comparative effects of ly3020371, a potent and selective...

1521-0103/361/1/68–86$25.00 http://dx.doi.org/10.1124/jpet.116.238121THE JOURNAL OF PHARMACOLOGY AND EXPERIMENTAL THERAPEUTICS J Pharmacol Exp Ther 361:68–86, April 2017Copyright ª 2017 by The American Society for Pharmacology and Experimental Therapeutics

Comparative Effects of LY3020371, a Potent and SelectiveMetabotropic Glutamate (mGlu) 2/3 Receptor Antagonist, andKetamine, a Noncompetitive N-Methyl-D-Aspartate ReceptorAntagonist in Rodents: Evidence Supporting the Use of mGlu2/3Antagonists, for the Treatment of Depression.

J.M. Witkin, S.N. Mitchell, K.A. Wafford, G. Carter, G. Gilmour, J. Li, B.J. Eastwood,C. Overshiner, X. Li, L. Rorick-Kehn, K. Rasmussen, W.H. Anderson, A. Nikolayev,V.V. Tolstikov, M.-S. Kuo, J.T. Catlow, R. Li, S.C. Smith, C.H. Mitch, P.L. Ornstein,S. Swanson, and J.A. MonnLilly Research Laboratories, Eli Lilly and Company, Indianapolis, IN (J.M.W., C.O., X.L., L.R.-K., K.R., W.H.A., A.N., V.V.T.,M.-S.K., J.T.C., R.L., S.C.S., C.H.M., P.L.O., S.S., J.A.M.); and Lilly Research Laboratories, Eli Lilly and Company, Windlesham,Surrey, United Kingdom (S.N.M., K.A.W., G.C., G.G., J.L., B.J.E.)

Received October 1, 2016; accepted January 5, 2017

ABSTRACTThe ability of the N-methyl-D-aspartate receptor antagonistketamine to alleviate symptoms in patients suffering fromtreatment-resistant depression (TRD) is well documented. In thispaper, we directly compare in vivo biologic responses in rodentselicited by a recently discovered metabotropic glutamate(mGlu) 2/3 receptor antagonist 2-amino-3-[(3,4-difluorophenyl)sulfanylmethyl]-4-hydroxy-bicyclo[3.1.0]hexane-2,6-dicarboxylicacid (LY3020371) with those produced by ketamine. BothLY3020371 and ketamine increased the number of spontane-ously active dopamine cells in the ventral tegmental area ofanesthetized rats, increased O2 in the anterior cingulate cortex,promoted wakefulness, enhanced the efflux of biogenic aminesin the prefrontal cortex, and produced antidepressant-relatedbehavioral effects in rodent models. The ability of LY3020371 toproduce antidepressant-like effects in the forced-swim assay inrats was associated with cerebrospinal fluid (CSF) drug levelsthat matched concentrations required for functional antagonist

activity in native rat brain tissue preparations. Metabolomicpathway analyses from analytes recovered from rat CSF andhippocampus demonstrated that both LY3020371 and ketamineactivated common pathways involving GRIA2 and ADORA1. Adiester analog of LY3020371 [bis(((isopropoxycarbonyl)oxy)-methyl) (1S,2R,3S,4S,5R,6R)-2-amino-3-(((3,4-difluorophenyl)thio)methyl)-4-hydroxy-bicyclo[3.1.0]hexane-2,6-dicarboxylate(LY3027788)] was an effective oral prodrug; when given orally, itrecapitulated effects of intravenous doses of LY3020371 in theforced-swim and wake-promotion assays, and augmented theantidepressant-like effects of fluoxetine or citalopram withoutaltering plasma or brain levels of these compounds. The broadoverlap of biologic responses produced by LY3020371 andketamine supports the hypothesis that mGlu2/3 receptor block-ade might be a novel therapeutic approach for the treatment ofTRD patients. LY3020371 and LY3027788 represent moleculesthat are ready for clinical tests of this hypothesis.

IntroductionThe naturally occurring amino acid glutamate plays a

preeminent role in excitatory neurotransmission within themammalian central nervous system. Rapid cellular responsesto glutamate aremediated bymembrane-associated, glutamate-gated ion channels (ionotropic glutamate receptors), which

consist of four pharmacologically defined receptor types[a-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA),kainate, N-methyl-D-aspartate (NMDA), and delta] each ofwhich are formed by heteromeric assemblies of multiple sub-unit proteins (AMPA: GluA1-4; kainate: GluK1-5; NMDA:GluN1, GluN2A-D, GluN3A,B; and delta: GluD1,2) (Trayneliset al., 2010). Glutamate also mediates slower, second-messenger-dependent processes through its activation of afamily of G-protein-coupled, seven transmembrane receptorsdx.doi.org/10.1124/jpet.116.238121.

ABBREVIATIONS: ACC, anterior cingulate cortex; AMPA, a-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid; ANOVA, analysis ofvariance; AP, anterior posterior; AUC, area under the curve; CSF, cerebrospinal fluid; DV, dorsal ventral; EEG, electroencephalogram; EMG,electromyogram; LC-MS/MS, liquid chromatography–tandem mass spectrometry; LY3020371, (1S,2R,3S,4S,5R,6R)-2-amino-3-[(3,4-difluorophenyl)sulfanylmethyl]-4-hydroxy-bicyclo[3.1.0]hexane-2,6-dicarboxylic acid; LY3027788, bis(((isopropoxycarbonyl)oxy)-methyl)(1S,2R,3S,4S,5R,6R)-2-amino-3-(((3,4-difluorophenyl)thio)methyl)-4-hydroxy-bicyclo[3.1.0]hexane-2,6-dicarboxylate; mGlu, metabotropic glutamate;LY354740.H2O, (1S,2S,5R,6S)-2-aminobicyclo[3.1.0]hexane-2,6-dicarboxylic acid monohydrate; ML, medial lateral; NBQX, 1,2,3,4-tetrahydro-6-nitro-2,3-dioxo-benzo[f]quinoxaline-7-sulfonamide; NMDA, N-methyl-D-aspartate; STR, striatum; VTA, ventral tegmental area.

68

at ASPE

T Journals on July 12, 2020

jpet.aspetjournals.orgD

ownloaded from

http://dx.doi.org/10.1124/jpet.116.238121

http://dx.doi.org/10.1124/jpet.116.238121

http://jpet.aspetjournals.org/

[metabotropic glutamate (mGlu) receptors]. These receptors(mGlu1–mGlu8) have historically been divided into threegroups on the basis of amino acid sequence homology, agonistpharmacology, and preferred G-protein-coupling partners (Pinand Duvoisin, 1995).Considerable scientific attention has been given to agents

that enhance mGlu2/3 receptor signaling, primarily owing tothe observation that ligands that either directly or allosteri-cally enhance the activity of these receptors can producerobust efficacy in preclinical tests that are controlled byelevated synaptic glutamate drive, including those associatedwith psychiatric disorders (Monn et al., 1997; Helton et al.,1998; Moghaddam and Adams, 1998; Cartmell et al., 1999,2000; Kłodzi�nska et al., 1999; Nakazato et al., 2000; Schoeppet al., 2003; Takamori et al., 2003; Swanson et al., 2005;Rorick-Kehn et al., 2007; Jones et al., 2011; Ago et al., 2012)and neurologic conditions (Neugebauer et al., 2000; Simmonset al., 2002; Jones et al., 2005; Du et al., 2008; Kumar et al.,2010; Caraci et al., 2011). Heightened interest came from thedemonstration of clinical efficacy in both anxiety (Dunayevichet al., 2008) and schizophrenia patients (Patil et al., 2007),although in the latter case antipsychotic efficacy may berestricted to patient subpopulations (Kinon et al., 2015;Nisenbaum et al., 2016).As part of a research effort aimed at the identification of

potent and selective antagonists for mGlu2/3 receptors, wediscovered (1S,2R,3S,4S,5R,6R)-2-amino-3-[(3,4-difluorophenyl)-sulfanylmethyl]-4-hydroxy-bicyclo[3.1.0]hexane-2,6-dicarboxylicacid (LY3020371) (Fig. 1) (Smith et al., 2012; Chappell et al.,2016). This molecule was characterized as a highly potent andselective orthosteric antagonist of recombinant human mGlu2/3receptors, endogenously expressed rat cortical and hippocampalmGlu2/3 receptors, and native mGlu2/3 receptors present insynaptosomes prepared from surgically resected human braintissue (Witkin et al., 2016b).We further demonstrated that whenadministered by the intravenous route at doses between 0.3 and10 mg/kg, exposures of LY3020371 in the cerebrospinal fluid(CSF) fully covered concentrations required to effectively blockmGlu2/3 receptor activation in native tissues (Witkin et al.,2016b). Based on these data, LY3020371 was considered anexemplary tool with which to probe the biologic signifi-cance of mGlu2/3 receptor antagonism in vivo by parenteraldosing. In this paper, we describe the results of a compara-tive assessment of in vivo electrophysiological, neurochem-ical, behavioral, and metabolomic responses elicited byLY3020371 and ketamine in rodents, highlighting thebroadly overlapping biologic effects induced by these

agents in vivo. In addition, we describe the attributes ofbis(((isopropoxycarbonyl)oxy)-methyl) (1S,2R,3S,4S,5R,6R)-2-amino-3-(((3,4-difluorophenyl)thio)methyl)-4-hydroxy-bicyclo[3.1.0]hexane-2,6-dicarboxylate (LY3027788) (Fig. 1),a diester form of LY3020371, and its utility in producingbiologically relevant plasma concentrations of active moiety(LY3020371) following oral administration in rodents.

Materials and MethodsAll experiments were conducted according to the Guidelines for

Care and Use of Laboratory Animals under protocols approved by aninstitutional animal care and use committee andmonitored by animalcare and use groups at a local level. We only used male rodents in thepresent series of experiments for two reasons; first, laboratoryprescendent and our historical data set in males, and second, ease ofanimal housing in our vivarium, which requires isolated rooms formales and females. Experiments performed in the United Kingdomwere in accordance with the 1986 United Kingdom Animals ScientificProcedures Act and Lilly UK ethical review (https://www.gov.uk/government/publications/consolidated-version-of-aspa-1986).

Compounds

LY3020371and its diester prodrug,LY3027788 (Smith et al., 2012), and(1S,2S,5R,6S)-2-aminobicyclo[3.1.0]hexane-2,6-dicarboxylic acid monohy-drate (LY354740.H2O) (Monn et al., 1997) were synthesized at Eli Lillyand Company. Imipramine was purchased from Sigma Chemical Co. (St.Louis,MO) and1,2,3,4-tetrahydro-6-nitro-2,3-dioxo-benzo[f]quinoxaline-7-sulfonamide (NBQX), ketamine, and (1)-ketamine were purchased fromTocrisBioscience (Ellisville,MO).LY3020371,LY354740, andNBQXweredissolved in water and titrated to solution with dilute NaOH. LY3027788.HCl was suspended in 10% acacia and dosed orally, whereas all othercompoundswere given by i.v. (LY3020371), i.p. (LY3020371, LY354740) ors.c. injection (NBQX). All compounds and vehicle were administered in avolume of 10 ml/kg (mice) or 1 or 2 ml/kg (rats).

Pharmacodynamic Effects of LY3020371 in Comparisonwith Ketamine

Dopamine Neuron Activity in the Ventral Tegmental Area(VTA) of Rats. Male Sprague-Dawley rats (230–350 g) from TaconicFarms (Germantown, NY) or Harlan Laboratories (Indianapolis, IN)were anesthetized using chloral hydrate (400 mg/kg, i.p.), followed bysupplemental doses as needed. Rats were placed on a heating pad tomaintain body temperature at 37°C throughout the procedures. An i.v.catheter, consisting of PE10 tubing (Becton Dickinson, Sparks, MD)connected to a 1 ml tuberculin syringe (Becton Dickinson, FranklinLakes, NJ) via a 30 gauge hypodermic needle (1 inch length; ExelInternational, Los Angeles, CA), was placed in the jugular vein foradministration of test compounds and supplemental anesthesia. The

Fig. 1. Structures of LY3020371 and an orally biovailable prodrug LY3027788.

mGlu2/3 Receptor Antagonism as Antidepressant 69

at ASPE



ownloaded from

https://www.gov.uk/government/publications/consolidated-version-of-aspa-1986

https://www.gov.uk/government/publications/consolidated-version-of-aspa-1986


catheter was secured into the vein using suture thread (3-0 Black BraidedSilk; Roboz Surgical, Rockville, MD), tied tightly enough to preventslippage but not so tight as to constrict blood flow. Following catheterimplantation, rats were positioned tightly into a stereotaxic apparatus(David Kopf Instruments, Tujunga, CA), a hole was drilled in the skulloverlying the target VTA coordinates [anterior posterior (AP): 25.3 to25.7mm,medial lateral (ML): 10.5 to 10.9mm, dorsal ventral (DV): 26.0 to28.5 mm, relative to bregma], and the dura was carefully resected. Asingle-barrel micropipette (Radnoti Starbore (Radnoti Glass Technol-ogy, Monrovia, CA) capillary tube, 1.8 mm outside diameter; pulledusing a Narishige (Tokyo, Japan) PE-2 vertical puller; broken back to atip diameter approximately 2–6 mm and filled with 2 M NaCl solution;impedance 3–10 MV) was mounted into a Burleigh (Fishers, NY)Inchworm 8200 microdrive for single-unit extracellular recordings.

Electrophysiological signalswere amplified byaDagan (Minneapolis,MN) 2400 preamplifier (low-cut 300 Hz, high-cut 3000 Hz; gain 1-fold).Data were recorded and analyzed using a Micro1401 data acquisitionsystem with Spike2 software (Cambridge Electronic Design, Ltd.,Cambridge, United Kingdom). Electrophysiological properties of spon-taneously active VTA dopamine cells were characterized by loweringrecording electrodes slowly (3 mm/s) through the dorsoventral extent ofthe VTA along 9–12 predefined tracks (separated by 200 mm) in a gridpattern throughout the entire nucleus. Dopamine cells were identifiedand distinguished from nondopamine cells based on their unique long-duration (2.5–4 ms), triphasic waveform, and slow irregular firing rate(2–10 spikes/s). For each track, the number of spontaneously activedopamine cells, the average firing rate, and the percentage of spikesoccurring in bursts were analyzed. A burst was defined as a period ofrapid cell firing, with burst initiation identified by an interspikeinterval ,80 ms between two consecutive action potentials, withinterspike interval .160 ms indicating the end of the burst.

For each dose level, spontaneously active dopamine cells werecharacterized throughout 2 to 3 tracks. Doses were administered(i.v., 10-minute pretreatment) to n 5 6 rats in ascending, cumulativeorder, such that vehicle was administered 10minutes prior to the firsttrack; 0.3mg/kgLY3020371was administered prior to track 4; 1mg/kgLY3020371was administered 10minutes prior to track 7; and 3mg/kgLY3020371 was administered 10 minutes prior to track 10. In adifferent cohort, the AMPA antagonist, NBQX (10 mg/kg, i.v.), wastested for its effects on VTA dopamine cells when administered alone,and also its ability to block the increase in spontaneously activedopamine cells produced by LY3020371 (3 mg/kg, i.v.) in n 5 8 rats.Finally, the NMDA antagonist ketamine HCl (tested as the racemate,17 mg/kg i.v., 10-minute pretreatment) was tested alone in a separatecohort of n 5 6 rats (formulations not corrected for salt weight).LY3020371 and NBQX were dissolved in sterile water with theaddition of 1 N NaOH (titrated to pH 7 to 8 using 8.5% lactic acid inwater). Ketamine was dissolved in sterile water.

The percentage of spikes occurring in bursts was determined usingSpike2 software (version 5.21, CED Ltd., Cambridge, United King-dom). The mean number of spontaneously active dopamine cells, themean firing rate, and the mean percentage of spikes occurring inbursts were analyzed using separate repeated-measures analysis ofvariance (ANOVA), followed byDunnett’s post hoc comparisons versusrespective vehicle (SAS version 9.3, SAS Institute, Cary, NC). Forcomparison purposes, effects of LY3020371, NBQX, and ketaminewere calculated as percent change from vehicle by dividing each druggroup mean by its respective vehicle group mean, multiplying by 100,and plotting together on the same graph.

Effects on Brain Oxygen. Male, Wistar rats (n 5 36) wereimplanted with carbon paste electrodes in the anterior cingulatecortex (ACC) (AP12.0mm,ML60.5mm,DV22.0mm fromdura) andthe striatum (STR) (AP 11.2 mm, ML 63.4 mm, DV 24.2 mm fromthe dura). At time of surgery, animals weighed 250–320 g. At the timeof testing, animals weighed 420–700 g. Animals were cabled for15–30 minutes predosing to establish a stable tissue oxygen baselinesignal. At time 0, animals were treated with either vehicle (saline i.p.),LY3020371 (1, 3, or 10mg/kg, i.p.), or (S)-(1)-ketamineHCl (10mg/kg,

s.c.) and oxygen signals recorded for 90 minutes. Animals received alltreatments, dosed once weekly per animal in a randomized within-subjects design. At the conclusion of the session they were returned totheir home cage. The study lasted 5 weeks.

Four-channel potentiostats (EA164 Quadstat) and 16-channelE-Corder data acquisition systems with chart software (all from eDAQ,New South Wales, Australia) were used to monitor and record oxygensignals. Data were sampled at 1 KHz and down-sampled to 1 Hz for thepurposes of analysis. Oxygen response-time courses from ACC and STRwere normalized to a time point immediately before injection time andpostinjection area under the curve (AUC) calculated for statisticalanalysis. All statistics were calculated using STATISTICA version 9(Dell Software, Round Rock, TX). A general linear model approach wasused in a repeatedmeasures within-subjects design. Amain effect of dosewas followedwith Tukey’s post hoc comparisons to determine the effect ofeach treatment group relative to the vehicle group. Following exclusionsdue to O2 signal quality, AUC statistical outliers, and histologic verifica-tion of carbon paste electrode placements, 11 animals from theACCgroupand 17 animals from the STR group remained in the final analysis.

Effects on Neurochemical Efflux in the Rat Medial PrefrontalCortex. Microdialysis probes (MAB 4.7.4 Cu 6 kDa cutoff; RoYemScientific Ltd., Luton, United Kingdom) were implanted under isoflur-ane anesthesia into the medial prefrontal cortex of male Wistar rats(approximately 300–400 g; Charles River Laboratories, Margate,United Kingdom) using the following coordinates (from bregma anddura surface; probe angled at 12°): AP 5 12.8 mm, ML 5 11.5 mm,DV 5 25.0 mm. The day after probe implantation, animals wereconnected to a liquid swivel suspended on a counterbalanced arm,allowing the probes to be perfused with artificial CSF containing NaCl(141 mM), KCl (5 mM), MgCl2 (0.8 mM), and CaCl2 (1.5 mM) via aninfusion pump flowing at 1.5 ml/min. Following a 90-minute presamplewashoutperiod, dialysate sampleswere collected at 20-minute intervalsand immediately frozen on dry ice prior to analysis by liquidchromatography–tandem mass spectrometry (LC-MS/MS). After sixbaseline samples (120 minutes), animals were injected with drug(LY3020371, 10 mg/kg i.p. dissolved in water buffered to pH 6 to7 with 1 N NaOH; S-(1)-ketamine hydrochloride 10 mg/kg free bases.c. dissolved in 5% glucose), or corresponding vehicle.

To each thawed dialysis sample (29 ml) was added: 20 ml buffer (1 M2-[bis(2-hydroxyethyl)amino]-2-(hydroxymethyl)propane-1,3-diol,pH 10), 20 ml mixed deuterated standard, and 260 ml 0.1% w/v dansylchloride (in acetone). The samples were vortexed and heated at 65°Cfor 30 minutes, and then dried under N2 and resuspended in 40 ml 50:50 (v/v) acetonitrile:water (containing 10mMammonium formate and0.06% formic acid). The samples were then centrifuged at 13000 rpmfor 10 minutes at ambient temperature and 35 ml pipetted into03-FIVR vials; 10 ml was injected onto the LC-MS/MS using a CTCPALHTC-xt Autosampler (CTCAnalytics AG, Zwingen, Switzerland).

Chromatographic separation of dansylated samples (including drugstandards) was performed under a 13-minute gradient (including washoutand a re-eqilibration step) using Shimadzu (Kyoto, Japan) LC-20AD XRbinary pumps (plus a CBM-20 controller) and a 2.6 mm PhenomenexKinetexXB-C18HPLCcolumn (Phenomenex,Torrance,CA).MobilephaseA consisted of acetonitrile/water 5:95 (v/v), 2 mM ammonium formate, and0.06% formic acid, and mobile phase B consisted of acetonitrile/water 95:5 (v/v), 2 mM ammonium formate, and 0.06% formic acid.

An AB Sciex API5500 (Sciex, Framingham, MA) was operated intwo periods: 1) positive turbo ion spray mode for detection ofacetylcholine, and 2) with positive and negative turbo ion spraymodesfor all other analytes (negative for glycine, glutamate, and glutamine)using a 20 ms dwell time. For most dansylated analytes, the dansylfragment (m/z 170 or 171) was the abundant product ion and wassubsequently used for multiple reactions monitoring. Additional frag-ment ions were selected for the multiple reactions monitoring of5-hydroxytryptamine (m/z 380), dopamine (m/z 619), glutamate (m/z144, negative mode), glutamine (m/z 234, negative mode), glycine (m/z234, negative mode), 3,4-dihydroxyphenylglycine (m/z 386), telemethylhistamine (m/z 188), and nondansylated acetylcholine (m/z 87).

70 Witkin et al.

at ASPE



ownloaded from


Data were expressed as a percentage of a preinjection control period,obtained by averaging three samples prior to drug delivery (5 100%)and expressing values as a percentage of this value. The amount ofanalyte in each microdialysate sample was recorded as a peak area orheight. Calibration curves were also constructed to allowmeasurementof LY3020371 or S-(1)-ketamine in the same microdialysate sample.Statistical analyses were undertaken using either ANOVA with re-peated measures (fit), to compare response profiles following drug orvehicle administration, or by one-way ANOVA (using JMP software,SAS Institute) for comparison of the AUC. A probability value ofP , 0.05 was considered to be statistically significant.

Metabolomic Analysis. We used an in-house mass spectrometry–based metabolomics platform, which allowed routine detectionof .5000 metabolic features. In this study the resulted data setsyielded statistically significant abundant changes in .240 metabo-lites, excluding peptides and intact lipids (Xia et al., 2012; Tolstikovet al., 2013). Statistical analysis was performed with JMP version 9.03(SAS Institute) and MetaboAnalyst version 2.0 (Cary, NC). Linearregression and curve clustering techniques to the time course expres-sion profile of all of the metabolites under different treatment schemeswere applied to extract information related to altered metabolism atdifferent time points. Ingenuity Systems (http://www.ingenuity.com/)was used to carry out pathway and network analyses.

Animals were dosed (i.p.) with vehicle, 10 mg/kg LY3020371, or10 mg/kg ketamine. Doses were based on comparable effects in the ratforced-swim assay. Tissue from hippocampus and CSF were harvestedat 1 hour postdosing. The harvest of these samples was synchronized tominimize the influence of diurnal fluctuations of metabolites on theprimary dependent measures of drug and dose. Tissue and biofluidsamples were processed and analyzed in randomized fashion undervalidated standard operating procedures. Identification of the potentialbiomarkers, which were not present in gas chromatography/liguidchromatography-mass spectrometer libraries, was performed using ahydrophilic-interaction chromatography - liquid chromatography massspectrometer discovery platform and further validated with the au-thentic standards comparisons. Data from multiple platforms werenormalized to fresh weights and/or to biofluid volumes taken formetabolite extraction and subjected to rigorous biostatistical analyses.

We used an ANOVA model to compare the effect of treatments.Multiplicity adjusted P values were calculated for each contrast ofinterest. Additionally, to identify biomarkers and discern the patternof strong diurnal effect on many metabolites, we applied curveclustering techniques to the expression profile of all of the metabolitesunder different treatment schemes. The significantly altered metab-olites were then used to discern the metabolic pathway changes atdifferent time points by bioinformatics analytic methods.

Sleep-Wake Assessments. The parallel groups study design wascarried out across five treatment days, each separated by at least 1 week.Animals were assigned in no particular order to a dose group, and noanimal received a given dose more than once. In general, animals werenot naive to drug treatments. At least 7 days washout preceded andfollowed any treatment. Adult, male Wistar rats (approximately250–300 g at time of surgery, Charles River Laboratories) wereanesthetized (2% isoflurane in 100% oxygen) and surgically preparedwith a cranial implant that permitted chronic electroencephalogram(EEG) and electromyogram (EMG) recording. The body temperature andlocomotor activity were monitored via a miniature transmitter (Minimit-ter PDT4000G, Philips Respironics, Bend, OR) surgically placed in theabdomen during the same anesthetic event in which the cranial portionwas implanted. The cranial implant consisted of stainless steel screws(2 frontal [13.5 AP frombregma,62.0ML] and 2 occipital [26.5 AP,65.2ML]) for EEG recording. Two Teflon-coated stainless steel wires werepositioned under the nuchal trapezoid muscles for EMG recording. Ananalgesic (buprenorphine 0.05 mg/kg s.c.) was administered preopera-tively, at the end of the surgery day, and the morning of the firstpostoperative day. To provide additional pain relief, Metacam (melox-icam) 0.15 mg/kg by mouth was administered for 6 days postsurgery. Anantibiotic [Ceporex (cefalexin) 20 mg/kg, by mouth] was administered

24 hours before and again immediately before surgery, and for 7 daysafter surgery. At least 3 weeks were allowed for recovery.

Animals were individually housed with a custom engineered flexibletether connected at one end to the commutator and at the other end tothe animal’s cranial implant. Animals were maintained on a 24-hourlight-dark cycle (12:12) at 21°C6 2°C with ad libitum available food andwater. Light intensity averaged 35–40 lux at midlevel inside the cage.Relative humidity averaged 50% approximately. Animals were un-disturbed for 48 hours before and after each treatment.

Amplified EEG 10,000-fold (band-pass filtered 1–30 Hz: GrassCorp., Quincy, MA) with an initial digitization rate of 400 Hz andamplified EMG (band-pass 10–100Hz, rms integration) were collectedfrom the fixed electrodes. Concurrently, body temperature, locomotoractivity, and food and drink-related activity were recorded.SCORE2004 (Lilly Research Labs, Windlesham, Surrey, UK), an

Fig. 2. LY3020371 and ketamine (17 mg/kg, i.v.) increased the number ofactively firing dopamine neurons in the VTA of anesthetized rats (A)[F(3,15) = 6.26, P = 0.005]. The effect of LY3020371 is significantlyattenuated by the AMPA receptor antagonist NBQX (10 mg/kg, i.p.) (A).Neither LY3020371 or ketamine affected the firing rate (B) or thepercentage of spikes occurring in bursts (C). Each bar represents themean 6 S.E.M. of n = 6 rats.


at ASPE



ownloaded from

http://www.ingenuity.com/


automated sleep-wake and physiologic monitoring systemwas used torecord and determine vigilance state, which is described and docu-mented elsewhere (Van Gelder et al., 1991; Seidel et al., 1995; Oliveet al., 1998; Phillips et al., 2012). Vigilance states were classifiedonline as non-rapid eye movement sleep, rapid eye movement sleep,wake, or theta-dominated wake every 10 seconds using the EEGperiod and amplitude feature extraction and ranked membershipalgorithms. A fast-Fourier transform was used to calculate thespectral power of the EEG in each 10-second epoch in 0.1 Hz bins.Offline, individually taught EEG vigilance-state templates and EMGcriteria differentiated states of arousal.

The post-treatment observation time was divided into three post-dosing periods. Time of dosing was defined as hour5 0, and each hourincluded measurements up to and not including the following hour.LY3020371 was dosed by the i.v. route and LY3027788.HCl was dosedby the oral route at 5 hours after the lights-on period. Period 1 wasdefined as the first 7 hours postdosing, denoted as hours 0–6, andduring which the lights were on; period 2 was defined as thesubsequent 12 hours, i.e., 7–18, during which time the lights wereoff. The measured outcomes were summarized in each period bycomputing either the mean hourly, maximum hourly, or the cumula-tive value across each period. All outcomes were analyzed by a mixed-model repeated measures ANOVA using treatment (drug dose) andtreatment date as factors. Adjusted mean and S.E. values weresummarized for each treatment group.

Antidepressant-Like Effects

Forced-Swim Assay. All animals were experimentally and drugnaive at the time of testing and were used for only one experiment.

Rats. Male Sprague-Dawley rats (250–275 g, from Harlan Sprague-Dawley, Indianapolis, IN) were received 7 days prior to testing. Theywere housed four rats per cage. Animals weighed about 300 g whentested. Animals were brought to the testing room at least 1 hour prior totesting. Rats were placed in clear plastic cylinders (diameter: 18 cm;height: 40 cm) filled with water (22°C–25°C) to a depth of 16 cm for15 minutes. On the next day, the same procedure was followed in thepresence of vehicle or drug (i.v.); immobility time of each rat wasrecorded for the first 5minutes of the experiment and the ratswere thenremoved from the chamber, dried, and warmed. A rat was regarded asimmobile when floating motionless or making only those movementsnecessary to keep its head above the water. The data—immobility(seconds)—was copied into a JMP data sheet, and analyzed by one-wayANOVA followed by post hoc Dunnett’s tests if significant at P , 0.05.

Mice. Male, National Institutes of Health Swiss mice (20–25 g,Harlan Sprague-Dawley) were used. In other experiments, mice withreceptor deletions were studied (mGlu22/2, see the subsequentdiscussion); these mice were bred by heterozygote � heterozogotebreeding and used as littermates for2/2 and1/1mouse comparisons(Taconic Farms). Mice were placed in clear, plastic cylinders (di-ameter: 10 cm; height: 25 cm) filled to 6 cm with 22°C–25°C water for6minutes. Different groups ofmicewere pretreatedwith either vehicle

Fig. 3. Both ketamine and LY3020371 dose dependently increase tissue oxygen in the ACC of rats after i.p. dosing. In contrast, ketamine but notLY3020371 enhance tissue oxygen in the dorsal STR of rats. Each point represents mean6 S.E.M. data from 16 (cingulate) or 17 rats (STR). Analysis ofthe AUC revealed a significant effect of treatment [ACC: F(4,40) = 12.79, P , 0.001; STR: F(4,64) = 15.02, P, 0.001), and post hoc comparisons showedthat the 10 mg/kg LY3020371 group was significantly different from vehicle in the ACC (P = 0.0007), while the 10 mg/kg ketamine group wassignificantly different from vehicle in both the ACC (P = 0.0001) and STR (P = 0.0001). *P, 0.05, **P, 0.01, ***P, 0.001 in comparison with the vehicletreatment group.

72 Witkin et al.

at ASPE



ownloaded from


or LY3020371 (i.p., 30 minutes prior to testing) or LY3027788.HCl (bymouth, 60 minutes prior to testing). The duration of immobility wasrecorded during the last 4 minutes of a 6-minute trial. A mouse wasregarded as immobile when floating motionless or making only thosemovements necessary to keep its head above the water. Data wereanalyzed by post hoc Dunnett’s test with the alpha level set at 0.05. Theamount of time spent immobile was measured. The mean 6 S.E.M.values were subjected to ANOVA followed by Dunnett’s test with P ,0.05 set as the error rate for statistical significance.

mGlu22/2 mice. mGlu22/2 mice were generated by homologousrecombination as described in detail in the literature (Zhai et al., 2002;Linden et al., 2005). Separate mGlu2 and mGlu3 receptor knockoutmouse strains were generated by homologous recombination as pre-viously described for the mGlu8 receptor knockout mice (Zhai et al.,2002). Briefly, DNA fragments containing mouse mGlu2 and mGlu3receptor coding regions were isolated from a 129SVJ genomic library(Adams et al., 2005). In both cases, an expression cassette of theneomycin resistance gene was inserted into the third exon. Eachtargeting construct was injected into the R1 line of mouse embryonicstem cells. The homologous recombinants were confirmed by Southernhybridization analysis and these embryonic stemcellswere injected intomurine C57Bl/6 blastocysts. The resulting chimeric males were matedwith ICR (CD-1) females. Male offspring carrying the null allele werebackcrossed for three generations (N3) with ICR (CD-1) females.Heterozygous offspring were then interbred to obtain age-matchedwild-type and receptor knockout mice with genotyping as previouslydescribed (Linden et al., 2005).

Enhancement of Quinpirole-Induced Locomotor Activity

Sensitization within the dopamine pathways can be measured in anumber of ways, one of which is to evaluate the increased sensitivity to

the locomotor effects of dopaminergic agonists (Willner, 1997, D’Aquilaet al., 2000). A mouse model has been disclosed that was used toevaluate the impact of antidepressant treatments on the behavioraleffects of the dopamine D3/2 agonist, quinpirole (Marsteller et al., 2009).Male balbC mice were used in these studies. Locomotor activity wasmeasured with a 20-station Photobeam Activity System (San DiegoInstruments, SanDiego, CA) with seven photocells per station. Animalswere weighed, injected with compound (i.p. for LY3020371 and bymouth for LY3027788.HCl), and returned to their home cages for 0.5hours. They were then injected with quinpirole and placed back in thehome cages for an additional 2 hours, after which they were placed inthe locomotor activity boxes (40.6 � 20.3 � 15.2 cm). Locomotion wasassessed for a 2-hour period post dosing with quinpirole. Data werecollected as total ambulation (where ambulation was defined as thesequential breaking of adjacent photobeams) in 10-minute intervals forthe entire period. Themean of the total ambulations for the fourth hourof the experimental session was summarized. ANOVA with post hocDunnett’s test was used to evaluate the dose-response functionsseparately of drug alone or drug plus quinpirole. Student’s t test wasused to compare each drug dose alone versus drug dose given withquinpirole. Ten mice were used per group and each animal was usedonly once.

Drug Combination Studies. Doses of fluoxetineHCl, citalopramHCl, and LY3027788.HCl were selected based on prior dose-effectstudies and were as follows: fluoxetine (10 mg/kg, i.p. 30 minutes);citalopram (0.3mg/kg, i.p., 30minutes); and LY3027788.HCl (10mg/kg,by mouth, 60 minutes prior). These doses were again shown in thisstudy to be inactive (no statistically significant difference comparedwith vehicle controls) when given alone when studied under the forced-swim assay inmice (methods as described previously) andwere used fordrug combination experiments. Doses of each drug alone and the drug

Fig. 4. Increases in monoamine efflux in the medial prefrontal cortex of freely moving rats by S-(+)-ketamine (10 mg/kg, s.c.). Each point representsmean 6 S.E.M. data from six rats. Data were analyzed by ANOVA; *P , 0.05; **P , 0.01; ***P , 0.001.


at ASPE



ownloaded from


combinationwere evaluatedbyposthocDunnett’s testwithP,0.05beingassigned as statistically significant. For studies evaluating drug synergy, asynergy analysis was conducted using the method of Bliss Independence(Grecoetal., 1995;Fitzgeraldetal., 2006); using theendpoint of percentageinhibition, an ANOVAwas applied to test the coefficient of the interactionterm in a 2 � 2 full factorial model of the two compounds.

Plasma and Brain Analysis under Drug Combination. Threemice from each group were sacrificed immediately postexperiment andplasma and brain were collected on ice and dry ice, respectively. Plasmaand brain levels of each drug alone were analyzed. Brain samples wereweighed and a 3-fold volume of water/methanol (4:1, v/v) was added priorto homogenization with an ultrasonic tissue disrupter. Control (naive)brain tissue was also homogenized to generate control homogenate forpreparation of calibration standards. Stock solutions containing 1 mg/mlof each of the analytes were diluted to produce working solutions, whichwere then used to fortify control plasma or control brain homogenate toproduce calibration standards with concentrations ranging from 1 to5000 ng/ml. Aliquots of each study sample, calibration standard, andcontrol samples were then transferred to 96-well plates, mixed withacetonitrile and internal standard to precipitate sample proteins,centrifuged to pellet insoluble material, and the supernatant evaporated.Extracts were reconstituted in 0.5% heptafluorobutyric acid in water.Study sampleswereanalyzedbyLC-MS/MSusingaSciexAPI 4000 triplequadrupole mass spectrometer (Applied Biosystems/MDS, Foster City,CA) equipped with a TurboIonSpray interface, and operated in positiveionmode.Citalopramand fluoxetinewere chromatographically separatedusing a Betasil C18 Dash, 5 mm 20 � 2.1 mm (Thermo Fisher,Waltham, MA). LY3020371 was chromatographically separated using a

2.1 mm� 50 mm, Atlantis T3 column (Waters, Milford, MA). The pumpswere Shimadzu LC-10AD units with a SCL-10A controller, and a Gilson(Middleton, WI) 215 liquid or a Leap CTC liquid handler (LEAPTechnologies, Carrboro, NC) was used as the autosampler. Water/1 MNH4HCO3 (2000:10, v/v) (mobile phase A), and MeOH/1 M NH4HCO3

(2000:10, v/v) (mobile phaseB)wasused for citalopramand fluoxetine and0.2% formic acid in water (mobile phase A), and MeOH/Acetic Acid (3:1,v/v) was used for LY3020371. The selected reactionmonitoring (M1H)1transition m/z values were 360.1 . 159.1 (LY3020371), 325.2 . 262.2(citalopram), and 310.1 . 148.1 (fluoxetine).

Prodrug Pharmacokinetics. Male CD-1 mice, male NationalInstitutes of Health Swiss mice, and male Sprague-Dawley rats wereemployed (all from Harlan Industries, Indianapolis, IN). Animals had freeaccess to food and water at all times, except overnight prior to dosing and4hours after dosing. LY3027788.HCl (Smith et al., 2012)was administeredorally (by mouth) as a suspension in hydroxyl-ethyl cellulose (1%), Tween80 (0.25%), and Dow Antifoam (0.05%) (Midland, MI). LY3020371.HCldissolved in saline was employed for studies involving intravenousadministration. For studies requiringassessment of pharmacokinetics overan extended period of time, blood was collected at 0.08, 0.25, 0.5, 1, 2, 4, 8,12, and 24 hours postadministration (i.v.) and at 0, 0.25, 0.5, 1, 2, 4, 8, 12,and 24 hours postadministration (oral); in the mouse brain/plasmaexperiment, a single time point (1 hour) was used. Samples werecentrifuged and the plasma was stored at 270°C until analyzed. CSFwas collected from the cisterna magna by syringe with up to four samplesper animal and stored at270°C until analyzed.

Aliquots of thawed plasma were mixed with acetonitrile and ananalog internal standard solution. Samples were centrifuged at

Fig. 5. Increases in monoamine efflux in the medial prefrontal cortex of freely moving rats by LY3020371 as a function of time postinjection (10 mg/kg,i.p.). Each point represents mean 6 S.E.M. vehicle n = 6, LY3020371 n = 7. Open symbols: LY3020371; filled symbols: vehicle. For all analyses, (F1,11),increases in extracellular levels of the following neurochemicals were significantly different than vehicle control values: 5-hydroxytryptamine (5-HT)(F = 14.5; P , 0.01); 5-hydroxyindolacetic acid (5-HIAA) (F = 42.5; P , 0.0001); dopamine (DA) (F = 63.5; P , 0.0001); 3,4-dihydroxyphenylacetic acid(DOPAC) (F = 93.1; P , 0.0001); homovanilic acid (HVA) (F = 73.3; P , 0.0001); noradrenaline (NA) (F = 70.5; P , 0.0001); 3-methoxy-4-hydroxyphenylglycol (MHPG) (F = 85.3; P , 0.0001); 3,4-dihydroxyphenylglycine (DHPG) (F = 20.1; P , 0.001); acetylcholine (Ach) (F = 54.1; P , 0.0001);histamine (F = 17.2; P , 0.01); tele-methyl-histamine (TMH) (F = 20.8; P , 0.001); GABA (F = 67.9; P , 0.0001); and glutamate (F = 13.5; P , 0.01).

74 Witkin et al.

at ASPE



ownloaded from


3200 rpm for 10 minutes and the supernatant evaporated. Extractswere reconstituted in 0.5% heptafluorobutyric acid and analyzed byLC-MS/MS using two Shimadzu LC-20AD pumps, a LEAP Technolo-gies PAL autosampler, and a Sciex API 4000 (monkey and mousesamples) or API 5000 (rat samples) triple quadrupole mass spectrom-eter (Applied Biosystems/MDS) equipped with a TurboIonSpray in-terface, and operated in positive ionmode. Chromatographic separationwas accomplished on a 2.1 mm � 50 mm, Atlantis T3 column (Waters)using a binary mobile phase. The initial mobile phase system wascomposed of 0.2% formic acid in water (mobile phase A) and MeOH/Acetic Acid (3:1, v/v; mobile phase B). The selected reaction monitoring(M 1 H) transition was m/z 360.1 . 159.2 (API 4000) or 360.1 . 159.0(API 5000). Pharmacokinetic parameters were calculated by noncom-partmental analysis using Watson 7.4 (Thermo Fisher Scientific).

ResultsPharmacodynamic Effects of LY3020371 in Comparisonwith Ketamine

Several studies were conducted in rats to compare theeffects of LY3020371 and ketamine in different brain areas.

Dopamine Neuron Activity in the VTA of Rats

LY3020371 was characterized for its ability to modulatedopamine cell activity in the VTA and compared with the

effects of ketamine. Acute administration of LY3020371(0.3–3 mg/kg, i.v.) significantly increased the number ofspontaneously active dopamine cells in the VTA (Fig. 2A),an effect similar to ketamine (Fig. 2A) (Witkin et al.,2016a). LY3020371 did not significantly affect the firingrate (Fig. 2B) or the percentage of spikes occurring inbursts (Fig. 2C). The AMPA receptor antagonist, NBQX(10 mg/kg, i.v.), prevented the LY3020371-induced in-crease in the number of spontaneously active VTA dopa-mine cells (P , 0.05), without producing any effects on itsown (P . 0.05) (Fig. 2A).

Effects on Brain Oxygen

These experiments examined the changes in brain oxygenin two brain areas induced by LY3020371 or ketamine.LY3020371 (i.p.) produced dose-dependent increases in tissueoxygen in the ACC (Fig. 3) but not in the STR (Fig. 3). The doseof ketamine tested also robustly increased tissue oxygen levels inboth regions.

Effects on Neurochemical Efflux in the Rat Medial PrefrontalCortex

In vivo microdialysis studies were undertaken in ratsto investigate the sensitivity of neurotransmitter systems to

Fig. 6. Both ketamine and LY3020371 increase the time spent awake in freely moving rats. Each point represents mean6 S.E.M. data from 8 to 12 rats.(A) Increases in accumulated time awake over baseline (minutes) after S-(+)-ketamine (3, 10, and 30 mg/kg, i.p., designated as green, blue, and red lines,respectively). (B) Increases in accumulated time awake over baseline (minutes) after LY3020371 (1, 3, 10, and 30 mg/kg, i.v., designated as green, blue,and red lines, respectively). (C) AUC for accumulated wake over baseline for S-(+)-ketamine. (D) AUC for accumulated wake over baseline forLY3020371. (E) Changes in rapid eye movement (REM) and non-REM (NREM) sleep at 7–19 hours for S-(+)-ketamine. (F) Changes in REM and NREMsleep at 7–19 hours for LY3020371. *P , 0.05; **P , 0.01; ***P , 0.001.


at ASPE



ownloaded from


modulation by ketamine and to LY3020371 to explore neuro-chemical commonalities in theiractions. (1)-Ketamine (10mg/kg,s.c.) significantly increased the efflux of multiple neurochemicalswith time courses and overall response profile (AUC) for the3-hour postdosing as shown in Fig. 4.The time course for effects and the overall response

profile (AUC) for the 3-hour postdosing period forLY3020371 on efflux of brain neurochemicals are shown inFig. 5. LY3020371 (10 mg/kg, i.p.) produced a significantincrease in the medial prefrontal cortex compared withvehicle controls for the neurochemicals shown. Other ana-lytes, including glycine and glutamine, failed to changesignificantly after drug administration (data not shown).Measurement of LY3020371 in the same microdialysatesamples revealed an average concentration of approxi-mately 35 nM at 1 hour after administration (Fig. 5).Recovery across the probe was not calculated; therefore,the true extracellular concentration can only be approxi-mated. Based on a 10%–20% recovery (which may beexpected for this particular flow rate and molecule size),this value would be 10–20 times the dialysate concentration(i.e., 350–700 nM).

Wake Promotion

LY3020371 was characterized using the SCORE2004bioassay system (Lilly Research Labs, Windlesham, Surrey,

UK) that can detect wake-promoting effects via EEGelectrodes. Effects were compared with those observed withketamine. Of n 5 115 animal dosings, n 5 113 yieldedusable data. The average age at the time of dosing was148 days and the average weight was 517 g. Ketamine(10 mg/kg, s.c.) produced increases in wake promotion ofrats followed by a rebound hypersomnolence starting atabout hour 16 (Fig. 6). LY3020371 increased the cumulativewake time of rats in a dose- and time-dependent mannerwithout rebound hypersomnolence, with significant in-creases being observed at the lowest dose tested of 1 mg/kg,i.v. (Fig. 6).

Antidepressant-Like Effects in Serotonin UptakeInhibitor–Sensitive and –Insensitive Models: EffectsComparable to Those of Ketamine

Forced-Swim Assay. Both ketamine and LY3020371 pro-duced an antidepressant-like behavioral signature in rats,significantly decreasing immobility time, with a minimaleffective dose of 17 mg/kg for each compound when dosed i.p.(28.6% decrease for ketamine; 19.2% decrease for LY3020371).Imipramine, given as a positive control, was as efficacious at15 mg/kg, i.p. (21.7% decrease).When dosed intravenously, the potency of LY3020371 (but

not ketamine) increased, with doses as low as 0.3 mg/kg beingactive in the forced-swim assay in rats (Fig. 7). The plasma

Fig. 7. (Upper panels) Intravenous ketamine and LY3020371 decrease the time rats are immobile in the forced-swim test in the rat forced-swim assay.Ketamine: F(3, 28) = 14.82, P , 0.0001; LY3020371: F(5,38) = 12.38, P , 0.0001. The mGlu2/3 receptor agonist, LY354740 (3 mg/kg, i.p.), was given 60min prior to testing. The AMPA receptor antagonist NBQX (10 mg/kg, i.p., 60 minutes prior) attenuated the effects of LY3020371. For the drugcombination studies, ketamine was dosed 60 minutes prior and LY3020371 was given 120 minutes prior to testing. Each point represents the mean 6S.E.M. of 6–8 rats. V: vehicle; I: Imipramine (30 mg/kg, i.p.). Open symbols above the dose effect curves are molecule + agonist or antagonist. *P , 0.05compared with vehicle control values by post hoc Dunnett’s test. (Lower panels) Attenuation of the antidepressant-like effects of ketamine andLY3020371 by an inhibitor of mTOR, AZD8055 in the mouse forced-swim assay. (Left) The mTor inhibor AZD8055 was dosed at 10 mg/kg, by mouth,60 minutes prior to testing. Ketamine and LY3020371 were given i.p., 30 minutes prior to testing. Each bar represents the mean + S.E.M. results from7 to 8 mice. *P , 0.05 compared with vehicle (veh) (Dunnett’s test); $P , 0.05 compared with ketamine alone (t test). (Right) The antidepressant-likeeffects of LY3020371 (10 mg/kg, i.p.) in wild-type (WT) mice are absent in mGlu22/2 mice [knockout (KO)]. Each bar represents the mean + S.E.M. ofeight mice. *P , 0.05 compared with WT-veh (Dunnett’s test).

76 Witkin et al.

at ASPE



ownloaded from


concentration of ketamine at 10 mg/kg 5 200 ng/ml, which isequal to plasma exposures associated with clinical efficacy indepressed patients (Zarate et al., 2012)The dependence of the effects of i.v. ketamine or LY3020371

on mGlu2/3 receptors was evaluated in the presence of themGlu2/3 receptor agonist LY354740 (3 mg/kg, i.p.). Underthese conditions, themGlu2/3 receptor agonist fully preventedthe antidepressant-like effects of ketamine and (Fig. 7)LY354740 (3mg/kg, i.p.) did not significantly affect immobilityin rats when given alone (2726 16 seconds,P. 0.05 comparedwith vehicle control). To further evaluate the role of mGlu2/3receptors in the antidepressant-like effects of LY3020371, theeffects of LY3020371.HCl were tested in mGlu2 receptor 2/2mice. In this study, the antidepressant phenotype wasproduced in mGlu1 /1 mice but not in mGlu22 /2 mice (Fig.7, bottom panel).

Effects of ketamine in the forced-swim assay have pre-viously been shown to be dependent on AMPA receptors (c.f.,Maeng at al., 2008; Witkin et al., 2016a). In the present study,we now document a comparable dependence on AMPA recep-tors of the antidepressant-like effects of LY3020371. Underthese conditions, NBQX (10 mg/kg) was able to completelyprevent the antidepressant-like effects of LY3020371 withoutsignificant effects on its own (Fig. 7). Given alone, NBQXproduced immobility times of 275 616 seconds (P . 0.05compared with vehicle control).The risk of tolerance development to the antidepressant-

like effects of LY3020371 was evaluated by dosingLY30200371 or vehicle for four consecutive days, and thentesting on day 5 with LY3020371 (i.v.). In this study, 4 days ofprior treatment with LY3020371 did not significantly alter theanti-immobility effects of LY3020371 when tested on day 5.

Fig. 8. Time course of action of LY3020371 in rats under the forced-swim test after intravenous dosing. 1 mg/kg: F(5,32) = 4.836, P , 0.01; 3 mg/kg:F(4,27) = 4.334, P, 0.01; 10 mg/kg: F(5,32) = 10.42, P, 0.0001. Each point represents the mean 6 S.E.M. of 6–8 rats. *P, 0.05 compared with vehiclecontrol values. Each bar represents the mean 6 S.E.M. of six rats. *P , 0.05 compared with vehicle control values by post hoc Dunnett’s test.

Fig. 9. CSF concentrations predict efficacy in rat forced-swim assay (R2 =0.85, P, 0.05) (lower-right panel). The numbers adjacent to data points (lower-left panel) are the percent decrease in immobility in the forced-swim assay. The hashed line (lower-left panel) represents the IC50 value for LY3020371 inrat hippocampal slice preparation (IC50 = 46 nM) (Witkin et al., 2016a).


at ASPE



ownloaded from


The percent decrease in immobility at 1 mg/kg was 84.8% (4-day vehicle treated) versus 87.9% (4-day LY3020371 treated)and 78.3% (versus 4-day vehicle treated) versus 81.2% (4-dayLY3020371 treated).The potential impact of a priori exposure to other antide-

pressant medications similar to the selective serotonin uptakeinhibitors on the antidepressant-like effects of LY3020371was tested in rats. In this study, rats were dosed for14 consecutive days with either drug vehicle or citalopram(10 mg/kg) and then tested with LY3020371 two days later.Both 1mg/kg (84.8% versus 83.9% decrease for vehicle-treatedversus citalopram-treated rats, respectively) and 3 mg/kg

(78.3% versus 81.4% decrease for vehicle-treated versuscitalopram-treated rats, respectively) were active in the ratforced-swim assay after 14 days of prior exposure to citalo-pram, the same as was observed after 14-day dosing withvehicle. Time-course experiments were conducted in rats inthe forced-swim assay from 0.5 to 24-hour postdosing (i.v.) at1, 3, and 10 mg/kg. Both the maximal effect of LY3020371and the duration of action were increased as a function of dose(Fig. 8).Relationship of LY3020371 CSF Drug Levels to

Antidepressant-Like Effects. CSF concentrations ofLY3020371 were collected in a group of rats exposed to 1, 3,

Fig. 10. Analytes differentiating ketamine from LY3020371 Analytes are from hippocampus or CSF from rats treated with either ketamine orLY3020371 at 10 mg/kg, i.p.

78 Witkin et al.

at ASPE



ownloaded from


or 10 mg/kg, i.v. at times from 0.5 to 24 hours postdosing(Witkin et al., 2016a). These were previously shown toaccount for 3%–6% of plasma drug concentrations. CSFconcentrations of drug were a function of both dose andtime (Fig. 9, bottom left). With all doses tested, LY3020371(i.v.) produced CSF concentrations at some time pointsthat were above the IC50 value for functional antago-nism in rat hippocampal slices (Witkin et al., 2016a).Antidepressant-like effects (reduction in immobility timein the forced-swim assay) of LY3020371, i.v., were linearlyand positively associated with CSF drug levels (Fig. 9,bottom right).Metabolomic Analysis. Based on the dose-effect func-

tions defining potency and efficacy of ketamine andLY3020371 (i.p.), we selected doses of 10 mg/kg as pro-ducing effects most comparable to one another. Under theseconditions, ketamine significantly decreased immobilitytime, whereas LY3020371 just missed statistical signifi-cance. We then used these doses to compare the metabolo-mic profiles of ketamine and LY3020371. We detected morethan 240 polar metabolites, excluding peptides and intactlipids. Of these, several analytes displayed significantdifferentiation from those detected in vehicle-treatedanimals in both CSF and hippocampus (Fig. 10). In hippocampal

tissue, both ketamine and LY3020371 significantly alteredlevels of pyruvic acid and hydroxylamine relative to vehicle-treated rats; methylnicotinamide was positively alteredin both hippocampus and the CSF of rats treated withLY3020371 (Fig. 10).Utilizing pathway analysis of the analytes detected,

predicted metabolic changes in hippocampus were uncov-ered for the ketamine-treated (Fig. 11) and LY3020371-treated (Fig. 12) rats that both overlapped and weredistinct from one another. Specifically, predicted activa-tion or upregulation of the following analytes were detectedfor both ketamine and LY3020371: ADORA1, GRIA2,MAP2, prostaglandin D2, and phosphocreatine. Pathwayanalysis predicted inhibition or downregulation of thefollowing analytes for both ketamine and for LY3020371:kynurenic acid, glycine, natriuretic peptide A, SLC1A1,and EPO.Plasma Pharmacokinetics Following Oral Adminis-

tration of Prodrug LY3027788.HCl. Oral administrationof diester prodrug LY3027788.HCl (Fig. 1) led to the rapidand dose-proportionate appearance of the pharmacologi-cally active species LY3020371 in plasma of both mouse(Fig. 13A; Table 1) and rat (Fig. 13B; Table 2). No detectableplasma levels of either diester LY3027788 or individual

Fig. 11. Metabolomics analysis of CSF samples from rats after 10 mg/kg, i.p. ketamine.


at ASPE



ownloaded from


monoesters were observed. In a separate study (Fig. 14C),oral bioavailability of LY3020371 in rats was determined bycomparing plasma exposures of LY3020371 following ad-ministration of prodrug LY3027788.HCl (8.74 mg/kg, bymouth, a dose equivalent to 5 mg/kg of LY3020371) withthose achieved via i.v. administration of LY3020371.HCl(1.1 mg/kg, a dose equivalent to 1 mg/kg of LY3020371)(Table 3). From this study, the bioavailability of LY3027788was calculated to be 42.9% 6 12%. Plasma and CSF concentra-tions of LY3020371 following single oral doses of LY3027788.HCl are shown in Fig. 14D.CSF Levels of LY3020371 Following Oral Administration

of Prodrug LY3027788.HCl in Rat. Mean plasma andCSF levels of LY3020371 were determined following oraladministration of LY3027788.HCl (10 and 60 mg/kg). CSFlevels of LY3020371 were apparent at the first assessment(1 hour), peaked at 2 hours, and persisted through the full(24-hour) course of the experiment (Fig. 13D; Table 4). Meanconcentrations of LY3020371 in CSF were significantlylower than those observed in plasma at each assessed time

point [CSF/plasma ratios based on mean AUC were 5.4%(10 mg/kg) and 3.7% (60 mg/kg)], while CSF/plasma ratiosbased on Cmax were 1.4% (10 mg/kg) and 1.8% (60 mg/kg).The CSF/plasma ratio appeared to increase over time(from 0.7% to 230% at 10 mg/kg dose; from 1.3% to 13% at60 mg/kg dose), suggesting that clearance of LY3020371from the central compartment is slow compared withthat from plasma. These results are similar to previouslydisclosed findings for rat CSF/plasma pharmacokineticsfollowing i.v. dosing of LY3020371 (Witkin et al., 2016b) aswell as to those associated with mGlu2/3 receptor agonistspossessing similar physiochemical characteristics (Monnet al., 2015a,b).Brain Levels of LY3020371 Following Oral Adminis-

tration of Prodrug LY3027788.HCl in Mouse. Meanbrain and plasma levels of LY3020371 were determined inthe mouse 1 hour following oral dosing of LY3027788.HCl(5.6, 17.6, and 29.9 mg/kg) (Table 5). Similar to theCSF/plasma ratios found in the rat, the brain/plasma ratiosin the mouse were determined to be 3.7% and 2.3% following

Fig. 12. Metabolomics analysis of CSF samples from rats after 10 mg/kg, i.p. LY3020371.

80 Witkin et al.

at ASPE



ownloaded from


oral doses of 17.6 and 29.9 mg/kg, respectively. Brain levelsof LY3020371 in the mouse following an oral dose of5.6 mg/kg of LY3027788.HCl were below levels of detection(14 nM).Behavioral Effects of the Oral Prodrug LY3027788.

HCl. In the mouse forced-swim assay, LY3027788.HCl waspotent and efficacious with a minimal effective dose of16 mg/kg, by mouth (equivalent to 10 mg/kg active moietyLY3020371) (Fig. 14A). The ED60 value in this assay was8.2 mg/kg (equal to 4.7 mg/kg of active moiety LY3020371). In

the locomotor activity assay in mice, a single dose ofLY3027788.HCl (16 mg/kg, bymouth) enhanced the locomotorstimulant effects of quinpirole (Fig. 14B). Furthermore, aswith LY3020371, oral administration of the prodrug form,LY3027788.HCl, dose dependently increased the wake time ofrats without engendering rebound hypersomnolence (Fig.14C). The amounts of time spent awake comparedwith vehiclecontrol that were induced by single oral doses of LY3027788.HCl (10, 20, and 30 mg/kg) were 40 6 16, 44 6 13, and 106 68 minutes, respectively).

Fig. 13. Pharmacokinetics of LY3020371 following oral dosing of diester prodrug LY3027788.HCl in mouse and rat. (A). Plasma exposures ofLY3020371 in themouse following single oral doses of LY3027788.HCl (j 4.8mg/kg,u 16mg/kg,d 27mg/kg). (B) Plasma exposures of LY3020371 in therat following single oral doses of LY3027788.HCl (j 3 mg/kg,u 10 mg/kg,d 30 mg/kg). (C) Plasma exposures of LY3020371 following either a single 1.1.mg/kg i.v. dose of LY3020371 (j) or 8.74 mg/kg oral dose of LY3027788.HCl (u). (D) Plasma and CSF concentrations of LY3020371 following single oraldoses of LY3027788.HCl. m 10 mg/kg, Plasma; D 60 mg/kg, Plasma; j 10 mg/kg, CSF, u 60 mg/kg, CSF). Data points are means 6 S.D. of 4–12animals/time point.

TABLE 1Mean plasma pharmacokinetics of LY3020371 following single oral doses of diester prodrug LY3027788.HCl in male CD-1mice

LY3027788.HCl Dose (by Mouth) LY3020371 Equivalent Dose (by Mouth)LY3020371 Mean Plasma Pharmacokinetic Parameter

Cmax Tmax AUC0–24 T1/2

mg/kg mg/kg nM hour nM/h hour

4.8 2.7 1990 0.25 2830 1.616 9.1 10,800 0.25 10,900 1.327 15.4 15,500 0.25 17,100 1.8


at ASPE



ownloaded from


Drug Combination Studies. Individual doses of fluoxe-tine, citalopram, and LY3027788.HCl were selected based ontheir lack of effect when given alone in the forced-swim assayin mice. Fluoxetine (10 mg/kg, i.p., 30 minutes prior), citalo-pram (0.3 mg/kg, i.p., 30 minutes prior), and LY3027788.HCl(10 mg/kg, by mouth, 60 minutes prior) were also withouteffect when given alone in the present study (Fig. 15, toppanels). However, when given in combination withLY3027788.HCl, citalopram produced antidepressant-likeeffects in this assay that were greater than those producedby either drug alone (Fig. 15). In contrast, the effects of thefluoxetine/LY3027788 combination was additive, having just

missed statistical significance for synergy. Under theseconditions, giving both compounds in combination did notalter the plasma or brain levels of either drug alone (P, 0.05).Concentrations of the active/parent moiety LY3020371 afteroral dosing of LY3027788.HCl with or without fluoxetine were2992.8 6 347.2 (alone) and 3963.7 6 1036.6 nM (drugcombination) for plasma and 63.6 6 6.4 (alone) and 70.4 67.0 (drug combination) for brain. Concentrations of the active/parent moiety LY3020371 after oral dosing of LY3027788.HClwith or without citalopram were 3294.2 6 418.6 (alone) and3043.46 558.6 nM (drug combination) for plasma and 173.1696.9 (alone) and 188.16 45.4 nM (drug combination) for brain.

TABLE 2Mean plasma pharmacokinetics of LY3020371 following single oral doses of diester prodrug LY3027788.HCl in male Sprague-Dawley rats

LY3027788.HCl dose (by Mouth) LY3020371 Equivalent Dose (by Mouth)Mean Rat Plasma Pharmacokinetic Parameter

Cmax Tmax AUC0–24 T1/2

mg/kg mg/kg nM hour nM/h hour

3 1.7 1540 0.5 4420 1.510 5.7 5590 0.8 14,700 2.930 17.2 9160 1.0 30,900 5.4

Fig. 14. Some behavioral effects of an oral prodrug form of LY302071, LY3027788.HCl when given orally to rodents. (A) Antidepressant-like efficacy in themouse forced-swim assay. Each bar represents themean6 S.E.M. of 7 to 8mice. F(3,27) = 13.21,P, 0.0001. *P, 0.05 comparedwith vehicle control valuesby post hoc Dunnett’s test. (B) Enhancement of the locomotor stimulant effects of quinpirole in mice after oral administration of LY3027788.HCl. F(2,54) =21.31, P , 0.0001 (dose); F(1,54) = 18.24, P , 0.0001 (treatment, quinpirole + or 2); F(2,54) = 10.60, P , 0.0001 (dose � treatment interaction). Each barrepresents themean6S.E.M. of 9 to 10mice. *P, 0.05 comparedwith vehicle control values, # comparedwith quinipirole alone, $ comparedwith same dosewithout quinpirole. (C) Wake-promoting effects of LY3027788.HCl in freely moving rats. Each point represents the mean 6 SEM of 8–12 rats.

82 Witkin et al.

at ASPE



ownloaded from


The concentrations of fluoxetine were 2675.9 6 220.8 (alone)and 2394.5 6 1170.6 nM (drug combination) for plasma and37659 6 5579 (alone) and 33718 6 14203 nM (drug combina-tion) for brain. The concentrations of citalopram were 22.6 610.1 (alone) and 29.3 6 3.8 nM (drug combination) for plasmaand 499.3 6 236.6 (alone) and 704.5 6 132.0 nM (drugcombination) for brain.

DiscussionA new orthosteric antagonist of mGlu2/3 receptors,

LY3020371, was recently reported (Smith et al., 2012) andcharacterized as a potent and selective orthosteric antago-nist in vitro (Witkin et al., 2016b). The pharmacology ofLY3020371 from human clonal systems to native tissuefunctional preparations along with its pharmacokinetic prop-erties suggested that LY3020371 was a well-behaved orthos-teric mGlu2/3 receptor antagonist that would block mGlu2/3receptors in vivo upon systemic dosing (Witkin et al., 2016b).In the present paper we documented the in vivo effects ofLY3020371and its diester prodrug, LY3027788 (Smith et al.,2012) that predict efficacy in treatment-resistant depression.The basis for this therapeutic prediction comes from theconvergent biologic effects of LY3020371 and that of ketamine,an NMDA receptor antagonist that produces rapid andpersistent relief of depression symptoms in treatment-resistant depression patients (Zarate et al., 2006; Abdallahet al., 2015). Prior studies have also suggested the efficacy ofmGlu2/3 receptor antagonists in depression (Chaki et al.,2004; Witkin and Eiler, 2006; Witkin et al., 2016a), and twomGlu2/3 antagonists have entered clinical development: BCI-838 (Yasuhara et al., 2006; Nakamura et al., 2006), an oralprodrug of orthosteric antagonist BCI-632, also known asMGS0039 (Nakazato et al., 2004; Yoshimizu et al., 2006) andRO4995819, also known as RG1538 (Decoglurant), anmGlu2/3 negative allosteric modulator (Gatti et al., 2006;clinicaltrials.gov, study NCT01457677).

In the present paper, we first documented the commonpharmacodynamic effects of ketamine and LY3020371.Both molecules increased the probability of dopamine cellfiring in the VTA (see also Witkin et al., 2016a), a brain areaknown to regulate hedonic valuation and regulate a neuro-transmitter system well characterized for its control ofmood (Price and Drevets, 2010). This enhancement ofdopamine neurotransmission was also observed at theneurochemical level where extracellular levels of dopaminewere markedly enhanced in the medial prefrontal cortex ofrats with both LY3020371 and ketamine. In behavioralstudies, LY3027788.HCl provided immediate enhancementof the functional effects of dopamine as measured in theaugmentation of the locomotor stimulant effects of quinpir-ole, an effect that also occurs with ketamine but notcitalopram (Witkin et al., 2016a). LY3020371 also increasedoxygen levels in the ACC, an effect also observed withketamine. The ACC is a brain area controlling mood and isknown to be modulated by ketamine in patients (Lally et al.,2015). In addition to dopamine, multiple and overlappingneurotransmitters were enhanced by both LY3020371 andketamine in the medial prefrontal cortex of behaving ratsincluding the wake- and cognition-regulating monoamines,histamine, and acetylcholine.LY3020371 was also wake promoting in rats, an effect

reported earlier for the mGlu2/3 receptor antagonist (2S)-2-amino-2-[(1S,2S)-2-carboxycycloprop-1-yl]-3-(xanth-9-yl)propanoic acid (Feinberg et al., 2005). Although ketamine alsoengenders wake promotion, the striking feature of the effectsof blockade of mGlu2/3 receptors is that there is no reboundnon-rapid eye movement hypersomnolence.In rodent models that detect conventional antidepressant

drugs, such as the forced-swim assay (Cryan et al., 2002), wedemonstrated potent and efficacious antidepressant-like ef-fects of LY3020371 similar to that of ketamine in both ratsand mice. These behavioral effects of LY3020371 were pre-vented by an mGlu2/3 receptor agonist and were absent in

TABLE 3Bioavailability of LY3020371 following oral dosing of prodrug LY3027788.HClMean plasma levels of LY3020371 were determined following a single i.v. dose of LY3020371.HCl and single oral dose of diester prodrug LY3027788.HCl in male Sprague-Dawley rats.

Compound Administered (Dose, Route)LY3020371 Equivalent Dose

Mean LY3020371 Rat Plasma Pharmacokinetic Value

LY3020371.HCl (i.v.) LY3027788.HCl (p.o.) Cmax Tmax AUC0–24 T1/2 Vdss CL F

mg/kg mg/kg mg/kg nM hour nM/h hour l/kg ml/min/kg %

1.1 — 1 6867 N/A 5930 0.77 0.458 8.7 —

— 8.74 5 4130 0.5 12,500 1.5 N/A N/A 42.9

—, not applicable.CL, clearance; N/A, not applicable; Vdss, volume of distribution at steady state.

TABLE 4Mean plasma and CSF concentrations of LY3020371 following single oral doses of diester prodrug LY3027788.HCl in Sprague-Dawley rats

LY3027788.HCl dose (by Mouth) LY3020371 Equivalent Dose (by Mouth)

Mean LY3020371 Rat Plasma and CSF Pharmacokinetic Value

Plasma CSF

Cmax Tmax AUC0–24 T1/2 Cmax Tmax AUC0–24 T1/2

mg/kg mg/kg nM hour nM/h hour nM hour nM/h hour

10 5.7 7980 1.0 17,400 3.0 109 2.0 942 5.260 34.3 13,700 1.0 46,500 3.6 240 2.0 1730 4.8


at ASPE



ownloaded from


mGlu22/2 mice, attesting to the on-target actions of thismolecule. We also demonstrated that the antidepressant-related signature of LY3020371 was positively associatedwith drug levels in the CSF of rats and that these drug levelswere above the IC50 value for inhibition reported in rat braintissue preparations (∼40 nM) (Witkin et al., 2016a).One of the compelling features of ketamine as an antide-

pressant is its sustained effects (Zarate et al., 2006). Thequestion arises in the current context as to whether mGlu2/3receptor antagonists also produce effects that outlast theirkinetics. Our laboratory has not been able to achieve effectseven 24 hours postdosing with either ketamine or themGlu2/3 receptor antagonist (2S)-2-amino-2-[(1S,2S)-2-carboxycycloprop-1-yl]-3-(xanth-9-yl) propanoic acid. Othergroups have observed longer lasting effects of ketamine and

mGlu2/3 receptor antagonists (see Pilc et al., 2013; Donget al., 2016). However, effects that last beyond a day have notbeen universally observed (Popik et al., 2008; Pilc et al.,2013). We did not evaluate the effects of LY3020371 forthis effect. Recently, a metabolite of ketamine was report-ed to be active and to engender sustained activity (Zanoset al., 2016).As with ketamine (Karasawa et al., 2005; Maeng et al.,

2008; Fukumoto et al., 2016), we showed that theantidepressant-like efficacy of LY3020371 was prevented byblockade of AMPA receptors. AMPA receptor dependence hasbeen reported for other mGlu2/3 receptor antagonists(Gleason et al., 2013; Fukumoto et al., 2016; Witkin et al.,2016a) and has been suggested to be (Alt et al., 2006) andremains (Abdallah et al., 2015) a leading hypothesizedtransducer of antidepressant efficacy. We independentlyconfirmed this idea with the metabolomics profiling of ket-amine and LY3020371. Doses of these molecules (10 mg/kg,i.p.) that were close to or at the dose level producingantidepressant-like effects in rats produced a series ofhippocampal analyte changes that predicted enhancedAMPA receptor signaling. These are the first data to compareketamine and mGlu2/3 receptor antagonism at the metab-olomic level and provide additional evidence supporting acritical role for AMPA receptor signaling in the antidepres-sant responses produced by these agents in rodents as well asan additional reason to believe that mGlu2/3 antagonistsmight produce ketamine-like antidepressant efficacy inpatients.Another protein predicted to be amplified by both ket-

amine and by LY3020371 from pathway analysis of themetabolomics data was ADORA1 (Adenosine A1). AdenosineA1 receptors have previously been shown to have a likely rolein driving antidepressant efficacy (see Ortiz et al., 2015;Serchov et al., 2015). Therefore, the present findings suggestthat A1 receptor activation might be another importantdownstream pathway in the actions of rapidly acting anti-depressants. Of interest in this regard is a recent reportthat the potent noncompetitive NMDA receptor antagonist(1)-MK-801 might induce antidepressant-like effects inzebra fish through A1 receptors (da Silva et al., 2015). It isnoteworthy that the metabolomics analysis did not uncovermTOR as a potential target for either ketamine or LY3020371despite reports of mTOR involvement (Li et al., 2010; Dwyeret al., 2012) and the blockade of effects of ketamine andLY3020371 reported here (Fig. 7). A recent paper has outlineddifficulties in recapitulating the data on mTOR expression byketamine (Popp et al., 2016).LY3020371 is a highly polar amino diacid with restricted

intestinal absorption, resulting in poor oral bioavailability inrats (F , 10%, data not shown). This limitation has been

TABLE 5Plasma and brain concentrations of LY3020371 following single oral doses of diester prodrug LY3027788.HCl in NationalInstitutes of Health Swiss mice

LY3027788.HCl Dose (by Mouth) LY3020371 Equivalent Dose (by Mouth)Mean LY3020371 Concentration at 1 hour

Plasma Brain

mg/kg mg/kg17.6 10 1365 50.929.9 17 3316 76.3

Fig. 15. (Top panels) LY3027788.HCl enhances the antidepressant-likeeffects of fluoxetine and of citalopram without altering plasma or brainlevels of either drug alone. Each bar represents the mean 6 S.E.M. of 6–8mice. Veh: vehicle; Fluox: Fluoxetine (10 mg/kg, i.p. 30 minutes); Cital:citalopram (0.3 mg/kg, i.p., 30 minutes); LY: LY3027788.HCl (10 mg/kg, bymouth, 60 minutes prior). *P , 0.05 compared with vehicle control valuesby post hoc Dunnett’s test. By synergy analysis, only the drug combinationof LY3027788 and citalopram was significant (see Materials and Methodsfor statistical methods).

84 Witkin et al.

at ASPE



ownloaded from


overcome with LY3027788.HCl, a diester form of LY3020371(Smith et al., 2012) that demonstrates rapid absorption andbioconversion to LY3020371 following oral dosing in rodents.Oral doses at or above 10 mg/kg LY3027788.HCl in ratsgenerated CSF levels of LY3020371 that either reached orexceeded the measured antagonist potency for this compoundin the rat hippocampal slice (IC505 46 nM) (Witkin et al., 2016b)and doses at or above 10 mg/kg enhanced wakefulness in rats.Similarly, in mice, functional mGlu2/3 antagonist-associatedbrain levels of LY3020371 (i.e., concentrations at or above46 nM) measured 1 hour after single oral doses of LY3027788.HCl were attained following oral doses at or above 17.6 mg/kg,while statistically significant responses in both themouse forced-swim and quinpirole locomotor sensitization assays wereachievedwith oral doses of LY3027788.HCl at or above 16mg/kg.These results relating central drug concentrations with behav-ioral responses are consistent with those previously noted forLY3020371 when administered by the i.v. route.In summary, we have shown here that LY3020371 and oral

prodrug LY3027788.HCl engender a host of antidepressant-relevant biologic effects that are in common with the knownefficacious drug for treatment-resistant depression patients, ket-amine, and that these effects can be predicted by drug concen-trations measured in the central compartment. Multipleadditional commonalties in biologic activities have been foundbetween ketamine and other mGlu2/3 receptor antagonists (e.g.,Dwyer et al., 2012, 2013; Dong et al., 2016; Fukumoto et al., 2016Witkin et al., 2016a), providing an overall consistent profile acrosslaboratories, biologic readouts, and molecules. Finally, whileantidepressant efficacy for LY3020371 and its oral prodrug havebeen demonstrated, evidence that mGlu2/3 receptor antagonismmight overcome the side effects that characterize ketamine hasnot been addressed. A direct comparison of potentially limitingside effects engendered by these two mechanisms will be thesubject of a future disclosure to be communicated in due course.

Authorship Contributions

Participated in research design: Witkin, Mitchell, Wafford, Gilmour,J. Li, Eastwood, X. Li, Rorick-Kehn, Rasmussen, Nikiolayev,Tolsikov, Swanson.

Conducted experiments: Witkin, Carter, J. Li, Overshiner, X. Li,Rorick-Kehn, Nikolayev, Tolskivov, Catlow.

Contributed new reagents or analytic tools: Kuo, R. Li, Smith,Mitch, Ornstein, Monn.

Performed data analysis: Witkin, Mitchell, Wafford, Carter,Gilmour, J. Li, Eastwood. Overshiner, X. Li, Rorick-Kehn, Anderson,Nikolayev, Tolstikov, Catlow, Swanson.

Wrote or contributed to the writing of the manuscript: Witkin,Mitchell, Wafford, Carter, Gilmour, J. Li, Eastwood, Overshiner, X.Li, Rorick-Kehn, Rasmussen, Anderson, Nikiolayev, Tolsikov, Kuo,Catlow, R. Li, Smith, Mitch, Ornstein, Swanson, Monn.

References

Abdallah CG, Sanacora G, Duman RS, and Krystal JH (2015) Ketamine and rapid-acting antidepressants: a window into a new neurobiology for mood disordertherapeutics. Annu Rev Med 66:509–523.

Adams DJ, Quail MA, Cox T, van der Weyden L, Gorick BD, Su Q, Chan WI, DaviesR, Bonfield JK, Law F, et al. (2005) A genome-wide, end-sequenced 129Sv BAClibrary resource for targeting vector construction. Genomics 86:753–758.

Ago Y, Araki R, Yano K, Kawasaki T, Chaki S, Nakazato A, Onoe H, Hashimoto H,Baba A, Takuma K, et al. (2012) The selective metabotropic glutamate 2/3 receptoragonist MGS0028 reverses isolation rearing-induced abnormal behaviors in mice.J Pharmacol Sci 118:295–298.

Alt A, Nisenbaum ES, Bleakman D, and Witkin JM (2006) A role for AMPA receptorsin mood disorders. Biochem Pharmacol 71:1273–1288.

Caraci F, Molinaro G, Battaglia G, Giuffrida ML, Riozzi B, Traficante A, Bruno V,Cannella M, Merlo S, Wang X, et al. (2011) Targeting group II metabotropic glu-tamate (mGlu) receptors for the treatment of psychosis associated with Alzheimer’s

disease: selective activation of mGlu2 receptors amplifies b-amyloid toxicity incultured neurons, whereas dual activation of mGlu2 and mGlu3 receptors is neu-roprotective. Mol Pharmacol 79:618–626.

Cartmell J, Monn JA, and Schoepp DD (1999) The metabotropic glutamate 2/3 re-ceptor agonists LY354740 and LY379268 selectively attenuate phencyclidine ver-sus d-amphetamine motor behavior in rats. J Pharmacol Exp Ther 291:161–170.

Cartmell J, Monn JA, and Schoepp DD (2000) The mGlu2/3 receptor agonistLY379268 selectively blocks amphetamine ambulations and rearing. Eur J Phar-macol 400:221–224.

Chaki S, Yoshikawa R, Hirota S, Shimazaki T, Maeda M, Kawashima N, YoshimizuT, Yasuhara A, Sakagami K, Okuyama S, et al. (2004) MGS0039: a potent andselective group II metabotropic glutamate receptor antagonist with antidepressant-like activity. Neuropharmacology 46:457–467.

Chappell MD, Li R, Smith SC, Dressman BA, Tromiczak EG, Tripp AE, Blanco MJ,Vetman T, Quimby SJ, Matt J, et al. (2016) Discovery of (1S,2R,3S,4S,5R,6R)-2-Amino-3-[(3,4-difluorophenyl)sulfanylmethyl]-4-hydroxy-bicyclo[3.1.0]hexane-2,6-dicarboxylic Acid Hydrochloride (LY3020371·HCl): A Potent, MetabotropicGlutamate 2/3 Receptor Antagonist with Antidepressant-Like Activity. J MedChem 59:10974–10993.

Cryan JF, Markou A, and Lucki I (2002) Assessing antidepressant activity in rodents:recent developments and future needs. Trends Pharmacol Sci 23:238–245.

D’Aquila PS, Collu M, Gessa GL, and Serra G (2000) The role of dopamine in themechanism of action of antidepressant drugs. Eur J Pharmacol 405:365–373.

da Silva RB, Siebel AM, and Bonan CD (2015) The role of purinergic and dopami-nergic systems on MK-801-induced antidepressant effects in zebrafish. PharmacolBiochem Behav 139:149–157.

Dong C, Zhang JC, Yao W, Ren Q, Ma M, Yang C, Chaki S, and Hashimoto K (2016)Rapid and sustained antidepressant action of the mGlu2/3 receptor antagonist MGS0039in the social defeat stress model: comparison with ketamine. Int J NeuropsychopharmacolDOI:https://doi.org/10.1093/ijnp/pyw089 [published ahead of print].

Du J, Zhou S, and Carlton SM (2008) Group II metabotropic glutamate receptoractivation attenuates peripheral sensitization in inflammatory states. Neurosci-ence 154:754–766.

Dunayevich E, Erickson J, Levine L, Landbloom R, Schoepp DD, and Tollefson GD(2008) Efficacy and tolerability of an mGlu2/3 agonist in the treatment of gener-alized anxiety disorder. Neuropsychopharmacol 33:1603e-1610.

Dwyer JM, Lepack AE, and Duman RS (2012) mTOR activation is required for the an-tidepressant effects of mGluR2/3 blockade. Int J Neuropsychopharmacol 15:429–434.

Dwyer JM, Lepack AE, and Duman RS (2013) mGluR2/3 blockade produces rapidand long-lasting reversal of anhedonia caused by chronic stress exposure. J MolPsychiatry 1:15.

Feinberg I, Schoepp DD, Hsieh KC, Darchia N, and Campbell IG (2005) Themetabotropic glutamate (mGLU)2/3 receptor antagonist LY341495 [2S-2-amino-2-(1S,2S-2-carboxycyclopropyl-1-yl)-3-(xanth-9-yl)propanoic acid] stimulates wakingand fast electroencephalogram power and blocks the effects of the mGLU2/3 re-ceptor agonist LY379268 [(2)-2-oxa-4-aminobicyclo[3.1.0]hexane-4,6-dicarboxylate]in rats. J Pharmacol Exp Ther 312:826–833.

Fitzgerald JB, Schoeberl B, Nielsen UB, and Sorger PK (2006) Systems biologyand combination therapy in the quest for clinical efficacy. Nat Chem Bio 2:458–466.

Fukumoto K, Iijima M, and Chaki S (2016) The antidepressant effects of an mGlu2/3receptor antagonist and ketamine require AMPA receptor stimulation in themPFC and subsequent activation of the 5-HT neurons in the DRN. Neuro-psychopharmacology 41:1046–1056.

Gatti S, Goetschi E, Palmer WS, Wichmann J, and Woltering TJ (2006) investors,Hoffman-La Roche AG, assignee. Preparation of acetylenyl-pyrazolo-pyrimidinederivatives for use a mGluR2 antagonists treating CNS disorders. PCT Int. Appl.WO2006099972:A1. 2006 Sep 28.

Gleason SD, Li X, Smith IA, Ephlin JD, Wang XS, Heinz BA, Carter JH, Baez M,Yu J, Bender DM, et al. (2013) mGlu2/3 agonist-induced hyperthermia in mice:An in vivo assay for detection of mGlu2/3 receptor antagonism and its relationto antidepressant-like efficacy in mice. CNS Neurol Disord Drug Targets 12:554–566.

Greco WR, Bravo G, and Parsons JC (1995) The search for synergy: a critical reviewfrom a response surface perspective. Pharmacol Rev 47:331–385.

Helton DR, Tizzano JP, Monn JA, Schoepp DD, and Kallman MJ (1998) Anxiolyticand side-effect profile of LY354740: a potent, highly selective, orally active agonistfor group II metabotropic glutamate receptors. J Pharmacol Exp Ther 284:651–660.

Jones CA, Brown AM, Auer DP, and Fone KCF (2011) The mGluR2/3 agonistLY379268 reverses post-weaning social isolation-induced recognition memorydeficits in the rat. Psychopharmacology (Berl) 214:269–283.

Jones CK, Eberle EL, Peters SC, Monn JA, and Shannon HE (2005) Analgesic effectsof the selective group II (mGlu2/3) metabotropic glutamate receptor agonistsLY379268 and LY389795 in persistent and inflammatory pain models after acuteand repeated dosing. Neuropharmacology 49 (Suppl 1):206–218.

Karasawa J, Shimazaki T, Kawashima N, and Chaki S (2005) AMPA receptorstimulation mediates the antidepressant-like effect of a group II metabotropicglutamate receptor antagonist. Brain Res 1042:92–98.

Kinon BJ, Millen BA, Zhang L, and McKinzie DL (2015) Exploratory analysis for atargeted population responsive to the metabotropic glutamate 2/3 agonist poma-glumetad methionil in schizophrenia. Biol Psych 78:754–62.

Kłodzi�nska A, Chojnacka-Wójcik E, Pałucha A, Bra�nski P, Popik P, and Pilc A (1999)Potential anti-anxiety, anti-addictive effects of LY 354740, a selective group IIglutamate metabotropic receptors agonist in animal models. Neuropharmacology38:1831–1839.

Kumar N, Laferriere A, Yu JSC, Poon T, and Coderre TJ (2010) Metabotropic glu-tamate receptors (mGluRs) regulate noxious stimulus-induced glutamate releasein the spinal cord dorsal horn of rats with neuropathic and inflammatory pain.J Neurochem 114:281–290.


at ASPE



ownloaded from


Lally N, Nugent AC, Luckenbaugh DA, Niciu MJ, Roiser JP, and Zarate CA, Jr (2015)Neural correlates of change in major depressive disorder anhedonia following open-label ketamine. J Psychopharmacol 29:596–607.

Li N, Lee B, Liu RJ, Banasr M, Dwyer JM, Iwata M, Li XY, Aghajanian G,and Duman RS (2010) mTOR-dependent synapse formation underlies the rapidantidepressant effects of NMDA antagonists. Science 329:959–964.

Linden AM, Shannon H, Baez M, Yu JL, Koester A, and Schoepp DD (2005)Anxiolytic-like activity of the mGLU2/3 receptor agonist LY354740 in the elevatedplus maze test is disrupted in metabotropic glutamate receptor 2 and 3 knock-outmice. Psychopharmacology (Berl) 179:284–291.

Maeng S, Zarate CA, Jr, Du J, Schloesser RJ, McCammon J, Chen G, and Manji HK(2008) Cellular mechanisms underlying the antidepressant effects of ketamine:role of a-amino-3-hydroxy-5-methylisoxazole-4-propionic acid receptors. Biol Psy-chiatry 63:349–352.

Marsteller DA, Gerald CPG, Kong R, Cajina M, Craig DA, and Swanson CJ (2009)The MCH1 receptor antagonist SNAP 94847 induces sensitivity to dopamine D2/D3receptor agonists in rats and mice. Eur J Pharmacol 602:66–72.

Moghaddam B and Adams BW (1998) Reversal of phencyclidine effects by a group IImetabotropic glutamate receptor agonist in rats. Science 281:1349–1352.

Monn JA, Prieto L, Taboada L, Hao J, Reinhard MR, Henry SS, Beadle C, Walton L,Man T, Rudyk H, et al. (2015a) Synthesis and pharmacological characterization ofC4-(thiotriazolyl)-substituted-2-aminobicyclo[3.1.0]hexane-2,6-dicarboxylates.In vitro and in vivo evidence of a highly potent, functionally selective mGlu2 re-ceptor agonist. J Med Chem 58:7526–7548.

Monn JA, Prieto L, Taboada L, Pedregal C, Hao J, Reinhard MR, Henry SS, Goldsmith PJ,Beadle CD, Walton L, et al. (2015b) Synthesis and pharmacological characterization ofC4-disubstituted analogs of 1S,2S,5R,6S-2-aminobicyclo[3.1.0]hexane-2,6-dicarboxylate:identification of a potent, selective metabotropic glutamate receptor agonist and de-termination of agonist-bound human mGlu2 and mGlu3 amino terminal domainstructures. J Med Chem 58:1776–1794.

Monn JA, Valli MJ, Massey SM, Wright RA, Salhoff CR, Johnson BG, Howe T, AltCA, Rhodes GA, Robey RL, et al. (1997) Design, synthesis, and pharmacologicalcharacterization of (1)-2-aminobicyclo[3.1.0]hexane-2,6-dicarboxylic acid (LY354740):a potent, selective, and orally active group 2 metabotropic glutamate receptor agonistpossessing anticonvulsant and anxiolytic properties. J Med Chem 40:528–537.

Nakamura M, Kawakita Y, Yasuhara A, Fukasawa Y, Yoshida K, Sakagami K,and Nakazato A (2006) In vitro and in vivo evaluation of the metabolism andbioavailability of ester prodrugs of MGS0039 (3-(3,4-dichlorobenzyloxy)-2-amino-6-fluorobicyclo[3.1.0]hexane-2,6-dicarboxylic Acid), a potent metabotropic glutamatereceptor antagonist. Drug Metab Dispos 34:369–374.

Nakazato A, Kumagai T, Sakagami K, Yoshikawa R, Suzuki Y, Chaki S, Ito H,Taguchi T, Nakanishi S, and Okuyama S (2000) Synthesis, SARs, and pharmaco-logical characterization of 2-amino-3 or 6-fluorobicyclo[3.1.0]hexane-2,6-dicarboxylicacid derivatives as potent, selective, and orally active group II metabotropic gluta-mate receptor agonists. J Med Chem 43:4893–4909.

Nakazato A, Sakagami K, Yasuhara A, Ohta H, Yoshikawa R, Itoh M, Nakamura M,and Chaki S (2004) Synthesis, in vitro pharmacology, structure-activity relation-ships, and pharmacokinetics of 3-alkoxy-2-amino-6-fluorobicyclo[3.1.0]hexane-2,6-dicarboxylic acid derivatives as potent and selective group II metabotropic glutamatereceptor antagonists. J Med Chem 47:4570–4587.

Neugebauer V, Chen PS, and Willis WD (2000) Groups II and III metabotropic glu-tamate receptors differentially modulate brief and prolonged nociception in pri-mate STT cells. J Neurophysiol 84:2998–3009.

Nisenbaum LK, Downing AM, Zhao F, Millen BA, Munsie L, Kinon BJ, Adams DH,Gomez JC, and Penny MA (2016) Serotonin 2A receptor SNP rs7330461 associationwith treatment response to pomaglumetad methionil in patients with schizo-phrenia. J Pers Med 6:9.

Olive MF, Seidel WF, and Edgar DM (1998) Compensatory sleep responses to wake-fulness induced by the dopamine autoreceptor antagonist (2)DS121. J PharmacolExp Ther 285:1073–1083.

Ortiz R, Ulrich H, Zarate CA, Jr, and Machado-Vieira R (2015) Purinergic systemdysfunction in mood disorders: a key target for developing improved therapeutics.Prog Neuropsychopharmacol Biol Psychiatry 57:117–131.

Patil ST, Zhang L, Martenyi F, Lowe SL, Jackson KA, Andreev BV, Avedisova AS,Bardenstein LM, Gurovich IY, Morozova MA, et al. (2007) Activation of mGlu2/3receptors as a new approach to treat schizophrenia: a randomized Phase 2 clinicaltrial. Nat Med 13:1103–1108.

Phillips KG, Bartsch U, McCarthy AP, Edgar DM, Tricklebank MD, Wafford KA,and Jones MW (2012) Decoupling of sleep-dependent cortical and hippocampalinteractions in a neurodevelopmental model of schizophrenia. Neuron 76:526–533.

Pilc A, Wiero�nska JM, and Skolnick P (2013) Glutamate-based antidepressants:preclinical psychopharmacology. Biol Psychiatry 73:1125–1132.

Pin JP and Duvoisin R (1995) The metabotropic glutamate receptors: structure andfunctions. Neuropharmacology 34:1–26.

Popik P, Kos T, Sowa-Ku�cma M, and Nowak G (2008) Lack of persistent effects ofketamine in rodent models of depression. Psychopharmacology (Berl) 198:421–430.

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-amine’s antidepressant effects: How robust is the evidence behind the mTor acti-vation hypothesis. F1000 Res 5:634.

Price JL and Drevets WC (2010) Neurocircuitry of mood disorders. Neuro-psychopharmacology 35:192–216.

Rorick-Kehn LM, Johnson BG, Knitowski KM, Salhoff CR, Witkin JM, Perry KW,Griffey KI, Tizzano JP, Monn JA, McKinzie DL, et al. (2007) In vivo pharmacologicalcharacterization of the structurally novel, potent, selective mGlu2/3 receptor agonistLY404039 in animal models of psychiatric disorders. Psychopharmacology (Berl) 193:121–136.

Schoepp DD, Wright RA, Levine LR, Gaydos B, and Potter WZ (2003) LY354740, anmGlu2/3 receptor agonist as a novel approach to treat anxiety/stress. Stress 6:189–197.

Seidel WF, Maze M, Dement WC, and Edgar DM (1995) Alpha-2 adrenergic modu-lation of sleep: time-of-day-dependent pharmacodynamic profiles of dexmedeto-midine and clonidine in the rat. J Pharmacol Exp Ther 275:263–273.

Serchov T, Clement HW, Schwarz MK, Iasevoli F, Tosh DK, Idzko M, Jacobson KA,de Bartolomeis A, Normann C, Biber K, et al. (2015) Increased signaling viaadenosine A1 receptors, sleep deprivation, imipramine, and ketamine inhibitdepressive-like behavior via induction of homer1a. Neuron 87:549–562.

Simmons RMA, Webster AA, Kalra AB, and Iyengar S (2002) Group II mGluR re-ceptor agonists are effective in persistent and neuropathic pain models in rats.Pharmacol Biochem Behav 73:419–427.

Smith SC, Li R, Mitch CH, and Vetman TN (2012) inventors, Eli Lilly, as-signee. 4-Substituted-3-phenylsulfanylmethyl-bicyclo[3.1.0]hexane compoundsas mGluR 2/3 antagonists. International patent application WO 2012/0680678.2016 Apr 5.

Swanson CJ, Bures M, Johnson MP, Linden AM, Monn JA, and Schoepp DD (2005)Metabotropic glutamate receptors as novel targets for anxiety and stress disorders.Nat Rev Drug Discov 4:131–144.

Takamori K, Hirota S, Chaki S, and Tanaka M (2003) Antipsychotic action of se-lective group II metabotropic glutamate receptor agonist MGS0008 and MGS0028on conditioned avoidance responses in the rat. Life Sci 73:1721–1728.

Tolstikov V, Nikolayev A, Laska D, Kuo MS, and Duffin K (2013) Metabolomics inputin a search for chronic kidney disease targets utilizing clinical cross-platform omicsdata integration, in Proceedings of the 61st ASMS Conference on Mass Spec-trometry and Allied Topics; 2013 Jun 9–13, Minneapolis MN. American Society forMass Spectrometry, Santa Fe, NM.

Traynelis SF, Wollmuth LP, McBain CJ, Menniti FS, Vance KM, Ogden KK,Hansen KB, Yuan H, Myers SJ, and Dingledine R (2010) Glutamate receptorion channels: structure, regulation, and function. Pharmacol Rev 62:405–496.

Van Gelder RN, Edgar DM, and Dement WC (1991) Real-time automated sleepscoring: validation of a microcomputer-based system for mice. Sleep 14:48–55.

Willner P (1997) The mesolimbic dopamine system as a target for rapid antide-pressant action. Int Clin Psychopharmacol 12 (Suppl 3):S7–S14.

Witkin JM and Eiler WJA (2006) Antagonism of metabotropic glutamate group IIreceptors in the potential treatment of neurological and neuropsychiatric disor-ders. Drug Dev Res 67:757–769.

Witkin JM, Monn JA, Schoepp D, Li X, Overshiner C, Mitchell SN, Carter G, JohnsonB, Rasmussen K, and Rorick-Kehn LM (2016a) The rapidly-acting antidepressantketamine rapidly engages dopaminergic mood circuits. J Pharmacol Exp Ther 358:71–82.

Witkin JM, Ornstein PL, Mitch CH, Li R, Smith SC, Heinz BA, Wang XS, Xiang C,Carter JH, Anderson WH, et al. (2016b) In vitro pharmacological and ratpharmacokinetic characterization of LY3020371, a potent and selectivemGlu2/3 receptor antagonist. Neuropharmacology, in press DOI: 10.1016/j.neuropharm.2015.12.021.

Xia J, Mandal R, Sinelnikov I, Broadhurst D, and Wishart DS (2012) MetaboAnalyst2.0—a comprehensive server for metabolomic data analysis. Nucl Acids Res 40:W127–W133.

Yasuhara A, Nakamura M, Sakagami K, Shimazaki T, Yoshikawa R, Chaki S, OhtaH, and Nakazato A (2006) Prodrugs of 3-(3,4-dichlorobenzyloxy)-2-amino-6-fluorobicyclo[3.1.0]hexane-2,6-dicarboxylic acid (MGS0039): a potent and orallyactive group II mGluR antagonist with antidepressant-like potential. Bioorg MedChem 14:4193–4207.

Zanos P, Moaddel R, Morris PJ, Georgiou P, Fischell J, Elmer GI, Alkondon M, YuanP, Pribut HJ, Singh NS, et al. (2016) NMDAR inhibition-independent antide-pressant actions of ketamine metabolites. Nature 533:481–486.

Zarate CA, Jr, Brutsche N, Laje G, Luckenbaugh DA, Venkata SL, Ramamoorthy A,Moaddel R, and Wainer IW (2012) Relationship of ketamine’s plasma metaboliteswith response, diagnosis, and side effects in major depression. Biol Psychiatry 72:331–338.

Zarate CA, Jr, Singh JB, Carlson PJ, Brutsche NE, Ameli R, Luckenbaugh DA,Charney DS, and Manji HK (2006) A randomized trial of an N-methyl-D-aspartateantagonist in treatment-resistant major depression. Arch Gen Psychiatry 63:856–864.

Zhai J, Tian MT, Wang Y, Yu JL, Köster A, Baez M, and Nisenbaum ES (2002)Modulation of lateral perforant path excitatory responses by metabotropic gluta-mate 8 (mGlu8) receptors. Neuropharmacology 43:223–230.

Address correspondence to: J.M. Witkin, Lilly Research Laboratories, LillyCorporate Center, Indianapolis, IN 46285-0510. E-mail: [email protected],S.N. Mitchell, Lilly Research Laboratories, Eli Lilly and Company, Wind-lesham, Surrey, U.K., [email protected]

86 Witkin et al.

at ASPE



ownloaded from

mailto:[email protected]

mailto:[email protected]


comparative effects of ly3020371, a potent and selective...

Documents