(T1030-08-49) Leveraging Automation to Develop a Robust High Throughput Thermodynamic Solubility Assay Using Neat Drug Powder

Purpose: Thermodynamic solubility values in various media are important during drug discovery for informing early dose predictions, differentiation of compounds based on developability, and early formulation strategy development. Traditionally, these solubility assays are time consuming and have lower throughput when starting material is neat powder as opposed to a solvated stock solution of compound. Here we leverage automation to increase throughput, improve data quality, expand data collection, decrease compound consumption, and reduce human error and excessive hands-on time to develop a robust high-throughput thermodynamic solubility assay using neat drug powder.

Methods: A solid compound distribution technique called “ChemBeads” was leveraged to dilute small quantities of solid compounds to improve ease of weighing and improve handling. Compound batching on 96-well plates, short UPLC gradient methods, UPLC plate organizers, and plate-based positive pressure filtration techniques were utilized to increase throughput and decrease turnaround time. Plate-based filtration using positive pressure was used to eliminate manual filtration and time-consuming ultracentrifugation processes. An Opentrons OT2 liquid handling robot was used to eliminate manual creation of standard curves, neat sample dilution, and repetitive pipetting of diluents. Data-rich experimentation was enabled via implementation of solid form analysis by inverted, plate-based microscopy of neat starting material along with post-equilibration residual solids.

Results: The introduction of ChemBead technology improved consistency to within +/-10% error in dispensing sub milligram quantities of API (Active Pharmaceutical Ingredient) across sample tubes. This technology allows for quick and easy dispensing of small amounts of API without manual weighing or powder dispensing robots, which are often difficult to implement early in drug discovery due to minimum material requirements. Batching compounds has allowed for fewer runs, and improved UPLC methodologies have cut run times in half. Positive pressure filtration is 70% faster in solid removal compared to sequential manual filtration and removes the step of transferring to and from centrifuge tubes, improving assay efficiency. This method of plate-based solubility also requires only 5mg of solid API, while traditional vial-based or shake flask methods of powder solubility requires around 2mg solid compound per replicate. In this case, nine replicates of sample would require 18mg of neat drug powder API, leading to an overall savings of roughly 13 mg of compound, which is critical in drug discovery. The implementation of automated liquid handling with the Opentrons OT2 robot resulted in 30%-time savings over manual 12-channel pipetting of standard curves, diluents, and filtrates. Implementation of the OT2 also resulted in improved linearity of standard curves, evidenced by a five-fold drop in necessary outlier point removal. This increase in linearity is indicative of a significant reduction in human error and has eliminated the need for sample re-runs. Adding plate-based microscopy to the assay has enabled crystallinity information to aid in early decision making without significant assay modifications, and has eliminated manual, individual slide preparation that often exceeds 60 samples per run. There is also an opportunity for completely automating this process with the addition of automated stage adjustment and image collection.

Conclusion: An automation focused approach to development of a high-throughput thermodynamic solubility assay has positively impacted the efficiency, speed, API cost, and quality of data for discovery compounds advancing toward pharmacokinetic testing. Overall, development of a 96 well plate-based, method using neat drug powder allowed for full analysis (solubility and crystallinity) of 28 compounds in 3 media with 3 replicates in 4 days compared to 2.5 weeks for traditional methods.

References: Tu, et al. High-Throughput Reaction Screening with Nanomoles of Solid Reagents Coated on Glass Beads. Angewandte Chemie 2019. https://onlinelibrary.wiley.com/doi/full/10.1002/anie.201900536