University of Groningen

The medicinal chemistry of aryl triflates Barf, Tjeerd Andries

IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below. Document Version Publisher's PDF, also known as Version of record

Publication date: 1996

Link to publication in University of Groningen/UMCG research database

Citation for published version (APA): Barf, T. A. (1996). The medicinal chemistry of aryl triflates: as applied to 5-HT1A and 5-HT1D ligands. s.n.

Copyright Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).

Take-down policy If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.

Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.

Download date: 23-09-2021 The Medicinal Chemistry of Aryl Triflates as applied to 5-HT 1A and 5-HT 1D Receptor Ligands RIJKSUNIVERSITEIT GRONINGEN

The Medicinal Chemistry of Aryl Triflates as applied to 5-HT 1A and 5-HT 1D Receptor Ligands

PROEFSCHRIFT

ter verkrijging van het doctoraat in de Wiskunde en Natuurwetenschappen aan de Rijksuniversiteit Groningen op gezag van de Rector Magnificus, dr. F. van der Woude, in het openbaar te verdedigen op vrijdag 4 oktober 1996 des namiddags te 2.45 uur

door

Tjeerd Andries Barf

geboren op 18 april 1967 te Groningen Promotor Prof. Dr. H. V. Wikström

Co-promotor Dr. C. J. Grol Life is like a box of chocolate.... You never know what you gonna get

Forrest Gump

Voor mijn ouders, Arend & Wimmy Promotiecommissie Prof. Dr. D. Nichols Prof. Dr. B. Olivier Prof. Dr. B.L. Feringa

Paranimfen Lynet Rooks Arwin Ridder

ISBN 90 367 0681 5 NUGI 746

Cover design : Tjeerd Barf / “The only antimigraine agent that gives you the headache.” The 3-D stereogram was generated with Crosseye, available on Internet Printing : PrintPartners Ipskamp bv

An electronic version of this thesis in Adobe PDF-format is available on the Internet: WWW : http://docserver.ub.rug.nl/eldoc/dis/science/t.a.barf Gopher : gopher://docserver.ub.rug.nl/11/eldoc/dis/science/t.a.barf Chapter 1...... 1

Introduction

1.1 History...... 1 1.2 Receptors...... 1 Classification...... 1 Evolutionary Perspectives...... 3

1.3 5-HT1A Receptors...... 4 Distribution and Function ...... 4 Structural Aspects...... 5

Pharmacophore of the 5-HT1A Receptor...... 6

1.4 5-HT1A Receptor and Antagonists (SAR)...... 7 Indolealkylamines ...... 7 Aminotetralins and Analogues ...... 8 Arylpiperazines and Analogues ...... 11 Aryloxyalkylamines...... 12

1.5 5-HT1D Receptors...... 13 Distribution and Function ...... 13 Structural Aspects...... 15

The 5-HT1D Receptor Pharmacophore...... 16

1.6 5-HT1A Receptor Agonists and Antagonists (SAR)...... 16 1.7 Objective and Outline...... 20 1.8 References...... 23

Chapter 2...... 27

Synthesis and Pharmacological Evaluation of 2-Aminotetralin Derivatives

2.1 Introduction...... 27 2.2 Chemistry...... 29

Preparation and Resolution of (±)-8-OSO2CF3-PAT...... 29

Preparation and Resolution of Cis-(±)-8-OSO2CF3-MPAT...... 31

Preparation of Cis- and Trans-8-OSO2CF3-MMAT...... 32 Cis- and Trans-assignment...... 34 2.3 ...... 35 Receptor Binding...... 35

In Vivo Biochemistry of (R)- and (S)-8-OSO2CF3-PAT ...... 35 Locomotor Activity and Gross Behavioural Observations ...... 37

Oral Bioavailability and In Vitro of (R)-8-OSO2CF3-PAT ...... 38 Hypothermia...... 38 2.4 Results and Discussion...... 39 Structure-Affinity Relationships...... 39 Structure-Activity Relationships...... 39 2.5 Experimental Section...... 42 2.6 References...... 52

Chapter 3...... 53

Potential Properties of ( R)-8-SO 2CF 3-PAT

3.1 Introduction...... 53 3.2 Results...... 54 Conditioned Defensive Burying...... 54 Conditioned Fear of Footshock...... 55 Elevated Plus-maze ...... 55 Effect on the 5-HIAA/5-HT ratio in Various Rat Brain Regions...... 56 3.3 Discussion ...... 56 3.4 Conclusion...... 59 3.5 Experimental Section...... 60 3.6 References...... 63

Chapter 4...... 65

5-HT 1D Receptor Properties of Novel 5- [[(Trifluoromethyl)sulfonyl]oxy]indolylethylamines

4.1 Introduction...... 65 4.2 Chemistry...... 67 4.3 Pharmacology...... 69 Receptor Binding...... 69 cAMP Assay ...... 69 In Vivo Biochemistry...... 69 Hypothermia...... 71 4.4 Results and Discussion...... 71 Structure-Affinity and Structure-Activity Relationships...... 71 Pharmacology...... 72 4.5 Experimental Section...... 74 4.6 References...... 83

Chapter 5...... 85

5-(Sulfonyl)oxy-tryptamines and Ethylamino Side Chain Restricted Derivatives

5.1 Introduction...... 85 5.2 Chemistry...... 87 Preparation of 5-Sulfonyloxytryptamines...... 87 Preparation of 3-Aminocarbazoles...... 90 Preparation of Indol-3-ylpiperidines...... 91 5.3 Pharmacology...... 91 Receptor Binding...... 91 5.4 Results and Discussion...... 91 5-Sulfonic Acid Ester Derivatives of 5-HT ...... 91 Alkylamino Side Chain Restriction ...... 93 5.5 Experimental Section...... 95 5.6 References ...... 10 2

Chapter 6...... 103

Structure-Affinity and Structure-Activity Relationships of Ortho- Substituted Phenylpiperazines

6.1 Introduction...... 103 6.2 Chemistry...... 105 6.3 Pharmacology...... 107 Receptor Binding...... 107 cAMP Assay ...... 107 In Vivo Inhibition of Lower Lip Retraction...... 108 6.4 Results and Discussion...... 108 6.5 Experimental Section ...... 11 3 6.6 References ...... 11 5 Chapter 7...... 117

Selctive 5-HT 1A Receptor Ligands for PET; A Comparative Study of [11C]ORG13502 and [ 11C]WAY100635 in Normal and Adrenalectomized Rats

7.1 Introduction...... 117 7.2 Chemistry...... 118 7.3 Pharmacology...... 119 Tissue Distribution Studies...... 119 Metabolism of [11C]ORG13502 ...... 119 7.4 Results and Discussion...... 119 Chemistry...... 119 Pharmacology...... 120 7.5 Experimental Section ...... 12 4 7.6 References ...... 12 6

Conluding Remarks ...... 12 7

Samenvatting ...... 12 9 Toelichting voor Niet-Farmacochemici ...... 13 3

List of Publications ...... 13 7

Tot Slot ...... 13 9 Chapter 1

Introduction

1.1 History

Serotonin (5-hydroxytryptamine, 5-HT, 1), a neurotransmitter present in most species including man, plays a role in a variety of physiological functions such as pain, appetite, sex, emotion, sleep, memory and their associated pathological states.1 5-HT was isolated for the first time from the in 1948 and characterized shortly thereafter by Rapport et al.,2 who named the endogenous compound after its source (serum) and action (tonus). This ‘peripheral ’ was also found in the intestinals but had yet another accommodation, the brain. This finding, and the synthesis of 5-HT in 1951, initiated an extensive research on the functioning and dysfunctioning of (central) systems. Although less than 5% of the total amount of 5-HT in the body resides in the central (CNS), 5-HT is an important factor in normal brain function. Serotonin was recognized to be a neurotransmitter substance.3 In 1957, Gaddum and Picarelli4 found indications of more than one 5-HT OH receptor site. According to different antagonizing effects of dibenzyline and morphine on the action of 5-HT in smooth muscle NH2 cells in guinea-pig ileum on the one hand, and acetylcholine

H release on the other, the receptors were initially designated as D- H and M-receptors, respectively. The introduction of in vitro Serotonin (1) radioligand binding techniques in 1979 by Peroutka and Snyder allowed accurate discrimination between various 5-HT receptor subtypes on the basis of different ligand binding characteristics.5

1.2 Serotonin Receptors

Classification. Molecular biology techniques accounted for the cloning of 5-HT receptor genes, and it became evident that multiple subtypes of this receptor protein exist. Recently, Hoyer et al.6 proposed operational (selectivity and affinity for agonists and antagonists), structural (protein homology) and transductional (intracellular mechanisms) criteria which a receptor should meet in order to become part of the 5-HT

1 Chapter 1 receptor superfamily. To date, 5-HT receptors can be classified to at least three, possibly seven, groups of receptors. They comprise the 5-HT1, 5-HT2 and 5-HT3 classes, as well as the uncloned 5-HT4 receptor. The 5-ht5, 5-ht6 and 5-ht7 receptor genes have been cloned but these so-called ‘orphan’ receptors have yet to be fully characterized with respect to their pharmacological function and selectivity for certain . All 5- HT receptor (sub)types belong to the G-protein coupled receptor superfamily, except the 5-HT3 receptor, which is a ligand gated ion channel. In Table 1.1, the nomenclature and characteristics of the 5-HT receptor subtypes are summarized.

2 Introduction

Table 1.1. Current Status of 5-HT Receptor Characteristics

Subtype Location Response Agonists Antagonists Clinical Implication

5-HT 1A Mainly CNS Neuronal 8-OH-DPAT WAY100635 Anxiety, hyperpolarization, buspirone, 5-CT depression hypotension

5-HT 1B CNS and some Neurotransmitter CP93129, 5-CT SDZ21009 Appetite peripheral nerves release ↓ disorders

5-HT 1D α Mainly CNS Neurotransmitter Sumatriptan GR127935 Migraine, release ↓ L694247, 5-CT GR55562 depression

5-HT 1D β Mainly CNS Neurotransmitter Sumatriptan GR127935 Migraine, release ↓ L694247, 5-CT depression

5-HT 1E Only CNS cAMP ↓ 5-HT None ??

5-HT 1F Mainly CNS cAMP ↓ 5-HT None ??

5-HT 2A Vascular smooth Vasoconstriction, α-Methyl-5-HT Ketanserine Sexual and muscle, , DOI cinanserine sleep disorders lung, CNS, gastro aggregation, pirenperone intestinal tract bronchoconstrictio n

5-HT 2B Mainly peripheral? Rat stomach fundic α-Methyl-5-HT SB200646 ?? muscle contraction DOI

5-HT 2C CNS Phosphoinositide α-Methyl-5-HT Mesulergine Anxiety, turnover ↑ DOI migraine

5-HT 3 Peripheral and Depolarization 2-Methyl-5-HT Ondansertron, Emesis, anxiety, central neurones m- tropisetron depression, chlorophenyl- memory biguanide disorders

5-HT 4 Gastrointestinal acetylcholine Metoclopramid GR113808 ?? tract, CNS, heart, release in gut↑, erenzapride SB204070 urinary bladder tachycardia, cAMP Cisapride ↑ in CNS neurones

5-ht5A CNS ?? 5-HT Methiothepin ?? and

5-ht 5B

5-ht6 CNS cAMP ↑ 5-HT Methiothepin Anxiety?

5-ht7 CNS cAMP ↑ 5-HT Methiothepin ??

Adapted with minor modifications from refs.1 , 5 and 6. Orphan receptors are denoted in small characters (see text).

Evolutionary Perspectives. Protein receptors that mediate the actions of 5-HT have existed in the membranes of a variety of animal cell types for millions of years, which likely explains the multiplicity of these receptors. Peroutka and Howell7

3 Chapter 1 performed a molecular evolution analysis which was based on the amino acid sequence homology of 5-HT and other G-protein coupled receptors. By correlating the percent amino acid homology between various species and the dates of evolutionary divergence of the species, the rate of molecular evolution was estimated to be approximately 1% every 10 million years.

5-HT1B.rat

5-HT1Dβ.human

5-HT1Dα.rat

5-HT1Dα.human

5-HT1F.rat

5-HT1F.human

5-HT1E.human

5-HT1A.rat1

5-HT1A.rat2

5-HT1A.human

5-HT7.rat

5-HT5B.rat

5-HT5A.rat

5-HT2A.rat

5-HT2A.human

5-HT2C.rat

5-HT2C.human

5-HT2B.rat

5-HT6.rat

700 600 500 400 300 200 100 0 Millions of Years Ago

Figure 1.1. A phylogenetic tree of 5-HT receptors, adapted from Peroutka and Howell.7 The length of each ‘branch’ of the tree correlates with the evolutionary distance between receptor subpopulations. 5-HT subpopulations of other species have been excluded for the sake of clarity

The sequences, which were aligned according to the method of Feng and Doolittle,8 were used to construct a phylogenetic tree (Figure 1.1). The lenght of each

4 Introduction

‘branch’ correlates with the evolutionary distance between receptor subpopulations. Data indicate that the ‘primordial’ 5-HT receptor evolved over 750 million years ago. During such a long period of time, there has been ample opportunity for mutation and consequent evolutionary acceptance of multiple variants of receptors for the neurotransmitter 5-HT. The number of 5-HT receptor subtypes partly explains the amount of pathological pathways 5-HT is involved in. Although each receptor can be potently activated by 5-HT itself, the differences in protein structure offer medicinal chemists opportunities to design selective ligands for each receptor subtype. The challenge for (molecular) pharmacologists is to define the role of each receptor and to elucidate their function and distribution. These efforts should provide (selective) 5-HT receptor ligands with therapeutic utility and a better understanding of the clinical relevance of each receptor variant and vice versa.

1.3 5-HT 1A Receptors

Distribution and Function. The distribution and function of central 5-HT1A receptors has been extensively studied in a number of vertebrates the past two decades.

Among 5-HT receptors, the 5-HT1A receptor became the first and the best characterized 9,10 receptor due to the development of high affinity 5-HT1A receptor ligands. Autoradiography studies revealed common distribution patterns in the brain of rats, guinea-pigs, cats, primates and humans.11 Recently, these patterns were confirmed with PET-studies utilizing radiolabelled WAY100635,12 which is the first silent and selective 13 5-HT1A (for chemical structure see section 1.4). High densities of

5-HT1A receptors were shown to be located in the raphe nuclei and in limbic structures such as hippocampus, lateral septum and amygdala. The cerebellum was reported to be 14 essentially devoid of 5-HT1A receptors.

The presence of 5-HT1A receptor populations in the dorsal and median raphe nuclei indicates that 5-HT can modulate the activity of serotonergic neurones.

Activation of these somatodendritic 5-HT1A receptors causes inhibition of the cell firing activity and consequently reduction of 5-HT synthesis and neurotransmission in 15 terminal brain areas. On the other hand, activation of the postsynaptic 5-HT1A receptors results in neuronal inhibition of particular parts of the limbic system, an area which has been implicated in the modulation of emotion. This ‘double-face’ character of the 5-HT1A receptor makes this receptor an interesting therapeutic target in the treatment of mood-disorders such as anxiety and depression (Figure 1.2). It has been suggested that the presence of a large receptor reserve at somatodendritic receptors and the lack of a receptor reserve at postsynaptic receptors,16 combined with the intrinsic

5 Chapter 1

activity of 5-HT1A receptor ligands at these receptors, may determine the anxiolytic/ profile.17

Trp 5-HTP 5-HT

5-HT1A

5-HT1D

Somatodendritic Presynaptic Postsynaptic

Figure 1.2. Schematic representation of a 5-HT neurone with 5-HT1A and 5-HT1D receptors

Typically, activation of central 5-HT1A receptors causes a lower lip retraction (LLR)18 and the so-called ‘5-HT syndrome’ in rats, which is characterized by flat body posture, abducted hind-limbs, fore-paw treading (piano playing) and head weaving.19

Several pharmacological studies performed with 5-HT1A receptor agonists showed effects on body temperature, feeding behaviour and sexual activity, which all could be effectively blocked by antagonists. A variety of 5-HT1A receptor agonists produce anxiolytic effects in animals models of anxiety, although the clinically effective compounds (such as buspirone) show additional anti- activity.1b

Structural Aspects. The gene encoding the human 5-HT1A receptor TM 4 was successfully cloned in 1987 by TM 3 20 Kobilka et al., although its identity (Asp116) H remained unelucidated until 1988.21 O TM 2 O 198 O (Ser ) H H (Asp82) The corresponding protein consists of NH3 TM 5 O H O 421 amino acid residues, which were (Thr199) O O shown to contain seven hydrophobic H TM 1 N I regions, reflecting seven O (Ser392) transmembrane spanning regions H TM 6 TM 7 (TMRs), in analogy with the structure of the G-protein coupled receptor 22 bacteriorhodopsin. The rat 5-HT1A Figure 1.3. Schematic representation of possible receptor gene has also been cloned and neurotransmitter-receptor interactions.

6 Introduction expressed,23 having 99% sequence homology with the human equivalent in the putative TMRs. Presumably, the TMRs are arranged in α-helices, which are connnected with intra- and extracellular loops. According to mutagenicity experiments, certain conserved carboxylate and hydroxy bearing residues are important for the binding of a 82 82 5-HT1A receptor agonist. Amino acid substitution in the TMRs, Asp →Asn (TM2), Asp116→Asn116 (TM3) and Ser198→Ala198 (TM5) all resulted in a decrease in affinity for 5-HT by 60-100 fold, whereas mutant Thr199→Ala199 (TM5) showed virtually no binding (Figure 1.3).24 Another mutation study revealed that Ser392 and Asn395 on TM7 may also be crucial for ligand binding.25 In summary, the agonist binding is proposed to be facilitated through ion-pair formation between the protonated amino group of the ligand and the carboxylate groups of the aspartate residues in the 2nd and 3rd helix, and an interaction between the hydrogen-bonding group (hydroxy) and the hydroxy group of a serine or threonine residue (Figure 1.4). In addition, the interaction is probably stabilized by (π-π interactions with) surrounding aromatic groups.26

Extracellular

I Q C G D C P T S Y D L L H G F T L L L A F M P Y P L L L A I P T V G L G M A L I I L A I T P L D S I Y S A V I L I V V L T F I N W F L L C F G F F L C V S C G I A P G A V T L F L Y V L M S S W Y I W C S N G L D I T L P L S N T L S L I L A C V H L I L F L V A L C L A L M T G N P V A S A I R L M V I A G I A P P V L I I Y I L L T Y G A TM1 TM2 TM3 TM4 TM5 TM6 TM7

Intracellular

Figure 1.4. Representation of the human 5-HT1A receptor showing 7 putative trans-membrane regions, embedded in the lipid bilayer (gray). The TM-regions are connected by intra- and extracellular loops. The dark-grey amino acid residues have been implicated in receptor-ligand interactions.

Pharmacophore of the 5-HT 1A Receptor. Taking these considerations into

+β -β 5.37 Å N α 0.20 Å 2.1-2.6 Å

5.2-5.7 Å

Figure 1.5. Two pharmacophoric models for 5-HT1A receptor agonists proposed by Hibert et al. (left) and Mellin et al. (right). Adapted from refs.28 and 29, respectively.

7 Chapter 1

account, several groups have attempted to define the pharmacophore of a 5-HT1A receptor agonist or antagonist.27,28,29,30 Mostly, common structural elements of minimized (or crystal) conformations of known semi-rigid 5- O α 2.6 Å N HT1A receptor ligands, such as 2-aminotetralins (see Section 1.4), were fitted in order to find this N pharmacophore. According to the resulting models, pharmacophoric elements are an aromatic plane and (a specific direction of) the lone pair of a nitrogen atom at a fixed distance Figure 1.6. Model proposed by Nilsson et al. of approximately 5-6 Å from the midpoint of the aromatic plane (Figure 1.5). Nilsson et al. stressed the importance of the hydrogen bond accepting oxygen atoms by fitting several hydroxy and methoxy analogues of 8-OH-DPAT,31 and developed a model with two different nitrogen lone-pairs or nitrogen dummy sites at a distance of 2.6 Å (Figure 1.6). This model resembles the Mellin model but additionally explains the affinity of some partial 5-HT1A receptor agonists. However, the structure- activity relationship (SAR) is complicated and no model has been able to accommodate

5-HT1A receptor ligands from different structural classes.

1.4 5-HT 1A Receptor Agonists and Antagonists (SAR)

A variety of compounds, representing different structural classes, have been 32 presented in the literature to have considerable affinity for 5-HT1A sites with varying degrees of intrinsic efficacy. The following section deals with the most important structures, including some structure-affinity-relationships (SAR).

Indolealkylamines. Modification of serotonin (1), the endogenous 5-HT1A receptor ligand, was the first approach for a number of groups. Removal of the hydroxy moiety resulted in a 50 fold decrease in affinity and moving the hydroxy from the 5- to the 4- or 6-position had a similar effect.33 Dimethyl- and di-n-propylation of the primary amine group resulted in slightly lower affinities but neither compound displayed 34 selectivity for 5-HT1A vs 5-HT2 sites. Several tryptamine derivatives were shown to level the affinity of 5-HT as exemplified by 5-carboxamidotryptamine (5-CT, 4) and its N,N-di-n-propyl analogue DP-5-CT (5). Conformationally restricted tryptamine analogues such as RU28253 (6) and RU24969 (7) were shown to have good affinity for

8 Introduction

the 5-HT1A receptor, but also were reported to show considerable affinity for the 5-HT1B sites.35 All tryptamine derivatives reported to date are agonists.

H2N O OMe X 5 6 4 NR 2 NR2 7 X 3 Y 1N 2 N N H H H

1 X = 5-OH, R = H 4 R = H 6 X = NH, Y = CH2 2 X = 4-OH, R = H 5 R = n-Pr 7 X = CH2, Y = NH 3 X = 6-OH, R = H The tetracyclic ergolines constituted a special class of serotonergic agents, which possess high affinity but low selectivity for 5-HT receptor subtypes.36 (+)-LSD

(8) shows a Ki of 2.6 nM for 5-HT1A sites and, when tritiated, has proven its use as a 5 radioligand for 5-HT1 and 5-HT2 receptor types. Molecular modification of the ergoline framework has led to the development more selective compounds such as 10, showing 37 an IC50 value of 5 nM for 5-HT1 sites. Interestingly, trans-(±)-2,3- dihydrofestuclavines, such as compound 11, are essentially inactive stressing the importance of the double bond in these type of structures. Omission of the D-ring of the ergoline skeleton led to the development of the (S)-enantiomer of compound 12 (LY228729),38 which is a fused structure of an ergoline and DP-5-CT and displays high affinity (Ki = 0.13 nM) and selectivity for the 5-HT1A receptor but was recently withdrawn from the clinic due to adverse effects.39

O NR1R2 Me Me

H2N O

D N N N N Me Me Me A C

B N N N N

R3 H H H 11 12 8 R1 = R2 = Et; R3 = H 10 9 R1 = 2-(butanol); R2 = H; R3 = Me

Aminotetralins and Analogues. Beside an indolealkylamine structure, the above compounds also possess a 2-aminotetralin moiety within their multicyclic framework. In 1981, Arvidsson et al. provided a breakthrough in the search for selective 5-HT receptor ligands with 8-hydroxy-N,N-di-n-propyl-2-aminotetralin (8-

9 Chapter 1

OH-DPAT, 13), which was the first nonindole-containing agent with full agonist properties.40 It was reported to induce the 5-HT behavioural syndrome and to decrease the cerebral 5-HT turnover very potently. But it was not until 1984, after extensive pharmacological evaluations, that the 5-HT1A receptor was found to be the mediator of these effects.41 Ever since this finding, 8-OH-DPAT (Ki = 1.2 nM) has been serving as an important pharmacological and structural tool in the development of novel 5-HT1A receptor agonists. Due to poor pharmacokinetic properties, 8-OH-DPAT itself failed to be of any clinical interest. Thorberg et al. prepared racemic 3-aminochromanes, which were predicted to have better brain penetration than 8-OH-DPAT. 5-OH-DPAC (14) levelled the affinity of 8-OH-DPAT for 5-HT1 sites but was less potent in the inhibition of 5-HTP accumulation.42 The (R)-enantiomer of its orally available (21% in the cat) carboxamido congener (Ebaltozan,15) displays a Ki value of 8.5 nM for the 5-HT1A site and is subjected to clinical trials.43

OH R1 R2 OH N N Me

O N 13 14 R1 = OH, R2 = n-Pr 16 Me 15 R1 = CONH(i-Pr), R2 = i-Pr

Although aporphines are typically associated with dopaminergic activity, (R)-(− )-10-methyl-11-hydroxyaporphine (16) unexpectedly was shown to be a high affinity 44 5-HT1A receptor agonist. It is difficult to reconcile the affinity of 16, since it bears the hydroxy group in the ‘wrong’ position, compared to 13. (1S,2R)-Cis-1-methylated 2-aminotetralin derivative (17) was shown to be equipotent to 13. The (R)-enantiomer of 13 is only twice as potent as the (S)- enantiomer, whereas the (1R,2S)-antipode of 17 and the respective trans-isomers are inactive.45,46 This improvement in stereoselectivity has prompted considerable structure activity work. Comparison of semi-rigid cis- and trans-octahydrobenzoquinoline (OHBQ) derivatives in binding, biochemical assays and conformational calculations (MM2) led to the observation that trans-(4aS,10bS)-isomer 18 was the most active one (Ki = 3.87 nM).47 The nitrogen lone-pair is in the opposite direction, as compared to compound 17, but this fits in the pharmacophore model of Nilsson et al. (see Section 1.3). In 5-membered fused ring-systems, the cis-isomers were the most active, as exemplified by benz[e]indole derivative 19 and the orally active analogue 20 (Ki values 0.1 and 1.9 nM, respectively). The activities of these compounds reside in the

10 Introduction enantiomers which have the same configuration at the carbon resembling the 2-position of compound 17.48 It should be noted that the C-1 methyl group of compound 17 and the C-1 methylene groups of compounds 18 and 19, respectively, coincide and occupy the same space in the receptor.49 In the above mentioned studies, methylation or ring fusion at the 4-position provided inactive 8-OH-DPAT analogues.29

OH OH R N N N

17 18 19 R = OH 20 R = CONH2

Interestingly, substitution of the aromatic ring had quite dramatic consequences for the potency and/or intrinsic activity of the resulting compounds. Introduction of a C-5-fluoro substituent into the S-enantiomer of 8-OH-DPAT ((S)-UH301, 21) was reported to abolish the intrinsic efficacy and to some extent the affinity (Ki = 52 nM), presumably due to electronic effects.50 Liu et al. described a method to prepare a variety of 8-substituted 2-aminotetralins of 8-OH-DPAT via palladium-catalyzed reactions, utilizing the triflate analogue (22) as the key-intermediate.51 The highest affinities for the 5-HT1A receptor were observed for the R-enantiomers, except for the derivatives containing an acetyl or methylester moiety at the 8-position. The stereoselectivity of the latter compounds was reversed when cis-1-methyl analogues were tested.52

R O OH OSO2CF3 N N N

F 21 22 23 R = Me 24 R = OMe

Benz[e]indole derivative 25 was reported to exhibit mixed 5-HT1A and D2 53 receptor stimulating properties. Extremely potent 5-HT1A receptor agonists with high oral bioavailabilities were obtained by introducing a formyl (OSU191, 26)54 on the C-1 position, or a nitrile group on the C-1 (27) or C-2 (28) position of the N,N-di-n-propyl derivative, respectively.55 Stjernlöf et al. postulated that the electronic density of the

11 Chapter 1 nitrile groups and the lone-pairs of the formyl moiety interact with the hydroxy groups 56 of a serine or threonine residue on TM5 of the 5-HT1A receptor protein (Figure 1.6).

R2 R1 H N Me H CO N H 3 198 N N Ser H N H N Thr199 O Me O O O H H H 25 26 R1 = CHO, R2 = H N O 27 R = CN, R = H N 1 2 C 28 R1 = H, R2 = CN C H

N Figure 1.6. Possible hydrogen-bonding interactions between H compound 26, 27 or 28 and the 5-HT1A receptor. Adapted from ref.56 .

The examples given so far, among other things, suggest an important positive contribution of linear hydrophobic N-substituents to binding, giving optimal affinity for the 5-HT1A receptor in case of two n-propyl groups. Lengthening of unfunctionalized alkyl chains results in loss of affinity and, in case of the N,N-dibutyl analogue of 8-OH- DPAT, inversion of stereoselectivity, as was reported by Björk et al.57 However, a number of functionalized alkyl chains (such as ethylene, propylene and butylene chains) were shown to retain high affinity for this receptor subtype. Naiman et al.58 were the first to show that a phenyl substituent at one of the n-propyl terminals did not attenuate the affinity. Recently, Podona et al.59 prepared a number of aminochromane derivatives and explored the length of alkyl spacers and their substituents. Their best compounds, exemplified by compounds 30 and 31, possess imido or sulfonamido functional groups with a preferential length of four methylenes for the side chain and were proven to be full agonists. In line with this observation Ennis et al.60 presented a number of viable (heteroaromatic) ring substituents, such as 2-thiophene and 2- methoxy- or 3-chloro-benzene, on the alkyl chain terminal employing the basic structure of compound 26, resulting in subnanomolar Ki values ranging from 0.02-2.8 nM. Taken together, these observations suggest the presence of at least two lipophilic pockets in the receptor, which can accommodate an n-propyl group and a reasonable flat moiety, respectively. The latter may contribute in stabilizing the ligand-receptor complex by means of hydrogen-bonding and/or π-π-interactions.26

12 Introduction

O

OH H OMe 30 R = N N N R O

O 31 R = NSO2 Me 29

Arylpiperazines and Analogues. The prototypical arylpiperazine buspirone (32, Ki = 30 nM) and the more selective ipsapirone (33, Ki = 7 nM) and gepirone (34, Ki = 181 nM),61 were the first anxiolytic agents that directly stimulated the serotonergic 62 system. Of these partial 5-HT1A receptor agonists buspirone, recently classified as anti-depressant, is widely used as an anti-anxiety agent.

O O O R Me N 32 R = N 33 R = N 34 R = N N N S Me N O O O O

The spacer length and the nature of the aryl substituent and the terminal moiety play important roles in determining the affinity, the selectivity and the degree of 32 intrinsic efficacy for 5-HT1A receptors. NAN190 (35), a ligand which was initially reported to be an antagonist (Ki = 0.6 nM),63 was later shown to have agonistic effects in some assays.64 All above-mentioned compounds behave as partial agonists or antagonists at receptors localized postsynaptically, but stimulate the somatodendritic receptors. Evaluation of compounds belonging to the class of benzodioxyn-5- ylpiperazines led to the discovery of 5-HT1A/1B receptor agonist (36) and the selective 5-HT1A receptor agonist flesinoxan (37), which is investigated in clinical trials 65 in depression. Another potent, selective and full agonist at 5-HT1A receptors in vitro and in vivo is tetrahydropyridine SR57747A (38), having a Ki of 2.0 nM.66 Until recently, no full antagonists were known at both somatodendritic and postsynaptic 5- 12 HT1A receptors. In 1994, Fletcher et al., reported WAY100635 (39) to be a silent and selective antagonist, which is now being frequently used as a pharmacological tool and as PET-ligand in clinical studies (see Section 1.3 and Chapter 7).

13 Chapter 1

R1 O 36 R1 = H, R2 = H

37 R1 = CH2OH; R2 = OMe O O O N H F N N N N N R2

35 O

F3C OMe N N N N N O 38 39

Aryloxyalkylamines. Pindolol (40, Ki = 35 nM) and propanolol (41, Ki = 90 nM) belong to a whole different, but important structural class of compounds. In addition to being β1-adrenergic antagonists, these compounds are known to bind at 5- 67 HT1A sites, having low intrinsic efficacies. The β-hydroxy groups of 40 and 41 do not contribute to 5-HT1A receptor binding and their removal actually enhances 5-HT1A affinity, whereas it reduces β-adrenergic affinity.68 Guan et al.69 reported that amino acid residue Asn386 is responsible for the binding of these aryloxyalkylamines, while mutation of this residue produced only minor changes in the binding of other 5-HT receptor agonists. Conformationally constrained 1,4-benzodioxanes, such as the antagonist spiroxatrine (42),70 possess good affinity (Ki = 1.9 nM) and reasonable selectivity for 5-HT1A receptors. Indolodioxane derivative U86192A (43) is another representative of this type of compounds and was shown to have antihypertensive effects in the cat.71 The latter examples, including S14063 (44),72 have a shorter alkyl spacer in between the oxygen and nitrogen atom, compared to 40 and 41, which enhances the 5-HT1A receptor affinity.

OH H OH H O N N H O N Me O N Me N O Me Me O N 40 41 42 H O OMe MeO O N H N H N O S N H N O 43 44

14 Introduction

1.5 5-HT 1D Receptors

Distribution and Function. 5-HT1D receptors were first defined in bovine caudate and subsequently in the brain of a variety of other species, including man.73,74

Initially, the anatomical distribution of 5-HT1D receptors has been studied using 3 quantitative autoradiography using nonselective 5-HT1 receptor ligands, such as [ H]5- 3 CT (4) or [ H]5-HT, which required saturation of the non-5-HT1D sites. The highest densities have been found in the substantia nigra, basal ganglia and nigrostriatal pathway, whereas lower densities were reported in the hippocampus, raphé nuclei and cortex.75 The introduction of serotonin-5-O-carboxylmethyl-glycyl[125I]tyrosinamide 125 ([ I]GTI, 50) allowed for the direct visualization of 5-HT1D sites, confirming the 76 distribution patterns previously reported. The 5-HT1D receptor was shown to exist as a presynaptic heteroreceptor or a terminal autoreceptor, activation of which inhibits neurotransmitter release.77,78 Starkey and Skingle79 were the first to demonstrate the presence of functional 5-HT1D autoreceptors in the guinea-pig dorsal raphé nucleus, using the technique of fast-cyclic voltametry. The cloning of two distinct human genes 80 encoding for two highly homologous proteins, designated 5-HT1Dα and 5-HT1Dβ (alias 81 5-HT1B), accounted for a temporary confusion. Which of the two receptors was responsible for the pharmacological effects reported? In 1992, Adham et al.,82 shed light on this problem by showing that the human 5-HT1Dβ receptor, althought operationally distinct, constitutes the counterpart of the rodent 5-HT1B receptor. The CNS distribution of the latter receptor, notably absent in mammals and birds, indeed 83 parallels the regional distribution of 5-HT1D receptors in non-rodent species. Growing evidence suggests that the 5-HT1Dβ receptors are predominant, for instance, the distribution of 5-HT1Dβ receptor mRNA is consistently more widespread than that of the 84 co-distributing 5-HT1Dα receptor mRNA. Activation of 5-HT1D receptors induces inhibition of forskolin-stimulated adenylyl cyclase in the substantia nigra of calf and guinea pig. This observation is substantiated with studies performed in cells transfected 85 with either 5-HT1Dα or 5-HT1Dβ receptors. Since 5-HT1B receptors are implicated in the regulation of 5-HT release,86 and on the basis of the topographical similarities, centrally acting 5-HT1D receptor antagonists may well produce both antidepressant and anxiolytic effects, alone or in combination with SSRIs, and thus may constitute an attractive new research target.

5-HT1D receptors seem to have a prominent position within the final common pathway of the mechanisms involved in migraine, which is presumably manifested 87 through dilation of cerebral arteries. Stimulation of these receptors by 5-HT1D receptor

15 Chapter 1 agonists, such as sumatriptan (GR43175, 45), rapidly relieve the symptoms of the headache phase. Four mechanism have been suggested for the anti-migraine action of 88 5-HT1D receptor agonists (Figure 1.8): (1) Vasoconstriction of cranial blood vessels; (2) inhibition of release of vasoactive neuropeptides;89 (3) blockade of trigeminal nerve terminal depolarization;90 and (4) central inhibition with the trigeminal nucleus caudatus in the brainstem.91 pain

cortex unknown trigger sumatriptan

thalamus

(1) nausea (4) vomiting

(2) trigeminal photophobia blood vessel (3) neuron phonophobia

c-fos dilation trigeminal nucleus caudalis peptide release

“Peripheral” “Central”

Figure 1.8. Proposed intervention pathways of sumatriptan. Adapted from refs.87 and 89.

92 Hamel et al. reported the presence mRNA for the 5-HT1Dβ receptor in cerebral arteries of humans, suggesting that constriction of these vessels results from 5-HT1Dβ receptor activation. However, selective gene-expression for the 5-HT1Dα receptor was found in the human trigeminal ganglia,93 implying that subtype-selective agonists are still needed to determine the contribution of each receptor subtype in the abortion of migraine-attacks.

Structural Aspects. The 5-HT1Dα and the 5-HT1Dβ receptor subtypes were shown to contain 377 and 390 amino acids, respectively. Both receptors are G-protein-coupled and consist of seven transmembrane spanning segments connected by extra- and intracellular loops (Figure 1.9).94 The amino acid sequence identity in the membrane

16 Introduction

spanning domain of both receptors is approximately 77%, whereas the human 5-HT1Dβ receptor differs only 4% from its rodent homologue, represented by eight amino acid residues (Table 1.2).

Extracellular

S T C D W Q C H L A Y A I F L I S I P A V I W P Y P L L V S L P T V F L S I P S I I L D V M S D S I Y S S V F F I V I C T V T W T L L T C I S C F L A I S C I G A F P G T V T A F L Y V L L A W I Y W C L N S L D S I V I P I S N T L A S I L A F T H L I V F I V A L C M T L L A G N P L S V A I L I T T G I I A A I I I Y I L L H G L Y I G T TM1 TM2 TM3 TM4 TM5 TM6 TM7

Intracellular

Extracellular

L T C D W H C H L V Y M F F F I L I L A M T W P Y P M I L S L P T V F L A I P S L V L D L M S D S I Y S S I F F I V I S T I T W T L L T C I S V F L A I S C F G A F P G T V T V F L Y T L L A W Y F W C L N S L D S I V L P V S N T L A T I L A F V H L I L F I V A L C V M L L A G N P I S A A I L I A T A I I A A A I I I Y V L L R K L Y G T TM1 TM2 TM3 TM4 TM5 TM6 TM7

Intracellular

Figure 1.9. Top: Representation of the human 5-HT1Dα receptor. Dark-grey circles indicate amino acids that are different from corresponding positions in the human 5-HT1Dβ receptor. Bottom: Representation of the human 5-

HT1Dβ receptor.

355 355 By replacing Thr of the human 5-HT1Dβ receptor with a corresponding Asn 67 found in rodent 5-HT1B receptors, Oksenberg et al. showed that the major pharmacological difference between these species homologues confers a one single amino acid residue. This implies that the 5-HT1Dβ and 5-HT1B receptors are likely to have the same biological functions, while exhibiting distinct binding profiles for various compounds. Other important residues for binding to 5-HT1Dα and 5-HT1Dβ receptors are

17 Chapter 1 likely the ones which are conserved in most 5-HT receptor proteins, such as an Asp (TM3) and Ser/Thr on TM5. Additionally, the serine residue on helix 4, which is suggested to be important for the binding of 5-HT to the 5-HT2 receptor, is also present 95 in the 5-HT1D receptor subtypes.

Table 1.2. Amino acid sequence identities (%) in the TM domain of cloned 5-HT receptors.

Species Receptor

Human 5-HT1A - 53 54 53

Rat 5-HT1B - 74 96

Human 5-HT1Dα - 77

Human 5-HT1Dβ -

The 5-HT 1D Receptor H Pharmacophore. To date, only one pharmacophore model has been proposed. N 6.7 96 5.2 Glen et al. superimposed computed 4.2-4.7 conformations of active molecules using H-bond 4.8 donor/acceptor known ligands, such as methysergide (9), as 5.1-7.1 template molecules. The resulting H-bond acceptor Hydrophobic pharmacophore hypothesis is composed of a site protonated amine site, an aromatic region, a Figure 1.9. Pharmacophore model as proposed by hydrophobic pocket and two hydrogen- Glen et al. Distances are given in Å. Adapted from bonding sites (Figure 1.9). However, this ref. 96. pharmacophore has yet to be challenged by (future) conformationally restricted analogues which bind to the 5-HT1D receptor.

1.6 5-HT 1D Receptor Agonists and Antagonists (SAR)

Due to the recent discovery of the 5-HT1D receptor subtypes, the collection of known ligands that bind to these receptors is much smaller, for instance, compared to 5-

HT1A receptor ligands. Many of the tryptamine derivatives mentioned in section 1.4, such as 5-CT (4; Ki = 1.1 nM), show considerable affinity for the 5-HT1D receptor subtypes. Simple modifications of 5-HT (Ki = 2.2 nM), such as removal or methylation of the hydroxy group and dimethylation of the primary amine function resulted in an approximately ten, two and four fold attenuation of the affinity, respectively.97 Larger

18 Introduction lipophilic moieties in the 5-position of the indole nucleus, such as a p-chlorobenzyloxy group, are well-tolerated. The non-selective compound (+)-LSD (8) also binds to the 5-

HT1D receptor with a Ki of 11 nM. However, none of the above-cited compounds can discriminate between the 5-HT1A and 5-HT1D receptor subtypes and were superceded by sumatriptan (GR43175, 45), which was identified as the first 5-HT1D receptor agonist with reasonable selectivity.98 This compound was recently successfully introduced for the treatment of migraine. From then on, novel compounds marched swiftly after each other. Mostly hydrogen-bond accepting, (aromatic) heterocycles in the 5-position of tryptamine have proven to be viable moieties, exemplified by compounds 46 (MK462)99 and 47 (311C90).96 Both compounds are clinically effective in the treatment of migraine, thus confirming the therapeutic utility of this class of compounds.

H O O N N O S O MeNH N N Me Me H Me N N N Me Me Me

N N N H H H 45 46 47

Soon, it became evident that a 5-substituent with considerable length can easily 100 be accommodated by the 5-HT1D receptor. This was demonstrated by Glennon et al, who introduced hydrophobic tails with various lenghts on this position resulting in 5- nonyloxy-tryptamine (48), a compound which binds with higher affinity to 5-HT1Dβ receptors than 5-HT1Dα receptors (Ki 1.2 vs 16 nM, respectively). Other representatives bearing a large group in this direction are L694,247 (49),101 GTI (50)76 and the arylpiperazide derivative 51.102

19 Chapter 1

48 R = O(CH2)8CH3 NHSO2Me

O N R 49 R = CH2 N

NH2 O O OH 50 R = OCH CNHCH CNHCH N 2 2 H NH2 O O

51 R = O(CH2)4CN N NHSO2Me

Others embarked on the synthesis of conformationally restricted tryptamine analogues. Both, the sumatriptan analogue 52 (CP122,288)103 and the (3-nitro-pyridin- 2-yl)amino derivative 53 (CP124,439)104 are constrained in the ethylamino side chain by the introduction of a pyrrolidine ring. King et al.105 developed the 3- aminotetrahydrocarbazole derivative BRL56905 (54), which is a conformationally restricted analogue of 5-CT, exhibiting a Ki of 10 nM for the 5-HT1Dβ receptor. However, this carbon skeleton seems to be valid only with the carboxamido substituent but not for alkyloxy substituents.106

NO2 O O

S H2N O MeNH N

N Me Me N H

N N N NH2 H H H 52 53 54

Although most 5-HT1D receptor agonists are of the tryptamine type, few compounds were reported having a structurally different composition. LeBoulluec et al. demonstrated that bivalent indoles, represented by compound 55, bind in the low- 107 nanomolar range to the 5-HT1D receptor subtype (IC50 = 0.05 nM). The optimal chain linkage was found to be seven methylene units but was also found to have high affinity for the 5-HT1A site.

20 Introduction

H2N O O NH2

(CH2)n N N H H N N H H 55

Arylpiperazine 56 was shown to have a high affinity for 5-HT1D receptors (Ki = 2 108 nM), but is a high affinity ligand for 5-HT1A receptors as well (Ki = 3.3 nM). Alniditan (57) is a structurally different antimigraine agent which binds in de nanomolar range to

5-HT1D and 5-HT1A receptors, possessing full agonist properties for both 5-HT1Dα and 5- 109 HT1Dβ receptors in vitro.

H N MeO N N O N N N H N

56 57

Several of the benz[e]indoles 26, 27 and 28 and their derivatives (see Section

1.4), in addition to being 5-HT1A receptor ligands, were reported to have considerable affinity for both 5-HT1D receptor subtypes, showing a preference of 20-37 fold for the 56,60 5-HT1Dα over the 5-HT1Dβ site. Few compounds, such as compound 58 (Ki 13 nM 5-

HT1A; 18 nM 5-HT1Dα) and 58 (Ki 0.6 nM 5-HT1A; 13 nM 5-HT1Dα) were shown to be inactive for the 5-HT1Dβ receptor subtype.

SMe

H N N H N N

S

58 59

The first disclosures of selective 5-HT1D antagonists were made by researchers at Glaxo, who based their structures upon the phenylpiperazinylbenzanilide moiety. The antagonism was determined by the inhibitory response of the 5-HT-induced contraction

21 Chapter 1 of the dog saphenous vein and hypothermia in guinea-pigs. This strategy provided the oxadiazole derivative 60 (GR127935), which now serves as a widely used 110 pharmacological tool (Ki 5-HT1Dα = 1.3 nM, Ki 5-HT1Dβ = 0.13 nM). Despite its recognized antagonist profile in animal isolated tissues and behavioural models,

GR127935 was shortly thereafter shown to be a partial agonist at cloned human 5-HT1Dα receptors.111 Unlike GR127935, the (dimethylamino)propyl benzanilide 61 (GR55562) 112 behaved as a silent antagonist at the 5-HT1Dα and 5-HT1Dβ receptors in a similar study.

H Me N N Me N H O N N OMe OH O N N Me N O H Me 60 61 1.7 Objective and Outline

The trifluoromethanesulfonate (triflate) group is known for its electron- withdrawing properties indicated by the positive Hammett σp (+0.37) and Taft σI (+0.84) constants.113 Consequently, the aliphatic triflate functionality is an excellent leaving group, and therefore widely used in Organic Chemistry. However, the properties of aryl triflates are considerably different as indicated by its chemical stability with respect to solvolysis.114 Chemical stablity is an advantageous property in drug-development and therefore, the aryl triflate group may be employed as a bioisostere of some of the aryl substituents discussed in this chapter. In addition, the electron-withdrawing effect of aryl triflates may prevent aromatic in vivo hydroxylation, resulting in particularly metabolic stable compounds. The research presented in this thesis describes a survey of the aryl triflate concept, as applied to 5-

HT1A, 5-HT1Dα and 5-HT1Dβ receptor ligands. Consequently, this class of compounds may have potential therapeutic applications in treatment of depression, anxiety disorders or migraine. The emphasis is put on the structure-affinity relationships (SAFIR) and the structure-activity relationships (SAR). In addition, the bioavailability of some of the newly synthesized compounds is considered. As indicated in this chapter, much is known about the brain function and localization of 5-HT1A receptors and moreover, selective and potent 5-HT1A receptor agonists such as 8-hydroxy-2-(di-n-propylamino)tetralin (8-OH-DPAT) are available.

22 Introduction

However, due to poor pharmacokinetic properties the clinical potential of this compound is low. Chapter 2 deals with the efforts to develop structural analogues of 8-

OH-DPAT which retain selectivity and potency for the 5-HT1A receptor but display an improved (oral) bioavailability. The hydroxy group is sensitive to O-glucuronidation and additionally, the low oral bioavailability of 8-OH-DPAT was found to be caused by N-depropylation. For this reason, the phenol portion was masked as an aryl triflate and the N-monopropyl substituted 2-aminotetralins were selected as a starting-point. The newly synthesized compounds were screened for their affinity to 5-HT1A receptors and the most promising compounds were evaluated by means of 5-hydroxytryptophan (5- HTP) accumulation in the rat brain, behavioural experiments and hypothermia.

Especially, the (R)-enantiomer of 8-OSO2CF3-PAT was found to be very potent, however, the oral availability was comparatively low (7.6%). In addition, a drastic increase in affinity for 5-HT1D receptors was observed, as compared to 8-OH-DPAT.

Methylation of the C1-position of the tetralin system (cis-8-OSO2CF3-MPAT) resulted in a slight decrease in affinity for 5-HT1A and 5-HT1Dα receptors. Cis-(1S,2R)-8-

OSO2CF3-MPAT was shown to be the most potent enantiomer for 5-HT1A sites in the rat after subcutaneous or oral administration. The (1R,2S)-enantiomer exhibited a low intrinsic efficacy but an increased selectivity towards 5-HT1A receptors, as compared to its optical antipode. The cis-N-methylamino derivative (cis-8-OSO2CF3-MMAT) was found to be a nonselective ligand, whereas the trans-analogues were shown to be inactive.

The pharmacological profile of the 5-HT1A receptor agonist (R)-8-OSO2CF3-PAT gave rise to the investigation of its potential anxiolytic properties by means of animal models. Chapter 3 describes the effects of acute administration of (R)-8-OSO2CF3-PAT on rats in the conditioned defensive burying, the elevated plus-maze and the inescapable footshock model. In addition, the 5-HT turnover was determined in homogenates of various brain areas after administration of (R)-8-OSO2CF3-PAT at the doses that were used in the behavioural models. (R)-8-OSO2CF3-PAT was found to be active in the burying model and the plus-maze but not in the footshock paradigm. The 5-HT turnover significantly decreased in parts of the limbic system of the rat brain.

Inspired by the positive influence of the triflate group on the 5-HT1D receptor affinity of 2-aminotetralins the SAFIR and SAR of tryptamines was investigated. In Chapter 4 , the N-methylaminosulfonylmethylene group of the effective anti-migraine agent sumatriptan, was replaced by a triflate group. A series of N,N-dialkyl substituted

5-triflated tryptamines was prepared and screened for affinity and activity at 5-HT1Dα en

5-HT1Dβ sites. Forskolin-stimulated cAMP inhibition was employed as a measure for the

5-HT1D receptor-agonist properties of the compounds. The primary amines and small

23 Chapter 1

N,N-dialkyl substituents were well-tolerated by the 5-HT1Dα and 5-HT1Dβ receptor subtype, resulting in fairly potent compounds. All derivatized tryptamines displayed moderate affinity for the 5-HT1A receptor. The most promising compound, the N,N- dimethyl-5-triflate-substituted tryptamine, induced hypothermia and a decreased 5-HT turnover in the brain of the guinea pig. The inactivity of this compound for 5-HT1A sites in the rat was confirmed by means of 5-HTP accumulation and intracerebral microdialysis. The receptograms of the 2-aminotetralins described in Chapter 2 indicate that selectivity may be induced by ethylamino side chain restriction of serotonin analogues. Chapter 5 deals with other rigidification possibilities, exemplified by the synthesis of a triflate-substituted 3-aminocarbazole and 4-indol-3-ylpiperidines. Both classes of compounds displayed a strong preference for 5-HT1D receptors. This chapter also describes the preparation and testing of other sulfonic acid ester derivatized tryptamines, however, all compounds were found to have a lower affinity relative to the triflate analogue.

The 5-HT1A receptor antagonists ORG13502 and WAY100635 both possess an ortho-methoxyphenylpiperazine structure. In Chapter 6 , the affinity and intrinsic activity for 5-HT1A receptors of ortho-methoxy, hydroxy and triflate substituted phenylpiperazines are compared. The triflate analogues were found to have a comparatively lower affinity than ORG13502 and WAY100635, along with an strongly enhanced intrinsic activity. With the aid of molecular modelling and a crystal database search we tried to find an explanation for this phenomenon. Finally, Chapter 7 describes the radiochemical synthesis and biodistribution studies of [11C]ORG13502 and the previously reported [11C]WAY100635 by means of positron emission tomography (PET) in the rat brain. The regional uptake of 11 11 [ C]WAY100635, but not of [ C]ORG13502, reflected the known 5-HT1A receptor density. In addition, the experiments were repeated with adrenalectomized (ADX) animals which are known to have an increased 5-HT1A receptor density, as compared to normal animals. Small differences were observed in the uptake of [11C]WAY100635 between normal and ADX rats, however, these differences were not significant.

Taken together, this thesis presents some interesting 5-HT1A and 5-HT1D receptor agonists. Follow-up studies will have to determine whether these compounds have any therapeutic potential. In the Concluding remarks the author states that the electron- withdrawing aryl triflate group is an interesting bioisostere for a number of aryl substituents. Depending on the nature of the ligand and the receptor, the aryl triflate group may successfully be applied in future drugs.

24 Introduction

1.8 References

[1] a) Ferrari, M.D. Serotonin (5-HT) in neurologic and psychiatric disorders, 1994, Leiden. b) Peroutka, S.J. Serotonin receptors subtypes. Basic and clinicalaspects, 1991, Wiley-Liss, New York. [2] a) Rapport, M.M.; Green, A.A.; Page, I.H.J. Biol. Chem. 1948, 176, 1237. b) Rapport, M.M.; Green, A.A.; Page, I.H. Science 1948, 108, 329. [3] Twarog, B.M.; Page, I.H.Am. J. Physiol. 1953,175, 157. [4] Gaddum, J.H.; Picarelli, Z.P.Br. J. Pharmacol. 1957, 12, 323. [5] Peroutka, S.J.; Snyder, S.H.Mol. Pharmacol. 1979, 16, 687. [6] Hoyer, D.; Clarke, D.E.; Fozard, J.R.; Hartig, P.R.; Martin, G.R.; Mylecharane, E.J.; Saxena, P.R.; Humprey, P.P.A. Pharmacol. Rev. 1994, 46, 157. [7] Peroutka, S.J; Howell, T.A.Neuropharmacol . 1994, 33, 319. [8] Feng, D.F.; Doolittle, R.F.Meth. Enzymol. 1990, 183, 375. [9] Arvidsson, L.-E.; Hacksell, U.; Nilsson, J.L.G.; Hjorth, S.; Carlsson, A.; Lindberg, P.; Sanchez, D.; Wikström, H. J. Med. Chem. 1981, 24, 921. [10] Gozlan, H.; El Mestikawy, S.; Pichat, L.; Glowinski, J.; Hamon, M.Nature 1983, 305, 140. [11] a) Hoyer, D.; Pazos, A.; Probst, A.; Palacios, J.M. Brain Res. 1986, 376, 97. b) Pazos, A.; Probst, A.; Palacios, J.M. Neurosci. 1987, 21, 97. c) Radja, F.; Daval, G.; Hamon, M.; Vergé, D. J. Neurochem. 1992, 58, 1338. [12] a) Hume, S.P.; Ashworth, S.; Opacka-Juffry, J.; Ahier, R.C.; Lammertsma, A.A.; Pike, V.W.; Cliffe, I.A.; Fletcher, A.; White, A.C.Eur. J. Pharmacol. 1994, 271, 515. b) Pike, V.W.; Hme. S.P.; Opacka-Juffry, J.; McCarron, J.A.; Cliffe, I.A.; Fletcher, A.Third IUPHAR Meeting on Serotonin, 1994, Abstr. 73. [13] Fletcher, A.; Bill, D.J.; Cliffe, I.A.; Forster, E.A.; Jones, D.; Reilly, Y.Br. J. Pharmacol. 1994, 112, 91P. [14] Mathiessen, L.; Daval, G.; Bailly, Y.; Gozlan, H.; Hamon, M.; Vergé, D.Neurosci . 1992, 51, 475. [15] Dourish, C.T.; Hutson, P.H.; Curzon, G.Trends. Pharmacol. Sci. 1986, 7, 212. [16] Yocca, F.D.; Iben, L.; Meller, E.Mol. Pharmacol. 1992, 41, 1066. [17] De Vry, J. Psychopharmacology-Berl. 1995, 121, 1. [18] Berendsen, H.H.; Broekkamp, C.C.; Van Delft, A.M.Eur. J. Pharmacol. 1990, 187, 97. [19] Tricklebank, M.D.; Forler, C.; Fozard, J.R.Eur. J. Pharmacol. 1984, 106, 271. [20] Kobilka, B.K.; Frielle, T.; Collins, S.; Yang-Feng, T.; Kobilka, T. S.; Francke, U.; Lefkowitz, R.J.; Caron, M.G. Nature, 1987, 329, 75. [21] Fargin, A.; Raymind, J.R.; Lohse, M.J.; Kobilka, B.K.; Caron, M. G.; Lefkowitz, R.J.Nature , 1988, 335, 358. [22] Dixon, R.A.F.; Strader, C.D.; Sigal, I.S.Annu. Rep. Med. Chem. 1988, 23, 221. [23] Albert, P.R.; Zhou, Q.-Y.; Van Tol, H.H.M.; Bunzow, J.R.; Civelli, O.J. Biol. Chem. 1990, 265, 5825. [24] Ho, B.Y.; Karschin, A.; Branchek, T.; Davidson, N.; Lester, H.A.FEBS Letters 1992, 312, 259. [25] a) Guan, X.-M.; Peroutka, S.J.; Kobilka, B.K.Mol. Pharmacol. 1992, 41, 695. b) Chanda, P.K.; Minchin, M.C.W.; Davis, A.R.; Greenberg, L.; Reilly, Y.; McGregor, W.H.; Bhat, R.; Lubeck, M.D.; Mizutani, S.; Hung, P.P. Mol Pharmacol. 1993, 43, 516. [26] Trumpp-Kalmeyer, S.; Hoflack, J.; Bruinvels, A.; Hibert, M.J. Med. Chem. 1992, 35, 3448. [27] Arvidsson, L.-E.; Karlén, A.; Norinder, U.; Kenne, L.; Sundell, S.; Hacksell, U.J. Med. Chem. 1988, 31, 212. [28] Hibert, M.F.; Gittos, M.W.; Middlemiss, D.N.; Mir, A.K.; Fozard, J.R.J. Med. Chem. 1988, 31, 1087. [29] Mellin, C.; Vallgårda, J.; Nelson, D.L.; Björk, L.; Yu, H.; Andén, N.-E.; Csoregh, I.; Arvidsson, L.E.; Hacksell, U. J. Med. Chem. 1991, 34, 497. [30] Chidester, C.G.; Lin, C.-H.; Lahti, R.A.; Haddsma-Svensson, S.R.; Smith, M.W.J. Med. Chem. 1993, 36, 1301. [31] Nilsson, J.; Lilljefors, T.; Wikström, H.1992 , Unpublished results. [32] a) Glennon, R.A. J. Med. Chem. 1987, 30, 1. b) Glennon, R.A.Drug Dev. Res. 1992, 26, 251. [33] a) Hoyer, D.; Engel, G.; Kalkman, H.Eur. J. Pharmacol. 1985, 118, 13. b) Engel, D.; Gothert, M.; Hoyer, D.; Schlicker, E.; Hillenbrand, K.Naunyn-Schmied. Arch. Pharmacol. 1986, 332, 1 [34] Glennon, R.A.; Titeler, M.; Lyon, R.A.; Slusher, R.M. J. Med. Chem. 1988, 31, 867. [35] Robertson, D.W.; Fuller, R.W.Ann. Rep. Med. Chem. 1988, 23, 49.

25 Chapter 1

[36] Cassady, J.M.; Li, G.S.; Spitzner, E.B.; Floss, H.G.; Clemens, J.A.J. Med. Chem. 1974, 17, 300. [37] Ward, J.S.; Fuller, R.W.; Merritt, L.; SNoddy, H.D.; Paschal, J.W.; Mason, N,R,; Horng, J.S.J. Med. Chem. 1988, 31, 1512. [38] Foreman, M.M.; Fuller, R.W.; Leander, J.D.; Benvenga, M.J.; Wong, D.T.; Nelson, D.L.; Calligaro, D.O.; Swanson, S.P.; Lucot, J.B.; Flaugh, M.E.J. Pharmacol. Exp. Ther. 1993, 267, 58. [39] Goldberg, M.J.; Rowe, H.M.; Cerimele, B.J.; Woodworth, J.R.; Bowsher, R.R.; Santa, P.F. 5-HTThird IUPHAR Satellite Meeting on Serotonin 1994, Abst. 89. [40] Arvidsson, L.-E.; Hacksell, U.; Nilsson, J.L.G.; Hjorth, S.; Carlsson, A.; Lindberg, P.; Sanchez, D.; Wikström, H. J. Med. Chem. 1981, 24, 921. [41] Tricklebank, M.D.; Middlemiss, D.N.; Fozard, J.R.Trends Pharmacol. Sci. 1984, 5, 415. [42] Thorberg, S.O.; Hall, H.; Åkesson, C.; Svensson, K.; Nilsson, J.L.G. Acta Pharm. Suec. 1987, 24, 169. [43] Hammarberg, E.M.; Johansson, L.G.; Ross, S.B.; Thorberg, S.O.Eur. Pat. Appl. 538222A (21-3-93). [44] a) Cannon, J.G.; Mohan, P.; Bojarski, J.; Long, J.P.; Bhatnagar, R.K.J. Med. Chem. 1988, 31, 313. b) Hedberg, M.; Johansson, A.M.; Hacksell, U.J. Chem. Soc., Chem. Commun. 1992, 845. [45] Arvidsson, L.-E.; Johansson, A.M.; Hacksell, U.; Nilsson, J.L.G.; Svensson, K.; Hjorth, S.; Magnusson, T.; Carlsson, A.; Andersson, B.; Wikström, H.J. Med. Chem. 1987, 30, 2105. [46] Hjorth, S.; Sharp, T.; Liu, Y.Naunyn-Schmied. Arch. Pharmacol. 1990, 341, 149. [47] a) Wikström, H.; Andersson, B.; Elebring, T.; Jacyno, J.; Allinger, N.L.; Svensson, K.; Carlsson, A.; Sundell, S. J. Med. Chem. 1987, 30, 1567. b) Mellin, C.; Vallgårda, J.; Nelson, D.L.; Björk, L.; Yu, H.; Andén, N.-E.; Csoregh, I.; Arvidsson, L.E.; Hacksell, U.J. Med. Chem. 1991, 34, 497. [48] a) Lin, C.-H.; Haadsma-Svensson, S.R.; Lahti, R.A.; McCall, R.B.; Piercey, M.F.; Schreur, P.J.K.D.; Von Voigtlander, P.F.; Smith, M.W.; Chidester, C.G.J. Med. Chem. 1993, 36, 1053. b) Lin, C.-H.; Haadsma- Svensson, S.R.; Phillips, G.; McCall, R.B.; Piercey, M.F.; Smith, M.W.; Svensson, K.; Carlsson, A.; Chidester, C.G.; Von Voigtlander, P.F.J. Med. Chem. 1993, 36, 2208. [49] Chidester, C.G.; Lin, C.-H.; Lahti, R.A.; Haadsma-Svensson, S.R.; Smith, M.W.J. Med. Chem. 1993, 36,1301. [50] Hillver, S.-E.; Björk, L.; Li, Y.-L.; Svensson, B.; Ross, S.; Andén, N.-E.; Hacksell, U.J. Med. Chem. 1990, 33, 1541. [51] a) Liu, Y. Yu, H.; Svensson, B.E.; Cortizo, L.; Lewander, T.; Hacksell, U. J. Med. Chem. 1993, 36, 4221. b) Liu, Y. Svensson, B.E.; Yu, H.; Cortizo, L.; Ross, S.B.; Lewander, T.; Hacksell, U.Bioorg. Med. Chem. Letters 1991, 1, 257. [52] Liu, Y.; Yu, H.; Mohell, N.; Nordvall, G.; Lewander, T.; Hacksell, U. J. Med. Chem. 1995, 38, 150. [53] Wikström, H.; Andersson, B.; Svensson, A.; Humber, L.G.; Asselin, A.A.; Svensson, K.; Ekman, A.; Carlsson, A.; Nilsson, I.; Chidester, C.J. Med. Chem. 1989, 32, 2273. [54] a) Stjernlöf, P.; Gullme, M.; Elebring, T.; Andersson, B.; Wikström, H.; Lagerquist, S.; Svensson, K.; Ekman, A.; Carlsson, A.; Sundell, S. J. Med. Chem. 1993, 36, 2059. b) Lin, C.-H.; Ennis, M.D.; Hoffman, R.L.; Phillips, G.; Haadsma-Svensson, S.R.; Ghazal, N.B.; Chidester, C.G.J. Heterocyclic Chem. 1994, 31, 129. [55] Stjernlöf, P.; Elebring, T.; Nilsson, J.; Andersson, B.; Svensson, K.; Lagerkvist, S.; Ekman, A.; Carlsson, A.; Wikström, H. J. Med. Chem. 1994, 37, 3263. [56] Stjernlöf, P.; Ennis, M.D.; Hansson, L.O.; Hoffman, R.L.; Ghazal, N.B.; Sundell, S.; Smith, M.W.; Svensson, K.; Carlsson, A.; Wikström, H. J. Med. Chem. 1995, 38, 2202. [57] Björk, L.; Backlund-Höök, B.; Nelson, D.L.; Andén, N.-E.; Hacksell, U.J. Med. Chem. 1989, 32, 779. [58] Naiman, N.; Lyon, R.; Bullock, A.E.; Rydelek, L.T.; Titeler, M.; Glennon, R.A.J. Med. Chem. 1989, 32, 253. [59] Podona, T.; Guardiola-Lemaître, B.; Caigard, D.-H.; Adam, G.; Pfeiffer, B.; Renard, P.; Guillaumet, G.J. Med. Chem. 1994, 37, 1779. [60] Ennis, M.D.; Stjernlöf, P.; Hoffman, R.L.; Ghazal, N.B.; Smith, M.W.; Svensson, K.; Wikström, H.; Haadsma-Svensson, S.R.; Lin, C.-H.J.Med. Chem. 1995, 38, 2217. [61] Yocca, F.D.; Smith, D.W.; Hyslop, D.K.; Maayani, S.Soc. Neurosci. 1986, 12, Abstr. 422. [62] a) Hjorth, S.; Carlsson, A. Eur. J. Pharmacol. 1982, 83, 299. b) Traber, J.; Glaser, T. Trends in Pharmacol. Sci. 1987, 8, 432. [63] Glennon, R.A.; Naiman, N.A.; Lyon, R.A.; Titeler, M.J. Med. Chem. 1988, 31, 1968.

26 Introduction

[64] Rydelek-Fitzgerald, L.; Teitler, M.; Fletcher, P.W.; Ismaiel, A.M.; Glennon, R.A. Brain Res. 1990, 532, 191. [65] Van Steen, B.J.; Van Wijngaarden, I.; Tulp, M. Th. M.; Soudijn, W.J. Med. Chem. 1994, 37, 2761. [66] a) Bachy, A. Steinberg, R.; Santucci, V.; Fournier, M.; Landi, M.; Hamon, M.; Manara, L.; Keane, P.E.; Soubrié, P.; Le Fur, G. Fundam. Clin. Pharmacol. 1993, 7, 487. b) Cervo, L.; Bendotti, C.; Tarizzo, E.; Cagnotto, A.; Skorupska, M.; Mennini, T.; Samanin, R.Eur. J. Pharmacol. 1994, 253, 139. [67] Oksenberg, D.; Peroutka, S.J.Biochem. Pharmacol. 1988, 37, 3429. [68] Pierson, M.E.; Lyon, R.A.; Titeler, M.; Kowalski, P.; Glennon, R.A.J. Med. Chem. 1989, 32, 859. [69] Guan, X.-M.; Peroutka, S.J.; Kobilka, B.K.Mol. Pharmacol. 1992, 41, 695. [70] Nelson, D.L.; Taylor, E.W.Eur. J. Pharmacol. 1986, 128, 207. [71] Ennis, M.D.; Baze, M.E.; Smith, M.W.; Lawson, C.F.; McCall, R.B.; Lahti, R.A.; Piercey, M.F.J. Med. Chem. 1992, 35, 3058. [72] Dabiré, H.; Bajjou, R.; Chaoche-Teyara, K.; Fournier, B.; De Nanteuil, G.; Laubie, M.; Safar, M.; Schmitt, H. Eur. J. Pharmacol. 1991, 203, 323. [73] Heuring, R.E.; Peroutka, S.J.; J. Neurosci. 1987, 7, 894. [74] Waeber, C.; Schoeffter, P.; Palacios, J.M.; Hoyer, D.Naunyn-Schmiedeberg’s Arch. Pharmacol. 1988, 337, 595. [75] a) Waeber, C.; Dietl, M.M.; Hoyer, D.; Palacios, J.M.Naunyn-Schmiedeberg’s Arch. Pharmacol. 1989, 340, 486. b) Waeber, C.; Dietl, M.M.; Probst, A.; Hoyer, D.; Palacios, J.M.Neurosci. Lett. 1988, 88, 11. [76] a) Boulengez, P.; Segu, L.; Chauveau, J.; Morel, A.; Lanoir, J.; Delaage, M. J. Neurochem. 1992, 58, 951. b) Palacios, J.M.; Waeber, C.; Bruinvels, A.T.; Hoyer, D.Brain Res. Mol. Brain Res. 1992, 13, 175. [77] Hoyer, D.; Middlemiss, D.N.Trends. Pharmacol. Sci. 1989, 10, 130. [78] a) Maura, G.; Thellung, S.; Andreoli, G.C.; Ruelle, A.; Raiteri, M. J. Neurochem. 1993, 60, 1179. b) Price, G.W.; Roberts, J.; Watson, J.; Burton, M.; Mulholland, K.; MIddlemiss, D.N.; Jones, B.J.Behav. Brain Res. 1996, 73, 79. [79] Starkey, S.J.; Skingle, M.Neuropharmacology 1994, 33, 393. [80] Hamblin, M.; Metcalf, M.Mol. Pharmacol. 1991, 40. 143. [81] Weinshank, R.L.; Zgombick, J.M.; Macchi, M.J.; Branchek, T.A, Hartig, P.R.Proc. Natl. Acad. Sci. USA 1992, 89, 3630. [82] Adham, N.; Romanienko, P. Hartig, P.; Weinshank, R.L.; Branchek, T.Mol. Pharmacol. 1992, 41, 1. [83] Hoyer, D.; Schoeffter, P.; Waeber, C.; Palacios, J.M.Ann. NY Acad. Sci. 1990, 600, 168. [84] Bruinvels, A.T.; Landwehrmeyer, B. Gustafson, E.L., Durkin, M.M.; Mengod, G.; Branchek, T.A.; Hoyer, D.; Palacios, J.M. Neuropharmacology 1994, 33, 367. [85] Zgombick, J.M.; Borden, L.A.; Cochran, T.L.; Kucharewicz, S.A.; Weinshank, R.L.; Branchek, T.A.Mol. Pharmacol. 1993, 44, 575. [86] Chopin, P.; Moret, C.; Briley, M.Pharmacol. Ther. 1994, 62, 385. [87] Ferrari, M.D.; Saxena, P.R.Eur. J. Neurol. 1995, 2, 5. [88] Humprey, P.P.A.; Feniuk, W.Trends Pharmacol. Sci. 1991, 12, 444. [89] Moskowitz, M. Trends Pharmacol. Sci. 1992, 13, 307. [90] Ferrari, M.D.; Saxena, P.R.Trends Pharmacol. Sci. 1993, 14, 129. [91] Kaube, H.; Hoskin, K.L.; Goadsby, P.J.Br. J. Pharmacol. 1993, 109, 788. [92] Hamel, E.; Fan, E.; Linville, D.; Ting, V.; Villemure, J.-G.; Chia, L.-S.Mol. Pharmacol. 1993, 44, 242. [93] Rebeck, G.W.; Maynard, K.I.; Hyman, B.T.; Moskowitz, M.A.Proc. Natl. Acad. Sci. 1994, 91, 3666. [94] Hartig, P.R.; Branchek, T.A.; Weinshank, R.L.Trends Pharmacol. Sci. 1992, 13, 152. [95] Rippmann, F.; Böttcher, H.Kontakte (Darmstadt) 1994, 1, 30. [96] Glen, R.C.; Martin, G.R.; Hill, A.P.; Hyde, R.M.; Woollard, P.M.; Salmon, J.A.; Buckingham, J.; Robertson, A.D. J. Med. Chem. 1995, 38, 3566. [97] Glennon, R.A.; Ismaiel, A.M.; Chaurasia, C.; Titetler, M.Drug Dev. Res. 1991, 22, 25. [98] Peroutka, S.J.; McCarthy, B.G.Eur. J. Pharmacol. 1989, 163, 133. [99] Street, L.J.; Baker, R.; Davey, W.B.; Guiblin, A.R.; Jelley, R.A.; Reeve, A.J.; Routledge, H.; Sternfeld, F.; Watt, A.P.; Beer, M.S.; Middlemiss, D.N.; Noble, A.J.; Stanton, J.A.; Scholey, K.; Hargreaves, R.J.; Sohal, B.; Graham, M.I.; Matassa, V.G. J. Med. Chem. 1995, 38, 1799. [100] Glennon, R.A.; Hong, S.-S.; Dukat, M.; Teitler, M.; Davis, K.J. Med. Chem. 1994, 37, 2828.

27 Chapter 1

[101] Beer, M.S.; Stanton, M.A.; Bevan, Y.; Heald, A.; Reeve, A.J.; Street, L.J.; Matassa, V.G.; Hargreaves, R.J.; Middlemiss, D.N.Br. J. Pharmacol. 1993, 110, 1196. [102] Perez, M.; Fourrier, C.; Sigogneau, I.; Pauwels, P.J.; Palmier, C.; John, G.W.; Valentin, J.-P.; Halazy, S. J. Med. Chem. 1995, 38, 3602. [103] a) Macor, J.E.; Blank, D.H.; Post, R.J.; Ryan, K.Tetrahedron Lett. 1992, 33, 8011. b) Beattie, D.T.; Connor, H.E. Eur. J. Pharmacol. 1995, 276, 271. [104] Macor, J.E.; Blank, D.H.; Fox, C.B.; Lebel, L.A.; Newman, M.E.; Post, R.J.; Ryan, K.; Schmidt, A.W.; Schulz, D.W.; Koe, B.K.J. Med. Chem. 1994, 37, 2509. [105] King, F.D.; Brown, A.M.; Gaster, L.M.; Kaumann, A.J.; Medhurst, A.D.; Parker, S.G.; Parsons, A.A.; Patch, T.L.; Raval, P.J. Med. Chem. 1993, 36, 1918. [106] Glennon, R.A.; Hong, S.-S.; Bondarev, M.; Law, H.; Dukat, M.; Rahkit, S.; Power, P.; Fan, E.; Kineau, D.; Kamboj, R.; Teitler, M.; Herrick-Davis, K.; Smith, C.J. Med. Chem. 1996, 39, 314. [107] LeBoulluec, K.L.; Mattson, R.J.; Mahle, C.D.; McGovern, R.T.; Nowak, H.P.; Gentile, A.J.Bioorg. Med. Chem. Letters 1995, 5, 123. [108] Schoeffter, P.; Hoyer, D.Naun. Schmied. Arch. Pharmacol. 1989, 339, 675. [109] (a) Van Lommen, G.; Bruyn, M.; Schroven, M.; Verschueren, W.; Janssens, W.; Verrelst, J.; Leysen, J. Bioorg. Med. Chem. Lett. 1995, 5, 2649. (b) Van de Water, A. D’Aubioul, J.; Van Gerven, W. Van Ammel, K.; De Clerck, F.Eur. J. Pharmacol. 1996, 299, 127. [110] a) Clitherow, J.W.; Scopes, D.I.C.; Skingle, M.; Jordan, C.C.; Feniuk, W.; Campbell, I.B.; Carter, M.C.; Collington, E.C.; Connor, H.E.; Higgins, G.A.; Beattie, D.; Kelly, H.A.; Mitchell, W.A.; Oxford, A.W.; Wadsworth, A.H.; Tyers, M.B. J. Med. Chem. 1994, 37, 2253. b) Clitherow, J.W.; Scopes, D.I.C.; Beattie, D.T.; Skingle, M. Exp. Opin. Invest. Drugs 1995, 4, 323. [111] Pauwels, P.J.; Colpaert, F.C.Neuropharmacol . 1995, 34, 235. [112] Walsh, D.M.; Beattie, D.T.; Connor, H.E.Eur. J. Pharmacol. 1995, 287, 79. [113] Stang, P.J.; Anderson, A.G.J. Org. Chem. 1976, 41, 781. [114] Streitwieser, A.; Dafforn, A.Tetrahedron Lett. 1976, 18, 1435.

28 Chapter 2

Synthesis and Preliminary Pharmacological Evaluation of 8- * OSO2CF3-2-aminotetralin Derivatives

Abstract

A series of (enantiopure) 8-triflate-substituted 2-(n-propylamino)tetralins has been synthesized and evaluated for in vitro binding to 5-HT1A, 5-HT1Dα and 5-HT1Dβ receptors and in in vivo biochemical and behavioural assays. Consequently, (R)-8- [[(trifluoromethyl)sulfonyl]oxy]-2-(n-propylamino)tetralin ((R)-3; Ki = 1.3 nM) was found to be a potent and selective 5-HT1A receptor agonist inducing a full-blown 5-HT behavioural syndrome and a decrease of 3.9 °C in body temperature, while (S)-3 appeared to be a partial 5-HT1A receptor agonist. The oral bioavailability of (R)-3 was low (7.6%), probably as the result of a relatively high clearance. In an attempt to improve the oral bioavailability the C1-methylated analogues cis-1-methyl-8- [[(trifluoromethyl)sulfonyl]oxy]-2-(n-propylamino)tetralin (cis-9), and its enantiomers were prepared. The activity was found to reside in the cis-(1S,2R)-9 enantiomer which displayed fairly high binding to 5-HT1A receptors (Ki = 7.1 nM) but moderate potency in postsynaptic 5-HT1A receptor agonist assays after subcutaneaous or oral administration. The optical antipode cis-(1R,2S)-9 seemed to be a selective 5-HT1A receptor ligand with low intrinsic efficacy. The cis-1-methyl-N-monomethyl derivative

(cis-10) displayed an enhanced affinity for 5-HT1Dα (Ki = 3.4 nM) and 5-HT1Dβ receptors (Ki = 10 nM), whereas trans-10 was inactive for the receptor subtypes tested.

2.1 Introduction

8-Hydroxy-2-(di-n-propylamino)tetralin (8-OH-DPAT, 1) is a potent, selective 1,2,3 and centrally active 5-HT1A receptor agonist. The clinical potential is low, due to extensive first-pass elimination via O-glucuronidation and N-depropylation, as was shown in the rat.4 The electron-withdrawing aryl trifluoromethanesulfonate (triflate) group is known as a chemically5 and biologically6 stable entity. Accordingly, 8-

* This chapter is partially based on : Sonesson, C.; Barf, T.; Nilsson, J.; Dijkstra, D.; Carlsson, A.; Svensson, K.; Smith, M.W.; Martin, I.J.; Duncan, J.N.; King, L.J.; Wikström H.J. Med. Chem. 1995, 38, 1319.

27 Chapter 2

[[(trifluoromethyl]sulfonyl]oxy)-2-(di-n-propylamino)tetralin (8-OSO2CF3-DPAT, 2) was found to retain high affinity for this receptor subtype, but displayed low potency in vivo.7 Interestingly, the absolute oral bioavailability proved to be higher than that of 8- OH-DPAT (11.2% vs 2.4%; Table 1). In addition to this observation, compound 2 was found to be more potent after oral (po) than subcutaneous (sc) administration, in the in vivo biochemistry assays, suggesting the formation of (an) active metabolite(s). The monopropyl analogue (8-OSO2CF3-PAT, 3) was reported to be the major metabolite (in rat hepatocytes) and was subsequently found to be more potent in vivo than 8-OSO2CF3- DPAT.8 In order to explore further the structure-affinity relationships (SAFIR) and pharmacology of this series of 2-aminotetralins, we prepared the enantiomers of 3.

Table 2.1. Absolute Oral Bioavailabilities of 2-Aminotetralins

R2 R3 8 1 N 7 2 R1 6 3 5 4

5-HT1A, oral avail,

R1 R2 R3 Ki (nM) % ref. 1a 8-OH H n-Pr 0.5 2.4 [1] 2 8-OTf H n-Pr 0.8 11.2 [7] 3 8-OTf H H 3.8 - [8] 4 8-OH Me n-Pr 2.1 - [9] 5b 5-OMe Me n-Pr - 3.7 [10] 6c 5-OMe Me H - 1.6 [10] 7 5-OTf Me n-Pr - 9.4 [10] 8 5-OTf Me H - 62.2 [10]

(a) 8-OH-DPAT; (b) UH232; (c) AJ76.

The triflate concept can also be implemented for the cis- and trans-C1- methylated 2-aminotetralins, which were shown to exhibit interesting SAFIR for the 5-

HT1A receptor (see Chapter 1). The C1-methylated 2-aminotetralin cis-(1S,2R)-4 retains the selectivity and potency of the parent compound 8-OH-DPAT.11 This observation, and extrapolation of the substitution patterns of autoreceptor antagonists UH232 (5) and AJ76 (6) and their respective triflated analogues 7 and 8 led us to

28 Synthesis and Pharmacological Evaluation of 8-OSO2CF3-2-aminotetralin Derivatives

* predict 2-aminotetralin cis-8-OSO2CF3-MPAT (cis-9) to have superior oral bioavailability, without the loss of affinity or in vivo activity compared to the non- methylated congener 3. The relevant 5-HT1A receptor binding data and absolute oral bioavailablities of the known 2-aminotetralins are listed in Table 2.1.

The molecular structure of the selective 5-HT1D receptor agonist sumatriptan 12 (11) and the improvement in affinity for the 5-HT1D receptor by substituting the phenolic group of 1 with a triflate group (see Chapter 4), prompted us to replace the N- monopropyl substituent of compounds cis- and trans-9 (cis- and trans-8-SO2CF3- MMAT)* by an N-methyl substituent, enabling us to investigate the influence on the 5-

HT1A and 5-HT1D receptor affinity and selectivity.

OTf Me H OTf Me H CH2SO2NHMe N N Me Me N Me

cis-9 cis-10 N trans-9 trans-10 H 11

2.2 Chemistry

Preparation and Resolution of ( ±)-8-OSO 2CF 3-PAT (3). The syntheses of the pure enantiomers of (R)-3 and (S)-3 are outlined in Scheme 2.1. 8-Methoxy-2-tetralone (12) was prepared according to literature procedures from commercially available 1,7- dihydroxynaphthalene by O-methylation followed by a Birch reduction.13 Racemic 13 was prepared by a reductive amination reaction using either n-propylamine with sodium cyanoborohydride (Borch reduction) or Dean-Stark conditions through the formation of an enamine and subsequent catalytic hydrogenation. The chiral cyclic phosphoric acid, 2-chlocyphos (14), was reported to be an efficient resolving agent for the resolution of amines.14 Thus, by using (R)-14 the separation of (R)-13 (15%) was effected by two recrystallizations from 2-propanol, after which enriched (S)-13 was recovered from the mother liquor and optically purified by recrystallization with (S)-14 in the same yield. The enantiomeric excesses (e.e.) were >98% as determined by HPLC analysis using a chiral column (Chiracel OD, Daicel). In addition, the optical rotations matched those reported in the literature.15 The enantiomers of the phenol derivative 15, which were obtained by refluxing (R)- and (S)-13 in 48% aqueous HBr, were triflated

* MPAT = 1-methyl-(n-propylamino)tetralin. * MMAT = 1-methyl-(methylamino)tetralin.

29 Chapter 2 using the mild triflating agent N-phenyltrifluoromethanesulfonimide.16 Phase-transfer conditions were employed in order to prevent substitution of the secondary amine. Our efforts to obtain suitable crystals of the salt of (S)-3 and (S)-14 for single X-ray crystallography succeeded, but due to the poorly resolved atomic positions of the triflate functionality the fluorine atoms could not be found. Either conformational flexibility of the triflate group in the crystal packing or the fact that the triflate group is fixed in various positions may contribute to this observation. This may be the reason why X-ray data of aromatic triflate containing molecules have not been reported previously.

OMe H c,d N

OMe OMe H S-(−)-13 O a,b N e

OMe H 12 13 c,d N

OH H OTf H R-(+)-13 N f N

Me Me S-(−)-15 S-(−)-3 R-(+)-15 R-(+)-3

O O Cl P O OH 14

Scheme 2.1. Reagents and conditions: a)n -PrNH2, p-TsOH, toluene, ∆; (b) 10% Pd/C, H2, EtOH; (c) Resolution with 2-chlocyphos ((R)-16 or (S)-16); (d) NaOH, CH2Cl2; (e) 48% HBr, ∆; (f) PhN(Tf)2, TBAHSO4, 10%

NaOH, CH2Cl2.

Another approach, in which 14 may be applied for the resolution of 2- aminotetralins, is outlined in Scheme 2.3. Chiral phosphorinane derivatives were shown to be successful as derivatizing agents of amino acids, alcohols and amines and in the enantiomeric excess determination of these nucleophiles by means of 1H and 31P NMR.17 We envisaged derivatization of a small amount of racemic 2-aminotetralin with the phosphorinane analogue of 2-chlocyphos, which subsequently might be separated by

30 Synthesis and Pharmacological Evaluation of 8-OSO2CF3-2-aminotetralin Derivatives means of column chromatography. In addition, these chiral phosphorinane derivatives may serve as protective groups since the phosphamidates can easily be deprotonated and N-alkylated giving (enantiopure) N-monosubstituted 2-aminotetralins after deprotection.18 According to a method described by Hulst et al. for the deschloro derivative of 14, compound 17 is readily obtained by reduction of compound 14 to the diol 16 using LiAlH4 in THF, followed by treatment with PCl3. An Arbuzov rearrangement19 using ethanol yields (R)-2H-2-oxo-4-(R)-(2-chlorophenyl)-5,5- dimethyl-1,3,2-dioxaphosphorinane ((2R,4R)-17; Scheme 2.2).

Me Me Me Me Me Me a b,c

O O Cl O O Cl P OH OH Cl P HO O H O (R)-14 (R)-16 (2R,4R)-17

0 Scheme 2.2. Reagents and conditions: (a) LiAlH4, THF, ∆; (b) PCl3, benzene, 0 C-RT; (c) ethanol.

The coupling of 8-methoxy-2-aminotetralin (18) and (2R,4R)-17 is effected by using the Atherton-Openshaw-Todd coupling (Scheme 2.3).20 The reaction proceeds with inversion of configuration at the phosphorus atom via putative formation of a trichloromethylphosphonate derivative, which is substituted by the aminotetralin. The 31P NMR of the diastereomeric mixture of 19 showed a small chemical shift of 0.096 ppm resulting in an overlapping signal with a ratio of 67:37 which suggests that the formation of one of the diastereomers is favored. Unfortunately, our efforts to separate the isolated diastereomeric mixture using TLC (on SiO2; eluting with various combinations of solvents) did not succeed.

OMe Me Me Me Me

NH2 a b separated + diastereomers O O Cl O O Cl P P H O N O OMe H (±)-18 (2R,4R)-17 19

Scheme 2.3. Reagents and conditions: (a) CCl4, Et3N, ethanol, 0 °C; (b) TLC on SiO2.

Preparation and Resolution of cis-(±)-8-OSO 2CF 3-MPAT ( cis-9). The synthesis of cis-8-[[(trifluoromethyl)sulfonyl]oxy-1-methyl-2-(n-propylamino)tetralin (cis-9) is outlined in Scheme 2.4. The Stork enamine reaction was employed to

31 Chapter 2 introduce the methyl group on the C1-position of the 2-tetralone skeleton according to the method of Arvidsson et al.11 with the exception of the purification procedure (column chromatography instead of distillation), affording cis-21 in a 55% yield after recrystallization as the HCl salt. The reductive amination proceeded with a cis/trans ratio of 90:10 as determined by GC-MS. The demethylation and triflation reactions were accomplished as described for the preparation of the enantiomers of 3. The resolution of cis-9 was effected at the stage of cis-21 according to the method of Arvidsson and co-workers using di-p-toluoyl-tartaric acid.11a The optical antipodes cis- (1S,2R)-9 and cis-(1R,2S)-9 were prepared as described above. The enantiomeric purity was determined on the hydroxy derivatives by HPLC using a chiral column (Chiralpak AD, Daicel), eluting with n-hexane/ethanol/diethylamine (98/2/0.1 v/v/v). The e.e. of cis-(1S,2R)-22 was 99% whereas cis-(1R,2S)-22 was shown to have an e.e. of 98.6%.* Interestingly, the enantiomers of the triflate derivatives exhibited reversed optical rotations.

OMe Me H f N

OMe OMe Me cis-21 O a,b,c O d,e g

OMe Me H 12 20 f N

OH Me H OTf Me H trans-21 N h N

cis-22 cis-9 trans-22 trans-9

Scheme 2.4. (a) pyrrolidine,p -TsOH, benzene, ∆; (b) MeI, dioxane, ∆; (c) H2O, acetic acid, ∆; (d) n-PrNH2, p-

TsOH, toluene, ∆; (e) 10% Pd/C, H2, EtOH; (f) column chromatography on SiO2 eluting with CH2Cl2/MeOH

(20:1); (g) 48% HBr,∆ ; (h) PhN(Tf)2, triton-B, 10% NaOH, CH2Cl2.

Preparation and cis- and trans-(±)-8-OSO 2CF 3-MMAT ( cis- and trans-10). Analogous to the chemistry as utilized for the synthesis of cis- and trans-9, methylamine can be condensed with the 2-tetralone giving compounds cis- and trans- 10 after O-demethylation and triflation (Scheme 2.5). We chose benzylamine for the

* A previous batch ofcis -(1R,2S)-22 displayed an e.e. of 82%.

32 Synthesis and Pharmacological Evaluation of 8-OSO2CF3-2-aminotetralin Derivatives reductive amination in order to have the primary amine within reach via debenzylation.21 The cis/trans ratio obtained after the first step was 50:50 as was determined by GC-MS analysis. After separation using column chromatography, the cis- and trans-isomers of 23 underwent successive N-methylation, N-debenzylation and O-demethylation. The triflation was carried out employing triton-B (benzyltrimethyl ammonium hydroxide) as the phase-transfer catalyst since tetrabutyl ammonium hydrogen sulfate gave an unexpected by-product containing a butyl group. However, the exact structure of this compound has not been elucidated.

OMe Me H

c N

OMe Me OMe Me Me cis-23 O a,b d N

OMe Me H 20 cis-24 c N trans-24

trans-23

OMe Me H OH Me H OTf Me H

e N f N g N Me Me Me

cis-25 cis-26 cis-10 trans-25 trans-26 trans-10

Scheme 2.5. (a) benzylamine,p -TsOH, benzene, ∆; (b) PtO2, H2, MeOH; (c) column chromatography on SiO2 eluting with CH2Cl2/MeOH (20:1); (d) 37% formaldehyde, NaCNBH3, pH 5, CH3CN; (e) 10% Pd/C, H2, EtOH; (f) 48% HBr, ∆; (g) PhN(Tf)2, triton-B, 10% NaOH, CH2Cl2.

N-methylation of cis- and trans-23 resulted in the conformationally restricted cis- and trans-24. The presence of the C1-methyl causes steric hindrance which prevents proper rotation of the dialkylamino group around the C2-N bond. Surprisingly, the 1H NMR and 13C NMR spectra of each of the cis- and trans-isomers of 24 showed two populations. The ratio of these populations, determined by the integration of a number of individual signals, appeared to be solvent-depended for both cis- and trans-

24. In CD3OD the population ratio was approximately 59:41 for cis-24 and 42:58 for trans-24, but displayed a different ratio in DMSO-d6 (cis-24; 49:51 and trans-24;

33 Chapter 2

37:63). We performed heating experiments in DMSO-d6 to investigate whether the populations represented different conformations that could isomerize at higher temperatures or that the two populations would stay locked in a certain conformation. Normally, the C1-methyl group gives one doublet in case of secondary amines. At 25 °C the C1-methyl group of cis-24 showed two doublets at 1.27 ppm and a coupling constant of J = 13.91 Hz separated by 0.047 ppm. At 100 °C, the two doublets coincided (J = 5.85 Hz) at 1.15 ppm and upon cooling to 25 °C the signal pattern resembled that of the first 1H NMR spectrum, displaying the same shift and coupling constant (Figure 2.1). Importantly, the conformational populations adopted the same ratio as before the heating experiment, implying that these two populations most probably have an equilibrium at room temperature. Similar observations were found with trans-24, however, the two doublet signals were shown to have a non-complete coalescence at 100 °C in DMSO-d6 (not shown).

25 °C 100 °C 25’ °C

Figure 2.1. 1H NMR signals of the C1-methyl of cis-24 at 25, 100 and 25 °C, respectively (DMSO-d6)

Cis- and Trans-assignment. Compounds cis- and trans-21 were prepared according to literature procedures and for the O-demethylated 2-(di-n- propylamino)tetralins the stereochemistry has been resolved by means of X-ray crystallography and NMR spectroscopy.22 The spectroscopic data could be used in order to assign the correct stereochemistry to cis- and trans-23, which are new compounds. The pattern of 1H NMR signals, corresponding to specific protons of cis and trans compounds, is consistent throughout the C1-methylated 2-aminotetralin series. In line with the observations of Arvidsson et al., the H1β signals of the cis compounds usually exhibit a higher chemical shift than that of the H1α signals of their trans isomers (Table 2.2). In addition, the coupling constants J1β,2β of the cis compounds ∼ ∼ (J 5 Hz) are larger than the coupling constants J1α,2β of the trans compounds (J 1

34 Synthesis and Pharmacological Evaluation of 8-OSO2CF3-2-aminotetralin Derivatives

Hz). When CDCl3 was employed as the solvent, the coupling constant J1α,2β was generally measureable as exemplified by trans-10, being 1.09 Hz. Additional evidence was generated by the independent preparation of cis-21 from cis-23 via n-propylation and N-debenzylation, providing identical GC-chromatograms for the two compounds (Scheme 2.6).

OMe Me H OMe Me OMe Me H

N a N b N

cis-23 cis-27 cis-21

Scheme 2.6. (a) propionaldehyde, NaCNBH3, pH 5, CH3CN; (b) 10% Pd/C, H2, EtOH. Table 2.2. 1H NMR Spectral Data of Compounds21 and 23. chemical shift, δ (ppm)

compound H1α H1β H2β cis-21 Me 3.61 3.46

R'O R1β R1α trans-21 3.42 Me 3.51 H2β NHR" cis-23 Me 3.46 2.94a trans-23 3.20 Me 2.95a coupling constant, J (Hz) cis : R1α = Me; R1β = H α β α β β β trans: R1 = H; R1 = Me compound J1 ,2 J1 ,2 cis-21 - 5.13 trans-21 b - cis-23 - 5.49 trans-23 b -

(a) Obscured. (b) Too small to determine.

2.3 Pharmacology

Receptor Binding. Compounds (R)- and (S)-3 were evaluated for their in vitro 3 binding affinity at 5-HT1A receptors using [ H]8-OH-DPAT, at 5-HT1Dα and 5-HT1Dβ 3 3 using [ H]5-HT, at dopamine D2 receptors using either the antagonist [ H]spiperone or 3 3 the agonist [ H]U86170, and at dopamine D3 receptors using [ H]spiperone (Table 2.3). The 1-methyl substituted 2-aminotetralins 9 and 10 were tested for their abilities to

35 Chapter 2

3 3 compete with the radioligand [ H]8-OH-DPAT (5-HT1A) and [ H]5-CT (5-HT1Dα and 5-

HT1Dβ). All above receptors, except the dopamine D2 receptor (rat), are human clones.

In Vivo Biochemistry of ( R)- and ( S)-8-OSO 2CF 3-PAT (( R)- and ( S)-3). The in vivo biochemical test utilizes the well-established phenomenon of receptor-mediated inhibition of the presynaptic neuron. The synthesis rate of 5-HT is inhibited by 5-HT1A receptor agonists. 5-Hydroxytryptophan (5-HTP) accumulation, following decarboxylase inhibition by (3-hydroxybenzyl)hydrazine (NSD 1015), was used as an indicator of the 5-HT turnover in three different brain areas (Table 2.4). For this study we used both nonpretreated and reserpine-pretreated rats (5 mg/kg sc, 18 h). This model is designed to detect directly acting agonists (with various degrees of intrinsic activity) at central 5-HT receptors through both biochemical and behavioural effects (Tables 2.4 and 2.5, respectively).

Table 2.3. Affinities at cloned 5-HT1A, 5-HT1Dα, 5-HT1Dβ, and D2 Receptors In Vitro

Ki ± SEM (nM)a

Compound 5-HT1A 5-HT1Dα 5-HT1Dβ D2 1 0.5 ± 0.02 164 ± 30 638 ± 75 90 ± 4 2 0.8 ± 0.1 12 ± 4 127 ± 26 62 ± 7 3 2.8 ± 0.4 15 ± 1 169 ± 17 108 ± 12 (R)-3 1.3 ± 0.3 6.7 ± 0.5 138 ± 22 69 ± 4 (S)-3 13 ± 0.7 157 ± 15 1255 ± 344 225 ± 14 cis-9 6.1b 15.7b 125b NT cis-(1S,2R)-9 7.1b 12b 60b NT cis-(1R,2S)-9c 7.9b >1000b 200b NT cis-10 5.7b 3.4b 10b NT trans-10 >1000b >1000b >1000b NT

3 (a) Ki values for displacement of 5-HT1A receptor agonist [ H]8-OH-DPAT, 5-HT1Dα and 5-HT1Dβ receptor 3 3 agonist [ H]5-HT, and dopamine D2 receptor agonist [ H]U86170. Data from dopamine D2 and D3 antagonists binding were higher than 300 nM and are not shown. Data from cloned mammalian receptors expressed in CHO-K1 cells (for experimentals see Section 4.5; receptor binding - method A). Compound1- 3 and the enantiomers of 3 were tested at the Upjohn Company, whereas compound cis-9, 10 and the enantiomers ofcis - 9 were tested at Centre de Recherche Pierre Fabre. (b) Value obtained from a single experiment (for the experimentals see Section 5.5). (c) e.e. 82%.

Table 2.4. Effects on Rat Brain 5-HT Synthesis Rates (5-HTP Accumulation)In vivo in Reserpine-Pretreated and Nonpretreated Rats. 5-HTP accumulationa

36 Synthesis and Pharmacological Evaluation of 8-OSO2CF3-2-aminotetralin Derivatives

b reserpine-pretreated rats, ED50 (µmol/kg) nonpretreated rats, % of ctrl Compound limb striatum hemi limb striatum (R)-1 sc 0.036 0.047 0.05 50 ± 3***c 48 ± 4***c 2 sc 8.3 26.9 13.8 NTd NT 2 poe 1.2 1.5 1.2 NT NT 3 sce 1.2 0.8 1.1 NT NT (R)-3 sc 0.50 0.62 0.93 39 ± 4***f 44 ± 3***f (R)-3 po NT NT NT 63 ± 6***f 62 ± 4***f (S)-3 sc 24.0 24.0 28.2 52 ± 3***g 53 ± 3***g

(a) The animals were treated with the test drug 60 min and NSD1015 30 min before decapitation. Reserpinized animals received reserpine 18 h before drug treatment. Shown are the values producing a half-maximal decrease in the accumulation of 5-HTP in the limbic, striatal and hemispheral brain areas. (b) The values are the percent of control, means± SEM (n = 16 and n = 4 in control and tested groups, respectively). (c) Dose 0.25 µmol/kg. (d) NT means Not Tested. (e) Taken from ref8 . (f) Dose 25 µmol/kg. (g) Dose 50 µmol/kg. ***P ≤ 0.005.

Locomotor Activity and Gross Behavioural Observations. Postsynaptic agonistic effects of the test compounds were assessed in normal rats and by reversal of reserpine-induced hypokinesia. Postsynaptically acting dopamine agonists induce locomotor activity while 5-HT1A receptor agonists induce the 5-HT behavioural syndrome (flat body posture, forepaw treading (piano playing), abducted hind limbs and straub tail; Table 2.5).23 Compound cis-9 and its enantiomers were additionally screened for their ability to induce the lower lip retraction.24 Oral administration was performed by gavage to animals that had been fasted for 18 h.

37 Chapter 2

Table 2.5. Effects on 5-HT Behavioural Syndrome in Normal and Reserpinized Rats.

5-HT syndromea normal (dose)b normal (dose) reserpine (dose) reserpine (dose) Compound sc po sc po Vehicle 0/25 - 0/12 - (R)-1 4/4 (0.25) NTc 4/4 (0.25) NT (R)-3 4/4 (25) 3/4 (25)d 4/4 (1.6) NT (S)-3 0/4 (50) NT 4/4 (50) NT cis-9 0/4 (25) 0/4 (25) 0/4 (25) 1/4 (25)d cis-(1S,2R)-9 3/3 (25)d 1/3 (25)d 4/4 (25)d 4/4 (25)d cis-(1R,2S)-9 0/4 (25) 0/4 (25) 0/4 (25) 0/4 (25)

(a) Shown is the number of rats displaying the 5-HT syndrome (flat body posture, reciprocal forepaw treading “pianoplaying”, straub tail. (b) Dose inµ mol/kg. (c) NT means Not Tested. (d) Rats only displayed flat body posture and lower lip retracion.

Table 2.6. Locomotor Activity in Reserpinized-Pretreated and Nonpretreated Rats.

Counts/ 30 mina normal (dose)b normal (dose) reserpine (dose) reserpine (dose) Compound sc po sc po Vehicle 226±33 - 18±9 - (R)-1 330±70 NTc 235±82** NT cis-9 229±67 132±42 95±46 58±26 cis-(1S,2R)-9 288±46 175±64 52±6 87±22 cis-(1R,2S)-9 285±58 247±43 6±5 10±6

(a) The animals were treated with the test drug 30 min before the 30-min motility test. Reserpinized animals received reserpine 18 h before drug treatment. (b) Dose inµ mol/kg. (c) NT means Not Tested. **P ≤ 0.01.

Oral Bioavailability and In Vitro Metabolism of ( R)-8-OSO 2CF 3-PAT (( R)-3). The absolute oral bioavailability of (R)-3 was determined by measuring the plasma concentrations after both oral and intraveneous administration. Blood samples were collected at various time intervals up to 12 h after drug administration. The doses were 25 µmol/kg (po, n = 5) and 5 µmol/kg (iv, n = 3). The test compound was administered orally by gavage to animals that had been fasted for 18 h. The metabolism of (R)-3 was studied following incubation with suspensions of rat isolated hepatocytes. The metabolic profiles were examined by thermospray (TSP) LC/MS with or without β-

38 Synthesis and Pharmacological Evaluation of 8-OSO2CF3-2-aminotetralin Derivatives glucuronidase/sulfatase treatment of incubates. Structural information on metabolites was obtained by the MS/MS daughter ions analysis (Table 2.7).

Table 2.7. Pharmacokinetic Data for Compounds1 , 2 and (R)-3 in the Rat.

AUC ratio half-life, clearance, Compound po/iv, %a minb mL/min kg 1 2.4 ± 0.9c 72 NT 2 11.2 ± 5.2d 90 57 (R)-3 7.6 ± 1.1e 110 75

(a) Blood samples were taken via artherial catheters. The absolute oral bioavailability was estimated by comparing the areas under the curves (AUC) in graphs in which the drug concentrations were plotted against time (n = 4 for both administration routes). (b) The half-lives were estimated graphically from the elimination phase of the blood-concentration curves after oral administration. (c) Dose 20 (po) and 1 (iv) µmol/kg. Number taken from ref.4 . (d) Dose 40 (po) and 5 (iv) µmol/kg. Number taken from ref.8 . (e) Dose 25 (po) and 5 (iv) µmol/kg.

Hypothermia. Postsynaptic activation of 5-HT1A receptors elicits hypothermia in the rat.25 Compound (R)-3 and cis-(1S,2R)-9 were administered in a dose of 25 µmol/kg to normal rats. The animals that received the test compound orally were fasted for 18 h. Table 2.8 gives the effect after 30 min, as well as the maximal effect of each of the test compounds. In order to compare the relative effects, the area under the curve (AUC) for each compound was estimated in the range from the control body temperature until the maximal hypothermic effect (Figure 2.2).

Table 2.8. Effects on Body Temperature in Rats ∆T (°C)a max. ∆T (°C) ∆T (°C)a max. ∆T (°C) AUCb compound sc sc po po po/sc (R)-3 −2.0±0.3** −3.9±0.3** −2.0±0.2** −2.4±0.4** 32 % cis-(1S,2R)-9 −1.7±0.3** −2.3±0.3** −1.7±0.3** −1.7±0.3**c 47%

(a) Change in body temperature ±( SEM) measured 30 min after administration of the test compound (n = 4). Dose of 25 µmol/kg. (b) Estimated from the curves obtained after polynomial regression. (c) n = 3.** P<0.01 2.4 Results and Discussion

Structure-Affinity Relationships. Obviously, the affinity of racemic 3 resides in the R-enantiomer, which exhibited a Ki of 1.3 nM for 5-HT1A sites. The same is true for the 5-HT1Dα and 5-HT1Dβ receptor subtype as well as the dopamine D2 agonist

39 Chapter 2

binding. The S-enantiomer of 3 was found to be 10-fold less potent in 5-HT1A receptor binding, improving the two-fold stereoselectivity observed for the enantiomers of 8- OH-DPAT. As predicted, compound cis-9 is shown to have a comparatively similar binding profile with 3 exhibiting a two fold lower affinity for 5-HT1A receptors (Table 2.3). Surprisingly, the enantiomers of cis-9, unlike (R)- and (S)-3, displayed no stereoselectivity al all for the 5-HT1A receptor subtype, whereas the 5-HT1D receptor subtypes clearly discriminate between the two antipodes. When it comes to a comparison of the affinities of (S)-3 and cis-(1R,2S)-9 for the receptor subtypes considered, it is obvious that the introduction of a methyl group on the C1-position has a dramatic influence. The absence of affinity of cis-(1R,2S)-9 for 5-HT1Dα sites (Ki

>1000 nM) and the low affinity for the 5-HT1Dβ site (Ki = 200 nM) results in a fairly selective 5-HT1A receptor ligand (Ki = 7.9 nM). The replacement of the N-n-propyl substituent of cis-9 by a N-methyl group expectedly improved the binding by approximately 5- and 13-fold for the 5-HT1Dα and 5-HT1Dβ sites, respectively. This leads to the assumption that the propyl group is too large in order to be properly accommodated by both 5-HT1D receptor subtypes and moreover, that the 5-HT1Dβ receptor is more sensitive to bulk at N-substituents than is the 5-HT1Dα receptor (see also

Chapter 4). No major change was observed in the affinity for the 5-HT1A receptor, resulting in a rather non-selective compound. The trans-isomer of 10 is essentially inactive at the receptor subtypes tested. Structure-Activity Relationships. On the basis of the data presented in Tables

2.3-2.8, it may be concluded that the enantiomers of compound 3 are 5-HT1A receptor agonists, similar in profile to 8-OH-DPAT. The (R)-enantiomer of 3 displays a potent and selective interaction with 5-HT1A receptors and is approximately 10 times less potent than 1 (reserpinized animals Table 2.4). Interestingly, in nonpretreated rats, a high dose (25 µmol/kg, sc) of (R)-3 induced a full-blown 5-HT1A behavioural syndrome along with a maximal decrease in 5-HTP accumulation, indicative of a 5-HT1A receptor agonist with full intrinsic activity (Tables 2.4 and 2.5). In contrast, the (S)-enantiomer was found to be a weak agonist (> 40 times less potent in vivo compared to (R)-3) and did not induce the 5-HT behavioural syndrome in nonpretreated rats (Tables 2.4 and 2.5). The fairly high affinity in vitro (Ki of 13 nM; Table 2.3), along with the low potency and intrinsic activity at 5-HT1A receptors, may reflect possible antagonistic properties of (S)-3. Consequently, this compound was tested for the ability to antagonize the behavioural action induced by (R)-3. Interestingly, the forepaw treading of (R)-3 (1 µmol/kg, sc) in nonpretreated rats was nearly completely blocked by (S)-3 (50 µmol/kg, P < 0.05), while the flat body posture was not affected at all. No

40 Synthesis and Pharmacological Evaluation of 8-OSO2CF3-2-aminotetralin Derivatives antagonism of the biochemical effects was observed, suggesting that (S)-3 is a partial 5-

HT1A receptor agonist. The pharmacological profile of (R)-3 sharply contrasts with the absence of activity reported for compound 2. This is intriguing, since both compounds exhibited 8 similar in vitro affinities for the 5-HT1A receptor (Table 2.3). Also Liu et al. reported the inability of both enantiomers of 2 to produce the 5-HT syndrome, hypothermia, or changes in the 5-HT turnover after sc administration.7 The lack of central effects were attributed to putative formation of inactive metabolites, or to the inability to penetrate the brain. The predicted logD values for compounds 1 and 3 were 1.8 and 2.0, respectively. However, the logD value for compound 2 was calculated to be 3.8, suggesting that the latter compound may be too lipophilic, allowing greater penetration of fat tissue in the rat. The first hypothesis, the formation of inactive metabolites, was opposed by Sonesson et al., who demonstrated that the major metabolite of 2 is the monopropyl analogue 3, and that 3 was more potent in vivo than 2. Indeed, 2 was more active after oral (po) than after subcutaneous (sc) administration, which supports the notion of first-pass metabolism to a pharmacologically more active compound. N- Dealkylation of 2, to yield 3, was shown to be the major metabolic pathway, as well as further metabolism to the primary amine. Oxidation was another important pathway, although the relative responses of the various metabolites were unknown, making quantification speculative. The biochemical and behavioural data from Table 2.3 and 2.4 reveal that (R)-3 elicits a maximal effect when the compound is administered sc, but not po, to normal rats. When (R)-3 was incubated with rat isolated hepatocytes, the major metabolite was the primary amine resulting from N-dealkylation. Minor oxidized metabolites were also observed. It is therefore likely that (R)-3 is metabolized by hepatic N-dealkylation when administered orally. The corresponding primary amine has not yet been synthesized and its effects remain to be tested. The oral bioavailability of (R)-3 is lower than that of 2 (7.6 vs 11.2%; Table 2.7), which may reflect the slightly higher clearance value obtained for (R)-3. In a semi-quantitative assay, the metabolism of (R)-3 in vitro was comparatively slower than that of 2, providing indirect evidence that (R)-3 would undergo less extensive first-pass metabolism in vivo. However, the oral bioavailability is slightly lower, indicating that other factors, such as absorption or inhibition of metabolism by a metabolite, play an important role in determining the bioavailability of these compounds.

Although very similar in binding profile, much of the efficacy of cis-9 for 5-HT1A receptors is lost, as compared to 3. In line with previously reported results, the behavioural pharmacology (Table 2.5) clearly indicates that cis-(1S,2R)-9 is the most active enantiomer. However, unlike (R)-3, cis-(1S,2R)-9 was not able to induce a full-

41 Chapter 2 blown 5-HT syndrom in normal rats at the 25-µmol/kg dose (sc), indicating its weaker

5-HT1A receptor agonist activity. A parameter which has been proposed as an index of

5-HT1A receptor-mediated activity is the hypothermic response to systemic 25 administration of 5-HT1A receptor agonists. The maximal hypothermic response to cis- (1S,2R)-9 was 2.3 °C, and comparatively lower than the lowering of the body temperature induced by (R)-3 (3.9 °C). Interestingly, the maximal decrease of core temperature in rats induced by 8-OH-DPAT was reported to be approximately 2.2 °C (2 µmol/kg, sc) after 30 min.25 t (min)

0 30 60 90 120 150 180 0.0 (R)-3 (25 µmol/kg sc ) -0.5 (R)-3 (25 µmol/kg po) cis-(1S,2R)-9 (25 µmol/kg sc ) -1.0 cis-(1S,2R)-9 (25 µmol/kg po)

-1.5

-2.0

T (°C) -2.5 ∆

-3.0

-3.5

-4.0

-4.5

Figure 2.2. The effects of (R)-3 and cis-(1S,2R)-9 on the body temperature in rats after sc and po administration.

Figure 2.2 shows the curves that were obtained when following the hypothermic response in time. We graphically estimated the area under the curve (AUC) of the individual compounds and administration routes, enabling us to compare the relative availabilities in brain. The AUC following po administration of (R)-3 was 32% of that obtained via the sc administration route, which is much greater than the oral bioavailablity of 7.6%. Interestingly, the oral availability of cis-(1R,2S)-9 improved to approximately 47%, as compared to (R)-3, although the relative maximal response of cis-(1R,2S)-9 after sc administration was only 43%. Supportive information is provided by the behavioural profiles, since cis-(1R,2S)-9 elicited a partial 5-HT syndrome when administered via the sc or po route to reserpinized rats. In normal animals, the po administration route was less efficacious in evoking the 5-HT syndrome as compared to

42 Synthesis and Pharmacological Evaluation of 8-OSO2CF3-2-aminotetralin Derivatives sc administration. It is expected that cis-9 is able to penetrate the brain in sufficient amounts, as is indicated by its calculated logD value of 2.4. Taken together this suggests that cis-(1S,2R)-9 is a 5-HT1A receptor agonist with moderate potency but with increased oral availablity relative to (R)-3. Despite the fairly high Ki value of 7.9 nM for 5-HT1A sites, cis-(1R,2S)-9 was devoid of any postsynaptic 5-HT1A-receptor agonist activity. Either, the intrinsic efficacy is low or this compound behaves as a 5-HT1A receptor antagonist. If this is the case, this would contribute to the inability of racemic cis-9 to induce the 5-HT behavioural syndrome. Any possible antagonistic effects of cis-(1R,2S)-9 need to be investigated by means of behavioural and neurobiochemical experiments.

In summary, (R)-3 behaved as a full 5-HT1A receptor agonist when administered subcutaneously, but not orally, to normal rats. As indicated by in vitro hepatocyte experiments, N-depropylation seems to be the major metabolism pathway after po administration, resulting in the corresponding primary amine. This observation and the higher clearance value, as compared to 2, presumably account for the relatively low oral availability of 7.6%. In an attempt to improve the oral bioavailability cis-9, and its enantiomers were prepared and subjected to preliminary behavioural pharmacology studies. The activity resides in the cis-(1S,2R)-9 enantiomer which displayed fairly high binding to 5-HT1A receptors but moderate potency in postsynaptic 5-HT1A receptor agonist assays. In addition, the optical antipode cis-(1R,2S)-9 may turn out to be a 5-

HT1A receptor ligand with low intrinsic activity. Both enantiomers of cis-9 deserve further investigation on the basis of the presented results. Moreover, the C-1 methyl substituent may interfere with the site which is responsible for the extensive first-pass metabolism of compounds 2 and (R)-3. This may result in drastically improved pharmacokinetic properties for (+)- and (−)-cis-9, compared to the non-C1 substituted 2-aminotetralins.

2.5 Experimental Section

General. 1H and 13C NMR spectra were recorded at 200 and 50.3 MHz, respectively, on a Varian Gemini 200 spectrometer. CDCl3 was employed as the solvent unless otherwise stated. Chemical shifts are given in δ units (ppm) and relative to TMS or deuterated solvent. The splitting patterns are designated as follows: s (singlet), d (doublet), t (triplet), q (quartet), m, (multiplet), br (broad), dd (double doublet) and ddd (double double doublet). The heating experiments with cis- and trans-24 were recorded

43 Chapter 2 with a 300 MHz Varian VXR-300 spectrometer. IR spectra were obtained on a ATI- Mattson spectrometer. Elemental analyses were performed in the Microanalytical department of the University of Groningen or at Parke-Davis (Ann Arbor, MI). The chemical ionization (CI) mass spectra were obtained on a Finnegan 3300 system. GC- MS (EI) mass spectra were recorded on a Unicam 610/ Automass 150 GC/MS system. Melting points were determined on a Electrothermal digital melting point apparatus and are uncorrected. Specific optical rotations were measured in methanol (c 1.0 if not stated otherwise) at 23 °C on a Perkin Elmer 241 polarimeter. Materials . 1,7-dihydroxynaphtalene was purchased from Tokyo Kasei Kogyo Co, Ltd (Japan). 8-Methoxy-2-tetralone was prepared according to literature procedures. (R)- and (S)-2-chlocyphos were obtained from Syncom B.V. (The Netherlands). All further chemicals used were commercially available (Aldrich) and were used without further purification. (±)-8-Methoxy-2-( n-propylamino)tetralin (13). 8-Methoxy-2-tetralone (16.6 g, 94.0 mmol), n-propylamine (15.0 mL, 183.0 mmol) and a spatula of p-TsOH were refluxed under N2-atmosphere in toluene (400 mL) under Dean-Stark conditions. After 5 h the volatiles were evaporated in vacuo leaving a brown oil, which was immediately dissolved in dry THF (400 mL). The resulting solution was acidified with ether/HCl until pH 5 after which methanol (30 mL) and NaCNBH3 (8.45 g, 134.0 mmol) were added. The reaction mixture was magnetically stirred for 18 h, evaporated to dryness and taken up in saturated aqueous Na2CO3 (500 mL). The aqueous layer was extracted with ether (3 × 150 mL) which was dried over Na2SO4 and evaporated in vacuo affording 22.3 g of a brown oil which converted to the HCl salt and recrystallized from MeOH/ether (off-white material, 75%). Resolution of ( ±)-8-Methoxy-2-( n-propylamino)tetralin. A mixture of racemic amine (11.1 g, 50.7 mmol) and (R)-(+)-2-chlocyphos (14.0 g, 50.7 mmol) in abs. ethanol (200 mL) was refluxed until all material was dissolved after which the solvent was removed in vacuo giving an off-white solid. The salt (24.1 g, 48.7 mmol) was recrystallized from 2-propanol yielding 5.16 g (10.42 mmol, 21%) of white crystals with [α]D +53.1°. A second recrystallization gave 3.74 g (7.56 mmol, 16%) salt with [α

]D +60.1°. This salt (3.65 g, 7.37 mmol) was converted to the free base by stirring in

10% KOH (50 mL), extraction with ether and drying over Na2SO4. Evaporation of the solvent in vacuo yielded (R)-(+)-13 (1.58 g, 15%) as a colorless oil with [α]D +76.4° 15 22 (lit [α] D +78.3° (c 1.05)). The residual salt (11.2 g, 22.7 mmol) was converted to the free base described as above using 10% KOH (100 mL). Repeating the above procedure with the enriched (−)-enantiomer of 13 (4.57 g, 21,7 mmol) with (S)-(−)-2-chlocyphos

44 Synthesis and Pharmacological Evaluation of 8-OSO2CF3-2-aminotetralin Derivatives

(5.99 g, 21.7 mmol) gave (S)-(−)-13 (1.65 g, 15%) as a colorless oil with [α]D −77.6° 15 22 (lit [α] D −77.0° (c 1.03)). (R)-(+)-8-Hydroxy-2-( n-propylamino)tetralin HBr (( R)-(+)-15). (R)-13.HCl (1.74 g, 6.82 mmol) was refluxed in 48% aqueous HBr (50 mL, freshly distilled) for 2 h under N2-atmosphere. The reaction mixture was allowed to cool to room temperature and evaporated to dryness giving 1.88 g (97%) of a pale-brown solid, of which 445 mg was recrystallized from ethanol/ether for purification (374 mg, 78%): mp 283-286 °C; −1 1 IR (KBr) 3275 cm ; H NMR (CD3OD) δ 1.07 (t, J = 7.69, 3H), 1.70-1.91 (m, 3H), 2.33

(m, 1H), 2.60 (dd, J1 = 10.25, J2 = 16.23, 1H), 2.91 (m, 2H), 3.08 (m, 2H), 3.26-3.37 (m, 1H), 3.43-3.58 (m, 1H), 6.61 (d, J = 7.7, 1H) 6.62 (d, J = 8.12, 1H) 6.97 (dd, J1 = 7.69, J2 13 = 8.12, 1H); C NMR (CD3OD) δ 11.0, 20.7, 26.6, 27.2, 28.3, 47.5, 55.8, 112.6, 120.0, +1 120.3, 127.8, 136.9, 156.0; MS (CI with NH3) m/e 206 (M ); Anal Calcd (Obsd) for

C13H19NO.HBr: C: 54.55 (54.59), H: 7.04 (7.11), N: 4.89 (4.82); [α]D +63.5°. (S)-(−)-8-Hydroxy-2-( n-propylamino)tetralin HBr (( S)-(−)-15). Demethylation of (S)-13.HCl (1.92 g, 7.53 mmol) was performed according to procedure as described for (R)-15 as above giving (S)-15 in a quantitative yield. Part of the salt (1.06 g) was recrystallized from ethanol/ ether yielding 0.80 g (76%) of off- −1 1 white crystals: mp 273-277 °C; IR (KBr) 3275 cm ; H NMR (CD3OD) δ 1.07 (t, J =

7.69, 3H), 1.70-1.91 (m, 3H), 2.33 (m, 1H), 2.60 (dd, J1 = 10.25, J2 = 16.23, 1H), 2.91 (m, 2H), 3.08 (m, 2H), 3.26-3.37 (m, 1H), 3.43-3.58 (m, 1H), 6.61 (d, J = 7.7, 1H) 6.62 (d, J 13 δ = 8.12, 1H) 6.97 (dd, J1 = 7.69, J2 = 8.12, 1H); C NMR (CD3OD) 11.1, 20.7, 26.6,

27.2, 28.3, 47.5, 55.8, 112.6, 120.0, 120.3, 127.8, 136.9, 156.0; MS (CI with NH3) m/e +1 206 (M ); Anal Calcd (Obsd) for C13H19NO.HBr: C: 54.55 (54.46), H: 7.04 (7.03), N:

4.89 (4.98); [α]D −64.5°. (R)-(+)-8-[[(Trifluoromethyl)sulfonyl]oxy]-2-( n-propylamino)tetralin HCl ((R)-(+)-3). A mixture of (R)-15 (200 mg, 0.70 mmol), N- phenyltrifluoromethanesulfonimide (376 mg, 1.05 mmol) and tetrabutyl ammoniumhydrogensulfate (24 mg, 10 mol%) in dichloromethane (8 mL) and 10% NaOH (3 mL) was stirred at room temperature for 24 h. The reaction mixture was quenched with 5% HCl solution (v/v) until pH 1, diluted with H2O (25mL) and washed with ether (50 mL). The ether layer was extracted with H2O and 5% HCl solution (20 mL). The combined aquous layers were basified with solid Na2CO3 until pH 9, extracted with ether (3 × 30 mL) after which the organic phase was washed with brine and dried over Na2SO4. Evaporation in vacuo yielded a colorless oil which was converted to the HCl salt and recrystallized from methanol/ether (177 mg, 68%): mp 238-240 °C; IR −1 1 (KBr) 1217, 1421 cm (O-SO2); H NMR δ 0.96 (t, J = 7.5, 3H), 1.35 (br s, NH), 1.55 (m, 2H), 1.63 (m, 1H), 2.05 (m, 1H), 2.53 (dd, J1 = 8.55, J2 = 16.24, 1H), 2.69 (t, J = 7.5,

45 Chapter 2

2H), 2.83-3.04 (m, 3H), 3.12 (dd, J = 4.71, J = 16.24, 1H), 7.05-7.18 (m, 3H); 13C NMR 1 2 δ 11.8, 23.4, 27.8, 28.6, 30.7, 49.0, 52.5, 118.3, 118.6 (q, J = 321, CF3), 126.7, 128.6, +1 128.7, 139.9, 148.4; MS (CI with NH3) m/e 338 (M ); Anal Calcd (Obsd) for

C14H18NO3SF3.HCl: C: 44.98 (45.18), H: 5.12 (5.12), N: 3.75 (3.86); [α]D +61.5° (HCl). (S)-(−)-8-[[(Trifluoromethyl)sulfonyl]oxy]-2-( n-propylamino)tetralin HCl ((S)-(−)-3). Triflation of (S)-15 (880 mg, 3.08 mmol) was performed according to the procedure given for the synthesis of (R)-3 above giving an oil after extractive workup. Conversion to the HCl salt and subsequent recrystallization from methanol/ether gave −1 760 mg (66%) of a white crystals: mp 235-238 °C; IR (KBr) 1217, 1419 cm (O-SO2); 1H NMR δ 0.95 (t, J = 7.7, 3H), 1.32 (br s, NH), 1.54 (m, 2H), 1.64 (m, 1H), 2.06 (m, 1H),

2.53 (dd, J1 = 8.55, J2 = 16.24, 1H), 2.69 (t, J = 7.69, 2H), 2.82-3.02 (m, 3H), 3.12 (dd, J1 = 4.7, J = 16.24, 1H), 7.04-7.20 (m, 3H); 13C NMR δ 11.7, 23.4, 27.7, 28.6, 30.7, 49.0, 2 52.5, 118.3, 118.6 (q, J = 321, CF3), 126.7, 128.6, 128.7, 139.9, 148.4; MS (CI with NH3) +1 m/e 338 (M ); Anal Calcd (Obsd) for C14H18NO3SF3.HCl: C: 44.98 (44.94), H: 5.12

(5.14), N: 3.75 (3.74); [α]D −61.5° (HCl). (R)-1-(2-chlorophenyl)-2,2-dimethyl-1,3-propanediol (( R)-16). To a refluxed solution of LiAlH4 (4.0 g, 105.3 mmol) in dry THF (100 mL), (R)-2-chlocyphos (8.2 g, 29.7 mmol) was slowly added as a solid leading to a violent reaction. The mixture was subsequently refluxed for 5 h followed by overnight stirring at room temperature. The excess LiAlH4 was destroyed by the slow additon of aqueous 1N aqueous KOH (4.0 mL)

(caution: this yields the very poisonous PH3 gas). The resulting mixture was stirred with

Celite (8.8 g) for 30 min and filtered. The ether layer was dried over Na2SO4 and concentrated in vacuo, yielding 3.93 g (62%) of a colorless oil. IR (KBr) 3337 cm−1 (OH); 1H NMR δ 0.86 (s, 3H), 0.91 (s, 3H), 3.53 (m, 2H), 3.84 (br s, 2H), 5.25 (s, 1H), 13 δ 7.18-7.33 (m, 4H); 7.57 (d, J = 7.69, 1H); C NMR 18.7, 22.4, 40.0, 72.2, 76.7, 126.5, + 128.5, 129.2, 129.5, 133.3, 139.2; MS (EIPI) 196 (M −H2O). (R)-2H-2-oxo-4-( R)-(2-chlorophenyl)-5,5-dimethyl-1,3,2- dioxaphosphorinane (( 2R,4R)-17). A magnetically stirred solution of diol 16 (3.8 g,

14.6 mmol) in benzene (30 mL) was cooled to 0 °C under N2-atmosphere. Over a 15-min period, PCl3 (3.0 g, 17.4 mmol) was added carefully, while the solution was degassed regularly. After this addition, the solution was stirred at room temperature for 1 h. Subsequently, ethanol (2.5 mL) was added slowly to the mixture following stirring for another hour at room temperature. Evaporation of the solvent yielded a colorless oil, which was crystallized from ether to afford pure 17 (0.39 g, 10%) as a white solid. The mother liquor was evaporated to dryness and stored in the freezer (−18 °C) until further use. IR (KBr) 1036, 1252, 1268, 1585, 2931 cm−1; 1H NMR δ 0.85 (s, 3H), 1.16 (s, 3H), 4.00 ((dd, J = 26.06, J = 11.54, 1H), 4.30 (dd, J = 11.54, J = 3.42, 1H), 5.77 (d, J = 1 2 1 2

46 Synthesis and Pharmacological Evaluation of 8-OSO2CF3-2-aminotetralin Derivatives

13 δ 3.42, 1H), 7.06 (d, J = 671.74, 1H), 7.24-7.39 (m, 4H); 7.54-7.59 (m, 1H); C NMR 18.0, 20.7, 37.4, 72.3 (d, J = 6.38), 81.6 (d, J = 3.20), 126.8, 129.4, 130.0, 130.1, 132.9, 133.1; 31P NMR δ 3.61; MS (EIPI) 261 (M+). (±)-N-(8-methoxy-tetralin-2-yl)- O,O’-[1-( R)-2-chlorophenyl)-2,2-(dimethyl)- prop-1,3-yl]-( S)-phosponamide (( ±)-19). A suspension of 18 (120 g, 0.68 mmol) in

Et3N (0.2 mL) and ethanol (0.2 mL) was cooled to 0 °C and treated dropwise with a solution of phosporinane derivative (2R,4R)-17 (204 mg, 0.78 mmol) in CCl4 (135 µL) via a syringe. The resulting reaction mixture was stirred at room temperature for 2 h. The reaction was quenched by acidification with 10% aqueous HCl (2 mL), diluted with

H2O (10 mL) and extracted with CH2Cl2 (3 × 10 mL). The combined organic layers dried over Na2SO4 and removed in vacuo yielding 410 mg (>100%) of a sticky colorless oil. Attempts to separate the enantiomers of 19 via TLC, trying a variety of eluent combinations did not succeed: 1H NMR δ 0.83 (s, 3H), 1.07 (s, 3H), 1.71-1.83 (m, 1H), 2.11-2.16 (m,1H), 2.54 (dd, J = 17.21, J = 8.05, 1H) 2.92 (m, 2H), 3.20 (ddd, J = 32.22, 1 2 1 J = 17.21, J = 5.13, 1H), 3.39 (t, 1H) 3.79 (s, 3H) 3.83 (dd, J = 24.17, J = 11.35, 1H), 2 3 1 2 4.15-4.31 (m, 1H) 4.55 (d, J = 11.35, 1H), 6.03 (d, J = 1.47, 1H), 6.64-6.73 (m, 2H), 7.10 (t, 1H) 7.26-7.59 (m, 4H); 31P NMR δ 5.84 (d, ∆δ 0.096; ratio 67:36); MS (EIPI) 435 (M+). 8-Methoxy-1-methyl-2-tetralone (20). A mixture of 12 (7.1 g. 40.2 mmol), pyrrollidine (6.7 mL, 80.5 mmol) and a spatula of p-TsOH were refluxed in benzene (150 mL) under Dean-Stark conditions. After 18 h the volatiles were removed in vacuo giving a brown oil. The enamine was dissolved in dioxane (35 mL) and stirred at 40 °C together with iodomethane (12.0 mL, 187.8 mmol) for 3 h. The temperature was raised to 75 °C for 24 h after which an additional portion of iodomethane (4.0 mL, 62.5 mmol) was added and the heating continued for another 24 h. H2O (15 mL) and acetic acid (0.7 mL) were added and the reaction mixture was refluxed for 7 h. Evaporation in vacuo afforded a brown residue which was taken up in CHCl3 (100 mL), washed with 10% aqueous HCl (100 mL). The aqueous phase was extracted with CHCl3 (2 × 100 mL), and the combined organic layers were dried over MgSO4 and reduced to dryness. The obtained oil was purified on a silica column eluting with CH2Cl2, collecting 30 mL fractions. Pure fractions were pooled and evaporated in vacuo giving 7.2 g (94%) of a yellow oil: IR (KBr) 1716 cm−1 (C=O); 1H NMR δ 1.37 (d, J = 7.69, 3H), 2.37-2.54 (m, 1H), 2.69-2.82 (m, 1H), 2.88-3.01 (m, 1H), 3.09-3.25 (m, 1H), 3.82 (q, J = 7.69, 1H), 13 δ 3.84 (s, 3H), 6.80 (d, J = 8.12, 2H) 7.18 (d, J = 8.12, 1H); C NMR 18.2, 27.8, 38.1, 41.9, 55.3, 108.6, 120.2, 127.3, 127.6, 136.8, 156.8. cis- and trans-(±)-8-Methoxy-1-methyl-( n-propylamino)tetralin ( cis- and trans-21). A solution of 20 (3.2 g, 16.8 mmol), n-propylamine (3.0 mL, 36.5 mmol) and

47 Chapter 2 a spatula of p-TsOH in dry benzene (60 mL) was refluxed for 24 h in a Dean-Stark apparatus. At this time, an additonal amount of n-propylamine (1.5 mL) was added and the heating continued for 48 h. Removal of the solvent in vacuo gave a brown oil which was immediately dissolved in MeOH (100 mL), transferred to a Parr-apparatus and hydrogenated under H2-atmosphere (4 atm) using 10% Pd/C. After 1 h the reaction was complete (GC/ TLC). The mixture was filtered over Celite, rinsed with MeOH and evaporated to dryness yielding a brown oil. GC-MS analysis revealed a cis/ trans ratio of 90:10. The oil was chromatographed on a short silica column (6 × 6 cm), eluting with

CH2CL2/MeOH (20:1). Pure fractions were pooled and evaporated in vacuo, after which the combined intermediate fractions were subjected to another column affording 2.15 g (55%) in total of pure of cis-21 and 0.17 g (4%) of trans-21. A portion of the free bases were treated with ethereal HCl and recrystallized from EtOH/ether. cis-21HCl: mp 244- 11a −1 1 246 °C (lit 243-245 °C ); IR (KBr) 1258, 1584 cm ; H NMR (CD3OD) δ 1.06 (t, J = 7.69, 3H), 1.18 (d, J = 6.84, 3H), 1.71-1.89 (m, 2H), 1.97-2.16 (m, 2H), 2.92-3.16 (m,

4H), 3.37-3.49 (m, 1H), 3.56-3.66 (dq, J1 = 6.41, J2 = 5.13, 1H), 3.82 (s, 3H), 6.71 (d, J = 13 δ 7.69, 1H), 6.78 (d, J = 8.12, 1H), 7.13 (dd, J1 = 8.12, J2 = 7.69, 1H); C NMR (CD3OD) 11.0, 14.0, 20.5, 20.8, 28.7, 29.4, 48.1, 55.5, 59.3, 108.6, 120.3, 121.7, 128.3, 135.7, 157.8; MS (EIPI) 233 (M+). trans-21.HCl: mp 166-167 °C (lit 175-176 °C11a); IR (KBr) −1 1 1248, 1580 cm ; H NMR (CD3OD) δ 1.02 (t, J = 7.27, 3H), 1.30 (d, J = 7.26, 3H), 1.63- 1.83 (m, 2H), 2.03-2.31 (m, 2H), 2.80-2.93 (m, 2H), 2.96-3.14 (m, 2H), 3.42 (br q, J = 13 6.83, 1H), 3.85 (m, 1H), 3.85 (s, 3H), 6.80 (t, J = 8.98, 2H), 7.17 (t, J = 8.12, 1H); C

NMR (CD3OD) δ 11.0, 20.2, 20.3, 21.0, 23.8, 31.0, 48.2, 55.5, 59.6, 109.0, 122.2, 125.9, 128.3, 136.0, 158.6; MS (EIPI) 233 (M+). cis-(±)-8-Hydroxy-1-methyl-2-( n-propylamino)tetralin HBr ( cis-22). Demethylation of cis-21. HCl (200 mg, 0.74 mmol) was performed according to procedure as described for (R)-15 as above giving cis-22 as a pinkish solid in a quantitative yield. The salt was recrystallized from EtOH/ether yielding 197 mg (89%) −1 1 off-white crystals: mp 251-253 °C; IR (KBr) 3242 cm (OH); H NMR (CD3OD) δ 1.07 (t, J = 7.69, 3H), 1.22 (d, J = 6.84, 3H), 1.76-1.88 (m, 2H), 1.98-2.11 (m, 2H), 2.89-2.96

(m, 2H), 3.07-3.16 (m, 2H), 3.30-3.49 (m, 1H), 3.54-3.66 (dq, J1 = 6.41, J2 = 5.13, 1H), 13 δ 6.60 (t, J = 8.98, 2H) 6.96 (t, J = 8.98, 1H); C NMR (CD3OD) 11.1, 13.9, 20.5, 21.1,

28.8, 29.6, 48.1, 59.5, 113.0, 120.5, 126.7, 128.0, 135.8, 155.5; MS (CI with NH3) 220 +1 (M ); Anal Calcd (Obsd) for C14H21NO.HBr: C: 56.01 (55.74), H: 7.39 (7.32), N: 4.67 (4.50). cis-(±)-8-[[(Trifluoromethyl)sulfonyl]oxy]-1-methyl-2-( n- propylamino)tetralin HCl ( cis-9). Triflation of cis-22 (215 mg, 0.72 mmol) was performed according to the procedure given for the synthesis of (R)-3 using Triton-B

48 Synthesis and Pharmacological Evaluation of 8-OSO2CF3-2-aminotetralin Derivatives

(50 µL, 10 mol%) as the phase-transfer catalyst giving a colorless oil after extractive workup. Column chromatography on silica eluting with CH2Cl2/MeOH (10:1) afforded pure cis-9 (200 mg, 79%). Conversion to the HCl salt and subsequent recrystallization from acetone gave 171 mg (61%) of a white solid: mp 194-195 °C; IR (KBr) 1211, 1414 −1 1 cm (O-SO2); H NMR δ 0.95 (t, J = 7.32, 3H), 1.11 (d, J = 7.09, 3H), 1.29 (br s, NH), 1.45-1.69 (m, 2H), 1.73-1.83 (m, 2H), 2.65-2.71 (m, 2H), 2.86-2.96 (m, 3H), 3.28-3.40 13 (dq, J1 = 7.08, J2 = 4.88, 1H), 7.06-7.21 (m, 3H); C NMR δ 11.5, 13.4, 23.2, 23.5, 28.7,

30.4, 48.5, 55.6, 118.1, 118.4 (q, J = 320.5, CF3), 126.8, 128.8, 134.7, 138.9, 148.0; MS +1 (CI with NH3) 352 (M ). Resolution of cis-( ±)-8-Methoxy-1-methyl-2-( n-propylamino)tetralin. The resolution of cis-21 was performed according to the method of Arvidsson et al.,11a affording 63 mg (25%) of (+)-21. HCl and 67 mg (26%) of (−)-21. HCl. (1S,2R)-cis-(+)-

21: [α]D +29.6° (MeOH, c 1.08; lit. [α]D +30.4° (MeOH, c 1.02). (1R,2S)-cis-(−)-21: [α]D

−29.0° (MeOH; lit. [α]D −31.1° (MeOH, c 1.02). cis-(1S,2R)-(+)-8-Hydroxy-1-methyl-2-( n-propylamino)tetralin HBr ( cis- (1S,2R)-(+)-22). The title compound was prepared as described for the synthesis of (R)- 15 as above giving cis-(+)-22. The crude product was stirred in acetone and collected on a glass sintered funnel as a pinkish solid (51 mg, 85%): mp 255 °C (decomp); [α]D +27.9° (MeOH, c 0.95); E.e. 99%. cis-(1R,2S)-(−)-8-Hydroxy-1-methyl-2-( n-propylamino)tetralin HBr ( cis- (1R,2S)-(−)-22). The title compound was prepared as described for the synthesis of cis- (1S,2R)-22 as above giving cis-(−)-22 as a pinkish solid (60 mg, 80%): mp 254 °C +1 (decomp); MS (CI with NH3) 220 (M ); Anal Calcd (Obsd) for C14H21NO.HBr: C: 56.01

(55.71), H: 7.39 (7.41), N: 4.67 (4.57); [α]D −28.0° (MeOH, c 1.07). cis-(1S,2R)-(−)-8-[[(Trifluoromethyl)sulfonyl]oxy]-1-methyl-2-( n- propylamino)tetralin HCl ( cis-(1S,2R)-(−)-9). The title compound was prepared as described for the synthesis of (R)-3 using triton-B as the phase-transfer catalyst, affording cis-(1S,2R)-9 as a colorless oil (57 mg, 85%) after column chromatography. Conversion to the HCl salt and recrystallization from acetonitrile gave 41 mg (61%) colorless needles: mp 240-245 °C; Anal Calcd (Obsd) for C15H20NO3SF3.HCl: C: 46.45

(46.34), H: 5.46 (5.32), N: 3.61 (3.62); [α]D −3.7° (MeOH). cis-(1R,2S)-(+)-8-[[(Trifluoromethyl)sulfonyl]oxy]-1-methyl-2-( n- propylamino)tetralin HCl ( cis-(1R,2S)-(+)-9). The title compound was prepared as described for the synthesis of (R)-3 using triton-B as the phase-transfer catalyst, affording cis-(1R,2S)-9 as a colorless oil (55 mg, 78%) after column chromatography. Conversion to the HCl salt and recrystallization from i-propyl acetate gave 43 mg

49 Chapter 2

(55%) of a white yelly, which was difficult to filter: mp 240-245 °C; [α]D +3.5° (MeOH, c 1.15).

cis- and trans-(±)-8-Methoxy-1-methyl-2-(benzylamino)tetralin HCl ( cis- and trans-23). These compounds were made essentially according to literature procedures.21 A solution of 8-methoxy-1-methyl-2-tetralone (4.85 g, 25.5 mmol), benzylamine (4 mL) and p-TsOH. H2O (0.09 g) in benzene (70 mL) was refluxed for 72 hours under continuous removal of H2O using a Dean-Stark apparatus. The benzene and the excess benzylamine were removed in vacuo and the residue was dissolved in MeOH

(150 mL). After transferring the solution to a Parr hydrogenation flask, PtO2 (80 mg) was added as a catalyst and the mixture was hydrogenated for 2.5 hours under a H2- pressure of 3 atmosphere. After filtration over Celite and evaporation in vacuo, the brown residual oil (cis/ trans ratio 50:50 according to GC-MS) was subjected to column chromatography (SiO2) eluting with CH2Cl2/ MeOH (20:1) affording pure trans- (eluted 1st) and cis-23 (eluted 2nd). Both compounds were converted in the HCl salt and recrystallized from EtOH/ ether: cis-23 (0.91 g, 11%, white crystals): mp 239-240 °C; IR (KBr) 1253, 1582 cm−1; 1H NMR δ 1.17 (d, J = 6.96, 3H), 1.79-1.87 (m, 2H), 2.87-2.99

(m, 3H), 3.46 (dq, J1 = 6.69, J2 = 5.49, 1H), 3.86 (s, 3H), 3.91 (dd, J1 = 44.31, J2 = 12.81, 2H), 6.71 (t, J = 7.69, 2H), 7.11 (t, J = 7.69, 1H), 7.26-7.42 (m, 5H); 13C NMR δ 13.5, 24.1, 29.3, 29.9, 51.0, 55.2, 55.9, 107.2, 121.2, 126.3, 126.9, 128.2, 128.4, 130.7, 136.7, + 140.7, 157.3; MS (EI) M 281; Anal Calcd (Obsd) for C19H23NO.HCl: C: 71.80 (71.55), H: 7.61 (7.50), N: 4.41 (4.37). trans-23 (0.81 g, 10%; white crystals): mp 223-224 °C; IR (KBr) 1254, 1583 cm−1; 1H NMR δ 1.23 (d, J = 6.96, 3H), 1.90-2.01 (m, 2H), 2.69-2.76

(ddd, J1 = 16.84, J2 = 5.49, J3 = 2.56, 1H), 2.90-3.00 (m, 2H), 3.22 (br q, J = 6.96, 1H), 3.84 (s, 3H), 3.98 (dd, J1 = 21.97, J2 = 13.55, 2H), 6.70 (d, J = 8.06, 1H), 6.75 (d, J = 7.32, 1H), 7.11 (t, J = 8.06, 1H), 7.26-7.37 (m, 5H); 13C NMR δ 20.9, 21.7, 24.2, 32.8, 32.9, 51.1, 55.2, 56.4, 107.4, 121.3, 126.0, 126.8, 128.2, 128.3, 128.8, 136.8, 140.8, + 158.0; MS (EI) M 281; Anal Calcd (Obsd) for C19H23NO.HCl: C: 71.80 (71.57), H: 7.61 (7.53), N: 4.41 (4.33). cis-8-Methoxy-1-methyl-2-( N-methyl-benzylamino)tetralin HCl ( cis-24). Compound cis-23 (1.01 g, 3.18 mmol) was dissolved in acetonitrile (15 mL), then an aquaous solution of formaldehyde (2.5 mL, 37 %) was added. Subsequently, NaCNBH3 (608 mg, 9.68 mmol) and glacial acetic acid (340 µL, pH 5) were added. The resulting mixture was stirred under N2-atmosphere for 1 h, after which time another amount of acetic acid (340 µL) was added. Stirring continued for 1 hour. The reaction mixture was taken up in 10% NaOH (30 mL) and extracted with ether (3 × 20 mL). The organic layers were dried over MgSO4, filtered and the solvent was removed in vacuo giving

50 Synthesis and Pharmacological Evaluation of 8-OSO2CF3-2-aminotetralin Derivatives

0.92 g (98 %) of a colorless oil. The HCl salt was recrystallized from EtOH/ether giving 0.67 g (64%) of a white powder: mp 207-208 °C; IR (KBr) 1252, 1584 cm−1; 1H NMR

(CD3OD) δ 1.30/1.37 (d; J = 6.59/6.59, 3H), 2.12-2.64 (m, 2H), 2.88/2.90 (s, 3H), 2.94- 3.11 (m, 2H), 3.41-3.63 (m, 1H), 3.73-3.95 (m, 1H), 3.83/3.85 (two s, 3H), 4.54/4.58

(AB, J1 = 108.17/74.47, J2 = 13.19/13.19, 2H), 6.71-6.82 (m, 2H), 7.11-7.20 (m, 1H),

7.49-7.63 (m, 5H); Anal Calcd (Obsd) for C20H25NO.HCl. 0.4H2O: C: 70.84 (70.42), H: 7.97 (7.88), N: 4.13 (4.35). trans-8-Methoxy-1-methyl-2-( N-methyl-benzylamino)tetralin HCl ( trans- 24). The title compound was prepared according to the procedure as described for compound cis-24, starting from trans-23 (1.01 g, 3.18 mmol). After extractive work-up and removal of the solvent a colorless oil was obtained (0.89 g, 95%). Conversion to the HCl salt and recrystallization from EtOH/ether gave 0.70 g (66%) of a white powder: −1 1 mp 192-193 °C; IR (KBr) 1254, 1584 cm ; H NMR (CD3OD) δ 1.21/1.30 (d, J = 6.83/7.08, 3H), 1.84-2.07 (m, 1H), 2.33-2.52 (m, 1H), 2.71 (s, 3H), 2.64-2.90 (m, 2H),

3.57-3.80 (m, 2H), 3.84/3.87 (s, 3H), 4.25/4.41 (AB, J1 = 30.03/26.37, J2 = 12.94/13.19, 2H), 6.67-6.89 (m, 2H), 7.11-7.22 (m, 1H), 7.47-7.63 (m, 5H); Anal Calcd (Obsd) for

C20H25NO.HCl. 0.1H2O: C: 71.99 (71.69), H: 7.91 (7.85), N: 4.20 (4.22). cis-8-Methoxy-1-methyl-2-(methylamino)tetralin HCl ( cis-25). Cis-24 (0.60 g, 1.80 mmol) was dissolved in abs EtOH (50 mL), then 10% Pd/C (0.4 g) was added and the solution was hydrogenated under a H2-pressure of 3 atmosphere in a Parr apparatus for 1.5 h at ambient temperature. The catalyst was filtered off (celite) and the solvent was evaporated in vacuo yielding 0.35 g of a white solid. Crystallization from EtOH/ ether gave 0.33 g (76 %) off-white crystals: mp 222-223 °C; IR (KBr) 1256, 1583 cm−1; 1H NMR δ 1.17 (d, J = 6.84, 3H), 1.87-2.20 (m, 2H), 2.80 (s, 3H), 2.91-2.99 (m, 2H),

3.30-3.40 (m, 1H), 3.60 (dq, J1 = 6.59, J2 = 5.37, 1H), 3.83 (s, 3H), 6.71 (d, J = 7.81, 1H), 13 δ 6.74 (d, J = 8.06, 1H), 7.13 (t, J1 = 8.06, J2 = 7.81, 1H); C NMR (CD3OD) 12.5, 19.1, 27.3, 28.1, 29.9, 54.2, 59.1, 107.4, 120.5, 126.9, 127.0, 134.5, 156.6; MS (EI) M+ ; Anal

Calcd (Obsd) for C13H19NO.HCl. 0.1H2O: C: 64.11 (63.99), H: 8.36 (8.40), N: 5.75 (5.76). trans-8-Methoxy-1-methyl-2-(methylamino)tetralin HCl ( trans-25). The title compound was prepared as was described for the synthesis of cis-25, starting from trans-24 (0.60 g, 1.80 mmol). This procedure gave 0.39 g of a white solid which was crystallized from EtOH/ether yielding 0.30 g (69 %) of white crystals. mp 252-253 °C; IR (KBr) 1247, 1580 cm−1; 1H NMR δ 1.29 (d, J = 6.84, 3H), 2.06-2.35 (m, 2H), 2.72 (s, 3H), 2.79-2.94 (m, 2H), 3.40 (q, 1H, obscured), 3.84 (s, 3H), 6.77 (d, J = 8.06, 1H), 6.82 13 (d, 1H), 7.17 (t, J1 = 8.06, J2 = 7.81, 1H); C NMR (CD3OD) δ 18.3, 19.3, 22.1, 29.5, 30.0, 54.3, 59.3, 107.8, 121.1, 124.5, 127.1, 134.5, 157.5; MS (EI) M+ ; Anal Calcd

(Obsd) for C13H19NO.HCl.: C: 64.59 (64.29), H: 8.34 (8.56), N: 5.79 (5.78).

51 Chapter 2

cis-8-Hydroxy-1-methyl-2-(methylamino)tetralin HBr ( cis-26). Demethylation of cis-25 (200 mg, 0.83 mmol) was performed according to procedure as described for (R)-15 as above giving the title compound as a brown solid, which was dissolved in hot MeOH and treated with activated charcoal. After filtration using Celite, the salt was recrystallized from MeOH/ether yielding 179 mg (79%) off-white crystals: −1 1 mp 248-250 °C; IR (KBr) 3323 cm (OH); H NMR (CD3OD) δ 1.11 (d, J = 6.96, 3H), 1.84-2.04 (m, 2H), 2.72 (s, 3H), 2.78-2.86 (m, 2H), 3.16-3.30 (m, 1H), 3.46-3.51 (m, 1H), 13 δ 6.46-6.53 (m, 2H), 6.87 (t, J = 7.69, 1H); C NMR (CD3OD) 10.9, 18.0, 26.1, 26.9, 28.6, 57.9, 110.5, 117.8, 123.9, 125.4, 133.1, 152.9; MS (EI) M+ ; Anal Calcd (Obsd) for

C12H17NO.HBr. 0.2H2O: C: 52.26 (52.25), H: 6.73 (6.77), N: 5.08 (5.05). trans-8-Hydroxy-1-methyl-2-(methylamino)tetralin HBr ( trans-26). Demethylation of trans-25. HCl (237 mg, 0.98 mmol) was performed according to procedure as described for (R)-15 as above giving trans-26 as a brownish solid, which was dissolved in hot MeOH and treated with activated charcoal. After filtration using Celite, the salt was recrystallized from MeOH/ ether yielding 155 mg (58%) off-white −1 1 crystals: mp 218-219 °C; IR (KBr) 3299 cm (OH); H NMR (CD3OD) δ 1.24 (d, J = 6.96, 3H), 1.98-2.10 (m, 1H), 2.11-2.21 (m, 1H), 2.64 (s, 3H), 2.66-2.83 (m, 2H), 3.27- 13 δ 3.38 (m, 2H), 6.56 (d, J = 7.69, 2H), 6.91 (t, J = 7.69, 1H); C NMR (CD3OD) 17.2, 17.7, 20.8, 28.3, 28.8, 58.1, 110.7, 118.3, 121.5, 125.4, 132.9, 153.8; MS (EI) M+ ; Anal

Calcd (Obsd) for C12H17NO.HBr: C: 52.95 (53.21), H: 6.67 (6.79), N: 5.15 (5.18). cis-8-[[(Trifluoromethyl)sulfonyl]oxy]-1-methyl-2-(methylamino)tetralin HCl ( cis-10). Triflation of cis-26 (104 mg, 0.38 mmol) was performed according to procedure as described for (R)-3, using triton-B (20 µL, 10 mol%) as the phase-transfer catalyst. After extractive work-up the title compound was obtained in 107 mg (87%): −1 1 mp 226-227 °C; IR (KBr) 1206, 1418 cm (O-SO2); H NMR δ 1.12 (d, J = 6.83, 3H), 1.69-2.03 (m, 2H), 2.52 (s, 3H), 2.7 (br s, NH), 2.83-3.03 (m, 3H), 3.42 (dq, J1 = 6.83, J2 = 4.73, 1H), 6.96-7.21 (m, 3H); 13C NMR δ 13.5, 22.6, 28.5, 30.0, 33.0, 57.6, 118.3, 118.4 + (q, J = 320, CF3), 126.9, 128.8, 134.2, 138.6, 147.8; MS (EI) M 323; Anal Calcd (Obsd) for C13H16NO3SF3.HCl: C: 43.40 (43.27), H: 4.76 (4.61), N: 3.89 (3.87). trans-8-[[(Trifluoromethyl)sulfonyl]oxy]-1-methyl-2-(methylamino)tetralin HCl ( trans-10). Triflation of trans-26 (124 mg, 0.46 mmol) was performed according to procedure as described for cis-10, giving the title compound in 100 mg (67%) yield: mp −1 1 212-213 °C; IR (KBr) 1208, 1397 cm (O-SO2); H NMR δ 1.25 (d, J = 7.32, 3H), 1.93- 2.04 (m, 2H), 2.16 (br s, NH), 2.48 (s, 3H), 2.69-2.78 (ddd, J1 = 17.58, J2 = 5.86, J3 = 13 2.56, 1H), 2.87-3.00 (m, 2H), 3.17 (dq, J1 = 7.33, J2 = 1.09, 1H), 7.06-7.26 (m, 3H); C NMR δ 20.8, 20.9, 23.8, 32.4, 33.6, 58.7, 118.5 (q, J = 320, CF3), 118.6, 126.8, 129.3, 132.7, 138.8, 148.8; MS (EI) M+ 323.

52 Synthesis and Pharmacological Evaluation of 8-OSO2CF3-2-aminotetralin Derivatives

Pharmacology. The behavioural pharmacology, biochemistry experiments and pharmacokinetic experiments for compound 1, 2, 3, (R)-3 and (S)-3 were performed according to ref 26. Animals. Male Wistar rats weighing 200-250 g were used for the gross behaviour and the motility experiments. The rats were housed eight per cage with free access to food and water. The experiments were performed between 10:00 and 16:00 h. Lights were on from 7:30−18:30 h. Materials . All substances to be tested were dissolved in a physiological saline/solutol (90/10 v/v) solution with moderate heating in order to obtain complete dissolution. Reserpine was dissolved in a few drops of glacial acetic acid, made up to volume with 5.5% glucose (w/v) and neutralized before use. Receptor Binding. Compound 1, 2, 3, (R)-3 and (S)-3 were tested at Upjohn (Kalamazoo, MI). For experimental details see Chapter 4, section 4.5, Method A; Compound cis-9, cis-(1S,2R)-9, cis-(1R,2S)-9, cis-10 and trans-10 were screened at Centre Recherche Pierre Fabre, Castres, France (Chapter 5, section 5.5). Gross Behavioural Observations. The 5-HT behavioural syndrom (flat body posture, reciprocal forepaw treading, straub tail, hindlimb abduction) and the lower lip retraction (LLR) were scored between zero and 30 min after drug-treatment, prior to the motility test. The test compounds were given subcutaneously in the neck-region or orally via gavage. The animals that were treated orally were fasted for 18 h before the experiments. Reserpinized animals received reserpine (5mg/kg, sc) 18 h prior drug- treatment. Locomotor Activity. 30 Min after drug-treatment (described as above) the rats were placed in the test cages (1 rat/cage) on the motility meters (Automex II locomotor boxes, Columbus Instruments, Columbus, Ohio). Motor activity was recorded for 30 min. Hypothermia. The core temperature was determined by insertion of a digital temperature probe (CMA 150 Temperature Controller, Microdialysis AB, Stockholm, Sweden) into the rectum for 30 sec (n = 4). In all studies, the basal values were determined immediately after removal of animals from their home-cage. The time course experiment was stopped beyond the maximal effect. The ∆T (°C) that were obtained at each time-point in every rat were fitted via polynomial regression after which the AUC was estimated from the beginning of the experiment until the maximal effect. Statistics. Differences between the saline- and drug-treated group in the locomotor activity test were analyzed with one-way ANOVA followed by a Bonferroni t-test. The differences between the control tempertures and the treated-group

53 Chapter 2 temperatures were analyzed by one-way ANOVA with repeated measures followed by Tukey’s protected t-test.

Acknowledgments. We are grateful to Josanne Schellekens and Ulrike Selditz for the e.e. determinations. Dr. Peter de Boer is gratefully acknowledged for his help with the behavioural pharmacology and hypothermia experiments. We thank Wouter Brink for preparing the cis- and trans-10 series. Dr. Ron Hulst is acknowledged for his helpful discussions concerning the phoshamidate chemistry and we thank Dr. Wim Kruizinga (Department of Organic and Molecular Inorganic Chemistry, Groningen, The Netherlands) for performing the NMR heating experiments.

54 Synthesis and Pharmacological Evaluation of 8-OSO2CF3-2-aminotetralin Derivatives

2.6 References

[1] Arvidsson, L.-E.; Hacksell, U.; Nilsson, J.L.G.; Hjorth, S.; Carlsson, A.; Lindberg, P.; Sanchez, D.; Wikström, H..J. Med. Chem. 1981, 24, 921. [2] Middlemiss, D.N.; Fozard, J.R.Eur. J. Pharmacol. 1983, 90, 151. [3] Arvidsson, L.-E.; Hacksell, U.; Johansson, A.; Nilsson, J.L.G.; Lindberg, P.; Sanchez, D.; Wikström, H.; Svensson, K.; Hjorth, S.; Carlsson, A..J. Med. Chem. 1984, 27, 45. [4] Stjernlöf, P.; Gullme, M.; Elebring, T.; Andersson, B.; Wikström, H.; Lagerquist, S.; Svensson, K.; Ekman, A.; Carlsson, A.; Sundell, S.J. Med. Chem. 1993, 36, 2059. [5] Streitwieser, A.; Dafforn, A.Tetrahedron Lett. 1976, 18, 1435. [6] Sonesson, C.; Boije, M.; Romero, A.; Stjernlöf, P.; Andersson, B.; Hansson, L.; Waters, N.; Svensson, K.; Carlsson, A.; Wikström, H.Pat. Appl. WO 92-18475. [7] (a) Liu, Y.; Svensson, B.E.; Yu, H.; Cortizo, L.; Ross, S.B.; Lewander, T. Hacksell, U.Bioorg. Med. Chem. Letters 1991, 1, 257. (b) Liu, Y.; Yu, H.; Svensson, B.E.; Cortizo, L. Lewander, T.; Hacksell, U.J. Med. Chem. 1993, 36, 4221. [8] Sonesson, C.; Boije, M.; Svensson, K.; Ekman, A.; Carlsson, A.; Romero, A.G.; Martin, I.J.; Duncan, J.N.; King, L.J.; Wikström, H.J. Med. Chem. 1993, 36, 3409. [9] Mellin, C.; Vallgårda, J.; Nelson, D.L.; Björk, L.; Yu, H.; Andén, N.-E.; Csöregh, I.; Arvidsson, L.-E.; Hacksell, U. J. Med. Chem. 1991, 4, 497. [10] Haadsma-Svensson, S.R.; Smith, M.W.; Lin, C.-H.; Duncan, J.N.; Sonesson, C.; Wikström, H.; Carlsson, A.; Svensson, K. Bioorg. Med. Chem. Letters 1994, 4, 689. [11] (a) Arvidsson, L.-E.; Johansson, A.M.; Hacksell, U.; Nilsson, J.L.G.; Svensson, K.; Hjorth, S.; Magnusson, T.; Carlsson, A.; Andersson, B.; Wikström, H.J. Med. Chem. 1987, 30, 2105. (b) Hjorth, S.; Sharp, T.; Liu, Y. Naunyn-Schmiedeberg’s Arch. Pharmacol. 1990, 341, 149. [12] Peroutka, S.J.; McCarthy, B.G.Eur. J. Pharmacol. 1989, 163, 133. [13] Cornforth, J.W.; Cornforth, R.H.; Robinson, R.J. Chem. Soc. 1942, 689. [14] Ten Hoeve, W.; Wynberg, H.J. Org. Chem. 1985, 50, 4508. [15] Arvidsson, L.-E.; Hacksell, U.; Johansson, A.; Nilsson, J.L.G.; Lindberg, P.; Sanchez, D.; Wikström,H.; Svensson, K.; Hjorth, S.; Carlsson, A.J. Med. Chem. 1984, 27, 45. [16] Hendrickson, J.B.; Bergeron, R.Tetrahedron Letters 1973, 4607. [17] Hulst, R.; Zijlstra, R.W.J.; De Vries, N.K.; Feringa, B.L.Tetrahedron: Assymmetry 1994, 5, 1701. [18] (a) Zwierzak, A.; Brylikowska-Piotrowicz, J.Angew. Chem. Int. Ed. Eng. 1977, 16, 107; (b) Zwierzak, A. Synthesis, 1984, 332. [19] Arbuzov, B.A. Pure Appl. Chem. 1964, 44, 3035. [20] Atherton, F.R.; Openshaw, H.T.; Todd, A.R.J. Chem. Soc. 1945, 660. [21] (a) McDermed, J.D.; McKenzie, G.M.; Phillips, A.P.J. Med. Chem. 1975, 18, 362. (b) Horn, A.S.; Grol, C.J.; Dijkstra, D. Mulder, A.H.J. Med. Chem. 1978, 21, 825. [22] Arvidsson, L.-E.; Karlén, A.; Norinder, U.; Kenne, L.; Sundell, S.; Hacksell, U.J. Med. Chem. 1988, 31, 212. [23] Tricklebank, M.D. Trends. Pharmacol. Sci. 1985, 6, 403. [24] Berendsen, H.H.; Broekkamp, C.L.; Van Delft, A.M.Eur. J. Pharmacol. 1990, 187, 97. [25] Millan, M.J.; Rivet, J.-M.; Canton, H.; Le Marouille-Girardon, S.; Gobert, A.J. Pharmacol. Exp. Ther. 1992, 264, 1364. [26] Sonesson, C.; Barf, T.; Nilsson, J.; Dijkstra, D.; Carlsson, A.; Svensson, K.; Smith, M.W.; Martin, I.J.; Duncan, J.N.; King, L.J.; Wikström H.J. Med. Chem. 1995, 38, 1319.

55 Chapter 3

* Potentential Anxiolytic Properties of (R)-8-OSO2CF3-PAT

Abstract

The anxiolytic property of (R)-8-OSO2CF3-PAT ((R)-8-

[[(trifluoromethyl)sulfonyl]oxy]-2-(n-propyl-amino)tetralin), a 5-HT1A receptor agonist, was evaluated in Wistar rats by means of animal models of anxiety, the conditioned defensive burying model and the conditioned stress-induced freezing response followed by the elevated plus-maze test, respectively. In addition, the 5- HIAA/5-HT ratio (5-hydroxyindole acetic acid/5-hydroxytryptamine) of rat brain homogenates was studied. Acute drug administration resulted in abolition of the burying behaviour (3 mg/kg, ip), a dose-dependent decrease of rearing and induction of hyperphagia. (R)-8-OSO2CF3-PAT had no effect on conditioned footshock-induced freezing behaviour but increased open-arm activity in the rats on the plus-maze. The 5- HIAA/5-HT ratio was decreased in the lateral septum (1 and 3 mg/kg), dorsal hippocampus (3 mg/kg) and somatosensory cortex (3 mg/kg), implying that (R)-8-

OSO2CF3-PAT affects particularly the limbic system in anxiety-inducing situations.

3.1 Introduction

A number of studies have shown that drugs that reduce serotonin (5-hydroxy- tryptamine; 5-HT) function produce anxiolytic-like effects. Acute administration of 5-

HT1A receptor ligands such as 8-hydroxy-2-(di-n-propylamino)tetralin (8-OH-DPAT, Figure 3.1) are thought to reduce serotonergic function by acting as agonists at somatodendritic autoreceptors, which inhibit the firing of 5-HT neurones in raphe 1 nuclei. It is of importance that direct infusion of these 5-HT1A receptor agonists into the dorsal raphe gives reproducible anxiolytic effects, which supports this hypothesis.2 Previous investigations revealed that (R)-8-[[(trifluoromethyl)sulfonyl]oxy]-2-(n- propyl-amino)tetralin ((R)-8-OSO2CF3-PAT, Figure 3.1), an analogue of 8-OH-DPAT, is

* This chapter is based on: Barf, T.; Korte, S.M.; Korte-Bouws, G.; Sonesson, C.; Damsma, G.; Bohus, B.; Wikström, H. Eur. J. Pharmacol. 1996, 297, 205.

53 Chapter 3 a chemically and biologically stable compound possessing the pharmacological profile 3 of a 5-HT1A receptor agonist (see Chapter 2).

OH OSO2CF3 H N N

8-OH-DPAT R-(+)-8-OSO2CF3-PAT

Figure 3.1. Chemical structures of 8-hydroxy-2-(di-n-propylamino)tetralin and 8- [[(trifluoromethyl)sulfonyl]oxy]-2-(n-propylamino)tetralin.

The behavioural pharmacology of the drug has not been explored as yet.

Therefore, the effects of (R)-8-OSO2CF3-PAT were studied in different animal models of anxiety. In the elevated plus-maze test, rats restrict their activity to the enclosed areas, avoiding the two open arms.4 This behaviour can be reversed and enhanced by anxiolytic and anxiogenic drugs, respectively. The plus-maze model is relatively insensitive to drugs other than and anxiogenics.5 Prior exposure to an emotional stressor produces higher anxiety in the animals, which is reflected by reduced exploration of the open maze arms in favor of the enclosed maze arms.6 A second animal model of anxiety is the 'conditioned defensive burying test'.7 In this test, rats are exposed to an electrified shock probe, and the duration of burying behaviour is the major index of anxiety. Standard antianxiety agents suppress the burying response in a dose-related manner. It has been suggested that acute administration of partial agonists such as buspirone,8 ipsapirone9 and gepirone10 decreased 5-HT turnover or lowered levels of 5- hydroxy-indole acetic acid (5-HIAA), the major metabolite of 5-HT, which may be a feature of particular brain regions. The 5-HIAA/5-HT ratios are indicative of changes in 11 the 5-HT turnover rate. Due to the expected 5-HT1A receptor agonist property of (R)-8-

OSO2CF3-PAT, it was of interest to examine the 5-HIAA/5-HT ratios in several brain regions after administration of this compound. In the present study, the ability of (R)-8-

OSO2CF3-PAT to produce anxiolytic effects after acute administration was evaluated in these animal models.

3.2 Results

Conditioned Defensive Burying. Figure 3.2A shows that (R)-8-OSO2CF3-PAT dose dependently reduced the burying behaviour until complete abolition at the 3-

54 Potential Anxiolytic Properties of (R)-8-OSO2CF3-PAT mg/kg dose (F(2,18) = 3.46; P ≤ 0.05). The time spent rearing was also significantly decreased, both at the 1-mg/kg (F(2,18) = 10.95; P ≤ 0.05) and 3-mg/kg dose (P ≤ 0.01), as depicted in Figure 3.2B. Interestingly, food intake was dramatically increased at the doses applied (Figure 3.2C).

Conditioned Defensive Burying

20 20 70 A B * C 60

15 15 50 *

40 10 10 * 30 % Eating % Burying % Rearing 20 5 5

** 10

0 * 0 0 Vehicle 1 mg/kg 3 mg/kg Vehicle 1 mg/kg 3 mg/kg Vehicle 1 mg/kg 3 mg/kg

Figure 3.2. Percent time spent on different types of behaviour during 10-min conditioned defensive burying test,

30 min after intraperitoneal (ip) administered vehicle (n = 7) or (R)-8-OSO2CF3-PAT (1 mg/kg, n = 6; 3 mg/kg, n = 6). The different types of behaviour: defensive burying (A); rearing (B); eating (C). Data are expressed as means ± S.E.M. *P ≤ 0.05; **P ≤ 0.01, significantly different from control.

Conditioned Fear of Footshock. The percentage of time spent immobile in the inescapable footshock compartment was not significantly affected either by the 1- mg/kg or by the 3-mg/kg dose (Figure 3.3).

55 Chapter 3

Footshock

control stressed 100 90 80 70 60 50 40

% Immobility 30 20 10 0 Vehicle Vehicle 1 mg/kg 3 mg/kg

Figure 3.3. Percent time spent immobile during exposure to the former footshock compartment, 30 min after ip administered vehicle (control, n = 6; stressed, n = 8) or (R)-8-OSO2CF3-PAT (1 mg/kg, n = 7; 3 mg/kg, n = 7). Data are expressed as means± S.E.M. *P ≤ 0.05, significantly different from control.

Elevated Plus-maze. The percentage time L/L+D (L = time spent in open arms; D = time spent in enclosed arms) showed a non-significant trend towards increased open-arm activity after increasing of the dose (Figure 3.4A). One-way ANOVA revealed a significant effect of (R)-8-OSO2CF3-PAT on the number of open-arm entries at 3 mg/kg (F(2,21) = 15.27; P ≤ 0.01), as shown in Figure 3.4B. In addition, the numbers of open-arm entries for the non-stressed and stressed animals were significantly different (P ≤ 0.05; t-test). Figure 3.4C shows that the number of enclosed- arm entries was not significantly changed at the doses applied.

56 Potential Anxiolytic Properties of (R)-8-OSO2CF3-PAT

Plus-maze

60 20 20 control stressed A control stressed B control stressed C 50 15 15 40 * 30 10 10

20

Entries Open 5 5 % Time L/L+D * Entries Closed 10

0 0 0 Vehicle Vehicle 1 mg/kg 3 mg/kg Vehicle Vehicle 1 mg/kg 3 mg/kg Vehicle Vehicle 1 mg/kg 3 mg/kg

Figure 3.4. The effect of (R)-8-OSO2CF3-PAT (1 and 3 mg/kg, ip) on the (A) percent time spent in the open arms relative to cumulative time in all four arms; (B) number of open-arm entries; (C) number of closed-arm entries in rats given a 5-min test in the elevated plus-maze, directly after the rats' exposure to the conditioned emotional stressor. For further explanations see Fig. 3.2.

Effect on the 5-HIAA/5-HT Ratio in Various Rat Brain Regions.

Administration of 3 mg/kg of (R)-8-OSO2CF3-PAT ip reduced the 5-HIAA/5-HT ratio significantly in the somatosensory cortex (F(2,19) = 5.11; P ≤ 0.05), dorsal hippocampus (F(2,26) = 3.99; P ≤ 0.05) and lateral septum (F(2,25) = 4.91; P ≤ 0.05). In the case of the lateral septum, the maximum effect was achieved at a dose of 1 mg/kg (P ≤ 0.05). The 5-HIAA/5-HT ratio of the paraventricular nucleus of the hypothalamus and the ventral median hypothalamus showed a trend towards a decrease, whereas the ratios in other brain regions were not altered (Tabel 3.1).

3.3 Discussion

The treatment with (R)-8-OSO2CF3-PAT led to an increase in time spent in the open-arm area of the plus-maze and decreased defensive burying behaviour. Therefore, it is suggested that this new compound possesses anxiolytic properties. The same dose range of this drug induced a reduction of the 5-HIAA/5-HT ratio in several parts of the limbic system.

57 Chapter 3

Table 3.1. The Effect of (R)-8-OSO2CF3-PAT on the 5-HIAA/5-HT Ratio in Rat Brain.

Control (R)-8-OSO2CF3-PAT (R)-8-OSO2CF3-PAT Brain area (1 mg/kg) (3 mg/kg) MRN 1.809 ± 0.182 2.260 ± 0.384 2.444 ± 0.504 DRN 1.155 ± 0.074 1.338 ± 0.127 0.979 ± 0.052 SC 1.701 ± 0.191 1.345 ± 0.122 1.009 ± 0.032* PC 1.834 ± 0.203 1.935 ± 0.233 1.386 ± 0.183 DH 1.487 ± 0.166 1.249 ± 0.064 1.061 ± 0.050* LS 1.101 ± 0.099 0.755 ± 0.091* 0.784 ± 0.070* MS 1.081 ± 0.124 0.893 ± 0.137 0.977 ± 0.130 Striatum 1.188 ± 0.083 1.052 ± 0.124 1.124 ± 0.087 PVN 0.902 ± 0.086 0.916 ± 0.208 0.722 ± 0.117 VMH 1.341 ± 0.116 1.278 ± 0.114 1.051 ± 0.130 CEA 0.929 ± 0.128 0.828 ± 0.126 0.827 ± 0.091 NAc 1.047 ± 0.060 1.417 ± 0.344 1.367 ± 0.245

The ratio of 5-HIAA/5-HT in median raphe nucleus (MRN), dorsal raphe nucleus (DRN), somatosensory cortex (SC), prefrontal cortex (PC), dorsal hippocampus (DH), lateral septum (LS), medial septum (MS), striatum, paraventricular nucleus of the hypothalamus (PVN), ventral median hypothalamus (VMH), central amygdala (CEA) and accumbens (NAc), 30

min after ip administration of saline (n= 8 or 9) or (R)-8-OSO2CF3-PAT (1 mg/kg, n = 5 - 9; 3 mg/kg, n = 6 - 9). Data are expressed as means± S.E.M. *P ≤ 0.05, significantly different from control.

Together, these behavioural animal models give a particularly strong indication of the anxiolytic properties of a drug. In the burying model the major index of anxiety is expressed as an active coping behaviour (burying),12 whereas in the plus-maze test reduced exploration of the open arms is the major index of anxiety.4, 13 Decreases in defensive burying behaviour are evoked by classical benzodiazepine anxiolytics14,15 12 15 and 5-HT1A receptor agonists (e.g. ipsapirone, buspirone) and are often interpreted as anxiolytic actions of these two classes of drugs. In the burying test, (R)-8-OSO2CF3- PAT produced a dose-dependent decrease in burying behaviour and similar results were obtained for rearing activity. Interestingly, (R)-8-OSO2CF3-PAT at the doses applied induced hyperphagia during the conditioned defensive burying experiment. It is known that agonistic action at somatodendritic autoreceptors of e.g. 8-OH-DPAT, buspirone and gepirone reduces the synthesis and release of brain serotonin and thereby enhances food intake in freely feeding rats.16 In addition, there is evidence that 5-HT levels in the hypothalamus mediate the regulation of food intake via 5-HT1B receptors. Direct infusion of 5-HT1B receptor agonists into the paraventricular nucleus induces

58 Potential Anxiolytic Properties of (R)-8-OSO2CF3-PAT hypophagia.16f,17 This suggests that lowering of the 5-HT level in the paraventricular nucleus may cause hyperphagia. However, this hypothesis cannot be supported or contradicted by our findings, since systemically administered (R)-8-OSO2CF3-PAT failed to significantly reduce the 5-HIAA/5-HT ratio in the paraventricular nucleus of the hypothalamus. Contradictory results have been reported for the effects of acutely administered 18 partial or full 5-HT1A receptor agonists in the elevated plus-maze model. Some of these ligands were found to be anxiolytic, whereas other workers classed the same compounds as anxiogenic. Changes in experimental conditions may have dramatic effects, e.g., increasing the light intensity from 170 to 785 Lux causes inversion of the previously found anxiogenic effect. In our elevated plus-maze model, the anxious behaviour was enhanced by prior exposure to a conditioned stressor, i.e. re-exposure to a compartment associated with an inescapable, uncontrollable stressor.6 This was manisfested as diminished exploration of the open arms of the plus-maze, as compared to the non-stressed controls. Acute treatment with 3 mg/kg of (R)-8-OSO2CF3-PAT not only prevented this effect but even resulted in enhanced exploration behaviour in the open arms, as compared to the non- stressed control group. The number of entries to enclosed arms of the non-stressed, stressed and treated groups remained unchanged.

Obviously, (R)-8-OSO2CF3-PAT could not block the shock-induced immobility in the inescapable footshock compartment at the doses tested. Footshock-induced freezing provides one way of examining the anxiolytic potential of drugs from a number of different classes. At the dose of 2.5 mg/kg, but not at the dose of 5.0 mg/kg, buspirone reduced footshock-induced freezing.19 Ipsapirone (12.5 mg/kg) reduced the conditioned immobility behaviour, not only in stressed, but also in non-stressed animals.12 Therefore, in the latter case, it remained questionable whether ipsapirone has an anxiolytic action or whether it acts as on behavioural activity in general.

So far, effects of other (partial) 5-HT1A receptor agonists in this model have not been reported. Considering the anxiolytic efficacy of (R)-8-OSO2CF3-PAT in the other two anxiety models, it is possible that a different form of anxiety is generated in this model. The differential effect in the footshock compartment suggests that its anxiolytic effects may depend on the type of stimulus used to induce fear. Furthermore, it is postulated that different forms of experimental anxiety may be modulated in different ways by specific serotonin receptor subtypes.13 So far, only scarce data have been presented on the relationship between

(partial) 5-HT1A receptor agonists, 5-HT turnover in particular brain areas and stress. Saphier and Welch20 examined the effects of 8-OH-DPAT on neurochemical responses

59 Chapter 3 in various stress models. Conditioned fear-induced increases in the 5-HIAA/5-HT ratio in the prefrontal cortex were attenuated by 8-OH-DPAT, however, stress-induced excitatory output as a result of footshock, could not be inhibited by 8-OH-DPAT. The potential anxiolytic and partial 5-HT1A receptor agonist, ipsapirone (5 mg/kg, ip), reduces 5-HIAA/5-HT ratios* in various brain areas (brainstem, hypothalamus, striatum, hippocampus, anterior cortex and posterior cortex) by 40-50%.21 Buspirone, another partial 5-HT1A receptor agonist, mimics the inhibitory activity of 5-HT to supress neuronal activity in the dorsal raphe nuclei after systemic application8 and produces a decrease in cortical 5-HIAA levels (40%), comparable to decreases evoked by the same 21 dose of ipsapirone. However, like (R)-8-OSO2CF3-PAT, but unlike ipsapirone, buspirone had no effect on striatal 5-HIAA and 5-HT levels.22 The limbic, striatal and cortical, but not hippocampal, 5-HIAA/5-HT ratios were decreased by intra-DRN 8-OH- DPAT infusion.23 Liu et al. found a decreased 5-HIAA/5-HT ratio (up to 49%) in hippocampal rat brain homogenates after s.c. administration of 5-HT1A receptor agonists.24 The reference compound, 8-OH-DPAT, lowered 5-HT turnover by 39% at a dose of 1 µmol/kg. In the present study, administration of (R)-8-OSO2CF3-PAT induced a significant reduction of the 5-HIAA/5-HT ratio in the somatosensory cortex, dorsal hippocampus and lateral septum (41, 29 and 31%, respectively). The latter result can be compared with findings of Treit et al., who reported a decrease in defensive burying after septal and raphe lesions.13 Similarly, we observed abolition of defensive burying after administration of (R)-8-OSO2CF3-PAT. This suggests that the projection of the raphe to the septum might be particularly important. Trends towards a decrease of the 5- HIAA/5-HT ratio were found in paraventricular nucleus of the hypothalamus, the ventral median hypothalamus and central amygdala. Together, these findings suggest that serotonergic function after (R)-8-OSO2CF3-PAT administration seems to be reduced in the limbic system in general. The limbic structures are predominantly innervated by the dorsal raphe nucleus and the median raphe nucleus.25 Therefore, it is assumed that the observed anxiolytic effects are mediated by both nuclei, and not exclusively by either of these brain areas. Direct stimulation of dorsal hippocampal 5-

HT1A receptors may also produce anxiolytic effects, e.g., intra-dorsal hippocampal administration of 8-OH-DPAT increases exploration of open arms in the plus-maze.26 Przegaliñski et al. found anxiolytic-like effects of ipsapirone in the conflict drinking test after injection into the dorsal hippocampus.27

* Calculated from separately determined 5-HIAA and 5-HT levels

60 Potential Anxiolytic Properties of (R)-8-OSO2CF3-PAT

3.4 Conclusion

In conclusion, the present results suggest that acute administration of the 5-HT1A receptor agonist, (R)-8-OSO2CF3-PAT, induces anxiolytic effects by stimulation of the presynaptic 5-HT1A somatodendritic receptors in the raphe nuclei, resulting in a decrease of 5-HT neurotransmission in somatosensory cortex, dorsal hippocampus and lateral septum. Therefore, it is assumed that these brain regions may play an important role in the behavioural expression of anxiety. However, it cannot be excluded that a direct agonistic action on postsynaptic 5-HT1A receptors in the dorsal hippocampus contributes to this anxiolytic effect.

61 Chapter 3

3.5 Experimental Section

Animals and Drug Treatment. Male Wistar rats weighing 300 - 495 g at the beginning of the experiments were used. They were housed individually in transparent Plexiglass cages (25 × 25 × 30 cm) with a 12-h light-dark regime (light on between 08:00 - 20:00 h). All animals had free access to standard rat chow (Hope Farms) and tapwater. The experiments were carried out between 10:00 - 14:00 h. (R)-8-OSO2CF3- PAT was synthesized in the Department of Medicinal Chemistry, University Centre for Pharmacy in Groningen, The Netherlands. The compound was dissolved in saline and given intraperitoneally (ip), 30 min before the test session in a dose range of 1 - 3 mg/kg. Cited doses refer to the HCl salt and do not produce the 5-HT syndrome in the rats (see chapter 2). The control group received saline. Conditioned defensive burying. The shock-probe defensive burying test was performed in the animals' home cage. The floor was covered with wood shavings (height 2 cm). A removable teflon probe (10 cm long, 1 cm in diameter) was positioned 2 cm above the bedding. The probe was inserted through a small hole in the center of the wall of the cage. Two exposed wires (0.5 mm in diameter) were each wrapped (25 times) independently around the probe. Whenever the animal touched both wires simultaneously with some part of its body an electric current of 1.5 mA was delivered to the animal. During the entire period the shock circuit was left on, i.e. "repeated shock probe procedure" was used.15 Shock intensity was adjusted with a variable resistor in series with a 1000-V shock source. On day 2 vehicle or drug was injected 30 min before the introduction of the non-electrified probe in the home cages of the rats. Thus the procedure investigated the conditioned emotional consequence of former punishment rather than the direct effect of shock. All animals were observed for 10 min. To avoid false negative results, only animals burying for more than 25% of total time on day 1 were tested for day 2. Conditioned fear of footshock. The rats were exposed to a dark compartment (50 × 50 × 50 cm) equipped with a grid floor,12 where they were allowed to stay for 5 min. During the trial an inescapable scrambled footshock (0.6 mA, AC for 3 s) was given after the 1st and the 4th min. On the next day the animals were re-exposed to the dark footshock compartment in which no further shock was given. Elevated plus-maze. Directly after the 5-min re-exposure to the shock compartment the animals were placed in the elevated plus-maze. The apparatus consisted of two open (50 × 10 cm) and two enclosed arms (50 × 10 × 40 cm), arranged so that the two arms of the same type were opposite to each other,4 connected by an open central area (10 × 10 cm). The maze was elevated to a height of 50 cm. Light

62 Potential Anxiolytic Properties of (R)-8-OSO2CF3-PAT intensity in the open arms and closed arms was 200 - 350 Lux and <1 Lux, respectively. The rats were placed individually in the centre of the maze facing one of the enclosed arms. Each rat was tested for 5 min on the elevated plus-maze. The maze was cleaned with ‘Glassex’ after each rat had occupied it. Behavioural measurements. The behaviour in the defensive burying model was classified in five categories: (a) defensive burying - moving toward the probe and spraying or pushing bedding material toward the probe with rapid movements of the snout or forepaws as described by Pinel and Treit;28 (b) eating - chewing chow or faeces; (c) rearing - standing or sitting on hindlegs, mostly making sniffing movements, with the nose up into the air; (d) resting - the rat's hindlimbs, forelimbs, and belly touched the floor and supported its weight; (e) exploring - investigation of any part of the home cage. During the exposure to the former shock compartment, time spent immobile - i.e. animal completely motionless - was measured. On the elevated plus- maze, the activity scores - i.e. number of arm entries - the closed entries and the open entries were used as indices of the anxiolytic or anxiogenic effects.4 The observations were recorded by trained observers who were blind to the treatment order. Measurement of 5-HT and 5-HIAA. The rats were anaesthetized in an ether chamber and killed by rapid decapitation, 30 min after injection. The brains were immediately frozen in a dry ice precooled tube containing n-heptane and stored at −70 ° C until the assays were performed. For the assay, a brain was cut in slices of 2 mm (homemade brain-slicer) at −4 °C after which particular brain areas were punched out on a frozen surface. The tissue samples were homogenized in icewater in a 150-µL solution containing 5 µM clorgyline, 5 µg/mL glutathione and 20 ng/mL N-ω- methylserotonin (internal standard) with a High Intensity Ultrasonic Processor (Sonics

& Materials Inc., U.S.A.). Thereafter, 12.5 µL 2 M HClO4 and 10 µL 2.5 M potassium acetate was added to 50 µL of the homogenate. After 15 min the tissue samples were centrifuged for 10 min at 15,000 g (−10 °C). Thereafter, 30 µL of the supernatant was diluted with 450 µL UP. The samples were injected onto a reverse-phase/ion-pair High Performance Liquid Chromatography (HPLC) setup with electrochemical detection for the measurement of 5-HT and 5-HIAA. The chromatographic system consisted of a LKB 2150 HPLC pump (Pharmacia, Sweden), a Promis II autosampler (Spark, The Netherlands) with a 100-µL loop and a column (150 mm × 4.6 mm i.d.) packed with Hypersil ODS, 5 µm particle size (Alltech Associates Inc., U.S.A.). The mobile phase consisted of 0.051 M citric acid monohydrate, 0.063 M

NaH2PO4.2H2O, 0.403 mM EDTA, 0.356 mM sodium octyl sulphonate, 0.265 mM di- N,N-n-butylamine and 13% methanol. This buffer was set to pH 3.8 with 1 M HCl and

63 Chapter 3 then filtered through a 0.22-µm membrane filter (Schleicher & Schuell, Germany). Separation was done at room temperature using a flow rate of 1 mL/min. Detection of the 5-HT and 5-HIAA was performed using an electrochemical detector (Antec, Leiden, The Netherlands) with a glassy carbon working electrode set at −0.75 V (2nA/V) versus an Ag/AgCl reference electrode. The data were recorded with a chart recorder (Model BD41, Kipp & Zn., The Netherlands), and peak heights of samples were compared with those of standards determined each day for quantification. The limit of detection (signal/noise ratio 3:1) was 9.5 fmol/100 µL. Statistics. The data were analyzed with a one-way analysis of variance (ANOVA). The ANOVA was followed by Dunnett's t-test in order to compare the vehicle group to each of the drug groups.

64 Potential Anxiolytic Properties of (R)-8-OSO2CF3-PAT

3.6 References

[1] Dourish, C.T.; Hutson P.H.; Curzon, G.Trends Pharmacol. Sci. 1986, 7, 212. [2] a) Higgins, G.A.; Bradbury, A.J.; Jones, B.J.; Oakley, N.R.Neuropharmacology 1988, 96, 829. b) Higgins, G.A.; Jones, B.J.; Oakley, N.R.Psychopharmacology 1992,106, 261. c) Hogg, S.N.; Andrews, N.; File, S. Neuropharmacology 1994, 33, 343. [3] Sonesson, C.; Boije, M.; Svensson, K.; Ekman, A.; Carlsson, A.; Romero, A.G.; Martin, I.J.; Duncan, J.N.; King, L.J.; Wikström, H.;J. Med. Chem. 1993, 36, 3409. Sonesson, C.; Barf, T.; Nilsson, J.; Dijkstra, D.; Carlsson, A.; Svensson, K.; Smith, M.W.; Martin, I.J.; Duncan, J.N.; King, L.J.; Wikström, H.;J. Med. Chem. 1995, 38, 1319. [4] Pellow, S.; Chopin, P.; File, S.E.; Briley, M.J. Neurosci. Method. 1985, 144, 149. [5] Handley, S.L.; McBlane, J.W.J. Pharmacol. Toxicol. Meth. 1993, 29, 129. [6] Korte, S.M.; De Boer, S.F.; De Kloet, E.R.; Bohus, B.Psychoneuroendocrinology 1995, 20, 385. [7] Treit, D.; Pinel, J.P.J.; Fibiger, H.C.Pharmacol. Biochem. Behav. 1981, 15, 619. [8] Vandermaelen, C.P.; Matheson, G.K.; Wilderman, R.C.; Patterson, L.A.Eur. J. Pharmacol. 1986, 129, 123. [9] Sprouse, J.S.; Aghajanian, G.K.Synapse 1987, 1, 9. [10] Blier, P.; De Montigny, C.Synapse 1987, 1, 470. [11] Blanchard, R.J.; Shepherd, J.K.; Armstrong, J.; Tsuda, S.F.; Blanchard, D.C.Psychopharmacology 1993, 112, 55. [12] Korte, S.M.; Bohus, B.Eur. J. Pharmacol. 1990, 181, 307. [13] Treit, D.; Robinson, A.; Rotzinger, S.; Pesold, C.Behav. Brain Res. 1993, 54, 23. [14] De Boer, S.F.; Slangen, J.C.; Van der Gugten, J.Physiol. Behav. 1990, 47, 1089. [15] Treit, D.; Fundytus, M.Pharmacol. Biochem. Behav. 1988, 31, 1071. [16] a) Dourish, C.T.; Hutson, P.H.; Curzon, G. Psychopharmacology 1985, 86, 197. b) Hutson P.H.;

Donohoe, T.P.; Curzon, G.Eur. J. Pharmacol. 1986, 138, 215. c) Cooper, S.J. in Brain 5-HT1A receptors, 1987, 233-242, Dourish, C.T.; Ahlenius, S.; Hutson, P.H. (Eds) Ellis Horwood, Chichester. d) Hutson, P.H.; Donohoe, T.P.; Curzon, G.Psychopharmacology 1988, 95, 550. e) Cooper, S.J. Trends Pharmacol. Sci. 1989, 10, 56. f) Curzon, G. Ann. N.Y. Acad. Sci. 1990, 600, 521. [17] a) Hutson, P.H.; Dourish, C.T.; Curzon, G.Eur. J. Pharmacol. 1988, 129, 347. b) Macor, J.E.; Burkhart, C.A.; Heym, J.H.; Ives, J.L.; Lebel, L.A.; Newman, M.E.; Nielsen, J.A.; Ryan, K.; Schulz, D.W.; Torgersen, L.K.; Koe, B.K.J. Med. Chem. 1990, 33, 2087. [18] Handley, S.L.; McBlane, J.W.Psychopharmacology 1993, 112, 13. [19] Conti, L.H.; Maciver, C.R.; Ferkany, J.W.; Abreu, M.E.Psychopharmacology 1990, 102, 492. [20] Saphier D.; Welch, J.E.J. Neurochem. 1995, 64, 767. [21] Hamon, M.; Fattaccini, C.-M.; Adrien, J.; Gallissot, M.-C.; Martin, P.; Gozlan, H.;J. Pharmacol. Exp. Ther. 1988, 246, 745. [22] Cimino, M.; Ponzio, F.; Achilli, G.; Vantini, G.; Perego, C.; Algeri, S.; Garattini, S.Biochem. Pharmacol. 1983, 32, 1069. [23] Hjorth S.; Magnusson, T. Naunyn-Schmiedeberg’s Arch. Pharmacol. 1988, 338, 463. [24] a) Liu, Y.; Yu, H.; Svensson, B.E.; Cortizo, L.; Lewander, L.; Hacksell, U,J. Med. Chem. 1993, 36, 4221. b) Liu, Y.; Yu, H.; Mohell, N.; Nordvall, G.; Lewander, T.; Hacksell, U.J. Med. Chem. 1995, 38, 150. [25] Azmitia, E.C.; Segal, M.J. Comp. Neur. 1978, 179, 641. [26] Guimaraes, F.S.; Del Bel, E.A.; Padovan, C.M.; Mendonca Netto, S.; De Almeida, R.T.Brain Res. 1994, 58, 133. [27] Przegaliñski, E.; Tatarczyñska, E.; Klodziñska, A.; Chojnacka-Wõjcik, E. Neuropharmacology1994 , 33, 1109. [28] Pinel, J.P.J.; Treit, D. The conditioned defensive burying paradigm and behavioral neuroscience, Behavioral approaches to brain research 1983, 212-234, Robinson, T.E. (Ed) Oxford, Oxford University Press.

65 Chapter 4

5-HT1D Receptor Agonist Properties of Novel 5- [[(Trifluoromethyl)sulfonyl]oxy]indolylethylamines*

Abstract

2-[5-[[(Trifluoromethyl)sulfonyl]oxy]-1H-indol-3-yl]ethylamine (18) and its N,N-di-n-propyl (12), N,N-diethyl (13), N,N-dimethyl (14) derivatives, and 4-[3-[2- (N,N-dimethylamino)ethyl]-1H-indol-3-yl]-N-(p-methoxybenzyl)acrylamide

(GR46611, 19) were synthesized and tested for binding affinities to cloned 5-HT1A, 5-

HT1Dα, 5-HT1Dβ and D2 receptors. In addition, the intrinsic efficacy was measured as the reduction of forskolin stimulated cAMP in cells transfected with 5-HT1Dα and 5-HT1Dβ receptors in vitro. The 5-substituted indolylethylamines investigated displayed agonist activity at the 5-HT1D receptors with varying degrees of preference for the 5-HT1Dα vs the 5-HT1Dβ receptors. The primary amine and N,N-dimethyl substitution seemed to be optimal for 5-HT1Dα affinity. Furthermore, the N,N-diethyl (13) and N,N-dimethyl (14) derivatives showed a 10-25 times preference for the 5-HT1Dα vs the 5-HT1Dβ receptor. In addition, all of the novel compounds showed affinity for the 5-HT1A receptor in vitro (Ki values ranging from 18 to 40 nM). The most promising derivative 14, was virtually devoid of central 5-HT1A agonist activity in rats, as determined by in vivo biochemical assays. Paradoxically, 14, like 19, induced a hypothermic response and a decrease in 5- HIAA levels in the prefrontal cortex and hypothalamus in guinea pigs after systemic administration. Sumatriptan failed to produce either of these effects due to a poor brain penetration.

4.1 Introduction

5-HT1D receptors were first defined in bovine caudate and subsequently in the 1,2 brains of other species, including man. The 5-HT1D receptor is the most abundant 5-

HT1 receptor subtype in the mammalian CNS, existing as a presynaptic heteroreceptor or a terminal autoreceptor, activation of which inhibits serotonin release.3,4 Human 5,6 genes encoding for the 5-HT1Dα and 5-HT1Dβ receptor have recently been cloned,

* This chapter is based on: Barf, T.A.; De Boer, P.; Wikström, H.; Peroutka, S.J.; Svensson, K.A.; Ennis, M.D.; Ghazal, N.B.; McGuire, J.C.; Smith, M.W.J. Med. Chem. 1996, In press.

65 Chapter 4 raising questions about which of the two receptors is relevant to reported pharmacological effects. The mRNAs for the 5-HT1Dα and 5-HT1Dβ receptors appear to codistribute in the brain of non-rodent species, however, the density of the 5-HT1Dα 7 receptor mRNA is much lower. It seems that the 5-HT1Dβ receptors constitute the human counterpart of rodent 5-HT1B receptors, and have also been identified in vascular smooth muscle mediating contraction.8,9 Stimulation of the former receptors by 10 selective 5-HT1Dβ receptor agonists such as sumatriptan (1) and newer 5-C-substituted tryptamine derivatives such as MK-462 (2)11 and 311C90 (3)12 may account for the clinical effectiveness of these agents in the treatment of migraine. In addition, the anti- migraine action of these agents has been attributed to other, both peripheral and central 13 mechanisms mediated by 5-HT1D receptors. Moreover, the functional distinction between the 5-HT1Dα and 5-HT1Dβ receptor subtypes is unclear. Obviously, selective agonists and antagonists are needed to unravel the role of 5-HT1D receptors in the CNS 14 and in the periphery. 5-HT1D receptor antagonists, such as GR127935 (4) and GR55562 (5)15 are cautiously suggested to serve as centrally acting , alone or in combination with SSRIs (Selective Serotonin Reuptake Inhibitors).16

H O O N N O S O MeNH N N H

Me Me Me N N N N Me N Me N Me H H H

Sumatriptan 1 MK-462 2 311C90 3

Me Me N N Me N H O N N OMe OH O N N Me N O H Me

GR127935 4 GR55562 5

66 5-HT1D Receptor Agonist Properties of Novel 5-[[(Trifluoromethyl)sulfonyl]oxy]indolylethylamines

In almost all 5-HT1D receptor agonists reported to date, a 5-substituted tryptamine serves as a structural template. It seems that selectivity for either one of the

5-HT1 receptor subtypes can be achieved by modifying substituents at the 5-position of the indole portion and at the basic nitrogen atom of the ethylamino side-chain.17 Certain, mostly hydrogen bond accepting, (aromatic) heterocycles have proven to be viable replacements for the C-5 hydroxy substituent of serotonin itself.18 Other groups embarked on the synthesis of conformationally restricted tryptamine analogues.19

However, few data have been published on the affinity of ligands for 5-HT1Dα and 5-

HT1Dβ receptors individually. Sumatriptan, and to date reported 5-HT1D ligands, possess no, or at most, limited selectivity for 5-HT1Dα or 5-HT1Dβ receptor binding sites. Recently, the trifluoromethylsulfonyloxy (triflate) group has been succesfully applied as a bioisostere of a hydroxy or a methoxy functionality in a number of 2- aminotetralins.20 These dopaminergic and serotonergic ligands generally displayed improved pharmacokinetic properties compared to their hydroxy/methoxy analogues while exhibiting a similar pharmacological profile. The enhanced 5-HT1D receptor affinities of triflate substituted 2-aminotetralins prompted us to replace the N- methylaminosulfonylmethylene group of sumatriptan by a triflate group, enabling us to investigate the affinity of these new tryptamine analogues to 5-HT1A, 5-HT1Dα and 5-

HT1Dβ receptor subtypes. N,N-Dialkyl substituents were introduced in order to gain more insight into the structure-affinity relationships (SAFIR) and structure-activity relationships (SAR) of tryptamine derivatives. Furthermore, aromatic triflates may serve as key intermediates in the synthesis of phenyl ring substituted compounds, as was shown in a number of triflated 2-aminotetralins21 and phenylpiperidines.22 Correspondingly, this chemistry is applicable on triflated tryptamines, as is exemplified by the one-step synthesis of the potent 5-HT1A/1D receptor agonist GR46611 (19) from compound 14. The 5-HT1Dα and 5-HT1Dβ receptor-mediated inhibition of forskolin- stimulated cAMP formation was measured for a series of prepared compounds. In case of compounds 14 and 19, these data are substantiated by neurochemical data and hypothermic effects.

67 Chapter 4

OBz OH OSO2CF3

a or b c

NR2 NR2 NR2 N N N H H H 6: R = n-Pr 9: R = n-Pr 12: R = n-Pr 7: R = Et 10: R = Et 13: R = Et 8: R = Me 11: R = Me 14: R = Me

Scheme 4.1. (a) ammonium formate, 10% Pd/C, 96% EtOH; (b) 2H (4 atm), Pd/C, MeOH; (c) PhN(SO2CF3)2, Et3N, CH2Cl2

4.2 Chemistry

The triflates 12-14 were synthesized according to Scheme 4.1, starting from the N,N-dialkylated 5-benzyloxyindolylethylamines, which were prepared according to literature procedures.23 Catalytic debenzylation in the presence of ammonium formate or H2 atmosphere afforded the 5-hydroxytryptamines. The triflates were synthesized by treatment with N-phenyltrifluoromethanesulfonimide and a base.24 Since serotonin itself could not be directly triflated the amine functionality was protected first by a phthalimido group.25 After triflation conversion to the amine was conducted by deprotection with hydrazine (Scheme 4.2).26 Interestingly, the aromatic triflates can be used as synthetic intermediates to other 5-substituted indolylethylamine derivatives. This is exemplified by the synthesis of 19, which was effected via coupling of the triflate analogue 14 and p-methoxybenzylacrylamide in the presence of Pd(OAc)2 and 1,3-diphenylphosphinopropane (dppp) (Scheme 4.3).21, *

* This step was performed by Micheal D. Ennis at Pharmacia & Upjohn Inc.

68 5-HT1D Receptor Agonist Properties of Novel 5-[[(Trifluoromethyl)sulfonyl]oxy]indolylethylamines

OH OH OSO2CF3

a b

NH2 NPhth NPhth N N N H H H 15 16 17

OSO2CF3 OSO2CF3 O

c Phth = d Me

NH2 N N N Me O H H 18 14

Scheme 4.2. (a) N-EtCO2-phth, 10% NaHCO3 (pH 8), THF/H2O; (b) PhN(SO2CF3)2, Et3N, CH2Cl2; (c) H2NNH2.H2O, EtOH; (d) 37% formaldehyde, NaCNBH3, pH 5, CH2Cl2.

Me Me OSO2CF3 N O a Me N N H N N Me MeO H H 14 GR46611 19

Scheme 4.3. (a) p-Methoxybenzylacrylamide, Pd(OAc)2, dppp, DMF, 85 °C.

69 Chapter 4

4.3 Pharmacology

Receptor Binding. The abilities of the test compounds to displace the 3 3 radioactively labelled ligand [ H]8-OH-DPAT (5-HT1A), [ H]5-HT (5-HT1Dα and 5- 3 HT1Dβ) and [ H]U-86170 (D2) were assessed in mammalian receptor clones expressed in CHO cells (Table 4.1A, Method A). In addition, the compounds were evaluated for their in vitro binding affinity at 5-HT1Dα and 5-HT1Dβ human receptor clones expressed in a human embryonic kidney (HEK 293) cell line (Table 4.1B, Method B).

Table 4.1A. Affinities at Cloned 5-HT1A, 5-HT1Dα/β and D2 receptors in Vitro

Ki (nM)a

b b b c compd. 5-HT1A 5-HT1Dα 5-HT1Dβ D2 Select. 5-HT1Dα

vs 5-HT1Dβ 12 23 (19-27) 190 (161-224) 246 (198-306) 502 (354-712) 1.3 13 27 (20-35) 12 (11-14) 171 (140-210) 637 (479-994) 14 14 40 (32-50) 3.2 (2.8-3.6) 32 (28-36) 538 (379-764) 10 18 18 (15-22) 2.8 (2.5-3.1) 14 (11-16) 658 (472-918) 5.0 19 1.3 (0.8-2.1) 0.3 (0.2-0.5) 0.2 (0.1-0.3) >1000 1.5 1 341 (283-522) 5.7 (2.9-9.5) 22 (19-27) >218 3.8

3 (a) Ki values for displacement of the 5-HT1A receptor agonist [ H]8-OH-DPAT, the 5-HT1Dα/β 3 3 agonist [ H]5-HT and the dopamine D2 receptor agonist [ H]U86170. (b) Method A; data from cloned human receptors expressed in CHO-K1 cells. Mean values (n=3). Parentheses contain 95% confidence intervals. (c) Data from cloned rat receptors expressed in CHO-K1 cells.

cAMP Assay. The functional cAMP assay using the cloned human 5-HT1Dα and 27,28 5-HT1Dβ receptors was employed as previously described. Compounds 12, 13, 14, 18 and 19 were tested at 10 µM and the agonist inhibition was calculated as a percent of the 5-HT control (Table 4.2). In Vivo Biochemistry. The synthesis rate of 5-HT in terminal brain areas is inhibited by 5-HT1A receptor agonists due to stimulation of the somatodendritic 5-HT1A receptors in the raphe nuclei.29 The effect of compound 14 on 5-HT synthesis was measured in four brain areas in reserpinized rats. 5-Hydroxytryptophan (5-HTP) accumulation, following decarboxylase inhibition by (3-hydroxybenzyl)hydrazine (NSD 1015), was used as an indicator of the 5-HT synthesis rate (Table 4.3).30 The extracellular 5-HT levels in the hippocampus were measured by in vivo microdialysis after sc administration of 14 in normosensitive rats (Table 4.3). In addition, the 5-HT

70 5-HT1D Receptor Agonist Properties of Novel 5-[[(Trifluoromethyl)sulfonyl]oxy]indolylethylamines and 5-HIAA levels after administration of 14 and 19 were determined in the prefrontal cortex and the hypothalamus of guinea pigs (Table 4.4).

Table 4.1B. Affinities at Cloned 5-HT1Dα and 5-HT1Dβ receptors in Vitro a IC50 (nM) b b c compd 5-HT1Dα 5-HT1Dβ log D 6 470±100 550±200 3.08 7 110±7 260±60 2.18 8 25±4 76±10 2.28 9 820±100 2600±500 1.24 10 77±20 690±100 0.28 12 170±10 660±50 2.21 13 21±2 530±30 1.31 14 8.4±1 98±6 1.41 18 NTd NT −0.24 19 NTd NT 1.71

(a) IC50 values for displacement of the 5-HT1Dα and 5-HT1Dβ receptor agonist [3H]5-HT. (b) Method B; data from human5-

HT1Dα and 5-HT1Dβ receptor clones, expressed in a human embryonic kidney (HEK 293) cell line. Mean value± SEM (n=3). (c) Calculated with Pallas 1.2 (CompuDrug Chemistry Ltd. 1994). (d) NT means Not Tested.

71 Chapter 4

Table 4.2. Intrinsic Efficacy in Cells Transfected with Human 5-HT1Dα or

5-HT1Dβ Receptors cAMP (pmol)a

Compound 5-HT1Dα (% 5-HT) 5-HT1Dβ (% 5-HT) Forskolin control 64±3 (0) 166±6 (0) 5-HT (1µM) 27±1 (100) 13±1 (100) 1 (1µM) 30±1 (92) 24±1 (93) 12 (10 µM) 36±4 (75) 88±11 (51) 13 (10 µM) 29±3 (94) 37±3 (84) Forskolin control 61±4 (0) 232±8 (0) 5-HT (1µM) 20±1 (100) 17±1 (100) 1 (1µM) 31±4 (75) 32±2 (93) 14 (10 µM) 30±1 (75) 28±1 (95) 18 (10 µM) 19±2 (102) 18±2 (100) 19 (1 µM) (104)b NTc

(a) Values are expressed as means± SEM (n=3). Data within parentheses denote percent of 5-HT’s response. (b) Tested in a separate assay. (c) NT means Not Tested.

Hypothermia. 5-HT1D receptors are implicated in the regulation of body temperature of guinea pigs.31 The ability of compounds 14 and 19 to induce hypothermia in guinea pigs after sc administration was tested in a 60-min reading (Table 4.4).

4.4 Results and Discussion

Structure-Affinity and Structure-Activity Relationships . From the data presented in Tables 4.1A and 4.1B two clear trends can be observed. Firstly, the size of the N,N-dialkyl group dramatically influences the affinity for both 5-HT1D receptor subtypes and secondly, the nature of the 5-O-substituent is of major importance for both the affinity and the selectivity for these subtypes. N,N-Dimethyl substitution (IC50 values 8.4-25 nM) improves the affinity for 5-HT1Dα sites by approximately 20-fold as compared to the N,N-di-n-propyl derivatives (Ki values 170-820 nM), but only 7-fold in case of 5-HT1Dβ receptors (Table IIB). Notably, bulk at the protonated amine site of the tryptamine is poorly tolerated by the latter receptor, resulting in moderate binding affinities even for the N,N-dimethyl derivatives. Within the N,N-di-n-propyl derivative series the triflate group enhances the affinity for 5-HT1Dα sites by 3- and 5-fold relative to the benzyloxy and hydroxy substituents, respectively. Similar trends are found in the

72 5-HT1D Receptor Agonist Properties of Novel 5-[[(Trifluoromethyl)sulfonyl]oxy]indolylethylamines

N,N-diethyl and N,N-dimethyl substituted tryptamines. The comparatively low affinities of the 5-benzyloxy and 5-hydroxy derivatives for the 5-HT1D receptor subtypes suggest that the 5-oxygen is not of crucial importance for the 5-HT1D receptor interaction. Obviously, the triflate derivatives benefit from their two additional sulfonyl oxygens which are capable of accepting a hydrogen bond. In addition, the electron-withdrawing effect of the triflate substituent may contribute to a putative interaction of the indole portion with the active-site surrounding amino acid residues. Since the novel triflate derivatives 12-14 and 18 displayed the most interesting binding profiles, these compounds were subjected to the in vitro 5-HT1D receptor agonist assays. In the in vitro binding assay, the N,N-di-n-propyl substituted analog 12 showed a comparatively weak affinity (Ki values around 200 nM) for both 5-HT1D receptor subtypes, while the affinity for the 5-HT1A site was higher (23 nM, Table 4.1A).

The IC50 values (Table 4.1B) revealed a 4-fold preference for the the 5-HT1Dα vs 5-HT1Dβ binding site (170 nM and 660 nM, respectively). Despite these moderate affinities, compound 12 reduced the formation of cAMP, both in the 5-HT1Dα and 5-HT1Dβ receptor assays; however, the intrinsic activity appeared to be higher in the former (Table 4.2).

As already indicated, the 5-HT1Dα receptor affinity was greatly improved in the N,N- diethyl substituted analogue 13. This compound showed approximately a 15- to 25-fold preference for the 5-HT1Dα vs 5-HT1Dβ receptor, but only a 2-fold preference vs the 5-

HT1A receptor subtype. In the 5-HT1Dα and 5-HT1Dβ receptor agonist assay similar and slightly lower intrinsic activity was observed, respectively, compared to sumatriptan when tested at 10 µM. Increased affinity at the expense of selectivity was observed in moving from N,N-diethyl to N,N-dimethyl substituents. Compound 14 exhibited a 10- to

12-fold preference for the 5-HT1Dα vs 5-HT1Dβ receptor subypes (Ki values (IC50 values) of 3.2 nM (8.4 nM), 32 nM (98 nM), respectively). Interestingly, 14 showed both higher affinity for the 5-HT1Dα receptor subtype and a higher 5-HT1Dα vs 5-HT1Dβ preference than sumatriptan (1). As expected, both 1 and 14 were found to be agonists with similar intrinsic efficacy in both cAMP assays. In addition, the primary amine 18 showed higher affinity for the 5-HT1Dα site (Ki = 2.8 nM). However, the affinities for the 5-HT1A and 5-HT1Dβ sites were also increased (Ki values of 18 and 14 nM, respectively). Compound 18 displayed a maximal intrinsic efficacy, similar to that of 5-HT in the cAMP assay. Introduction of a hydrophobic tail on the 5-position of the indole nucleus seems 32 to favor 5-HT1Dβ affinity as exemplified by 5-(nonyloxy)tryptamine. This 5-HT1D receptor ligand binds with higher affinity at human 5-HT1Dβ receptors than 5-HT1Dα receptors (Ki = 1.2 and 16 nM, respectively). Similarly, the p- methoxybenzylacrylamido group of 19, allowing for hydrogen bond formation, is well-

73 Chapter 4

tolerated by the 5-HT1A/1D receptor subtypes. Table 4.1A shows that 19 potently displaced the radioligands at these receptor subtypes exhibiting Ki values of 1.3 nM (5-

HT1A), 0.3 nM (5-HT1Dα), and 0.2 nM (5-HT1Dβ), and was found to behave as a full 5-

HT1Dα receptor agonist in the cAMP assay. Obviously, both 5-HT1D receptor binding sites contain a (lipophilic) pocket which can accommodate bulky, hydrogen-bond accepting groups. None of the compounds tested showed appreciable affinity for the dopamine D2 receptor (Table 4.1A).

Table 4.3. Effect of 14 on Rat Brain 5-HT Synthesis (5-HTP Accumulation)in Vivo in Reserpinized and Nonpretreated Rats Reserpinized (5-HTP Accumulation)a Normal (5-HT Levels)b striatum accumbens frontal cortex hippocampus 10 µmol/kg 50 µmol/kg saline 0.37±0.03 0.49±0.07 0.47±0.09 0.28±0.02 - - 14 0.28±0.01 0.40±0.02 0.37±0.01 0.28±0.03 - - % 75.7 81.6 78.7 100.0 108±10.5 75±4.0

(a) The animals received reserpine 24 h before decapitation. 45 min and 30 min before the test drug (10µmol/kg) and NSD 1015 (100mg/kg) were administered, respectively. Shown are the [5-HTP] inµgram/gram wet tissue (means ± SEM) in striatum (n=8), accumbens (n=4), frontal cortex (n=4) and hippocampus (n=4), aftersc administration of saline or compound14 (10 µmol/kg). (b) Determined byin vivo microdialysis. The values are percent of control 5-HT levels, means± SEM (n=3), 1 h aftersc administration of compound14 .

Pharmacology. The 5-HT1D receptor agonist (14) of this series was additionally screened for its (unwanted) central 5-HT1A receptor activity by means of in vivo biochemical models. In rats, 14 failed to significantly inhibit the 5-HTP accumulation at a dose of 10 µmol/kg in the investigated brain areas, although slight decreases were observed in striatum, accumbens and frontal cortex. The inactivity at the 5-HT1A receptor is substantiated by the microdialysis data, which showed no differences compared to the control 5-HT levels in the hippocampus after sc administration of 14 at a dose of 10 and 50 µmol/kg (Table 4.3). However, like 19, compound 14 has a pronounced effect on the 5-HT turnover in the prefrontal cortex and the hypothalamus of the guinea pig brain using similar doses. A lower dose of 14 tested (3.1 µmol/kg, sc) did not produce a statistical significant effect. As previously reported, sumatriptan failed to affect guinea pig brain 5-HT turnover after systemic administration. This is likely explained by the fact that sumatriptan has a poor penetration across the blood brain barrier, as indicated by its low calculated log D value (−0.53). This hypothesis is further strengthened by the fact that sumatriptan failed to produce hypothermia, while 19 and 14 significantly produced a hypothermic response at the 12.5- and 50-µmol/kg dose, sc.31a

74 5-HT1D Receptor Agonist Properties of Novel 5-[[(Trifluoromethyl)sulfonyl]oxy]indolylethylamines

Table 4.4. Effects on Guinea Pig Rectal Temperature and Brain 5-HT Turnover

Dose Hypothermiaa Neurochemistry (% of vehicle controls)b Hypothalamus Prefrontal Cortex µmol/kg, sc 0-60 min, ∆ °C 5-HT 5-HIAA 5-HT 5-HIAA saline - +0.18±0.15 100±6 100±5 100±5 100±9 14 12.5 −1.70±0.43* 97±13 66±16 104±8 59±9*** 50 −1.72±0.36** 123±16 54±16* 129±1 46±6*** 19 12.5 −1.62±0.35** NTc NT NT NT 50 −2.62±0.05*** 121±11 49±7** 94±6 37±7*** 1 25 −0.14±0.13 122±5 133±13 119±6 129±14 8-OH-DPAT 6.25 +0.22±0.15 134±17 37±7** 126±2 34±8**

(a) The hypothermia is presented as the difference in°C from control. The data represent a sixty-minute reading and are expressed as mean± SEM (n=4-5). (b) The animals were decapitated, 60 min after test drug treatment. Shown are the 5-HT and 5-HIAA levels expressed as mean± SEM (n=4-5), after sc administration of the test drugs. (c) NT means Not Tested. * P < 0.05, ** P < 0.01 and *** P < 0.001 vs vehicle treated controls.

The ability of 19 and 14 to lower the central 5-HT turnover is likely the result from activation of inhibitory presynaptic 5-HT1D or somatodendritic 5-HT1A and/or 5-HT1D receptors. Interestingly, in contrast to rats and mice, the selective 5-HT1A receptor agonist 8-OH-DPAT (20) failed to induce hypothermia in the guinea pig, indicating that

5-HT1D receptors are of higher importance for the temperature regulation in this species.

The levels of 5-HIAA were reduced probably as a result of a stimulation of 5-HT1A cell body autoreceptors. OH N

8-OH-DPAT 20

75 Chapter 4

In conclusion, the triflate substituted tryptamines investigated all display agonist activity at the 5-HT1D receptors with varying degrees of preference for the 5-HT1Dα vs the 5-HT1Dβ receptors in vitro. The primary amine and N,N-dimethyl substitution seemed to be optimal for 5-HT1Dα affinity. Furthermore, the N,N-dimethyl and N,N-diethyl analogs showed a 10 to 25-fold selectivity for the 5-HT1Dα vs 5-HT1Dβ receptor. In addition, all of the compounds showed substantial affinity for the 5-HT1A receptor in vitro (Ki values ranging from 18−40 nM). Compound 14 seems to have a weak inhibitory effect on the 5-HT turnover at doses up to 50 µmol/kg (sc) in rats, but induces pronounced decreases of 5-HT turnover in guinea pigs. This contrasting observation may be explained by the fact that in these experiments, the 5-HT turnover in rats is predominantly mediated by 5-HT1A cell body autoreceptors, whereas the inhibitory presynaptic 5-HT1D receptors have a major contribution in the observed effect in guinea pigs. The potential antimigraine action of compound 14 has been evaluated by means of a porcine carotid blood flow model and appears to be equipotent with sumatriptan.33 The question is whether brain penetration is a desireable property of antimigraine agents or not. Animal studies have provided evidence that if putative antimigraine drugs, such as 3, get access to the CNS, they display central neuronal actions.13 Whether this provides increased clinical efficacy in the treatment of migraine is still in debate.

4.5 Experimental Section

General. For general remarks See section 2.4. Log D values were calculated with Pallas version 1.2.34 Materials. All 5-benzyloxyindolylethylamines were prepared according to literature procedures.23 GR46611 (19) was synthesized at the Pharmacia & Upjohn Inc. Chemicals were commercially available (Aldrich) and used without further purification. N,N-Di-n-propyl-2-[5-benzyloxy-1 H-indol-3-yl]ethylamine oxalate (6). N,N- Di-n-propyl-2-[5-benzyloxy-1H-indol-3-yl]glyoxalylamide (2.20 g, 5.82 mmol) was dissolved in dry Et2O (80 mL) and dry THF (20 mL) at room temperature. LiAlH4 (2.40 g, 11 eq.) was added portionwise and the reaction mixture was refluxed for 6 h under N2 (g). After cooling the reaction mixture to room temperature, the reaction was quenched with the addition of water (2.4 mL), 10% NaOH (2.4 mL) and water (7.2 mL) under N2 (g). This mixture was stirred until the Li-salts had turned white. These salts were filtered and washed with Et2O. The filtrate was evaporated under reduced pressure resulting in a yellow oil (1.83 g, 90%). Conversion into the oxalate and recrystallization from acetonitrile gave pale brown crystals (1.68 g, 66%): mp 128-130 °C; IR (KBr) cm-1 1196

76 5-HT1D Receptor Agonist Properties of Novel 5-[[(Trifluoromethyl)sulfonyl]oxy]indolylethylamines

1 (C−O); H NMR (CD3OD) δ 0.94 (t, J = 7.27, 6H), 1.57-1.76 (m, 4H), 3.03-3.14 (m, 6H),

3.27-3.34 (m, 2H), 5.09 (s, 2H), 6.87 (dd, J1 = 8.98, J2 = 2.14, 1H), 7.12 (d, J = 2.14, 13 1H), 7.15 (s, 1H), 7.26-7.47 (m, 6H); C NMR (CD3OD) δ 10.9, 17.8, 20.8, 53.7, 55.1, 71.7, 102.5, 109.7, 113.1, 113.5, 124.8, 128.1, 128.3, 128.4, 129.1, 133.3, 154.0, 166.5; +1 MS (CI with NH3) m/e 351 (M ); Anal Calcd (Obsd) for C23H30N2O.C2H2O4: C: 68.16 (67.87, H: 7.32 (7.23), N: 6.36 (6.50). N,N-Diethyl-2-[5-benzyloxy-1 H-indol-3-yl]ethylamine oxalate (7). Reduction of N,N-diethyl-2-[5-benzyloxy-1H-indol-3-yl)glyoxalylamide (1.93 g, 5.51 mmol) was performed according to the procedure given for the synthesis of 6, resulting in a brown oil after evaporation of the solvents (1.70 g, 96%). Conversion to the oxalate and recrystallization from acetone gave pale green crystals (1.75 g, 77%): mp 154-156 °C (lit mp 161-162 °C)22; IR (KBr) cm-1 1186 (C−O); 1H NMR δ 1.14 (t, J = 7.26, 6H), 2.72 (q, J = 7.27, 4H), 2.79-2.99 (m, 4H), 5.14 (s, 2H), 6.98 (m, 2H), 7.16-7.55 (m, 7H), 8.24 (br s, NH); 13C NMR δ 11.7, 22.8, 46.6, 53.3, 70.9, 103.4, 111.9, 112.8, 114.0, 122.4, +1 127.6, 127.7, 127.9, 128.5, 131.5, 137.6, 152.8; MS (CI with NH3) m/e 323 (M ); Anal

Calcd (Obsd) for C21H26N2O.C2H2O4: C: 66.26 (66.20), H: 6.90 (6.92), N: 6.72 (6.71). N,N-Dimethyl-2-[5-benzyloxy-1 H-indol-3-yl]ethylamine oxalate (8). N,N- Dimethyl-2-[5-benzyloxy-1H-indol-3-yl]glyoxalylamide (1.78 g, 5.53 mmol) was converted to 8 according to the procedure given for the synthesis of 6, resulting in a colorless oil (1.53 g, 95%). This product was converted to its oxalate salt and recrystallization from MeOH/Et2O gave a white solid (1.79 g, 85%): mp 167-170 °C (lit mp 178-179 °C)22; IR (KBr) cm-1 1186 (C−O); 1H NMR δ 2.38 (s, 6H), 2.66 (t, J = 7.32, 2H), 2.95 (t, J = 7.32, 2H), 5.11 (s, 2H), 6.90 (s, 1H) 6.96 (d, J = 8.79, 1H), 7.17 (d, J = 5.12, 2H), 7.41 (m, 5H), 9.03 (br s, NH); 13C NMR δ 23.3, 45.0, 59.9, 70.7, 102.1, 111.7, 112.2, 113.0, 122.5, 127.3, 127.4, 128.2, 131.6, 137.5, 152.5; MS (EIPI) m/e 294 (M+);

Anal Calcd (Obsd) for C19H22N2O.C2H2O4: C: 65.61 (65.24), H: 6.29 (6.24), N: 7.29 (7.49). N,N-Di-n-propyl-2-(5-hydroxy-1 H-indol-3-yl)ethylamine (9). The crystals of 6 (1.51 g, 3.43 mmol) were dissolved in 95% EtOH (50 mL) after which ammonium formate (2.16 g, 10 eq.) and Pd/C (10%, 100 mg) were added. The reaction mixture was stirred at room temperature for 2 days. The solids were filtered over Celite and the filtrate was evaporated in vacuo leaving a brown oil. 10% Aqueous Na2CO3 (50 mL) was added and the product amine was extracted with EtOAc (3 × 30 mL). The organic phases were separated, pooled, dried (MgSO4) and filtered. The solvent was removed under reduced pressure yielding a pale brown solid (0.79 g, 89%). Part of this solid (0.28 g) was recrystallized from acetonitrile giving 0.24 g of brownish crystals: mp 135-136 °C; -1 1 IR (KBr) cm 3236 (OH); H NMR (CD3OD) δ 0.92 (t, J = 7.31, 6H), 1.49-1.61 (m, 4H),

77 Chapter 4

2.49-2.57 (m, 4H), 2.71-2.88 (m, 4H), 6.66 (dd, J1 = 8.55, J2 = 2.42, 1H), 6.91 (d, J = 13 2.56, 1H), 6.97 (s, 1H), 7.14 (d, J = 8.45, 1H); C NMR (CD3OD) δ 12.0, 20.3, 22.8, 55.4, 56.8, 103.1, 112.0, 112.4, 112.9, 123.6, 129.0, 132.8, 150.7; MS (EIPI) m/e 260 (M+);

Anal Calcd (Obsd) for C16H24N2O: C: 73.81 (73.69), H: 9.29 (9.21), N: 10.76 (10.84). N,N-Diethyl-2-(5-hydroxy-1 H-indol-3-yl)ethylamine oxalate (10). The crystals of compound 7 (5.00 g, 12.14 mmol) were dissolved in dry MeOH (200 mL) and hydrogenated over 10% Pd/C in a Parr apparatus under a H2 pressure of 4 atm. After 4 h the reaction mixture was filtered over Celite by suction and after evaporating the solvent under reduced pressure a tarry pinkish oil (3.77 g) was obtained.

Recrystallization from MeOH/Et2O gave pale brown crystals (2.57 g; 66%): mp 226-228 -1 1 °C; IR (KBr) cm 3300 (OH); H NMR (CD3OD) δ 1.09 (t, J = 7.17, 6H), 2.64 (q, J =

7.27, 4H), 2.72-2.86 (m, 4H), 6.67 (dd, J1 = 8.65, J2 = 2.38, 1H), 6.92 (d, J = 2.23, 1H), 13 6.96 (s, 1H), 7.15 (d, J = 8.68, 1H); C NMR (CD3OD) δ 11.0, 22.6, 47.4, 54.1, 103.1, +1 112.0, 112.4, 112.8, 123.6, 129.0, 132.8, 150.8; MS (CI with NH3) m/e 233 (M ); Anal

Calcd (Obsd) for C14H20N2O.0.6 C2H2O4: C: 63.76 (63.87), H: 7.46 (7.59), N: 9.78 (9.80). N,N-Dimethyl-2-(5-hydroxy-1 H-indol-3-yl)ethylamine oxalate (11). Compound 8 (1.51 g, 3.93 mmol) was converted to 11 according to the procedure given for the synthesis of 10, resulting in a purple oil (1.01 g) which solidified upon standing. Pink crystals (70 mg) were obtained whilst stirring in MeOH (5 mL) and were collected by filtration on a glass-sintered funnel. The filtrate was taken up in 10% aqueous

Na2CO3 (50 mL), then the product amine was extracted into EtOAc (3 × 30 mL). The organic phases were separated, pooled, dried (MgSO4) and filtered. The filtrate was evaporated under reduced pressure leaving a brown oil (0.47 g, 65%): mp (oxalate) 90- 22 -1 1 93 °C (lit 93-94 °C) ; IR (KBr) cm ; H NMR (CD3OD) δ 2.23 (s, 6H), 2.54-2.82 (m, 13 4H), 6.71 (d, J = 8.64, 1H), 6.94 (m, 2H), 7.15 (d, J = 8.64, 1H); C NMR (CD3OD) δ 23.9, 45.0, 60.9, 103.3, 112.2, 112.5, 123.7, 129.0, 132.8, 150.8; MS (EIPI) m/e 204 + (M ); Anal Calcd (Obsd) for C12H16N2O.C2H2O4: C: 53.82 (53.71), H: 6.46 (6.37), N: 8.96 (8.82). N,N-Di-n-propyl-2-[5-[[(trifluoromethyl)sulfonyl]oxy]-1 H-indol-3- yl]ethylamine oxalate (12). N,N-di-n-propyl-5-hydroxytryptamine (9, 268 mg, 1.03 mmol), Et3N (290 mL, 2.52 mmol) and PhN(SO2CF3)2 (550 mg, 1.55 mmol) were dissolved in CH2Cl2 (10 mL) and stirred at room temperature. After 3 h the mixture was diluted with CH2Cl2 (20 mL) and washed with 10% aqueous Na2CO3 (2 × 25 mL). The aqueous layers were once more extracted with CH2Cl2 (30 mL) after which the organic layers were pooled, washed with brine and dried over MgSO4. The solvent was removed in vacuo leaving a yellow oil which was chromatographed (SiO2, eluting with

CH2Cl2/MeOH (5:1). Pure fractions were pooled and evaporated to dryness yielding a

78 5-HT1D Receptor Agonist Properties of Novel 5-[[(Trifluoromethyl)sulfonyl]oxy]indolylethylamines pale yellow oil (483 mg). The residual oil was converted to the oxalate and recrystallized from MeOH/Et2O giving white crystals (251 mg, 51%): mp 148-150 °C; -1 1 IR (KBr) cm 1225, 1396 (O-SO2); H NMR (base) δ 0.92 (t, J = 7.36, 6H), 1.46-1.65 (m,

4H), 2.46-2.61 (m, 4H), 2.76-2.97 (m, 4H), 7.04 (dd, J1 = 8.83, J2 = 2.37, 1H), 7.09 (s, 1H), 7.34 (d, J = 8.83, 1H), 7.46 (d, J = 2.37, 1H), 8.93 (br s, NH); 13C NMR (base) δ

11.8, 19.6, 22.2, 54.2, 55.9, 111.1, 112.2, 114.6, 114.9, 118.8 (q, J = 321, CF3), 124.4, +1 127.7, 135.0, 143.2; MS (CI with NH3) m/e 393 (M ); Anal Calcd (Obsd) for

C17H23N2O3SF3.C2H2O4: C: 47.30 (47.20), H: 5.22 (5.19), N: 5.81 (5.61). N,N-Diethyl-2-[5-[[(trifluoromethyl)sulfonyl]oxy]-1 H-indol-3-yl]ethylamine oxalate (13). Triflation of the free base of 10 (200 mg, 0.86 mmol) was performed according to the procedure given for the synthesis of 12 yielding an oil (317 mg, quant.) after chromatography (SiO2, eluting with CH2Cl2/MeOH (5:1)). This oil was converted to the oxalic acid salt and recrystallized from MeOH/Et2O giving white -1 1 crystals (247 mg, 63%): mp 142-145 °C; IR (KBr) cm 1221, 1415 (O-SO2); H NMR

(base) δ 1.16 (t, J = 7.27, 6H), 2.84 (q, J = 7.26, 4H), 2.93 (s, 4H), 6.96 (dd, J1 = 8.97, J2 = 2.14, 1H), 7.05 (s, 1H), 7.31 (d, J = 8.97, 1H), 7.43 (d, J = 2.13, 1H), 9.80 (br s, NH); 13C NMR (base) δ 10.2, 21.5, 46.8, 52.6, 110.7, 112.5, 112.7, 114.7, 118.8 (q, J = 321, +1 CF3), 125.0, 127.3, 135.2, 143.2; MS (CI with NH3) m/e 365 (M ); Anal Calcd (Obsd) for C15H19N2O3SF3.C2H2O4: C: 44.93 (44.60), H: 4.66 (4.65), N: 6.16 (6.15). N,N-Dimethyl-2-[5-[[(trifluoromethyl)sulfonyl]oxy]-1 H-indol-3- yl]ethylamine oxalate (14 from 11). Triflation of the free base of 11 (257 mg, 1.26 mmol) was performed as above, yielding a colorless oil (394 mg, 93%) after column chromatography (SiO2, eluting with CH2Cl2/MeOH (5:1). This oil (369 mg) was converted to its oxalate salt with oxalic acid and recrystallized from MeOH/Et2O yielding white crystals (189 mg, 38%): mp 176-177 °C; IR (KBr) cm-1 1221, 1415 (O- 1 SO2); H NMR (CD3OD) δ 2.64 (s, 6H), 3.05 (s, 4H), 7.07 (d, J = 8.78, 1H), 7.33 (s, 1H), 13 7.47 (d, J = 8.79, 1H), 7.58 (s, 1H); C NMR (CD3OD) δ 22.4, 44.1, 59.6, 111.4, 112.3,

113.4, 115.2, 120.0 (q, J = 321, CF3), 126.6, 128.2, 136.6, 144.3; MS (CI with NH3) m/e +1 337 (M ); Anal Calcd for C13H15N2O3SF3.C2H2O4: C: 42.25 (41.99), H: 4.02 (3.89), N: 6.57 (6.35). N,N-phthalimido-2-(5-hydroxy-1 H-indol-3-yl)ethylamine (16). A stirred solution of serotonin creatine sulphate monohydrate (15, 5.00 g, 12.35 mmol) in H2O

(20 mL) and THF (20 mL) was basified until pH 8 with 10% NaHCO3 after which N- carbethoxyphthalimide (2.75 g, 12.35 mmol) was added. The reaction mixture was stirred 8 h. during which time a bright yellow precipitate had formed. The organic solvent was evaporated in vacuo and the yellow solid (3.52 g, 94%) was collected on a sintered glass funnel (P3) and rinsed with Et2O. Recrystallization from EtOH (abs.) gave

79 Chapter 4

3.01 g (80%) of yellow crystals: mp 213-216 °C (lit. 210 °C)25; IR (KBr) cm-1 1690 (C=O), 3370 (OH); 1H NMR (acetone-d6) δ 3.06 (t, J = 8.06, 2H), 3.92 (t, J = 8.06, 2H), 6.71 (d, J = 8.42, 1H), 7.18 (d, J = 8.79, 1H), 7.10 (d, J = 8.42, 2H), 7.53 (br s, 1H), 7.81 (m, 4H), 9.63 (br s, NH); 13C NMR (acetone-d6) δ 25.2, 39.3, 103.4, 103.5, 111.7, 112.6, +1 123.7, 124.2, 129.3, 133.3, 134.9, 151.7, 168.7, 206.0; MS (CI with NH3) m/e 393 (M ). N,N-Phthalimido-2-[5-[[(trifluoromethyl)sulfonyl]oxy]-1 H-indol-3- yl)ethylamine (17). Triflation of 16 (1.33 g, 4.35 mmol) was performed according to the procedure given for the synthesis of 12 affording a white solid after chromatography (SiO2, eluting with CH2Cl2). The solid was recrystallized from EtOH yielding colorless needles (1.48 g, 78%): mp 165-166 °C; IR (KBr) cm-1 1206, 1398 (O- 1 SO2), 1719 (C=O); H NMR δ 3.11 (t, J = 7.69, 2H), 3.97 (t, J = 7.69, 2H), 7.04 (d, J = 8.79, 1H), 7.13 (s, 1H), 7.30, (d, J = 8.78, 1H), 7.55 (s, 1H), 7.67-7.81 (m, 4H), 8.38 (br s, 13 NH); C NMR δ 21.6, 35.7, 108.7, 109.6, 110.7, 112.8, 116.3 (q, J = 321, CF3), 120.7, 122.0, 125.2, 129.5, 131.4, 132.4, 141.0, 165.6; MS (EIPI) m/e 438 (M+); Anal Calcd

(Obsd) for C19H13N2O5SF3: C: 52.06 (51.94), H: 2.99 (3.09), N: 6.39 (6.27). 2-[5-[[(Trifluoromethyl)sulfonyl]oxy]-1 H-indol-3-yl]ethylamine oxalate (18). The N,N-phthalimide 17 (1.07 g, 2.44 mmol) was dissolved in absolute EtOH (50 mL) after which hydrazine hydrate (2.0 mL) was added. The reaction mixture was stirred for 0.5 h at room temperature after which time the volatiles were removed in vacuo. The residue was refluxed in CHCl3 for 0.5 h, cooled to ambient temperature and filtered in order to remove the solid phthalimidohydrazine. The filtrate was evaporated in vacuo leaving a colorless oil which was converted to the oxalate. The oxalate salt was recrystallized from EtOH/Et2O giving 0.86 g (89%) of white crystals: mp 166-167 -1 1 °C; IR (KBr) cm 1210, 1412 (O-SO2), H NMR (base) δ 3.10-3.26 (m, 4H), 7.08 (dd, J1 13 = 8.79, J2 = 2.19, 1H), 7.36 (s, 1H), 7.48 (d, J = 8.79, 1H), 7.57 (d, J = 2.19, 1H); C

NMR (base) δ 24.9, 41.1, 111.6, 111.7, 113.8, 115.6, 120.3 (q, J = 320, CF3), 127.4, 128.4, 137.1, 144.7, 166.7, 194.4; MS (EIPI) m/e 308 (M+); Anal Calcd (Obsd) for

C11H11N2O3SF3.C2H2O4.H2O: C: 37.50 (37.80), H: 3.63 (3.38), N: 6.73 (6.96). N,N-Dimethyl-2-[5-[[(trifluoromethyl)sulfonyl]oxy]-1 H-indol-3- yl]ethylamine oxalate (14 from 18). To a magnetically stirred solution of 18 (0.46 g,

1.5 mmol), 37% formaldehyde (aq., 1.2 mL) in acetonitrile (6 mL), was added NaCNBH3 (0.29 g, 4.5 mmol). The mixture was acidified until pH 5 with acetic acid and stirring continued for 3 h. 10% NaOH (30 mL) was added after which the aquous layer was extracted with CH2Cl2 (3 × 20 mL). The organic layers were combined and dried over

MgSO4. Filtration and removal of the solvent in vacuo gave an oil, which was subjected to column chromatography (SiO2, eluting with CH2Cl2/MeOH (5:1) affording 193 mg (38%) of a colorless oil. Recrystallization of the oxalate salt from MeOH/ether afforded

80 5-HT1D Receptor Agonist Properties of Novel 5-[[(Trifluoromethyl)sulfonyl]oxy]indolylethylamines

164 mg (26%) of white crystals: mp 176-177 °C; all further spectroscopic data were analoguous to that of 14 prepared from 11.

Pharmaco logy. Animals. Adult male albino rats of a Wistar-derived strain (Harlan, Zeist, The Netherlands) weighing 275-325 g were used. Until experiments, the rats were housed in groups of six animals in plastic cages under conditions of constant temperature (20 °C) and humidity with lights on 06:30 and lights off 17:00. Food and water was available ad libitum. Animal procedures were conducted in accordance with guidelines published in the NIH Guide for the Care and Use of Laboratory Animals and all protocols were approved by the Groningen University Institutional Animal Care and Use Committee. Dunkin-Hartley guinea pigs were ordered in from Kuipers Rabbit Ranch (Gary, IN, USA) with a weight range of 225-275 g approximately one week before testing. On arrival the animals were placed in standard guinea-pig cages, six to seven animals per cage. The ambient temperature of the housing room and the testing room is 22.2 ± 2.0 °C. The humidity is kept at 45-55 percent and a 12-h light-dark regimen is employed (lights on between 06:00-18:00). Materials. Sumatriptan was obtained from Glaxo (UK) and serotonin, forskolin and reserpine were purchased from RBI (Natick, MA). The RIA kit for the cAMP assay was purchased from Biomedical Technologies (Stroughton, MA). The radioligand and other compounds were obtained from the following sources: [3H]5-HT (28.2 Ci/mmol) from New England Nuclear (Boston, MA) 5-HT, pargyline and Tris-HCl from Sigma Chemical co. (St.Louis, MO); STV, DME H-21, gentamicin from UCSF (San Francisco, CA); geneticin (G-418 sulfate), fetal bovine serum, penicillin, streptomycin and HAMS #l2 DMEM (50:50) from GIBCO Laboratories (Grand Island, NY); ascorbic acid from Mallinckrodt Inc. (Paris, KY). All substances to be tested in rats were dissolved in saline (0.9% NaCl in distilled water) and administered sc in a volume of 1.0 mL/kg. In guinea-pig experiments, the compounds are made up in a 0.25% methyl cellulose in water solution with the addition of equimolar amounts of citric acid in case the test compound was a free base. The volume of injection is 5 mL/kg for all injections. The guinea pigs that are dosed subcutaneous are injected with a Becton Dickinson 3 mL syringe with a 25 gage 5/8th inch Precision Glide needle both being disposable. Receptor binding assay. Method A. Competition radioligand binding experiments employed 11 drug concentrations run in duplicate. Radioligands used were 3 3 [ H]8-OH-DPAT (5-HT1A, 85 Ci/mmol, 1.2 nM), [ H]5-HT (5-HT1Dα and 5-HT1Dβ, 85 3 Ci/mmol, 2.6 nM) and [ H]U-86170 (D2-dopamine, 62 Ci/mmol, 2 nM). Non-specific binding (75-95% of total) was defined with the following cold compounds added in

81 Chapter 4

excess: lisuride (5-HT1A), serotonin (5-HT1D), and haloperidol (D2). Total binding was determined with buffer. Buffers (pH 7.4) used were 50 mM TRIS, 5 mM MgCl2 (5-

HT1A), the same with 0.1% ascorbic acid (5-HT1D) and 20 mM HEPES, 10 mM MgSO4

(D2). Cloned human receptors permanently expressed in CHO cells were the source of the 5-HT binding sites, except for the dopamine D2 receptor which was cloned from the rat.27,35,36 Binding mixtures were made in 96-deep well titer dishes by the addition of 50 µL of drug, 50 µL of radioligand and 800 µL of membranes (20-60 µg protein) in binding buffer. After room temperature incubation for 1 h (5-HT1D reactions were protected from light), reaction were stopped by vacuum filtration with a TomTec harvester. Counting was with a 1205 Betaplate using Meltilex as scintillant. IC50 values were estimated by fitting the data to a one-site competition model: Y = T/(1 + 10logX-logIC50) where Y is the specific CPM bound at the concentration X, and T is the specific bound CPM in the absence of the competitor. Inhibition constants (Ki) were calculated with the Cheng-Prushoff equation.37

Method B. Human 5-HT1Dα and 5-HTlDβ receptor clones were expressed in a human embryonic kidney 293 (HEK 293) cell line.38,39 The HEK 293 cells were grown as a monolayer in 10 mL HAMS #12 Dulbecco's Modified Eagle Medium (50:50) supple- mented with 10% fetal bovine serum, penicillin G (100 U/mL) and streptomycin (10 mg/mL). Confluent monolayers of the cell lines were harvested (PBS containing 5 mM EDTA) and centrifuged at 480 × g for 10 minutes at 40 °C. The cells were lysed in ice-cold buffer (50 mM Tris-HCl, pH 7.4 containing 5 mM EDTA), homogenized and sonicated for 10 s. Nuclei and intact cells were removed by centrifugation at 1000 × g for 10 min. The supernatant was spun at 35,000 × g for 30 min and the pellet, containing the microsomal membrane fraction, was resuspended binding buffer containing 50 mM

Tris-HCl, 4 mM CaCl2, 0.1 % ascorbic acid, 10 mM pargyline and 1 mM leupeptine. The microsomal membrane suspension was stored at −70 °C. Radioligand binding assays consisted of 0.1 mL of radioligand (final concentrations: [3H]-5HT, 0.01-150 nM, 0.8 mL of tissue suspension (50 mg protein) and 0.1 mL of assay buffer or displacing drug. All drugs were diluted in assay buffer. After an incubation of 30 minutes at 25 °C, assay mixtures were rapidly filtered through 132 glass fiber filters (Schleicher and Schuell; Keene, NH) and washed 2 times with 5 mL of 50 mM Tris-HCl buffer (pH 7.8). The filters were transferred to plastic counting vials and radioactivity was measured by liquid scintillation spectroscopy in 2.5 mL of Bio-Safe II Scintillation Cocktail (Research Products International Corp.; Mount Prospect, IL). Specific binding was defined as the excess over blanks taken in the presence of 10-5 M 5-HT. Radioligand binding data were analyzed by the EBDA40 and

82 5-HT1D Receptor Agonist Properties of Novel 5-[[(Trifluoromethyl)sulfonyl]oxy]indolylethylamines

LIGAND41 programs that utilize the non-linear least squares curve fitting technique with the Marquardt-Levenberg modification of the Gauss-Newton method. Forskolin stimulated cAMP-inhibition. The funcional cAMP assay using the 26,27 cloned human 5-HT1Dα or 5-HT1Dβ receptors was employed as previously described. Briefly, confluent cells were pre-incubated with α-MEM 10 mM/HEPES/1 mM IBMX for 10 min, then stimulated with 25 mM forskolin with or without test drug (either 1 µM 5-HT or 10 µM test compound) for 20 min. The reaction was quenched with TCA and an aliquot assayed for cAMP using a RIA kit. The results were expressed as pmol cAMP/well (n=3) and agonist inhibition calculated as a percent of 5-HT response. 5-HTP Measurements. 30 Rats were reserpinized (5 mg/kg) 24 h prior to administration of the test compound. The test compound was administered 15 min prior to the administration of 100 mg/kg NSD. After 30 min the rats were decapitated and the brain quickly dissected on ice. Striatum, accumbens, frontal cortex and hippocampus were stored at −70 °C until analysis. Before analysis the samples were homogenized in TCA and centrifuged. The supernatant was analyzed by means of HPLC with electrochemical detection for 5-HTP and L-DOPA (not shown). 5-HTP accumulation was expressed as [5-HTP] in µgram/ gram wet tissue. Surgery and Microdialysis Experiments. The microdialysis probes that were used were of a ventrical, concentric design.42 The exposed tip of the dialysis membrane was 4 mm. The dialysis tube (ID: 0.22 mm; OD: 0.31 mm) was prepared from polyacrylonitrile/sodium methyl sulfonate copolymer (AN 69, Hospal, Bologna, Italy). The microdialysis probes were implanted under chloral hydrate anesthesia (400 mg/kg ip) at the following coordinates: AP −5.2, ML ± 4.8 relative to bregma, and V −8.0 below dura (hippocampus). During surgery, lidocaine HCl salt (6% in saline, brought to pH 6.0 with 1 N NaOH) was used as an adjuvant local anesthesic. Probes were secured to the skull with two set-screws and fast-securing dental cement. Microdialysis experiments were carried out 24-48 h after implantation of the probe. Samples were collected on-line every 15 min in a 20-µL sample loop of an HPLC system. In brief, the inlet of the microdialysis probe was connected to a piece of polyethylene tubing (450 × 0.28 mm) whereas the outlet of the microdialysis probe was connected to a piece of peek tubing (450 × 0.12 mm). The inlet tube was connected to the perfusion pump, and the peek tube directly into the injection valve of the HPLC apparatus. The connection with the HPLC equipment introduced a lag time of about 8 min, for which the presented data are corrected. With the help of an electronic timer, the injection valve was held in the load position for 15 min, which was the time needed to record a complete chromatogram. The perfusion was carried out with an artificial cerebrospinal fluid (aCSF) solution at a flow rate of 1.5 µL/min using Carnegie CMA (Stockholm, Sweden)

83 Chapter 4 perfusion pump. The composition of the aCSF solution was (in mM): NaCl, 147.0; KCl,

4.0; CaCl2, 1.2; and MgCl2, 1.0. After finishing the experiment, the rat was terminated with an overdose pentothal and the brain was fixed with 4% paraformaldehyde via intracardiac perfusion. Coronal sections (40 µm thick) were cut, and dialysis probe placement verified with the help of the atlas of Paxinos and Watson.43 Analysis of the Dialysates. 5-HT was quantified by HPLC with electrochemical detection. A Shimadzu LC10-AD pump was used in conjunction with an electrochemical detector (Antec, Leiden, The Netherlands) set at 650 mV vs an Ag/AgCl reference electrode. A reversed-phase C18 Supelco LC18DB column (150 × 4.6 mm; 5 µm) was used. The mobile phase consisted of an aquous solution of 2.0 g/L of citric acid, 5.0 g/L of sodium acetate, 100 mg/L of EDTA, 300 mg/L of TMA, 300 mg/L of MSA, 10% methanol (v/v) and 5% acetronitile (v/v) and was delivered at a flow rate of 1 mL/min. 5-HT eluted after 8 min. Guinea-pig Brain Neurochemistry. Male Dunkin-Hartley guinea pigs were decapitated 60 min after test drug administration by means of guillotine. Their brains were rapidly removed and put on an ice-chilled petri dish. The prefrontal cortex and the hypothalamus were dissected and the tissue parts were stored at −80 °C until further analysis. The levels of 5-HT and 5-HIAA were measured by means of HPLC with electrochemical detection according to methods described in the literature with minor modifications.30 Hypothermia. At least one hour before animals are to be tested they are removed from the gang cages and placed in individual plastic cages (20 × 30 × 15 cm) and then taken to the testing room. The guinea pigs are tested in groups of five animals per group. The zero rectal temperature is taken using a Digital Thermometer VWR Scientific Inc. with a range of −40 degrees to a 300 degrees Fahrenheit, or −40 to a 150 °C. The probe used in this study is a Yellow Springs Instruments 423-series probe. The probe was lubricated with a drop of silicon and then inserted 3-4 centimeters in the rectum and left until the reading on the Digital Thermometer is stable, usually around 10 seconds. The temperature is recorded to the nearest 1/10th °C. This is the control measurement for the other time intervals used which are 30, 60 and 120 minutes after dosing. In the present study we present data from the 60-minute readings. Expression of Results and Statistics. Differences between 5-HTP concentration of the control- and drug-treatment were analyzed with one-way ANOVA followed by a post-hoc t-test. In microdialysis, the average of the last four stable samples (less than 20% variation) before the drug-treatment was considered as the control value and was defined as 100%. Values given are expressed as percentages of controls. Differences between the average dialysate concentrations of the control- and drug-treatment were

84 5-HT1D Receptor Agonist Properties of Novel 5-[[(Trifluoromethyl)sulfonyl]oxy]indolylethylamines compared with Friedman’s one-way ANOVA with repeated measures (p ≤ 0.05) followed by Dunn’s post-hoc test. The data from the guinea-pig brain neurochemistry are expressed as mean ± SEM (n = 4-5) % of vehicle treated controls. Statistical analyis was performed by means of ANOVA followed by Fishers’ PLSD. In the hypothermia test, the mean difference, the SEM, and the probablility value between the zero minute control and the sixty minute were calculated via the RS-1 statistical program.

85 Chapter 4

4.6 References

[1] Heuring, R.E.; Peroutka, S.J. J. Neurosci. 1987, 7, 894. [2] Waeber, C.; Schoeffter, P.; Palcios, J.M.; Hoyer, D.Naunyn-Schmiedeberg’s Arch. Pharmacol. 1988, 337, 595. [3] Hoyer, D.; Middlemiss, D.N. Trends. Pharmacol. Sci. 1989, 10, 130. [4] Maura, G.; Thellung, S.; Andreoli, G. C.; Ruelle, A.; Raiteri, M. J. Neurochem. 1993, 60, 1179. [5] Hartig, P.R.; Branchek, T.A.; Weinshank, R.L.Trends Pharmacol. Sci. 1992, 13, 152. [6] Weinshank, R.L.; Zgombick, J.M.; Macchi, M.J.; Branchek, T.A.; Hartig, P.R.Proc. Natl. Acad. Sci. U. S. A. 1992, 89, 3630. [7] Hoyer, D.; Clarke, D.E.; Fozard, J.R.; Hartig, P.; Martin, G.R.; Mylecharane, E.J.; Saxena, P.R.; Humprey, P.P.A. Pharmacol. Rev. 1994, 46, 157. [8] Hamel, E.; Fan, E.; Linville, D.; Ting, V.; Villemure, J.G.; Chia, L.S.Mol. Pharmacol. 1993, 44, 242. [9] Hamel, E.; Gregoire, L.; Lau, B. Eur. J. Pharmacol. 1993, 242, 75. [10] (a) Ferrari, M.D.; Saxena, P.R. Trends Pharmacol. Sci. 1993, 14, 129. (b) Peroutka, S.J.; McCarthy, B.G. Eur. J. Pharmacol. 1989, 163, 133. [11] Street, L.J.; Baker, R.; Davey, W.B.; Guiblin, A.R.; Jelley, R.A.; Reeve, A.J.; Routledge, H.; Sternfeld, F.; Watt, A.P.; Beer, M.S.; Middlemiss, D.N.; Noble, A.J.; Stanton, J.A.; Scholey, K.; Hargreaves, R.J.; Sohal, B.; Graham, M.I.; Matassa, V.G.J. Med. Chem. 1995, 38, 1799. [12] Glen, R.C.; Martin, G.R.; Hill, A.P.; Hyde, R.M.; Woollard, P.M.; Salmon, J.A.; Buckingham, J.; Robertson, A.D. J. Med. Chem. 1995, 38, 3566. [13] For a review see: Ferrari, M.D.; Saxena, P.M. Eur. J. Neurol. 1995, 2, 5. [14] Clitherow, J.W.; Scopes, D.I.C.; Skingle, M.; Jordan, C.C.; Feniuk, W.; Campbell, I.B.; Carter, M.C.; Collington, E.W.; Connor, H.E.; Higgins, G.A.; Beattie, D.; Kelly, H.A.; Mitchell, W.L.; Oxford, A.W.; Wadsworth, A.H.; Tyers, M.B.J. Med. Chem. 1994, 37, 2253. [15] Walsh, D.H.; Beattie, D.T.; Connor, H. E. Eur. J. Pharmacol. 1995, 287, 79. [16] Clitherow, J.W.; Scopes, D.I.C.; Beattie, D.T.; Skingle, M.Exp. Opin. Invest. Drugs. 1995, 4, 323. [17] Glennon, R.A.; Ismaiel, A.M.; Chaurasia, C.; Titeler, M.Drug Dev. Res. 1991, 22, 25. [18] (a) Castro. J.L.; Baker, R.; Guiblin, A.R.; Hobbs, S.C.; Jenkins, M.R.; Russell, M.G.N.; Beer, M.S.; Stanton, J.A.; Scholey, K.; Hargreaves, R.J.; Graham, M.I.; Matassa, V.G.J. Med. Chem. 1994, 37, 3023. (b) Perez, M.; Fourrier, C.; Sigogneau, I.; Pauwels, P.J.; Palmier, C.; John, G.W.; Valentin, J.-P.; Halazy, S. J.Med. Chem. 1995, 38, 3602. [19] (a) Macor, J.E.; Blank, D.H.; Post, R.J.; Ryan, K.Tetrahedron Lett. 1992, 33, 8011. (b) King, F.D.; Brown, A.M.; Gaster, L.M.; Kaumann, A.J.; Medhurst, A.D.; Parker, S.G.; Parsons, A.A.; Patch, T.L.; Raval, P. J. Med. Chem. 1993, 36, 1918. (c) Macor, J.E.; Blank, D.H.; Fox, C.B.; Lebel, L.A.; Newman, M.E.; Post, R.J.; Ryan, K.; Schmidt, A.W.; Schulz, D.W.; Koe, B.K.J. Med. Chem. 1994, 37, 2509. [20] (a) Sonesson, C.; Boije, M.; Svensson, K.; Ekman, A.; Carlsson.; Romero, A.G.; Martin, I.J.; Duncan, J.N.; King, L.J.; Wikström, H. J. Med. Chem. 1993, 36, 3409. (b) Sonesson, C.; Barf, T.; Nilsson, J.; Dijkstra, D.; Carlsson, A.; Svensson, K.; Smith, M.W.; Martin, I.J.; Duncan, J.N.; King, L.J.; Wikström, H. J. Med. Chem. 1995, 38, 1319. [21] (a) Liu, Y.; Yu, H.; Svensson, B.E.; Cortizo, L.; Lewander, T.; Hacksell, U.J. Med. Chem. 1993, 36, 4221. (b) Liu, Y.; Yu, H.; Mohell, N.; Nordvall, G.; Lewander, T.; Hacksell, U.J. Med. Chem. 1995, 38, 150. [22] Sonesson, C.; Lin, C.-H.; Hansson, L.; Waters, N.; Svensson, K.; Carlsson, A.; Smith, M.W.; Wikström, H. J. Med. Chem. 1994, 37, 2735. [23] (a) Speeter, M.E.; Anthony, W.C. J. Am. Chem. Soc. 1954, 76, 6208. (b) Kondo, H.; Kataoka, H.; Hayashi, Y.; Dodo, T.C.A . 1960, 54, 492. [24] Hendrickson, J.B.; Bergeron, R. Tetrahedron Lett., 1973, 4607. [25] De Silva, S.O.; Snieckus, V.Can. J. Chem. 1978, 56, 1621. [26] Kukolja, S.; Lammert, S.R. J. Am. Chem. Soc. 1975, 97, 5582. [27] Veldman, S.A.; Bienkowski, M.J.Mol. Pharmacol. 1992, 42, 439. [28] McCall, R.B.; Romero, A.G.; Bienkowksi, M.J.; Harris, D.W.; McGuire, J C.; Piercey, M F.; Shuck, M.E.; Smith, M W.; Svensson, K.A.; Schreur, P.J.K.D.; Carlsson, A.; VonVoightlander, P.F.J. Pharmacol. Exp. Ther. 1994, 271, 875.

86 5-HT1D Receptor Agonist Properties of Novel 5-[[(Trifluoromethyl)sulfonyl]oxy]indolylethylamines

[29] Dourish, C.T.; Hutson, P.H.; Curzon, G. Trends Pharmacol. Sci. 1986, 7, 212. [30] Shum, A.; Sole, M.J.; van Loon, G.R. J. Chromatogr. 1982, 228, 123. [31] (a) Skingle, M.; Higgins, G.A.; Feniuk, W.J. J. Psychopharmacol. 1995, 8, 14. (b) Kalkman, H.O.; Neumann, V. Eur. J. Pharmacol. 1995, 285, 313. [32] Glennon, R.A.; Hong, S.-S.; Dukat, M.; Teitler, M.; Davis, K.J .Med. Chem. 1994, 37, 2828. [33] Saxena, P R.; De Vries, P.; Heiligers, J P C.; Maassen-Van Den Brink, A.; Barf, T.; Wikström, H.Eur. J. Pharmacol. In press. [34] Pallas version 1.2 is commercially available software of CompuDrug Chemistry Ltd. (c) (1994). [35] Chio, C.L.; Hess, G.F.; Graham, R.S.; Huff, R.MNature 1990, 343, 266. [36] Fargin, A.; Raymond, J.R.; Lohse, M.J.; Kobilka, B.K.; Caron, M.G.; Lefkowitz, R.J.Nature 1988, 335, 358. [37] Cheng, Y.C.; Prushoff, W.H.;Biochem. Pharmacol. 1973, 22, 3099. [38] Oksenberg, D.; Marsters, S.A.; O’Dowd, B.F.; Jin, H.; Havlik, S.; Peroutka, S.J.; Ashkenazi.Nature , 1992, 360, 161. [39] Jin, H.; Oksenberg, D.; Ashkenazi, A.; Peroutka, S.J.; Duncan, A.M.V.; Rozmahel, R.; Yang, Y.; Mengod, G.; Palacios, J.M.; O’Dowd, B.F.J. Biol. Chem. 1992, 267, 5735. [40] McPherson, G.A. Comput. Prog. Biomed. 1983, 17, 107. [41] Munson, P.J.; Rodbard, D. Anal. Biochem. 1980, 107, 220. [42] Santiago, M.; Westerink, B.H.C. J. Neurochem. 1990, 55, 169. [43] Paxinos, G.; Watson, C.Rat Brain in Stereotaxic Coordinates, Academic press,New York. 1982.

87 Chapter 5

5-(Sulfonyl)oxy-tryptamines and Ethylamino Side Chain Restricted Derivatives

Abstract

A number of 5-substituted sulfonic acid ester derivatives of 5- hydroxytryptamine (5-HT) were prepared and their affinities are compared to that of the reference compound 5-OSO2CF3-tryptamine (6). The structure-affinity relationships

(SAFIR) are discussed in terms of in vitro binding for cloned human 5-HT1A, 5-HT1Dα and 5-HT1Dβ receptors. The 5-tosylated tryptamine (9) exhibited the best profile for 5-

HT1Dα receptors (Ki = 4.8 nM) but still showed a comparatively lower affinity than compound 6. Other tryptamine derivatives displayed moderate binding to 5-HT1A and 5- − HT1Dβ receptors, along with Ki values ranging from 9 22 nM for the 5-HT1Dα sites. In addition, the syntheses of two ethylamino side chain restricted derivatives are described. The 6-triflated 3-aminotetrahydrocarbazole 19, as well as the 5-triflated indolepiperidines 22 and 23 induced a shift in affinity in favor of the 5-HT1Dβ receptors. The relatively longer N-O distance of 19, 22 and 23 as compared to tryptamines or 2- aminotetralins, is likely responsible for this observation.

5.1 Introduction

The electron withdrawing aryl triflate group was previously shown to be a group 1 which (i) improves the pharmacokinetic properties of 5-HT1A receptor ligands and (ii) 2 enhances the affinity for 5-HT1D receptor ligands, compared to the hydroxy analogues.

The latter improvement seemed to be more pronounced at 5-HT1Dα receptors relative to

5-HT1Dβ sites (see Chapter 2, Table 2.3 and Chapter 4, Table 4.1B). The question arises whether the triflate group is the optimal sulfonic acid ester for 5-HT1D affinity.

Especially, the active site of both 5-HT1D receptor subtypes are known to contain a pocket which can accommodate large groups, located at the 5-position of indolealkylamines.3 The nature of this pocket may be explored by using sulfonate substituents with different electronic and steric properties. Thus, it was of interest to examine the effects of readily available sulfonic acid ester derivatives of 5-HT on the selectivity and affinity for the 5-HT1A and 5-HT1D receptors.

85 Chapter 5

H

H2N O N S Me O O OTf Me H Me N N Me NH N 2 N H H (±)-cis-1 (±)-2 Naratriptan,3

Another approach which may improve the 5-HT1D versus 5-HT1A receptor affinity is rigidification of the tryptamine moiety. From the data presented in Chapters 2 and 4 it seems that ethylamino side chain restriction in the 2-aminotetralins produces marked effects on selectivity and affinity at 5-HT1 receptor subtypes. Recently, the enantiomers of 8-OH-DPAT were reported to have nanomolar affinities for the 5-HT1D receptor 4 subtypes. The affinity and agonist efficacy for the 5-HT1Dα and 5-HT1Dβ receptor were shown to reside in the R-enantiomer (Ki: 28.8 nM; EC50: 30 nM and Ki: 75.5; EC50: 415 nM, respectively). (±)-Cis-8-[[(trifluoromethyl)sulfonyl]oxy]-1-Methyl-2-

(methylamino)tetralin (cis-1) displayed Ki values of 3.4 and 10 nM for 5-HT1Dα and 5-

HT1Dβ receptors, respectively (Chapter 2). The antipodes of the monopropyl analogue of cis-1, like the enantiomers of 8-OH-DPAT, exhibited marked stereoselectivity for the 5-

HT1D sites. Obviously, low nanomolar affinities of 2-aminotetralins for the 5-HT1D receptor subtypes are feasible when proper substituents are employed and, moreover, orientation A of 5-HT seems to be the active conformation (Figure 5.1).

OH OH 8 1 NR2 7 2 NH2 6 3 5 4 N H OrientationA 2-aminotetralins

OH OH 6 7 5

8 4

NH2 3 N 9 N NR2 H H 1 2 OrientationB 3-aminocarbazoles

86 5-(Sulfonyl)oxy-tryptamines and Ethylamino Side Chain Restricted Derivatives

Figure 5.1. Ethylamino side chain orientations of 5-HT Orientation B of 5-HT is captured in the 3-aminocarbazole skeleton. King and co-workers reported aminocarbazole derivative (2) to have a Ki of 10 nM and high 5,* intrinsic activity for 5-HT1D receptors. This finding suggests that the binding conformation of the side chain of 5-CT at the 5-HT1D receptor may approximate orientation B. However, in the pharmacology experiments, the authors did not discriminate between the two 5-HT1D receptor subtypes, which makes it difficult to draw conclusions regarding the orientation of the ethyl amino group in each of these subtypes. Another type of indolealkylamine is represented by the semi-rigid Naratriptan (GR85548A, 3),6 which possesses a piperidine ring instead of an ethylamino side chain. Naratriptan is reported to be clinically effective in the treatment of migraine, with superior potency as compared to sumatriptan (24) in binding and functional studies.7

This compound showed a Ki of approximately 8 nM for 5-HT1D receptors, measured in guinea pig striatal membranes. In the present study we describe the synthesis and SAR of sulfonic acid ester substituted tryptamines. In addition, the influence of rigidification of tryptamine analogues on selectivity for the 5-HT1A and 5-HT1D receptor subtypes is examined, which is the second objective of this chapter. Throughout this series of compounds, the aryl triflate group is used as the reference substituent.

5.2 Chemistry

Preparation of 5-Sulfonyloxytryptamines. The sulfonic acid ester derivatives were prepared in moderate to high yields by treating N,N-phthalimido protected 5-HT (see Section 4.5) with the appropriate sulfonyl or sulfamoyl chloride. The coupling was effected using; Et3N as a base and dioxane as the solvent (method A); phase-transfer conditions with tetrabutyl ammonium iodide as the phase-transfer catalyst (method B); or NaH as the base and DMF as the solvent (method C). The phthalimides were converted into the primary amines upon treatment with hydrazine in ethanol (Scheme 5.1).

* Measured in Piglet caudate.

87 Chapter 5

O O OH O O Me S S N O F3C O NH2 Me NH2 NH2 N H 4 N N 12 H 6 H a h b O O O O Me S S N O OH Me O H NH2 g NPhth c NH2

N N N H 11 H 5 H 7 f d

O O e O O S S S O O O O S NH2 O NH2

Me NH N 2 N H 10 H 8 N H 9

Scheme 5.1. (a) N-Et-CO2-phth, 10% NaHCO3 (pH 8), THF/H2O; (b) PhN(SO2CF3)2, Et3N, CH2Cl2; (c) MeSO2Cl, method A; ; (d) PhSO2Cl, method A; (e) p-TolSO2Cl, method A; (f) 2-ThSO2Cl, method B; (g) MeNHSO2Cl, method A; (h) (Me)2NSO2Cl, method C. Steps b-h were succeeded by deprotection with hydrazine in abs. ethanol.

Upon crystallization from H2O/MeOH, the mesylate derivative 7 yielded crystals that were suitable for a single crystal X-ray spectroscopy determination (Figure 5.2). The compound crystallized as the hemi-oxalate in a monoclinic C2/c spacegroup with 8 molecules per unit cell (a = 22.158; b = 5.791; c = 24.172 Å). In Table 5.1, selected bond distances, bond angles and torsional angles are listed. The crystal structure is stabilized by complexation between the NH of the indole portion and the oxalic acid via H-bonds at a distance of 1.834 Å (Figure 5.2B). This distance is comparable with that of the normal ionic interaction of the primary amine and the oxalic acid, being 1.826 Å. Furthermore, clear intermolecular H-bonds (2.034 Å) can be observed between O16 of one molecule and the ethylamino N+H of another molecule. This is also reflected by the

88 5-(Sulfonyl)oxy-tryptamines and Ethylamino Side Chain Restricted Derivatives relative long bond length of S14-O16 (1.426 Å) as compared that of S14-O17 (1.359 Å).

Figure 5.2A. Molecular structure of7 .

Figure 5.2B. Stereoview of the intermolecular interactions of7 (indicated by a dashed line).

89 Chapter 5

16 17 O O 15 S 13 Me 14 O 10 9 11 1 2 NH2 8 12 4 7 3 6 N 5 H

Figure 5.2C. Numbering of the atoms of3 .

Table 5.1. Selected interatomic distances, angles and torsional angles of compound3

Distance (Å) Angle (deg) Torsional Angle (deg)

N1-C2 1.498 N1-C2-C3 110.2 N1-C2-C3-C4 178.4

S14-C15 1.744 C2-C3-C4 110.3 C2-C3-C4-C5 108.9

O13-S14 1.550 C10-O13-S14 120.9 C11-C10-O13- −97.3

C10-O13 1.486 O13-S14-C15 101.1 C10-O13-S14- −84.7

S14-O16 1.426 O13-S14-O16 104.9 C10-O13-S14- 28.2

S14-O17 1.359 O13-S14-O17 114.3 C10-O13-S14- 158.9

OMe OMe NPhth a b + O NPhth 15 NHNH2.HCl N H d 13 14 16

OH OSO2CF3 OSO2CF3

c d

N NPhth N NPhth N NH2 H H H 17 18 19

Scheme 5.2. (a) EtOH, ∆; (b) BBr3, CH2Cl2, –78 °C–rt; (c) PhN(SO2CF3)2, Et3N, CH2Cl2; (d) H2NNH2.H2O, EtOH.

Preparation of 3-Aminocarbazoles. N-Phthalimido protected 4- aminocyclohexanol was oxidized with pyridinium chlorochromate to give the

90 5-(Sulfonyl)oxy-tryptamines and Ethylamino Side Chain Restricted Derivatives cyclohexanone derivative 14. The 3-aminocarbazole skeleton was prepared via the Fischer-indole synthesis by refluxing 14 with p-methoxyphenylhydrazine (13) in ethanol (83%).8 Intermediate 15 was either N-deprotected to give 16, or O-demethylated to give 17 (31%), which subsequently was triflated and converted into the primary amine to yield carbazole derivative 19 employing conditions described as above (Scheme 5.2). Preparation of Indol-3-ylpiperidines. The t-BOC-protected 5-hydroxy- indolepiperidine (20) was triflated in 86% yield as described before and deprotected using TFA in CH2Cl2, resulting in compound 22 (44% after purification). This secondary amine was N-methylated to give 23 (Scheme 5.3).

OH OSO2CF3 OSO2CF3

a b N Z N Z N R

N N N H H H 20 21 22: R = H c 23: R = Me Z = t-BOC

Scheme 5.3. (a) PhN(SO2CF3)2, Et3N, CH2Cl2; (b) TFA, CH2Cl2. 0 °C; (c) 37% Formaldehyde, NaCNBH3, acetic acid (pH 5), acetonitrile.

5.3 Pharmacology

Receptor Binding. The compounds were tested for the inhibition of [3H]8-OH- 3 3 DPAT (5-HT1A) or [ H]5-carboxamidotryptamine ([ H]5-CT) (5-HT1Dα and 5-HT1Dβ) binding to cloned human receptors expressed in Cos-7 cells (Table 5.2).

5.4 Results and Discussion

5-Sulfonic acid ester derivatives of 5-HT. As indicated by the resolved crystal structure and the intermolecular interactions of mesylate derivative 7, aryl sulfonic acid esters are capable of accepting a hydrogen bond (Figure 5.2B). This observation may be of importance with respect to interactions with H-bond donating amino acid residues in a particular receptor. Compound 7 binds with modest affinity to 5-HT1A and 5-HT1D receptors, displaying a slight preference for the 5-HT1Dα subtype (Ki = 18 nM; Table 5.2). Since the steric interactions of 7 and 6 with the receptors are similar, the relatively

91 Chapter 5 lower affinities of 7 have to be discussed in terms of electronic effects, which are commonly expressed as Hammett σp and Taft σI parameters. Both sulfonate groups display an electron withdrawing character, indicated by the positive signs of the σp and

σI values for the mesylate (+0.33 and +0.61, respectively) and the triflate (+0.37 and +0.84, respectively).9

Table 5.2. Affinities at 5-HT1A, 5-HT1Dα, and 5-HT1Dβ, Receptors In Vitro R R R

N R'

N NH2 N NH2 N H H H A B C

Ki (nM)a

5-HT1Dβ/

Compd Type R 5-HT1A 5-HT1Dα 5-HT1Dβ 5-HT1Dα

6 A OSO2CF3 50 2.4 11 4.6

7 A OSO2Me 71.4 18 60 3.3

8 A OSO2Ph 32 22 40 1.8

9 A OSO2(p-Tol) 12.1 4.8 30 6.3

10 A OSO2(2-Th) 34.3 18 20 1.1

11 A OSO2NHMe 64.3 9 18 2

12 A OSO2N(Me)2 47 12 26 2.2 16 B OMe >1000 >1000 600 -

19 B OSO2CF3 >1000 56 40 0.7

22 C OSO2CF3; R’ = H 71.4 24 7 0.3

23 C OSO2CF3; R’ = Me 57.1 14 6 0.4

3 (a) Ki values for displacement of 5-HT1A receptor agonist [ H]8-OH-DPAT and 5-HT1Dα/5-HT1Dβ receptor agonist [3H]5-CT. Data from cloned mammalian receptors expressed in Cos-7 cells. The values were obtained from a single experiment and were generated at Centre de Recherche Pierre Fabre.

The different binding properties may be attributed to the polarizability of the triflate group, which allows participation in hydrogen bonding in the drug-receptor interaction. 9 A phenylsulfonate (σp = +0.33) in this position results in complete loss of selectivity.

Compared to the mesylate, it improves the affinity by two-fold for the 5-HT1A receptor

(Ki = 32 nM) and 1.5-fold for the 5-HT1Dβ (40 nM). Interestingly, a methyl substituent on the para-position of the phenyl ring induces a pronounced increase in affinity for

92 5-(Sulfonyl)oxy-tryptamines and Ethylamino Side Chain Restricted Derivatives

the 5-HT1A (Ki = 12.1 nM) and the 5-HT1Dα receptor (Ki = 4.8 nM), and to a lesser σ σ extend for the 5-HT1Dβ receptor (Ki = 30 nM). The Hammett p and Taft I values of a tosylate group are +0.28 and +0.54 respectively, and thus comparable with those of the other sulfonates.9 The increased affinities suggest that additional favourable drug- receptor interactions may be provided by extension in the direction of para-methyl substituent of compound 9. The 2-thiophene ring has little influence on the affinity or selectivity as compared to the phenyl ring, except for a two-fold increase in affinity for the 5-HT1Dβ receptor (Ki = 20 nM), which indicates that no extra hydrogen bonding interaction in this position is to be expected. However, a thiophene sulfonic acid ester bearing the sulfur atom in the 3-position has not been examined as yet.

O O S MeNH Me N Me

N H 24

Structurally, the sulfamate derivatives 11 and 12 are in close resemblance to sumatriptan (24),10 differing only in the 5-oxygen and the unsubstituted amine functionality. The sulfamoyl moiety has been widely utilized as an activity-modifying substituent in various classes of drugs with therapeutic potential in for instance the treatment of cancer11 or psychosis.12 Compound 11, as well as 12, display a two-fold preference for 5-HT1Dα receptors over 5-HT1Dβ receptors. Both sulfamate substituted tryptamines show a similar binding profile as compared to sumatriptan, having higher affinity for the 5-HT1A site (for comparison see Table 4.2A and Table 5.1). Presumably, the 5-oxygen of 11 and 12, as in most of the other sulfonic acid ester derivatives, participates in hydrogen bond formation with the 5-HT1A receptor. Obviously, the 5- oxygen is much more important for 5-HT1A receptor binding than for the 5-HT1D receptor subtypes, since sumatriptan and other 5-HT1D receptor agonists lack this particular oxygen atom. This suggests that truly selective 5-HT1D receptor ligands are not to be expected when sulfonate substituted serotonin analogues are employed. The homology of the transmembrane spanning regions (TM) between the human

5-HT1A, 5-HT1Dα and 5-HT1Dβ receptors is considerable (ranging from 53-96 %, see Table 1.2). Alignment of the amino acid residues of putative helix 5 shows us that the three proteins bear a serine and threonine residue in a similar position.13 Discriminatory

93 Chapter 5 properties of the receptors for the tryptamine derivatives with relatively small 5- substituents may be accounted for by differential active site-surrounding amino acid residues. In addition, the potential differences in distant helical environments, in other words, the distance between the serine and/or threonine on TM-5 and the aspartate on TM-3, will be of importance. The affinities of conformationally restricted indolealkylamines may provide useful information regarding the size of these distances. Alkylamino Side Chain Restriction. While our work was progress, Glennon and co-workers14 reported the synthesis and binding results of 6-methoxy-3- aminocarbazole 16 for the 5-HT1Dβ receptor population (Ki = 342 nM). In line with their result we found a Ki of 600 nM for this receptor along with Ki values of >1000 nM for the 5-HT1A and 5-HT1Dα sites. These authors also reported Ki values of 5- methoxytryptamine for the 5-HT1A (3.2 nM), 5-HT1Dα (5.4 nM) and 5-HT1Dβ (3.5 nM) receptors. This means that orientation B (Figure 5.1) is most probably not the binding conformation of serotonin at these receptor subtypes. Obviously, the N-O distance of compound 16 is too long for a proper interaction with the aspartate and one of the hydrogen bond donating residues on TM-5. This distance seems to be partially restored in carboxamido derivative 2 (Ki = 10 nM) in case of the 5-HT1D receptors but is still too long for the 5-HT1A receptor. Replacing the 6-carboxamido substituent by a triflate group (19) confirms this hypothesis. Compound 19, like 2, was still inactive at the 5-

HT1A receptor (Ki = >1000 nM) but displays moderate affinities for the 5-HT1Dα (56 nM) and 5-HT1Dβ (40 nM) sites. Taken together, this leads to the assumption that in case of compound 19 one of the sulfonyl oxygens serves as a hydrogen bond acceptor, whereas the ester-oxygen is no longer available for such an interaction, as compared to the non- restricted triflate derivative 6. Noteably, the 3-aminocarbazole derivatives, unlike tryptamines, seem to have a preference for the 5-HT1Dβ receptor. Ofcourse, it will be of interest to prepare the enantiomers of 19 and include them in the SAFIR discussion. The semi-rigid naratriptan derivatives 22 and 23 both show an interesting receptogram. Displaying Ki values of 7 and 6 nM for the 5-HT1Dβ receptor, respectively, these compounds exhibited a clear preference for the 5-HT1Dβ site over the 5-HT1Dα site of about 3-fold. The 5-HT1A receptor affinities, being 71.4 nM for 22 and 57.1 nM for 23, are comparable with that of compound 6, but much higher than that of 19. The 5-

HT1Dβ receptor preference again may be explained by an increased N-O distance relative to 6. Other factors, such as a positive lipophilic interactions of the piperidine ring with a hydrophobic part of the 5-HT1Dβ receptor may also contribute to this observation. The binding dat are comparatively similar to naratriptan, which displayed a Ki of 8 nM for 3 5-HT1D receptors (against [ H]5-HT binding in guinea pig striatal membranes), and an

94 5-(Sulfonyl)oxy-tryptamines and Ethylamino Side Chain Restricted Derivatives

7 IC50 value of approximately 80 nM for 5-HT1A sites. Again, no binding data were provided for 5-HT1Dα and 5-HT1Dβ receptors individually. In summary, it can be concluded that the previously reported 2-aminotetralins

(orientation A) have a marked selectivity for 5-HT1A and 5-HT1Dα sites, whereas 3- aminocarbazole derivatives (orientation B) have a tendency to prefer the 5-HT1Dβ receptor. Likewise, the indolepiperidines display a marked preference for the latter receptor. Thus it seems that the N-O distance, roughly defined by a (semi)-rigid skeleton, primarily determines the receptor selectivity, whereas proper substituent selection will optimize the affinity for each of the receptor subtypes. In our hands, the aryl triflate group seemed to be the optimal sulfonic acid ester, at least for 5-HT1Dα and

5-HT1Dβ receptors, however, the positive contribution to binding of the para-methyl group of tosylate 9 suggests that further extension is possible in this direction, providing a ‘handle’ for future improvements.

5.5 Experimental Section

General. For general remarks see Section 2.4. The chemical ionization (CI) mass spectra were obtained on a Unicam Automass 150 system using a direct-inlet probe. Materials. The synthesis of N,N-phthalimido protected serotonin (5) and compound 6 is described in Chapter 4. N-Boc-4-[(5-hydroxy)-1H-indol-3-yl]piperidine (20) was kindly provided by Merck KGaA (Darmstadt, Germany). N-methylsulfamoyl chloride was prepared according a literature procedure.15 Method A. N,N-Phthalimido-2-[5-[[(methyl)sulfonyl]oxy]-1 H-indol-3- yl]ethylamine. Methanesulfonyl chloride (170 µL, 2.20 mmol) was added dropwise to a solution of 5 (0.56 g, 1.83 mmol) and Et3N (0.5 mL) in dioxane (10 mL). After 2 h of stirring anhydrous ether (30 mL) was added after which the formed precipitate was removed by filtration. The filtrate was evaporated to dryness leaving 0.71 g (100%) of a colorless oil, which was recrystallized from i-PrOH yielding white crystals (0.35 g, 50%): mp 167-168 °C; IR (KBr) cm-1 3046 (NH), 1705 (C=O), 1398, 1360, 1180 (O- 1 SO2); H NMR δ 3.14 (t, J = 7.69, 2H), 3.17 (s, 3H), 3.87 (t, J = 7.79, 2H), 7.16 (m, 2H), 7.34 (d, J = 8.74, 1H), 7.61 (s, 1H), 7.69-7.85 (m, 4H), 8.21 (br s, NH); 13C NMR δ 24.2, 36.8, 38.2, 111.6, 112.1, 113.0, 116.7, 123.2, 124.0, 127.7, 132.0, 133.9, 134.7, 143.2, + 168.3; MS (EIPI) m/e 384 (M ); Anal. Calcd (Obsd) for C19H16N2O5S: C: 59.4 (59.1), H: 4.2 (4.3) N: 7.3 (7.3). Method A. N,N-Phthalimido-2-[5-[[(phenyl)sulfonyl]oxy]-1 H-indol-3- yl]ethylamine. Using benzenesulfonyl chloride afforded a crude yellow solid quantitatively. Recrystallization from acetone gave 0.63 g (86%) of colorless plates: mp

95 Chapter 5

-1 1 166-168 °C; IR (KBr) cm 3423 (NH), 1706 (C=O), 1395, 1356, 1192 (O-SO2); H NMR

δ 2.98 (t, J = 7.74, 2H), 3.85 (t, J = 7.74, 2H), 6.82 (dd, J1 = 8.88, J2 = 2.18, 1H), 7.06 (d, J = 2.33, 1H), 7.17 (s, 1H), 7.20 (d, J = 5.79, 1H), 7.45-7.88 (m, 9H), 8.33 (br s, NH); 13C NMR δ 24.1, 38.1, 111.7, 112.1, 112.8, 116.8, 123.2, 123.9, 127.4, 128.6, 129.0, 132.0, 134.0, 134.6, 135.4, 143.2, 168.2; MS (EIPI) m/e 446 (M+); Anal. Calcd (Obsd) for

C24H18N2O5S: C: 64.4 (64.4), H: 4.1 (4.1) N: 6.3 (6.3). Method A. N,N-Phthalimido--2[5-[[(4-toluoyl)sulfonyl]oxy]-1 H-indol-3- yl]ethylamine. Using tosyl chloride afforded a crude yellow oil in a quantitative yield. Recrystallization from acetone gave 0.67 g (89%) of white crystals: mp 178-179 °C; IR -1 1 (KBr) cm 3416 (NH), 1718 (C=O), 1395, 1353, 1175 (O-SO2); H NMR δ 2.40 (s, 3H),

3.00 (t, J = 7.69, 2H), 3.86 (t, J = 7.79, 2H), 6.86 (dd, J1 = 8.74, J2 = 2.28, 1H), 7.08-7.32 (m, 5H), 7.70-7.87 (m, 6H), 8.15 (br s, NH); 13C NMR δ 21.6, 24.1, 38.1, 111.6, 112.1, 113.0, 117.0, 123.2, 123.7, 127.4, 128.7, 129.6, 132.0, 132.5, 133.9, 134.5, 143.3, 144.9, + 168.1; MS (EIPI) m/e 460 (M ); Anal. Calcd (Obsd) for C25H20N2O5S: C: 65.2 (65.0), H: 4.4 (4.5) N: 6.1 (6.0). Method B. N,N-Phthalimido-2-[5-[[(2-thienyl)sulfonyl]oxy]-1 H-indol-3- yl]ethylamine. A mixture of 5 (0.45 g, 1.47 mmol), 2-thiophene-sulfonyl chloride (0.32 g, mmol) and tetrabutylammonium iodide (50 mg) was magnetically stirred in CH2Cl2 (10 mL). 10% NaOH (20 mL) was added after which the reaction mixture was stirred for

30 min. The product was extracted with CH2Cl2 (3 × 30 mL), the combined organic layers were washed with brine, dried over MgSO4 and evaporated in vacuo. The residual colorless oil was purified by medium-pressure liquid chromatography on SiO2, by eluting with a gradient from n-hexane to EtOAc/ n-hexane (1:4). Pure fractions were pooled and the resulting white solid (0.38 g) was recrystallized from EtOH/H2O yielding a white crystalline material (0.26 g, 39%): mp 170-172 °C; IR (KBr) cm-1 3420 (NH), 1 1701 (C=O), 1398, 1366, 1183 (O-SO2); H NMR δ 3.00 (t, J = 8.02, 2H), 3.87 (t, J =

7.88, 2H), 6.87 (dd, J1 = 8.83, J2 = 2.23, 1H), 7.06 (m, 2H), 7.22, (m, 2H), 7.55 (dd, J1 = 13 3.79, J2 = 1.38, 1H), 7.68-7.84 (m, 5H), 8.40 (br s, NH); C NMR δ 24.1, 38.2, 111.8, 112.8, 116.6, 123.2, 124.0, 127.4, 132.0, 134.0, 134.4, 134.6, 134.7, 135.5, 143.3, 168.2; + MS (EIPI) m/e (M ); Anal. Calcd (Obsd) for C22H16N2O5 S2: C: 58.4 (57.5), H: 3.6 (3.1) N: 6.2 (5.9). Method A. N,N-Phthalimido-2-[5-[[(methylamino)sulfonyl]oxy]-1 H-indol-3- yl]ethylamine. A similar procedure as above was employed using N-methylsulfamoyl chloride affording a crude yellow solid. Recrystallization from EtOH gave 0.40 g (61%) of yellow crystals: mp 184-187 °C; IR (KBr) cm-1 3388, 3247 (NH), 1704 (C=O), 1400, 1 1359, 1184 (O-SO2); H NMR (DMSO) δ 2.74 (d, J = 4.65, 3H), 3.00 (t, J = 6.84, 2H),

3.38 (br s, NH), 3.84 (t, J = 6.84, 2H), 7.00 (dd, J1 = 8.79, J2 = 2.20, 1H), 7.30 (d, J =

96 5-(Sulfonyl)oxy-tryptamines and Ethylamino Side Chain Restricted Derivatives

2.20, 1H), 7.38 (d, J = 8.79, 1H), 7.46 (d, J = 2.20, 1H), 7.78-7.88 (m, 4H), 8.02 (m, NH); 13C NMR δ 24.0, 29.5, 38.3, 111.3, 111.5, 112.4, 116.1, 123.3, 125.3, 127.4, 131.9, 134.7, + 134.8, 143.2, 168.2; MS (EIPI) m/e (M ); Anal Calcd (Obsd) for C19H17N3O5S: C: 57.13 (56.98), H: 4.29 (3.91), N: 10.52 (10.41). Method C. N,N-Phthalimido-2-[5-[[(dimethylamino)sulfonyl]oxy]-1 H-indol- 3-yl]ethylamine. NaH (0.13 g; 60% oil dispersion) was washed with n-hexane and taken up in dry DMF (5 mL). To this magnetically stirred suspension, 5 (0.50 g, 1.63 mmol) was added. After H2-evolution had ceased, N,N-dimethylaminosulfamoyl chloride (210 µL, 1.96 mmol) was added dropwise to the red solution. The resulting reaction mixture was stirred for 30 min at room temperature after which time the reaction was quenched with H2O (50 mL) and extracted with CH2Cl2 (3 × 30 mL). The combined organic layers were dried over MgSO4, filtered and evaporated at the rotavapor. The residual colorless oil was chromatographed on silica gel eluting with

CH2Cl2/MeOH (40:1). Identical TLC-fractions were pooled and evaporated to dryness affording the desired product (0.26 g, 38%) as a white solid and a minor amount of di- substituted product (0.04 g, 5%): mp 179-180 °C; IR (KBr) cm-1 3339 (NH), 1704 1 (C=O), 1396, 1356, 1178 (O-SO2); H NMR δ 3.01 (s, 6H), 3.12 (t, 2H), 3.99 (t, 2H), 7.17 (m, 2H), 7.32 (d, J = 8.54, 1H), 7.59 (s, 1H), 7.69-7.86 (m, 4H), 8.17 (br s, NH); 13C NMR δ 24.3, 38.2, 38.8, 111.5, 111.8, 112.9, 116.6, 123.2, 123.8, 127.6, 132.1, 133.2, 133.9, + 143.8, 168.3; MS (EIPI) m/e 413 (M ); Anal. Calcd (Obsd) for C20H19N3O5S: C: 58.1 (57.8), H: 4.6 (4.7) N: 10.2 (10.1).

General Procedure for Deprotection of N,N-phthalimido-tryptamines. The N,N-phthalimide derivative (1.0 mmol) was dissolved in absolute EtOH (10 mL) after which hydrazine hydrate (1.0 mL) was added. The reaction mixture was stirred for 0.5 h at room temperature after which time the volatiles were removed in vacuo. The residue was refluxed in CHCl3 for 0.5 h, cooled to ambient temperature and filtered in order to remove the solid phthalimidohydrazine. The filtrate was evaporated in vacuo leaving the product which was converted to the oxalate and recrystallized from the appropriate solvent. 2-[5-[[(Methyl)sulfonyl]oxy]-1 H-indol-3-yl]ethylamine oxalate (7). The oxalate salt was recrystallized from MeOH/H2O giving 201 mg (75%) of colorless needles, suitable for single X-ray analysis: mp 178-179 °C; IR (KBr) cm-1 1345, 1178 1 (O-SO2); H NMR (DMSO-d6) δ 2.97-3.04 (m, 4H), 3.29 (s, 3H), 7.06 (d, J = 8.78, 1H), 7.31 (s, 1H), 7.43 (d, J = 8.78, 1H), 7.52 (s, 1H); 13C NMR (DMSO-d6) δ 23.9, 36.8, 39.1, 110.7, 111.9, 115.1, 125.1, 126.8, 134.5, 142.1, 163.9; MS (EIPI) m/e 254 (M+); Anal.

Calcd (Obsd) for C11H14N2O3S.C2H2O4: C: 45.4 (42.4), H: 4.7 (5.1) N: 8.1 (11.6).

97 Chapter 5

2-[5-[[(Phenyl)sulfonyl]oxy]-1 H-indol-3-yl]ethylamine oxalate (8). The oxalate salt was recrystallized from MeOH/H2O giving 143 mg (53%) of white -1 1 crystalline material: mp 136-138 °C; IR (KBr) cm 3430 (NH), 1372, 1191 (O-SO2); H

NMR (CH3OD) δ 3.03 (br s, 4H), 6.68 (d, J = 7.32, 1H), 7.11 (s, 1H), 7.22 (m, 2H), 7.50- 13 7.78 (m, 5H); C NMR (CD3OD) δ 22.6, 39.3, 109.7, 111.0, 111.5, 115.6, 125.2, 126.6, 128.2, 128.8, 133.9, 135.1, 142.8; MS (EIPI) m/e (M+); Anal. Calcd (Obsd) for

C16H16N2O3S.C2H2O4: C: 53.2 (), H: 4.5 () N: 6.9 (). 2-[5-[[(4- Toluoyl)sulfonyl]oxy]-1 H-indol-3-yl]ethylamine oxalate (9). The oxalate salt was recrystallized from MeOH giving 385 mg (94%) of a white powder: mp -1 1 122-123 °C; IR (KBr) cm 3423 (NH), 1364, 1191 (O-SO2); H NMR (DMSO-d6) δ 2.40

(s, 3H), 2.92 (br s, 4H), 6.64 (dd, J1 = 8.73, J2 = 2.28, 1H), 7.20 (d, J = 2.28, 1H), 7.29 (d, J = 8.73, 1H), 7.31 (s, 1H), 7.43 (AB, J = 8.31, 2H) 7.69 (AB, J = 8.31, 2H); 13C NMR (DMSO-d6) δ 21.1, 22.1, 39.2, 110.2, 111.3, 112.2, 115.2, 125.7, 126.7, 128.2, 130.0, 131.7, 134.6, 142.1, 145.4, 164.8; MS (EIPI) m/e (M+); Anal. Calcd (Obsd) for

C17H18N2O3S.C2H2O4: C: 54.3 (), H: 4.8 () N: 6.7 (). 2-[5-[[(2-Thienyl)sulfonyl]oxy]-1 H-indol-3-yl]ethylamine oxalate (10). The oxalate salt was recrystallized from MeOH/Et2O giving 101 mg (56%) of white crystals: -1 1 mp 172-173 °C; IR (KBr) cm 3420 (NH), 1365, 1184 (O-SO2); H NMR (CD3OD) δ

2.97-3.16 (m, 4H), 6.76 (dd, J1 = 8.97, J2 = 2.14, 1H), 7.16 (m, 2H), 7.27 (s, 1H), 7.29 (d, 13 J = 8.97, 1H), 7.54 (dd, J1 = 3.42, J2 = 1.28, 1H), 7.95 (dd, J1 = 5.13, J2 = 1.28, 1H); C

NMR (CD3OD) δ 24.0, 40.7, 111.0, 112.0, 112.8, 116.6, 126.4, 127.9, 128.5, 136.1, +1 136.5, 136.6, 144.1; MS (CI with NH3) m/e (M ); Anal. Calcd (Obsd) for

C14H14N2O3S2.C2H2O4: C: 46.4 (48.3), H: 3.9 (4.6) N: 6.8 (7.2). 2-[5-[[(Methylamino)sulfonyl]oxy]-1 H-indol-3-yl]ethylamine oxalate (11). -1 1 mp °C; IR (KBr) cm 3420 (NH), 1347, 1184 (O-SO2); H NMR (DMSO) δ 2.72 (s, 3H), 13 3.01 (br s, 4H), 7.01 (dd, J1 = 8.79, J2 = 2.20, 1H), 7.32-7.46 (m, 3H); C NMR (DMSO) δ 23.2, 29.5, 38.8, 110.5, 111.2, 112.5, 116.1, 125.7, 127.2, 134.8, 163.3; MS (EIPI) m/e (M+). 2-[5-[[(Dimethylamino)sulfonyl]oxy]-1 H-indol-3-yl]ethylamine oxalate (12).

The oxalate salt was recrystallized from MeOH/Et2O giving 73 mg (46%) of a white -1 1 powder: mp 185-187 °C; IR (KBr) cm 3294 (NH), 1357, 1181 (O-SO2); H NMR

(CH3OD) δ 2.50 (s, 6H), 2.69 (t, J = 7.32, 2H), 2.83 (t, J = 7.32, 2H), 6.63 (d, J = 8.78, 13 1H), 6.68 (s, 1H), 6.99 (d, J = 8.78, 1H), 7.05 (s, 1H); C NMR (CD3OD) δ 24.7, 39.6, 41.6, 111.7, 112.2, 113.8, 117.3, 127.1, 128.9, 145.5, 168.9; MS (EIPI) m/e (M+); Anal.

Calcd (Obsd) for C12H17N3O3S.0.8 C2H2O4: C: 45.96 (45.79), H: 5.28 (5.10) N: 11.89 (11.97).

98 5-(Sulfonyl)oxy-tryptamines and Ethylamino Side Chain Restricted Derivatives

(±)-N,N-Phthalimido-3-amino-6-methoxy-1,2,3,4-tetrahydrocarbazole (15). trans-N-Phthalimido-4-aminocyclohexanone (1.92 g, 7.9 mmol) and (p- methoxyphenyl)hydrazine. HCl salt (1.37 g, 7.9 mmol) were refluxed in abs EtOH (25 mL). The title compound precipitated from solution as an off-white solid. After 1 h, the reaction mixture was cooled to ambient temperature and the solid material (2.26 g, 83%) was collected by filtration on a glass-sintered funnel: mp 213-214 °C (lit.14 211- 213 °C); IR (KBr) cm-1 1702 (C=O); 1H NMR δ 2.01-2.11 (m, 1H), 2.86-2.98 (m, 4H),

3.43-3.56 (m, 1H), 3.83 (s, 3H), 4.62-4.76 (m, 1H), 6.80 (dd, J1 = 8.55, J2 = 2.56, 1H),

6.88 (d, J = 2.56, 1H), 7.19 (d, J = 8.55, 1H), 7.74 (dd, J1 = 5.55, J2 = 2.99, 2H), 7.78 (s, 13 1H), 7.88 (dd, J1 = 5.55, J2 = 2.99, 2H); C NMR δ 23.1, 24.9, 26.8, 48.3, 55.9, 100.2, 108.5, 110.9, 111.1, 123.1, 127.9, 131.4, 132.0, 133.5, 134.0, 153.9, 168.4; MS (EIPI) m/e 346 (M+). (±)-3-Amino-6-methoxy-1,2,3,4-tetrahydrocarbazole (16). The deprotection was conducted according to the general method described for the the N,N-phthalimido- tryptamines, yielding a browhish solid which was taken up in ethylacetate (25 mL) and washed with saturated K2CO3. The aqueous layer was twice extracted with ethylacetate

(25 mL) and the combined organic layers were dried over Na2SO4 and evaporated in vacuo leaving 226 mg (80%) of an off-white solid. Part of this material (120 mg) was converted to the HCl salt and recrystallized from EtOH/ ether giving an off-white solid (103 mg; 58%): mp 108-110 °C IR (KBr) cm-1 ; 1H NMR δ 1.72-1.87 (m, 1H), 1.92-2.08

(m, 3H), 2.45 (dd, J1 = 14.96, J2 = 8.12, 1H), 2.75 (m, 2H), 3.00 (dd, J1 = 14.96, J2 = 4.70,

1H), 3.23-3.33 (m, 1H), 3.87 (s, 3H), 6.80 (dd, J1 = 8.55, J2 = 2.56, 1H), 6.94 (d, J = 2.56, 1H), 7.13 (d, J = 8.55, 1H), 8.38 (br s, NH); 13C NMR δ 21.5, 31.1, 32.5, 47.8, 56.0, 100.2, 107.9, 110.6, 111.2, 128.0, 131.4, 134.1, 153.7. (±)-N,N-Phthalimido-3-amino-6-hydroxy-1,2,3,4-tetrahydrocarbazole (17).

Compound 15 (1.23 g; 3.55 mmol) was dissolved in CH2Cl2 (25 mL) and cooled to –78

°C. This solution was treated dropwise with BBr3 (6.0 mL of 1.0 M in CH2Cl2, 6.0 mmol) under N2-atmosphere and magnetically stirred for 2 h at –78 °C and 18 h at room temperature. The dark brown reaction mixture was poured into H2O (100 mL) and extracted with EtOAc (3 × 100 mL). The combined organic layers were washed with brine, dried over MgSO4 and reduced to dryness at the rotavapor yielding a brown residue. Purification was effected by column chromatography using silica gel eluting with EtOAc/ n-hexane (1:1). Pure fractions were pooled affording a yellow solid (0.36 g, 31%): mp >270 °C (dec; lit.14 270 °C dec); IR (KBr) cm-1 1693 (C=O), 3352 (OH); 1H NMR δ 2.01-2.11 (m, 1H), 2.86-2.98 (m, 4H), 3.43-3.56 (m, 1H), 3.83 (s, 3H), 4.62-4.76

(m, 1H), 6.80 (dd, J1 = 8.55, J2 = 2.56, 1H), 6.88 (d, J = 2.56, 1H), 7.19 (d, J = 8.55, 1H), 13 7.74 (dd, J1 = 5.55, J2 = 2.99, 2H), 7.78 (s, 1H), 7.88 (dd, J1 = 5.55, J2 = 2.99, 2H); C

99 Chapter 5

NMR δ 23.1, 24.9, 26.8, 48.3, 55.9, 100.2, 108.5, 110.9, 111.1, 123.1, 127.9, 131.4, + 132.0, 133.5, 134.0, 153.9, 168.4; MS (CI with NH3) m/e 333 (M ). (±)-N,N-Phthalimido-3-amino-6-[[(trifluoromethyl)sulfonyl]oxy]-1,2,3,4- tetrahydrocarbazole (18). A solution of 17 (227 mg, 0.68 mmol), Et3N (0.2 mL) and

PhN(SO2CF3)2 (300 mg, 0.84 mmol) in CH2Cl2 (10 mL) was magnetically stirred until the reaction mixture became colorless. After 24 h, the organic layer was washed with

10% Na2CO3 (2 × 20 mL) which layers were extracted with CH2Cl2 (2 × 30 mL). The combined organic phases were dried over MgSO4, filtered and evaporated to dryness.

The residual yellow oil was chromatographed on SiO2 eluting with CH2Cl2. Pure fractions were pooled and evaporated in vacuo yielding a white solid (399 mg). This solid was recrystallized from EtOH (195 mg, 62%): mp 105-108 °C; IR (KBr) cm-1 3338 1 (NH), 1705 (C=O), 1398, 1208 (O-SO2); H NMR δ 2.00-2.06 (m, 1H), 2.75-2.88 (m,

4H), 3.37-3.51 (m, 1H), 4.43-4.70 (m, 1H), 6.97 (dd, J1 = 8.54, J2 = 2.44, 1H), 7.15-7.31 m, 2H), 7.70-7.87 (m, 4H), 8.31 (br s, NH); 13C NMR δ 22.7, 24.3, 26.3, 47.6, 109.1,

109.9, 111.2, 114.0, 118.7 (q, J = 321, CF3), 123.1, 127.6, 131.7, 134.0, 135.0, 135.5, + 143.3, 168.4; MS (EIPI) m/e (M ); Anal Calcd (Obsd) for C21H15N2O5SF3: C: 54.31 (52.44), H: 3.26 (3.18), N: 6.03 (5.64). (±)-3- Amino-6-[[(trifluoromethyl)sulfonyl]oxy]-1,2,3,4-tetrahydrocarbazole (19). The title compound was prepared as described for the synthesis of 16, affording 47 mg of a white foam (100%): mp 135-136 °C; 1H NMR δ 1.65-1.83 (m, 1H), 1.92-2.12

(m, 1H), 2.41 (dd, J1 = 14.16, J2 = 8.55, 1H), 2.76 (m, 2H), 2.86-2.99 (m, 1H), 3.22-3.29

(m, 1H), 6.96 (dd, J1 = 8.79, J2 = 2.44, 1H), 7.19 (d, J = 8.79, 1H), 7.28 (d, J = 2.44, 1H), 8.52 (br s, NH); 13C NMR δ 21.1, 30.3, 31.7, 47.1, 109.9, 111.0, 113.7, 118.7 (q, J = 320, + CF3), 127.8, 134.9, 135.0, 143.2; MS (EIPI) m/e 344 (M ).

N-Boc-4-[5-[[(trifluoromethyl)sulfonyl]oxy]-1 H-indol-3-yl]piperidine (21). The triflate derivative of compound 20 was prepared according to the procedure used for 18 giving 3.86 g (86%) after recrystallization from ether/n-hexane: mp 187-188 °C -1 1 IR (KBr) cm 3309 (NH), 1657 (C=O), 1422, 1209 (O-SO2); H NMR δ 1.51 (s, 9H), 1.65

(dt, J1 = 12.49, J2 = 3.84, 2H), 2.01 (br d, J = 12.30, 2H), 2.86-2.99 (m, 3H), 4.19 (br d, J = 12.49, 2H), 7.04-7.10 (m, 2H), 7.37 (d, J = 8.74, 1H), 7.50 (d, J = 2.26, 1H), 8.70 (br s, 13 NH); C NMR δ 28.5, 32.7, 33.4, 44.3, 79.6, 111.4, 112.2, 115.0, 118.7 (q, J = 321, CF3), +18 + 121.5, 122.4, 126.7, 135.2, 143.1, 155.0; MS (CI with NH3) m/e 466 (M NH4 ); Anal

Calcd (Obsd) for C19N23N2O5SF3: C: 50.89 (50.77), H: 5.17 (4.98), N: 6.25 (6.10). 4-[5-[[(Trifluoromethyl)sulfonyl]oxy]-1 H-indol-3-yl]piperidine (22).

Compound 21 (3.4 g, 7.6 mmol) was dissolved in CH2Cl2 (30 mL) and deprotected by adding TFA (3.5 mL) at 0 °C. The reaction mixture was allowed to warm to ambient

100 5-(Sulfonyl)oxy-tryptamines and Ethylamino Side Chain Restricted Derivatives temperature. After 7 h, the volatiles were removed in vacuo giving an off-white solid which was taken up in 10% NaHCO3 and extracted with CH2Cl2 (3 × 50 mL; the funnel broke, lost some material). The combined organic layers were dried (Na2SO4) and reduced to dryness affording 1.95 g (74%) of a light brown solid. Recrystallization from ethyl acetate/n-hexane gave an off-white solid (1.16 g, 44%). A small portion was converted in the HCl salt and recrystallized from acetonitrile: mp 249-250 °C (HCl); IR -1 1 (KBr) cm 3389 (NH), 1425, 1200 (O-SO2); H NMR (base) 2.08-2.20 (m, 2H), 2.35 (d, J

= 12.82, 2H), 3.24 (t, J = 12.09, 3H), 3.69 (d, J = 12.09, 2H), 7.36 (dd, J1 = 8.79, J2 = 1.83, 1H), 7.49 (s, 1H), 7.64 (d, J = 8.79, 1H), 7.75 (d, J = 1.83, 1H); 13C NMR (base) δ

28.4, 29.3, 44.6, 108.5, 109.8, 112,6, 116.3 (q, J = 321, CF3), 123.9, 126.4, 132.5, 140.7; +1 MS (CI with NH3) m/e 349 (M ); Anal Calcd (Obsd) for C14H15N2O3SF3.HCl: C: 43.70 (43.47), H: 4.19 (4.06), N: 7.28 (7.32). N-Methyl-4-[5-[[(trifluoromethyl)sulfonyl]oxy]-1 H-indol-3-yl]piperidine (23). To a magnetically stirred solution of compound 22 (0.73 g, 2.1 mmol) in acetonitrile (10 mL), 37% aquous formaldehyde (1.2 mL) and NaCNBH3 (0.4 g) were added. The reaction mixture was acidified until pH 5 with glacial acetic acid, stirred for 2 h at room temperature and quenced with 10% NaOH (50 mL). After extraction

(CH2Cl2, 2 × 50 mL), the organic layers were dried (MgSO4) and evaporated in vacuo yielding 0.53 g (73%) of an oil. Conversion to the HCl salt and recrystallization from acetonitrile gave an off-white solid (165 mg, 20%): mp 247-248 °C; IR (KBr) cm-1 3134 1 (NH), 1416, 1204 (O-SO2); H NMR (CD3OD) δ 1.93-2.31 (m, 4H), 2.92 (s, 3H), 3.11-

3.32 (m, 3H), 3.61 (br d, J = 12.21, 2H), 7.07 (dd, J1 = 8.79, J2 = 2.45, 1H), 7.29 (s, 1H), 13 7.46 (d, J = 8.79, 1H), 7.61 (d, J = 2.44, 1H); C NMR (CD3OD) δ 30.0, 30.1, 42.4, 54.4,

110.5, 112.1, 113.9, 121.9, 123.1, 126.0, 143.0 (two carbons missing); MS (CI with NH3) +1 m/e 363 (M ); Anal Calcd (Obsd) for C15H17N2O3SF3.HCl: C: 45.17 (45.08), H: 4.55 (4.50), N: 7.02 (7.29).

Pharmacology. Materials. The HeLa/HA7 cell line was obtained from Tulco (Duke university, Durham, NC, USA). Cos-7 cells were purchased from ATCC (Rockville, USA). [3H]5-CT (51.3 Ci/mmol) and [3H]8-OH-DPAT (228 Ci/mmol) were obtained from New England Nuclear (Les Ulis, France)

Receptor Binding Assay. Membrane preparations of the Hela/HA7 cell line transfected with the 5-HT1A receptor gene and stably transfected Cos-7 cells expressing either 5-HT1Dα or 5-HT1Dβ receptors were prepared in 50 mM Tris-HCl pH 7.7 containing 4 4 mM CaCl2, 10 µM pargyline and 0.1% ascorbic acid as previously described. Binding assays were performed with 1 nM [3H]8-OH-DPAT or 0.5 nM [3H]5-CT. Incubation of mixtures consisted of 0.4 mL cell membrane preparation [200 µg (5-HT1A), 20 to 100 µg µ (5-HT1Dα) and 20 to 80 g (5-HT1Dβ) protein], 0.05 mL radioligand and 0.05 mL

101 Chapter 5 compound for inhibition or 10 µM 5-HT to determine non-specific binding. The reaction were stopped after 30 min incubation at 25 °C by adding 3.0 mL ice-cold 50 mM Tris-HCl pH 7.7 and rapid filtration over Whatman GF/B glass fiber filters using a Brandel Harvester, washed and counted as previously described.4 Data were analyzed graphically with inhibition curves and IC50 values were derived. Ki values were calculated according to the equation:

Ki = IC50/(1 + C/Kd) with C the concentration and Kd the equilibrium dissociation constant of the radioactively labelled ligand. The corresponding Kd values are: 2.5 nM (5-HT1A); 0.22 nM (5-HT1Dα) and 0.12 nM (5-HT1Dβ).

Acknowledgments. Dr. Peter Pauwels, Christiane Palmier and Stephanie Tardif (Centre de Recherche Pierre Fabre, Castres, France) are gratefully acknowledged for performing the binding experiments. We thank Dr. Max Lundmark and Dr. Staffan Sundell (Department of Structural Chemistry, University of Gothenburg, Sweden) for solving the X-ray structure of compound 7.

102 5-(Sulfonyl)oxy-tryptamines and Ethylamino Side Chain Restricted Derivatives

5.6 References

[1] Sonesson, C.; Boije, M.; Svensson, K.; Ekman, A.; Carlsson, A.; Romero, A.G.; Marin, I.J.; Duncan,J.N.; King, L.J.; Wikström, H.J. Med. Chem. 1993, 36, 3409. [2] Sonesson, C.; Barf, T.; Nilsson, J.; Dijkstra, D.; Carlsson, A.; Svensson, K.; Smith, M.W.; Marin, I.J.; Duncan, J.N.; King, L.J.; Wikström, H.J. Med. Chem. 1995, 38, 1319. [3] See for instance: Perez, M.; Fourrier, C.; Sigogneau, I.; Pauwels, P.J.; Palmier, C.; John, G.W.; Valentin, J.-P.; Halazy, S. J. Med. Chem. 1995, 38, 3602. [4] Pauwels, P.J.; Colpaert, F.C.Eur J. Pharmacol. 1996, 300, 137. [5] King, F.D.; Brown, A.M.; Gaster, L.M.; Kaumann, A.J.; Medhurst, A.D.; Parker, S.G.; Parsons, A.A.; Patch, T.L.; Raval, P. J. Med. Chem. 1993, 36, 1918. [6] Oxford, A.W.; Butina, D.; Owen, M.R.Eur. Patent Appl. 88-307499. [7] Mealy, N.; Castañer, J.Drugs of the future 1996, 21, 476. [8] King, F.D.; Gaster, L.M.; Kaumann, A.J.; Young, R.C.Pat. Appl. WO 93-00086. [9] Stang, P.J.; Anderson, A.G.J. Org. Chem. 1976, 41, 781. [10] Humphrey, P.P.A.; Feniuk, W.; Perren, W.; Oxford, A.W.; Brittain, R.T.Drugs of the Future 1989, 14, 35. [11] Howarth, N.M.; Purohit, A.; Reed, M.J.; Potter, B.V.L.J. Med. Chem. 1994, 37, 219. [12] Yamada, I.; MIzuta, H.; Ogawa, K.; Tahara, T.Chem. Pharm. Bull. Tokyo 1990, 38, 2552. [13] Rippmann, F.; Böttcher, H.Kontakte (Darmstadt) 1994, 1, 30. [14] Glennon, R.A.; Hong, S.-S.; Bondarev, M.; Law, H.; Dukat, M.; Rahkit, S.; Power, P.; Fan, E.; Kinneau, D.; Kamboj, R.; Teitler, M.; Herrick-Davis, K.; Smith, C.J. Med. Chem. 1996, 39, 314. [15] Kloek, J.A.; Leschinsky, K.L.J. Org. Chem. 1976, 41, 4028.

103 Chapter 6

Structure-Affinity and Structure-Activity Relationships of Ortho-Substituted Phenylpiperazines

Abstract

The hydroxy and trifluoromethylsulfonic acid ester derivatives of (N-{2-[4-(2- methoxyphenyl)-1-piperazinyl]ethyl}-N-(2-pyridinyl)cyclohexane carboxamide (WAY100635, 13) and (6-{4-[4-(2-methoxyphenyl)-1-piperazinyl]butyl}-N-methyl- benzoxazolinone (ORG13502, 14) were prepared and their pharmacology is evaluated in terms of affinity and intrinsic activity for 5-HT1A receptors. The phenylpiperazines 14, 16 and 18 were all found to be high affinity receptor ligands (Ki values ranging from 0.13-4.0 nM). Interestingly, the ortho-hydroxy and -methoxyphenylpiperazines were shown to be low intrinsic activity receptor ligands. In contrast, 17 and 18 behaved as partial 5-HT1A receptor agonists, showing intrinsic activities around 0.7-0.8. The structure-activity relationships (SAR) are discussed in terms of electronic properties of the arylpiperazine moieties. A crystal database search provided supportive structural information which is included in the discussion.

6.1 Introduction

During the last decade, 2-aminotetralins have been particularly useful in formulating the structure-affinity relationships (SAFIR) of 5-HT1A receptor ligands and, more recently, the accumulation of data on arylpiperazines seems to have provided a basis for the formulation of structure-activity relationships (SAR) for this receptor.1 Simple arylpiperazines are partial agonists and were shown to have non-selective binding profiles and moderate affinities for 5-HT1A sites (Table 6.1). However, substitution of the more basic amine functionality of the piperazine with a carbon-chain separated amide function usually enhances both the selectivity and affinity for this receptor subtype.1 Consequently, special attention has been paid to the modification of the terminal amide moiety and the spacer length. A number of compounds that represent such modifications are depicted in Table 6.1. Obviously, the optimal chain length depends highly on the type of aryl substituent used (compare entries 6 and 8, and 9 and 10, respectively). Furthermore, the intrinsic activity seemingly depends on the type of

103 Chapter 6 aryl subtituent, the spacer length and the amide terminus employed. Notably, the 1-(2- pyrimidinyl)piperazines 6-8 all behave as agonists with variable intrinsic activities (I.A.), whereas the use of ortho-methoxyphenylpiperazines may give rise to receptor ligands with low or no intrinsic activity.

104 SAFIR and SAR of Ortho Substituted Phenylpiperazines

Table 6.1. SAFIR and SAR of Arylpiperazines for 5-HT1A receptors.

Aryl N N (CH2)n amide terminus

Entry Name Aryl n Amide Ki (nM) I.A.a ref.

1 2-MeO-PP o-MeO-Ph - - 68 0.7 [2] 2 m-CPP m-Cl-Ph - - 1950 0.7 [3]

3 TMFPP m-CF3-Ph - - 2400 0.4 [3] 4 1-PP 2-pyrimidinyl - - 1410 - [2]

NH2 5 PAPP m-CF3-Ph 2 6 1.0 [4] O

6 buspirone 2-pyrimidinyl 4 N 30 0.5 [5]

O

O

7 ipsapirone 2-pyrimidinyl 4 N 7 0.9 [5] S O O

O

8 BMY9075 2-pyrimidinyl 2 N 150 1.0 [5]

O

O

9 BMY8227 o-MeO-Ph 4 N 2.5 0.1 [5]

O

O

10 BMY7078 o-MeO-Ph 2 N 1.8 0.2 [5]

O

O

11 NAN190 o-MeO-Ph 4 N 0.6 0 [6]

O

O

N 12 SDZ216-525 2-CO2Me-(indol-4- 4 0.6 0 [7] S O yl) O

(a) Intrinsic activities determined by cAMP assays.

105 Chapter 6

Supportive data were produced by Millan and co-workers, who utilized the postsynaptic 5-HT1A receptor-mediated hypothermia in rats as an in vivo model to evaluate the intrinsic activity of 5-HT1A receptor ligands. Indeed, the 8-OH-DPAT- induced hypothermia (DIH) was potently blocked by NAN190 (max inhibition 88%), BMY7378 (96%) and SDZ216-525 (100%).8 Thus it seems that the antagonist properties of 5-HT1A receptor ligands can be modulated by varying the aryl moiety. Recently, WAY100635 (13) was reported to be a potent and selective antagonist 9 at pre- and postsynaptic 5-HT1A receptors. This ligand also consists of an amide terminus which is separated by an ethylene unit from the piperazine ring. ORG13502

(14) is an extremely potent and selective 5-HT1A receptor ligand which consists of an o- MeO-phenylpiperazine, an alkyl chain of four methylene units and a benzoxazolinone moiety. It exhibits a Ki of 0.16 nM for 5-HT1A receptors and has a low intrinsic activity profile (I.A. = 0.2; cAMP assay). This chapter focusses on the effect of O-demethylation and subsequent triflation of 13 and 14 on the intrinsic activity for 5-HT1A receptors.

6.2 Chemistry

The O-demethylation of the ortho-methoxy phenylpiperazines could not be effected with refluxing in 48% HBr, with or without acetic acid (partial conversion), or by employing BBr3 in CH2Cl2 (no conversion). The demethylation step went smoothly and in high yields upon treatment with 5–6 equivalents of AlCl3 in refluxing benzene for both compounds, ORG13502 and WAY100635.10 Complete conversion was necessary since the starting materials and the products were difficult to separate by column chromatography. Triflation of the ortho-hydroxy phenyl piperazines provided compounds 17 and 18 and required phase-transfer-conditions using 10% NaOH and

CH2Cl2 in order to be successful (see also Section 2.2). Tetrabutyl ammonium iodide was used as phase-transfer catalyst. The crystallization of the free base of 16 from i-PrOAc yielded colorless needles, which were suitable for single crystal X-ray spectroscopy (Figure 6.1). The phenylpiperazine derivative crystallized in the monoclinic P21/c space group with 4 molecules per unit cell (a = 15.515; b = 8.823; c = 27.032 Å). Some selected bond distances, angles and torsional angles are given in Table 6.2. The torsional dihedral angle of C20-N19-C22-C27 is –62.1°, which indicates that the relative orientation of the phenyl ring and the piperazine ring lies between the coplanar and perpendicular conformation. As expected, both piperazino N-substituents are oriented in a equatorial fashion resulting in a ‘stretched’ molecule.

106 SAFIR and SAR of Ortho Substituted Phenylpiperazines

Figure 6.1A. Molecular structure of16 .

Figure 6.1B. Stereoview of the molecular structure of16 .

28 OH 26 27 20 21 14 13 7 8 11 16 6 9 Me 25 N19 N 22 15 12 N1 24 23 18 17 5 4 2 O O10 3 Figure 6.1C. Numbering of the atoms of16 .

107 Chapter 6

Table 6.1. Selected interatomic distances, angles and torsional angles of compound16

Distance (Å) Angle (deg) Torsional Angle (deg)

N1-C2 1.357 O3-C2-O10 122.0 C11-N1-C2-O10 –2.8

N1-C9 1.400 N1-C2-O10 129.9 C7-C6-C12-C13 74.4

N1-C22 1.446 C6-C12-C13 114.4 C6-C12-C13-C14 178.4

C2-O3 1.383 C14-C15-N16 113.3 C12-C13-C14-C15 66.7

C2-O10 1.215 C15-N16-C17 108.0 C13-C14-C15- –170.6

N16

O3-C4 1.393 C18-N19-C22 117.3 C14-C15-N16- 175.1

C17

N16-C15 1.479 N19-C22-C27 119.4 C13-N3-C16-C17 165.6

N19-C22 1.421 C18-N19-C20 109.9 N19-C22-C27- 5.5

O28

O28-C27 1.374 C18-N19-C22 115.6 C20-N19-C22- –62.1

C27

OMe OH OSO2CF3 a b N N R N N R N N R

13 (WAY100635; R = R1) 15 (R = R1) 17 (R = R1) 14 (ORG 13502; R = R2) 16 (R = R2) 18 (R = R2)

N O O

R = (CH ) N R = (CH ) N 1 2 2 2 2 4 Me

O

Scheme 6.1. (a) AlCl3, benzene, ∆; (b) PhN(SO2CF3)2, TBAI, 10% NaOH, CH2Cl2.

6.3 Pharmacology

Receptor Binding . The test compounds were evaluated for their in vitro binding affinities at human cloned 5-HT1A, 5-HT2A and 5-HT2C, expressed in NIH-3T3 cells (Table 6.2). The displacement of the radioactively labelled ligands [3H]8-OH-DPAT (5- 3 3 HT1A), [ H]ketanserine (5-HT2A) and [ H]5-HT (5-HT2C) was measured.

108 SAFIR and SAR of Ortho Substituted Phenylpiperazines

Table 6.2. Affinities at 5-HT1A, 5-HT2A and 5-HT2C Receptor Subtypesin Vitro

Ki in nM(pKi)a

Compound 5-HT1A 5-HT2A 5-HT2C 13 2 (8.7)b - - 14 0.16 (9.8) 794 (6.1) 1259 (5.9) 16 0.13 (9.9) 4,000 (5.4) 10,000 (5.0) 18 4.0 (8.4) 400 (6.4) 5,000 (5.3)

(a) Ki in nM (pKi’s in parentheses) values for displacement of the 5-HT1A 3 receptor agonist [ H]8-OH-DPAT, the 5-HT2A receptor antagonist 3 3 [ H]ketanserine and the 5-HT2C receptor agonist [ H]5-HT. Data from cloned human receptors expressed in NIH-3T3 cells. (b) IC50 (pIC50). Taken from ref 9.

cAMP Assay . The forskolin-stimulated cAMP inhibition, using the cloned human 5-HT1A receptor expressed in NIH-3T3 cells, was assessed by measuring the test drugs’ effective concentration which produced 50% activation (EC50; agonist assay) in the presence of 1µM forskolin. The antagonist assays were carried out in the presence of 1 µM forskolin and 3 × 10-7 M 5-HT, and were expressed as the test drugs’ concentration which induced 50% inhibition of the 5-HT response (IC50). In Vivo Inhibition of Lower Lip Retraction . The abilities of the compounds to block the 8-OH-DPAT-induced (0.22 mg/kg) lower lip retraction (LLR) were tested in 11 rats and are expressed as the effective dose producing 50% inhibition (ID50).

109 Chapter 6

a Table 6.3. Intrinsic Efficacy and Intrinsic Activity in Cells Transfected with 5-HT1A Receptors.

(Ant)agonist Properties

b c d e Compound IC50 (pIC50) EC50 (pEC50) ID50 I.A. 13 2.5 (8.6) >10,000 (<5) 0.04 ≥0.2 15 3.2 (8.5) >10,000 (<5) NTf ≥0.4 17 10 (8.0) ≥3,200 (≤5.5) NT 0.7 14 8.5 (8.3) 0.5 (9.3) 0.3 0.2 16 32 (7.5) 39.8 (7.4) 0.16 0.2 18 7.9/316 (8.1/6.5)g ≥3,200 (≤5.5) NT ≥0.8

(a) Cloned human 5-HT1A receptors are expressed in 5-HT1A-NIH-3T3 cells. The forskolin stimulated extracellular cAMP-accumulation (nM) is measured. (b) Antagonist assay: The concentration of the test

drug at which 50% of the 5-HT response is blocked. IC50 in nM. (c) Agonist assay: Concentration of the test drug which induces 50% activation. EC50 in nM. (d) The dose (mg/kg,sc ) at which 50% of the 8- OH-DPAT-induced LLR in rats is inhibited. (e) The intrinsic activities (I.A.) are deduced from the cAMP agonist assays. (f) NT means Not Tested. (g) Two determinations gave different numbers for unknown reasons.

6.4 Results and Discussion

When applied on 2-aminotetralins, the aryl triflate group is an excellent bioisostere of the phenol group (see Chapter 2). However, this concept seems to be somewhat less successful in case of the more flexible arylpiperazines. The ortho- hydroxy and -methoxy substituted phenylpiperazines 14 and 16 seemed to be approximately 30 times more potent 5-HT1A receptor ligands than the triflated analogue 18 (Table 6.2). The observation that compound 18 displays a Ki of 4.0 nM, although comparatively lower than the hydroxy- and methoxyphenylpiperazines, suggests that a triflate group in this position is tolerated if the proper type of amide terminus is employed. None of the compounds showed considerable affinity for the 5-HT2A and 5-

HT2C receptor subtypes. The data from Table 6.3 demonstrate that the methoxy substituted phenylpiperazines are the most potent 5-HT1A receptor antagonists, whereas the phenol analogues exhibit altered antagonist properties, depending on the amide terminus used. Interestingly, 13 exhibited an intrinsic activity of ≥0.2, which challenges 9 the reported ‘silence’ of this compound in the in vivo assays. However, the EC50 of >10,000 nM of 13 shows that this compound was virtually inactive in the agonist assay.

This observation constitutes the major difference with 14, which exhibited an EC50 value of 0.5 nM. All three low intrinsic activity compounds, 13, 14 and 16, blocked the 8-OH-DPAT-induced LLR in rats very potently. Strikingly, both triflate derivatives loose much of their antagonist properties and become partial 5-HT1A receptor agonists,

110 SAFIR and SAR of Ortho Substituted Phenylpiperazines as is shown by the intrinsic activities of 0.7 and 0.8 for compounds 17 and 18, respectively. Direct interactions of the ortho-substituents with the 5-HT1A receptor may contribute to this observation. The possibility that the bulkiness of the triflate group causes unfavorable interactions in the receptor as compared to the hydroxy and methoxy substituents cannot be ruled out, since compound 18 displayed a Ki of 4 nM for 5-HT1A sites. In addition, it may well be that the electronic properties of these ortho- substituents have considerable consequences for the individual intrinsic activity and efficacy of these phenylpiperazines. Molecular modelling studies have shown that the electronic character of the aryl group and its substituents determines the degree of conjugation between the anilino lone pair and the aromatic π-electrons.12 This conjugation directly influences the relative orientation between the aryl and piperazine ring (Figure 6.2). Electron- donating groups, such as the hydroxy and methoxy substituents, decrease the conjugation and thus direct both rings to adopt a more perpendicular (T0) conformation. Electron-withdrawing substituents, such as the triflate group, have the opposite effect and favor a more coplanar (T90) orientation. OR 4 1 3 2 N N R

Figure 6.2. Relative orientation defined by the torsional angle 1234T ; coplanar: T1234 = 90° (T90) and perpendicular: T1234 = 0° (T0).

By using the SYBYL13 molecular modelling package, we have calculated the rotational energy barrier on the free bases derived from the X-ray structure of 16 presented in Figure 6.1. Thirty six conformations were generated by a stepwise rotation of 10° around Lp1-N2-C3-C4, and energy minimized while keeping the torsional angle, which defines the relative orientation between the ring systems, fixed. Figure 6.4 shows comparable energy plots for the three molecules, which implicates that the rotation barriers around the C3-N2 bond are similar. The predicted energy barriers (∆E) between the absolute minimum and the absolute maximum for the o-methoxy, o-hydroxy and the o-triflate phenylpiperazines are 14.3, 12.7 and 14.2 kcal/mol, respectively. Each of the structures have two predicted minima with torsional angles T1234 of approximately 0 and 180°, corresponding to perpendicular conformations, in which the anilino lone pair and the ortho-substituent point towards the same and opposite direction, respectively. The

X-ray conformation of compound 16 displays a T1234 of 50°, which is not a calculated

111 Chapter 6 absolute or local minimum, according to the energy plot of o-OH-PP. Dijkstra demonstrated that the electronic effects are not accounted for by the Tripos force field of SYBYL, implying that the rotation barriers presumably represent steric interactions.12 The conformations in which the steric interactions are maximal form the tops of the energy plots.

O CF3 H3C H S O O O O 4 1 3 2 N NH N NH N NH

o-OMe-PP o-OH-PP o-OTf-PP

Figure 6.3. Structures of the compounds subjected to rotational energy barrier calculations.

Unfortunately, the electronic influence could not be investigated since we were not able to perform MOPAC14 calculations on these phenylpiperazines due to the lack of parameters for the triflate functionality. Instead, we analyzed crystallographic conformations of 41 phenylpiperazines from the Cambridge Crystallographic Database (CDB). Only X-rays of tertiary amines were considered. In all these structures the piperazine rings are found in the chair conformation with both N-substituents in the equatorial position. We measured the C3-N2 distance (Å), the total of the three angles (°) around the anilino nitrogen atom and the T1234 (°) of each phenylpiperazine. The phenylpiperazines without aryl substituents (17 examples) randomly exhibit T1234 angles ranging from 55–90°, whereas substitution has a marked effect on this particular torsional angle. A meta-chloro or meta-methyl substituent (3 and 2 examples, respectively) results in angles in the range of 60–90°, which contrasts the effect of an ortho-methoxy group giving T1234’s ranging from 45–55° (5 examples). We did not observe a clearcut correlation between the sp3-hybridization of the anilino nitrogen

(total C-N-C angle or C3-N2 distance) and T1234, which contradicts the results of Gilli and Bertolasi,15 who demonstrated that the C-N bond distance, which reflects the C(sp2)- N(sp3) bond order, depends on the torsional angle. Although no straighforward answers can be given on the effects of individal substitution patterns, some trends are indicated.

A single example of a protonated N2-atom revealed a T1234 of 9°, which sharply contrasts with the torsional angles measured for 17 nonprotonated anilino-nitrogen atoms. Thus, the absence of the anilino lone pair in the former compound strongly favors the perpendicular orientation. This is also seen in six retrieved phenylpiperidines having a

112 SAFIR and SAR of Ortho Substituted Phenylpiperazines sp3-carbon atom (instead of an anilino-nitrogen atom), which all showed a nearly perpendicular conformation. Methyl-substituted (weak electron-releasing) and a chloro-substituted (mesomeric electron-releasing; inductive electron-withdrawing) phenylpiperazines exhibit torsional angles comparable to unsubstituted phenylpiperazines. 16

14 o-OMe-PP

12

10

8

E 6 ∆

4

2

0

-2 -180 -150 -120 -90 -60 -30 0 30 60 90 120 150 180 T 1234 16

14 o-OH-PP

12

10

8

E 6 ∆

4

2

0

-2 -180 -150 -120 -90 -60 -30 0 30 60 90 120 150 180

T1234

113 Chapter 6

16

14 o-OTf-PP

12

10

8

E 6 ∆

4

2

0

-2 -180 -150 -120 -90 -60 -30 0 30 60 90 120 150 180

T1234

Figure 6.4. The rotational energy barriers of o-OMe-PP (top), o-OH-PP (middle) and o-OTf-PP (bottom) calculated with the Tripos force field.

F3C

N

19 (SR57746A)

Interestingly, an unsubstituted phenyltetrahydropyridine and 1-(2- pyrimidinyl)piperazine in the CDB both exhibited nearly coplanar conformations, which is probably the result of a strong conjugation of the aryl π-electrons with the double bond π-electrons and the anilino lone-pair, respectively. The meta- (trifluoromethyl)phenyltetrahydropyridine SR57746A (19, Ki = 2 nM) possesses the 16 profile of a full 5-HT1A receptor agonist, whereas pyrimidinylpiperazines (Table 6.1) are all (partial) agonists. This suggests that coplanar conformations favor 5-HT1A receptor agonism. The electron-withdrawing CF3 group on the meta position of PAPP (5) may have a coplanar-orientation inducing effect, giving rise to the full intrinsic activity. It it tempting to say that ortho-substitution, favoring a perpendicular orientation, results in a low intrinsic activity profile for 5-HT1A receptors. However, the 17 high intrinsic activity benzodioxynylpiperazines, such as eltoprazine (CDB; T1234 = 50°) and flesinoxan,18 are also ortho-substituted phenylpiperazines. The agonist properties may be due to the increased electron-donating effect of the di-oxo substituent compared to a single methoxy group or to a direct interaction of the meta- positioned oxygen atom with the receptor. Thus, it is difficult to separate the influence of the relative conformation and the electrostaic potential of the aryl moiety on the

114 SAFIR and SAR of Ortho Substituted Phenylpiperazines intrinsic activity. Whether the inductive electron-withdrawing property of an aryl triflate group contributes in a slightly lower binding affinity of phenylpiperazines 17 and 18 compared to their hydroxy/methoxy congeners remains to be determined. It seems that (i) the electrostatic potential of the aryl ring and (ii) the positioning of this aryl ring relative to that of the piperazine moiety and specific aromatic amino acid residues in the receptor are parameters that define the agonist/antagonist properties of an arylpiperazine. These two parameters are governed primarily by the substitution pattern of the aryl ring. In addition, (iii) the nature of the N-substituent determines whether the relative position of the aryl moiety will be optimized in order to become a full agonist or antagonist. According to the energy plots generated with the Tripos force field, the steric effects of the ortho-substituents on the conformational behaviour of the phenylpiperazine portion in compounds 13-18 are similar. The dramatic increase of intrinsic activity of the triflated phenylpiperazines compared to the hydroxy/methoxy congeners thus must be explained by electronic properties or a differential direct interaction of the various ortho-substituents with the 5-HT1A receptor. Whether a preferential coplanar conformation as a result of the electon-withdrawing character of the triflate group accounts for the diminished 5-HT1A receptor antagonist properties remains to be investigated.

6.5 Experimental Section

General. For general remarks see Section 2.4. Materials. ORG13502 was kindly provided by N. V. Organon (Oss, The Netherlands). WAY100635 was synthesized by Marquerite Mensonides in our laboratory according to published procedures.19 6-[4-[4-(2-hydroxyphenyl)-1-piperazinyl]butyl]- N-methyl-benzoxazolinone

(16). ORG13502 (60 mg, 128 µmol) and AlCl3 (86 mg, 5 equivalents) were suspended in dry benzene (5 mL) and refluxed for 6 h. The reaction mixture was cooled to room temperature, quenched with H2O (5 mL) and neutralized with solid NaHCO3. The mixture was extracted with CH2Cl2 (3 × 20 mL) after which the combined organic layers were washed with brine (30 mL) and dried over MgSO4. Filtration and evaporation in vacuo gave a white solid (46 mg; 94%), which was recrystallized from ethylacetate yielding 34 mg (70%) of colorless needles: mp 146-148 °C; (mono-HCl salt) mp 255- ° -1 1 δ 257 C; IR (KBr) 1772 cm (C=O); H NMR 1.67-1.81 (m, 4H), 2.54 (dd, J1 = 7.33, J2 = 7.69, 2H), 2.72 (m, 4H), 2.80 (dd, J1 = 7.69, J2 = 6.95, 2H), 3.02 (t, J = 4.76, 4H), 3.50 (s,

3H), 6.94-7.29 (m, 7H); HRMS Calcd (Obsd) for C22H27N3O3 381.205 (381.205); Anal.

Calcd (Obsd) for C22H27N3O3.HCl: C 63.23 (62.94), H 6.75 (6.74), N 10.05 (9.87).

115 Chapter 6

N-[2-[4-(2-hydroxyphenyl)-1-piperazinyl]ethyl]- N-(2-pyridinyl)cyclohexane carboxamide (15). WAY100635 (50 mg, 118 µmol) was demethylated as described for

ORG13502 employing 6 eq. of AlCl3 and refluxing for 2 h yielding 45 mg (93%) of a colorless oil which was converted to the oxalate and recrystallized from i-PrOH yielding 43 mg (73%) of white crystals; (mono-oxalate salt) mp 204-207 °C; IR (KBr) cm-1 1659 (C=O); 1H NMR δ 0.98-1.26 (m, 4H), 1.47-1.79 (m, 6H), 2.18-2.32 (m, 1H),

2.60-2.67 (m, 6H), 2.74 (t, J = 4.56, 4H), 4.00 (t, J = 7.02, 2H), 6.81-7.31 (m, 6H), 7.75- 7.84 (dt, J1 = 7.45, J2 = 2.09, 1H), 8.53-8.56 (dd, J1 = 5.18, J2 = 2.00, 1H); HRMS Calcd

(Obsd) for C24H32N4O2 408.253 (408.253). Anal. Calcd (Obsd) for C24H32N4O2.C2H2O4.0.5

H2O: C 61.52 (61.24), H 6.95 (6.68), N 11.04 (10.60). 6-[4-[4-[2-(trifluoromethyl)sulfonyl]oxy]phenyl]-1-piperazinyl]butyl]- N- methyl-benzoxazolinone (18). Under a N2-atmosphere, a mixture of (16, 114 mg, 0.30 mmol), PhN(SO2CF3)2 (130 mg, 0.36 mmol) and tetrabutylammonium iodide (11 mg, 10 mol%) were magnetically stirred in CH2Cl2 (2 mL) and 10% aquous NaOH (1 mL) for 20 h. H2O (10 mL) was added to the reaction mixture which was subsequently extracted with ether (3 × 15 mL). The organic layers were washed with brine (30 mL), dried over

MgSO4 and evaporated in vacuo. The resulting oil was purified on a silica column eluting with CH2Cl2/MeOH (30:1), affording a colorless oil which solidified on standing. After conversion to the oxalate, the title compound was recrystallized from acetonitrile yielding 91 mg (50%) off-white material: mp 169-170 °C; IR (KBr) 1772 cm-1 (C=O); 1H NMR δ 1.59-1.69 (m, 4H), 2.51-2.56 (m, 2H), 2.66-2.74 (m, 6H), 3.10 (m,

4H), 3.39 (s, 3H), 6.86 (d, J = 6.81, 1H), 7.02-7.35 (m, 6H); MS (CI with NH3) m/e 514 +1 (M ); Anal. Calcd (Obsd) for C23H26N3O5SF3. C2H2O4: C 49.75 (50.04), H 4.68 (4.71), N 6.96 (7.12). N-[2-[4-[2-[[(trifluoromethyl)sulfonyl]oxy]phenyl]-1-piperazinyl]ethyl]- N- (2-pyridinyl)cyclohexane carboxamide (17). This compound was prepared according to the procedure as described for compound 18 starting from 100 mg of compound 15 (0.25 mmol). After extractive workup, the product was purified by column chromatography (silica gel eluting with CH2Cl2/MeOH 50:1). The oxalate of the title compound was recrystallized from ethanol yielding 46 mg (36%) white material: mp -1 1 182-184 °C; IR (KBr) cm 1665 (C=O), 1414, 1207 (SO2); H NMR δ 0.96-1.27 (m, 3H), 1.47-1.73 (m, 7H), 2.18-2.31 (m, 1H), 2.64 (m, 6H), 2.94 (m, 4H), 4.00 (t, J = 6.93, 2H), 7.07-7.36 (m, 6H), 7.73-7.81 (dt, J1 = 7.45, J2 = 2.04, 1H), 8.52-8.55 (dd, J1 = 5.55, J2 = 2.04, 1H); MS (EIPI) m/e 540 (M+).

Acknowledgments. Dr. Ton van Delft and Dr. Dirk Leysen (N.V. Organon, Oss, The Netherlands) are gratefully acknowledged for providing ORG13502 (14) and for

116 SAFIR and SAR of Ortho Substituted Phenylpiperazines generating the binding and (ant)agonist data presented in this chapter. We thank Dr. Max Lundmark and Dr. Staffan Sundell (Department of Structural Chemistry, University of Gothenburg, Sweden) for solving the X-ray structure of compound 16. We are also grateful to Marguerite Mensonides, who synthesized WAY100635 and its triflate analogue.

117 Chapter 6

6.6 References

[1] Glennon, R.A. Drug. Dev. Res. 1992, 26, 251. [2] Glennon, R.A.; Naiman, N.A.; Lyon, R.A.; Titeler, M. J. Med. Chem. 1988, 31, 1968. [3] Tricklebank, M.D.; Forler, C.; Middlemiss, D.N.; Fozard, R.Eur. J. Pharmacol. 1985, 117, 15. [4] Schoeffter, P.; Hoyer, D.Naunyn-Schmiederbergs Arch. Pharmacol. 1989, 339, 675. [5] Yocca, F.D.; Smith, D.W.; Hyslop, D.K.; Maayani, S.Soc. Neurosci. 1986, 422, Abstract 12. [6] Rydelek-Fitzgerald, L.; Teitler, M.; Fletcher, P.M.; Ismaiel, A.M.; Glennon, R.A.Brain Res. 1990, 532, 191. [7] Lanfumey, L.; Haj-Dahmane, S.; Hamon, M.Eur. J. Pharmacol. 1993, 249, 25. [8] Millan, M.J.; Rivet, J.-M.; Canton, H.; Le Marouille-Girardon, S.; Gobert, A. J. Pharmacol. Exp. Ther. 1993, 264, 1364. [9] Fletcher, A.; Bill, D.J.; Cliffe, I.A.; Forster, E.A.; Jones, D.; Reilly, Y.Br. J. Pharmacol. 1994, 112, 91P. [10] Lednicer, D.; Grostic, M.F.J. Org. Chem. 1967, 32, 3251. [11] Berendsen, H.H.; Broekkamp, C.J.; Van Delft, A,M.Eur. J. Pharmacol. 1990, 187, 97. [12] Dijkstra, G.D.H. Recl. Trav. Chim. Pays-Bas, 1992, 112, 151. [13] Tripos Associates, Inc., 1699 S. Hanley Rd.,Suite 303, St. Lious, Missouri 63144. [14] Stewart, J.J.P. J. Comp. Chem. 1990, 4, 1. [15] Gilli, G.; Bertolasi, V.J. Am. Chem. Soc. 1977, 101, 7704. [16] (a) Bachy, A. Steinberg, R.; Santucci, V.; Fournier, M.; Landi, M.; Hamon, M.; Manara, L.; Keane, P.E.; Soubrié, P.; Le Fur, G. Fundam. Clin. Pharmacol. 1993, 7, 487. (b) Cervo, L.; Bendotti, C.; Tarizzo, E.; Cagnotto, A.; Skorupska, M.; Mennini, T.; Samanin, R.Eur. J. Pharmacol. 1994, 253, 139. [17] Olivier, B.; Mos, J.; Rasmussen, D.Eur .J. Pharmacol. 1990, 8, 31. [18] Van Steen, B.J.; Van Wijngaarden, I.; Tulp, M. Th. M.; Soudijn, W.J. Med. Chem. 1994, 37, 2761. [19] Zhuang, Z.-P.; Kung, M.-P.; Kung, H.F.J. Med. Chem. 1994, 37, 1406.

118 Chapter 7

Selective 5-HT1A Receptor Ligands for PET; A Comparative Study of [11C]ORG13502 and [11C]WAY100635 in Normal and Adrenalectomized Rats*

Abstract

ORG13502 (6-{4-[4-(2-methoxyphenyl)-1-piperazinyl]butyl}-N-methyl- benzoxazolinone, 1) and WAY100635 (N-{2-[4-(2-methoxyphenyl)-1- piperazinyl]ethyl}-N-(2-pyridinyl)cyclohexane carboxamide, 2) are highly potent and 11 selective 5-HT1A receptor ligands. We prepared the C-analogues by methylation with 11 [ C]CH3I of the corresponding phenol piperazino precursors. The specific activities of [11C]ORG13502 and [11C]WAY100635 were >300 and 1000 Ci/mmol, respectively, after HPLC purification. Total synthesis times were 45 and 30 min, respectively, and the 11 radiochemical yields were ∼60% (from [ C]CH3I and decay-corrected). Tissue distribution studies in male Wistar rats revealed that the regional uptake of 11 11 [ C]WAY100635 after 60 min, but not of [ C]ORG13502 reflected the known 5-HT1A receptor distribution in the rat brain. Pretreatment with the selective 5-HT1A receptor agonist 8-OH-DPAT resulted in substantial blockade of [11C]WAY100635 uptake in 5-

HT1A receptor-rich brain regions (70-78% in raphe nuclei, frontal cortex, septum, hippocampus). Adrenalectomy (ADX, 1 or 6 days), which is known to cause 5-HT1A receptor upregulation in rats, had no significant effect on the uptake of [11C]WAY100635. However, the brain uptake of 11C after 24 h ADX was more sensitive to pretreatment with 8-OH-DPAT than in control animals in all examined brain areas, except for cerebellum.

7.1 Introduction

Central 5-HT1A receptors, existing both as somatodendritic autoreceptors at the raphe nuclei and post-synaptically, have been implicated in the pathogenesis of anxiety 1,2 and depression. Thus, acutely or chronically administered 5-HT1A receptor agonists all

* This chapter is based on: Barf, T.; Van Waarde, A.; Visser, G.M.; Medema, J.; Postema, F.; Korf, J.; Mensonides, M.M.; Wikström, H.; Korte, S.M.; Bohus, B.; Leysen, D.; Van Delft, A.M.L.; Vaalburg,W. Submitted.

117 Chapter 7 have therapeutic potency in treating these disorders. When given chronically they may 3 alter the 5-HT1A receptor density in various brain areas. Definitive conclusions on their mode of action can not be drawn because of the lack of appropriate pharmacological tools. A suitable procedure for visualization and quantification of central (and peripheral) 5-HT1A receptors as may be achieved with positron emission tomography (PET) is of great clinical interest. The radioligands developed and evaluated so far mostly were (partial) agonists which, due to unfavourable in vivo kinetic properties, failed as PET-ligands.

WAY100635 (2; Figure 7.1), a selective and silent 5-HT1A receptor antagonist 4 (IC50 value of 1.6 nM), has succesfully been labelled and evaluated as a potential in 5 vivo PET-imaging agent. Quantification of 5-HT1A receptors can be highly valuable for elucidating their role in the pathogenesis of various diseases such as depression and anxiety disorders.

OMe O O

N N N Me ORG13502 1 ()

N OMe N N N O WAY100635 2 ()

Figure 7.1. Chemical structures of ORG13502 and WAY100635

ORG13502 (1) is a highly potent and selective 5-HT1A receptor agonist (Ki = 0.25 nM) with low intrinsic activity (I.A. of 0.2) and therefore was considered a candidate ligand for labelling with a positron emitter.6 In order to compare ORG13502 and WAY100635, the 11C-labelled congeners were prepared by methylation with [11C]methyliodide of the corresponding ortho-hydroxyphenylpiperazines (Figure 7.2).

The second objective was to study the effect of changed 5-HT1A receptor densities on the biodistribution of 5-HT1A receptor ligands. In principle, brain Bmax in rats can be altered by adrenalectomy (ADX), which is known to cause an upregulation of the 5- 7,8 11 HT1A receptor subtype. Here we report the synthesis of [ C]ORG13502 and

118 Selective 5-HT1A Receptor Ligands for PET

[11C]WAY100635 and the results of biodistribution studies in rat brain in normal and adrenalectomized animals.

7.2 Chemistry

11 11 14 11 [ C]CH3I was produced from [ C]CO2 ( N (p,α) C nuclear reaction with 17 11 MeV protons) using an Anatech robotic system, yielding 15 GBq [ C]CH3I with a specific activity of more than 1000 Ci/mmol. According to a modified procedure as described by Elsinga et al,9 the 11C analogues of ORG13502 and WAY100635 were prepared by methylation with [11C]methyliodide of the corresponding phenols. The specific activities were >300 and >1000 Ci/mmol, respectively. In brief, a mixture of 11 [ C]CH3I, the phenol and potassium-t-butoxide in acetetonitrile was heated for 5 min in a cap-sealed tube at 110 °C (Figure 7.2). After purification by reversed-phase HPLC the desired compounds were obtained in a radiochemical yield of about 60% (from 11 [ C]CH3I, corrected for decay).

14 11 N (p,α) C

11 CO2

1) LiAlH4 2) HI 11 OH O CH3

11 CH3I N NR N NR tBuOK, CH3CN

Figure 7.2. Radiosyntheses of [11C]-o-methoxy-phenylpiperazines.

7.3 Pharmacology

Tissue distribution studies. A tail vein was catheterized with the rat under anaesthesia and after recovery the animals were kept under light restraint. The radioligands (100 µCi) was injected via the tail vein. Rats were killed after 60 min after injection. Brains were rapidly removed, nine regions sampled and the radioactivity was measured (expressed as a differential absorption ratio [DAR = (counts per min recovered/g tissue)/(counts per min injected/g body weight)]). Blocking experiments

119 Chapter 7 were performed with 8-OH-DPAT (Table 7.1). If required, animals were adrenalectomized (ADX) 1 or 6 days before the experiments (Table 7.3). Metabolism of [ 11C]ORG13502. Blood-samples (200 µL) were taken at different time intervals. After removal of the proteins, the supernatant was injected onto an HPLC-system. HPLC-samples were collected every 30 s and the radioactivity content was determined (Table 7.2).

7.4 Results and Discussion

Chemistry. The methylation reaction of desmethyl ORG13502 and WAY100635 could be performed conveniently in reasonable radiochemical yields of about 60%. Only minor amounts of by-products (probably due to N-alkylation) were observed and the radioligands could be separated easily from the precursors by HPLC. An unexpected difference of about 700 Ci/mmol in specific activities between the radiolabelled products was found in favor of [11C]WAY100635, although the reaction conditions and the position of labelling of both precursors were identical. No obvious explanation can be given for this observation. Advantageous in the synthesis of [11C]WAY100635 was the direct application of the reaction mixture on the semi- preparative column. The use of an ethanol/water mixture as the eluent, instead of methanol/ phosphate buffer, saved an additional evaporation step which had to be performed in the synthesis of [11C]ORG13502. These two ‘short-cuts’ resulted in a 15 min reduction of the total synthesis time of [11C]WAY100635.

120 Selective 5-HT1A Receptor Ligands for PET

Table 7.1. Distribution Studies in Rat Brain after 100µ Ci [11C]ORG13502 (1) or [11C]WAY100635 (2) Injection and Pretreatment with 8-OH-DPAT. [11C]-1 [11C]-1 [11C]-2 [11C]-2 % Reduction + + of [11C]-2 brain area 8-OH-DPAT 8-OH-DPAT binding Cerebellum 1.77±0.15 1.06±0.17* 0.11±0.01 0.12±0.02 - Striatum 2.03±0.21 1.31±0.15** 0.22±0.03 0.14±0.02* 36 Thalamus 1.80±0.22 1.21±0.28*** 0.29±0.06 0.20±0.09 40 Med. ol. 2.07±0.40 1.56±0.32*** 0.50±0.08 0.20±0.04* 57 Oc. cortex 2.34±0.40 1.16±0.23*** 0.65±0.08 0.23±0.02** 62 Fr. cortex 1.99±0.21 1.24±0.46*** 0.78±0.09 0.20±0.02** 74 Raphe Nuclei 2.27±0.58 1.56±1.34 0.91±0.15 0.25±0.09* 70 Septum 1.93±0.22 1.33±0.33* 1.39±0.19 0.37±0.08*** 74 Hipp. 1.73±0.30 1.07±0.25 1.55±0.37 0.16±0.06*** 78

The uptakes are expressed as differential absorbtion ratios (D.A.R.) , 60 min post-injection. Errors are in SEM. P<0.05 is denoted with *, P<0.01 with ** and P<0.005 with *** (Vehicle vs blocked). % of reduction of [11C]WAY100635 after 8-OH-DPAT pretreatment calculated from the brain area/cerebellum ratios (Table 7.3)

Pharmacology. The regional uptake of [11C]WAY100635, but not of 11 [ C]ORG13502, correlated with the known 5-HT1A receptor distribution in the rat brain (Tables 7.1 and 7.3).10 The ex vivo data obtained with rat brain membranes showed that 11 brain uptake of [ C]ORG13502, a highly potent and selective 5-HT1A receptor ligand, was homogeneous throughout the brain and partially reduced upon pretreatment with the 5-HT1A agonist 8-OH-DPAT. This reduction was also observed in cerebellum which 11 is a brain area essentially devoid of 5-HT1A receptors. At first, rapid formation of radioactive metabolites was thought to be the main cause of strong non-specific binding. Preliminary data on rat plasma, however, revealed that [11C]ORG13502 was only slowly metabolized; more than 50% of the parent compound still being present in plasma after 20 min (Table 7.2). Apparently, rapid metabolism is not the cause of the failure of [11C]ORG13502 as a radioligand. Other pharmacokinetic aspects of this compound have not been investigated and may be difficult to tackle. The lipophilicity of ORG13502 (logP value of 3.6* (3.4)#) is comparable with that of WAY100635 (3.3)#. The calculated logD values at pH 7.4 of both compounds were 3.0. All in all, this suggests that the lipophilicity of ORG13502 has no major contribution to the observed non-specific binding.

* Experimentally determined by N. V. Organon. # Calculated with Pallas 1.2 (CompuDrug Chemistry Ltd. (c) 1994)

121 Chapter 7

11 O 11 CH CH3 3 11 Table 7.2. Rate Nof Metabolism of H 11 OH CH2 [ C]ORG13502 N N N Time post inj. % Intact H Me O(min)H Parent 1 98.9 [11C]8-OH-DPAT3 () [11C]OSU191 4() [11C]HYMAP 5() 2 96.5 5 90.5 Figure 7.3. Chemical structures of 11[ C]8-OH-DPAT, 10 74.8 [11C]OSU191 and [11C]HYMAP. 20 51.8

It has been speculated that only pure 5-HT1A receptor antagonists are suitable as 12 11 ligands for PET, since radiolabelled 5-HT1A receptor agonists, such as [ C]8-OH- DPAT (3) 13 and [11C]OSU191 (4),14 are thought to compete unsuccesfully with the endogenous neurotransmitter (5-HT) for binding sites.15 However, contrasting results were disclosed by Thorell and co-workers who presented the 5-HT1A receptor agonist [11C](R)-10-methyl-11-hydroxyaporphine ([11C]HYMAP, 5) as a promising PET- ligand,16 although no biodistribution data were given. The agonist-receptor interaction seems to be relatively short in case of G- protein-coupled receptors (Figure 7.4). The formation of a ternary complex of an agonist, a receptor and an inactive, GDP-bound, G-protein facilitates the exchange of GDP by GTP. Hereafter, the agonist, receptor and G-protein rapidly dissociate resulting in free receptor subunits having a low affinity state, which can no longer accommodate agonist ligands.17 In contrast, antagonist radioligands have proven to be efficient through long duration and high affinity binding therefore allowing autoradiographic visualization and quantification of the specific labelling. In chapter 6, ORG13502 was found to be more potent in inducing the agonist effect (EC50 of 0.5 nM) than the antagonist effect (IC50 of 5 nM). WAY100635 was totally inactive in the agonist assay. This suggests that at the low concentration employed in these PET-studies, ORRG13502 rather may behave as an 5-HT1A receptor agonist. Therefore, it can not be excluded that [11C]ORG13502 failed as an in vivo PET-ligand due to its agonist properties and (consequently) the pharmacodynamic properties.

122 Selective 5-HT1A Receptor Ligands for PET

antag. ag.

inactive active

Figure 7.4.. Simple representation of the equilibrium between ‘active’ and ‘inactive’ receptors.

We found similar uptake patterns as Hume et al.5a and Pike et al.5b in rat brain after injection of [11C]WAY100635. At 60 min after injection, the ratio of radioactivity in 5-HT1A receptor-rich regions (e.g. septum and hippocampus) to that in cerebellum reached ca. 13 and 15, respectively. Substantial blockade of 11C uptake was achieved by pretreatment of rats with 8-OH-DPAT (Table 7.1). The higher the 5-HT1A receptor density in a particular brain area, the more effective was 8-OH-DPAT in blocking the [11C]WAY100635 uptake. Moderate blockade was observed in areas with low receptor density, such as striatum (36%), whereas a greater reduction of [11C]WAY100635 binding was observed in receptor-rich areas of the brain (78% in hippocampus). In order to study changes in receptor densities, rats were adrenalectomized 7,8 (ADX), which is known to cause upregulation of the 5-HT1A receptor. These authors found a ∼30% increase of hippocampal 5-HT1A receptor density after 1−7 days ADX utilizing autoradiography with [3H]8-OH-DPAT and in situ hybridization techniques. Surprisingly, in our experiments, none of the studied brain areas showed an increased uptake of [11C]WAY100635, as compared to normal rats after 1 or 6 days ADX (Table 7.3).

123 Chapter 7

Table 7.3. Distribution Studies in Rat Brain with [11C]WAY100635.

Vehicle Normal + ADX (24 h) ADX (24 h) + ADX (6 d) brain area 8-OH-DPATa 8-OH-DPATb Striatum 2.06±0.21 1.32±0.30 1.96±0.38 0.99±0.14 2.90±0.81 Thalamus 2.50±0.29 1.49±0.35 2.31±0.26 0.80±0.08* 3.16±0.54 Med. obl. 4.50±0.43 1.93±0.55 5.36±1.15 1.06±0.22* 5.10±0.52 Oc. cortex 5.92±0.41 2.23±0.44 6.60±0.95 1.27±0.22* 7.00±1.09 Fr. cortex 7.24±0.59 1.87±0.27* 6.57±0.81 1.44±0.20* 7.28±1.00 Raphe Nuclei 8.05±0.71 2.42±1.05* 5.78±0.71 1.69±0.38* 6.06±0.54 Septum 12.91±1.27 3.42±0.90* 10.34±2.11 1.66±0.46* 10.50±1.30 Hippocampus 14.58±1.13 3.22±0.28* 12.78±1.83 2.59±0.48* 14.79±2.73

The uptakes are expressed as brain area/cerebellum ratios, 60 min post-injection. Errors are in SEM. P<0.05 is denoted with * (a Vehicle vs blocked andb ADX (24 h) vs blocked ADX (24 h))

Interestingly, the brain uptake of 11C in animals after 24 hours ADX seemed more sensitive to pretreatment with 8-OH-DPAT than that in control animals in all examined brain regions, except for cerebellum, which suggests a shift from a low affinity state to a high affinity state, rather than an upregulation of the 5-HT1A receptor. However, the effect of ADX on brain uptake did not reach statistical significance due to large individual variances. Hume et al.5a checked the ‘specific’ signal of [3H]WAY100635 by pre-dosing the rats with compounds of known selectivity. Pretreatment with 8-OH- DPAT resulted in an average 77% reduction of the specific signal, however, the 8-OH- DPAT-insensitive binding corresponded regionally with both the specific signal and the 8-OH-DPAT-sensitive binding. Additionally, autoradiography studies revealed that [3H]WAY100635 could not discriminate between G-protein-coupled and G-protein- 18 3 uncoupled 5-HT1A receptors. The Bmax of [ H]WAY100635 specific binding sites was 50-60% higher than that of [3H]8-OH-DPAT in the same membrane preparations from various regions (hippocampus, septum, cerebral cortex).18a Furthermore, the relationship between the [3H]WAY100635 binding (total receptor density) and of [3H]8-

OH-DPAT binding (high affinity 5-HT1A binding sites only) in rat brain seems to depend upon the brain region.18b In other words: if adrenalectomy (or another disease state) causes a shift from a low affinity state to a high affinity state of 5-HT1A receptors, this can be detected with a radiolabelled 5-HT1A receptor agonist, but not with [3H]WAY100635 or [11C]WAY100635. 11 Unfortunately, the partial 5-HT1A receptor agonist, [ C]ORG13502, was found to be unsuitable for in vivo imaging of (central) 5-HT1A receptors. Probably, the intrinsic activity, or unfavourable in vivo kinetic properties, of this compound undermine its

124 Selective 5-HT1A Receptor Ligands for PET

ability to exert specific binding. In the present study, changed 5-HT1A receptor densities due to adrenalectomy did not result in altered brain uptake of the radioligand [11C]WAY100635. This may be due to the inability of [11C]WAY100635 to discriminate between the high affinity and low affinity state of 5-HT1A receptors. A radiolabelled 5-

HT1A receptor agonist may differentiate between these two affinity states.

7.5 Experimental Section

General. For general remarks see Section 2.4. Log D values were calculated with Pallas version 1.2. 11 11 14 11 Materials. [ C]CH3I was produced from [ C]CO2 ( N (p,α) C nuclear reaction 11 with 17 MeV protons) using an Anatech robotic system, yielding 15 GBq [ C]CH3I with a specific activity of more than 1000 Ci/mmol. The O-desmethyl precursors of ORG13502 and WAY100635 were prepared as described in Chapter 6 (section 6.5). 8- OH-DPAT (8-hydroxy-2-(N,N-di-n-propylamino)tetralin) was synthesized in our laboratory according to published procedures.19 Preparation of [ 11C]ORG13502. A solution of 1 mg (2.6 µmol; free base) desmethyl precursor and 0.3 mg (2.7 µmol) t-BuOK in 0.5 mL acetonitrile was prepared 11 at least 30 min before adding [ C]CH3I. The methyliodide was trapped in the reaction vessel at 0 °C after which the reaction mixture was heated at 110 °C for 5 min in an oil bath. After 1 min of cooling an aliquot of 50 µL of the solution was evaporated to dryness under reduced pressure at 50 °C. The residue was dissolved in 1.0 mL HPLC- eluent, which consisted of methanol/10 mM phosphate buffer pH 7.4 65/35 (v/v). The reaction mixture was applied on a C-18 Reversed Phase column (Chrompack; 150 × 4.6 mm, 5 µm). Using a flow rate of 2 mL/min, [11C]ORG13502 was eluted after 8 min. After evaporation of the eluent under reduced pressure at 50 °C, [11C]ORG13502 was dissolved in saline to prepare it for injection. The radioligand was obtained in a 11 radiochemical yield of ∼60% (from [ C]CH3I, corrected for decay) with a radiochemical purity > 99%. The total synthesis time was 45 min and the specific activity was > 300 Ci/mmol. Preparation of [ 11C]WAY100635. A similar procedure as for the synthesis of [11C]ORG13502 was employed. After heating the reaction mixture in an oil bath, an aliquot of 50 µL was dissolved in 1.0 mL HPLC-eluent (ethanol/water 55/45 (v/v)). The reaction mixture was applied on a semi-preparative C-8 Reversed Phase Column (Waters µBondapak; 300 × 7.8 mm, 5 µm) and [11C]WAY100635 was collected after 8 min, using a flow rate of 5 mL/min. [11C]WAY100635 was diluted with saline in order to prepare it for injection. The desired compound was obtained in a radiochemical yield of

125 Chapter 7

∼60% (from [11C]MeI, corrected for decay) with a radiochemical purity > 99%. The specific activity of [11C]WAY100635 was > 1000 Ci/mmol at the end of the 30-min radiosynthesis. Tissue distribution studies. Protocols of the animal experiments were approved by a local ethics committee as is prescribed by the Law on Animal Experiments of The Netherlands. Male Wistar rats weighing 200-250 g were used. If required, rats were adrenalectomized (ADX) 1 or 6 days before the experiments. Before administration of the radioligand, the rats were treated either with saline (control group n = 8) or with 0.5 mg/kg 8-OH-DPAT (blocking experiments, n = 4) by intravenous (iv) injection in the tail vein. After ca. 2 min, 100 µCi of the radioligand was injected in a volume of 0.3 mL saline. Rats were killed by decapitation 60 min after injection. The brain was rapidly removed and the uptake of 11C was measured in striatum, frontal cortex, occipital cortex, hippocampus, thalamus, medulla oblongata, cerebellum, raphe nuclei and septum. The amount of radioactivity was expressed as a differential absorption ratio [DAR = (counts per min recovered/g tissue)/(counts per min injected/g body weight)]. The concentration of the radioligand that was specifically bound was calculated as the [radioactivity content (brain tissue)]/[radioactivity content (cerebellum)]. Metabolism of [ 11C]ORG13502. A heart-catheterized rat was injected with 100 µCi [11C]ORG13502 under anaesthesia. Blood-samples (200 µL) were taken at different time intervals, diluted with acetonitrile (1/1 v/v) and centrifuged (10.000 g, 2 min). The supernatant was injected onto a HPLC-system using a Waters RCM C-18 column (100 × 8 mm, 5 µm) which was eluted with acetonitrile/ 65 mM acetate buffer pH 6.5 55/45 (v/v), flow rate 2 mL/min. HPLC-samples were collected every 30 s and the radioactivity content was determined with a LKB Compu Gamma counter (Table 7.2) Statistics. Differences between the 8-OH-DPAT treated group and the vehicle treated group were analyzed with the Student’s t-test (Table 7.1). Differences between the control groups and 8-OH-DPAT pretreated groups, in normal and ADX animals were analyzed using One Way Analysis of Variance (ANOVA) followed by Bonferroni’s t-test (Table 7.3).

Acknowledgments. Dr. Philip Elsinga and Ton Visser (PET-centre, University Hospital, Groningen, The Netherlands) are gratefully acknowledged for their assistance in operating the robotic system. We thank Dr. Durk Dijkstra for synthesizing 8-OH-DPAT.

126 Selective 5-HT1A Receptor Ligands for PET

7.6 References

[1] Traber, J.; Glaser, T.Trends Pharmacol. Sci. 1987, 8, 432. [2] Barrett, J.E.; Vanover, K.E. Psychopharmcol. 1993, 112, 1. [3] Fletcher, A.; Cliffe, I.A.; Dourish, C.T.Trends Pharmacol. Sci. 1993, 14, 441. [4] Forster, E.A.; Cliffe, I.A.; Bill, D.J.; Dover, G.M.; ones,J D.; Reilly, Y. Fletcher, A.Eur. J. Pharmacol. 1995, 281, 81. [5] (a) Hume, S.P.; Ashworth, S.; Opacka-Juffry, J.; Ahier, R.C.; Lammertsma, A.A.; Pike, V.W.; Cliffe, I.A.; Fletcher, A.; White, A.C.Eur. J. Pharmacol. 1994, 271, 515. (b) Pike, V.W.; Hume, S.P.; Ashworth, S.; Opacka-Juffry, J.; McCarron, J.A.; Cliffe, I.A.; Fletcher, A.Third IUPHAR Satellite Meeting on Serotonin, Chicago, USA, 1994, Poster no. 73. (c) Mathis, C.A.; Simpson, N.R.; Mahmood, K.; Kinahan, P.E.; Mintun, M.A.Life Sciences 1994, 55, 403. [6] Moussavi, Z.; Bonte, J.P.; Lesieur, D.; Leinot, M.; Lamar, J.C.; Tisne-Versailles, J.Farmaco. Ed. Sci. 1989, 44, 77. [7] Mendelson, S.D.; McEwen, B.S.Neuroendocrinol. Lett. 1990, 12, 353. [8] Chalmers, D.T.; Kwak, S.P.; Mansour, A.; Akil, H.; Watson, S.J.J. Neuroscience 1993, 13, 914. [9] Elsinga, P.H.; Van Waarde, A.; Visser, G.M.; Vaalburg, W.Nucl. Med. Biol. 1994, 21, 211. [10] Radja, F.; Daval, G.; Hamon, M.; Vergé, D.J. Neurochem. 1992, 58, 1338. [11] Matthiessen, L.; Daval, G.; Bailly, Y.; Gozlan, H.; Hamon, M.; Vergé D.Neuroscience 1992, 51, 475. [12] Laporte, A.-M.; Lima, L.; Gozlan, H.; Hamon, M.Eur. J. Pharmacol.1994 , 271, 505-514. [13] Thorell, J.-O.; Stone-Elander, S.; Ingvar, M. J. Label. Compounds Radiopharm. 1994, 35, 496. [14[ Halldin, C.; Wikström, H.; Swahn, C.-G.; Sedvall, G.; Stjernlöf, P.; Farde, L.J. Label. Compounds Radiopharm. 1994, 35, 494. [15] Kung; M.-P.; Zhuang, Z.-P.; Frederick, D.; Kung, H.F.Synapse 1994, 18, 359. [16] Thorell, J.-O.; Hedberg, M.H.; Johansson, A.M.; Hacksell, U.; Stone-Elander, S.; Eriksson, L.; Ingvar, M. J. Label. Compounds Radiopharm. 1996, 37, 314. [17] (a) Emerit, M.B.; El Mestikawy, S.; Gozlan, H.; Rouot, B.; Hamon, M.Biochem. Pharmacol. 1990, 39, 7. (b) Kobilka, B. Annu. Rev. Neurosci. 1992, 15, 87. [18] (a) Gozlan, H.; Thibault, S.; Laporte, A.-M.; Lima, L.; Hamon, M. Eur. J. Pharmacol., Mol. Pharmacol. Section 1995, 288, 173. (b) Khawaja, X; Brain Res. 1995, 573, 217. (c) Khawaja, X.; Evans, N.; Reilly, Y.; Ennis, C.; Minchin, M.C.W. J. Neurochem. 1995, 64, 2716. [19] Arvidsson, L.-E.; Hacksell, U.; Nilsson, J.L.G.; Hjorth, S.; Carlsson, A.; Lindberg, P.; Sanchez, D.; Wikström, H. J. Med. Chem. 1981, 24, 921.

127 Concluding Remarks

The novel, synthesized and tested aryl triflates constitute an interesting class of

5-HT1A and 5-HT1D receptor ligands, mostly exhibiting a high intrinsic activity profile. Although the carbon-skeletons of the compounds presented were not new, the aryl triflate concept seems to be valid for the type of compounds examined. The 5-triflate substituted tryptamines (Table 1; type A) were potent 5-HT1D receptor agonist with 1 preference for the 5-HT1Dα site. The N,N-dimethyl derivative 6 was active in a pig model, predictive of anti-migraine activity, and was indicated to have less propensity for coronary side effects, as compared to sumatriptan.2 Unfortunately, these compounds also displayed a pronounced affinity for 5-HT1A receptors, which may underlie the hypotensive effects of 6. Other 5-sulfonic acid ester derivatized tryptamines were found to have lower affinities for 5-HT1D receptors. Ethylamino side chain restriction, giving 2-aminotetralins (type B), resulted in fairly selective 5-HT1A receptor ligands, although still considerable affinity for 5-HT1Dα receptors was observed.3 The R-enantiomer of compound 8 proved to be the most potent

5-HT1A receptor agonist, inducing a full-blown 5-HT behavioural syndrome, along with a strong hypothermic effect in rats. In addition, (R)-8 possessed anxiolytic properties after acute administration to rats, however, (R)-8 also was found to have a low oral availability (7.6%). In an attempt to improve the oral bioavailability, the cis-1- methylated analogue (cis-9) and its enantiomers were prepared. These compounds were shown to be less efficacious 5-HT1A receptor ligands, with respect to their intrinsic efficacy, as compared to (R)-8. Other types of rigidifications are exemplified by compounds 10-12. The longer N-O distance is likely responsible for the observation that the 3-aminocarbazole (type

C) and the 4-indol-3-ylpiperidines (type D) exhibited a strong preference for the 5-HT1D receptor subtypes. Notably, the latter compounds were found to have a comparatively higher affinity for the 5-HT1Dβ receptors. It is of interest to subject compounds of type C and D to pharmacological assays, which are predictive of 5-HT1D receptor activity. ortho-Triflate substituted phenylpiperazine derivatives were shown to have higher intrinsic activity profiles than the ortho-methoxy and -hydroxy congeners. Presumbly, changes in the relative conformation between the phenyl and the piperazine

127 Concluding Remarks ring or an altered electrostatic potential of the aryl moiety, induced by the electron- withdrawing effect of the triflate accounts for this observation.

Table 1. Binding Results and Behavioural Pharmacology of the Novel Compounds. OTf OTf R H OTf OTf N NR2 NR

N N NH2 N H H H A B C D

Receptor Binding (Ki in nM)a Assay 5-HT Hypo-

Compd. Type R 5-HT1A 5-HT1Dα 5-HT1Dβ syndrome thermia 4 A propyl 23 190 246 5 A ethyl 27 12 171 6 A methyl 40 3.2 32 ++b 7 A H 18 2.8 14 8 B H 2.8 15 169 ++ (R)-8 B H 1.3 7.6 138 ++ ++c (S)-8 B H 13 157 1255 + cis-9 B methyl 6.1 15.7 125 − −c (1S, 2R)-9 B methyl 7.1 15 60 + +c (1R,2S)-9 B methyl 7.9 >1000 200 − −c 10 C - >1000 56 40 11 D H 71.4 24 7 12 D methyl 57.1 14 6

(a) Affinities for cloned mammalian receptors. (b) In guinea pigs. (c) In rats.

Taken together, concerning the affinity for the receptor subtypes examined, the triflate concept has proven to be successful for a number of classes of compounds. Future investigations will dictate the fate of these compounds. Depending on the nature of the drug-target, we believe that aryl triflates in themselves may provide a suitable basis for the development of novel drugs.

128 Concluding Remarks

References

[1] Barf, T.; De Boer, P.; Wikström, H.; Peroutka, S.J.; Svensson, K.A.; Ennis, M.D.; Ghazal, N.B.; McGuire, J.C.; Smith, M.W. J. Med. Chem. 1996, In press. [2] Saxena, P.R.; De Vries, P.; Heiligers, J.P.C.; MaassenVanDenBrink, A.; Bax, W.A.; Barf, T.; Wikström, H. Eur. J. Pharmacol. 1996, In press. [3] Sonesson, C.; Barf, T.; Nilsson, J.; Dijkstra, D.; Carlsson, A.; Svensson, K.; Smith, M.W.; Martin, I.J.; Duncan, J.N.; King, L.J.; Wikström, H.J. Med. Chem. 1995, 38, 1319.

129 Samenvatting

In 1948 werd er uit menselijk bloed een endogene stof geisoleerd. Deze stof bleek bloedvatvernauwende eigenschappen te bezitten en werd naar de bron (serum) en de activiteit (tonus), serotonine genoemd. Kort daarna, in het begin van de jaren ‘50, werd serotonine (5-hydroxytryptamine, 5-HT) in de hersenen gevonden en het werd duidelijk dat het daar een prominente rol als neurotransmitter vervulde. Ofschoon minder dan 5% van de totale hoeveelheid in het menselijke lichaam in het centrale zenuwstelsel (CZS) voorkomt, is 5-HT het werkpaard van de hersenen gebleken. Het speelt een rol in de regulatie van stemming, pijn, slaap, geheugen, sex, eetlust en emotie. In 1957 vonden wetenschappers, op basis van verschillende antagonerende effecten van dibenzyline en morfine, indicaties dat 5-HT niet op één, maar op meerdere receptor subtypen kon aangrijpen. Radioligand bindingstechnieken en de moleculaire biologie hebben er voor gezorgd dat er nu tenminste 15 subtypen 5-HT receptoren bekend zijn, die allemaal hun eigen plek in het CZS innemen. Op basis van farmacologie, moleculaire structuur en intracellulaire mechanismen zijn deze receptoren onderverdeeld in 7 families (5-HT1 t/m 5-HT7). Elke familie kan weer bestaan uit subtypen (bv. 5-HT1A t/m 5-HT1F). De meeste 5-HT receptor subtypen zijn gekloneerd en de aminozuurvolgorde is opgehelderd. Begrijpelijkerwijs wordt door de verscheidenheid aan 5-HT receptor subtypen het ontrafelen van de farmacologische functie van elk van deze receptoren bemoeilijkt. Er zijn selectieve 5-HT receptor agonisten en antagonisten nodig om elk van deze receptor subtypen functioneel te karakteriseren. Over één van de eerst ontdekte subtypen, de 5-HT1A receptor, is redelijk veel bekend. Deze receptor is veelvuldig in verband gebracht met depressie en angststoornissen, die mogelijk samenhangen met een verlaagde 5-HT neurotransmissie. De 5-HT1Dα en 5-HT1Dβ receptoren zijn recentelijk gekloneerd en ofschoon een ware inhaalrace begonnen is, valt er nog veel op te helderen. De 5-HT1D receptoren spelen een rol in de effectieve behandeling van migraine met de 5-HT1D receptor agonist sumatriptan (Imigran), maar over de exacte mechanismen wordt nog gespeculeerd. Tevens zou voor deze receptoren een rol in de regulatie van de eerder genoemde gemoedstoestanden weggelegd kunnen zijn.

129 Samenvatting

Het onderzoek in dit proefschrift beschrijft de synthese en de farmacologische evaluatie van nieuwe 5-HT1A en 5-HT1Dα/β receptor liganden, met potentiële therapeutische toepassingen in bovengenoemde afwijkingen. De structuur-affiniteit relaties (SAFIR) en structuur-activiteit relaties (SAR) staan centraal maar er wordt ook aandacht besteed aan de biologische beschikbaarheid van de nieuwe verbindingen. Een selectie van reeds bekende liganden die, al dan niet selectief, aangrijpen op de 5-HT1A,

5-HT1Dα en 5-HT1Dβ receptoren is beschreven in Hoofdstuk 1 .

Over de functie en de ligging van 5-HT1A receptoren in de hersenen is erg veel opgehelderd, bovendien bestaan er reeds selectieve en zeer potente 5-HT1A receptor agonisten, zoals 8-hydroxy-2-(di-n-propylamino)tetraline (8-OH-DPAT). Deze verbinding is echter niet klinisch toepasbaar omdat de farmacokinetische eigenschappen niet toereikend zijn. Hoofdstuk 2 behandelt de inspanningen om op basis van 8-OH-DPAT een verbinding te ontwikkelen die de selectiviteit en de potentie voor de 5-HT1A receptor behoudt, maar die een verbeterde orale beschikbaarheid bezit. De hydroxy-groep, die gevoelig is voor glucuronidering, werd gemaskeerd als de stabiele trifluoromethyl sulfonaat ester (triflaat). Een bijkomend voordeel van deze substituent is de elektronenzuigende eigenschap waardoor de aromatische ring t.o.v. 8- OH-DPAT moeilijker te oxideren is. Een andere reden voor de lage orale beschikbaarheid van de 2-di-n-propylaminotetralines was de afsplitsing van de één van de n-propyl staarten van het stikstof atoom. Om deze reden hebben we in onze serie verbindingen de N-monopropyl substitueerde 2-aminotetralines als uitgangspunt genomen. Van de nieuw gesynthetiseerde verbindingen werd de affiniteit bepaald, waarna de verbindingen met het interessantste profiel werden geëvalueerd aan de hand van 5-hydroxytryptofaan (5-HTP) accumulatie in de hersenen en met behulp van gedragsexperimenten. Met name het (R)-enantiomeer van 8-OSO2CF3-PAT bleek erg potent, maar de orale beschikbaarheid bleef laag (7.6%). Tevens werd een drastische toename in affiniteit voor de 5-HT1D receptoren waargenomen. Methylering van de 1- positie van het tetraline systeem (cis-8-OSO2CF3-1-Me-PAT) had een kleine afname van affiniteit voor de 5-HT1A en de 5-HT1Dα tot gevolg. Cis-(1S,2R)-8-OSO2CF3-1-Me- PAT bleek in de rat, zowel via de subcutane als orale toedieningsroute, de meest potente enantiomeer voor 5-HT1A receptoren te zijn. Het (1R,2S)-enantiomeer vertoonde weliswaar een lage effectiviteit maar een veel grotere selectiviteit voor de 5-HT1A receptor. De trans-analoga waren inactief.

Op basis van het farmacologisch profiel van (R)-8-OSO2CF3-PAT, werd dit 5-

HT1A receptor ligand getoetst op angstremmende eigenschappen d.m.v. gedragsmodellen met ratten. Hoofdstuk 3 beschrijft de acute werking van (R)-8-

OSO2CF3-PAT op het gedrag van de rat in het zgn. conditioned defensive burying, de

130 Samenvatting elevated plus-maze en het inescapable footshock model. Tevens werd het effect van de gebruikte doses op de 5-HT turnover in een tal van gehomogeniseerde hersengebieden gemeten. (R)-8-OSO2CF3-PAT vertoonde activiteit in de eerste twee modellen maar had geen invloed in het laatste model. De 5-HT turnover liet een significante daling zien in het limbische gedeelte van de hersenen. Hoofdstuk 4 behandelt de resultaten die worden verkregen als de N- methylaminosulfonylmethylene-groep van het antimigraine middel sumatriptan vervangen wordt door een triflaat-groep. Mede geïnspireerd door het positieve effect van de triflaat-substituent op de 5-HT1D affiniteit van 2-aminotetralines werden de SAFIR van tryptamines onderzocht. Een serie N,N-dialkyl gesubstitueerde 5-triflaat tryptamines werd gesynthetiseerd en getest op 5-HT1D receptor affiniteit en activiteit, dit laatste d.m.v. het forskoline gestimuleerde cAMP-inhibitie model. Primaire amines en verbindingen met kleine substituenten, zoals de N,N-dimethyl-groep, werden het beste getolereerd op het 5-HT1Dα en 5-HT1Dβ subtype en leverden ook de meest potente verbindingen op. Alle verbindingen vertoonden een middelmatige affiniteit voor de 5-

HT1A receptor. De meest selectieve verbinding, het N,N-dimethyl-5-triflaat- gesubstitueerde tryptamine, induceerde hypothermie en een verlaging van de 5-HT turnover in de hersenen van de cavia. De inactiviteit van deze stof voor de 5-HT1A receptor werd gestaafd middels 5-HTP accumulatie en intracerebrale microdialyse in ratten. De 2-aminotetralines die beschreven zijn in Hoofdstuk 2 geven al een indicatie dat selectiviteit kan worden verkregen door de ethylamino-keten van serotonine in een bepaalde positie te fixeren. In Hoofdstuk 5 worden andere restrictiemogelijkheden van de ethylamino-groep onderzocht. Zo zijn de synthesen beschreven van een triflaat- gesubstitueerde 3-aminocarbazool en 4-indol-3-ylpiperidine die, wat receptorbinding betreft, een sterke voorkeur voor de 5-HT1D receptor subtypen bleken te hebben. In dit hoofdstuk komen ook andere sulfonzure ester-gesubstitueerde tryptamines aan de orde, die allemaal een lagere affiniteit voor de 5-HT1D receptoren bleken te hebben dan het triflaat analoog.

ORG13502 en WAY100635 zijn 5-HT1A receptor antagonisten die beide een ortho-methoxyfenylpiperazine-structuur bevatten. In Hoofdstuk 6 worden de effecten van een methoxy-, hydroxy- of triflaat-groep op de ortho-posities van fenylpiperazines op de 5-HT1A receptor affiniteit en de intrinsieke activiteit vergeleken. Het getrifleerde analoog bleek een lagere affiniteit te hebben dan ORG13502 en WAY100635, alsmede een sterk verhoogde intrinsieke activiteit. Met behulp van molecular modelling en een crystal database search werd gepoogd een verklaring te vinden voor dit fenomeen.

131 Samenvatting

Tot slot behandelt Hoofdstuk 7 de radioactieve synthese van [11C]ORG13502 en in de literatuur eerder beschreven [11C]WAY100635 voor evaluatie in positron emissie tomografie (PET) studies. Het distributiepatroon van beide 11C gelabelde liganden werd onderzocht in rattehersenen, waarbij de [11C]ORG13502, anders dan [11C]WAY100635, een niet-specifieke opname liet zien. De biodistributiestudies werden herhaald met bijnierloze ratten, die een verhoogde 5-HT1A receptordichtheid zouden moeten vertonen. Ofschoon er kleine verschillen waarneembaar waren tussen de opname van [11C]WAY100635 in normale en bijnierloze ratten, bleken deze niet significant.

Dit onderzoek heeft een aantal zeer interessante 5-HT1A en 5-HT1D receptor agonisten opgeleverd. Uit vervolgstudies moet blijken of deze verbindingen een therapeutische toepassing hebben. Resumerend wordt in de Conclusies gesteld dat de electronenzuigende triflaat-groep een interessante bioisosteer is voor een aantal substituenten op een aryl-groep, die afhankelijk van het ligand en de receptor met succes toegepast zou kunnen worden in toekomstige medicijnen.

132 Toelichting voor niet-Farmacochemici

Elk menselijk lichaam wordt aangestuurd door de hersenen via zenuwbanen waardoor electrische prikkels lopen. De zenuwbanen zijn niet oneindig lang maar hebben vertakkingen en onderbrekingen. Als een electrische prikkel bij zo’n onderbreking (ook wel synaps genoemd) aankomt moet het over worden gedragen naar de zenuwcel. Serotonine (de chemische naam is 5-hydroxytryptamine en de afkorting is 5-HT) is een lichaamseigen stof die o.a. werkzaam is in de hersenen. Deze chemische moleculen (neurotransmitter) fungeren als een soort overslagbedrijf van electrische prikkels. Op het moment dat er een prikkel arriveert worden er serotonine moleculen losgelaten, die aan de andere kant van de onderbreking een aangrijpingspunt, in de vorm van een eiwit, op een andere zenuwcel vinden. Hierdoor vindt er weer een activering van een electrische prikkel of van een biochemisch proces plaats. Die aangrijpingpunten zijn de ontvangers van de neurotransmitters en heten dan ook receptoren. Als de hoeveelheden van de neurotransmitters en de receptoren in de hersenen goed op elkaar afgestemd zijn kan een lichaam normaal functioneren. Echter, als er een tekort of een overschot ontstaat van de één ten opzichte van de ander dan kan het zijn dat ook de overslag van electrische prikkels in het gedrang komt, waardoor bepaalde ziektebeelden kunnen ontstaan. We zouden kunnen proberen om een extra hoeveelheid 5-HT in de vorm van een pil via de bloedbaan naar de hersenen te loodsen, en zo de verstoorde verhouding recht te trekken. Maar de hersenen zijn verpakt in een beschermend vlies (de bloed-hersen- barriere) en die laat 5-HT niet door. De chemische structuur van 5-HT is gelukkig bekend. We zijn in staat om chemische afgeleiden te maken die de zelfde werking hebben als serotonine, maar wel de bloed-hersen-barriere kunnen passeren. Het probleem is dat de receptoren waar serotonine precies op past niet allemaal identiek zijn. Afhankelijk van de plaats in de hersenen zijn er tenminste vijftien familieleden (serotonerge (5-HT) receptoren) waar serotonine goed aan bindt. 5-HT is dus niet selectief. Chemische afgeleiden, mits goed gekozen, kunnen wel onderscheid maken binnen de 5-HT familie van receptoren en kunnen in principe beter specifieke balansverstoringen, en dus bepaalde ziektebeelden, verhelpen.

133 List of publications

Sonesson, C.; Barf, T.; Nilsson, J.; Dijkstra, D.; Carlsson, A.; Svensson, K.; Smith, M.W.; Martin, I.J.; Duncan, J.N.; King, L.J.; Wikström, H. Synthesis and Evaluation of Pharmacological and Pharmacokinetic Properties of Monopropyl Analogs of 5-, 7- and 8-[[(Trifluoromethyl)sulfonyl]oxy]-2-aminotetralins. J. Med. Chem. 1995 , 38, 1319- 1329.

Barf, T.; Korte, S.M.; Korte-Bouws, G.; Sonesson, C.; Damsma, G.; Bohus, B.; Wikström,

H. Potential anxiolytic properties of R-(+)-8-SO2CF3-PAT, a 5-HT1A receptor agonist. Eur. J. Pharmacol. 1996 , 297, 205-211.

Osman, S.; Lundkvist, C.; Pike, V.W.; Halldin, C.; McCarron, J.A.; Swahn C.G.; Ginovart, N.; Luthra, S.K.; Bench, C.J.; Grasby, P.M.; Wikström, H.; Barf, T.; Cliffe, I.A.;

Fletcher, A.C.; Farde, L. Characterization of the metabolites of the 5-HT1A receptor radioligand, [O-methyl-11C]WAY-100635, in monkey and human plasma - A comparison of the behaviour of an identified radioactive metabolite with parent radioligand using PET. Nucl. Med. Biol. 1996 , 23, 627-634.

Barf, T.; De Boer, P.; Wikström, H.; Peroutka, S.J.; Svensson, K.A.; Ennis, M.D.; Ghazal,

N.B.; McGuire, J.C.; Smith, M.W. 5-HT1D Receptor Agonist Properties of 5- [[(Trifluoromethyl)sulfonyl]oxy]indolylalkylamines and their use as Synthetic Intermediates. J. Med. Chem. 1996 , In press.

Saxena, P.R.; De Vries, P.; Heiligers, J.P.C.; MaassenVanDenBrink, A.; Barf, T.; Wikström, H. Investigation with GMC2021, a triflated analogue of sumatriptan, in experimental models predictive of antimigraine activity and coronary side-effect potential. Eur. J. Pharmacol. 1996 , In press.

Barf, T.; Van Waarde, A.; Visser, G.M.; Mensonides, M.M.; Medema, J.; Postema, F.; Korte, S.M.; Bohus, B.; Korf, J.; Wikström, H.; Leysen, D.; Van Delft, A.J.M.; Vaalburg, 11 W. Selective 5-HT1A receptor ligands for PET: A comparative study of [ C]WAY- 100635 and [11C]ORG-13502 in normal and adrenalectomized rats. Submitted.

Hall, H.; Lundkvist, C.; Halldin, C.; Farde, L.; Pike, V.W.; McCarron, J.A.; Fletcher, A.;

Cliffe, I.A.; Barf, T.; Wikström, H.; Sedvall, G. Autoradiographic localization of 5-HT1A

137 receptors in post-mortem using [3H]WAY-100635 and [11C]WAY-100635. Submitted.

Abstracts

Barf, T.; Wikström, H.; Sonesson, C.; Svensson, K. R-(+)-8-OTf-PAT, a new 5-HT1A agonist with good oral availability. 9th Noordwijkerhout-Camerino Symposium, May 23-27, 1993 , Noordwijkerhout, The Netherlands.

Barf, T.A.; Wikström, H.V.; Sonesson, C.; Svensson, K.; Martin, I.J.; Duncan, J.N. Trifluoromethanesulfonic acid esters of phenols increase metabolic stability, as exemplified by R-8-TfO-PAT, a potent 5-HT1A agonist. Third IUPHAR satellite meeting on serotonin, July 30-August 3, 1994 , Chicago, U.S.A.

Barf, T.A.; Wikström, H.V.; Peroutka, S.J. Novel serotonergic ligands with selectivity for the 5-HT1D receptor subtype. XIIIth international symposium on medicinal chemistry, September 19-23, 1994 , Paris, France.

11 Barf, T. Selective 5-HT1A receptor ligands for PET: A comparative study of [ C]WAY- 100635 and [11C]ORG-13502. SON Medicinal Chemistry meeting, April 19-20, 1995 , Lunteren, The Netherlands (oral presentation).

Barf, T.; Visser, G.M.; Van Waarde, A.; Korte, S.M.; Postema, F.; Leysen, D.; Van Delft, A.J.M.; Wikström, H.; Bohus, B.; Korf, J.; Vaalburg, W. Synthesis and biodistribution of 11 th [ C]ORG-13502, a high-affinity serotonin (5-HT1A) receptor ligand. 42 Annual meeting of the society of nuclear medicin, June 12-15, 1995 , Minneapolis, Minnesota, U.S.A. ( J. Nucl. Med. 1995 , 36, 163P).

Barf, T.A.; Visser, G.M.; Van Waarde, A.; Medema, J.; Mensonides, M.M.; Korte, S.M.; Postema, F.; Leysen, D.; Van Delft, A.J.M.; Wikström, H.; Bohus, B.; Korf, J.; Vaalburg, 11 W. Selective 5-HT1A receptor ligands for PET: A comparative study of [ C]WAY- 100635 and [11C]ORG-13502. 11th international symposium on radiopharmaceutical chemistry, August 13-18, 1995 , Vancouver, Canada. (J. Label. Compd. Radiopharm. 1995 , 37, 280).

138 Tot slot....

.....wil ik een aantal mensen bedanken voor de goede tijd of omdat ze op enigerlei wijze hebben bijgedragen tot de totstandkoming van dit proefschrift. Ik heb een stel fijne mensen leren kennen en ik hoop een aantal van jullie nog vaak te zien. De Farmacochemie is een multidisciplinair vakgebied en ik denk dat ik hieraan te danken heb, dat ik de afgelopen vier jaar met zoveel mensen heb mogen samenwerken. Allereerst mijn promotor Prof. Håkan Wikström. Jij hebt mij de eerste stappen in de Farmacochemie leren zetten. Ik heb je enthousiaste begeleiding als zeer prettig ervaren. De discussies gingen altijd in goede sfeer en je hebt mij de ruimte gegeven om mijn eigen weg te bewandelen. Met name dit laatste heb ik erg belangrijk gevonden. I’m very grateful to the reading-committee, constituted by Prof. Ben Feringa, Prof. Berend Olivier and Prof. David Nichols, for the rapid and thorough correction of this manuscript. Hier wil ik ook co-promotor Cor Grol noemen die in zijn eentje de alternatieve leescommissie belichaamde. Van Cor en de andere ‘doctors’, Ben Westerink, Durk Dijkstra, wijlen Geert Damsma, Wia Timmerman en Peter de Boer, heb ik de afgelopen jaren met betrekking tot de Farmacochemie zeer veel wetenwaardigheden mogen vernemen, waarvoor mijn dank. Peter de Boer krijgt een eervolle vermelding omdat hij er voor gezorgd heeft dat een deel van de farmacologische evaluatie (de statistiek incluis) van mijn verbindingen bij huis gedaan kon worden. Dat ik daarbij mocht helpen was alleen maar leuk. Ook Jan de Vries wil ik bedanken voor het aandragen en uitvoeren van allerlei ad hoc oplossingen op dit gebied. Het grootste gedeelte van de tijd heb ik natuurlijk op de synthese-zaal doorgebracht. Al mijn zaalgenoten van voormalig 1-32 en het gloednieuwe 436 krijgen een grote pluim voor de gezellige uurtjes binnen en buiten het lab. Met dat laatste doel ik vnl. op de afsluiter van de week in “De Toeter.” Met Jonas, Evert en Ulrike hoop ik in de toekomst ook nog eens een keutje te kunnen leggen. Mijn ‘zuurkastbuur’ en kamergenoot Sander wil ik speciaal bedanken voor zijn onderhoudendheid en de nuttige commentaren op de publicaties in spé. Mijn andere twee wisselende kamergenoten, Nienke en Eytan, waren eveneens aangename aanspreekpunten. Marguerite ben ik zeer erkentelijk voor haar bijdrage aan Hoofdstuk 6 en 7, nl. de synthese van WAY100635 en het triflaatanaloog. Pieter verdient een grote schouderklop voor zijn inbreng in het reilen en zeilen van de labzaal. Yi dank ik voor zijn vele tips op synthesegebied. Speciale dank ben ik verschuldigd aan de keuze- en bijvakstudenten Marianne Deinum, Wouter Brink en Arjen Bouter. De één had wat meer geluk dan de ander in de synthese, maar ik heb jullie inzet en aanwezigheid in ieder geval erg op prijs gesteld. De hulp van computer-monteuren, zoals Evert, Jonas en Willem Jan, is in een modern laboratorium onontbeerlijk, bedankt! Janita en Janneke van het secretariaat ben ik zeer erkentelijk voor hun hulp bij alle administratieve rompslomp. Ook alle collega’s die reeds ‘afgezwaaid’ zijn, wil ik bedanken voor de prettige dagelijkse omgang. Andries Bruins en Margot Jeronimus van de afdeling Massa Spectrometrie bedank ik voor het nemen van de massa’s. Ook Albert Kiewiet van OMASCH wil ik hieraan toevoegen. Het NMR-service team dank ik voor het verzorgen van de NMR-spectra. Alle mensen van de ondersteunende diensten (een beetje te veel om allemaal persoonlijk te noemen) wil ik bedanken voor hun bijdrage op velerlei gebied. Wel wil ik vermelden dat ik de persoonlijke ‘touch’ van Connie van Donselaar zeer op prijs heb gesteld. Harm Metting dank ik voor het aanreiken van de ‘last minute’ oplossing voor de omslag. Tijdens het promotieonderzoek heb ik, dicht bij huis, twee uitstapjes gemaakt die bewijzen dat ‘er niets boven Groningen gaat.’ De immer enthousiaste Mechiel Korte, Gerdien Korte-Bouws en Prof. Béla Bohus wil ik bedanken voor de onvergetelijke tijd bij de afdeling Dierfysiologie in Haren. Op het PET-centrum in Groningen kreeg ik ‘heet’ onthaal in de vorm van een gezonde dosis radiochemie. Voor deze zeer leerzame en leuke periode wil ik Geb Visser, Aren van Waarde, Folkert Postema, Jitze Medema, Philip Elsinga, Ton Visser, Prof. Wim Vaalburg en de rest van hun PET-collega’s bedanken. Met betrekking tot dit uitstapje wil ik ook noemen dat ik Dirk Leysen en Ton van Delft van de N.V. Organon zeer dankbaar ben voor de verstrekking van de nodige farmacologische data en de back-up vanuit Oss. With respect to the long-distance collaborations I would like to thank Class Sonesson from the Department of Pharmacology in Göteborg for the pleasant exchange of pharmacological data on our 2-aminotetralins (nästa gång vi syns i Göteborg, ska vi svinga en bägare ihop!), Stephen Peroutka of Spectra Biomedical Inc. (San Francisco, CA) for providing binding data on the indolealkylamines, and with respect to the same class of compounds, Kjell Svensson and his colleagues at Pharmacia & Upjohn Inc. (Kalamazoo, MI) for their very collaborative attitude. Also, I would like to acknowledge Peter Pauwels and his co-workers at Centre de Recherche Pierre Fabre

(Castres, France) for providing many of the 5-HT1D binding data. I’m also grateful to Shelly Glase at Parke-Davis (Ann Arbor, MI) for providing the elemental analyses of my compounds. Prof. Pramod Saxena van de Erasmus Universiteit te Rotterdam ben ik zeer erkentelijk voor het testen van GMC2021 in zijn antimigraine-model. Tijdens de vrije momenten en de moeilijke periodes heb ik de respectievelijke afleiding en steun van mijn vrienden enorm gewaardeerd. Eigenlijk had ik jullie allemaal als paranimf willen hebben maar uiteindelijk kan ik alleen Arwin en Lynet bedanken voor het feit dat jullie mijn ‘secondanten’ wilden zijn. Rian wil ik speciaal bedanken voor haar steun tijdens het promotieonderzoek, dat zij voor een groot gedeelte van dichtbij heeft meegemaakt. Ik mag wel zeggen dat ik mij geen betere familie had kunnen wensen. Gert, Annemarie, Hans en Edith, bij jullie voel ik mij thuis en dat is steeds erg belangrijk voor mij geweest. Een hele speciale plek wordt ingenomen door mijn ouders. Ook al hebben jullie weinig invloed gehad op wat ik doe, ik heb grotendeels aan jullie te danken wie ik ben. Het idee dat jullie achter mij staan, is een onbeschrijflijk goed gevoel! Tot slot Uli, jou wil ik niet alleen bedanken voor je wetenschappelijke bijdrage, maar temeer voor het feit dat ik de laatste tijd de ‘stressor’ was en jij mijn ‘serenic’ wilde zijn!

Tjeerd

141 Toelichting voor niet Farmacochemici

Het onderzoek, beschreven in dit proefschrift, is toegespitst op het maken van chemische stoffen die selectief zouden moeten aangrijpen op één van deze receptoren. De binding van deze gesynthetiseerde stoffen werd d.m.v. weefsel kweken getest op drie receptoren (de 5-HT1A, 5-HT1Dα en 5-HT1Dβ receptor), die grote structurele overeenkomsten vertonen. Een verstoring van de balans met elk van deze receptoren wordt in verband gebracht met o.a. depressie en angststoornissen. Het 5-HT1Dβ receptor subtype speelt waarschijnlijk een rol in de effectieve behandeling van migraine met Imigran®. Dit antimigraine-middel werkt net buiten de hersenen, maar er zijn aanwijzingen dat een extra werking kan worden verkregen met medicijnen die ook de hersenen penetreren.

OH OH OH

NH2 NH2

N N NH2 H H 5-Hydroxytryptamine,1 8-Hydoxy-2-aminotetraline,2 3-Amino-6-hydroxycarbazool,3 5-Hydroxytryptamine ( 1) 8-Hydroxy-2-aminotetraline ( 2) 3-Amino-6- hydroxycarbazool ( 3)

Hierboven zijn de chemische structuren van respectievelijk 5-HT (1), 8-hydroxy- 2-aminotetraline (2) en 3-amino-6-hydroxycarbazool (3) afgebeeld. De chemische stoffen bezitten allen een aminogroep (NH2) en een zesring met een hydroxygroep (OH). Het verschil zit in de ethylaminostaart (vet weergegeven), die vrij kan bewegen in 5-HT maar gefixeerd is in molecuul 2 en 3. Men heeft gevonden dat o.a. de afstand van het O-atoom (zuurstofatoom) naar het N-atoom (stikstofatoom) bepaalt, hoe goed een 5- HT-afgeleide aan een bepaald receptor subtype bindt. Door deze afstanden vast te leggen kan selectiviteit voor één van de receptor subtypen verkregen worden. Simpel gezegd komt het er op neer dat structuur 2 een O O voorkeur heeft voor de 5-HT receptor en structuur 3 F S 1A OH C O F voor het 5-HT1Dα en 5-HT1Dβ receptor subtype. Het F probleem is echter dat dit type chemische stoffen te snel afgebroken (gemetaboliseerd) wordt in het lichaam. Door de voor metabolisme gevoelige stukken 'aryl hydroxy' 'aryl triflaat' van het molecuul te beschermen of er een andere groep voor in de plaats te zetten kunnen in principe chemische stoffen ontwikkeld worden die niet zo snel metaboliseren en dus effectiever als medicijn zouden kunnen

134 Toelichting voor niet Farmacochemici werken. Zo’n beschermgroep moet chemisch en biologisch stabiel zijn en een voorbeeld daarvan is de aryl trifluoromethaansulfonaatgroep (triflaatgroep). Deze groep hebben we vnl. onderzocht door bestaande chemische structuren te synthetiseren en van een triflaatgroep te voorzien. Daarna werden ze getest op binding voor de verschillende 5-HT receptor subtypen. Om een idee te krijgen hoe deze stoffen in de mens zullen werken, zijn deze stoffen getest op proefdieren (rat en cavia). Chemische stoffen, zoals 2, die de 5-HT1A receptor activeren (agonisten genoemd) veroorzaken bij de rat het zgn. 5-HT gedragssyndroom, wat gekenmerkt wordt door een platte lichaamshouding (een Citroën-DS in de parkeerstand), een teruggetrokken onderlip (pruillip) en trappelen met de voorpootjes (piano spelen). Tevens veroorzaken deze stoffen een verlaging van de lichaamstemperatuur (hypothermie) bij de rat.

Daarentegen induceren 5-HT1D receptor agonisten weer hypothermie bij de cavia. Alleen de gesynthetiseerde 5-HT-afgeleiden die een goed bindingsprofiel lieten zien werden geëvalueerd in bovenstaande experimenten. De meest interessante verbindingen, beschreven in dit proefschrift, staan in Tabel 1. De triflaatgroep is als ‘OTf’ weergegeven en ‘R’ is een willekeurige groep.

Tabel 1. Bindingsresultaten en Gedragsfarmacologie van 5-HT-afgeleiden OTf OTf R H OTf OTf N NR2 NR

N N NH2 N H H H A B C D

Receptor Bindinga Test 5-HT Hypo-

Stof Type R 5-HT1A 5-HT1Dα 5-HT1Dβ syndroom thermie 4 A propyl 23 190 246 5 A ethyl 27 12 171 6 A methyl 40 3.2 32 ++b 7 A H 18 2.8 14 8 B H 2.8 15 169 ++ (+)-8 B H 1.3 6.7 138 ++ ++c (−)-8 B H 13 157 1255 + 9 B methyl 6.1 15.7 125 − (+)-9 B methyl 7.1 12 60 + +c (−)-9 B methyl 7.9 >1000 200 −

135 Toelichting voor niet Farmacochemici

10 C - >1000 56 40 11 D H 71.4 24 7 12 D methyl 57.1 14 6

(a) De getallen geven de concentratie (in nM) weer die nodig was om een radioactief gemerkte stof van de receptor te verdringen. Hoe lager het getal, hoe beter de binding. (b) Bij de cavia. (c) Bij de rat.

In de serie verbindingen die structureel overeenkomt met 5-HT (4-7; type A), is te zien dat de grootte van de substituent op het N-atoom van invloed is op de bindingsresultaten. De propyl-groep is groter dan de ethyl en deze is weer groter dan de methyl, enz. Kleine groepen werden dus goed getolereerd in de 5-HT1D receptor subtypen. Verbinding 6 was de meest selectieve verbinding van deze serie en werd actief bevonden in een hypothermie-experiment met cavia’s. De verbindingen 10-12 (type C en D) zijn niet op gedragsfarmacologie getest, maar lieten wel een sterke voorkeur voor de 5-HT1D receptor subtypen zien. Verbinding 10 vertoonde zelfs helemaal geen binding met de 5-HT1A receptor. De verwachting is dat op basis van deze structuren 10-12 selectieve 5-HT1Dα en 5-HT1Dβ receptor agonisten ontwikkeld kunnen worden.

Moleculen van het type B bezitten C- spiegel atomen (koolstofatomen) met vier verschillende groepen eraan, gemarkeerd met OTf OTf een sterretje in de figuur hiernaast (verbinding H R R H H H 8 heeft er één en verbinding 9 heeft er twee). C* *C N N C C Dat betekent dat verbindingen 8 en 9 uit twee H * * H spiegelbeelden (enantiomeren, (+) en (−)) bestaan, die chemisch gelijk, maar farmacologisch verschillend zijn. Omdat de (+) (−) ene enantiomeer een gunstige werking en de andere een bijwerking zou kunnen hebben, hebben we deze spiegelbeelden van elkaar gescheiden. Verbinding (+)-8 had een sterkere voorkeur voor de 5-HT1A receptor dan (− )-8 en werd onderworpen aan gedragstesten met de rat. Structuur (+)-8 was zeer potent in het induceren van het 5-HT gedragssyndroom en van hypothermie. Helaas bleek de orale beschikbaarheid van deze stof (te meten door de concentraties in het bloed na orale en intraveneuze toediening te vergelijken) slechts 7.6%, wat betekent dat deze verbinding niet erg geschikt is om in een pil te verwerken. Uit deze studies bleek wel dat de aryl triflaatgroep intact bleef. Om de orale beschikbaarheid te verbeteren, moet dit soort moleculen dus elders worden gemodificeerd. In een poging, hebben we een extra methyl-groep op het molecuul gezet (9). Wederom was (+)-9 de meest potente van

136 Toelichting voor niet Farmacochemici de twee enantiomeren. Uit hypothermie-experimenten bleek dat de orale beschikbaarheid vergeleken met (+)-8 wel iets toegenomen was, maar de binding en de potentie voor de 5-HT1A receptor was lager. Op basis van de bindingsresultaten kunnen we concluderen dat de gesynthetiseerde aryl triflaten een interessante groep stoffen vormt. De toekomst zal uitwijzen of een aantal van deze stoffen zijn weg naar de kliniek vindt. Of het ‘triflaatconcept’ werkt voor andere potentiële medicijnen zal afhangen van de aard van de doelmoleculen en hun receptoren.

137