Mitigation of off-target adrenergic binding and effects on cardiovascular function in the discovery of novel ribosomal S6 kinase 2 inhibitors.

We previously reported the discovery of a novel ribosomal S6 kinase 2 (RSK2) inhibitor, (R)-5-Methyl-1-oxo-2,3,4,5-tetrahydro-1H-[1,4]diazepino[1,2-a] indole-8-carboxylic acid [1-(3-dimethylamino-propyl)-1H-benzoimidazol-2-yl]-amide (BIX 02565), with high potency (IC(50) = 1.1 nM) targeted for the treatment of heart failure. In the present study, we report that despite nanomolar potency at the target, BIX 02565 elicits off-target binding at multiple adrenergic receptor subtypes that are important in the control of vascular tone and cardiac function. To elucidate in vivo the functional consequence of receptor binding, we characterized the cardiovascular (CV) profile of the compound in an anesthetized rat CV screen and telemetry-instrumented conscious rats. Infusion of BIX 02565 (1, 3, and 10 mg/kg) in the rat CV screen resulted in a precipitous decrease in both mean arterial pressure (MAP; to -65 ± 6 mm Hg below baseline) and heart rate (-93 ± 13 beats/min). In telemetry-instrumented rats, BIX 02565 (30, 100, and 300 mg/kg p.o. QD for 4 days) elicited concentration-dependent decreases in MAP after each dose (to -39 ± 4 mm Hg on day 4 at T(max)); analysis by Demming regression demonstrated strong correlation independent of route of administration and influence of anesthesia. Because of pronounced off-target effects of BIX 02565 on cardiovascular function, a high-throughput selectivity screen at adrenergic α(1A) and α(2A) was performed for 30 additional RSK2 inhibitors in a novel chemical series; a wide range of adrenergic binding was achieved (0-92% inhibition), allowing for differentiation within the series. Eleven lead compounds with differential binding were advanced to the rat CV screen for in vivo profiling. This led to the identification of potent RSK2 inhibitors (cellular IC(50) <0.14 nM) without relevant α(1A) and α(2A) inhibition and no adverse cardiovascular effects in vivo.

Molecular features crucial to the activity of pyrimidine benzamide-based thrombopoietin receptor agonists.

The identification of small molecule modulators of biological processes mediated via protein-protein interactions has generally proved to be a challenging endeavor. In the case of the thrombopoietin receptor (TPOr), however, a number of small molecule types have been reported to display biological activity similar to that of the agonist protein TPO. Through a detailed analysis of structure-activity relationships, X-ray crystal structures, NMR coupling constants, nuclear Overhauser effects, and computational data, we have determined the agonism-inducing conformation of one series of small molecule TPOr agonists. The relationship of this agonism-inducing conformation to that of other series of TPO receptor agonists is discussed.

A hybrid mixture discriminant analysis-random forest computational model for the prediction of volume of distribution of drugs in human.

A computational approach is described that can predict the VD(ss) of new compounds in humans, with an accuracy of within 2-fold of the actual value. A dataset of VD values for 384 drugs in humans was used to train a hybrid mixture discriminant analysis-random forest (MDA-RF) model using 31 computed descriptors. Descriptors included terms describing lipophilicity, ionization, molecular volume, and various molecular fragments. For a test set of 23 proprietary compounds not used in model construction, the geometric mean fold-error (GMFE) was 1.78-fold (+/-11.4%). The model was also tested using a leave-class out approach wherein subsets of drugs based on therapeutic class were removed from the training set of 384, the model was recast, and the VD(ss) values for each of the subsets were predicted. GMFE values ranged from 1.46 to 2.94-fold, depending on the subset. Finally, for an additional set of 74 compounds, VD(ss) predictions made using the computational model were compared to predictions made using previously described methods dependent on animal pharmacokinetic data. Computational VD(ss) predictions were, on average, 2.13-fold different from the VD(ss) predictions from animal data. The computational model described can predict human VD(ss) with an accuracy comparable to predictions requiring substantially greater effort and can be applied in place of animal experimentation.