Purpose: Early-stage stability screening assays with high predictive power can influence compound design and enable medicinal chemistry teams to produce compounds with better stability profiles from the beginning, and greatly decrease the possibility of later stage failures. Degradation of drug compounds mainly occurs via hydrolysis or oxidation; light absorption can also induce photolytic degradation. In early high-throughput hydrolytic stability analyses, aggressive conditions are typically applied for faster screening. However, it can be challenging to extrapolate the real stability risk and rank compounds due to the over-prediction and limited discriminative window. By investigating the critical parameters that impact the predictive data quality, a highly discriminative stability screen with optimized assay parameters and experimental data quality is proposed. Knowledge from predictive science about oxidation potential and photolytic stability can also help assess the overall stability risk. Together these provide early guidance on the potential stability risk of a drug molecule and enable more confident decision-making in the compound design, selection and development.
Methods: Five model compounds with ranges of known stability profiles (Aspirin, Esomeprazole, Felodipine, Metoprolol, and Rosuvastatin) were selected to investigate multiple hydrolytic screening conditions. The model compounds were prepared at 25 and 200 µM in three different pH (1, 7.4 and 10) solutions and placed in thermos-stated autosampler blocks at 40 C and 70 C. The samples were injected hourly over a twenty-four hour timecourse and monitored for peak area decrease by liquid chromatography with ultraviolet (UV) and mass (MS) detection (LC-DAD-MS/MS).
A set of more than ten compounds with known UV spectra and molar extinction coefficient at 290nm or above were selected for the validation of predictive photostability methodology. Molar extinction coefficients at UV-A and UV-B regions were calculated through Gaussian functions to convolute calculated oscillator strengths and spectra. The methodology is based on quantum mechanical calculations.
Results: The selected five model compounds with known stability profiles all followed first order reaction kinetics. The Arrhenius equation was used for stability prediction and activation energy calculation. The prediction data quality was examined by linear regression coefficient. Metoprolol and Rosuvastatin are stable under pH 10 conditions; the limit of differentiation window was calculated based on standard derivation of these compounds over the time course of the analysis. The impact and interplay of the critical assay parameters (temperature, concentration, and detection technique) on the predictive power were assessed. Better data quality was achieved by using higher concentration, reduced temperature, and UV detection. However, for compounds with limited solubility, lower concentration and multiple reaction monitoring (MRM) mass detection yielded more data points and better results. For the best predictivity of real time stability, extrapolation from two temperature screens is recommended. For compounds with the fastest degradation rates, use of higher concentration is recommended to yield sufficient prediction and differentiation. Careful consideration of these factors in designing stability screens can significantly increase the prediction/differentiation power and more effectively guide compound design.
Compound degradation rate by photodegradation is affected by its UV energy absorption. For compounds which exhibited strong UV absorption in UVA and UVB regions with higher molar extinction coefficient, it was suggested that the compound would absorb much more photoenergy resulting in easier photoexcitation and hence photodegradation products upon exposure to UV light. The calculated molar extinction coefficients are in a good agreement with the experimental values.
Conclusion: In early drug discovery stages, often limited material is available to evaluate stability risk. Hence it is valuable to apply predictive sciences to calculate intrinsic molecular properties and use highly discriminative stability screens. By focusing the experimental screen on hydrolysis, compounds will be designed with acceptable hydrolytic stability. Oxidation can be predicted, and photolysis potential can be assessed from a predicted UV curve by using quantum mechanical tools. This combined approach allows early identification of stability risk and assesses the need for further stability evaluation of specific chemical series/scaffolds. It also ensures sufficient stability in the biological assay buffers for reliable measurement of biological activity. In later pharmaceutical development, mitigation options are available to reduce oxidation and photolysis of the drug compound, thus early compound design can allow greater uncertainty inherent to oxidative and photolytic instability. The presented approach for early influence on compound design will reduce significant surprises in later development.
Jenny Ottosson– Pharmaceutical Sciences, IMED Biotech Unit, AstraZeneca, Gothenburg, Sweden
Sharon Tentarelli– Oncology, IMED Biotech Unit, AstraZeneca, Boston, USA, Massachusetts
Anders Broo– Pharmaceutical Sciences, IMED Biotech Unit, AstraZeneca, Gothenburg, Sweden