Several research groups have disclosed potent
Several research groups have disclosed potent and selective DGAT-1 inhibitors from several chemically-distinct series. Pre-clinical studies with these compounds have confirmed that small molecule DGAT-1 inhibitors can elicit metabolic outcomes comparable to those observed in DGAT-1−/− mice.19, 20, 21 We have recently disclosed an orally-active, novel DGAT-1 inhibitor PF-04620110 (1) which was advanced to human clinical trials for the treatment of type 2 diabetes. This pyrimidooxazepinone-based series arose from specific design criteria aimed at eliminating the potential photochemical and reactive metabolite risks associated with the pyrimidinooxazine-based DGAT-1 inhibitor 2. As compound 1 advanced into human clinical studies there were questions as to whether the poor passive permeability associated with agent would result in limited distribution into key in vivo compartments, potentially leading to a reduced efficacy maximum. This report details the discovery of neutral, potent and orally-active DGAT-1 inhibitors with high passive permeabilities and their resulting efficacy in preclinical rodent models. Analysis of the key pharmacophore elements of 1 will also be discussed as part of a broader structure–activity-relationship (SAR) analysis.
Results and discussion As previously reported, the bicyclic core of 1 arose from analyses of alternative cores that retained what were assumed to be the key pharmacophore elements of 2 and the spatial orientation of these in three-dimensional space. Incorporation of the fused, seven-membered lactam moiety in 1 properly orients the cyclohexylacetic Aurora Kinase Inhibitor III sidechain relative to 2 (Fig. 1). In addition, it eliminates the conjugated imine-based bicyclic present in 2, which is likely responsible for the photochemical instability associated with this lead. While our design resulted in a potent DGAT-1 inhibitor, with a pharmacokinetic/safety profile supporting its advancement to human clinical trials, there were a number of unanswered SAR questions relating to this novel chemotype. The combination of limited bicyclic core SAR disclosed around 2 and the inherent chemical differences between the two core structures, led us to pursue a systematic investigation of the key pharmacophore elements present in 1. While the in-plane orientation of the phenycyclohexyl sidechain relative to the heterobicylic cores of 1 and 2 (Fig. 1) suggested this to be a necessary requirement for potent DGAT-1 inhibition, alternative bicyclic cores that led to alternative spatial arrangements of these two pharmacophore elements were evaluated prior to further optimization of 1. An additional goal was to improve on its low passive permeability (apparent passive permeability (Papp)= 1mm10−6cm/s) associated with 1. This property became the focus of our back-up effort since there was a concern that this low passive permeability might limit exposure to key target tissues in vivo. Computational models suggested that conversion of the lactam functionality in 1 to the corresponding bicyclic amine-based core could lead to a substantial increase in passive permeability (Papp=7×10−6cm/s). Toward this end, global reduction of 1 afforded amine 3. Modeling of the predicted low energy conformation of 3 (Fig. 1) suggested a significantly altered relationship between the core and sidechain relative to 1. This modification led to a loss in DGAT-1 potency of >110-fold relative to 1 (Table 1). It is important to note that the loss in potency of 3 is not driven by reduction of the carboxylic acid functionality, given that a propyl alcohol analog of 1 is a potent inhibitor of DGAT-1 (unpublished data). An alternative lactam motif evaluated was the regioisomeric lactam 4. Modeling of this lactam suggested its low energy conformation would be similar to amine 3, while retaining a neutral bicyclic core (Fig. 1). The lack of DGAT-1 inhibitory activity associated with 4 was consistent with the hypothesis that an in-plane relationship between the bicyclic core and phenycyclohexyl sidechain found in 1 was critical. These results led to a focusing of efforts on analogs containing the pyrimidooxazepinone core present in 1, with a goal of improving passive permeability.