ABSTRACTBoth central and peripheral cannabinoid receptor type 1 (CB1R) have been considered to be among the key targets for obesity treatment. First generation CB1R antagonists/inverse agonists such as rimonabant and taranabant exhibited severe CNS side effects such as anxiety and depression, which are considered to be related to the compounds' ability to access central CB1R. Recently, several compounds have been developed as second generation antagonists with a profile of restriction to peripheral CB1R. We evaluated the distribution of TM38837, a second generation CB1R antagonist, using brain and whole body PET in three cynomolgus monkeys, and established the relationship between CB1R occupancy and dose/plasma concentration of TM38837 in comparison with rimonabant. A brain PET study was performed using [11C]MePPEP, a PET radioligand for CB1R, to evaluate the brain CB1R occupancy of TM38837 at various plasma concentrations in comparison with rimonabant at known efficacious plasma concentrations. A whole body PET study was performed to investigate the change of peripheral distribution of [11C]MePPEP by TM38837 administration, which indirectly estimated the effects to the peripheral CB1R by TM38837. CB1R occupancy by both TM38837 and rimonabant increased in a dose/plasma concentration‐dependent manner. However, in vivo affinity by plasma level was more than 100 times lower for TM38837. Peripherally, [11C]MePPEP accumulation decreased in gall bladder and brown adipose tissue by TM38837 administration. TM38837 showed rather lower CB1R occupancy than rimonabant at the expected therapeutic plasma level, which is expected to reduce CNS side effects in clinical situations. Further clinical development of TM38837 is warranted. Synapse 68:89–97, 2014. © 2013 Wiley Periodicals, Inc.