Elsevier Editorial System(tm) for European Journal of Medicinal Chemistry Manuscript Draft Manuscript Number: Title: Discovery of non-oxime reactivators using an in silico pharmacophore model of reactivators for tabun-inhibited acetylcholinesterase. Article Type: Original Paper Keywords: In silico pharmacophore model, virtual screening, WRAIR-CIS database, non-oxime reactivators, DFP-inhibited AChE Corresponding Author: Dr. Apurba Krishna Bhattacharjee, Ph.D. Corresponding Author's Institution: Walter Reed Army Institute of Research First Author: Apurba K Bhattacharjee, Ph.D. Order of Authors: Apurba K Bhattacharjee, Ph.D.; Apurba Krishna Bhattacharjee, Ph.D.; Elizabeth Marek, M.S.; Ha T Le; Ruthie Ratcliff; James C DeMar; Dmitry Pervitsky; Richard K Gordon Abstract: Utilizing our previously reported in silico pharmacophore model for reactivation efficacy of oximes, we present here a discovery of twelve new non-oxime reactivators of DFP (diisopropylfluorophosphate)-inhibited acetylcholinesterase (AChE). The non-oximes were obtained through virtual screening of an in-house database that showed rate constant (kr) efficacy values within ten-fold of pralidoxime (2-PAM) in an in vitro assay and one them showed in vivo efficacy comparable to 2-PAM against brain symptoms for DFP-induced neuropathology in guinea pigs. Short listing of the identified compounds were performed on the basis of fit score to the pharmacophore model, conformational energy requirement for the fit, and in silico evaluations for favorable blood brain barrier penetrability, octanol-water partition (log P), toxicity (rat oral LD 50) and binding affinity to the active site of the crystal structure of inhibited AChE.
Cover Letter
DEPARTMENT OF THE ARMY WALTER REED ARMY INSTITUTE OF RESEARCH 503 ROBERT GRANT AVENUE, ROOM 2S04 SILVER SPRING MARYLAND 20910-7500
Dr. Apurba K. Bhattacharjee Ex-Chief Molecular Modeler Department of Regulated Laboratories Division of Regulated Activities Walter Reed Army Institute of Research 503 Robert Grant Avenue Silver Spring, MD 20910-7500 Email:
[email protected]
August 26, 2013
The Editor-in-Chief European Journal of Medicinal Chemistry
Dear Sir, I have attached online a manuscript titled “Discovery of non-oxime reactivators using an in silico pharmacophore model of reactivators for tabun-inhibited acetylcholinesterase – A.K. Bhattacharjee, E. Marek, H.T. Le, R. Ratcliff, J.C. DeMar, D. Pervitsky, R.K. Gordon” for your kind perusal and consideration regarding its publication in the esteemed European Journal of Medicinal Chemistry. The manuscript embodies a rational strategy for discovery of novel non-oxime reactivators for organophosphate (OP) inhibited AChE (acetylcholinesterase). Utilizing our previously reported in silico pharmacophore model and discovery of non-oxime reactivators (EJMC 49 (2012) 229-238), we present here a discovery of twelve new non-oxime reactivators of DFP (diisopropylfluorophosphate)-inhibited acetylcholinesterase (AChE) that not only showed rate constant (kr) efficacy values within ten-fold of pralidoxime (2-PAM) in an in vitro assay but one them showed in vivo efficacy comparable to 2-PAM against brain symptoms for DFP-induced neuropathology in guinea pigs. These non-oximes were never reported before to have reactivation efficacy against any OP-inhibited AChE. To assist you in choosing referees for review of the manuscript, I have attached the following list of individuals who may be considered: 1. Prof.dr.Kornelia Tekes Semmelweis University Department of Pharmacodynamics Nagyvarad ter 4. H-1089 Budapest, Hungary Email:
[email protected]
REPLY TO ATTENTION OF
2. Dr Pascal Houzé Biochemistry laboratory St Louis Hospital 1 avenue C. Vellefaux 75010 Paris, France, Email:
[email protected] 3. Professor Ramesh C. Gupta Murray State University Hopkinsville, KY 42241-2000 (U.S.A.) Email:
[email protected]. 4. Dr Alex Piao CHIN Senior Member of Technical Staff Chemical Biological Radiological & Explosives Programme DSO National Laboratories 11, Stockport Road, Block 6, Singapore 117605 Tel: +65-6871-2143 Fax: +65-6873-0742 Email:
[email protected]. 5. Dr. Suman S. Thakur, Ph.D Proteomics and Cell Signaling, Lab E409 Centre for Cellular & Molecular Biology Habsiguda, Uppal Road Hyderabad - 500 007, India Email :
[email protected];
[email protected] Should you require any further assistance, please let me know.
Best regards. Sincerely, Apurba K Bhattacharjee
*Graphical Abstract (for review) Click here to download Graphical Abstract (for review): Bhattacharjee etal EJMC-2013 Graphical Abstract.docx
Discovery of non-oxime reactivators using an in silico pharmacophore model of reactivators for tabun-inhibited acetylcholinesterase.
Apurba K. Bhattacharjeea,*, Elizabeth Mareka, Ha Thu Lea, James C. DeMara, Ruthie Ratcliffa, Dmitry Pervitskyb, and Richard K. Gordona a
Department of Regulated Laboratories, Division of Regulated Activities, Walter Reed Army Institute of Research, Silver Spring, MD 20910, U.S.A. b Division of Experimental Therapeutics, Walter Reed Army Institute of Research, Silver Spring, MD 20910, U.S.A. c U.S. Army Medical Research and Material Command, Fort Detrick, MD 21702, U.S.A.
Virtual screening of WRAIR-CIS database with the generated in silico pharmacophore model led us to identify 12 non-oxime reactivators against DFP-inhibited AChE showing efficacy within 10-fold of 2-PAM, one of which showed in vivo efficacy comparable to 2-PAM against brain symptoms for DFP-induced neuropathology in guinea pigs.
*Highlights (for review)
Discovery of non-oxime reactivators using an in silico pharmacophore model of reactivators for tabun-inhibited acetylcholinesterase. Apurba K. Bhattacharjeea,*, Elizabeth Mareka, Ha Thu Lea, Ruthie Ratcliff a , James C. DeMara, Dmitry Pervitskyb, and Richard K. Gordona,c,# a Department of Regulated Laboratories, Division of Regulated Activities, Walter Reed Army Institute of Research, Silver Spring, MD 20910, U.S.A. b Division of Experimental Therapeutics, Walter Reed Army Institute of Research, Silver Spring, MD 20910, U.S.A. c U.S. Army Medical Research and Material Command, Fort Detrick, MD 21702, U.S.A.
Highlights: The manuscript embodies a rational strategy for discovery of novel non-oxime reactivators for organophosphate (OP) inhibited AChE (acetylcholinesterase). Utilizing our previously reported in silico pharmacophore model, we present here a discovery of 12 new non-oxime reactivators of DFP (diisopropylfluorophosphate)inhibited AChE These non-oximes showed rate constant (kr) efficacy values within ten-fold of pralidoxime (2-PAM) in an in vitro assay. One them showed in vivo efficacy comparable to 2-PAM against brain symptoms for DFP-induced neuropathology in guinea pigs. These non-oximes were never reported before to have reactivation efficacy against any OP-inhibited AChE.
*Manuscript Click here to view linked References
Discovery of non-oxime reactivators using an in silico pharmacophore model of reactivators for tabun-inhibited acetylcholinesterase.
Apurba K. Bhattacharjeea,*, Elizabeth Mareka, Ha Thu Lea, Ruthie Ratcliff a , James C. DeMara, Dmitry Pervitskyb, and Richard K. Gordona,c,# a
Department of Regulated Laboratories, Division of Regulated Activities, Walter Reed Army Institute of Research, Silver Spring, MD 20910, U.S.A. b Division of Experimental Therapeutics, Walter Reed Army Institute of Research, Silver Spring, MD 20910, U.S.A. c U.S. Army Medical Research and Material Command, Fort Detrick, MD 21702, U.S.A.
*Corresponding author: Tel: 301-438-8196 Email address:
[email protected] or #Alternative corresponding author
[email protected] (Richard K. Gordon)
1
ABSTRACT Utilizing our previously reported in silico pharmacophore model for reactivation efficacy of oximes, we present here a discovery of twelve new non-oxime reactivators of DFP (diisopropylfluorophosphate)-inhibited acetylcholinesterase (AChE). The non-oximes were obtained through virtual screening of an in-house database that showed rate constant (kr) efficacy values within ten-fold of pralidoxime (2-PAM) in an in vitro assay and one them showed in vivo efficacy comparable to 2-PAM against brain symptoms for DFP-induced neuropathology in guinea pigs. Short listing of the identified compounds were performed on the basis of fit score to the pharmacophore model, conformational energy requirement for the fit, and in silico evaluations for favorable blood brain barrier penetrability, octanol-water partition (log P), toxicity (rat oral LD 50) and binding affinity to the active site of the crystal structure of inhibited AChE. Keywords: In silico pharmacophore model, virtual screening, WRAIR-CIS database, nonoxime reactivators, DFP-inhibited AChE
2
1. Introduction Organophosphorus (OP) compounds such as tabun, soman, DFP, sarin, cyclosarin and pesticides (paraoxon, chlorpyrifos, tetraethyl pyrophosphate: TEPP) are highly toxic group of compounds [1]. Due to their high intrinsic toxicity, these agents pose potential threats for warfare in battlefields, terrorist attacks and sudden accidents due to potentially demilitarization efforts of an army unit [2]. For example, VX (O-ethyl S-[2-(diisopropylamino) ethyl] methylphosphonothioate) was used in Iran-Iraq war) and sarin was used by terrorists in Tokyo subway [2]. The acute toxicity of OP compounds in mammals is due to the inhibition of acetylcholinesterase (AChE; EC 3.1.1.7). AChE hydrolyzes the neuron-mediator acetylcholine (ACh) at the synaptic clefts. Inhibition of the ChE results in a cholinergic crisis by building up more and more AChE which can ultimately lead to death if untreated. AChE is a serine hydrolase responsible for hydrolyzing (removing) the neurotransmitter acetylcholine in the synaptic junction of nerve terminals in humans and animals. The enzyme has a catalytic triad consisting of Ser203, His447, and Glu334 at the active site in a narrow deep gorge (~ 20 Å), the lining of which contains mostly aromatic residues that form a narrow entrance to the catalytic Ser203 [3]. A peripheral anionic site (PAS) comprising another set of aromatic residues, Tyr72, Tyr124, Trp286, and Tyr341 and the acidic Asp74 [4], is located at the rim of this gorge. They provide a binding site for allosteric modulators and inhibitors. AChE is inhibited by a phosphorous group originating from the OP that is conjugated to the catalytic serine residue at the active site [3].
3
To counter the effects of OP poisoning, anticholinergics (as functional drugs) and AChE reactivators (as causal drugs) are usually used as first step for treatments [1, 5, 6]. Mechanism of the reactivation is believed to proceed through a rapid hydrolysis of the phosphorylated esteratic site of AChE. The conjugate of the inhibited AChE may either undergo a process known as “aging” via an elimination reaction involving dealkylation/deamidation, or prior to the “aging”, reactivation by nucleophiles such as oximes [7, 8]. These compounds contain a quaternary nitrogen atom which promotes binding in the catalytic site of the AChE. Oximes function as nucleophiles to displace the phosphate moiety of the OP’s which reacted with the enzyme’s active site serine hydroxyl group, thereby reactivating AChE [5]. Oxime based cholinesterase reactivators are known in the literature [1, 2, 9] since the 1950’s documenting both in vitro and in vivo efficacy against OP and pesticide poisoned cholinesterases in animals and humans. Although oximes have been investigated for many years for treatments of OP agent poisoning, only five pyridinium oximes, 2-PAM, TMB-4, HI-6, obidoxime, and MMC-4 have been clinically considered for standard treatments so far [1]. This includes treatments for both insecticides and nerve agents, usually in combination with the muscarinic ACh receptor antagonist, atropine, to block the overstimulation of cholinergic receptors by ACh. However, these oximes have several disadvantages including toxicity, CNS penetrability and limited reactivation efficacy against the nerve agent tabun [8, 9]. In fact, none of the above oximes can be regarded as a broad spectrum antidote to all nerve agents. Despite many efforts for discovery of improved reactivation therapeutics, little success has been made so far to find truly effective broad-spectrum AChE reactivators based on oxime chemistry and structure theory. Efforts for reactivator discovery have only led to the development of derivatives of preexisting oxime chemical structures. Compounds from non4
oxime chemical classes having good blood brain barrier (BBB) penetrability have barely been explored [10, 11]. Although to date a few BBB-penetrable oximes have been reported in the literature, none of them were further developed because of low functional activity and/or significant toxicity [12]. Oximes devoid of charges, such as monoisonitrosoacetone (MINA) and diacetylmonooxime (DAM) were reported to have better BBB penetration abilities, but these compounds showed much lower reactivation in blood and peripheral tissue ChEs compared to 2PAM and other quaternary oximes [13, 14]. An alternative approach reported to circumvent the problem was through use of pro-drugs that undergo oxidation in vivo in the CNS to produce an active quaternary oximes [15, 16]. However, the pro-drug approach has several disadvantages including difficulty of synthesis and instability in moisture due to auto-oxidation [17]. Therefore, new strategies are needed for development of neurologic therapeutics to counter post-exposure nerve agent cholinesterase poisoning. In recent years, in silico methods have emerged as powerful tools for identification of novel compounds with improved properties [18]. These methods have made remarkable progress in mechanistic drug design and discovery of novel bioactive chemical entities [19]. To pursue these objectives, we adopted in silico pharmacophore modeling and virtual screening strategies to identify potential non-oxime reactivators. Pharmacophores are viewed as ensemble of steric and electronic properties that are necessary for optimal interaction with a specific receptor to trigger or inhibit its biological response [20]. It can be represented by a geometric distribution of chemical features such as hydrogen bond acceptors & donors, aliphatic & aromatic hydrophobic sites, ring aromaticity, and ionizable sites in 3D space of a molecule. Since pharmacophore transcends the structural class and captures only the features responsible for activity, use of pharmacophores for virtual screening of databases has the advantage for 5
identification of new chemical class or chemo-types of compounds. Initiating this approach, we first reported an in silico pharmacophore model for reactivation of oximes from published binding affinity data on tabun-inhibited AChE [21]. Recently, utilizing the model we discovered five novel non-oximes reactivators from two commercial databases, Maybridge and ChemNavigator that showed reactivation efficacy within ten-fold of pralidoxime (2-PAM) against in a DFP-inhibited AChE in vitro assay [23]. However, in vivo efficacies of the compounds were not studied. In the present study, we report both in vitro and in vivo results for discovery of twelve additional non-oximes reactivators of DFP -inhibited AChE reactivators. These non-oximes belong to different chemical classes of compounds. The kr values for these non-oximes were found to be within ten-fold of 2-PAM in an in vitro DFP inhibited AChE assay and significantly, one of these compounds showed comparable efficacy to 2-PAM against brain symptoms of OPagent poisoning in animal studies. None of these compounds were earlier reported to have reactivation of any OP-inhibited AChE agents.
2. Materials and methods 2.1. Procedure for generation of the pharmacophore model. The pharmacophore model used as a template for database searches was developed using the CATALYST methodology [24] as reported earlier [21]. Briefly, the methodology allows the use of known structure and activity data of compounds to create a hypothesis (pharmacophore) which characterizes activity of those compounds. Conformational models of known structures were generated by creating a training set of oximes that emphasized representative coverage within a range of the permissible Boltzman population with significant abundance (within10.0 6
kcal/mol) of the calculated global minimum. We selected this conformational model for pharmacophore generation within CATALYST, which aimed to identify the best threedimensional arrangement of chemical features, such as hydrophobic regions, hydrogen bond donor, hydrogen bond acceptor, and positively and/or negatively ionizable sites distributed over a three dimensional space explaining the activity variations among the training set. The hydrogen bonding features are vectors, whereas all other functions are points. 2.2.Procedure for virtual screening of databases using the pharmacophore model. The above pharmacophore model was used for conducting a targeted virtual assay of approximately 290,000 compounds from the in-house WRAIR-CIS database [25]. First, we generated multi-conformers for all the compounds in the WRAIR-CIS database [25] using the catDB algorithm of CATALYST [24] and stored them. The model was used as a template for conduct the virtual screening in an iterative manner to identify compounds and in vitro evaluations successively. The database search algorithm in CATALYST [24] accounts for molecular flexibility of compounds by considering each compound as an ensemble of conformers. For each compound, conformational energies were assigned with respect to an energy-minimized structure. This screen evaluated the goodness of fit between the compounds in the chemical database and the pharmacophore model. 2.3.Procedure for in silico ADME/Toxicity evaluation of compounds. The down selection of the identified compounds from database searches was performed on the basis of in silico evaluations for favorable ADME (absorption, distribution, metabolism, and excretion)/toxicity properties, particularly the blood brain barrier (BBB) penetrability property, polar surface area (PSA), octanol-water partition (Clog P), and toxicity (rat oral LD50) using the empirical methods of Discovery Studio DS 2.5 [26] and TOPKAT [27], respectively. The 7
procedures incorporated in the software are based on 2D structure information of the compounds. Computed data and models were assessed on the basis of known compound data stored in the databases of the software. 2.4.Procedure for 3D protein-ligand (3DPL) docking. We performed additional down selection of the identified compounds by calculating the binding energy (BE) of each compound at the active site of the known X-ray crystallographic structure of tabun-inhibited AChE using a three dimensional protein-ligand (3DPL) docking protocol of the software, ChemNavigator [28]. This protocol allows analyzing the potential ligand-protein interactions across all potential sites on a protein surface. The receptor structure is usually obtained from publicly accessible RCSB Protein Databank on the internet (http://www.rcsb.org/pdb/) and the ligands to search may be in the SDF or SLN or MOL2 format. The ligands can have the pre-calculated 3D coordinates or can be generated by 3DPL. The procedure by default adjusts the protonation of the receptor and the ligands to simulate the physiological pH. It may also automatically calculate the electrostatic fields and charges. If the electrostatic fields are selected and the molecule does not have charges, these will be calculated using the Gasteiger-Marselli method in 3DPL. The protocol also provides an option for recalculating charges. It uses pre-calculated energy and derivative fields derived from the receptor structure to achieve rapid searching of ligands from databases. The fields generated by 3DPL include steric interactions, hydrogen-bonding interactions, hydrophobic interactions and electrostatic interactions. As the hydrogen-bonding interactions are often used instead of the steric interactions for hydrogen-bonding pairs, and the hydrophobic interactions are usually treated in terms increased steric energies, all these three interactions (steric, hydrogen-bonding, and hydrophobic) are encoded in a single type of field in the protocol. The location of the 8
binding site may be either known or can be indicated, or 3DPL can automatically find the potential binding sites. The structures that are found as hits are extracted in the output file in SDF, SLN, or MOL2 format along with the corresponding binding energies (BE). This docking protocol has been successfully used to find many lead compounds in pharmaceutical discovery programs by inspecting millions of potential ligands, and selecting a small subset for biological screening [23, 28]. 2.5. General methods of preparation for the identified non-oxime compounds. Non-oxime compounds were obtained from the WRAIR-CIS database on contract [25] and their respective data sheets available on request. The non-oximes belong to different classes of compounds (Chart 1) and were not synthesized by us. However, a procedure for synthesis of each of these compounds is included in their data sheets which are also available on request from ref [25]. 2.6. Structure confirmation of the identified compounds Structure and purity of all the 12 compounds were confirmed by the submitter to WRAIR-CIS [25]. However, we re-determined the structures for 10 of the 12 available compounds in 1g quantity by NMR and the results were found to be consistent with the structures shown in Chart 1. The 1H NMR spectra of the compounds were obtained in DMSO-d6 or Chloroform-d using a Bruker Advance 300 or 600 MHz NMR spectrometer. NMR spectra of all compounds showed a good match with that predicted by ACD NMR software spectra. Data for the non-oxime compounds are shown below: N-(3-((2-hydroxyethyl)amino)-1,4-dioxo-1,4-dihydr onaphthalen-2-yl)acetamide, 2
9
1
H NMR (DMSO-d6) 2.01 (s, 3H, -CH3), 3.44 (q, 2H, -CH2CH2OH), 3.53 (q, 2H, -CH2CH2OH ),
4.89 (m, 1H, - CH2CH2OH), 6.87 (br.s., 1H, -NHC2H4OH), 7.98 (d, 1H, C6H4 ring), 7.95 (d, 1H, C6H4 ring), 7.83 (t, 1H, C6H4 ring), 7.74 (t, 1H, C6H4 ring), 9.06 (s, 1H, -NHCOCH3). 4-amino-N-(2,3-dihydroxypropyl)benzenesulfonamide, 3 1
H NMR (DMSO-d6) 2.76 (dd, 1H, -NHCH2CH-), 3.24 (d, 2H, -CH2OH), 3.44 (m, 2H, -NCH2-),
4.55 (br.s., 1H, -CH2OH), 4.75 (br.s., 1H, -CHOH), 5.91 (s, 2H, -C6H4NH2), 6.59 (d, 2H, C6H4), 6.96 (br.s., 1H, -SO2NH-), 7.40 (d, 2H, C6H4). 1-(4-(phenylsulfonyl)phenyl)ethanone, 4 1
H NMR (DMSO-d6) 2.61 (s, 3H, -COCH3), 7.64 (t, 2H, C6H5), 7.72 (t, 1H, C6H5), 7.99 (d, 2H,
C6H5), 8.11 (m(para), 4H, C6H4COCH3). N1-(6-methoxy-4-methylquinolin-8-yl)butane-1,4-diamine, 5 1
H NMR (DMSO-d6) 1.68 (m, 4H, -NHCH2CH2 CH2 CH2NH2), 2.81 (t, 2H, -NHCH2CH2 CH2
CH2NH2), 2.55 (s, 3H, -CH3), 3.24 (br.s., 2H, -NHCH2CH2 CH2 CH2NH2), 3.85 (s, 3H, -OCH3), 6.26 (d, 1H, H-7, quinoline ring), 6.45 (d, 1H, H-5, quinoline ring), 6.53 (br.s., 1H, -NHCH2CH2 CH2 CH2NH2), 7.29 (d, 1H, H-3, quinoline ring), 8.05 (br.s., 2H, -NHCH2CH2CH2CH2NH2), 8.41 (d, 1H, H-2, quinoline ring). 4-chloro-N-(4-nitrophenyl)benzenesulfonamide, 6 1
H NMR (DMSO-d6) 0.86 (t, 3H, -(CH2)5CH3), 1.14 (d, 3H, -CHCH3), 2.21-1.33 (m, 8H, -
CHCH2(CH2)4CH3), 1.38 (m, 1H, -CHCH2-), 1.52 (m, 1H, -CHCH2-), 3.14 (m, 1H, -CH-), 6.82 (d, 2H, C6H5Cl), 7.47 (d, 2H, C6H5NO2), 7.62 (br.s., 1H, -SO2NH-), 7.73 (d, 2H, C6H5Cl), 7.84 (m, 2H, C6H5NO2). N-(2-carbamylethyl)-N-(2-thiolbenzoylethyl) ammonium tosylate, 7
10
1
H NMR (DMSO-d6) 2.28 (s, 3H, -C6H4CH3), 3.13-3.23 (m, 4H, -NHCH2CH2CO-), 3.30-3.41
(m, 4H, SCH2CH2NH-), 7.11 (d, 2H, -C6H4CH3), 7.16 (br.s., 1H, -CH2NHCH2-), 7.47 (m, 2H, C6H5), 7.59 (m, 2H, -C6H4CH3), 7.73 (m, 1H, C6H5), 7.94 (d, 2H, C6H5-), 8.54 (br.s., 2H, -NH2). 2-chloro-4-nitrophenyl benzenesulfonate, 8 1
H NMR (DMSO-d6) 7.58 (d, 1H, C6H3ClNO2), 7.72 (t, 2H, C6H5), 7.90 (t, 1H, C6H5), 7.95 (d,
2H, C6H5), 8.28 (dd, 1H, C6H3ClNO2), 8.44 (d, 1H, C6H3ClNO2). S-(2-((phenylcarbamoyl)oxy)ethyl) methanesulfonothioate, 9 1
H NMR (DMSO-d6) 3.53 (t, 2H, -CO2CH2CH2-), 3.59 (s, 3H, -OCH3), 4.38 (t, 2H, -
CO2CH2CH2-), 7.00 (t, 1H, C6H5), 7.28 (t, 2H, C6H5), 7.46 (d, 2H, C6H5). S-(2-(4-methoxyphenyl)-2-oxoethyl) benzenesulfonothioate, 10 1
H NMR (DMSO-d6) 3.85 (s, 3H, -OCH3), 4.78 (s, 2H, -SCH2O-), 7.04 (d, 2H, C6H4OCH3),
7.65 (t, 2H, C6H5), 7.75 (t, 1H, C6H5), 7.91 (m, 4H, benzene rings). 2-(3-(2-oxooxazolidin-3-yl)propyl)isoindoline-1,3-dione, 12 1
H NMR (DMSO-d6) 1.84 (m, 2H, -NCH2CH2CH2-), 3.18 (t, 2H, -NCH2CH2CH2-), 3.54 (m, 2H,
2-oxazolidinone), 3.58 (t, 2H, -NCH2CH2CH2-) 4.23 (m, 2H, 2-oxazolidinone), 7.81-7.91 (m, 4H, benzene ring). 2.6. Procedure for development of in vitro reactivation assay We followed published literature [29, 30] procedure for developing the DFP-inhibited reactivation assay. An experimental solution was prepared with 1.125µg/mL AChE, 0.05% BSA, and 9.85µM DFP (diisopropyl fluorophosphate) and a control solution was composed of 1.125µg/mL AChE (eel acetylcholinesterase not inhibited with DFP) and 0.05% BSA. Both solutions were incubated at room temperature for 45 minutes. During the incubation, each compound was dissolved in DMSO to a concentration of 50mM while pralidoxime (2-PAM) was dissolved in a 11
sodium phosphate buffer, pH 8 to a concentration of 15mM. Following the incubation, excess DFP was removed from the solutions through centrifugation in separate C18 columns for two minutes. In a 96 well microplate, the non-oxime compound, varying in concentrations from 8.33x10-4 µM to 1.63X10-6 µM, and 2-PAM, varying in concentrations from 2.5x10-4uM to 4.88x10-7uM, were added to the experimental solution and controlled solution with acetylthiocholine to act as a substrate for AChE, DTP to act as a chromogen for AChE, and 0.05M sodium phosphate buffer, pH 8. The SpectraMax Plus 384 then measured the change of optical density per minute at 324nm for 60 minutes. 2.7. Procedure for in vivo study The methodology adopted for determining efficacy of the non-oximes against DFPinduced neuropathology in guinea pigs is described below: 2.7.1. Animals Two-month old adult male guinea pigs were utilized which were previously implanted with radiotelemetry probes (Data Science International, DSI; St. Paul, MN) one week prior to experiments. The radiotelemetry probes are used to record brain activity, as EEG (electroencephalogram) traces, of the animals during our experiments. Animals are given an i.p. injection of pyridostigmine bromide (PB; 0.026mg/kg), which is a reversible antagonist of OPagent binding to cholinesterases in the peripheral nervous system but doesn’t enter the brain. After
a
20
min
delay,
the
animals
were
s.c.
injected
with
8
mg/kg
DFP
(diisopropylphosphofluorate), followed 1 min later by i.m. injections of 2 mg/kg atropine methyl bromide, which doesn’t cross the blood brain barrier, and 5 mg each (~ 16 mg/kg) of the selected WRAIR-CIS non-oximes were dissolved in 200 µl neat DMSO prior to injection. As controls, animals were instead given an injection of 13 mg/kg 2-PAM dissolved in saline (1.5 human auto12
injector equivalents). 2-PAM is the standard oxime used to treatment OP-agent exposure in the United States; however it is a highly charged molecule and not able to cross the blood brain barrier to restore OP-agent inhibited AChE within the central nervous system. During all drug injections and up to 24 h later, EEG response was continuously measured by radiotelemetry. Animals that survived 24 h were euthanized and blood, diaphragm, and whole brain collected. The forebrain was removed for AChE activity assay; and the remainder of the brain was subjected to immersion fixation in 4% formaldehyde prior to histopathology analysis. 2.7.2. AChE activity assay: WRAIR Assay was performed in 96-well microtiter plates; the final concentrations of substrates were 1 mM each of acetylthiocholine, propionylthiocholine, butyrylthiocholine iodides, and 0.2 mM 4,4’ dithiodipyridine, the indicator for the hydrolyzed thiocholine, and UV absorbance measured at 324 nm, which avoids the hemoglobin interference observed with Ellman’s reagent [31]. Guinea pig blood was collected with EDTA and heparin and tissues were frozen at -80°C. To perform the ChE assays, a small aliquot of blood, typically 10 μL, was diluted 20-fold in distilled water. Brain samples were homogenized (on ice) in T-PER buffer using a ground glass homogenizer until completely emulsified. Frozen diaphragm was powdered using an electric tissue pulverizer (Covaris, Inc; Woburn, MA). Homogenates were centrifuged for 10 min at 15,000 x g at 4oC and the supernatants assayed for ChE activity in triplicate (final volume of 300 μL using 50 mM sodium phosphate buffer, pH 8.0). A four-minute kinetic assay was performed at 25°C using a Molecular Devices SpectraMax Plus384 microtiter spectrophotometer (Sunnyvale, CA). Data were subjected to linear least squares analysis from
13
which the activities of AChE (U/mL) were calculated using SoftMax v5.2 and an Excel spreadsheet. Samples were done in duplicate and the averages reported. 2.7.3. Analysis of EEG traces for seizure activity The EEG of the animals, recorded during the 24 h period after exposure to DFP, was quantified by digitalizing the analog signals through the computer program, Dataquest Art 4.1 (Data Science International, DSI; St. Paul, MN). Analysis of the signals was performed using the computer software, Neuroscore v2.0 (DSI), which permits power spectral analysis based on Fast Fourier Transformation. The EEG data obtained from each animal was extracted into varying wave frequencies (Delta, 0.1-4 Hz; Theta, 4.1-8 Hz; Alpha, 8.1-13 Hz; Beta 1, 13.1-21 Hz, and Beta 2, 21.1-32 Hz) and the power of each wave expressed in units of mV2 over time. Total power of the original EEG trace over time was also determined. The percent of total power accounted for by each frequency range was determined by dividing the power of each of the five wave frequencies by the total power of the EEG; this value was then graphed as the percent of total power. 2.7.4. Histopathology Formaldehyde preserved brains were transverse sectioned using a rodent brain matrix (model: RMB-5000C; ASI Instruments, Inc., MI). Transverse sections of 2mm thickness were cut from each brain, using microtome blades hand dropped into the matrix. A section of primary interest was cut out near the distal end of the hippocampus, adjacent to the midbrain. This brain slice contains the entire dorsal to ventral arch of the hippocampus, including the dentate gyrus, CA1, CA2 and CA3 regions. Brain sections were submitted to an outside contract company for 14
processing into microscope slides, containing paraffin embedded 6 µm transverse sections (microtome cut) with hematoxylin and eosin (H & E) and fluoro-jade staining in duplicate (FD Neurotechnologies, Inc, MD). H&E stain is reactive towards membrane lipids and proteins, and highlights the general structural morphology of cells. Fluoro-jade stain, however, penetrates only leaky membranes and thus highlights dead cells a bright-green color. Prepared slides are examined at 40x magnification under an axial light microscope equipped with an image capture camera (Olympus Provis AX80; Olympus DP70).
Standard bright field and fluorescence
illumination (FITC filter) were used on the H&E and fluoro-jade slides, respectively. On the slides the outside pyramidal layer of the ventral hippocampus (CA1-CA2 regions) was chosen for examination, and photographic images were captured for neurons and granular cells comprising this region.
3. Results and discussion Virtual screening of the in-house WRAIR-CIS database [25] using our previously reported pharmacophore model [21] led to discover 12 new non-oxime reactivators (Chart 1) showing kr values between 2.3 to 13.6 mmol-1 min-1 in in vitro DFP-inhibited AChE assay (Table 1). One of these non-oxime reactivators showed in vivo efficacy comparable to 2-PAM against brain symptoms for DFP-induced neuropathology in guinea pigs. None of these compounds reported before as reactivators for OP-inhibited AChE. The pharmacophore model contained a hydrogen bond acceptor, a hydrogen bond donor, and an aromatic ring (Figure 1). Although the model was originally developed to describe the feature requirements for reactivation efficacy against tabun (GA)-inhibited AChE, it proved to be quite satisfactory for identification of non-oxime reactivators of DFP-inhibited AChE [23], 15
probably because of similar simulation characteristics of the two G-agents. Notably, this similarity of pharmacophore led us to discover five recently reported [23] non-oxime reactivators of DFP-inhibited AChE from Maybridge and ChemNavigator databases. Since quantum chemically calculated stereo-electronic properties provided important guidance for developing this pharmacophore as we earlier demonstrated [21, 23], these properties were also assessed for the new non-oxime reactivators. Consistency of several electronic features were observed, such as large localized electrostatic potential regions by the most electronegative atoms (nucleophilic sites) and large lowest unoccupied molecular orbitals (LUMOs) signifying hydrogen bond acceptor regions, weakly distributed potential regions by the aromatic ring regions signifying the hydrophobic regions and localized positive electrostatic regions by the acidic hydrogen atoms signifying hydrogen bon donor regions in the molecules. Stereo-electronic profiles of a few non-oximes were reported by us earlier [23]. The most consistent model (Figure 1) developed from the above calculated stereo-electronic profiles and pharmacophore was used for virtual screening of the database. The identified compounds were then subject to in silico ADME/Toxicity evaluations for down selection prior to in vitro evaluations. The iterative process of in silico searches and in vitro evaluations led us to down select 67 compounds initially from the database [25] containing over 290,000 compounds. Although majority of the 67 compounds showed moderate to weak reactivation efficacy for DFP-inhibited AChE, we short-listed 12 non-oxime compounds (Chart 1) with potent in vitro reactivation efficacy compared to 2-PAM (Table 1) based on in silico evaluations of drug like properties and favorable toxicity profiles. Surprisingly, one of the shortlisted non-oxime showed better in vitro efficacy than 2-PAM and the rest eleven were within 10 folds of 2-PAM. The shortlisted non-oximes are (Chart 1): 1, Thioformylphosphonous acid compound with 16
formaldehyde and 3-(methylamino) pyrrolidine-2,5-dione and toluene (1:1:1:1); 2, N-(3-((2hydroxyethyl)amino)-1,4-dioxo-1,4-dihydronaphthalen-2-yl)acetamide; 3, 4-amino-N-(2,3dihydroxypropyl)benzenesulfonamide; 4, 1-(4-(phenylsulfonyl)phenyl)ethanone; 5, N1-(6methoxy-4-methylquinolin-8-yl)butane-1,4-diamine; 6, 4-chloro-N-(4nitrophenyl)benzenesulfonamide; 7, S-(2-((3-amino-3-oxopropyl)amino)ethyl) benzothioate; 8, 2-chloro-4-nitrophenyl benzenesulfonate; 9, S-(2-((phenylcarbamoyl)oxy)ethyl) methanesulfonothioate; 10, S-(2-(4-methoxyphenyl)-2-oxoethyl) benzenesulfonothioate; 11, 2,2'-(2-(carboxymethyl)(phenethyl)amino)ethyl)azanediyl)diacetic acid; and 12, 2-(3-(2oxooxazolidin-3-yl)propyl)isoindoline-1,3-dione. In vitro activities along with a few selected in silico ADME/toxicity evaluation data are shown in Table 1. Out of the 12 potent non-oximes, four compounds were selected for animal study from which one showed in vivo efficacy comparable to 2-PAM against brain symptoms against DFPinduced neuropathology in guinea pigs The identified non-oxime showing in vivo reactivation efficacy for DFP-inhibited AChE, is significant as no prior report relating to the reactivation activity of this compounds is known in published literature. 3.1. Pharmacophore model and its application in the identification of the non-oximes: The first in silico three dimensional pharmacophore for reactivation efficacy of the oximes [21] was reported by us from binding affinity data on tabun –inhibited AChE [22]. Pharmacophores may be derived in several ways, for example, by analogy to a natural substrate or known ligand, by inference from a series of dissimilar active analogs, or by direct analysis of the structure of a target protein [20, 32-34]. However, a pharmacophore can be used in two ways to identify new compounds that share its features, and thus may exhibit a desired biologic response. In the first approach, de novo design can be performed that link the disjointed parts of 17
the pharmacophore together with fragments in order to generate hypothetical structures that are chemically reasonable but completely novel. The second approach is the three dimensional compound database searching with the pharmacophore model to identify new potent compounds. One key advantage of 3D database searching over de novo design is that it allows the identification of existing compounds that are either readily available or have a known synthetic procedure. In the later case, pharmacophores generated by multiple conformations from a set of structurally diverse molecules enables rapid screening of virtual molecules/libraries to identify new compounds that share its features and may thus exhibit a desired biologic response [32]. Although the pharmacophore model developed for this study contained only three chemical functions localized in space, it proved to be quite predictive. We found that the pharmacophore was instrumental in conducting a targeted in silico screening of about 290,000 compounds from the in-house WRAIR-CIS database [25]. Even though we had developed a high throughput in vitro reactivation assay for determining the AChE reactivation efficacy, time and resource costs prohibited an exhaustive search of the entire chemical database. Thus, a rational strategy for development of a predictive pharmacophore tool was essential to identify a limited number of down selected compounds for in vitro screening. First, we used the model template to conduct the virtual screening by generating multi-conformer format of the set of compounds from the database [25] and iteratively identified and evaluated compounds before finally down selecting the above mentioned twelve good compounds. This in silico screen not only evaluated the goodness of the fit by scoring the fit-score and conformational energy requirement for the fit but also estimated activity for each of the identified compounds to aid down selection. Despite the model [21] used in this study was developed from affinity data on oximes for the tabun (GA) inhibited AChE, it was useful for identifying the non-oxime reactivators for DFP 18
inhibited AChE. However, it is important to note here that although the model was generated from binding affinity data, it depends on a variety of physical features that have modeled, including electrostatic effects, hydrophobic interactions (or more specifically aromatic hydrophobic or aliphatic hydrophobic interactions), hydrogen bond donors, hydrogen bond acceptors, hydrogen bond acceptors (lipid), ionizable sites, and ring aromatic sites. All these features were found to be predictive for the non-oxime reactivators of DFP-inhibited AChE. Binding affinity (KR) also takes into account the shape, size, surface area, volume, and functional groups of the oxime. This is an important parameter for overall reactivating efficacy because reactivation is related to binding of the reactivator to the inhibited enzyme, followed by breakdown of the inhibited complex (rate constant of reactivation) [35]. Mappings of the pharmacophore model onto four of the twelve short-listed non-oxime reactivators of DFPinhibited AChE are shown in Figures 2 to validate the predictive power of the model. 3.2. Assessment of relevance of the pharmacophore model to known X-ray crystallographic structures. To assess the relevance of the pharmacophore features for binding at the active site, we analyzed an earlier reported X-ray crystal structure of tabun-inhibited mouse AChE (mAChE) bound with HI-6 and obidoxime [8]. The crystallographic observations were found to be consistent with the pharmacophore model and show the importance of the generated features, such as the ring aromatic, hydrogen bond acceptor, and hydrogen bond donor features toward binding at the active site of OP-inhibited AChE. It is commonly believed that despite possible intermolecular interactions at the peripheral site of the binding pocket of mAChE, the pyridinium ring of the oxime reactivators interact with the catalytic site in a similar fashion as the OPs [36]. The main feature of the HI-6 bound mAChE crystal structure was the strong cation-pi interaction 19
between the 4-carboxylamide-pyridinium ring and the side chains of Tyr124 and Trp 286 [8]. Thus, the ring aromatic feature of our pharmacophore model (Figure 1) incorporates cation-pi interaction. The carboxylamide moiety of HI-6 is reported to form a hydrogen bond with the main chain Ser298 and water molecule. The hydrogen bond acceptor feature of our pharmacophore model (Figure 1) accounts for this interaction. In addition, the 2-hydroxyiminomethylpyridinium ring of HI-6 moves toward the catalytic site interacting with the side chains of Tyr337, Phe338 and Tyr341 via non-bonded contacts and hydrogen bonds [8]. This crystallographic observation also supports the hydrogen bond donor feature of the pharmacophore model (Figure 1). Taken together, the above observations of HI-6 bound to the mAChE crystal structure were clearly consistent with our pharmacophore model (Figure 1) and support the possible interactions of the newly identified non-oximes at the active site. In addition, we performed a 3D protein-ligand (3DPL) docking calculation [28] to assess binding affinity of the down selected non-oximes at the active site of the crystal structure of tabun-inhibited AChE by calculating the binding energy of each identified compounds (Table 1). For this docking protocol, we used the 2GYU protein structure (PDB code 2GYU, mAChE-HI6) [12] to have a comparative assessment of binding affinity of the non-oximes to the similar Gsimulating OP-inhibited AChE active site. The binding energies were determined after energy minimized the non-oximes at the active site. For comparison, the binding energies (BE) of the twelve short listed non-oximes docked at the active site of OP-inhibited AChE are listed in Table 1. Inspection of the binding energy difference between the two AChE active sites indicates that a few of the identified non-oximes had stronger affinity for the OP-inhibited AChE, such as 1, 5, 7, 10, and 11. Although the compounds seem to have a relationship between in vitro activity and binding affinity for OP-inhibited AChE, no quantitative relationship could be established. 20
However, a trend for competitive inhibitory and reactivation activities for the non-oximes is indicated, though this assessment may not be the perfect match for binding affinity at the active site for DFP-inhibited AChE. Further studies are required to establish the exact nature of interaction at the active site for determining both inhibitory and reactivation efficacies of these non-oximes. 3.3. In silico ADME/Toxicity evaluations of compounds: Down selection of the identified compounds from database searches was not only on the basis of fit score to the model and conformational energy requirement for the fit but also in silico evaluation for favorable blood brain barrier (BBB) properties, octanol-water partition (Clog P), and toxicity (rat oral LD50) using the methods as implemented in Discovery Studio, DS 2.1 [26] and TOPKAT [27], respectively. In silico evaluations of absorption, distribution, metabolism and excretion (ADME) properties of the 12 non-oximes were computed, listing particularly the predictive blood-brain penetration (BBB), lipophilicity (ClogP), and polar surface area (PSA) along with respective molecular weights in Table 1 to assess the overall drug-like character of the compounds. In general, these non-oximes show strong basic characteristics as evidenced from the hydrogen bond acceptor sites in all the non-oximes. This characteristic indicates nucleophilic ability of these compounds that are essential for reactivation efficacy. Since it has been documented [37] that molecules of base- like character with a molecular weight < approximately 400 and ClogP < 4 should have higher/average solubility, permeability, bioavailability and CNS penetration, all the twelve identified non-oximes seem to have these characteristics and thus, possibly have the potential for therapeutics. Furthermore, these non-oximes are devoid of charges unlike the oximes. The presence of quaternary charges in oximes is believed to be the key reason for 21
unfavorable CNS penetration properties [8, 9]. Inspection of Table 1 indicates that ClogP [38] of these non-oximes range from -0.035 to 3.66 indicating varying ability for BBB penetration. In general, as ClogP of a molecule increases, the CNS penetration on average increases along with lower PSA values [12, 37]. However, the effect of molecular weight is reported to be stronger [37] and this suggests that lower molecular weight compounds with a correspondingly lower PSA (or higher ClogP) are desired for improved CNS penetration. We also performed an in silico toxicity evaluation of the non-oximes and compared the results with 2-PAM using models implemented in TOPKAT 6.2 [27]. The results are presented in Table 1. The calculated rat oral LD50 values indicate that except 5 and 8, toxicity of the identified non-oximes is lower or comparable to 2-PAM. Thus, iterative database searches, in silico ADME/toxicity evaluations and in vitro tests allowed us to down select 67 compounds. Each compound’s rate of reactivation was determined using the data’s second order polynomial equation. Majority of these compounds showed weak to moderate reactivation efficacy but 12 of these compounds showed better potent in vitro efficacies within ten fold range of 2-PAM including one compound showing in vivo efficacy comparable to 2-PAM demonstrating the predictive power of the model. 3.4. In vivo study of the down selected compounds: Four of the twelve shortlisted non-oxime compounds, 6, 7, 8, and 9 (Chart 1) were evaluated for testing efficacy against DFP-induced (OP-agent induced) neuropathology in guinea pigs. The AChE activities of blood, brain, and diaphragm; relative strength of seizure activity, and survival duration of the animals after exposure to DFP and drug treatments are shown in Table 2. Animals that were treated with the non-oxime reactivators, compounds 8 and 9 showed 22
severe seizure activity (status epilepticus; SE) that was not alleviated by the treatments; and these animals all died from severe neurological dysfunction within a 7 h period after DFP exposure. Consistent with this, the AChE activity of their tissues was markedly depressed with no major differences between the treatments. Thus, both 8 and 9 at the current given doses were apparently not shown favorable response to the symptoms of OP-agent poisoning. In contrast, animals given the compounds, 6 and 7 survived for 24 h post-exposure to DFP. As shown by the raw EEG recordings in Figure 3, animals treated with compound 7 had continuous SE seizures over the entire 24 h test period; whereas, that given with compound 6 displayed only trace disturbances if any in the brain activity during this time. The animal given the standard oxime 2-PAM as a comparative control, however, showed periodic bursts of moderate seizure activity over 24 h, which is consistent with its inability to cross the blood brain barrier and protect the central nervous system. Shown in Figure 4, the EEG recordings were processed by algorithm transformation to pull out the total power of the trace and the percent power of the underlying waveforms (alpha, beta, delta, and theta). Treatment with the non-oxime 6 showed EEG waveforms consistent with a normal animal; while 7 and 2-PAM both had intensified and abnormal EEG components, especially demonstrating increased delta waves that are hall mark of seizure activity. Tissue AChE activities (Table 2) showed both 6 and 7 to be ~2-fold weaker AChE reactivators compared to 2-PAM in peripheral tissues (blood and diaphragm); whereas, there were no major differences between these compounds in brain restoration of AChE activity. Lack of higher brain AChE activity for 6 may be due to having taken the measurements 24 h after DFP exposure, where residual DFP maintained in fatty tissues had a chance to re-inhibit brain AChE 23
in the absence of the non-oxime reactivator. AChE activities of tissues should be re-examined at different intervals after DFP exposure and drug treatments, through which the peak of non-oxime reactivator entry and action in the brain can be determined. As shown in Figure 5, brain histopathology for the pyramidal neuronal cell layer of the hippocampus supported the favorable EEG findings for 6, where 24 h after DFP exposure there was a complete absence of cellular damage under both H&E and fluoro-jade stains. In contrast, the animal given 7 showed severe degradation of the neuronal cell layer (H&E) and likewise large numbers of apoptotic cells (fluoro-jade). While the neuronal cell layer for the 2-PAM control animal had a relatively normal morphological appearance (H&E), under fluoro-jade stain there was faintly highlighted neurons (pre-apoptosis) and a distinct separation of the cell layer from adjacent tissues. We have further examined the efficacy of these non-oximes by investigating the autofluorescence intensity of the hippocampus cell lines. The non-oximes, 6 and 7 along with 2-PAM (control) were given to the G-pigs after DFP exposure for the auto-fluorescence intensity of the hippocampus cell line study. Out of eight images shot for the two slices mounted on each slide, starting at the bottom and working up to the middle of the pyramidal neuron cell line with 4 shots taken on each side (left and right). This yielded a total of 16 auto-fluorescence intensities for each brain, which were averaged and plotted out as a bar graph with the standard deviation. As it can be seen in the bar graph (Figure 6) that 7 has the highest auto-fluorescence intensity (i.e., an indication of neuronal apoptosis), followed by 6 and then 2-PAM. All differences are statistically significant as judged by simple t-test. Therefore, it can be concluded that for all intensive purposes the compound 6 showed an overall appearance similar to the 224
PAM treated animal. However, the study could not be repeated at higher doses (2 to 3-fold) after DFP exposure to look for accentuation of therapeutic effects and any side effects due to toxicity of the compounds alone due to other limitations. Thus, the in vivo results clearly suggests that one of the WRAIR-CIS compound 6 has the potential for further study as a therapeutic candidate against brain symptoms of OP-agent poisoning.
25
Conclusion Our study clearly demonstrates a rational approach for discovery of new reactivators against OP-inhibited AChE starting from the development of a model to identification of new compounds and testing for their efficacies in both in vitro and in vivo assays. Stereo-electronic property analysis of the oximes led to the generation of a pharmacophore model that allowed us to identify twelve potent non-oxime reactivators including one showing in vivo efficacy comparable to 2-PAM against brain symptoms for DFP-inhibited AChE. Although the model was solely developed from structure-activity relationships, it is consistent with the observations of X-ray crystal structures of AChE with bound reactivators. However, despite perfect mapping onto the model, many of the 67 down selected non-oxime compounds showed poor efficacy. Thus, it is important to note that although the directed approach of 3D database search is an efficient tool for extracting potential bioactive compounds, it is necessary to experimentally test the compounds and iteratively refine the model. Perfect mapping of any molecule to the pharmacophore model does not guarantee its experimental activity despite reflecting receptor complementarity. There may be several factors lacking, such as the perfect fit to the active site due to steric hindrance, electrostatics, overall lipophilicity, and other unforeseen parameters. Thus, in our study, despite reasonably well mapping of the pharmacophore model, not all the compounds were found to have similar reactivation efficacy. Nonetheless, the model provided a useful tool for identification of new compounds that will eventually allow the development of a robust test set of OP-inhibited AChE non-oxime reactivators.
26
Acknowledgments. The opinions or assertions contained herein are the private views of the authors and are not to be construed as official or reflecting true views of the Department of the Army or the Department of Defense. Funding from DTRA (#1E0057_08_WR_C) is acknowledged.
27
References [1] T.C. Marrs, Organophosphate poisoning. Pharmacol Ther. 58 (1993) 51-66. [2] K.S. Bakshi, S.N.J. Pang, R. Snyder, Review of the US Army’s health risk assessment for oral exposure to six chemical-warfare agents – Introduction to the special issue. J. Toxicol. Environ. Health A. 59 (2000) 282-283. [3] J.L. Sussman, M. Harel, F.Frolow, C. Oefner, A.Goldman, L. Toker, I. Silman, Atomic structure of acetylcholinesterase from torpedo californica: a prototypic acetylcholine-binding protein. Science 253 (1991) 872-879. [4] D. Barak, C. Kronman, A. Ordentlich, N. Ariel, A. Bromberg, D. Marcus, A. Lazar, B. Velan, A. Shafferman, Acetylcholinesterase peripheral anionic site degeneracy conferred by amino acid arrays sharing a common core. J. Biol. Chem., 269 (1994) 62966305. [5] J. Bajgar. Organophosphates/nerve agent poisoning: mechanism of action, diagnosis, prophylaxis, and treatment. Adv. Clin. Chem., 38 (2004) 151-216. [6] J. Patocka, K. Kuca, D. Jun. Acetylcholinesterase: crucial enzyme of human body. Acta Medica (Hradec Kralove) 47 (2004) 215-230. [7] F. Ekstrom, C. Akfur, A.K. Tunemalm, S. Lundberg, Structural changes of phenylalanine 338 and histidine 447 revealed by the crystal structures of tabun inhibited murine acetylcholinesterase. Biochemistry 45 (2006) 74-81. [8] F. Ekstrom, Y.P. Pang, M. Boman, E. Artursson, C. Akfur, S. Borjegren, Crystal structures of acetylcholinesterase in complex with HI-6, ortho-7 and obidoxime: Structural basis for differences in the ability to reactivate tabun conjugates. Biochem. Pharmacol. 72 (2006) 597- 607. 28
[9] P. Taylor in: A.G. Gilman, T.W. Rall, A.S. Neis, P. Taylor. (Eds.) Goodman & Gilman’s The Pharmacological Basis of Therapeutics, 8th edition, 1990, pp 131-149. [10] C.D. Bedford, R.A. Howd, O.D. Dailey, A. Miller, H.W. Nolen, R.A. Kenley, J.R. Kern., J.S. Winterle. Nonquaternary cholinesterase reactivators. 3. 3(5)- Substituted 1,2,4oxadiazol-5(3)-aldoximes and 1,2,4-oxadiazole-5(3)-thiocarbohydroximates as reactivators of organophosphonateinhibited eel and human acetylcholinesterase in vitro. J. Med. Chem. 29 (1986) 2174-83. [11] J. Kassa, Review of oximes in the antidotal treatment of poisoning by organophosphorus nerve agents. J. Toxicol Clin Toxicol 40 (2002) 803-816. [12] S. Okuno, K. Sakurada, H. Ohta, H. Ikegaya, Y. Kazui, T. Akutsu, T. Takatori, K. Iwadate. Blood-brain barrier penetration of novel pyridinealdoxime methiodide (PAM)-type oximes examined by brain microdialysis with LC-MS/MS. Toxicol. Appl. Pharmacol. 227 (2008) 8-15. [13] T.-M. Shih, J.W. Skovira, J. C. O’Donnell, J. H. McDonough. Central acetylcholinesterase reactivation by oximes improves survival and terminates seizures following nerve agent intoxication. Advanced Studies in Biology, 1 (2009) 155-196. [14] J.W. Skovira, J. C. O'Donnell, I. Koplovitz, R.K. Kan, J.H. McDonough, T.M. Shih. Reactivation of brain acetylcholinesterase by monoisonitrosoacetone increases the therapeutic efficacy against nerve agents in guinea pigs. Chem. Biol. Interact., 187 (2010) 318-324. [15] E. Shek, T. Higuchi, N. Bodor. Improved delivery through biological membranes. 3. Delivery of N-methylpyridinium-2-carbaldoxime chloride through the blood brain barrier. J. Med. Chem., 19 (1976) 113-117. 29
[16] J.C. DeMar, E.D. Clarkson, R.H. Ratcliffe, A.J. Campbell, S.G. Thangavelu, C.A. Herdman, H. Leader, S.M. Schulz, E. Marek, M.A. Medynets, T.C. Ku, S. A. Evans, F. A. Khan, R.R. Owens, M.P. Nambiar, R.K. Gordon. Pro-2-PAM therapy for central and peripheral cholinesterases Chem. Biol. Interact, 187 (2010) 191-198. [17] N. Bodor, E. Shek, T. Higuchi. Improved delivery through biological membranes. 1. Synthesisand properties of 1-methyl-1,6-dihydropyridine-2-carbaldoxime, a pro-drug of Nmethylpyridinium-2-carbaldoxime chloride. J. Med. Chem., 19 (1976) 102-107. [18] M.T.H. Khan. Molecular interactions of cholinesterases inhibitors using in silicomethods: current status and future prospects. N. Biotechnol. 25 (2009) 331-346. [19] I.M. Kapetanovic. Computer-aided drug discovery and development (CADDD): In silicochemico-biological approach. Chem. Biol. Interact. 171(2008) 165-176. [20] A.R. Leach, V.J. Gillet, R.A. Lewis, R. Taylor. Three-dimensional pharmacophore methods in drug discovery. J. Med. Chem. 53 (2010) 539-558.
[21] A.K. Bhattacharjee, K. Kuca, K. Musilek, R.K. Gordon. In Silico pharmacophore model for tabun-inhibited acetylcholinesterase (AChE) reactivators: a study of their stereoelectronic properties. Chem. Res. Toxicol. 23 (2010) 26-36.
[22] K. Musilek, D. Jun, J. Cabal, J. Kassa, F. Gunn-Moore, K. Kuca. Design of a Potent Reactivator of Tabun-Inhibited Acetylcholinesterase - Synthesis and evaluation of (E)-1-(4-carbamoylpyridinium)-4-(4-hydroxyiminomethylpyridinium)-but-2ene Dibromide (K203). J. Med. Chem. 50 (2007) 5514-5518. [23] A.K. Bhattacharjee, E. Marek, H. T. Le, R.K. Gordon. Discovery of non-oxime reactivators using an in silico pharmacophore model of oxime reactivators of OP-inhibited 30
acetylcholinesterase. European J. Med. Chem. 49 (2012) 229-238. [24] CATALYST version 4.10, Accelrys Inc., San Diego, CA. [25] Chemical Information System; Division of Experimental Therapeutics, Walter Reed Army Institute of Research, Silver Spring, MD. U.S.A. Archives of the Chemical Information System, Division of Experimental Therapeutics, Washington, DC; Data Sheet #: [26] Discovery Studio, DS Version 2.5, Accelrys Inc., San Diego, CA. [27] TOPKAT Version 6.2, Accelrys Inc., San Diego, CA. [28] ChemNavigator-3DPL, ChemNavigator, Inc., San Diego, CA, Email:
[email protected]. [29] F. Worek, N. Aurbek, M. Koller, C. Becker, P. Eyer, H. Thiermann. Kinetic analysis of reactivation and aging of human acetylcholinesterase inhibited by different phosphoramidates. Biochem Pharmacol.73 (2007) 1807-1817.
[30] G.A. Petroianu, D.E. Lorke. Pyridinium oxime reactivators of cholinesterase inhibited by diisopropyl fluorophosphates (DFP): predictive value of in vitro testing for in vivo efficacy. Mini-Reviews in Medicinal Chemistry, 8 (2008) 1328-1342. [31] J.R. Haigh, L. J. Lefkowitz, B.R. Capacio, B.P. Doctor, R.K. Gordon. Advantages of the WRAIR whole blood cholinesterase assay: comparative analysis to the micro-Ellman, Testmate ChETM, and Michel (ΔpH) assays. Chem. Bio. Interact. 175 (2008) 417-420. [32] P. Gund, Evolution of the pharmacophore concept in pharmaceutical research. In Pharmacophore Perception, Development and Use in Drug Design (Güner, O.F., Eds.) pp 5-11, International University Line, La Jolla, CA (2000). [33] A.K. Bhattacharjee, J.A. Geyer, C.L. Woodard, A.K. Kathcart, D.A. Nichols, 31
S.T. Prigge, Z. Li, B.T. Mott, N.C. Waters. A three dimensional in silico pharmacophore model for inhibition of plasmodium falciparum cyclin dependent kinases and discovery of different classes of novel pfmrk specific inhibitors. J. Med. Chem., 47 (2004) 5418-5426. [34] A.K. Bhattacharjee, Virtual screening of compound libraries using in silico three dimensional pharmacophores to aid the discovery and design of antimalarial and antileishmanial Agents. Frontiers in Drug Design and Discov. 3 (2007) 257-292. [35] K. Kuca, J. Cabal, J. Kassa, D. Jun, M. Hrabinova. A comparison of the potency of the oxime HLö-7 and currently used oximes (HI-6, pralidoxime, obidoxime) to reactivate nerve agent-inhibited rat brain acetylcholinesterase by in vitro methods. Acta. Medica. (Hradec Kralove), 48 (2005) 81-86. [36] F.J. Ekstrom, C. Astot, Y.P. Pang. Novel nerve-agent antidote design based on crystallographic and mass spectrometric analyses of tabun-conjugated acetylcholinesterase in complex with antidotes. Clin. Pharmacol. Ther. 82 (2007) 282-293. [37] M.P. Gleeson. Generation of a set of simple, interpretable ADMET rules of thumb. J. Med. Chem. 51 (2008) 817- 834. [38] ChemDraw Ultra 12.0, CambridgeSoft Corporation, USA.
32
Table 1: In vitro reactivation efficacy data, a few selected in silico ADME/Toxicity evaluation results and binding energies at the active site for the 12 short listed compounds from the WRAIR-CIS compared with 2-PAM. BE Kr Compound (ID)
-1
(mmol min-1)
Mol. Wt.
Clog P
PSA
136.3
2PAM 2.61
-3.66
37.5
Rat oral
(kcal/mol)
LD50 (mg/kg)
OPInhibited AChE
477.4
-188.2
(pralidoxime) 1
5.2
362.3
2.64
119.1
480.0
-421.7
2
12.1
274.3
-0.035
98.3
9440.0
-326.3
3
12.5
246.3
-0.68
115.6
4360.0
-406.7
4
12.5
260.3
2.28
51.9
630.0
-409.8
5
10.1
259.0
2.78
59.5
260.0
-398.3
6
10.6
441.9
3.56
93.7
9560.0
-307.2
7
2.3
424.6
0.91
73.9
2440.0
-363.9
8
9.3
313.7
3.66
86.3
230.0
-302.4
9
4.6
275.0
1.59
73.6
850.0
-322.4
10
5.2
322.4
3.05
86.3
1700.0
-405.3
11
13.6
338.0
-0.54
97.4
7900.0
-367.8
12
8.8
274.2
1.60
72.4
10000.0
-401.2
33
Table 2. Tissue AChE activity, seizure intensity, and survival duration of animals after DFP exposure and drug treatments.
Activity (U/mL)
Activity (U/mg)
Activity (U/mg)
Blood
Diaphragm
Brain
9
PB, DFP, atr MeBr, 1.5 AI 9
0.179
1.92
7.54
Severe
6.3
8
PB, DFP, atr MeBr, 1.5 AI 8
0.106
1.58
7.75
Severe
6.5
6
PB, DFP, atr MeBr, 1.5 AI 6
0.223
3.61
15.9
Trace
24.0
7
PB, DFP, atr MeBr, 1.5 AI 7
0.109
3.18
17.6
Severe
24.0
2-PAM
PB, DFP, atr MeBr, 1.5 AI 2-PAM
0.231
6.59
17.0
Moderate
24.0
Compound
Treatment
Seizure
Post-DFP
Strength Survival, h
PB = pyridostigmine bromide @ 0.026 mg/kg; DFP = diisopropyl-fluorophosphate @ 8 mg/kg Atr MeBr = atropine methyl bromide @ 2 mg/kg; AI = human auto-injector equivalents, 13 - 16 mg/kg
34
Chart 1 O
O C6H5
O
NHCH 3CHS
P
O
C
NH
C
CH3
C
NH
C2H4OH
HN OH
OH O
1. Thioformylphosphonous acid compound with formaldehyde and 3-(methylamino)pyrrolidine-2,5-dione and toluene (1:1:1:1)
O
2.N-(3-((2-hydroxyethyl)amino)-1,4-dioxo-1,4-dihydr onaphthalen-2-yl)acetamide O C
H2N
CH3
O S
O
O
O
OH
S
NHCH 2CH
CH2OH
3. 4-amino-N-(2,3-dihydroxypropyl)benzenesulfonamide
4. 1-(4-(phenylsulfonyl)phenyl)ethanone
NHCH2CH2CH2CH2NH2
O
O N
S Cl
NH
O N+ O–
6. 4-chloro-N-(4-nitrophenyl)benzenesulfonamide
H3CO CH3
5. N1-(6-methoxy-4-methylquinolin-8-yl)butane-1,4-diamine
35
Chart 1 contd. (page 2).
O
S
N H
O N+
NH2
O–
7. S-(2-((3-amino-3-oxopropyl)amino)ethyl) benzothioate
H N
Cl
O O S O
O
O S
8. 2-chloro-4-nitrophenyl benzenesulfonate
O2 S
O O
O
S
C S
OCH3
CH2
O
9. S-(2-((phenylcarbamoyl)oxy)ethyl) methanesulfonothioate
10. S-(2-(4-methoxyphenyl)-2-oxoethyl) benzenesulfonothioate
O O
CH2CH2NCH 2CH2N(CH 2COOH)2 O
N
CH2COOH N
11. 2,2'-(2-(carboxymethyl)(phenethyl) amino)ethyl)azanediyl)diacetic acid
O
12. 2-(3-(2-oxooxazolidin-3-yl)propyl)isoindoline-1,3-dione
36
Figure legends
Figure 1. Pharmacophore for binding affinity of the oximes to tabun-inhibited AChE. Figure 2. Mapping of the pharmacophore onto four selected, 1, 2, 5, & 6, potent non-oxime reactivators of DFP-inhibited AChE. Figure 3. EEG traces of animals over initial 16 h after DFP exposure and treatments with the two non-oximes, 6, 7. Figure 4. Power extraction for EEG traces of animals over full 24 h after DFP exposure and treatments with the two non-oximes, 6, 7 and 2-PAM. Figure 5. Histopathology for Pyramidal Neuronal Cell Layer of the Hippocampus after treatments with the two non-oximes, 6, 7 and 2-PAM. Figure 6. Auto-fluorescence intensity of the hippocampus cell line for the two non-oximes, 6, 7, and 2-PAM (control) given to the G-pigs after DFP exposure.
37
Figure(s)
Fig. 1
Ring aromatic
H-bond donor H-bond acceptor
Fig. 2
1 (Kr = 5.2)
5 (Kr = 10.1)
2 (Kr = 12.1)
6 (Kr = 10.6)
Fig. 3
7
6
2-PAM Control
= Time of DFP, atropine, and non-oxime reactivator or 2-PAM administration Each line in the traces represents 60 min of recording time
Fig. 4
7
2-PAM Control
6
Total power
Total power
Total power 60000
600000
60000
400000 40000
200000
40000
30000 20000
20000
20000
10000 0
0 0
0 240
480
720 960 Time (minutes)
1200
1440
0
240
480
720
960
1200
0
Time (Minutes)
Percent total power
240
1440
480
720 960 Time (Minutes)
Percent total power Delta Theta Alpha Beta 1 Beta 2
80
Delta Theta Alpha Beta 1 Beta 2
100
80
Delta Theta Alpha Beta 1 Beta 2
80
60 60
1440
Percent total power
100 100
1200
60
40
40 40
20
20 20
0 0 0
240
480
720
Time (minutes)
960
1200
1440
0
0
240
480
720 960 Time (minutes)
1200
1440
0
240
480
720
Time (minutes)
960
1200
1440
Fluoro-jade stain H&E stain
Fig. 5
7 6 2-PAM Control
Fig. 6
Relative Auto-Fluorescence Intensity
3000
2500
2000
1500
7 6 2-PAM
1000
500
0
