This work describes the optimization and scale-up of a Buchwald−Hartwig amination reaction for the preparation of a pharmaceutical intermediate. This C− N bond formation is challenged by the use of a chiral primary amine, which both adds cost and favors formation of biaryl byproducts. In order to develop a scalable process, a number of factors had to be investigated including catalyst selection and stoichiometry of the chiral amine. These all needed to be optimized while maintaining low palladium levels in the isolated product. The reaction was found to be most effective using Pd(dba) 2 with BINAP and Cs 2 CO 3 in THF. When executed on 2.5 kg scale, these conditions provided 2.06 kg of the desired product in 80% yield with only 73 ppm residual palladium. To date, this process has been successfully executed to produce more than 12 kg of compound (S)-3.

This is the first in a series of three papers describing commercial manufacturing process development for palbociclib (1). This manuscript focuses on the SNAr coupling between aminopyridine 3 and chloropyrimidine 7. The regioselectivity of the SNAr coupling was studied from a synthetic and mechanistic perspective. Grignard bases were identified as the preferred class of bases for this reaction, allowing for a simplified process and reduced usage factor for aminopyridine 3. The development of this SNAr reaction into a scalable commercial manufacturing process is also described.

The original synthesis of the irreversible epidermal growth factor receptor (EGFR) T790 M inhibitor 1 was enabled by successful application of ammonium hydroxide to cleanly cleave the N-hydroxymethyl group and by development of high yielding conditions for the subsequent amidation reaction. Furthermore, a protection-free and regioselective new synthetic route was developed that shortened the synthesis from the original 8 steps to 6 steps and improved the overall yield from 5% to 34% on scale. Crystallizations of 1 and intermediates were correspondingly developed to control the quality en route.

Indole acids 1, 2, and 3 are potent 5′-adenosine monophosphate-activated protein kinase (AMPK) activators for the potential treatment of diabetic nephropathy. Compounds 1–3 were scaled to supply material for preclinical studies, and indole 3 was selected for advancement to first-in-human clinical trials and scaled to kilogram quantities. The progression of the synthesis strategy for these AMPK activators is described, as routes were selected for efficient structure–activity relationship generation and then improved for larger scales. The developed sequences employed practical isolations of intermediates and APIs, reproducible cross-coupling, hydrolysis, and other transformations, and enhanced safety and purity profiles and led to the production of 40–50 g of 1 and 2 and 2.4 kg of 3. Multiple polymorphs of 3 were observed, and conditions for the reproducible formation of crystalline material suitable for clinical development were identified.

A scalable process for PF-06651600 (1) has been developed through successful enabling of the first generation syntheis. The synthesis highlights include the following: (1) replacement of costly PtO 2 with a less expensive 5% Rh/C catalyst for a pyridine hydrogenation, (2) identification of a diasteroemeric salt crystallization to isolate the enantiomerically pure cisisomer directly from a racemic mixture of cis/trans isomers, (3) a high yielding amidation via Schotten−Baumann conditions, and (4) critical development of a reproducible crystallization procedure for a stable crystalline salt (1•TsOH), which is suitable for long-term storage and tablet formulation. All chromatographic purifications, including two chiral SFC chromatographic separations, were eliminated. Combined with other improvements in each step of the synthesis, the overall yield was increased from 5% to 14%. Several multikilogram batches of the API have been delivered to support clinical studies.