Purpose: Formation of diastereomeric salts is one of the most useful methods to separate enantiomers from racemic compounds. In this method, differential solubility between diastereomeric salts strongly affects optical resolution efficiency and yield. Therefore, optimization of solvents, their ratio and volume is effective for optical resolution, particularly, when some problems occur in the scale up process from milligram to kilogram.
JNJ-54861911, which Shionogi licensed out to Janssen Pharmaceuticals, Inc., was a drug candidate for Alzheimer’s disease. In the manufacturing process of JNJ-54861911, key chiral intermediate compound having an amidine group was obtained by optical resolution by using diastreomeric salt with L-tartric acid (LTA) having two carboxyl groups (Fig. 1). In the miligram scale trial of this process, diastreomeric crystal (R-LTA) consisted of R-isomer and LTA was selectively obtained against to S-isomer and LTA crystal (S-LTA) to achieve high diastreomer excess (de%). However, new crystal consisted of R-isomer, S-isomer, and LTA (R,S-LTA) appeared in the optimization of the process for kilogram scale manufacture. Once R,S-LTA appeared, R-LTA and S-LTA were not obtained from the solution to decrease the de% to 0%. We tried to optimize the process from the view point of differential solubility between these crystals but the de% was not improved. Therefore, we investigated the detail of R-LTA and R,S-LTA crystal structures in order to solve the problem and improve the process.
Methods: Molar ratio of R-isomer, S-isomer, and LTA in R-LTA, S-LTA, and R,S-LTA was determined by HPLC analysis. Crystal forms of these bulk crystals were characterized by X-ray powder diffraction. Crystal structures of R-LTA and R,S-LTA were determined by single crystal structure analysis. Single crystals of R-LTA and R,S-LTA for single crystal structure analysis were prepared by recrystallization from organic solvents.
Results: HPLC analysis revealed that the molar ratio of R-isomer, S-isomer, and LTA in R-LTA, S-LTA, and R,S-LTA, which were 2:0:1, 0:2:1, 1:1:1, respectively. The result implied that R,S-LTA has potential for three patterns as shown in Fig. 2. Pattern I is crystal mixture as shown in Fig. 1a and has possibility for optical resolution if these crystals have differential solubility to some solvents. On the other hand, pattern II (Fig. 2b) and pattern III (Fig. 2c) are in principle impossible to achieve optical resolution because R-isomer and S-isomer exist in racemic ratio in unit cell. To the best of our knowledge, pattern II and III crystals have not been reported. Single crystal structure analysis revealed that R,S-LTA was pattern II. Therefore, the process was optimized as described below to avoid the crystallization of R,S-LTA. In the previous manufacture process, LTA solution was added to racemic intermediate compound solution. On the other hand, racemic intermediate compound solution was added to LTA solution in optimized manufacture process. While the process, stoichiometry of LTA was excess in comparison with racemic intermediate compound to disturb crystallization of R,S-LTA. As a result, de% increased from 0% to 99%.
Conclusion: Salts of racemic intermediate compound and LTA provided different patterns of crystals in responsive to stoichiometry. R-isomer/S-isomer/LTA = 1:1:2 provided R-LTA and S-LTA crystals. In this stoichiometry, optical resolution was achieved effectively due to the differential solubility of these crystals. On the other hand, R-isomer/S-isomer/LTA = 1:1:1 provided R,S-LTA which disturbed optical resolution. Crystallization of R,S-LTA was avoided by optimization of the stoichiometry and method to form diastreomeric salts. As a result, optical resolution efficiency was dramatically improved. These results strongly suggest that optimization of stoichiometry between racemic compounds and counter molecules is as important as that of solvent ratio and volume in optical resolution process.
Shigeru Ando– Pharmaceutical Research Division, Shionogi & Co., Ltd
Kenichi Ishibashi– CMC R＆D Division, Shionogi & Co., Ltd.
Norihito Sato– Pharmaceutical Research Division, Shionogi & Co., Ltd.