THE ANTI-EPILEPSY POTENTIAL AND ANTAGONISM OF THE N-METHYL-D-ASPARTATE RECEPTOR BY DIAMINODIPHENYLS

Epilepsy is a disorder characterized by the occurrence of seizures, which are periods of abnormally excessive synchronous neuronal activity in the brain. Affecting over 70 million people worldwide, many of whom do not respond to pharmacotherapy, there is a need for novel anticonvulsant compound discovery. Diaminodiphenyl compounds, a class of compounds shown to present anticonvulsive effects in vivo have been purported to exert their effects on the N-methyl-D-aspartate receptor (NMDAr); an excitatory, ionotropic receptor that is a key player in the functions of the glutamatergic system. The glutamatergic system is vital in the promotion of synaptic plasticity and has been implicated in a myriad of mental health issues, including epilepsy. We hypothesize that diaminodiphenyl compounds interact with NMDAr at an allosteric binding site on the NR2 subunit which is activated by the agonist glutamate. This project is intended to characterize the diaminodiphenyl binding interactions with NMDAr, elucidate a structure activity relationship between diaminodiphenyl compounds and NMDAr and create novel diaminodiphenyl compounds employing rational drug design to improve the therapeutic index of the compounds. The data presented in the following manuscripts characterizes a novel binding motif between diaminodiphenyl compounds and NMDAr using computer based modeling techniques. A structure activity relationship was derived by examining the anticonvulsant effects of several different diaminodiphenyl compounds in animal models of epilepsy. Employing the computationally derived binding motif and structure activity relationship, several diaminodiphenyl derivatives have been designed in an effort to alleviate metabolic toxicity and improve anticonvulsive potency. Rational drug design targeting the described diaminodiphenyl binding site could offer novel anticonvulsants which act through a novel mechanism on NMDAr. Antagonism of NMDAr is not limited to epilepsy alone, NMDAr has been implicated in a number of disease states and diaminodiphenyls could serve multiple indications such as: neuropathy, mood disorders and post stroke outcomes.


LIST OF TABLES
Nevertheless, a number of literature reviews of clinical studies have concluded that, even with the availability of these new AEDs, the proportion of patients who are seizure free after drug therapy amounts to only about 30%. 8 In addition, the adverse effects of AEDs can be serious enough to negatively affecting a patient's quality of life, leading to issues surrounding patient compliance with the medication dosing regimen. 9 These side-effects include dizziness, drowsiness, mental slowing, weight gain, metabolic acidosis, nephrolithiasis, glaucoma, skin rash as well as movement and behavioral disorders, among many others. 9,10 The problems posed by the side-effects are compounded by the requirement for multiple-drug therapy for adequate seizure control in many patients. 11 There remains a pressing need for the development additional AEDs with novel mechanism of action that can provide patients and physicians with alternative treatment options. Furthermore, it is clear that numerous and diverse underlying pathophysiological mechanisms contribute to the epileptic syndrome. Therefore, access to a spectrum of AEDs, with different modes of action, might facilitate the optimization of drug therapy to suit the pharmacological and toxicological profiles of individual patients so as to improve tolerability and long-term treatment success. The global AED market is a multibilliondollar enterprise that is still growing and the market for pharmaceutical treatment of epilepsy generated $12 billion in 2008. 12 Given this background, the development of an AED with a novel mechanism of action is likely to be economically successful if it can show superior efficacy, even in a subset of patients. While conventional antiepileptic drugs have acted by blocking sodium channels or enhancing the function of γaminobutyric acid (GABA), the major inhibitory neurotransmitter in the brain. 11 The newer drugs have more novel actions including binding to presynaptic vesicle proteins and calcium channel subunits.
N-methyl-D-aspartate receptors (NMDARs) are important excitatory receptors and play a key role in the pathophysiology of several neurological diseases, including epilepsy, stroke, traumatic brain injury, dementia and schizophrenia which make it an ideal candidate as an anticonvulsant drug target. 13 The NMDAR is a hetero-tetrameric cation channel formed as a complex between two NR1 and two NR2 subunits. 14 (Fig 1). Furthermore, the model also discriminated between active and inactive compounds as defined by previous in-vivo studies.
To investigate whether the interaction between NMDA receptors and TDA predicted by in the in silico modeling studies could result in functional modulation of NMDAR activity, we functionally expressed NMDAR in Xenopus oocytes by co-injecting cRNA for the most ubiquitously expressed NMDA channel subunits NR1A and NR2B. 4-5 days after cRNA injection, electrophysiological recordings using the two electrode voltage clamp configuration were used to characterize NMDAR responses. Oocytes were clamped at -70 mV and perfused with 1 µM glycine and 50 µM glutamate which resulted robust positive inward currents only in oocytes injected with both channel subunits. Coapplication of TDA with glutamate inhibited the glutamate-elicited NMDAR conductance (Fig 2). NMDAR inhibition was readily reversible upon TDA wash out.
Rational drug design may be an effective approach for producing potent targeted drug therapies and alleviate know toxic mechanisms. In order to transition from inefficient high throughput screening protocol to a systematic rational drug design protocol a well defined binding pocket, key amino acids and binding interactions must be identified. By combining a full range of methods, including flexible side chain models and homology modeling, with the established electrophysiology based techniques, it is hypothesized that the binding pocket and key interactions between DADPs and NMDAR can be elucidated.
NMDAR is a complex receptor whose dysfunction in noted in a variety of disorders.
This study proposes to characterize a novel binding site on NMDAR which may prove to be a viable druggable target. The resulting data could be used to develop new lead compounds and provide viable treatment options for epilepsy and a variety of other glutamate related disorders.

Introduction
The N-Methyl D-aspartate receptor (NMDAr), a ligand gated non-specific cation channel, is essential to neuronal plasticity, chronic pain, and may also be involved in seizure development 1  Diaminodiphenyl sulfide has been shown as a potent anticonvulsant. However, the potential toxicity of any aromatic amine must always be addressed before a compound could be investigated as an investigational API. Based on the toxic mechanism of aniline it can be assumed that N-oxidation will lead to the production of a key toxic metabolite which can be further metabolized leading to blood toxicity and DNA adduct formation 7 (Fig 1). In order to reduce the proposed toxic effects of diaminodiphenyl sulfide, rational drug design must be used to eliminate the potential for the formation of the key hydroxyl amine metabolite.

In Silico Ligands
Ligands were created using ChemBio3D Ultra 12. All ligands were drawn using program defaults, and then the energy of each ligand was minimized using MM2 energy minimization function before being saved in a format suitable for docking in AutoDock4.2. (Fig 2). In Silico Docking to NMDA NR2a

In Silico
Grid Box Selection The grid box was initially centered on the ligand glutamate and further refined, using the coordinates determined previously in Discovery Studio 3.1, in order to encompass the cavity of the protein in which glutamate binding was characterized.
The grid was centered at (21.5, 21.4, 36.1) with dimensions of: Grid map x-dimension : 22.5 Angstroms Grid map y-dimension : 22.5 Angstroms Grid map z-dimension : 24.0 Angstroms The gridbox was overlaid on the water-free ligand-free structure 2A5S produced in Discovery Studio 3.1 and processed using default AutoGrid parameters in AutoDockTools 4.2.

AutoDock Parameters
AutoDock parameters were held constant when docking each ligand with the prepared NMDAr NR2a structure 2A5s. Each ligand was docked starting from a computer generated seed site to the protein within the gridbox previously described.
Using default Lamarckian Algorithm settings set to default short run in AutoDock 4, the energy was minimized. Optimum binding energy and ligand conformation was then determined and recorded by AutoDock. (Table 1) This process was repeated 256 times for each ligand and the data was compiled. The conformations are then grouped based on similar location and orientation with a 2 angstrom deviation to the mean of the population as the grouping criteria.

Results and Discussion
Diaminodiphenyl compounds appear to bind to NMDAr NR2A subunit in a specific orientation, while hydrogen bonding with THR174 which may contribute to the reported NMDAr effects in epilepsy. (Fig 4) The ligands evaluated addressed several key variables which may affect the interactions between ligand and NMDAr; they were: linker element, linker substitution, linker length, N-substitution, aromatic substitution and aromaticity.
The diaminodiphenyl derivatives evaluated are rationally designed to prevent the toxic metabolism of the diaminodiphenyls. These derivatives examine the effects of new functional groups and availability of H-bond donors/acceptors on NMDAr interaction.
In an effort to elucidate the binding interactions of diaminodiphenyl compounds with NMDAr the crystal structure of the NR2A subunit of NMDA was examined and the agonist, glutamate, binding site was identified and the binding cavity was characterized. N-substitution resulting in secondary and tertiary amines does not appear to significantly affect the binding energy of the diaminophenyl compounds. (Fig 10) The     C .

I .
A .

K .
A .

L .
A .

N .
A .        climbed to only about 30%. 5 In addition, adverse effects of AEDs can be serious enough to negatively affecting a patient's quality of life leading to issues surrounding patient compliance with the medication dosing regimen. 6 These side-effects include dizziness, drowsiness, mental slowing, weight gain, metabolic acidosis, nephrolithiasis, glaucoma, skin rash as well as movement and behavioral disorders among many others. 6,7 The problems posed by the side-effects are compounded by the requirement for multiple-drug therapy for adequate seizure control in many patients. 8 Thus there is still a pressing need for the development additional AEDs with novel mechanism of action that can provide patients and physicians with alternative treatment options. Furthermore, it is clear that numerous and diverse underlying pathophysiological mechanisms contribute to the epileptic syndrome. Therefore, access to a spectrum of AEDs, with different modes of action, will allow for optimizing therapy to suit the pharmacological and toxicological profiles of individual patients so as to improve tolerability and long-term treatment success. The global AED market is a multibilliondollar enterprise that is still growing and the market for pharmaceutical treatment of epilepsy generated $12 billion in 2008. 9 Given this background, the development of an AED with a novel mechanism of action is likely to be economically successful if it can show superior efficacy even in a subset of patients.

4,4'-ethylenedianiline
Diaminodiphenyl compounds have been characterized as potent anticonvulsants with little notable acute toxicity. 10 Though the activity of diaminodiphenyl compounds has been described, little is known about the mechanism of action of these compounds. In order to better understand the mechanism by which diaminodiphenyl compounds exert their anticonvulsive properties, a series of diaminodiphenyl compounds were examined in animal models of epilepsy in an attempt to elucidate a structure activity relationship. With a majority of the compounds tested presenting anticonvulsive properties at similar doses, it is also beneficial to explore the link between diaminodiphenyl structure and toxicity. (Table 3 Table 2. 6Hz model of epilepsy in mice. Results displayed show the lowest dose at which neuroprotection was observed, as well as, the dose required to exhibit neuroprotection for the duration of the experiment. All doses were delivered by IP injection.   Upon further evaluation in computer based models and results gathered in animal models of epilepsy, it was determined that hydrogen bonding appears to be an important factor in a diaminodiphenyl compound's anticonvulsive potential. For this reason the consideration and synthesis of several proposed azacyclo derivatives, was discontinued ( Fig 5).
Chemical characterization of the diacetyl-derivative of diamindiphenyl sulfide revealed product decomposition. Unlike previously characterized diaminodiphenyl derivatives, product separation was not easily accomplished by flash chromatography using a silica stationary phase. Using liquid chromatography and solid phase extraction techniques the mono-and di-substituted derivatives of diaminodiphenyl sulfide were separated across a C-18 column in acetonitrile/ 0.1% aqueous acetic acid mobile phase (Appendix 1). The mass spectrometry data suggests that the compound may be breaking down from the diacetylated derivative to the monoacetylated derivative of diaminodiphenyl sulfide (Fig 6). The extracted ion trace of the diacetyl-derivative of diaminodiphenyl sulfide, purified by solid phase extraction, support the conclusion that the stability of N,N'-diacetyl-diaminodiphenyl sulfide needs to be evaluated further. The LCMS data shows a very small, though detectable, peak with a higher retention time in SPE purified product, which can be characterized as the monoacetyl-derivative (Fig 7).
Further method development and compound stability studies would be necessary to create a stable diacetyl-derivative of diaminodiphenyl sulfide. The ruthenium catalyzed reaction between dibromoethane and diaminodiphenyl sulfide (Scheme 4) also produced zero yield of the diindole derivate. However, the reaction did take place using this method. The first step of the reaction may have proceed resulting in the formation of N-bromoethane substituted diaminodiphenyl sulfide derivative. The mass spectroscopy data suggests that the intended indole product was not formed (Fig 8). The mono-substituted product is likely a result of the molar ratio of the reagents, which is low in order to prevent piperazine polymer formation.        The separation of the mono-and di-acetyl products of diaminodiphenyl sulfide was not completed successfully by silica based flash chromatography, as was the case with other diaminodiphenyl derivatives were. The use of a C-18 column was necessary and an appropriate gradient method was elucidated in order to efficiently separate the products.
All liquid chromatographic separations were performed on a Thermo Scientific Accela pump using a Thermo Scientific ODS Hypersil C-18 2.1x100 mm, 5um column. Initial studies using a neutral aqueous mobile phase resulted in poor compound separation. Addition of 0.1% acetic acid separated mono-and di-acetyl derivatives of diaminodiphenyl efficiently with a 7 minute difference in retention time.
Solid phase extraction across Waters Sep-Pak C-18 syringe cartridges was considered as a means to readily purify the desired diacetylated product. Separations were completed using manufacturers recommended procedures. Cartridges were conditioned with 500µL 100% acetonitrile, followed by 500 µL 100% 0.1% aqueous acetic acid. Samples were taken up in 0.1% aqueous acetic acid, 200 µL and injected onto the cartridge. The cartridge was then washed with 200 µL 0.1% aqueous acetic acid. The product was then eluted in 3 -300µL fractions of 40% acetonitrile in 0.1% aqueous acetic acid.
Each fraction was evaluated by LCMS and it was determined that the first two fractions contained the best ratio of diacetyl-to monoacetyl-diaminodiphenyl sulfide.

Mass Spectrometer Conditions
Thermo Scientific Electron Exactive Orbitrap Mass Spectrometer was run in ESI positive ion mode flowing ultra pure nitrogen gas. All acquisitions and data analysis were completed using Thermo Scientific XCalibur Software Version 2.2 SP1.48. The results of this study confirm the assertion the diaminodiphenyl compounds modulate the functionof the N-methyl-D-aspartate receptor. Antagonism of this receptor may be the mechanism by which diaminodiphenyl compounds exert their anticonvulsive properties.