Characterization of structure-activity relationships for serotonin receptors coupled to adenylate cyclase

Item Type text; Dissertation-Reproduction (electronic)

Authors Cornfield, Linda Joan.

Publisher The University of Arizona.

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

Download date 07/10/2021 21:57:33

Link to Item http://hdl.handle.net/10150/185101 INFORMATION TO USERS

The most advanced technology has been used to photograph and reproduce this manuscript from the microfilm master. UMI films the text directly from the original or copy submitted. Thus, some thesis and dissertation copies are· in typewriter face, while others may be from any type of computer printer.

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

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

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

Photographs included in the original manuscript have been reproduced xerographically in this copy. Higher quality 6" x 9" black and white photographic prints are available for any photographs or illustrations appearing in this copy for an additional charge. Contact UMI directly to order.

U·M·I UnlV£:~rslly Microfilms Inlernatlonal A Bell & Howell Informal Ion Company 300 NOrlh Zeeb Road. Ann Arbor. M148106-1346 USA 3131761-4700 800/521-0600

Order Number 9032525

Characterization of structure-activity relationships for serotonin receptors coupled to adeny late cyclase

Cornfield, Linda Joan, Ph.D.

The University of Arizona, 1990

V·M·I 300 N. Zccb Rd. Ann Arbor, MI 48106

CHARACTERIZATION OF STRUCTURE-ACTIVITY RELATIONSHIPS FOR

SEROTONIN RECEPTORS COUPLED TO ADENYLATE CYCLASE

by

Linda Joan Cornfield

A Dissertation Submitted to the Faculty of the

COMMITTEE ON PHARMACOLOGY AND TOXICOLOGY (GRADUATE)

in Partial Fulfillment of the Requirements For the Degree of

DOCTOR OF PHILOSOPHY

in the Graduate College

THE UNIVERSITY OF ARIZONA

1 990 2

THE UNIVERSITY OF ARIZONA GRADUATE COLLEGE

As members of the Final Examination Committee, we certify that we have read the dissertation prepared by ------Linda Joan Cornfield entitled Characterization of Structure-Activity Relationships for

Serotonin Receptors Coupled to Adenylat~ Cyclase.

and recommend th~t it be accepted as fUlfilling the dissertation requirement for the Degree of Doctor of Philosophy

Date

Date 2j?Jd()

Date ' s-b//;-() Dat~ I

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

I hereby certify that I have read this dissertation prepared under my direction and recommend that it be accepted as fulfilling the dissertation requirement. I~i /)(1-// / ft~~"'" Dissertation Director David L. Nelson Date 3

STATEMENT BY AUTHOR

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

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

Signed: 4

ACKNOWLEDGEMENTS

I would like to express my sincere appreciation to Dr. David Nelson

for all his guidance, support and encouragement. I would like to thank

Drs. Hugh E. LaIrd II, Dan C. Liebler, James R. Halpert and IIenry I.

Yamamura, the original members of my committee, for their time and

advice. I would also like to thank Dr. John W. Regan for substituting as a committee member during the last few months leading up to my

defense. would lIke to express my deep personal gratitude to Georgina

Lambert for all she has done for me, both in and outside the l~b. I would like to thank Drs. Arnold Martin, Will Taylor, Sham Nikam, Torsten

Dahlgren, Youhua Yang and Uli Hacksell for synthesizing many interesting novel cumpounds. Special thanks go out to the following: Drs. Glen[}

Sipes, Phil Monroe and Anne Killam, as well as Janet Firriolo, Alice

Villalobos, Pam Murray, Richard Wiedhopf, Judith Stanfield, Scott

Mobley, Dale Vander Hamm and Peter Gabriel. Last but not least, would

like to thank my parents and brother for their love, support and encouragement over the years. 5

TABLE OF CONTENTS

PAGE

LIST or ILLUSTRATIONS ...... 7

LIST OF TABLES ...... 10

ABSTRACT...... 11

1 INTRODUCTION ...... 13 1.1 The Discovery of Serotonin ...... 13 1.2 5-HT Receptor Heterogeneity ...... 15 1.3 5-HT and Adenylate Cyclase - The Early years ...... 16 1.4 Current Status of 5-HT Receptors Coupled to Adenylate Cyc lase...... 22 1.4.1 5-HT lA receptors ...... 23 1.4.2 5-HT lB receptors ...... 29 1.4.3 5-HTlD receptors ...... 30 1.4.4 5-HT4 receptors ...... 31 1.5 Determination of Structure-Activity Relationships for 5-HTIA Receptors - Goals of This Research ...... 32

2 USE OF THE FORSKOLIN-STIMULATED ADENYLATE CYCLASE ASSAY AS A SCREEN FOR COMPOUNDS POSSESSING POTENTIAL 5-HT 1A ACTIVITY. . .. 36 2.1 Introduction ...... 36 2.2 Methods...... 37 2.2.1 Chemicals ...... 37 2.2.2 Tissue preparation ...... 38 2.2.3 Forskolin-stimulated adenylate cyclase (FSC) assay ...... 40 2.2.4 5-HTIA receptor binding assay ...... 41 2.2.5 Data analysis ...... 42 2.3 Results ...... 44 2.3.1 Response of known 5-HT 1A agonists ...... 44 2.3.2 Effects of 5-HT antagonists ...... 47 2.3.3 Response of aryltryptamines ...... 52 2.3.4 Response of 2-AMBD ...... 61 2.3.5 Response of 5MEO-dlmer ...... 61 2.3.6 Response of spiroxatrine ...... 64 2.3.7 Species differences ...... 68 2.4 Discussion ...... 76

3 MDL 73005EF: ANALYSIS OF A PURPORTED 5-HT 1A ANTAGONIST USING THE PSC ASSAY...... 85 3.1 Introduction ...... 85 3.2 Methods...... 88 3.2.1 Chemicals ...... 88 3.2.2 FSC assay...... 88 3.3 Results ...... 88 3.4 Discussion ...... 90 6

PAGE

4 STRUCTURE-ACTIVITY RELATIONSHIPS OF ENANTIOMERS OF 8-0H-OPAT ANALOGS ...... 93 4.1 Introduction ...... 93 4.2 Methods...... 94 11.2.1 Chemicals ...... 94 4.2.2 FSC assay...... 96 4.3 Results ...... 97 4.3.1 8-0H-DPAT...... 99 4.3.2 ALK-3 ...... 99 4.3.3 CM-12 ...... 102 4.3.4 SE-61 ...... 102 4.3.5 LEA-146 ...... 105 4.3.6 JV-26 ...... 105 4.4 Discussion ...... 108

5 DETERMINATION OF STRUCTURE-ACTIVITY RELATIONSHIPS FOR A SERIES OF 3- AND 4-TETRAHYDROPYRIDYL INDOLES ...... 119 5.1 Introduction ...... 119 5.2 Methods ...... 121 5.2.1 Chemicals ...... 121 5.2.2 FSC assay ...... 121 5.3 Results ...... 122 5.3.1 4-methyl-THPI analogs ...... 122 5.3.2 3-methyl-THPI analogs ...... 125 5.3.3 3-propyl-THPI analogs ...... 127 5.3.4 4-propyl-THPI analogs ...... 130 5.4 Discussion ...... 130

6 SUMMARy...... 134 6.1 The use of the FSC assay as a functional screen for 5-HTlA activity ...... 134 6.2 Future directions ...... 141

APPENDIX A - COLUMN PREPARATION, MAINTENANCE AND ELUTION PROFILE DETERMINATION ...... 143

APPENDIX B - PROTOCOL FOR THE 5-HTlA BINDING ASSAy ...... 149

APPENDIX C - EFFECT OF ASSAY CONDITIONS ON 5-HT POTENCY IN THE FSC ASSAY ...... 150

REFERENCES...... 152 7

LIST OF ILLUSTRATIONS

F IGORE PAGE

1 Structure of serotonin ...... 1.1

2 A. Structures of reference 5-HT1A agonists ...... 45 8. Inhibition of PSC by the reference 5-HT1A agonists .. 45

3 A. Structure of spiperone ...... 48 8. Inhibition of PSC by spiperone ...... 48

4 A. Structure of pindolol ...... 50 B. Inhibition of PSC by pindolol ...... 50

5 A. Structure of BMY 7378 ...... 51 B. Inhibition of PSC by BMY 7378 ...... 51

6 Ability of pindolol to shift the 5-HT inhibition curve. 54

7 Structures of aryltryptamines ...... 55

8 Ability of AHR 1709 to shift the 5-HT inhibition curve. 57

9 A. Inhibition of PSC by 5CAHR ...... 59 B. Inhibition of PSC by 5MEOAHR ...... 59 C. Inhibition of PSC by BOT ...... 59

10 A. Structure of 2-AMBD ...... 62 B. Inhibition of PSC by 2-AMBD ...... 62

11 A. Structure of 5-MEO dimer ...... 63 B. Inhibition of PSC by 5-MEO dimer ...... 63

12 Structure of spiroxatrine enantiomers ...... 65

13 A. Inhibition of PSC by (±)spiroxatrine ...... 67 8. Inhibition of PSC by (-) and (+) spiroxatrine ...... 67

14 PSC in rabbit hippocampus: A. Inhibition of PSC by 5-HT ...... 72 8. Inhibition of PSC by (±)spiroxatrine ...... 72

15 PSC in bovine hippocampus: A. Inhibition of PSC by 5-HT ...... 73 B. Inhibition of PSC by (±)spiroxatrine ...... 73 8

FIGURE PAGE

16 FSC in bovine hippocampus: A. Inhibition of PSC by 5MEOAHR ...... 75 B. Inhibition of FSC by BOT ...... 75

17 Example of variation of response to (t)spiroxatrine observed In the FSC assay among batches of bovine hippocampus...... 78

18 Structures of buspirone, MOL 73005EF and BOT ...... 87

19 A. Inhibition of FSC by MOL 73005EF ...... 89 B. Comparison of the activities of MOL 73005EF, buspirone and BOT in the FSC assay ...... 89

20 Structures of the enantiomers of 8-0H-OPAT and its analogs ...... 95

21 A. Inhibition of FSC by (-)8-0H-OPAT ...... 100 B. Inhibition of FSC by (+)8-0H-OPAT ...... 100

22 A. Inhibition of FSC by (-)ALK-3 ...... 101 B. Inhibition of FSC by (+)ALK-3 ...... 101

23 A. Inhibition of FSC by (-)CM-12 ...... 103 B. Inhibition of FSC by (+)CM-12 ...... 103

24 A. Inhibition of FSC by (-)SE-61 ...... 104 B. Inhibition of FSC by (+)SE-61 ...... 104

25 A. Inhibition of FSC by (-)LEA-146 ...... 106 B. Inhibition of FSC by (+)LEA-146 ...... lOB

26 A. Inhibition of FSC by (-)JV-26 ...... 107 B. Inhibition of FSC by (+)JV-2B ...... 107

27 Figures adapted from Mellin (1990): A. A recently proposed 5-HT 1A agonist pharmacophore ... 114 B. Superposition of some enantiomers of 8-0H-OPAT analogs to illustrate the proposed alignment of these structures on the 5-HT1A receptor ...... 115

28 Generic structures of: A. 3-tetrahydropyridyl indoles (3-THPI) ...... 120 B. 4-tetrahydropyridyl indoles (4-THPI) ...... 120 9

FIGURE PAGE

29 Inhibition of FSC by 4-methyl-THPI analogs with varying 5-indole substituellts: A. -Br ...... 124 B. -OCH3...... 124 C. -CH3 ...... 124

30 Inhibition of FSC by 3-methy-THPI analogs with varying 5-indole substituents: A. -Br ...... 126 B. -OCH3 ...... 12() C. -CH3 ...... 126

31 Inhibition of FSC by 3-propyl-THPI analogs with varying 5-indole substituents: A. -Br ...... 128 B. -OCH3 ...... 128 C. -NHCOCH 3 ...... '.' 129 D. -CONH 2 ...... 129

32 Inhibition of FSC by 4-propyl-THPI analogs with varying 5-indole substituents: A. -NHCOCH 3 ...... 131 B. -CONH2 ...... 131

33 Example of typical elution profile for Dowex columns .. 146

34 Example of typical elution profile for alumina columns. 148 10

LIST OF TABLES

TABLE PAGE

1 Current 5-HT receptor classification and some proposed functional characteristics ...... 17

2 Comparison of 5-HTIA binding affinities and FSC assay potencies for the reference 5-HTIA agonists ...... 46

3 Comparison of pKi and pA2 values for the 5-HT 1A antagonists spiperone and pindolol ...... 53

4 Comparison of 5-HTIA binding affinities and FSC assay potencies for the aryltryptamines 5CAHR, 5MEOAHR and BOT ...... 58

5 Comparison of 5-HTIA binding affinities and FSC assay potencies for spiroxatrine and its enantiomers ...... 69

6 Species comparison of prop08ed 5-HT 1A intrinsic activities in the FSC assay...... 77

7 Comparison of 5-HTIA binding affinities and FSC assay potencies for enantiomers of 8-0H-DPAT and its analogs .. 98

8 Comparison of 5-HTIA binding affinities and PSC assay potencies for 3- and 4-THPI analogs ...... 123

9 Comparison of EC 50 values for 5-HT under differing assay condi tions in the FSC assay ...... 151

10 Effect of tissue concentration on EC 50 values for 5-HT in the FSC assay ...... lSI 1 1

ABSTRACT

The. need to develop selective compounds for 5·-IIT receptors rt~quires

further elucidation of structure-activity relationships involved with

recepto' recognition and activation. In particular, there is great

interest in the involvement of the 5-HTIA receptor in therapeutic

treatment of conditions such as anxiety, depression and hypertension.

Despite previous attempts to define the pharmacophore for 5-HT 1A receptors, the structural requirements for 5-HT 1A receptor activity remain largely unexplored. In fact, no selective 5-HT 1A antagonists currently exist. Binding assays provide important information concerning receptor affinity, but do not reflect functional consequences of receptor interaction. One of the functional correlates of the 5-HI'lA receptor is the inhibition of forskolin-stimulated adenyl ate cyclase

(PSe). The objective of this dissertation was to assess the merits of using the FSC assay as a functional screen to determine 5-HT 1A intrInsic activity of novel compounds. From the variety of compounds tested, it was apparent that the FSC assay can be used as an in vitro functional screen for assessing 5-HTIA agonistic and antagonistic, as well as partial agonistic, activity. All putative 5-HTIA agonists tested in the

FSC assay had potencies (EC50) that were less than their respective

5-HTIA binding affinities (Ki)' The use of pindolol to block 5-HT 1A receptors, and therefore cause a rightward shift in any observed inhibition curve, provided qualitative evidence of 5-HTIA interaction.

Quantitative evaluation of 5-HT 1A activity had to be approached 12

cautiously, due to possible complications arising from non-5-HT1A - mediated inhibJtioll. The data for a series of enanatiomers of 8-0H-OPAT

and its analogs supported a recently proposed model for the 5-HTIA agonist phnrmacophore. Also, within the enantiomeric pairs exhibiting stereoselectivity, the compound with the higher 5-HTIA affinity also possessed greater 5-HTIA agonistic activity. Within a series of tetru­ hydropyridylindoles (THPI), the 4-THPI analogs generally had both greatest 5-HTIA affinity and agonistic activity, although there were exceptions to this trend. Several compounds, whose structures were based on tile small molecules of the 5-HTIA agonist 8-hydroxy-2-(di-n­ propylaminotetralin) (O-OH-DPAT) and 5-HT itself, demonstrated potential antagonistic activity in the PSC assay. Further analysis of an extended series of these compounds in the FSC assay is required to establish specific conclusions concerning the 5-HTIA pharmacophore. 13

CHAPTER 1

INTRODIJCTION

1.1 The Discovery of Serotonin

Over 100 years ago, an endogenous vasoconstrictor substance was detected in blood. This substance was later named "serotonin" due to its "serum tonic" abilities (Rapport et al., 1947), and its structure was first identified by Rapport in 1949 (Figure 1) as 5-hydroxy­ tryptamine (5-HT). Concomitantly, researchers in Italy identified a substance "enteramine" in enterochromaffin cells of gastrointestinal mucosa, which also turned out to be serotonin (Erpsamer and Asero,

1952). Because 5-HT was initially found in the intestine and blood serum, it was originally thought to function mainly as a peripheral hormone. Subsequent studies of the localization of 5-HT in the body demonstrated that it was also present in the brain (Twarog and Page,

1953). The observation that regional variations existed in 5-HT concentrations in the brain (Amin et al., 1954) was the first suggestion that 5-HT may have a physiological role in the central nervous system

(eNS). Use of histochemical, autoradiographic and immunohistochemical techniques provided further information concerning the 5-HT neuronal system in the eNS. 14

HO

Figure 1: Structure of serotonin (5-hydroxytryptamine. 5-HT). 15

1.2 5-HT Receptor Heterogeneity

Over the years, serotonin has been proposed to be involved in many physiological actions, such as temperature regulation. nociception. feeding. sleeping. sexual activity. blood pressure, anxiety and depression (Pascual. 1990; Arvidsson et al .. 1986; Frazer et al .. 1990).

Consequently, there has been growing interest in the quest to better define the actions of 5-HT, both peripherally and in the central nervous system, such that therapeutic benefit can be ultimately achieved. A major outgrowth of this increased interest in 5-HT has been the discovery of multiple receptor subtypes. Gaddum and Picarelli (1957) were the fIrst to suggest the possibility of the existence of more than one 5-HT receptor in isolated guinea pig ileum. They termed the two putative receptors "0" and "M", based on the pl'efel'ential blockade by dibenzyline (phenoxybenzamine) and morphine, respectIvely. Although these drugs are of mInimal use in the CUl'rent classification of 5-HT receptors (Humphrey, 1984). this was the first demonstration of at least two distinct 5-HT receptors.

The introduction of the technique of radioligand receptor binding contributed greatly to the CUl'l'ent knowledge of 5-HT receptor classifi­ cations. Early receptor binding assays were carried out using

[3 H]lysergic acid diethylamide (LSD) and [3H]5-HT (Bennett and Snyder.

1976). Shortly thel'eaftel'. the existence of at least two classes of

5-HT receptors in the mammalian central nervous system was identified

(Peroutka and Snydel'. 1979). one with high affinity fol' 5-HT (5-HTl) and 16

the other with low affinity (5-HT2)' By 1981, the 5-HT1 class had been further divided into 5-HT1A and 5-HTIB subtypes, usIng spiperone to

discriminate r3 H]5-HT binding sites (Pedigo et al., 1981; Nelson et

al .. 1981). This marked the beginning of an explosion of repol'ted

information concerning 5-HT receptors and receptor subtypes. The

current classification of putative 5-HT receptors and related binding

sites includes the 5-HT1 (with at least seven subtypes, 5-HTIA to

5-HT 1E , 5-HT1P and 5-HT 1R ), 5-HT2' 5-HT3 and 5-HT4 classes. Table 1 briefly summarizes some characteristics, including proposed functions.

associated with each of these receptor types and subtypes (For a current

review of 5-HT receptor classification, refer to Frazer et al .. 1990).

Many of the proposed actions associated with 5-HT receptor activation

involve interactions with signal transduction mechanisms. The focus of

this dissertation will be the interaction of 5-HT receptors with the adenyl ate cyclase second messenger system.

1.3 5-HT and Adenylate Cyclase - The Early Years

This section will discuss the relationship betw~en 5-HT and adenylate cyclase, prior to the revelation of 5-HT receptor heterogeneity. Adenylate cyclase is a ubiquitous enzyme, and its highest activity among mammalian organs investigated was reported to be in the brain (Sutherland et al., 1962). Although little was known about the biochemical mechanism of action of endogenous amines in animal and human brain at that point in time, 5-HT was initially reported to cause a small increase in cAMP production in slices of rabbit cerebellum 17

Table 1: CURRENT 5-HT RECEPTOR CLASSIFICATION AND SOME PROPOSED FUNCTIONAL CHARACTERISTICS.

RECEPTOR PROPOSED FUNCTION(S)

5- HT IA -AC 1 modulation -PI2 turnover -K+ channel activation -behavior ("5-HT syndrome": tremor, forepaw treading, head weaving) -hypotension -hypothermia -auto receptor

5- HT IB -AC l inhibi tion -autoreceptor

5- HT IC -PI 2 turnover

5- HT ID -AC 1 inhibition

5- HT IP -slow EPSP's3 (in the gut)

-vascular smooth muscle contraction -platelet aggregation -behavior (head twitches)

-ligand-gated ion channel -anti-emetic (associated with cancer therapy)

5- HT 4 -AC1 stimulation

1 adenyl ate cyclase 2 phosphatidylinositol 3 excitatory postsynaptic potential Adapted from Mawe et al. (1986), Peroutka (1988a), Hoyer (1988b), Saxena and Ferrari (1989), Kilpatrick et al. (1987) and Frazer et al. (1990) . Other proposed binding sites, as yet not associated with any function: 5-HTIE (Leonhardt et al., 1989) and 5-HT 1R (Xiong and Nelson, 1989). 18

and cerebral cortex (Kakiuchi and RaIl, 1968a. 1968b). Evidence began to mount in support of the speculation that the central effect of some biogenic amines could be. at least in part. mediated by cyclic adenosine monophosphate (cAMP). By 1971, the in vitro stimulation of adenylate cyclase by 5-HT in brain homogenates and slices was well established.

Chou et a!. (1971) then demonstrated that 5-HT significantly stimulated cAMP production in rat cerebral cortex, in vivo. A 1.5-2 fold increase in adenylate cyclase activity in response to 5-HT was later observed in slices of human frontal lobe (Kodama et al., 1973), and in squirrel monkey cortex (Skolnick et al .. 1973). 5-HT was shown to stimulate adenyl ate cyclase in horse brain striatum synaptosomal membranes in a

GTP-dependent manner, with a EC50 of 20-60 nM (Fillion et al., 1979).

In addition. glial cell membranes from horse brain striatum were reported to contain an adenylate cyclase positively coupled to a low affinity 5-HT receptor (EC50 = 1 pM; Fillion et al., 1980).

Von IIungen et a!. (1974, 1975) reported that cell-free preparations of new-born rat colliculi contained 5-HT sensitive adenylate cyclase systems, and that the stimulation of cAMP production in response to 5-HT diminished significantly with age. A similar age-dependent response to

5-HT was also observed with cAMP levels in rat hypothalamus slices, and was antagonized by methiothepin and metergoline (Daszuta and Cadilhac,

1979). However, there was no difference between newborn and adult guinea pigs in the percentage of 5-HT-stimulated adenylate cyclase activity, thus suggesting species differences in the development of 19

serotonergic responses (Enjalbert et al., 1978a). Metergoline competitively inhibited the 5-HT sensitive adenylate cyclase in various regions of newborn rat brains (Enjalbert et al., 1978b).

In an early attempt to compare the 5-HT receptor linked to adenylate cyclase and the putative receptor measured by [3H]5-HT binding, Bockaert and colleagues (1981) found a 500-1000 fold difference between the respective affinities of 5-HT for the receptor coupled with adenylate cyclase, and for that labelled with [3H]5-HT, using brain areas from newborn rats. Although this observation could have been attributed to different conditions in the cyclase and binding assays, a more rigorous characterization of the 5-HT sensitive adenyl ate cyclase in collicular homogenates from newborn rats suggested that the receptor linked to adenylate cyclase was distinct from the putative receptor measured by [3H]5-HT binding. Of the 5-HT antagonists that blocked the

5-HT-stimulation of adenylate cyclase, only metergoline acted as a competitive inhibitor. Even under comparable assay conditions, there was poor correlation of the affinities of 5-HT and associated compounds at binding sites, with corresponding potencies of the compounds in the cyclase assay.

In 1980, Nelson et al. (1980a, 1980b) characterized a 5-HT receptor positively linked to adenyl ate cyclase that was not pharmacologically identical to the site labeled by the binding of [3H]5-HT. In so doing, they provided evidence for the possible heterogeneity of 5-HT receptors in newborn rat brain: one serotonin receptor positively linked to 20

adenylate cyclase in newborn rat colliculi that was sensitive to ~M concentrations of 5-HT. and another that had high affinity for 5-HT. which was apparently not linked to adenylate cyclase.

In adult rat hippocampal membranes. slight (yet statistically significant) 5-HT-stimulation of adenyl ate cyclase was detected.

However. no significant 5-HT sensitive cyclase activity was observed in membranes from adult rat cortex. hypothalamus and coiliculi (Barbaccia et al. 1983j Hotta and Yamawaki. 1986). Adenylate cyclase activity in rat hippocampal membranes was also stimUlated by tryptamine. 5-methoxy­ tryptamine. bufotenine. and LSD. which acted as a partial agonist.

Ketanserin and mianserin had no effect. The rank order of potencies of antagonists in this system correlated well with the order of potencies for the inhibition of [3H]5-HT binding (Barbaccia et al .. 1983).

MacDermot (1979) described a 5-HT sensitive adenylate cyclase that was GTP-sensitive in NCB-20 cells (hybrid of mouse neuroblastoma x 18- day-fetal Chinese hamster brain explant). but not in NG108-I5 cells

(hybrid of mouse neuroblastoma x rat glioma cells). In homogenate preparations of NCB-20 cells. the 5-HT stimulation of cA}IP production was antagonized by mianserine and cyproheptadine. whereas LSD acted as a mixed agonist/antagonist (MacDermot et al .• 1979). The EC so for 5-HT activation of adenyl ate cyclase was reported to be 400-800 nM (MacDermot et al .• 1979; Berry-Kravis and Dawson. 1983). Although the rank order of potencies was comparable to [3H]5-HT binding data, the actual affinity for 5-HT was much reduced in the cyclase linked receptors. 21

This discrepancy may reflect differences due to coupling to guanine

nucleotide binding proteins (G proteins), and will be discussed further

in Chapter 2 (Section 2.4). Cultured calf aorta smooth muscle cells

were also shown to contain 5-HT receptors positively linked to adenylate

cyclase, which were antagonized by methysergide but not by cianserin,

molindone or LSD (Luchins and Makman, 1980). In addition, an

indoleamine-stimulated adenyl ate cyclase activity detected in rabbit

retina was proposed to possibly be due to a 5-HT receptor (Blazynski et

al., 1985).

5-HT receptors coupled to adenyl ate cyclase have also been studied

in invertebrate systems, although not as extensively as with mammalIan

systems. The liver fluke, Fasciola hepatica, was the initial

invertebrate system used to demonstrate an interaction between 5-HT and adenylate cyclase (Mansour et al., 1960). Since then, there have been further reports describing the ability of 5-HT to stimulate adenylate cyclase actIvIty in the liver fluke (Northup and Mansour, 1978a, 1978b;

McNall and Mansour, 1984, 1985; Mansour et al., 1983), as well as

tissues from other invertebrates such as the annelid (Robertson and

Osborne, 1979), aplysia (Cedar and Schwartz, 1972; Levitan and Benson,

1981; Ocorr and Byrne, 1986; Ram et al., 1983), blowfly (Heslop and

Berridge, 1980; Berridge and Heslop, 1981; Litosch et al., 1982), platyhelminth (Ribeiro and Webb, 1987) and cockroach (Nathanson and

Greengard, 1973; Downer et al., 1985). For a more comprehensive discussion of invertebrate 5-HT sensitive adenylate cyclase, refer to the book chapter by Cornfield and Nelson (in press). 22

1.4 Current Status of 5-HT Receptor's Linked to Adenylate Cyclase

The early attempts at characterizing 5-HT receptors linked to

adenylate cyclase had often tried to compare their pharmacology to that

of [3H]5-HT-Iabelled binding sites. However. with the revelation that

[3H]5-HT bound to a heterogeneous population of 5-HTI receptors with

high affinity. it became apparent that the affinity values of compounds

in r3H]5-HT binding studies could not necessarily be expected to correlate with the corresponding potencies in the cyclase assay. Now.

of course. it is known that r3H]5-HT labels a whole family of receptor subtypes. which further illustrates the difficulty of trying to compare

5-HT receptors linked to adenyl ate cyclase with those defined by ligand

binding.

Following the revelation of 5-HT receptor heterogeneity. a dramatic amount of progress has been accomplished in the study of 5-HT receptors coupled to adenylate cyclase in mammalian systems. To date. the 5-HTIA'

5-HT18. 5-HTID and 5-HT4 receptors have been reported to be coupled to adenyl ate cyclase. while no reports have provided convincing evidence

that the 5-HTIC' 5-HT2 or 5-HT3 receptors are linked to this system.

The following section will discuss the progression of knowledge regarding the individual 5-HT receptors that are linked to adenylate cyclase systems. 23

1.4.1 5-HT1A receptors

The discovery that r3H]8-0H-DPAT provided a selective method to pharmacologically define 5-HT1A binding sites has certainly aided the study of 5-IIT receptors linked to adenylate cyclase (Arvidsson et al.,

1981; Gozlan et al., 1983; Schlegel and Peroutka, 1986). To date, the

5-HT1A receptor subtype has been proposed to be linked to the stimulation of adenylate cyclase (Shenker et al., 1985; Markstein et al., 1986), the inhibition of adenylate cyclase (DeVivo and Maayani,

1986), the inhibition of carbachol-induced stimulation of phosphoino­ sitide (PI) turnover (Claustre et al., 1988), as well as the activation of potassium channels (Andrade and Nicoll. 1987a, 1987b; Colino and

Halliwell, 1987). In addItion, the somatodendritic autoreceptor for

5-HT in cells of the dorsal raphe nucleus is likely to be the 5-HTIA

SUbtype, since these cells are inhibited potently and selectively by

5-HT 1A selective agonists (Weissmann-Nanopoulos et al., 1985).

The ability of 5-HT to inhibit adenylate cyclase is currently one of the more widely accepted functional correlates of the 5-I1T 1A receptor. Over the years, technical difficulties have been associated with measuring the inhibition of basal cAMP levels. The use of forskolin, a diterpene that directly stimulates the catalytic subunit of adenylate cyclase (Laurenza et al., 1989), has made it possible to detect inhibition of cAMP production ordinarily too low to detect under basal conditions. The premise of the forskolin-stimulated adenylate cyclase (FSC) assay involves the direct stimulation of adenyl ate cyclase 24

with forskolin, and then quantitation of adenylate cyclase activity following exposure to the compound under investigation.

DeVivo and Maayani (1985) reported that 5-HT, 5-carboxamidotrypt­ amine (5-CT) and 8-0H-DPAT inhibited FSC activity in guinea pig hippocampal membranes. Spiperone competitively antagonized this response. Based on the comparison of the EC50 values of the three agonisls and the KB for spiperone to their affinities in [3 H]8-0H-DPAT binding studies, the inhibition of FSC was proposed to be mediated via the 5-HTIA receptor. DeVivo and Maayani (1986) then reported further pharmacological characterization that confirmed that the inhibition of adenylate cyclase is a functional correlate of the 5-HTIA binding site in guinea pig and rat hippocampal membranes.

Additional evidence to confirm that 5-HTIA receptors are negatively coupled with adenylate cyclase was the demonstration that 5-HT IA receptors in rat hippocampus were linked to a pertussis-sensitive G protein, likely Gi (Harrington et al., 1988). There is also recent evidence that common pertussis toxin-sensitive G protein(s) couple adenosine Al and 5-HTIA receptors to the inhibition of adenyl ate cyclase activity in rat hippocampus (Zgombick et al., 1989).

5-HT I receptors that inhibit vasoactive intestinal polypeptide (VIP)- and forskolin-stimulated cAMP production have also been subjected to extensive pharmacological characterization in primary cultures of mouse hippocampal, cortical and striatal neurons (Dumuis et al., 1988a;

Weiss et al., 1986). Based on these pharmacological characterizations, 25

the 5-HT receptor negatively coupled to adenylate cyclase in cortical

neurons appeared to be different from the 5-HTIA receptor found in

hippocampal neurons. Thus, although the 5-HT receptors inhIbiting cAMP

formation in mouse embryonic cortical neurons shared many

pharmacological characteristics with the 5-HT1A receptor (such as

inhibition of VIP-stimulated adenylate cyclase activity by 5-HT.

RU 24969 and 8-0H-D~AT), the overall pharmacology was sufficiently

different to conclude that the response was mediated by a non-typical

5-HT 1A receptor (Weiss et al., 1986).

There have been reports that the stimulation of adenylate cyclase

activity might also be linked to the 5-HTIA receptor, despite the fact

that previous work has also linked 5-HT receptors to the inhibition of

adenylate cyclase in a variety of species and tissues. In 1985, Shenker

et al. proposed that the stimulation of adenylate cyclase in adult guinea pig hippocampal membranes was likely mediated by a heterogeneous

population of 5-HT receptors, based on the shallow and non-parallel dose

response curves for 5-HT, tryptamine, 5-methoxytryptamine and

bufotenine. In addition, 5-carboxamidotryptamine (5-CT) produced a distinctly biphasic dose response curve. Spiperone was a potent

competitive antagonist at the high affinity component, but not the low affinity component, of the response to 5-CT. In this system, the EC50

for 5-HT was 0.1 ~M (Shenker et al., 1985). Although Markstein et al.

(1986) reported further evidence supporting the concept that the 5-HTIA

receptor was positively coupled to adenyl ate cyclase in adult rats,

there was no evidence of a biphasic response to 5-CT. In contrast to 26

the observations of Shenker et al. (l985), 5-HT sensitive adenyl ate cyclase activity (EC50 = 0.5 pM) in primary cultures of mouse purified strjatal and cortical neurons was not stimulated by the 5-HT1 selective agonists RU 24969 and 8-0H-DPAT (Weiss et al., 1986). In the absence of

VIP, cAMP production was stimulated through a 5-HT receptor which did not share the same pharmacological profile as the 5-HT 1A receptor. In 1987, Shenker and assocIates reported that two 5-HT receptors, a high affjnity 5-HT 1A receptor and a low affinity novel 5-HT receptor (recently classified as 5-HT4' Dumuis et al., 1988b, 1988c, 1989), mediated stimulation of adenylate cyclase in guinea pig hippocampal membranes.

Most of the current knowledge of 5-HTIA receptors, including interactions with adenylate cyclase, is based on classical biochemical, pharmacological and autoradiographic analyses (Maayalli and Sherman,

1990). The recent advent of methods capable of isolating. cloning and expressing specific receptor genes has added new dimensions to the elucidation of the mechanisms involved jn the coupling of the 5-HT1A receptor to adenylate cyclase. For example, a genomic clone for the

5-HT1A receptor, G-21, was recently obtained from a human genomic library by cross-hybridization with a full-length ~2-adrenergic probe

(Kobilka et al., 1987; Fargin et al., 1988). G-21 was very recently transfected and expressed in two mammalian cell lines which do not normally express the 5-HT1A receptor: COS-7 (transient expression) and

He La (permanent expression), such that the second messenger systems linked to the cloned 5-HT1A receptor could be examined (Fargin et al .. 27

1989). In both cell lines, 5-HT inhibited FSC (up to 90% in HeLa

cells). Metitepine and spiperone competitively antagonized this effect

(Ki = 81 and 31 nM, respectively), which was also blocked via

pretreatment with pertussis toxin.

[t is interesting to note that 5-HTIA-stimulation of adenylate

cyclase was not detected in either cell line, although p-adrenergic­ stimulated adenyl ate cyclase was observed in the COS-7 cells (Fargin et al., 1989). This suggests the possibility that the cell lines could be deficient in appropriate stimulatory G protein(s) for 5-HT-stimulated adenylate cyclase, or more likely, that another receptor, which is pharmacologically similar to the 5-HTIA/G-21, is positively coupled to adenylate cyclase.

Although this human 5-HT 1A receptor, as expressed in HeLa cells, was shown to be primarily negatively linked to adenyl ate cyclase, it hAS also recently been shown to stimulate sodium-dependent phosphate uptake via the activation of protein kinase C (Raymond et al., 1989). This

function was also ~hown to be sensitive to pertussis toxin. This observation suggests that a single receptor subtype may mediate several second messenger systems as well as ion channels, thus providing many potential points of feedback for amplification or attenuation of signal

transduction. Such proposed "crosstalk" among second messenger systems further complicates the task of studying signal transduction mechanisms.

Currently, it is not definitively known whether one 5-HT receptor subtype is linked exclusively to one second messenger system, or is able 28

to link to multiple transduction mechanisms, perhaps via different G proteins.

Hopefully, molecular techniques will help to resolve much of the controversy associated with the signal transduction mechanisms linked to the 5-HT 1A receptor. The 5-HTIA receptor may be coupled to several effector systems, through different G proteins, or there may be other

5-HTI-like receptors positively linked to adenylate cyclase which are pharmacologically similar to 5-HTIA' The question of whether the 5-HTIA receptor that mediates inhibition of FSC in hippocampus is pharmacologically identical to the 5-HT receptor that stimulates adenylate cyclase in that same tissue remains unanswered. The exact nature of the coupling of adenylate cyclase to the 5-HT1A receptor currently continues to be a controversial issue.

Although much information has been gained concerning the identification and classification of 5-HT receptors linked to adenylate cyclase in mammalian systems, comparable research progress has yet to be achieved with invertebrates. With respect to invertebrate systems, the pharmacological profile for the 5-HT receptor linked to adenylate cyclase in Schistosoma mansoni did not match that for any of the reported mammalian 5-HT receptors, based on data from a limited number of compounds (Mansour and Mansour, 1986). The ligand PAPP, p-amino-­ phenylethyl-m-trifluoromethylphenyl piperazine, which has previously been shown to selectively bind to the 5-HTIA receptor in rat brain, was reported to activate adenyl ate cyclase in Aplysia neuronal membranes, 29

suggesting that the 5-HTIA receptor may be responsible for stimulatIon

of adenylate cyclase in the Aplvsia (Saitoh and Shih, 1987). Further

work is necessary to determine whether or not the 5-HTIA receptor (as

defIned In mammalian systems) is actually responsible for the

stimulation of cAMP production observed in Aplysia neuronal membranes.

1.4.2 5-HT1B receptors

It has long been known that the 5-HTIB binding site was GTP-

sensitive (Sills et al., 1984a), which gave rise to the idea that it . must be a G protein coupled receptor, likely linked to some second

messenger system. Bouhelal et al. (1988) reported that 5-HT1B receptors

mediated the inhibition of FSC in homogenates of rat substantia nigra

homogenates, an area containing a high density of 5-HTIB receptors. The

5-HT1A agonist, 8-0H-DPAT, which is very potent in inhibiting FSC in the

hippocampus, was very weak and ineffective in rat substantia nigra. The

other agonists tested, RU 24969, tryptamine and 5-CT, did not

discriminate between 5-HTIA and 5-HT1B receptor subtypes and gave nanomolar potencies in their ability to inhibit FSC in the substantia

nigra. To date, no stimulation of adenylate cyclase by 5-HT agonists

has been reported in rat substantia nigra.

The 5-HT1B binding site is located both postsynaptically and presynaptically as an autoreceptor (Engel et al., 1986). However, the

location of the 5-HT1B-mediated inhibition of FSC activity is not known at present. 5-HT1B receptors seem to be limited to rodent species, such 30

as the rat and mouse. and may correspond to another 5-HT receptor subtype (i.e .• 5-HTIO) in other species. There exists close parallel between the 5-HT 1B sites of rat substantia nigra and 5-HT1D sites of calf substantia nigra. since each is the predominant site in the respective species (Hoyer and Schoeffter. 1988). Hoyer and Middlemiss

(1989) have proposed that the 5-HT1B and 5-HT1D receptor subtypes that are negatively coupled to adenylate cyclase activity share similar functions in different species.

1.4.3 5-HT 1D receptors

5-HT 1D receptor-mediated inhibition of FSC activity has been reported in calf substantia nigra by Hoyer and Schoeffter (1988). Th.is inhibition of adenylate cyclase was GTP-dependent. and is likely mediated by Gi' although this has not yet been definitively demonstrated.

The novel compound sumatriptan (GR 43175) has been shown to have high affinity for 5-HT1B and 5-HT1D binding sites (Ki = 27 and 17 nM. respectively; Peroutka and McCarthy, 1989). Further studies are required to determine whether sumatriptan evokes similar functional responses with the adenylate cyclases linked to these two receptor subtypes. 31

1.4.4 5-HT4 receptors

Very recently, 5-HT receptors positively coupled to adenyl ate cyclase in primary cultures of mouse embryo colliculi neurons were identified and characterized. These 5-HT receptors possessed properties that did not correspond with either the 5-HT1 or 5-HT2 receptor classifications. In addition, although this novel receptor was antagonized by the selective 5-HT3 antagonist, ICS 205 930, it was not related to 5-HT3 receptors (Dumuis et al .. 1988c). ICS 205 930 had low affinity for this non-classical receptor (Ki = 500-1000 nM - Dumuis et aI, 1989). The novel pharmacologic properties of this receptor has led to its assignment to a new class of 5-HT receptors, the 5-HT4 . Another pharmacologically unique feature of the 5-HT4 receptor is that substituted benzamide derivatives such as metoclopramide and BRL 24924 are agonists, and stimulate cAMP production in primary cultures of mouse embryo colliculi neurons. It is interesting to note that BRL 24924 also stimulates gut motility, most likely through an enteric 5-HT receptor

(Dumuis et al., 1989; Craig and Clarke, 1990).

In earlier reports, a biphasic response of 5-HT-stimulated adenyl ate cyclase activity was observed with the stimulation of adenylate cyclase by 5-CT in guinea pig hippocampal membranes (Shenker et aI., 1987), and it now seems likely that the low affinity component of this response was actually due to the 5-HT4 receptor. In fact, very recent work has demonstrated that the potency of benzamides tested in guinea pig hippocampal membranes was similar to their effect on 5-HT4 32

receptors positively coupled to adenyl ate cyclase in fetal mouse colliculi neurons (Bockaert et al., 1990).

1.5 Determination of Structure-Activity Relationships for 5-HT 1A receptors - Goals of This Research

Since 5-HT is proposed to be involved in such a diversity of physiological actions, the use of compounds with highly selective actions on 5-HT receptor subtypes shows clear potential for therapeutic benefit. Moreover. such compounds could potentially provide important research tools which could be used to improve overall understanding of the role played by 5-HT both with normal physiology and in disease states.

In particular, 5-HTIA receptors have been proposed to be clinically relevant in the treatment of disorders such as depression, hypertension and anxiety (Cooper and Abbott, 1988; Traber and Glaser. 1987). For example. the activity of the non-benzodiazepine anxiolytic agents buspirone and ipsapirone has been attributed to their activity as partial agonists at 5-HTIA receptors. Consequently, there is great interest in the continued development of 5-HT1A selective agonists and antagonists for potential therapeutic use, as well as for research tools. To this end, defining structure-activity relationships (SAR) is of fundamental importance to the overall understanding and ultimate development of selective agonists and antagonists for 5-HTIA receptors. 33

For a receptor with unknown 3-dimensional (3-D) nature, such as the

5-HT 1A receptor, one can only attempt to deduce an operational model that gives a consistent explanation of the known data. Ideally, such information would provide predictive value when considering new compounds for synthesis and biological testing (Marshall and Cramer,

1988). Therefore, one approach is to iook at a series of novel 3--0 chemical structures and attempt to interpret the response of the receptor in a consistent manner. Such an approach is based on the premise that structural complementarity exists between the receptor and the compounds that bind to it. Also, It is useful to assume that the receptor site is constant and that structurally different drugs nevertheless bind in conformations that present a similar steric and electronic pattern -otherwise known as the pharmacophore. This is a complicated task, since most drugs, due to inherent conformational freedom, are capable of presenting a multitude of 3-~ patterns to a receptor.

Hibert et al. (1987, 1988a, 1989) have used a graphics computer­ assisted receptor mapping technique to attempt to define the pharmacophore associated with 5-HT1A receptor interaction. They hypothesized that the two reference structural features defining the pharmacophore for the 5-HT1A antagonist recognition site were an aromatic ring and one strongly basic nitrogen atom. The resultant 3-~ map of their proposed model specifically consisted of a basic nitrogen which was 5.6 (and 5.37) angstroms from the centre of an aromatic ring and 1.6 (and 0.2) angstroms below (or above) the plane of this ring, for 34

the 5-HT1A antagonist (and 5-HT1A agonist) recognition site, respectively. Although their approach has led to the discovery of somp.

interesting compounds, often with high affinity and selectivIty for

5-HT 1A receptors. it has yet to produce any selective 5-HTIA antagonists (see Chapter 3). The potent 5-HT1A agonist, 8-0H-DPAT, has also been

the subject of previous SAR studies which attempted to determine

structural features required for 5-HT1A agonist and antagonist binding

(Glennon et al .. 1988a; Arvidsson et al., 1987. 1988; 8j6rk et al.,

1989). Very recently, Mellin (1990) has proposed another pharmacophore

for 5-HT 1A agonists (see Chapter 4 for further discussion).

Most studies investigating SAR for 5-HTIA receptors have tried to

determine trends between structure and optimal receptor affinity.

Although binding experiments provide important information concerning

the affinity of compounds for receptors, some measure of the functional

consequences associated with receptor interaction (if any) is also

essential. A reliable functional assay is required to investigate

structural features that confer agonistic, partial agonistic and

antagonistic activity. Much remains to be learned concerning SAR for

the 5-HT1A receptor site, and the elucidation of the pharmacophore for

5-HT 1A agonistic and antagonistic activity is consequently an area of ongoing research.

The determination of SAR associated with intrinsic activity of

5-HT1A receptors negatively linked to adenyl ate cyclase is the focus of

this dissertation. The underlying hypothesis of this research is that 35

the inhibition of FSC, a recognized functional correlate of 5-HT IA receptors, can be used as a screen to determine SAR for compounds possessing potential 5-HTIA agonistic, partial agonistic and antagonistic activity. Using several series of novel compounds, this is the first time that the 5-HTIA-cyclase linked receptor has beell characterized in this fashion.

Results of the dissertation are organized in the following manner:

Chapter 2 describes the initial characterization of the FSC assay as a functional screen for 5-HT1A activity, including a detailed description of the methods employed, and approaches to data interpretation. Chapter 3 specifically addresses the utility of using the FSC assay to assess the activity of MDL 73005EF, a purported 5-HT 1A antagonist (developed using the computer-modelling graphics of Hlbert et al., 1988a). Chapter 4 discusses the evaluation of SAR for a series of enantiomers of 8-0H-DPAT analogs. Chapter 5 discusses the assessment of

SAR for a novel series of 3- and 4-tetrahydropyridyl indole analogs which have affinity for the 5-HTIA receptor. Chapter 6 summarizes the conclusions made in the preceding chapters and suggests future directions. Several appendices have also been included. Appendix A discusses column preparation, maintenance and elution profile determination. Appendix B describes the protocol for 5-HTIA binding studies. Appendix C discusses a small study to investigate. the effect of assay conditions on the potency of 5-HT in the FSC assay. 36

CHAPTER 2 USE OF THE FORSKOLIN-STIMULATED ADENYLATE CYCLASE ASSAY AS A SCREEN FOR COMPOUNDS POSSESSING POTENTIAL 5-HT1A ACTIVITY

2.1 Introduction

Due to the clinical relevance of 5-HT1A receptors in treatment of disorders such as hypertension, depression and anxiety (File, 1987;

Cooper and Abbott, 1988; Frazer et al., 1990), there is continuing interest in the design of agents possessing selective activity at the

5-HT1A receptor. To date, the development of a selective 5-HT 1A antagonist has yet to be accomplished. Consequently, there exists a real need to not only assess affinity of compounds for 5-HT1A receptors

(as determined with radioligand binding), but also functional activity

(or lack thereof). Currently, (3H]8-0H-DPAT is the potent and selective

5-HT1A ligand used to define affinity at the 5-HTIA receptor (Glennon et al., 1988b; Arvidsson et al., 1984). Binding data can supply valuable information concerning receptor affinity, but do not provide insight into functional consequences of receptor interaction. A reliable functional assay is therefore essential to determine intrinsic activity of novel compounds. Currently, behavioral studies, such as the drug discrimination paradigm and 5-I1T syndrome, as well as biochemical tests such as the indirect measurement of monoamine synthesis, are commonly used to assess the activity of compounds at the 5-HT1A receptor

(Arvidsson et aI., 1986). These in vivo experiments are not without limitations, and another method to provide supporting evidence of 5-HTIA activity is needed. Although the inhibition of FSC is an accepted 37

functional correlate for the 5-HTIA receptor (DeVivo and Maayani, 1986,

1988; Frazer et al. 1990), the use of this FSC assay as an In vitro

screen for compounds possessing potential 5-HTIA intrinsic activity has

yet to be evaluated.

This chapter will discuss the characterization of the FSC assay for

use as such a scr-een. In this study, we investigated the potential use

of the FSC assay as a screen to differentiate possible 5-HT1A agonists,

par-tial agonists and antagonists. To this end, compounds with

relatively high 5-HT 1A binding affinity, such as the aryl tryptamine AHR 1709, novel AHR 1709 analogs (BDT, FPPEI, 5MEOAHR and 5CAHR),

2-AMBD, 5MEO-dimer- and spiroxatr-ine were examined in the FSC assay using

rat hippocampus. We then compared their responses to known 5-HTIA

agonists, namely buspirone, 8-0H-DPAT and 5-HT itself. The effects of

some 5-HT antagonists were also investigated. In addition. species

differences were explored, using rabbit and bovine hippocampal membranes, in order to compare with rat, and ultimately deter-mine the

species best suited for screening purposes.

2.2 Methods

2.2.1 Chemicals

8-0H-DPAT (8-hydroxy-2-(di-n-propylamino)tetralin HBr) was

purchased fr-om Resear-ch Biochemicals, Inc. (Natick, MA). [a-32 P}ATP.

[3H]CAMP and [3H]8-0H-DPAT were obtained from New England Nuclear (Boston. MA). (±)Pindolol and 5-hydroxytryptamine creatinine sulphate> 38

we~e purchased from Sigma Chemical Co. (St. Louis. MO). AHR 1709 (3-(2-

(4-phenyl-l.2.3.6-tetrahydropyrid-l-yl)ethyl)indole) was a gift from

A.H. Robbins Company (Richmond. VA). The novel aryltryptamines 5CAHR

(5-carboxamido-3-(2-(4-phenyl-l.2.3,6-tetrahydropyrid-l-yl)ethyl)­

Indole). 5MEOAHR (5-methoxy-3-(2-(4-phenyl-l.2.3.6-tetrahydropyrid-l­

yl)ethyl)-indole). BDT (N-(2-(5-methoxyindol-3-yl)ethyl)-2-aminomethyl-

1.4-benzodloxane) and FPPEI (5-methoxy-3-(2-(4-(p-fluorophenyl)­

piperazin-l-yl)ethyl)indole). as well as 2-AMBD (2-aminomethyl-l.4-

benzodloxane) and the 5MEO-dimer (N-[2-(5-methoxyindol-3-yl)ethyl]-5-

methoxytryptamine) were synthesized by Dr. E.W. Taylor. (±)Spiroxatrine

(1-oxo-4-phenyl-8-(1.4-benzodloxan-2-ylmethyl)-2.4.8-t~lazaspiro[4.5]­

decane) and its enantiomers were synthesized by Dr. S.S. Nikam.

Splperone was a gift from Janssen Pharmaceutica (Beerse. Belgium).

BMY 7378 (8-(2-[4-(2-methoxyphenyl)-1-piperazinyl]-methyl]-8-

aZBsplrol[4.5]-decane-7.9-dione dlhydrochloride) was a gift from Bristol

Meyers. Dowex (AG 50W-X4, 200-400 mesh) and neutral alumina (AG 7, 100-

200 mesh) were purchased from Bio-Rad (Richmond. CAl. All other chemicals were purchased from standard commercial sources.

2.2.2 Tissue preparation

The conditions used for the tissue preparation were adapted from

DeVivo and Maayani (1986) and are outlined according to species:

RAT: Male Sprague-Dawley rats (150-250 g) were rendered unconscious

with carbon dioxide (in a chamber containing dry ice) and then 39

decapitated. Brains were removed, chilled in saline (0.9%) and

the hippocampi were rapidly dissected out. One rat yielded

approximately 0.120 grams of hippocampal tissue. The hippocampi

were then homogenized in buffer containing 300 mM sucrose, 1 mM

EGTA, 5 mM EDTA, 20 mr.1 Tris--HCl (pH 7.4),5 mM dithiothreitol

(1:9, original wet weight:volume (ww:v)) using a glass/teflon

homogenizer (20 strokes by hand). The tissue suspension was

further diluted 8-fold and centrifuged at 500 x g for 5 minutes at

4oC. The resulting supernatant was centrifuged at 39,000 x g for

10 minutes. The pellet was stored on ice for no more than one

hour prior to resuspension in buffer (1:9, ww:v) for use in the

FSC assay.

RABBIT: Male New Zealand White rabbits (2 to 4 pounds) were killed with

an overdose of pentobarbital. One rabbit brain yielded

hippocampal tissue comparable to five rats. The tissue

preparation was the same as described above for the rat.

BOVINE: Bovine brains were obtained from local slaughterhouses. One

bovine brain yielded about 75-times the hippocampal weight of one

rat. The dissection and tissue preparation were carried out as

described above for the rat, with the following differences. The

hippocampal membranes were mechanically homogenized using a

glass/teflon homogenizer. After the final centrifugation step,

the pellet was resuspended (1:9, ww:v), divided into 1 ml

aliquots. quickly frozen in an ethanol/dry ice bath, and stored at 40

-70°C. For individual experiments, a sufficient volume of

aliquots was thawed just prior to starting the incubation.

2.2.3 Forskolin-stimulated adenylate cyclase (FSC) assay

The conditions used for the FSC assay components were adapted from

DeVivo and Maayani (1986). The conversion of [a-32 P]ATP to [a-32P]cAMP was quantitated by the method of Salomon (1979). The procedures outlining the preparation and maintenance of the dowex and alumina columns, as well as the determination of column elution volumes, are described in Appendix A.

Some drugs were dissolved in water with one of the following pretreatments: 0.5% glacial acetic acid (for pindolol, spiroxatrine, spiperone, 2-AMBD, 5MEO-dimer), 1% glacial acetic acid with 20% DMSO

(for AHR 1709, FPPEI, 5MEOAHR), 20% methanol (for 5CAHR) or 20% DMSO

(for BDT), with subsequent dilutions in water. All other drugs were dissolved directly in water. The final composition of the assay medium was 100 mM NaCI, 2 mM magnesium acetate, 10 ~M forskolin, 80 mM Tris-Hel

(pH 7.4), 2 mM magnesium acetate, 100 mM sodium chloride, 0.2 mM ATP,

1 mM cAMP, 10 ~M GTP, 60 mM sucrose, 4 mM theophylline, 0.2 mM EGTA,

1 mM EDTA, 1 mM dithiothreitol, 0.1 mg/ml creatinine phosphokinase, 5 mM creatine phosphate and 1-2 ~Ci [a-32P]ATP (specific activity 20-40

Ci/mmole), per sample. Each drug concentration was tested in triplicate. 41

Tubes containing 150 ~l of the cyclase assay medium and 50 ~l of

drug dilution (or water) were pre incubated at 30 0 e for five minutes.

staggering samples by 15 second intervals. The assay was then started

with the addition of 50 ~l of the tissue suspension (rat and rabbit:

200-250 ~g protein per tube: bovine: 100-125 ~g protein per tube).

Following a 5.25 minute incubation at 30oe. the reaction was terminated

with the addition of 100 ~l of a solution containing 2% sodium lauryl

sulphate. 45 mM ATP and 1.3 mM cAMP. The internal standard [311]cAMP

(diluted in water to give approximately 30.000 dpm) was then added to

the samples. which were subsequently put in a boiling water bath for

three minutes. After cooling the samples to room temperature. the conversion of [a-32P]ATP to [a-32 P]cAMP was quantitated by sequential chromatography. Samples were loaded onto Dowex columns. and then eluted with water onto neutral alumina columns. The final fraction of interest

(as determined previously from column elution profiles - see Appendix A) was eluted from the alumina columns with 0.1 mM imidazole (pH 7.3) directly into scintillation vials containing 10 ml of Safety Solve scintillation fluid (RPI Corp .. Mount Prospect. IL). These samples were then counted in a Packard scintillation counter (B460e) using a program for 3H/ 32 p dual labelling.

2.2.4 5-HT 1A Receptor Binding Assay

The r3H]8-0H-DPAT binding assay was performed by Georgina Lambert. using cortex from male Sprague-Dawley rats. Details of this protocol are outlined in Appendix B. 42

2.2.5 Data analysis

Adenylate cyclase activity was corrected for relative recovery and expressed as "fmole cAMP/minute/~g protein" or as a fraction of 5-HT sensitive adenyl ate cyclase, as defined by the maximal inhibition of FSC achieved with 10 ~M 5-HT (routinely 25-30%). Maximal stimulation of adenylate cyclase by 10 ~M forskolin was 700-800% over basal levels.

Protein concentration was measured by the method of Lowry et al. (1951) using bovine serum albumin as the standard.

The [C50 value is the concentration of inhibitor causing 50% inhibition of [3H]8-0H-DPAT binding. The ECSO value refers to the agonist concentration that yielded SO% of maximal S-HT sensitive inhibition of FSC. IC50 values from the binding assays and ECSO values from the FSC assay were calculated using the nonlinear regression analysis software PC-NONLIN (Statistical Consultants Inc., Kentucky).

Ki values were calculated from binding data using the Cheng-Prusoff equation (1973),

1 + L

such that L is the concentration of [3H]8-0H-DPAT (1 nM), and Kd is the dissociation constant for [3H]8-0H-DPAT (4 nM). 43

Affinity constants for competItive antagonists were calculated by comparing the EC50 of the control 5-HT inhibition curve to the EC50 of the 5-HT inhibition curve in the presence of the antagonist with the following equation, as modified from Schild regression analysis

(Arunlaskshana and Schild, 1959; Ott et al., 1981):

[B)

[(A'/A) -1] where KB is the dissociation constant for the antagonist, [B] is the concentration of the antagonist, A' is the EC50 value of the agonist inhibition curve in the presence of antagonist and A Is the EC50 of the agonist inhibition curve in the absence of antagonist.

For a more rigorous assessment of antagonists, 3-dose pA2 studies were performed and the pA2 values were calculated according to Schild regression analysis, such that Log(A'/A)-l] Log(B] - Log(KS)' The plot of Log[(A'/A)-I] vs Log(B] yields a straight line whose x­ intercept is equivalent to the pA 2 value, which equals -Log(KS)'

Nonlinear regression analysis was used to estimate EC50 values (from a four parameter logistic equation -DeLean et al., 1978) for most inhibition curves from the FSC assay. Additional analysis for complex multiphasic inhibition curves involved comparing the fit of one­ site and two-site models with the F-test. Other statistical analysis included the T-test, as well as ANOVA, followed by the Newman-Keul's ad hoc test (Number Cruncher Statistical System, Utah). 44

2.3 Results

The data in the following sections (2.3.1 to 2.3.6) involved only

rut hippocampus. Species differences will be addressed in Section

2.3.7.

2.3.1 Response of known 5-HT1A agonists

5-HT and two other known 5-HT1A agonists, (±)8-0H-DPAT and buspirone (Figure 2A) were tested in our FSC system using rat hippocampus to compare their responses to previous reports. Tbese three compounds inhibited FSC, as shown in Figure 2B. The rank order of potencies of these compounds in our system (8-0H-DPAT > 5-HT > buspirone) was consistent with the literature (DeVivo and Maayani, 1986;

Bockaert et al., 1987). Buspirone, a well known 5-HT 1A partial agonist, inhibited only 50% of maximal 5-IIT sensitive adenylate cyclase.

Although 8-0H-DPAT has been previously reported to act as a full 5-HT1A agonist in a variety of systems, (Bockaert et al., 1987; DeVivo and

Maayani, 1986), it only inhibited up to 84% of the maximal 5-HT sensitive adenyl ate cyclase in our system. This observation will be discussed in further detail in Chapter 4.

The EC50 values of these three compounds in the cyclase assay were greater than their corresponding 5-HT1A binding Ki values in [3H]8-0H­

DPAT binding studies (Table 2). Nonetheless, the correlation between the cyclase EC 50 values and binding Ki values was reasonably good. This finding is consistent with the idea that the inhibition of FSC is H H

8-0H-OPRT 5-HT

BUSPIRONE A

-ui 1.2 ffi en 1.0 ...l: I III 0.8 ~ O.S -~ 0.4 ~ 0.2 ~ B lIJ 0.0 (.)~ ~ -0.2 -10 -9 -8 -7 -6 -5 -4 LOG [DRUG] . M

Figure 2A: Structures of the ('efercnce 5-HT1A ar,olliRtS. B: Inhibition of FSC by the reference 5-II1'lA al!Olljtll~: !j III' ( A ; n=l1), 8-0U-UrAT ( • : 11=:1), hus(lit'orw ( .; 11-1). 46

Table 2: COMPARISON OF 5-HTIA BINDING AFFINITIES (Ri) AND FSC ASSAY POTENCIES (EC50) FOR THE REFERENCE 5-HT1A AGONISTS.

COMPOUND AFFINITY FSC POTENCY RATIO Ki (nM)a EC50 (nM)b EC50/Ki

5-HT 7.4 ± 1.6 106 ± 16 (11 ) 14

8-0H-DPAT 5.5 ± 0.1 34.9 ± 6.9 (3) 6

Buspirone 21 ± 5.1 107 ± 9.8 (4 ) 5

a Values are mean ± SEM for at least three independent [3H]8-0H-DPAT binding experiments. b Values from the FSC assay are mean ± SEM (n value). 47

linked to 5-HTIA receptors. The ratio of EC50 (cyclase potency) to Ki

(binding affinity) values ranged from 5-14. and is in agreement with other reports comparing 5-HT 1A binding affinity to potency in the FSC assay (Frazer et al .. 1090).

2.3.2 Effects of 5-HT antagonists

Since a discrepancy existed between binding Ki values and EC 50 values for the reference 5-HT1A agonists tested (Section 2.3.1). the potencies of 5-HTIA antagonists in the FSC assay were compared to their respective 5-HT 1A binding affinities. The unfortunate lack of 5-HTIA selective antagonists is a definite limitation associated with research involving 5-HTIA receptors. and consequently the list of antagonists to choose from is very small. We have looked at three: spiperone. pindolol and BMY 7378. Another putative 5-HTIA antagonist. MDL 73005EF. wili be addressed separately in Chapter 3.

Spiperone (Figure 3A) has high affinity for 5-HT1A' 5-HT1C and

5-HT2 receptors. but has low affinity for 5-HTIB receptors (Pedigo et al .. 1981; Peroutka. 1988b). As shown in Figure 38. increasing concentrations of spiperone (up to 3 ~M) did not inhibit FSC. and reversed the inhibition produced by a fixed concentration of 1 ~M 5-HT.

At concentrations greater than 3 ~M. spiperone inhibited FSC through an apparently unknown mechanism.

Pindolol (Figure 4A) has high affinity for 5-HTIA receptors (Hoyer.

1988a; Herrick-Davis and Titeler. 1988) and is known to antagonize the 48

F HN-t~ z..(~~~ 6 o A

ui 1.2 -ffi en 1.0 &' t- ...-=-6.....- - _- ~---&--,..------:t: A-.-- - ... I . It) 0.8 r/ ~ 0.6 -~ 0.4 :> I ~ 0.2 ~ lIJ ~J. 5 0.0 B ~ -0.2 u -10 -9 -8 -7 -6 -5 -4 LOG [SPfPERONE]. M

Figure 3A: Structure of spiperone. B: Inhibition of FSC by spiperone ( A ; n=5) alone. and in the presence of 1 \.1M 5-HT ( 0 ; n=3). 49

pharmacological effects of 8-0H-DPAT (Gudelsky et al., 1986). In addition, binding of pindolol to 5-HTIA receptors is not GTP-sensitive, which is characteristic of antagonist binding (Oksenberg and Peroutka,

1988). In the FSC assay, pindolol had very low Intrinsic activity at concentrations up to 10 ~M, inhibiting less than 15% of 5-HT sensitive

FSC. Increasing concentrations of pindolol also reversed the inhibi t i Oil produced in the presence of 1 ~M 5-HT (Figure 48), suggesting blockade of the 5-HT1A receptor. At concentrations greater than 10 ~M, non­ specific (i.e., likely non-5-HT1A -mediated) inhibition of FSC was observed.

BMY 7378 (Figure 5A) is a relatively new compound with high affinity for 5-HTIA binding (Ki 2.4 nM) that was initially reported tu act as a 5-HTIA antagonist (Vocca et al., 1987). However, BMY 7378 appeared to act as a very weak partial agonist in our system

(Figure 58). Like pindolol, 8MY 7378 also reversed the inhibition produced by 10 ~M 5-HT in a dose-dependent manner. At best, BMY 7378 is a weak partial agonist, with low intrinsic activity at the 5-HT 1A receptor. The limited availability of BMY 7378 precluded its further use to characterize the FSC as a screen for compounds possessing potential 5-HT1A activity.

To further investigate whether the 5-HTIA binding affinities of spiperone and pindolol corresponded with their potencies in the FSC assay, pA 2 values were calculated using Schild plots (Table 3). There was no significant difference between binding Ki values and the 50 NHCH(CH 3 )Z

H

H A

1.2 1.0 ~~~~~T-I----r----­ ~ 1 A~.1 T ~ T ~ ~A"l--A 0.8 .1.1 1

0.6

0.4

~ 0.2 ~ ~ S 0.0 ~ -0.2+-----~----_T----~------T_----~----~--- -10 -9 -8 -7 -6 -5 -4-

LOG [PINDOLOL] t M

Figure 4A: Structure of pindolol. B: Inhibition of FSC by pindolol (A ; n=4) alone. and In the presence of 1 pM 5-HT ([J; n=3). 51

0 CH 3 0 \ N \ / 0

A

"""':" UJ 1.2 z lU UJ ...... T .... 1.0 --~------J: .I. &-...-..+-.- T I 0.8 ~ ~l · &n. -A u 1 ~ 0.6 - 0.4 ~ ~ 0.2 I ~ w 0.0 B u~ t; -0.2 -10 -9 -8 -7 -6 -5 -4 LOG [BMY 7378]. M

f1gure 5A: Structure of BMY 7378. B: Inhibition of FSC by BMY 7378 ( A ; n=5) alone. and in the presence of 10 11M 5-HT ( 0 ; n=2). 52

corresponding pA2 values for either compound. In addition to be being potent at 5-HT1A receptors, pindolol also has affinity for 5-HT18 and p-adrenergic receptors (Herrick-Davis and Titeler, 1988). As determined by Schild analysis, the slope value of the ability of pindolol to prevent 5-HT-induced inhibition of FSC was 1.05, thus suggesting interaction with only one receptor', namely the 5--HT 1A . In addition, the 5-HT-induced inhibition of FSC was shifted to the right in the presence of a fixed concentration of 10 ~M pindolol, suggesting competitive antagonism at the 5-HT1A receptor (Figure 6). Thus the blockade of

5-HT 1A receptors in this system by 10 pM pindolol resulted in the observed rightward shift in the 5-HT inhibition curve. We consequently chose to use pindolol in all subsequent experiments using the FSC assay system, in order to verify that any observed inhibition of FSC by novel compounds was actually due to 5-HTIA interaction. The fixed concentration of 10 pM pindolol was used, since at higher concentrations pindolol caused inhibition in a non-specific manner.

2.3.3 Response of aryltryptamines

The novel aryltryptamines tested were based on the prototype compound AHR 1709 (Figure 7). [3H]8-0H-DPAT binding studies have demonstrated that this series of compounds has high affinity for the

5-HT 1A receptor (Taylor, 1985). AHR 1709 and FPPEI did not significantly inhibit FSC up to a concentration of 10 pM. Further testing demonstrated that AHR 1709 appeared to act as a competitive antagonist in this system, due to its ability to cause a significant 53

Table 3: COMPARISON OF pKi VALUES FROM r3H]8-0H-DPAT BINDING AND pA2 VALUES FROM THE FSC ASSAY FOR THE 5-HTIA ANTAGONISTS SPIPERONE AND PINDOLOL.

COMPOUND BINDING ASSAY PSC ASSAY PKja pA2 slopeb

Spiperone 7.09 ± 0.08 6.96 1.19 (3) (2 ) (2)

Pindolol 6.66 ± 0.09 6.79 ± 0.16 1.05 ± 0.09 (4 ) (4) (4 )

All values are the mean ± SEM (n value). a As determined in (3H]8-0H-DPAT binding experiments. b As determined from Schild plot analysis. 54

~en 1.2 z I.&Jen T .... 1.0 J: ! V-t::~'?-¢--~ I II) 0.8 , '0 .l 0.6 ~ -" A -~ 0.4 \ A CJ~ 0.2 c( I.&J \. °'-6 0.0 ------4_~--- ~ .. A. __ • CJ >- -0.2 CJ -10 -9 -8 -7 -6 -5 -4 LOG [S-HT], M

Figure 6: Ability of pindolol to shift the 5-HT Inhibition curve. 5-HT ( • ; n=3) alone. and in the presence of 10 ~M pindolol ( 0 ; n=3). 55

COUPOUNo R 1 R2 R2 AHR 1709 -H -HJ-©

r-', FPPEI - OCH 3 -H\...... /H -©-\ / F R 1 -CONH SCAHR 2 -HJ-© SUEOAHR - OCH 3 -HJ-©

BOT - OCH 3 -N~,(MH 0

figure 7: Structures of aryltryptamines. 56

shifl to the right of the 5-HT inhibition curve (Figure 8).

Interestingly. the calculated Ks of 232 nM for AHR 1709 was significantly different from its 5-HTIA binding Ki of 44 nM.

Furthermore. comparable to the previous KS calculation. Schild plot analysis with AHR 1709 yielded a pA2 value of 6.66 ± 0.19 (n=6) that was also significantly different from its corresponding pKi of 7.36 ± 0.05

(n=3) from [3H]8-0H-DPAT binding (p < 0.025). Consequently. further work is needed to identify the effect of AHR 1709 on 5-HT 1A intrinsic activity. FPPEI has yet to undergo further testing for antagonistic activity. since it has lower selectivity. as well as affinity (Ki

188 ± 12 nM) for the 5-HTIA receptor. as compared to AHR 1709.

The remaining novel aryltryptamines. namely 5CAHR. 5MEOAHR and BDT. exhibited potential 5-HT1A agonistic activity by inhibiting PSC with varying degrees of efficacy. The corresponding 5-HT1A binding affinities were compared with cyclase EC50 values in Table 4. The inhibition of PSC produced by 5CAHR was shifted in the presence of pindolol (Figure 9A), suggesting that 5CAHR is a 5-HT1A agonist.

Nonlinear analysis of the inhibition curve for 5CAHR resolved it into two components. and the ratio of EC 50 (of the high affinity component) to the Ki (from [3H]8-0H-DPAT binding studies) for 5CAHR was within the range of 5-14, as observed for the 5-HTIA reference compounds previously tested (namely 5-HT. 8-0H-DPAT and buspirone). 57

en 1.2 -z lIJ 111 1.0 I­ J: I I/) 0.8 CJ ~ 0.6 -?= 0.4 ~ 0.2 lIJ 0.0 ~ ~ -O.2+------r------~----~----~------~----~---10 -9 -8 -7 -6 -5 -4 LOG [5-HT], M

Figure 8: Ability of AHR 1709 to shift the 5-HT inhibition curve. 5-HT (4 n=3) alone. and In the presence of 10 ~M AHR 1709 (0; n=3). 58

Table 1: COMPARISON OF 5 HTIA BINDING AFFINITY (Ki) AND FSC ASSAY POTENCY (EC50) VALUES FOR THE NOVEL ARYL TRYPTA.~ I NES 5C,\HR. 5MEOAHR AND BOT.

COMPOlIND BINDING AFFINITY FSC ASSAY POTENCY RATIO K i (nM) a EC50 (nM)b EC50,Ki

5CAHR 9.6 ± 4.7 8

5MEOAHR 9.2 1: 0.7 35

BOT 28 ± 4 1290 ± 290 46

a Values are the mean 1: SEM for at least three independent [3H)8-0H-DPAT binding assays. b Values are the mean ± SEM for at least three independent experiments. c EC50 of the high potency component of the inhibition curves. 59

i :~ !: 0.1 J, 0.1 -i 0.4 I ~ -0.2 ~ -0.4 B 8 -o.l~--~--~~--~--~----~--~-- -to ... -7 -. -5 LOG (sa.cEOAHRJ. u i 1.2 51 1.0 ~ J, 0.1 t 0.. i 0.4 ~ 0.2 I ~~--~----~--~----~--__ ---~-~. -10 -0 -t -1 -. -5 LOG (SOT]. y

Figure 9A: Inhibition of FSC by 5CAHR ( A : n~3-6) alone. and in the presence of 10 liM pindolol (0 : n=3). B: Inhibition of FSC by 5MEOAHR ( A : n=3-6) alone. and in the presence of 10 liM pindolol (0 : n=4). c: Inhibi tion of FSC by BOT ( A ; n=3-7) alone. clnd in the presence of 10 liM pindolol (0 : n=3). 60

Although 5MEOAHR inhibited FSC. pindolol did not produce i\

significant shift in the 5MEOAHR-inducerl inhibition curve (Figure 98).

suggestillp, t.hcll. the observt'd inhibition was iil.:tuclily due 1.0 il lion

;'-HTlA-mediatp.d mPI'hilOism(s). The nonlinear analysis resolvpd thp

5MEOAHR inhibition curve into high and low potenr.y components with n

proportion of high affinity si t.es (Bill was calculatl~d to be only 22% ot"

the total sitp.s. T1w ratio of the Ee50 of the high potency component. t,l

the binding Ki was 35. Tn an attempt to determine if 5MEOAIIR might be a

5-HT 1A antagonist. 5-HT dose-response curves were examined in the absence and presence of fixed concentrations (i.e .. 10 ~M and 0.1 ~Ml of

5MEOAHR. The intrinisic activity of 5MEOAHR has yet to be fully established. however. since the non-5-HT 1A -mediated inhibition produced by 5MEOAHR complicated data analysis. Thus. the ability nf ~ compoulld. such as 5MEOAHR. to interact with lIon-5-HT 1A -mediated mechanisms to produce inhibition uf FSC in this system reinforces the importance of using a S-HTIA antagonist, such as pindolol, to verify

5-HT1A interactions ..

The BOT-induced inhibition curve reached a plateau of about 70%

inhibition at a concentration of 30 ~M. This inhibition curve was also shifted to the right in the presence of pindolol (Figure gel, suggesting that BOT is a 5-HTIA partial agonist. However, the ECSO/Ki ratio was much greater than the ratios for the reference 5-HT 1A agonists (namely 5-HT, 8-0H-DPAT and buspironel. and no correlation seemed to exist between the potency in the cyclase data and [3H]8-0H-DPAT binding 6 I

affinity data. despite the ability of pindolol to affect the BDT-induced inhibition. This lack of correlation has yet to be resolved.

2.3.1 2 ·AMBD

The structllre of 2-·,\MBD (Figure lOA) was designed as a high rlffinity 5HT IA ligand. using the 5-·HT1A pharmacophol'e propospd 11>'

Hi bert et al. (1988a). Al though this compound was proposed to fit th~' i I' model. it was never reportedly tested to confirm its putative high

5-HT 1A binding affinity. Consequently. we were interested to test both its 5-HTIA affinity and intrinsic activity. 2-AMBD !I~t~allY had only moderate binding affinity for the 5-HTIA receptor, with a Ki of

421: 82 nM (n=5). It inhibited FSC in a dose-dependent manner. producing almost full inhibition of 5-HT sensitive adenylate cycJadP. which was pindoJol sensitive (Figure lOB). Although 2-A,."fBD had weak potency in the FSC assay (EC 50 = 29.500 ~ 4300 nM). it appeared tn be almost a full 5-HT lA agonist. The ratio of cyclase EC50 to binding I\j was 70 in this case.

2.3.5 5-methoxytryptamine dimer (5MEO-dimer)

The 5MEO-dimer (FIgure 11A) was synthesized in order to determille whether the accessibility of an indole at either end of the molecule would facilItate 5-HT1A binding. 5-Methoxytryptamine, itself. had high

5-HT 1A affinity (Ri = 6 nM; Taylor. 1985). The 5MEO-dimer also had high 5-HTIA affinity, with a Ki of 17.2 ± 0.7 nM (n=3), and inhibited FSC in a dose-dependent manner (Figure lIB), with an EC50 of 739 ! 241 nM. As 62

A

1.2 -uiz l&J .-en 1.0 l: I ~>

Figure lOA: Structure of 2-AMBD. B: Inhibition of FSC by 2-AMBD ( A ; n=4) alone, and in the presence of 10 llM pindolol (0 ; n=3). 63

OCMS

OCMS

A

1.2 -uiz L&.I • A• - en +-4-~!-..1 T l- %: '" .I.-~T I A 111 0.6 ~ '\1 A -~ !" T.---A. ~ 0.0 ------1- ~T - -- ~ L&.I t'T S B A (,J >-. 1 (,J. -0.6 -10 -9 -8 -7 -6 -5 -4 LOG [SMEO-OIMER], M

figure 11A: Structure of 5MEO-dimer. B: Inhibition of PSC by 5MEO-dimer (A n=4). 64

obsprv~d with other compounds In the FSC assay. th~ ratio of ECso!Ki'

which was 43. was larger than the rutios calculated for thp. reference

5-HT 1A ngonists (see Section 2.3.1). This compound was testpd prior t~

the st.andard practice of using pindolol to verify 5-HTIA inter."lctloll of

tes t:i ng.

2.3.6 Spiroxatrine

(~)Spiroxatrine (Figure 12) is structurally related to the S-HT antagonist spiperon~, and was investigated In a search [01' analogs of spiperone that retained high affinity for S-HTIA receptors and low affinity for S-HTIB receptors. but which had decreased affinity for

5-HT 2 receptors (Nikam et al .. 1988; Nelson and Taylor. 1986).

(:)Gpiluhatrine has previously been shown to interact with 5-HT 1A binding sites. in a GTP-sensitive manner (Nelson et al .. 1987). suggesting that it possesses some S-HT1A agonistic activity. In the FSC assay, (:)spiroxatrine produced a complex inhibition curve with a shallow slope, and inhibited cAMP production to an extent greater than

S-HT itself (Figure 13A). Although a fixed concentration of 1 ~M plndolol did not produce a significant shift in the inhibition curve.

10 ~M pindolol successfully caused a very slight, yet statistically significant (p < 0.05) shift to the right, suggesting that only a small f'lS

R ( + )

s ( - )

Figure 12: Structure of spiroxatrine enantiomers. 66

proportion of the observed inhihition induced by spiroxatrine was actually due to 5-HTIA interaction.

Nonlinear regression analysis of this complex inhibition curve t'cs()ll:l~d it into high .Jnd low potency components, providing furt.hl'l' evidence that only a small proportion af the inhibition induced b~

{.f.):;piroxatrine could he iiltrihuted to S·HT1A illt.eract.io/l (pro()(H,tillll lit' high affinity sites, or RH' was 47'1». Although the ['atio of thp. EC50 O! the r~ntire (':)spiroxat.t'ine inhiuition curve, to its I{i from [3 I1 j8-0/l­

DPAT binding studies (Table ~) did /lot correlate with the ratios for the reference 5-HT 1A agonists (namely 5-HT, 8-0H-DPAT and buspirone), the ratio of the EC50 of the high potency component to the binding Ki was consistent wi th the ratios for 5-HT, 8-0H-DPAT and buspirone. Based ')11 these results, (")spir-I)xatrine at best can be considered a weak 5-IIT 1A pur-tial agonist.

At concentrations greater than 10 ~M, (~)spiroxatrine inhibited FSC to an extent greater than 5-HT itself, suggesting the involvement of non-5-HT 1A -mediated mechanisms. (!)Spiroxatrine has high affinity at

D2-dopamine (Nikam et al., 1988) and opiate receptors (Leysen et al., 1977), but the addition of the antagonists haloperidol and naloxone hud no effect on the abi! i ty of spiroxatr ine to inhibit FSC, thus sugges ling that dopaminergic and opiate systems were not involved in the non-

5-HTlA-mediated effects.

Enantiomers of spiroxatrine were also tested in the FSC assay and like the racemate, produced complex inhibition curves (Figure 13B). 67

- 1.2 '"~ !C o.a I 1ft ! 0.4 '~~ 0.0 ------~~--- ~ -<1.4 A ~i I -0.1 -10 -9 -8 -7 -6 -5 -~ LOG [(+/-)SPIROXATRINE] .... i 1.2 !C 0.8 I 1ft ~ 0.4 - 0.0 ~ -0.4 B I -0.8 -10 -9 -8 -7 -6 -5 -~ LOG [ENANnOUER]. U

figure 13A: Inhibition of FSC by (±)spiroxatrine ( 6 ; n=6) alone, and in the presence of 10 ~M pindolol ( () ; n=~), B: Inhibition of FSC by (-) spiroxatr ine (. ; n=7) and (+)spiroxatrine (. ; n=7), 68

Nonlinear analysis rpsolved the inhibitiun curve for the ()erlflntiomp.l·

into two components, while (+)spiroxatrine produced a less potent, yet paraIlpj inhihition curve. It is interp.::;ting t.o note that the

~~Ilantinmel's displayed both stereoselectlve 5-HTIA binding and intl'insi,' activity (Table 5). The t.rend wherehy t.he ( )p.nt.llltiumer was mol'l? pntr~r\! th'ln the (·)enantiomer was consistent at other receptors te::;ted (nampjy al·adrenergic, D2-dopaminergic and 5 -HT 2 - Nikam et ill., 1988), and demonstrates that the effects observed in the FSC assay were likely l·ecept.or mediated, as opposed to being non-specific. Thus, the

(+)enantiomer gave a Ki value that was 11-times greater than the

(-)enantiomer, and an EC 50 that was 13-times greater than the

(-)enantiomer. The ratios of EC50/Ki for the (-) and (t) p.nantiomers were 30 and 36, respectively. These enantiomers were examined prior to the routine use of pindolol to verify the involvement of 5-HT 1A ill t.he observed inhibition of FSC. Due to extremely limited quantitip.s of these compounds, the effect of pindolol, and thus the extent of 5-HT1A interaction, has yet to be established, barring separation of more of these spiroxatrine enantiomers. 8ased on the results for (±)spiroxa­ trine, it is likely that only a small proportion of the inhibition curves produced by (+) and (-)spiroxatrine would be pindolol sensitive.

2.3.7 Species DIfferences

Concerns existed about the apparent lack of correlation between cyclase EC50 and binding Ki values for some of the compounds tested in the PSC assay using rat hippocampus, and possible alternatives were T;lble 5: COMPAHISON OF 5HTl,\ IH~/)ING AFFINITY (Kj) AND FSC ASSAY POTENCY (EC50 J VALIJES rem SPIR()XATRINE AND [TS ENANTIOMERS.

SP[ROX. BTNDING AFFINITY FSC ASSAY POTENCY RATIO Ki (nM)Cl EC50 (nM)b EC50/Ki

(:: ) 3.9 - 0.4 31 8

( ) 1.9 .t 0 . .1 58 30

( ~ ) 21.1 ,)'" . -I 754 36

a All values are the mean ± SEM for at least three independent t 3 H]8-0H-DPAT binding experiments. b EC50 of high potency component of the average inhibition curve from at least six experiments. ;0

considered in an attt!mpt to improve the system and facilitate assessment of functional activity associated with the 5-HTIA receptor. Based on previuus \\'o['k demonstrating that (!::)Spil'oxatrine inhibited FSC ill itll apparently monophasic manner in calf hippucampus (Schoeffter- alld Hoyel'.

1988). we cons0quently investigated species other lhan rat (namely bovine and rabbit) in an attempt to ascertain which species would bp best to use fur our screening purpOst!s. It appears t.hat the cOllplillg or

5-HT receptors to adenyate cyclase is very much dependent on the species. as well as tissue. under investigation (DeVivo and Maayani,

1988) .

We also were interested to investigate whether these othp[' species contained a greater proportion of 5-HT sensitive adenylate cyclasl! in the total FSC activity. as compared to a proportion of about 25-30% 5-HT sensitive adenylate cyclase in rat hippocampus. It was pussible that u better signal-ta-noise ratio (SNR), as well as a larger 5-HT1A interaction. could further improve the ~creening potential of the fSC assay. The SNR can be calculated as the ratio of the magnitude of the response to the standard error of the mean measurement of baseline activity (DeVivo and Maayanl. 1988). It is reportedly difficult to obtain accurate EC50 values for concentration response data obtained in systems with a SNR of less than 10 (DeVivo and Maayani. 1988). The following Is an example calculation of the SNR using typical data from the FSC assay carried out with rat hippocampal membranes: The difference between forskolin-stimulated cyclase (233 fmole cAMP/~g protein Iminute) and maximally inhibited cyclase in the presence 71

of 10 11M 5-HT (167 fmole ci\,"1P/lll~ protein/minute) was 66 fmole cAMP/ill{ protein/minute. The SEM of the forskolin-stimulated cy~lase activity

W.1S (J.tl4. thu!) yielding a SNR of 66/0.64 = 103.

In rabbit hippocampus. the 5-HT-lnduced inhibition of Fsr was similar' to that seell in the rat. with an EC50 of 50 nM (Figure 14'\).

The maximal inhibition of FSt: ill the presence of 10 11M 5··HT was routinely 15-20% and the SNR was typically around 25. As demonstrated by' the shj ft ill the presence of pindolol. (t)spiroxatrine produced a complex inhibition curve that contained a larger 5-HT 1A component than observed with rat hippocampus (Pigure 14D). In addition. rab~it caudate nucleus was tested in the FSC. Although total cyclase actIvlty was very high in the caudate nucleu!). none of it appeared to be sensitive tl)

10 11M 5-HT. To our knowledge. this is the first report where the FSC assay has been used to detect 5-HT sensitive adenylate cyclase ill r~bbit brain.

Using bovine hippocampus. the dose-response curve for the 5-HT­ induced inhibition of FSC was comparable to that seell in the rat and rabbit (Figure 15A). The maximal inhibition of FSC in the presence of

10 11M 5-HT varied between 10-20%. and at best. the SNR was about 62.

Nonetheless. the profile for (±)spiroxatrine was significantly different

(Figure 158). A much larger shift to the right of the (t)splroxatrine­ induced inhibition curve was observed in the presence of pindolol. as compared to either the rat or the rabbit. Nonlinear analysis of the

(±)spiroxatrine inhibition curve fit a one-site model for concentrations i2

1.2 i 1.0 A !i: I 0.8 ." ~~\ l 0.8 ,,~ 0.4 I 0.2 \\ 0.0 - - - -~- ; ,~;--h:0 -0.2 -10 -9 -6 -7 -6 -5 -4 LOG [5-HT]. M

1.2 i B % 0.8 I ." ! 0.4 " ~ 0.0 "

~ -0.4

~ -0.8 -10 -9 -8 -7 -6 -5 -4 LOG (SPIROXATRINE]. M "

Figure 14: Inhibition of FSC in rabbit hippocampus. A: 5-HT ( A ; n=4) alone. and in the presence of 10 ~M pindolol ( 0 ; n=l). 8: (±)Spiroxatrine (A ; n=4) alone. and in the presence v! 10 11M pindolol (0 ; n=3). 73

1.2

1.0

0.8 l 0.8 0.4 I 0.2 ; 0.0 -i -8 -1 -5 LOG (5-HT], ...

1.2

0.8

0.0

-a.4

-0.8 ...... --..--....,...--.,...... ---,.------,.-- -10 -9 -8 -7 -6 -5 LOG [SPIROXATRINE] ....

Figure 15: Inhibition of FSC in bovine hippocampus. A: 5-HT ( • ; n=3) alone. and in the presence of 10 ~M pindolol ( 0 ; n=3). B: (:!:)Spiroxatrine ( A ; n=4) alone. and in the presence of 10 ~M pindolol (0 ; n=3). 74

up to 10 ~M. From the three species examined. (!)spiroxatrine appeared

to have the greatest 5-HT 1A interaction in bovine hippocampus. based on the pindolol shifts. We therefore decided to focus our species

comparisoll helw~en rat and bovine. despite the fact that bovine

hippocampus had a lower proportion of 5-HT sensitive FSC. as well as a

lower SNR. than observed ill the rat.

Since the nuvel aryltryptamines 5MEOAHR and BOT had EC50Ki ratins

thilt did not appear to correspond with the reference 5-HT 1A agonists (i.e .. 5--HT. 8-0H-DPAT and buspirone) in rat hippocampus. these two

compounds were tested. both in the absence and presence of 10 ~M

pindolol. in bovine hippocampal membranes. Corresponding to previous observations in the rat. 5MEOAHR produced H complex inhibition curve

that was not significantly shifted in the presence of 10 ~M pindoloi in bovine hippocampus (Figure 16A). In addition. at concentrations greater

than 10 ~M. 5MEOAHR inhibited FSC to an extent greater than 5-HT ilself. as seen in the rat. The inhibition curve for 5MEOAHR appeared to contain high and low potency components. with an EC 50 of 19 nM (SH 39%) for the high potency component. As with rat hippocampus. BOT was a partial agonist in bovine (Figure 16B), but the magnitude of the pindolol-induced shift was less than that observed in the rat. suggesting comparatively less 5-HT 1A interaction by BOT in bovine hippocampus. The proposed activities of (±)spiroxatrine. 5MEOAHR and

BOT in bovine hippocampus are summarized in Table 6. 75

1.2

0.8

0.4

-0.4 -0.8 A -1.2

-us ..... -~~---,.---,..--....---__ ---4- -10 -9 -8 -1 -6 -5 LOG [5MEOAHR]. M

1.2 i 1.0 !i: I 0.8 ." _~ 0.8 0.4 0.2 ~~ ~ 0.0 -----~- ~ -0.2 B G -0.4+---'~-"",,---"'--"""--""------10 -9 -8 -7 -6 -5 -4 LOG [BOT]. M

FIgure 16: Inhibition of FSC in bovine hippocampus. A: 5MEOAHR (A ; n=3-4) alone. and in the presence of 10 lJM pindolol ( 0 ; n=3-4). B: BOT ( A ; 0=3) alone. and in the presence of 10 lJM pindolol ( 0 ; n=3). 76

Each time additional bovine brains were collected, the resultant batch of newly prepared hippocampal membranes were tested in the FSC assay to check their responses to 5-HT. Although the 5-HT inhibition curves among batches of bovine hippocampal homogenates were comparable. disconcerting inconsistencies were observed in the ability of

(±)spiroxatrine to inhibit FSC (Figure 17). This raised concerns associated with using cattle. which consist of highly inbred strains.

From batch to batch. the bovine brains obtained from slaughterhouses were from unknown strain, sex and age of cattle. We also had no control over the length of time that passed from time of death to plac~ment of the brains in ice-cold saline, thus leading to variable times of commencement of the ensuing tissue preparation. Consequently, we decided to forego any further investigation using bovine tissue. and continue using rat hippocampal tissue for all subsequent SAR work with the PSC assay.

2.4 Discussion

The PSC assay, which has been previously been used to demonstrate that S-HTIA receptors are negatively coupled to adenylate cyclase, was used to screen novel compounds of varying chemical structures for potential 5-HTIA intrinsic activity. Although the rank order of potencies of reference 5-HT1A agonists (namely 5-HT, 8-0H-DPAT and buspirone) in the PSC assay had the same rank order as their corresponding 5-HTIA binding affinities, the ECSO values in the FSC assay were greater than their corresponding binding Ki values. 77

Table 6: SPECIES COMPARISON OF PROPOSED 5-HTIA ACTIVrTIES IN THE FSC ASSAY.

COMPOUND PROPOSED ACTIVITY

Rat hippocampus Bovine hippocampus

5-HT agonist agonist

5MEOAHR antagonista . b antagonista . b

BOT partial agonist partial agonist

(± )SPfROX. partial agonist agonist or antagonista . b . c

a Although the inhibition curves were insensitive to 10 pM plndolol. further testing is required to confirm putative antagonism. b Based on the inability of pindolol to produce significant shift in inhibition curve. Further testing required to confirm 5-HTIA antagonism. C Response varied among batches of tissue. 78

-ui 1.2 ~ (IJ 0.8 ...J: I 111 0.4 -~ 0.0 J:=~ -0.4 ~ ~ -0.8 A 1 ~ -1.2 +---..,.---r---___ ---r--~-___ .___ -10 -9 -8 -7 -6 -5 -4 LOG (SPIROXATRINE] • M

Figure 17: Example of variation of response to (±)splroxatrine ubserved in the FSC assay among batches of bovine hippocampus. Inhibition of FSC by (=)spiroxatrine ( A ; n=3) alone. and In the presence of 10 ~M pindolol (0 ; 0=3). 79

Nonetheless, the ratios of EG50/Ki among these reference compounds were within an order of magnitude, ranging from 5-14.

The lower potency of 5-HT 1A agonists in the FSC assay, as compared to 5-HT 1A binding affinity, did not appear to be due to re-uptake or breakdown of 5-HT in the FSC assay system, nor could it be attributed to tissue concentration effects (see Appendix C for details). Although this discrepancy betw~en EC SO and Ki values for S-HTIA agonists has yet to be fully understood, there have been previous reports of this situation, such that agonist ECSO values measured in adenylate cyclase assays were consistently greater than their respective affinity (Ki) values in binding experiments (Hoyer et al., 1985; DeVivo and Maayani,

1988; Cornfield et al., 1988). The trend, whereby the average EC50/Ki ratios were around 10, has been previously documented for 5-HT 1A , as well as other, receptors (Frazer et al., 1990; Hoyer et al., 1989).

This generally observed discrepancy between cyclase EC50 potencies and Ki binding affinities may reflect differences due to coupling of

G proteins, especially caused by tissue preparation, or possibly be attributed to different assay conditions. For example, binding affinIty is usually measured under equilibrium conditions, while potencies determined in the cyclase assay typically are not. Another factor which could contribute to this discrepancy is that the binding and FSG assays employ differing buffer compositions. The affinity for S-HT1A sites is determined in the absence of GTP, which favors formation of a state of the receptor different from that present in the FSG assay, where GTP is 80

an absolute requirement (Sills et al., 1984b). In fact, agonist EC 50 values from the FSC assay were apparently In much better quantitative

agreement with their respective Ki values at the 5-HTIA receptor when

measured in the presence of GTP than with values calculated when GTP was

not present (Frazer et al., 1990).

GTP. which is present in the FSC assay but not typically in the

5-HTIA binding assay, has been shown to decrease agonist affinities for a receptor without affecting the binding of antagonists. The fact that

the pA2 values calculated in the FSC assay for the 5-HT 1A antagonists (pindolol and spiperone) were not statistically different from their respective binding affinities suggests that G protein coupling is a

likely factor involved in the discrepancy between functional potency and binding affinity for the agonists tested.

This FSC assay provides data that challenges the proposed 5-HT 1A pharmacophore of Hibert et al. (1988a). 2-AMBD, which was proposed to be a high affinity 5-HT1A ligand based on this model, possessed only relatively low 5-HTIA binding affinity, and acted as a weak 5-HTIA agonist in the FSC assay. These data reinforce both the importance of having a functional system with which to test compounds with putative

5-HTIA activity, and the usefulness of the FSC assay to investigate the structural features involved in 5-HT1A intrinsic activity.

The use of functional potencies of agonists to classify receptors can be confounded by different degrees of "spare receptors" in various tissues. Very recently, studies utilizing the irreversible receptor 81

antagonist EEDQ (N-ethoxycarbonyl-2-ethoxy-l,2-dihydroquinoline)

demonstrated a large receptor reserve for 5-HTIA agonist-induced

inhibition of 5-HT synthesis in rat brain. which appeared to be mediated

by the somatodendritic 5-HTIA autoreceptor (considered to be

"presynaptic") in rat hippocampus (Meller et al., 1990; Hjorth and

Sharp. 1990). Consequently. partial agonists can elicit maximal

apparent response in the presence of receptor reserve. It was

postulated that postsynaptic 5-HT1A receptors in terminal areas have little receptor reserve for agonists, and that a differential receptor

reserve at pre- and postsynaptic 5-HT1A receptors may underlie.the pharmacological differences observed at these sites with 5-HTIA

agonists. The FSC assay reflects activity at the postsynaptic 5-HTIA

receptor. Although the hippocampus has been proposed to contain little

or no receptor reserve. the issue of calculating receptor reserve in the

FSC assay (i.e .. with agents such as irreversible antagonists) has yet

to be addressed. Since partial agonistic activity can be assessed with

this FSC assay system, it is a powerful tool with which to differentiate agonistic, partial agonistic and antagonistic activity.

In conclusion, the FSC assay is an in vitro biochemical model with which to study functional consequences of interaction with 5-HT 1A receptors negatively linked to adenylate cyclase. Rat hippocampus contained the greatest proportion of 5-HT sensitive adenyl ate cyclase and the highest SNR. as compared to bovine and rabbit tissue. Although it is a feasible method with which to screen compounds for 5-HTIA activity. data obtained from this assay must nevertheless be interpreted 82

cautiously, due to some limitations associated with this system. The ability of some compounds to inhibit FSC via non-5-HTIA-mediated mechanisms warrants the use of a selective 5-HT1A antagonist, such as plndolol. to verify a 5-HTIA interaction. The possibility exists that if a compound interacts with only a relatively small proportion of

5-HT 1A receptors (i.e., high affinity binding sites (BH) < 45%), the shift could be too small to be detectable. The ability of some compounds to inhibit PSC via a combination of 5-HTIA and non-5-HT1A - mediated mechanisms complicates the interpretation of data, and in this situation, only a tentative assessment of intrinsic activity can be made.

With these limitations in mind, the PSC assay can be used as a functional screen to assess intrinsic activity at 5-HT1A receptors, and ultimately contribute important information about structural features associated with 5-HT1A intrinsic activity. Suggested guidelines for analyzing data from the FSC assay are as follows:

A) Qualitative assessment of 5-HTIA intrinsic activity:

1. If the compound inhibits FSC in a dose-dependent manner, the

inhibition must be verified for 5-HTIA interaction by use of a

fixed concentration of 10 ~M pindolol. A shift to the right in

the inhibition curve would suggest competitive antagonism at the

5-HT 1A receptor. such that the compound possesses 5-HT1A agonistic activity. In addition. if the inhibition is due to 5-HTIA­

mediated mechanisms, the'Ks calculated for pindolol with the 83

compound under investigation should be the same for 5-HT. If

pindolol does not produce a detectable shift in the inhibition

curve, it is likely that the observed inhibition can be attributed

to non-5-HT 1A mediated mechanism(s).

2. If the compound does not appear to significantly inhibit PSC,

the ability of the compound to reverse the inhibItion caused by a

fixed concentration of 5-HT (i.e., 1 ~M, but choice of 5-HT

concentration ultimately depends on the Ki of the compound under

investigation) should be assessed. Full reversal of 5-HT-induced

inhibition of FSC suggests full antagonistic activity.

B) QuantitatIve assessment of 5-HTIA Intrinsic activity:

3. Once the Inhibition curves have been assessed in steps 1 or 2,

data can usually be expressed as "% inhibition of 5-HT sensitive

cyclase", as determIned from the minimum calculated from nonlinear

analysis, using a four parameter logistic equation (DeLean et al.,

1978). The 5-HT1A intrinsic activity can then be categorized according to the magnitude of inhibition produced (i.e., full

agonist, partial agonist or potential antagonist).

The calculation of "% inhibition" can be complicated by the

presence of interfering non-5-HT1A mechanisms that cause

inhibition (usually at high concentrations above 10 ~M), or by

inhibition curves that do not reach a plateau within the

concentration range tested. In this case, it may not be possible 84

to assign a value to "% inhibition", and assignment of 5-HT 1A activity must be based on the relative magnitude of shift produced in the presence of 10 ~M pindolol.

4. For agonists, nonlinear analysis can generally be used to estimate EC50 values (from the four parameter logistic equation), although it is more difficult to apply this approach to weak partial agonists. Analysis of shallow, complex inhibition curves can be attempted, and the fit of one- and two-site models to the data set can be compared. Inhibition curves that do not appear to reach a plateau by the highest concentration tested, or appear to be mainly due to non-5-HT1A interaction, are also more difficult to assess using nonlinear analysis, and accurate EC50 values are unlikely to be calculated in such cases.

5. Further tests of antagonistic activity using the FSC assay

Include: the ability of a fixed concentration of the compound to shift the 5-HT-induced inhibition curve (KB determination), as well as the determination of pAa values using Schild analysis. 85

CHAPTER 3

MOL 73005EF: ANALYSIS OF A PURPORTED 5-HTIA ANTAGONIST USING THE FSC ASSAY

3.1 Introduction

This chapter has been adapted from Cornfield et al. (1989) and will specifically discuss the purported 5-HTIA antagonist, MOL 73005EF. in order to illustrate the importance and practical application of using the forskolin-stimulated adenylate cyclase (FSC) assay as a functional screen for compounds possessing affinity for the 5-HTIA receptor.

The structure of MOL 73005EF (8-(2-[(2,3-dihydro-l.4-benzodioxin-2- yl)methylamino]ethyl]-8-azaspiro[4,4]-decan-7.9-dione methane suI phon- ate) was previously determined using computer-assisted molecular modelling (Hibert et al .. 1988b) in an attempt to rationally design

5-HT1A antagonists. MDL 73005EF was recently described as a potent. highly selective 5-HTIA ligand, with an rC50 of 2.5 nM (Hibert et al .•

1988a). Based on several behavioral tests, MOL 73005EF was proposed to act predominantly as a 5-HTIA antagonist. Owing to the current dearth of selective 5-HTIA antagonists. there is great interest in their development. and we were consequently interested to test MOL 73005EF for potential antagonistic activity in the FSC assay.

Of the in vivo functional tests recently used to evaluate

MOL 73005EF, all but one demonstrated its potential 5-HTIA antagonistic activity. For example. in the punished drinking model, low doses of 86

MDL 73005EF caused a dose related increase in open arm entries. while

8-0H-DPAT and buspirone caused ~ decrease In both the percentage of open arms entered. and the total number of arms entered (Moser et al .. 1988).

In addition. unlike 8-0H-DPAT and buspirone. MDL 73005EF demonstrated putative anxiolytic activity In the plus-maze test (Mosl~r et a1.. 19138).

MDL 73005EF also blocked the 8-0H-DPAT-induced inhibition of contractions of transmurally stImulated guinea pig ileum. Forepaw treading and flat body posture were induced by 8-0H-DPAT. while

MOL 73005EF had minimal agonist effects. or antagonized the 8-0H-DPAT­ induced response. depending on the dose of MOL 73005EF used (Hioert et al .. 1988b). The one exception to all these observations suggesting antagonistic activity was the inability of MOL 73005EF to antagonize the

S·Off ··DPAT response in the 8··0H-DPAT discrimination paradigm. In fact.

MDL 73005EF generalized to 8-0H-DPAT (Hibert et al., 1988b).

Owing to potential problems associated with the interpretation of results from behavioral tests, we have used the inhibition of FSC to examine the functional consequence of the interaction of MDL 73005EF with 5-HT1A receptors negatively coupled to adenylate cyclase using rat hippocampal membranes. In addjtion, we compared the activity of

MDL 73005EF (Figure 18) to two structurally related 5-HT 1A partial agonists (previously discussed in Section 2.3), namely buspirone and the novel aryltryptamine, BDT (N-(2-(5-methoxyindol-3-yl)ethyl)-2-amino­ methyl-1,4-benzodioxane). 87

~ NyN (") O~ 051 N ~O yo N N < ~ CH 0 °LJO °LJO 3 H

BUSPIRONE UDL 7300SEf BOT

Figure 18: Structures of buspirone. MOL 73005EF and BDT. 88

3.2 Methods

3.2.1 Chemicals

MOL 73005EF was a gift from the Merrell Dow Research Institllte

(CIncInnati. Ohio). All other compounds and chemicals were obtained as

described previously in Chapter 2 (Section 2.2.1).

3.2.2 FSC assay

The inhibition of FSC was performed as described previously using

male Sprague-Dawley rats (Section 2.2.2). Data were analysed as

previously described.

3.3 Results

Despite previous reports emphasizing its predominant antagonistic

activity. MDL 73005EF interacted with 5-HT 1A receptors negatively coupled to adenyl ate cyclase as an efficacious partial agonist. in rat

hippocampal membranes (Figure 19A). In the FSC assay. MOL 73005EF had

an EC50 of 39.5 ± 6.7 nM (mean ± SEM. n=3). A fixed concentration of

10 pM pindolol shifted the MOL 73005EF inhibition curve to the right.

suggesting competitive antagonism at the 5-HT1A receptor. The resultant

KB value calculated for pindolol was 108 ± 21.2 nM (n=3). which was

consistent with its 5-HT1A binding Ki' as well as the KS calculated for pindolol in the FSC assay using Schild analysis (see Table 3). 89

1.2 -iii i 1.0 ~ ~~--o ~ .,I o.a '-"-, l 0.1 '-\ 0.4 'I-.. +-6--~~, 0.2 A i • 0.0 - -- I -0.2 -10 -i -8 -7 -6 -5 - .. LOG [MOL 73005EF]. M

-iii 1.2 ~ 1.0 !i: .,I 0.8 ~ 0.8 - 0.4 ~ 0.2 ~ B 0.0 -0.2 ~ -10 -9 -8 -7 -6 -5 -4 LOG (DRUG] • M

Figure 19A: Inhibition ot' FSC hy MnL 7:l005EF ( Ai. ; n=3) aIOIlI!. or ill the presence of 10 ~M pindolol (0 ; n=3). B: Comparison of the acL:Lvltles of MDl. 73005EF (t::,.; n~-:n. buspi rone ( 0 ; n.,-4) nllc! nOT ( 0 ; n"'3-7) in til!' FSC assay. 90

MDL 73005EF shares structural similarities with buspirone and BDT.

As discussed in Chapter 2. buspirone and BOT acted as partial 5-HTIA agonists in the FSC assay. The rank order of potencies of these compounds in the FSC assay (MOL 73005EF > buspirione z BOT) corresponded with their affinities for the 5-HT 1A receptor. as determin(~d previously with [3H)8-0H-OPAT binding (Hibert et al .. 1988b; Cornfield et al ..

1988). Up to a concentration of 10 pM. the intrinsic activities of

MOL 73005EF. buspirone and BOT were 72 ± 2.3. 54 ± 4.4 and 55 ± 1.9%. respectively. of the maximal 5-HT sensitive adenylate cyclase activity

(Figure 19B). In fact. MOL 73005EF was significantly more efficacious than both buspirone and BOT (p < 0.05 .• ANOVA followed by Newman-Keuls post hoc test). At concentrations higher than 10 pM. the inhibition due to buspirone reached a plateau. while MDL 73005EF and BDT appeared to further inhibit cyclase activity via non-5-HTIA -mediated mechanism(s).

3.4 Discussion

The structure of MOL 73005EF was developed using a computer­ graphics model proposed to define the pharmacophore for 5-HT 1A antagonists (Hibert et al .. 1988a). As predicted. MDL 73005EF certainly had high affinity for 5-HT1A receptors (as well as selectivity). as previously reported. Despite recent reports predominantly describing

MDL 73005EF as a 5-HTIA antagonist (based on several in vivo tests). however. we have demonstrated that MOL 73005EF has SUbstantial agonistic activity at the 5·HTIA receptor negatively linked to adenylate cyclase. 91

Since the drug discrimination paradigm was the only reported behavioral test that did not demonstrate antagonistic activity of MOL 73005EF at the 5--HT1A receptor, this test may provide a better index of fllnctional activity of 5-HTIA selective compounds, as observed in the FSC assay. than the other behavioral tests used previously.

MOL 73005EF is structurally quite similar to buspit'one and the novel aryltryptamine, BOT, two compounds which have previously been shown to have partIal agonist activity at 5-HTIA receptors (DeVivo and Maayani, 1986; Cornfield et al., 1988). Both buspirone and DDT contain a heterocyclic group in common with MOL 73005EF. The rank order of potencies of these three compounds in the FSC assay was consistent with their corresponding affinities at the 5-HTIA receptor. Interestingly,

MOL 73005EF was a more efficacious 5-HT1A partial agonist than either buspirone or BOT, and consequently MOL 73005EF may have potential clinical anxiolytic activity.

Another compound designed using the computer-graphics generated

5-HTIA antagonist model of Hibert et al. (1988a) was MOL 72832, which also has high affinity for 5-HTIA receptors. The structure of

MOL 73005EF is identical to MOL 72832, with the exception that

MOL 73005EF has an ethyl (instead of a butyl) group separating the two heterocyclic moieties. Since MOL 72832 has previously been shown to be a potent ligand with mixed agouist and antagonist properties at 5-HT 1A . receptors (Mir et al., 1988), it is not surprising to find that

MDL 73005EF has similar properties. 92

Despite previous reports that MOL 73005EF was predominantly a

5-HT 1A antagonist, the mixed agonistic and antagonistic properties of MOL 73005EF should not be overlooked when considering future studies with this compound, and need to be clarified to assess the functional consequences associated with the interaction of MOL 73005EF at the

5-HTIA receptor. In fact, recent work reporting that MOL 73005EF, like buspirone, is best described as a partial agonist at 5-HTIA receptors

(Moser et al., 1990) confirms our data. These results concerning

MOL 73005EF demonstrate the usefulness of the FSC assay as a functional screen for 5-HTIA activity. 93

CHAPTER 4

STRUCTURE-ACTIVITY RELATIONSHIPS OF ENANTIOMERS OF 8-0H-DPAT AND ITS ANALOGS

4.1 Introduction

A substantial number of reported analogs of the selective 5-HTIA agonist 8-0H-DPAT have been used to investigate structural features required for 5-HT1A agonist binding (Glennon. 1987; Arvidsson et al ..

1984. 1987; Bjork et al .. 1989; Mellin et al .. 1988; Cossery et al ..

1987; Wikstrom et al .. 1982. 1988. 1987). The two n-propyl groups of

8-0H-DPAT were previously shown to be necessary for optimal activity in a biochemical measure of serotonergic activity (specifically. the receptor-mediated feedback inhibition of 5-hydroxytryptophan (5-HTP) accumulation). Aminotetralin analogs with alkyl groups smaller than n- propyl were less active. whereas those with larger substituents were essentially inactive (Arvidsson et al .. 1984). The di-n-propyl portion of 8-0H-DPAT has since been proposed to make a significant (> 50-fold) contribution to its affinity for 5-HT1A sites (Hoyer et al .. 1985). In contrast. other researchers have proposed that it was not the N.N- dipropyl groups that accounted for 5-HT1A selectivity, but rather some feature associated with the pyrrole portion of the indolylalkanamines

(Glennon et al., 1988b). Ring-expanded 8-0H-DPAT analogs have also been investigated (Liu et al .. 1989). Despite all this work, structure- activity relationships (SAR) to define the pharmacophore for the 5-HT IA 94

receptor, including stereochemical requirements, have yet to be fully

elucidated.

In order to assess the effect of structural differences on

illteraction with 5-HTIA receptors, a serIes of structural analogs of

8-0H-DPAT were synthesized. The enantlomers of 8-0H-DPAT, as well as

enantiomers of these 8-0H-DPAT analogs, have already been evaluated for

both 5-HTIA binding affinity and in vivo 5-HTIA activity (Hjorth et al.,

1989; Hillver et al., 1990; Mellin et al., in preparation). From this

data, Mellin (1990) has proposed a pharmacophore model for 5-HT 1A agonists, which will be discussed further in Section 4.4. We have

further analyzed some of the enantiomeric pairs of novel 8-0H-DPAT

analogs for 5-HT 1A intrinsic activity with the in vitro FSC assay. Thus, this chapter provides a unique perspective concerning the

assessment of SAR for 5-HT 1A ligands, and is the first time that the 5-HTIA receptor negatively linked to adenylate cyclase has been

characterized in this manner.

4.2 Methods

4.2.1 Chemicals

The enantiomers of 8-0H-DPAT and its analogs (Figure 20) were synthesized under the direction of Dr. Uli Hacksell (University of

Uppsala, Sweden). These compounds were all water-soluble salts. The compounds tested in the FSC assay were: (+)S-OH-DPAT [(+)-(2R)-S- OH OH ':," J

~ .•••• \",.( CHzCHzCH,) Z C H, C H, CH,) I 00 ... "". ( 95 ~ (+) 1I-0H-OPAT (+) ALK-l

OH

(-) I-OH-OPAT (-) ALK-l

OH OH \,\ II ( e HIe HI eH, ) I ~ " ( CH t CIi t CII ) ) 1 .. ", ." .\'

CH, (+) eW-12 r ( .. ) H-II OH OH II (CHI CHI CH, >z "(CHtCHtCH,)l

(-) CW-IZ r (-) H-If NoH ~ (+) L£A-141 i~ (+) JV-U.%

(-) JY-ZI. I

Figure 20: Structures of enantiomers of 8-0H-DPAT and structural analogs. 96

hydroxy-2-(di-n-propylamino)tetralln.HBr), (-)8-0H-DPAT [(-)-(25)-8-

hydroxy-2-(di-n-propylamino)tetralln.HBr], (+)JV-26 [(+)-(4aR,10bR)-

1,2,3,4,4a,5,6,lOb-octahydro-lO-hydroxy-4-propylbenzo[f]quinoline.HCl).

(-)JV-26 [(-)-(4aS,10bS)-1,2,3,4,4a,5.6,lOb-octahydro-l0-hydroxy-4-

propy lbenzo (f)quinoline. HC 1], (+) CM--12 [( i-) - (2S. 3S) - trans-·8 -hydroxy -:l­

methyl-2·-(di-n-propylamino)tetralln.HCl], (-)CM-12 [(-)-(2R,3RI-trans-8-

hydroxy-3-methyl-2-(di-n-propylamino)tetralin.HCl], (+)ALK-3 [(t)­

(lS,2R)-cis-8-hydroxy-l-methyl-2-(di-n-propylamino)tetralin.HCI],

(-)ALK-3 [(-)-(lR,2S)-cis-8-hydroxy-l-methyl-2-(di-n-propylamino)­

tetral in. HCI], (+ )LEA-146 [( +) - (1S, 2R) -trans-2- (2-hydroxyphenyt) -N, N-di­

n-propylcyclopropylamine.HBr], (-)LEA-146 [(-)-(IR,2S)-trans-2-(2-

hydroxyphenyl)-N,N-di-n-propylcyclopropylamine.H8r], (+)SE-61 [(+)-(R)-

5-fluoro-8-hydroxy-2-(di-n-propylamino)tetralin.HBr] and (-)SE-61 [(-)­

(S)-5-fluoro-8-hydroxy-2-(di-n-propylamino)tetralin.HBr). All remailling

compounds were obtained and solubilized as previously described (see

Section 2.2.1).

4.2.2 FSC Assay

The FSC assay was carried out with rat hippocampal membranes, as

described previously (Section 2.2), using 100-125 ~g of protein per sample. Affinities of these novel compounds at the 5-HTIA receptor were determined by Georgina Lambert, using the [3H)8-0H-DPAT binding assay

(Appendix 8). 97

As discussed in Chapter 2. the screening of these compounds for

5-HT 1A activity was carried out as follows: Any compound that inhibited FSC in a dose-dependent manner and whose inhibition curve was shifted to the right in the presence of 10 ~M pindolol. was thus considered to possess 5-HT 1A agonistic activity. A compound was considered to possess full 5-HT 1A agonistic activity if it inhibited 100% of the maximal 5-HT sensitive adenyl ate cyclase (as defined with 10 ~M 5-HT). Anything less would indicate partial agonism. If a compound did not appear to significantly inhibit FSC in a dose-dependent fashion. it was then tested in the presence of a fixed concentration (i.e .• 1 ~M) of 5-HT. to determine the extent to which the compound was able to reverse the 5-HT­ induced inhibition of FSC. Full reversal of the inhibition of FSC produced by 5-HT would suggest antagonistic activity. Anything less would suggest weak partial agonistic activity.

4.3 Results

All 5-HT1A binding affinity data for the enantiomers of 8-0H-DPAT and its analogs are summarized in Table 7. along with the corresponding estimates of EC50 for the compounds demonstrating agonistic activity ill the FSC assay. For compounds where an accurate EC50 value could be determined. the majority of the calculated EC50/Ki ratios ranged between

12 and 22 (as compared to a range of 5 to 14 for the reference 5-HT 1A agonists 5-HT. 8-0H-DPAT and buspirone in Chapter 2). although the

EC50/Ki ratio for (+)SE-61 was 109. Nonetheless. all cyclase EC50 values were greater than the corresponding binding Ki values. Owing to 98

Table 7: COMPARISON OF 5-HT1A BINDING AFFINITIES TO FSC ASSAY POTENCIES FOR ENANTIOMERS OF 8-0H-DPAT AND ITS ANALOGS.

COMPOUND AFFINITY POTENCY RATIO % c Ki(nMl a EC50b EC50/Ki INHI BITION

(-)8-0H-DPAT 6.1 ± 0.6 135 ± 51 22 47 ± 3

(+)8-0H-DPAT 4.1 :!: 0.3 57.4± 11 14 101 ± 8

(-)ALK-3 2920±219 NOd NOd 35 1: 9

(+)ALK-3 2.9 ± 0.3 ND d ND d 22 ± 5

(-)CM-12 1389± 31 NOd NOd NS e

(+)CM-12 49.6:!: 4.2 643 ± 63 13 78 ± 2

(-)SE-61 127 ND d NOd NS e

(+)SE-61 3.3 356 ± 198 109 47 ±3

(-)LEA-146 8.0 ± 1.1 176 ± 79 22 90 1: 5

(+)LEA-146 923 ± 66 ND d NOd NS e

(-)JV-26 3.9 ± 0.2 47.1± 8.3 12 82 1: 3

(+ )JV-26 32.3± 3.2 458 ± 245 14 28 ± 9

All values are the mean ± SEM for at least three independent experiments. a 5-HT1A binding affinity, as determined using [3U]8-0H-DPAT. Apparent Kl values determined using the Cheng-Prusoff equation (1973). b Potency in the FSC assay, as calculated using nonlinear regression analysis. c % inhibition of 5-HT sensitive adenylate cyclase, as determined from the minimum value calculated using nonlinear regression analysis. d ND - not determined. e NS - no significant inhibition of FSC observed. 99

the difficulty in quantitatively assessing 5- HT I A potency for the weak partial agonists (+) and (-)ALK-3, only a qualitative assessment of activity was assigned to these compounds. None of the compounds

inhibIted FSC to an extent greater than 5-HT itself.

The effect of each enantiomeric pair of compounds on the inhibition of FSC will be discussed indivIdually:

4.3.1 8-0H-DPAT

Although the enantiomers of 8-0H-DPAT had virtually identical affinities for the 5-HTIA receptor, there was a dramatic dIfference between the activities of these two enantiomers at the 5-HT1A receptor negatively coupled to adenylate cyclase. Although (-)8-0H-DPAT was only a partial agonist, inhibiting FSC about 50% of the maximal 5-HT sensitive adenylate cyclase (Figure 21A), (+)8-0H-DPAT was a full 5-HT 1A agonist (Figure 21B). Both inhibition curves contained pindolol sensitive components.

4.3.2 ALK-3

The only difference between the structures of ALK-3 and 8-0H-DPAT

is that the former compound has a cis-methyl substituent in the 1-carbon position. Such a simple modification resulted in tremendous stereo­ selectivity in 5-HT1A binding, as compared to the parent compound,

8-0H-DPAT. Although the magnitude of did not appear to be stereoselective, the activities of the (-)ALK-3 and (+)ALK-3 were 100

-~ 1.2T------~ ~ !E, '.0 ~ t ,....,,~:::-~\------1ft 0.8. / t 1 • ~ 0 \ 0, _ o.e· ~.1 T ~i_· ,-, • A ~ 0.4· . o-a ~ 0.2 A

0.0+-----~--~----~----~----~--~~ ~ -10 -9 -8 -7 -6 -5 LOG [(-)8-0H-OPAT]. M

1.2~------~o 1.0·~ -~ ------~ 0 ~o-a, 0.8 ., ~ "I 0, 0.6 -" _~ 0.4· '"A 0 \ ~ 0.2 . \T 0, ~ t--.i-..,. 0, ~ 0.0· . ~l T f B l ..... ·-A a-0.2 1 1 u -O.4+-----~--~----~------~--~~ -10 -9 -8 -7 -6 -5 -4 LOG [( + )8-0H-OPAT]. M

Figure 21A: InhIbition of FSC by (-)8-0H-DPAT ( A ; 11=3) alone and in the presence of 10 ~M pindolol ( 0; n=I). B: InhIbition of FSC by (+)8-0H-DPAT (A ; n=3) alone and in the presence of 10 ~M pindolol ( () ; n=1). 101

-ui 1.2 ~ 1.0 ~ I ~~~tt~~- '" 0.8 ,. A-A?6\. 1 0.8 ... ~ 0.4 ~ 0.2 A

0.0 I -10 -9 -8 -7 -6 -5 -4 LOG [( - )AlK-J). ~

_~ O.S·

0.4· B

O.O+----t----,..--..,....--,----..,..----.---J -10 -9 -8 -7 -6 -5 LOG [( + )ALK-J). M

FIgure 22A: InhibItion of FSC by (-)ALK-3 ( A ; n=4) alone and In the presence of 10 ~M pindolol (() ; n=3) B: Inhibition of FSC by (f-)ALK-3 (A ; n=4) alone and in the presence of 1 ~M 5-HT ( () ; n=3). 102

dramatically reduced, compared to the 8-0H-DPAT enantiomers, rendering

both of these enantiomers weak partial agonists at the 5-HTIA receptor.

Pindolol produced a significant shift (p < 0.05) in the (-)ALK-3-induced

inhibition curve (Figure 22A). (+)ALK-3 dose-dependently reversed about

80% of 5-HT-induced inhibition (Figure 228).

4.3.3 CM-12

The structure of CM-12 differs from 8-0H-DPAT by only a trans­

methyl group in the C3 position. Although such a minor change in the

prototypical structure evoked stereoselectivity with respect to 5-HT1A

binding affinity, the overall affinities of the CM-12 enantiomers was

decreased, as compared to the 8-0H-DPAT enantiomers. The CM-12 enantlomeric pair also displayed stereochemical differences in

functional activity at the 5-HTIA receptor negatively coupled to adenylate cyclase. (-)CM-12 produced no measurable agonist activity and appeared to fully reverse 5-HT-induced inhibition (Figure 23A). In contrast, (+)CM-12 had high intrinsic activity, inhibiting almost 78% of

the 5-HT sensitive adenylate cyclase, and this inhibition was very pindolol sensitive (Figure 238).

4.3.4 SE-61

SE-61 is identical in structure to 8-0H-DPAT, with the except jon of the presence of a fluorine in the 5-carbon position. This enantiomeric pair demonstrated stereoselectivity jn both 5-HTIA affinity and 103

-o.s+---~----~----~----~--~----~~ -10 -9 -8 -7 -6 -5 -4 LOG [(-)CM-12], M

O.O+---~----~----~----~--~----~~ -10 -9 -8 -7 -6 -5 -4 LOG [( + )CU-12], M

Figure 23A: Inhibition of FSC by (-)CM-12 (. ; n=3) alone and In the presence of 1 pM 5-HT (0 ; 0=2). B: Inhibition of FSC by (+)CM-12 (A ; n=3) alone and in the presence of 10 pM pindolol ( 0 ; n=3). 104

-en 1.2 ffi T T l' s:111 1.0 i"--;..:-....:./~1 --- .. I . ~ 1 1ft 0.8 1 0.1 l: 0.4 ~ 0.2 IIJ 0.0 A g-0.2 -10 -9 -8 -7 -6 -5 -4 LOG [( - )SE-61]. M

-en 1.2 T ....~ 1.0 .... t.--~~ ------::c 1 6-6 I 1ft 0.8 ·'1 "T T 0.8· t---A_~.t. ~ ! .u- T "0 -~ 0.4 / § ~c 41( 0.2' cr '" 0.0 ~o-c/ B g-0.2 +-_-,...._---,-_...,..-.._-,...._---,~-..,..-.-I -10 -9 -8 -7 -6 -5 -4 LOG ((+)SE-61]. M

Figure 24A: Inhibition of FSC by ('-)SE-61 (4 ; n=4) alone and in the presence of 1 ~M 5-HT ( 0 ; n=2). B: Inhibition of FSC by (I-) SE-61 (A ; n=3) alone and in the presence of 1 ~M 5-HT ( 0 ; 11=2). 105

intrinsic activity. (-)SE-61 was a putative 5-HTIA antagonist since it

did not significantly inhibit FSC. and produced full reversal of 5-HT­

induced inhibition of FSC (Figure 24A). (+)SE-61 was a weak partial

agonist. since its reversal of 5-HT-induced inhibition converged to its

level of inhibition of FSC (Figure 248).

4.3.5 LEA-146

LEA-146 was constructed as a less conformationally constrained

compound. based on the aminotetralin structure. and is loosely considered a 8-0H-DPAT analog based on its spatial configuration. This analog showed stereoselectivity with respect to both 5-HT1A binding affinity and intrinsic activity. (-)LEA-146 was an efficacious 5-HT 1A agonist. inhibiting 90% of maximal 5-HT sensitive adenylate cyclase

(Figure 25A). This inhibition curve was shifted to the right in the presence of pindolol. In contrast. (+)LEA-146 was only a very weak partial agonist. as demonstrated by its ability to almost fully reverse

the inhIbitory effect of 10 ~M 5-HT (Figure 258). Since the intrinsic activity of (+)LEA-146 was so low. no quantitative measure of cyclase potency was attempted.

4.3.6 JV-26

Although the enantiomers of JV-26 had moderate stereoselectivity for 5-HTIA binding sites. there was greater difference between enantiomers with respect to intrinsic activity. The (-)enantiomer of 106

-en 1.2 • ~ 1.0 ------!:, If) 0.8 0-0" l 0.8 I 0, ~ 0.4 l \ r ". 0 0.2 I A-a-...• &-AT~ l"~ 0.0 A A l -0.2 ~ -10 -9 -8 -7 -6 -5 -04- LOG [( - )LEA-146]. M

1.2~------~

~ 0.6· - 0.4- ! 0.2· ~¥'~ ~ 0.0 1 1 B ~ -0_2+---~-----~~---~--~~--~--~~ -10 -9 -8 -7 -6 -5 -04- LOG [(+)LEA-146]. M

Figure 25A: Inhibition of FSC by (-)LEA-146 ( A ; n=3) alone and in the presence of 10 ~M pindolol (0 ; n=2). B: Inhibition of FSC by (+)LEA-146 ( A ; n=3) alone and in the presence of 1 ~M 5-HT ([] ; n=3). 107

.*_~ ------l: . ' J. T I ~~ 111 0.8 f-..l ~ ~ 0.& 1\ 0"0 -~ 0.4 +" .,'0 i 0.2 ~ ' •..A-~.-+ '" 0.0 A 0--0 aU -0.2 -10 -9 -8 -7 -6 -5 -04- LOG [(-)JV-26]. M

-ui 1.2.,.....------... ~ !i: I 111 ~ 0.& -~ 0.4 ~ 0.2 ~ \AI B ~ 0.0

~ -0.2+---~----~----~--~----~----~-10 -9 -8 -7 -6 -5 -04- LOG [( + )JV-26]. M

Figure 26A: Inhibition of PSC by (-)JV-26 ( 4 ; n=3) alone and in the presence of 10 ~M pindolol ( () ; n=2), B: Inhibition of FSC by (+ )JV-26 ( ... ; n=<3) alone and in the presence of 10 ~M pindolol ( (); n=2), 108

JV-26 was an efficacious agonist with almost 85% of intrinsic activity of 5-HT in this system (Figure 26A). This inhibition curve had a large

5-HT1A component, based on the ability of pindolol to shift the curve.

(+)JV-26 produced a much weaker inhibition of FSC (Figure 26B), which was completely blocked by 10 ~M plndolol.

4.4 Discussion

The FSC assay has provided some interesting results concerning SAR associated with this series of enantiomeric pairs of 8-0H-DPAT analogs.

The stereoselective activity of some of the tested enantiomers emphasizes the importance of studying enantiomers when evaluating the pharmacology of analogs of a compound, since erroneous conclusions may sometimes be drawn from results obtained with the racemates (Bjork et al., 1989).

The discrepancy between cyclase potency and binding affinity has yet to be resolved, and as a consequence, caution must be exercised when trying to quantitatively assess 5-HTIA intrinsic activity. The use of pindolol to block 5-HT1A receptors and evoke rightward shifts of inhibition curves clearly demonstrated 5-HTIA interaction as did the reversal of 5-HT-induced inhibition.

Generally, it appeared that for each pair of enantiomers tested, the enantiomer with the higher 5-HT 1A binding affinity also had the greater level of 5-HTIA agonistic activity in the FSC assay. The compounds excluded from this trend are 8-0H-DPAT, whose enantiomers did 109

not have stereoselective 5--HT1A binding affinity, and ALK-3, where both enantiomers displayed minimal stereoselective differences with respect to 5-HT 1A intrinsic activity.

The 5-HT 1A activity of these compounds was previously evaluated using two in vivo tests (Hjorth et al., 1989; Hillver et al., 1990;

Mellin et al., in preparation). The in vivo biochemical test determined activity at the presynaptic 5-HT1A receptor by indirectly measuring 5-HT synthesis. Due to receptor-mediated feedback inhibition of presynaptic neuronal activity, 5-HT1A agonists inhibit the rate of 5-HT synthesis

(Arvidsson et al., 1987). Following decarboxylase inhibition, the accumulation of the 5-HT precursor 5-hydroxytryptophan (5-HTP) was used as an indicator of the rate of synthesis of 5-HT. The second in vivo test monitored behavior associated with the "5-HT syndrome". Specific components of the 5-HT behavioral syndrome that reflect activity at the postsynaptic 5-HTIA receptor are forepaw treading, headweaving and tremor (Tricklebank et al., 1985; Middlemiss et al., 1985). In general. the 5-HTIA binding affinities of these compounds correlated with their agonist potencies in these two in vivo tests.

Overall. the 5-HT 1A activity of these enantiomers of 8-0H-DPAT and its analogs in the FSC assay corresponded well with their respective in vivo actions. Based on results from the in vivo tests, (-)CM-12,

(+)LEA-146 and (-)SE-61 were described as inactive analogs. since they did not significantly affect 5-HTP levels in non-reserpine-treated rats, and did not cause any signs of 5-HT syndrome behavior in reserpinized 110

rats. These analogs also did not significantly inhibit FSC. (+)CM-12, considered to possess moderate 5-HT1A potency in the in vivo tests,

inhibited 78% of 5-HT sensitive FSC.

Although (+)8-0H-DPAT was considered to be only slightly more active than its antipode based on in vivo data, the difference in activities was much more dramatic in the FSC assay. This discrepancy could sjmply be a reflection of differing receptor densities between the in vivo and in vitro systems. As mentioned in Chapter 2, the FSC assay measures activity at the postsynaptic 5-HT1A receptor, where there is

likely little or no receptor reserve. Consequently, the FSC assay provides a means of observing partial agonistic activity that would otherwise be seen as full agonistic activity in the presence of large receptor reserve.

Also, although the in vivo tests demonstrated (+)ALK-3 to be a fairly potent 5-HT1A agonist and (-)ALK-3 to be inactive, both these enantiomers appeared as weak partial agonists in the FSC assay. Since

the (+)ALK-3 enantiomer has the greater 5-HT 1A affinity, the (-)enantiomer may have such low potency relative to its antipode that no agonistic activity could be determined, under the concentrations tested in the in vivo tests.

The stereoselective difference in intrinsic activity, but not

5-HT1A affinity, for the enantiomers of 8-0H-DPAT itself are of particular interest. Previous data with (±)8-0H-DPAT (Section 2.3.1) suggested that, at best, it was a partial agonist with about 85% of the 111

intrinsic activity of 5-HT itself. This is in contrast to other reports

that 8-aH-DPAT was a full agonist at the 5-HT1A receptor linked to

adenylate cyclase (DeVivo and Maayani, 1986). This partial agonistic

activity for racemic 8-aH-DPAT might be consistent with the

(-)enantiomer antagonizing the (t)enantiomer at the higher

concentrations (up to 100 pM).

The introduction of a single cis-methyl-substituent in the I-carbon

position (see ALK-3, Section 4.3.2) of the reference 5-HT1A agonist,

8-aH-DPAT, dramatically improved the stereoselectivity of 5-HT 1A binding, as compared to the parent compound. The main effects due to

the pseudoaxial I-methyl group are a decrease in mobility of the di-n­

propylamino moiety and an increase in steric bulk on one face of the

tetralin ring system (Hjorth et al., 1989). This structural

manipulation also resulted in a significant reduction of 5-HT 1A agonistic efficacy.

Placing a trans-methyl substituent in the 3-carbon position (as with CM-12) also resulted in stereoselectivity, although with relatively

lower affinity, for 5-HT1A receptors (as compared to enantiomers of

ALK-3, as well as 8-aH-DPAT itself). Instead, there was a greater stereoselectivity with respect to functional activity, as compared to a cis-methyl group being in the I-carbon position.

Fluorine differs only slightly in size compared to hydrogen, and its effect on the acidity of the phenol group of SE-61 has been proposed

to 'be minimal (Hillver et al., 1990). Preliminary NMR-experiments and 112

molecular mechanics calculations do not indicate that the fluorine substituent induces any conformational changes. Therefore, major differellces between 8-0H-DPAT and SE-61 in 5-HTIA affinity and activity may be related to electronic distribution. Ultimately, th~ efficacy of

8-0H-DPAT may be altered by changes in electronic properties of the aromatic ring. It is interesting that the 5-fluoro-substituted analog of 8-0H-DPAT abolished efficacy in only one of the enantiomers (namely

(-)SE-61). It is very possible that further alterations of the sUbstitution pattern of this series of 8-0H-DPAT-based analogs could lead to further development of selective and potent 5-HT 1A antagonists, which would possess high and stereospecific 5-HT 1A affinity.

The enantiomers of JV-26 were considered to be potent 5-HT 1A agonists, based on their ability to reduce 5-HTP accumulation in non­ reserpinized rats. It is interesting to note that the reduction of

5-HTP accumulation was more pronounced after administration of (-)JV-26 than after the same dose of (+)JV-26. This is consistent with the activity of these enantiomers in the FSC assay, where (-)JV-26 was a more efficacious 5-HT1A agonist than its antipode.

As alluded to in the introduction to this chapter, Mellin (1990) has reported an elegant study in which the stereochemical and conformational characteristics of these, and other 8-0H-DPAT analogs, were calculated using X-ray crystallography and molecular mechanics. A relatively simple model for the 5-HT1A pharmacophore was proposed, which consisted of a flexible pharmacophore and partial 5-HTIA rec~ptor 113

excluded and essential volumes (Figure 27A). The compound JV-26 , with

its potent 5-HT 1A agonistic activity, was instrumental in the development of this model. This three-dimensional model accounted for

the obser'ved in vivo pharmacological activities and in vitro 5-HT1A binding affinities of these compounds based on molecular mechanIcs­ derived energies and geometries of potentially bioactive conformatlolls of the compounds discussed. The compounds tested were divided Into three groups, based on whether they were potent 5-HT1A agonists (i.e.,

Ki < 32.3 nM), moderate or weak 5-HT1A agonists (i.e., 49.6

The 5-HT 1A agonist pharmacophore (Figure 27A) proposed by Mellin (1990) consists of an aromatic site and a vector between the nitrogen and the acceptor, or dummy atom, in the 5 - lIT 1 A receptor. The pharmacophore (Figure 27A) is defined by:

1) the distance (y) from the dummy atom to the plane of the

aromatic ring,

2) the distance (x) from the normal of the aromatic centre to the

dummy atom,

3) the angle (a) between a vector connecting the nitrogen and the

dummy atom, and a vector resulting from summation of the

corresponding vectors for the enantiomers of JV-26, and

4) the angle (P) between the aromatic plane and that described by

the enantiomers of JV-26.

Note: the enantiomers of JV-26 were used to define the outer limIts of 114

y

<:

'I

y = 2.1 - 2.6 A x=5.2 - 5.7 A a = (-) 28 - (+) 28°

Cl ~ = (-) 4 - (+) OA

rigure 27: Pigures adapted from Mellin (1990). A: A 5-HT 1A agonist pharmacophore recently proposed by Mellin (1990), such that: X = the distance from the normal of the aromatic center to the dummy atom of the acceptor; Y = the distance from the dummy atom to the plane of the aromatic ring; a = the angle between a vector connecting the nitrogen and the dummy atom; ~ = the angle between the aromatic plane and the plane described by the enantiomers of JV-26. 115

OH

1 CCJ"~""

2 C):)· ... N/

B

Figure 27: Figures adapted from Mellin (1990) - continued. B: Superposition of some enantiomers of 8-0H-DPAT bilalogs. to illustrate the proposed alignment of these structures on the 5-HTIA receptor. using a dummy atom (corresponding to an oxygen atom on the 5-HT1A receptor) and the endpoints of a two angstrom vector passing through the centre of the aromatic ring and perpendicular to it. The analogs shown are (+)8-0H-DPAT (1). (-)8-0H-DPAT (2). (+)JV-26 (3). (-)JV-26 (4) and (+)ALK-3 (5). 116

3) and 4), due to their "extreme" geometries (Mellin, 1990).

The ranges of distances and angles defined by the compounds

considered to be potent 5-HTIA agonists may correspond to limits for all

optimal fit to this pharmacophore model for potent agonism. Moderately

or weakly potent compounds were not able to fulfill all the requirements of this model based on potent agonists. Consequently, low agonist

potencies are likely due to the inability of low-energy conformations to

fIt optimally to the pharmacophore, as defined by the high potency agonists. Another rationale for low agonistic activity could be the production of excess volume(s), as compared to the partial receptor­ excluded volume defined above.

At present, too few enantiomers of aminotetralins have been studied

in the FSC assay to develop a precise definition of the criteria for agonism, partial agonism and antagonism. Nonetheless, there are some

interesting features that these compounds present if Mellin's model is applied. In order to fit this model, the enantiomers have to be aligned at the receptor, and in so doing, different faces of the structures are exposed to the 5-HT 1A receptor. As already discussed, the enantiomers of 8-0H-DPAT were dramatically different with respect to 5-HT 1A intrinsic activity in the FSC assay, despite possessing virtually identical 5-HT 1A affinities. Using the Mellin's model, (+)8-0H-DPAT fits at the receptor such that its 8-hydroxyl group is aligned in the same position of the hydroxyl group of 5-HT itself (Figure 278). In contrast, (-)8-0H-DPAT must "flip" over, such that its 8-hydroxyl group 117

mimics the position of the pyridyl nitrogen of 5-HT, in order to satisfy this proposed pharmacophore. Therefore, the position of the hydroxyl, and not the pyrole, might be important for optimal 5-HT 1A intrinsic activity, and would account for the observed difference in intrinsic activity of the 8-0H-DPAT enantiomers.

When considering the ALK-3 enantiomers, the (+)enantiomer sits b~st on the receptor since its has the higher 5-HT1A affinity. Although the presence of the methyl group caused such as large decrease In measurable intrinsic activity in the F8C assay, no distinction could be made concerning the relative activities of the enantiomers. Based on the in vivo tests, (-)ALK-3 was considered an inactive compound, but it nevertheless fit well to the extended pharmacophore model, suggesting that its inactivity was related entirely to steric factors due to the methyl group.

With respect to (-)CM-12, conformations of fairly high energy fit to the extended pharamcophore model, but produced excess volumes.

Therefore it was not possible to unambiguously identify one single factor that rendered it inactive. JV-26 cannot be directly compared to these other aminotetralins since its structure has a third fixed ring.

Further extensive screening is required to determine whether any compounds of particular interest (i.e., the putative 5-HTIA antagonist

(-)8E-61) are truly selective for 5-HTIA receptors. In the past, the development of agonists was generally based on structures related to the transmitter they mimicked, while antagonists were large molecules that 118

usually belonged to many different chemical classes and were generally

nonspecific in action (Ariens, 1971; Lloyd and Andrews, 1986; Hibert et

al .. 1989). Compounds that were devoid of measurable intrinsic activity

in the PSC assay, yet had reasonably good 5-HTIA binding affinity

(such as (-)SE-61 and (-)CM-12), could lead t.he way in the development

of simple small molecules, based on agonist structures, that are

selective and potent 5-HT1A antagonists. Conformational preferences and

electrostatic potentials of these compounds must also be studied in order to find out if conformational and/or electronic factors may be of

importance for the 5-HTIA receptor activities of these compounds

(Arvidsson et al., 1988).

This new model for 5-HT1A agonism will undoubtedly be of some predictive value. The pharmacophore accommodates a relatively large set of compounds with different stereoselectivities and agonist potencies. and therefore appears to have some generality. Despite challenges associated with quantitatively assessing potency for some compounds with the PSC assay, the qualitative assignment of 5-HTIA activity through use of pindolol shifts and 5-HT reversals was generally in good agreement with the in vivo functional data, and supports this newly proposed

5-HTIA pharmacophore. Consequently, the PSC assay is a useful in vitro biochemical screen to assess 5-HT1A activity of these enantiomers of

8-0H-DPAT analogs and has contributed important information concerning the SAR associated with 5-HT 1A activity. 1 1 9

CHAPTER 5

DETERMINATION OF STRUCTURE-ACTIVITY RELATIONSHIPS FOR A SERIES OF 3- AND 4-TETRAHYDROPYRIDYL INDOLES

5.1 Introduction

A previous study of a series of 3-(1,2.5,6-tetrahydropyridin-4-yll-

indoles examined structural features that determined high binding

affinity and selectivity at both 5-HT 1A and 5-HT2 receptors (Taylor et

al., 1988). The structural feature that almost exclusively determined

optImal 5-HT1A binding affinity was the volume of the 5-indole

substituent. The optimal size was calculated to be about 24 cubic

angstroms, which is approximately the size of a carboxamido (-CONH 2) group.

The work discussed in this chapter extends the previous study by

examining relationships between structure and intrinsic activity at

5-HT 1A receptors in a larger series of tetrahydropyridinyl analogs, specifically 3-(1.2,5,6-tetrahydropyridin-4-yl)indoles (4-THPI) and 3-

(1,2,5,6-tetrahydropyridin-3-yl)indoles (3-THPI). These analogs have

either methyl or propyl sUbstituents on the tetrahydropyridyl nitrogen

and include a variety of substituents at the 5-indole position

(Figure 28). In this way, comparisons can be made concerning the

activity of compounds with only a subtle difference in structure, namely

the position of the pyridyl nitrogen. These compounds represent semi-

rigid analogs of 5-HT and as such, provide fewer possible conformations

for recognition by receptors than does 5-HT itself, which has a 120

x A

x

B

Figure 28: Generic structures of A) 3-THPI and B) ~-rHpr analogs. X 5-indole substituent y ~ pyridyl nitrogen substituent 12 1

relatively flexible aminoethyl side chain. This property of relative conformational constraint is useful in studies to determine structural features of compounds required for high potency and selectivity betwcp.n the different subtypes of 5-HT receptors (Taylor et al., 1988). The intrinsic activity of these analogs was assessed at the 5-HTIA receptor in the FSC assay, using rat hippocampus.

5.2 Methods

5.2.1 Chemicals

The 3- and 4-THPI analogs were synthesized by Drs. S.S. Nikam, Y.

Yang and T. Dahlgren, and were solubilized in water after pretreatment with 0.5% glacial acetic acid. All remaining compounds involved in the

FSC assay were obtained and solubilized as previously described (Section

2.2.1).

5.2.2 FSC Assay

The FSC assay was carried out with rat hippocampal membranes, as described previously (Section 2.2.2), using 100-125 ~g of protein per sample. Affinities of these compounds at the 5-HT 1A receptor were determined by Georgina Lambert, using the [3H]8-0H-DPAT binding assay

(Appendix 8). The screening of these compounds for 5-HTIA activity was carried out as described in Chapter 4 (4.2.2). 122

5.3 Results

The 5-HT 1A binding affinities (Ki values) of the THP! analogs and observed activities in the FSC assay (% inhibition values) are

summarized in Table 8. At concentrations greater than 10 pM. all the

compouJlds tested appeared to produce inhibition via non-5-HTIA

mechanisms. The activity of the individual THP! analogs in the FSC assay will be discussed according to the following categories:

4-methyl-. 3-methyl-, 3-propyl- and 4-propyl-THPI analogs.

5.3.1 4-methyl-THPI Analogs

Regardless of the 5-indole substituent [X = bromo (-Br), methoxy

(-OCH3) and methyl (-CH3)]' the 4-methyl-THPI analogs behaved similarly in the FSC assay. They all inhibited FSC in a dose-dependent manner. such that the inhibition curves were complex and did not reach a plateau in the concentration ranges tested (Figure 29). In fact. all three analogs inhibited FSC to an extent greater than 5-HT itself. suggestiJlg that these compounds either exert a non-specific effect at adenylate cyclase at high concentrations. or are acting at another receptor(s). in addition to 5-HTIA' to produce inhibition. These characteristics precluded non-linear analysis to obtain quantitative measures of potency. Nonetheless. the presence of 10 pM plndolol caused a large shift in the inhibition curve for each of these three analogs. suggesting that these analogs are at least partial agonists at the

5-HT 1A receptor negatively coupled to adenylate cyclase. 123

Table 8: SUMMARY OF 5-HTIA BINDING AFFINITY AND INTRINSIC ACTIVITY FOR 3- and 4-THPI ANALOGS.

THPI 5- INDOLE AFFINITY PROPOSED SUBSTITUENT Ki (nM)a 5-HTIA ACTIVITy b

4-methyl -Br 15 ± 1.2 Partial agonist - OCH3 21 ± 4.9 Partial agonist - CH 3 32 ± 2.0 Partial agonist 3-methyl -Br 47 ± 9.1 Partial agonist - OCH3 81 ± 16 Partial agonist - CH 3 73 ± 15 Partial agonist

3-propyl -Br 194 ± 48 Partial agonist - OCH3 215 ± 24 Partial agonist -NHCOCH3 199 ± 55 AntagonistC -CONH2 4.5 ± 0.3 Agonist

4-propyl -NHCOCH:J 35 ± 3.6 Partial agonist -CONH2 16 ± 2.7 partial agonist

a 5-HT IA affinity determined using [3H]8-0H-DPAT. Values are the mean ± SEM for at least three independent experiments. b Proposed 5-HT IA activity based on ability of pindolol to shift inhibition curve. or ability to reverse 5-HT-induced inhibition. c Observed inhibition not 5-HT IA-mediated. 124 i 1.3 ! U I u i III ---- ..... A -Br I -u -tl -e -e -7 ...... -4 LOG(DRUGJ.u

i U l U lo.o '------L. C -CH3

------lfl-.... 1-,.201------10 -t ... -1 ~..... -II -4 LOG (aWG). Y

Figure 29: Inhibition of FSC by 4-methyl-THP[ ~naiogs with varying 5-indole substituents: A. -Br ( A ; n=3) alone. and in the presence of 10 \.1M pindolol (0 ; n= 1) ; B. -OCH3 ( A ; n=3) alone. and in the presence of 10 \.1M pindolol (0; n=1); C. -CH3 ( .... ; n=4) alone. ilnd in the presence of 10 \.1M pindolol (0 ~ n=1). Each point represents the mean ~ SEM. 125

5.3.2 3-methyl-THPI Analogs

These three analogs were identical in structure to the 4-methyl­

THPI compounds [X = bromo (-Br). methoxy (-OCH3). methyl (-CH3)J. with the exception of the postiton of the pyridyl nitrogen. Similar results were obtained for all three analogs. These analogs weakly inhibited

FSC. producing multiphasic inhibition curves (Figure 30). At concentrations above 10 pM. the inhibition greatly increased. suggesting potential non-specific action or interaction with receptor(s) in addition to 5-HTIA receptors. The fact that the inhibition increased greatly above 10 pM. even in the presence of pindolol. suggests that it is a non-5-HT1A-mediated effect that causes inhibition above 10 pM.

Pindolol effectively blocked the inhibition produced by the methoxy­ analog (Figure 308). suggesting that this analog produced its slight inhibition via 5-HT1A receptors. The other two analogs (X = bromo and methyl) were capable of reversing 5-HT-induced inhibition. up to 10 pM

(Figure 30A&C). At higher concentrations. further inhibition occurred. thus supporting the theory that a non-serotonergic factor is involved in the general inhibition observed with concentrations greater than 10 pM.

It should be noted that in addition to the fact that the 3-THPI analogs were not only primarily weaker than the 4-THPI analogs with respect to

5-HT1A agonistic activity. the 3-THPI analogs also generally had lower

5-HT 1A affinity. 126 - 'JI A -Br I ,.. -- -- 1 u u 1 u i u

I.:~,. .. ~ -7 ...... -4 LOG (DRUG). U

- 1.2 a o.t

I CUI 1 IU I 0.0 B -OCH3 I~.. -t ... -1 -f -I .... LOG (JJRUG). U

C -CH3 - 1.2 a toG 1 GJI .4~~~t:+::!\ l GJI 0.4 I 0.2

I~,. -f ... -7 -t -5 .... LOG (ORUG). U

Figure 30: Inhibition of FSC by 3-methyl-THPI analogs with varying 5-indole substituents: A. -Br ( A ; n=3) alone, and in the presence of 1 ~M 5-HT (0 ~ n=1); B. - OCH 3 ( ~ ; n=3) alone, and in the presence of 10 ~M pindolol (0 ; n=1); C. -CH3 ( A ; 11=5) alone, and in the presence of 1 ~M 5-HT ([J ; n=3). Each point represents the mean ± SEM. 127

5.3.3 3-p~opyl-THPI Analogs

Like the previously mentioned 3- and 4-methyl-THPI analogs, the

3-propyl-THPI analogs with b~omo and methoxy substltuents in the

5-indole position produced inhibition curves that did not reach a plateau within the tested concentration range (Figure 3IA&S). PindoloL produced smaller shifts for these 3-propyl-THPI analogs than it did for the 4-methyl-THPI analogs. This difference suggests that perhaps the

3-propyl compounds had comparably weaker 5-HT1A agonistic activity.

Although the acetamido-substituted analog (Figure 3le) produced inhibition of FSC, pindolol did not produce a significant shift, thus demonstrating that the observed inhibition was due to non-5-HTIA mechanisms. The 5-HT IA binding affinity of this compound Is relatively weak and consequently reversal of 5-HT-induced inhibition was undetectable. Further testing is required to confirm that this analog is a 5-HTIA antagonist. The carboxamido-substituted 3-propyl analog had ve~y high 5-HT1A binding affinity and produced a multicomponent inhibition curve that had a large pindolol-sensitive portion

(Figure 310). This high degree of apparent 5-HT 1A activity is in direct contrast to the relatively weaker (or lack of) activity of the other three analogs in this 3-propyl series. 128

1.2 i 1.0 !i: ..I o.s l 0.8 0.4 I 0.2 A -Br 0.0 -0.2 I -10 -i -a -7 -8 -~ -4 LOG [DRUG]."

-.,; 1.2 § 1.0 !i: ..I 0.8 -~\-- 0.8 ! 0.4 '~'\~ 0.2 I 0.0 -0.2 8 -OCH3 \ -0.4 I -10 -9 -a -7 -8 -~ -4 LOG [DRUG] ...

Figure 31: Inhibition of FSC by 3-propyl-THPI analogs with varying 5-indole sUbstituents: A. -Br ( A ; n=3) alone, and in the presence of 10 llM pindolol (0 ; 11=2); 8. -OCH3 ( A ; n=4) alone, and in the presence of 10 pM plndolol ( 0 ; n=2); continued on next page. 129

-vi 1.2 &'i 1.0 .,f 0.1 0.8 l 0.4 0.2 ~ 0.0 ------0.2 C -NHCOC H3 -0.4 ~ -10 -I -3 -7 -6 -5 -4 LOG (DRUG] ....

Figure 31 - continued: C. -NHCOCH3 ( ... ; 11=:3) alone. in the presence of 10 11M i>indolol ( 0 ; n=2). as well as in the presence of 111M 5-HT (0; 11=1). D. -CONH2 ( ... ; n=3) :llone. and in the presence of 10 11M pindolol (0 '; n=3). Each point represents the mean ± SEM. 130

5.3.4 4-propyl-THPI analogs

These two 4-THPI analogs [X = acetamido (-NHCOCH3) and carboxamido

(-CONH2)] also produced complex inhibition curves in the PSC assay, although they did not inhibit to an extent greater than 5-HT itself, as seen with the 3- and 4-methyl-THPI a/lalogs. Both 4-propyl analogs appeared to be weak 5-HTIA partial agonists, as demonstrated by the ability of pindolol to shift the inhibition curve to the right

(Figure 32A), or the reversal of 5-HT--induced inhibition (Figure 32B).

The carboxamido-substituted 4-THPI analog was a dramatically weaker

5HTtA partial agonist, as compared to its 3-THPI homolog, des~ite less than a four-fold difference in 5-HTIA binding affinity. The general observation of non-5-HT1A -mediated inhibition occurring at concen­ trations greater than 10 ~M also applies to the 4-propyl-THPI analogs.

5.4 Discussion

The volume of the 5-indole SUbstituent has been previously determined to be important for 5-HT 1A binding affinity (Taylor et al.,

1988), and may also influence 5-HT1A agonistic activity. The 3-methyl analogs had lower 5-HTIA binding affinities than did the corresponding

4-methyl analogs. Similarly, the 3-methyl analogs appeared to possess comparatIvely lower intrinsic activity at the 5-IIT 1A receptor. This conclusion is based on the fact that the inhibition curves for the 4- methyl-THPI analogs experienced larger shifts in the presence of pindolol. Therefore, the 4-methyl-THPI analogs appeared to be 5-HTIA 131

1.2 B 1.0 !i: I o.a 1ft 0.8 I 0.. 0.2 I 0.0 ------0.2 A -NHCOC H3 0 -0.4 I -10 -a -8 -7 -a -5 -4 LOG (DRUG), ...

-wi 1.2 ~ , 1.0 t:..t= til +~------1ft 0.8 ~i ! 0.8 0.4 I 0.2 I O;G B -CONH2 -0.2 -10 -9 -8 -7 -8 -5 -4 LOG [DROO), u

Figure 32: Inhibition of FSC by 4-propyl-THPI analogs with varying 5-indole substituents: A. -NHCOCH3 ( ... ; n=3) alone. in the presence of 10 IlM pindolol (0 ; n=2); B. -CONH2 ( A ; n=3) alone. and in the presence of 1 IlM 5-HT () ; n=1). Each point represents the mean ± SEM. 132

partial agonists with relatively higher intrinsic activity, while the

3-methyl-THPI analogs appeared to be comparatively weaker, based on the pindolol shifts.

The 3-propyl analogs generally had lower 5-HTIA binding affinities compared to the 4-propyl analogs, as well as compared to the 3-methyl­

THPI analogs, with the exception of the carboxamido-substituted compounds. It is interesting to note that there are similarities between these THPI compounds and analogs of tryptamine. For example,

5-carboxamidotryptamine has very high affinity for 5-HT 1A binding sites (Ki = 0.21 nM - Hoyer et al., 1985), and was a full agonist in the FSC assay (DeVivo and Maayani, 1986), much like the 5-carboxamido-3-propyl­

THPI analog. It certainly appears that the carboxamido-substitutcd 4- propyl-THPI analog is a special case.

At concentrations greater than 10 pM, all the THPI analogs tested inhibited adenyl ate cyclase via non-5-HTIA-mediated mechanisms. Due to the presence of this interfering and undefined non-5-HTIA mechanism causing inhibition, assignment of quantitative measures of potency to these compounds would not be prudent.

Ideally, a very large number of compounds needs to be examined in order to perform rigorous SAR studies. Although there was no good correlation between 5-HTIA affinity and apparent intrinsic activity when considering all the analogs together. there appeared to be a trend, such that the 5-HTIA binding affinities within each of the four groups of

THPI analogs generally correlated with corresponding 5-HTIA activities 133

in the FSC assay. The higher the 5-HTIA affinity, tile greater the

5--HTIA involvement in the inhibition of FSC. Overall, the 4-THPI analogs generally had higher 5-HTIA affinity and intrinsic activity.

One major exception to this trend was the carboxamido-substituted propyl-THPl analogs. The 3-propyl-THPI analog had higher 5-HT 1A affinity and intrinsic activity than the corresponding 4-THPI, whicll had relatively low 5-HT 1A agonistic activity despite its high 5-HTIA binding affinity. A dramatic difference in 5-HT1A activity existed between these two analogs, despite their similar 5-HT1A affinities.

Given the varying results obtained with this series of compounds, no firm conclusions concerning structural features and 5-HTIA activity can be made. At the minimum, one can qualitatively assess agonistic and antagonistic activity with the use of the 5-HTIA antagonist pindolol.

Tills series of THPI analogs needs to be extended further with more homologs in both the 3- and 4-THPI series in order to further investigate the association of 5-HTIA affinity and activity. In addition, a procedure needs to be developed to eliminate non-5-HT 1A - mediated inhibition that complicates quantitative analysis of data from the FSC assay. 134

CHAPTER 6

SUMMARY

H.1 The Use of The FSC Assay as a Functional Screen for 5-HT 1A Activity.

One of the most significant problems currently facing 5-HT research

is the lack of agonists and antagonists selective for the myriad of 5-HT

receptor types and subtypes. Previous studies have demonstrated that

the inhibition of forskolin-stimulated adenylate cyclase (FSC) is

considered to be a functional correlate of the 5-HTIA receptor. The

focus of this dissertation has been to assess the feasibility of using

the FSC assay to determine activity of novel compounds at the 5-HTIA

receptor. It is hoped that the results of this research will aid in the

future design of compounds selective for 5-HT1A receptors.

This FSC assay system was extensively characterized in Chapter 2,

both using reference 5-HT 1A agonists (namely 5-HT, 8-0H-DPAT and buspirone), and a variety of structurally diverse compounds possessing

affinity for the 5-HT1A binding site. The assessment of putative

5-HT 1A intrinsic activity of the series of novel aryltryptamines in the FSC assay extends previous work by Taylor (1985), which investigated

the effect of structure on binding affinity.

All compounds tested in the FSC assay that appeared to possess some

5-HTIA agonistic activity also had EC 50 values in the FSC assay that were greater than their respective 5-HT1A binding affinity Ki values.

Some of these compounds had EC50/Ki ratios that were consistent with the 135

5-HTIA reference agonists (i.e., between 5 and 15), although this trend

did not hold for all compounds. The fact that some compounds had

EC50/Ki ratios that were not consistent suggests that the direct

comparison of cyclase potencies with binding affinities is not

necessarily the best method of confirming 5-HTIA interaction ill this

functional system.

The use of pindolol. currently one of the few relatively selective

5-HTIA antagonists available. provided a better method of confirming

5-HT 1A interaction. and thus provided a qualitative assessment of 5-HT1A activity based on the magnitude of the resultant shift in the Inhibition

curve. In order to detect a shift in the presence of 10 pM pindolol for

compounds producing multiphasic inhibition curves. it appeared that at

least 45% of the effect must be 5-HT 1A related (as determined by two site analysis).

In contrast to the cyclase data for putative 5-HT1A agonists. the

pA2 values of known 5-HTIA antagonists (namely spiperone and pindolul)

in the FSC assay were not significantly differe~t from their 5-HT 1A binding affinities. This supports the theory that the discrepancy

between cyclase potency and binding affinity is likely associated with

G proteins. which do not affect antagonist interactions.

Faced with the current lack of selective 5-HT1A antagonists. pindolol was nonetheless an acceptable choice of 5-HTIA antagonist to demonstrate 5-HTIA interaction in the ability of a compound to inhibit

FSC. The ability to demonstrate that the observed inhibition of FSC 136

possessed a pindolol-sensitive component was crucial in the assignment of 5-HT1A agonistic activity to a compound that inhibited FSC. The importance of using a 5-HTIA antagonist to verify 5-HTIA interaction was reinforced with some of the compounds tested in Chapter 2. For example,

5MEOAHR was not considered to be a 5-HTIA agonist. despite its ability to inhibit FSC. based on the inability of 10 pM pindolol to produce a significant shift in the inhibition curve. Although spiroxatrine also produced full inhibition of FSC (in fuct to an extent greater than 5-HT itself). only a small proportion of its inhibition curve was shifted in the presence of pindolol.

The possibility exists such that the detection of a significant shift in the inhibition curve in the presence of a fixed concentration of pindolol requires a certain proportion of 5-HTIA receptor interaction. Anything less would not be detected. alld would therefore be below the limits of detection of this system. For example. based on two site analysis. (±)spiroxatrine was calculated to possess about 47% high affinity sites. and pindolol-induced blockade of such a proportion of 5-HTIA receptors was detectable as a significant shift in the inhibition curve. On the other hand, two site analysis of the 5MEOAHR­ induced inhibition curve calculated the proportion of high affillity sites to be 40%, and no significant shift in the inhibition curve in the presence of pindolol was detected for this compound. Further analysis of 5MEOAHR is required to determine whether it is a 5-HT1A antagonist or just a very weak partial agonist. 137

Quantitative determination of potency in the FSG assay is

complicated for compounds that produce complex inhibition curves, that

Inhibit to an extent greater than 5-HT itself, and/or do not reach a plateau within the concentration range tested. For example, despite the fact that one of the novel compounds tested, BOT, produced an inhibition curve that was shifted in the presence of pindolol, the ratIo of EG50 to

Ki was still not consistent with the majority of other compounds tested.

In addition, some compounds produced complex inhibition curves that appeared to be multicomponent, and for some compounds, non-5-HT 1A - mediated inhibItIon occurred at concentrations above 10 ~M. In such cases, cyclase potency was difficult to quantitatively evaluate using nonlinear regression analysis. As well, the smaller the 5-HT1A component in the ensuing inhibition curve, or the weaker the overall inhibition of FSC, the greater the difficulty in assigning an EC50 value for potency. Consequently, the quantitative assessment of intrinsic activity was not necessarily clear-cut for some of the compounds tested in the FSC assay. Despite this limitation, the qualitative assessment of 5-HT 1A activity, using pindolol, could be applied to all compounds tested in the PSG assay.

Generally, the data obtained from the FSC assay must be interpreted cautiously. The qualitative assessment of 5-HT 1A activity involves using the 5-HT1A antagonist pindolol to block 5-HT 1A receptors and can be widely applied to all compounds tested. The observed inhibition curve will be shifted to the right in the presence of pindolol, the degree of shift reflecting the relative amount of 5-HTIA interaction 138

involved in the observed inhibition. If no shift occurs. the compound is producing the inhibition via non-5-HTIA mechanism(s). If the compound tested does not appear to inhibit FSC. its ability to reverse inhibition induced by a fixed concentration of 5-HT can thell be evaluated.

The quantitative assessment of potency at the 5-HTIA receptor in the FSC assay is far from a straight-forward exercise. Some compounds produce complex inhibition curves that do not reach a plateau by the highest concentration tested. The ability of some compounds to inhibit

FSC via non-5-HT1A mechanisms further complicates the issue of data analysis. In addition. compounds that only weakly inhibit FSC (i.e .. % inhibition of 5-HT sensitive adenylate cyclase < 45%) can be difficult to quantify with respect to 5-HT1A intrinsic activity. Therefore. it is important to use a 5-HTIA antagonist. such as pindolol. to qualitatively assess intrinsic activity.

Despite the above limitations associated with data analysis. especially quantitatively. the FSC assay can be used to assess SAR for compounds possessing 5-HT1A affinity. as long as the data are treated cautiously. Once the feasibility of using this system was established.

SAR associated with several series of compounds was investigated. One proposed 5-HT IA pharmacophore (Hibert et al .• 1988a. 1989) resulted in the design of compounds. such as 2-AMBD and MDL 73005EF. which did not possess the predicted 5-HTIA affinity and intrinsic activity

(respectively). Despite reports claiming its predominant 5-HTIA 139

antagonistic activity. MOL 73005EF was shown to be an efficacious partial 5-HTIA agonist in the FSC assay (Chapter 3). Since the structure of MOL 73005EF was initially designed using a proposed 5-HT1A antagonist pharmacophore (Hibert et al .• 1988a. 1988b). it is apparent that much remains to be learned about the SAR associated with 5-HTIA intrinsic activity, and emphasizes the need for a functional test tllal can be used to assess structural features of compounds that determine

5-HTIA intrinsic activity.

Clearly. a large series of compounds must be evaluated in order to obtain extensive information concerning overall trends. The series of enantiomers of aminotetralin analogs (Chapter 4) exhibited the trend such that the compound with the highest 5-HT 1A affinity in each enantiomeric pair also possessed relatively greater 5-HT1A agonist activity in the FSC assay. The exceptions to this observation were

8-0H-DPAT and ALK-3. 8-0H-DPAT displayed stereoselective differences with respect to 5-HT1A intrinsic activity. but not binding affinity. while in contrast, there was no measurable difference in the level of intrinsic activity for (+) and (-)ALK-3, despite strong enantlomeric differences with respect to binding affinity.

In addition, the enantiomers tested in Chapter 4 provided supporting data for the newly proposed 5-HT 1A agonist pharmacophore model (Mellin. 1990; Mellin et al .• in preparation). In vivo data for these compounds included assessment of activity at the somatodendritic

5-HT 1A autoreceptor. which has been reported to possess a large receptor 140

reserve. Consequently, even weak partial agonists would appear to be agonists. The PSC assay demonstrates activity at postsynaptic 5-HT1A receptors in the hippocampus, which are proposed to possess little or no receptor reserve. Consequently,the F~C assay is a powerful tool for detecting partial agonists that cannot be detected in systems where there is a large 5-HT1A receptor reserve. In this way, the PSC assay provides a unique ability to distinquish varying levels of agonism at the postsynaptic 5-HT1A receptor.

The series of 3- and 4-tetrahydropyridyl indoles (THPI) analogs also exhibited a similar trend, such that for each group (i.e.,

3-methyl-, 3-propy-, 4-methyl- and 4-propyl-THPI analogs), the higher the 5-HT1A binding affinity, the greater the ability to inhibit FSC (Chapter 5). Overall, no trend existed when all the compounds were considered together. The 5-carboxamido-substituted-THPI analogs were the major exceptions to this general observation and appeared to be a special case.

The compounds examined in Chapters 4 and 5 were analogs of 5-HTIA agonists, specifically 8-0H-DPAT and 5-HT, respectively. These compounds are all relatively small molecules whose structures were based on compounds that bind to the active site of the 5-HT 1A receptor. It is interesting to note that some of the compounds tested did not appear to significantly inhibit FSC, but showed promise as putative 5-HTIA antagonists. In the past, the development of antagonists has mainly consisted of relatively large molecules that cover binding sites with 141

"accessory receptor areas" (Ariens, 1971). However. the SAR from these

two series of compounds. namely the aminotetralins (Chapter 4) and the

THPI analogs (Chapter 5) suggests that the use of small compounds, whose

structures are based on high affinity 5-HTIA agonists. can possibly lead

to the development of selective 5-HTIA antagonists (i.e., (-")SE-61.

(-lCM-12, (+)LEA-146). To broadly generalize, it appears that for many

of these novel compounds, the higher the 5-HT1A binding affinity, the greater the level of 5-HTIA intrinsic activity. There are exceptions to

this general trend. of course, and testing of a more extensive series of

compounds will help to shed light into possible reasons as to why they

do not fit the general trend.

6.2 Future Directions

The use of a cell line in which only the 5-HT 1A receptor is negatively coupled to adenylate cyclase would be a preferable system In which to assess the SAR of the 5-HTIA receptor negatively coupled to adenylate cyclase compared to the cell-free brain homogenates used in

this dissertation. The major advantage would be to eliminate

interfering non-5-HT1A cyclase-linked systems, and therefore have a higher proportion of 5-HT sensitive adenylate cyclase. This type of system could also provide information concerning partial agonists at

this cyclase-linked receptor and spare receptors, since the activity of compounds in the FSC assay could be assessed in cell lines seeded with different concentrations of 5-HTIA receptor. As discussed previously in

Chapter 2, the presence of receptor reserve, not normally present 142

postsynaptically, would allow partial agonists to exhibit greater agonistic activity.

The development of selective 5-HT 1A antagonists is likely to occur in the near future. The qualitative assessment of 5-HTIA intrinsic activity in the FSC assay, currently appears to require illteraction with a proportion of 5-HT1A receptors of at least 47%, when using 10 ~M pindolol. Since the qualitative assessment of 5-HTIA intrinsic activity using the PSC assay relies heavily upon the 5-HTIA antagonist used, it would be interesting to determine whether any newly designed purported

5-HT1A antagonists could improve on the current sensitivity of this FSC assay system.

The interpretation of cAMP-response data can also be complicated, however, by the possibility that receptors may activate multiple independent effector pathways. Such "cross-talk" could result in interaction with other signal transduction mechanisms in addition to the system directly coupled to the receptor in question.

Finally, it would be very interesting to perform computer graphics modeling on the THPI analogs from Chapter 5, and compare their 5-HT 1A activities to the 5-HT 1A pharmacophore proposed by Mellin and colleagues. 143

APPENDIX A

COLUMN PREPARATION, MAINTENANCE AND ELUTION PROFILE DETERMINATION

A.I Column Preparation

A.1.1 Materials

Dowex (AG 50W-X4, 200-400 mesh) and neutral alumina (AG7, 100-200 mesh) were purchased from BioRad. Column materials were ohtained from Sarstedt.

Columns - Cat# 91.787 (disposable tips for macro/micro pipettes) Funnels - Cat# 91.1465 Caps - Cat# 65.784 Tips - Cat# 65.781

Racks were constructed of Plexiglas and each held 48 columns, under which could be placed either a collecting tray or a Plexiglas rack holding 48 scintillation vials.

A.I.2 Procedure for pouring Dowex columns

Dowex must first be activated prior to use, according to the following scheme:

For 500 g Dowex, prepare 2 litres of 3N HCI (538 ml concentrated HCI / 2 litres water), and 2 litres of 3N NaOH (240g NaOH / 2 litres water).

1. Add 1 litre of 3N NaOH to a beaker containing Dowex. Let stir for 30 minutes. Let settle and decant. 2. Add 1 litre of water and stir 10 minutes. Let settle and decant. 3. Repeat Step 2 until the pH of the rinsing water is neutral. 4. Add 1 litre of 3N HCI to the Dowex and stir for 30 minutes. Let settle and decant. 5. Add 1 litre of water and stir for 10 minutes. Let settle and decant. 6. Repeat Step 5 until pH of rinsing water is neutral. 7. Repeat Steps 1 through 6. 8. Pour columns. Store remaining Dowex in IN HCI at 4°C to prevent degradation. 144

To pour columns, pipette 3 ml of Dowex (50:50 mixture with water, neutral pH) into glass wool stoppered columns. Wash with 2 ml of IN HCI. Store in IN HCI and cap columns until used.

A.l.3 Procedure for Pouring Alumina Columns

Fill glass wool stoppered columns with water and immediately add 0.5-0.6 g alumina. Once the alumina settles, wash the columns with 12- 15 ml of 1.0 M imidazole. Then wash the columns with 15--20 ml 0.1 M imidazole (pH 7.3). Store columns in 0.1 M imidazole, and cap tightly when not in use.

A.2 Column Recycling

A.2.1 Dowex Columns

During the FSC assay, the Dowex columns must be rinsed with 20-30 ml water prior to loading on samples. After running samples througll, the Dowex columns must be rinsed with 10 ml of IN HCI and allowed to drain completely. To store columns until their next use, place a small volume of IN HCI on each column and cap it as the effluent emerges. Be sure that a small aliquot of HCI covers the resin. Columns will have drastically reduced recoveries if allowed to run dry overnight.

A.2.2 Alumina Columns

For the FSC assay, alumina columns must be prerinsed with about 15 ml 0.1 M imidazole. After use in the FSC assay, each column must be washed with about 10 ml of 0.1 M imidazole (pH 7.3), and allowed to draIn completely prior. To store columns until next use, place a small volume of 0.1 M imidazole on each column and cap off column as the effluent emerges. Be sure to leave an aliquot of imidazole covering the resin. Columns will have drastically reduced recoveries if allowed to run dry for several hours. 145

A.3 Determination of Column Elution Profiles

For elution profile determination, prepare samples as usual according to the FSC protocol, with the following dIfferences: Suhstitute water for the drug and use tIssue homogenizing buffer alone (without tissue). Also add a standard concentration of [ 32 pJATP (approx. 10,000 dpm/ 50 ~l) and [3 H]cAMP (approx. 30,000 dpm/ 50 ~l) to each sample.

1.4.1 For Dowex columns:

Load above samples onto a few selected Dowex columns (at least three). Collect fractions as follows:

1. sample 2. 0.5 ml water 3. 0.5 ml water 4. 0.5 ml water 5. 0.5 ml water B. 0.5 ml wuter 7. 0.5 ml water 8. 0.5 ml water 9. 0.5 ml water 10. 0.5 ml water 11. 0.5 ml water 12. 0.5 ml water 13. 0.5 ml water 14. 0.5 ml water 15. 0.5 ml water 16. 0.5 ml water 17. 0.5 ml water 18. 0.5 ml water 19. 1.0 ml water 20. 3.0 ml water

Count fractions and determine profile for Dowex. A typical elution profile for a Dowex column is shown in Figure 33. 146

DOWEX COLUMN 7500 _[3H]cAMP ~[32p]ATP

5000 oi. 2 Cl. o 2500 ..

o • I I L. - l 111ft J o 5 10 15 20 FRACTION NUMBER

Figure 33: Typical elution profile for a Dowex column under described conditions. Total recovery: [3HlcAMP = 98%, [32P1ATP = 99%. 147

1.4.2 For alumina columns:

Prepare new samples for Dowex. Load samples on Dowex and elute according to the profile (ie - 2 X 2 ml aliquots of water). Elute directly onto alumina with water and collect fractions as follows:

1. Sample eluted from Oowex. 2. 0.5 ml 0.1 M imidazole 3. 0.5 ml 0.1 M imidazole 4. 0.5 ml 0.1 M imidazole 5. 0.5 ml 0.1 M imidazole 6. 0.5 ml 0.1 M imidazole 7. 0.5 ml 0.1 M imidazole B. 0.5 ml 0.1 M imidazole 9. 0.5 ml 0.1 M imidazole 10. 0.5 ml 0.1 M imidazole 11. 0.5 ml 0.1 M imidazole 12. 0.5 ml 0.1 M imidazole 13. 0.5 ml 0.1 M imidazole 14. 1.0 ml 0.1 M imidazole 15. 3.0 ml 0.1 M imidazole

Count fractions and determine elution profile. Since cAMP is univalent at a of pH 7.3, it elutes from alumina almost Immediately, while the multivalent ATP adsorbs more firmly to the alumina (White and Zenser, 1971). This second chromatographic step further purifies the cAMP. A typical elution profile for a alumina column is shown In Figure 34. 148

ALUMINA COLUMN 1E4 _[3H]cAMP ~[32p]ATP 8000 ~ 0.... o 4000

o I ~ -I. o 5 10 15 FRACTION NUMBER

Figure 34: Typical elution profile for an alumina column under described conditions. Total recovery: [3H]cAMP = 90%. [32pJATP ~ 5%. 149

APPENDIX B

PROTOCOL FOR THE 5-HT1A BINDING ASSAY

This section was taken from Tayloe et al. (1986).

B.1 Tissue Preparation

Male Sprague-Dawley rats (150-250g) were decapitated, and the brains were rapidly removed and chilled in ice-cold 0.9% saline and tliA cortices were dissected out. The cortices were then homogenized in 40 volumes of ice-cold Tris-Hel (50 mM, pH 7.4) buffer, using a polytron (Brinkmann Instruments Inc., Westbury, NY) on setting 5 for 15 seconds. The homogenate was centrifuged at 48,000 x g for 10 minutes. The resultant pellet was then resuspended in fresh Tris-HCI buffer, and the centrifugation and resuspension process was repeated three additional times to wash the membranes. Between the second and third washes, the resuspended membranes were incubated for 10 minutes at 370 C to facilitate the removal of endogenous 5-HT (Nelson et al., 1978). The final pellet was resuspended in the Tris-HCI buffer to a final concentration of "10 mg tissue (original wet weight) / ml" for use in the binding assay.

B.2 [3H]8-0H-DPAT Binding Assay

To each assay tube, the following were added: 0.1 ml drug dilution (or water if no competing drug was added~, 0.9 ml of binding buffer (containing Tris, CaCl2, pargyline and [ H]8-0H-DPAT to achieve final assay concentrations of 50 mM, 3 mM, 100 ~M and 1.0 nM. respectively. pH 7.4) and 1.0 ml of resuspended membranes. The tubes were incubated for 15 minutes at 37 0 C, and the incubations were terminated by vacuum filtration (Brandel Cell Harvester, Gaithersburg, MD) through Whatman GF/B filters (pretreated by soaking for 2 hours in a 0.1% (v/v) solution of polyethylenimine and then dried), followed by two 4-ml rinses with ice-cold 50 mM phosphate buffer. The filters were then dried and the bound radioactivity was determined by liquid scintillation spectrometry. Specific [3H]8-0H-DPAT binding was defined as the difference between binding inthe presence and absence of 10 ~M 5-HT. The free ligand concentration was determined by directly counting 50 ~l aliquots of the binding buffer which contained (3H]8-0H-DPAT. 150

APPENDIX C

EFFECT OF ASSAY CONDITIONS ON 5-HT POTENCY IN THE FSC ASSAY

For all 5-HTIA agonists tested in the FSC assay, the EC50 values from cyclase data were greater than the corresponding Ki values determined from r3H]8-0H-DPAT binding data. In order to further investigate this phenomenon, several factors were investigated.

Also, other investigators use slightly different assay conditIons. To further investigate the fact that the cyclase values were less potellt than the 5-HT1A binding affinities, we looked at two possibilities: 1. The actual 5-HT concentration was less than the theoretical concentration, such that the 5-HT added was undergoing breakdown or reuptake. To pursue these possiblities, the ability of 5-HT to inhibit FSC was determined under the following conditions:

a) normal conditions b) 10 pM paroxetine (a 5-HT reuptake inhibitor) c) 0.6 mM ascorbic acid (to protect 5-HT from oxidation) d) 10 11M pargyline (an inhibi tor of monoamine oxidase (MAO), which is an endogenous enzyme that degrades 5-HT) i) added to the assay components ii) added diredctly to the tissue homogenate

The results from this small study are out! ined in Table 9. All of the treatments resulted in EC50 values that were greater than the Ki of 5-HT for 5-HT1A receptors. Since none of these treatments yielded EC50 values that were not significantly different from the 5-HTIA binding affinity values, this avenue was not pursued any further.

2. The discrepancy could also have been due to an artifact resulting from tissue concentration effects. To test this possibility, the ability of 5-HT to inhibit FSC was tested with a range of tissue concentrations. However, despite varying the tissue concentration, the EC50 for 5-HT remained unchanged (Table 10).

In conclusion, the discrepancy between binding Ki and cyclase EC50 values did not appear to be attributed to breal

Table 9: COMPARISON OF EC50 VALUES FOR 5-HT UNDER DIFFERING ASSAY CONDITIONS.

TREATMENT EC50 (nM)

normal 160 paroxetlne 184 ascorbic acid 126 pargyline i) assay 160 i i ) ti::lsue 168

Table 10: COMPARISON OF EC 50 VALUES FOR 5-HT IN THE FSC ASSAY USING DIFFERING CONCENTRATIONS OF RAT HIPPOCAMPUS.

TISSUE EC50 CONCENTRATION (nM) (J.lg per sample)

225 196 170 168 112 168 55 133 23 168 152

REFERENCES

Amin A. H., Crawford T. B. B., and Ceadlum J. n. (1954) The distribution of substance P and 5-hydroxytryptamine in the central nervous system of the dog. J. Physiol. (London) 126, 596-618.

Andrade R., and Nicoll R. A. (1987a) Novel anxiolytics discriminate between postsynaptic serotonin receptors mediating different pllysiolo~ical responses on single neurons of the rat hippocampus. Naunyn-Schmiedeberg's Arch. Pharmacol. 336, 5-10.

Andrade R., and Nicoll R. A. (1987b) Pharmacologically distinct actions of serotonin on single pyramidal neurones of the rat hippocampus recorded in vitro. J. Physiol. 394, 99-124.

Ariens E. J. (1971) Chapter 1: A general introduction to the field of drug design, in Drug Design, (Ariens E. J., ed), pp. 172-193. Academic Press, Inc., New York.

Arunlaskshana 0., and Schild H. O. (1959) Some quantitative uses of drug antagonists. Br. J. Pharmacal. 14, 48-58.

Arvidsson L., Karlen A., Norinder U., Kenne L., Sundell S., and Hacksell U. (1988) Structural factors of importance for 5-hydroxytrypt­ aminergic activity. Conformational preferences and electrostatic potentials of 8-hydroxy-2-(di-n-propylamino)tetralin (8-0H-DPAT) and some related agents. J. Med. Chern. 31, 212-221.

Arvidsson L., Johansson A. M., Hacksell U., Nilsson J. L. G., Svensson K., Hjorth S., Magnusson T., Carlsson A., Andersson B., and Wikstrom H. (1987) (+)-Cis-8-hydroxy-1-methyl-2-(di-n-propylamino)tetralin: a potent and highly stereoselective 5-hydroxytryptamine receptor agonist. J. Med. Chern. 3D, 2105-2109.

Arvidsson L.-E., Hacksell U., and Glennon R. A. (1986) Recent advances in central 5-hydroxytryptamine receptor agonists and antagonists. Prog. Drug Research 3D, 365-471.

Arvidsson L.-E., Hacksell U., Johansson A. M., Nilsson J. L. G., Lindberg P., Sanchez D., Wikstrom H., Svensson K., Hjorth S., and Carlsson A. (1984) 8-nydroxy-2-(alkylamino)tetralins and related compounds as central 5-hydroxytryptamine receptor agonists. J. Med. Chern. 27, 45-51.

Arvidsson L., Hacksell U., Nilsson J. L. G., Hjorth S., Carlsson A., Lindberg P., Sanchez D., and Wikstrom H. (1981) 8-Hydroxy-2-(di-n­ propylamino)tetralin, a new centrally acting 5-hydroxytryptamine receptor agonist. J. Med. Chern. 24, 921-923. 153

Barbaccia M. L., Brunella N., Chuang D. M., and Costa E. (1983) Serotonin-elicited amplification of adenylate cyclase activity in hippocampal membranes from adult rat . .1. Neurochem. 40, 1671-1679.

Bennett .1. P., and Snyder S. H. (1976) Serotonin and lysergic acid diethylamide binding to rat brain membranes: relationship to postsynaptic serotonin receptors. Mol. Pllarmacol. 12, 373-389.

Berridge M. J., and Heslop .1. P. (1981) Separate 5-hydroxytryptamine receptors on the salivary gland of the blowfly are linked to the generation of either cyclic adenosine 3' ,5'-monophosphate or calcium signals. Br. J. Pharmacol. 73, 729-738.

Berry-Kravis E., and Dawson G. (1983) Characterization of an adenylatp cyclase-linked serotonin (5-HT1) receptor in a neuroblastoma x brain explant hybrid cell line (NCB-20). J. Neurochem. 40, 977-985.

Bjork L., Hook B. B., Nelson D. L., Anden N., and Hacksell U. (1989) Resolved N,N-dialkylated 2-amino-8-hydroxytetralins: stereoselective interactions with 5-HTIA receptors in the brain. J. Med. Chern. 32, 779-783.

Blazynski C., Ferrendelli J. A., and Cohen A. I. (1985) Indoleamine­ sensitive adenylate cyclase in rabbit retina: characterization and distribution. J. Neurochem. 45, 440-447.

Bockaert J., Sebben M., and Dumuis A. (1990) Pharmacological characterization of 5-hydroxytryptamine4 (5-HT4 ) receptors positively coupled to adenylate cyclase in adult guinea pig hippocampal membranes: effect of substituted benzamide derivatives. Mol. Pharmacol. 37, 408-411.

Bockaert J., Dumuis A., Bouhelal R., Sebben M., and Cory R. N. (1987) Piperazine derivatives including the putative anxioytic drugs, buspirone and ipsapirone, are agonists at 5-HT 1A receptors negatively coupled with adenylate cyclase in hippocampal neurons. Naunyn Schmiedeberg's Arch. Pharmacal. 335, 588-592.

Bockaert J. (1981) General characteristics, localization and adaptative responsiveness of neurotransmitter-sensitive adenylate cyclases in the central ner.vous system. Adv. Cyclic Nucl. Res. 14, 611-628.

Bockaert J., Nelson D. L., Herbet A., Adrien J., Enjalbert A., and Hamon M. (1981) Serotonin-receptors coupled with an adenyl ate cyclase in the rat brain: Non-identity with 3H- 5- HT binding sites, in Serotonin - Current aspects of neurochemistry and function (Haber B., Gabay S., Issidorides M.R., Alvisatos S. G. A., eds.), pp 327-345. Plenum Press, New York. 154

Bouhelal R., Smounya L., and Bockaert J. (1988) 5-HTIB receptors are negatively coupled with adenylate cyclase in rat sUbstantia nigra. Eur. J. Pharmacal. 151, 189-196.

Cedar H., and Schwartz J. H. (1972) Cyclic adenosine monophosphate in the nervous system of Aplvsia californica. II. Effect of serotonin and dopamine. J. Gen. Physiol. 60, 570-587.

Cheng Y., and Prusoff W. H. (1973) RelationshIp betw~en the inhibition constant (Ki) and the concentration which causes 50 percent inhibition (ICsol of an enzymatic reaction. Biochem. Pharmacol. 22, 3099-3108.

Chou W, S., Ho A. K. S., and Loh H. H. (1971) Neurohormones on brain adenyl cyclase activity in vivo. Nature New BioI. 233. 280-281.

Claustre Y.. Benavides J .• and Scatton B. (1988) 5-HTIA receptor agonists inhibit carbachol-induced stimulation of phosphoinositide turnover in the rat hippocampus, Eur. J. Pharmacal. 149, 149-153.

Colina A.. and Halliwell J. V. (1987) Differential modulation of three separate K-conductances In hippocampal CAl neurons by serotonin. Nature 328. 73-77.

Cooper S. J" and Abbott A. (1988) Clinical psychopharmacology may benefit from new advances III 5-HT pharmacology. Trends in Pharmacnl. Sci. 9, 269-271.

Cornfield L. J .. and Nelson D. L. Biochemistry of 5-hydroxytryptamine receptor subtypes: coupling to second messenger systems. in Serotonin Receptor Subtypes: Basic and Clinical Aspects (Peroutka. S. J., ed). Alan R. Liss, Inc. New York. in press.

CornfIeld L. J., Nelson D. L .. Taylor E. W., and Martin A. R. (1989) MOL 73005EF: partial agonist at the 5-HT 1A receptor negatively coupled to adenylate cyclase. Eur. J. Pharmacal. 173. 189-192.

~ornfjeld L. J., Nelson 0: L., Monroe P. J., Taylor E. W., and Nikam S. S. (1988) Use of forskolin-stimulated adenylate cyclase in rat hippocampus as a screen for compounds that act through 5-HT 1A receptors. Proc. West. Pharmacol. 31, 265-267.

Cossery J. M., Gozlan H., Spampinato U., Peridicakis C., Guillaumet G.. Pichat L., and Hamon M. (1987) The selective labelling of central 5-HT1A receptor binding sites by [3H]5-hydroxy-3(di-n-propylamino)­ chroman. Eur. J. Pharmacol. 140, 143-155.

Craig D. A., and Clarke D. E. (1990) Pharmacological characterization of a neuronal receptor for 5-hydroxytryptamine In guinea pig ileum with properties similar to the 5-hydroxytryptamine4 receptor. J. Pharmacal. Exp. Ther. 252, 1378-1386. 155

Daszuta A., F. P., and Cadilhac J. (1979) Effect of serotonin on cyclic AMP level in rat hypothalamus slices during development. Eur. J. Pharmacal. 56, 397-401.

DeLean A., Munson P. J., and Rodbard D. (1978) Simultaneous analysis of fami! les of sigmoidal curves: appl ication to bioassay, radiol igand assay, and physiological dose-response curves. Am . .1. Physiol. 235, E97-EI02.

DeVivo M., and Maayani S. (1988) 5-HT receptors coupled to adenylate cyclase, in The Serotonin Receptors (Sanders-Bush E., ed), pp. 141- 179. The Humana Press, Clifton, N.J.

DeVivo M., and Maayani S. (1986) Characterization of the 5-hydroxy­ tryptaminelA receptor-mediated inhibition of forskolin-stimulated adenyl ate cyclase activity in guinea pig and rat hippocampal membranes. J. Pharmacal. EXp. Ther. 238, 248-253.

DeVivo M., and Maayani S. (1985) Inhibition of forskolin-stimulated adenylate cyclase activity by 5-HT receptor agonists. Eur. J. Pharmacal. 119, 231-234.

Downer R. G. H., Gale J. W., and Orr G. L. (1985) Interaction of formamidines with octopamine-, dopamine- and 5-hydroxytryptamine­ sensitive adenylate cyclase in the nerve cord of Periplaneta americana. Pestic. Sci. 16, 472-478.

Dumuis A., M. S., and Bockaert J. (1989) BRL 24924: a potent agonist at a non-classical 5-HT receptor positively coupled with adenylate cyclase in colliculi neurons. Eur. J. Pharmacol. 162, 381-384.

Dumuis A., Sebben M., and Bockaert J. (1988a) Pharmacology of 5-hydroxytryptaminelA receptors which inhibit cAMP production in hippocampal and cortical neurons in primary "'11 ture. Mol. Pharmacol. 33, 178-186.

Dumuis A., Bouhelal R., Sebben M., Cory R., and Bockaert J. (1988b) A nonclassical 5-hydroxytryptamine receptor positively coupled to adenylate cyclase in the central nervous system. Mol. Pharmacal. 34, 880-887.

Dumuis A., Bouhelal R., Sebben M., and Bockaert J. (1988c) A 5-HT receptor in the central nervous system positively coupled with adenylate cyclase is antagonized by ICS 205 930. Eur. J. Pharmacal. 146, 187-188.

Engel G., Gothert M., Hoyer D., Schlicker E., and Hillenbrand K. (1986) Identity of inhibitory presynaptic 5-hydroxytryptamine (5-HT) autoreceptors in the rat brain cortex with 5-HT 1B binding sites. Naunyn Schmiedeberg's Arch. Pharmacol. 332, 1-7. 156

Enjalbert A., Bourgoin S., Hamon M.. Adrien J., and Bockaert J. (1978a) Postsynaptic serotonin-sensitive adenyl ate cyclase in the central nervous system. I Development and distribution of serotonin and dopamine-sensitive adenylate cyclases in rat and guinea pig brain. Mol. Pharmacal. 14, 2-10.

Enjalbert A., Hamon M., Bourgoin S., and Oockaert J. (1978b) Postsynaptic serotonin-sensitive adenylate cyclase in the central nervous system. II Comparison with dopamine- and isoproterenol­ sensitive adenyl ate cyclase in rat brain. Mol. Pharmacol. 14. 11-23.

Erpsamer V., and Asero B. (1952) Identification of enteramin~, the specific hormone of the enterochromaffin cell system. as 5-hydroxy­ tryptamine. Nature 169, 800-801.

Fargin A., Raymond J. R., Regan J. W., Cotecchia S., Lefkowitz R. J., and Caron M. G. (1989) Effector coupling mechanisms of the cloned 5-HT1A receptor. J. BioI. Chem. 264, 14848-14852.

Fargin A., Raymond J. R., Lohse M. J., Kobilka O. K., Caron M. G., and Lefkowitz R. J. (1988) The genomic clone G-21 which resembles a p-adrenergic receptor sequence encodes the 5-HT1A receptor. Nature 335, 358-360.

File S. E. (1987) The search for novel anxiolytics. Trends in Neurol. Sci. 10, 461-463.

Fillion G., Beaudoin D., Rousselle J. C., and Jacob J. (1980) [3H]5-HT binding sites and 5-HT-sensitive adenyl ate cyclase in glial cell membrane fraction. Brain Res. 198, 361-374.

Fillion G., Rousselle J. C., Beaudoin D.. Pradelles P., Golny M.. Dray F., and Jacob J. (1979) Serotonin sensitive adenylate cyclase in horse brain synaptosomal membranes. Life Sci. 24, 1813-1822.

Frazer A., Maayani S., and Wolfe B. B. (1990) Subtypes of receptors for serotonin. Ann. Rev. Pharmacol. Toxicol. 30, 307-348.

Gaddum J. H., and Picarelli Z. P. (1957) Two kinds of tryptamine receptor. Brit. J. Pharmacol. 12, 323-328.

Glennon R. A., Naiman N., Lyon R. A., and Titeler M. (1988a) Arylpiperazine derivatives as high-affinity 5-HT 1A serotonin ligands. J. Med. Chem. 31, 1968-1971.

Glennon R. A., Titeler M., Lyon R. A., and Slusher R. M. (1988b) N,N'di­ n-propylserotonin: Binding at serotonin binding sites and a comparison with 8-hydroxy-2-(di-n-propylamino)tetralin. J. Med. Chem. 31, 867-870. 157

Glennon R. A. (1987) Central serotonin receptors as targets for drug research. J. Med. Chern. 30, 1-12.

Gozlan H., EI Mestikawy S., Pichar L., Glowbuski J., and Hamon M. (1983) Identification of presynaptic serotonin autoreceptors using a new ligand [3HJ-PAT. Nature 305, 140-142.

Gudelsky G. A., Koenig J. I., and Meltzer H. Y. (1986) Thermoregulatory responses to serotonin (5-HT) receptor stimulation in the rat: evidence for opposing roles of 5-HT2 and 5-HT 1A receptors. Neuropharmacol. 12, 1307-1313.

Harrington M. A., Oksenberg D., and Peroutka S. J. (1988) 5-HydroxytryptaminelA receptors are linked to a Gj-adenylate cyclase complex in rat hippocampus. Eur. J. Pharmacol. 154, 95-98.

Herrick-Davis K., and Titeler M. (1988) Detection and characterization of the serotonin 5-HT 1D receptor in rat and human brain. J. Neurochem. 50, 1624-1631.

Heslop J. P., and Berridge M. J. (1980) Changes in cyclic AMP and cyclic GMP concentrations during the action of 5-hydroxytryptamine on an insect salivary gland. Biochem. J. 192, 247-255.

Hibert M. F., McDermott I., Middlemiss D. N., Mir A. K., and Fozard J. R. (1989) Radioligand binding study of a series of 5-HTIA receptor agonists and definition of a steric model of this site. Eur . .J. Med. Chern. 24, 31-37.

Hibert M. F., Gittos M. W., Middlemiss D. N., Mir A. K., and Fozard J. R. (1988a) Graphics computer-aided receptor mapping as a predictive tool for drug design development of potent, selective, and stereospecific ligands for the 5-HTIA receptor. J. Med. Chern. 31, 1087-1093.

Hibert M., Mir A. K., Maghioros C., Moser P., Middlemiss D. N., Tricklebank M. D., and Fozard J. R. (1988b) The pharmacological properties of MDL 73005EF: a potent and selective ligand at 5-HTIA receptors. Br. J. Pharmacol. 93, 2P.

Hibert M. (1987) Graphics computer-aided receptor mapping as a predictive tool for drug design: application to the serotonin 5-HTIA and 5-HTIB receptors, in Topics in Pharmaceutical Sciences (Breimer D. D., and Speiser P., eds), pp 367-377. Elsevier Science Publishers, The Netherlands.

Hillver S.-E., Bjork L., Li Y.-L., Svensson B., Ross S., Anden N.-E., and Hacksell U. (1990) (S)-5-Fluoro-8-hydroxy-2-(dipropylamino)­ tetralin: a putative 5-HTIA-receptor antagonist. J. Med. Chern. 33, 1541-1544. 158

Hjorth S., and Sharp T. (1990) Mixed agonist/antagonist properties of NAN-190 at 5-HT1A receptors: behavioural and in vivo brain microdialysis studies. Life Sci. 46, 955-963.

Hjorth S., Sharp T., and Hacksell U. (1989) Partial postsynaptic 5-HT 1A Rgonist properties of the novel stereoselective 8-0H-DPAT analogue (t)cis-8-hydroxy-l-methyl-2-(di-n-propylamino)tetralin, (t)ALK-3. Eur. J. Pharmacol. 170, 269-274.

Hotta I., and Yamawaki S. (1986) Lithium decreases 5-HT1 receptors but increases 5-HT-sensitive adenylate cyclase activIty in rat hippocampus. BioI. Psychiatry 21, 1382-1390.

Hoyer D., and Middlemiss D. N. (1989) Species differences in the pharmacology of terminal 5-HT autoreceptors in mammalian brain. Trends in Pharmacal. Sci. 10, 130-132.

Hoyer D., Waeber C., Schoeffter P., Palacios J. M., and Dravid A. (1989) 5-HT1C receptor-mediated stimulation of inositol phosphate production in pig choroid plexus. Naunyn Schmiedeberg'sArch. Pharmacal. 339, 252-258.

Hoyer D. (1988a) Functional correlates of serotonin 5-HT1 recognition sites. J. Receptor Research 8, 59-81.

Hoyer D. (1988b) Molecular pharmacology and biology of 5-HT 1C receptors. Trends in Pharmacol. Sci. 9, 89-94.

Hoyer D., and Schoeffter P. (1988) 5-HT 1D receptor-mediated inhibition of forskolin-stimulated adenylate cyclase activity in calf substantia nigra. Eur. J. Pharmacol. 147, 145-147.

Hoyer D., Engel G., and Kalkman H. O. (1985) Molecular pharmacology of 5-HT1 and 5-HT2 recognition sites in rat and pig brain membranes: radiol J.~and binding studies wi th [3H]5--HT, [3Hl8-0U-DPAT, (_)[l~~I]iodocyanopindolol, [3U]mesulergine and [3 U]ketanserin. Eur. J. Pharmacol. 118, 13-23.

Humphrey P. P. A. (1984) Peripheral 5-hydroxytryptamine receptors and their classification. Neuropharmacol. 23, 1503-1510.

Kakiuchi S., and RaIl T. W. (1968a) Studies on adenosine 3' ,5'-phosphate in rabbit cerebral cortex. Mol. Pharmacol. 4, 379-388.

Kakiuchi S., and RaIl T. W. (1968b) The influence of chemical agents on the accumulation of adenosine 3' ,5'-phosphate in slices of rabbit cerebellum. Mol. Pharmacal. 4, 367-378.

KIlpatrick G. J., Jones B. J., and Tyers M. B. (1987) Identification and distribution of 5-HT3 receptors in rat brain using radioligand binding. Nature 330, 746-748. 159

Kobilka 8. K., Frielle T., Collins S., Yang-Feng T., Kobilka T. S., Francke U., Lefkowitz R. J., and Caron M. G. (1987) An intronless gene encoding a potential member of the family of receptors coupled to guanine nucleotide regulatory proteins. Nature 329, 75-79.

Kodama T., Matsukado Y., and Shimizu H. (1973) The cyclic AMP system of human brain. Brain Res. 50, 135-146.

Laurenza A., Sutkowski E. M.. and Seamon K. B. (1989) Forskolin: a specific stimulator of adenylyl cyclase or a diterpene with multlple sites of action? Trends in Pharmacol. Sci. 10. 442-447.

Leonhardt S., Herrick-Davis K.. and Titeler M. (1989) Detection of a novel serotonin receptor subtype (5-HT1E) in human brain: Interaction with a GTP-binding protein. J. Neurochem. 53, 465-471.

Levitan I. 8., and Benson J. A. (1981) Neuronal oscillators in Aplysia: modulation by serotonin and cyclic AMP. Trends Neurosci. Feb, 38-41.

Leysen J., Tollenaere J. P., Koch M. H. J .• and Laduron P. (1977) Differentiation of opiate and neuroleptic receptor binding in rat brain. Eur. J. Pharmacol. 43, 253-267.

Litosch I., Fradin M., Kasaian M., Lee H. S., and Fain J. N. (1982) Regulation of adenylate cyclase and cyclic AMP phosphodiesterase by 5-hydroxytryptamine and calcium ions in blowfly salivary-gland homogenates. Biochem. J. 204, 153-159.

Liu Y., Mellin C., 8jfirk L., Svensson B., Csoregh T., Helander A., Kenne L., Anden N. -E., and Hacksell U. (1989). (R)- and (S)-5,6,7,8- tetrahydro-1-hydroxy-N,N-dipropyl-9H-benzocyclohepten-8-ylamlne. Stereoselective interactions with 5-HT 1A receptors in the brain. J. Med. Chem. 32, 231i-2318.

Lloyd E. J., and Andrews P. R. (1986) A common structural model for central nervous system drugs and their receptors. J. Med. Chern. 29, 453-462.

Lowry O. H., Rosebrough N. J., Farr A. L., and Randall R. J. (1951) Protein measurement with the Falin phenol reagent. J. 8iol. Chern. 193, 265-275.

Luchins A. R., and Makman M. H. (1980) Presence of histamine and serotonin receptors associated with adenylate cyclase in cultured calf-aorta smooth muscle cells. Biochem. Pharmacal. 29, 3155-3161. 160

Maayani S .. and Sherman M.R. (1990) Adenylate cyclase-linked 5-hydroxy­ tryptamine receptors in the brain. in Serotonin: From cell biology to pharmacology and therapeutics (Paoletti R.. Vanhoutte P. M.. Brunello N.• and Maggi F. M.. eds.). pp 39-51. Kluwer Academic Publishers. Dordrecht. The Netherlands.

MacDermot J. (1979) Independent regulation of receptor-medIated cell depolarization and adenylate cyclase by 5-hydroxytryptamine (sl!rotonin) in neuroblastoma hybrid cells. Biochem. Soc. Trans. 7. 670-671.

MacDermot J .• HIgashida H.• Wilson S. P .. Matsuzawa H.. Minna J .. and Nirenberg M. (1979) Adenylate cyclase and acetylcholine release regulated by separate serotonin receptors of somatic cell hybrids. Proc. Natl. Acad. Sci. USA 76. 1135-1139.

Mallsour J. M.. and Mansour T. E. (1986) GTP-bindlng proteins associated with serotonin-actIvated adenylate cyclase in Fasciola hepatica. Mol. BIochem. Parasitol. 21. 139-149.

Mansour J. M.• Ehrlich A.. and Mansour T. E. (1983) The dual effects of aluminum as activator and inhibitor of adenylate cyclase in the liver fluke Fasciola hepatica. Biochem. Biophys. Res. Comm. 112. 911-918.

Mansour T. E., Sutherland E. W., RaIl T. W., and Bueding E. (1960) The effect of serotonin (5-hydroxytryptamine) on the fromation of adenosine 3' ,5'-phosphate by tissue particles from the liver fluke. Fasciola hepatica. J. BioI. Chern. 235, 466-470.

Markstein R.. Hoyer D., and Engel G. (1986) 5-HTIA-receptors mediate stimulatIon of adenylate cyclase in rat hippocampus. Naunyn SchmIedeberg's Arch. Pharmacol. 333, 335-341.

Marshall G. R., and Cramer R. D. (1988) Three-dimensional structure­ activity relationships. Trends in Pharmacol. Sci. 9, 285-289.

Mawe G. M., Branchek T. A., and Gershon M. D. (1989) Blockade of 5-HT­ mediated enteric slow EPSP's by BRL 24924: gastrokinetic effects. Am. J. Physiol. 257, G386-G396.

McNall S. J., and Mansour T. E. (1985) Forskolin activation of serotonin-stimulated adenylate cyclase in the liver fluke Fasciola hepatica. Biochem. Pharmacol. 34, 1683-1688.

McNall S. J., and Mansour T. E. (1984) Desensitization of serotonin­ stimulated adenyl ate cyclase in the liver fluke Fasciola hepatica. Biochem. Pharmacol. 33, 2799-2805. 161

Meller E., Goldstein M., and Bohmaker K. (1990) Receptor reserve for 5-hydroxytryptamineIA-mediated inhibition of serotonin synthesis: possible relationship to anxiolytic properties of 5-hydroxy­ tryptaminelA agonists. Mol. Pharmacal. 37, 231-237.

Mellin C. (1990) 5-HT1A-receptor agonists. Synthesis, conformational analysis and structure-activity relationships, in Comprehensive Summaries of URRsaia Dissertations from the Faculty of Pharmacy 65, pp 1-54. Uppsala.

Mellin C., Bjbrk L., Karlen A., Johansson A. M., Sundell S., Kenne L .. Nelson D. L., Anden N. E., and Hacksell U. (1988) Central dopaminergic and 5-hydroxytryptaminergic effects of C3-methylated derivatives of 8-hydroxy-2-(di-n-propylamino)tetralin. J. Med. Chem. 31, 1130-1140.

Mellin C., Vallgarda J., Nelson D. L., Bjbrk L., Yu H., Anden N.-E., Csdregh r., Arvldsson L.-E., and Hacksell U. A 3-D model for 5-HT 1A -receptor agonists which accamodates stereoselective methyl­ substituted and conformationally restricted analogues af 8-hydraxy- 2-(dipropylamino)tetralin. Submitted to .1. Med. Chem.

Middlemiss D. N., Neill J., and Tricklebank M. D. (1985) Subtypes af the 5-HT receptor involved in hypothermia and forepaw treading induced by 8-0H-DPAT. Br. J. Pharmacal. 85, 251P.

Mir A. K., Hibert M., Tricklebank M. D., Middlemiss D. N., Kidd E. J., and Fozard J. R. (1988) MDL 72832: a potent and stereoselective ligand at central and peripheral 5-HT 1A receptors. Eur. J. Pharmacol. 149, 107-120.

Moser P. C., Tricklebank M. D., Middlemiss D. N., Mir A. K., Hibert M. F., and Fazard J. R. (1990) Characterization af MOL 73005EF as a 5-HTIA selective ligand and its effects in animal models of anxiety: comparison with buspirone, 8-0H-DPAT and diazepam. Br. J. Pharmacal. 99, 343-349.

Moser P., Hibert M., Middlemiss D. N., Mir A. K., Tricklebank M. D., and Fozard J. R. (1988) Effects of MOL 73005EF in animal models predictive of anxiolytic activity. Br. J. Pharmacol. 93, 3P.

Nathanson J. A., and Greengard P. (1973) Octopamine-sensltive adenylate cyclase: Evidence for a biological role of octopamine in nervous tissue. Science 180, 308-310.

Nelson D. L., Monroe P. J., Lambert G., and Yamamura H. I. (1987) [3H]Spiroxatrine labels a serotoninlA-like site in the rat hippocampus. Life Sci. 41, 1567-1576.

Nelson D. L., and Taylor E. W. (1986) Spiroxatrine: a selective serotoninlA receptor antagonist. Eur. J. Pharmacol. 124, 207-208. 162

Nelson D. L., Pedigo N. W., and Yamamura H. I. (1981) Multiple 3H-5-hydroxytryptamine binding sites in rat brain. J. Physiol. Paris 77, 369-372.

Nelson D. L., Herbet A., Enjalbert A.. Bockaert J., and Hamon M. (1980a) Serotonin-sensitive adenylate cyeJ,.:e and (3H]serotonin binding sites in the CNS of the rat-I. Kinetic parameters and pharmacological properties. Biochem. Pharmacol. 29, 2445-2453.

Nelson D. L., Herbet A., Adrien J., Bockaet't H., and M. (1980b) Serotonin-sensitive adenylatp. cyclase and (3H]serotonin binding sites in the CNS of the rat-II. Respective regional and subcellular distributions and ontogenetic developments. Biochem. Pharmacol. 29, 2455-2463.

Nikam S. S., Martin A. R., and Nelson D. L. (1988) Serotonergic properties of spiroxatrine enantiomers. J. Med. Chern. 31, 1965-1968.

Northup J. K., and Mansour T. E. (1978a) Adenylate cyclase from Fasciola hepatica. 1. Ligand specificity of adenylate cyclase-coupled serotonin receptors. Mol. Pharmacol. 14, 804-819.

Northup J. R., and Mansour T. E. (1978b) Adenylate cyclase from Fasciola hepatica 2. Role of guanine nucleotides in coupling adenylate cyclase and serotonin receptors. Mol. Pharmacol. 14, 820-833.

Ocorr R. A., and Byrne J. H. (1986) Evidence for separate receptors that mediate parallel effects of serotonin and small cardiocactive peptide B (SCPB) on adenylate cyclase in Aplysia californiea. Neurosci. Letters 70, 283-288.

Oksenberg D., and Peroutka S. J. (1988) Antagonism of 5-hydroxy­ tryptaminelA (5-HT1A ) receptor-mediated modulation of adenylate cyclase activity by pindolol and propranolol isomers. Biochem. Pharmacol. 37, 3429-3433.

Ott L., Weiner D., Cheng H., and Woodward J. (1981) Estimating pA 2 values for different designs. J. Pharmacol. Methods 5, 75-92.

Pascual J. (1990) Serotonin, serotonin receptors, serotonin receptor subtype agonists and pain. Pain 40, 115-116.

Pedigo N. W., Yamamura H. I., and Nelson D. L. (1981) Discrimination of multiple [3H]5-hydroxytryptamine binding sites by the neuroleptic spiperone in rat brain. J. Neurochem. 36, 220-226.

Peroutka S. J. (1988a) 5-Hydroxytryptamine receptor subtypes: molecular, biochemical and physiological characterization. Trends in Neurol. Sci. 11, 496-500. 163

Peroutka S. J. (1988b) 5-hydroxytryptamine receptor subtypes. Ann. Rev. Neurosci. II, 45-60.

Peroutka S. J., and McCarthy B. G. (1989) Sumatriptan (GR 43175) interacts selectively with 5-HT1B and 5-HT1D binding sites. Eur. J. Pharmacal. 163, 133-136.

Peroutka S. J., and Snyder S. H. (1979) Multiple serotonin receptors differential binding of 3H-serotonin, 3H-Iysergic acid diethylamide & 3H-spiroperidol. Mol. Pharmacol. 16, 687-699.

Ram J. L., Gole D., Shukla U., and Greenberg L. (1983) Serotonin­ activated adenylate cyclase and the possible role of cyclic AMP ill modulation of buccal muscle contraction in Aplysia. J. Neurobiol. 14, 113-121.

Rapport M. M. (1949) Serum vasoconstrictor (serotonin). V. Presence of creatinine in the complex. A proposed structure of the vasoconstrictor principle. J. BioI. Chern. 180, 961-969.

Rapport M. M., Green A. A., and Page I. H. (1947) Purification of the substance that is responsible for vasoconstrictor activity of serum. Federation Proc. 6, 184.

Raymond J. R., Fargin A., Middleton J. P., Graff J. M., Haupt D. M.. Caron M. G., Lefkowitz R. J., and Dennis V. W. (1989) The human 5-HTIA receptor expressed in HeLa cells stimulates sodium--dependent phosphate uptake via protein kin~se C. J. BioI. Chern. 264, 21943- 21950.

Ribeiro P., and Webb R. A. (1987) Characterization of a serotonin transporter and an adenylate cyclase-linked serotonin receptor in the cestode Hymenolepis diminuta. Life Sci. 40, 755-768.

Robertson H. A., and Osborne N. N. (1979) Putative neurotransmitters in the annelid central nervous system: presence of 5-hydroxytryptamine and octopamine-stimulated adenylate cyclases. Compo Biochem. Physiol. 64C, 7-14.

Saitoh T., and Shih J. C. (1987) Photoaffinity labeling of adenylate cyclase-linked serotonin receptors in Aplysia neurons. J. Neurochem. 49, 1361-1366.

Salomon Y. (1979) Adenylate cyclase assay. Adv. Cyclic Nucleotide Res. 10, 35-·55.

Saxena P. R., and Ferrari M. D. (1989) 5-HT1-like receptor agonists and the pathophysiology of migraine. Trends in Pharmacol. Sci. 10, 200- 204. 164

Schlegel J. R., and Peroutka S. J. (1986) Nucleotide interactions with 5-HTIA binding sites directly labeled by [3H]-8-hydroxy-2-(di-n­ propylamino)tetralin ([3H]-8-0H-DPAT). Biochemical Pharmacol. 35, 1943-1949.

Schoeffter P., and Hoyer D. (1988) Centrally acting hypotensive agents with affinity for 5-HT1A bindIng sites inhibit forskalin-stimulated adenyl ate cyclase activity in calf hippocampus. Br. J. Pharmacol. 95, 975-985.

Shenker A., Maayani S., Weinstein H., and Green.1. P. (1987) PharmacologIcal characterization of two 5-hydroxytryptamine receptors coupled to adenylate cyclase in guinea pig hippocampal membranes. Mol. Pharmacol. 31, 357-367.

Shenker A., Maayani S., Weinstein H., and Green J. P. (1985) The 5-HT receptors linked to adenylate cyclase in guinea pig hippocampus are discriminated by 5-carboxamidotryptamine and spiperone. Eur. J. Pharmacol. 109, 427-429.

Sills M. A., Wolfe B. B., and Frazer A. (1984a) Determination of selective and nonselective compounds for the 5-HT 1A and 5-HT 1B receptor subtypes in rat frontal cortex. J. Pharmacal. Exp. Ther. 231, 480-487.

Sills M. A., Wolfe B. B., and Frazer A. (1984b) Multiple states of the 5-hydroxytryptamine1 receptor as indicated by the effects of GTP on [3H]5-hydroxytryptamine binding in rat frontal cortex. Mol. Pharmacol. 26, 10-18.

Skolnick P., Huang M., Daly J., and Hoffer B. J. (1973) Accumulation of adenosine 3' ,5'-monophosphate in incubated slices form discrete regions of squirrel monkey cortex: effect of norepinephrine, serotonin and adenosine. J. Neurochem. 21, 237-240.

Sutherland E. W.o RaIl T. W., and Menon T. (1962) Adenyl cyclase. 1. Distribution, preparation and properties. J. BioI. Chern. 237, 1220- 1227.

Taylor E. W., Nikam S. S., Lambert G., Martin A. R., and Nelson D. L. (1988) Molecular determinants for recognition of RU 24969 analogs at central 5-hydroxytryptamine recognition sites: Use of a bilinear function and substituent volumes to describe steric fit. Mol. Pharmacol. 34, 42-53.

Taylor E. W., Duckles S. P., and Nelson D. L. (1986) Dissociation constants of serotonin agonists in the canine basilar artery correlate to Ki values at the 5-HT1A binding site. J. Pharmacol. EXp. Ther. 236, 118-125. 165

Taylor E. W. (1985) The development of indoleamine derivatives selective for subtypes of serotonin receptors. University of Arizona, PhD Dissertation.

Traber J., and Glaser T. (1987) 5-HTIA receptor-related anxiolytics. Trends In Pharmacol. Sci. 8, 432-437.

Tricklebank M. D., Farler C., and Fozard J. R. (1985) The involvement of subtypes of the 5-HT1 receptor and of catecholaminergic systems in the behavioural response to 8-hydroxy-2Idi-n-propylamino)tetralin in the rat. Eur. J. Pharmacol. 106, 271-282.

Twarog B. M., and Page I. H. (1953) Serotonin content of some mammalian tissues and urine and a method for its determination. Am. J. Physiol. 175, 157-161.

Von Hungen K., Roberts S. and Hill D. F. (1975) Serotonin-sensitive adenylate cyclase activity in immature rat brain. Braill Research 84, 257-267.

Von Hungen K., Roberts S. and Hill D. F. (1974) Developmental and regional variations in neurotransmitter-sensitive adenylate cyclase systems in cell-free preparations from rat brain. J. Neurochem. 22, 811-819.

Weiss S., Sebben M., Kemp D. E., and Bockaert J. (1986) Serotonin 5-HTI receptors mediate inhibi tion of cyclic AMP production in neurons. Eur. J. Pharmacal. 120, 227-230.

Weissmann-Nanopoulos D., Mach E., Magre J., Demassey Y., and Pujol J.-F. (1985) Evidence for the localization of 5-HT 1A bInding sites on serotonin containing neurons in the raphe dorsalis and raphe centralis nuclei of the rat brain. Neurochem. Int. 7, 1061-1072.

White A. A., and Zenser T. V. (1971) Separation of cyclic 3' ,5'­ nucleotide monophosphates from other nucleotides on aluminum oxide columns. Anal. Biochem. 41, 372-396.

Wikstrom H., Elebring T., Hallnemo G., Andersson B., Svensson K., Carlsson A., and Rollema H. (1988) Occurrence and pharmacological significance of metabolic ortho-hydroxylation of 5- and 8-hydroxy-2- (di-n-propylamino)tetralin. J. Med. Chern. 31, 1080-1084.

WIkstrom H., Andersson B., Elebring T., Jacyno J., Allinger N. L., Svensson K., Carlsson A., and Sundell S. (1987) Resolved cis-10- hydroxy-4-n-propyl-1,2,3,4,4a,5,6,10b-octahydrobenzo[f]quinoline: central serotonin stimulating properties. J. Med. Chern. 30, 1567- 1573. 166

Wikstrom H., Sanchez D., Lindberg P., Arvidsson L.-E., Hacksell U., Johansson A., Nilsson J. L. G., Hjorth S., and Carlsson A. (1982) Monophenolic octahydrobenzo[f]quinolines: central dopamine- and serotonin-receptor stimulating activity. J. Med. Chem. 25, 925-931.

Xiong W., and Nelson D. L. (1989) Characterization of a [3 H]-5-hydroxy­ tryptamine binding site in rabbit caudate nucleus that differs from the 5-HT1A' 5-HT1B' 5-HTIC and 5-HT1D subtypes. Life Sci. 45, 1433- 1442.

Yocca F. D., Hyslop D. K., Smith D. W., and Maayani S. (1987) BMY 7378, a buspirone analog with high affinity, selectivity and low intrinsic activity at the 5-HT1A receptor in rat and guinea pig hippocampal membranes. Eur. J. Pharmacol. 137, 293-294.

Zgombick J. M.. Beck S. G., Mahle C. D., Craddock-Royal B., and Maayani S. (1989) Pertussis toxin-sensitive guanine nucleotide-binding proteins(s) couple adenosine At and 5-hydroxytryptaminelA receptors to the same effector systems in rat hippocampus: biochemical and electrophysiological studies. Mol. Pharmacol. 35, 484-494.