In order to develop improved radioligands for imaging brain CB(1) receptors with positron emission tomography (PET) based on rimonabant (5-(4-chlorophenyl)-1-(2,4-dichlorophenyl)-4-methyl-N-(piperidin-1-yl)-1H-pyrazole-3-carboxamide, 1), we synthesized compounds 9a-s in which the N-piperidinyl ring was replaced with a 4-(4-cyanotetrahydro-2H-pyranyl) or 1-cyanocyclohexyl ring. Such changes were expected to be almost isosteric with 1, confer greater metabolic resistance, and in the case of the 4-(4-cyanotetrahydro-2H-pyranyl) compounds, substantially reduce lipophilicity. One derivative, 1-(2-bromophenyl)-N-(1-cyanocyclohexyl)-5-(4-methoxyphenyl)-4-methylpyrazole-3-carboxamide (9n), showed high affinity (K(i) = 15.7 nM) and selectivity for binding to CB(1) receptors. The corresponding 4-(4-cyanotetrahydro-2H-pyranyl) derivative (9m) also showed quite high affinity for CB(1) receptors (K(i) = 62 nM) but was found to have even higher affinity (K(i) = 29 nM) for the structurally unrelated 18 kDa translocator protein (TSPO). Some other minor structural changes among 9a-s were also found to switch binding selectivity from CB(1) receptors to TSPO or vice versa. These unexpected findings and their implications for the development of selective ligands or PET radioligands for CB(1) receptors or TSPO are discussed in relation to current pharmacophore models of CB(1) receptor and TSPO binding sites.