Speaker
Description
Accurately investigating the polymorphism of a given compound is an important part of material science in various areas. The number of polymorphs a compound can form and the relationship between them greatly influences the applicability of the compound for industrial usage for example. However, polymorphism has even greater implications for the field of pharmaceuticals and drug forms, both in terms of patient safety and finances or intellectual property, as patents usually extend only to specified polymorphs. The phase behaviour of an active pharmaceutical ingredient (API) by itself can be complex enough, but other solid forms are also permitted for use as the basis for drug forms, including salts and co-crystals.
In co-crystals, two or more different chemical species bind together to form a single crystal structure. In terms of pharmaceutical co-crystals, they consist of either multiple APIs or more commonly of an API and some other excipient molecule commonly referred to as a co-former. These co-formers, like the API, must be safe to consume by humans. The usage of co-crystals over pure solid API forms can be facilitated by several factors – namely that a co-crystal is regarded as a different solid form and therefore does not constitute patent litigation and more importantly, while the medicinal effect of the API stays the same, the co-crystals solid form properties are different than that of the original pure API, such as solubility and bioavailability. In other words, these co-crystal forms offer immense flexibility in how a solid drug form is constructed, provided a given co-former truly forms a co-crystal with the target API.
In this work, the polymorphism potential of two known carboxylic acid co-crystals of the API agomelatine is studied via the methods of crystal structure prediction (CSP) in combination with a computational thermodynamic stability assessment for selected generated solid phases. In the basic setup of CSP, the immense phase space available in crystalline (co-)crystals is searched through efficiently using a particular structure generation methodology, followed by a cheap calculation of their energy. The best structures found during this process are then subjected to more rigorous re-optimizations and calculations to determine their stability in relation to each other. In our case, we chose the mol-CSPy program for structure generation, which performs a quasi-random algorithm to pack molecules it treats as rigid bodies into periodic structures based on predetermined space group symmetry. Since the molecules within these solid forms are flexible, we introduced a pre-generation step, identifying combinations of different conformations of API and co-former molecules that have a higher likelihood of forming dense co-crystals before performing a full detailed rigid body search only within these best-performers. Afterwards, we re-optimized the best individual structures using periodic density functional theory (DFT-D) calculations. To compute the relative thermodynamic stability, we used an approach combining DFT-D and density functional tight binding calculations within the so-called quasi-harmonic approximation to calculate solid-state vibrational contributions to the total thermodynamic properties, allowing us to rank the structures by stability at temperatures higher than 0 K.