The route selection and process research and development of a practical synthesis for JAK2 inhibitor LY2784544 is described. The first-generation synthesis route, similar to that used in discovery for derivatization of a benzylic amine moiety, was 14 overall steps and possessed several steps that required extensive development for large-scale production. Route selection considerations led to a modified synthesis that utilized a novel vanadium-catalyzed carbon–carbon bond-forming arylation reaction for incorporation of the key benzylic morpholine moiety. A protecting group used to mask an amino pyrazole unit was modified from PMB to tert-butyl, resulting in a dramatic reduction in the overall length of the route. These two major changes resulted in an eight-step synthesis, which was six steps shorter than the first-generation synthesis. In the pilot plant, the new synthesis was scaled to produce >100 kg of LY2784544 in high yield and purity under GMP conditions. The overall development including the vanadium-catalyzed C–C bond-forming methodology, a ketone reductive deoxygenation, and a palladium-catalyzed amination is described.

The design, development, and implementation of a pilot-scale continuous Schotten−Baumann amide bond formation and reactive crystallization to afford LY2886721 is described. The material met all API quality attributes and was comparable to material produced by a defined batch process. The scalability of the reaction and crystallization processes was confirmed during the development process. The pilot-scale equipment set was contained in a walk-in fume hood and operated at a production rate of 3 kg/day in a 72 h continuous run. Significant technical and business drivers for running the process in continuous flow mode were proposed and examined during development. The continuous process provided for lab hood commercialization and provided for minimal material at risk in the process. The demonstration also confirmed the risk inherent to operation of a tubular reactor under supersaturated conditions, and fouling occurred in the plug flow reactor. Fouling also occurred in the crystallizer. Recognizing these deficiencies, the process operated within the footprint of a standard walk-in fume hood, providing a successful demonstration of the opportunities afforded by continuous processing for low volume pharmaceuticals.

Efficient continuous Grignard and lithiation processes were developed to produce one of the key regulatory starting materials for the production of edivoxetine hydrochoride. For the Grignard process, organometallic reagent formation, Bouveault formylation, reduction, and workup steps were run in continuous stirred tank reactors (CSTRs). The lithiation utilized a hybrid approach where plug flow reactors (PFRs) were used for the metal halogen exchange and Bouveault formylation steps, while the reduction and workup steps were performed in CSTRs. Relative to traditional batch processing, both approaches offer significant advantages. Both processes were high-yielding and produced the product in high purity. The two main processes were directly compared from a number of perspectives including reagent and operational safety, fouling potential, process footprint, need for manual operation, and product yield and purity.

The development of an alternative synthetic route to a functionalized imidazopyridazine which strategically streamlines the synthesis and avoids a number of problematic reagents is described. Key to the success of this alternative route is the use of two C–H functionalization reactions: a Pd-catalyzed direct benzylation reaction to functionalize a C–H bond with a substituted benzyl group and a V-catalyzed NMO addition reaction to install a benzylic morpholine moiety.

No abstract