Discovery and Characterization of the Potent and Highly Selective (Piperidin-4-yl)pyrido[3,2- d]pyrimidine Based in Vitro Probe BAY-885 for the Kinase ERK5.

The availability of a chemical probe to study the role of a specific domain of a protein in a concentration- and time-dependent manner is of high value. Herein, we report the identification of a highly potent and selective ERK5 inhibitor BAY-885 by high-throughput screening and subsequent structure-based optimization. ERK5 is a key integrator of cellular signal transduction, and it has been shown to play a role in various cellular processes such as proliferation, differentiation, apoptosis, and cell survival. We could demonstrate that inhibition of ERK5 kinase and transcriptional activity with a small molecule did not translate into antiproliferative activity in different relevant cell models, which is in contrast to the results obtained by RNAi technology.

Characterization of a Novel Hafnium-Based X-ray Contrast Agent.

OBJECTIVE: Characterization of BAY-576, a new x-ray contrast agent which is not based on iodine, but rather on the heavy metal hafnium. Compared with iodine, hafnium provides better x-ray absorption in the energy range of computed tomography (CT) and allows images of comparable quality to be acquired at a significantly reduced radiation dose. MATERIALS AND METHODS: A range of standard methods were used to explore the physicochemistry of BAY-576 as well as its tolerability in in vitro assays, its pharmacokinetics and toxicology in rats, and its performance in CT imaging in rabbits. RESULTS: BAY-576 is an extraordinarily stable chelate with a metal content of 42% (wt/wt) and with excellent water solubility. Formulations of 300 mg Hf/mL exhibited viscosity (3.3-3.6 mPa) and osmolality (860-985 mOsm/kg) in the range of nonionic x-ray agents. No relevant effects on erythrocytes, the coagulation, or complement system or on a panel of 87 potential biological targets were observed. The compound did not bind to plasma proteins of a number of species investigated. After intravenous injection in rats, it was excreted fast and mainly via the kidneys. Its pharmacokinetics was comparable to known extracellular contrast agents. A dose of 6000 mg Hf/kg, approximately 10 to 20 times the expected diagnostic dose, was well tolerated by rats with only moderate adverse effects. Computed tomography imaging in rabbits bearing a tumor in the liver demonstrated excellent image quality when compared with iopromide at the same contrast agent dose in angiography during the arterial phase. At 70% of the radiation dose, BAY-576 provided a contrast-to-noise ratio of the tumor, which was equivalent to iopromide at 100% radiation dose. CONCLUSIONS: The profile of BAY-576 indicates its potential as the first compound in a new class of noniodine x-ray contrast agents, which can contribute to the reduction of the radiation burden in contrast-enhanced CT imaging.

Estimating the domain of applicability for machine learning QSAR models: a study on aqueous solubility of drug discovery molecules.

We investigate the use of different Machine Learning methods to construct models for aqueous solubility. Models are based on about 4000 compounds, including an in-house set of 632 drug discovery molecules of Bayer Schering Pharma. For each method, we also consider an appropriate method to obtain error bars, in order to estimate the domain of applicability (DOA) for each model. Here, we investigate error bars from a Bayesian model (Gaussian Process (GP)), an ensemble based approach (Random Forest), and approaches based on the Mahalanobis distance to training data (for Support Vector Machine and Ridge Regression models). We evaluate all approaches in terms of their prediction accuracy (in cross-validation, and on an external validation set of 536 molecules) and in how far the individual error bars can faithfully represent the actual prediction error.

Estimating the domain of applicability for machine learning QSAR models: a study on aqueous solubility of drug discovery molecules.

We investigate the use of different Machine Learning methods to construct models for aqueous solubility. Models are based on about 4000 compounds, including an in-house set of 632 drug discovery molecules of Bayer Schering Pharma. For each method, we also consider an appropriate method to obtain error bars, in order to estimate the domain of applicability (DOA) for each model. Here, we investigate error bars from a Bayesian model (Gaussian Process (GP)), an ensemble based approach (Random Forest), and approaches based on the Mahalanobis distance to training data (for Support Vector Machine and Ridge Regression models). We evaluate all approaches in terms of their prediction accuracy (in cross-validation, and on an external validation set of 536 molecules) and in how far the individual error bars can faithfully represent the actual prediction error.

Machine learning models for lipophilicity and their domain of applicability.

Unfavorable lipophilicity and water solubility cause many drug failures; therefore these properties have to be taken into account early on in lead discovery. Commercial tools for predicting lipophilicity usually have been trained on small and neutral molecules, and are thus often unable to accurately predict in-house data. Using a modern Bayesian machine learning algorithm--a Gaussian process model--this study constructs a log D7 model based on 14,556 drug discovery compounds of Bayer Schering Pharma. Performance is compared with support vector machines, decision trees, ridge regression, and four commercial tools. In a blind test on 7013 new measurements from the last months (including compounds from new projects) 81% were predicted correctly within 1 log unit, compared to only 44% achieved by commercial software. Additional evaluations using public data are presented. We consider error bars for each method (model based error bars, ensemble based, and distance based approaches), and investigate how well they quantify the domain of applicability of each model.