D Modulator with Potential in Major Depressive Disordersjpet.aspetjournals.org/content/jpet/363/3/377.full.pdfAlthough this unique efficacy profile is widely accepted, ketamine also

1521-01033633377ndash393$2500 httpsdoiorg101124jpet117242784THE JOURNAL OF PHARMACOLOGY AND EXPERIMENTAL THERAPEUTICS J Pharmacol Exp Ther 363377ndash393 December 2017Copyright ordf 2017 by The American Society for Pharmacology and Experimental Therapeutics

Preclinical Characterization of (R)-3-((3S4S)-3-fluoro-4-(4-hydroxyphenyl)piperidin-1-yl)-1-(4-methylbenzyl)pyrrolidin-2-one (BMS-986169) a Novel Intravenous GlutamateN-Methyl-D-Aspartate 2B Receptor Negative AllostericModulator with Potential in Major Depressive Disorder s

Linda J Bristow Jyoti Gulia Michael R Weed Bettadapura N Srikumar Yu-Wen LiJohn D Graef Pattipati S Naidu Charulatha Sanmathi Jayant Aher Tanmaya BastiaMahesh Paschapur Narasimharaju Kalidindi Kuchibhotla Vijaya Kumar Thaddeus MolskiRick Pieschl Alda Fernandes Jeffrey M Brown Digavalli V Sivarao Kimberly NewberryMark Bookbinder Joseph Polino Deborah Keavy Amy Newton Eric ShieldsJean Simmermacher James Kempson Jianqing Li Huiping Zhang Arvind MathurRaja Reddy Kallem Meenakshee Sinha Manjunath Ramarao Reeba K VikramadithyanSrinivasan Thangathirupathy Jayakumar Warrier Imadul Islam Joanne J BronsonRichard E Olson John E Macor Charlie F Albright Dalton King Lorin A ThompsonLawrence R Marcin and Michael SinzNeuroscience Discovery Biology (LJB MRW Y-WL JDG TM RP AF JMB DVS KN MB JP DK ANCFA) Neuroscience Discovery Chemistry (JJB REO JEM DK LAT LRM) and Preclinical Candidate Optimization(ES JS MiS) Bristol-Myers Squibb Company Wallingford Connecticut Discovery Synthesis Bristol-Myers SquibbCompany Lawrenceville Princeton New Jersey (JK JL HZ AM) and Biocon Bristol-Myers Squibb Research CenterBangalore India (JG BNS PSN CS JA TB MP NK KVK RRK MeS MR RKV ST JW II)

Received May 12 2017 accepted September 18 2017

ABSTRACT(R)-3-((3S4S)-3-fluoro-4-(4-hydroxyphenyl)piperidin-1-yl)-1-(4-methylbenzyl)pyrrolidin-2-one (BMS-986169) and the phosphateprodrug 4-((3S4S)-3-fluoro-1-((R)-1-(4-methylbenzyl)-2-oxopyrrolidin-3-yl)piperidin-4-yl)phenyl dihydrogen phosphate (BMS-986163)were identified from a drug discovery effort focused on thedevelopment of novel intravenous glutamate N-methyl-D-as-partate 2B receptor (GluN2B) negative allosteric modulators(NAMs) for treatment-resistant depression (TRD) BMS-986169showed high binding affinity for the GluN2B subunit allo-steric modulatory site (Ki 5 403ndash63 nM) and selectivelyinhibited GluN2B receptor function in Xenopus oocytes express-inghumanN-methyl-D-aspartate receptor subtypes (IC505241nM)BMS-986169 weakly inhibited human ether-a-go-gondashrelatedgene channel activity (IC50 5 284 mM) and had negligibleactivity in an assay panel containing 40 additional pharmaco-logical targets Intravenous administration of BMS-986169or BMS-986163 dose-dependently increased GluN2B receptoroccupancy and inhibited in vivo [3H](1)-5-methyl-1011-dihydro-

5H-dibenzo[ad ]cyclohepten-510-imine ([3H]MK-801) binding con-firming target engagement and effective cleavage of the prodrugBMS-986169 reduced immobility in the mouse forced swim testan effect similar to intravenous ketamine treatment Decreasednovelty suppressed feeding latency and increased ex vivohippocampal long-term potentiation was also seen 24 hoursafter acute BMS-986163 or BMS-986169 administrationBMS-986169 did not produce ketamine-like hyperlocomotionor abnormal behaviors in mice or cynomolgus monkeys butdid produce a transient working memory impairment inmonkeys that was closely related to plasma exposure FinallyBMS-986163 produced robust changes in the quantitativeelectroencephalogram power band distribution a transla-tional measure that can be used to assess pharmacodynamicactivity in healthy humans Due to the poor aqueous solubilityof BMS-986169 BMS-986163 was selected as the leadGluN2B NAM candidate for further evaluation as a novelintravenous agent for TRD

IntroductionMajor depressive disorder (MDD) is a prevalent disease and

a leading cause of global disability (Ustuumln et al 2004) In theUnited States alone recent reports show a prevalence of 68with 154 million individuals affected and an overall economicburden of sim$200 billion (Greenberg et al 2015) Although

This work was supported by Bristol-Myers Squibb Company At the timethese studies were conducted all authors were either employees of Bristol-Myers Squibb Company or employees of the Biocon Bristol-Myers SquibbResearch Center Some authors own Bristol-Myers Squibb Company stockAdditionally some authors are inventors on patents related to the subjectmatter

httpsdoiorg101124jpet117242784s This article has supplemental material available at jpetaspetjournalsorg

377

httpjpetaspetjournalsorgcontentsuppl20170927jpet117242784DC1Supplemental material to this article can be found at

at ASPE

T Journals on January 17 2021

jpetaspetjournalsorgD

ownloaded from

therapeutics targeting the monoaminergic systems have beenavailable for many years many patients fail to show anadequate treatment response In this regard results fromthe STARD (Sequenced Treatment Alternatives to RelieveDepression) study showed that only one-third of patientstreated with a selective serotonin reuptake inhibitor achievedremission with 40 of those requiring more than 8 weeks oftreatment for resolution of their symptoms (reviewed byGaynes et al 2009) An additional 50 of patients showedpartial improvement but still exhibited residual symptoms(McClintock et al 2011) which increases the probability offuture relapse and poorer functional and psychosocial out-comes (Paykel et al 1995 Fava 2006) Finally almost onethird of patients were resistant to treatment and failed torespond to additional antidepressant drug switching andaugmentation approaches Clearly a high unmet medicalneed exists for new agents with novel mechanisms that showrapid and improved efficacy in patients with MDD andtreatment-resistant depression (TRD)In the search for novel mechanismsN-methyl-D-aspartate

(NMDA) receptor antagonists have been intensely investi-gated following the first demonstration that the nonselectiveNMDA receptor channel blocker ketamine is efficacious inTRD patients (Berman et al 2000) The clinical response toketamine was remarkable in that a single iv subanestheticdose produced rapid symptom relief within hours of dosingthat was sustained for up to 7 days Subsequent studies haveconfirmed that acute ketamine treatment rapidly achieves a66ndash77 response rate and 31 remission rate in TRDpatients and have also extended these findings to patientswith bipolar depression (reviewed by Lener et al 2017)Although this unique efficacy profile is widely acceptedketamine also produces undesirable effects including tran-sient dissociative and psychotomimetic effects that emergeduring the infusion but resolve within sim2 hours of dosingand may therefore be manageable within the treatmentenvironment However additional concerns including thepotential for neuronal injury and abuse have led to thesearch for alternative approaches that may retain theefficacy profile of ketamine but with fewer liabilities In thisregard agents that selectively inhibit the glutamate NMDA2B (GluN2B) receptor subtype have received attention Thistarget has been the focus of extensive drug developmentefforts aimed at the treatment of conditions such as strokeParkinsonrsquos disease and chronic pain (Layton et al 2006)GluN2B receptors possess an amino terminal domain allo-steric binding site that can be targeted to develop subtype-selective negative allosteric modulators (NAMs) (Monyet al 2009) In a small proof-of-concept study in selective

serotonin reuptake inhibitorndashresistant MDD patients asingle iv administration of the GluN2B NAM CP-101606[(1S2S)-1-(4-hydroxyphenyl)-2-(4-hydroxy-4-phenylpiperidino)-1-propanol] achieved a 60 response rate and 30 remission rate5 days after dosing with an infusion regimen that producedminimal dissociative effects (Preskorn et al 2008)Preclinical results also show that GluN2B NAMs have a

ketamine-like antidepressant profile and that the effects ofketamine may involve the GluN2B receptor subtype Acutetreatment with the GluN2B NAM Ro 25-6981 (4-[(1R2S)-3-(4-benzylpiperidin-1-yl)-1-hydroxy-2-methylpropyl]phenol) acti-vates the mechanistic target of rapamycin (mTOR) signalingpathway and increases synaptic protein expression effectsthought to underlie the rapid antidepressant effect of ketamine(Li et al 2010) Direct infusion of themTOR inhibitor rapamycininto the medial prefrontal cortex prevents the antidepressant-like behavioral effects of Ro 25-6981 and treatment with thea-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA)receptor antagonist NBQX (23dihydroxy-6-nitro-7-sulfamoyl-benzo[f]quinoxaline-23-dione) inhibits antidepressant activityand mTOR pathway activation after acute ketamine or Ro25-6981 administration suggesting common mechanisms(Maeng et al 2008 Li et al 2010) On the other hand geneticdeletion of GluN2B subunits in cortical pyramidal neuronsmimics and occludes the effects of ketamine on mTOR pathwayactivation and synaptic protein synthesis excitatory synaptictransmission and depression-like behavior (Miller et al 2014)Although it should be acknowledged that not all groups are ableto demonstrate themTOR pathway activation effect of ketamine(Popp et al 2016) these results have led to renewed interest inGluN2B NAMs and their potential as rapid-acting antidepres-sant agents for TRD (Miller et al 2016)The present studies describe the pharmacological character-

izationofBMS-986169 [(R)-3-((3S4S)-3-fluoro-4-(4-hydroxyphenyl)-piperidin-1-yl)-1-(4-methylbenzyl)pyrrolidin-2-one] anovelGluN2BNAM identified during a drug discovery program focused on theidentification of an iv therapeutic for TRD We also present re-sults forBMS-986163 [4-((3S4S)-3-fluoro-1-((R)-1-(4-methylbenzyl)-2-oxopyrrolidin-3-yl)piperidin-4-yl)phenyl dihydrogen phosphate]the phosphate prodrug of BMS-986169 which was developed toaddress the poor aqueous solubility of the parent molecule

Materials and MethodsAnimals Studies were conducted at either Bristol-Myers Squibb

(BMS Wallingford CT) or at the Biocon Bristol-Myers Squibb Re-search Center (BBRC Bangalore India)

BMS Studies Animals were housed in an Association for Assess-ment andAccreditation of Laboratory Animal Carendashaccredited facilityand maintained in accordance with the guidelines of the Animal Care

ABBREVIATIONS aCSF artificial cerebrospinal fluid AMPA a-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptor ANOVA analysisof variance BBRC Biocon Bristol-Myers Squibb Research Center BMS Bristol-Myers Squibb BMS-986163 4-((3S4S)-3-fluoro-1-((R)-1-(4-methylbenzyl)-2-oxopyrrolidin-3-yl)piperidin-4-yl)phenyl dihydrogen phosphate BMS-986169 (R)-3-((3S4S)-3-fluoro-4-(4-hydroxyphenyl)-piperidin-1-yl)-1-(4-methylbenzyl)pyrrolidin-2-one CANTAB Cambridge Neuropsychological Test Automated Battery CL confidence limitCP-101606 (1S2S)-1-(4-hydroxyphenyl)-2-(4-hydroxy-4-phenylpiperidino)-1-propanol CSF cerebrospinal fluid fEPSP field excitatory postsynapticpotential FST forced swim test GluN glutamate N-methyl-D-aspartate hERG human ether-a-go-gondashrelated gene potassium channel HFS high-frequency stimulation list-DMS list delayed match to sample LMA locomotor activity LTP long-term potentiation MDD major depressivedisorder MK-0657 (CERC-301) 4-methylbenzyl-(3S4R)-3-fluoro-4-[(2-pyrimidinylamino)methyl]-1-piperidinecarboxylate MK-801 (1)-5-methyl-1011-dihydro-5H-dibenzo[ad]cyclohepten-510-imine mTOR mechanistic target of rapamycin NAM negative allosteric modulator NBQX 23dihydroxy-6-nitro-7-sulfamoyl-benzo[f]quinoxaline-23-dione NMDA N-methyl-D-aspartate NSF novelty suppressed feeding Occ50 dose producing50 occupancy PEI polyethylenimine qEEG quantitative electroencephalogram RM repeated measures Ro-04-5595 1-[2-(4-chlorophenyl)ethyl]-1234-tetrahydro-6-methoxy-2-methl-7-isoquinolinol Ro 25-6981 4-[(1R2S)-3-(4-benzylpiperidin-1-yl)-1-hydroxy-2-methylpropyl]phenol TRD treatmentresistant depression

378 Bristow et al

at ASPE



ownloaded from

andUse Committee of the Bristol-Myers Squibb Company the ldquoGuidefor Care and Use of Laboratory Animalsrdquo and the guidelinespublished in the National Institutes of Health Guide for the Careand Use of Laboratory Animals Research protocols were approved bythe Bristol-Myers Squibb Company Animal Care and Use CommitteeIn vivo [3H](+)-5-methyl-1011-dihydro-5H-dibenzo[ad]cyclohepten-510-imine ([3H]MK-801) binding studies were conducted in singlyhoused male Sprague Dawley rats (170ndash250 g sim6ndash8 weeks of age)supplied with a previously implanted jugular vein catheter (HilltopLaboratory Animals Inc Scottsdale PA) Ex vivo GluN2B occupancystudies were conducted in male Sprague Dawley rats housed tworatscage purchased from Charles River Laboratories (Raleigh NC5ndash7 weeks of age) or Harlan (Indianapolis IN 7ndash9 weeks of age)Ex vivo long-term potentiation (LTP) studies were conducted in maleSprague Dawley rats housed two ratscage purchased from Harlan(4ndash6 weeks of age) Cage size for both single- and pair-housed rats was1025 14 inches and animals were housed in temperature- (216 2degC)and humidity-controlled (50 6 10) rooms maintained under a12-hour lightdark cycle (lights on at 0600 hours) with food and wateravailable ad libitum Cognitive performance behavioral observationalstudies and quantitative electroencephalogram (qEEG) studies wereconducted in separate cohorts of male cynomolgus monkeys (Macacafascicularis 5ndash7 years of age 50ndash85 kg) Monkeys were typically pairhoused [cage size (widthdepthheight) 64 31 33 inches] and fedstandard monkey chow (Teklad global 20 protein Primate Diet 2050Harlan) Subjects for cognitive testing were fed sufficient quantities toensure normal growth while maintaining motivation to performcognitive tasks (Weed et al 1999) subjects for behavioral observationwere food restricted overnight prior to compound administration For allsubjects water was continuously available except during studies andfresh fruit or dietary enrichment was provided twice weekly Toys andforaging devices were routinely provided and television programs wereavailable in the colony rooms Subjects were fitted with plastic or metalneck collars (Primate Products Immokalee FL)

BBRC Studies All experimental procedures were conducted inaccordancewith the guidelines set by theCommittee for thePurpose ofControl and Supervision on Experiments on Animals Ministry ofEnvironment and Forests Government of India An institutionalanimal ethics committee approved the experimental protocols and allanimals were housed in an Association for Assessment and Accredi-tation of Laboratory Animal Carendashaccredited animal facility FemaleXenopus laevis were obtained from Nasco (Fort Atkinson WI) andhoused two frogscage in Tecniplast (Buguggiate Varese Italy)XenopLus standalone housing Water temperature and pH were main-tained at 18 6 1degC and 68ndash74 respectively with regular monitoring ofnitrogenous wastes Frogs were fed with Zeigler adult Xenopus diet(Xenopus Express Brooksville FL) three times per week and each cagewas provided with enrichment tubes and pads Male Sprague Dawleyrats (7ndash10 weeks of age Harlan Livermore CA) were used for in vitroGluN2B binding studies and were housed three ratscage (cage size126 12 inches) Male CD-1mice were obtained directly fromHarlan(Horst Limburg Netherlands) for GluN2B occupancy (5ndash8 weeks ofage) and locomotor activity (LMA) studies (6ndash7 weeks of age) or fromin-house breeding colonies derived from Harlan CD-1 mice for theforced swim test (FST 6ndash7 weeks of age) Novelty suppressed feeding(NSF) studies were conducted in male BALBc mice (13ndash14 weeks ofage Taconic Vivo-bio Hyderabad India) Mice were housed fivecage(cage size 128 66 inches) and all rodents were held in colony roomswith controlled temperature (22ndash24degC) and humidity (48ndash54)under a 12-hour lightdark cycle (lights on at 0700 hours) with foodand water available ad libitum unless specified otherwise

Reagents BMS-986169 BMS-986163 CP-101606 [Traxoprodilfree base (in vitro studies) or methane sulfonic acid salt] Ro 25-6981MK-0657 (CERC-301 4-methylbenzyl-(3S4R)-3-fluoro-4-[(2-pyrimi-dinylamino)methyl]-1-piperidinecarboxylate) [3H]Ro 25-6981 and[3H]MK-801 were synthesized by the BMS or BBRC chemistrygroups or the BMS radiosynthesis group The chemical structures ofBMS-986169 and the phosphate prodrug BMS-986163 are shown in

Fig 1 Sources of other reagents were as follows (1)MK-801 hydrogenmaleate (Sigma-Aldrich St Louis MO) ketamine HCl [Ketaset (FortDodge Animal Health Fort Dodge IA) or Aneket (Neon Pharmaceu-ticals Mumbai India)] 1-[2-(4-chlorophenyl)ethyl]-1234-tetrahydro-6-methoxy-2-methyl-7-isoquinolinol hydrochloride (Ro-04-5595) (TocrisBioscience Bristol UK)

GluN2B Binding Ki Determinations In Vitro Studies todetermine in vitro binding affinity at rat brain GluN2B receptorswere conducted at BBRC For membrane preparation rat forebrainswere thawed on ice for 20 minutes in homogenization buffer composedof 50 mM KH2PO4 (pH adjusted to 74 with KOH) 1 mM EDTA0005 Triton X-100 and protease inhibitor cocktail (11000 Sigma-Aldrich) Thawed brains were homogenized using a Dounce homoge-nizer and centrifuged at 48000g for 20 minutes The pellet wasresuspended in cold buffer and homogenized again using a Douncehomogenizer Finally the protein concentrationwas determined usinga BCA kit (Sigma-Aldrich) and the homogenate was aliquoted flashfrozen and stored at 280degC for 3ndash4 months To perform competitionbinding experiments thawed membrane was passed through a24-gauge needle and 98 ml (20 mg) of the membrane protein wasadded to eachwell of a 96-well plate Test compounds (2ml) were addedto achieve the desired concentration in each well of the assay plateincubated with brain membrane for 15 minutes followed by theaddition of [3H]Ro 25-6981 (50 ml 4 nM final concentration) andfurther incubation for 1 hour Nonspecific binding was determinedusing MK-0657 (40 mM) Membranes were then harvested onto GFBfilter plates using a Filtermate universal harvester (PerkinElmerWaltham MA) Prior to membrane transfer filter plates werepresoaked in 05 polyethylenimine (PEI) for 1 hour at roomtemperature and washed three times using cold assay buffer Aftermembrane transfer filter plates were washed eight times with coldassay buffer and dried at 50degC for 20 minutes followed by addition of50 ml of Microscint20 scintillation cocktail (PerkinElmer) Radioac-tivity was counted using a TopCount (PerkinElmer) after 10 minutesof Microscint addition Counts per minute were converted to percent-age of inhibition and the concentration-response curves plotted andfitted to calculate Ki values using custom-made software Eachexperiment had a plate duplicate and was repeated at least twice toobtain an average binding Ki value

Studies to determine in vitro binding affinity at human (AnalyticalBiologic Services Inc Wilmington DE) and cynomolgus monkeyGluN2B receptors were conducted at BMS Frontal cortical mem-branes were prepared from frozen brain sections by homogenizingsamples at 4degC in hypotonic lysis buffer consisting of 10 mM Tris(pH 74) 5 mM EDTA and protease inhibitors and centrifugation at32000g for 20 minutes The pellet was washed once in buffer

Fig 1 Chemical structures of BMS-986169 and the phosphate prodrugBMS-986163

Preclinical Characterization of the GluN2B NAM BMS-986169 379

at ASPE



ownloaded from

consisting of 50 mM Tris (pH 74) 1 mM EDTA and proteaseinhibitors and centrifuged at 32000g for 20 minutes This pelletwas then resuspended in buffer containing 50 mM KH2PO4

(pH 74 at 25degC) 1 mM EDTA 0005 Triton X-100 and 01 (vv)Sigma Protease Inhibitor Cocktail Aliquots were then frozen on dryiceethanol and kept at280degC On the day of the assay frozen aliquotsof membrane homogenate were thawed homogenized and resus-pended to provide 15 mgwell cynomolgus brain protein or 35 mgwellhuman brain protein in assay buffer (5 mM Tris-HCl pH 74) Insaturation binding experiments the membrane preparation wasincubated at room temperature for 120 minutes in the presence ofincreasing concentrations of [3H]Ro 25-6981 Nonspecific binding wasdefined with 10 mM Ro-04-5595 Competition binding experimentswere performed using a single concentration of [3H]Ro 25-6981(15 nM) in the presence of 10 increasing concentrations (in duplicate)of test compound The reaction was terminated by the addition of 5 mlof ice-cold assay buffer and rapid vacuum filtration through a BrandelCell Harvester (Brandel Gaithersburg MD) using Whatman GFBfilters (PerkinElmer) presoaked in 05 PEI The filtration andwashing was completed in less than 30 seconds The filter was thenpunched onto a 96-wellmicrobeta sample plate 200mlwell of PackardUltima Gold XR scintillation fluid (PerkinElmer) was added and thefilters were soaked overnight The filter pads were then counted in anLKB Trilux liquid scintillation counter (PerkinElmer) Ki values frominhibition curves were determined using the ldquosigmoidal dose-response(variable slope)rdquo non-linear curve fit by GraphPad Prism (version 501GraphPad Software La Jolla CA)

Functional Inhibition of NMDA Receptor SubtypesIn Vitro Studies to demonstrate functional inhibition of GluN2Breceptorndashmediated currents in Xenopus oocytes were conducted atBBRC Prior to surgery frogs were anesthetized using 3-amino-benzoic acid ethyl ester (1 gl pH 70 adjusted using sodium bi-carbonate) the ovarian lobes were removed using aseptic surgicaltechniques and the animals were then singly housed for a 24-hourrecovery period with postoperative monitoring for up to a weekpostsurgery Following removal the ovarian lobes were incubated incold Xenopus oocyte buffer A (Biopredic International Saint GregoireFrance) for 1 hour The oocytes were then mechanically isolated insmall clusters followed by collagenase treatment (15ndash2 mgml) for atleast 1 hour on a shaking platform at 18degC to remove the follicularlayer The defolliculated oocytes were treated with Xenopus oocytebuffer A B and C (Biopredic International) for 15 minutes each inseries at 18degC and then maintained in Barthrsquos solution (pH 74)composed of 88 mM NaCl 1 mM KCl 24 mM NaHCO3 10 mMHEPES 082 mM MgSO4 033 mM Ca(NO3)2 091 mM CaCl2 andsupplemented with gentamycin (100 mgml) penicillin (10 mgml) andstreptomycin (10 mgml) To determine functional inhibition ofGluN2B receptors oocytes were injected with 8ndash15 ng each of humanGluN1a andGluN2B cRNAwithin 24 hours of isolation Two electrodevoltage clamp recordings were made 2ndash7 days postinjection Oocyteswere placed in a plexiglass chamber and impaled with glass micro-electrodes filled with 3 M KCl The oocytes were clamped at 240 mVand perfused with buffer containing 90 mM NaCl 1 mM KCl 10 mMHEPES 001 mM EDTA and 05 mM BaCl2 pH 74 NMDA receptorcurrents were activated by the application of 50 mM glutamate and30 mM glycine and recorded using an Axoclamp 900A amplifier andpClamp 10 data acquisition software (Molecular Devices LLCSunnyvale CA) The concentration response was determined usingfour to seven concentrations of the test compound and each concen-tration was applied for 15ndash20 minutes on a different oocyte Thebaseline leak current at 240 mV was recorded at the beginning andthe end of the recording and the full current trace was linearlycorrected for any change in leak current The level of inhibition wasexpressed as the percentage of the initial glutamateglycine responsein the absence of compound and the IC50 values were obtained byconcentration-response curve fitting using the following equation inGraphPad Prism (version 501) Y 5 bottom 1 (top-bottom)[1 1 10^(LogIC50 2 X) HillSlope)] where X 5 log of concentration

Y5 percentage of inhibition and the top and bottomwere constrainedto 100 and 0 respectively To determine functional inhibition at otherNMDA receptor subtypes isolated oocytes were injected with02ndash05 ng of human GluN1a and 05ndash1 ng of human GluN2A or45ndash60 ng of human GluN1a and 35ndash55 ng of human GluN2C orGluN2D The voltage clamp recording procedure was as describedearlier except that oocytes were clamped at280mV to record GluN2Cor GluN2D receptor currents and compoundwas applied only at 3 mMfor 10-15 minutes The level of inhibition at 3 mM was calculated as apercentage of the initial glutamateglycine-mediated current

Ex Vivo GluN2B Occupancy in Mice and Rats Studies todetermine ex vivo GluN2B occupancy in rats were conducted at BMSFor dose-response studies rats (n 5 4ndash5group) were randomlyassigned to receive an iv injection via the tail vein of vehicle (09saline pH 74 for BMS-986163 or 30 hydroxypropyl-b-cyclodex-trin70 citrate buffer pH 4 for BMS-986169 2 mlkg) BMS-986169(003 03 1 or 3 mgkg) or BMS-986163 (01 03 1 3 or 10 mgkg)15 minutes prior to decapitation and blood and brains were collectedIn a separate study the time course of GluN2B occupancy wasdetermined from 5 to 120 minutes after iv dosing with 3 mgkgBMS-986169 (n 5 4group) In addition to collection of plasma andbrain a terminal cerebrospinal fluid (CSF) sample was also collectedfrom the cisterna magna for measurement of drug concentrationsFinally in parallel with [3H]MK-801 binding studies satellite groupsof rats with jugular vein catheters were randomly assigned to receiveiv BMS-986163 or vehicle treatment (n 5 3group) to determineGluN2B occupancy and drug exposure Following collection the brainwas placed on ice the cerebellumwas removed and the forebrain wasdissected along the midline into left and right hemispheres snapfrozen in isopentane on dry ice and stored at 280degC Blood sampleswere centrifuged at 3500 rpm for 5 minutes at 4degC to separate theplasma and plasma and CSF samples were stored at 280degC On theday of occupancy determinations one forebrain hemisphere from eachanimal was thawed and homogenized in 7 volumes of an assay buffercontaining 50 mMKH2PO4 1 mMEDTA 0005 Triton X-100 and 11000 dilution of Sigma protease inhibitor P3843 (pH 74) using aPolytron homogenizer (Fisher Scientific Hampton NH) In a 96-wellplate 100 mg of tissue (04 mgml tested in triplicate) was incubatedwith 5 nM [3H]Ro 25-6981 in the assay solution at 4degC for 5 minutesThe nonspecific binding was defined by inclusion of 10mMRo 25-6981At the end of the incubation the reaction was stopped by filtrationthrough FPXLR-196 filters (Fisher Scientific) that had been soaked in05ndash10 PEI for 1 hour at 4degC The filters were washed twice withice-cold assay solution and the radioactivity was measured using aWallac Microbeta liquid scintillation counter (PerkinElmer) For eachsample specific binding was calculated by subtracting the value of thenonspecific binding from that of the average total binding andpercentage of occupancy calculated as (1 2 specific binding in drugtreatedspecific binding in vehicle treated) 100 For estimation ofthe dose producing 50 GluN2B occupancy (Occ50) a one-sitebinding model using nonlinear regression was fitted to the meanoccupancy achieved at each dose (GraphPad Prism version 702) Toestimate the Occ50 drug concentration in plasma brain or CSF thesame curve fit was applied to plots of occupancyexposure results forindividual subjects

Additional GluN2B occupancy determinations were also conductedat BBRC in mice completing behavioral assessment in the FST(n 5 4group) or in satellite groups dosed in parallel with NSF andLMA subjects (n5 4group) Mice were rapidly decapitated and brainand blood samples were collected Forebrain membranes were pre-pared as described earlier except for the addition of 3ml of assay bufferfor 160mg of tissue and homogenization by the Polytron for 10 secondsfollowed by 30 strokes using a Dounce homogenizer On the day of theexperiment brain homogenates were thawed on ice and passed threetimes through a 24-gauge needle In a 96-well plate 200ml (64 mgml)of tissue homogenate (tested in triplicate) was incubated with 6 nM[3H]Ro 25-6981 in assay buffer at 4degC for 5 minutes on a shakingplatform The membrane was harvested onto a GFB filter plate

380 Bristow et al

at ASPE



ownloaded from

(treated with 05 PEI for 1 hour at 4degC) then the filter plate wasdried at 50degC for 20 minutes and 50 ml of Microscint20 (PerkinElmer)was added After shaking of the filter plate for 10 minutes theradioactivity was measured using a TopCount (Perkin Elmer) Non-specific binding was determined by incubating the membrane fromvehicle-treated animals with 10 mM Ro 25-6981

In Vivo [3H]MK-801 Binding in Rats Studies to determineinhibition of in vivo [3H]MK-801 bindingwere conducted at BMSusingmethods previously described (Fernandes et al 2015) For the dose-response study rats with previously implanted jugular vein catheterswere randomly assigned (n5 4group) to treatmentwith either vehicle(09 saline 2 mlkg iv) BMS-986163 (01 03 1 3 10 or 30 mgkgiv) or MK-801 (5 mgkg iv) to define the specific binding windowAll rats received [3H]MK-801 (02 mCig body weight) 5 minutes laterand were decapitated 10 minutes after [3H]MK-801 administration(ie 15 minutes after BMS-986163 treatment) For the time-coursestudy rats (n 5 4group) were dosed with BMS-986163 (3 mgkg iv)5ndash120 minutes prior to euthanasia with all subjects receiving [3H]MK-801 10 minutes prior to sacrifice At the appropriate time pointthe brain was collected the cerebellum and brain stem were removedand one forebrain hemisphere was weighed and homogenized in30 volumes of ice-cold 5 mM Tris-acetate buffer (pH 70) A 600-mlsample of the homogenate was filtered through GFBWhatman filters(presoaked in ice-cold Tris-acetate buffer) on a filter box connected to avacuum source Filters were washed four times with ice-cold Tris-acetate buffer and then placed in a scintillation vial with 10 ml ofscintillation fluid (Optiphase supermix PerkinElmer) overnightRadioactivity in the scintillation vial was counted using a WallacMicrobeta liquid scintillation counter The total [3H]MK-801 bindingwas determined by measuring radioactivity in the forebrain ofanimals treated with vehicle whereas nonspecific binding was de-termined by measuring the radioactivity in animals treated withMK-801 For each sample specific binding was calculated by sub-tracting the value of the nonspecific binding from that of the totalbinding and percentage of inhibition was calculated as (1 2 specificbinding in drug treatedspecific binding in vehicle treated) 100For estimation of the dose producing 50 inhibition of MK-801binding data were fitted to a one-site binding model using nonlinearregression as described previously

Mouse FST Studies to evaluate GluN2B NAM treatment effectson immobility in the mouse FST were conducted at BBRC All studieswere conducted between 0900 and 1330 hours and were performedunder dim light and low noise conditions Behavior was recorded andquantified using the automated CleverSys forced swim test apparatus(CleverSys Reston VA) and investigators were unblinded to dosingsolutions Animals (n 5 8ndash11group) were randomly assigned totreatment with either vehicle (30 hydroxypropyl-b-cyclodextrin70 citrate buffer pH 4 5 mlkg iv) CP-101606 (3 mgkg iv) orBMS-986169 (03 056 1 or 3 mgkg iv) and 15 minutes laterwere placed in plexiglass swim tanks (20-cm diameter 40-cm height)filled with water up to a height of 20 cm at a temperature of 256 1degCIn a separate study mice were randomly assigned to treatment(n 5 9ndash10group) with either vehicle (09 saline 10 mlkg iv) orketamine (1 3 or 10 mgkg iv) and tested 30 minutes after dosingSwim tanks were positioned inside a box made of plastic andseparated from each other by opaque plastic sheets to the height ofcylinders Individual animals were placed in one of the three swimtanks and behavior was recorded in a 6-minute testing session Thetotal immobility duration defined as the time spent passively floatingwith only small movements necessary to keep the nosehead abovewater and remain afloat during the 6-minute test was measuredusing the automated CleverSys software Results were analyzed byone-way analysis of variance (ANOVA) followed by Dunnettrsquos post-hoctest (GraphPad Prism version 702) Upon completion of testing(sim25 minutes after drug treatment) a subset of animals (n5 4group)were euthanized by rapid decapitation and blood and brain sampleswere collected for GluN2B occupancy and drug exposure measure-ments as previously described

NSF in Mice Studies to examine the effect of GluN2B NAMtreatment on NSF were conducted at BBRC Prior to treatment allmice were briefly (sim1 minute) placed in a restrainer to acclimatizethem to the iv dosing conditions for 2 consecutive days Animalswere then randomly assigned to treatment (n 5 12ndash15group) withCP-101606 (10 mgkg iv) BMS-986163 (1 3 or 10 mgkg iv) ortheir respective vehicles (25 hydroxypropyl-b-cyclodextrin or phos-phate buffer pH 74 respectively 5 mlkg iv) and then food deprivedfor 21 hourswithwater available ad libitum On the following day (ie24 hours after treatment) individual mice were placed in a novelPerspex arena (40 40 40 cm Coulbourn Instruments AllentownPA) covered on three sides with white paper with the fourth side lefttransparent to enable behavioral evaluation A light source waspositioned on the top to produce a light intensity of sim1000 lux in thecenter of the arena One food pellet (standard chow feed) was placed onthe top of a 60-mm Petri plate placed in the center of the arena Thesession began with the animal placed in one corner of the arena facingaway from the food pellet The latency to begin eating the pellet (ielatency to feed) was recorded by an observer blinded to drugtreatment Feeding was defined as the mouse holding the food pelletwith its forepaws and biting it and testing was terminated eitherwhen the mouse started to feed or after 6 minutes had elapsed Afterevaluation in the novel arena the mouse was transferred to its homecage and food consumption was monitored for 30 minutes (home cagefeeding) Results were analyzed by either two-tailed unpaired t test(CP-101606 vs vehicle) or one-way ANOVA followed by Dunnettrsquospost-hoc test (GraphPad Prism version 702) Satellite groups of mice(n 5 4treatment) were dosed in parallel with the NSF subjects andplasma and brain samples were collected 15 minutes later forexposure and GluN2B occupancy determinations

During the initial pharmacological characterization of the NSFassay studies were conducted to examine the effect of ketaminetreatment following ip administration a dosing route commonlyused to examine the antidepressant-like effect of this agent Subjects(n5 12ndash15group) were treated with vehicle (Milli-Q water 10 mlkg)or ketamine (3 10 or 30 mgkg) and tested in the NSF assay 24 hourslater as described earlier

Ex Vivo Hippocampal LTP in Rats Studies to examine theeffects of GluN2B NAM treatment on hippocampal LTP were con-ducted by BMS Rats were handled and acclimatized to the iv dosingrestrainer for up to 5 minutes each day for 3 days prior to dosing Onthe morning of drug treatment rats were restrained and dosed viathe tail vein by an investigator unblinded to treatment with eitherBMS-986169 (1 or 3 mgkg) or vehicle (30 hydroxypropyl-b-cyclo-dextrin70 citrate buffer 2 mlkg) Typically either two rats (onetreated one control) or four rats (two treated two control) wererandomly assigned to treatment each day for subsequent LTPmeasurements 24 or 72 hours later Each dose of BMS-986169 ortime point was examined as a separate study with independentvehicle-treated controls (n 5 5ndash8group) At the appropriate timepoint rats were rapidly decapitated and brain tissue was removedand quickly submerged into ice-cold cutting solution containing100 mM sucrose 60 mM NaCl 3 mM KCl 7 mM MgCl2bull6H2O125mMNaH2PO4 05mMCaCl2 5mMD-glucose 06mMascorbateand 28 mM NaHCO3 The cerebellum and frontal cortex wereremoved and the brain was cut down the midline Transverse slices(350 mm) were cut from the middle of the hippocampal formation witha vibratome (Leica VT1000S Leica Microsystems Wetzlar Germany)in ice-cold cutting solution bubbled with 95 O25 CO2 Brain sliceswere transferred to a 5050 mixture of cutting solution and artificialcerebrospinal fluid (aCSF) containing 124mMNaCl 3 mMKCl 1 mMMgSO4bull7H2O 125mMNaH2PO4 2 mMCaCl2 10 mM D-glucose and36 mM NaHCO3 and oxygenated with 95 O25 CO2 at roomtemperature (26degC) for 30 minutes Slices were then switched to100 aCSF and continued to be oxygenated with 95 O25 CO2 at30degC for at least 30 minutes of recovery Up to eight slices wererecorded simultaneously using a modified multislice recording con-figuration (Graef et al 2013) In brief two separate tissue slice


at ASPE



ownloaded from

chambers each with a capacity for four isolated tissue slices wereperfused with warm (32ndash34degC) oxygenated aCSF using a gravityperfusion system A flow rate of 2 mlmin for each four-slice recordingchamber was regulated by maintaining a constant fluid level in asecondary solution chamber through the use of a solenoid valve(Harvard Apparatus Holliston MA) a float and an IR beam detector(NV-253SD Med Associates Fairfax VT) Four vehicle slices wereplaced in one of the four-slice recording chambers and four slices froma rat treated with test agent were placed in the other four-slicerecording chamber with the chambers used for each condition beingalternated on subsequent experimental days Once all slices wereoriented in the recording chambers custom-made recording andstimulating electrodes (02-mm diameter platinumiridium wire) werepositioned in the stratumradiatum layer of theCornuAmmonis area 1(CA1) region of the hippocampus Field responses were elicited with asingle monophasic stimulation from two twisted wires (01 ms) every20 seconds (3 Hz) using individual IsoFlex stimulus isolator unitsdriven by a Master 8 stimulus generator (AMPI Jerusalem) Thestrength of each stimulus was adjusted to elicit 30ndash50 of themaximum response Baseline signals were recorded for 20 minutesthen LTP was induced by the following high-frequency stimulation(HFS) protocol two trains of 80 stimulations at 100 Hz at the samestimulus intensitywith a 60-second interval between trains ThisHFSprotocol was repeated two more times at 20 and 40 minutes followingthe first bout of HFS Following this conditioning stimulus a 1-hourtest period was recorded where responses were again elicited by asingle stimulation every 20 seconds (3 Hz) at the same stimulusintensity Signals were amplified with a differential amplifier andwere digitized and sampled at 10 kHzwith aDigiData 1440 (MolecularDevices Sunnyvale CA) recorded using Clampex 100 acquisitionsoftware and analyzed with the Clampfit 100 software package(Molecular Devices) Data from all four slices from each animal(control and treated) on each experimental day were averaged toobtain mean LTP values for each subject Only data from slices wherethe baseline responses were greater than 01 mV and did not vary bymore than 10 and demonstrated at least a 10 potentiation of thebaseline responsewere used for analysis These inclusion criteriawereprespecified prior to the LTP recordings and resulted in averagevalues from two to four slices per subject The total numbers of slicesexcluded from analysis across all studies were as follows LTPamplitude vehicle 5 978 (115) BMS-986169 5 979 (114) LTPslope vehicle 5 878 (103) BMS-986169 5 979 (114) Thenumbers of slices excluded from analysis for individual studies wereas follows study 1 vehicle (24 hours) 5 127 slices 3 mgkg BMS-986169 (24 hours)5 327 slices (slope) or 427 slices (amplitude) study2 vehicle (72 hours) 5 431 slices 3 mgkg BMS-986169 (72 hours) 5432 slices study 3 vehicle (24 hours)5 320 slices (slope) or 420 slices(amplitude) 1 mgkg BMS-986169 (24 hours) 5 220 slices (slope) or120 slices (amplitude) One-tailed paired t tests comparing corre-sponding test periods between control and treated animals were usedto test for statistical significance

Locomotor Activity in Mice Studies to examine the effect ofGluN2B NAM treatment on mouse locomotor activity were con-ducted at BBRC and performed using an automated LMA apparatusAll testing was conducted between 0900 and 1330 hours in a darkroom with low noise conditions and investigators were unblindedto dosing solutions Mice were habituated to individual chambers(40 40 40 cmCoulbourn Instruments) for 2 hours and then randomlyassigned to treatment Inan initial study thedose response to ketaminewas examined in mice (n 5 8ndash9group) treated with vehicle (Milli-Qwater 5mlkg iv) or ketamine (3 10 30mgkg iv) In the subsequentstudy mice (n 5 7ndash8group) were treated with either vehicle (25hydroxy-b-cyclodextrin75 water 5 mlkg iv) ketamine (30 mgkgiv) or BMS-986169 (3 10 or 30 mgkg iv) Animals were returned tothe chamber and activity data were acquired every 250 ms from thefloor panel sensor and expressed as ambulatory distance (centimeters)recorded in 10-minute time bins using TruScan 20 software (CoulbournInstruments) Results were analyzed by two-way repeated-measures

(RM)ANOVA followedbyDunnettrsquos post-hoc analysis (GraphPadPrismversion 702) Separate satellite groups (n5 4treatment) were dosed inparallel and plasma and brain tissue was collected 15 minutes later todetermine drug concentrations and GluN2B occupancy

Abnormal Behaviors in Cynomolgus Monkeys Studies toexamine the effect of GluN2B NAM treatment on spontaneousbehaviors in monkeys were conducted by BMS Prior to treatmentbaseline behavioral profiles were established by observing home cagebehavior for 15 minuteswk for 3 weeks Behaviors were evaluatedusing a 12-item behavioral checklist previously used to demonstratepsychotomimetic-like behaviors following treatment with amphet-amine or ketamine in macaques (Castner and Goldman-Rakic 1999Roberts et al 2010 Supplemental Table 1) When a given behaviorwas observed a mark was made in the appropriate time bin Eachstudy was organized using a Latin-square design and the observerwas blinded to treatment A 5-minute predose observation was made30 minutes before treatment with vehicle (30 hydroxypropyl-b-cyclodextrin70 water 04 mlkg im) (6)ketamine (1 mgkg im) orBMS-986169 (1 3 or 56mgkg im) and behavior was evaluated 0ndash55ndash10 10ndash15 30ndash35 and 60ndash65 minutes post-treatment (n 5 8) Toexamine a higher dose of 10 mgkg BMS-986169 a second blindedstudy was conducted using a Latin-square design and a two-injectionprotocol whereby 2 04-mlkg injections were given to achieve a totaldose of 10 mgkg BMS-986169 1 mgkg ketamine or two injections ofvehicle (n5 8) The dose of ketamine used for comparisonwas selectedfrom a separate study showing a dose-dependent increase in behaviorafter ketamine administration (n5 12) Behaviors were summed overthe entire session and expressed in terms of change from preinjectionbaseline Composite measures of ldquoall abnormal behaviorsrdquo (behaviors2141617110111 Supplemental Table 1) and ldquohallucinatorydissociativerdquo (behaviors 7111 Supplemental Table 1) were calculatedand analyzed by one-way RM ANOVA followed by Dunnettrsquos post-hoctest (GraphPad Prism version 702) On completion of the evaluationsubjectswere placed in a primate restraint chair and a plasma sample(sim75 minutes postdose) was collected from the saphenous vein formeasurement of drug concentrations

Cognitive Impairment in Cynomolgus Monkeys Studies toexamine the effect of GluN2B NAM treatment on cognitive perfor-mance were conducted at BMS using a list delayed match to sample(list-DMS) procedure described previously (Weed et al 2016) Sub-jects (n 5 9) received prior treatment with other pharmacologicalagents and were given at least a 2-month drug-free period prior to thestart of these studies During the procedure animals were seated inrestraint chairs and placed within a sound-attenuated chambercontaining a touch-sensitive computer monitor controlled by theMonkey CANTAB (Cambridge Neuropsychological Testing Auto-mated Battery) software (Lafayette Instruments Lafayette IN)Following correct responses a dispenser delivered 190-mg bananaflavored pellets (Bioserv Frenchtown PA) Animals were first trainedto proficiency on the standard CANTAB DMS task in which anabstract stimulus is presented in the ldquosamplerdquo phase and after amemory delay the sample stimulus plus three distracter stimuli arepresented in the ldquochoicerdquo phase Animals then progressed to a list-DMS procedure in which three different sample stimuli are presentedfirst followed by their choice phases which occur sequentially based ontheir respective memory delays Thus the three trials are nestedwithin the longest memory delay thereby reducing the overall sessionlength The memory delays used in this study were 2 20 and45 seconds Completion of the list was followed by a 10-second screenblank afterwhich a new list was presented for a total of 20 lists in eachtest session During both the sample and choice phases the subjecthad 5 seconds to respond to the stimulus otherwise the trial wasconsidered an omission If the omission was in the sample phasethe choice phase for that sample stimulus was not presented butthe timing of the sample and choice phases for other stimuli remainedunchanged All subjects were habituated to receiving injectionsprior to testing Evaluation of performance after vehicle treatment(25 hydroxypropyl-b-cyclodextrin75 water 04 mlkg im)

382 Bristow et al

at ASPE



ownloaded from

occurred prior to during and at the end of the dose-response function(total of four vehicle determinations) The effect of BMS-986169treatment was tested twice weekly (Monday and Thursday) withbaseline performance sessions on Tuesday and Friday On test daysthe cohort was divided roughly in half and one half received a givendose of BMS-986169 and the other half received a different dose Drugdoses were thus administered in a mixed order across test days andacross subjects by investigators unblinded to treatment In the firststudy the effect of BMS-986169 (01 03 1 or 3 mgkg im)administered 30 minutes prior to testing was examined to determinethe dose-effect relationship In a follow-up study vehicle or 1 mgkgBMS-986169was administered 3 or 5 hours prior to testing to examinethe time course of the effect List-DMS test sessions were sim30minuteslong and on completion blood samples were collected from thesaphenous vein for analysis of plasma drug concentrations Samplecollection times therefore correspond to 1 hour postdose for the dose-response study and 35 or 55 hours for the time-course study Plasmaand CSF drug concentrations were also investigated after dosing withBMS-986169 (03ndash56 mgkg im) in a separate cohort of cynomolgusmonkeys previously implanted with vascular and CSF lumbar accessports (n5 2group) Blood sampleswere collected intoK2EDTA tubes onice plasma was separated by centrifugation at 1330g for 10 minutesand plasma and CSF samples were stored at 280degC until analyzed

The primary measure for these studies was performance accuracyexpressed as the percentage of correct trials ( accuracy 5 [correcttrials][(correct trials1 errors)] 100) Secondary measures included1) latency to respond in the choice phase (milliseconds between onsetof stimulus presentation and touching of monitor) and 2) percentage oftask completed (trials with a response in both sample and choicephasetrials presented) Omissions were not included in the percent-age correct measure but contributed to percentage of task completedPercentage correct accuracy was analyzed by two-way RM ANOVAfollowed by Dunnettrsquos post-hoc test For the time-course studyperformance after vehicle treatment was similar regardless of thepretreatment time (05 3 or 5 hours) and the average was thereforeused in the analysis Latencies to respond in the choice phase(averaged across delays) and percentage of task completed for theentire session were analyzed by one-way RM ANOVA followed byDunnettrsquos post-hoc test In addition for the long delay condition thedifference between performance after vehicle and drug treatment wascalculated for each subject as follows difference 5 correct at longdelay(drug treatment) 2 correct at long delay(vehicle treatment) Allstatistical analyseswere conductedusingGraphPadPrism (version 702)

qEEG Studies in Cynomolgus Monkeys Studies to examinethe effect of GluN2B NAM treatment on the qEEG response wereconducted by BMS using methods described previously (Keavy et al2016) Six cynomolgus monkeys served as subjects for the qEEGstudies and each had been used previously in pharmacologicalstudies with a drug-free period of at least 1 month prior to testingSubjects were surgically implanted under isoflurane anesthesia withradiotelemetry transmitters (Konigsberg Instruments Inc Monro-via CA) attached to two sets of EEG leads placed over the dura one setwith leads over the frontal cortex (with a reference over parietalcortex roughly F6-P6 in the 10ndash20 system) and the second set withleads over the auditory cortex (roughly C6-CP6 in the 10ndash20 system)Buprenorphine (001ndash003 mgkg im) was administered after sur-gery and for a further 2ndash3 days twice daily with extended treatment asneeded if subjects showed signs of discomfort After full recovery(approximately 3ndash4 weeks) the implanted radiotelemetry device wastested to ensure a good EEG signal and suitable subjects wereenrolled in the study On each test day prior to the start of qEEGrecordings subjects were seated in a primate restraint chair and aniv catheter was placed in a saphenous vein for drug administrationAnimalswere allowed to acclimate to a quiet testing chamber outfittedwith an antenna to receive telemeter signals and a camera to monitorthe animal during the session After a 20-minute baseline recordinganimals were treated with either vehicle (09 saline 04 mlkg iv)or BMS-986163 (012 036 or 12 mgkg iv) by investigators

unblinded to treatment and the qEEG was measured for a further90minutes The iv catheter was removed at the end of the session andanimals were tested once weekly with treatments assigned using aLatin square design Analog signals from the telemeters were digitizedat 1000 Hz by PowerLab DAQ systems (ADInstruments ColoradoSprings CO) Raw EEG waveforms were Fourier transformed using aHan (cosine bell) window with 50 overlap between data blocks and anFFT block size of 1024 Overall total power (05ndash55 Hz) as well as powerwithin the different frequency bands was recorded for delta (05ndash4 Hz)theta (4ndash9 Hz) alpha (9ndash13 Hz) beta 1 (13ndash19 Hz) beta 2 (20ndash30 Hz)and gamma (30ndash55 Hz) bandwidths The relative power (power in abandoverall total power) was determined as the average value from1-minute bins across the 90-minute session and analyzed by two-wayRM ANOVA followed by Holm-Sidak post-hoc test (GraphPad Prismversion 702) Area under the curve over the entire 90-minute period foreach powerbandwas calculated fromthe time-course data andanalyzedby RM ANOVA followed by Dunnettrsquos post-hoc test (GraphPad Prismversion 702) Finally using the area under the curve resultsCohenrsquos d estimate of effect sizewas calculated using unpaired statistics(httpwwwuccsedusimlbecker)

Plasma and CSF drug concentrations were also investigated afterdosing with BMS-986163 (12 mgkg iv) in cynomolgus monkeyspreviously implanted with vascular and CSF lumbar access ports(n5 2time point) Sample collection processing and storage were aspreviously described

Quantification of BMS-986169 and BMS-986163 in Samplesfrom In Vivo Studies For exposure analysis of BMS-986169 andBMS-986163 in plasma brain and CSF rats andmonkeyswere dosedand samples were collected as described BMS-986163 was stabilizedex vivo by collecting blood samples into tubes containing 4 K2EDTAdissolved in 1M phosphate buffer pH 74 (5 of the blood volume) andmaintained at 4degC Plasma was gathered within 15 minutes of bloodcollection and stored at 280degC until analysis For brain samplestissue was homogenized and diluted with blank plasma CSF sampleswere prepared in the same manner as plasma An aliquot of 50 ml ofcalibration standard or study sample (plasmatissueCSF) was addedto 200 ml of acetonitrile containing internal standard (Ro 25-6981 at100 nM) and transferred to a Strata protein precipitation filter plate(Phenomenex Torrance CA) Plates were vortexed on a plate shakerfor 5 minutes at 300 rpm centrifuged at 4000 rpm for 5 minutes andthe eluent was collected into a 96-well collection plate (WatersCorporation Milford MA) A volume of 4 ml was injected for liquidchromatographyndashmass spectrometrymass spectrometry analysisLiquid chromatographyndashmass spectrometrymass spectrometry anal-ysis was performed on a Waters Acquity UPLC (Waters Corporation)interfaced to an API4000 triple quadrupole mass spectrometer (SciexToronto Canada) equipped with a Turboionspray source The sourcetemperature was set at 550degC and the ionspray voltage was set to45 kV Samples were maintained at 10degC for the duration of theanalysis Separations were performed on a Waters Acquity BEH C18(21 50 mm 17 mm) columnmaintained at 60degC The mobile phasewhich consisted of 01 (vv) formic acid in water (A) and 01 (vv)formic acid in acetonitrile (B) was delivered at a flow rate of700 mlmin Separation was achieved using a gradient elution startingat 95 aqueous mobile phase and going to 95 organic mobile phaseNitrogen was used as nebulizer and auxiliary gas with a pressure andflow rate of 80 psi and 7 lmin respectively Data acquisition usedselected reaction monitoring operating mode Standard curves usedfor quantification were fitted with linear regression weighted byreciprocal concentration 1x2 Data and chromatographic peaks werequantified using Analyst version 142 software (Sciex)

ResultsBMS-986169 Potently Inhibits [3H]Ro 25-6981 Bind-

ing In Vitro Saturation binding experiments showed that[3H]Ro 25-6981 labeled a single population of receptors in


at ASPE



ownloaded from

membranes prepared from rat forebrain cynomolgus monkeyfrontal cortex or human frontal cortex with Kd values of 223(n 5 3) 093 (n 5 1) and 199 nM (n 5 2) respectively BMS-986169 CP-101606 and MK-0657 potently and completelydisplaced [3H]Ro 25-6981 specific binding in all species(Supplemental Figs 1 and 2 Table 1)BMS-986169 Selectively Inhibits GluN2B Receptor

Function To demonstrate that BMS-986169 inhibits GluN2Breceptor function electrophysiological recordings were made inXenopus oocytes expressing human GluN1aGluN2B receptorsApplication ofBMS-986169produceda concentration-dependentinhibition of glutamateglycine activated currents the IC50 valuewas 241 nM [95 confidence limits (CL) 201ndash289] determinedfrom a 7-point concentration-response curve (n 5 3ndash5concen-tration) (Fig 2) In comparison the IC50 values determinedfollowing 15-minute incubation with CP-101606 or Ro 25-6981were 106 (95 CL 77ndash146) and 93 nM (95 CL 84ndash104)respectively (Fig 2) BMS-986169 showed minimal inhibition ofglutamateglycine-activated currents at other NMDA receptorsubtypes themean6SD inhibition at 3mMwas22462686 6 64 and 7 6 38 at human GluN1aGluN2AGluN1aGluN2C or GluN1aGluN2D receptors respectively (n5 2ndash3) BMS-986169 was also examined in a broad panel ofin vitro assays at 40 additional pharmacological targets with noadditional activity of relevance identified (Supplemental Ta-ble 2) Finally BMS-986169 showed weak functional inhibitionathumanether-a-go-gondashrelated genepotassiumchannel (hERG)channels the IC50 determined by patch-clamp electrophysiologyin HEK293 cells expressing hERG was 284 mMIntravenous Administration of BMS-986169 or BMS-

986163 Achieves High Levels of GluN2B ReceptorOccupancy in Rat Brain Intravenous administration ofBMS-986169 produced a dose-dependent increase in ex vivoGluN2B occupancy determined in rat brain homogenates15 minutes after dosing the dose achieving 50 occupancywas 022 mgkg (95 CL 013ndash037) with 97 occupancyachieved at the highest dose tested (3 mgkg Fig 3A)Treatment with the phosphate prodrug BMS-986163 alsoproduced a dose-dependent increase in GluN2B occupancydetermined 15 minutes after iv dosing in rats (Fig 3A)Again maximal occupancy was achieved at 3 mgkgBMS-986163 (101) and the dose achieving 50 occupancywas 042 mgkg (95 CL 028ndash061) As expected GluN2Boccupancy was closely correlated with the total plasmaconcentration of BMS-986169 following dosing of eitherBMS-986169 or BMS-986163 (Fig 3B) These results confirmthat the prodrug is rapidly converted to the parent drug afteriv administration and are consistent with other pharmaco-kinetic studies showing little to no detectable levels of prodrugin the systemic circulation Brain tissue concentrations ofBMS-986169 also showed a strong relationship with GluN2Boccupancy (Supplemental Fig 3A Supplemental Table 3) the

average brain tissuetotal plasma concentration ratio across allGluN2B occupancy studies was 3 Investigation of the occupancytime course after dosing with 3 mgkg iv BMS-986169 showedthat peak occupancy (sim93)wasmaintained for up to 30minutesand then declined rapidly to a level of 57 2 hours aftertreatment (Fig 3C) Analysis of BMS-986169 concentrationsin plasma (total and free drug concentration) brain tissue andCSF collected from these subjects showed a strong relationshipbetween drug exposure and occupancy in all compartments(Fig 3D Supplemental Table 3) BMS-986169 was detected inCSF at levels similar to the free plasma drug concentration atall time points examined (Supplemental Fig 3B) The CSFconcentration to achieve 50 GluN2B occupancy was 35 nM

TABLE 1Displacement of [3H]Ro 25-6981 binding to native GluN2B receptors in membranes prepared from ratforebrain cynomolgus monkey frontal cortex or human frontal cortex

Compound Rat (Mean Ki 6 SD) Monkey (Mean Ki 6 SD) Human (Mean Ki 6 SD)

nM nM nMBMS-986169 403 6 139 (n = 13) 42 6 08 (n = 3) 63 6 15 (n = 3)CP-101606 696 6 056 (n = 3) 22 6 05 (n = 4) 31 6 09 (n = 4)MK-0657 252 6 042 (n = 3) 32 6 05 (n = 3) 44 6 14 (n = 3)

Fig 2 BMS-986169 inhibits glutamateglycine currents in xenopusoocytes expressing human GluN1aGluN2B receptors (A) Representativerecording showing inhibition of 50 mMglutamate + 30 mMglycine currentsfollowing application of 100 nM BMS-986169 (B) Concentration-responsecurve for inhibition of glutamateglycine-activated currents Each pointshows the mean 6 SEM percentage of inhibition following 15-minuteincubations with BMS-986169 (n = 3ndash5concentration) CP-101606 (n =1ndash4concentration) or Ro 25-6981 (n = 1ndash2concentration)

384 Bristow et al

at ASPE



ownloaded from

and was consistent with the in vitro GluN2B binding affinityIn addition GluN2B occupancy predicted from CSF concentra-tions and the rat in vitro GluN2B binding Ki was similar to themeasured occupancy in these subjects (Supplemental Fig 3C) Itshould be noted that BMS-986169 shows poor oral bioavail-ability in rats the oral bioavailability determined after oraldosing with 5 mgkg BMS-986169 or the prodrug BMS-986163was 25 and 17 respectivelyBMS-986163 Functionally Inhibits NMDA Receptors

after iv Dosing in Rats Intravenous administration of thephosphate prodrug BMS-986163 dose-dependently inhibitedin vivo [3H]MK-801 binding with a maximal 41 inhibitionachieved at 10 mgkg (Fig 4A) Satellite groups of ratsreceiving jugular vein injections of the same BMS-986163dosing solutions showed 97 GluN2B occupancy at 10 mgkgand an Occ50 dose of 055 mgkg (95 CL 02ndash147) (Fig 4A)Further examination of the [3H]MK-801 binding responseconfirmed that inhibition saturates at sim40 after treatmentwith BMS-986163 (mean 6 SEM percentage of inhibition15 minutes after dosing 10 mgkg 5 384 6 12 30 mgkg 5436 6 27) Inhibition of [3H]MK-801 binding was also time-dependent the maximal inhibition achieved after dosing with3 mgkg BMS-986163 was 37 and occurred at 15 minutespost-treatment (Fig 4B) Thereafter inhibition of [3H]MK-801binding declined rapidly over time and was negligible(6 inhibition) by 2 hours post-treatment (Fig 4B) Again

GluN2B occupancy determined in satellite groups was high($89) for the first 30 minutes after dosing and then declinedrapidly to a level of 28 2 hours after treatment (Fig 4B) Therelationship between total plasma and brain BMS-986169concentrations and GluN2B occupancy from these combinedstudies was similar to previous results (Supplemental Fig 4Supplemental Table 3)BMS-986169 Reduces Immobility in the Mouse FST

Treatment with BMS-986169 produced a significant dose-dependent decrease in immobility [ANOVA F(553) 5 3547P 5 00077] (Fig 5A) in mice tested in the FST assayThe minimum effective dose was 1 mgkg which achievedan average total plasma BMS-986169 concentration andGluN2B occupancy of 268 nM and 733 respectively de-termined immediately after completion of the test (Supple-mental Table 4) Decreased immobility was also seen in micetreated with CP-101606 (3 mgkg iv) which achieved amean (6SEM) GluN2B occupancy of 85 6 045 Asexpected BMS-986169 produced a dose- and exposure-dependent increase in ex vivo GluN2B receptor occupancythe dose achieving 50 occupancy was 04 mgkg (95 CL016ndash082) and the total plasma and brain drug concentra-tions achieving 50 occupancy were 102 and 404 nM re-spectively (Supplemental Fig 5 Supplemental Table 3) Theaverage BMS-986169 brain tissuetotal plasma concentrationratio in mice was 4 A dose-dependent decrease in immobility

Fig 3 Ex vivo GluN2B occupancy in rat brain following tail vein administration of BMS-986169 or the phosphate prodrug BMS-986163 Occupancydose response (A) and total plasma concentrationoccupancy relationship (B) determined 15 minutes after BMS-986169 or BMS-986163 treatmentOccupancy time course (C) and relationship between occupancy and total plasma free plasma brain tissue and CSF BMS-986169 concentration (D) inrats treated with 3 mgkg BMS-986169 Results are presented as the mean 6 SEM percentage of GluN2B occupancy (n = 4ndash5dose) orexposureoccupancy results for individual subjects


at ASPE



ownloaded from

was also seen after iv ketamine administration [F(334) 55212 P 5 00045] which was significant at doses of 3 and10 mgkg and similar in magnitude to the effects seen withBMS-986169 (Fig 5B)BMS-986163 Reduces NSF 24 Hours after iv Dosing

in Mice Treatment with BMS-986163 produced a dose-dependent decrease in the latency to feed in a novel environ-ment determined 24 hours after dosing in mice [ANOVA F(352) 5 4635 P 5 0006] (Fig 5C) A similar reduction inlatency was also observed in CP-101606ndashtreated animals(unpaired t test P 5 00255) (Fig 5C) Home cage foodconsumption measured on completion of testing in the NSFassay was not significantly different from vehicle-treatedsubjects inmice treated with BMS-986163 [ANOVA F(352)51976 P5 01291] or CP-101606 (unpaired t test P5 09731)(Fig 5D) Measurement of GluN2B occupancy 15 minutesafter dosing in satellite groups of mice showed that 101

occupancy was achieved at both 3 and 10 mgkg BMS-986163(Supplemental Table 4) The mean 6 SEM GluN2B occu-pancy determined 15minutes after CP-101606 (10mgkg iv)administration was 92 6 1The effect of intravenous ketamine administration was

not tested however ip dosing was examined during ourinitial characterization of the NSF assay Ketamine signifi-cantly reduced the latency to feed in the novel environment[ANOVA F(351) 5 3788 P 5 00158] 24 hours afteradministration of 10 and 30 mgkg ip with no effect observedon home cage food consumption [ANOVA F(351) 5 1749P 5 01687] (Supplemental Fig 6)BMS-986169 Enhances Ex Vivo Hippocampal LTP

24 Hours after Dosing in Rats Recordings from hippo-campal slices taken from vehicle-treated rats showed thatapplication of HFS increased both the amplitude and the slopeof the field excitatory postsynaptic potential (fEPSP) responsecompared with baseline recordings consistent with the in-duction of LTP A single administration of BMS-986169(3 mgkg iv) produced a robust enhancement of ex vivohippocampal LTP measured 24 hours after dosing (Fig 6)This elevation was apparent for both the amplitude (Fig 6A and B) and the slope (Fig 6 C and D) of the fEPSP responsewith a significant enhancement over baseline observed fol-lowing application of all three HFS stimuli The elevation ofLTP was dose-related with a small but significant elevationof the fEPSP slope observed 24 hours after treatmentwith BMS-986169 (1 mgkg iv) (Supplemental Table 5)Finally the enhancement of LTP was time-dependent andwas no longer observed 72 hours after dosing with 3 mgkgBMS-986169 (Supplemental Table 5)BMS-986169 Has Marginal Effects on LMA in Mice

Treatment with ketamine produced a significant increase inLMA that lasted 50 minutes after dosing with 10 or 30 mgkg[treatment effect F(331) 5 1981 P 00001 time effectF(11341)5 9919 P 00001 interaction F(33341)5 8232P 00001] (Fig 7A) In comparison treatmentwithBMS-986169hadmarginal effects on LMA [treatment effect F(431)5 997P 00001 time effect F(11341) 5 6262 P 00001interaction F(44341) 5 676 P 00001] (Fig 7B) Treat-ment with 3 mgkg BMS-986169 had no effect on LMA at anytime point At higher doses a small but significant increasewas observed in the first 10 minutes after dosing which wassubstantially lower than the ketamine response (Fig 7B) Asmall increase was also observed at the 50ndash60-minute timepoint in mice treated with 30 mgkg All doses of BMS-986169were associated with high levels of GluN2B occupancy (range5 964ndash997) determined 15 minutes after iv dosing insatellite groups of mice (Supplemental Table 4)BMS-986169 Does Not Produce Ketamine-Like Dis-

sociativeHallucinogenic-Like Effects in CynomolgusMonkeys In a previous study im ketamine administrationproduced a dose-dependent increase in the incidence ofabnormal and dissociativehallucinogenic-like behaviors adose of 1 mgkg was chosen as the comparator for furtherinvestigation (Supplemental Fig 7) Two independent stud-ies were conducted to examine the effects of intramuscularadministration of BMS-986169 on spontaneous behaviorsin cynomolgus monkeys the first examined BMS-986169at doses of 1 3 or 56 mgkg and the second examinedBMS-986169 at a dose of 10 mgkg Results from repeated-measures ANOVA showed a significant effect of treatment on

Fig 4 Inhibition of in vivo [3H]MK-801 binding and relationship toex vivo GluN2B occupancy in rats treated with the phosphate prodrugBMS-986163 (A) Dose response determined 15 minutes after iv dosing(B) Time course determined in rats treated with 3 mgkg (iv) Results arepresented as the mean 6 SEM percentage of inhibition of [3H]MK-801binding (n = 4group) or percentage of GluN2B occupancy (n = 3group)

386 Bristow et al

at ASPE



ownloaded from

the incidence of abnormal behaviors [study 1 F(428)5 1206P 00001 study 2 F(214) 5 1443 P 5 00004] anddissociativehallucinogenic behaviors [study 1 F(428) 56533 P 5 00008 study 2 F(214) 5 1108 P 5 00013]Dunnettrsquos post-hoc analysis showed that the incidence of thesebehaviors was significantly increased following ketaminetreatment (1 mgkg im) whereas BMS-986169 had no effectat any dose tested (Fig 8) Analysis of total plasma BMS-986169 concentrations on completion of testing (sim75 minutespost-treatment) showed dose-dependent exposure rangingfrom 620 to 10260 nM (Supplemental Table 6)BMS-986169 Transiently Impairs Working Memory

in Cynomolgus Monkeys Vehicle-treated monkeys per-forming the list-DMS task showed a significant decrease inperformance accuracy as the delay between the initial pre-sentation of the stimulus and the choice phase increased(mean6 SEM percentage of accuracy short delay5 956613mediumdelay5 746 41 long delay5 6376 46one-way RM ANOVA [F(216) 5 333 P 00001] followed byDunnettrsquos post-hoc test compared with short delay) Intra-muscular injection of BMS-986169 produced a significantdose- and delay-dependent decrease in performance accuracyin subjects tested 30 minutes after dosing [treatment F(432)5 3921 P 00001 delay F(216) 5 1162 P 00001interaction F(864) 5 8533 P 00001] Further post-hoc

analysis showed that BMS-986169 had no effect on perfor-mance accuracy at short delays at any dose tested (Fig 9A) Incontrast a significant impairment in accuracywas observed atboth the medium and long delays in subjects treated with1 and 3mgkg BMS-986169 (Fig 9A) The lack of effect at shortdelays indicates that performance of the task per se is notaltered by drug treatment Consistent with these resultstreatment with BMS-986169 did not alter percentage of taskcompletion indicating no effect on motivation to perform thetask at any dose (Supplemental Fig 8A) Although a signifi-cant increase in latency to respond was seen at the highestdose this increase (sim300 ms) is only a minor slowing ofresponse time (Supplemental Fig 8B) The effect of treatmentwith 1 mgkg BMS-986169 on performance accuracy was alsotime-dependent [treatment F(324)5 187 P 00001 delayF(216) 5 1031 P 00001 interaction F(648) 5 8112P 00001] with maximal impairment observed 30 minutespostdose and full recovery achieved by 5 hours after treatment(Fig 9C) Again there was no effect on percentage of taskcompletion or response latency consistent with a selectivememory impairment (Supplemental Fig 8 C and D) Acrossboth studies the degree of impairment as represented bythe difference in the percentage of accuracy from vehicletreatment at the long delay condition was related to totalplasma BMS-986169 concentrations measured on completion

Fig 5 Effect of iv administration of BMS-986169 (A) or ketamine (B) on immobility in the mouse FST assay Results are presented as the mean 6SEM immobility duration determined 15 minutes after treatment with BMS-986169 or CP-101606 (CP 3 mgkg iv n = 8ndash11group) (A) or30 minutes after treatment with ketamine (n = 9ndash10group) (B) Results were analyzed by ANOVA followed by Dunnettrsquos post-hoc test P 001P 005 compared with vehicle (Veh) Effect of iv administration of BMS-986163 or CP-101606 (10 mgkg) on novelty-suppressed feeding (C) andhome cage feeding (D) determined 24 hours after dosing in mice (n = 12ndash15group) Results are presented as the mean6 SEM latency to feed in thenovel environment or quantity of food consumed in the home cage during a 30-minute period beginning immediately after testing in the novelenvironment Treatment effects were compared with their respective vehicle by either unpaired t test (CP-101606 vs Veh-1) or ANOVA followed byDunnettrsquos post-hoc test (BMS-986163 vs Veh-2) P 005 P 001


at ASPE



ownloaded from

of testing in experimental subjects (Fig 9 B and D Supple-mental Table 6) with large impairments seen at plasmaconcentrations $483 nM Investigation of plasma exposures ina separate group of subjects confirmed that total plasma BMS-986169 concentrations are stable across the duration of the list-DMS task (ie 30ndash60minutes postdose Supplemental Fig 9A)CSF concentrations of BMS-986169 increased in a dose-dependent manner ranging from 185 to 79 nM at 30 minutespostdose (Supplemental Fig 9B) The predicted GluN2B occu-pancy calculated using CSF concentrations and the monkeyin vitroGluN2B bindingKi value (42 nM) ranged from815 to95 across the dose range examined (Supplemental Fig 9B)BMS-986163 Alters qEEG Power Band Distribution

in Cynomolgus Monkeys Intravenous administration ofBMS-986163 produced robust and dose-related changes in theqEEG power distribution in the beta 1 alpha and delta powerbands Analysis of time-course data showed that all doses ofBMS-986163 reduced the relative power in the beta 1 powerband (Supplemental Table 7) At the highest dose (12 mgkg)a significant increase in relative power in the delta band anddecrease in the alpha band were also observed (SupplementalTable 7) Time-course results show that treatment effectsemerged rapidly and were sustained throughout the90-minute recording period (Fig 10A Supplemental Figs10ndash12) Analysis of results expressed as the area under therelative power curve confirmed these results except for thelow-dose (012 mgkg) effect on beta 1 (Fig 10B SupplementalFig 13 Supplemental Table 8) Cohenrsquos d analysis showedthat the relative power changes produced at the highest doseof BMS-986163 were moderate (05) to large (08) witheffect sizes of 077 (delta) 2088 (alpha) and 2101 (beta 1)respectively (Fig 10C) Investigation of plasma exposure in

separate subjects dosed with BMS-986163 (12 mgkg iv)showed that maximal total plasma BMS-986169 concentra-tions were achieved within 5 minutes of dosing confirmingrapid conversion of the prodrug to parent (Fig 10D) inmonkeys Consistent with these results the prodrug concen-trations in 5-minute plasma samples were negligible andbelow the assay limits of quantification (156 nM) CSFconcentrations of BMS-986169 were stable over the90-minute qEEG recording period (Fig 10D) and associatedwith a high level of predicted GluN2B occupancy of 97ndash98

DiscussionSelective inhibition of the GluN2B receptor subtype has

been proposed as a strategy to deliver novel therapeutics withketamine-like antidepressant efficacy and improved tolerabil-ity Using well established in vitro approaches (Mutel et al1998 Malherbe et al 2003 Ng et al 2008 Risgaard et al2013) BMS-986169 was identified as a novel agent with highaffinity for the GluN2B amino terminal domain allosteric site(Ki 5 4ndash6 nM) that potently and selectively inhibits GluN2Breceptor function (IC50 5 241 nM) BMS-986169 showednegligible activity in a broad assay panel consisting of40 additional targets Importantly BMS-986169 weaklyinhibited hERG channels (IC50 5 284 mM) an off-targetliability that has been challenging to address in chemicalscaffolds targeting the GluN2B NAM site (Brown et al 2011Layton et al 2011 Muumlller et al 2011)The in vivo characterization of BMS-986169 and the

phosphate prodrug BMS-986163 focused on 1) confirmationof occupancy and functional inhibition of GluN2B receptors2) testing in assays of antidepressant drug-like effects and side

Fig 6 BMS-986169 (3 mgkg iv) en-hances ex vivo hippocampal LTP 24 hoursafter dosing in rats (A and C) Timecourse of the normalized fEPSP ampli-tude (A) or slope (C) (B and D) Percent-age increase over baseline of the fEPSPamplitude (B) or slope (D) during the last5 minutes of recording following eachHFS application Results are presentedas the mean 6 SEM (n = 7group)Results at each HFS (B and D) wereanalyzed by one-tailed paired t test P005 compared with vehicle

388 Bristow et al

at ASPE



ownloaded from

effects and 3) testing on translational measures that can beinvestigated in healthy humans GluN2B occupancy in ro-dents was measured by ex vivo [3H]Ro 25-6981 binding(Fernandes et al 2015) and showed a dose-dependent in-crease that fully saturated at doses$3 mgkg Occupancy wastightly correlated with exposure and occupancyexposurerelationships were similar after iv dosing with BMS-986169or BMS-986163 confirming rapid and effective cleavage ofthe prodrug Functional inhibition of GluN2B receptors wasdemonstrated by inhibition of in vivo [3H]MK-801 binding(Fernandes et al 2015) [3H]MK-801 is a radiotracer thatlabels all NMDA receptor subtypes present in an openconformational state The assay can be used to show bothdirect binding site competition following treatment with otherNMDA receptor open-channel blockers and binding inhibitiondue to the functional effects of agents acting at othermodulatory sites (Murray et al 2000 Fernandes et al2015) Specifically the conformational change produced by

binding to the GluN2B allosteric site closes channels contain-ing the GluN2B subunit prevents [3H]MK-801 from accessingits binding site in the channel pore and reduces [3H]MK-801binding Previous studies have shown that GluN2B NAMsmaximally inhibit [3H]MK-801 binding by sim60 suggestingthat GluN2B-containing NMDA receptors represent 60of the total NMDA receptor population labeled in vivo(Fernandes et al 2015) In a similar manner treatment withBMS-986163 also inhibited [3H]MK-801 binding in ratsconfirming functional NMDA receptor inhibition in vivoHowever in contrast to other agents the maximal inhibitionachieved at doses fully occupying the GluN2B NAM bindingsite was sim40 suggesting a subpopulation of GluN2B-containing receptors that are less sensitive to functionalinhibition by BMS-986169 in vivo Lower inhibition couldreflect functional differences in discrete brain regions express-ing GluN2B receptors the current homogenate assay does notaddress this question and GluN2B functional inhibition is notdetectable using ex vivo [3H]MK-801 autoradiography (Lordet al 2013) Different receptor subtypesmdashnamely dihetero-meric GluN1GluN2B and triheteromeric GluN1GluN2A-GluN2B receptorsmdashmay also be involved since differencesin functional potency and maximal inhibition at tri- versus

Fig 7 (A) Increased LMA after iv ketamine treatment in mice (n =8ndash9group) (B) Effect of BMS-986169 (iv) or ketamine (30 mgkg iv) onLMA in mice (n = 7ndash8group) Results are presented as the mean6 SEMambulatory distance in 10-minute time bins recorded for 2 hours afterdosing Results were analyzed by two-way RM ANOVA followed byDunnettrsquos post-hoc test P 005 P 001 P 0001 comparedwith vehicle

Fig 8 BMS-986169 does not produce abnormal or dissociativehalluci-nogenic-like behaviors in cynomolgus monkeys (n = 8) Results arepresented as the mean 6 SEM total score for all abnormal behaviors (A)or dissociativehallucinogenic behaviors (B) observed up to 65 minutesafter im dosing with vehicle ketamine (Ket 1 mgkg im) or BMS-986169 (1ndash10 mgkg) Results were analyzed by RM ANOVA followed byDunnettrsquos post-hoc test P 001 P 0001 compared with vehicle


at ASPE



ownloaded from

diheteromeric receptors have been reported for CP-101606(Hansen et al 2014) Additional studies including in vivoapproaches such as pharmacological magnetic resonanceimaging may provide further insight into this interestingresult (Chin et al 2011)To provide evidence of antidepressant drug-like effects we

first tested BMS-986169 in the mouse FST a simple behav-ioral assay sensitive to many antidepressant drug classes andmultiple NMDA receptor agents (Porsolt et al 1977 Trullasand Skolnick 1990 Pilc et al 2013) As expected BMS-986169 reduced immobility 15 minutes after dosing consis-tent with results for other GluN2B NAMs (Li et al 2010Kiselycznyk et al 2015 Refsgaard et al 2017) We nextshowed that BMS-986169 robustly elevated ex vivo hippo-campal LTP measured 24 hours after treatment As evidenceof synaptic strengthening this finding is consistent with thehypothesis that rapid-acting antidepressants enhance synapticplasticity (Wohleb et al 2017) and with results for ketamineand the structurally distinct GluN2B NAM CP-101606 in thisassay (Burgdorf et al 2013 Graef et al 2015) The effect is

thought to arise from increased expression of synaptic pro-teins such as GluR1 and PSD95 and at least twomechanismsdownstream of NMDA receptor inhibition may be involved1) disinhibition of pyramidal glutamate neurons leading toactivity-dependent brain-derived neurotrophic factor re-lease and activation of extracellular signal-regulated kinaseAKTmTORC1 signaling andor 2) activity-independentblockade of extrasynaptic postsynaptic NMDA receptorsleading to reduced eEF2 kinase activity decreased eEF2phosphorylation and desuppression of brain-derived neuro-trophic factor and synaptic protein translation (Miller et al2016 Murrough et al 2017) Finally we showed thatBMS-986163 reduced latency to feed in the NSF assay amodel of stress-induced anxiety that is sensitive to chronicbut not acute treatment with classic antidepressant agents(Dulawa and Hen 2005) Importantly we demonstratedefficacy 24 hours after dosing a result consistent with thesustained efficacy seen after acute ketamine treatment in thisassay (Iijima et al 2012) With respect to target engagementLTP and NSF results suggest that high levels of GluN2B

Fig 9 BMS-986169 produces a transient impairment in working memory after im dosing in cynomolgus monkeys (n = 9) Effect on list-DMS accuracydetermined 30 minutes after dosing with BMS-986169 (01ndash3 mgkg im) (A) or at various time points after dosing with 1 mgkg (im) BMS-986169 (C)Results are presented as the mean 6 SEM percentage correct responses at short medium or long delays and were analyzed by two-way RM ANOVAfollowed by Dunnettrsquos post-hoc test P 0001 compared with vehicle Relationship between total plasma BMS-986169 concentration and impairmentat long delays determined 30 minutes after dosing with BMS-986169 (B) and at various time points after dosing with 1 mgkg BMS-986169 (D) Resultsare presented as the mean 6 SEM difference from vehicle in percentage correct responses at the long delay condition or the mean + SD total plasmaBMS-986169 concentration (n = 7ndash9group)

390 Bristow et al

at ASPE



ownloaded from

occupancy sustained for a critical duration are necessary todeliver efficacy 24 hours after dosing Robust LTP enhance-ment was seen at 3 mgkg which achieves sim95 peakoccupancy and maintains $80 occupancy for sim1 hourEfficacy in NSF was seen only at 10 mgkg whereas peakoccupancy was saturated at doses $3 mgkg suggesting theduration of occupancy is important This initial characteriza-tion suggests that BMS-986169BMS-986163 has potential asa rapid-acting antidepressant and supports progression tofurther testing in more complex animal models and investi-gation of the molecular mechanisms involved (Ramaker andDulawa 2017)While BMS-986169 showed low potential to produce

ketamine-like hyperactivity or abnormal behaviors impairedworking memory was observed in monkeys performing theCANTAB list-DMS task We chose im dosing for thesestudies to ensure stable plasma levels during the 30-minutetesting session and to facilitate demonstration of the effectImportantly this impairment was transient fully resolving by5 hours postdose and was closely related to plasma exposureswith subsequent pharmacokinetic modeling predicting amore rapid resolution after iv administration This profileis similar to other GluN2B NAMs tested in this model

(Weed et al 2016) and consistent with electrophysiologystudies showing that persistent activation of dorsolateralprefrontal cortical delay neurons is dependent upon GluN2Breceptors in monkeys performing working memory tasks(Wang et al 2013 Wang and Arnsten 2015) The temporaldifferences between the working memory and hippocampalLTP effects should be noted working memory deficits coincidewith the primary pharmacological effect ie GluN2B receptorinhibition whereas enhanced 24-hour LTP is measured at atime when downstream events leading to synaptic strength-ening have occurred Given the overlapping exposures and(predicted) occupancy across these assays the separation ofdoses providing antidepressant-like effects from those pro-ducing transient cognitive impairment within the first hour(s)of dosing may not be possible for this mechanismThe present results raise important questions about the

optimal GluN2B NAM candidate profile From this perspec-tive both intravenous (CP-101606) and oral agents (CERC-301 also known asMK-0657) have been examined in patientsIn the case of CP-101606 a single iv infusion improveddepression symptoms in a small cohort of TRD patientswith minimal dissociative effects (Preskorn et al 2008) Incontrast although CERC-301 initially appeared promising

Fig 10 Robust effects on the qEEG power band distribution after iv treatment with BMS-986163 in cynomolgus monkeys (n = 6) (A) Time course ofthe effect of 12 mgkg iv BMS-986163 on relative power in the beta 1 power band Results are presented as the mean 6 SEM relative power in1-minute time bins and were analyzed by two-way RM ANOVA followed by Holm-Sidak post-hoc test (B) Effect of BMS-986163 on relative power in thebeta 1 power band Results are presented as the mean 6 SEM area under the beta 1 relative power curve (AUC) and were analyzed by RM ANOVAfollowed by Dunnettrsquos t test P 005 P 001 P 0001 compared with vehicle (C) Treatment effect size for changes in the qEEG power banddistribution Results show Cohenrsquos d values calculated using the relative power AUC for each power band after treatment with BMS-986169 Statisticscorrespond to results from the AUC analysis as described in (B) (D) Mean (6SD) total plasma and CSF concentration of BMS-986169 measured inseparate subjects after iv administration of 12 mgkg BMS-986163 (n = 2group)


at ASPE



ownloaded from

(Ibrahim et al 2012) a recent study showed that treatmentwith 8 mg daily was not effective in severely depressed TRDpatients (wwwCerecorcom) It is possible that target engage-ment is inadequate at this dose of CERC-301 since preclinicalresults suggest that exposures reported in MDD patientsreceiving a 4ndash8 mg titration (Ibrahim et al 2012) only briefly(sim2 hours) achieve levels predicted to deliver sim50 occupancy(Addy et al 2009) Management of cognitive impairment maybe difficult for oral agents with a long pharmacokinetic half-life if human results confirm the high target engagementrequirement for antidepressant effects suggested by preclin-ical studies Thus an iv agent may be preferred sinceexposure can be tightly controlled to deliver transient highlevels of GluN2B occupancy while enabling monitoring forresolution of any transient cognitive impairment within aclinical settingTo facilitate the progression of BMS-986163 we examined

effects on the qEEG Previous studies have shown robustqEEG power band changes in monkeys treated with GluN2BNAMs (Keavy et al 2016) and BMS-986163 produced thesame pattern of response As an approach changes in qEEG-related measures for NMDA receptor antagonists generallyalign well across species and although not tested here othermeasures such as auditory evoked potentials also have trans-lational utility for GluN2B NAMs (Nagy et al 2016) Suchapproaches may provide valuable information for examplethe sustained qEEG effect was not anticipated after iv dosingbut is likely explained by the relatively stable CSF levels ofBMS-986169 during the recording window Thus qEEG-related measures provide a noninvasive way to monitor targetengagement in humans and potentially aide the optimizationof iv infusion regimensIn conclusion the present results support the further

investigation of BMS-986163 the phosphate prodrug of thenovel GluN2B NAM BMS-986169 as an iv agent withpotential for a rapid antidepressant effect

Acknowledgments

We acknowledge Keely Walsh Sharon Aborn James Brennan andMargaret Batchelder for veterinary science assistance and studysupport Michelle Nophsker for drug formulation support DavidLuchetti for dosing and sample collection support and Peng LiDauh-Rurng Wu and Richard Rampulla for chemistry synthesis andscale-up support

Authorship Contributions

Participated in research design Bristow Gulia Weed SrikumarY-W Li Graef Naidu Kumar Sivarao Keavy Shields J Li ZhangMathur Ramarao Vikramadithyan Thangathirupathy WarrierIslam Bronson Olson Macor Albright King Thompson MarcinSinz

Conducted experiments Gulia Weed Srikumar Graef NaiduSanmathi Aher Bastia Paschapur Kalidindi Kumar MolskiPieschl Fernandes Newberry Bookbinder Polino Keavy NewtonShields Simmermacher Kempson J Li Sinha Ramarao Thanga-thirupathy Warrier Thompson

Contributed new reagents or analytic tools Brown SivaraoShields Kempson Kallem Thangathirupathy Warrier KingThompson

Performed data analysis Bristow Gulia Weed Srikumar Y-WLi Graef Naidu Sanmathi Aher Paschapur Kalidindi KumarPieschl Fernandes Sivarao Keavy Shields Simmermacher J LiRamarao Thangathirupathy Warrier Islam Thompson

Wrote or contributed to the writing of the manuscript BristowGulia Weed Shields Simmermacher Olson Sinz

References

Addy C Assaid C Hreniuk D Stroh M Xu Y Herring WJ Ellenbogen A Jinnah HAKirby L Leibowitz MT et al (2009) Single-dose administration of MK-0657 anNR2B-selective NMDA antagonist does not result in clinically meaningful im-provement in motor function in patients with moderate Parkinsonrsquos disease J ClinPharmacol 49856ndash864

Berman RM Cappiello A Anand A Oren DA Heninger GR Charney DSand Krystal JH (2000) Antidepressant effects of ketamine in depressed patientsBiol Psychiatry 47351ndash354

Brown DG Maier DL Sylvester MA Hoerter TN Menhaji-Klotz E Lasota CCHirata LT Wilkins DE Scott CW Trivedi S et al (2011) 26-Disubstituted pyr-azines and related analogs as NR2B site antagonists of the NMDA receptor withanti-depressant activity Bioorg Med Chem Lett 213399ndash3403

Burgdorf J Zhang XL Nicholson KL Balster RL Leander JD Stanton PK GrossAL Kroes RA and Moskal JR (2013) GLYX-13 a NMDA receptor glycine-sitefunctional partial agonist induces antidepressant-like effects without ketamine-like side effects Neuropsychopharmacology 38729ndash742

Castner SA and Goldman-Rakic PS (1999) Long-lasting psychotomimetic conse-quences of repeated low-dose amphetamine exposure in rhesus monkeys Neuro-psychopharmacology 2010ndash28

Chin CL Upadhyay J Marek GJ Baker SJ Zhang M Mezler M Fox GB and Day M(2011) Awake rat pharmacological magnetic resonance imaging as a translationalpharmacodynamic biomarker metabotropic glutamate 23 agonist modulation ofketamine-induced blood oxygenation level dependence signals J Pharmacol ExpTher 336709ndash715

Dulawa SC and Hen R (2005) Recent advances in animal models of chronic antidepres-sant effects the novelty-induced hypophagia test Neurosci Biobehav Rev 29771ndash783

Fava M (2006) Pharmacological approaches to the treatment of residual symptomsJ Psychopharmacol 20(3 Suppl)29ndash34

Fernandes A Wojcik T Baireddy P Pieschl R Newton A Tian Y Hong Y Bristow Land Li YW (2015) Inhibition of in vivo [(3)H]MK-801 binding by NMDA receptoropen channel blockers and GluN2B antagonists in rats and mice Eur J Pharmacol7661ndash8

Gaynes BN Warden D Trivedi MH Wisniewski SR Fava M and Rush AJ (2009)What did STARD teach us Results from a large-scale practical clinical trial forpatients with depression Psychiatr Serv 601439ndash1445

Graef JD Newberry K Newton A Pieschl R Shields E Luan FN Simmermacher JLuchetti D Schaeffer E Li YW et al (2015) Effect of acute NR2B antagonisttreatment on long-term potentiation in the rat hippocampus Brain Res 160931ndash39

Graef JD Wei H Lippiello PM Bencherif M and Fedorov N (2013) Slice XVIvotrade anovel electrophysiology system with the capability for 16 independent brain slicerecordings J Neurosci Methods 212228ndash233

Greenberg PE Fournier AA Sisitsky T Pike CT and Kessler RC (2015) The eco-nomic burden of adults with major depressive disorder in the United States(2005 and 2010) J Clin Psychiatry 76155ndash162

Hansen KB Ogden KK Yuan H and Traynelis SF (2014) Distinct functional andpharmacological properties of triheteromeric GluN1GluN2AGluN2B NMDA re-ceptors Neuron 811084ndash1096

Ibrahim L Diaz Granados N Jolkovsky L Brutsche N Luckenbaugh DA HerringWJ Potter WZ and Zarate CA Jr (2012) A Randomized placebo-controlledcrossover pilot trial of the oral selective NR2B antagonist MK-0657 in patientswith treatment-resistant major depressive disorder J Clin Psychopharmacol 32551ndash557

Iijima M Fukumoto K and Chaki S (2012) Acute and sustained effects of a metab-otropic glutamate 5 receptor antagonist in the novelty-suppressed feeding testBehav Brain Res 235287ndash292

Keavy D Bristow LJ Sivarao DV Batchelder M King D Thangathirupathy S MacorJE and Weed MR (2016) The qEEG signature of selective NMDA NR2B negativeallosteric modulators a potential translational biomarker for drug developmentPLoS One 11e0152729

Kiselycznyk C Jury NJ Halladay LR Nakazawa K Mishina M Sprengel R GrantSG Svenningsson P and Holmes A (2015) NMDA receptor subunits and associatedsignaling molecules mediating antidepressant-related effects of NMDA-GluN2Bantagonism Behav Brain Res 28789ndash95

Layton ME Kelly MJ III and Rodzinak KJ (2006) Recent advances in the devel-opment of NR2B subtype-selective NMDA receptor antagonists Curr Top MedChem 6697ndash709

Layton ME Kelly MJ III Rodzinak KJ Sanderson PE Young SD Bednar RADilella AG McDonald TP Wang H Mosser SD et al (2011) Discovery of3-substituted aminocyclopentanes as potent and orally bioavailable NR2B subtype-selective NMDA antagonists ACS Chem Neurosci 2352ndash362

Lener MS Kadriu B and Zarate CA Jr (2017) Ketamine and beyond Investigationsinto the potential of glutamatergic agents to treat depression Drugs 77381ndash401

Li N Lee B Liu RJ Banasr M Dwyer JM Iwata M Li XY Aghajanian Gand Duman RS (2010) mTOR-dependent synapse formation underlies the rapidantidepressant effects of NMDA antagonists Science 329959ndash964

Lord B Wintmolders C Langlois X Nguyen L Lovenberg T and Bonaventure P(2013) Comparison of the ex vivo receptor occupancy profile of ketamine to severalNMDA receptor antagonists in mouse hippocampus Eur J Pharmacol 71521ndash25

Maeng S Zarate CA Jr Du J Schloesser RJ McCammon J Chen G and Manji HK(2008) Cellular mechanisms underlying the antidepressant effects of ketaminerole of alpha-amino-3-hydroxy-5-methylisoxazole-4-propionic acid receptors BiolPsychiatry 63349ndash352

Malherbe P Mutel V Broger C Perin-Dureau F Kemp JA Neyton J Paoletti Pand Kew JN (2003) Identification of critical residues in the amino terminal domain

392 Bristow et al

at ASPE



ownloaded from

of the human NR2B subunit involved in the RO 25-6981 binding pocketJ Pharmacol Exp Ther 307897ndash905

McClintock SM Husain MM Wisniewski SR Nierenberg AA Stewart JW TrivediMH Cook I Morris D Warden D and Rush AJ (2011) Residual symptoms indepressed outpatients who respond by 50 but do not remit to antidepressantmedication J Clin Psychopharmacol 31180ndash186

Miller OH Moran JT and Hall BJ (2016) Two cellular hypotheses explaining theinitiation of ketaminersquos antidepressant actions direct inhibition and disinhibitionNeuropharmacology 10017ndash26

Miller OH Yang L Wang CC Hargroder EA Zhang Y Delpire E and Hall BJ (2014)GluN2B-containing NMDA receptors regulate depression-like behavior and arecritical for the rapid antidepressant actions of ketamine eLife 3e03581

Mony L Kew JN Gunthorpe MJ and Paoletti P (2009) Allosteric modulators ofNR2B-containing NMDA receptors molecular mechanisms and therapeutic po-tential Br J Pharmacol 1571301ndash1317

Muumlller A Houmlfner G Renukappa-Gutke T Parsons CG and Wanner KT (2011)Synthesis of a series of g-amino alcohols comprising an N-methyl isoindolinemoiety and their evaluation as NMDA receptor antagonists Bioorg Med Chem Lett215795ndash5799

Murray F Kennedy J Hutson PH Elliot J Huscroft I Mohnen K Russell MGand Grimwood S (2000) Modulation of [3H]MK-801 binding to NMDA receptorsin vivo and in vitro Eur J Pharmacol 397263ndash270

Murrough JW Abdallah CG and Mathew SJ (2017) Targeting glutamate signallingin depression progress and prospects Nat Rev Drug Discov 16472ndash486

Mutel V Buchy D Klingelschmidt A Messer J Bleuel Z Kemp JA and Richards JG(1998) In vitro binding properties in rat brain of [3H]Ro 25-6981 a potent andselective antagonist of NMDA receptors containing NR2B subunits J Neurochem702147ndash2155

Nagy D Stoiljkovic M Menniti FS and Hajoacutes M (2016) Differential effects of anNR2B NAM and ketamine on synaptic potentiation and gamma synchrony rele-vance to rapid-onset antidepressant efficacy Neuropsychopharmacology 411486ndash1494

Ng FM Geballe MT Snyder JP Traynelis SF and Low CM (2008) Structural in-sights into phenylethanolamines high-affinity binding site in NR2B from bindingand molecular modeling studies Mol Brain 116 DOI 1011861756-6606-1-16

Paykel ES Ramana R Cooper Z Hayhurst H Kerr J and Barocka A (1995) Residualsymptoms after partial remission an important outcome in depression PsycholMed 251171ndash1180

Pilc A Wieronska JM and Skolnick P (2013) Glutamate-based antidepressantspreclinical psychopharmacology Biol Psychiatry 731125ndash1132

Popp S Behl B Joshi JJ Lanz TA Spedding M Schenker E Jay TM SvenningssonP Caudal D Cunningham JI et al (2016) In search of the mechanisms of ket-aminersquos antidepressant effects how robust is the evidence behind the mTOR ac-tivation hypothesis F1000 Res 5634 DOI 1012688f1000research82361

Porsolt RD Bertin A and Jalfre M (1977) Behavioral despair in mice a primaryscreening test for antidepressants Arch Int Pharmacodyn Ther 229327ndash336

Preskorn SH Baker B Kolluri S Menniti FS Krams M and Landen JW (2008) Aninnovative design to establish proof of concept of the antidepressant effects of theNR2B subunit selective N-methyl-D-aspartate antagonist CP-101606 in patientswith treatment-refractory major depressive disorder J Clin Psychopharmacol 28631ndash637

Ramaker MJ and Dulawa SC (2017) Identifying fast-onset antidepressants usingrodent models Mol Psychiatry 22656ndash665

Refsgaard LK Pickering DS and Andreasen JT (2017) Investigation ofantidepressant-like and anxiolytic-like actions and cognitive and motor side effectsof four N-methyl-D-aspartate receptor antagonists in mice Behav Pharmacol 2837ndash47

Risgaard R Nielsen SD Hansen KB Jensen CM Nielsen B Traynelis SFand Clausen RP (2013) Development of 2ꞌ-substituted (2S1ꞌR2ꞌS)-2-(carboxycyclopropyl)glycine analogues as potent N-methyl-d-aspartic acid receptoragonists J Med Chem 564071ndash4081

Roberts BM Shaffer CL Seymour PA Schmidt CJ Williams GV and Castner SA(2010) Glycine transporter inhibition reverses ketamine-induced working memorydeficits Neuroreport 21390ndash394

Trullas R and Skolnick P (1990) Functional antagonists at the NMDA receptorcomplex exhibit antidepressant actions Eur J Pharmacol 1851ndash10

Ustuumln TB Ayuso-Mateos JL Chatterji S Mathers C and Murray CJ (2004) Globalburden of depressive disorders in the year 2000 Br J Psychiatry 184386ndash392

Wang M and Arnsten AF (2015) Contribution of NMDA receptors to dorsolateralprefrontal cortical networks in primates Neurosci Bull 31191ndash197

Wang M Yang Y Wang CJ Gamo NJ Jin LE Mazer JA Morrison JH Wang XJand Arnsten AF (2013) NMDA receptors subserve persistent neuronal firing duringworking memory in dorsolateral prefrontal cortex Neuron 77736ndash749

Weed MR Bookbinder M Polino J Keavy D Cardinal RN Simmermacher-Mayer JCometa FN King D Thangathirupathy S Macor JE et al (2016) Negative allo-steric modulators selective for the NR2B subtype of the NMDA receptor impaircognition in multiple domains Neuropsychopharmacology 41568ndash577

Weed MR Taffe MA Polis I Roberts AC Robbins TW Koob GF Bloom FE and GoldLH (1999) Performance norms for a rhesus monkey neuropsychological testingbattery acquisition and long-term performance Brain Res Cogn Brain Res 8185ndash201

Wohleb ES Gerhard D Thomas A and Duman RS (2017) Molecular and cellularmechanisms of rapid-acting antidepressants ketamine and scopolamine CurrNeuropharmacol 1511ndash20

Address correspondence to Linda J Bristow Bristol-Myers SquibbCompany 5 Research Parkway Wallingford CT 06492 E-mail Lindabristowsunovioncom or Lawrence Marcin Bristol-Myers Squibb Company5 Research Parkway Wallingford CT 06492 E-mail lawrencemarcinbmscom


at ASPE



ownloaded from

therapeutics targeting the monoaminergic systems have beenavailable for many years many patients fail to show anadequate treatment response In this regard results fromthe STARD (Sequenced Treatment Alternatives to RelieveDepression) study showed that only one-third of patientstreated with a selective serotonin reuptake inhibitor achievedremission with 40 of those requiring more than 8 weeks oftreatment for resolution of their symptoms (reviewed byGaynes et al 2009) An additional 50 of patients showedpartial improvement but still exhibited residual symptoms(McClintock et al 2011) which increases the probability offuture relapse and poorer functional and psychosocial out-comes (Paykel et al 1995 Fava 2006) Finally almost onethird of patients were resistant to treatment and failed torespond to additional antidepressant drug switching andaugmentation approaches Clearly a high unmet medicalneed exists for new agents with novel mechanisms that showrapid and improved efficacy in patients with MDD andtreatment-resistant depression (TRD)In the search for novel mechanismsN-methyl-D-aspartate

(NMDA) receptor antagonists have been intensely investi-gated following the first demonstration that the nonselectiveNMDA receptor channel blocker ketamine is efficacious inTRD patients (Berman et al 2000) The clinical response toketamine was remarkable in that a single iv subanestheticdose produced rapid symptom relief within hours of dosingthat was sustained for up to 7 days Subsequent studies haveconfirmed that acute ketamine treatment rapidly achieves a66ndash77 response rate and 31 remission rate in TRDpatients and have also extended these findings to patientswith bipolar depression (reviewed by Lener et al 2017)Although this unique efficacy profile is widely acceptedketamine also produces undesirable effects including tran-sient dissociative and psychotomimetic effects that emergeduring the infusion but resolve within sim2 hours of dosingand may therefore be manageable within the treatmentenvironment However additional concerns including thepotential for neuronal injury and abuse have led to thesearch for alternative approaches that may retain theefficacy profile of ketamine but with fewer liabilities In thisregard agents that selectively inhibit the glutamate NMDA2B (GluN2B) receptor subtype have received attention Thistarget has been the focus of extensive drug developmentefforts aimed at the treatment of conditions such as strokeParkinsonrsquos disease and chronic pain (Layton et al 2006)GluN2B receptors possess an amino terminal domain allo-steric binding site that can be targeted to develop subtype-selective negative allosteric modulators (NAMs) (Monyet al 2009) In a small proof-of-concept study in selective

serotonin reuptake inhibitorndashresistant MDD patients asingle iv administration of the GluN2B NAM CP-101606[(1S2S)-1-(4-hydroxyphenyl)-2-(4-hydroxy-4-phenylpiperidino)-1-propanol] achieved a 60 response rate and 30 remission rate5 days after dosing with an infusion regimen that producedminimal dissociative effects (Preskorn et al 2008)Preclinical results also show that GluN2B NAMs have a

ketamine-like antidepressant profile and that the effects ofketamine may involve the GluN2B receptor subtype Acutetreatment with the GluN2B NAM Ro 25-6981 (4-[(1R2S)-3-(4-benzylpiperidin-1-yl)-1-hydroxy-2-methylpropyl]phenol) acti-vates the mechanistic target of rapamycin (mTOR) signalingpathway and increases synaptic protein expression effectsthought to underlie the rapid antidepressant effect of ketamine(Li et al 2010) Direct infusion of themTOR inhibitor rapamycininto the medial prefrontal cortex prevents the antidepressant-like behavioral effects of Ro 25-6981 and treatment with thea-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA)receptor antagonist NBQX (23dihydroxy-6-nitro-7-sulfamoyl-benzo[f]quinoxaline-23-dione) inhibits antidepressant activityand mTOR pathway activation after acute ketamine or Ro25-6981 administration suggesting common mechanisms(Maeng et al 2008 Li et al 2010) On the other hand geneticdeletion of GluN2B subunits in cortical pyramidal neuronsmimics and occludes the effects of ketamine on mTOR pathwayactivation and synaptic protein synthesis excitatory synaptictransmission and depression-like behavior (Miller et al 2014)Although it should be acknowledged that not all groups are ableto demonstrate themTOR pathway activation effect of ketamine(Popp et al 2016) these results have led to renewed interest inGluN2B NAMs and their potential as rapid-acting antidepres-sant agents for TRD (Miller et al 2016)The present studies describe the pharmacological character-

izationofBMS-986169 [(R)-3-((3S4S)-3-fluoro-4-(4-hydroxyphenyl)-piperidin-1-yl)-1-(4-methylbenzyl)pyrrolidin-2-one] anovelGluN2BNAM identified during a drug discovery program focused on theidentification of an iv therapeutic for TRD We also present re-sults forBMS-986163 [4-((3S4S)-3-fluoro-1-((R)-1-(4-methylbenzyl)-2-oxopyrrolidin-3-yl)piperidin-4-yl)phenyl dihydrogen phosphate]the phosphate prodrug of BMS-986169 which was developed toaddress the poor aqueous solubility of the parent molecule

Materials and MethodsAnimals Studies were conducted at either Bristol-Myers Squibb

(BMS Wallingford CT) or at the Biocon Bristol-Myers Squibb Re-search Center (BBRC Bangalore India)

BMS Studies Animals were housed in an Association for Assess-ment andAccreditation of Laboratory Animal Carendashaccredited facilityand maintained in accordance with the guidelines of the Animal Care

ABBREVIATIONS aCSF artificial cerebrospinal fluid AMPA a-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptor ANOVA analysisof variance BBRC Biocon Bristol-Myers Squibb Research Center BMS Bristol-Myers Squibb BMS-986163 4-((3S4S)-3-fluoro-1-((R)-1-(4-methylbenzyl)-2-oxopyrrolidin-3-yl)piperidin-4-yl)phenyl dihydrogen phosphate BMS-986169 (R)-3-((3S4S)-3-fluoro-4-(4-hydroxyphenyl)-piperidin-1-yl)-1-(4-methylbenzyl)pyrrolidin-2-one CANTAB Cambridge Neuropsychological Test Automated Battery CL confidence limitCP-101606 (1S2S)-1-(4-hydroxyphenyl)-2-(4-hydroxy-4-phenylpiperidino)-1-propanol CSF cerebrospinal fluid fEPSP field excitatory postsynapticpotential FST forced swim test GluN glutamate N-methyl-D-aspartate hERG human ether-a-go-gondashrelated gene potassium channel HFS high-frequency stimulation list-DMS list delayed match to sample LMA locomotor activity LTP long-term potentiation MDD major depressivedisorder MK-0657 (CERC-301) 4-methylbenzyl-(3S4R)-3-fluoro-4-[(2-pyrimidinylamino)methyl]-1-piperidinecarboxylate MK-801 (1)-5-methyl-1011-dihydro-5H-dibenzo[ad]cyclohepten-510-imine mTOR mechanistic target of rapamycin NAM negative allosteric modulator NBQX 23dihydroxy-6-nitro-7-sulfamoyl-benzo[f]quinoxaline-23-dione NMDA N-methyl-D-aspartate NSF novelty suppressed feeding Occ50 dose producing50 occupancy PEI polyethylenimine qEEG quantitative electroencephalogram RM repeated measures Ro-04-5595 1-[2-(4-chlorophenyl)ethyl]-1234-tetrahydro-6-methoxy-2-methl-7-isoquinolinol Ro 25-6981 4-[(1R2S)-3-(4-benzylpiperidin-1-yl)-1-hydroxy-2-methylpropyl]phenol TRD treatmentresistant depression

378 Bristow et al

at ASPE



ownloaded from









at ASPE



ownloaded from







380 Bristow et al

at ASPE



ownloaded from








at ASPE



ownloaded from






382 Bristow et al

at ASPE



ownloaded from










at ASPE



ownloaded from









384 Bristow et al

at ASPE



ownloaded from







at ASPE



ownloaded from









386 Bristow et al

at ASPE



ownloaded from






at ASPE



ownloaded from








388 Bristow et al

at ASPE



ownloaded from






at ASPE



ownloaded from





390 Bristow et al

at ASPE



ownloaded from







at ASPE



ownloaded from




Acknowledgments








References


























392 Bristow et al

at ASPE



ownloaded from






























at ASPE



ownloaded from









at ASPE



ownloaded from







380 Bristow et al

at ASPE



ownloaded from








at ASPE



ownloaded from






382 Bristow et al

at ASPE



ownloaded from










at ASPE



ownloaded from









384 Bristow et al

at ASPE



ownloaded from







at ASPE



ownloaded from









386 Bristow et al

at ASPE



ownloaded from






at ASPE



ownloaded from








388 Bristow et al

at ASPE



ownloaded from






at ASPE



ownloaded from





390 Bristow et al

at ASPE



ownloaded from







at ASPE



ownloaded from




Acknowledgments








References


























392 Bristow et al

at ASPE



ownloaded from






























at ASPE



ownloaded from







380 Bristow et al

at ASPE



ownloaded from








at ASPE



ownloaded from






382 Bristow et al

at ASPE



ownloaded from










at ASPE



ownloaded from









384 Bristow et al

at ASPE



ownloaded from







at ASPE



ownloaded from









386 Bristow et al

at ASPE



ownloaded from






at ASPE



ownloaded from








388 Bristow et al

at ASPE



ownloaded from






at ASPE



ownloaded from





390 Bristow et al

at ASPE



ownloaded from







at ASPE



ownloaded from




Acknowledgments








References


























392 Bristow et al

at ASPE



ownloaded from






























at ASPE



ownloaded from








at ASPE



ownloaded from






382 Bristow et al

at ASPE



ownloaded from










at ASPE



ownloaded from









384 Bristow et al

at ASPE



ownloaded from







at ASPE



ownloaded from









386 Bristow et al

at ASPE



ownloaded from






at ASPE



ownloaded from








388 Bristow et al

at ASPE



ownloaded from






at ASPE



ownloaded from





390 Bristow et al

at ASPE



ownloaded from







at ASPE



ownloaded from




Acknowledgments








References


























392 Bristow et al

at ASPE



ownloaded from






























at ASPE



ownloaded from






382 Bristow et al

at ASPE



ownloaded from










at ASPE



ownloaded from









384 Bristow et al

at ASPE



ownloaded from







at ASPE



ownloaded from









386 Bristow et al

at ASPE



ownloaded from






at ASPE



ownloaded from








388 Bristow et al

at ASPE



ownloaded from






at ASPE



ownloaded from





390 Bristow et al

at ASPE



ownloaded from







at ASPE



ownloaded from




Acknowledgments








References


























392 Bristow et al

at ASPE



ownloaded from






























at ASPE



ownloaded from










at ASPE



ownloaded from









384 Bristow et al

at ASPE



ownloaded from







at ASPE



ownloaded from









386 Bristow et al

at ASPE



ownloaded from






at ASPE



ownloaded from








388 Bristow et al

at ASPE



ownloaded from






at ASPE



ownloaded from





390 Bristow et al

at ASPE



ownloaded from







at ASPE



ownloaded from




Acknowledgments








References


























392 Bristow et al

at ASPE



ownloaded from






























at ASPE



ownloaded from









384 Bristow et al

at ASPE



ownloaded from







at ASPE



ownloaded from









386 Bristow et al

at ASPE



ownloaded from






at ASPE



ownloaded from








388 Bristow et al

at ASPE



ownloaded from






at ASPE



ownloaded from





390 Bristow et al

at ASPE



ownloaded from







at ASPE



ownloaded from




Acknowledgments








References


























392 Bristow et al

at ASPE



ownloaded from






























at ASPE



ownloaded from







at ASPE



ownloaded from









386 Bristow et al

at ASPE



ownloaded from






at ASPE



ownloaded from








388 Bristow et al

at ASPE



ownloaded from






at ASPE



ownloaded from





390 Bristow et al

at ASPE



ownloaded from







at ASPE



ownloaded from




Acknowledgments








References


























392 Bristow et al

at ASPE



ownloaded from






























at ASPE



ownloaded from









386 Bristow et al

at ASPE



ownloaded from






at ASPE



ownloaded from








388 Bristow et al

at ASPE



ownloaded from






at ASPE



ownloaded from





390 Bristow et al

at ASPE



ownloaded from







at ASPE



ownloaded from




Acknowledgments








References


























392 Bristow et al

at ASPE



ownloaded from






























at ASPE



ownloaded from






at ASPE



ownloaded from








388 Bristow et al

at ASPE



ownloaded from






at ASPE



ownloaded from





390 Bristow et al

at ASPE



ownloaded from







at ASPE



ownloaded from




Acknowledgments








References


























392 Bristow et al

at ASPE



ownloaded from






























at ASPE



ownloaded from








388 Bristow et al

at ASPE



ownloaded from






at ASPE



ownloaded from





390 Bristow et al

at ASPE



ownloaded from







at ASPE



ownloaded from




Acknowledgments








References


























392 Bristow et al

at ASPE



ownloaded from






























at ASPE



ownloaded from






at ASPE



ownloaded from





390 Bristow et al

at ASPE



ownloaded from







at ASPE



ownloaded from




Acknowledgments








References


























392 Bristow et al

at ASPE



ownloaded from






























at ASPE



ownloaded from





390 Bristow et al

at ASPE



ownloaded from







at ASPE



ownloaded from




Acknowledgments








References


























392 Bristow et al

at ASPE



ownloaded from






























at ASPE



ownloaded from







at ASPE



ownloaded from




Acknowledgments








References


























392 Bristow et al

at ASPE



ownloaded from






























at ASPE



ownloaded from




Acknowledgments








References


























392 Bristow et al

at ASPE



ownloaded from






























at ASPE



ownloaded from






























at ASPE



ownloaded from

Documents

D Modulator with Potential in Major Depressive Disordersjpet.aspetjournals.org/content/jpet/363/3/377.full.pdfAlthough this unique efficacy profile is widely accepted, ketamine also