be observed that both [^-^[]biphalin and ['^^I][p-Cl-

Phe'^''^']biphalin have higher unidirectional transfer constants (Kin) initial volumes of distribution (Vi) when compared to [^^C]sucrose. The Kin values for biphalin into the

CSF is approximately equal to its rate of uptake into the brain, whereas, [l-^I][p-Cl-

Phe'^''^']biphalin into the CSF makes up only 55% of its rate of uptake into the brain.

Another interesting observation is that the Vj for [^^5l][p-Cl-Phe'^''^']biphalin is 1.4-fold and 1.9-foId higher than tlie Vi for [^^^qbiphalin and [l^C]sucrose, respectively.

Capillary depletion.

Figure 4.2 shows the contribution of the vascular component to brain uptake for both [^^^Ijbiphalin and [^2^I][p-Cl-Phe^'^']biphalin. [^^^IJBiphalin showed no significant difference in the amount of radioactivity detected in the brain homogenate and in the capillary-depleted supernatant (Figure 4.2a). In addition, there was a limited 121 accumulation of [^-^I]biphalin into the pellet. In contrast, [^^5l][p-Cl-Phe*^''^']biphalin did accumulate in the pellet and contributed approximately 10% to the total brain uptake

(Figure 4.2b). Although, the level of [^-^I][p-Cl-Phe^'^']biphalin in the capillary-depleted supernatant was not significantly different to that found in the brain homogenate, it only contained 86% of the total radioactivity.

Arterial inflow and venous outflow HPLC analysis.

Figure 4.3 illustrates the purity and stability of the samples from the arterial inflow and venous outflow. It is apparent that for both [^-^Hbiphalin and [p-Cl-

Phe^'^']biphalin the arterial inflow and venous outflow eluted as a single peak that co- migrated with the radioactive standard. Also, [^-^I][p-Cl-Phe^'^']biphalin was intact in the arterial inflow, but there was a small peak eluting at 10.5 min in the venous outflow that made up approximately 12% of the total area counts.

Extract ion of radiolabeled peptides.

Figure 4.4 shows the stability of iodinated biphalin and [p-Cl-Phe'^>'^']biphalin in the brain after a 20 min vascular perfusion. A typical chromatogram of the TFA extracted

[^^^Ijbiphalin showed that approximately half of the radioactivity co-eluted with the iodinated biphalin standard, a portion eluted with the solvent front (21%) and a metabolite eluted at about 4 min (31%). When [^25ij^p.(3i_phg4,4'jbiphalin was analyzed the 122 majority was intact (67%), and in similarity with [l-5l]biphalin a metabolite peak eluted at around 4 min that made up 27% of the total area counts.

In vitro brain stability studies.

The enzymatic stability of both biphalin and [p-Cl-Phe^'^']biphaIin was assayed by incubation with twice washed 15% w/v brain membrane homogenates over a time course of 360 min. As shown in Figure 4.5, the metabolic half-lives (T1/2) of biphalin and [p-Cl-

Phe'^'^ jbiphalin were calculated to be 173 and 310 min, respectively.

OctanoUsaline partition coefficient.

The octanol/saline partition coefficient represents a measurement of lipophilicity.

[125ijBjphalin was calculated to be 0.2011 ± 0.0675 compared to a value of 0.7832 ±

0.0268 for [^25m-p.Q_pijg4,4'j{jiphaIjj^ hydrophilic control, sucrose, mainly partitioned in the saline phase yielding a partition coefiBcient of 0.0005 ± 0.000029. All values were significantly greater than [^^cjsucrose (P < 0.01). 123

TABLE 4.1. In vitro BBB permeability coefficients (PC) determined for biphalin, [p-Ci- Phe*'--*']biphalin, [p-F-Phe''-"*']biphaIin and ["C]sucrose. PC ± S.E.M., n = 3-8 monolayers. * P < 0.01; Students t-test was used to compare the PCs determined for ['''C]sucrose with those determined for the peptide analogs.

[^••QSucrose Biphalin [p-CI-Phe^''*T [p-F-Phe'^'^'] Biphalin Biphalin

PC (cm/min x 10.43 ± 0.10 55.00 ± 92.00 ± 23.21 ± 10-^) 4.98* 5.88' 3.76- Students t-test Comparison with P < 0.01 - P < 0.01 P < 0.01 Biphalin 124

12

11

10 9 8

7 6

5

4

3 2 1

0 to 20 Timo (min)

12 -] 11 -

10 -

9 -

0 -

7 -

^ 0 —

Fig 4.1 Multiple time uptake plots of ['^^Ilbiplialin (open boxes), ['25lI[p-Cl-Phc'''"''lhiplialin (closed boxes) and [MCJsucrosc (closed triangles) into the brain (A) and CSF (D) of the in siiit perfused rat. Uptake is expressed as the % ratio of tis.sue to pla.sma radioactivities (ml/g or niL'ml). Each point represents mean ± SEM of 3-6 rats. Kin and Vi values were detennined as tlie slope and ordinate intercept of the computed regression lines. 125

Table 4.2

TIic cntculaicil uniilircctiunul tmnsfcr conMnnts (K^,), initial volumes ofdutributian (V^) anil ccrcbro\a\ciilar permeability constant* (P) for llbiphalin. |'-'II(p-CI-Plic-<,4'|biphalin ami |"C|sucrosc. Values ^ S.E.M. were Jctcnnincii from ^thc computer ijcncratcd lines of regression in Figure I and ''siniile-time uptake analysis at a pcriiision time of 20 .iiinutes (brain xalue !S currceicd for vascular space). Permeability constants (P r S.EM.) were determined assumini; a cerebrovasci^lar "lurfacearca uf IDO cm- f' (Bradburs-. lOTM).

CO.MPOLNIl BItVIN C.SF Kin (ml.It)(lfr') (ul-min'.a') (cm.min-' x 10 ') (ul.min '.^')

|"C"lSurrii*c 1.41 ilU4 0J2 i (l.02» l)J2 i 0.112 0.1)7 i I).II2''

I'-Milliphalin 1.89 ± 1.13 :J4 ± O..S5=" 2J4 ± (l.S.'i 1.95 ±0.10'' :.2t) i U.4ii"

|'-"l|l|i-n-Phc''|Riphalin 2.67 i: 0.S4 JJ9 ± 0.62^ 3J9 i ().fi2 J.76 ± 0.4y'' 2.117 ± 0.49" 126

WMi

iloaoatiitf l»p Ptiltl

10

9

0

7

f. 0 C 5 ff mlM. tE5 4' 3

2 1

0 HomogKiti Sa^*r>il«nt Ptllil B

Fig 4.2 Contribution of tlic vascular component (pellet) to tlie brain uptake of ['^^Ilbiphalin (A) and p-Cl- Phc*'-"' [125ijbipiialin (B). Rbnin % represents tlic ratio of homogenate, supernatant or pellet to plasma radioactivities. Supernatant represents brain homogenate depleted of tlic ccrebral capillary endotheliimi. Perfusion time was 15 niin. Values arc mean ± SEM, n=3.4 rats. * pellet was found to be significantly smaller tliat homogenate and supernatant in botli cases (Students' t test; p < 0.01). 127

/lilCIKJl IlllkJ-.v AjIciM ti iriov.-

•A. •->'ri"'-iVf^

Voiicui Culllo-.Y Vf:nc»j3 Culilow

10 :o Lij :*) J') Hrno (ixwhjIc:} Tbiio

Fig 4.3 riPLC Flo-One Radioactive Detector chromatograms of ['-Sijbjphaiin (A) and of p-Cl-Phe"*'"' ['25ij5iphalin (D) from the arterial inflow and the venous outflow. Venous outflow samples were taken after a 20-min brain perfusion. Retention times are identical to standards. 128

?.2 BIphalin Slandard

0 X

1 4.6-. TJ

Biphalin Brain Lxlracl

10 20 30 A

Fig 4.4 HPLC Flo-One Radioactive Detector chromatograms of [^-^I]biphaiin TFA extract after a 20-minute vascular perfusion. Peaks coeluted with radioactive standards and radiolabeled metabolites were detected. 129

1.2 1 [P-Cl-Plie'']Diplia!in Standard

E u. 0.6- XJ

10 20 30

1.51 [P-CI-Phe'-'lDIphclin Drain E;

E XJu. 0.0-.

10 20 30 B

Fig 4.4 liPLC Flo-One Radioactive Detector chromatograms of p-Cl-PIie'*''' [^-^Ijbiphalin TFA extract after a 20-minute vascular perfusion. Peaks coeluted with radioactive standards and radiolabeled metabolites were detected. I3U

110

100 eo 60 70

60

50

40

30 ZO

10

100 200 300 400 Tim* (min)

Fig 4.5 The percent recovery of intact biphalin (closed trianges) and p-[Cl- Phe4,4']biphalin (open triangles) over a 360 minute time course in brain homogenate. Ti,7 represents the lialf-time disappearance of the enkephalin analogs determined by HPLC analysis. 131

Discussion

The aim of this present study was to investigate the CNS uptake characteristics and stability of biphalin and its halogenated derivatives, [p-Cl-Phe^'^']biphaIin and [p-F-

Phe^'^']biphalin. The initial investigation examined peptide passage across an in vitro

BBB and a permeability coefRcient (PC) was determined. As can be seen in Table 4.1, the peptide analogs can cross the BMEC monolayers to a significantly greater e.xtent compared to the membrane impermeant marker molecule, [^''^CJsucrose. These data provide evidence that biphalin and its halogenated derivatives have the ability to cross the in vitro BBB. Biphalin had a BMEC permeability coefficient (cm/min x 10"^) of 55.00 ±

4.98, which compares well to 49.24 ± 2.78 previously determined for the delta selective enkephalin analog, DPDPE as shown in chapter 2 of this dissertation. It is interesting to note that while the molecular weight of biphalin (MW 1137) is significantly larger than

DPDPE (MW 646), the permeability coefficients for these two enkephalin analogs are within a similar range. This observation may be due to the increased liphophilicity, which can be seen in the octanol/saline partition coefficients of biphalin (0.2011 ± 0.0675) compared to DPDPE (0.076 ± 0.002; Chapter 3 of this dissertation). As shown in Table

4.1, chloro-halogenation causes a 1.7- fold increase in the ability of biphalin to cross the in vitro BMEC. This may also be explained, at least in part, by the increase in lipophilicity associated with chloro-halogenation of the Phe residues of biphalin (octanol/saline partition coefficient 0.7832 ± 0.0268). A previous study has shown that lipophilicity plays a critical role in the BBB penetration of certain peptides (Banks and Kastin, 1985). 132

Table 4.1 shows that fluoro-halogenation causes a 2.4- and 4.0- fold decrease in

the in vitro permeability of biphalin and [p-Cl-Phe^''^']biphalin, respectively. This can be

correlated with the increase in electronegativity associated with the attachment of fluoro-

versus chloro-halogens in the para-Phe^'^' position (CI = 3.0; F = 4.0) and therefore a

decrease in lipophilicity. This has been previously observed for halogenated para-Phe^

analogs (Toth et al., 1990).

This initial screening with the BMEC model, thus enabled identification of [p-Cl-

Phe'^>^']biphalin as the enkephalin analog with the best potential for greater CNS entry

(Table 4.1). In order to further characterize the uptake of this halogenated analog, a comparison of the blood-to-CNS pharmacokinetics of biphalin and [p-CI-Phe^'^']biphalin

was investigated using an in situ brain perfusion technique in the anaesthetised rat.

Figure 4.1 illustrates multiple-time uptake plots for [^^^IJbiphalin, [^~^I][p-CI-

Phe^''^']biphalin and [I'^Clsucrose into the CNS. It is apparent that the brain uptake of

both enkephalin analogs is linear with time and is significantly greater than that for the

vascular space marker, [^^Cjsucrose (P < 0.05 in both instances). Figure 4.2 shows the

proportion of [^^Sij^jphalin and [125ijj^p_d.phe4,4'j5iphalin that has actually traversed

the cerebral capillary endothelium and entered the brain. In both cases, the radioactivity

detected in the capillary-depleted supematants was not significantly different to that found in the brain homogenates. In addition, the low levels of radioactivity detected in the

pellets indicate that both enkephalin analogs had predominately crossed the BBB and entered the brain. Together the in situ brain perfusion (Figure 4.1) and capillary depletion 133 experiments (Figure 4.2), indicate that both [^^Sjjbiphalin and [^-5l][p-Cl-

Phe^'^'jbiphalin can enter the brain and confirms the results obtained using the in vitro

BBB model (Table 4.1). Previously, brain distribution and antinociceptition studies have indicated that biphalin has a reduced analgesic effect after both subcutaneous and intraperitoneal administration, (Silbert et al., 1991; Horan et al., 1993). However, intravenous biphalin has been shown to elicit significant analgesia, and is in agreement with the in vitro and in situ results of the present study, also indicating that biphalin can enter the CNS from the blood (Silbert et al., 1991).

Examination of the CSF uptake of the enkephalin analogs revealed a significantly greater uptake when compared to [l^C]sucrose (Figure 4.1b). Since sucrose entry into the CSF is representative of paracellular difilision, these results suggest that both

[l-^I]biphalin and [^-^I][p-CI-Phe'^'^']biphalin can also cross the choroidal epithelium transcellularly.

Solutes can enter the CNS through either the blood-brain and/or blood-CSF barriers and it has been demonstrated that these barriers have differential permeability characteristics (Pardridge, 1995). After considering vascular space, the actual uptake of

[^^Sijbiphalin into the brain was significantly smaller (P < 0.05) than uptake into the CSF compartment (Figure 4.1). However, the uptake rate (Kj^) of [125];]|jiphalin into the

CSF, only represented 75% of the rate of uptake into the brain (Table 4.1). In whole animal studies it is not possible to completely separate entry across either the BBB or the blood-CSF barrier, but one can make certain assumptions. It is commonly accepted that 134 the total surface area of the choroid plexuses is smaller (approximately 5000 times) in comparison to the surface area of the cerebral capillary endothelium (Bradbury, 1979).

Also, the CSF is more likely to act as a sink to the brain than the brain to act as a sink to the CSF (Davson et al., 1961). Considering the above statements, while it appears likely that the blood-CSF barrier plays an important role in the bioavailability of [^^5l]biphalin to the brain, the BBB must provide a significant route.

In contrast to the brain and CSF uptake of ^he entry of

Phe^''^']biphalin into these two compartments was found to be not significantly different fi-om each other. This may be due to the increased brain bioavailability related to the improved lipophilicity with chloro-halogenation of biphalin, and would suggest that the

BBB provides a major route for the CNS entry of [12^I][p-Cl-Phe'^'^']biphaIin.

Comparisons of brain uptake between [^-^IJbiphalin and [^~^I][p-Cl-

Phe^''^']biphalin revealed that chloro-halogenation significantly (P < 0.05) increased entry of biphalin (Figure 4.1a). This is in fiill agreement with the results obtained from the in vitro BBB study (Table 4.1) and the octanol/saline partition coefficients. It is interesting to note that capillary depletion analysis of the in situ perfused brains showed a significant accumulation of [125i][p.d_phe4,4']biphalin compared to [^2^I]biphalin in the pellet

(Figure 4.2). This larger accumulation within the cerebral capillary endothelial cells

(pellet) is also likely to be related to the higher lipophilicity of [^^^I][p-Cl-

Phe^'4']biphalin. Another interesting observation is that the Vj for [125q[-p_(;^i.

Phe^'^']biphalin is 1.4-fold and 1.9-fold higher than the Vj for [^^^IJbiphalin and 135

[^"^Clsucrose, respectively. This also suggests that chloro-halogenation may enhance the ability of biphalin to distribute to lipophilic compartments of the body. However, when comparing the CSF uptake of and [^2^I][p-CI-Phe^'^']biphaIin, no significant differences were observed (Figure 4. lb). Thus, lipophilicity may not be the only physicochemical characteristic to consider when examining the entry of biphalin into the CSF. In fact, these results do suggest the presence of a satureable uptake system.

When interpreting blood-to-CNS pharmacokinetics, it is important to assess the stability of the test drug and the integrity of the radiolabel to the drug in the blood and the brain. Figure 4.3 illustrates chromatographic analysis of [^-^I]biphalin and [^-5l][p-Cl-

Phe^'^']biphalin in the perfusion solution, before (arterial inflow) and after (venous outflow) it has passed through the cerebral circulation. It is apparent that for both

[^-^I]biphalin and [^25jj[^p.d_phe4j4'jj,iphalin, greater than 88% of the radioactivity eluted as a single peak that co-migrated with the radioactive, pure drug standards. Most endogenous enkephalins and related peptides contain the N-terminal amino acids, Tyr^-

Gly^-Gly^-Phe^, which are subject to enzymatic hydrolysis mainly by aminopeptidases between residues 1-2 and neutral endopeptidase (NEP) and angiotensin-converting enzyme (ACE) between residues 3-4 (Malfi-oy et al., 1978; Erdos et al., 1978; Shibanoki et al., 1991). In the case of the double-enkephalin analogs, [^^Sqijiphajin and [^^^I][p-

Cl-Phe^'^']biphalin, enzymatic resistance has been conferred (Figure 16), because of the incorporation of D-alanine instead of L-glycine in position 2 and the incorporation of the hydrazide bridge (Lipkowski et al., 1982; Dooley et al., 1994). This has previously been 136

observed in serum studies, where biphalin had a metabolic half-life of 87 minutes (Horan

et al., 1991), which is significantly longer than that observed for the endogenous

enkephalins (i.e. Met-enkephalin ~ 2 minutes; Meek et al., 1977). It must also be

considered that any metabolic products detected in the venous outflow samples could actually be due to non-CNS tissues and vascular beds, since not all of the common carotid inflow goes to the brain. Thus the venous outflow chromatograms (Figure 3a and b) could possibly represent an over-estimation of CNS-specific metabolism. In addition.

Figure 4.3 demonstrates that the radiolabel remained attached to the tyrosine^ residue of both analogs in the perfusion medium.

Analysis of the test drug in the rat brain matrix offers evidence that intact peptide has actually entered the brain and is available to bind to the supraspinal opioid receptors and is, therefore, capable of eliciting the desired pharmacological response. HPLC chromatograms of the radiolabelled test solutes in rat brain homogenates after an in situ perfusion are shown in Figure 4.4. Figure 4.4a demonstrates that there is intact

[125jjijiphalin in the brain homogenate, after a 20 minute perfiision. The [^-^IJbiphalin, which co-eluted with the pure drug standard at 14.2 minutes, contributed approximately

50% to the total area counts. Some metabolism had occurred, with radiolabelled peaks eluting at 2.0 and 3.8 minute retention times. It is unclear where these metabolic products were produced, however one could hypothesize that tliis metabolism has actually occurred at the level of the cerebral capillary endothelium due to its high density of peptidases (Pardridge and Mietus, 1981; Baranczyk-Kuzma and Audus, 1987; Brownlees 137

and Williams, 1993; Brownson et al., 1994). Figure 4.5 indicates that [^^Sij^jphalin has

a metabolic half-life of- 173 minutes in brain homogenate, this value is similar to that of

Horan et al. (1991; 112 minutes). In conclusion, the results illustrated in Figures 4.3a,

4.4a and 4.5 all indicate that biphalin is resistant to enzymatic degradation.

[^25qf^p_d_p}ie4,4']Biphalin was found to remain predominately intact in the brain, ahhough, a radioactive peak was observed at 3.8 minutes, contributing to 27% of the total area counts (Figure 4.4b). In addition, this peak was smaller than that detected at

3.8 minutes for [l-^Ijbiphalin. In similarity with [l-^I]biphalin, it is possible that the

metabolism shown in Figure 4.4b occurs at the level of the cerebral endothelium. This chromatographic analysis further emphasizes the importance of chloro-halogenation as a

means to maximize biphalins entry into the CNS. In addition, these experiments confirm that both iodinated biphalin and [p-CI-Phe^'^']biphaiin can reach the CNS intact and therefore would be capable of eliciting a pharmacological response such as analgesia.

The metabolic profiles observed for [^-^I]biphalin and [^-^I][p-Cl-

Phe^'^'jbiphalin, suggest that chloro-halogenation may introduce a structural change in the conformation of biphalin that masks it fi"om potential enzymatic recognition. This conformational alteration prevents and reduces the metabolism as shown in Figure 4.4.

This is also suggested by the in vitro stability of these two compounds in brain homogenates (Figure 4.5).

Several studies have extensively correlated transport of solutes across in vitro and in vivo models of the BBB (Pardridge et al., 1990; Chikhale et al., 1994). Assuming a 138 cerebrovascular surface area of 100 cm^.g'^ of brain tissue, a permeability constant (P) can be calculated from the unidirectional transport constants (Kjji) in Table 4.2.

Comparisons between the in vitro (Table 4.1) and the in situ (Table 4.2) studies, show that there is a 235- and 271- fold diflference between the permeability constants determined for biphalin and [p-Cl-Phe^'^']biphalin, respectively. This is higher than the

150-fold difference between in vitro and in vivo studies observed by Pardridge et al.

(1990) and may be related to the different whole animal and culturing methods of the present study. Another factor influencing the ability to make in vitro and in vivo comparisons is the use of unlabeled drugs for in vitro BMEC experiments and the use of radio-iodinated drugs for in situ brain perfusion experiments. It has been shown previously that the presence of an group in the tyrosine^ position of [p-Cl-

Phe^]DPDPE decreased the CNS entry of this analog when compared to [^HJ[p-CI-

Phe^]DPDPE (Williams et al., 1995b). Thus, the actual brain and CSF uptake values for the in situ brain perfusion experiments shown in Figure 4.1 may be higher due to the necessity of radio-iodinating both biphalin and [p-Cl-Phe^'^'jbiphalin. It is important to note that conclusions determined for in situ experiments would not be changed because both analogs were iodinated.

New drugs that have greater stability, potency and selectivity for the opioid receptors are essential for the improvement of nociceptive pain management. It has been shown that intrathecal and intracerebroventricular biphalin are significantly more potent than morphine (Silbert et al., 1991; Horan et al., 1993). In contrast, intravenous biphalin 139 elicited significant analgesia, but it was less potent or equipotent to that observed with morphine (Silbert et al., 1991, Horan et al., 1993). Therefore CNS entry and biological stability are critical in the development of opioid drugs. This present study has demonstrated that biphalin and [p-CI-Phe^'^']biphaIin can significantly enter the CNS through both the blood-brain and blood-CSF barriers. Chloro-halogenation was shown to significantly improve CNS entry, as well as biological stability, and would suggest that incorporation of chloro-halogens at the p-Phe'^''^ positions of biphalin is a promising structural modification in the development of this opioid enkephalin drug for the treatment of pain. 140

Chapter 5. Blood-brain barrier permeability and bioavailability of a highly potent and mu-selective opioid receptor antagonist, CTAP: Comparison with morphine.^

Introduction

We have previously shown, in Chapter 3, that DPDPE a cyclized enkephalin analog, enters the brain via a saturable uptake mechanism. The saturable mechanism for

DPDPE brain entry was not identified. This chapter will attempt to determine if another structurally dissimilar, cyclic peptide enters the brain through a saturable uptake mechanism. These studies will investigate whether or not the cyclized structure is a requirement for saturable uptake of a peptide molecule across either the blood-brain or blood-CSF barriers.

Since the discovery of multiple types of opioid receptors (Martin et al., 1976; Lord et al 1977) attempts have been made to elucidate the physiological function of these receptors (mu, kappa and delta). The development of potent, specific antagonists and agonists is essential for clarification of the multiple biological effects thought to be mediated by each receptor. Several receptor agonists have been developed that are selective for the various receptor subtypes, for example, DPDPE (delta) (Mosberg et al.,

1983), DAMGO (mu) (Handa et al., 1981) and U50,488H (kappa) (VonVoigtlander et al.,

1983). These agonists have been extensively characterized pharmacologically and a few have been evaluated for their ability to enter the brain. For example, DPDPE, has been 141 shown to be both enzymatically stable (Weber et al., 1991, 1992) and able to enter the

CNS through a saturable mechanism at the BBB as described in Chapters 2 and 3 of this dissertation. Opioid antagonists that are commercially available have historically been modeled from alkaloid opioid agonists, i.e. naloxone and , both of which are not receptor selective. This paper will describe the blood-to-CNS pharmacokinetics of a cyclic, peptidergic analog of somatostatin, D-Phe-Cys-Tyr-D-Trp-Arg-Thr-Pen-Thr-NH2

(CTAP), which has been shown to be extremely potent and selective for mu opioid receptors (BCramer et al., 1989).

Somatostatin is a 28 amino acid regulatory, peptide hormone that has numerous effects within the central and peripheral nervous systems such as controlling growth hormone, insulin and glucagon release. It is also postulated that, following neurosecretion of somatostatin, there is a metabolic interaction that occurs with brain capillary endothelial cells (Pardridge et al., 1985).

Several analogs of somatostatin have been developed that may provide clinical intervention for the treatment of endocrine disturbance such as acromegaly, diabetes mellitus (Karashima et al., 1988) and peptic ulcer disease (Laslo et al., 1989) Additionally cancer treatment has been shown to be another important application for the use of somatostatin analogs (Shally et al., 1986). A recently developed somatostatin analog that shows promise for treating abnormal hormone secretion by cancerous tumors is

Sandostatin™, D-Phe-Cys-Phe-D-Trp-Lys-Thr-Cys-Thr-OH. (Lamberts 1986, 1987).

Another analog, RC-121 (D-Phe-Cys-Tyr-D-Trp-Lys-Va!-Cys-Thr-NH2), has been shown 142

to be about 100 times more potent than somatostatin 1-14 in the inhibition of growth

hormone release but less than 5 times more potent in the inhibition of gastric acid release

(Cai et al., 1986, 1987).

Several years ago, somatostatin 1-14 was shown to display affinity for opioid

receptors, despite the apparent lack of structural similarity to endogenous opioid peptides or opiate alkaloids (Terem'us, 1976). Thus, interest within our research group focused on the development of opioid-receptor selective and enzymaticaily stable somatostatin analogs that could be used to characterize opioid receptors. Additionally, mu selective antagonists that can reverse the unwanted side effects of mu receptor-activated analgesia often seen with morphine and heroin: such as respiratory depression, convulsions, nausea, vomiting, decreased gastrointestinal motility, changes in mood, alterations in endocrine and autonomic nervous systems, tolerance and physical dependence are needed (Pasternak

1993) . The mu receptor has often been cited to play a vital role in the expression of central opiate dependence and the delta and kappa receptor appear to play a minor role

(Maldonado et al., 1992)

CTAP , CTOP and CTP (D-Phe-Cys-Tyr-D-Trp-Lys-Thr-Pen-Thr-NHo), were a series of conformationally constrained, penicillamine-containing octapeptides synthesized by Pelton et al., (1985; 1986). CTAP, CTOP and CTP are conformationally constrained peptides because they contain a disulfide linkage between the cysteine and the penicillamine which provides a useful approach to improving selectivity of flexible peptides (Kazmierski et al., 1988). This synthesis approach eliminates the low energy 143

conformations of the peptide and provides insight into the topological features that are

required for high affinity binding to a specific opioid receptor subtype. Another

advantage exists in that there is an elimination of activity at the natural receptor for the

peptide, (i.e. somatostatin receptor). CTAP was shown to display greater antagonist

potency and selectivity for mu opioid receptors when compared to the classical mu-

selective antagonist, CTOP (Kramer, et al., 1989). CTAP is 1200-fold more selective for

the mu vs. delta receptor binding sites and over 4000-fold selective for mu opioid receptor

binding vs. somatostatin binding in the rat brain (Pelton et al., 1986). CTAP has also been

shown to reduce the morphine tolerant state (antinociception) in mice and block the mu

receptor without causing severe withdrawal as measured by withdrawal jumping in

morphine-dependent mice (Wang et al., 1994). Furthermore, CTAP is a neutral antagonist, showing low intrinsic activity, and has considerable potential for the clinical

treatment of overdose, particularly in addicts where naloxone precipitates immediate withdrawal (Wang et al., 1994). In a model of acute morphine tolerance in

mice, CTAP has been shown to block the effects of both morphine and naloxone without

any effect on the mu receptor alone (Maldonado et al., 1992). This is advantageous since naloxone has been shown to elicit agonist-like effects at high doses (Nestler, E.J.,

1993; Crain et al., 1992). Based on the pharmacological profile, CTAP may be a promising and selective antagonist that can be used for both opiate overdose, addiction and additionally as a pharmacological tool. 144

Central nervous system (CNS) penetration and biological stability are deciding

factors for the clinical eflScacy of CTAP. The aim of this study was to characterize the

blood-to-CNS pharmacokinetics and biological stability of CTAP since only central

routes, i.e., (i.e.v.), have been examined for the related analogue CTP (Shook et al.,

1987). In the present study CNS entry of [^H]CTAP was compared to [^H]morphine, the

classical mu receptor agonist, and the vascular space marker [^^C]inulin. CNS uptake and

stability studies were also performed using a well characterized in situ brain perfusion

technique coupled to HPLC analysis (Takasato et al., 1986). Comparisons were made

between the brain and cerebrospinal fluid (CSF) uptake of [^HjCTAP, [^^CJinulin and

[^H]morphine after a 20 minute perfusion. The existence of saturable uptake mechanisms controlling the CNS entry of [^H]CTAP was also investigated.

If CTAP is able to cross the blood-brain and / or blood-CSF barriers then it may

provide a useful means to treat narcotic drug overdose and addiction in the clinic without

the unwanted precipitated withdrawal symptoms seen with the use of naloxone. CTAP

could also be used as a pharmacological tool with systemic administration for ftirther

understanding of opioid neurobiology or the potency of novel opioid agonists as shown in

Chapter 6 of this dissertation. 145

Methods

Supplies and Chemicals.

CTAP, [^HjCTAP (22.5 Ci/mmole) and [^Hjmorphine (50 mCi/mmoIe) were generous gifts from NIDA. [^^C]lnulin (2.7 mCi/gram) was purchased fr^om Dupont New

England Nuclear, Boston, MA.

In Situ Brain Perfusion Studies.

The protocol described below was approved by the Institutional Animal Care and

Use committee (lACUC) at the University of Arizona. Adult Sprague-Dawley rats (250-

300g) were anaesthetized with sodium pentobarbital (64.8 mg.kg'^) and heparinized

(10,000 U.kg"^). The jugular veins were located and the common carotid arteries were cannulated using fine silicone tubing connected to a perfusion system as previously described (Takasoto et al., 1986)

Perfusion was performed with a thoroughly oxygenated (p02 = 642-727 nunHg) mammalian Ringer (37°C) solution. Once the desired perfusion pressure and rate were achieved (approximately 100 mmHg and 3.1 mls.min"! respectively), the right jugular vein was cut and allowed to drain. The contralateral carotid was cannulated and perfused in a similar manner as described above. [^HjCTAP (M.W.= 1107) in the presence or absence of 100 loM CTAP, [^HJmorphine (M.W.= 758.8) or [l^qinulin (M.W.= 5000-5500) 146 were infused using a slow-drive syringe pump into the inflowing, mammalian Ringer.

Once the set perfusion time (2.5, 10, 15 or 20 min) was achieved, a cistema magna cerebrospinal fluid (CSF) sample was taken with a glass carmulae. The animal was then decapitated and the brain was removed. The choroid plexuses were excised and a portion of the brain was homogenized in 26% dextran and a capillary depletion buffer. The perfusion outflow was collected from the carotid cannulae at the end of the time point to serve as a reference. The brain and CSF samples were then weighed and prepared for liquid scintillation counting on a model LS 5000 TD beta counter; 43% efficiency for and 93% efficiency for (Beckman Instruments, Fullerton CA).

Capillary Depletion.

Measurement of the vascular contribution to total brain uptake was performed using a capillary depletion step as previously described (Zlokovic et al., 1992). Briefly, the brain was removed and choroid plexuses excised. The brain tissue (500 mg) was homogenized (Polytron homogenizer, Brinkmann Instruments, Westbury NY) in 1.5 mis of physiological buffer kept on ice (10 mM HEPES, 141 mM NaCl, 4 mM KCl, 2.8 mM

CaCl2, 1 itiM MgS04, 1 mM NaH2P04 and 10 mM D-glucose, pH 7.4) Two ml of ice- cold 26% dextran (MW 60,000) were then added and homogenization was performed again. Two aliquots of homogenate were taken and centrifuged at 5,400 x g for 15 min in a microfuge (Beckman Instruments Inc.). The capillary-depleted supernatant was then separated from the vascular pellet. All the above homogenization procedures were 147 performed within 2 min. The homogenate, supernatant, and pellet were then aliquoted for radioactive counting (Beckman 5500 beta counter).

Expression of Results.

The amount of radioactivity in the whole brain, CSF, homogenate, supernatant and pellet was expressed as the percentage ratio of tissue (Cjissue dpm.g"l or dpm.ml'^) to that in the perfusion fluid (Cpejf dpm.ml"^) and expressed as Rjissue or ml.ml'l).

^Tissue ~ ^Tissue^Cperf ^

The unidirectional transfer constant, (Kj^) and the initial volume of distribution

(Vj) were graphically determined from the multiple-time uptake data (2.5-20 min) using the following equation (Zlokovic et al. 1986).

C Tissue (T) / C Perf(T) = (Kin) (T) + Vj

Where C Tissue ^ Plasma radioactivities per unit weight of tissue and perfusion fluid at time T. The above equation describes a straight line, where the slope is

Kjn (ml.min'^g"^) and y-ordinate is Vj (ml.g"^). Any brain-to-blood movement of the test compound can be observed as a departure from linearity of the experimental points. us

To determine blood-to-CSF transfer constants a two compartment/single-time uptake analysis was used. This can be performed by using the following equation:

Kin = RCSF / T

Blood-to-brain unidirectional transfer constants were also determined by single-time uptake analysis. The vascular space was corrected for by subtracting [^^C]inulin (RBrain) from the test drug at the same time point.

Extraction of Radiolabeled Peptide. Brain extractions were performed using a modified method of Erchegyi et al. (1991). Briefly, rats were perfused with [^H]CTAP in as described previously. At the end of a 20 min perfusion period, the animal was perfiised with mammalian Ringer for 2 minutes to remove any remaining [^H]CTAP from the cerebral vasculature. The animal was decapitated and the brain removed and immediately placed in 7.5 ml ice-cold 10% trifluoroacetic acid (TFA). Each sample was then homogenized (Polytron homogenizer) and centrifiiged at 20000 x g for 20 min. The supematants were collected and an equal volume of ether was added. The ether phase was discarded and the remaining sample were then lyophilized to dryness. The samples were then diluted to 500 jil with 10% acetonitrile and stored for HPLC analysis.

In Vitro Brain Stability Studies. 149

Mouse brain homogenates were prepared by a modified method of Davis and

Culling-Berglund (1985). The protein concentration was determined to be 6.8 mg.ml"! by method of Lowry et al. (1951). Aliquots (180 mI) of re-suspended, twice-washed 15% rat brain homogenate were placed into 1.5 ml centrifuge tubes and, together with a buffer control, wanned to 37°C in a rolling water bath incubator. At time 0, CTAP was added to each tube to achieve a final concentration of 100 [oM and incubated for 0, 30, 60, 120, 240 and 360 min. At the end of the set incubation period, enzyme activity was terminated by the addition of 200 ul of acetonitrile with 0.5% acetic acid and the tubes were placed on ice. Each tube was then centrifiiged at 3000 x g, and 300 h1 of the supernatant was transferred to a clean 1.5 ml conical tube. An equal volume of water was added to reduce the final acetonitrile concentration to 25% and the sample was stored for HPLC analysis.

HPLC Analysis.

Brain extractions of [^H]CTAP were analyzed using a Series 410 HPLC gradient system (Perkin-Elmer, Norwalk, CT). Samples were eluted from an Inertsil ODS-2 column ( 4.6 x 150 mm; Metachem Technologies Inc., Torrance, CA) with a curvilinear gradient of 0.1% TFA in acetonitrile 20-50% vs. 0.1% aqueous TFA in 30 min at 1.5 ml/min and the column temperature was maintained at 37°C. After separation on the

HPLC column the outflow was routed to the on-line A200 Flo-One Radioactive Detector equipped with a 2.5 ml flow cell (Packard Radiomatic Instruments & Chemicals, Tampa

Bay, PL). 150

Peptide stability studies in rat brain homogenate and serum were analyzed using a reversed-phase Perkin Elmer 250 HPLC gradient system, a model 71OB WISP autoinjector (Waters Associates), a Perkin Elmer LC-15 UV absorbance detector, a

Hewlett-Packard model 3396A integrator and a Vydac 218TP54 column (4.6 x 250 mm).

Samples were eluted using a curvilinear gradient of 20-50% CH3CN vs. 0.1% NaH2P04 pH 7.4 in 30 min. The flow rate was 1.5 ml/min and the column temperature was O maintained at 37 C.

Protein Binding Studies.

The amount of [^H]CTAP binding to either bovine albumin in the perfusion medium or proteins in rat serum was determined by ultrafiltration centrifugal dialysis

(Paulus H., 1969). Rat serum was obtained by harvesting blood from Sprague-Dawley rats and allowing the blood to clot for 30 minutes on ice and 30 minutes at room temperature. The whole blood was then centrifliged (Sorvall RC2-B centrifuge; Dupont

Medical Products, Wilmington, DE) at 20,000 x g for 20 minutes to produce a serum supernatant. [3H]CTAP was dissolved in either perfusion medium or rat serum warmed to 370c and ultrafiltrated using a Centrifree^" micropartition device (Amicon, Beverly,

MA). The total concentration (T) of [^H]CTAP introduced into the system and found in the ultrafiltrate (F) was determined by liquid scintillation counting (Beckman 5500). The percentage of [^H]CTAP bound to either albumin in the perfusion medium or proteins in the rat serum was expressed as [(T-F) / T ] x 100. To verify that bovine albumin was not 151 found in the ultrafiltrate, the protein concentration was determined by the method of

Lowry et al. (1951).

Data Analysis.

All experiments were expressed as means ± S.E. of the mean. Analysis of variance

(ANOVA) was used to compare the slopes, determined by least squares linear regression analysis of the multiple-time uptake data. Student's t-test was used for the comparison of the two means and statistical significance was taken as ** P < 0.01 or * P < 0.05. 152

Results

In situ brain perfusion experiments.

Multiple time analysis was performed for both [^HjCTAP and [l^^CJinuiin into the brain and CSF. Uptake was expressed as Rxjssue^hich is the percent ratio of tissue to

plasma radioactivities (ml.g"^ or ml.ml"^). As shown in Figure 5.1a and 5.1b, the uptake of [^H]CTAP and [^"^Clinulin into the brain and CSF was linear with time. These results indicate that the brain uptake of [^H]CTAP was statistically greater than that of the vascular space marker [^^Cjinulin (P < 0.05). After considering the vascular space, the brain and CSF uptake values of [^HJCTAP were not statistically different.

Table 5.1 shows that the unidirectional transfer constant of [^H]CTAP into the brain and CSF was 5.96 and 2.43 fold higher than that calculated for [^^C]inulin. Also,

the initial volume of distribution into the brain for [-^HJCTAP was 1.62 fold higher than

that determined for [l^C]inulin.

Extraction of [^HJCTAP.

After a 20 minute vascular brain perfusion the majority (62.8%) of the [^H]CTAP co-eluted with the radioactive standard (Figure 5.2). Five metabolites were also observed

that comprised 37.2% of the total area counts. 153

In vitro brain and serum stability studies with CTAP.

The percent recovery of intact CTAP incubated for 240 minutes in 15% twice washed brain membranes or 100% plasma was determined using HPLC analysis. The

T1/2 of CTAP was > 500 minutes in both the brain and serum as determined by HPLC analysis (Figure 5.3).

Inhibition experiments with 100 fjA/I CTAP.

Entry into the brain and CSF was not statistically different after a 20 minute brain perfusion with [^H]CTAP in the presence and absence of 100 j.im CTAP (Figure 5.4).

Thus the entry into the brain of [^H]CTAP was not inhibited by the addition of cold

CTAP to the perfusion medium.

Protein binding studies with [^HJCTAP.

[3H]CTAP was found to be bound to protein in botli the perfusion medium

(68.2%) and rat serum (84.2%) (Table 5.2). No protein was detected in the ultrafiltrate after performing a Lowry protein assay.

Capillary depletion analysis.

The vascular component of the brain uptake of [^HJCTAP and [^H]morphine, 44

% and 32% respectively, contributed extensively to overall brain uptake, (Figure 5.5).

The homogenate and the supernatant were not statistically different in both cases. The 154 counts detected in the pellet were found to be significantly smaller than counts detected in the homogenate for both [^HJCTAP and [^H]morphine (P < 0.05).

^Tissue Percent and octanol / saline partition coefficients determined for [^HJCTAP,

[^HJmorphine and [^"^CJimdin.

Table 5.3 shows that a significantly greater amount of [^H]CTAP and

[^H]morphine entered the brain when compared to [^^Cjinulin after a 20 minute vascular brain perfusion (P < 0.01). In addition, a significantly greater amount of [^H]morphine entered the CSF when compared to [^^Cjinulin (P < 0.05). Octanol / saline partition coefficients for [^HJCTAP and [^H]morphine were higher and statistically different when compared to [I'^CJinulin (P < 0.01). Furthermore, the Rxissue values correlate well (R =

0.946) with the octanol / saline partition coefficients for [^H]CTAP, [^H]morphinc and

[^^C]inulin. 155

8 -

7 -

5 -

0 10 20 Time (tnin)

7 -|

G -

5 -

0 10 20 Time (min)

Fig 5.1 Multiple-time uptake plots of [^II]CT/VP (squares) and [^"^CJinulin (triangles) into brain and CSF of the in silu perfused rat. Uptake is expressed as the percent ratio of tissue to plasma radioactivities (ml.g'^ or ml.mh^). Each point represents the mean i S.E.M., n = 3-7 animals for each point. The brain uptake of [^HJCTAP vvas statistically greater than that of the vascular space marker [^^cjjnulin (P < 0.05). However, aller considering vascular space, the brain and CSF uptake values of [-^HJCTAP were not statistically different. 156

Table 5.1. The calculated unidirectional transfer constants (Kjn) and initial volumes of distribution (Vj) for [^H]CTAP and [^^C]inulin.

Compound Brain CSF

Vj (mI.lOOg-1) Kjn (ul.niin-l.g"l) Kjn (ul.min-l.g"l)

f^'^ClinuIin 1.18 ±0.45 0.27 ± 0.03a 0.88 ± 0.33b PHICTAP 1.91 ±0.09 1.61 ± 0.07a 2.14 ± 0.82b Kin values were determined as the slope and ordinate intercept of the computed regression lines shown above. Values ± S.E.M. were determined from the ^ the computer generated lines of regression and single-time uptake analysis at a perfiision time of 20 minutes (brain value is corrected for vascular space). 157

1.02

[^HICIAP olandatd

•r 0.51 E ou

L60

[^IIICIAP Brain Exiracl

a ~ 0.00 ull

0 10 20 30

Fig 5.2 HPLC Flo-One Radioactive Detector Chromatograms of TFA extracts of [^HjCTAP from the brain after a 20 minute vascular perfusion. Majority of sample co- eluted with purified radioactive standards. 15S

110

100 -IT\ Ql •< 90 f— o 80 "o >« 70 >(U o 60 1 GL 30 •- Brain 20 *-S0rum 1 10 0

100 200 Time (min)

Fig 5.3 The percent recovery of intact CTAP in rat brain and serum over a 240 minute time coursc. T|/2 denotes half time disappearance. Ty2 > 500 minutes in both brain and scrum as determined by HPLC analysis. 1_V)

ia • t3H|CTAP 9 • I3H1CTAP + 100 uM CTAf

8 ; 7 § 6 «o

0 Brain CSF

Fig 5.4 Uptake is expressed as a percentage ratio of tissue to perfusate radioactivities (^Tissue^ Of Perfusion time was 20 minutes and values are the mean i S.E.M. for 3 animals. The uptake of [3H]CTAP in the absence (black) and presence (grey) of 100 [.iM unlabeled CTAP into the brain and CSF was found not to be statistically difTerent. 160

Table 5.2. Percent of [3H]CTAP bound to protein in the perfusion medium or rat serum.

[3H]CTAP (dpm/O.lml) pinCTAP (dpm/O.lml) in Perfusion Medium in Rat Serum Total Sample 271127 1196935 Ultrafiltrate 8S841 189294 Percent Protein Bound 68.2 84.2 The levels of radioactivity found in the total sample and ultrafiltrate were determined by ultrafiltrational centrifugal dialysis using Amicon® Centrifree Micropartition Devices. lol

10 0 [3HICTAP 9 • [3H]Morphine

7 H 6 -I

-E 5 cS cn az 4 -J 3 •

2 •

1 •

0

Homogsnat* Suparnatainl Psllat

Fig 5.5 Rorain percent represents the ratio of homogenate, supernatant or pellet to plasma radioactivities. Supernatant represents brain homogenate depleted of the cercbral capillary endothelium. Perfusion time was 20 minutes. Values ± S.E.M., n=3-4 experiments each. *Pellet contained significantly smaller counts than homogenate for both [3H1CT.\P and plI]morphine. (P < 0.05). 162

Table 5.3. Rrissye octanol/saline partition coefficient for [^H]CTAP, [^Hlmorphine and p-'^C]inulin. pHlCTAP [^HlMorphine [l^cilnulin

'^Brain 5.10 ±0.66** 8.33 ± 1.21** 1.46 ±0.36

RCSF 4.27 ± 1.06 6.97 ± 0.76* 1.75 ±0.66

Octanol/Saline 0.0380 ± 0.0047** 0.0455 ± 0.0036** 0.0008 ±0.0001

^Tissue percent represents the ratio of tissue to plasma radioactivities x 100. Data are mean ± S.E.M. Perfusion time is 20 minutes and n = 3-5 animals per compound. Octanol / saline partition coefficients were calculated as the ratio of labeled substances in the octanol phase to that in the aqueous phase. For each compound triplicate determinations were made. Student's t-test was used and statistical significance was taken as ** P < 0.01 or * P < 0.05. 163

Discussion

The present study has led to two major findings. First, [^H]CTAP can enter the

brain by crossing the blood-brain barrier (BBB) and second [^H]CTAP is stable in the

brain and serum of the rat but remains extensively bound to albumin in the perfusion

medium.

My data demonstrates that very little [^^C]inulin actually enters the brain and CSF.

The unidirectional transfer constants for [l^C]inulin into the brain and CSF were 0.27 =

0.03 and 0.88 ± 0.33 ul.min"l.g"^. xhese values are quite low and compare well to the

values shown in Chapter 4 of this dissertation for [^^CJsucrose, 0.32 ± 0.02 and 0.07 i

0.02 ul.min-l.g-l into the brain and CSF, respectively. These data ensure that the overall

physiology of the BBB remains intact during the in situ brain perfusion experiments

because these large molecular weight compounds([^'^C]inuIin MW = 5000-5500 and

[^^C]sucrose MW = 342) were not detected at high levels in the CNS.

The data presented show that [^H]CTAP can enter the CNS. The unidirectional

transfer constant of [^H]CTAP into the brain and CSF was 5.96 and 2.43-fold higher than

that calculated for [^^CJinulin. These data also show that there is a greater amount of

radioactivity detected in both the brain and/or CSF at all time points for [^H]CTAP in comparison to [^^C]inulin and that [^H]CTAP is entering the CNS predominantly through the BBB, whereas the blood-CSF barrier plays a minor role. This can be explained by the fact that the CSF is more likely to act as a "sink" to the brain, than the brain acting as a 164

"sink" to the CSF (Davson et al., 1961) and that the surface area of the choriod plexus is

approximately 5000 times smaller than the surface area of the cerebral capillary

endothelium (Bradbury, 1979). The small amount of [^H]CTAP detected in the CSF is

most likely due to the diffusion of drug from the stagnant brain extracellular fluid to the

rapidly flowing CSF environment.

The measurement of intact pH^CTAP in the brain after a 20 minute in silii brain

perfusion ensured that we were measuring intact pH]CTAP in the brain and not just free

tritium due to water exchange. HPLC verification also allowed for the monitoring of

potential peptide metabolism due to peptidases that may be expressed in the brain or at the

blood-brain interface. [^HJCTAP remained predominantly intact (62.8%) in the brain

after a 20 min rat brain perfiision. The HPLC verification of detectable amounts of

[^HJCTAP measured in the brain ensures that this mu-selective antagonist can enter into

the brain intact and be available to elicit a pharmacological response. While other

metabolites produced by brain perfusion were not identified, they may represent enzymatic

metabolism either at the BBB interface or in the CNS after passage. The large amounts of

intact [^H]CTAP detected in the brain after a 20 minute brain perfusion may explain why

CTAP is such a potent antagonist. This peptide may actually enter into the brain via diffusion, then become trapped in the brain compartment.

Other important experiments were performed to ensure the biological stability of

this drug. In vitro stability studies were conducted in serum and brain homogenate of the

rat. [3H]CTAP was shown to be stable in the blood and serum of the rat (T1/2 > 500 165 min), showing that the stmcture of this peptide offers enzymatic resistance to blood-borne peptidases. The biological stability of CTAP is probably due to the penicillamine - cysteine disulfide linkage that allows the compound to become conformationally constrained and biologically active. The long metabolic half life may also be attributed to the high protein binding ability of CTAP (Table 5.2). The metabolic half-life values were quite long compared to another octapeptide analog of somatostatin, Sandostatin™.

Sandostatin™ has numerous clinical uses in the treatment of endocrine disturbances, especially those resulting from inappropriate hormone secretion by tumors. The pharmacokinetic half-life of Sandostatin™ was determined to be 113 min after subcutaneous administration (Lamberts 1986, 1987). These experiments confirm that

CTAP can overcome a problem that impedes the use of naturally occurring peptides in the clinic, which is mainly a short metabolic half-life.

Another component of CNS biodistribution that needs to be measured when evaluating the blood-to-CNS pharmacokinetics is the ability of a given test solute to bind to serum proteins. [^H]CTAP was found to be extensively protein bound to albumin in the perfusion medium (68.2%) and to rat serum proteins (84.2%) (Table 5.2). A protein binding component has also been observed with other analogs of somatostatin (Banks et al., 1992). This suggests that actual CNS uptake values may be higher without this protein binding component being taken into consideration. Extensive binding of

[3H]CTAP to albumin in the perfusion medium and rat serum proteins may actually be protecting CTAP fi-om enzymatic degradation by systemic peptidases. 166

It is apparent that [^H]CTAP can enter into the CNS based on in sitii brain perfiasion experiments coupled to HPLC analysis. The next question to be answered was to determine if the mechanism of entry was by means of passive diffusion or saturable transport. In situ brain perfusion experiments were performed with [^HjCTAP in the presence of 100 {xM CTAP. Entry into the brain and CSF was not inhibited by the addition of unlabeled CTAP (100 jxM) to the perfusion medium This suggests passage into the CNS was most likely directed through diffusion across the membranes that comprise the BBB rather than by saturable transport. These results concur with the findings of Banks et al., 1992, showing that somatostatin analogs can cross the murine

BBB guided by diffusion. This does not rule out the possibility of a saturable transport mechanism that may facilitate CTAP transport from the brain back into the blood which has been described previously as PTS-5 which is involved in the brain to blood transport of somatostatin and certain other analogs (Banks and Kastin 1992). This brain-to-blood transport system seems less likely to occur with CTAP since there was considerable amounts of intact [^H]CTAP detected in the brain after a 20 minute vascular brain perfusion.

Comparisons were made between CTAP and the classical, clinically efficacious, opioid agonist, morphine, in reference to the amount of intact compound that crossed either the blood-brain or blood-CSF barriers and the contribution of binding to the endothelial space. The vascular component contributes significantly to the uptake of both

[^H]CTAP and [^HJmorphine, 44 % and 32 % respectively. [^H]CTAP and 167

[^H]morphine may sequester in the endothelial cell component due either to high lipophilicity and / or binding to brain microvessels. Previously it has been shown that brain microvessels rapidly sequester and degrade somatostatin analogues (Pardridge et al.,

1985). This may represent one mecham'sm for the rapid inactivation of brain derived neuropeptides subsequent to neurosecretion. A potential reason for the high concentration of [^H]CTAP detected in the microvasculature pellet may be due to the binding of [^H]CTAP to a receptor on the cell membrane of the endothelial cells that comprise the vessel walls in order to achieve enzymatic degradation. High levels of peptidases are known to be expressed at the membranes of brain microvessel endothelial cells as shown in Chapter 2 of this dissertation.

A greater amount of [^H]morphine entered both the brain and CSF after a 20 minute brain perfusion when compared to [^'^CJinulin, P < O.OL and P < 0.05, respectively

(Table 5.3). The increased CNS penetration by [^H]morphine compared to [^H]CTAP is likely due to increased lipophilicity shown by the high octanol / saline partition coefficient.

In addition, the RTissue values correlate well with octanol / saline partition coefficients for [3H]CTAP, [3H]morphine and [^^qinulin. (R=0.946 for brain and R=0.926 for CSF).

Thus, lipophilicity may be a determining factor for CNS entry of these drugs. This also confirms the reliability of using our in situ brain perfusion technique to mimic or predict in vivo situations such as compounds attempting to traverse the blood-brain and / or blood-

CSF barriers. 168

This work supports the hypothesis that the mu-selective somatostatin analog,

CTAP, can cross the BBB at therapeutic levels. The actual amount of CTAP that crosses both the blood-brain and blood-CSF barriers is quantitatively comparable to that of the efficacious, mu-selective, agonist, morphine. It is surprising that a compound with the clinical efficacy of morphine does not enter the brain at large levels. The absolute percent of injected dose of morphine that enters into the brain has been calculated at 0.02 % / g of brain tissue (Banks and Kastin, 1994). The present study shows that CTAP may provide an important role in the clinic for treating narcotic addiction, dependence or overdose.

Since CTAP has excellent biological stability and blood-CNS penetration it may be an improvement over the classical , naloxone, for treating opioid crisis.

Naloxone has a relatively short duration of action and must be administered repeatedly or by infusion. Also, one must be precise in titrating the dose for fear of precipitating a severe withdrawal (Goodman and Oilman, 1996). CTAP may therefore provide improved antagonism at the mu receptor without intrinsic activity and a longer duration of action with less severe withdrawal. 169

Chapter 6. Brain and spinal cord distribution of biphalin: Correlation with opioid

receptor density and mechanism of CNS entry.

Introduction

A major challenge to cell biology, pharmacology and drug delivery is

understanding the ability and nature of diffusion and transport of peptides across a cell

membrane. It is commonly accepted that the degree of lipophilicity is primarily

responsible for the movement of molecules across a membrane as described in ^Hapters 2 and 4 of this dissertation, yet facilitated transport, which may involve association with a

protein expressed on the cell membrane, can contribute significantly to the transport of smaller molecules such as peptides as shown in Chapter 3 of this dissertation. Since

peptides can be very potent and efiRcacious pharmacological drugs, only a small amount of

the peptide actually needs to enter the CNS to bind to a receptor and exert a

pharmacological effect. Since small doses of peptides reach clinical efficacy, it is

important to explore the ability of potential peptide therapeutics to utilize this potential delivery vehicle for enhanced CNS entry.

These investigations will elucidate the CNS entry of biphalin,(Tyr-D-AIa-GIy-Phe-

NH)2 3- dimeric, peptide analog of enkephalin which contains two pharmacophores linked

by a hydrazide bridge. It is the most potent analgesic enkephalin analog we have studied to date. In mice, biphalin is more potent than most alkaloid opiates such as morphine and 170 etorphine when administered i.e.v. and morphine and biphalin were found to be equipoteni after ip administration (Horan et al., 1993). In rats, i.v. biphalin was found to produce significant analgesia in comparison to morphine and intrathecal biphalin was found to be more potent than morphine (Silbert et al., 1991). In addition, biphalin was found to be 90 times more potent than methionine enkephalin at inhibiting electrically induced contractions of guinea-pig ileum, suggesting it binds to mu-opioid receptors. It was also found to produce pronounced analgesia after peripheral administration (Lipkowski et al.,

1982). Development of potent and efficacious opioid analgesics is important for the clinical treatment of chronic pain because potent analgesics have been shown to produce less tolerance (Stevens et. al., 1994) and dependence (Stevens and Yaksh, 1989). The aim of this work was to determine and understand the mechanism of biphalin potency.

Initially, regional brain and spinal cord distribution studies with [^^^I-Tyr^jbiphalin v/ere performed in the rat after 5, 20 and 40 minute i.v. bolus injection. It is important to measure CNS uptake into the spinal cord since it has been shown that, although the vascularity of the spinal cord is half that of the brain (Craigie 1919), the spinal cord has higher vascular permeability than the brain (Daniel et al., 1985; Prockop et al., 1995).

With opioid receptor analgesia being mediated at spinal sites as well as brain sites, it is crucial to measure distribution into spinal areas. These results were then correlated with the measurement of delta and mu opioid receptor mRNA and binding sites in various brain regions (Mansour et al., 1995). The present experiments investigated if biphalin can reach the brain and spinal cord regions that express opioid receptors. Additional 171 pharmacological experiments were performed to examine whether biphalin detected in various brain and spinal cord regions could be displaced by pre-treating the animals with various selective and nonselective opioid antagonists. The delta-selective antagonist naltrindole was used because it has been shown through pharmacological methods to produce antagonism at delta-opioid receptors from a subcutaneous route of administration. Mu-receptor antagonism was produced in the brain using CTAP which has been shown to cross the BBB solely by diffusion (Chapter 5 of this dissertation).

Several techniques have been used over the past two decades to study the ability of peptides to gain access to the CNS. Since peptides are slowly permeating molecules, a technique that can asses the kinetics of CNS entry over extended time points is needed.

Vascular brain perfusion has been shown to be a more sensitive technique when compared to the BUI method in which a transit time of the test-molecules throughout the cerebral circulation is about one second (Zlokovic 1995). I performed in situ brain perfusion experiments to more fully examine the mechanism of biphalin CNS entry. Biphalin has already been shown to cross the BBB, yet the actual mechanism of CNS entry has not been investigated. These experiments will elucidate whether or not the CNS entry of biphalin is governed by diffusion only, or if a saturable component exists, since Chapter 4 of this dissertation suggests the presence of a saturable uptake mechanism at the BBB contributing to the CNS entry of biphalin. Investigations were made to characterize the saturable component that contributes to the CNS entry of this potent opioid agonist. I examined the acute modulation of this putative transport system by treating vsath an opioid 172 agonist (DPDPE), large neutral amino acid carrier substrates L-phenylalanine and 2 aminobicycIo-[2,2,l]heptane 2 carboxylic acid (BCH), a Na"^ K"^ ATPase inhibitor

(ouabain), a cellular microtubular system inhibitor of transcytosis (colchicine) and inhibitors of L-type Ca^"^ channels (verapamil and nifedipine).

Protein binding assays in rat serum and an artificial perfusate were also performed to determine the contribution of protein binding to the bioavailability of

Tyrl]biphalin. These studies will describe the mechanism of biphalin potency so that conformational changes can be explored in the future that may promote a more efficient transmembrane movement of peptides via either difiiision or facilitated transport. 173

Materials and Methods

lodination of biphalin.

Biphalin was synthesized by methods similar to those reported previously (Misicka et al., 1997). Biphalin was monoiodinated on the tyrosine^ residue using a standard chloramine-T procedure as described previously (Bolton, 1986). Purification of iodinated biphalin was performed using a reversed-phase Perkin Elmer 250 high-performance liquid chromatography (HPLC) gradient system, a Perkin Elmer integrator and a Beckman ODS

C18 Ultrasphere column (4.6 x 250 mm). Samples were eluted using a curvilinear gradient of 0.1% TFA in acetonitrile (20 to 50%) vs. 0.1% aqueous TFA over 30 min at

1.5 ml.min and the column temperature was maintained at 37°C. After separation on the column, the outflow from the HPLC was routed to an A200 Flo-One Radioactive

Detector with a 2.5 ml flow cell.

Regional Brain and Spinal Cord Distribution of [^-^I-Tyr^Jbiphalin.

After anesthesia with sodium pentobarbital (65 mg/kg), (2200

Ci/mmol) was administered by tail-vein injection (10 |iCi/animal). After 5, 20, 40 min, the scalp of the rat was cut allowing for a cistemae magna CSF sample with a glass cannula.

The animal was then cut along the ventral midline from the lower abdomen tlirough the chest cavity and a blood sample was immediately taken form the right ventricle. The 174 animal was then perfiised with 0.9% saline and the brain removed and dissected into brain stem (BS), cerebellum (CE), frontal cortex (FC), caudate-putamen (CP) and nucleus accumbens (NA). The pituitary (PT), choroid plexus (CHP) and spinal cord were also removed and the spinal cord was sectioned into cervical (CV), thoracic (TH), lumbar (LR) and Cauda equina (EQ). Each tissue region was weighed (smaller tissues were weighed on a CAHN 29 automatic microbalance) and the amount of radioactivity was determined by counting on a Beckman 5500 gamma counter. For additional 20 minute experiments the rat was pretreated with either naloxone (10 mg/kg, i.v.; a non-selective opioid antagonist), naltrindole (10 mg/kg, s.c.; a delta-selective antagonist dose; Portoghese et al., 1988;

Ayres et al., 1990), CTAP (lOmg/kg, i.p., a mu-selective antagonist dose; Wang et al.,

1994 and Chapter 5 of this dissertation) 10, 35 and 20 minutes before i.v. administration of the radiolabeled peptide, respectively.

Brain and spinal cord distribution of [^-^I-Tyr^]biphalin was expressed as the percentage of total radioactivity injected. The syringe was weighed before and after injections to ensure the correct injected dose was calculated. Statistical analyses of data was done using analysis of variance coupled to the Newman-Keuls test to determine the significance of difference between % injected dose / gram tissue for each brain region.

Also, Student's t-test was used for the comparison of regions to the CNS mean value calculated for each respective time point analyzed.

Braf/i Extraction of p~^I-Tyr^]Biphalin After 20 Minute i.v. Injection. 175

Brain extractions were performed using a modified method of Erchegyi et al.

(1991). Briefly, rats were admim'stered approximately 10 uCi of [^2^I-Tyr^]biphalin i.v. as described above. The animal was perfused with 0.9% saline and the brain was removed and placed immediately in 7.5 ml of ice-cold 10% trifluoroacetic acid (TFA). Each sample was then homogenized (Polytron homogenizer) and centrifiiged at 20,000 x g for 20 min.

The supematants were collected and an equal volume of ether was added. The ether phase was discarded and the samples were then lyophilized to dryness. The samples were diluted to 500 ul with 10 % acetonitrile and taken for HPLC analysis.

Brain extractions were then analyzed using a series 410 HPLC gradient system

(Perkin-Elmer). Samples were elated from a Beckman Ultrasphere column (4.6 x 250 mm: Beckman Instruments Inc.) with a curvilinear gradient of 20-50% 0.1% TFA in acetonitrile v.y. 0.1% TFA aqueous in 30 minutes at 1.5 ml/min and the column temperature was maintained at 37° C. Afl;er separation on the HPLC column, the outflow was routed to the on-line A200 Flo-One Radioactive Detector equipped with a 2.5-ml flow cell (Packard Radiomatic Instruments &. Chemicals, Meriden, CT).

In Situ Brain Perfusion Studies.

The protocol described below was approved by the Institutional Animal Care and

Use committee (lACUC) at the University of Arizona. Adult Sprague-Dawley rats (250-

300 g) were anaesthetized with sodium pentobarbital (64.8 mg.kg"^) and heparinized

(10,000 U.kg'l). The jugular veins were located and the common carotid arteries were 176 cannulated using fine silicone tubing connected to a perfusion system as previously described (Takasato et al., 1985)

Perfusion was performed with a thoroughly oxygenated (p02 = 642-727 mmHg) mammalian Ringer (37°C) solution. Once the desired perfusion pressure and rate were achieved (approximately 100 mmHg and 3.1 mls.min"^ respectively), the right jugular vein was cut and allowed to drain. The contralateral carotid was cannulated and perfused in a similar manner as described above. [^-^I-Tyr^]biphalin in the presence or absence of varying concentration of biphalin (0, 10, 25, 50 and 100 i-iM), was infused using a slow- drive syringe pump into the inflowing, mammalian Ringer. Once the set perfusion time

(20 min) was achieved, a cistema magna cerebrospinal fluid CSF sample was taken with a glass cannulae. The animal was then decapitated and the brain was removed. The choroid plexuses were excised and the brain was dissected- The perfusion outflow was collected from the carotid cannulae at the end of the time point to serve as a reference. The brain and CSF samples were then weighed and the amount of radioactivity was determined by counting on a Beckman 5500 gamma counter.

Additional 20 minute in situ brain perfusion experiments were performed in the same manner as above in the presence of 100 |.iM DPDPE (delta-selective agonist), 10 mM BCH (non-metabolized analog transported by L-amino acid system) 1 and 100 |.iM L- phenylalanine (Hargreaves and Pardridge 1988), 50 |aM leucine-5-enkephalin, 10 nM ouabain (Na"*" K"^ ATPase inhibitor), 10 nM colchicine (cellular microtubular system inhibitor), 10 nM verapamil (l-type Ca^"^ channel inhibitor) and 10 nM nifedipine (1-type 177

Ca^"*" channel inhibitor). A 2 minute pre-perfiision was performed to allow for maximal inhibition and radiolabel infusion started the 20 minute experiment.

Expression of Results.

The amount of radioactivity in the brain and CSF (Cxissuci dpm.g"i or dpm.ml"') was expressed as a percentage of that in the artificial perfusate (Cpi; dpm.ml"^) and termed the Rxissue

Rxissue ~ CTissue X 100 (1) Cpl

Unidirectional rate constants (Kjn f.il.min"^g"^) were determined by single time-point analysis, as previously described (Zlokovic et al., 1986), where T is the time in minutes;

Kin ~ CTissue CpiT (2)

Blood-to-brain unidirectional transfer constants determined in this manner were corrected for vascular space by subtracting [^"^CJsucrose (Rerain) [^^^I-Tyr^]biphalin (Rerain)-

Unidirectional transfer constants determined from these experiments, represent cerebrovascular permeability surface area products, PA (ml.min-l.g-1) (Gjedde, 1988;

Zlokovic et al., 1989;1990). Thus, if [^^^I-Tyr^jbiphalin is being studied in the presence of increasing unlabelled concentrations of biphalin, Kjn can be defined as:

PA = Kjjj = Vmax + Kj ^cap) (3) 17S

where Vj^ax is the maximal transport rate of the saturable component; is the half-

saturation constant; is the constant of non-saturable diffiision and Cgap is the mean

capillary concentration of DPDPE. Under the experimental conditions described above,

the difference between Ccap and the concentration of biphalin in the perfusion medium

(Cpi), becomes negligible, since the flow to the brain (F) is always greater than 1 ml.min*

^g"^ which is much greater than the highest measured kju and the equation can be

simplified (Gjedde, 1983,; Smith et al., 1984; Zlokovic et al., 1989; 1990);

Kjn = Vmax + (4) (Km Cpi) Unidirectional [^^^I-Tyr'jbiphalin flux (Jj^; nmol.min'^ml"^) into the brain and CSF can

then be calculated as;

Jin = F(l-e-Kin/F)Cp| (5) and since F »K(n (Takasato et al., 1985) this equation approximates to;

^in KinCpi (6)

Unidirectional Flux of [l^Si.Xy^-ljijiphalin can be related to K^, V^ax Kj by the following equation:

Jm= Vmax Cpl + KjCpi (7) CKm + Cpl)

Estimates of the best fit values for V^ax (nmol.min-l.g-1), (mM) and K^i (ml.min"

^g"^) were obtained by fitting this equation to the brain vascular perflision data by the 179 method of least squares with statistical weighting (Enzfitter program from Biosoft,

Cambridge, UK).

Protein Binding Studies.

The amount of [^25i_Tyi-l]5iphaIin binding to either bovine albumin in the perfusion medium or proteins in rat serum was determined by ultrafiltration centrifugal dialysis (Paulus H., 1969). Rat serum was obtained by harvesting blood from Sprague-

Dawley rats and allowing the blood to clot for 30 minutes on ice and 30 minutes at room temperature. The whole blood was then centrifliged (Sorvall RC2-B centrifuge; Dupont

Medical Products, Wilmington, DE) at 20,000 x g for 20 minutes to produce a serum supernatant. [^-^I-Tyr^]biphalin was dissolved in either perfusion medium or rat serum warmed to 21^C and ultrafiltrated using a Centrifiree™ micropartition device (Amicon,

Beverly, MA). The total concentration (T) of [^^Si.i-y^-ljbiphalin introduced into the system and found in the ultrafiltrate (F) was determined by counting on a Beckman 5500 gamma counter. The percentage of [^25];_7yrl]biphalin bound to either albumin in the perfusion medium or proteins in the rat serum was expressed as [(T-F) / T ] x 100. To verify that bovine albumin was not found in the ultrafiltrate, the protein concentration was determined by the method of Lowry et al. (1951). 180

Results

Regional brain and spinal cord distribution of Jbiphalin.

The calculated mean % injected dose values for the CNS regions from the 5, 20 and 40 minute experiments were 0.00451, 0.0534 and 0.0304 % injected dose / g tissue respectively (Figure 6.1). For the 5 minute i.v. administration a statistically greater amount of radioactivity was detected in the NA (P<0.05), PT (P<0.05) and CHP (P<0.01) when compared to the mean CNS value. A 20 minute i.v. administration showed statistically greater amounts of radioactivity detected in the NA (P<0.01), PT (P<0.01) and CHP (P<0.05). The longest time point, 40 minute, showed statistically greater amounts of radioactivity in the NA (P<0.05), PT (P<0.05) and CHP (P<0.05). The 20 minute time point yielded the highest mean CNS value. The uptake values of

Tyr^jbiphalin into various brain regions were compared to each other by analysis of variance coupled with the Newman-Keuls test (Table 6.1).

Administration experiments (20 minute i.v.) were then performed (Figure 6.2) pretreating with either naloxone (10 mg/kg, i.v.; a non-selective opioid antagonist), naltrindole (10 mg/kg, s.c.; a delta-selective antagonist dose; Portoghese et al., 1988:

Ayres et al., 1990), or D-Phe-Cys-Tyr-D-Trp-Arg-Thr-Pen-Thr-NH2 (CTAP) (lOmg/kg, i.p., a mu-selective antagonist dose; Wang et al., 1994) 10, 35 and 20 minutes before i.v. administration of the radiolabeled peptide, respectively. Pretreatment with the 181 nonselective opioid antagonist, naloxone, resulted in a statistically significant decrease in the following brain and spinal cord regions: FC (P<0.05), NA (P<0.01), CE (P<0.05) TH

(P<0.05) and LR (P<0.05). Naltrindole pretreatment showed a decrease in all brain and spinal cord regions, yet this was not statistically significant. Interestingly, there was a statistically significant increase in the uptake of [^-^I-Tyr^]biphalin into the CHP

(P<0.01) after naltrindole pretreatment. CTAP pretreatment caused a significant reduction in the % injected dose / g tissue in the following regions: NA (P<0.01), CSF

(P<0.05), CV (P<0.05), TH (P<0.05), LR (P<0.05) and PT (P<0.05).

Brain extractions of [^-^I-Tyr^Jbiphalin after a 20 minute i.v. administration.

The HPLC chromatogram (data not shown) confirmed that approximately 80% of the TFA-extracted [l^Sj.Xyrljbiphalin co-eluted with the intact iodinated biphalin standard after a 20 minute i. v. injection.

In situ brain perfusion studies.

Vascular brain perfusion of [I25i.xyr^]biphalin in the presence or absence of varying concentration of unlabeled biphalin (0, 10, 25, 50 and 100 joM) was performed as shown Figure 6.3. These experiments revealed the presence of a saturable uptake into the brain. The brain uptake of [^^^I-Tyr^]biphalin into the brain was inhibited by 21.6%,

34.5%, 50.5% and 52.0% at concentrations of 10, 25, 50 and 100 |.iM, respectively. The 182 uptake of j^q jjjg Qgp ^^5 jjot statistically affected by the presence of biphalin at any of the above concentrations studied.

Figure 6.4a illustrates the effect of unlabelled biphalin on the flux of

Tyr^Jbiphalin (Jju) into the brain. The influx values were calculated from Kj^ values determined by single time uptake analysis for each experimental point (vascular space removed). A small saturable component does exist that contributes to the total brain uptake of [^-^I-Tyr^]biphalin. The saturable and non-saturable components to the total influx of [^-^I-Tyr^]biphalin were estimated using the calculated values of the half- saturation constant (2.60 (.iM), the maximal transport rate (14.6 pmol.min'^.g"

^) and the difRision constant Kj (0.568 |j.l.min"l.g'l). (Table 6.2).

Additional 20 minute in situ brain perfusion experiments were performed in the same manner as above in the presence of 100 j.iM DPDPE (delta-selective agonist), 10 mM BCH (non-metabolized analog transported by L amino acid system) 1 and 100 jaM L- phenylalanine, 50 |iM leucine enkephalin (Table 6.3). A statistically significant decrease in the RBrain value was observed when 10 mM BCH and 100 |.iM L-phenylalanine were added to the perfusate (P<0.01). Experiments were also performed with 10 nM ouabain

(Na"^ K"*" ATPase inhibitor), 10 nM colchicine (cellular microtubular system inhibitor), 10 nM verapamil (1-type Ca^"^ channel inhibitor) and 10 nM nifedipine (1-type Ca^"*" channel inhibitor) added to the perfusate (Table 6.4). No inhibition in the brain accumulation of

was observed. 183

Protein binding studies.

[llSi.-fyrljbiphalin was found to be bound to protein in both the perfusion medium

(65.1%) and rat senmi (90.8%) (Table 6.5). No protein was detected in the ultrafiltrate after performing a Lowry protein assay. 1S4

Regional CNS distribution of [125l-Tyr1]biphalin after a 5 minute intravenous administration

0.35 • Remaining Brain 0) - ZJ 0.30 • Frontal Cortex CO cn a Caudata-Pulaman - 3 Nucleas Accumbens 0.25 -- —O) 1 — • Csrebsllum o CO • Brain Stem O 0.20 —•

Q - = CSF "O - CD 0-15 — o - Q Cervical CD cz O.'Q 3 Thoracic n Lumbar c J. CD 3 Equinao o 0.05

Fig 6.1 Regional brain, spinal cord and circumventricular organ (CVO) distribution of [l25[_-ryplj5iphai[n after a 5 minute i.v. injection. Plot represents the mean ± S.E.M. of pcrcent injected dose / gram tissue for each tissue region. Statistical significance was taken as compared to CNS mean percent injected dose / g tissue. (**P < 0.01 and "P <0.05) (n = 4-6 rats). 185

Regional CNS distribution of [125I-Tyr1]biphaIin after a 20 minute intravenous adminislralion

0.35 a Remaining Brain

Fig 6.1 Regional brain, spinal cord and circumventricular organ (CVO) distribution of [l25i_Tyrl]biphaIin after a 20 minute i.v. injection. Plot represents the mean ± S.E.M. of percent injected dose / gram tissue for each tissue region. Statistical significance was taken as compared to CNS mean percent injected dose / g tissue. (**P < 0.01 and *P <0.05) (n = 4-6 rats). I St'.

Regional CNS distribution of [125l-Tyr1]biphalin after a 40 minute intravenous administration

• Remaining Brain CD Q.30 • Frontal Cortex 00 CO o Caudats-Putamen 3 Nucleas Accumbens cn 3 Cerebellum • Cauda Equina o '-'•t. -J I a> Q- - • Pituitary -4 kMHi. 0.00 iB a Choroid Plexus Brain CSF Spinal CVO CNS Regions

Fig 6.1 Regional brain, spinal cord and circumventricular organ (CVO) distribution of [I25i_7ypl jbjpiiaiin af^ep ^ 40 minute i.v. injection. Plot represents the mean ± S.E.M. of percent injccted dose / gram tissue for each tissue region. Statistical significance was taken as compared to CNS mean percent injected dose / g tissue. (**P < 0.01 and <0.05) (n = •'1-6 rats) 187

Table 6.1. Analysis of variance coupled with the Newman-Keuls test for significance between CNS uptake values (percent injected dose / gram tissue) for 20 minute time point. ** denotes significance at P < 0.01 and * denotes significance at P < 0.05.

2 3 4 5 6 7 9 10 11 12 13 Region 1 8

1. Remaining Brain ns ns • ns ns ns ns ns ns ns •• • 2. Frontal Cortex •« • ns - ns ns ns ns ns ns ns ns ns 3. Caudate-Putamen •• • ns ns - ns ns ns ns ns ns ns ns 4.Nucleas Accumfaens • • • « •« ns ns - ns ns ns ns ns 5. Cerebellum •• « ns ns ns ns - ns ns ns ns ns ns

6. Brain Stem • •• • ns ns ns ns - ns ns ns ns ns 7. CSF •• « ns ns ns ns ns ns - ns ns ns ns

8. Cervical • •« • ns ns ns ns ns ns - ns ns ns 9. Thoracic ns ns ns ns ns ns ns ns ns ns • 10. Lumbar •• « ns ns ns ns ns ns ns ns ns - ns 11. Cauda Equina « • ns ns ns ns ns ns ns ns ns - 12. Pituitary • - 13. Choroid Ple.\us • • • • • • • • • • • ns - ISS

Regional CNS distribution of [125l-Tyr1]biphalin after a 20 minute i.v administration with naloxone pretreatment {10mg/kg) I.v.

0.35 -3 Remaining Brain id Frontal Cortex 0.30 CfJS mean = 0.0278 Caudate-Putamen ;0 Nucleus Accumbar !• Cerebellum 0.25 -f; O) iii Brain Stem o CO O 0.20 —; ;c CSF Q •o CD In Cervical O 0.15 O) Thoracic •o Lumbar 0.10 I ! in Equinae (D CJ ~ > >

! I 0.00 —^j-t 1 i' Brain CSf; Spinal CVO CNS Regions

Fig 6.2 Twenty minute regional brain, spinal cord and circumventricular organ distribution of [^^^I-Tyr^]biphalin after a 20 minute i.v. administration. (A) Nalo.xone (10 mg/kg; i.v.) was administered 10 minutes before labeled peptide administration (i.v ). Statistical significance was taken as compared to each specific region in the 20 minute experiment (figure 6.1), (n = 3-^1 rats) 18'J

Regional CNS distribution of [125I-Tyr1]biphalin after a 20 minute i.v. administration and naitrindole pretreatment (10mg/!

0.40 —

0.35 -| G Remaining Brain CNS mean = 0.042 Frontal Cortex 0.30 -I n Caudale-Putamen a Nucleus Accumbeni • Cerebellum 0.25 - n Brainstem CTJ 03 C/J 0.20 Q CSF O Q TD o Cervical (D 0.15 -i O 3 Thoracic (D 0.10 -i • Lumbar U Equinae O O 0.05 \ I w_ Pituitary

CNS Regions

Fig 6.2 Twenty minute regional brain, spinal cord and circumventricular organ distribution of [^^S^.jyj-Ijbiphalin after a 20 minute i.v. administration. Naitrindole (10 mg/kg, s.c.) was administered 35 minutes before labeled peptide administration (i.v ). Statistical significance was taken as compared to each specific region in the 20 minute experiment (figure 6.). (n = 3-4 rats) i';o

Regional CNS distribution of [125I-Tyr1]biphalin after a 20 minute i.v. aclministraton and CTAP pretreatment (lOmg/kg) i.p. 0.30 — i a Remaining Brain • Frontal Cortex o

CNS Region

Fig 6.2 Twenty minute regional brain, spinal cord and circumventricular organ distribution of ['-^I-Tyr']biphalin after a 20 minute i.v. administration. CTAP (10 mg/kg, i.p.) was administered 20 minutes before labeled peptide administration (i.v.). Statistical significance was taken as compared to each specific region in the 20 minute c.xperiment (figure 6.1). (n = 3-4 rats) 191

i Brain

7

6

O"' (D tn=> CS w j— OC

4

3

2

0 10 20 30 40 50 60 70 80 90 100 110 Concentration (micromolar)

Fig 6.3 The relative uptake of ['^^I-Tyr']biphalin into the brain and CSF measured as a function of unlabelled biphalin concentration. Values are mean ± S.E.M. (n = 4-5 animals / concentration) l'^2

0.10

0.09 Brain

0.00

0.07

.£ 0.06 Total E 1 0.05 Non-saturable § 0.04 LL- 0.03

0.02 Saturable 0.01

0.00

0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.10 0.11

Concentration (mM)

Fig 6.4 (A) The contribution of the saturable and non-saturable components to the brain influ.x plotted against unlabellcd biphalin concentrations. The measured values for total flu.x are the mean ± S.E.M. for 4-5 animals at 5 unlabelled biphalin concentrations. The saturable and non-saturable components to total influx were estimated using the calculated values of Kp-,, and K^j (Table 2) 193

0.20

0.15 -

I 0.10

X 3 0.05 -

0.00

0.00 0.01 0.02 0.03 0.04 0.05 O.OG 0.07 0.08 0.09 0.10 0.11 Concentration (mM)

Fig 6.4 (B) CSF influx plotted against unlabelled biphalin concentrations. The measured values for total flux are the mean ± S.E.M. for 3-4 animals at 5 unlabelled biphanin concentrations. 194

Table 6.2. The kinetic parameters for [^^Sf.Xyrljbiphalin influx into the brain and CSF determined from the vascular brain perfusion data.

[l25i_Xyrl]biphalin Km Vmax Kd (UM) (pmol.min'^.g"^) Brain 2,60±4.82 14.6±2.89 0.568±0.197

CSF - - 1.463±0.182 Values ± S.E.M. derived from the weighted non-linear regression analysis of 5 values based on 20 individual experiments. (N=4-5 animals per concentration) 195

Table 6.3. Competitive inhibition of saturable uptake of [^^Sixyrijbiphalin at the BBB.

Drug Inhibitor System ^Brain

[l^SiXyr^lbiphalin none control experiment 6.78 ±0.80 (n=10) [125iXyrl]biphalin 100 \jM DPDPE peptide carrier 6.55 ± 1.03 (n=3) [^25iTyi-l]biphaIin 50 (iM [Leu^lEnk leu-enkephalin carrier 7.28 ± 1.51 (n=3) [^25iXyrl]biphalin lOmMBCH large neutral amino **3.55 ± 0.54 (n=3) acid carrier [I25ii;-yri]biphalin 1 |.iM L-phenylalanine large neutral amino *4.91 ±0.41 (n=3) acid carrier l25iYyi-l jbiphalin 100 |jM L-phenylalanine large neutral amino **3.62 ±0.52 (n=3) acid carrier The expression Rflrain represents the ratio of tissue to plasma radioactivities x 100. Data are mean ± S.E.M. Perfusion time is 20 minutes. Student's t-test was used and statistical significance was taken as ** P < 0.01 and * P < 0.05. BCH is 2 aminobicyclo- [2,2, l]heptane-2-carboxylic acid. 196

Table 6.4. Inhibition of energy mechanisms at the BBB.

Drug Inhibitor System RBrain %

[^25ixyr^lbiphalin 10 nM ouabain Na"*" K"^ ATPase inhibitor 7.52 ± 1.09 (n=4) [l25iXyrl]biphalin 10 nM colchicine cellular microtubular system 7.95 ± 0.85 (n=3) (transcytosis) r^25iTyri]biphalin 10 nM verapamil L-type Ca^"*" channel 5.92 ±0.58 (n=3) [^25ixyrl]biphalin 10 nM nifedipine L-type Ca^"^ channel 6.15 ± 1.25 Cn-3) The expression RBrain represents the ratio of tissue to plasma radioactivities x 100. Data are mean ± S.E.M. Perfusion time is 20 minutes. 197

Table 6.5. Percent of [^^fj.-j-yi-ljbiphalin bound to protein in the perfusion medium or rat serum.

[125i.xyrl]biphalin [125i.Tyrl]biphalin (dpm/O.lml) in Perfusion (dpm/O.lml) in Rat Medium Serum Total Sample 219757 215554 Ultraflltrate 76726 19637 Percent Protein Bound 65.1 90.8 The levels of radioactivity found in the total sample and uitrafiltrate were determined by ultrafiltrational centrifugal dialysis using Amicon® Centrifi-ee Micropartition Devices, (n = 4 samples) 19S

Discussion

The aim of the present study was to examine biphalin potency from a drug delivery

and CNS biodistribution perspective. Initial investigations examined the regional brain

and spinal cord distribution of [^^^I-Tyr^lbiphalin after various timed i.v. injections into

the rat tail vein. As can be seen in Figure 6.1, the earliest time point tested, 5 minutes, showed minimal CNS entry of labeled peptide into most brain and spinal cord regions,

although a statistically greater amount of ['^^I-Tyr^Jbiphalin was detected in the CHP fP

<0.01) and PT (P < 0.05) when compared to the mean CNS value. It is not surprising

that uptake into these circumventricular organs (CVO) was high since they are not

protected by the traditional BBB and after the first pass of j-hg uptake values into these areas should be high. A 20 minute inflision seemed adequate time to allow [^^Sj.fyrljbiphalin to difRise to additional regions in the brain and spinal cord.

Figure 6.1b shows that uptake values into all brain and spinal cord regions were higher at

20 minutes than those shown for the 5 minute experiments. Interestingly the uptake into the NA was statistically greater (P < 0.01) than the CNS mean calculated for the 20 minute experiment. This shows that after only a 20 minute i.v. administration of [^-^I-

Tyr^jbiphalin this potent agonist can effectively reach brain and spinal cord sites that express delta and mu opioid receptors (Mansour et al, 1995). The overall CNS mean % injected dose / g tissue for [^^^I-Tyr^Jbiphalin was shown to be 0.053 %. This is actually significantly higher than that observed for morphine (0.02%) after an i.v. administration 199

(Banks and Kastin 1994). The CHP and PT also showed significantly higher uptake values when compared to the CNS mean value calculated for the 20 minute experiments

(P < 0.05 and P < 0.01, respectively) It is interesting to note that the 20 minute experiment revealed that the region with the highest amount of the injected dose of

Tyr^jbiphalin was the PT. Bzdega et al., 1993, have shown that the delta opioid receptor gene is expressed in large amounts in the PT and pineal glands. This is an interesting finding because our work shows that biphalin distributes highest to the PT after a 20 minute i.v. administration, suggesting that this opioid peptide analog could easily reach receptors in these tissues not protected by the blood-brain barrier. Figure 6.1c shows that after a 40 minute i.v. administration of [^25i_7yi-I]biphalin there was no corresponding increase in brain or spinal cord uptake and actually there was a slight decrease. This suggests two possibilities. One, with this later time point (40 minutes),

could be enzymatically degraded, but since the brain half life value is

173 minutes as shown in Chapter 4, this most likely does not contribute sigm'ficantly to the loss of brain accumulation after 40 minutes. Second, [^2^I-Tyr^]biphalin could potentially be transported out of the brain back into the blood at this later time point.

A 20 minute i.v. administration appears to yield the highest uptake values into the brain and spinal cord regions for [^^Sj.Xyr^lbiphalin. In addition, intact biphalin reaches both spinal and supra-spinal sites after the intravenous route of injection. It was surprising that [^2^I-Tyr^]biphalin was detected in most all spinal regions at a comparable level to the brain. This would suggest that biphalin permeability is higher across the blood-spinal 200 cord barrier when compared to the BBB, because the vasculature of the spinal cord is half that of the brain (Craigie 1919). This also has been shown for mannitol and inulin (Daniel et al., 1985), suggesting that the channels through which compounds move are larger in the spinal cord than those in the brain. This larger channel size increases the diffusional area and reduces restriction to free diSiision. Increased biphalin potency (Lipkowski et al., 1982; Silbert et al., 1991; Horan et. al., 1993), may be due to biphalin reaching spinal sites at a comparable level to brain sites shown in the present work. Also, all the time points tested in Figure 6.1 showed a comparable uptake of [^-^I-Tyr^]biphalin into the

CSF compared to most brain and spinal cord regions studied. This may be due to the CSF acting as a sink to the brain (Davson et al., 1961).

Stability of a test drug is important to measure since kinetic determinations are made from quantification of radioactivity in tissues of interest. HPLC analysis of TF.A.- extracted [^-^I-Tyr^Jbiphalin from the brain after a 20 minute i.v. administration revealed that approximately 80% of the radioactivity was intact [^-^I-Tyr^]biphalin and not metabolites. This ensures that we were measuring intact radiolabeled peptide in our experiments.

Additional experiments were performed to examine if the [^^Sj.jyj-Ijbiphalin detected in the brain and spinal cord regions was bound to opioid receptors. Figure 6.2 represents the regional distribution of [^2^I-Tyr^]biphalin after a 20 minute i.v. administration pretreating with selective and nonselective opioid receptor antagonists

(naloxone (nonselective) 10 mg/kg, 10 minutes pre, i.v.; naltrindole (delta-selective) 10 201 mg/kg, 35 minute pre, s.c.; and CTAP (mu-selective)lO mg/kg, 20 minutes pre, i.p.).

Preadministration of naloxone showed a significant brain reduction in the FC (P<0.05),

NA (P<0.01) and CE (P<0.05) compared to [125j.-ryr^]biphalin alone. This is direct evidence that a significant portion of the intact [^-^I-Tyr^]biphalin entering into the brain is predominantly bound to either mu or deha opioid receptors since it has been shown that high concentrations of mu and delta opioid receptors have been found in both the FC and

NA (Sharif and Hughes, 1989; Mansour et al., 1995). Using the delta-selective antagonist, naltrindole, we observed no significant decreases in any of the brain or spinal cord regions tested. An apparent increase in the uptake into the CHP was obser\'ed with naltrindole pretreatment. The greatest decrease in the overall CNS mean value was observed with rat pretreatment with the mu-selective antagonist CTAP, which has been shown to enter the CNS from a systemic administration purely by diffusion (Chapter 5). A statistically significant decrease was seen in the NA (P<0.01), CSF (P<0.05), CV

(P<0.05), TH (P<0.05) and LR (P<0.05). This suggests that the [^-^I-Tyr']biphaiin entering both spinal and supra-spinal sites is binding to mu opioid receptors. Interestingly, there was a significant decrease in the amount of [^2^I-Tyr^]biphalin detected in the PT with CTAP pretreatment. It is not known if the mu-opioid receptor is expressed in the

PT, but if this is true, [^^Si.i-yi-ijbjphalin could easily reach the PT from the blood to bind to opioid receptors.

In situ vascular brain perfiision experiments were performed to characterize the nature of the CNS uptake of [^-^I-Tyr']biphalin as difilisional or saturable. We 202

performed self-inhibition experiments using cold biphalin (0-100 fiM). A saturable component was identified that contributes to the brain entry of [^^Sj.-j-yj-ljbiphalin (Figure

6.3). Adding 100 |iM cold biphalin inhibited the RBrain value by 52.0%. This decrease was found to be significantly less at concentrations of 50 and 100 [oM biphalin

when compared to ['^-^I-Tyr^]biphalin perfused alone. Figure 6.4a shows that the brain

uptake consisted of both saturable and non-saturable components, which could be

described by Michaelis-Menten type kinetics with a K^i of 2.6±4.8 [.iM, Vmax

14.6±2.89 pmol"^min"^.g"^ and Kj of 0.568±.0.157 |al.min"^g"l. These kinetic

constants reflect the presence of a saturable system with a relatively high affinity and low capacity. These kinetic constants are similar to those observed for leucine-enkephalin

which has been found to enter the brain by a saturable uptake system with a relatively higii

affinity, 34-41 |iM, and a low capacity of 0.14 - 0.16 nmol.min'^.g"^. Interestingly, this

putative biphalin transporter, described in this manuscript, is not the same one described

by Zlokovic et al., 1989, for enkephalin, because 50 |jM leucine enkephalin added to the

perfusate showed no significant decrease in the RBrain value (Table 6.3). In contrast to

the brain uptake, the CSF entry of [^-^I-Tyr^]biphalin was found to be nonsaturable

diffusion (Figure 6.4b). Even with 100 |iM biphalin in the perfusate the R^SF

statistically decreased from that achieved by perfusing [^-^I-Tyr^Jbiphalin alone. The

diffusion constant, K^j, for [^^^I-Tyr^]biphalin into the CSF was 1.5±0.18 ul.min"'.g''.

Further investigations were performed to examine the acute modulation of the

putative biphalin transport system. The conformationally constrained, stable, enkephalin 203 analog, DPDPE (Toth et al., 1990; Weber at al., 1992) which has been shown to cross the

BBB in Chapters 2 and 3 of this dissertation, did not inhibit the CNS entry of

Tyr^Jbiphalin at a concentration of 100 (jM. DPDPE has been shown to enter the CNS by a saturable uptake mechanism with a of 46±28 fiM and a Vmax 51.13±13.23 pmoLmin~^.g"l (Chapter 3). The saturable component of [^-^l-Tyr^]biphalin CNS entry appears to have a higher affinity yet lower capacity when compared to the one described for DPDPE in Chapter 3, yet it does not share the same saturable transport system as

DPDPE. Additional experiments attempted to inhibit the Na"^ K"^ ATPase by adding 10 nmol ouabain to the perfusate. No significant inhibition of [l^Sj-Xyr^Jbiphalin brain entry was observed with 10 nmol ouabain. This suggests that this putative transport system at the BBB does not use the Na"^ K"^ ATPase to yield cellular energy for transport.

Attempts to interfere with cellular tubular proteins by adding 10 nmol colchicine resulted in no inhibition of transport, suggesting that the brain to blood transport of [^-^I-Tyr^Jbiphalin may not require the transcellular movement of a carrier or carrier-[l25j_Yyj.l]l3ip[jaIjn complex. This does not rule out the possibility of transcytosis, although, because this concentration of colchicine may not completely inhibit endothelial tubular proteins. Lastly, I attempted to inhibit L-type Ca^"*" channels with 10 nM verapamil or nifedipine, since calcium pools have been shown to affect tight junctions in endothelial cell membranes (Joo et al., 1992). Both experiments showed no significant inhibition in the brain entry of [^-^I-Tyrl]biphalin. These results suggest that intracellular calcium pools are not involved in the transport of [^-^I-Tyr^]biphalin across the tight 204 junctions of the BBB. It is still unclear if energy is necessary for the transport of this potent and efficacious enkephalin analog.

It was most interesting to observe that when I added BCH, a non-metabolized analog which is a substrate for the large neutral amino acid carrier, I saw a significant decrease (P<0.01) in the brain entry of [^^Sj.Xyrljbiphalin. I also observed a significant decrease (P<0.01) in the brain entry of [^^^I-Tyr^jbiphalin with 100 (iM L-phenylalanine added to the perfusate. This suggests that the large neutral amino acid carrier (Oldendorf

1971; Oldendorf and Szabo 1976) has aflfinity for this double enkephalin analog, biphalin.

This result may be due to the existence of two n-terminal tyrosines as well as two phenylalanine amino acids in the structure of this tandem enkephalin analog. Tyrosine and phenylalanine have been shown to be transported by the large neutral amino acid carrier system (Oldendorf and Szabo 1976). There was a 48 and 47 % decrease in the Rsrain for ['^2^I-Tyr^]biphalin in the presence of 10 mM BCH or 100 |.iM L-phenylalanine, respectively. Biphalin may also display affinity for the large neutral amino acid carrier because it is a relatively lipophilic peptide as shown by its relatively high octanol / saline partition coefficient of (0.2011 ± 0.0675) compared to DPDPE (0.076 ± 0.002) (Chapter

4 of this dissertation). It has been shown that the major determinant for affinity of antineoplastic drugs such as acivicin and melphalan to the large neutral amino acid carrier at the BBB is lipophilicity (Takada et al., 1991; Chikhale et al., 1995). These experiments were all conducted in the anesthetized rat which is a good model species for comparison 205 of the neutral amino acid transporter expressed at the human BBB (Hargreaves and

Pardridge 1988).

Two important characteristics of drug delivery are the ability of a given test drug to bind to serum proteins or become trapped in the endothelial cell component of the CNS vasculature. Based on protein binding studies (Table 6.4) it is apparent that

Tyr^jbiphalin is extensively protein bound to albumin in the perfusion medium (65.1%) and to rat serum proteins (90.8%). The extensive binding of [^-^I-Tyr^Jbiphalin to albumin in the perfusion medium and rat serum proteins may actually protect

Tyr^]biphalin from potential enzjmnatic degradation by systemic peptidases. Increased stability resulting in enhanced bioavailability may be another factor contributing to biphalin potency. Previous capillary depletion experiments show that the amount of

Tyr^jbiphalin accumulating in the endothelial cell component (pellet) contributes only

10% to the total brain uptake as shown in Chapter 4 of this dissertation. This shows that we are mostly measuring [^^^I-Tyr^Jbiphalin entering into the brain compartment and not the drug being trapped in the endothelial cell component of the CNS vasculature.

In conclusion, the present work has shown that the liighly potent and efficacious enkephalin analog, biphalin, can enter both spinal and supra-spinal sites that have been shown previously to express mu and delta opioid receptors. The uptake was shown to be displaceable in CNS regions using both selective and nonselective opioid antagonists, suggesting that the [^^Sj.Xyi-ljbiphalin entering the CNS is bound to both mu and delta opioid receptors. In silu brain perfusion experiments identified a saturable component that 206 contributes to the brain entry of [^^Sj.Tyj-ijjjjphaiin xhfs component can be described by

Michaelis Menten kinetics with a of 2.6 iiM, V^igx pmol"^.min"^.g"^ and Kj of 0.568 fil.niin"l.g~l. Entry into the CSF for [^25[_7yi-l]biphalin could not be self- inhibited. Further experiments revealed that [^^^I-Tyr^]biphalin was entering the CNS by the large neutral amino acid transporter and not by the leucine enkephalin uptake system or DPDPE transport system. Also the Na"^ K"^ ATPase, L-type CaP-'^ charmels, or the cellular microtubular system are not involved in generating energy to perform transcytosis.

Although it is unclear where the energy is derived to generate this saturable transport of

[I25i_TyrI]biphalin, it is clear that it is most likely similar to the L-type amino acid transport systems for large neutral amino acids which have been shown to be Na"^ independent. This work has further clarified the mechanisms contributing to biphalin potency from a drug deliver and CNS biodistribution perspective and hopefully conformational changes can be explored in the structure of opioid analogs that may promote a more efficient transmembrane movement facilitated by tliis putative transporter. 207

Chapter 7: General Discussion

The present studies demonstrate that the in vitro BMEC model of the BBB is quite reflective of the in vivo barrier that separates the blood from the central nervous system in reference to predicting a peptides relative lipophilicity. Table 2.3 showed that the overall rank order of in vitro BMEC permeability coefficient compared to the % total I.V. dose entering the brain was similar in value (r = 0.998) and when comparing in vitro permeability coefficient to capacity factor the correlation coefficient was 0.745.

However, relying solely on in vitro results to predict the in vivo situation should be performed with caution. Oftentimes, there is extensive de-differentiation of the brain capillary endothelial cell when a primary tissue culture is used. This results in some BBB characteristics potentially being lost such as specific transporters, carriers and endocytotic mechanisms. For example, several of the nutrient transport systems are downregulated by as much as 100-fold (Lin et al., 1987). Therefore, the BMEC model may be underestimating the BBB permeability of a given drug that utilizes carrier mediation to gain access to the brain. Another drawback to in vitro BMEC results is that the concentration used for HPLC quantification of a permeability coefficient is 500 ijM which is a much higher concentration than would be seen in the cerebral microvasculature.

Using such a high concentration of test drug in the BMEC model would most likely saturate any carrier present in the system. With the above limitations considered, the 208

BMEC is still useful for predicting apparent lipophilicity so that lead compounds can be identified for further analysis using more precise techniques.

The BMEC model was also characterized enzymatically. BMECs were shown to be quite active enzymatically as far as the peptidases known to be involved in the degradation of methionine enkephalin and other related peptides. High levels of membrane bound enzymatic activity were detected for total aminopeptidase, APM and

ACE and low amounts for NEP in confluent monolayers of endothelial cells. This work demonstrates that after primary culturing, cerebral endothelial cells still express the enzymes needed for the degradation of neuropeptides. Figure 2.4 shows that only a small amount of methionine enkephalin is detected in the receiver chamber after 120 minutes proving that the BMEC monolayer is active enzjmiatically. Using specific inhibitors of the en2:ymes involved in methionine enkephalin, the permeability coefficient could be improved dramatically (Table 2.6). This work clarified that the enzymes involved in peptide degradation were present at concentrations high enough to affect the permeabilit>' across the in vitro BBB. Also, the structurally stable methionine enkephalin analog,

DPDPE, had a permeability coefficient equivalent to methionine enkephalin incubated with

APM and ACE inhibitors. This proves that structural modifications that improve biological stability of a peptide also result in increased BBB permeability.

DPDPE was then further characterized for its ability to enter the CNS using the in situ brain perfusion technique. The CNS uptake of [^H]DPDPE can be separated into both brain and CSF uptake. After considering the vascular space, the brain uptake of 209

[^H]DPDPE was found to be -57% greater than the CSF uptake. The uptake of

[^H]DPDPE into the CSF was quite small in comparison the brain uptake although it was significantly greater than the CSF uptake of the vascular space marker, ['''Clsucrose The in situ brain perfusion technique also allows a researcher to investigate saturable uptake because a radiolabeled test drug is used at low concentrations. ["EijDPDPE was also found to enter the brain by both saturable and non-saturable uptake mechanisms that can be described by Michaelis-Menten type kinetics with a of 46 ± 28 uM, Vmax of 51.13 ±

13.23 pmoI.min'Vg'' and Kj of 0.56 ± 0.26 z/l.min"'.g'^ The CSF uptake of [^HJDPDPE was also shown to not be self-inhibited (Kj 0.89 ± 0.07) thus it enters the CSF purely through diffusion. The brain uptake of ["HjDPDPE was also shown not to be inhibited by either adding leucine enkephalin or BCH to the perfusate. This suggests that the saturable uptake mechanism that contributes to the CNS entry of [^H]DPDPE is not the enkephalin carrier or the large neutral amino acid carrier. It is possible that the saturable component that contributes to [^H]DPDPE CNS entry may be a specific endocytotic mechanism that was not evaluated in this work.

Structural modifications were also evaluated that have the goal of increasing the lipophilicity of enkephalin. Halogenation of enkephalin analogs has been shown to increase the brain uptake after systemic administration (Weber et al., 1991) and the specificity for delta-opioid receptors (Toth et al., 1990). In order to improve CNS entry of biphalin, a double enkephalin analog, our research group sjmthesized analogs, which had chloro- or fluoro- halogens on the para position of phenylalanine-4,4' residues. It was 210 observed that biphalin and [p-Cl-Phe^'^']biphalin can significantly enter the CNS through both the blood-brain and blood-CSF barriers. Chapter 4 showed that chloro-halogenation significantly improves CNS entry, as well as biological stability of biphalin. This would suggest that incorporation of chloro-halogens at the p-Phe'^'"^ positions of biphalin is a promising structural modification in the development of this opioid enkephalin drug for the treatment of pain.

Specific antagonists that can cross the BBB from systemic administration have not been studied as much as agonists. Chapter 5 showed that [^H]CTAP, a mu-selective antagonist can enter the CNS. The unidirectional transfer constant of [^H]CTAP into the brain and CSF was 6 and 2.4-foId higher than that calculated for [^^CJinuIin. In addition,

[•'HJCTAP was found to enter the CNS predominantly through the BBB, whereas the blood-CSF barrier played only a minor role. Also, the amount of CTAP that crosses both the blood-brain and blood-CSF barriers was quantitatively comparable to that of the efficacious, mu-selective, agonist, morphine. These observations were significant because they identified a mu-selective antagonist that crosses the BBB from a systemic administration, so that it could be used as a pharmacological tool in Chapter 6 to evaluate the basis for the analgesic potency of biphain.

Chapter 6 evaluated the highly potent and efficacious enkephalin analog, biphalin.

It was shown that biphalin can enter both spinal and supra-spinal sites that have been shown previously to express mu and delta opioid receptors. In situ brain perfusion experiments identified a saturable component that contributes to the brain entry of [^-^I- 211

Tyr'jbiphalin. This component can be described by Michaelis Menten kinetics with a of 2.6 {.iM, of 14.6 pmoI~l.min"l.g~^ and K^j of 0.568 ^il.min'^.g'^ Entry into the

CSF for [^25£_i-yi-lj5ipjjalin could not be self-inhibited. Further experiments revealed that

[i25i_Tyri]biphaIin was entering the CNS by the large neutral amino acid transporter and not by the leucine enkephalin uptake system or DPDPE transport system. Also the Na"*"

K"^ ATPase, L-type Ca^"'' channels, or the cellular microtubular system were not involved in generating energy to perform transcytosis.

In comparing the CNS entry of biphalin in Chapter 6 to the CNS entry of DPDPE in Chapter 3, several conclusions can be drawn about the saturable component to each of these compounds for CNS uptake. One, is that in both cases they do not share the same carrier as previously described for leucine enkephalin (Zlokovic et al., 1989). Second, is that the large-neutral amino acid carrier does not have affinity for DPDPE due to some component of the chemical structure. This may be the conformationally constrained component brought out by the cyclized penicillamine residues in the 2 and 5 positions.

Also, biphalin is a dimeric enkephalin analog, therefore there are two actual pharmacophores available for recognition by the large neutral amino acid carrier. In addition, biphalin is more lipophilic in comparison to DPDPE which may aid in its affinity for the large neutral amino acid carrier also.

The affinity constant of the large neutral amino acid carrier for most neutral amino acids is in the range of 10-50 |iM (Miller et al., 1985), which is approximately equal to the normal plasma concentration of large neutral amino acids. This has important clinical 212 implications, because when plasma amino acids rise after a high protein meal, the BBB large neutral mino acid carrier becomes highly saturated with substrate, and the centrally active analgesic, biphalin, may not be readily transported into the brain and have a decreased efficacy. This nutritional regulation of the brain amino acid uptake should be considered in the delivery of analgesics targeting the large neutral amino acid carrier.

Use of Xenopus oocyte expression systems for the large neutral amino acid carrier would allow for fiirther characterization of the kinetic values for affinity of biphalin to the large neutral amino acid carrier. This would allow us to more specifically characterize the affinity of biphalin for the large neutral amino acid carrier by eliminating the diffiision component. Also, it would be easier to compare the affinity of biphalin to the large neutral amino acid carrier with other large neutral amino acids.

This project has provided important preliminary work for the characterization of peptide transport into the brain. The importance of using neuropharmaceutical drug delivery vectors in modem medicine needs attention for the evolution of successful drug design targeted for CNS entry. 213

Conclusions of present research.

1. The in vitro BMEC model of the BBB is reflective of the in vivo BBB functionally. High levels of membrane bound enzymatic activity were detected for total aminopeptidase, APM and ACE and low amounts for NEP in confluent monolayers of endothelial cells. Thus, the BMEC monolayer is not only a physical barrier but it represents an enzymatic barrier. Also, peptidases active at specifc hydrolytic cleavage sites of enkephalin can be affected not only by chemical inhibitors but also by amino acid substitutions, leading to improved BBB permeability.

2. [^H]DPDPE can enter the brain by both saturable and non-saturable uptake mechanisms that can be described by Michaelis-Menten type kinetics with a of 46 ± 2S uM, Vnux of 51.13 ± 13.23 pmol.min"'.g"' and Kj of 0.56 ± 0.26 wl.min''.g''. The CSF uptake of [^H]DPDPE was also shown to not be self-inhibited (IQ 0.89 ± 0.07) thus it enters the CSF purely through diffusion.

3. Both biphalin and [p-Cl-Phe^''^']biphalin can significantly enter the CNS through both the blood-brain and blood-CSF barriers. Chloro-halogenation was shown to significantly improve CNS entry, as well as biological stability, and would suggest that incorporation of chloro-halogens at the p-Phe'^''^ positions of biphalin is a promising 214

Structural modification in the development of this opioid enkephalin drug for the treatment of pain.

4. There is a greater amount of [^HJCTAP detected in both the brain and/or CSF at all time points in comparison to [^4C]inuIin. [^H]CTAP is entering the CNS predominantly through the BBB, whereas the blood-CSF barrier plays a minor role. The amount of CTAP that crosses both the blood-brain and blood-CSF barriers is quantitatively comparable to that of the efficacious, mu-selective, agonist, morphine.

5. The highly potent and efficacious enkephalin analog, biphalin, can enter both spinal and supra-spinal sites that have been shown previously to express mu and delta opioid receptors. In situ brain perfusion experiments identified a saturable component that contributes to the brain entry of [^^^I-Tyr^]biphalin. This component can be described by

Michaelis Menten kinetics with a K^i of 2.6 (iM, of pmol"^.min"^.g"^ and K^j of 0.568 |il.min"^.g"^. Entry into the CSF for [^25i_i"yi-lj5ipf,alin could not be self- inhibited. Further experiments revealed that [^2^I-Tyr^]biphalin was entering the CNS by the large neutral amino acid transporter and not by the leucine enkephalin uptake system or DPDPE transport system. Also the Na"^ K"*" ATPase, L-type Ca^"^ channels, or the cellular microtubular system are not involved in generating energy to perform transcytosis. 215

References

Audus, K.L. and Borchardt, R-T. Characterization of an in vitro blood-brain barrier model system for studying drug transport and metabolism. Pharm. Res. 3: 81-87, 1986.

Audus, K.L. and Borchardt, R.T. Bovine brain microvessel endothelial cell monolayers as a model system for the blood-brain barrier. Ann. N.Y. Acad. Sci. 507: 9-18, 1987.

Audus, K.L. Chikhale, P.J., Miller, D.W., Thompson, S.E. and Borchardt, R.T. Brain uptake of drugs: the influence of chemical and biological factors. Adv. Drug. Res. 23: l- 64, 1992.

Ayres, E.A., Davis, P. and Burks, T.F. In vivo and in vitro investigation of naltrindole, a delta-opioid antagonist. Proc. West. Pharmacol. Soc. 33: 55-63, 1990.

Bai, J.P.F. Influences of regional differences in activities of brush-border membrane peptidases within the rat intestine on site-dependent stability of peptide drugs. Life Sciences 52: 941-947, 1993.

Banks, W.A. and Kastin, AJ. Peptides and the blood-brain barrier: Lipophilicity as a predictor of permeability. Brain Res. Bull. 15: 287-292, 1985.

Banks, W.A., Schally, A.V., Barrera, C.M. Fasold, MB., Durham, D.A., Csernus, V.J., Groot, K. and Kastin, A.J. Permeability of the murine blood-brain barrier to some octapeptide analogs of somatostatin. Proc. Natl. Acad. Sci. USA, 87: 6762-6766, 1990.

Banks, W.A. and Kastin, A.J. Bi-directional passage of peptides across the blood-brain barrier. In: Progress in Brain Research. 91: 139-148 1992.

Banks, W.A., Kastin, A.J. and Davis, T.P. Permeability of the blood-brain barrier to peptides: An approach to the development of therapeutically useful analogs. Peptides 13: 1289-1294, 1992.

Banks, W.A. and Kastin, AJ. Opposite direction of transport across the blood-brain barrier for T>T-MIF-1 and MIF-1: comparison with morphine. Peptides, 15: 23-29, 1994.

Banks, W.A. and Kastin, A.J. Passage of peptides across the blood-brain barrier: pathophysiological perspectives. Life Sciences 59: 1923-1943, 1996.

Baranczyk-Kuzma, A., Audus, K.L. and Borchardt, R.T. Catecholamine-metabolizing enzymes of bovine brain microvessel endothelial cell monolayers. J. Neurochem. 46: 1956-1960, 1986. 216

Baranczyk-Kuzma, A. and Audus, K.L. Characteristics of aminopeptidase activity from bovine brain microvessei endothelium. J. Cereb. Blood Flow Metab. 7: 801-805, 1987.

Baranczyk-Kuzma, A., Audus, K.L. and Borchardt, R.T. Substrate specificity of phenol sulfotransferase from primary cultrues of bovine brain microvessei endothelium. Neurochem. Res. 14: 1989.

Barchas, J.D. and Elliot, G.R. Neuropeptides in behavior and psychiatric syndromes: an overview. In: Neuropeptides in Neurological and Psychiatric Disease. J.B. Martina and J.D. Barchas, Editors, pp. 287-307. Raven Press, New York, 1986.

Begley, D.J. Strategies for delivering of peptide drugs to the central nervous system: exploiting molecular structure. J. Controll. Release 29: 293-306, 1994.

Benter, I.F., Hirsh, E.M., Tuchman, A.J. and Ward, P.E. N-terminal degradation of low molecular weight peptides in human cerebrospinal fluid. Biochem. Pharm. 40: 465-472, 1990.

Bertler, A., Falck, B., Owman, C.H. and Rosengrenn, E. The localization of monoaminergic blood-brain barrier mechanisms. Pharm. Rev. 18:369-385, 1966.

Bloom, F., Segal, D., Ling, N., Guillemin, R. ; profound behavioral effects in rats suggest new etiological factors in mental illness. Science 194: 630-692, 1978.

Blume, A.J., Boone G. and Lichtshtein, D. Regulation of the neuroblastoma glioma hybrid opiate receptors by Na+ and guanine nucleotides. Advances in Exp. Med. and Biology 116: 163-174.

Bolton, AE. Comparative methods for the radiolabeling of peptides. In Methods in Enzymology. Hormone Action: Part J, Neuroendocrine Peptides, ed. by P.M. Conn, 124: 18-29, 1986.

Borchardt, R.T. Assessment of transport barriers using cell and tissue culture systems. Drug Dev. and Ind. Pharm. 16: 2595-2612, 1990.

Bowman, P.D., Ennis, S.R., Rarey, K.E., Betz, A.L. and Boldstein, G.W. Brain microvessei endothelial cells in tissue culture: A model for study of the blood-brain barrier permeability. Ann. Neurol. 14: 396-402, 1983

Bradbury, M.W. The concept of a blood-brain barrier. Wiley-Interscience Publication, pp 22-27, 1979. 217

Bradford, M.M. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem. 72: 248- 254, 1976.

Brightman, M.W. The intracerebral movement of proteins injected into blood and cerebrospinal fluid of mice. Prog. Brain Res. 29: 19-40, 1968.

Brightman, M.W. and Reese, T.S. Junctions between intimately apposed cell membranes in the vertebrate brain. J. Cell. Biol. 40: 648-677, 1969.

Brightman, M.W. and Tao-Cheng, J.H. Tight junctions of brain endothelium ad epithelium. In: The Blood-Brain Barrier, W. M. Pardridge (Ed.), Raven Press, Ltd., New York, pp. 107-125.

Bunnett, N.E., Walsh, J.H. and Debas, H.T. Metabolism of enkephalin in stomach wall of rats. Am. J. Physiol. 258: G143-G151, 1990.

Bzdega, T., Chin, H., BCim, H., Jung, H.H., Kozak, C.A. and Klee, W.A. Regional expression and chromosomal localization of the delta opiate receptor gene. Proc. Natl. Acad. Sci. USA 90: 9305-9309, 1993.

Cai, R.-Z., Szoke, B., Lu, R, Fu, D., Ressing, T.W. and Schally, AV. Synthesis and biological activity of highly potent octapeptide analogs of somatostatin. Proc. Natl. Acad. Sci. USA 83: 1896-1900, 1986.

Cai, R.-Z., Karashima, T., Guoth, J., Szoke, B., Olsen, D., and Schally, AV. Superactive octapeptide somatostatin analogs containing tryptophan at position 1. Proc. Natl. Acad. Sci. USA 84: 2502-2506, 1987.

Chikhale, E.G., NG, Kay-Yun, Burton, P.S. and Borchardt, R.T. Hydrogen bonci.ig potential as a determinant of the in vitro and in situ blood-brain barrier of peptides. 11: 412-419, 1994.

Chikhale, E.G., Chikhale, P.J. and Borchardt, R.T. Carrier-mediated transport of the antitumor agent acivicin across the blood-brain barrier. Biochemical Pharmacology, 49: 941-945, 1995.

Clark, J.A., Itzhak, Y. Hruby, V.J., Yamamura, H.L and Pasternak, G.W. DPDPE: .A. delta-selective enkephalin with low aflBnity for opiate binding sites. Eur. J. Pharm. 128: 303-304, 1986.

Clough, G. and Michel, C.C. The role of vesicles in the transport of ferretin through frog endothelium. J. Physiol. 315: 127-142, 1981. 218

Collins, J.M., Klecker, R.W., Kelley, J.A., Roth, J.S., McCulIy, C.L., Balis, F.M. and Poplack, D.G. Pyrimidine dideoxynucleosides: selectivity of penetration into cerebrospinal fluid. J. Pharmacol. Exp. Ther. 245: 466-470.

Craigie, E.H. On the relative vascularity of various parts of the central nervous system on the albino rat. J. Comp. Neurol. 31:429-464, 1919.

Grain, S.M. and Shea, K.-F. .Ailer GMl ganglioside treatment of sensory neurons naloxone paradoxically prolongs the action potential but still antagonizes opioid inhibition. J. Pharmacol. Exp. Ther. 260: 182-186, 1992.

Crone, C. and Christensen, O. Electrical resistance of a capillary endothelium. J. Gen. Physiol. 77: 349-371, 1981.

Crone, C. and Olesen, S.P. Electrical resistance of brain microvascular endothelium. Brain Res. 241:49-55, 1982.

Daniel, P.M., Lam, D.K.C. and Pratt, O.E. Comparison of the vascular permeability of the brain and spinal cord to mannitol and inulin in the rat. J. Neurochem. 45: 647-649, 1985.

Davis, T.P. Methods of measuring neuropeptides and their metabolism in Hie Roles of Neuropeptides in Stress Pathogenesis and Systemic Disease (Kaufman P.G., McCubbin, J.A. andNemeroff, C.B., eds), pp 149-177. Academic Press, Orlando, Florida, 1991.

Davis, T.P. and Culling-Berglund, A. High-performance liquid chromatographic analysis of in vitro central neuropeptide processing. J. Chromatography 327: 279-292, 1985.

Davson, H., Kleeman, C.R. and Levin, E. Blood-brain barrier and extracellular space. J. Physiol. (London) 159: 67-68, 1961.

Dooley, C.T., Chung, N.N., Wilkes, B.C., Schiller, P.W. Bidlack, J.M., Pasternak, G.W. and Houghten, R.A. An all D-amino acid opioid peptide with central analgesic activity from a combinatorial library. Science 266: 2019-2022, 1994

Drouin, J and Goodman, H.M. Most of the coding region of rat ACTH3-LPH precursor gene lacks intervening sequences. Nature 288: 610-613, 1980.

Dyer, S.H., Slaughter, C.A., Orth, K., Moomaw, C.R. and Hersh, L.B. Comparison of the soluble and membrane-bound forms of the puromycin-sensitive enkephalin-degrading aminopeptidases from rat. J. Neurochem. 54: 547-554, 1990. 219

Eafhrlich P. Das SayerstofF-Bedurfins des Organismus Eine Farbenanalysiche Studie. Berline 69-72, 1885.

Erdos, E.G., Johnson, A.R. and Boyden, N.T. Inactivation of enkephalins; effect of purified peptididyl dipeptidase and cultured human endothelial cells. Advanc. Biochem. Psychopharmacol. 18: 45-49, 1978.

Erchegyi, J., Kastin, A.J., Zadina, J.E- and Qiu, X.-D. Isolation of a heptapeptide Val- Val-Tyr-Pro-Trp-Thr-GIn (Valorphin) and some opiate activity. Int. J. Pept. Protein Res. 39:477-488, 1991.

Ermisch, A. Brust, P., Ketzschmar, R. and Ruhle, H-J. Peptides and the blood-brain barrier. Phys. Rev. 73: 489-527, 1993.

Galligan, J.J., Mosberg, H.I., Hurst, R., Hruby, V.J. and Burks, T.F. Cerebral delta opioid receptors mediate analgesia but not the intestinal motility effects of intracerebroventricularly administered opioids. J. Pharm. Exp. Ther. 229: 641-648, 1984.

Gee, C.E., Chen, C-L.C., Roberts, J.L., Thompson, R. and Watson, S.J. Identification of proopiomelanocortin neurones in rat hypothalamus by in situ cDNA-mRNA hybridization. Nature 306: 374-376, 1983.

Gibson, A.M., Biggins. J.A., Lauffart, B., Mantle, D. and Mcdermott, J.R. Human brain leucyl aminopeptidase: isolation, characterization and specificity against some neuropeptides. Neuropeptides 19: 163-168, 1991.

Gilbert, P.E. and Martin, VV.R. The effects of morphine- and - like drugs in the nondependent, morphine-dependent, and -dependent chronic spinal dog. J. Pharmacol. Exp. Ther. 198: 66-82, 1976.

Gillespie, T.J., Konings, P.N.M., Merrill, B.J. and Davis, T.P. A specific enzyme assay for aminopeptidase M in rat brain. Life Sciences 51: 2097-2106, 1992.

Gjedde A. Modulation of substrate transport to the brain. Acta Neurol. Scand. 67: 3-25, 1983.

Gjedde, A. Exchange diffusion of large neutral amino acids between blood and brain. In Peptide and Amino Acids Transport Mechanisms in the Central Nervous System, ed. by L. Rakic, D.J. Begley, H. Davson and B.V. Zlokovic, 209-217, Macmillan Press, London, 1988.

Goldmann, E.E. Die Aussere und irmere Sekretion des gesunden und kranken Organismus im Lichte der 'vitalan Farbung'. Beitr. Kiin. Chir. 64: 192-265, 1909. 220

Goldmann, E.E. Vitalfarbung am Zentralenrvensystem. Abh. Preuss. Akad. Wiss. Phys. Math. KI.. 1; 1-60, 1913.

Goodman, M and Stueben, K.C. Peptide syntheses via amino acid active esters. J. Am. Chem. Soc. 81: 3980-3983, 1953.

Gulya, K., Pelton, J.T., Hruby, V.J. and Yamamura, H.I. Cyclic somatostatin octapeptide analogues with high affinity and selectivity toward mu opioid receptors. Life Sci. 38: 2221-2229, 1986.

Handa, B.K., Lane, A.C., Lord, J.A.H., Morgan, B.A., Ranee, M.J. and Smith, C.F.C. Analogues of B-LPH61-64 possessing selective agonist activity at mu-opiate receptors. Eur. J. Pharmacol. 70: 531-540, 1981.

Hargreaves, K.M. and Pardridge, W.M. Neutral amino acid transport at the human blood- brain barrier. J. Biol. Chem. 236: 19392-19397, 1988.

Hersh, L.B. Solubilization and characterization of two rat brain membrane-boimd aminopeptidases active on met-enkephalin. Am. Chem. Soc. 20: 232-237, 1981.

Heyman, J.S., Mulvaney, S.A., Mosberg, H.L, and Porreca F. Opioid 5-receptor involvement in supraspinal and spinal antinociception in mice. Brain Res. 420: 100-1 OS, 1987.

Horan, P.J., Mattie, A. Bilsky, E.J., Weber, S.J., Davis, T.P., Yamamura, H.L, Matynska, E., Appleyard, S.M., Slaninova, J., Misicka, A., Lipkowski, A.W., Hruby, V.J. and Porreca, F. Antinociceptive profile of biphalin, a dimeric enkephalin analog. J. Pharmacol. Exp. Ther. 265: 1446-1454, 1993.

Hruby, V.J., Toth, G., Gehring, C.A., Kao, L.-F., BCnapp, R., Lui, G.K., Yamamura, H.L, Galligan, J.J., Kramer, T.H., Davis, T.P., and Burks, T.F. Topographically designed analogues ofDPDPE. L Med. Chem. 34, 1823-1830, 1991.

Hughes, J., Smith, T.W., Kosterlitz, H.W., Fothergill, L.A., Morgan, B.A. and Morris, H.R. Identification of two related pentapeptides from the brain with potent opioid activity. Nature 258: 577-579, 1975.

Hussain, M.A., Rowe, S.M., Shenvi, A.B. and Aungst, B.J. Inhibition of leucine enkephalin metabolism in rat blood, plasma and tissues in vitro by an aminoboronic acid derivative. Drug Metab. Disposition 18: 288-291, 1990. 221 lannotti, F. Functional imaging of blood-brain barrier permeability by single photon emission computerized tomography and positron emission tomography. -Vdv. Tech. Stand. Neurosurg. 19: 103-119, 1992.

Janzer, R.C. and Raff, M.C. Astrocytes induce blood-brain barrier properties in endothelial cells. Nature 325: 253-256, 1987.

Joo, F., Lengyel, I., Kovacs, J. and Penke, B. Regulation of transendothelial transport in the cerebral microvessels; the role of second messenger-generating systems. Progr. Brain Res. 91; 177-187, 1992.

Karishima, T., Cai, R-Z. and Schally, A.V. Effects of highly potent octapeptide analogs of somatostatin on growth hormone, insulin and glucagon release. Life Sciences 41: 1011- 1019, 1987.

Kazmierski, W., Wire, W.S., Lui, G.K., Knapp, R.J. Shook, J.E., Burks, T.F., Yamamura, H.L and Hruby, V.J. Design and synthesis of somatostatin analogues with topographical properties that lead to highly potent and specific mu opioid receptor antagonist with greatly reduced binding at somatostatin receptors. J. Med. Chem. 31: 2170-2177, 1988.

Keep, R.F. and Jones, H.C. A morphometric study on the development of the lateral ventricle choroid plexus, choroid plexus capillaries and ventricular epyndyma in the rat. Deve. Brain Res. 56: 47-53, 1990.

Kest, B., Lee, C.E., Mogil, J.S. and Inturrisi, C.E. Blockade of morphine supersensitivity by antisense oligodeoxynucleotide targeting the delta opioid receptor (DOR-1). Life Science, 60: 155-159, 1997.

Klee, W.A. Endogenous opiate peptides. In: Peptides in Neurobiology, Current Topics in Neurobiology Series, ed. by H. Gainer. Plenum Press, New York, pp. 375-396, 1977.

Knapp, R.J. and Yamamura, H.L Delta opioid receptor radioligands. Biochem. Pharm. 44: 1687-1695, 1992.

Knapp, R.J., Malatynska, E., Fang, L., Li, X., Babin, E., Nguyen, M., Santoro, G., Varga, E.V., Hruby, V.J., Roeske, W.R. and Yamamura, H.L Identification of a human delta opioid receptor: Cloning and expression. Life Sciences 54: PL 463-469, 1994.

Koski, G. and Klee, W.A. Opiates inhibit adenylate cyclase by stimulating GTP hydrolysis. Proc. Natl. Acad. Sci. USA 78: 4185-4189, 1981. Kramer, T.H., Shook, J.E., Kazmierski, W., Ayres, E.A., Wire, W.S., Hruby, V.J.and Burks, T.F. Novel peptidic mu opioid antagonists: pharmacologic characterization in vitro and in vivo. J. Pharmacol. Exp. Ther. 249: 544-549, 1989.

Kuno, Y. and Oka, T. Estimation of relative importance of three enzymes in inactivation of Met-enkephalin-Arg6 in three isolated preparations by employing the inhibitor specific for each enzyme. Japan. J. Pharmacol. 44: 241-247, 1987.

Lamberts, S.W.J. Non-pituitary action of somatostatin. A review on the therapeutic role of SMS 201-995 (Sandostatin). Acta Endocrinol. 112; [Suppl276], 41-55, 1986.

Lamberts, S.W.J. A guide to the clinical use of the somatostatin analog SMS201-995 (Sandostatin). Acta Endocrinol. 116: [Suppl286], 54-66, 1987.

Laszlo, F., Pavo, 1., Penke, B. and Balint, G.A. Protective effect of an orally administered, highly potent somatostatin analog (RC-121) against absolute ethanol- induced hemorrhagic erosions of rat gastric mucosa. Life Sciences 44: 1573-1578, 1989.

Li, C.H. and Chung, D. Isolation and structure of an untriakontapeptide with opiate activity fi-om camel pituitary glands. Proc. Natl. Acad. Sci. USA 73: 1145-1148, 1976.

Lipkowski, A.W., Konecka, A.M. and Scoczynska, I. Double-enkephalins-syntliesis, activity on guinea-pig ileum and analgesic effects. Peptides. 3: 697-700, 1982.

Lipkowski, A.W., Konecka, A.M., Scoczynska, L, Przewlocki, R., Stala, L. and Tam, S.W. Bivalent opioid peptide analogues with reduced distances between pharmacophores. Life Sciences. 40: 2283-2288, 1987.

Llorens-Cortes, C., Huang, H., Vicart, P., Gasc, J., Paulin, D. and Corvol, P. Identification and characterization of neutral endopeptidase in endothelial cells from venous of arterial origins. J. Biol. Chem. 267: 14012-14018, 1992.

Lloyd, K.G., Davidson, L., and Homykiewicz, O. The neurochemistry of Parkinson's disease: Effect of L-dopa therapy. J. Pharmacol. Exp. Ther. 195: 453-464, 1975.

Lord, J.A.H., Waterfield, A.A,, Hughes, J. and Kosterlitz, H.W. Endogenous opioid peptides: Multiple agonists and receptors. Nature 267: 495-499, 1977.

Lowry, O.H., Rosenbrough, N.J., Parr, A.L. and Randal, R.J. Protein measurement with Folin phenol reagent. J. Biol. Chem. 193: 265-274, 1951. 223

Machen, T.E., Erlij, D. and Wooding, F3. Permeable junctional complexes. The movement of lanthanum across rabbit gall bladder and intestine. J. Cell. Biol. 54: 302- 312, 1972.

Maldonado, R., Negus, S. and Koob, G.F. Precipitation of morphine withdrawal syndrome in rats by administration of mu-, delta- and kappa-selective opioid antagonist. Neuropharm. 31; 1231-1241, 1992.

Malfroy, B., Swerts, J.P., Guyon, A., Roques, B.P. and Schwartz, J.C. High affinity enkephalin-degrading peptidase in brain is increased after morphine. Nature 2767: 523- 526, 1978.

Mansour A., Fox, C.A., Akil, H. and Watson S.J.: Opioid-receptor mRNA expression in the rat CNS: anatomical and functional implications. Trends in Neuroscience 18: 22-29, 1995.

Mansson, E., Bare, L. and Yang, D. Isolation of a human kappa opioid receptor cDNA from placenta. Biochem. and Biophys. Research Comm. 202; 1431-1437, 1994.

Martin, W.R., Eades, C.G., Thompson, J.A., Huppler, R.E. and Gilbert, P.E. The effects of morphine- and nalorphine-like drugs in the nondependent and morphine-dependent chronic spinal dog. J. Pharmacol. Exp. 197: 517-532, 1976.

Martin, W.R. and Sloan, J.W. Neuropharmacology and neurochemistry of subjective effects, analgesia, tolerance and dependence produced by narcotic analgesics. In, Handbook of Experimental Pharmacology. (Martin, W.R., ed.) 45: 43-158, 1977.

Masuzawa, T. and Sato, F. The enzyme histochemistry of the choroid plexus. Brain 106: 55-99, 1983.

Mathes, H.W.D., Maldonado, R., Simonin, F., Valverde, 0., Slowe, S., Kitchen, I., Befort, K., Dierich, A., Le Meurs, M., Dolle, P., Tzavara, E., Hanoune, J., Roques, B.P. and Kieffer, B.L. Loss of morphine-induced analgesia, reward effect and withdrawal symptoms in mice lacking the |4.-opioid-receptor gene. Nature 383: 319-323, 1996.

Matsas, R., Fulcher, I.S., Kenny, J. and Turner, A.J. and leu-enkephalin are hydrolyzed by an enzyme in pig caudate s^Tiaptic membranes that is identical with the endopeptidase of kidney microvilli. Proc. Natl. Acad. Sci. 80: 3111-3115, 1983.

Matsas, R., Kenny, A.J. and Turner, AJ. The metabolism of neuropeptides the hydrolysis of peptides, including enkephalins, tachykinins and their analogs, by endopeptidase-24.11. Biochem. J. 223; 433-440, 1984. 224

Meek, J.G., Yang, H. -Y.T. and Costa, E. Enkephalin catabolism in vitro and in vivo. Neurochemistry. 53: 1363-1371, 1977.

Miller, L.P., Pardridge, VV.M., Braun, L.D. and Oldendorf, W.H. Kinetic constants for blood-brain barrier amino acid transport in conscious rats. J. Neurochem. 45: 1427-1432, 1985.

Misicka, A., Lipkowski, A.W., Horvath, R., Davis, P., Porreca, F. Yamamura, H.I. and Hruby, V.J. Structure-activity relationships of biphalin. The synthesis and biological activities of new analogues with modifications in position 3 and 4. Life Sciences, 60: 1263-1269, 1997.

Mosberg, H.I. Hurst, R., Hruby, V.J., Gee, k., Yamamura, H.I., Galligan, J.J., and Burks, T.F. Bis-penicillamine enkephalins possess highly improved specificity toward delta opioid receptors. Proc. Natl. Acad. Sci. USA 80: 5871-5874, 1983.

Nabeshima, S., Rose, T.S., Landis, D.M. and Brightman, M.W. Junctions in the meninges and marginal glia. J. Comp. Neurol. 164: 127-170, 1975.

Nestler, E.J. Cellular responses to chronic treatment with drugs of abuse. Crit. Rev. Neurobiol. 7: 23-29, 1993.

Ohno K., Pettigrew, K.D. and Rappoport, S.I. Lower limits of cerebrovascular permeability to nonelectrolytes in the conscious rat. Am. J. Physiol. 235: H229-H307, 1978.

Oldendorf, W.H. Measurement of brain uptake of radiolabeled substances using a tritiated water internal standard. Brain Res. 24: 372-376, 1970.

Oldendorf, W. H. Brain uptake of radiolabeled amino acids, amines and hexoses after arterial injection. American Journal of Physiology 221, 1629-1639, 1971.

Oldendorf, W.H. and Brown, W.J. Greater number of capillary endothelial cell mitochondria in brain than in muscle. Proc. Sec. Exp. Biol. Med. 19: 736-738, 1975.

Oldendorf, W.H. and Szabo, J. Amino acid assignment to one of three blood-brain barreir amino acid carriers. American Journal of Physiology 230: 94-98, 1976.

Palade, G.E. Transport in quanta across the endothelium of blood capillaries. Anat. Rec. 136: 254, 1960.

Pardridge, W.M. and Mietus, L.J. Enkephalin and blood-brain barrier, studies of binding and degradation in isolated brain micro vessels. Endocrinology 109: 1138-1143, 1981. 225

Pardridge, W.M. Brain metabolism: a perspective from the blood-brain barrier. Physiol. Rev. 63: 1481-1535, 1983.

Pardridge, W.M. and Fierer, G. Blood-bram barrier transport of butanol and water relative to N-isopropyl-p-['^I] iodoamphetamine as the internal reference. J. Cereb. Blood FlowMetab. 5: 275-281, 1985.

Pardridge, W.M., Eisenberg, J. and Yamada, T. Rapid sequestration and degradation of somatostatin analogs by isolated brain microvessels. J. Neurochem. 44: 1178-1184, 1985.

Pardridge, W.M., Triguero, D., Yang, J. and Cancilla, P.A. Comparisons of in vitro and in vivo models of transcytosis through the blood-brain barrier. J. Pharmacol. Exp. Ther. 253: 884-891, 1990.

Pardridge, W.M, Boado, R.J. and Farrell, C.R. Brain-type glucose transporter (Glut-1) is selectively localized to the blood-brain barrier. J. Biol. Chem. 265: 18035-18040, 1990.

Pardridge, W.M. Transport of small molecules through the blood-brain barrier: biology and methodology. Advan. Drug Deliv. Rev. 15: 5-36, 1995.

Pardridge, W.M. Blood-brain barrier peptide transport and peptide drug delivery to the brain. In: Peptide-Based Drug Design. Eds: M.D. Taylor and G.L. Amidon. American Chemical Society 265-296, 1995.

Pasternak, G.VV. Pharmacological mechanisms of opioid analgesics. Clinical Neuropharm. 16: 1-18, 1993.

Paulus, H. A rapid and sensitive method for measuring the binding of radioactive ligands to proteins. Anal. Biochem. 32: 91-100, 1969.

Pelton, J.T., Gulya, K., Hruby, V.J., Duckies, S.P. and Yamamura, H.I. Conformationally restricted analogs of somatostatin with high mu-opiate receptor specificity. Proc. Natl. Acad. Sci. U.S.A. 82: 236-238, 1985.

Pelton, J.T., Kazmeirski, W., Gulya, K., Yamamura, H.I. and Hruby, V.J. Design and synthesis of somatostatin analogs with high potency and specificity for mu opioid receptors. J. Med. Chem. 29: 2370-2374, 1986.

Poduslo, J.F., Curran, G.L. and Berg, C.T. Macromolecular permeability across the blood-nerve and blood-brain barriers. Proc. Natl. Acad. Sci. USA. 91: 5705-5709, 1994. 226

Pollay, M.,Stevens, M., Estrada, E. and Kaplan, R. Extracorporeal perfusion of the choroid plexus. J. Appl. Physiol. 32: 612-617, 1972.

Porreca, F., Mosberg, H.I., Hurst, R., Hruby, V.J., and Burks, T.F. Roles of mu, delta, and kappa opioid receptors in spinal and supraspinal mediation of gastrointestinal transit effects and hot-plate analgesia in the mouse. J. Pharmacol. Exp. Ther. 230: 341-348, 1984.

Portoghese, P.S., Sultana, M. and Takemori, A.E. Naltrindole, a highly selective and potent non-peptide delta opioid receptor antagonist. Eur. J. Pharmacol. 146: 185-186, 1988.

Preston, J.E., Al-Sarra^ H., and SegaL, M.B. Permeability of the developing blood-brain barrier to 12C-mannitol using the rat in situ brain perfusion technique. Dev. Brain Res. 87: 69-76, 1995.

Prockop, L.D., Naidu, K.A., Binard, J.E. and Ransohoff, J. Selective permeability of [3H]-D-mannitol and [14C]-carboxyi-inulin across the blood-brain barrier and blood- spinal cord barrier in the rabbit. The Journal of Spinal Cord Medicine 18: 221-226, 1995.

Rapoport, S.I. Sites and functons of the blood-brain barrier. In: Blood-Brain Barrier in Physiology and Medicine. Raven Press, New York 43-86, 1976.

Rapoport, S.I., Pettibrew, K.D. and Ohno, K. Entry of opioid peptides into the central nervous system. Science, 207: 84-86, 1980.

Reese, T.S. and Kamovsky, M.J. Fine structural localization of a blood-brain barrier to exogenous peroxidase. J. Cell. Biol. 34: 207-217, 1967.

Reisine, T. and Bell, G.I. Molecular biology of opioid receptors. Trends Neurosci. 16: l- 18, 1993.

Saito, Y. and Wright, E.M. Bicarbonate transport across the frog and its control by cyclic nucleotides. J. Physiol. 336: 635-648, 1983.

Schally, A.V., Cai, R.-Z., Torres-Aleman, I., Redding, T.W., Szoke, B., Fu, D., Hierowski, M.T., Colaluca, J. and Konturek, S. Neural and Endocrine Peptides and Receptors, ed. Moody, T.W. (Plenum, New York), pp. 73-88, 1986.

Schweisflirth, H. and Schioberg-Schiegnitz, S. Assay and biochemical characterization of angiotensin-converting enzyme in cerebrospinal fluid. Enzyme 32: 12-19, 1984.

Scriba, G.K.E., Borchardt, R.T. J. Neurochem. 53: 610, 1989. 227

Sharif, N.A. and Huches, J. Discrete mapping of brain mu ad delta opioid receptors using selective peptides: Quantitative Autoradiography, species dififerences and comparison with kappa receptors. Peptides 10: 499-522, 1989.

Shibanoki, S., Weinberger, .B., Ishikawa, K. and Martinez, J.L., Jr. Further charaterization of the in vitro hydrolysis of [Leu] and [Met] enkephalin in rat plasma: HPLC-ECD measurement of substrate and metabolite concentrations. Regul. Peptides 32: 267-268, 1991.

Shook, J.E., Pelton, J.T., Lemcke, P.K., Porreca, F., Hruby, V.J. and Burks, T.F. VIu opioid antagonist properties of a cyclic somatostatin octapeptide in vivo: identification of mu receptor-related functions. J. Pharmacol. Exp. Ther. 242: 1-7, 1987.

Shook, J.E., Pelton, J.T., Hruby, V.J. and Burks, T.F. Peptide opioid antagonist seperates peripheral and central opioid antitransit effects. J. Pharmacol. Exp. Ther. 243: 492-500, 1987.

Silbert, B.S., Lipkowski, A.W., Cepeda, M.S., Szyfelbein, S.K., Osgood, P.F. and Carr. D.B. Analgesic activity of a novel bivalent opioid peptide compared to morphine via different routes of administration. Agents and actions 33: 382-287, 1991.

Simionescu, N., Simionescu, M. and Palade, G.E. Permeability of muscle capillaries to small hemepeptides. J. Cell Biol. 64: 586-607, 1975.

Silbert, B.S., Lipkowski, AW., Cepeda, M.S., Szyfelbein, S.K., Oscood, P.F. and Carr, D.B. Analgesic activity of a novel bivalent opioid peptide compared to morphine via different routes of administration. Agents Actions 33, 382-387, 199L

Smith, Q.R., Takasato, Y. and Rapoport, S.L Kinetic analysis of L-leucine transport across the blood-brain barrier. Brain Res. 311: 167-170, 1984.

Spector R. and Johanson, C.E. The mammalian choroid plexus. Sci. Amer. Nov: 68-74, 1989.

Stevens, C.W. Perspectives on opioid tolerance from basic research: Behavioral studies after spinal administration in rodents. Cancer Surveys 21: 25-47, 1994.

Stevens, C.W., and Yaksh, T.L. Magnitude of opioid dependence afler continuous intrathecal infusion of mu- and delta-selective opioids in the rat. European Journal of Pharmacol. 166, 467-472, 1989. 228

Takada, Y., Greig, N.H., Vistica, D.T., Rapoport, S.I. and Smith, Q.R. Affinity of antineoplastic amino acid drugs for the large neutral amino acid transporter of the blood- brain barrier. Cancer Chemother. Pharmacol. 29: 89-94, 1991.

Takasato, Y., Rapoport, S.I. and Smith, Q.R. An in situ brain perfusion technique to study cerebrovascular transport in the rat. Am. J. Physiol. 247: H484-H493, 1984.

Takasato, Y., Momma, S. and Smith, Q.R. Kinetic analysis of cerebrovascular isoleucine transport form saline and plasma. J. Neurochem. 45: 1013-1020, 1985.

Terenius, L. Somatostatin and ACTH are peptides with partial antagonist-like selectivity for opiate receptors. Eur. J. Pharmacol. 38: 211-215, 1976.

Terashima, H. Okamoto, A., Menozzi, D., Goetzl, E.J. and Bunnett, N.W. Identification of neuropeptide-degrading enzymes in the pancreas. Peptides 13: 741-748, 1992.

Thompson, S.E. and Audus, K.L. Leucine-enkephalin metabolism in brain microvessel endothelial cells. Peptides 15: 109-116, 1994.

Toth, G., Kramer, TJI., Knapp, R., Lui, G., Davis, P., Burks, T.F., Yamamura, H.I. and Hruby, V.J. DPDPE analogs with increased affinity and selectivity for delta opioid receptors. J. Med. Chem. 33: 249-253, 1990.

Triguero, D., Buciak, J. and Pardridge, W.M. Capillary depletion method for quantification of blood-brain barrier transport of circulating peptides and plasma proteins. J. Neurochem. 54: 1882-1888, 1990.

Vlasses, P.H., Ferguson, R.K., Chatteijee, K. Captopril: Clinical pharmacolog}' and benefit-to-risk ration in hypertension and congestive heart failure. Pharmacotherapy 2: l- 17, 1982.

Vonvoigtlander, P.P., Lahti, R.A. and Ludens, J.H. U-50,488: A selective and structurally novel non-mu (kappa) opioid agonist. J. Pharmacol. Exp. Ther. 224: 7-12, 1983.

Wang, Z., Bilsky, E.J., Porreca, F. and Sadee, W. Constitutive mu opioid receptor activation as a regulatory mechanism underlying narcotic tolerance and dependence. Life Sciences. 54 PL339-350, 1994.

Watson S. J., Barchas, J.D. and Li, C.H. P-: Locali2:ation of cells and axons in rat brain by immunocytochemistry. Proc. Natl. Acad. Sci. USA 11: 5155-5158, 1977.

Weber, S.J., Greene, D.L., Sharma, S.D., Yamamura, H.I., Kramer, T.H., Burks, T.F., Hruby, V.J., Hersh, L.B., and Davis T.P. Distribution and analgesia f [3H]DPDPE and 229 two halogenated analogs after intravenous administration. J. Pharm. Exp. Ther. 259; 1109-1117, 1991.

Weber, S.J., Greene, D.L., Hruby, V.J., Yamamura, H.I., Porreca, F., and Davis, T.P. Whole body and brain distribution of [3HJDPDPE after intraperitoneal, intravenous, oral and subcutaneous administration. J. Pharmacol. Exp. Ther. 263, 1308-1316, 1992.

Weston, P.G. Sugar content of the blood and spinal fluid of insane subjects. J. Med. Res. 35: 199-207, 1916.

Yoshida, T. and Nosaka, S. Some characteristics of a peptidyl dipeptidase (kinainase 11) from rat CSF. J. Neurochem. 55; 1861-1869, 1990.

Zadina, J.E., Heckler, L., Ge, L-J-, and Kastin, A.J. A potent and selective endogenous agonist for the |.i-opiate receptor. Nature 386; 499-502, 1997.

Zlokovic, B.V., Begley, D.J., Djuricic, B.M. and Mitrovic, D.M. Measurement of solute transport across the blood-brain barrier in the perfused quinea pig brain: Method and application to N-methyl-alpha-aminoisobutyric acid. J. Neurochem. 46, 1444-1451, 1986.

Zlokovic, B.V., Mackic, J.B., Djuricic, B.M. and Davson, H. Kinetic analysis of leucine- enkephalin cellular uptake at the luminal side of the blood-brain barrier of an in situ perfused quinea-pig brain. J. Neurochem. 53: 1333-1340, 1989.

Zlokovic, B.V., Hyman S., McComb, J.G., Lipovac, M.N., Tang, G. and Davson, H. Kinetics of arginine- uptake at the blood-brain barrier. Biochim. Biophys. Acta. 1025: 191-198, 1990.

Zlokovic, B.V., Banks, W.A., Kadi, H.E., Erchegyl J., Mackic, J.B., McComb, J.G. and Kastin, A. J. Transport, uptake and metabolism of blood-boume vasopressin by the blood- brain barrier. Brain Res. 590: 213-218, 1992.

Zlokovic, B.V. Cerebrovascular permeability to peptides: Manipulations of transport systems at the blood-brain barrier. Pharmac. Res. 12: 1395-1406, 1995.