Functional NMDA receptor-based target engagement biomarkers for - PowerPoint PPT Presentation

Functional NMDA receptor-based target engagement biomarkers for schizophrenia research Daniel C. Javitt, M.D., Ph.D. Professor & Director, Division of Experimental Therapeutics Columbia University Medical Center Director of Schizophrenia Research Nathan Kline Institute

Di Disclosures • Consultant : Pfizer, FORUM, Autifony, Glytech, Lundbeck, Concert, Cadence • Scientific Advisory Board : Promentis, NeuroRx, Phytecs • Equity : Glytech, AASI, NeuroRx • Intellectual property rights : Glycine, D-serine and glycine transport inhibitors in Sz; D-cycloserine, combined NMDAR/5-HT2AR antagonism in depression & PTSD; visual ERP for early diagnosis of Alzheimer disease • Off-label treatment : pomaglumetad

Functi tional tar arget en engagement bi biomarkers The challenge • Glutamatergic theories of schizophrenia have become increasingly established over the past 25 years • Many glutamate-base pharmacological approaches show encouraging effects in preclinical models • To date, none has translated into an effective medication The solution • Need target engagement biomarkers to permit better translation from animals to humans • Allow “ FAST fail ” decisions regarding mechanism of action: • No target engagement - “fail” drug • Target engagement but no beneficial clinical effect – “fail” mechanism • Permit informed dose selection

Ac Academia, Go Government, and and Phar harma con ontributions Sci Transl Med. 3:102mr2, 2011 TRANSLATION TYPE (Institute of Medicine, 2013)

NM NMDAR-based tr trea eatment de development Background • NMDAR antagonists such as phencyclidine (PCP) and ketamine induce symptoms, neurocognitive deficits and neurophysiological deficits that closely resemble schizophrenia • Acute NMDAR blockade is associated with an increase in presynaptic glutamate release (“glutamate surge”) • Hypothesis 1 : Symptoms/neurocognitive deficits results from NMDAR hypofunction • Hypothesis 2 : Symptoms/neurocognitive deficits result from “glutamate surge” Glutamate synapse N-methyl-D-aspartate (NMDA) receptor Treatment targets - Moghaddam & Javitt., - Javitt & Zukin, Am J Psychiatry , 148: 1301-8, 1991 Neuropsychopharmacol Rev , 2011

NM NMDAR-based tr trea eatment de development Background • NMDAR antagonists such as phencyclidine (PCP) and ketamine induce symptoms, neurocognitive deficits and neurophysiological deficits that closely resemble schizophrenia • Acute NMDAR blockade is associated with an increase in presynaptic glutamate release (“glutamate surge”) • Hypothesis 1 : Symptoms/neurocognitive deficits results from NMDAR hypofunction • Hypothesis 2 : Symptoms/neurocognitive deficits result from “glutamate surge” Glutamate synapse Potential treatment approaches Hypothesis 2 : Desired solution - ↓ presynaptic glutamate release Treatment targets Hypothesis 1 : Desired solution – ↑ postsynaptic NMDAR activity - Moghaddam & Javitt., Neuropsychopharmacol Rev , 2011

Effect ct of of metabotropic glutamate rece eceptors: pr precl clinical mod odels • Treatment with an NMDAR antagonist (PCP) leads to significant increases in glutamate & dopamine release in prefrontal cortex, along with increases in locomotion • Pretreatment with the mGluR2/3 agonist LY354740 blocks both the increases in glutamate and locomotion • PCP-induced increases in dopamine release also inhibited but not blocked Moghaddam et al., Science. 281:1349-52, 1998

Cl Clinical effects of of pom pomaglumetad (LY21400230) Initial positive result Subsequent failures to replicate % Change PANSS Total 15 80 mg % Change PANSS Total 10 5 0 -5 -10 -15 -20 -25 0 1 2 3 4 Week Week Downing et al., BMC psychiatry. 14:351, 2014 Patil et al., Nat Med. 13:1102-7, 2007

FAST-Fail: : Psychotic c di diso sorders Questions • Why don’t mGluR2/3 agonists work in people, even though they work in rodent models? • Were doses used in the clinical studies sufficient to adequately test the hypothesis? • What biomarkers can be used to test target engagement? NIMH FAST-psychosis spectrum (FAST-PS) biomarker project • Phase I : Biomarker selection • Selected measures potentially sensitive to the glutamate “surge” • PharmacoBOLD, Glutamate MRS, task-based fMRI • Phase II : Target engagement studies with pomaglumetad (healthy volunteers)

Pha hase I bi biom omark rker val alidation JAMA Psychiatry. 75:11-19, 2018 Design • 65 Healthy volunteers • Randomized 2:1 to ketamine infusion vs. placebo • 2 MRI sessions per subject • Session 1: Glutamate 1 H-MRS, task-based fMRI • Session 2: “ PharmacoBOLD ” • Psychosis ratings: CADSS, BPRS

Magn agnetic reso esonance spe pectroscopy (MRS): Methods Mechanism of action Example spectrum • Increase in presynaptic glutamate release leads to a rapid increase in glutamate synthesis from glucose • The increase in glutamate synthesis , in turn, leads to an increase in total glutamate+glutamine (“ Glx ”) in brain Abdallah et al., Neuropsycho-pharmacology 43 : 2154-2160, 2018.

Magn agnetic reso esonance spe pectroscopy (MRS): Results Mechanism of action Results • Increase in presynaptic glutamate release leads to a • Significant increase in Glx content at 15 min rapid increase in glutamate synthesis from glucose • BUT : Moderate effect size (d=.64) • The increase in glutamate synthesis , in turn, leads to an • Not sufficiently robust to be able to detect a increase in total glutamate+glutamine (“ Glx ”) in brain drug effect even if present * Abdallah et al., Neuropsychopharmacology 43 : 2154-2160, 2018.

Phar harmacoBOLD Mechanism of action Results • • Extremely large effect (d=5.4) Increase in presynaptic glutamate release leads to a • Sufficient to detect a change, if present rapid increase in glutamate synthesis from glucose • The increased metabolic rate leads to an acute increase in BOLD reponse Abdallah et al., Neuropsycho-pharmacology 43 : 2154-2160, 2018.

Pha hase 2 stu tudy: Design overview • 81 healthy volunteer completers across 4 sites (Columbia, NYU, UAB, UCLA) (100 total subjects randomized) • Randomized (1:1:1) double blind administration of a placebo, 40mg bid POMA, or 160mg bid POMA • 40 mg BID = dose used in prior successful clinical trial • 160 mg BID = maximum tolerated dose (limited by nausea, vomiting) • Subjects took POMA or placebo for 10 days • Administered ketamine or placebo on the final day of treatment • Ketamine dose - 0.23 mg/kg bolus over 1 minute • Ketamine-induced prefrontal glutamate activity as measured by resting BOLD fMRI (pharmacoBOLD) • Inclusion criteria : BOLD fMRI response in dACC-ROI > 0.5% at Screening

Pha hase 2 Pom omaglumetad stu tudy: : results Clinical ratings (BPRS) dACC pharmacoBOLD 0 0 Mean change in BPRS Total -0.2 %change dACC peak -1 -0.4 d=-.14, p=.52 vs. placebo d=-.33, p=.15 -0.6 vs. placebo -2 -0.8 Main effect of treatment : p=.33 Main effect of treatment : p=.80 -3 -1.0 Placebo Low-dose High-dose Placebo Low Dose High Dose Placebo Low Dose High Dose Placebo Low-dose High-dose Treatment Group Treatment Group (40 mg BID) (160 mg BID) (40 mg BID) (160 mg BID)

mGluR/3 ag agonist de development in n Sz Sz: : Conclusions Conclusions • The good news is that the bad news may be wrong • At doses used in prior clinical studies, pomaglumetad does not show evidence of significant functional target engagement • Other compounds are presently under investigation • The bad news is that the mechanism may not be viable • Further dose escalation may be precluded by high rates of nausea/vomiting • HOWEVER : Side effects are likely due to local effects in stomach; could be blocked by peripheral antagonists • Future studies with higher doses are needed • “ PharmacoBOLD ” can be used for T1 -type translation to healthy volunteers • Not suitable for T2-T4 translation involving patient groups • Biomarker studies should probably be implemented before, rather than after , intensive phase 3 development

T2 T2 tr translation: Mismatch ne negativity (MMN) • Elicited in the context of an auditory “oddball” paradigm • Reflects information processing dysfunction at the level of auditory sensory cortex • Consistent deficits in Sz related to impaired functional outcome • Can be assessed in parallel in rodents, monkeys, & humans • Inhibited by NMDAR antagonists (e.g., ketamine) • Improved by putative NMDAR agonists (e.g. D-serine) MMN Meta-analysis (Sz) ERP Biomarker Qualification Consortium www.erpbiomarkers.org - Avissar et al., Schizophr Res . 191:25-34; 2018

Col Collaborators Pomaglumetad MMN Consortium Biomarker validation Columbia UCLA Yale University Jeff Lieberman (PI) Steve Marder John Krystal Josh Kantrowitz Junghee Lee Phil Corlett Larry Kegeles Michael Green Graeme Mason Jack Grinband Douglas Rothman Ragy Girgis NYU Maolin Qui Melanie Wall Don Goff Tse Hswei Choo Fernando Boada UC Davis Marlene Carlson Erica Diminich Cameron Carter J. Daniel Ragland Jim Robinson (NKI) UA Birmingham Richard Maddock Adrian Lahti Costin Tanase David White Tyler Lesh Mark Bolding