Translational receptor occupancy for the 5-HT4 partial agonist PF-04995274 in rats, non-human primates and healthy volunteers

Abstract

PF-04995274 is a 5-HT4 partial agonist in clinical development as a potential therapeutic to improve cognition in patients with Alzheimer's disease. PF-04995274 has subnanomolar binding affinities for multiple human 5-HT4 receptor isoforms, as well as rat, dog and non-human primate (NHP) isoforms. We have used a selective 5-HT4 receptor tracer, [11C]PF-05127401 ([11C]SB207145; Marner et al., 2009), to determine receptor occupancy (RO) for PF-04995274 in NHP and human using positron emission tomography (PET), and [3H]PF-05127401 for rat in vivo binding studies. All animal study procedures conducted were approved by an institutional review committee (IACUC). PF-04995274 and [3H]PF-05127401 were administered to rats by subcutaneous and intravenous (IV) injection, respectively. Rats were euthanized 15 minutes after [3H]PF-05127401 and brain homogenates were rapidly filtered. For NHP imaging, two NHPs were studied at baseline and with PF-04995274 (0.0005 and 0.0021 mg/kg IV bolus). Drug levels were maintained constant by a lower maintenance dose throughout the 90 minutes dynamic PET scan with [11C]PF-05127401. The clinical study utilized the same tracer as the NHP. RO was determined in 8 healthy volunteers via three 120 minute scans: baseline, Tmax and 24 hours postdose [range: 0.25-5.0 mg]. Plasma samples were collected and analyzed for PF-04995274. RO was determined by Simplified Reference Tissue Model with cerebellum as the reference region. PF-04995274 displaced [3H]PF-05127401 binding in rat striatum in a dose-dependent manner with an ED50 of 0.008 mg/kg at 1 hour. In a time-course study, RO levels returned to baseline by 4 hours after PF-04995274 administration. NHP [11C]PF-05127401 microPET studies also showed displacement by PF-04995274. A simple Emax model was used to characterize the exposure-RO relationship. The model estimated a Kd value of 0.161±0.015 nM which was ∼ 7-fold lower than the in vitro Ki value. Human RO data were also characterized by an Emax model with a simple additive error structure. Emax was estimated at 88.0 (3.97% SE) and the IC50 (SE) was estimated at 119 (39.3) pg/mL total PF-04995274 plasma concentration. After correcting for species specific Ki and plasma protein binding the observed IC50 in human was similar (within 3 fold) to both rodent and NHP estimates.