A Structure Activity Investigation of Benzylidene Anabaseine Interactions with the Alpha7 Nicotinic Acetylcholine Receptor

Material Information

A Structure Activity Investigation of Benzylidene Anabaseine Interactions with the Alpha7 Nicotinic Acetylcholine Receptor
Copyright Date:


Subjects / Keywords:
Agonists ( jstor )
Calcium ( jstor )
Cell lines ( jstor )
Dyes ( jstor )
Fluorescence ( jstor )
Ligands ( jstor )
Nicotinic receptors ( jstor )
P values ( jstor )
Rats ( jstor )
Receptors ( jstor )

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright Susan Elizabeth Lefrancois. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Embargo Date:
Resource Identifier:
71781801 ( OCLC )


This item is only available as the following downloads:

Full Text




Copyright 2004 by Susan Elizabeth LeFrancois


This document is dedicated to my husband Ben.


iv ACKNOWLEDGMENTS I would like to thank my mentor Willia m Kem for his guidance, my husband Ben for all of his encouragement and support and my friend and colleague Kristin Wildeboer for her candor and kind spirit. I would al so like to thank my committee members for taking the time to help guide my studies, Dr s. Tu and Prokai for their experimental assistance, Debbie Otero, and the Ph armacology and Therapeutics staff.


v TABLE OF CONTENTS page ACKNOWLEDGMENTS.................................................................................................iv LIST OF TABLES............................................................................................................vii LIST OF FIGURES.........................................................................................................viii ABSTRACT....................................................................................................................... xi CHAPTER 1 INTRODUCTION........................................................................................................1 Nicotinic Acetylcholine Receptors...............................................................................1 Possible Relevance of the 7 nAChR to Human Health..............................................6 DMXBA as a Lead Compound.....................................................................................9 Study Objectives.........................................................................................................14 2 MATERIALS AND METHODS...............................................................................17 Materials.....................................................................................................................1 7 Chemicals and Bioreagents.................................................................................17 Methods......................................................................................................................18 Radioligand Binding Studies...............................................................................18 Octanol/Water Partition Coefficients..................................................................20 Spectrophotometric pKa Determination..............................................................22 Cell Culture.........................................................................................................23 Functional Assays................................................................................................24 Statistics...............................................................................................................26 3 COMPARISON OF THE AFFINITY OF EACH 3-BENZYLIDENE ANABASEINE FOR RAT BRAIN 4 2 AND 7 NACHRS.................................28 Introduction.................................................................................................................28 Results........................................................................................................................ .31 Discussion...................................................................................................................36


vi 4 THE INFLUENCE OF THE IONIZA TION STATE OF TH E LIGAND ON ITS AFFINITY FOR THE RECEPTOR...........................................................................49 Introduction.................................................................................................................49 Results........................................................................................................................ .53 Discussion...................................................................................................................56 5 INFLUENCE OF LIPOPHILICITY OF THE LIGAND ON IT S AFFINITY FOR THE RECEPTOR.......................................................................................................65 Introduction.................................................................................................................65 Results........................................................................................................................ .69 Discussion...................................................................................................................70 6 QUANTITATIVE STRUCTURE ACTIVI TY RELATIONSHIPS USING THE 3BENZYLIDENE ANABAS EINE DERIVATIVES...................................................75 Results........................................................................................................................ .79 Discussion...................................................................................................................80 7 FUNCTIONAL DATA FOR TH E 3-BENZYLIDENE ANABASEINE DERIVATIVES USING CELLS EXPR ESSING THE HUMAN AND RAT ALPHA7 NACHR AND CELLS EXPRES SING THE HUMAN FETAL MUSCLE NICOTINIC RECEPTOR..........................................................................................86 Introduction.................................................................................................................86 Results........................................................................................................................ .90 Discussion...................................................................................................................94 APPENDIX 3-BENZYLIDENE ANAB ASEINE COMPREHENSIVE TABLE..........103 LIST OF REFERENCES.................................................................................................113 BIOGRAPHICAL SKETCH...........................................................................................124


vii LIST OF TABLES Table page 3-1 Calculated Hammett and Hansch-Fujita values for the 3-benzylidene anabaseines...............................................................................................................47 4-1 Table for Figure 4-5 Displacement curves for DMXBA at three different pHs......60 4-2 Table for Figure 4-6 Representative Scatchard Plots at different pHs 6.6,7.4, and 8.2......................................................................................................................62 4-3 Table for Figure 4-7, Displacement curves for PTHP at three different pHs 6.6, 7.4, and 8.2...............................................................................................................63 4-4 Table for Figure 4-8, Displacement curves for 3-[4-Cyanobenzylidene]-anabaseine at three different pHs 6.6, 7.4, 8.2.................64 6-1 Training set of 29 3-benzylidene anabas eine (BA) compounds used to create the alpha7 and alpha4beta2 equations in the SciQSAR 3.0 software............................84 6-2 Prediction set of 2, 3-benz ylidene anabaseine (BA) derivatives that were used to test the validity of the two equations produced by the SciQSAR 3.0 software and the training set...................................................................................................85 7-1 TE-671 cell line EC50 and IC50 values upon addition of the 3-benzylidene anabaseine derivatives using the membrane potential assay..................................100


viii LIST OF FIGURES Figure page 1-1 The Craig plot...........................................................................................................16 3-1 Direct relationship between the published Hammett values for only the meta and para substitutions, representing the electronic influence of a substituent vs. experimentally derived pKa values determined for the imine nitrogen of each 3benzylidene anabaseine............................................................................................40 3-2 Direct relationship between the calculated Hammett values , representing the electronic influence of a substitu ent vs. experimentally derived pKa values determined for the imine nitrogen of th e 3-benzylidene anabaseine derivatives excluding the derivatives w ith an ortho substituent and the derivatives with multiple substituents.................................................................................................40 3-3 Experimentally determined KI values for the alpha7 nAChR vs. the calculated Hammett values ....................................................................................................41 3-4 Experimentally determined KI values for the alpha4beta2 nAChR vs. the calculated Hammett values ....................................................................................................41 3-5 Direct relationship between th e calculated Hansch-Fujita value and the experimental Log P values determined for each 3-benzylidene anabaseine derivative..................................................................................................................42 3-6 Experimentally determined KI values for the alpha7 nAChR vs. the calculated Hansch-Fujita value ..............................................................................................42 3-7 Experimentally determined KI values for the alpha4beta2 nAChR vs. the calculated Hansch-Fujita value ..............................................................................................43 3-8: Comparison of the alpha7(white bars) and alpha4beta2 (solid bars) KI values ( M) for the para substituted 3-benzylidene anabaseines........................................44 3-9 Comparison of the alpha7 (white ba rs) and alpha4beta2 (solid bars) KI values ( M) for the ortho, meta, trisubstituted, para/ortho substitu ted, fused ring, and unsubstituted 3-benzyl idene anabaseines.................................................................45 3-10 Comparison of the alpha7 (white ba rs) and alpha4beta2 (solid bars) KI values ( M) for the compounds with selectivity above 1.4 for the alpha7 nAChR............46


ix 4-1 Log KI (alpha7) for all 3-benzylidene an abaseines vs. the experimentally determined pKa.........................................................................................................58 4-2 Log KI (alpha4beta2) for all 3-benzylidene anabaseines vs. the experimentally determined pKa.........................................................................................................58 4-3 Predicted Log KI (alpha7) that are corrected for the percentage unionized (determined by the Henderson-Hasselbach e quation) at physiological pH for all 3-benzylidene anabaseines vs. the experimentally determined pKa.........................59 4-4 Predicted Log KI (alpha4beta2) that are corrected for the percentage unionized (determined by the Henderson-Hasselbach e quation) at physiological pH for all 3-benzylidene anabaseines vs. the experimentally determined pKa. ......................59 4-5 Displacement curves for DMXBA at three different pHs. .....................................60 4-6 Respresentative Scatchard Plots at the different pHs 6.6, 7.4, and 8.2....................61 4-7 Displacement curves for PTHP (permanently charged ligand at this pH range, pKa>10) at three different pHs. ..............................................................................63 4-8: Displacement curves for 3-[4-Cyanobenz ylidene]-anabaseine at three different pHs for three separate experiments. .......................................................................64 5-1 HPLC trace. 3-benzylidene anabasei ne backextracted octanol phasse sample injected using the ammonium acetate buffer system (pH = 4.5)..............................73 5-2 The Log P values determined using th e octanol/water partitioning system vs. the Log KI values determined through radi oligand binding studies for the alpha 7 nAChR.........................................................................................................74 5-3 The Log P values determined using th e octanol/water partitioning system vs. the Log KI values determined through ra dioligand binding studies for the alpha4beta2 nAChR.................................................................................................74 6-1 SciQSAR 3.0 equation for the alpha7 nAChR using the KI values obtained experimentally for the 29 co mpounds in the training set.........................................82 6-2 SciQSAR 3.0 equation for the alpha4beta2 nAChR using the KI values obtained experimentally for the 29 co mpounds in the training set.........................................83 7-1 Representative trace of a change in fluorescence in GH4C1 cells due to the addition of a 3-benzylidene anabasei ne derivative at three different concentrations...........................................................................................................95 7-2 Calcium flux of epibatidine in SHEP 1 cells (A), and GH4C1 cells (B) in the presence of 1 M atropine.........................................................................................96


x 7-3 Calcium flux curve of 3-[4-(Methy lamino)benzylidene]-anabaseine with percent maximum fluorescence relative to epibatidine in SH-EP1 cells expressing human alpha7 nAChRs..........................................................................97 7-4: Calcium flux curves with GH4C1 cells......................................................................98 7-5 Membrane potential curve using TE671 cells expressing the human fetal muscle receptor in the presence of epibatidine........................................................99 7-6: Membrane potential curve using TE671 cells expressing the human fetal muscle receptor in the presence of 3-[2-Trifluoromethoxybenzylidene]-anabaseine......................................................99


xi Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy A STRUCTURE ACTIVITY INVESTIGAT ION OF BENZYLIDENE ANABASEINE INTERACTIONS WITH THE ALPHA7 NICOTINIC ACETYLCHOLINE RECEPTOR By Susan Elizabeth LeFrancois December 2004 Chair: William R. Kem Major Department: Pharmacology and Therapeutics Neuronal nicotinic acetylchol ine receptors (nAChRs) ar e pentameric ligand-gated ion channels composed of homologous s ubunits. They participate in cholinergic transmission in autonomic ganglia, neuromus cular junctions and brain synapses. By adding a benzylidene moiety to the structure of a lead nicotinic compound, anabaseine, a selective alpha7 nAChR agonist DMXBA [3-(2,4-Dimethoxybenzylidene) anabaseine] was identified. Even though this compound sele ctively stimulates this receptor subtype, it also is a relatively potent antagonist at the alpha4beta2 nAChR. While an unsubstituted benzylidene group attached to anabaseine is sufficient to produce this spectrum of nicotinic activity, substitu tions on the benzylidene ring were shown to modulate the agonist and antagonist properties. How such substituents influence compound ionization, lipophilicity (Log P) and the com poundÂ’s interactions with nACh Rs is the subject of this dissertation. By directly co mparing a subset of 3-benzy lidene anabaseines that have different substituents on the benzylidene ring at the same position one can infer structure


xii activity relationships. Using th ese structure activity relations hips we have quantified the importance of certain types of structural ch ange for receptor binding. The data presented in this dissertation supports th e conclusions that the fused ring 3-benzylidene anabaseines are more selective for the alpha7 as compared to the alpha4beta2 nAChR indicating that there is unexploited space in th e alpha7 binding pocket that co uld be exploited to create a selective alpha7 ligand. The most important pr operties to take into consideration when designing a 3-benzylidene anabaseine deriva tive based on affinity appears to be the electronic influence of the substituent and the ionization state of the imine nitrogen. Also, lipophilicity does not appear to affect affinity for either receptor subtype. Finally, Quantitative Structure Activity Relationship (QSAR) models were obtained for both receptors in order to predict binding affinity of a 3-benzylidene anabaseine derivative without having to synthesize the compound.


1 CHAPTER 1 INTRODUCTION Nicotinic Acetylcholine Receptors Neurons that secrete the neurotransmitter acetylcholine (ACh) ar e referred to as cholinergic neurons. Central c holinergic neurons have been implicated in a number of neurodegenerative conditions including Alzhei mer’s disease (AD) and schizophrenia. AD affects an estimated 15 million people worldwide and accounts for approximately 5060% of the overall cases of dementia fo r people over the age of 65 (Doody, 1999). The characteristic pathology of AD includes extracellular -amyloid plaques, intracellular neurofibrillary tangles, and lo ss of neuronal synapses and pyram idal cells (Francis, 1999). The cholinergic dysfunction in Alzheimer’s di sease includes a reduction in the activity of the ACh-synthesizing enzyme cholineacetyltran sferase (ChAT) (Danielsson et al., 1988) and a loss of certain nAChRs. In schizophren ia, there is a disruption in the normal brain sensory “gating” mechanism that eliminates responsiveness to repetitive stimuli, thus suppressing distracting stimuli. This malfunction in the filter for sensory input may cause an overload of stimuli, and can lead to mi sperceptions, delusions, withdrawal from such stimuli and schizoid behavior (Freedman et al., 2002). In the mammalian central nervous system (CNS) acetylcholine receptors can be divided into muscarinic (mAChR) and nico tinic (nAChR) subtypes. These subtypes are distinguished based on their ability to be stimulated by either the mushroom toxin muscarine or the plant alkaloid nicotine. nAChRs are important in cholinergic transmission, autonomic ganglia, the neurom uscular junction and in brain synapses


2 (Clementi et al., 2000). nAChRs are cation sele ctive ligand-gated ion channels that form pentameric structures in the plasma membra ne. Each subunit of the pentamer contains four transmembrane domains. There are at least seventeen different nAChR subunit genes, including five found in striated muscle ( 1, 1, , , ) and twelve neuronal nAChR subunits ( 2-10, 2-4). These channels can be co mposed of a number of different combinations of subunits. For example, the most abundant subtypes in the brain are the 7 subtype ( -bungarotoxin sensitive) and the 4 2 subtype. Different combinations of the and subunits exhibit different physiological and pharmacological properties. Like many ion channels, the exact composition of the 7 nAChR is still a matter of debate. However, there is now strong evidence that supports the idea that most 7 receptors are expressed as homopentamers. Bungarotoxin-sensitive receptors from pheochromocytoma 12 cells and many differe nt tissue preparations contained only 7 subunits and not 3, 5, 2 or 4 subunits (Drisdel and Gr een, 2000; Chen and Patrick, 1997). Also, functional bungarotoxin-sens itive channels were expressed in Xenopus oocytes when only 7 cDNA had been injected (Couturie r et al., 1990). However, there is some heterogeneity in 7 nAChRs expressed in rat intr acardiac and superior cervical ganglion neurons. One subtype exhibits a slow rate of desensitizati on and incorporates a novel 87-base pair cassette exon in the N terminus. This 7 splice variant has been shown to form functional homomeric ion channe ls that are activate d by acetylcholine and blocked by -BTX when expressed in Xenopus oocytes. (Severance et al., 2004). Besides the ability to assemble into homomeric channels, other proper ties unique to the 7 nAChRs include their greater permeability to Ca+2 than the other nAChRs and the NMDA


3 glutamate receptor subtype (Seguela, 1993) a nd their rapid desensitization kinetics (Couturier et al, 1990). The amino-terminal domain of the neurona l nAChR contains the ACh binding site determinants (amino acid side chains). It encompasses approximately 50% of the nAChR protein mass. The binding site for the hete romeric nAChRs is proposed to be between nonand an subunit (including sites on the N-te rminal domains). Even though the 7 nAChRs are homopentamers, their binding si tes are proposed to be similar to their heteromeric counterparts. The 7 receptor has been shown to contain the characteristics of the binding site normally confer red by two different subunits, both nonand -like contributions. Affinity labeli ng studies that identified the subunit as the major component of the ACh binding site helped identify four sepa rate regions in the sequence of the subunit (loops A, B, C, and D) that cont ribute to the binding si te (Prince et al., 1998; Corringer et al., 2000). Each loop cont ains conserved aromatic residues that stabilize ACh through aromatic/ quaternary ammonium interac tions. The consensus on the ACh binding site is that it incl udes a conserved core of arom atic residues that aid in stabilization of the ligand-receptor complex with their electron-dona ting side chains. The residues forming the principal side of the binding site are Tyr89, Trp143, Tyr185, Cys187, Cys188, and Tyr192. The residues forming the complementary side of the binding site are Trp53, Arg104, Leu112, and Met114 (Karlin, 2004). In the 7 receptor alteration in residues W54, Y92, W149 and Y 188 modifies the apparent affinities of binding and activation of ACh, establishing th eir contribution to the ACh binding site (Corringer et al., 2000).


4 A pharmacophore is the minimal structural features necessary to define a given pharmacological action. It can be based on functional activity or on binding data. However, because activity typically repr esents a combination of pharmacodynamic, pharmacokinetic, and pharmacophoric fact ors, depending on the system being considered, it might be expected that pha rmacophore models based on binding are more reliable. When using binding data to develop an accurate pharmacophore a ligand that has all of the required pharmacophoric features might not bind with high affinity. For example, the ligand might have an added co mponent that disrupts the ability of the compound to bind to the receptor. Therefore, it is possible to have multiple pharmacophore models for a particular pharmaco logical effect. In other words parallel structural changes do not necessarily result in parallel shifts in affinity (Glennon et al., 2004). Several groups have proposed an nAChR pharmacophore that is viewed as having a three-point interaction with the receptor. The three inte ractions that are necessary include: a quaternary nitrogen, a pi bonded electronegative atom th at is capable of hydrogen bonding, and the third interaction is a positive dipole set up by the second interaction (Prince and Stevens, 1998). One of the proposed key elements in the binding of ligands to the nAChR is a hydrogen bond between a receptor hydrogen donor and a ligand hydrogen acceptor. This bond is formed 5.9 Ã… from the positively charged nitrogen atom (Beers and Reich, 1970). Howeve r, it is known that this portion of the classical Beers and Reich nicotinic phar macophore based on binding is not a strict requirement for activation, because simple quaternary ammonium compounds like tetramethylammonium activate nAChRs. Therefore, an ideal pharmacophore model should include both binding and functiona l data. However, a pharmacophore model


5 based on binding may produce completely di fferent predictions than a pharmacophore model based on functional data. Using a freezing process and electron micr oscopy with the muscle nAChR from the electric ray, Unwin (2003) has shown that th e binding of acetylcholin e initiates rotations of the -sheets of both the subunits. These rotations are conveyed to the inner, pore lining helices and cause them to cooperativel y change into an a lternate conformation. This change destabilizes th e gate, a hydrophobic girdle formed by weak side-to-side interactions between the inner he lices, and opens the channel. Most features of the nAChR binding site model are present in the binding site identified within the acetylcholine bi nding protein (AChBP), a homopentameric structural homolog of the N-terminal lig and binding domain of a nAChR alpha subunit (Brejc et al., 2001, Smit et al., 2001). Stru ctural models to analyze receptor-ligand interactions have been greatly advanced by crystallographic in vestigation of the AChBP. This water-soluble protein, isol ated from the glia cells in a freshwater snail, has become an established model for the extracellular domain of the pentameric ligand-gated ion channels (Cromer et al., 2002; Sixma and Sm it, 2003; Karlin, 2002; Reeves and Lummis, 2002). It has high sequence similarity to th e N-terminal extrace llular halves of the ionotropic acetylcholine, gamma-aminobutyric acid, serotonin and glycine receptor subunits belonging to this family. The amino acid sequence of AChBP is most similar to the extacellular domains of vertebrate neuronal 7 nAChRs (Karlin, 2002). Also, the pharmacological properties of the AChBP are similar to the homopentameric 7 nAChR subtype, with relatively weak affinity for ACh and an approximately 10fold higher affinity for nicotine (Hansen et al., 2002; Smit et al., 2001). AChBP has also been shown


6 to bind nAChR agonists such as carbamylcholine and epibatidine and competitive antagonists such as tubocurarine and -bungarotoxin with similar affinities as the 7 nAChR (Karlin, 2002). Celie et al. (2004) de monstrate that alignm ent of the AChBP, with nAChRs shows that residues involved in ligand binding from the principal subunit are generally conserved, whereas residues in the complementary part of the binding site show more variation. The crystal structures of AChBP contai ning nicotine, carbamylcholine/or HEPES in the binding pocket have been solved and show that both nicotine and carbamylcholine bind at similar positions and cause si milar protein conformational changes. Superpostition of AChBP bound by nicotine and carbamylcholine shoes that these ligands bind with their nitrogen atoms at almost the same position however the amino acids that they interact w ith are not identical. Thermodynamic data shows a gain in enthalpy of nicotine versus carbamylcholin e binding which indicates that there are stronger interactions in the AC hBP for nicotine. These intera ctions can be accounted for based on the presence of the hydrogen bonds to nicotine and to the larger number of contacts with the protein for nicotine as th an for carbamylcholine (Celie et al., 2004). Thus, AChBP provides a vital structural model for studying ligand recognition by the nAChRs (Karlin, 2002). Possible Relevance of the 7 nAChR to Human Health Neuronal nAChR deficits have been imp licated in several diseases including AlzheimerÂ’s disease (AD) and schizophrenia (Whitehouse et al., 1986; Freedmen et al., 1995). Until recently, the study of neurodegenerative diseases focused on the muscarinic type neuronal acetylcholine receptor becau se of its abundance in the brain when


7 compared to the population of neuronal nicotin ic AChRs. However, the discovery of a greater relative loss of nicotinic receptors than of muscarinic receptors in the Alzheimer’s brain, as well as evidence that nicotinic a gonists enhance cognition has spurred interest in nAChRs as drug targets (Arendash et al ., 1995a; Araujo et al., 1988; Kem, 2000). Additional evidence of enhan ced attentiveness and rapid information processing in humans because of nicotine treatment has been collected. Protecting the vulnerable neurons in the above-mentioned diseases has been a therapeutic approach to treating neurodegenerative conditions; agonists of nAChRs are among the growing arsenal of compounds reported to be neuroprotective (Mey er et al., 1998; Sullivan et al., 1995). The possibility that drugs acting at neuronal nA ChRs may exert favorable effects in central nervous system diseases has prompted a sear ch for compounds that can selectively affect their function. It is importan t to understand why some potential drug candidates interact potently with a particular nAChR subtype and not to another subtype. Instead of depending upon serendipity to find a drug, determining how a ligand’s structure determines its interaction with a defined r eceptor would provide a more rational basis for drug design. The most prominent nAChR subtypes in the brain can be separately examined using the selective radioligands, sn ake venom alpha-bungarotoxin for the 7 subtype and nicotine or cytisine for the 4 2 nAChR subtype. The 2 containing nAChR subtypes constitute the high affinity bi nding sites for (S)-nicotine, the natural stereoisomer (Flores et al., 1992). Another lab showed elimination of (S)-nicotine binding in brain slices using a transgenic “knock-out” mouse strain lacking a functional 2 gene. These mice experienced distorted learning and enhanced neurodegeneration during aging (Picciotto et


8 al., 1995; 1998). It has been documented that the 4 2 receptor is the main receptor population which is greatly reduced by Alzh eimerÂ’s disease when comparing AD brain samples to normal aging brain samples. The neocortex and the hippocampus, regions associated with higher mental functions, ar e the most affected areas of the brain in AD (Francis, 1999). Recent studies in AD brains sh ows that the loss of nicotine binding sites in the neocortex is associated with a marked reduction of the 4, but not 3 or 7 subunits (Martin-Ruiz et al., 1999) . Since the reduction in the 7 subtype in AlzheimerÂ’s disease patients is not as significant as the 4 2 subtype, this nicotini c receptor is still around to be pharmacologically manipulated (Kem, 2000). Also, the 7 subtype has been shown to cause long-term synap tic modulation through its influence on glutamatergic synapses. Strong, br ief stimulation of presynaptic 7 containing nAChRs can enhance hippocampal glutamatergic synapt ic transmission for a matter of minutes after a nicotinic agonist has been removed (Radcliffe and Dani, 1998). An interesting observation regarding schiz ophrenics is that th eir propensity to smoke is approximately three times highe r than the normal United States smoking population (Ripoll et al., 2004). Further, this presumed self-medication with nicotine has been demonstrated to transiently normalize a diminished suppression of auditory-evoked response in schizophrenic patients (Adl er et al.,1992; 1993). This auditory-evoked response is a way to measure th e inability of schizophrenic pa tients to filter excessive sensory stimulation from their environmen t. The incapability to filter out excess information could be a reason why schizophren ics have a diminished capacity to focus (Martin et al., 2004). Agonists of the 7 nicotinic receptor were found to increase sensory inhibition in both humans and in an animal model (Stevens et al., 1998). Also, the locus


9 of the human 7 gene on chromosome 15 has been associated with some forms of schizophrenia by the Freedman lab. In addition, radiolabeling of the 7 subtype and to a lesser extent, 4 2 subtype in the hippocampus was decreased in schizophrenics (Freedman et al., 1995). The results of thes e studies suggest an abnormal expression and function of 7 and possibly 4 2 nAChRs in schizophrenia. Besides their role in central nervous system disorders, nAChRs have recently been implicated in non-neuronal cells both w ithin and outside the nervous system. For example, systemic inflammation (sepsis) a nd associated tumour-n ecrosis factor (TNF) secretion cause morbidity and mortality in many human diseases. The nervous system, through the vagus nerve, can i nhibit the release of macrophage TNF. Wang et al. (2003) showed that ACh inhibits HMGB1, a m acrophage activator, release from human macrophages by signaling through the alpha7 nAChR. In 7 deficient mice vagal stimulation was unable to inhibit TNF rel ease from macrophages during the systemic inflammatory response to endotoxaemia. Since, there seems to be an ever-expandi ng role for nAChRs as drug targets, and there is a desire to reduce side effects by development of receptor subtype-selective drugs, there clearly is a need to find a more selective 7 agonist. DMXBA as a Lead Compound Anabaseine, a marine worm toxin, is a potent nicotinic acetylcholine receptor agonist which stimulates all nicotinic cho linergic receptors to some degree (Kem, 1971; Kem et al., 1997). Anabaseine contains a 2-tetrahydropyr idyl ring whose imine double bond is electronically conjugated with a 3-pyridyl ring. This -electron conjugation between the two rings causes its two rings to be approximately coplanar in their


10 orientation, unlike the tobacco alkaloids ni cotine and anabasine, whose two rings are almost perpendicular to each other. An abaseine expresses its highest potency on neuromuscular and neuronal 7 nAChRs (Kem et al., 1997) a nd is as toxic as nicotine when injected into mice (Kem et al., 1976). This toxin in its iminium active form is slightly more potent than ACh on the mouse embryonic neuromuscular nicotinic receptor. Also, the affinity and potency of anabaseine for the alpha7 nAChR is significantly higher than that of nicotine according to rat brain BTX binding displacement and Xenopus oocyte recordings. Anabaseine was reported to have a KI of ~60 nM for the rat 7 nAChR, while nicotineÂ’s affinity (KI, 400 nM) for this receptor subtype is much lower in rat (Kem et al., 1997). This toxin has been s hown to stimulate ACh release from rat brain cortical minces as well as elevate cortical ACh and norepinephrine levels in the intact rat (Meyer et al., 1987; Summers et al., 1997). Anabaseine as well as nicotine enhance passive avoidance behavior in nucleus basa lis-lesioned rats (Mey er et al., 1987). Since a benzylidene derivative of an abaseine, (DMXBA or GTS-21, (3-[2,4(dimethoxy)benzylidene]-anabaseine), has been shown to enhance cognition (Meyer et al., 1997; Woodruff-Pak et al., 1994; Arendash et al., 1995b) it was used as the lead compound in this dissertation. Altering the structure of a lead compound by modifying the substituents can change the various characteristics of that lead compound, including the electron density, conformati on, ionization, solubility, bioa vailability, its distribution to its target, and its in teraction with the receptor. Unfo rtunately, one cannot alter a single characteristic specifically. One approach is to add a substituent on a compound that affects only one of these characteristics dom inantly (Wermuth, 1996). Benzylideneand cinnamylidene-anabaseines were identified as analogs of the lead compound anabaseine


11 with selective agonism for the 7 receptor (de Fiebre et al ., 1995). Papke et al. (2004a) have shown that the 7 receptor has relatively relaxed requirements for activation via the agonist binding site; it seems that the benz ylidene group creates a steric hinderance, keeping the anabaseine moiety from being able to readily activate the 3 4, 4 2, and muscle-type nAChRs. If the benzylidene anabaseine does bind, the benzylidene group somehow prevents activation of the occupied receptor. DMXBA is a partial agonist on the 7 nAChR. It has perhaps the broadest range of cytoprotective effects compared to many other nicotinic agonists (RJR-1734, AB T-418, ABT-089) that primarily target 4 2 receptors. DMXBA bioavailability is only approximate ly 20% (Mahnir et al., 1997). Kem et al. (2004) have shown that DMXBA after oral administration is metabolized into three hydroxy metabolites which display similar binding affinities and partial agonist potencies (EC50Â’s) similar to their pare nt compound at rat brain 7 receptors. Each metabolite displayed a higher efficacy than DM XBA for activating both rat and human 7 receptors. Unfortunately, these metabolites are more pol ar than DMXBA and do not enter the brain as readily as DMXBA. Since DMXBA is a promising drug candidate, we shall search for chemical properties that make this t ype of compound selective for the 7 receptor by investigating a variety of benzylidene anabaseine analogs, in an attempt to understand how a change in the structure of DMXBA affects the ability of that altered compound to bind to the acetylcholine binding site. By testing a num ber of 3-benzylidene anabaseines that contain substituents which represent different ranges of values for chemical properties like lipophilicity, electronic influence, and size it should be possible to compile a


12 Quantitative Structure Activity Relationship (QSAR) model. This model could indicate the positions on the molecule which are the most pharmacologically sensitive to substitution and the types of substituents that give the most favorable effects. Apparently the first QSAR studies, where the chemical structure of a compound was proposed to have a direct effect on biologi cal activity were carried out in the mid-nineteenth century (Crum-Brown and Fraser, 1868-9). The pharm acological properties of six alkaloids (strychnine, morphine, codeine, brucine, th ebine and nicotine) and their methylated derivatives was examined. It was postulated that the physiological function of a compound, defined as , is a function of its chemical c onstitution, c. They utilized an equation: = (c), to illustrate how a change in the chemical structure of a molecule results in a predictable ch ange in biological action. The biological action of an agonist is dependent on both its affinity and its efficacy. Liapakis et al. ( 2004) studied the structural basis of ligand affinity by examining the interactions between a series of epinephrine anal ogs and the G-protein coupled receptor (GPCR) B2 adrenergic receptor (B2AR). By comparing the affinities of the various epinephrine analogs it was possible to determine the contribution of each chemical substituent to the binding affinit y. This group showed through binding studies that the derivatives bound to the B2AR through a series of conf ormational intermediates. They found that the unliganded state of the receptor was a minimal, low affinity binding state that bound a ligand based on a few general structural features (eg., aromatic ring and an amine group). However, once the ligand was bound there was a conformational transition of the receptor, stabilized by an interaction between the receptor and the substituent added to th e epinephrine ligand.


13 Another example of studies comparing stru ctural changes on a series of related ligands to the binding affinity was done by Sharples et al. (2002). This group investigated two classes of ligands synthesized fro m the lead structure ( )UB-165, a potent neuronal nicotinic acetylcholine ligand. They assessed the binding affinities of these compounds on three major nAChR subtypes ( 4 2, 3 4, 7). This group prepared and tested a rational series of compounds based on a lead structure to define nAChR subtype specific pharmacophores. Substituents on aromatic rings can be readily selected according to their electronic and lipophilic properties using the Craig Plot (Figure 1), which considers two substituent parameters: (1) the Hammett value , which estimates the electronic influence of a substituent, and (2) the Hansch-Fujita , which estimates the lipophilicity of a substituent. The Hammett sigma constants were originally derived from an investigation of substituent electronic effects on the ionization constants of meta and para substituted benzoic acids in aqueous medium at room te mperature. The more negative the sigma value the more electron donating the substitu ent, the more positive the sigma value the more electron withdrawing the substituent . For benzoic acids, the more electron withdrawing a substituent the higher the pKa, acidity increases, m eaning that the electron withdrawing groups added to the benzoic acid structure create a more acidic compound than benzoic acid itself. While the electr on donating groups added to the benzoic acid structure make the compound a weaker acid compared to benzoic acid. The published sigma values for substituents were determined using the Hammett equation: = log Ks/Ko, where Ks is the ionization constant for the substituted benzoic acid, and Ko is the ionization constant for benzoic acid itself (Hansch et al., 1995). Hammett showed that


14 this equation is quite reliable for a wide selection of reactions of meta and para substituted aromatic compounds. However, this equation is not readily applicable to ortho substituted aromatic compounds. Estimation of the electronic influence of an ortho substituent to an aromatic compound is compli cated by steric effects of the substituent. The Hansch-Fujita constant, , measures the hydrophobicity contribution of a single substituent. This parameter, , is defined as = log (PX/PH) where PH is the partition coefficient of a parent compound and PX is the partition coefficient of a derivative. The published values were calculated from experimentally determined octanol-water partition coeffici ents (Hansch et al., 1995). Study Objectives Determining what properties of a ligand are necessary to interact specifically with the 7 nAChR would be very valuable when designing drugs. DMXBA ( 3-[2,4 dimethoxybenzylidene]anabaseine), has been su ccessful in a clinical trial; however, this ligand also has weak antagonist properties on the 4 2 nAChR subtype and its bioavailability is only moderate. The 3-be nzylidene ring on the primary anabaseine structure has been shown to be nece ssary for selective activation of the 7 nAChR. Since, DMXBA is a relatively, unique 7 ligand, further exploration of the receptor binding properties of this family of compounds was desirable. The present study explored a series of 3-be nzylidene anabaseine derivatives to find answers to these questions: 1. What types of substituents on the benzylidene ring and in what position confer 7 affinity and/or 4 2 affinity? 2. Can greater selectivity for the 7 receptor be obtained by modifying the benzylidene ring at a particular position with a particular substituent?


15 3. Do the chemical properties of a partic ular substituent on the 3-benzylidene ring alter the affinity of the molecule for the 7 binding site? 4. Is ionization of the 3-benzylidene an abaseine imine nitrogen important for interaction with these nAChRs? 5. Does the lipophilicity of the entire mol ecule contribute to the affinity of the 3benzylidene anabaseines for the 7 and/or the 4 2 nAChR? The studies presented here evaluated the ch anges in affinity of the 3-benzylidene derivatives by using radioligand binding st udies with washed Sprague-Dawley brain membranes. The KIs for each of the 3-benzylidene anabaseine derivatives were assessed using [3H]cytisine for the 4 2 nAChR subtype and [125I] -bungarotoxin for the 7 nAChR subtype. The pKas of the imine nitrogen on the 3-benzylidene anabaseines were determined for each derivative using a spect rophotometeric method. The determination of lipophilicity for the derivatives, Log P valu es, were calculated us ing the octanol/water experimental shake-flask method in conjunc tion with Reverse Phase-High Perfomance Liquid Chromatography (RP-HPLC). Also, the st ructures of all of the derivatives were used in a Quantitative Structure Acitivity Relationship computer program, SciQSAR, to establish a database. Which will make it possible to create a compound related to 3benzylidene anabaseine in silico , and without having to pe rform any experiments, produce reliable predictions about the molecu lar and pharmacological properties of that compound. Finally, functional assays of the compounds were performed using a fluorescence signal workstation (Flexsta tion) and cell lines expressing human 7 (SHEP1), rat 7 (GH4C1) and human fetal muscle (TE-671) nAChRs.


16 Figure 1-1: The Craig plot. The Hammett value, representing the substituents electronic influence, on the y-axis and the Hansch value, representing the hydrophilicity of the substituent, on the x-axis (M.E. Wolff, 1980).


17 CHAPTER 2 MATERIALS AND METHODS Materials Chemicals and Bioreagents [125I] -bungarotoxin (142 Ci/mmole) and [3H]cytisine (1 Ci/mmole) were obtained from PerkinElmer Life and Analytical Scie nces (Billerica, MA). BCA protein reagent A and B as well as Albumin Standard was purch ased from Pierce (Rockford, IL). 1-Octanol and Acetonitrile HPLC grade were purchased from Fisher Scientific (Fair Lawn, NJ). Sodium Phosphate Monobasic and Dibasic were obtained through Fisher Scientific (Fair Lawn, NJ). Cell culture media was purchase d from American Tissue Culture Collection (ATCC) (Manassas, VA). Hygromycin B was obt ained from Calbiochem (La Jolla, CA). Penicillin/Streptomycin and Fetal Bovine Serum were purchased from Cellgro by Mediatech (Herndon, VA). Trypsin (1:250) soluti on was purchased from Irvine Scientific (Santa Ana, CA). The flexst ation calcium assay and memb rane potential kits were obtained from Molecular Devices (Sunnyvale, CA ). All other chemicals were ACS grade and were obtained from either Sigma Chemi cal Co. (St. Louis, MO) or from Fisher Scientific (Fair Lawn, New Jersey). The benzylidene anabaseines used in th is study were synthesized in the Kem laboratory by Dr. Ferenc Soti, generally according to published methods (Kem, 1971; Zoltewicz et al., 1993; Kem et al., 2004). Thei r purity was ascertained by HPLC, NMR and elemental analyses.


18 Methods Radioligand Binding Studies Assays were performed where the experi mental compound competes for its nAChR binding site with a radioligand that specifically labels either the 4 2 or the 7 nAChR using homogenized Sprague-Dawley rat brain membranes. These assays determine the IC 50 of the experimental compounds, which is then used to determine the KI with the Cheng-Prusoff equation. The rat brain memb rane was obtained frozen from Pel-Freez Biologicals (Rogers, AR). The protein con centration of the rat brain homogenate was assessed using the bicinchoninic acid (BCA) protein assay kit and the Standard Protocol from Pierce (Rockford, IL). The radioligand binding assays were pe rformed according to Marks and Collins (1982) for the [125I] -bungarotoxin experiments and a modified Pabreza et al. (1991) method for the [3H]cytisine experiments. To asse ss the binding affinity of the compounds for the 7 nAChR 1 nM [125I] -bungarotoxin was incubated with 0.2 mg of rat brain homogenate and a particul ar concentration (ranging from 5nM-50 M) of 3benzylidene anabaseine deriva tive or 1mM nicotine in orde r to determine non-specific binding. The final volume was brought up to 0.5ml with a 2mg/ml concentration of bovine serum albumin (BSA) suspended in bi nding saline (120 mM NaCl, 5mM KCl, 2 mM CaCl2, 1mM MgCl2, 50mM Tris buffer, pH of 7.4) in order to reduce non-specific binding. To assess total binding, only the radioligand and the membranes were incubated together in 2mg/ml BSA c ontaining binding saline. The 3-benzylidene anabaseine derivative was suspended either in binding saline , with 2mg/ml BSA if it was a salt or in methanol if it was a free base and diluti ons were made in 2m g/ml BSA containing


19 binding saline. The membrane su spensions were incubated at 37 C for approximately 2.5 hours. Once the reaction was assumed to be at equilibrium it was stopped by harvesting the membranes on GF/C filters, which were pre-soaked for 45 minutes in a 0.5% polyethylenamine solution to reduce non-sp ecific binding of the radioligand to the filter, using a Brandel Harvester. The membra nes were washed three times with 3.0 ml of ice-cold Tris buffer by vacuum filtration. The filters were then placed in gamma vials and counted on a Beckman Biogamma counter fo r 5 minutes per sample. The counts per minute were then assessed, the re sults were tabulated and the IC50s and KIs were determined using the software program Graphpad Prism . An experimentally determined -BTX Kd of 0.32 nM was used in the Cheng-Prusoff equation to calculate the displacing compound KI. The [3H]cytisine experiments were handled in a similar manner. First the addition of all of the components: 1nM [3H]cytisine, 3-benzylidene anabaseine derivative (5nM50 M), 0.2 mg rat brain homogenate, and 2mg/ ml BSA containing binding saline at a pH of 7.4 to bring the final volume up to 0.5 ml in each tube. To assess non-specific binding, 1mM nicotine was added in place of the be nzylidene derivative. The incubation time and the temperature had been altered in this protocol to produce th e greatest difference between total and non-specific bi nding and to maximize the affi nity of the radioligand for the receptor, respectively (Pabreza et al., 1991). The incubation time for [3H]cytisine was 4 hours at 4 C. The reaction was stopped using a Brandel Harvester and the membranes were washed three times with 3.0 ml of icecold tris buffer by vacuum filtration. The GF/C filters were placed in sc intillation vials with a 30% Scin tasafe scintillation cocktail overnight. The samples were then placed into a Beckman scinti llation counter for 5


20 minute counts per sample. The specific binding counts per minute were then entered into GraphPad Prism where the IC50s and the KIs were assessed. An experimentally determined [3H]cytisine Kd of 0.92 nM was used in the Ch eng-Prusoff equation in order to calculate the KI . Octanol/Water Partition Coefficients Log P values were determined for each 3benzylidene anabaseine derivative by RPHPLC analysis, essentially as described in Kem et al.(2004). Approximately 1mg of each compound was weighed out and placed in equal volumes, 3 mls, of a 10mM sodium phosphate buffer, pH 7.4 which contained 150 mM NaCl, and 1-octanol. Previously, equal volumes of the sodium phosphate buffer and 1-octanol were added to a separatory funnel and allowed to equilibrate overnight. The equilibrated s odium phosphate buffer was checked prior to its add ition to the compound for a pH of 7.4. Once these phases were added to the weighed out compound, thes e solutions were allowed to equilibrate overnight on a gentle shaker at room temperat ure. Next, the samples were centrifuged at 1xg for 5 minutes. The octanol phase was caref ully removed with a Pasteur pipette. The pH of the water phase was retaken in order to calculate the corrected Log P value. Since, octanol can not be directly injected into the HPLC the amount of compound which was in the octanol phase had to be extracted out through the use of an acidic 150mM NaCl solution. This solution was brought to a pH of approximately 2.6 with 100mM glacial acetic acid in order to drastically reduce solu bility of the compound in the octanol phase. This acid solution was added to the octanol phase and allowed to mix with the octanol phase for 20 minutes while gently shaking at room temperature. The samples were then centrifuged again at 1xg for 5 minutes. The octanol phase was then removed carefully


21 with a Pasteur pipette. This back extraction step was repeat ed two to three more times for a total of three or four separate back extractions of the co mpound from the octanol phase. Once these samples were collected a ll of the back extracted solutions for a particular compound were combined. Dilutio ns of the aqueous and back extracted octanol phases were then made in a 50 mM ammonium acetate buffer, pH 4.5. This ammonium acetate buffer was the buffer used as the mobile phase in the HPLC. For aqueous solutions 1:1 and 1:5 dilutions we re made of the aqueous phase to the ammonium acetate buffer and then these samples were centrifuged. For octanol solutions 1:5 and 1:10 dilutions of the back extracted octanol phase to ammonium acetate buffer were made and centrifuged. The dilutions were then moved by Pasteur pipette to autosampler compatible tubes and 350 -500 l of the diluted sample was injected into the HPLC. The area under the curve (AUC) was de termined for the peak which represented the highest change in absorbance for a pa rticular compound. The following equations were then used to determine Log P: (1) AUCwater 1:1 dilution factor = A [A+(AUCwater 1:5 dilution factor)]/2 = AUCwater average 5(volume injected, 500 l =5, 350 l=3.5) =[AUC]water (2) AUCoctanol 1:5 dilution factor=B [B+(AUCoctanol 1:10 dilution factor)]/2 = (AUC octanol average 4(number of back-extractions) 5 = [AUC]octanol (3) Pvalue= ([AUC]octanol/[AUC]water)(1/1) (4) = percentage of charged species pH (of water phase)=pKa + log [B/BH+] =Solve for BH+ (5) Log[Pvalue] = Log P


22 For some samples the aqueous phases contai ned such a small amount of compound that they had to be directly injected into the HP LC. These samples, which were not diluted in ammonium acetate buffer, had to have their pH brought down to 4.5 with glacial acetic acid. This pH was compatible with the bu ffer system being used in the HPLC. 500 l injections of the water phases were manua lly injected into the HPLC and the Log P values were calculated as above . Each 3-benzylidene anabasei ne had a total of three Log P values determined. Spectrophotometric pKa Determination The pKa of the most basic, imine, nitrogen of every 3-benzylidene anabaseine was determined by analysis of the pH depende nce of the imineÂ’s electronic absorbance spectrum at room temperature using a 50mM potassium phosphate buffer in the presence of 150mM NaCl. Thirteen different pH values in the titration region were evaluated (pH: 4, 5, 6, 7, 7.2, 7.5, 7.8, 8, 8.2, 8.5, 8.8, 9, 10). A specific concentration of 3-benzylidene anabaseine compound, 1.3 10-3 M, was added to the potassium phosphate buffers at varied pH values. Each tube was then vortexed and immediately read in a Beckman spectrophotometer. The same quartz cuvette was used with every sample. The cuvette was thoroughly rinsed with dist illed deionized water between different pH samples. The wavelength scan option was selected from the main menu of th e spectrophotometer software and the UV and visible lamps were turned on. The wavelength scan was set to a range of 250-600 nm. Each sample in a seri es was read and the pH with the highest change in absorbance was determined. Th e wavelength where the highest change in absorbance was found and the absorbance values for all of the different pH samples at this wavelength were recorded. All of the absorbance values at a particular wavelength


23 were then entered into the Enzfitter software (Elsevier-Biosoft, Cambridge, UK) in order to estimate each pKa value, according to the equation: Y= (Lim + (Lim2 Temp1)/(Temp1 +1)). Where Lim1 is the lower limit or the lowest absorbance value, Lim2 is the upper limit or the largest abso rbance value and Temp1 is defined as 10 (f (pH-pKa)). The f variable is e qual to 1, if the absorbance curve goes up, or –1 if the absorbance curve goes down. Cell Culture The human epithelial cell line SH-EP1 expresses the recombinant human 7 nAChR, obtained from RJ. Lukas (St. Jose ph's Hospital and Medical Center, Phoenix, AZ). This cell line is native nAChR-null. Th ey were maintained in Dulbecco’s Modified Eagles Medium supplemented with 5% (w/v) fetal bovine serum, 10% heat-inactivated horse serum, penicillin/streptomycin at 100 g/ml, 2 g /ml Amphotericin B, 0.4 mg/ml hygromycin B, and 2.2 mg/ml sodium bicarbonate in a humidified atmosphere containing 5% CO2 at 37 C. The rat pituitary GH4C1 cell line expresses the rat 7 nAChR, obtained from M. Quik (Parkinson’s Institut e in Sunnyvale, CA). This cell line is a clonal line that does not endogenously express nico tinic receptors. They were maintained in F-10 nutrient mixture supplemented with 10% (w/v) fetal bovine serum (FBS), penicillin/streptomycin at 100 g/ml, and 0.4 mg/ml hygromycin B, in a humidified atmosphere containing 5% CO2 at 37 C. The human rhabdomyosarcoma TE-671 cell line expressing the fetal muscle nAChR was obt ained from J.W. Daly (National Institutes of Health, Bethesda, MD). This cell lin e endogenously expresses the fetal muscle nicotinic receptor. They were maintained in Dulbecco’s Modified Eagles Medium supplemented with 10% (w/v) fetal bovine serum (FBS), and penicillin/streptomycin


24 100 g/ml, in a humidified atmosphere containing 5% CO2 at 37 C. Cells were harvested weekly using 0.25% trypsin and seeded at a dilution of 1:3-1:8. Medium was changed every 2-3 days. For experiments, cells were plated onto poly-d-lysine-coated (50 g/ml) 96 well, black-walled, transparent bottomed plat es. All experiments were initiated when cells reached confluency, which was usua lly after an overnight incubation. Functional Assays The Flexstation is a fluorometer system (Molecular Devices, Sunnyvale, CA) that allows high throughput screening of a large number of compounds. Its purpose is to assess the agonist or antagonist properties of the 3-benzyl idene anabaseine series of compounds by measuring changes in fluorescence (either changes in intacellular calcium concentration or changes in membrane poten tial). The Flexstation experiments were performed according to Fitch et al. (2003). The membrane potential assay was used for the muscle expressing cell line (TE671). When using cells for the Flexstation, 75 cm2 flasks of cells which had reached confluency were harvested usi ng 0.25% trypsin and diluted with 15 ml of their respective medium. After dilution, the cells were then added to a 96 well, black-walled, transparent bottomed poly-d-lysine coated (50 g/ml) plate at 75 l of diluted cell medium per well. These cells were then allowed to attach and grow to confluency overnight in a 37 C humidified atmosphe re containing 5% CO2. The next day, the medium was removed from the cells and the cells were washed with 100 l per well of a 20mM HEPES/1 Hanks Buffered Salt Solution (HBSS) at a pH of 7.4 in orde r to ensure all medium had been removed from the cells. On e bottle of the membrane potential dye was diluted in 36 ml of the same HEPES/HBSS so lution. Next, the cells were incubated for


25 45 minutes with 30 l of the diluted molecular devices membrane potential dye. While the cells were being incubated in the dye, th e compound plates were prepared using 96 well V-bottomed transparent plates. A 0.1M stock solution of the 3-benzylidene anabaseine compounds was made in either MeOH for free bases or HEPES/HBSS for salts. Six 10fold dilutions were made with this stock so lution and these diluti ons were added to the compound plate with final concentrations usually ranging from 0.001M to 3.9 nM. Concentration ranges were adjusted occasiona lly to obtain a full concentration-response curve. Also, in the compound plate a 3 calibrant solution was added in order to fully stimulate the cell and get the largest change in fluorescence. This cal ibrant is necessary in order to account for any discrepancies in dye loading or cell number in each well. In the TE-671 cells a 3 KCl (160mM) solution was added to have a final concentration of ~54 mM in each well of the cell plate. Finally, the IC50 was also determined by adding a 2 nicotine (300 M) solution to each well after addi tion of the possible agonist. Once the compound plate was finished and the dye ha d incubated for its allotted time, the cell plate was placed into the Flexstation and read at excitation wa velength of 465nm and emission wavelength of 545nm for the memb rane potential dye. The Softmax Pro software on the Flexstation and a template obt ained from Robert Fitch was used to read the change in fluorescence on the cell plate. The fluorometer system will read a basal change in fluorescence of a co lumn of wells in the cell plate for 16 seconds followed by addition of 30 l of the 3-benzylidene anabasei ne compound, followed by addition of 30 l of the 2 nicotine, and finally by 30 l of the 3 calibrant. The final change in fluorescence or the response to the adde d 3-benzylidene anabaseine compound was calculated as the change in fluorescence fo llowed by the addition of compound minus the


26 basal fluorescence divided by the change in fluorescence after th e addition of the calibrant. In every compound plate a standard was used and then three 3-benzylidene anabaseines. The standard compound was epibatidine, and the responses to the 3benzylidene anabaseines were normalized to the largest change in fluorescence upon addition of epibatidine. The result s were analyzed using Graphpad Prism . The calcium assay was used for the 7 expressing cell lines (GH4C1 and SH-EP1). The same procedure was used for the calcium assay however, the GH4C1 cell line grew up at a slower rate than the other cell lin es and therefore once the cells had reached confluency in 75cm2 flasks they were harvested and d iluted using 7 ml of its respective medium. One bottle of the calcium dye was diluted in 15 ml of a HEPES/HBSS solution. The cells were incubated with 30 l of the calcium dye for one hour. In the compound plates the calibrant first used for the 7 cell lines was an iono mycin, carbachol, FCCP solution, but because of the cost a 2% Triton X-100 solution was eventually used. Once the compound plate was finished and the dye ha d incubated for its allotted time, the cell plate was placed into the Flexstation and r ead at excitation wavelength of 465nm and an emission wavelength of 525nm. Statistics GraphPad Prism (San Diego,CA)was used for a ll of the statistics in this dissertation. Linear regression was used to find the correlation between two parameters an r2 value was determined. The r2 value is a measure of the goodness-of-fit of linear regression. A r2 value of 0.0 means that knowing X will not help to predict Y. When r2 equals 1.0, all points lie on a straight line. In this case knowing X lets you predict Y. One-way ANOVA was also performed in th is dissertation. One-way ANOVA was used


27 to compare three or more groups when the data was categorized in one way. Through this test a P value is determined. This P valu e shows if the data is significantly different. A P value that is less than 0.05 m eans that the data is significan t. If there is a significant difference a post test is performed after the one-way ANOVA. This post test, Tukey, compares the largest group mean with the smallest. All KI values, Log P values, and pKa values were reported as the MEAN SE.


28 CHAPTER 3 COMPARISON OF THE AFFINITY OF EACH 3-BENZYLIDENE ANABASEINE FOR RAT BRAIN 4 2 AND 7 NACHRS Introduction The affinity of the 3-benzylidene anabaseines for the 4 2 and the 7 nAChR subtypes was investigated because these two subtypes are the most prominent nAChRs in the brain and they have been implicated in many CNS disorders. Also, there is evidence that stimulating either recep tor enhances cognition and inhi biting either reduces cognitive function (Rueter et al., 2004; Meyer et al ., 1997; Levin et al., 2002) . Therefore these receptors are potential therapeutic targets. Since the 3-benzylidene anabaseines generally stimulate 7Â’s but inhibit 4 2s, compounds with less affinity for 4 2 would be expected to have a greater enhancement of cognition. Therefore these receptors are potential therapeutic targets. Alpha-bungarotoxin, a large polypeptid e snake toxin (experimental mass 7983.75 Da; Mebs et al., 1972), binds with high affinity to assembled 7 receptors. This toxin will also bind to nA ChRs containing 1, 8, 9, and 10 subunits; however, these subunits are not expressed in the mammalian CNS (Jone s et al., 2003). In 1982 Marks and Collins concluded from binding studies in rat brain membranes that there were two different kinds of receptors labeled by acetylcholin e and the snake toxin alpha-bungarotoxin( BTX). Using chick cerebellum the Clemen ti lab demonstrated with the use of monoclonal antibodies specific for the alpha7 subunit that -BTX labeled the alpha7 subtype and that the liga nd binding subunit was present in vivo in two copies per receptor


29 (Gotti et al., 1992). Since then it has been s hown that the toxin binds at the interface of two receptor subunits, wraps itself around the rece ptor binding site loop and that it inserts a “finger” into the ligand binding site whic h blocks access to the ACh binding site. This blocking action explains its strong antagonistic activity (Harel et al., 2001). The 4 2 nAChR subtype accounts for greater than 90% of the binding to brain membranes of [3H]cytisine, a plant alkaloid. It was shown that only anti sera generated against the alpha4 and beta2 subunits were able to immunoprecipitate r eceptors labeled by [3H]cytisine (Flores et al., 1992). Since then [3H]cytisine was determined to be a partial agonist for the beta2-containing ACh receptors. It has a high affinity but low efficacy at 4 2 receptors and a low affinity and efficacy at 3 2 receptors (Papke and Heinemann, 1994). [3H]cytisine binding in rat brain homogenates is higher in the thalamus, striatum, and cortex than in the hippocampus, cerebellu m, or hypothalamus (Pabreza et al., 1991). Localization of [3H]cytisine and [125I]alpha-bungarotoxin binding sites were recently determined in the brains of rhesus monkeys by receptor autoradiography. The [3H]cytisine binding correlated well with previous studies; however, there was a wider distribution of bungarotoxin bindi ng sites in primates, sugge sting that these receptors have a more important role in this species compared to rodents (Han et al., 2003). One would like to directly compare the affinities of a particular compound for identical states of the two receptors. Th is would make it easier to determine the selectivity of a compound for a particular r eceptor subtype since the receptors would be in the same conformation. However, the conf ormational states of the nAChRs in which [125I]alpha-bungarotoxin and [3H]cytisine bind with high affinity are unlikely to be the same. For example, it has been shown that -BTX will shift the conformations of a


30 population of nAChRs from a mixture of resti ng and desensitized entirely into the resting state (Moore and McCarthy, 1995). As for [3H]cytisine, the confor mation state in which this radioligand binds with the highest affi nity is still under debate. However, in radioligand binding assays where there is pr olonged incubation with an agonist like cytisine, the 4 2 nAChR will be converted into a high affinity desensitized state. This high affinity desensitized state measured in binding experiments for an agonist will be a value different from the agonistÂ’s EC50 value for activation. The KI value will be lower than the EC50 value for activation. Generally, thes e experimental binding affinities are typically 2-3 orders of ma gnitude higher (lower KI) than EC50 values for activation (Sharples and Wonnacott, 2001). As for [125I]-alpha bungarotoxin determined binding affinities since the radioligand prefer entially binds to the resting state, 7 agonists may have higher KI values (lower affinity) compared to a displacing radioligand that binds to an activated or desensitized state. However, agonists such as nico tine and cytisine do not bind to the 7 receptor with high affinity. Also, fo r activation in the presence of high agonist concentrations the alpha7-type recep tor has not been observed to have a high affinity desensitized state (Zhang et al ., 1994). Cytisine was used in the following experiments because it is an establishe d radioligand that reliably labels the 4 2 nAChR with high affinity. The KI value obtained for a displacing ligand is the steady state concentration of competing ligand which would occupy 50% of the receptors if no radioligand was present. The KI value is used instead of the IC50 value because the IC50 value for a compound may vary between experiments depe nding on the concentration of radioligand


31 used. Although the KI is independent of the concentrati on of the radioligand, it will still be affected by other factors such as bindi ng saline composition and possibly the pH. Results The substituents added to the benzylid ene ring were choosen based on the Craig Plot discussed in Chapter 1 of this dissertation. When designing a compound and choosing what substituent needs to be a dded at a particular position, the predicted affinities of these newly created 3-benzy lidene anabaseines might depend on the two parameters of this plot. The relationship be tween affinity of 3-benzylidene anabaseines and the Hammett and Hansch values needs to be addressed in order to fully understand how a substituent influences the ch emical properties of a compound. Since the published Hammett sigma constants were predicted based on the ionization constants of substituted aroma tic compounds, a comparison of the published values for particular para and meta substitutents on the 3-benzylidene anabaseines and their experimentally derived imine nitrogen ionization constants (pKa) was performed. A direct relationship between the imine pKas of the 3-benzylidene anabaseines and published values was found (Figure 3-1). The mo re negative the sigma value or the more electron donating the substituent the higher the pKa of the imine nitrogen of the 3benzylidene anabaseine. Based on this relationship between the values to the ionization constants of the 3-benz ylidene anabaseines, calculated values were obtained using the Hammett equation discussed previously (Figure 3-2). Calculated sigma values were obtained for single para, meta, and fu sed ring substituents on the 3-benzylidene anabaseines used in this study based on their experimental pKas. The ortho substituted, ortho/para substituted and the trisubstituted benzylidene anabaseines were not calculated


32 because of steric effects of th e ortho substituent with the rest of the molecule and because these values are representative values fo r a single substituent. Calculating a value for a dior trisubstituted compound would not be appropriate (Table 3-1). Considering the structure of the 3-benzylidene anabaseines, the ortho substituent could sterically affect the orientations of the pyridyl ring as well as the tetrahydropyridyl ring and because of this interaction affect th e binding properties of this derivative with the receptor. This steric in terference could make the struct ure more rigid in the binding pocket. Substitutions in the ortho position could contribute large steric effects because the 3-benzylidene anabaseine structure has been shown to exist in a twisted conformation unlike its parent compound (anabaseine) which is coplanar. The calculated values were then compared to the affinities determined experimentally for the alpha7 and the al pha4beta2 nAChR using radioligand binding experiments. A comparison of the sigma values versus the KI values shows a trend: benzylidene anabaseines with more electrondonating substituents tend to have a higher affinity for the receptor (Figure 3-3, 3-4). Besides using the Hammett constants or the electronic influence of the substituent as a guide for what types of com pounds to synthesize, the Hansch ( ) constants are another substituent property that wa s scrutinized. The more negative the value, the more hydrophilic the substituent; conversely, the more positive this value the more lipophilic the substituent. Since the Hansch ( ) constants were originally derived from experimental Log P values, a comparison of the calculated values, obtained using the Hansch-Fujita equation discussed in chapter 1 of this dissertation, for every derivative except for the dior trisubstituted 3-benzylidene anabaseines and their experimentally


33 derived Log P values was performe d (Figure 3-5). These calculated values were then compared to the KI values of the 3-benzylidene an abaseines for the alpha7 and the alpha4beta2 nAChRs obtained through bindi ng studies (Figure 3-6, 3-7). This comparison gave no trend between these two pa rameters for either the alpha7 or the alpha4beta2 nAChRs. The 3-benzylidene anabaseineÂ’s binding affinities for the 7 and the 4 2 nAChRs in rat brain membrane homogenate, listed in the appendix and shown graphically in Figures 3-8, 3-9 and 3-10 show that subs titution at the para, ortho, meta positions separately and substitution at both the para and ortho, as we ll as simultaneous substitution at the para, ortho and meta positions does not prevent the 3-benzylidene anabaseine from binding to these receptors. The unsubstituted 3-benzylidene anabaseine has a selectivity, the ratio of the KI values for 4 2/ 7 nAChRs, of 1.4. This value show s that the unsubstituted compound has a higher affinity for the 7 subtype. Compared to the other derivatives explored twenty out of the 47 compounds have an im proved selectivity for the alpha7 nAChR. Every group (para, para/ortho, or tho, meta and fused ring) in the appendix has at least one compound which has an improved selectivity for the 7 nAChR compared to the unsubstituted benzylidene anabaseine. Ho wever, the trisubstituted benzylidene anabaseines do not have a compound out of th is group that has an improved selectivity for the alpha7 nAChR. In fact a majority of these compounds have a higher affinity for the 4 2 nAChR with selectivities under 1. In general the ortho and the para positions were expected to be of most interest because they are more effec tive at altering the imine pKa since they act by a resonance


34 mechanism. The meta position is much less he lpful electronically in affecting the imine pKa. However, two meta 3-benzylidene anabaseines were explored. These two meta derivatives, 3-[3-Methoxybe nzylidene]-anaba seine and 3-[3-Hydroxybenzylidene]anabaseine have similar calculated Hamm ett sigma values of 0.05 for the methoxy compound and 0.19 for the hydroxy compound. Th eir alpha7 binding affinities were 3 fold different, 0.46 0.12 M for the methoxy compound and 0.15 0.02 M for the hydroxy compound. This difference in affinity could possibly be due to electrons being more available for hydrogen bonds in the 3[3-Hydroxybenzylidene]-a nabaseine due to the less bulky hydrogen group. This group would be less disruptive to the interaction between other hydrogens and the valence elect rons available on the oxygen compared to the bulkier methyl group. Alternatively, a H-bond donor is more optimal than a H-bond acceptor. Other 3-benzylidene anabasei nes that have the same pKaÂ’s and thus calculated Hammett values are 3-[4-(Dimethylamino)be nzylidene]-anabaseine and 3-{4-[bis(2Hydroxyethyl)amino] benzylidene}-anabaseine ( = -1.32). These compounds both have high affinities for the 7 receptor. The Dimethylamino compound has a slightly higher affinity of 0.08 0.0004 M while the bis(2-Hydroxyethyl)amino compound has an affinity of 0.19 0.04 M. However, when comparing two other compounds with approximately the same value, 3-[(Benzofuran-2-yl)methylene]-anabaseine and 3-[4Pivaloyloxybenzylidene]-anabaseine ( = 0.23) their 7 binding affinities are very different. The (Benzofuran-2-yl)methylene compound has a KI value of 0.15 0.003 M and the 4-Pivaloyloxy compound has a KI value of 0.69 0.11 M. Even though these compounds have similar values there is a 5 fold difference in affinity. The Hammett


35 values have been applied with success in a va riety of physiochemical studies. However, the original Hammett parameter is limited and problems in accurate assessment of electronic influence of a subs tituent can occur when there is a resonance interaction between the substituent and th e reaction center, or the cont ributions of inductive and mesomeric effects are not constant within th e series of compounds or when there are steric interactions between the substituen t and the reaction center as with the orthosubstituted compounds. The diortho/para or dimeta/para substitutions were produced to see if trisubstituted benzylidene anabaseines are even allowed in the binding pocket and to determine the influence of both ortho positions being substi tuted as apposed to both meta positions. As stated earlier these compounds were the most unselective compared to the unsubstituted benzylidene anabaseine for the alpha7 nAChR. There is a 10 fold diffe rence in binding of the 3-[2,6 Dimethyl-4-hydroxybenz ylidene]-anabaseine (5.29 2.97 M) compared to its dimeta equivalent 3-[3,5-Dimethyl-4 -hydroxybenzylidene]-anabaseine (0.53 2.97 M). This would make sense when taking into c onsideration the steric effects of the ortho positions on the rest of the molecule. Also, when comparing the diortho 3-[2,4,6Trimethylbenzylidene]-anabasein e to these two previous st ructures this compound has a higher affinity for the alpha7 (1.51 0.11 M) and alpha4beta2 (0.52 0.22 M) receptors than the 3-[2,6 Dimethyl-4-hydr oxybenzylidene]-anabaseine but does not have a higher affinity than the dimeta 3-[3,5Dimethyl-4-hydroxybenzylidene]-anabaseine, again showing that the diortho interaction might be adversely interacting with the rest of the molecule. However, the 3-[2,4,6-Trimethoxybenzylidene]-anabaseine has a much higher affinity for both the alpha7 (0.18 0.03 M) and the alpha4beta2 (0.06 0.03


36 M) receptors as compared to its dimeta equivalent, 3-[3,4,5-Trimethoxy benzylidene]anabaseine (alpha7/3.48 2.10 M and alpha4beta2/4.67 1.06 M). Apparently, two methoxy ortho substituents interfere less with ligand ring orientati ons than two methyl ortho substituents. Finally, the fused ring structures were de veloped in hopes of determining whether or not such a large structure on the benzy lidene ring would be a llowed in the binding pocket of the alpha7 or the alpha4beta2 receptors. These structures for the most part have a similar binding affinity to the alpha7 receptor with the 3-[(Benzofuran-2-yl)methylene]anabaseine having a KI of 0.15 0.003 M for the 7 nAChR and the best selectivity (39.5) of the entire group of 47 compounds i nvestigated. The four compounds in this group compared to the other 3-benzylidene anabaseine groups disp lay the best alpha7 versus alpha4beta2 nAChR selectivity. The lack of interference of these bulky substituents on benzylidene anabaseine bindi ng leads one to conclude that there is unexploited space in the receptor binding poc ket that, when occupied optimally, can enhance the selectivity of a 3-benzylidene derivative for the alpha7 nAChR. Discussion Based on the presented data the Hammett value seems in many cases to be correlated with the receptor affinity of the compound. Thus, selection of substituents with electron-donating values should lead to more pot ent compounds. The influence of the Hammett values are specific to the set of compounds and the receptor one is analyzing. For example, in 1998 a quantitative struct ure-activity relationship study was performed on 19 pyridine nicotine analogues. Th is group looked at the electronic ( ) and the lipophilic ( ) nature of the 6-position substituent and its influence on nAChR affinity.


37 They determined that affinity seems to be related to the lipophili city and is further influenced by the size of the 6position substituent. In cont rast to this dissertation, the electronic influence of the substituent in this group’s series of nicoti nic ligands appears to play a less significant role on the 6-positi on of the pyridine ring (Dukat et al., 1998). The Hansch-Fujita values do not seem to aid in pr edicting receptor affinity for the 3-benzylidene anabaseines. The influence of the Hansch-Fujita values are specific to the set of compounds and receptor on e is analyzing. As an exam ple of a previous HanschFujita study, in 1996 a quantitative structure activity relationship investigation of 34 pyrrolidine-modified nicotine agonists was compiled. This group looked at a number of physicochemical parameters (ex. Molar refr activity, hydrogen bond accepting property of the substituent) including the hydrophobic effects of the substituents on the binding affinity using literature values, but this variable was found to have a negligible influence (Kim et al., 1996). There are many properties that contribut e to the overall “personality” of a compound and when comparing the compound KI values to assess the electronic or lipophilic influence of a substi tuent, one should keep in mi nd that even though a property can be separated “on paper” for comparison purposes, other chemical properties like Hbonding ability and size of the substituent are also influencing the activity of a compound. The meta substitution did not abolish the affinity of the 3-benzylidene anabaseine for the 7 or the 4 2 nAChR. When looking at novel 4 2 neuronal nicotinic AChR ligands like the 2-(3-pyridyloxy-methyl)a zacyclic compounds, electron-withdrawing groups at the meta position improve KI values, whereas substituti on at the ortho and para


38 positions is generally unfavorable (Gotti et al .,1998). The data in this dissertation shows that the meta substitutions, even though ther e are only two derivatives in this group, have a high selectivity for the 7 receptors. Also, the differences in 7 affinities for the two meta-substituted 3-benzylidene anabaseine compounds might not be due to electronic influence of the substituent alone. Other meta-substituted benzylidene anabaseines should be tested to assess which prop erties are important at this site. The trisubstitued 3-benzylidene anabaseines appeared for the most part to have a higher affinity for the 4 2 nAChR as compared to the 7 receptor. This observation might provide insight into the 4 2 pharmacophore. Perhaps there is more room for the diortho or the dimeta substituted benzylidene anabaseine in the 4 2 binding pocket. Other trisubstituted compounds from the literature include a frog skin alkaloid, 5,6,8 trisubstituted indolizidine which was found to block both 4 2 and 7 nAChR-mediated currents in Xenopus oocytes in a concentration dependent manner with IC50 values of about 4 M (Tsuneki et al., 2004). Finally, the most promising substitution in this series of compounds seems to be the fused ring substitution. A definitive knowle dge of the affinity of a fused substitution on a 3-benzylidene anabaseine compound can not be extracted from the literature because this series of compounds is different in stru cture and in its target receptor. However, another group looked at bor on-containing nicotine analogue s which had a fused ring structure. These analogues were somewhat successful with one analog, ACME-B, having selectivity for the 7 nAChR (Xu et al., 2001). In the da ta presented in this dissertation the 3-[(Benzofuran-2-yl)methylene]-anabaseine has the best selectiv ity (39.5) out of all


39 of the 3-benzylidene anabaseine derivatives and every derivative from the fused ring group has a selectivity of 1.4 or higher (Figure 3-10). According to Papke et al. (2004a), co mpared to other nAChR subtypes the 7 homomeric receptor has a relatively relaxed requirement for activation via the agonist binding site. This conclusion was reached b ecause the large 3-benzylidene anabaseines are more efficacious for the 7 receptors and 7 receptors are more receptive to the ACh precursor, choline. Also, this group goes on to show that the activation of subunitcontaining nAChRs is largely eliminated with the addition of a hydroxyl group to ethyltrimethylammonium forming choline. This suggests that this hydroxyl group interacts with residues in the agonist binding site of -subunit containing receptors in such a way as to stabilize the resting state. Also, Gotti et al. (1998) investigated five 4oxystilbene derivatives and using binding assa ys determined that an N,N,N-triethyl ammonium moiety had a lower KI for the 7 receptor compared to an N,N,N-trimethyl ammonium moiety which was possibly due to be tter exploitation of the binding pocket as a result of the ammonium volume. These c onclusions support the idea that there is unexploited space in the binding pocket of the 7 receptor and that large 3-benzylidene anabaseine compounds like the 3-[(Benzof uran-2-yl)methylene]-anabaseine take advantage of the 7 nAChRs more relaxed binding requirements.


40 Figure 3-1: Direct re lationship between the published Hammett values for only the meta and para substitutions, represen ting the electronic influence of a substituent vs. experimentally derived pKa values determined for the imine nitrogen of each 3-benzylidene anabasei ne. The data was fitted by using linear regression analysis. Figure 3-2: Direct re lationship between the calculated Hammett values , representing the electronic influence of a substituent vs. experimentally derived pKa values determined for the imine nitrogen of th e 3-benzylidene anabaseine derivatives excluding the derivatives w ith an ortho substituent and the derivatives with multiple substituents. 6.0 6.5 7.0 7.5 8.0 8.5 9.0 -1.0 -0.5 0.0 0.5 1.0Experimental pKaPublished Hammett values 6.0 6.5 7.0 7.5 8.0 8.5 9.0 -2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0Experimental pKaCalculated Hammett values


41 Figure 3-3: Experimentally determined KI values for the alpha7 nAChR vs. the calculated Hammett values . All of the 3-benzylidene anabaseine compounds from Table 1 are represented on the graph, except for the compounds which included an ortho substituent and the compounds which had multiple substituents. -1.5 -1.0 -0.5 0.0 0.5 1.0 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0Calculated Hammett values KI ( M) alpha4beta2 RBM Figure 3-4: Experimentally determined KI values for the alpha4beta2 nAChR vs. the calculated Hammett values . All of the 3-benzylidene anabaseine compounds from Table 1 are represented on the graph, except for the compounds which included an ortho substituent and the compounds which had multiple substituents. -1.5 -1.0 -0.5 0.0 0.5 1.0 0.0 0.5 1.0 1.5 2.0 2.5 3.0Calculated Hammett values KI ( M) alpha7 RBM


42 Figure 3-5: Direct relationship between the calc ulated Hansch-Fujita value and the experimental Log P values determined for each 3-benzylidene anabaseine derivative. The values represent a single subs tituentÂ’s influence so the 3benzylidene anabaseines with multiple substituents were not included. Figure 3-6: Experimentally determined KI values for the alpha7 nAChR vs. the calculated Hansch-Fujita value . All 3-benzylidene an abaseine derivatives were represented except for derivatives with multiple substituents. -1.0 -0.5 0.0 0.5 1.0 0 1 2 3 4Hansch-Fujita calculated value KI ( M) alpha7 RBM 2.0 2.5 3.0 3.5 4.0 4.5 -0.2 -0.1 0.0 0.1Experimental Log PHansch-Fujita calculated value


43 -1.0 -0.5 0.0 0.5 1.0 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0Hansch-Fujita calculated value KI ( M) alpha4beta2 RBM Figure 3-7: Experimentally determined KI values for the alpha4beta2 nAChR vs. the calculated Hansch-Fujita value . All 3-benzylidene an abaseine derivatives were represented except for derivatives with multiple substituents.


44 Figure 3-8: Comparison of th e alpha7(white bars) and al pha4beta2 (solid bars) KI values ( M) for the para substituted 3-benzylidene anabaseines. Experimentally determined KI ( M)


45 Figure 3-9: Comparison of the alpha7 (white bars) and alpha4beta2 (solid bars) KI values ( M) for the ortho, meta, trisubstituted, para/ortho substitu ted, fused ring, and unsubstituted 3-benzyl idene anabaseines. Experimentally determined KI ( M)


46 Figure 3-10: Comparison of the alpha7 (wh ite bars) and alpha4be ta2 (solid bars) KI values ( M) for the compounds with selec tivity above 1.4 for the alpha7 nAChR. Experimentally determined KI ( M)


47 Table 3-1: Calculated Hammett and Hansch-Fujita values for the 3-benzylidene anabaseines Compound Name Hammett value Hansch-Fujita value 3-[4-Aminobenzylidene]anabaseine -0.99 -0.78 3-[4-Butoxybenzylidene]anabaseine -0.11 3-[4-Trifluoro methoxybenzylidene]-anabaseine 0.50 -0.77 3-[4-(Methylamino)benzylidene]anabaseine -1.20 -0.11 3-[4-Hydroxybenzylidene]anabaseine 0.12 -0.93 3-[4-Methylthiobenzylidene]anabaseine -0.15 -0.05 3-[4-Cyanobenzylidene]anabaseine 0.62 -0.11 3-[4-Propylbenzylidene]anabaseine -0.05 -0.40 3-[4-Acetoxybenzylidene]anabaseine 0.20 -0.65 3-[4-Morpholinobenzylidene]anabaseine -0.63 -0.50 3-[4-Dimethylaminobenzylidene]anabaseine -1.32 0.42 3-[4-Difluoromethoxybenzylidene]-anabaseine -0.01 0.18 3-[4-Chlorobenzylidene]anabaseine 0.16 0.28 3-[4-(1H-1,2,4-Triazol-1yl)benzylidene]-anabaseine 0.28 -0.81 3-[4-Acetamidobenzylidene]anabaseine -0.08 -0.85 3-{4-[bis(2Hydroxyethyl)amino]benzylidene}anabaseine -1.32 -0.65 3-[4-Pivaloyloxybenzylidene]anabaseine 0.22 -0.12 3-[4-Acetylbenzylidene]anabaseine 0.52 -0.70 3-[4-n-Butoxybenzylidene]anabaseine -0.32 -0.53 Compound Name Hammett value Hansch-Fujita value


48 3-[2-Trifluoromethoxybenzylidene]-anabaseine -0.71 3-[2-Difluoromethoxybenzylidene]-anabaseine -0.14 3-[2-Pivaloyloxybenzylidene]anabaseine 0.13 3-[3-Methoxybenzylidene]anabaseine 0.05 -0.04 3-[3-Hydroxy benzylidene]anabaseine 0.19 -0.49 3-[3,4-Ethylenedioxybenzylidene]anabaseine -0.21 -0.06 3-[3,4Methylenedioxybenzylidene]anabaseine -0.29 -0.21 3-[(6-Methoxynaphth-2yl)methylene]-anabaseine 0.06 0.17 3-[(Benzofuran-2-yl)methylene]anabaseine 0.23 0.65


49 CHAPTER 4 THE INFLUENCE OF THE IONIZATI ON STATE OF THE LIGAND ON ITS AFFINITY FOR THE RECEPTOR Introduction The ionizability of a drug has an important role in controlling its absorption, distribution within the body and its binding to its receptor. Ther efore the ionization constant, also referred to as the equilibrium dissociation constant or the acidity constant (pKa), is one of the most important charac teristics of an ionizable drug. The pKa is a numeric estimate of the likel ihood that an ionizable group will donate a proton when a drug is dissolved in water. The pKa of a chemical group is th e pH at which the ionized concentration of the compound equals the unionized concentration of the compound. A compound may potentially have several ionizable groups will different pKas. For example, all of the 3-benzylidene anabaseine s have an ionizable nitrogen on the pyridyl ring (pKa~3) as well as an ionizable imine on the tetrahydropyridyl ring. In addition, some contain other potentially io nizable substituents (-OH, -NH2, etc) on the benzylidene ring. Under physiological conditi ons (pH 7.4), the higher the pKa of a basic compound like a 3-benzylidene anabaseine, the more ionized the imine group on the tetrahydropyridyl ring, as predicted by th e Henderson-Hasselbach equation. If a compound is charged it is more water-soluble an d therefore it is less likely to passively diffuse across gastrointestinal membranes, the blood brain barrier (BBB) and placenta. If a compound is unionized it is more lipid sol uble and therefore more likely that it will passively diffuse across these membrane barr iers. The unionized form may still bind to


50 the receptor by hydrogen bonding or other chemical interactions with the receptor while the ionized form may enhance electrostatic in teractions with the receptor. Ion-dipole interactions between the ionized drug and th e receptor occur because the ionized drug has an asymmetric distribution of electrons and this produces electronic dipoles. The dipole in a compound can be attracted by ionizable grou ps in the receptor, provided that charges of opposite sign are properly aligned. This i on-dipole interaction is a stronger interaction than hydrogen bonding. Nevertheless, hydr ogen bonding can be quite important for biological activity, partly because there are usually more hydrogen bonding groups than ionizable groups. This type of bond is a dipole -dipole interaction in which the dipoles in a drug are attracted by dipoles in the receptor (Silverman, 1992). Also, ionization state can affect the me tabolism of a drug. Ideally a drug is metabolized appropriately, not too quickly bu t soon enough to excrete the drug when it is necessary, and this property can be influen ced by its ionization state and polarity. For example, if a drug is very lipophilic (neutral compound) it wi ll readily penetrate liver and other cells containing enzymes catalyzing its me tabolism. It will be rapidly absorbed and will likely have a large apparent volume of distribution. One of the most desirable characteristics of a drug is its ability to be water-soluble. The compartments of the body through which a drug must pass to reach its receptor are usually aqueous. Since ionization enhances water solubility a drug that is at least partially ionized would be ideal, at least fo r an otherwise water insoluble drug. However, the ionized form does not diffuse freely across me mbranes. For this reason it appears that the “best” basic drug to passively diffuse across the BBB would be one that has a low enough pKa such that some of it will be in a neutral form at physiological pH. Just


51 because a basic drug has a high pKa does not completely remove its chances of crossing the BBB. There is always an ongoing rapid equilibrium between th e unionized and the ionized forms of the drug; as long as some pe rmeating form is present then entry occurs even if at a slower rate. Albert and coworkers found, in vitro , that there was a correlation between ionization and biological activity for certain antibacterial drugs. They tested 101 aminoacridines, all having a variety of pKa values, against 22 species of bacteria and found a direct association between formation of the cation of the aminoacridines and antibacterial activity (Albert, 1951). Little is known about the pH sensitivity of the neuronal nAChRs as compared to the many reports about the muscle and electric organ nicotinic receptors. For example Barlow and Hamilton (1962) presented the pos sibility that the ionized form of a compound was the only species of the compound th at was active. They studied the effects of pH on the relative activities of nico tine and nicotine monomethiodide on a rat diaphragm (skeletal muscle) preparation. Th ey understood that their results could be altered by direct effects of pH on their tissue preparation and took this into consideration when performing their experiments. Their ove rall conclusion was that the nicotinium ion was the active species at the neuromuscular junction; however, they found some evidence to suggest that the unionized base might have some effect. They could not definitively conclude that the unionized form was involved because thei r rat diaphragm preparation was too sensitive to changes in pH. In 1980 Jeng and Cohen investigated the structural elements of cholinergic ligands that were esse ntial for a ligand to act as an agonist. At that time it was assumed that the presence of a positive charge was essential, however


52 both charged and potentially uncharged analogs of acetylcholine had been reported to act as nicotinic agonists in electr ophysiological studies. Using 22Na + efflux experiments on Torpedo electric tissue, they found that positive ch arge is essential fo r a ligand to act as an agonist at this neuromuscular type ni cotinic receptor (Cohen and Jeng, 1980). In a recent study, a constitutively active muscle receptor was used which contained amine agonist groups tethered to tyrosyl side chai ns in the ACh binding si te. If the protonated species of the ligand is the active form, then th is tethered receptor will be constitutively active only when the tethered amine group is protonated. This lab concluded that only the cationic form of the agonist would act ivate the receptor (Petersson et al., 2002). Finally in another recent study, 3/ 4 recombinant nAChRs expressed in human embryonic kidney 293 cells were subjected to rapid acidification from pH 7.4 to 6.0, and as a result displayed an enhanced current. The acidification accelerated both receptor activation and desensitization; with cytisine and nicotine evoked currents the greatest effect was upon desensitization kinetics. Howeve r, this group did not consider that both cytisine and nicotine are not to tally ionized at pH 7.4 so ther e may be a pH effect on them as well. They went on to state that increas ed proton concentration during desensitization enhanced the current and mimicked the reactivation of curre nt by increasing the concentration of agonist. This paper sugge sts that protons interact with multiple extracellular sites on the recep tor promoting agonist-induced block and that changes in pH at the synaptic cleft might play a vital role in the regulation of synaptic activity (Abdrakhmanova et al., 2002). It is important to determine if the cati onic form of a benzylidene anabaseine, like DMXBA, is the only form which binds effec tively to nAChRs. This is a necessary


53 question to answer because at physiological pH the various benzyidene anabaseine compounds could be in a more cationic or neutral form, depending on their pKas, and this could make a difference in the affinity of the compounds for the nAChR. Results To assess the importance of ionization fo r benzylidene anabaseine action it is necessary to know the imine pKa values for the compounds studied. These pKa values could not be predicted accura tely using the SciQSAR softwa re so it was necessary to determine them experimentally. Using the spectrophotometric tec hnique described in Chapter 2, the ionization constants or pKas were determined for the imine nitrogen of each 3-benzylidene anabaseine derivative. The compounds expresse d a wide array of pKa values which are presented in the a ppendix, ranging from 5.76 to 8.66. Values were relatively precise with all of th e standard errors less than or equal to 0.17. In Figures 4-1 and 4-2 the pKa values are plotted against the log KI values and in Figures 4-3 and 4-4 the pKa values are plotted against the ionized log KI values (corrected for percent ionization at physiological pH). The pKa vs. the log KI graphs suggest a relationship between these two parameters (r2 values, 0.4442 and 0.2785). The higher the value for the pKa the higher the affinity that the compound often has for its respective receptor, alpha7 or alpha4beta2. The pKa vs. log KI (ionized) graphs (Fi gures 4-3, 4-4) show th at after correcting for ionization, that relationship is no longer apparent (r2 values, 0.00406 and 0.0169). Correcting the KI values for the percent ionized at physiological pH makes the pKa value of the compound irrelevant. If the ionized form of the comp ound is the only form which is active the relationship between the pKa and the log KI(ionized) should be a straight horizontal line in Figures 4-3 and 4-4. These graphs show that the ionized form of the


54 compound is important to binding. To further explore the effect of ionization on the interaction of the ligand with the receptor, radioligand binding studies were performed using DMXBA, whose imine nitrogen pKa is 7.62. In these experiments it was shown that at a pH of 6.6 where 91.3% of the compound was found in the ionized form, DMXBA had the highest affinity for its receptor compared to the 7.4 (62.4 %) and 8.2 (20.8%) radioligand binding experime nts (Figure 4-5/ Table 4-1). However, by performing these radioligand bi nding studies at such a varied range in pH, these pH changes could be affecting the ionization state of th e amino acids in the binding pocket of the receptor or the ionization state of the radioligand, -BTX. If the pH is altering the ionization states of th ese amino acids, then the seemingly enhanced affinity might not be due to only a change in the ionization state of DMXBA. However, no ionizable side chains in th e immediate site of the bind ing pocket have so far been identified. Also, to ensure that the change in pH was not affecting the ionization state of the radioligand, [125I]alpha-bungarotoxin, Scatchard plots for binding of this toxin were performed at the three different pHs. By pe rforming Scatchard plots it can be determined if the affinity of this radioligand for the alpha7 receptor changes. As is shown by Figure 4-6, Table 4-2 there was not a si gnificant difference in the Kds (6.6/0.56 ± 0.18 nM, 7.4/ 0.48 ± 0.12nM, 8.2/ 0.46 ± 0.07nM) for the three different pHs according to a one-way ANOVA (P value = 0.8249). Also, there was not a significant difference in the Bmax values for the Scatch ard plots (6.6/37.14 3.77 fmol/mg, 7.4/35.74 2.75fmol/mg, 8.2/25.73 1.22 fmol/mg). The Bmax differences observed are lik ely due to variability of receptor expression in different Sprague-Dawley rat brains.


55 To further rule out any pH complications , a permanently charged ligand was used in the radioligand binding assays at pH 6.6,7.4,and 8.2 to exclude any changes in the ionization state of the amino acids in the binding pocket. At first acetylcholine was employed in these experiments; however, the presence of acetylcholine esterases in the rat brain homogenate would affect the concen tration of ACh in these experiments. The use of an acetylcholine es terase inhibitor, physostigmine, was considered; however, certain acetylcholine esterase inhibitors act as co-agonists at low concentrations and as antagonists at high concentra tions (Zwart et al., 2000). Instead of ACh, the highly ionized anabaseine derivativ e 2-(3-pyridyl)1,4,5, 6-tetrahydropyrimidine (PTHP) was utilized. PTHP (pKa>11) is essentially always charged at the pHs studied. Thus, the only variable would be the ionizat ion states of the amino acids in the acetylcholine binding pocket. Using this permanently ch arged compound, the data at pH 6.6 (KI, 0.74 0.11 M) 7.4 (KI, 0.91 0.23 M), and 8.2 (KI, 0.84 0.25 M) did not show significant pH dependence with a P value of 0.8070 (Figure 4-7/ Table 4-3). These experiments suggest that the ionization states of the amino acids in the binding pocket of the alpha7 receptor were not altered by pH changes, or th at they compensated for each other. Finally, a 3-benzylidene an abaseine with a lower pKa of 6.72, 3-[4Cyanobenzylidene]anabaseine, was used in pH dependent binding experiments to further test whether both the charged and th e uncharged forms bound to the receptor. By using a 3-benzylidene anabaseine with a lower pKa, less of the compound would be in its ionized form at physiological pH (Figure 4-8/ Table 4-4). At pH 6.6, 57% of the compound is in its ionized form, at pH 7.4 only 17% is ionized, and at 8.2, 3.2% is ionized. At the two lower pHs, it was found that the affinity of the compound for the


56 receptor was about the same (6.6 KI, 2.58 0.22 M/ 7.4 KI 2.43 0.18 M) while at pH 8.2 there was inhibition of [125I] -BTX at a significantly higher concentration (KI 6.91 0.37 M). Discussion The ionization state of the 3-benzylidene anabaseines seem to contribute to their respective affinities for rat 7 and rat 4 2 nAChRs. The higher the pKa, the higher the affinity of the compound for its receptor (Figur es 4-1 and 4-2). However, in a structureactivity relationship study performed on 6-subs tituted nicotine derivatives as cholinergic ligands the basicity of the pyridine nitroge n atom was examined by determining the pKa values of several representative analogs. It was shown that the pKa alone did not account for variations in affinity (Duka t et al., 1998). In the presen t study, the ionized form of the compound is important to binding, however other chemical properties could be contributing (eg. hydrogen bonding). The ionized and the unionized forms of th e 3-benzylidene anabaseines appear to bind to the 7 nAChR in rat brain membrane. This data is supported by the previously referenced Barlow and Hamilton study, but a ma jority of recent public ations show that the ionized form of the compound is the activ e form at the receptor (Jeng and Cohen, 1980; Petersson et al., 2002; Abdrakhma nova et al., 2002; Nettleton et al., 1990). Therefore, our conclusions need to be inves tigated further. It should be kept in mind when considering this data that we are assumi ng that the pH at the receptor is the same as the pH of the bulk phase. However, in the published AChBP structur e there is only one water molecule in the binding site. Thus, this assumption may be incorrect. For instance,


57 if the ACh binding site is electronegative, then the local pH may be less than the bulk pH. This could account for at least some of the observed pH dependence for DMXBA. When altering the structure of a lead compound by adding an electron donating or electron withdrawing group, besides altering ionization, steric fit and hydrogen binding within the binding site might be altered as well. Therefore, expecting a direct relationship between ionization (or any one property) and the affinity of that compound to its receptor would be naïve. Thus, da ta supporting the notion that ionization state affects the binding of the drug to the target recep tor is not meant to infer that this is the only property that is altered when adding a substituent to the benzylidene ring.


58 Figure 4-1: Log KI (alpha7) for all 3-benzylidene an abaseines vs. the experimentally determined pKa. Linear regression analysis produced an r2 value of 0.4442. Figure 4-2: Log KI (alpha4beta2) for all 3-benzylidene anabaseines vs. the experimentally determined pKa. Linear regression analysis produced an r2 value of 0.2785. 5.0 5.5 6.0 6.5 7.0 7.5 8.0 8.5 9.0 -1.5 -1.0 -0.5 0.0 0.5 1.0pKaLog KI (alpha 7) M 5.0 5.5 6.0 6.5 7.0 7.5 8.0 8.5 9.0 -1.5 -1.0 -0.5 0.0 0.5 1.0pKaLog KI (alpha4beta2) M


59 Figure 4-3: Predicted Log KI (alpha7) that are corrected for the percentage unionized (determined by the Henderson-Hasselbach equation) at physiological pH for all 3-benzylidene anabaseines vs. the experimentally determined pKa. Linear regression analysis produced an r2 value of 0.00406. Figure 4-4: Predicted Log KI (alpha4beta2) that are co rrected for the percentage unionized (determined by the Hende rson-Hasselbach equation) at physiological pH for all 3-benzylidene anabaseines vs. the experimentally determined pKa. Linear regression analysis produced an r2 value of 0.0169. 5.0 5.5 6.0 6.5 7.0 7.5 8.0 8.5 9.0 9.5 10.0 -2.0 -1.5 -1.0 -0.5 0.0 0.5pKaLog KI (alpha 7 ionized, predicted values) M 5.0 5.5 6.0 6.5 7.0 7.5 8.0 8.5 9.0 9.5 10.0 -2.0 -1.5 -1.0 -0.5 0.0 0.5pKaLog KI (alpha4beta2 ionized,predicted values) M


60 Figure 4-5: Displacement curves for DMXB A at three different pHs. At pH 6.6 DMXBA has a lower KI than at 7.4 or 8.2. At pH 6.6, 91.3% of DMXBA is found in the ionized form. The more ionized DMXBA is at physiological pH the higher the affinity of the compound for the alpha7 nAChR. Table 4-1: Table for Figure 4-5 Displacement curves for DMXBA at three different pHs. The data is the mean SE. pH Average KI ( M) % Ionized KI Ionized ( M) One-way ANOVA P value = 0.0029 6.6 (n=4) 0.37 0.06 91.3 0.34 0.05 Tukey: 6.6 vs. 7.4 P>0.05 7.4 (n=4) 0.80 0.11 62.4 0.50 0.07 Tukey: 7.4 vs. 8.2 P<0.05 8.2 (n=4) 3.69 0.32 20.8 0.77 0.07 Tukey: 6.6 vs. 8.2 P<0.01 4.5 5.0 5.5 6.0 6.5 7.0 7.5 0 25 50 75 100 6.6 7.4 8.2 -Log [DMXBA], M125I-BTX Bound (% of Total)


61 Figure 4-6: Respresentative Scatchard Plots at the different pHs 6.6, 7.4, and 8.2. 0 10 20 30 40 0.0 2.5×1010 5.0×1010 7.5×1010Kd= 0.43 0.10 nM Bmax= 29.73 1.911 fmol/mg Bound (fmol/mg) MOPS 6.6Bound/Free (fmol/mg/nM) 0 10 20 30 40 50 0.0 5.0×1010 1.0×1011 1.5×1011Kd= 0.40 0.07 nM Bmax= 41.76 1.98 fmol/mg Bound (fmol/mg) MOPS 7.4Bound/Free (fmol/mg/nM) 0 10 20 30 0.0 2.5×1010 5.0×1010 7.5×1010Kd= 0.43 0.10 nM Bmax= 25.30 1.676 fmol/mg Bound (fmol/mg) MOPS 8.2Bound/Free (fmol/mg/nM)


62 Table 4-2: Table for Figure 4-6 Representative Scatchar d Plots at different pHs 6.6,7.4, and 8.2. The data is the mean SE. pH Average KI (nM) Average Bmax (fmol/mg) One-way ANOVA P value for Kd= 0.8249 6.6 (n=4) 0.56 0.18 37.14 3.77 7.4 (n=4) 0.48 0.12 35.74 2.75 8.2 (n=3) 0.46 0.07 25.73 1.22


63 Figure 4-7: Displacement curves for PTHP (permanently charged ligand at this pH range, pKa>10) at three different pHs. Ther e are no significant differences in the KI values between the three pHs used. Table 4-3: Table for Figure 4-7, Displacemen t curves for PTHP at three different pHs 6.6, 7.4, and 8.2. The data is the mean SE. pH Average KI ( M) One-way ANOVA P value=0.8070 6.6 (n=4) 0.74 0.11 7.4 (n=4) 0.91 0.23 8.2(n=4) 0.84 0.25 4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0 0 25 50 75 1008.2 7.4 6.6 -Log [PTHP], M125I-BTX Binding (% of Total) N N N PTHP


64 4 5 6 7 8 -20 0 20 40 60 80 1007.4 6.6 8.2 Log [CNBA], M125I-BTX Bo und (% of Total) Figure 4-8: Displacement curves for 3-[4 -Cyanobenzylidene]-anabaseine at three different pHs for three separate experiments. At pH 6.6 and 7.4 the KI values are not significantly different. However, at pH 8.2 the KI value is significantly different from the 6.6 and the 7.4 values. Table 4-4: Table for Figure 4-8, Displ acement curves for 3-[4-Cyanobenzylidene]anabaseine at three different pHs 6.6, 7.4, 8.2. The data is the mean SE. pH Average KI ( M) % Ionized KI Ionized ( M) One-way ANOVA P value <0.0001 6.6 (n=3) 2.58 0.22 56.9 1.47 0.13 Tukey: 6.6 vs. 7.4 P>0.05 7.4 (n=3) 2.43 0.18 17.3 0.42 0.03 Tukey: 7.4 vs. 8.2 P<0.001 8.2 (n=3) 6.91 0.37 3.2 0.22 0.01 Tukey: 6.6 vs. 8.2 P<0.001


65 CHAPTER 5 INFLUENCE OF LIPOPHILICITY OF TH E LIGAND ON ITS A FFINITY FOR THE RECEPTOR Introduction The blood-brain barrier (BBB) and the bl ood-cerebrospinal fluid barrier (BCSFB) are diffusion barriers which effectively isol ate the brain from the periphery, thus affording protection for the CNS against th e entry of many toxins and diseases. Unfortunately, these barriers also keep many medications from freely diffusing from the blood stream into the brain. The BBB consists of a monolayer of endothelial cells that are connected to each other by tight junctions with no fenestrations ( pores). Besides the endothelial cells, the BBB also consists of the capillary basement membrane, pericytes embedded within the basement membrane and astrocyte end-feet which wrap around the vessels. Small neutral lipophilic molecules can freely diffuse across the BBB along their concentration gradient. Other larger molecu les including essential nutrients like glucose and amino acids can enter the brain through tr ansporters or through receptor mediated endocytosis. The BBB is a significant chal lenge to overcome during drug development because the vessel walls are fairly impermeable due to tight junctions and the presence of efflux pumps, such as the p-glycoprotein pump, keeps many non-essential molecules out of the brain (Waterhouse, 2003). The passive diffusion of drugs across the BBB seems to be the most general route of entry into the brain. However, entry of a drug into the cerebrospinal fluid (CSF) does not guarantee it s entry into the brain. The BCSFB is the second barrier a drug encounters before en tering the CNS. The CSF can exchange


66 molecules with the interstitial fluid of th e brain parenchyma and therefore needs some form of regulating what can pass into the brai n. This barrier is made up of the choroid plexus (whose main role is the secretion of CSF), the arachnoid membrane and the circumventricular organs. The BCSFB regulat es the composition of the CSF by transport mechanisms including the active organic acid transporter. This transporter keeps therapeutic organic acids, such as the antibio tic penicillin, out of the CSF and therefore the brain. Besides this barrier, another barr ier has been suggested to exist between the CSF and the brain. This barrier would help account for inconsistencies in components found in the CSF and in the interstitial fluid of the brain parenchyma (Misra et al., 2003). Lipophilicity is one of many molecular characteristics that is traditionally considered during drug development. Lipophili city is the membrane-seeking property of a molecule. This property plays an esse ntial role in the absorption, distribution, metabolism, and elimination of drugs. The mo re lipophilic a drug, the more likely that a drug will be absorbed from the GI tract a nd also be able to pass through the BBB and reach its target. However, a drug can be too lipophilic; in this case it may be nonspecifically bound to plasma proteins, cell me mbranes, or to other components in the blood (such as plasma albumins) before it can reach its proper destination. Also, the higher the lipophilicity of a dr ug the higher the affinity of that drug for certain efflux pumps in the BBB (Waterhouse, 2003). In a ddition, the higher the lipophilicity (within limits) of a drug, the more likely that it will be taken up in the liver and the better the access of the drug to cytochrome P450 and other intracellularly localized metabolic enzymes. Thus, a drug must have a ba lance between itÂ’s lipophilic and hydrophilic properties in order to successfully transverse the BBB (Waterhouse, 2003).


67 There has been data collected indicating that in the BBB changes in pH result in permeability changes. For example, nicotine, which enters the brain through passive diffusion, has reduced BBB permeability with increased ionizati on (Oldendorf et al., 1979). The ideal Log P value, a common term asso ciated with the li pophilicity assessment of a compound, has been reported to be in the range of 2.0-3.5 for many drugs which target the CNS. It has been established th at in order for a drug to enter the brain by passive diffusion the following criteria are ne cessary:(1) molecular weight of the drug must be less than 450g/mol, (2) the Log P value must be between 2.0 and 3.5, (3) the number of hydrogen bond donors should be less th an five or the sum of all nitrogens and oxygens must be less than 10 (Waterhouse, 2003) . Thus, the Log P value is a valuable parameter that will aid in predicting whether or not a co mpound can enter the brain. However, there are some compounds whic h have a direct relationship between lipophilicity and biological activity. Some of the few comp ounds that express this direct relationship include the general anesthetics; ether, chloroform, and halothane. The anesthetic potency of these compounds incr eases with the increasing Log P values of these compounds ( The experimental method most commonly used to determine the Log P of a compound is the liquid/liquid phase partitioni ng method. The “shake-flask” method uses two previously equilibrated yet immiscible liquid phases, 1-octanol with an aqueous phase (10mM sodium phosphate buffer). The qua ntity of derivative is measured in each layer, the octanol and the aqueous phases, using reverse phase high performance liquid chromatography (RP-HPLC). In RP-HPLC, a sample is injected into a C18 reversed


68 phase column. The octadecyl groups capture the more lipophilic compounds and hold on to them for a longer period of time than the more polar or less lipophilic compounds which elute off of the column at a faster rate. By elution with a gradient of increasing acetonitrile concentration the more lipophilic compounds will ev entually be stripped from the C18 support and be eluted at a later tim e. The lipophilic octanol phase is meant to approximate the environment within a cell memb rane lipid bilayer. The use of octanol as the best representation of the lipid bilayer is a matter of debate. Others believe that organic solvents like hexane, decane, and branched chain alcohols would be a better representation. The major advant age of using the liquid/liquid system is that it measures the true partitioning behavior of all sp ecies of the compound present. The compound normally exhibits hydrogen bonding and ioni zation with the liquid/liquid partitioning method in a manner similar to its behavior in plasma. The other available methods used to dete rmine the lipophilicity of a drug include: (1) liquid/solid partitioning, (2) immobilized artificial-m embrane chromatography (IAM), and (3) computational estimates of lipophilic ity. The liquid/solid method uses an HPLC analysis of lipophilicity which is based on the compounds retenti on time (or the time it takes the compound to emerge from the colu mn) compared to the retention times of standards with known values of Log P. IAM chromatography serves as a model of biological membranes by emulating the solute partitioning into fluid membranes that are like lipid bilayers. IAM chromatography is a liquid chromatographic separation on monolayers of phospholipids covale ntly immobilized to silica particles at high molecular density (Braddy et al., 2002). This method produces capacity factors whose logarithm correlates with in vivo penetration across the BBB. IAM chromatography can be


69 combined with mass spectrometry (MS) that would allow injection of the compounds on a silica column at a high dilution factor because MS allows a low limit of detection (Prokai et al., 2001). Computational methods include a fragmentation based and a whole molecule approach for determining the Log P values. The fragmentation method breaks up the compounds structure and summarizes th e contribution to lipophilicity from each atom, functional group, or fragment contained within a structure of interest. This method, however, normally estimates a valu e for Log P that is higher than the experimental values. This high assessment is due to the fact that the fragmentation method does not take into account the electron ic and resonance contributions and this leads to an underestimation of hydrogen bondi ng and ionization contribution. The whole molecule approach is a newer method than the fragmentation method and it takes into account the entire molecule, including the electronic inte ractions and the resonance contributions. With this particular method current research is op timizing the algorithms and database content in order to ensure accura te estimations. This approach can be quite powerful in high-throughput assays, providi ng an estimation of Log P before the synthesis of the compound. However, this es timation is limited by the information which is available in the program library (Waterhouse, 2003). The shake-flask liquid:liquid partitioni ng method, described in Chapter 2, was selected for use in the present st udy of benzylidene anabaseines. Results The Log P values for all compounds are f ound in the appendix. An example HPLC trace of a back extracted octa nol sample is shown in Figure 5-1. All but three of the 3benzylidene anabaseine derivatives have av erage Log P values which are in the ideal range of 2.0-3.5 to successfully cross the BBB by passive diffusion. The three outliers


70 include: DMXBA with a Log P value of 4.10 0.27, 3-[(Benzofuran-2-yl)methylene]anabaseine with a Log P value of 3.86 0.39, and 3-[4-(Dimethylamino) benzylidene]anabaseine with a Log P value of 3.63 (n=1). However, 3-[4Dimethylaminobenzylidene]-anabaseine, along with the 3-[4-Aminobenzylidene]anabaseine and 3-[(2-Acetoxy)-4-methoxybe nzylidene)-anabaseine could only be analyzed in the octanol/water partiti oning procedure once because there was an insufficient amount of compound available. As shown in Figures 5-2 (r2 value of 0.01636) and 5-3 (r2 value of 0.02975) there is no direct relationship between the KI of the compound for the alpha7 or alpha4beta2 receptor as measured in binding studies and th e experimentally measured Log P value. Discussion There does not seem to be a correlati on between the Log P values for the 3benzylidene anabaseines and the KI values determined through binding assays. However, even though Log P is not predictive of affin ity, it has been found by Papke et al. (2004a) that at least some 3-benzylidene anabaseine compounds that are li pophilic and have large substituents in the para pos ition on the benzylidene ring are unable to interact in the binding site in order to produ ce the agonist effect. Inst ead these compounds may be better accommodated in a hydrophobic pocket that leads to antagonism. Therefore, Log P might have an affect on the efficacy of the 3-benzylidene anab aseine compounds. The influences of lipophilicity are specif ic to the set of compounds one is analyzing. For example, a group in 1998 performed a QSAR study on 6-substituted nicotine derivatives and found that the propertie s of these derivatives that contributed to nAChR affinity were lipophilic ity and steric size of the s ubstituent (Dukat et al., 1998).


71 The evaluation of Log P values should be done with caution. Just because a compoundÂ’s Log P value is not in the ideal range does not rule out the po ssibility that the drug will successfully enter the brain in high concentrations after being administered. For example, DMXBA was previously determined to have a P octanol/water value of 3810 740 making a Log P value of 3.58 (Kem et al., 2004). This previously report ed value is at the high end of the ideal Log P scale and would be predicted to have trouble entering the brain through passive diffusion. However, in the same publication the brain:blood concentration ratio was measured resulting from an intraperitoneal injection in rats over a 6-hour time interval. DMXBA accumulated in w hole brain to a much greater extent than its metabolites with a ratio of 3.45 0.57. In the publication, the 4-OH metabolite of DMXBA (3-[2-Methoxy-4-hydroxybenzylidene]-a nabaseine) with a Log P of 2.39 had a brain:blood ratio of 0.28 0.03, the 2-OH metabolite (3-[2-Hydroxy-4methoxybenzylidene]-anabaseine) with a Log P of 1.88 had a ratio of 0.08 0.01, while the 2,4 DiOH metabolite (3-[2,4-Dihydroxybenzylid ene]-anabaseine) with a Log P value of 1.48 hardly was able to get into the brain with a ratio of <0.05. In the publication it was stated that the predictions for the mo st lipophilic (DMXBA) a nd the least lipophilic (2,4 DiOH) compounds were not satisfactor y because these measured brain:blood concentration ratios refer to total (bound plus free) brain and blood c oncentrations of the compounds. The relationship between the bound and free concentration of a compound may vary with the compartment, making a su ccessful prediction difficult. Even though the Log P value found for DMXBA was at the hi gh end of the ideal scale it was still able to penetrate the brain with much more success than the 4-OH metabolite which has a more favorable Log P value. The publication went on to state that the reasons for the poor


72 penetration of the metabolites into the br ain was not because of high plasma protein binding and not because of high pKa values which would make the compounds highly ionized at physiological pH. More like ly the hydroxyl metabolites created many hydrogen bonds making these metabolites less likel y to enter the brain (Kem et al., 2004). Plasma protein binding, ioni zation state, and hydrogen bonding are all factors that complicate the Log P value. By knowing the extent of plasma protein binding, the pKa of the compound, and the degree of hydrogen bonding along with the Log P values the ability of a compound to enter the brain becomes better understood. Also, even though the Log P values of th ese 3-benzylidene anabaseines are not related to the affinity of alpha7 or alpha4be ta2 receptors, such values are useful as a guide for therapeutics. For example, Wang et al. (2003) showed that alpha7 nAChRs are present on macrophages. If a 3-benzylidene anabaseine is very hydrophilic it may be localized within the blood and extracellu lar spaces where it could interact with macrophage alpha7 nAChRs and thereby redu ce systemic inflammation by inhibiting the release of tumor necrosis factor from the cel ls with fewer complicat ions by way of brain penetration.


73 Figure 5-1: HPLC trace. 3-benzylidene anab aseine backextracted octanol phasse sample injected using the ammonium acetate bu ffer system (pH = 4.5). The trace has the retention time and the area under the curve labeled.


74 Figure 5-2: The Log P values determined us ing the octanol/water pa rtitioning system vs. the Log KI values determined through ra dioligand binding studies for the alpha 7 nAChR. Linear regression analysis gave an r2 value of 0.01636. Figure 5-3: The Log P values determined us ing the octanol/water pa rtitioning system vs. the Log KI values determined through ra dioligand binding studies for the alpha4beta2 nAChR. Linear regr ession analysis gave an r2 value of 0.02975. 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 -1.0 -0.5 0.0 0.5 1.0Log PLog KI (alpha 7) M 1.5 2.0 2.5 3.0 3.5 4.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5Log PLog KI (alpha4beta2) M


75 CHAPTER 6 QUANTITATIVE STRUCTURE ACTIVITY RELATIONSHIPS USING THE 3BENZYLIDENE ANABASEINE DERIVATIVES Going from a lead compound with a biological profile of interest to a drug with an optimal activity profile can involve the te sting of many analogs with substituents combined in different ways. This structure-activ ity approach is labor and cost intensive. One traditional approach initially uses pair -wise comparisons of compounds differing in a single site of substitution. This approach is useful for initial identification of likely factors of importance. Use of the Craig plot for selection of arom atic compounds is an example of this approach. However, pair-wis e comparisons are not al ways successful in identifying the important factors, since s ubstitutions always change more than one property. Quantitative structure activity re lationships (QSAR) are increasingly used to identify important structural features since a variety of compounds are now usually available at the beginning of drug developmen t. QSAR can identify important properties of active compounds by using a linear equation in order to quantitatively predict in biological activity when seve ral changes in molecular pr operties occur simultaneously. The molecular properties evaluated can includ e: the Hammett (electronic) and Hansch (lipophilicity) parameters, molar refractivity (size), topological and shape indices, electronic and quantum-mechanical descriptors, etc. When using QSAR software, to get an accurate QSAR the analogs to be studied should have diverse activity values (eg. EC50 or KI values) in order to form a reliable data set. As measures of the reliability of the data produced using this technique one can dete rmine the residuals from the model, the


76 correlation coefficient (r), and the Fisher st atistic (F-value). The residuals are the differences between the calculated values and the experimental values of a compound for a particular property (ex. KI). The standard error calcula ted by taking the root-meansquare of the residuals should be smaller than the st andard error of the original data about the mean. Also, the correlation coefficient, r, measures how closely the observed data tracks the fitted regression line in a plot of measured vs. calculated values. Errors in either the model or the data will lead to a poor fit. r = Regression Variance/Original Variance Original variance is the sum-of -the-squares of the distances of the original data from the mean and regression variance is the orig inal variance minus the variance around the regression line. Possible values for r are between 0, whic h indicates no relationship between the activity and the property be ing evaluated, and 1, which would indicate a perfect correlation. The higher the value of the r the less likely the relationshi p calculated is due to chance. When using more than one variab le to determine the activity of the compounds in the data set, multiple linear regression is used in order to determine the importance of multiple properties (descriptors) to the overa ll fit of the data. Th is method adjusts the descriptors by adding or removing them to maxi mize the fit of the data to a regression equation. For example, if two properties are cros s-correlated one of th e properties will be removed. However, if one uses enough propertie s in the QSAR, any data set can be fit to a regression line. To avoid th is the regression analysis requires more compounds than properties to be determined. Finally, the F-valu e determines the statistical significance of


77 the multiple linear regression equation. The F-value is calculated from r and the number of data points in the data set (n). F1,n = (n-2)[ r/(1-r)] The higher this value, the more reliable the data (Richon, 1997). The QSAR approach was successfully used in the discovery and synthesis of FDL 64176, a new class of calcium channel activat or that does not act on any of the welldefined calcium channel modulator receptor site s, as typified by verapamil, diltiazem, and the dihydropyridines. Results from an initial training set of 8 related compounds were subjected to multiple linear regression analysis to guide the optimization of the structure activity relationships of this seri es of calcium channel agonists. The QSAR equation showed that most of the observed activity ch anges could be explained by considering the steric bulk of the substitu ent and its lipophilicity. The potent activity of FDL 64176 was predicted using QSAR on an in itial set of less pot ent benzoylpyrroles (Baxter et al., 1993). Also, Nicolotti et al. (2004) used QSAR analysis on 269 nAChR ligands taken from the literature. The ligands chosen were t hose which had structural modifications on the pyrrolidine and pyridine rings of nicotine or aryland heteroaryl-oxymethylcyclic amines. This group used only binding da ta to the alpha4beta2 nAChR on rat brain membranes. They used both a 2D and a 3D QSAR analysis. The bidimensional analysis used was the classical Hansch approach a nd it was used on several different congeneric series of nicotinic agonists in order to find the chemical features modulating the affinity to nAChRs. The 3D analysis was used in or der to better define how such interactions were localized in 3D space. This gr oup has developed a new pharmacophore model,


78 based on a wider set of active ligands showing a high homogene ity of the biological data and a great molecular diversity. So far this group has determined three geometrically independent elements which have been f ound to be critical for the ligand-receptor interaction and which have been used to es tablish ligand alignment in their 3D QSAR studies: a positively charged nitrogen atom, a lone pair of the pyridine nitrogen or a carbonyl oxygen, and a “dummy” atom indi cating a hydrophobic cen troid located in areas generally occupied by aliphatic cycles. The 2D results demonstrated that steric effects as modeled by molar refractivity were the most important one s in influencing the binding of nAChR ligands. The 3D QSAR allowed this group to merge all the congeneric series and to deve lop a global model which had good predictive ability. This group found that while the traditional Hansch approach affords valuable models which can be considered better than the 3D QSAR method, the 3D approach plays a fundamental role when different classes of bi oactive molecules have to be examined for the derivation of comprehensive models of ge neral validity (Nicolotti et al., 2004). Also, using QSAR should save time, materials and money. In the present study QSAR analysis ha s been done using the SciQSAR 3.0 (SciVision, Burlington, MA) program. In this program 17 properties can be analyzed (Lorand et al., 2002). The SciQSA R software uses 2D and 3D molecular parameters to create structure-activity rela tionships to explain the obs erved activities of compounds, and then predicts the activities of new compounds ( SciQSAR.html ). Geometry-optimised st ructures of the 29 experimental compounds have been inserted into the SciQSAR 3.0 program using semi -empirical quantumchemical method (PM3) available as a part of the Chem 3D Ultra 5.0 molecular modeling


79 software (Lorand et al, 2002). Besides us ing the multiple linear regression (MLR) method, some alternatives that are extensi ons of the MLR method, include the principal component regression analysis (PCA) or the partial least squares (PLS) method. PCA is a data reduction method where the main elem ent of this approach consists of the compilation of a small set of non-correlated variables derive d from a linear combination of the original variables. PLS is a multivar iate regression method using projections to summarize many potentially colinear variables ( Results Figures 6-1 and 6-2 show that, using the SciQSAR software, the predicted KI values based on the structure of the 3benzylidene anabaseines had a very good correlation when compared to the experimental KI values obtained from radioligand binding studies for both the 7 and the 4 2 nAChRs. This is shown in the multiple correlation coefficients and in the FisherÂ’s statistics calculated fo r each receptor (R= 0.80, F=15.14 7, and R=0.75, F=16.41 4 2). The equations that were produced by the software after some initial descriptors were cross-correlated and removed because of redundancy shows that for the 7 nAChR the most important chemical properties that relate to the affinity of the 3-benzylidene anabaseines for this r eceptor include the pKa, which shows how electronic influence and io nization are important, the dipole of the compound which is the 3D assessment of charge, and the ABSQon parameter which represents the char ges on the heteroatoms (nitrog ens and oxygens) of the compound. Most of these properties prove d to be important for the 4 2 nAChR as well. However,


80 the dipole variable was left out of the equation, because for this receptor it was a redundant parameter for predicting receptor affinity. The validity of the equations produced by the SciQSAR software was tested by using two 3-benzylidene anabas eines, the prediction set, th at were not included in the training set. In Table 6-2, (the so-c alled prediction set) included 3-[4(Acetamido)benzylidene]-anabaseine and 3-[4 -(Acetyl)benzylidene]-anabaseine. The KI values obtained by the equations, after th e dipole and the charges on the heteroatoms were calculated by the SciQSAR software, we re reasonably well correlated to the KI values determined experimentally for the 7 and the 4 2 receptors. The predicted KI value for the 3-[4-(Acetyl)benzylidene]-anabaseine for the 4 2 receptor departed the most from the experimental value. Since th is experimental value has a large standard error, this experimental valu e may be only approximate. Discussion The calculated QSARs indicate that ioniza tion of the imine nitrogen is important for binding to both receptor subtypes, which is consistent with our previous data on the pH dependence of DMXBA (GTS -21) binding to rat brain 7 receptors (Kem et al., 2004 and this dissertation Ch. 4). They also suggest that other electronic distribution properties are important for receptor binding. Finally, lipoph ilicity of the benzylidene substituents seem to have a relatively minor influence on receptor binding. Sin ce all the differences between the compounds presented here relate to substitutions on the benzylidene ring, it is likely that QSARs based on substitutions in the other two rings will be different. An important caveat is that QSARs for re ceptor activation may be different from QSARs for receptor binding. However, since most of the compounds in Table 6-1 have


81 been shown to be partial agonists at the 7 receptor and weak antagonists at 4 2 receptors, we anticipate that the derived QS ARS for the two nAChR subtypes will not be markedly state-dependent. These QSAR equations should allow a preliminary, approximate prediction of the receptor bindi ng properties of 3-benzylidene anabaseines one might synthesize for pharmacological test s. To our knowledge this would be the first 7 nAChR QSAR published.


82 Alpha 7: Log(1/Ki) = 1.42 + 0.57*pKa + 0.12*Dipole + 0.63*ABSQon Multiple Correlation Coefficient R= 0.80 FisherÂ’s Statistics F= 15.14 Figure 6-1: SciQSAR 3.0 equation for the alpha7 nAChR using the KI values obtained experimentally for the 29 compounds in the training set. Log(1/Ki)ex p Log(1/Ki)calc


83 Alpha4 beta2 :Log(1/Ki) = 2.44 + 0.42*pKa + 0.76*ABSQon Multiple Correlation Coefficient R= 0.75 FisherÂ’s Statistics F= 16.41 Figure 6-2: SciQSAR 3. 0 equation for the alpha4beta2 nAChR using the KI values obtained experimentally for the 29 compounds in the training set. Log(1/Ki)exp Log(1/Ki)calc


84 Table 6-1: Training set of 29 3-benzylidene anabaseine (B A) compounds used to create the alpha7 and alpha4beta2 equatio ns in the SciQSAR 3.0 software. 3-[4-Amino BA] 3-[4-n-Butoxy]BA 3-[4-Butoxy]BA 3-[(2-Propoxy)-4-methoxy]BA 3-[4-(Trifluoromethoxy)]BA 3[(2-Pentoxy)-4-methoxy]BA 3-[4-(Methylamino)]BA 3-[(2-H ydroxy)-4(diethylamino)]BA 3-[4-Hydroxy]BA 3-[2–Acetoxy]BA 3-[4-(Methylthio)]BA 3-[ 3,5-Dimethyl-4-hydroxy]BA 3-[4-Cyano]BA 3-[2,4,6-Trimethyl]BA 3-[4-Propyl]BA 3-[2,4,6Trimethoxy]BA 3-[4-Acetoxy]BA 3-[3-Methoxy]BA 3-[4-Morpholino]BA 3-[3 Hydroxy]BA 3-[4-(Dimethyl amino)]BA 3-[2-Pivaloyloxy]BA 3-[4-(Difluoromethoxy)]BA 3-[(6-Me thoxynaphth-2-yl)methylene]BA 3-[4-Chloro]BA Unsubstituted benzylidene anabaseine 3-[4-(1H-1,2,4-Triazol-1-yl)] BA 3-[2,4Dimethoxy]BA 3-[4-Pivaloyloxy]BA


85 Table 6-2: Prediction set of 2, 3-benzylidene anabaseine (BA) derivatives that were used to test the validity of the two equations produced by the SciQSAR 3.0 software and the training set. Compound Ki ( M) 7 Ki ( M) 4 2 Ki 4 2/ 7 Ratio 3-[4-Acetamido]BA 0.24 0.03 n=2 Calculated using QSAR: 0.13 0.38 0.013 n=2 Calculated using QSAR: 0.41 1.58 Calculated using QSAR: 3.15 3-[4-Acetyl]BA 0.92 0.04 n=2 Calculated using QSAR: 0.67 4.49 2.3 n=2 Calculated using QSAR: 1.09 4.88 Calculated using QSAR: 1.63


86 CHAPTER 7 FUNCTIONAL DATA FOR THE 3-BENZ YLIDENE ANABASEINE DERIVATIVES USING CELLS EXPRESSING THE HUM AN AND RAT ALPHA7 NACHR AND CELLS EXPRESSING THE HUMAN FE TAL MUSCLE NICOTINIC RECEPTOR Introduction Electrophysiological techniques, like w hole-cell patch-clamp, provide a rigorous and conventional assessment of the functional properties of ion channels; however, this method is very time consuming and costly a nd is not easily amenable to high throughput evaluation of compound libraries. High throughp ut screening has led to adaptations for screening large numbers of compounds in dr ug discovery. One method that is rapid and economical for the identification of compounds that modulate ion channel activity is a fluorescence-based assay. This assay incorporates a machine referred to as the Flexstation. Fluorescence has attracted much attention for the purpose of high throughput because it allows rapid screening of compounds for evaluation of structureactivity and subtype specificity relationships . The FlexstationÂ’s purpose is to assess agonist or antagonist properties of the experimental 3-benzylidene anabaseine compounds by measuring changes in fluorescence , which represent changes in membrane potential or intracellular calcium concentr ation, in cell lines that express different nicotinic receptor subt ypes (Figure 7-1). The dye used for the membrane potentia l assays (Molecular Devices, MD) is proprietary; however, previous publications have shown that this dye evokes the fluorescence response quite rapi dly (maximal response within two minutes) and results


87 obtained for potassium channels are comparable to results obtained with the same ligands on the same channels by whole-cell patch-cl amp studies. Upon depolarization of the cells, the MD membrane potential dye moves into the cell and subs equently binds to intracellular hydrophobic sites resulting in an en hanced quantum yield of the dye and an increase in the fluorescence intensity. Th e dye exits the cells during hyperpolarization and there is a decrease in fluorescence si gnal. When comparing the MD membrane potential dye used with the Flexstation on potassium channels directly to another membrane potential dye, DiBAC4, several a dvantages are apparent. For example, a larger signal-to-noise ratio, lower inci dence of interference from quenching and fluorescence artifacts, increased throughput, the ability to carry out the assays at room temperature, and not having to wash the ce lls with buffer before addition of the dye allowing more cells to be present on the assa y plate and simplifying the assay (Whiteaker et al., 2001). Using the membrane potential change as a m easure of activity is very useful for the receptors that do not pass a lot of calcium. When nicotinic receptor subtypes including the 4 2, 3 2, and human neuromuscular receptors were utilized by other groups using this high throughput assay the membrane potenti al dye proved to afford the most robust responses compared to the calcium res ponses. For example the TE-671 cells, which express the human fetal muscle type nicotinic receptor, ha ve a low calcium conductance and therefore the membrane pot ential dye is ideal. In this study the membrane potential dye served as a novel and sensitive measure of nicotinic activity which gave comparable results to studies using other calcium dye i ndicators and rubidium efflux experiments. However, the results obtained using this me mbrane potential dye were not completely

PAGE 100

88 identical, agonists were significantly more potent whereas antagonists were less potent (Fitch et al., 2003). Even though the membrane potential dye ha s proven useful for the other nicotinic receptor subtypes it has not been successful with the alpha7 nAChR subtype. In recent years fluorescence measurement of nicotinic receptor activity th rough calcium dynamics has been documented extensively (Chavez-No riega et al., 2000). Pr evious studies have indicated that 7 receptor activation alters calciu m flux across the membrane. This increase in intracellular calcium could be b ecause of an increase in calcium flux through the nicotinic receptors and/or it could be caused by an influx of calcium through voltagedependent calcium channels (Quik et al., 1997 ). This small incr ease in intracellular calcium is believed to support the short-te rm process of synap tic potentiation and facilitation as well as protect cells under c onditions that would othe rwise cause cell death (Zucker, 1996; Koike et al., 1989). If the intracellular calcium in flux was large this would stimulate synaptic releas e or it could contribute to th e induction of excitotoxicity. This delicate balance of intr acellular calcium concentrati on is an ideal job for the 7 receptor. The desensitization and the inward rectification (these channels pass more current when the membrane is hyperpolarized) of 7 receptor currents help to keep intracellular calcium levels produced by this receptor at a safe range (Clementi et al., 2000). The Molecular Devices calcium dye is al so proprietary, however its nature is similar to the Fluo-3,4 calcium chelating dyes, except in the MD cal cium dye there is a secondary dye incorporated to mask backgr ound fluorescence. The dye is practically nonfluorescent in free ligand form and the fluores cence increases dramatically (60-80 times)

PAGE 101

89 with calcium complexation. The dye enters the cell during incubation and upon addition of an agonist, which stimulates the openi ng of a nicotinic re ceptor and increases intracellular calcium, the dye complexes with the entering calcium and fluoresces. The rapidly desensitizin g character of endogenous 7 nAChRs have made them difficult to study. In hopes of better understanding the 7 nicotinic recept or structure, function and pharmacology single amino acid mu tations in the porelining TM2 domain of the alpha7 nAChR have been produced. These mutants which include the Leu250-toThr and the Valine274-Thr have been studied in Xenopus oocytes and both are gain-offunction models of the alpha7 receptor. Both mutations have an increased sensitivity to agonists, some antagonists on the wild-typ e receptor act as agonists on the mutant receptors, and both mutant receptors are slow lyor non-desensitizing (Bertrand et al., 1992; Briggs et al., 1999). Another mutant alpha7 nAChR was produced as a chimera comprising the N-terminal region of the hum an alpha7 receptor linked at valine 202 with the transmembrane C-terminal region of the mouse 5HT3 receptor. This chimera was functionally expressed in Xenopus oocytes and in HEK 293 cells. This chimera was found to be sensitive to ligands for nAChRs but not for 5-HT3 receptors. Also, in Xenopus oocytes, currents obtained using this chimer a desensitized more slowly (Craig et al., abstract). This slowing of the desensit ization would be a valuable change in the kinetics of the alpha7 receptor that would aid in the study for pot ential therap eutics. However, the pharmacology of the chimera and the other mutant receptors is completely different from the wild-type receptor and us ing these mutants as pharmacological tools for potential drugs would be unwise. These mutants would be much more valuable in

PAGE 102

90 understanding the structure of the receptor and determining how the structure relates to function. When using the Flexstation in experiments with different cell lines epibatidine was used in these assays as a standard agonist a nd all of the changes in fluorescence obtained upon addition of the 3-benzylidene anabas eine compounds were normalized to the maximum change in fluorescence obtain ed for epibatidine (Fitch et al., 2003). Epibatidine is an alkaloid is olated from the skin of a frog, Epipedobates tricolor , and has been shown to be a very potent analgesic. It is regarded as the most potent nicotinic acetylcholine receptor agonist (Qian et al ., 1993). In radioligand binding studies epibatidine displaces [3H]cytisine binding with a KI of 43pM to the 4 2 nAChR subtype and displaces [125I] -BTX binding with a KI of 230 nM to the 7 subtype in rat brain membrane (Sullivan et al., 1994). It has been shown that epibatidine acts as an extremely potent full agonist at the human neuronal 3 2, 3 4, and at the 7 nAChRs expressed in Xenopus oocytes. In this same study it was shown that the apparent affinity of epibatidine for these receptor subtypes was 100 to 1000-fold higher for epibatidine than for nicotine or acetylc holine (Gerzanich et al., 1995). Results The purpose of these flexstation assays is to determine the functionality of the 3benzylidene anabaseine compounds on an 7 nAChR expressing cell line and on a muscle expressing cell line. The first set of experiments were pe rformed on the SH-EP1 native nAChR-null epithelial cell line expressing the human al pha7 nAChR. The SH-EP1 cell line has been shown to exhibit many of the properties of naturally or heterologously expressed 7

PAGE 103

91 nAChRs made up of 7 subunits from different species. These cells have been shown to exhibit rapid activation a nd inactivation upon agonist exposure, responsiveness to choline, nicotine, and ACh, high calcium permeability, and high affinity competitive antagonism by MLA and BTX (Peng et al., 1994). The calcium dye protocol was used with these cells because previous attempts with the membrane potential dye were unsuccessf ul. With these cells the change in fluorescence obtained for the known agonist epibatidine (Figure 7-2) was not as significant as the change in fluorescence usi ng several of the 3-benzylidene anabaseine derivatives, eg. 3-[4-(Methylamino)benzylid ene]-anabaseine (Figure 7-3). This discrepancy in the apparent calcium flux make s it appear that some of the 3-benzylidene anabaseine derivatives are more efficacious than the full agonist epibatidine on the 7 nAChR. This is most likely not the ca se. The possibilities accounting for this discrepancy include: (1) the ba sicity of the 3-benzylidene anabaseines might be altering the stability of the cell membrane allowing mo re intracellular calcium entry that is not related to the stimulation of the 7 nAChR, (2) the 3-benzylidene anabaseine compounds could be fluorescent on their own, or (3) the cell line does not have the machinery required for post-translational m odification and assembly of the 7 receptors preventing the proper expression of functional homomeric 7 nAChRs on the cell membrane surface. This improper expression has been demonstrated in different strains of PC12 cells where dramatically different amounts of surface 7 receptors are expressed without there being a large diffe rence in the amounts of 7 mRNA expressed (B lumenthal et al., 1997).

PAGE 104

92 In an attempt to overcome this problem , of a larger change in fluorescence produced for the experimental 3-benzylidene anabaseine compounds as compared to the known agonist epibatidine, another 7 expressing cell line was employed. This next cell line, known as GH4C1, is a rat pituitary cell line and it expresses the rat 7 nAChR. The GH4C1 7 receptors are a proper mode l to study the function of 7 receptors because these receptors are similar to oocyte-expressed and endogenous 7 receptors. The GH4C1 cell line stably expresses 7 sites which have sim ilar binding affinities, pharmacology, kinetics, and mol ecular size as rat brain -BTX receptors. It has been reported with this cell line that nicotinic agon ists increase intracellular calcium levels in a dose-dependent manner using the fluorescent dye, Fura-2. Also, nicotinic receptor antagonists, in this cell line have inhibited agonist-induced increas es in intracellular calcium (Quik et al., 1997). Unfortunately, the same large change in fluorescence was found when using the 3-benzylidene anabasei nes as compared to epibatidine in the Flexstation assay (Figure 7-2,7-4). The 3-benzylidene anabaseines that produced these large changes in fluorescence in these cell lines did so reproducibly. However, when using epibatidine, acetylcholine, and nicotine on these cell lines reproducible curves were not able to be obtained. Also, the 3benzylidene anabaseine responses produced by the GH4C1 cell line were blocked by 150 M tubocurarine (Figure 7-4 ). This finding makes it seem that the calcium responses produced by the 3-benzylidene anabaseines ar e because of direct stimulation of the 7 nAChR. A successful assessment of the EC50 and IC50 values for epibatidine, ACh, and nicotine were obtained using the membrane potential dye protocol with the TE-671 cell

PAGE 105

93 line which expresses the huma n fetal muscle-type nAChR ( 1 ) (Figure 7-5). The purpose of this cell line is to determine what affect these 3-benzylidene anabaseines, have on the muscle nAChR. One problem with th e TE-671 cell line is that this cell line has been shown to express a finite number of hi gh-affinity binding sites for the muscarinic acetylcholine receptor antagonist quinuclidinyl benzilate, QNB (Bencherif and Lukas, 1991). Previously it had been published that ep ibatidine had little or no activity at a variety of other central receptors including the opioid, adrenergic, dopamine, serotonin, gamma-aminobutyric acid, and muscarinic re ceptors (Daly et al., 1994). However, a recent study has shown that (+/-) epibatidine, in micromolar concentr ations is a partial muscarinic receptor agonist (Kommalage and Hoglund, 2004). In order to reduce any complications from muscarin ic receptors present on the TE-671 cells all of the drugs delivered to these cells were suspended in a HankÂ’s Balanced Salt Solution/20mM Hepes buffer which contained 1 M atropine, a muscarinic receptor antagonist. A majority of the 3-benzylidene anabaseine derivatives ha d agonist and/or antagonist properties, EC50 and IC50 responses, on the fetal muscle nAChR used in the Flexstation assays (Figure 76;Table 7-1). The derivatives that had no ch ange in fluorescence, and therefore had no agonist or antagonist activity were: from the para substituted group, 3-{4-[bis(2Hydroxyethyl)amino]benzyliden e}-anabaseine, 3-[4-(Dimethylamino)benzylidene]anabaseine, 3-[4-Morpholinobenzylidene] -anabaseine, 3-[4-Cyanobenzylidene]anabaseine, 3-[4-(Methylamino)benzylidene]-ana baseine; from the para/ortho substituted group,3-[2,4-Dihydroxybenzylidene] -anabaseine and the 3-[2,4-Dimethoxybenzylidene]anabaseine; and from the trisubstitut ed derivatives, 3-[3,5-Dimethyl-4hydroxybenzylidene]-anabaseine, 3-[3,4,5-Trim ethoxybenzylidene]-anabaseine. Even

PAGE 106

94 though there were agonist and antagonist responses produced by the other 3-benzylidene anabaseine derivatives this does not mean that if used for treatment they would necessarily produce unwanted side effect s. The dose needed to activate the 7 nAChR with a particular deriva tive might be less than the dose used in the Flexstation assays to stimulate or block the muscle nAChR. Discussion Unfortunately, the SH-EP1 and the GH4C1 cell lines were unable to produce reliable results using the molecular devices calcium dye in the Flex station. The reasons for this disappointing outcome could be because the cell lines do not express the machinery required to allow proper assembly and transport of the receptors onto the surface of the cell membrane, the basicity of the 3-benzylidene anabaseines are compromising the cell membrane allowing calcium to enter the cell by other means, or the 3-benzylidene anabaseines could themselves be fluorescent. The hypothesis that the basicity of the compounds cause the large change in fluores cence due to influx of calcium from other sites of entry besides the nicotini c receptor might not be valid because in the GH4C1 cells the change in fluorescence upon addition of the 3-[4(Methylamino)benzylidene]-ana baseine was inhibited by 150 M tubocurarine a nonspecific nicotinic receptor antagonist. This i nhibition of the nicotinic receptor caused the calcium flux to be terminated. However, this was the only 3-benzylidene anabaseine tested using tubocurarine. Further ex periments blocking the activation of the 7 nAChR expressing cell lines needs to be done in or der to confidently conclude that this discrepancy between the epibatidine produ ced calcium flux and the 3-benzylidene

PAGE 107

95 anabaseine produced calcium flux is not due to entry of calcium by other means besides the nicotinic receptor. The results obtained using the TE-671 feta l muscle receptor cell line were much more reliable due to th e reproducibility of the EC50 and IC50 values for epibatidine, ACh, and nicotine with the membrane potential assay. For the 3-benzylidene anabaseine compounds these EC50 and IC50 values are important because they help to establish if there are any unwanted side effects by thes e derivatives on the muscle receptor. However, just because there may be activation or antagonism by these 3-benzylidene anabaseine derivatives does not mean that the dose required for stimulating the 7 receptor will be the same. Therefore, the derivatives might not have unwanted side effects on the muscle receptor. Figure 7-1: Representative trace of a cha nge in fluorescence in GH4C1 cells due to the addition of a 3-benzylidene anabasei ne derivative at three different concentrations. Time (secs) 0 20 40 60 80 100 120 140 160 180 200 6000 11000 16000 21000 26000 31000 36000 Well G4 G5 G6 Max Min249332930831477 Compound Nicotine (100 M) 4X ionomycin/carbachol/FCCP

PAGE 108

96 -10 -9 -8 -7 -6 -5 -4 -3 25 50 75 100Log (Epibatidine), M SH-EP1 (A)Percent Maximum Response Relative to Epibatidine -9 -8 -7 -6 -5 -4 -3 40 60 80 100Log (Epibatidine), M GH4C1 (B)Percent Maximum Response Relative to Epibatidine Figure 7-2: Calcium flux of epibatidine in SHEP1 cells (A), and GH4C1 cells (B) in the presence of 1 M atropine.

PAGE 109

97 Figure 7-3: Calcium flux curve of 3-[4 -(Methylamino)benzyliden e]-anabaseine with percent maximum fluorescence relative to epibatidine in SH-EP1 cells expressing human alpha7 nAChRs. -9 -8 -7 -6 -5 -4 -3 -2 0 500 1000 1500EC50= 146.0 M Hill slope= 0. 8631Log (3-[4-(Methylamino)benzylidene]-anabaseine), MPercent Maximum Response Relative to Epibatidine

PAGE 110

98 -10 -9 -8 -7 -6 -5 -4 -3 -10 15 40 65 90 115 140 165 190IC50 EC50 0 25 50 75 100Log (3-[4-(Methylamino)benzylidene]-anabaseine), M (B)Percent Maximum Response Relative to EpibatidinePercent Response Relative to 100 M Tubocurarine Figure 7-4: Calcium flux curves with GH4 C1 cells. (A) Calcium flux curve for 3-[4(Methylamino)benzylidene]-anabasein e with percent maximum fluorescence relative to epibatidine in GH4C1 cells expressing rat alpha7 nAChRs. (B) Calcium flux activation by 3-[4-(Methyl amino)benzylidene]-anabaseine with percent maximum fluorescence relative to epibatidine in GH4C1 cells and inhibition of the 3-[4-(Methylamino)be nzylidene]-anabaseine response by 150 M tubocurarine. -9 -8 -7 -6 -5 -4 -3 -2 0 100 200 300EC50= 85.88 M Hill slope= 2.552Log (3-[4-(Methylamino)benzylidene]-anabaseine), M (A)Percent Maximum Response Relative to Epibatidine

PAGE 111

99 Figure 7-5: Membrane poten tial curve using TE-671 cells expressing the human fetal muscle receptor in the presence of epibatidine. The EC50 was determined to be 0.30 0.45 M and the IC50 was determined to be 0.25 0.05 M. Previously published values from Fitch et al., 2003 were determined to be: EC50, 0.27 0.03 M , IC50, 0.52 0.01 M. -10 -9 -8 -7 -6 -5 -4 -3 -2 -10 15 40 65 90EC50 IC50 0 25 50 75 100Log 3-[2-Trifluoromethoxybenzylidene]-anabaseine, MPercent Maximum Response Relative to EpibatidinePercent Maximum Response Relative to 100 M Nicotine Figure 7-6: Membrane potenti al curve using TE-671 cells expressing the human fetal muscle receptor in the presence of 3-[2-Trifluoromethoxybenzylidene]anabaseine. The EC50 was determined to be 12.42 2.58 M and the IC50 was determined to be 7.89 1.38 M. -9 -8 -7 -6 -5 -4 -3 0 25 50 75 100IC50 EC50 0 25 50 75 100Log [Epibatidine], M% Maximum Response Relative to Epibatidine% Response Relative to 100 M Nicotine

PAGE 112

100 Table 7-1: TE-671 cell line EC50 and IC50 values upon addition of the 3-benzylidene anabaseine derivatives using the membrane potential assay. Compound Name EC50 ( M) IC50 ( M) 3-[4-Amino benzylidene]anabaseine No response (160 M) 18.42 3-[4-Butoxy benzylidene]anabaseine 62.25 4.55 1.79 0.23 3-[4-(Trifluoro methoxy) benzylidene]-anabaseine 1.35 0.32 1.31 3-[4-(Methylamino) benzylidene]anabaseine No response (500 M) No response (500 M) 3-[4-Hydroxy benzylidene]anabaseine No response 3.56 3-[4 Methoxy benzylidene]anabaseine Insufficient sample Insufficient sample 3-[4-Methyl benzylidene]anabaseine Insufficient sample Insufficient sample 3-[4-(Methylthio)benzylidene]anabaseine 90.74 26.18 3-[4-Cyano benzylidene]anabaseine No response (160 M) No response (160 M) 3-[4-Propyl benzylidene]anabaseine 23.54 1.78 2.46 0.36 3-[4-Acetoxybenzylidene]anabaseine No response (160 M) 5.92 2.16 3-[4-Morpholino benzylidene]anabaseine No response (500 M) No response (500 M) 3-[4-(Dimethyl amino) benzylidene]-anabaseine No response (160 M) No response (160 M) 3-[4-(Difluoromethoxy) benzylidene]-anabaseine 24.28 6.20 15.12 5.10 3-[4-Chlorobenzyidene]anabaseine 2.16 0.87 4.76 2.39 3-[4-(1H-1,2,4-Triazol-1yl)benzylidene]-anabaseine 21.97 6.58 36.20 11.59 3-[4-Acetamidobenzylidene]anabaseine 15.02 1.44 16.44 6.47 3-{4-[bis(2Hydroxyethyl)amino]benzylidene}anabaseine No response (160 M) No response (160 M) 3-[4-Pivaloyloxybenzylidene]anabaseine 5.24 3.53 10.17 0.62 3-[4-Acetylbenzylidene]anabaseine 6.49 4.48 70.22 24.85

PAGE 113

101 Table 7-1. Continued Compound Name EC50 ( M) IC50 ( M) 3-[4-n-Butoxybenzylidene]anabaseine 150.5 9.10 12.79 9.80 3-[2–Acetoxy benzylidene]anabaseine 10.89 1.59 45.35 21.19 3-[2-(Trifluoro methoxy) benzylidene]-anabaseine 12.42 2.58 7.89 1.38 3-[2-(Difluoro methoxy) benzylidene]-anabaseine 17.49 0.19 17.29 2.09 3-[2-Pivaloyloxy benzylidene] anabaseine 51.37 13.89 7.76 3-[3-Methoxybenzylidene]anabaseine 99.21 36.40 9.86 6.63 3-[3 Hydroxy benzylidene]anabaseine 96.62 47.57 3-[2,4Dimethoxy benzyl]anabaseine >926 43.06 6.84 3-[(2-Methoxy)-4glucuronide benzylidene]-anabaseine 0.73 0.10 1.22 3-[(2-Propoxy)-4-methoxy benzylidene]-anabaseine 99.21 0.1 0.01 3-[(2-Pentoxy)-4methoxy benzylidene]-anabaseine 7.94 2.03 0.40 0.18 3-[(2-Acetoxy)-4methoxybenzylidene)-anabaseine 16.85 4.81 3-[(2-Hydroxy)4(diethylamino)benzylidene]anabaseine No response (160 M) 15.22 4.34 3-[2,6-Dimethyl-4-hydroxy benzylidene]-anabaseine No response (160 M) 27.88 20.77 3-[3,5-Dimethyl-4-hydroxy benzylidene]-anabaseine No response (160 M) No response (160 M) 3-[2,4,6-Trimethyl benzylidene]-anabaseine 1.44 0.76 0.34 0.06 3-[2,4,6Trimethoxy benzylidene]-anabaseine 0.39 1.47 3-[3,4,5Trimethoxy benzylidene]-anabaseine No response (160 M) No response (160 M) 3-[3,4(Ethylenedioxy)benzylidene]anabaseine No response (500 M) 4.97 0.53

PAGE 114

102 Table 7-1. Continued Compound Name EC50 ( M) IC50 ( M) 3-[3,4(Methylenedioxy)benzylidene]anabaseine No response (500 M) 5.22 3.92 3-[(6-Methoxynaphth-2yl)methylene]-anabaseine 5.84 5.21 2.66 3-[(Benzofuran-2-yl)methylene]anabaseine 41.62 36.27 12.84 8.44 Unsubstituted benzylidene anabaseine 7.59 1.02 8.41 3.30 3-[2,4Dimethoxy benzylidene]-anabaseine No response (160 M) No response (160 M) 3-[2Methoxy-4hydroxy benzylidene]-anabaseine 0.28 3.32 3-[2-Hydroxy-4methoxy benzylidene]-anabaseine Insufficient sample Insufficient sample 3-[2,4-Dihydroxy benzylidene]anabaseine No response (160 M) No response (160 M)

PAGE 115

103 APPENDIX 3-BENZYLIDENE ANABASEINE COMPREHENSIVE TABLE Compound Name Structure Ki ( M) 7 Ki ( M) 4 2 Ki 4 2 / 7 Ratio pKa % ionize d Log P 3-[4-Amino benzylidene]anabaseine N N NH2 0.05 0.003 n=2 0.12 0.03 n=5 2.40 8.33 0.06 89.49 2.43 n=1 3-[4-Butoxy benzylidene]anabaseine N N OC4H9 2.23 0.48 n=3 2.17 0.24 n=2 0.97 7.45 0.04 52.88 Insufficient sample 3-[4Trifluoromethoxy benzylidene]anabaseine N N O CF3 0.93 0.19 n=3 0.93 0.09 n=3 1.0 6.84 0.05 21.58 2.44 0.33 3-[4(Methylamin o) benzylidene]anabaseine N N N CH3 H 0.05 0.001 n=2 0.05 0.03 n=2 1 8.54 0.03 93.28 3.11 0.03 3-[4-Hydroxy benzylidene]anabaseine N N OH 0.36 0.02 n=2 0.45 0.04 n=2 1.25 7.22 0.05 39.79 2.29 0.01

PAGE 116

104 Compound Name Structure Ki ( M) 7 Ki ( M) 4 2 Ki 4 2 / 7 Ratio pKa % ionize d Log P 3-[4 Methoxy benzylidene]anabaseine N N OMe 0.57 0.03 n=2 0.49 0.03 n=2 0.86 Insufficient sample --Insufficient sample 3-[4-Methyl benzylidene]anabaseine N CH3 N 17.06 1.52 n=2 0.99 0.12 n=4 0.06 Insufficient sample --Insufficient sample 3-[4Methylthiobe nzylidene]anabaseine N N S CH3 0.25 0.003 n=2 0.90 0.03 n=2 3.6 7.49 0.03 55.14 3.17 0.57 3-[4-Cyano benzylidene]anabaseine N N CN 2.30 0.08 n=2 1.23 0.40 n=2 0.53 6.72 0.06 17.28 3.11 0.26 3-[4-Propyl benzylidene]anabaseine N N C3H7 0.28 0.02 n=2 0.89 0.13 n=2 3.18 7.39 0.04 49.34 2.82 0.34 3-[4Acetoxybenz ylidene]anabaseine N N O Ac 0.17 0.006 n=2 0.25 0.01 n=2 1.47 7.14 0.06 35.47 2.56 0.01

PAGE 117

105 Compound Name Structure Ki ( M) 7 Ki ( M) 4 2 Ki 4 2 / 7 Ratio pKa % ionize d Log P 3-[4Morpholino benzylidene]anabaseine N N N O 0.30 0.0002 n=2 0.28 0.09 n=2 0.93 7.97 0.04 78.79 2.72 0.06 3-[4Dimethyl amino benzylidene]anabaseine N N N CH3 C H3 0.08 0.0004 n=2 0.17 0.005 n=2 2.13 8.66 0.05 94.79 3.63 n=1 3-[4Difluorometh oxy benzylidene]anabaseine N N OCHF2 1.01 0.18 n=3 1.16 0.25 n=2 1.15 7.35 0.01 47.12 3.39 0.10 3-[4Chlorobenzyl idene]anabaseine N N Cl 1.12 0.19 n=2 0.97 0.38 n=2 0.87 7.18 0.04 37.60 3.50 0.47 3-[4-(1H1,2,4-Triazol1yl)benzyliden e]-anabaseine N N N N N 1.18 0.13 n=3 0.52 0.19 n=2 0.47 7.06 0.04 31.37 2.40 0.04 3-[4Acetamidobe nzylidene]anabaseine N N N H Ac 0.24 0.03 n=2 0.38 0.013 n=2 1.58 7.42 0.06 50.86 2.36 0.08

PAGE 118

106 Compound Name Structure Ki ( M) 7 Ki ( M) 4 2 Ki 4 2 / 7 Ratio pKa % ionize d Log P 3-{4-[bis(2Hydroxyethy l)amino]benz ylidene}anabaseine N N N OH O H 0.19 0.04 n=2 0.16 0.01 n=2 0.84 8.66 0.04 94.79 2.57 0.07 3-[4Pivaloyloxyb enzylidene]anabaseine N N O O C H3 CH3 CH3 0.69 0.11 n=2 0.83 0.25 n=2 1.20 7.12 0.03 34.42 3.09 0.08 3-[4Acetylbenzyl idene]anabaseine N N O C H3 0.92 0.04 n=2 4.49 2.3 n=2 4.88 6.83 0.04 21.21 2.51 0.07 3-[4-nButoxybenzy lidene]anabaseine N N O C4H9 0.91 0.02 n=2 1.96 0.83 n=2 2.15 7.67 0.04 65.06 2.68 0.50

PAGE 119

107 Compound Name Structure Ki ( M) 7 Ki( M) 4 2 Ki 4 2/ 7 Ratio pKa % ionized Log P 3-[2–Acetoxy benzylidene]anabaseine N N O O 0.57 0.04 n=2 0.70 0.36 n=2 1.23 6.28 0.10 7.03 2.95 0.23 3-[2-Trifluoromethoxy benzylidene]anabaseine N N CF3O 3.09 1.88 n=2 1.76 0.75 n=2 0.57 5.76 0.17 2.23 2.50 0.53 3-[2-Difluoro methoxy benzylidene]anabaseine N N F2HCO 1.07 0.04 n=4 1.41 0.01 n=2 1.32 6.94 0.06 25.75 3.07 0.36 3-[2Pivaloyloxybenzylidene]anabaseine N N C H3 CH3 O C H3 0.65 0.11 n=4 1.44 0.05 n=2 2.22 6.48 0.07 10.73 3.34 0.22

PAGE 120

108 Compound Name Structure Ki ( M) 7 Ki( M) 4 2 Ki 4 2/ 7 Ratio pKa % ionized Log P 3-[3Methoxybenzy lidene]anabaseine N N MeO 0.46 0.12 n=4 1.18 0.10 n=4 2.54 7.30 0.02 44.27 3.17 0.07 3-[3-Hydroxy benzylidene]anabaseine N N O H 0.15 0.02 n=2 1.17 0.10 n=2 7.80 7.16 0.09 36.58 2.72 0.03

PAGE 121

109 Compound Name Structure Ki ( M) 7 Ki ( M) 4 2 Ki 4 2/ 7 Ratio pKa % ionize d Log P 3-[2,4Dimethoxy benzyl]anabaseine N OMe N MeO 4.34 0.63 n=2 >20 n=2 --Can not be determi ned --W/O pKa Can not determi ne 3-[(2-Methoxy)4glucuronide benzylidene]anabaseine N N O O O H O H HOOC OH MeO 0.18 0.02 n=4 1.25 0.47 n=2 6.94 Can not be determi ned --Insufficient sample 3-[(2-Propoxy)-4methoxy benzylidene]anabaseine N N OMe O 0.07 0.03 n = 2 0.17 0.02 n=2 2.43 7.71 0.05 67.12 3.33 0.47 3-[(2-Pentoxy)-4methoxy benzylidene]anabaseine N N OMe O 0.05 0.002 n=2 0.43 0.16 n=2 8.60 7.48 0.08 54.59 3.67 0.36 3-[(2-Acetoxy)-4methoxybenzylid ene)-anabaseine N N O OMe O CH3 0.30 0.09 n=2 0.77 0.05 n=2 2.57 7.00 0.06 28.49 2.46 n=1 3-[(2-Hydroxy)4(diethylamino)b enzylidene]anabaseine N N O H N 0.05 0.002 n=2 0.13 0.04 n=2 2.60 7.51 0.12 56.29 2.96 0.29

PAGE 122

110 Compound Name Structure Ki ( M) 7 Ki( M) 4 2 Ki 4 2/ 7 Ratio pKa % ionized Log P 3-[2,6Dimethyl-4hydroxy benzylidene]anabaseine N N OH CH3 C H3 5.49 2.97 n=2 0.67 0.04 n=2 0.06 7.14 0.05 35.47 3.02 0.13 3-[3,5Dimethyl-4hydroxy benzylidene]anabaseine N N OH CH3 C H3 0.53 0.11 n=2 0.36 0.01 n=2 0.68 7.29 0.08 43.67 2.78 0.36 3-[2,4,6Trimethyl benzylidene]anabaseine N N CH3 CH3 C H3 1.51 0.11 n=3 0.52 0.22 n=2 0.34 7.19 0.03 38.10 Insufficient sample 3-[2,4,6Trimethoxy benzylidene]anabaseine N N OMe OMe MeO0.18 0.03 n=4 0.06 0.03 n=2 0.38 7.15 0.09 35.99 Could not obtain peak 3-[3,4,5Trimethoxy benzylidene]anabaseine N N OMe OMe MeO3.48 2.10 n=2 4.67 1.06 n=2 1.34 7.23 0.03 40.34 3.07 0.18

PAGE 123

111 Compound Name Structure Ki ( M) 7 Ki( M) 4 2 Ki 4 2/ 7 Ratio pKa % ionized Log P 3-[3,4Ethylenedio xybenzylide ne]anabaseine N N O O 0.20 0.01 n=2 0.52 0.01 n=2 2.60 7.56 0.03 59.11 3.16 0.05 3-[3,4(Methylened ioxy)benzyli dene]anabaseine O O N N 0.23 0.002 n=2 0.70 0.07 n=2 3.04 7.64 0.04 63.48 2.99 0.29 3-[(6Methoxynap hth-2yl)methylen e]anabaseine N N MeO0.37 0.03 n=2 0.60 0.10 n=2 1.62 7.29 0.12 43.67 3.38 0.40 3[(Benzofura n-2yl)methylen e]anabaseine O N N 0.15 0.003 n=2 5.92 1.06 n=2 39.50 7.11 0.03 33.90 3.86 0.40 Compound Name Structure Ki ( M) 7 Ki( M) 4 2 Ki 4 2/ 7 Ratio pKa % ionize d Log P Unsubstituted benzylidene anabaseine N N 0.60 0.08 n=2 0.85 0.24 n=2 1.42 7.34 0.04 46.55 3.21 0.20

PAGE 124

112 Compound Name Structure Ki ( M) 7 Ki( M) 4 2 Ki 4 2/ 7 Ratio pKa % ioniz ed Log P 3-[2,4Dimethoxy benzylidene]anabaseine N N OMe MeO0.13 0.01 0.25 0.04 1.9 7.89 0.05 75.56 4.10 0.27 3-[2Methoxy-4hydroxy benzylidene]anabaseine N N OH MeO0.24 0.01 0.07 0.030 0.29 7.09 0.04 32.87 2.46 0.10 3-[2Hydroxy-4methoxy benzylidene]anabaseine N N OMe O H 0.32 0.07 0.39 0.03 1.22 7.07 0.02 31.85 2.70 0.31 3-[2,4Dihydroxy benzylidene]anabaseine N N OH O H 0.20 0.01 0.21 0.016 1.05 7.41 0.06 50.58 1.75 0.14 (The data is the mean SE.)

PAGE 125

113 LIST OF REFERENCES Abdrakhmanova G, Dorfman J, Xiao Y, Morad M. (2002) Protons enhance the gating kinetics of the alpha3/beta4 neuronal ni cotinic acetylcholine receptor by increasing its apparent affinity to agonist s. Mol Pharmacol. 61(2):369-78. Adler LE, Hoffer LJ, Griffith J, Waldo MC, Freedman R. (1992) Normalization by nicotine of deficient auditory sensory gating in the relatives of schizophrenics. Biol Psychiatry. 32(7):607-16. Adler LE, Hoffer LD, Wiser A, Freedman R. (1993) Normalization of auditory physiology by cigarette smoking in schizophr enic patients. Am J Psychiatry. 150(12):1856-61. Albert, A. (1951) The Acridines . Edward Arnold, London, 2nd edition. Araujo DM, Lapchak PA, Robitaille Y, Ga uthier S, Quirion R. (1988) Differential alteration of various cholinergic markers in cortical and subc ortical regions of human brain in Alzheimer's diseas e. J Neurochem. 50(6):1914-23. Arendash GW, Sanberg PR, Sengstock GJ. ( 1995a) Nicotine enhances the learning and memory of aged rats. Pharmacol Biochem Behav. 52(3):517-23. Arendash GW, Sengstock GJ, Sanberg PR , Kem WR. (1995b) Improved learning and memory in aged rats with chronic admini stration of the nicotin ic receptor agonist GTS-21. Brain Res. 674(2):252-9. Barlow RB, Hamilton JT. (1962) Effects of pH on the activity of nicotine and nicotine monomethiodide on the rat diaphragm pr eparation. Brit. J. Pharmacol 18, 543-549. Baxter AJ, Dixon J, Ince F, Manners CN, Tea gue SJ. (1993) Discovery and synthesis of methyl 2,5-dimethyl-4-[2(phenylmethyl )benzoyl]-1H-pyrrole-3-carboxylate (FPL 64176) and analogues: the first examples of a new class of calcium channel activator. J Med Chem. 36(19):2739-44. Beers WH, Reich E. (1970) Structure and Activity of Acetylcho line. Nature. 228:917922. Bencherif M, Lukas RJ. (1991) Ligand bindi ng and functional char acterization of muscarinic acetylcholine receptors on the TE671/RD human cell line. J Pharmacol Exp Ther. 257(3):946-53.

PAGE 126

114 Bertrand S, Devillers-Thiery A, Palma E, Buisson B, Edelstein SJ, Corringer PJ, Changeux JP, Bertrand D. (1997) Paradoxical allosteric effect s of competitive inhibitors on neuronal alpha7 nicotinic receptor mutants. Ne uroreport. 8(16):35916. Braddy AC, Janaky T, Prokai L. (2002) Immobilized artificial membrane chromatography coupled with atmospheric pressure ionization mass spectrometry. J Chromatogr A. 966(1-2):81-7. Brejc K, van Dijk WJ, Klaassen RV, Schuurma ns M, van Der Oost J, Smit AB, Sixma TK. (2001) Crystal structure of an AChbinding protein reveal s the ligand-binding domain of nicotinic rece ptors. Nature. 411(6835):269-76. Briggs CA, McKenna DG, Monteggia LM , Touma E, Roch JM, Arneric SP, Gopalakrishnan M, Sullivan JP. (1999) Gain of function mutation of the alpha7 nicotinic receptor: distinct pharmacology of the human alpha7V274T variant. Eur J Pharmacol. 366(2-3):301-8. Celie PH, van Rossum-Fikkert SE, van Dijk WJ, Brejc K, Smit AB, Sixma TK. (2004) Nicotine and carbamylcholine binding to ni cotinic acetylcholine receptors as studied in AChBP crystal st ructures. Neuron. 41(6):907-14. Chavez-Noriega LE, Gillespie A, Stauderman KA, Crona JH, Claeps BO, Elliott KJ, Reid RT, Rao TS, Velicelebi G, Harpold MM, Johnson EC, Corey-Naeve J. (2000) Characterization of the recombinant human neuronal nicotinic acetylcholine receptors alpha3beta2 and alpha4beta2 stably expressed in HEK293 cells. Neuropharmacology. 39(13):2543-60. Chen D, Patrick W. (1997) The alpha-bunga rotoxin-binding nicotinic acetylcholine receptor from rat brain contains only the alpha 7 subunit. J Biol Chem. 19 (272):24024-24029. Clementi F, Fornasari D, Gotti C, eds. (2000) Neuronal Nicotinic Receptors – Handbook of Experimental Pharmacology. Vol 144. Corringer PJ, Le Novere N, Changeux JP. ( 2000) Nicotinic receptors at the amino acid level. Annu Rev Pharmacol Toxicol. 40:431-58. Couturier S, Bertrand D, Matter JM, Herna ndez MC, Bertrand S, Millar N, Valera S, Barkas T, Ballivet M. (1990) A neuronal nicotinic acetylcholine receptor subunit (alpha 7) is developmentally regulate d and forms a homo-oligomeric channel blocked by alpha-BTX. Neuron. 5(6):847-56. Craig PG, Zwart R, Bose S, Beattie RE, Folly EA, Johnson LR, Bell, E, Evans, NM, Volsen SG, Sher E., Millar NS and Broa d, LM. Construction and characterization of a chimeric human 7 nicotinic acetylcholine / mouse 5HT3 receptor. P008 abstract, Biochemical Society Meeting.

PAGE 127

115 Cromer BA, Morton CJ, Parker MW. (2002) Anxiety over GABA(A) receptor structure relieved by AChBP. Trends Biochem Sci. 27(6):280-7. Crum-Brown, A. and Fraser, T.R. (1868–9) On the connection between chemical constitution and physiological action. Part 1. On the physiologi cal action of the salts of the ammonium bases, derived fr om Strychnia, Brucia, Thebia, Codeia, Morphia, and Nicotia. Trans. R. Soc. Edinburgh 25, 151–203. Daly JW. (1995) Alkaloids from frog skin s: selective probes for ion channels and nicotinic receptors. Braz J Med Biol Res. 28(10):1033-42. Danielsson E, Eckernas SA, Westlind-Danielsson A, Nordstrom O, Bartfai T, Gottfries CG, Wallin A. (1988)VIP-sensitive ade nylate cyclase, guanylate cyclase, muscarinic receptors, choline acetyltransf erase and acetylcholinesterase, in brain tissue afflicted by Alzheimer's disease/se nile dementia of the Alzheimer type. Neurobiol Aging. 9(2):153-62. de Fiebre CM, Meyer EM, Henry JC, Muraskin SI, Kem WR, Papke RL. Characterization of a series of anabasei ne-derived compounds reveals that the 3(4)-dimethylaminocinnamylidine derivative is a selective agonist at neuronal nicotinic alpha 7/125I-alpha-bungarot oxin receptor subtypes. (1995) Mol Pharmacol. 47(1):164-71. Doody RS. Therapeutic standards in Alzheimer disease. (1999) Alzheimer Dis Assoc Disord. Suppl 2:S20-6. Drisdel RC, Green WN. (2000) Neuronal al pha-bungarotoxin receptors are alpha7 subunit homomers. J Neurosci. 20(1):133-9. Dukat M, Dowd M, Damaj IM, Martin BR, El-Zahabi MA, Glennon RA. (1998) Synthesis, receptor binding and QSAR studies on 6-substituted nicotine derivatives as cholinergic ligands. Eur J Pharmacol. 33:1-10. Fitch RW, Xiao Y, Kellar KJ, Daly JW. (2003) Membrane potential fluorescence: a rapid and highly sensitive assay for nicotinic receptor channel function. Proc Natl Acad Sci U S A. 100(8):4909-14. Flores CM, Rogers SW, Pabreza LA, Wolfe BB, Kellar KJ. (1992) A subtype of nicotinic cholinergic receptor in rat brain is compos ed of alpha 4 and beta 2 subunits and is up-regulated by chronic nicotine trea tment. Mol Pharmacol. 41(1):31-7. Francis PT, Palmer AM, Snape M, Wilcoc k GK. (1999) The cho linergic hypothesis of Alzheimer's disease: a review of progr ess. J Neurol Neurosurg Psychiatry. 66(2):137-47. Freedman R, Adler LE, Olincy A, Waldo MC, Ross RG, Stevens KE, Leonard S. (2002) Input dysfunction, schizotypy, and genetic models of schizophrenia. Schizophr Res. 54(1-2):25-32.

PAGE 128

116 Freedman R, Hall M, Adler LE, Leonard S. (1995) Evidence in postmortem brain tissue for decreased numbers of hippocampal nicotinic receptors in schizophrenia. Biol Psychiatry. 38(1):22-33. Gerzanich V, Peng X, Wang F, Wells G, An and R, Fletcher S, Lindstrom J. (1995) Comparative pharmacology of epibatidine: a potent agonist for neuronal nicotinic acetylcholine receptors. Mol Pharmacol. 48(4):774-82. Glennon RA, Dukat M. (2004) alpha4beta2 nACh receptor pharmacophore models. Bioorg Med Chem Lett. 14(8):1841-4. Gohlke H, Schwarz S, Gundisch D, Tilotta MC, Weber A, Wegge T, Seitz G. (2003) 3D QSAR analyses-guided rational design of novel ligands for the (alpha4)2(beta2)3 nicotinic acetylcholine recepto r. J Med Chem. 46(11):2031-48. Gotti C, Balestra B, Moretti M, Rovati GE, Maggi L, Rossoni G, Berti F, Villa L, Pallavicini M, Clementi F. (1998) 4-O xystilbene compounds are selective ligands for neuronal nicotinic alphaBungarotoxin receptors. Br J Pharmacol. 124(6):1197206. Gotti C, Hanke W, Schlue WR, Briscini L, Moretti M, Clementi F. (1992) A functional alpha-bungarotoxin receptor is present in chick cere bellum: purification and characterization. Neur oscience. 50(1):117-27. Gurley DA, Harris EW, Li C, Johnson EC, Lanthorn TH. (2000) 5-Hydroxyindole potentiates the nicotinic acetylcholine receptor 7 subtype. Society for Neuroscience Abstracts 26, 1916-. Han ZY, Zoli M, Cardona A, Bourgeois JP, Changeux JP, Le Novere N. (2003) Localization of [3H]nicotine, [3H]cytisine, [3H]epibatidine, and [125I]alphabungarotoxin binding sites in the brain of Macaca mulatta . J Comp Neurol. 461(1):49-60. Hansen SB, Radic' Z, Talley TT, Molles BE, Deerinck T, Tsigelny I, Taylor P. (2002) Tryptophan fluorescence reveal s conformational changes in the acetylcholine binding protein. J Biol Chem. 277(44):41299-302. Hansch C, Leo A, Hoekman D. (1995) E xploring QSAR: Hydrophobi c, Electronic, and Steric Constants. American Chemical Society. Harel M, Kasher R, Nicolas A, Guss JM, Ba lass M, Fridkin M, Smit AB, Brejc K, Sixma TK, Katchalski-Katzir E, Sussman JL, Fuchs S. (2001) The binding site of acetylcholine receptor as visualized in th e X-Ray structure of a complex between alpha-bungarotoxin and a mimotope peptide. Neuron. 32(2):265-75. Homer LD, Nielsen TB. (1987) Spare receptors , partial agonists, and ternary complex model of drug action. Am J Physiol. 253(1 Pt 1):E114-21.

PAGE 129

117 Jeng AY, Cohen JB. (1980) Agonists of Torpedo nicotinic receptors: es sential role of a positive charge. Ann N Y Acad Sci. 358:370-3. Jones IW, Barik J, O'Neill MJ, Wonnacott S. (2004) Alpha bungarotoxin-1.4 nm gold: a novel conjugate for visualising the precise subcellular distribution of alpha 7* nicotinic acetylcholine receptors . J Neurosci Methods.134(1):65-74. Kalderon D, Oostra BA, Ely BK, Smith AE. (1982) Deletion loop mu tagenesis: a novel method for the construction of point muta tions using deletion mutants. Nucleic Acids Res.10(17):5161-71. Karlin A. (2004) A touching picture of nicotinic binding. Neuron. 41(6):841-2. Karlin A. (2002) Emerging structure of the nicotinic acetylcholine receptors. Nat Rev Neurosci. 3(2):102-14. Kem WR. (1971) A study of the occurrence of anabaseine in Paranemertes and other nemertines. Toxicon. 9(1):23-32. Kem WR. (1997) Alzheimer's drug design based upon an invertebrate toxin (anabaseine) which is a potent nicotinic receptor agoni st. Invert Neuros ci. 3(2-3):251-9. Kem WR. (2000) The brain alpha 7 nicotinic receptor may be an important therapeutic target for the treatment of Alzheimer's disease: studies wi th DMXBA (GTS-21). Behav Brain Res. 113(1-2):169-81. Kem WR, Mahnir VM, Papke RL, Lingle CJ. (1997) Anabaseine is a potent agonist on muscle and neuronal alpha-bungarotoxin-sens itive nicotinic receptors. J Pharmacol Exp Ther. 283(3):979-92. Kem WR, Mahnir VM, Prokai L, Papke RL, Cao X, LeFrancois S, Wildeboer K, ProkaiTatrai K, Porter-Papke J, Soti F. (200 4) Hydroxy metabolites of the Alzheimer's drug candidate 3-[(2,4-dimethoxy)benzylid ene]-anabaseine di hydrochloride (GTS21): their molecular properties, interactions with brain nicotini c receptors, and brain penetration. Mol Pharmacol. 65(1):56-67. Kem WR, Scott KN, Duncan JH. (1976) Hoplonemertine worms – a new source of pyridine neurotoxins. Experientia. 32(6):684-6. Kim KH, Lin NH, Anderson DJ. (1996) Quanti tative structure-activity relationships of nicotine analogues as neuronal nicotinic acetylcholine receptor ligands. Bioorg Med Chem. 4(12):2211-7. Koike T, Martin DP, Johnson EM Jr. (1989) Role of Ca 2+ channels in the ability of membrane depolarization to prevent neuronal death induced by trophic-factor deprivation: evidence that levels of internal Ca 2+ determine nerve growth factor dependence of sympathetic ganglion cel ls. Proc Natl Acad Sci U S A. 86(16):6421-5.

PAGE 130

118 Kommalage M, Hoglund AU. (2004) (+/-) Epib atidine increases ace tylcholine release partly through an action on muscarinic re ceptors. Basic Clin Pharmacol Toxicol. 94(5):238-44. Liapakis G, Chan WC, Papadokostaki M, Javi tch JA. (2004) Synergis tic contributions of the functional groups of epinephrine to its affinity and efficacy at the beta2 adrenergic receptor. Mol Pharmacol. 65(5):1181-90. Levin ED, Rezvani AH. Nicotinic treatment for cognitive dysfunction. (2002) Curr Drug Targets CNS Neurol Disord. 1(4):423-31. Lorand T, Kocsis B, Sohar P, Nagy G, Jozsef P, Kispal G, Laszlo R, Prokai L. (2002) Synthesis and antibacterial activity of fused Mannich ketones. Eur J Med Chem. 37(10):803-12. Mahnir V, Lin B, Prokai-Tatrai K, Kem WR . (1998) Pharmacokinetics and urinary excretion of DMXBA (GTS-21), a co mpound enhancing cognition. Biopharm Drug Dispos. 19(3):147-51. Marks MJ, Collins AC. (1982) Characterizati on of nicotine binding in mouse brain and comparison with the binding of alpha-bunga rotoxin and quinuclid inyl benzilate. Mol Pharmacol. 22(3):554-64. Martin LF, Kem WR, Freedman R. (2004) Alph a-7 nicotinic receptor agonists: potential new candidates for the treatment of schizophrenia. Psychopharmacology (Berl). 174(1):54-64. Martin-Ruiz CM, Court JA, Molnar E, Lee M, Gotti C, Mamalaki A, Tsouloufis T, Tzartos S, Ballard C, Perry RH, Perry EK. (1999) Alpha4 but not alpha3 and alpha7 nicotinic acetylcholine receptor subunits are lost from the temporal cortex in Alzheimer's disease. J Neurochem. 73(4):1635-40. Mebs D, Narita K, Iwanaga S, Samejima Y, Lee CY. (1972) Purifi cation, properties and amino acid sequence of -bungarotoxin from the venom of Bungarus multicinctus. Hoppe Seylers Z Physiol Chem. 353(2):243-62. Meyer EM, Arendash GW, Judkins JH, Ying L, Wade C, Kem WR. (1987) Effects of nucleus basalis lesions on the muscarinic and nicotinic modulation of [3H]acetylcholine rele ase in the rat cerebral cort ex. J Neurochem. 49(6):1758-62. Meyer EM, King MA, Meyers C. (1998) Neuroprotective effects of 2,4dimethoxybenzylidene anabaseine (DMXB) and tetrahydroaminoacridine (THA) in neocortices of nucleus basalis lesion ed rats. Brain Res. 786(1-2):252-4. Meyer EM, Tay ET, Papke RL, Meyers C, Huang GL, de Fiebre CM. (1997) 3-[2,4Dimethoxybenzylidene]anabaseine (DMXB) selectively activates rat alpha7 receptors and improves memory-related behaviors in a mecamylamine-sensitive manner. Brain Res. 768(1-2):49-56.

PAGE 131

119 Misra A, Ganesh S, Shahiwala A, Shah SP. (2003) Drug delivery to the central nervous system: a review. J Phar m Pharm Sci. 6(2):252-73. Monash University. QSAR Determining the relationship between structure and activity. Moore MA, McCarthy MP. (1995) Snake venom toxins, unlike smaller antagonists, appear to stabilize a resting state conf ormation of the nicotinic acetylcholine receptor. Biochim Biophys Acta. 1235(2):336-42. Mullen G, Napier J, Balestra M, DeCory T, Ha le G, Macor J, Mack R, Loch J 3rd, Wu E, Kover A, Verhoest P, Sampognaro A, Phillips E, Zhu Y, Murray R, Griffith R, Blosser J, Gurley D, Machulskis A, Z ongrone J, Rosen A, Gordon J. (2000) (-)Spiro[1-azabicyclo[2.2.2]oc tane-3,5'-oxazolidin-2'-one], a conformationally restricted analogue of acetylch oline, is a highly selective full agonist at the alpha 7 nicotinic acetylcholine recepto r. J Med Chem. 43(22):4045-50. Nettleton J, Wang GK. (1990) pH-dependent binding of local an esthetics in single batrachotoxin-activated Na+ channels. Co caine vs. quaternary compounds. Biophys J. 58(1):95-106. Nicolotti O, Altomare C, Pellegrini-Calace M, Carotti A. (2004) Neuronal nicotinic acetylcholine receptor agonists: pharmac ophores, evolutionary QSAR and 3DQSAR models. Curr Top Med Chem. 4(3):335-60. Oldendorf W, Braun L, Cornford E. (1979) pH dependence of blood-brain barrier permeability to lactate and nicotine. Stroke. 10(5):577-81. Pabreza LA, Dhawan S, Kellar KJ. (1991) [3H]cytisine binding to nicotinic cholinergic receptors in brain. Mol Pharmacol. 39(1):9-12. Papke RL, Heinemann SF. (1994) Partial a gonist properties of cytisine on neuronal nicotinic receptors contai ning the beta 2 subunit. Mol Pharmacol. 45(1):142-9. Papke RL, Meyer EM, Lavieri S, Bollampally SR, Papke TA, Horenstein NA, Itoh Y, Porter Papke JK. (2004a) Effects at a dist ance in alpha 7 nAChR selective agonists: benzylidene substitutions that regulat e potency and efficacy. Neuropharmacology. 46(7):1023-38. Papke RL, Porter Papke JK, Rose GM. (2004b) Activity of alpha7-se lective agonists at nicotinic and serotonin 5HT3 receptors expressed in Xenopus oocytes. Bioorg Med Chem Lett. 14(8):1849-53. Papke RL, Schiff HC, Stokes C, Horenstein NA. (2004c) Molecular dissection of the 7 nAChR partial agonist, 5HT3 antagonist, tropisetron. Neuroscience abstract 956.20.

PAGE 132

120 Peng JH, Lucero L, Fryer J, Herl J, Leonard SS, Lukas RJ. (1999) Inducible, heterologous expression of human alpha7nicotinic acetylchol ine receptors in a native nicotinic receptor-null human cl onal line. Brain Res. 825(1-2):172-9. Petersson EJ, Choi A, Dahan DS, Lester HA, Dougherty DA. (2002) A perturbed pK(a) at the binding site of the nicotinic acetyl choline receptor: implications for nicotine binding. J Am Chem Soc. 124(43):12662-3. Picciotto MR, Zoli M, Rimondini R, Lena C, Marubio LM, Pich EM, Fuxe K, Changeux JP. (1998) Acetylcholine receptors containi ng the beta2 subunit are involved in the reinforcing properties of nico tine. Nature. 391(6663):173-7. Picciotto MR, Zoli M, Lena C, Bessis A, La llemand Y, Le Novere N, Vincent P, Pich EM, Brulet P, Changeux JP . (1995) Abnormal avoidance learning in mice lacking functional high-affinity nicotine recept or in the brain. Nature. 374(6517):65-7. Prince RJ and Stevens S. (1998) The lig and binding domains of the nicotinic acetylcholine receptor. In: Barrantes, F. editor. The Nicotinic Acetylcholine Receptor: Current Views and Future Trends. Springer-Verlag and R.G. Landes Company, p.33-39. Prokai L, Zharikova A, Janaky T, Li X, Br addy AC, Perjesi P, Matveeva L, Powell DH, Prokai-Tatrai K. (2001) Integration of mass spectrometry into early-phase discovery and development of central ne rvous system agents. J Mass Spectrom. 36(11):1211-9. Qian C, Li T, Shen TY, Libertine-Garahan L, Eckman J, Biftu T, Ip S. (1993) Epibatidine is a nicotin ic analgesic. Eur J Pharmacol. 250(3):R13-4. Quik M, Philie J, Choremis J. (1997) Modul ation of alpha7 nicotinic receptor-mediated calcium influx by nicotinic agonists. Mol Pharmacol. 51(3):499-506. Radcliffe KA, Dani JA. (1998) Nicotinic stimulation produces multiple forms of increased glutamatergic synaptic tr ansmission. J Neurosci. 18(18):7075-83. Reeves DC, Lummis SC. (2002) The molecular basis of the structure and function of the 5-HT3 receptor: a model ligand-gated i on channel (review). Mol Membr Biol. 19(1):11-26. Rueter LE, Anderson DJ, Briggs CA, Donnelly-Roberts DL, Gintant GA, Gopalakrishnan M, Lin NH, Osinski MA, Reinhart GA, Buckley MJ, Martin RL, McDermott JS, Preusser LC, Seifert TR, Su Z, Cox BF, Decker MW, Sullivan JP. (2004) ABT-089: pharmacological properties of a neuronal nicotinic acetylcholine receptor agonist for the potential treatm ent of cognitive disorders. CNS Drug Rev.10(2):167-82. Richon, AB. An introduction to QS AR Methodology. Network Science.

PAGE 133

121 Ripoll N, Bronnec M, Bourin M. (2004) Nicoti nic receptors and schizophrenia. Curr Med Res Opin. 20(7):1057-74. Seguela P, Wadiche J, Dinele y-Miller K, Dani JA, Patric k JW. (1993) Molecular cloning, functional properties, and dist ribution of rat brain alpha7: a nicotinic cation channel highly permeable to calcium. J Neurosci. 3(2):596-604. Severance E, Zhang H, Cruz Y, Pakhlevaniant s S, Hadley SH, Amin J, Wecker L, Reed C, Cuevas J. (2004) The {Alpha}7 Nicotinic Acetylcholine Receptor Subunit Exists in Two Isoforms that Contri bute to Functional Ligand-Gated Ion Channels*Mol Pharmacol. 66(3):420-9. Sharples CG, Wonnacott S. (2001) Tocris Literature: Neuronal Nicotinic Receptors. . October. Sharples CG, Karig G, Simpson GL, Spencer JA, Wright E, Millar NS, Wonnacott S, Gallagher T. (2002) Synthesis and pharm acological characterization of novel analogues of the nicotinic acetylcholin e receptor agonist (+/-)-UB-165. J Med Chem. (15):3235-45. Shimomura M, Yokota M, Okumura M, Mats uda K, Akamatsu M, Sattelle DB, Komai K. (2003) Combinatorial muta tions in loops D and F st rongly influence responses of the alpha7 nicotinic acetylcholine r eceptor to imidacloprid. Brain Res. 991(12):71-7. Silverman, RB. (1992) The Organic Chem istry of Drug Design and Drug Action. Academic Press Inc. San Diego. Sixma TK, Smit AB. (2003) Acetylcholine bi nding protein (AChBP): a secreted glial protein that provides a high-resolution model for the extracellular domain of pentameric ligand-gated ion channels . Annu Rev Biophys Biomol Struct. 32:31134. Smit AB, Syed NI, Schaap D, van Minnen J, Klumperman J, Kits KS, Lodder H, van der Schors RC, van Elk R, Sorgedrager B, Br ejc K, Sixma TK, Geraerts WP. (2001) A glia-derived acetylcholine-binding protei n that modulates synaptic transmission. Nature. 411(6835):261-8. Stevens KE, Kem WR, Mahnir VM, Freedma n R. (1998) Selective alpha7-nicotinic agonists normalize inhibition of a uditory response in DBA mice. Psychopharmacology. (Berl) 136(4):320-7. Sullivan JP, Decker MW, Brioni JD, D onnelly-Roberts D, Anderson DJ, Bannon AW, Kang CH, Adams P, Piattoni-Kaplan M, Buckley MJ. (1994) (+/-)-Epibatidine elicits a diversity of in vi tro and in vivo effects mediat ed by nicotinic acetylcholine receptors. J Pharmacol Exp Ther. 271(2):624-31.

PAGE 134

122 Sullivan JP, Decker MW, Donnelly-Roberts D, Brioni JD, Bannon AW, Holladay MW, Anderson DJ, Briggs CA, Williams M, Ar neric SP. (1995) Cholinergic channel activators: novel oppor tunities for the treatment of CNS disorders. Proc West Pharmacol Soc. 38:127-30. Summers KL, Kem WR, Giac obini E. (1997) Nicotinic agonist modulation of neurotransmitter levels in the rat frontopa rietal cortex. Jpn J Pharmacol. 74(2):13946. Tsuneki H, You Y, Toyooka N, Kagawa S, K obayashi S, Sasaoka T, Nemoto H, Kimura I, Dani JA. (2004) Alkaloids indolizidin e 235B', quinolizidine 1-epi-207I, and the tricyclic 205B are potent and selective noncompetitive inhibitors of nicotinic acetylcholine receptors. Mol Pharmacol. 66(4):1061-9. Unwin N. (2003) Structure and action of the nicotinic acetylcholine receptor explored by electron microscopy. FEBS Lett. 555(1):91-5. Wang H, Yu M, Ochani M, Amella CA, Tanovi c M, Susarla S, Li JH, Wang H, Yang H, Ulloa L, Al-Abed Y, Czura CJ, Tracey KJ . (2003) Nicotinic acet ylcholine receptor alpha7 subunit is an essen tial regulator of inflamma tion. Nature. 421(6921):384-8. Waterhouse RN. (2003) Determination of lipophili city and its use as a predictor of bloodbrain barrier penetration of molecular imaging agents. Mol Imaging Biol. 5(6):376-89. Wermuth, CG. (1996) The Practice of Medicinal Chemistry . (Chapter 17) Academic Press Limited, pgs. 312-344. Whiteaker KL, Davis-Taber R, Scott VE, G opalakrishnan M. (2001) Fluorescence-based functional assay for sarcolemmal ATP-sens itive potassium cha nnel activation in cultured neonatal rat ventricular my ocytes. J Pharmacol Toxicol Methods. 46(1):45-50. Whitehouse PJ, Martino AM, Antuono PG, Lowe nstein PR, Coyle JT, Price DL, Kellar KJ. (1986) Nicotinic acetylch oline binding sites in Alzhei mer's disease. Brain Res. 371(1):146-51. Wolff ME, ed. (1980) BurgerÂ’s Medicinal Chemistry . 4th edition, Part 1, p.343. Woodruff-Pak DS, Li YT, Ke m WR. (1994) A nicotinic ag onist (GTS-21), eyeblink classical conditioning, and nicotinic recep tor binding in rabbit brain. Brain Res.645(1-2):309-17. Xu R, Dwoskin LP, Grinevich VP, Deaciuc G, Crooks PA. (2001) Neuronal nicotinic acetylcholine receptor binding affinities of boron-containing nicotine analogues. Bioorg Med Chem Lett. 11(9):1245-8.

PAGE 135

123 Zhang ZW, Vijayaraghavan S, Berg DK. (1994) Neuronal acetylcholine receptors that bind alpha-bungarotoxin with high affinity function as ligand-gated ion channels. Neuron. 12(1):167-77. Zoltewicz JA, Prokai-Tatrai K, Bloo m LB, and Kem WR. (1993) Long range transmission of polar effects in choliner gic 3-arylidene anabaseines. Conformations calculated by molecular modelli ng. Heterocycles. 35: 171-179. Zucker RS. (1996) Exocytosis: a molecula r and physiological perspective. Neuron. 17(6):1049-55. Zwart R, De Filippi G, Broad LM, McPhie GI, Pearson KH, Baldwinson T, Sher E. (2002) 5-Hydroxyindole potentiates human alpha 7 nicotinic receptor-mediated responses and enhances acet ylcholine-induced glutamat e release in cerebellar slices. Neuropharmacology. 43(3):374-84. Zwart R, van Kleef RG, Gotti C, Smulders CJ, Vijverberg HP.(2000) Competitive potentiation of acetylcholine effect s on neuronal nicotinic receptors by acetylcholinesterase-inhibiting drugs. J Neurochem. 75(6):2492-500.

PAGE 136

124 BIOGRAPHICAL SKETCH Susan Elizabeth LeFrancois was born July 20, 1978 in Terre Haute, Indiana to Dennis Henry Michalski and Ma rgaret Eleanor Michalski. She spent her childhood in Lakeland, Florida. She attended University of South Florida wher e she completed her undergraduate degree in Chemistry. Upon gra duation she entered the Interdisciplinary Program in Biomedical Sciences at the Univer sity of Florida College of Medicine. She joined the laboratory of Dr. William R. Ke m in the department of Pharmacology and Therapeutics to explore her interests in ne w drug treatments for cognitive disorders. Susan is the proud wife of Benjamin Ja meson LeFrancois who graduated in 2002 from the University of Florida College of Law. Her husband is a successful attorney in Lakeland. Susan and Ben are also the happy owners of two disobe dient and beautiful beagles, Jerry and Buddy. After graduation, Susan hopes to start a fam ily and continue her research at the University of South Florida in Tampa, Florida.