Process Development and Characterization for Impurity Control of a Dynamic Kinetic Resolution Reaction Using Data Rich Experimentation

Telescoped processes with multiple reactions can streamline manufacturing in the commercial syntheses of APIs and their intermediates. However, robustness challenges may arise, as there tends to be no normalization like that of isolated intermediates. To mitigate risks caused by these challenges, Data Rich Experimentation (DRE) is utilized to characterize telescoped processes to better enhance understanding around process factor sensitivity. This presentation covers the development of a telescoped process involving Fluorination and Dynamic Kinetic Resolution (DKR) reactions in the commercial synthesis of an API intermediate before scaling up and tech transferring to larger scales.

The current filed process for Belzutifan involves a chemocatalytic reduction with Ru. To be more sustainable and avoid the price volatility of precious metals, a post approval route that employed KRED was developed, streamlining the process to utilize a greener, safer, and more robust chemo-enzymatic through-process over the previous Ru-catalyzed route. Employing biocatalysis introduced new process challenges, including new reaction impurities and sensitivities to incoming materials from upstream reactions. Prior to scaling up to a commercial site, this DKR reaction was meticulously studied to assess the operable design space around applicable process variables. After designing a baseline procedure that maintained enzyme stability and provided the necessary reaction conversion, the operating ranges of the reaction parameters were refined through risk assessment for quality. Range finding experiments were performed using DRE systems to quickly and efficiently identify the boundaries of the design space. Results from these DRE screens reduced the number of parameters needed in subsequent experimentation. Process characterizations were performed according to a Design of Experiment (DoE) structure to statistically quantify the relationship between operating factors and reaction responses. These experiments were also performed with elements of a DRE strategy, using process analytical tools (PAT) and automated reactor sampling to generate reaction profiles to understand kinetics as well as end of reaction conditions. Data analysis generated practical understandings regarding pH control and KRED loading on the production of several reaction impurities. Laboratory data was bridged to the production environment through successful scale-up to a pilot plant facility. Key takeaways from the risk assessment work instilled confidence in scaling up to the manufacturing production plant.